Digital Method Development Principles, Benefits, and Potential Pitfalls

March 2018 Volume 31 Number 3 www.chromatographyonline.com

Digital Method Development Principles, benefits, and potential pitfalls

LC TROUBLESHOOTING GC CONNECTIONS BIOPHARMACEUTICAL Eluent preparation for HILIC The past and present of GC PERSPECTIVES detection Peptide mapping using micro-pillar array columns All other trademarks are the property of their respective owners. WHOSE TRIPLE QUAD CAN GIVE YOU 15% MORE TIME? Copyright © 2016 PerkinElmer, Inc. 400358A_03. All rights reserved. PerkinElmer® is a registered trademark of SIMPLE: PERKINELMER.

Whether you’re performing environmental, food, or industrial testing, your lab needs to analyze more and increasingly complex samples under tighter regulatory standards. Because our new QSight™ triple-quad LC/MS/MS needs no routine cleaning and maintenance, you get up to 15% more uptime (that’s up to 35 more days per year) to take on your increased testing workload. Plus, our patented technology delivers exceptional sensitivity and reliability along with all that productivity. QSight: A triple quad that simply knows the value of time.

Learn more at www.perkinelmer.com/QSight Put Our Insight to Work For You

Imagine how productive you could be if you had access to a global team of Real Stories from the Lab experts who strive to deliver insight in every interaction with your lab. With True Story 44: Agilent CrossLab, that’s exactly what you get. Lab-Wide Data Analytics Consider us for compliance services, complete lab relocations, lab analytics System data insight leads to to optimize performance, on-site and on-demand training, as well as complete improved lab performance. service coverage for your Agilent and non-Agilent instruments. We will also See this story and more at consult with you to develop methods and recommend the best columns and agilent.com/chem/story44 supplies to optimize your results.

Find out more at www.agilent.com/crosslab

Volume 31 Number 3 COVER STORY 120 A Practical Approach to Modelling of Reversed-Phase Liquid Chromatographic Separations:arations: Advantages,Advantages, Principles, and Possible Pitfalls Patrik Petersson, Bernard O. Boateng, Jennifer K. Field, and Melvin R. Euerby Chromatographic principles and best practices for obtaining highly precise retention time, peak width, and resolution predictions for the optimization ofof reversed-phase LC separations using commercially available retention modelling software will be discussed.

Columns

144 LC TROUBLESHOOTING Eluent Preparation for Hydrophilic Interaction Liquid Chromatography, Part 2: pH, Buffers, and Gradient Elution Dwight R. Stoll and Claudia Seidl When preparing buffers for HILIC separations, is the buffer concentration considered relative to the aqueous portion of the eluent only, or the aqueous/organic mixture?

150 GC CONNECTIONS A Compendium of GC Detection, Past and Present John V. Hinshaw This instalment reviews GC detectors, both past and vanished, as well as current and relevant to today’s separation challenges.

155 BIOPHARMACEUTICAL PERSPECTIVES Peptide Mapping of Monoclonal Antibodies and Antibody–Drug Conjugates Using Micro-Pillar Array Columns Combined with Mass Spectrometry Koen Sandra, Jonathan Vandenbussche, Isabel Vandenheede, Bo Claerebout, Jeff Op de Beeck, Paul Jacobs, Wim De Malsche, Gert

VISIT US! • Hall B2 • Stand 128 Desmet, and Pat Sandra This article reports on the use of micro-pillar array columns for peptide mapping of monoclonal antibodies and antibody–drug conjugates.

Departments

167 Products Editorial Policy: 169 The Applications Book All articles submitted to LC•GC Europe 178 Events are subject to a peer-review process in association with the magazine’s Editorial Advisory Board.

Cover: Original materials courtesy: echo3005/ Shutterstock.com

116 LC•GC Europe March 2018 Delivering Smart Solutions The GCMS-QP2020 and the GCMS Insight software package dramatically improve the efficiency of daily analysis procedures

The Smart SIM creates optimal methods automatically and enables a batch analysis of more than 400 components, thereby reducing the number of measurements.

Simultaneous analysis of 434 pesticides using Smart SIM With LabSolutions Insight software, quantitative results for a complete series of data files can be displayed side-by-side, enabling easy identification of any outliers, which shortens analysis time.

www.shimadzu.eu/gcms-qp2020 Published by UBM Americas Editorial Advisory Board

Daniel W. Armstrong Kiyokatsu Jinno University of Texas, Arlington, Texas, USA School of Materials Sciences, Toyohasi Vice President/ Digital Production Manager Günther K. Bonn University of Technology, Japan Group Publisher Sabina Advani Institute of Analytical Chemistry and Huba Kalász Radiochemistry, University of Innsbruck, Semmelweis University of Medicine, Mike Tessalone [email protected] Austria Budapest, Hungary [email protected] Deirdre Cabooter Hian Kee Lee Managing Editor Special Department of Pharmaceutical and National University of Singapore, Editorial Director Projects Pharmacological Sciences, University of Singapore Leuven, Belgium Wolfgang Lindner Laura Bush Kaylynn Chiarello-Ebner Peter Carr Institute of Analytical Chemistry, [email protected] kaylynn.chiarello.ebner@ Department of Chemistry, University University of Vienna, Austria of Minnesota, Minneapolis, Minnesota, ubm.com Henk Lingeman Editor-in-Chief USA Faculteit der Scheikunde, Free University, Jean-Pierre Chervet Amsterdam, The Netherlands Alasdair Matheson Art Director Antec Scientific, Zoeterwoude, The Tom Lynch [email protected] Dan Ward Netherlands Analytical Consultant, Newbury, UK [email protected] Jan H. Christensen Ronald E. Majors Managing Editor Department of Plant and Environmental Analytical consultant, West Chester, Sciences, University of Copenhagen, Kate Mosford Subscriber Customer Service Pennsylvania, USA Copenhagen, Denmark Debby Mangelings [email protected] Visit (chromatographyonline.com) Danilo Corradini Department of Analytical Chemistry to request or change a Istituto di Cromatografia del CNR, Rome, and Pharmaceutical Technology, Vrije Associate Editor Italy Universiteit, Brussels, Belgium subscription or call our customer Lewis Botcherby Hernan J. Cortes Phillip Marriot service department on H.J. Cortes Consulting, Monash University, School of Chemistry, [email protected] Midland, Michigan, USA +001 218 740 6877 Victoria, Australia Gert Desmet David McCalley Sales Manager Transport Modelling and Analytical Department of Applied Sciences, Oliver Waters Hinderton Point, Separation Science, Vrije Universiteit, University of West of England, Bristol, UK Brussels, Belgium [email protected] Robert D. McDowall Lloyd Drive, John W. Dolan McDowall Consulting, Bromley, Kent, UK Cheshire Oaks, LC Resources, McMinnville, Oregon, Mary Ellen McNally Sales Executive USA Cheshire, DuPont Crop Protection,Newark, Liz Mclean Anthony F. Fell Delaware, USA [email protected] CH65 9HQ, UK Pharmaceutical Chemistry, Imre Molnár Tel. +44 (0)151 353 3500 University of Bradford, Bradford, UK Molnar Research Institute, Berlin, Germany Attila Felinger Senior Director, Digital Fax +44 (0)151 353 3601 Luigi Mondello Professor of Chemistry, Department of Dipartimento Farmaco-chimico, Facoltà Media Analytical and Environmental Chemistry, di Farmacia, Università di Messina, Michael Kushner University of Pécs, Pécs, Hungary Messina, Italy [email protected] Francesco Gasparrini Peter Myers Dipartimento di Studi di Chimica Department of Chemistry, e Tecnologia delle Sostanze Webcast Operations University of Liverpool, Liverpool, UK Biologicamente Attive, Università “La Janusz Pawliszyn Manager Sapienza”, Rome, Italy Department of Chemistry, University of Kristen Moore Joseph L. Glajch Waterloo, Ontario, Canada Momenta Pharmaceuticals, Cambridge, Colin Poole [email protected] Massachusetts, USA Wayne State University, Detroit, Davy Guillarme Michigan, USA Project Manager School of Pharmaceutical Sciences, Fred E. Regnier Vania Oliveira University of Geneva, University of Department of Biochemistry, Purdue Lausanne, Geneva, Switzerland [email protected] University, West Lafayette, Indiana, USA Jun Haginaka Harald Ritchie School of Pharmacy and Pharmaceutical Trajan Scientific and Medical, Milton Sciences, Mukogawa Women’s Subscribe on-line at Keynes, UK University, Nishinomiya, Japan www.chromatographyonline.com Koen Sandra Javier Hernández-Borges Research Institute for Chromatography, Department of Chemistry Kortrijk, Belgium (Analytical Chemistry Division), University of La Laguna Pat Sandra SUBSCRIPTIONS: LC•GC Europe is free to qualified readers in Europe. Canary Islands, Spain Research Institute for Chromatography, Kortrijk, Belgium To apply for a free subscription, or to change your name or address, go to John V. Hinshaw www.chromatographyonline.com, click on Subscribe, and follow the prompts. Serveron Corp., Hillsboro, Oregon, Peter Schoenmakers To cancel your subscription or to order back issues, please email your request to USA Department of Chemical Engineering, [email protected], putting LCE in the subject line. Universiteit van Amsterdam, Amsterdam, Tuulia Hyötyläinen The Netherlands Please quote your subscription number if you have it. VVT Technical Research of Finland, MANUSCRIPTS: For manuscript preparation guidelines, visit www.chromatographyonline.com or Finland Robert Shellie call the Editor, +44 (0)151 353 3500. All submissions will be handled with reasonable care, but Trajan Scientific and Medical, Hans-Gerd Janssen the publisher assumes no responsibility for safety of artwork, photographs or manuscripts. Every Ringwood, Victoria, Australia Van’t Hoff Institute for the Molecular precaution is taken to ensure accuracy, but the publisher cannot accept responsibility for the Sciences, Amsterdam, The Netherlands Yvan Vander Heyden accuracy of information supplied herein or for any opinion expressed. Vrije Universiteit Brussel, Brussels, DIRECT MAIL LIST: Telephone: +44 (0)151 353 3500. Belgium Reprints: Reprints of all articles in this issue and past issues of this publication are available (250 minimum). Contact Brian Kolb at Wright’s Media, 2407 Timberloch Place, The Woodlands, TX 77380. Telephone: 877-652-5295 ext. 121. Email: [email protected]. © 2018 UBM (UK) all rights reserved. No part of the publication may be reproduced in any material form (including photocopying or storing it in any medium by electronic means and whether or not transiently or incidentally to some other use of this publication) without the written Follow us @ LC_GC ‘Like’ our page LCGC Join the LCGC LinkedIn group permission of the copyright owner except in accordance with the provisions of the Copyright Designs & Patents Act (UK) 1988 or under the terms of the license The Publishers of LC•GC Europe would like to thank the members of the Editorial Advisory Board issued by the Copyright License Agency’s 90 Tottenham Court Road, for their continuing support and expert advice. The high standards and editorial quality associ- London W1P 0LP, UK. ated with LC•GC Europe are maintained largely through the tireless efforts of these individuals. Applications for the copyright owner’s permission to reproduce any part of this publication outside of the Copyright Designs & Patents Act (UK) 1988 provisions, LCGC Europe provides troubleshooting information and application solutions on all aspects of should be forwarded in writing to Permission Dept. fax +1 732-647-1104 or email: separation science so that laboratory-based analytical chemists can enhance their practical 10% Post [email protected]. Warning: the doing of an unauthorized act in relation to a knowledge to gain competitive advantage. Our scientific quality and commercial objectivity Consumer provide readers with the tools necessary to deal with real-world analysis issues, thereby copyright work may result in both a civil claim for damages and criminal prosecution. Waste increasing their efficiency, productivity and value to their employer.

118 LC•GC Europe March 2018 7 8 9 4 5 6 If your standards and your samples DAWN 1 23 HELEOS-II 0 .

TAB ESC DEL are the same size, they must be ENTER the same mass, right?

With DAWN you can determine absolute molecular weights and sizes without That’s the problem with calibration-based measuring techniques. You don’t using reference standards made in someone else’s lab. know—you can only assume. Which is why every major pharmaceutical and biotechnology company, as well as most federal regulatory agencies are switching from relative methods to Wyatt Technology’s absolute measurements. Our DAWN® multi-angle light scattering (MALS) instruments allow you to determine absolute molecular weights and sizes without relying on so-called standards, or measurements made in someone else’s lab. Wyatt instruments measure all of the quantities required for determining absolute molar masses directly.

Learn how to end you dependence on reference standards forever at wyatt.com/NoRef

wyatt.com

© 2017 Wyatt Technology. All rights reserved. All trademarks and registered trademarks are properties of their respective holders. A Practical Approach to Modelling of Reversed-Phase Liquid Chromatographic Separations: Advantages, Principles, and Possible Pitfalls

Patrik Petersson1, Bernard O. Boateng2, Jennifer K. Field2, and Melvin R. Euerby2,3, 1Novo Nordisk A/S, Måløv, Denmark, 2Strathclyde Institute of Pharmacy and Biomedical Sciences, University of Strathclyde, Glasgow, United Kingdom, 3Shimadzu UK Limited, Milton Keynes, United Kingdom

Chromatographic principles and best practices for obtaining highly precise retention time, peak width, and resolution predictions for the optimization of reversed-phase liquid chromatography (LC) separations using retention modelling software will be discussed. The importance of fully characterizing the LC instrumentation, how to generate accurate input data, the selection of appropriate models, and peak tracking will be addressed along with a suggested workflow. Adhesion to a few basic rules and simple precautions and the use of modern retention modelling software programmes can assist the rapid development of highly accurate retention models to enable the development of robust and optimized reversed-phase LC separations using either ultrahigh-pressure liquid chromatography (UHPLC) or high performance liquid chromatography (HPLC) conditions. Examples of retention modelling for small and large molecules will be highlighted.

The use of simulation software (1–4) based on chromatographic requires the definition of one or several response functions to theory, to predict retention behaviour and to optimize describe the quality of the separation with a single number, a chromatographic separations has now become a pivotal tool in far from trivial task. In our opinion, the most efficient approach method development strategies for traditional small molecule is to use retention modelling based on chromatographic theory and the ever-expanding biopharmaceutical drug market for optimization and subsequently apply statistical DoE models (5,6). The main driver for using retention modelling in method (that is, reduced factorial designs) for method validation and development strategies is that it only requires limited input data robustness testing. to rapidly obtain accurate, optimum, and robust separation Without doubt, the most ubiquitous use of retention modelling conditions for the chromatographer’s particular problem. is in the reversed-phase liquid chromatography (LC) arena to The prediction accuracy for analyte retention time and separate small molecules such as pharmaceutically-active resolution is good (6) and the software is flexible enough to allow the chromatographer to model isocratic or gradient separations as a function of variables such as percentage organic, gradient time, gradient shape, pH, temperature, ion-pairing reagent or salt concentration, flow, and column dimensions in a continuous way. In addition to the one-dimensional modelling described above, KEY POINTS two-dimensional modelling—a simultaneous variation of any • The benefits and limitations of modelling isocratic and two-separation variables for a chromatographic procedure— gradient reversed-phase LC separations of small and can be accurately modelled. Examples include gradient time large molecules is discussed. versus pH, percentage organic versus pH, gradient time versus • Practical advice on how to perform accurate temperature, and salt concentration versus temperature (6). modelling of reversed-phase LC separation is given. Software is also available that can perform three-dimensional • The importance of instrument characteristics, modelling (7). operating parameters, input conditions, The use of more general optimization software based on chromatographic performance parameters, and entirely empirical models and factorial designs, often referred peak tracking on the accuracy of retention modelling, to as design of experiments (DoE), requires significantly more separation optimization, and method robustness is input data for optimization (8). Another drawback is that it does discussed.

not simulate the predicted chromatography. This approach Photo Credit: echo3005/Shutterstock.com

120 LC•GC Europe March 2018 THe 6 sHarpest perspectives for focused science and technology solutions in life science

Proven Preparation, Tailored Pharma Expert Pharma & State-of-the-Art Lab & Pioneering Lab Trusted Analytical Separation, Filtration & Biopharma Raw Biopharma Manufacturing Production Materials Water Solutions Products & Testing Products Material Solutions & Testing Services

The life science business of Merck operates as MilliporeSigma in the U.S. and Canada.

Merck has brought together the world‘s leading Merck, the vibrant M, Milli-Q, Millipore, SAFC, BioReliance, Supelco and Sigma-Aldrich are trademarks of Merck KGaA, Darmstadt, Life Science brands, so whatever your life science *HUPDQ\RULWVDɡOLDWHV$OORWKHUWUDGHPDUNVDUHWKHSURSHUW\ SUREOHP\RXFDQEHQHɟWIURPRXUH[SHUWSURGXFWV RIWKHLUUHVSHFWLYHRZQHUV'HWDLOHGLQIRUPDWLRQRQWUDGHPDUNVLV DQGVHUYLFHV DYDLODEOHYLDSXEOLFO\DFFHVVLEOHUHVRXUFHV j0HUFN.*D$'DUPVWDGW*HUPDQ\DQGRULWVDɡOLDWHV $OO5LJKWV5HVHUYHG 7RɟQGRXWKRZWKHOLIHVFLHQFHEXVLQHVV RI0HUFNFDQKHOS\RXZRUNYLVLW SigmaAldrich.com/advancinglifescience

#howwesolve Petersson et al.

Figure 1: (a) An illustration of the determination of dwell method development tool in their own method development volume. Solid line represents the detector signal. Dashed strategies to efficiently generate high-quality LC methods that are line represents the programmed gradient. The dwell fit for purpose. = + 2 volume is defined as Vd (t50% - [t2 t1] / 2)F - πLr where t50% is the time when the signal reaches 50% of Experimental the maximal signal, t1 programmed start, and t2 finish of Experimental work was performed on a Nexera X2 UHPLC 2 the linear gradient. F is the flow rate and πLr defines the system (Shimadzu) equipped with LC-30AD pumps, volume of the restrictor capillary that replaces the column DGU-20A5R degassers, SIL-30AC autosampler, CTO-20AC during the dwell volume determination. (b) Examples of column oven, and SPD-M30A photodiode array detector traces for pumps displaying a nonlinear behaviour. equipped with a 10 μL/10 mm pathlength flow cell, 40 μL mixer (Shimadzu UK Ltd). The system was controlled and data (a) collected by means of LabSolutions software (Shimadzu UK × t Ltd, version 5.86). A 50 4.6 mm, 3-μm ACE SuperC18 column 1.0 2 (Advanced Chromatography Technologies Ltd) was used in the 0.8 study. pH measurements were recorded in the aqueous fraction of the mobile phase. For the dwell volume investigation, a range 0.6 of differing UHPLC configurations from Shimadzu, Agilent, 0.4 Waters, and Thermo were evaluated. Modelling was performed t Normalised UV signal (-) 0.2 50% using ACD Lab’s LC Simulator (version 2016.2.2). t 1 0.0 Results and Discussion 0 5 10 15 20 25 This article will focus on the critically important parameters Time [min] (b) common to all commercial retention modelling software programmes, characterization of the LC system, and the number and type of input experiments that are required (dependant on the retention model used). Peak tracking and the selection of the most appropriate retention models will also be investigated, all of which are required to generate accurate retention predictions. The generation of suitable samples (for example, forced degradants, mother liquours) will not be covered in this article. Characterization of the LC System: Most method development strategies, and hence retention modelling, is performed using gradient chromatography, therefore it is vital to compounds, their synthesis impurities and degradation establish that the LC system being used is capable of generating products, peptide and tryptic digests and protein mixtures, drug a reproducible linear gradient. This can be rapidly established by metabolites, complex mixtures of active compounds from plant dwell volume determinations (Figure 1[a] and Figure 2). As can origin, food safety, environmental pollutants, polymer analysis, be seen from Figure 1(b), the resultant gradient profile from these drugs of abuse, and to estimate the robustness of LC methods three LC systems is unacceptable. If temperature modelling is to (6,9–12). Retention modelling has also been used for translations be used, the authors recommend that the column compartment between ultrahigh-pressure liquid chromatography (UHPLC) to should be checked using a calibrated thermocouple and that high performance liquid chromatography (HPLC) and vice versa there is sufficient preheating of the mobile phase before it enters (13), and in Quality by Design (QbD) approaches (14). the column, which can be achieved by the use of a preheater or Retention modelling is now being successfully applied to a sufficiently long piece of tubing. The flow rate accuracy must characterize proteins, monoclonal antibodies, their charge be established at the flow rate of the input experiments (using a variants, and antibody–drug conjugates using chromatographic flow rate meter or simply from the measurement of the weight of modes such as hydrophobic interaction chromatography (HIC), water delivered during a certain time at a certain temperature). reversed phase, hydrophilic interaction chromatography (HILIC), The dwell volume should be determined as shown in Figure 2 and ion-exchange chromatography (IEC) (15–19). for the LC configuration that is to be used for the modelling Retention models have also been applied in numerous input experiments. The detector sampling rate is not critical for separation techniques, including gas chromatography (GC) (20), modelling, but it is recommended to record no less than 25 ion-pair chromatography (IPC) (21), HILIC (22), micellar liquid points for each peak so that the resolution is not compromised. chromatography (MLC) (23), chiral chromatography (24), ion Determination of Dwell Volume: chromatography (IC) (25), and supercritical fluid chromatography The dwell volume of an instrument is defined as the volume from (SFC) (9). the point at which the mobile phases first mix in the pump to This article will describe, in a stepwise manner, how to the head of the column. The dwell volume can be determined perform successful and accurate retention modelling using in different ways. In many earlier text books on HPLC and reversed-phase LC examples and the pitfalls to be avoided pharmacopoeia publications it was suggested that a wide to generate accurate predictions. The advice given is equally linear gradient range should be employed, accompanied applicable to all types of retention models and applications using with a high flow rate (26). However, several instrument any of the commercial software programmes. manufacturers use a step gradient with a low flow rate and It is hoped that this discussion will encourage and promote a narrow organic range. We have found that the gradient chromatographers to adopt this highly accurate and rapid type (step or linear), flow rate, and gradient range are all

122 LC•GC Europe March 2018 6 powerful propellers to accelerate expert science and technology solutions in life science

The life science business of Merck operates as MilliporeSigma in the U.S. and Canada.

#howwesolve Figure 2: Determination of dwell volume.

• Replace the column with a 100 cm × 0.1 mm capillary to generate enough pressure to ensure proper function of check valves. • Use water as solvent A and 10 mg/L of uracil in water as solvent B. • Set the flow rate to the flow that should be modelled. SPEX CertiPrep Pesticide Mixes and Kit • Programme a 1 μL injection of water to ensure that the autosampler is in the flow path and European Pesticide Mix contributes to the dwell volume. • Run a gradient with a 5-min isocratic hold We are introducing a new pesticide mix to address followed by a 10-min linear gradient with the the European Commission’s Regulation 2017/170. same slope as the average gradient used for collection of the retention modelling data, The Commission is amending Annexes II, III and V for example 10%B at 0 min, from 10 to 20%B to Regulation (EC) No 396/2005 of the European between 5 and 15 min, from 20 to 10%B Parliament and of the Council as regards to maximum between 25 and 25.1 min. residue levels for bifenthrin, carbetamide, cinidon- • UV detection at 259 nm at 20 Hz to ensure a ethyl, fenpropimorph, and triflusulfuron in or on well-defined curve. • The dwell volume is defined using the gradient certain products. start and stop time and time point for 50% of maximal signal as described in Figure 1(a). Premixed Pesticide Multi-Compound • In order to evaluate the linearity (Figure 1) an extra run is made where the range is extended Certified Reference Materials to 0–100%B. An extra gradient is necessary since flow rate and gradient range influence We have designed a pesticide residue testing kit the determination of dwell volume. which includes 144 of the most commonly analyzed pesticides per EPA, AOAC, FDA and other international testing methods. The kit is structured to maximize critical for the determination of dwell volumes (differences of stability and solubility while minimizing unwanted up to 80% have been observed [27]). We have also found that micro-fabricated “maze”-type mixers display a larger analyte interaction and interference; enjoy shorter difference between high and low flow than a traditional type of calibration times, fewer injections and money savings, mixer. Thus, conditions that previously have been suggested as compared to purchasing individual pesticide for the determination of dwell volumes appear not to be standards. suited for certain UHPLC systems. It is the authors’ opinion that the dwell volume should be determined using gradient conditions that are appropriate for the type of analyte and LC For details and purchasing information visit instrumentation that will be used. This will be discussed in spexcertiprep.com/lcgc or contact us for additional more detail in the section called “Gradient Separations”. information. Figure 2 describes a procedure that provides an estimate of dwell volumes suitable for modelling purposes. This procedure is based on a linear gradient and it can also be used to ensure SPEX CertiPrep is the industry leader for over 60 years in the CRM that linear gradients can be generated by the system. The marketplace, meeting the needs of laboratories worldwide with procedure displays a good agreement with step gradients at innovation and research. Accredited by A2LA to ISO/IEC 17025:2005 & Δ ISO 17034:2016. Certified by DQS to ISO 9001:2008. the same flow rate (| Vd | <4% 90th percentile for nine UHPLC configurations with dwell volumes ranging from 202 to 718 μL [27]). It also displays a reasonable agreement between measured mixer volumes and nominal volumes specified by the instrument producers ( | Δ mixer volume | <16% for three types of UHPLC systems and nine mixers ranging from 35 μL Inorganic and Organic Certified Reference Materials to 380 μL). US ADDRESS An alternative approach to determine dwell volumes, as well /PSDSPTT"WFOVFt.FUVDIFO /+t5FM as to compensate for errors in other parameters, is to iteratively 'BYt$3.4BMFT!TQFYDPNtXXXTQFYDFSUJQSFQDPN try a few different dwell volumes while fitting the model, compare the residuals obtained, and, based on this, select the UK ADDRESS dwell volume that gives the lowest residual (28). %BMTUPO(BSEFOTt4UBONPSF .JEEMFTFYt)"#2t6, An error in dwell volume impacts on absolute retention 5FM t'BY 41&9&VSPQF!TQFYDPNtXXX41&9&VSPQFDPN predictions more than relative retention and resolution (29). Fortunately, even a relatively large error in dwell volume of

LC•GC Europe March 2018 DISCOVER NEW SOLUTIONS FOR LIQUID CHROMATOGRAPHY NEW UHPLC COLUMNS FOR ANTIBODY SEPARATION ADVANCED GPC/SEC SOLUTIONS FOR POLYMER ANALYSIS HILIC, RPC, IEC, HIC & SEC BIOPOLYMER ANALYSIS HIGH CAPACITY PROTEIN A, IEC, HIC & MIXED-MODE MEDIA TO FIND OUT MORE ABOUT OUR SOLUTIONS FOR SEC/GPC, HPLC, UHPLC AND DOWNSTREAM PROCESSING, TALK TO ONE OF OUR TECHNICAL SPECIALISTS AT ARABLAB, BOOTH 120, OR ANALYTICA, HALL 2, BOOTH 412. Petersson et al.

Figure 3: Experimentally determined dead volumes been changed by an introduction of a larger mixer for example). and chromatographic profiles obtained at 214 nm using Column Dead Volume: water and uracil as the dead time markers at different pH Dead volume is defined as the retention volume for a nonretained values and with different proportions of acetonitrile. The analyte. The determination may seem trivial but is actually mobile phase contained 10 mM of ammonium formate quite complicated (28). As can be seen in Figure 3, for two pH 3.0, 10 mM ammonium acetate pH 6.8, or 18.6 mM of frequently used dead volume markers, uracil and water, the value ammonium hydroxide pH 10.7. determined depends both on the mobile phase composition and the marker used. At pH 10.7 both water and, even more 0.6 pronounced, uracil display markedly different dead volumes

0.5 compared to pH 3 and 6.8. For uracil this is probably a result of repulsion between deprotonated and, therefore, negatively 0.4 pH 3.0 water charged uracil (pKa 8.8) and silanol groups (pKa range (mL) pH 6.8 water pH 10.7 water approximately 3.5–6.8 [31]) resulting in a reduced retention. In 0.3 pH 3.0 uracil the case of water, the opposite is observed. This may be a result volume U U U pH 6.8 uracil pH 10.7 uracil ead 0.2 of water penetrating deeper into the silica surface or pores, D 10% 35% 70% which, at pH 10.7, are negatively charged. W 0.1 W Fortunately, it has been found that relatively large errors W in dead volume only have a small impact on the quality of 0.0 10 20 3040506070 predictions. According to previous studies (28–30,32) and our

[MeCN] (% v/v) own experience, even an error of ±20% in the dead volume will only result in <1% error in the modelling of isocratic as well as gradient retention. Uracil is more affected by pH changes than water, and we ±20% will only have a small impact on absolute retention; propose that water should be used as the dead volume marker. <<1% according to literature (30) as well as our observations At a wavelength of 214 nm water usually produces a well-defined (27). negative peak (Figure 3). For isocratic modelling, we recommend It should be stressed that once a dwell volume has been that the dead volume is determined for the average amount determined it is not necessary to determine it again unless a of organic modifier used for the generation of the models. For significantly different flow rate or gradient slope is used for the gradient modelling however, we have found that the dead volume generation of models (or the configuration of the system has for initial gradient conditions gives a better fit than the average

Illuminating Chemistry Trusted Innovation

Visit booth 312 at Analytica to see the latest Swedish Innovations. dead volume. The determination should be done at the flow rate to be modelled and at the average temperature used during calibration experiments. Determination of Extra Column Band Broadening (ECBB): This parameter is a measure of the contribution to peak dispersion that takes place outside the column. Its effect on peak width can become very significant in UHPLC separations. It can be rapidly determined as described in reference 33. In the opinion of the authors, the retention modelling programme should be able to predict the effect of changing this value on the resultant peak width of the analyte. The chromatographer may wish to investigate this if it is necessary to assess the chromatographic performance of a separation when converting from a standard HPLC to a UHPLC configuration (or vice versa). Selection of Retention Models: The retention models used in today’s optimization software were to a large extent developed during the 1980s by Snyder et al. (34) and Jandera et al. (35). For isocratic reversed-phase LC separations of small molecules, it is often sufficient to use a first order polynomial retention model (equations 1 and 3). For very high levels of organic modifier where other retention mechanisms come into play, it may be necessary to add a second order term to account for curvature (equation 2). The use of a second order term is also necessary for peptides and proteins where polar and electrostatic interactions are more important and also because their secondary and higher structures may be affected by the organic modifier content and temperature (equations 2 and 4) (36). After optimizing the organic modifier content and temperature for protein separations, it is often advantageous to optimize the trifluoroacetic acid (TFA) concentration. This can be conveniently modelled by a second Analytica 2018 order model such as equation 2. The effect of buffer concentration in reversed-phase LC can Visit us at be modelled using the log–log relationship as described by hall A1, booth #117 equation 5. It is possible—but less common—to optimize pH by retention modelling. The reason for this is that the range covered by pH models is often quite narrow compared to the large pH ranges that the chromatographer may wish to exploit. In addition, the prediction errors can be relatively large. pH models typically Kromasil SFC cover pH ranges close to the pKa of the analytes of interest and, therefore, big selectivity differences may be observed Designed for as the species change their ionization state. The downside to green technology this approach is that peak shape is often poor at a pH close to their pKa. The robustness and reproducibility are usually poorer Introducing SFC-XT because of the high sensitivity of the analyte’s retention to small changes in pH. Kromasil SFC-XT, a fused organo-silane Commercial software for retention modelling is based on numerical calculations, and, as mentioned earlier, this allows phase for alternative selectivity and the combination of models for two variables such as amount of retention, is the latest addition to the organic modifier and temperature, which has been found to be a Kromasil SFC family of columns. very efficient method development strategy (6, 34 p. 92). Kromasil SFC, your ﬁrst choice for Φ Φ reliable, consistent and reproducible log(k) = a – m or log(k) = q1 + q2 [1] SFC analysis and puriﬁcation.

Φ Φ2 log(k) = q1 + q2 + q3 [2] www.kromasil.com q log(k) = q + 2 1 T [3]

www.chromatographyonline.com Petersson et al.

q2 q3 also k ≈ 6). An empirical trial-and-error approach can be used to log(k) = q + + 1 T T 2 [4] find conditions that provide such k-values. As a rule of thumb, k increases by a factor of three for a 10% reduction in amount of organic modifier (34 p. 237). By having k-values for two or more log(k) = q + q log(C) 1 2 [5] conditions it is possible to use equation 1 to calculate suitable conditions. In our experience it is more efficient to predict the Where a, m, and q1 – q3 are analyte and system-specific suitability of isocratic conditions for isocratic modelling based on constants, T is the temperature of the column, Φ the fraction of a simple gradient model (see “Gradient Separations” section). organic modifier, and C the salt concentration. Peak width models are based on chromatographic theory (28, R = 0.25 [k / (k + 1)][(α-1)/α] √ N 35, 34 p. 378) or, alternatively, on entirely empirical models (22). s [6] The latter approach can also be used to model peak asymmetry. How the peak width and peak asymmetry models are defined Where α is the selectivity factor and N the number of is usually not visible in the software to the user. The software theoretical plates for an isocratic separation. employed in the current study uses empirical peak width or Gradient Separations: asymmetry models. The resolution equation (equation 7) is also valid for gradient Although it is possible to model the peak asymmetry of the elution chromatography provided that k, α, and N are defined as main component, it should be noted that there are no suitable instantaneous values as the analyte passes the mid-point of the models to adequately describe the shape of the main peak at column, k*, α*, and N* (34 p. 39 and p 90). low level. For this reason, the resolution between a main peak and adjacent impurities are typically poorer than predicted. The R = 0.25 [k* / (k* + 1)][(α*-1)/α*] √ N* workaround is to keep this in mind while locating alternative s [7] optimal conditions that give the best possible resolution within an acceptable time and subsequently evaluate these experimentally. Assuming that isocratic retention can be described by equation Models are fitted to a calibration data set and the residuals 1 it is possible to define an expression for which gradient time, tG, give an assessment of the quality of the model. However, it is gives a certain retention k*. good practice to subsequently confirm the model against a validation data set as illustrated in the following examples. As a 1.15V ΔΦ mk* t = m general rule of thumb, the simplest model should be selected G F [8] that still gives an acceptable error for the validation data set. A ΔΦ first order model (equation 1) is more robust and allows for more Where Vm is the column dead time, the gradient range, m extensive extrapolation than a second order model (equation 2). a parameter in the retention equation (equation 1), and F the flow In order to save time, the collection of a validation data set can rate. be skipped and instead a direct prediction of optimal conditions An approximation for the m term can be calculated using the made based on a first and second order model. Subsequently, molecular weight (M) of the analyte m ≈ 0.25M0.5 (34 p. 18). both optima are experimentally evaluated and the appropriate Using equation 8 and the m-value approximation it is possible model as well as optimal conditions are thereby confirmed. to calculate gradient times that give k*-values which cover Input Runs Required for Isocratic and Gradient Modelling: the range of interest 1

128 LC•GC Europe March 2018 Petersson et al.

Table 1: Calculation of gradient times corresponding to k*= 3, 6, and 9 for a small molecule and a 50 kDa protein Molecular weight, M (Da) <1000 50,000 Retention parameter, m (-) (equation 1) 4 56 Column length, L (mm) 150 150

Column i.d., dc (mm) 2.1 2.1 Flow, F (mL/min) 0.3 0.3

Dead volume, Vm (mL) 0.3 0.3 \ Initial fraction of organic modifier, 0 (-) 0.03 0.4 \ Final fraction of organic modifier, max (-) 1 0.6 6\ = \ \ Gradient range, max - min (-) 0.97 0.2 k = t Gradient time for * 3 (equation 8), G 14 40 Gradient time for k* = 6 28 80 Gradient time for k* = 9 42 120 somewhat reduced resolution because of the steep van Deemter gradient is required to maximize resolution around the curve displayed by large molecules. main peak. This is bracketed with a steep initial and final In order to fit models, it is important to ensure that the gradient in order to capture very hydrophilic and hydrophobic gradient range is defined so that the analytes elute on the degradation products or process impurities. When generating linear part of the gradient that is not in the isocratic dwell models, it is usually only this shallow part that is changed to volume or at the hold at the end of the gradient. It is also an achieve different k*. advantage to keep the buffer concentration the same in mobile ΔΦ phases A and B to reduce superimposed salt and organic Φ = Φ + ( t – t – t ) e 0 t g m d modifier gradients, which can be difficult to model (this also G [9] minimizes drifting baselines). When working with peptides and proteins it is common Isocratic or Gradient Separations at Different Temperature: practice to use segmented gradients where a shallow As previously mentioned, large differences in selectivity can be

the micro-Chip Chromatography Company

Changing the ART of analytical chromatography with μPAC™ Pillar Array Columns. These micro-Chip columns feature a perfectly-ordered separation bed of free-standing pillars ensuring: • excellent separation power • unprecedented reproducibility • unrivalled robustness Enhance the data productivity of your nano-LC/MS system for tiny, complex biological samples.

Discover our products on www.pharmafluidics.com

www.chromatographyonline.com 129 Petersson et al.

Figure 4: Experimentally-derived retention behaviour for obtained by simultaneously optimizing the amount of organic carboxylic acids chromatographed in their (a) unionized modifier and temperature (34 p. 92). In order to fit such models, and (b) ionized state at pH 3 and 6.8, respectively. it is necessary to collect data at different temperatures for each mobile phase composition (or gradient slope). For small (a) molecules it is usually sufficient with two temperatures, for 1.6 4-Chlorocinnamic acid example, 30 °C and 60 °C. However, for peptides and proteins 1.4 Ketoprofen Flurbiprofen whose secondary structure changes with temperature it is Indomethacin 1.2 typically necessary to use three temperatures, such as 30 °C, Meclofenamic acid 45 °C, and 60 °C. Thus, these types of modelling require 2 × 2 1.0 = 4 or 3 × 3 = 9 experiments. (-) k 0.8 Collection of Input Data for Fitting of Models: Before retention log 0.6 times are collected for generation of retention models, it is

0.4 important to ensure that the instrument delivers linear gradients as well as stable retention times. Also, it is often forgotten that 0.2 a sufficiently long equilibration time should be used so that 0.0 45 50 55 60 65 70 the column is at a steady state prior to the next gradient. For (b) [MeCN] at pH 3.0 (%v/v) reversed-phase LC it is typically recommended that at least 10 column volumes should be used (34 p. 170). Once a method has 1.5 been developed, it is usual to investigate whether the number

1.0 of column volumes can be reduced (that is, before the retention of early peaks is affected) hence increasing productivity. Other 0.5 types of chromatography such as HILIC and IEC typically (-) k often require > 20 column volumes to establish a steady log 0.0 state. Changing between different %B isocratic conditions in reversed-phase LC also typically requires equilibrating the -0.5 column with 10 column volumes. -1.0 It should be noted that some proteins may require priming or conditioning of the column with the protein prior to establishing 20 3040506070 [MeCN] at pH 6.8 (%v/v) stable retention times; the heat of friction in the column may also require that a couple of “dummy runs” should be performed to

130 LC•GC Europe March 2018 DON’T PAY TO WASTE, INVEST IN THE BEST

Unmatched Stability for long-term enzyme usage

Abalonase™ UltraSXULILHGǂJOXFXURQLGDVH when kept independent of the SURYLGHGRapid Hydrolysis Buffer, possesses a shelf life of RQH\HDU

Contact us at [email protected] or 800-385-3153 to try a sample today! Figure 5: Prediction error ([predicted – actual] 100/ actual) for isocratic retention times calculated based on gradient retention models for acidic and basic (methoxydiphenidine “MXP” isomers) analytes chromatographed at low and intermediate pH.

20 Meclofenamic acid pH 6.8 11 – 15 June 2018 Indomethacin pH 6.8 Flurbiprofen pH 6.8 10 Ketoprofen pH 6.8 Frankfurt am Main 4-Chlorocinnamic acid pH 6.8 Meclofenamic acid pH 3 0 Indomethacin pH 3 Flurbiprofen pH 3 Ketoprofen pH 3 4-Chlorocinnamic acid pH 3

Prediction error (%) -10 2-MXP pH 6.8 4-MXP pH 6.8 3-MXP pH 6.8 -20

0 5 10 15 20 25 30 Retention factor (-)

Figure 6: (a) Critical resolution as a function of temperature and gradient slope for a mixture consisting of a crude peptide and its degradation products. Optimal conditions (circle) at 43 °C and a slope of 0.23%B/min (that is, 23.5%B at point Z in the gradient as indicated in [b]). (b) Predicted (red) and experimental (black) chromatograms corresponding to optimal conditions in (a).

(a) Rsmin (-) 50 1.1

0.84 45 0.58 ) C ° 40 0.31

0.050

35 -0.21 Temperature (

-0.47 30 -0.74

-1.0 BE INFORMED. BE INSPIRED. BE THERE. 25 18 20 22 24 26 28 30 (b) Amount of B solvent at point Z (% v/v) › World Forum and Leading Show 11 100 for the Process Industries 10 9 80 › 3,800 Exhibitors from 50 Countries 8 60 7 40 › 6 170,000 Attendees from 100 Countries X 20 5 Z

4 0

3 -20 Amount of B solvent (% v/v) Absorbance at 215 nm (mAU) LaboratoryTechniques @ ACHEMA 2 Exp. -40 Calc. 1 THE LAB IS WHERE IT %B -60 0 ALL STARTS. 0 2 4 6 8 10121416182022 Time (min) #labtechniques obtain a stable column temperature. The same principle must be used when changing other variables such as temperature. The chromatographer must be assured that the system is at a steady state—it is a paradox that rapid and accurate retention modelling cannot be rushed! It is recommended that the peak width is determined at half height and subsequently the peak width calculated at base line (using equation 10) if that is what is required by the software used for modelling. The reason is that the latter cannot be directly determined for partially coeluting www.achema.de peaks. LC•GC Europe March 2018 The FFF - MALS Platform Next Level Nano, Bio and Polymer Analysis

NEW With SEC Option!

NovaFFF Software

MALS DLS UV RI ICP-MS

Table 2: Gradient to gradient prediction for acids at pH 3.0 (equation 1) Calibration Data Set Gradient Time Peak Name 5 (min) 10 (min) 15 (min) t t t g (min) w (min) g (min) w (min) g (min) w (min) 4-Chlorocinnamic acid 3.52 0.069 5.43 0.119 7.04 0.164 Ketoprofen 3.73 0.065 5.88 0.110 7.74 0.152 Flurbiprofen 4.28 0.066 6.94 0.115 9.31 0.160 Indomethacin 4.42 0.064 7.25 0.110 9.78 0.152 Meclofenamic acid 4.87 0.067 8.05 0.117 10.92 0.161 Validation Data Set Gradient Time 7.5 min t Δt w Δw R ΔR Peak Name g (min) g* (%) (min) * (%) s (-) s* (%) 4-Chlorocinnamic acid 4.53 0.11 0.095 2 -- Ketoprofen 4.85 0.10 0.088 2 3.6 -2 Flurbiprofen 5.67 0.11 0.091 3 9.1 -3 Indomethacin 5.89 0.08 0.087 3 2.5 -4 Meclofenamic acid 6.52 0.09 0.093 1 7.0 -2 Gradient Time 12.5 min t Δt w Δw R ΔR Peak Name g (min) g* (%) (min) * (%) s (-) s* (%) 4-Chlorocinnamic acid 6.25 0.22 0.142 -1 -- Ketoprofen 6.83 0.21 0.132 -2 4.2 1 Flurbiprofen 8.15 0.18 0.137 -1 9.8 1 Indomethacin 8.53 0.18 0.132 -2 2.9 1 Meclofenamic acid 9.51 0.15 0.14 -1 7.2 1 *Prediction error = (predicted - actual) 100/actual Column: 50 × 4.6 mm, 3-μm Vm: 0.457 mL Vd: 0.227 mL F: 1 mL/min Φ 0: 0.1 - ΔΦ: 0.8 -

4w w = 50% 13.4% 2.35 [10] It is also recommended to determine and model the asymmetry of these peaks. It is essential to ensure that the σ Where w13.4% is the peak width at baseline (4 which chromatographic data system’s peak asymmetry calculation actually is defined at 13.4%) and w50% is the peak width is the same as the required retention modelling software at 50% (2.35σ). If a linear gradient is used and peaks are definition of peak asymmetry. symmetrical it is usually not necessary to determine the It is not necessary to determine peak areas to generate width of each peak in the chromatogram. It is then often models. Peak areas do not affect the resolution or the quality acceptable to assume that all peaks have a similar peak width of the separation. Peak areas are only used to facilitate visual and therefore use an average peak width determined for well comparisons of calculated and experimental chromatograms. separated early and late peaks. Isocratic separations require Therefore, it is sufficient to determine peak areas for one of that the width of all peaks be determined because peak width the experiments where most of the peaks are well separated increases with increasing retention. If the chromatograms and then use these areas as inputs for the other conditions. contain overloaded or asymmetric peaks it is necessary to If peaks are not well separated, areas from different peaks in determine the width of each of these peaks. The width of different experiments can be combined. such peaks is determined at 4.4% (5σ) and then recalculated To obtain a good agreement between calculated and (equation 11) to 4σ to better reflect the relatively broader peak experimental chromatograms, it is important to use the same in the modelling. equipment, column, and batch of solvents that was used to obtain data for the generation of models. It is also an advantage

4w 4.4% if the time between building and verification of the models is w13.4% = 5 [11] minimized.

134 LC•GC Europe March 2018 MAXIMISE YOUR CANNABIS PURIFICATION

Advantages of CPC Equipment: Ɣ Isolate any type of cannabinoids or terpenes Ɣ No product loss Ɣ High purity achievable

www.gilson.com/cpc Petersson et al.

Table 3: Gradient to isocratic prediction for alkyl phenones (equation 2) Calibration Data Set Gradient Time 5 (min) 10 (min) 15 (min) Peak Name t w t w t w g (min) (min) g (min) (min) g (min) (min) 2-Acetylfuran 1.96 0.066 2.33 0.092 2.53 0.109 Acetanilide 2.17 0.060 2.77 0.090 3.16 0.116 Acetophenone 3.15 0.078 4.44 0.131 5.39 0.177 Propiophenone 3.86 0.080 5.81 0.140 7.38 0.194 Butylparaben 4.06 0.063 6.42 0.107 8.47 0.151 Benzophenone 4.57 0.071 7.28 0.128 9.62 0.182 Valerophenone 4.86 0.075 7.82 0.133 10.40 0.191 Validation Data Set Gradient Time 7.5 min t Δt w Δw R ΔR Peak Name g (min) g* (%) (min) * (%) s (-) s* (%) 2-Acetylfuran 2.18 -0.28 0.078 5 - - Acetanilide 2.52 -0.28 0.075 6 4.4 -5 Acetophenone 3.86 -0.10 0.105 5 14.9 -5 Propiophenone 4.91 -0.16 0.109 8 9.8 -6 Butylparaben 5.30 -0.17 0.085 8 4.1 -8 Benzophenone 6.00 -0.13 0.100 8 7.5 -7 Valerophenone 6.41 -0.12 0.104 9 4.1 -8 Gradient Time 12.5 min t Δt w Δw R ΔR Peak Name g (min) g* (%) (min) * (%) s (-) s* (%) 2-Acetylfuran 2.44 0.04 0.100 -1 - - Acetanilide 2.98 0.00 0.104 -2 5.3 1 Acetophenone 4.94 0.12 0.155 -2 15.2 2 Propiophenone 6.63 0.02 0.166 -1 10.5 1 Butylparaben 7.47 0.00 0.129 -2 5.7 1 Benzophenone 8.48 0.02 0.156 -3 7.1 2 Valerophenone 9.14 0.02 0.161 -1 4.2 2 Isocratic at 45%B t k Δt w Δw R ΔR Peak Name R (min) R* (%) (min) * (%) s (-) s* (%) 2-Acetylfuran0.911.06.1---- Acetanilide 0.92 1.0 -5.2 - - - - Acetophenone 1.76 2.8 -6.3 0.100 5 - - Propiophenone 3.05 5.6 -2.9 0.195 0 8.7 0 Butylparaben 3.96 7.6 -1.2 0.248 -6 4.1 9 Benzophenone 6.45 13.0 1.3 0.459 -11 7.0 16 Valerophenone 8.84 18.3 1.6 0.674 -15 4.2 18 *Prediction error = (predicted - actual) 100/actual Column: 50 × 4.6 mm, 3-μm Vm: gradient and isocratic: 0.459 mL Vd: 0.227 mL F: 1 mL/min \ 0: 0.1 - Δ\: 0.8 -

Peak Tracking and Peak Identification: Effective and remains the “Achilles heel” of retention modelling rapid assignment of peaks in the input chromatograms and optimization. If not done properly, incorrect peak

136 LC•GC Europe March 2018 Perfect separation solutions for polymers, proteins and particles. We go that extra yard!

PSS is your partner for GPC/SEC, Interaction Polymer Chromatography IPC or 2-Dimensional Chromatography 2D – whether with our SECcurity² turnkey solutions, or with our expert contract analysis services. For the ultimate in macromolecular characterization, contact PSS and stay one step ahead of the game.

P ERFECT S EPARATION S OLUTIONS www.pss-polymer.com Petersson et al.

Table 4: Isocratic to isocratic prediction for MXPs (equation 2) Calibration Data Set %B 40% 55% 70% Peak Name t w t w t w R (min) (min) R(min) (min) R (min) (min) 2-MXP 4.21 0.250 1.85 0.094 1.18 0.061 4-MXP 6.16 0.387 2.65 0.145 1.62 0.083 3-MXP 10.86 0.673 4.29 0.235 2.37 0.117 Validation Data Set 35%B (extrapolation!) k t Δt w Δw R ΔR Peak Name R (min) R* (%) (min) * (%) s (-) s* (%) 2-MXP 14.48 6.33 -4.0 0.38 -8 -- 4-MXP 21.65 9.20 -3.3 0.566 -4 6.1 4 3-MXP 39.65 16.47 -3.0 1.099 -12 8.7 7 45%B k t Δt w Δw R ΔR Peak Name R (min) R* (%) (min) * (%) s (-) s* (%) 2-MXP 6.77 3.01 1.6 0.178 -1 -- 4 - MXP 10.34 4.41 1.1 0.282 - 4 6.1 3 3-MXP 18.51 7.60 0.9 0.481 -4 8.4 4 50%B k t Δt w Δw R ΔR Peak Name R (min) R* (%) (min) * (%) s (-) s* (%) 2-MXP 4.91 2.32 0.2 0.117 15 -- 4-MXP 7.56 3.35 0.5 0.187 10 6.8 -9 3-MXP 13.32 5.59 0.7 0.313 9 9.0 -7 60%B k t Δt w Δw R ΔR Peak Name R (min) R* (%) (min) * (%) s (-) s* (%) 2-MXP 2.89 1.53 -0.2 0.082 7 -- 4 - MXP 4.51 2.18 - 0.4 0.121 9 6.3 -9 3-MXP 7.64 3.41 -0.6 0.183 13 8.2 -11 65%B k t Δt w Δw R ΔR Peak Name R (min) R* (%) (min) * (%) s (-) s* (%) 2-MXP 2.35 1.32 -0.5 0.068 12 -- 4-MXP 3.69 1.85 -0.7 0.100 12 6.3 -12 3-MXP 6.09 2.81 -0.9 0.141 20 8.0 -15 *Prediction error = (predicted - actual) 100/actual Column: 50 × 4.6 mm, 3-μm Vm at 55%B: 0.387 mL F: 1 mL/min

assignment will obviously result in incorrect models being origin, retention times, peak widths, peak area, and possibly developed. peak symmetry for each experimental input condition, which can There are many approaches that can be used, from simple then be entered into the selected retention model. manual ones to fully automated techniques. The authors have Peak assignment or identity is best performed by a practical experience of many of these approaches and will briefly combination of mass spectrometry (MS), diode-array UV spectra describe them and their advantages and disadvantages. (DAD), or peak area ratios. Occasionally standards of the In the development of a LC method, the chromatographer impurities are available that can be individually chromatographed is often faced with several samples from various forced to aid peak assignment, but this approach can result in more degradation studies of the drug substance, drug formulations, samples being run, extending the sequence duration. or mother liquors from the drug synthesis, hence, each input run Many of the commercial automated and semi-automated may require results from multiple chromatograms. Therefore, a peak-tracking approaches of software manufacturers do way is needed to assign the identity of each peak and its origin, not handle multiple samples because they assume that all before constructing a combined table of peak identities and their components of interest are present in the same sample—in a

138 LC•GC Europe March 2018 www.mn-net.com Chromatography chromatography media Manufacturer ofpremium Find whatyouare lookingfor! MACHEREY-NAGEL

OriginalNUCLEOSIL CHROMABOND and highlyefficientNUCLEOSHELL polyester sheets RobustTLC glassplates, economicalaluminumand innovative SPEphases ® columnswithclassicaland ® , professional NUCLEODUR ® HPLC columns ®

Petersson et al.

Figure 7: Flow diagram for modelling based on the manual approach.

Optional determination of Vm, Vd, ECBB and confirm linearity of gradient profile ? Select appropriate samples (that is, forced degradation, mother liquors) ? Select suitable retention model ?

Select appropriate input experiments (typically, tG, T) ? Write methods and sequence (allowing sufficient time to equilibrate the new conditions) ? Collect input data for fitting and evaluation of models (typically performed overnight) ? Perform peak assignment or tracking (using MS, UV DAD, and peak areas), determine peak area, width, and asymmetry ? Produce Excel tables for each input experiment, copy and paste into retention software input section ? Enter chromatographic conditions used ? Combine each individual input run into one large combined table ? Transfer input data to modelling software and fit models ? Evaluate resolution as a function of the operating variable examined (that is typically, gradient shape and T) ? Establish operational parameters fit for purpose (that is, identify robust conditions) ? Collection of data for confirmation of optimal conditions ? Optional in silico evaluation of other column formats, particle size, and flow

real-life situation this is rarely the case. Hence, it is necessary rely on any peak tracking technique; the obvious disadvantage is either to perform pooling of samples, which can result in a that it is time-consuming and prone to operator error. dilution of the peaks, or to rely on a manual peak tracking Examples: Modelling can be used in different ways to approach. It should be stated that retention modelling software assist method development. The most common approach is requires that there should not be any missing data for any of the comprehensive modelling where retention and peak width are analytes that are modelled. modelled for all the peaks of interest. A simpler approach involves Automated Peak Tracking Software: only peak tracking and modelling of the main peak along with the Automated peak tracking software based on MS or UV DAD first and last peaks of interest. The retention models are then used peak-tracking algorithms are available (37,38). The advantages to define conditions in different parts of the experimental domain of using automated peak tracking software are reduced operator that give an acceptable retention and separation window for the error, rapid peak assignment, and fully documented method modelled peaks. Subsequently these conditions are screened and development history. The downside is a significant investment in the one that empirically gives the best separation is selected as both software as well as operator skill because the programme is optimal conditions. aimed at the experienced chromatographer who performs method Conducting a 3 × 3 experiment, which is necessary to development tasks on a routine basis. construct the models, and then performing peak tracking, affords Manual Approach: a better understanding of the separation compared to optimization In laboratories that do not use automated peak tracking software strategies, which simply screen and count the number of peaks it is still common for chromatographers to literally stick the input obtained. Peak tracking and modelling also reduces the risk that a chromatograms onto a wall and then, based on a combination peak could move in and out from under the main peak unnoticed. of UV DAD spectra, MS and peak area ratio, to manually identify Modelling can be used to locate optimal conditions, to assess and annotate each peak on the chromatograms. This can be and optimize robustness, and facilitate the definition of system a laborious process but it is still by far the one most favoured suitability tests by the identification of critical peak pairs. by many chromatographers. An Excel spreadsheet is then Typical prediction errors have previously been reported to Δ Δ constructed, which can be copied and pasted directly into the be accurate to | tg | <1% (34 p. 399), | w | <17% (28, 34 p. Δ retention modelling software. The advantages of this approach 400), and | Rs | <10% (34 p 119, p. 399). As can be seen in include the low cost and simple implementation, and it does not the two examples shown (Tables 2–3) using modern UHPLC

140 LC•GC Europe March 2018 Petersson et al. instrumentation, prediction errors can constructed from gradient to isocratic isocratic predictions generated gradient be observed that are significantly better. predictions of acidic and basic analytes, retention time errors of only <0.3%. This could possibly be related to a better when chromatographed at different Considering the relatively poor performance of the latest generation of pH, highlights that acceptable retention prediction accuracy for gradient to UHPLC equipment, resulting in better time predictions are only achievable for isocratic predictions, we recommend repeatability and reproducibility. retention factors in the range of ~5 to that gradient models be only used to In this article we have defined the ~25. This can also be seen for isocratic predict approximate isocratic conditions, prediction error as the 90th percentile for predictions based on gradient models which can then be used to generate the prediction error, that is | (predicted made for alkyl phenones in Table 3 where accurate isocratic models (See “Isocratic – actual) | 100/actual. In our opinion, the predictions for 6

Conclusions Adhesion to a few basic rules, simple precautions, and the use 'LPHQVLRQV of modern retention-modelling software programmes can assist [[PP the rapid development of highly accurate retention models %$%<'$'GHWHFWRU [[LQ enabling the creation of robust and optimized reversed-phase LC VRVPDOO\RXPLJKW separation using either UHPLC or HPLC conditions (Figure 7). ORVHLW The accuracy of the retention, peak width, and resolution predictions today appears to be better than those quoted in earlier papers. This may simply be a result of improved LC systems, more accurate linear gradients, improved chromatographic reproducibility from run to run, and a better understanding of what input runs and models are required.

Acknowledgements The authors would like to thank Advanced Chromatography Technologies Ltd for supplying the columns used in this work, the Chromatographic Society for support through a summer Visit us at Anallyytiicca 2018, booth A2.407 studentship to B. Boateng, Novo Nordisk for funding J. Field’s PhD Americká 3, 120 00 Praha 2, Czech Republic studies, and Dr O.B. Sutcliffe (Manchester Metropolitan University, www.ecomsro.com M15GD) for kindly synthesizing and supplying the individual 2-, 3-, and 4-methoxydiphenidine hydrochloride isomers.

References (1) http://www.acdlabs.com/products/com_iden/meth_dev/lc_sim/ (accessed 6/10/2017). (2) http://molnar-institute.com/drylab/ (accessed 6/10/2017). (3) http://www.chromsword.com/products/ (accessed 6/10/2017). (4) http://www.datalys.net/ (accessed 6/10/2017). (5) S. Fekete, R. Kormány, and D. Guillarme, LCGC Special Issue 30(6), 14–21 (2017). (6) J.W. Dolan, L.R. Snyder, N.M. Djordjevic, D.W. Hill, D.L. Saunders, L. Van Heukelem, and T.J. Waeghe, J. Chromatogr. A 803, 1–31 (1998). (7) M.R. Euerby, G. Schad, H-J. Rieger, and I. Molnár, Chromatography

142 LC•GC Europe March 2018 Petersson et al.

Today, 13–20 (2010). Strasbourg, France, 7.0th Ed). (8) http://www.smatrix.com/fusion_lc_method_dev.html (accessed 6/10/17) (27) M.R. Euerby and P. Petersson, Personal communication/unpublished (9) D. Spaggiari. V. Desfontaine. A.G-G. Perrenoud, S. Fekete. S. Rudaz. results (2017). and D. Guillarme, J. Chromatogr. A 1371, 244–56 (2014). (28) N. Lundell, J. Chromatogr. 639, 97–115 (1993). (10) A. Tölgyesi, R. Berky, K. Békési, S. Fekete, J. Fekete, and V.K. Sharma, (29) L.R. Snyder and M.A. Quarry, J. Liquid Chromatography 10, 1789–1820 J. Liq. Chrom. Rel. Techn. 36, 1105–1125 (2013). (1987). (11) R. Kormány, J. Fekete, D. Guillarme, and S. Fekete, J. Pharm. Biomed. (30) B.F.D. Ghrist, B.S. Cooperman, and L.R. Snyder, J. Chromatogr. A 459, Anal. 89, 67–75 (2014). 1–23 (1988). (12) R. Hanafi, H. Spahn-Langguth, L. Mahran, O. Heikal, A. Hanafy, H. (31) A. Méndez, E. Bosch, M. Rosés, and U.D. Neue, J. Chromatogr. A 986, Rieger, I. Molnár, and H.Y. Aboul-Enein, Chromatographia 79, 469–477 33–44 (2003). (2012). (32) M.A. Quarry, R.L. Grob, and L.R. Snyder, J. Chromatogr. 285, 19–51 (13) R. Kormány, I. Molnár, and J. Fekete, J. Pharm. Biomed. Anal. 135, 8–15 (1984). (2017). (33) https://www.sigmaaldrich.com/content/dam/sigma-aldrich/docs/ (14) R. Kormány, I. Molnár, and J. Fekete, LCGC North America 32, 354–363 Supelco/General_Information/t408143.pdf (accessed 6/10/2017). (2014). (34) L.R. Snyder and J.W. Dolan, High performance gradient elution: The (15) P. Petersson, J. Munch, M.R. Euerby, A. Vazhentsev, S.K. Bhal, and K. practical application of the liner-solvent strength model (John Wiley & Kassam, Chromatography Today 7, 15–18 (2014). Sons, Hoboken, New Jersey, USA, 2007). (16) E. Tyteca, J.-L. Veuthey, G. Desmet, D. Guillarme, and S. Fekete, (35) P. Jandera, J. Chromatogr. A 1126, 195–218 (2006). Analyst 141, 5488–5501 (2016). (36) M.T.W. Hearn and G. Zhao, Anal. Chem. 71, 4874–4885 (1999). (17) S. Fekete, J.-L. Veuthey, A. Beck, and D. Guillarme, J. Pharm. Biomed. (37) G.A. Von Wald and M.T. Vagnini, “Evaluation of ACD/Autochrom Anal. 130, 3–18 (2016). Software for LC Method Development,” paper presented at HPLC, San (18) S. Fekete, S. Rudaz, J. Fekete, and D. Guillarme, J. Pharm. Biomed. Francisco, California, USA, 2016. Anal. 70, 158–168 (2012). (38) K. Jayaraman, A.J. Alexander, Y. Hu, and F.P. Tomasella, Analytica (19) S. Fekete, I. Molnár, and D. Guillarme, J. Pharm. Biomed. Anal. 137, Chimica Acta 696, 116–124 (2011). 60–69 (2017). (20) D.E. Bautz, J.W. Dolan, and L.R. Snyder, J. Chromatogr. 541, 1–21 Melvin Euerby is the Principal of Shimadzu’s Centre of (1991). Excellence for Liquid Chromatography and Professor at the (21) R.C. Kong, B. Swachok, and S.N. Deming, J. Chromatogr. 199, 307–316 (1980). University of Strathclyde and the Open University. (22) M.R. Euerby, J. Hulse, P. Petersson, A. Vazhentsev, and K. Kassam, Bernard Boateng was a MSc student at the University Analytical and Bioanalytical Chemistry 407, 9135–9152 (2015). of Strathclyde and is now a PhD student at the National (23) J.R-Montano, C.O-Bolsico, M.J. Ruiz-Angel, and M.C. Garcia-Alvarez- University of Ireland in Galway. Coque, J. Chromatogr. A 1344, 31–41 (2014). (24) R.M. El-Nashar and H.Y. Aboul-enein, Chirality 17, 506–513 (2013). Jennifer Field is a PhD student at the University of (25) I. Molnár, LCGC Europe 14(4), 231–237 (2001). Strathclyde, UK. (26) Section 2.2.4.6, “Chromatographic Separation Techniques,” European Patrik Petersson is a Principal Scientist at Novo Nordisk A/S Pharmacopoeia (European Directorate for the Quality of Medicines, and Associate Professor at Uppsala University, Sweden.

Leverage the outstanding inertness, low bleed, and high reproducibility of Rxi® 3-in-1 technology to gain: • Accurate Data • The Right Results Fast • Maximized Instrument Uptime Put Rxi® columns to work in your lab. Visit www.restek.com/rxi and order yours today.

Rxi® columns come with an unbeatable guarantee. That is Restek® Pure Satisfaction.

Pure Chromatography www.restek.com/rxi

www.chromatographyonline.com 143 LC TROUBLESHOOTING

Eluent Preparation for Hydrophilic Interaction Liquid Chromatography, Part 2: pH, Buffers, and Gradient Elution

Dwight R. Stoll1 and Claudia Seidl2, 1LC Troubleshooting Editor, 2University of São Paulo, Brazil

When preparing buffers for hydrophilic interaction chromatography (HILIC) separations, is the buffer concentration considered relative to the aqueous portion of the eluent only, or to the aqueous–organic mixture?

In the January instalment of LC analyte retention, especially if the in the eluent that matters, or the pH Troubleshooting (1), we addressed analyte contains one or more ionizable of the aqueous–organic mixture. In several practical issues related to the functional groups (2). The same is other words, if we are aiming to use preparation of eluents for hydrophilic true for HILIC separations, and the an ammonium acetate buffer at pH interaction chromatography (HILIC) effects can be even more pronounced 6, should we determine that pH in separations. We discussed how than those in reversed-phase LC the aqueous component of the eluent different methods of preparing if a significant component of the before mixing it with acetonitrile eluents composed of acetonitrile retention of an analyte is dependent or after mixing? The fundamental and aqueous components can lead on electrostatic (that is, ion–ion) aspects underlying answers to this to very different retention times and interactions. Figure 1 shows the question have been discussed at selectivities, and that we can help our impact of pH on the retention of two length over the years in LCGC (3–5) colleagues by providing more detail compounds on two different stationary and elsewhere (6). Here, we briefly when documenting eluent preparation phases under HILIC conditions. We review the main points, because they procedures, whether for standard see that 5-methylcytidine, which is remain very important to the general operating procedures (SOPs) or journal not ionized in water at pH values topic of using buffers in HILIC articles. We also discussed the effects below about 12, is similarly retained separations. of different buffer cations (for example, on the two columns, and the retention The dissociation constants of acids + + Na and K ) on retention and selectivity is nominally unaffected by pH. (that is, pKas) in water are influenced in HILIC separations. In this second Nortriptyline, on the other hand, is by the presence of organic solvents instalment on the topic of eluent much more retained on the bare like acetonitrile and methanol. The preparation for HILIC separations, we silica column, and the dependence changes in pKa with the addition of discuss some of the other practical of retention on mobile-phase pH is different levels of acetonitrile are factors to consider concerning buffers much stronger than it is on the OH5 shown in Table 1 for some common used in the aqueous component of the column. This result can be rationalized buffering agents and simple analytes. eluent. Here too, different methods of by understanding that nortriptyline We see that for some compounds preparing buffers can have a significant is present primarily in the protonated the pKa values increase upon impact on retention, selectivity, and and positively charged form in water addition of acetonitrile, meaning the reproducibility of HILIC separations. at pH values below about 9, and that compounds become less acidic, Dwight Stoll deprotonated and negatively charged whereas for others the pKa values silanol groups on the exposed silica decrease. This differential effect on pH for HILIC Separations: surface can interact strongly with the pKa that depends on the chemistry Its Importance and How to positively charged analyte through of the compound can have a Measure It ion–ion interactions. significant effect on the behaviour In reversed-phase liquid As a practical matter, then, we are of a separation. Suppose we use an w chromatography (LC), the pH of the faced with the question of whether it acetate buffer with a measured wpH w eluent can significantly influence is the pH of the aqueous buffer used of 6.0 (the wpH nomenclature means 144 LC•GC Europe March 2018 LC TROUBLESHOOTING that the pH is being measured in a completely aqueous Figure 1: Retention time under gradient elution conditions for solution, and that the pH electrode has been calibrated 5-methylcytidine and nortriptyline on two different HILIC columns: using pH standards that are aqueous solutions), and (a) bare silica (Ascentis Express HILIC); and (b) OH5 (Ascentis Express OH5). Mobile-phase A: 100 mM ammonium acetate pyridine is one of our analytes. Given that the pKa of the w pyridinium ion in water is about 5.2, roughly 80% of the adjusted to wpH 6.0 with acetic acid, 100 mM ammonium w pyridine injected into a column containing only the acetate formate adjusted to wpH 3.0 with formic acid, or 100 mM ammonium bicarbonate adjusted to w pH 9.0 with ammonium buffer will be present in the column as pyridine and 20% w hydroxide; mobile-phase B: acetonitrile; mobile-phase gradient: will be present as the protonated pyridinium ion. However, 95–80% B in 5 min with a hold for 1 min at 80% B; flow rate: if we add acetonitrile to the eluent such that only 20% of 0.6 mL/min; column temperature: 35 °C. s the eluent is acetate buffer, the wpH (this nomenclature (a) means that the pH is being measured in a solution 5 containing organic solvent, while the pH standards are still 4 pH 3 pH 6 pH 9 in aqueous solution) will increase to about 8, whereas the pKa of the pyridinium ion will decrease to about 4. Under 3 these conditions, virtually all of the pyridine injected into the mobile phase will be present as the deprotonated 2 pyridine free base, which in turn will significantly affect

Retention time (min) 1 the way the analyte interacts with the stationary and mobile phases, especially relative to other analytes. 0 Clearly, these effects are important to HILIC separations 5-Methylcytidine Nortriptyline (b) particularly because much higher levels of acetonitrile are 5 used in the eluent compared to what is normally used in 4 reversed-phase LC. When possible, one should anticipate these effects during method development and decisions 3 involving eluent pH. However, given that the change in pKa because of the organic solvent is unknown, or at least not 2 available in the literature, for many compounds of interest,

Retention time (min) 1 simple scouting experiments are the most reliable way to understand the impact of eluent pH on retention of an 0 ionizable compound under HILIC conditions. As described 5-Methylcytidine Nortriptyline by John Dolan in an article on buffers recently, systematic variation of the eluent pH should certainly be a part of method robustness testing to make sure that the method does not use a pH where analyte retention is sensitive to small changes in pH (2). BULK SILICA Now, let’s come back to the question we started with: should the pH be measured in the aqueous component of the eluent, or in the aqueous–organic mixture? Our view is s From 40Å up to that measuring or even using wpH calibrants in aqueous– s organic mixtures to obtain the spH is definitely useful 300Å pore s for mechanistic studies of HILIC retention. When wpH can be measured reproducibly it is helpful for estimating size the ionization state of the analyte in the eluent. However, these steps are not necessarily required for practical work. Calibrating the pH electrode using aqueous calibrants and then measuring the pH of the aqueous buffer alone to w obtain wpH provides a useful reference point for the pH of the eluent.

Aqueous Buffers: Preparation and Impact on Retention and Selectivity As we stated at the beginning of this series of articles, we have been motivated to discuss the details of eluent preparation for HILIC separations in part because of the ambiguity we find in many descriptions of eluent conditions in the chromatographic literature. It is common to see a description of an eluent in the experimental section of a paper describing a HILIC separation that goes something like this: Description: 90:10 acetonitrile–buffer, upag agtWannenhofstrasse 1 CH-5726 Unterkulm with 10 mM ammonium acetate at pH 6. The problem U tG with descriptions like these is that they are ambiguous. XXXVQBHDIFNDPNtFNBJMTBMFT!VQBHDIFNDPN Assuming we are preparing 1 L of eluent, this description www.chromatographyonline.com 145 LC TROUBLESHOOTING

Table 1: Change in pKa of common buffering agents and simple analytes with addition Table 2: Two different approaches to of acetonitrile to aqueous solutions preparing A and B solvents for HILIC separations with gradient elution Acetonitrile Fraction of Solution (% by Volume) w Compound wpKa Scenario 1 40 60 80 90 100 mM ammonium A solvent Acetic acid 4.76 +1.2 +2.2 acetate, pH 6 + + Benzoic acid 4.21 1.0 2.1 B solvent Acetonitrile Methylamine 10.6 -0.3 -0.3 Scenario 2 Pyridine 5.17 -0.6 -0.9 -1.2 -0.4 80:20 acetonitrile–25 mM *Data are adapted from reference 6 A solvent ammonium acetate, pH 6 could reasonably lead to at least w pH to 6.0 by addition of acetic w 95:5 two different solutions that would be acid, and then add water to bring the acetonitrile–100 mM B solvent very different in composition. Case buffer to a final volume of 1 L. Then, ammonium acetate, A: We could prepare 1 L of buffer by we would combine 100 mL of this pH 6 adding 10 millimoles of ammonium buffer with 900 mL of acetonitrile to acetate to 900 mL of water, adjusting produce the 90:10 organic–aqueous w the wpH to 6.0 by addition of acetic eluent—the same as in Case A. Now, combining the aqueous buffer acid, and then add water to bring the the eluent produced in Case A is with the acetonitrile there will be buffer to a final volume of 1 L. Then, consistent with the description in the 10 millimoles of ammonia–ammonium we would combine 100 mL of this sense that the ammonia–ammonium present in 1 L of eluent (that is, the buffer with 900 mL of acetonitrile to concentration in the buffer before ammonia–ammonium has been produce the 90:10 organic–aqueous adding it to the acetonitrile would diluted by a factor of 10 by addition eluent. Case B: In this case we could be 10 mM. However, one could of the acetonitrile). Clearly, there prepare 1 L of buffer by adding argue that the eluent produced is a difference in the properties of 100 millimoles of ammonium acetate in Case B is consistent with the the eluents produced in these two to 900 mL of water, adjusting the description in the sense that after cases. And, the consequences of

SERVING ROYALTY. EXCEEDING EXPECTATIONS. EVERY MOMENT.

• Provider of top brand HPLC instrumentation products

• Equivalent to corresponding OEM products

• Serving customers for over 30 years

• Reduce product repair expenses by 15% to 30%

• Lifetime Warranty on manufacturing defects

www.sciencix.com 800.682.6480 [email protected]

146 LC•GC Europe March 2018 LC TROUBLESHOOTING the difference can be significant in the case of HILIC Figure 2: Retention time under gradient elution conditions for separations. Figure 2 shows the retention times obtained for different probe compounds on (a) bare silica (Ascentis Express neutral, acidic, and basic probe compounds on two different HILIC); and (b) OH5 (Ascentis Express OH5) HILIC columns with w HILIC columns using either 10 or 100 mM ammonium either 10 or 100 mM ammonium acetate, wpH 6 buffer used in acetate, w pH 6.0 in the preparation of a 95:5 organic– the preparation of an acetonitrile–buffer eluent. Mobile-phase A: w w aqueous eluent. We see that whereas there is very little 100 mM or 10 mM ammonium acetate, adjusted to wpH 6.0 with acetic acid; mobile-phase B: acetonitrile; mobile-phase gradient: impact of the buffer concentration on the neutral compounds 95–80% B in 2.5 min for the bare silica column, and 3.27 min for inosine and guanosine, the effect on the retention of the OH5 column; flow rates: 1.2 mL/min (bare silica) and 0.9 mL/ amitriptyline is most pronounced with the bare silica column. min (OH5); column temperature: 35 °C. This is expected given that ion–ion interactions are expected (a) to contribute substantially to the retention of amitriptyline 2.5 under these conditions (7). We note here that these retention 2.0 times were obtained under gradient elution conditions, and 10 mM buffer so the near doubling of retention time for amitriptyline is 1.5 100 mM buffer even more striking. Our point here is not that the method of eluent preparation described in Case A or B is better per se, 1.0 but that the method used can have a significant impact on retention, and that descriptions of experimental conditions Retention time (min) 0.5 used for these separations should be written explicitly to 0.0 avoid ambiguity. Inosine Benzoic acid Amitriptyline (b) 2.5 Gradient Elution 2.0 The last topic we’d like to touch on here concerns gradient elution for HILIC separations, particularly the composition 1.5 of the two solvents used in the gradient elution program. Suppose the eluent delivered to our HILIC column is 1.0

nominally 90:10 acetonitrile–buffer—where the buffer Retention time (min) 0.5 alone is 100 mM in ammonium acetate—but that we would like to do gradients from more to less acetonitrile 0.0 Guanosine Benzoic acid Amitryptiline to accommodate analytes in our sample with a range of hydrophilicity. The question here is, should we prepare A and B solvents such that the ammonium acetate concentration of the eluent delivered to the column is constant over the duration of the gradient, or can we LUCK HAS NOTHING just have the ammonium acetate in the A solvent, which will result in not only an acetonitrile gradient but also an TO DO WITH IT ammonium acetate gradient? Table 2 shows two possible scenarios that reflect this difference in operation. With decades of experience and the widest range of In Scenario 1, the number of moles of ammonium– chiral phases, you don’t need luck to know DAICEL ammonia per litre of eluent delivered to the column will columns will find the best method for your molecule. actually increase from 5 to 20 millimoles per litre over the • ULTRAFAST SEPARATIONS course of a solvent gradient program that runs from 95% • HIGHER RESOLUTION POWER to 80% B solvent. On the other hand, with Scenario 2 the • EASE OF METHOD TRANSFER ammonium acetate concentration delivered to the column stays constant at 5 millimoles per litre of eluent over a Don’t take any chances with your method gradient that runs from 100% B to 100% A. Clearly the development. Success requires the best. chemical dynamics inside the column will be much more DAICEL and Chiral Technologies – complex in Scenario 1, but the upside of this approach is the Chiral Chromatography Experts. that the A and B solvents can be used very flexibly during Contact us today. method development. On the other hand, in Scenario 2 WWW.CHIRALTECH.COM there are fewer mobile phase variables changing inside the column during the gradient, but making adjustments to the A and B solvents during method development is very tedious. In our own work, we have compared the repeatability of HILIC separations carried out under the two scenarios described here. Figure 3 shows three replicate gradient elution separations for each scenario. Visually it is obvious that while the two scenarios produce different retention patterns, the repeatability of each separation is really very good. With this result in hand, we most commonly do our method development work using the scheme described in Scenario 1. It is likely that © 2018 CHIRAL TECHNOLOGIES www.chromatographyonline.com LC TROUBLESHOOTING

Figure 3: Representative chromatograms from replicate separations of a mixture of seven probe compounds. Note that the two conditions were run at different times and the differences in peak height are simply a result of different preparations of the sample mixture. Column: 100 mm × 2.1 mm, 2.7-μm Supelco Ascentis Express HILIC. Scenario 1: mobile-phase A: 100 mM ammonium acetate at pH 6 adjusted w to wpH 6.0 with acetic acid; mobile-phase B: acetonitrile; mobile-phase gradient: 95–80% B in 2.5 min; flow rate: 1.2 mL/min; column w temperature: 35 °C. Scenario 2: mobile-phase A: 5:95 (v/v) 100 mM ammonium acetate, wpH 6–acetonitrile; mobile-phase B: 20:80 (v/v) w 25 mM ammonium acetate wpH 6–acetonitrile; mobile-phase gradient: 0–100% B in 2.5 min; flow rate: 1.2 mL/min; oven temperature: 35 °C. Peaks: 1 = acenaphtalene, 2 = pyridine, 3 = deoxyuridine, 4 = benzoic acid, 5 = inosine, 6 = amitriptyline, 7 = methylguanosine.

30 60 Scenario A 2 Scenario B 3 25 AU) 50 AU) m ( m 20 6

40 ( 3 4 ance 30 15 ance rb rb

so 20 1 10

4 so 1 Ab 5 6 5 7 10 2 7 Ab 5 0 0 0 1 2 3 0 1 2 3 Time (min) Time (min) 60 35

50 AU) 30 m ( AU) 40 25 m (

ance 20 30 rb

ance 15 so

rb 20

Ab 10 so 10 5 0 0 0 1 2 3 0 1 2 3 Time (min) Time (min) 60 40 35 50 AU) AU) Ab

m 30 m ( ( 40 25

30 ance 20 ance rb

rb 15

20 so so 10 Ab Ab 10 5 0 0 0 1 2 3 0 1 2 3 Time (min) Time (min)

this type of comparison carried out method of buffer preparation is used (7) A. Kumar, J.C. Heaton, and D.V. McCalley, on different high performance liquid for your HILIC separation, be sure to J. Chromatogr. A 1276, 33–46 (2013). doi:10.1016/j.chroma.2012.12.037. chromatography (HPLC) systems describe your work using a complete, with different pump and eluent mixer unambiguous description of how exactly Dwight Stoll is the editor of “LC designs will yield different retention the buffer was prepared so that others can Troubleshooting”. Stoll is an associate patterns and different degrees of reproduce the results of your work (5). professor and co-chair of chemistry variation between the two scenarios. at Gustavus Adolphus College in St. Understanding these differences will Acknowledgements Peter, Minnesota, USA. His primary be especially important in situations The columns used to obtain the research focus is on the development of where a method will be transferred data shown here were provided by 2D-LC for both targeted and untargeted to a laboratory running different Dave Bell at MilliporeSigma. C. Seidl analyses. He has authored or instrumentation. acknowledges support from the São coauthored more than 50 peer-reviewed Paulo Research Foundation - FAPESP publications and three book chapters in Summary - Process number 2016/02941-5 for her separation science and more than 100 Some details related to mobile phase contributions to this article. conference presentations. He is also a preparation that can be assumed or member of LCGC ’s editorial advisory taken for granted in reversed-phase LC References board. Direct correspondence to: LCGC separations with little consequence can (1) C.A. Lucy, C.B. Craven, C. Seidl, and D.R. [email protected] Stoll, LCGC Europe 31(1), 22–27 (2018). be much more consequential in HILIC (2) J.W. Dolan, LCGC Europe 30(1) 30–33 Claudia Seidl is a postdoctoral fellow separations. In many cases there is no (2017). in the Chemistry Department at the single “right” answer, but we think data (3) G.W. Tindall, LCGC North Am. 20(11), University of São Paulo, Brazil. She of the type shown here can inform the 1028–1032 (2002). developed a study focused on (4) G.W. Tindall, LCGC North Am. 20(12), adoption of a set of best practices that 1114–1118 (2002). column re-equilibration in HILIC under will make method development for HILIC (5) G.W. Tindall, LCGC North Am. 21(1), Professor Dwight R. Stoll’s supervision separations more robust, simple, and 28–32 (2003). while she was a visiting postdoctoral (6) M. Rosés, J. Chromatogr. A 1037, repeatable. We close by reiterating the 283–298 (2004). doi:10.1016/j. fellow at Gustavus Adolphus College, in suggestion of Tindall that no matter what chroma.2003.12.063. St.Peter, Minnesota.

148 LC•GC Europe March 2018 HILICON

iHILIC® -Fusion advances HILIC separations in UHPLC and HPLC

iHILIC® -Fusion iHILIC® -Fusion(+) iHILIC® -Fusion(P) Silica based Silica based Polymer based

Charge modulated hydroxyethyl amide/amide HILIC columns Complementary selectivities for polar/hydrophilic analytes Excellent durability and ultra-low bleeding for LC-MS Versatile columns for ”Omics” studies iHILIC®® -Fusion and iHILIC -Fusion(+): pH 2-8; 1.8, 3.5, and 5 μm iHILIC® -Fusion(P): pH 1-10; 5 μm

A Compendium of GC Detection, Past and Present

John V. Hinshaw, GC Connections Editor

Gas chromatography makes use of a wide variety of detection methods. In addition to the most often used flame-ionization detection (FID), electron-capture detection (ECD), thermal conductivity detection (TCD), and mass-selective detection (MSD), the list of other detection methods is long. They really shine when deployed properly, but their properties and applications can be a bewildering alphabet soup. This instalment presents a compendium of gas chromatography (GC) detection methods, both past and vanished as well as those that are current and relevant to today’s separation challenges.

In the six and a half decades since methods, the sheer number that are chromatographic detector. A universal its inception, gas chromatography in active use or have been in the detector, such as the thermal (GC) has seen a wide variety of past is remarkable—nearly 30 are conductivity detector, responds to detection methods. Four of them listed here in Tables 1 and 2. This any compound in the column effluent arguably account for greater than is not a comprehensive list. Some that is different than the carrier gas. 90% of applications today (1): chromatographers have chosen to A specific detector responds only to flame-ionization detection (FID), use other names and abbreviations, certain chemically related materials. thermal conductivity detection and certainly other varieties may The electron-capture detector with (TCD), electron-capture detection be found that are not as visible to halogenated compounds, or the (ECD), and mass-selective detection literature searches. aptly named nitrogen–phosphorus detector with nitrogen or phosphorus GC detection methods In the six and a compounds, are both specific cover a wide range of half decades since detectors. Selective detectors respond to groups of compounds sensitivity and selectivity its inception, gas that possess a common measurable that is unsurpassed by any chromatography (GC) has characteristic such as mass or other separation method. seen a wide variety of spectral absorbance. MSD falls into this group along with VUV and detection methods. Four infrared detection (IRD or GC–IR). (MSD). Many more detectors are of them arguably account found in modern chromatography for greater than 90% of A universal detector, laboratories in smaller quantities, and applications today (1): a few have found their way into the such as the thermal dusty closet of retirement. Ranging flame-ionization detection conductivity detector, from FID to electroantennographic (FID), thermal conductivity responds to any detection (EAD), which uses insect detection (TCD), compound in the column antennae as the sensing elements, electron-capture detection GC detection methods cover a wide effluent that is different range of sensitivity and selectivity (ECD), and mass-selective than the carrier gas. that is unsurpassed by any other detection (MSD). separation method. In 2015, The boundaries between these McNair and Schug, writing in “GC Detector Taxonomy classification are not always clearly Connections” (2), addressed the The International Union of Pure and defined. Flame photometric detection history and capabilities of eight major Applied Chemistry (IUPAC) recently (FPD), for example, responds to GC detection methods, ranging from published updated recommendations selected spectral emission lines of TCD to the newest member: vacuum regarding separation science eluted compounds, and might be ultraviolet (VUV) detection. Along terminology (3). The publication considered a selective detection with these mainstream detection defines three general types of method, but the spectral lines are

150 LC•GC Europe March 2018 Conserve bench space while providing on-demand hydrogen to 1000 mL/min

and zero air to 5 lpm

VICI DBS FID Tower Hydrogen + Air Generator

Continuous GC gas supply (no helium, hydrogen and zero air cylinders) Deionized water is automatically loaded from smart internal system tanks with a 3 liter capacity, providing up to 3000 liters of H2 before a refill is required.

Improves workplace safety and lab efficiency Ultra high purity carrier grade gas with convenient software control and safety alarm capability. It gives the GC user the option for on-demand hydrogen gas while managing safety concerns. Flow rate from 100 to 1000 mL/min

Increases GC performance and sensitivity Electrolytic dissociation separates the water into its two main components: hydrogen ready for analytical Fits neatly use, and oxygen that is released into the air. next to GC on bench Ultra high gas purity without the use of dessicants, caustics or palladium No acid or alkaline solutions are used in the hydrogen generation cycle and the patented automatic drying system ensures the maximum grade of hydrogen purity.

[email protected] www.vicidbs.com +41 (41) 925 62 00 GC CONNECTIONS

Table 1: GC detection methods

Detection Method Abbreviation Description

Atomic emission detection excites eluted compounds in a helium microwave-induced plasma. The resulting atomic emission is detected with an optical spectrometer in the Atomic AED 160–800 nm range. AED is element-specific by observing selected emission lines, much emission detection like FPD but with simultaneous multiple emission monitoring. It has sensitivity on par with FID.

Barrier BID uses near-UV light from a dielectric-barrier discharge plasma to ionize eluted BID ionization detection compounds. It has sensitivity similar to FID while exhibiting near-universal response.

An electron-capture detector ionizes solutes by collision with metastable carrier-gas Electron-capture molecules produced by β-emission from a radioactive source such as 63Ni. ECD is one ECD detection of the most sensitive detection methods, and responds strongly to halogenated solutes and others with a high electron-capture cross-section.

A flame-ionization detector ionizes hydrocarbon solutes in a hydrogen–air flame. The resulting electrons are collected and measured with a sensitive electrometer. FID is a Flame-ionization FID nearly universal detection method that responds strongly to most classes of organic detection compounds. Little to no response occurs for CO, CO2, water, and other compounds that lack C-H bonds.

The flame-photometric detector burns eluted solutes in a hydrogen–air flame. The Flame-photometric resulting atomic emission lines for sulphur, tin, or phosphorus are selected with an FPD detection optical interference filter and detected with a photomultiplier. Different optical filters are substituted to observe the emission lines of each specific element.

In its reductive mode, the electrolytic-conductivity detector catalytically reacts (Hall) Electrolytic- halogen-containing solutes with hydrogen to produce strong acid by-products that conductivity HECD, ElCD are dissolved in a working fluid. The acids dissociate, and the increased electrolytic detection conductivity of the solution is measured. Other operating modes modify the chemistry for response to nitrogen- or sulphur-containing substances.

The helium ionization detector operates by creating a helium plasma using Helium ionization radio-frequency excitation; the plasma emits energetic photons that ionize eluted HeID, HID, detection, discharge compounds. Additional electrons and metastable helium atoms may also contribute to DID ionization detection the response. Earlier versions of these detectors used a radioactive beta particle source similar to ECD. See also PDD.

A GC–IR detector obtains mid-infrared spectra of eluted solutes either by direct absorption in a light pipe for gas-phase transmission spectra, or by cryogenic solute IRD or Infrared detection trapping on a rotating gold-plated drum or Zn-Se disk for solid-state spectra. IRD GC–IR distinguishes and identifies eluants by their spectra and by library search. Some peak deconvolution is possible but good peak resolution is preferable.

MSD provides searchable mass spectra of chromatographic peaks. A variety of mass analyzers have been used, including quadrupole, electric and magnetic sector, ion trap, and time of flight (TOF). Various characteristic mass-fragmentation patterns are provided Mass-selective by sources such as electron ionization (EI) and chemical ionization (CI), both positive detection, and negative (NCI). Total ion current (TIC) chromatograms resemble those from other MSD mass-spectral ionization detectors like the flame-ionization detector. Single-ion monitoring (SIM) and detection multiple-ion monitoring (MIM) measure selected ions to deduce structural information or to deconvolve coeluted peaks. A second mass analyzer can be added for tandem GC–MS/ MS, which engenders further differentiation via selected-reaction monitoring (SRM) and multiple-reaction monitoring (MRM).

Nitrogen–phosphorus NPD catalytically ionizes N- or P-containing solutes on a heated rubidium or cesium detection, thermionic NPD, TSD, surface in a reductive atmosphere. NPD is highly specific with sensitivity somewhat specific detection, TID better than FID. Other modes of operation give selectivity for a variety of other thermionic ionization heteroatoms. detection

The photoionization detector ionizes solute molecules with photons in the ultraviolet (UV) Photoionization energy range from a discharge lamp. PID is a specific detection method that responds PID detection to aromatics and olefins when operated in the 10.2 eV range, and can respond to other materials with a more energetic light source.

Postcolumn reactors convert eluted compounds to others that have different detection characteristics. The most familiar postcolumn device is a nickel-based reducing catalytic converter that produces methane from CO and CO2, commonly known as a methanizer. Postcolumn reactors PCR The device makes sensitive detection of these compounds possible with a FID. A recent development, the Polyarc reactor (Activated Research Company) converts all carbon-containing peaks to methane in a two-step process of oxidation to CO2 followed by reduction to CH4.

152 LC•GC Europe March 2018 GC CONNECTIONS

Table 1: (continued) GC detection methods

Detection Method Abbreviation Description

In its helium ionization mode, PDD uses a pulsed, high-voltage direct current ionization source and helium gas to create photons that ionize eluted peaks. The resulting electrons Pulsed discharge are collected across biased electrodes. In this mode, PDD is a universal detection method PDD detection with sensitivity in the low parts-per-billion (ppb, 10-9) range. The addition of a noble gas (Ar, Xe, Kr) can produce specific responses to aromatics and other chemical species. PDD also can be operated in a halogen-specific electron-capture mode, similar to ECD.

A specific detection method that responds to sulphur-containing compounds by Sulphur generating and measuring light from chemiluminescence. Compounds are combusted at chemiluminescence SCD high temperature to form SO, which then reacts with ozone to produce chemiluminescent detection emission in the 300–400 nm range.

TCD measures the differential thermal conductivity of column effluent with reference Thermal-conductivity to pure carrier gas. TCD is a universal detection method with moderate sensitivity. detection, also TCD Katharometer is an older name that refers to the use of heated filaments to respond katharometer to changes in thermal conductivity. Some thermal conductivity detectors make use of thermistor beads for this function.

The vacuum ultraviolet detector measures the near-UV absorption spectrum of eluted Vacuum ultraviolet compounds at wavelengths from 115 to 240 nm. It responds to compounds that FID VUV detection does not, such as CO, O2, and water, while yielding unique spectra that can deconvolve difficult-to-separate peaks such as m- and p-xylene.

Table 2: Less common or obsolete GC detection methods

Detection Method Abbreviation Description

The acoustic flame detector is a unique device built to monitor the oscillation Acoustic flame detection AFD frequency of an unstable flame jet as compounds are eluted through it. AFD also has found application in supercritical fluid chromatography (SFC).

Perhaps the most unique GC detection method, EAD has been used to identify the pheromones of moths, bees, beetles, and other insects (5). A single insect Electroantennographic antenna or single sensilla is attached to electrodes and exposed to humidified EAD detection column effluent. Elution of an active compound results in a neuroelectrical response. Compound identification can then be performed with MSD or other selective detection methods.

The gas density balance was an early GC detector that used a thermistor-based anemometer to measure differences in the density of pure reference carrier Gas density balance GDB gas and the GC column effluent. It was supplanted by the thermal conductivity detector.

LOD used the photoacoustic effect with a tunable CO2 laser to produce Laser optoacoustic limited-range IR spectra of eluted compounds. The sensitivity of LOD was up LOD detection to 10 times better than FID (4), but it lacked sufficient selectivity to be useful for discrimination of coeluted peaks.

A number of researchers have interfaced GC with NMR, either by stopped-flow Nuclear magnetic gas-phase spectral measurement or by semipreparative liquid-phase collection GC–NMR resonance in NMR tubes. Both 1H and 13C NMR spectral data have been used to elucidate ancillary structural information for unknown compounds.

The ultrasonic detector was an early GC detector that used a pair of acoustical cavities resonant at ultrasonic frequencies to produce a differential beat signal Ultrasonic detection USD stemming from changes in the velocity of sound in the carrier gas as compounds were eluted. Sensitivity was relatively poor. emitted only by molecules containing A specific detector not its spectrally selective nature, certain elements, and thus FPD also responds only to certain so it is best considered a specific is a specific detection method. In detection method. The same logic a practical sense, FPD is used for chemically related can be applied to other detection its element-specific characteristics, materials. methods. www.chromatographyonline.com 153 GC CONNECTIONS

There is no standard for naming have their own unique set that reaction detector. The bulk of GC chromatographic detection methods. fortunately are related almost detectors continues to see significant GC detection method names most often one-to-one with GC devices, application, while only a few have really reflect modes of selectivity and specificity. detectors, inlets, columns, and so-on. fallen away into disuse. FID, photoionization detection (PID), and For the new (gas) chromatographer many others are generally named after the sheer number of terms is References their operating principles. NPD, named bewildering. Perhaps this list can be (1) Author’s estimate, not based on pub- for its element specificity, has an alias of assistance navigating the detector lished data. that refers to its physics: thermionic bazaar. (2) H.M. McNair and K.A. Schug, LCGC specific detection (TSD). The latter name North Am. 33(1), 24–33 (2015). is broader and encompasses other There is no standard for (3) T.A. Maryutina, E.Y. Savonina, P.S. operating modes of thermionic detection naming chromatographic Fedotov, R.M. Smith, H. Siren, and that are sensitive to other heteroatoms. detection methods. D.B. Hibbert, Pure Appl. Chem. 90(1), GC detection method 181–231 (2018). Gas chromatography (4) L.B. Kreuzer, Anal. Chem. 50(6), names most often reflect continues to evolve. Every 597A–606A (1978). modes of selectivity and (5) J.A. Byers, J. Neurosci. Methods 135, year new GC-related specificity. 89–93 (2004). devices appear in publications and in the “GC Connections” editor John V. Gas chromatography continues to Hinshaw is a senior scientist at marketplace. evolve. Every year new GC-related Serveron Corporation in Beaverton, devices appear in publications and Oregon, USA, and a member of Scientists have a love–hate affair in the marketplace. Three new GC LCGC Europe’s editorial advisory with acronyms and abbreviations. detectors have appeared in recent board. Direct correspondence about They are convenient, short, and easy years—vacuum ultraviolet and barrier this column to the author via e-mail: to misuse. Gas chromatographers ionization detectors, and a postcolumn [email protected]

Visit us! Hall A3, Booth 116A BIOPHARMACEUTICALBIOPHARMACEUTICAL PERSPECTIVES PERSPECTIVES

Peptide Mapping of Monoclonal Antibodies and Antibody–Drug Conjugates Using Micro-Pillar Array Columns Combined with Mass Spectrometry

Koen Sandra1,2, Jonathan Vandenbussche1, Isabel Vandenheede1, Bo Claerebout3, Jeff Op de Beeck3, Paul Jacobs3, Wim De Malsche4, Gert Desmet4, and Pat Sandra1,2 1Research Institute for Chromatography (RIC), Kortrijk, Belgium, 2 anaRIC biologics, Ghent, Belgium, 3 PharmaFluidics, Ghent, Belgium, 4 Vrije Universiteit Brussel, Brussels, Belgium

Monoclonal antibodies are becoming a core aspect of the pharmaceutical industry. Together with a huge therapeutic potential, these molecules come with a structural complexity that drives state-of-the-art chromatography and mass spectrometry (MS) to its limits. This article discusses the use of micro-pillar array columns in combination with mass spectrometry for peptide mapping of monoclonal antibodies (mAbs) and antibody–drug conjugates (ADCs). Micro-pillar array columns are produced by a lithographic etching process creating a perfectly ordered separation bed on a silicon chip. As a result of the order existing in these columns, peak dispersion is minimized and highly efficient peptide maps are generated, providing enormous structural detail. Using examples from the author’s laboratory, the performance of these columns is illustrated.

The discovery of monoclonal antibodies that highly toxic drugs can selectively with the variability associated with (mAbs) in the 1970s and the realization be delivered to tumour cells, thereby the conjugation. Conjugation typically of their therapeutic potential in the substantially lowering side effects as takes place on the amino groups of following decades has been a key typically experienced with classical lysine residues or on the sulphydryl milestone in medicine (1–3). Today, chemotherapy. Currently two ADCs are groups of interchain cysteine residues. over 70 mAbs have received regulatory marketed and over 30 are in clinical With 80 to 100 lysine and only eight approval in the US and Europe, of trials (12). interchain cysteine residues available, which over 20 display blockbuster Together with a huge therapeutic lysine conjugation yields a more status, and over 50 are in late-stage potential, mAbs come with an heterogeneous mixture of species clinical development (4–6). With the enormous structural complexity compared to cysteine conjugation. top-selling mAbs evolving out of patent, (13–15). Opposed to small-molecule Peptide mapping is particularly there has been a growing interest in drugs, mAbs are large (ca. 150 kDa) powerful for the detailed structural the development of biosimilars (5,7,8). and heterogeneous (as a result of the characterization of these products In 2013, we witnessed the European biosynthetic process and subsequent and has proven to be of enormous approval of the first mAb biosimilars manufacturing and storage) making value in demonstrating comparability, and in 2016, the first mAb biosimilar their analysis very challenging. Despite for example, between mAb originator also reached marketing authorization the fact that only a single molecule and biosimilar (14,15). Characteristics in the US. Since then, a growing is cloned, hundreds of possible such as amino acid sequence number of mAb biosimilars have variants differing in post-translational and modifications like N- and reached approval in both Europe and modifications (N-glycosylation, O-glycosylation, glycation, N- and the US. The successes of mAbs have asparagine deamidation, aspartate C-terminal processing, deamidation furthermore triggered the development isomerization, methionine oxidation), (asparagine, glutamine), aspartate of various next-generation formats. In amino acid sequence, and higher isomerization, succinimide, oxidation oncology, antibody–drug conjugates order structures may coexist, all (methionine, tryptophan), clipping, (ADCs) are particularly promising, contributing to the safety and efficacy sequence variants, cysteine variants because they synergistically combine of the product. Compared to naked (S-S bridges, thioether, free cysteine), a specific mAb linked to a biologically mAbs, ADCs further add to the and drug conjugation sites can readily active cytotoxic drug via a stable complexity because the heterogeneity be extracted out of the generated

Photo Credit: kentoh/Shutterstock.com linker (9–12). The promise of ADCs is of the initial antibody is superimposed peptide map data and at low levels.

www.chromatographyonline.com 155 BIOPHARMACEUTICAL PERSPECTIVES

Figure 1: Micro-pillar array column. From left to right: (left) Top view of two parallel 315-μm wide separation channels that have been interconnected with proprietary flow distributor structures, (middle) SEM image showing a transverse section of a separation channel containing 5-μm diameter cylindrical pillars, (right) HR-SEM image of the 300-nm porous-shell layer incorporated into a 5-μm diameter pillar.

Figure 2: LC–MS chromatograms of a Herceptin tryptic digest. (a) Total compound peak collection (16). Today, columns chromatogram, (b) Compound chromatograms corresponding to the identified peptides packed with sub-2-μm porous presented in Table 1. and superficially porous particles operated at system pressures up to 1500 bar and electrospray ionization (ESI) can be used to introduce peptides into high resolution mass spectrometers equipped with a variety of fragmentation modes. Peptide mapping is particularly powerful for the detailed structural characterization of these products and has proven to be of enormous value in demonstrating comparability between mAb originator and biosimilar.

A more recent addition to the chromatographers toolbox are micro-pillar array columns. The origin of this technology dates back to the late 1990s when Regnier et al. addressed the problem of miniaturizing capillary electrochromatography (CEC) columns and introduced microfabricated supports as an alternative for the conventional packed beds (17,18). The theoretical benefit (reduction of the van Deemter A-term) of such supports was elucidated only a few years later by Knox (19). In the years to follow, Desmet et al. Looking back to the peptide mapping chromatography (HPLC) separations conducted several quantitative studies of the first monoclonal antibodies in were performed on columns packed on Knox’s argumentation, taking a the late 1980s and early 1990s, one with 5–10 μm porous particles and column filled with an array of pillars will notice a substantial leap forward in pumps were operated at 400 bar. as a representative example (20). technical capabilities. Chromatography Fast atom bombardment (FAB) was In 2007, the first micromachined LC and mass spectrometry were at used to introduce peptides into low columns operated by pressure-driven that time of modest performance resolution mass spectrometers and liquid flow, later termed micro-pillar compared to the current state-of-the-art peptide identity was further confirmed array columns, were reported (21). technology. High performance liquid using Edman degradation following The inherent high permeability and

156 LC•GC Europe March 2018 BIOPHARMACEUTICAL PERSPECTIVES

Figure 3: LC–UV 214 nm chromatograms corresponding to the replicate analysis (n = 4) of a Herceptin tryptic digest.

Expertise in all aspects of GC sample analysis

Intr od u c t i

o n n tio ra Figure 4: LC–MS total compound chromatograms of a Herceptin originator and a candidate biosimilar tryptic digest. p

S Det ec t i o

sis ly a n A

t a

D Ap pl ic a t i

n s low “on-column” dispersion obtained characterization of mAbs and ADCs by the perfect order of the separation for the first time. bed makes micro-pillar array-based chromatography unique and offers Materials and Methods several benefits for chromatographers. Materials: Water and acetonitrile Petrochemicals. Food. Fragrance The peak dispersion originating were purchased from Biosolve. allergens. Environmental forensics. from heterogeneous flow paths in Trifluoroacetic acid, dithiothreitol Our application range is extensive, the separation bed is eliminated (DTT), and 2-iodoacetamide (IAA) and so is our experience. (no A-term contributions) and were from Sigma-Aldrich. Tris-HCl pH Experience that we use to help our therefore components remain more 7.5 was purchased as a 1 M solution customers achieve better results, concentrated (sharp peaks) during (Thermo Fisher Scientific). Porcine separation. The freestanding nature sequencing-grade modified trypsin more quickly. Find out about us of the pillars also leads to much lower was acquired from Promega and and the range of manufacturers backpressure allowing the use of Rapigest from Waters. Herceptin and we represent... very long columns. These properties Kadcyla were obtained from Roche, result in excellent chromatographic Remicade from Johnson & Johnson, www.sepsolve.com performance with high resolution and Adcetris from Seattle Genetics, and high sensitivity. the candidate biosimilars from a local

This article describes the use of biotechnology company. A company of the SCHAUENBURG International Group micro-pillar array columns in the Sample Preparation: To a volume www.chromatographyonline.com 157 BIOPHARMACEUTICAL PERSPECTIVES

16 h at 37 °C using trypsin as protease Figure 5: Extracted ion chromatograms of selected modified peptides measured in Herceptin originator and candidate biosimilar. Shown are asparagine deamidation, lysine added at an enzyme to substrate truncation, and methionine oxidation, which are elevated in the biosimilar compared to ratio of 1/25 (w/w). Lyophilized trypsin the originator. The location of the modifications is highlighted in the antibody structure. (20 μg) dissolved in 100 mM Tris-HCl (50 μL) was added in a volume of 10 μL giving rise to a final sample volume of 210 μL. LC–MS: An Ultimate 3000 RSLCnano system (Thermo Fisher Scientific) was used for LC–ultraviolet (UV)– MS measurements. Tryptic digests were analyzed on a 200 cm C18 μPAC column (PharmaFluidics) at 50 °C. Elution was performed with a linear gradient of (A) 0.1% TFA in H2O and (B) 0.1% TFA in 80:20 (v/v) acetonitrile–water, from 2% B to 70% B in 120 min. The flow rate was 500 nL/ min. An injection program allowed the introduction of 100-nL sample in between two plugs of mobile phase A. Loop size was 1 μL and samples were kept at 10 °C in the autosampler tray while waiting for injection. UV detection was performed at 214 nm using a 3 nL detection cell. High-resolution accurate mass measurements were obtained on a 6530 QTOF mass spectrometer (Agilent Technologies) equipped with a dual nano-ESI source operated in positive electrospray ionization mode. The micro-pillar array column was connected via a 30-μm internal diameter (i.d.), 280-μm outside diameter (o.d.) fused silica capillary to a PicoTip emitter (8-μm tip i.d. [New Objective]) via a true zero dead volume conductive union (UH-634 stainless steel adapter from Idex). To connect the PicoTip emitter to the union, an F-330 blind fitting equipped with an M-125 perfluoroelastomer conductive ferrule was used (Idex). Capillary voltage was set at +2.4 kV, drying gas flow rate and temperature were set at 4 L/min and 300 °C, and fragmentor voltage at 175 V. The TOF was calibrated on a daily basis and subsequently operated at high accuracy (<5 ppm) without using reference masses. Data were collected in centroid mode at a rate of 1 spectrum/s in the extended dynamic range mode (2 GHz) offering a resolution of corresponding to 100 μg of protein, at 60 °C for 30 min by the addition 9000 at mass-to-charge ratio (m/z) 105 μL of 0.1% Rapigest in 100 mM of 5 mM DTT (2.5 μL of 400 mM DTT 622.0296. MS/MS experiments were Tris- HCl pH 7.5 was added followed by in 100 mM Tris-HCl) and alkylated performed in the data-dependent the addition of 100 mM Tris-HCl pH 7.5 at 37 °C for 1 h by adding 10 mM acquisition (DDA) mode. One survey to a final volume of 192.5 μL. The IAA (5 μL of 400 mM IAA in 100 mM MS measurement was complemented sample was subsequently reduced Tris-HCl). Digestion proceeded for with three data-dependent MS/MS

158 LC•GC Europe March 2018 BIOPHARMACEUTICAL PERSPECTIVES measurements. Double, triple, and Because of the high permeability, 750 nL/min, efficiencies up to 200,000 quadruple charged precursor ions the 200-cm column used in this theoretical plates could be achieved with the Averagine isotope model were study can be operated at moderate within 30 min at operating pressures selected based on abundance. After LC pump pressures (50 to 300 bar) well below 250 bar (mobile phase 50% being fragmented twice, a particular over a wide range of flow rates acetonitrile with 0.1% [v/v] FA) m/z value was excluded for 30 s. (100–1000 nL/min). Van Deemter Peptide Mapping of Herceptin Selecting the same m/z value twice measurements with heptyl-phenyl Originator and a Candidate increases the chance of measuring a ketone demonstrated that a total of Biosimilar: Herceptin (scientific particular precursor at its maximum 400,000 theoretical plates could be international nonproprietary names intensity while an exclusion time of generated at the optimal linear solvent [INN] name trastuzumab) is a 30 s allows MS/MS information on velocity, corresponding to a flow rate humanized IgG1 mAb recombinantly chromatographically resolved isomers of 200–250 nL/min and generating a produced in Chinese Hamster Ovary to be obtained. The quadrupole was column backpressure of only 70 bar. (CHO) cells and used in the treatment operated at medium resolution and the By increasing the flow rate up to of human epidermal growth factor collision energy varied based on the following equation:

(3.6*m/z)/100-4 [1]

All-ions MS/MS experiments were performed at a collision energy of 25 20 eV. LC data were acquired in years Chromeleon (Thermo Fisher Scientific) and MS data in MassHunter (Agilent Technologies). Data analysis was performed in MassHunter Qualitative Analysis with BioConfirm add-on.

Results and Discussion Characteristics of Micro-Pillar Array Columns: The separation beds of micro-pillar array columns are fabricated by carefully etching the interstitial volumes out of a silicon substrate following lithographic definition of an array of pillars. This creates a stationary phase support structure that is organized in a reproducible and ordered pattern. Concatenation of several of these channels allows long column lengths to be fabricated on a small footprint (22). The most important characteristics of the micro-pillar array separation bed design are: pillar diameter, 5 μm; inter pillar distance, Visit us in Munich 2.5 μm; pillar length or bed depth, Hall A1, Booth 405 18 μm; external porosity (Vinterstitial/ Vtotal), 59%; bed channel width, 315 μm; and bed length, 200 cm. To increase the retentive surface, the pillars are rendered superficially Reproducibility... YMC porous with a typical porous shell thickness of 300 nm and pore sizes in the nanometre range. The porous Robustness Scalability Selectivity surface has been uniformly modified · pH · (U)HPLCļHPLCļPREP · RP, NP, HILIC with octadecyl alkyl chains to create a · temperature · Chiral, SFC hydrophobic stationary phase suited · easy method transfer for reversed-phase LC separations. · 100% aqueous eluents · IEX, SEC Figure 1 shows some relevant characteristics of the micro-pillar array column. www.chromatographyonline.com Discover more at www.ymc.de BIOPHARMACEUTICAL PERSPECTIVES

Figure 6: LC–UV 214 nm peptide map of (a) a Remicade originator and (b) a candidate companies are actively developing biosimilar tryptic digest. Unzoomed and zoomed views. a Herceptin biosimilar. In developing these products, similarity to the originator has to be demonstrated and for that, an enormous weight is placed on analytics; both the biosimilar and originator need to be characterized and compared in great detail. Peptide mapping is a powerful method to characterize and compare mAbs in great detail. When digesting Herceptin with the enzyme trypsin, which cleaves the protein next to arginine and lysine residues, 62 identity peptides are formed. Taking into account post-translational modifications and incomplete and aspecific cleavages taking place, over 100 peptides with varying physicochemical properties present in a wide dynamic concentration range are to be expected. This is quite a complex sample demanding the best in terms of chromatographic resolution. Figure 2(a) shows the LC–MS total compound chromatogram of a Herceptin tryptic digest on a micro-pillar array column. The compound chromatograms corresponding to the identified peptides presented in Table 1 are shown in Figure 2(b). Approximately 95% of the Figure 7: LC–MS total compound chromatograms of (a) a Remicade originator and (b) sequence is covered by this a candidate biosimilar tryptic digest. MS/MS spectra associated with peaks 4 and 5 are peptide map and post-translational presented in Figure 8. modifications such as glycosylation, asparagine deamidation, aspartate isomerization, methionine oxidation, N-terminal cyclization (pyroglutamate), and C-terminal lysine truncation, amongst others, are revealed. Peptides that are not detected are typically small and their signal might be suppressed in the column flow through. The UV 214 nm chromatograms corresponding to replicate analyses (n = 4) of the Herceptin tryptic digest are demonstrated in Figure 3. These measurements are highly precise making this technology attractive to compare different mAb production batches and to compare mAb originator products to biosimilars. The LC–MS peptide maps of a Herceptin originator and candidate biosimilar are shown in Figure 4. While both peptide maps are receptor 2 (HER2) positive breast patent in the US in 2019 (5). Given highly comparable, differences in cancer. Herceptin is open to the its market potential (global sales post-translational modifications are European market and evolves out of of $6.6 billion in 2015), dozens of detected. This is illustrated in

160 LC•GC Europe March 2018 Figure 8: LC–MS/MS spectra of peak 4 and 5 (Figure 7) acquired on the fly in data-dependent MS/MS mode confirming the threonine to serine substitution in the variable part of the heavy chain.

the extracted ion chromatograms presented in Figure 5. The resolving power offered by a micro-pillar array column allows an +DOO$6WDQG in-depth study of ADC conjugation sites www.analytica.de/en and accommodates the separation of isomeric conjugated peptides. •

Peptide Mapping of Remicade Originator and Candidate Biosimilar: Remicade (scientific INN name infliximab) is a chimeric IgG1 mAb on the market since • 1998 targeting tumour necrosis factor alpha (TNF-α) and consequently used in the treatment of autoimmune diseases. Remicade reached global sales of $8.4 billion in 2015. Several Remicade biosimilars have already • been approved both in Europe and the US and many more are in development (8). Figures 6 and 7 show the UV and MS total compound chromatograms of a Remicade originator and a candidate biosimilar tryptic digest, respectively. • Chromatograms are very similar but some striking differences are noted (peaks 1–5), which can be explained upon consulting the corresponding MS data. Peaks 1 and 2 in the originator chromatogram, corresponding to C-terminal heavy chain peptides SLSLSPGK and SLSLSPG respectively, are replaced by peak 3, corresponding to SLSLSPGI in the biosimilar. The origin of the two peaks SLSLSPG and SLSLSPGK in the originator mAb can be explained by the knowledge that heavy chains are historically cloned with a C-terminal lysine but during cell culture production, www.chromatographyonline.com BIOPHARMACEUTICAL PERSPECTIVES

Table 1: Identified peptides in the Herceptin tryptic digest analyzed by LC–MS RT (min) Mass Sequence Seq Loc. Theor. Mass ∆m (ppm) Modifications 61.29 1880.99676 EVQLVESGGGLVQPGGSLR Hc(001-019) 1880.9956 0.62 66.83 1862.98605 EVQLVESGGGLVQPGGSLR Hc(001-019) 1862.98503 0.55 Pyroglutamate 51.04 1166.57794 LSCAASGFNIK Hc(020-030) 1166.5754 2.18 Cys-Alkylation 71.21 2237.10621 LSCAASGFNIKDTYIHWVR Hc(020-038) 2237.10516 0.47 Cys-Alkylation 55.35 1088.54237 DTYIHWVR Hc(031-038) 1088.54033 1.87 52.54 829.44557 GLEWVAR Hc(044-050) 829.44464 1.12 39.97 1084.52052 IYPTNGYTR Hc(051-059) 1084.51893 1.47 Deamidation 40.39 1083.53637 IYPTNGYTR Hc(051-059) 1083.53491 1.35 41.46 1084.52241 IYPTNGYTR Hc(051-059) 1084.51893 3.21 Deamidation 28.81 681.33403 YADSVK Hc(060-065) 681.33336 0.99 44.14 968.48309 FTISADTSK Hc(068-076) 968.48148 1.66 43.48 1366.66793 NTAYLQMNSLR Hc(077-087) 1366.66634 1.17 Met-Alkylation 53.78 1310.63608 NTAYLQMNSLR Hc(077-087) 1310.62889 5.49 Deamidation 54.96 1309.6462 NTAYLQMNSLR Hc(077-087) 1309.64487 1.02 40.75 1333.56246 AEDTAVYYCSR Hc(088-098) 1333.56087 1.19 Cys-Alkylation WGGDGFYAMDYWGQ 70.51 2840.27436 Hc(099-124) 2840.2752 -0.3 Met-Alkylation GTLVTVSSASTK WGGDGFYAMDYWGQ 79.37 2783.25248 Hc(099-124) 2783.25374 -0.45 GTLVTVSSASTK 58.45 1185.64103 GPSVFPLAPSSK Hc(125-136) 1185.63938 1.4 54.84 1320.67286 STSGGTAALGCLVK Hc(137-150) 1320.67075 1.59 Cys-Alkylation DYFPEPVTVSWNSGALTSG 92.81 6712.27146 VHTFPAVLQSSGLYSLSSVVTVP Hc(151-213) 6712.3072 -5.32 Cys-Alkylation SSSLGTQTYICNVNHKPSNTK 24.21 599.36222 KVEPK Hc(217-221) 599.36426 -3.41 23.25 471.26917 VEPK Hc(218-221) 471.2693 -0.27 21.56 508.19936 SCDK Hc(222-225) 508.19515 8.28 Cys-Alkylation 82.24 2618.30392 THTCPPCPAPELLGGPSVFLFPPK Hc(226-249) 2618.30254 0.53 Cys-Alkylation 80.29 2843.4514 THTCPPCPAPELLGGPSVFLFPPKPK Hc(226-251) 2843.45027 0.4 Cys-Alkylation 41.18 891.44858 DTLMISR Hc(252-258) 891.44841 0.19 Met-Alkylation 41.64 850.42241 DTLMISR Hc(252-258) 850.42186 0.65 Oxidation 45.43 834.42764 DTLMISR Hc(252-258) 834.42694 0.84 64.07 2138.02125 TPEVTCVVVDVSHEDPEVK Hc(259-277) 2138.02016 0.51 Cys-Alkylation

61.55 1676.79346 FNWYVDGVEVHNAK Hc(278-291) 1676.79471 -0.75 Isomerization 62.47 1676.7955 FNWYVDGVEVHNAK Hc(278-291) 1676.79471 0.47

32.86 2957.13923 EEQYNSTYR Hc(296-304) 2957.14427 -1.7 G2F 32.92 2795.08868 EEQYNSTYR Hc(296-304) 2795.09144 -0.99 G1F 33.05 2633.03635 EEQYNSTYR Hc(296-304) 2633.03862 -0.86 G0F

81.52 1807.98677 VVSVLTVLHQDWLNGK Hc(305-320) 1807.98324 1.95 Deamidation 82.11 1807.0017 VVSVLTVLHQDWLNGK Hc(305-320) 1806.99922 1.37

162 LC•GC Europe March 2018 BIOPHARMACEUTICAL PERSPECTIVES

Table 1: (Continued) Identified peptides in the Herceptin tryptic digest analyzed by LC–MS

82.92 1807.98553 VVSVLTVLHQDWLNGK Hc(305-320) 1807.98324 1.27 Deamidation 84.33 1789.97519 VVSVLTVLHQDWLNGK Hc(305-320) 1789.97267 1.4 Succinimide 46.50 837.49753 ALPAPIEK Hc(330-337) 837.49601 1.82 23.71 447.26932 TISK Hc(338-341) 447.2693 0.04 50.08 1285.66847 EPQVYTLPPSR Hc(348-358) 1285.66665 1.41 24.31 636.2778 EEMTK Hc(359-363) 636.27888 -1.69 58.87 1160.62392 NQVSLTCLVK Hc(364-373) 1160.62235 1.36 Cys-Alkylation 68.04 2544.11037 GFYPSDIAVEWESNGQPENNYK Hc(374-395) 2544.10812 0.88 Deamidation 68.51 2543.12268 GFYPSDIAVEWESNGQPENNYK Hc(374-395) 2543.12411 -0.56 69.70 2526.09392 GFYPSDIAVEWESNGQPENNYK Hc(374-395) 2526.09756 -1.44 Succinimide 70.26 1872.90824 TTPPVLDSDGSFFLYSK Hc(396-412) 1872.91455 -3.37 Isomerization 71.48 1872.91517 TTPPVLDSDGSFFLYSK Hc(396-412) 1872.91455 0.33 32.14 574.33213 LTVDK Hc(413-417) 574.33263 - 0.86 66.07 2800.25791 WQQGNVFSCSVMHEALHNHYTQK Hc(420-442) 2800.25984 -0.69 Cys-Alkylation 44.81 659.34888 SLSLSPG Hc(443-449) 659.34901 -0.19 Lys-Truncation 43.20 787.44611 SLSLSPGK Hc(443-450) 787.44397 2.72 46.55 1934.90169 DIQMTQSPSSLSASVGDR Lc(001-018) 1934.90038 0.68 Met-Alkylation 54.09 1877.87963 DIQMTQSPSSLSASVGDR Lc(001-018) 1877.87891 0.38 35.39 748.39074 VTITCR Lc(019-024) 748.39016 0.78 Cys-Alkylation 53.62 1707.82255 ASQDVNTAVAWYQQK Lc(025-039) 1707.82165 0.53 54.73 1708.80311 ASQDVNTAVAWYQQK Lc(025-039) 1708.80567 -1.5 Deamidation 51.50 1990.97492 ASQDVNTAVAWYQQKPGK Lc(025-042) 1990.97486 0.03 Deamidation 53.37 1989.99336 ASQDVNTAVAWYQQKPGK Lc(025-042) 1989.99084 1.27 54.43 1990.97614 ASQDVNTAVAWYQQKPGK Lc(025-042) 1990.97486 0.64 Deamidation ASQDVNTAVAWYQQKPG 83.35 4040.11375 Lc(025-061) 4040.116 -0.55 KAPKLLIYSASFLYSGVPSR 76.25 1771.95123 LLIYSASFLYSGVPSR Lc(046-061) 1771.95088 0.2 SGTDFTLTISSLQPEDFAT 81.29 4186.90432 Lc(067-103) 4186.91062 -1.5 Cys-Alkylation YYCQQHYTTPPTFGQGTK 30.68 487.3006 VEIK Lc(104-107) 487.3006 -0.01 72.96 2101.12277 RTVAAPSVFIFPPSDEQLK Lc(108-126) 2101.1208 0.94 74.60 1945.02035 TVAAPSVFIFPPSDEQLK Lc(109-126) 1945.01968 0.34 80.21 1796.88935 SGTASVVCLLNNFYPR Lc(127-142) 1796.88796 0.77 Cys-Alkylation 21.45 346.18747 EAK Lc(143-145) 346.18524 6.46 34.98 559.31157 VQWK Lc(146-149) 559.31183 -0.48 39.99 2134.96295 VDNALQSGNSQESVTEQDSK Lc(150-169) 2134.96146 0.7 59.63 1501.75181 DSTYSLSSTLTLSK Lc(170-183) 1501.75118 0.42 24.33 624.27451 ADYEK Lc(184-188) 624.27551 -1.6 50.86 1874.92219 VYACEVTHQGLSSPVTK Lc(191-207) 1874.91965 1.35 Cys-Alkylation 26.74 522.25451 SFNR Lc(208-211) 522.25505 -1.03 www.chromatographyonline.com 163 BIOPHARMACEUTICAL PERSPECTIVES

Figure 9: LC–UV 214 nm peptide map of Herceptin and Kadcyla. host cell carboxypeptidases act on the antibody resulting in the partial removal of these lysine residues. A missense mutation (LysIle) in the CHO clone producing the candidate biosimilar explains peak 3 (SLSLSPGI). Peaks 4 and 5 corresponding to, respectively, heavy chain peptides NYYGSSYDYWGQGTTLTVSSASTK in the candidate biosimilar and NYYGSTYDYWGQGTTLTVSSASTK in the originator again result from a point mutation (ThrSer) in the biosimilar CHO clone. The corresponding MS/MS spectra are shown in Figure 8. According to US and European regulatory authorities, identical primary sequence is primordial to similarity Figure 10: LC–MS peptide map of Herceptin and Kadcyla. thereby ruling out this candidate biosimilar from further development. Peptide Mapping of the Antibody– Drug Conjugate Kadcyla: The antibody–drug conjugate Kadcyla (ado-trastuzumab emtansine) has been used in the treatment of HER2 positive breast cancer since 2013. It combines the anti-HER2 antibody trastuzumab (Herceptin) with the cytotoxic microtubule- inhibiting maytansine derivative DM1 conjugated to lysine residues via a non-reducible thioether linker. With a drug distribution of 0 to 8, a drug-to-antibody ratio (DAR) of 3.5, and various lysine residues available for conjugation, thousands of species can be generated. To obtain insight in the drug conjugation sites, peptide mapping is the gold standard Figure 11: All-ions MS/MS chromatograms of the Herceptin and Kadcyla tryptic digest. technology. The ion at m/z 547.2206 was extracted at high mass accuracy. Figures 9 and 10 show, respectively, the LC–UV and MS total compound chromatograms of a Herceptin and Kadcyla tryptic digest. Since Herceptin and Kadcyla have the same amino acid sequence, the majority of the peptide map is identical. Differentiating peaks, corresponding to the DM1 conjugated peptides, are nevertheless observed and are located in the late eluting part of the chromatogram. Indeed, the conjugation of DM1 makes the peptides more hydrophobic explaining the elution behaviour. Upon collision-induced dissociation (CID), DM1 conjugated peptides give rise to specific fragments originating from the cytotoxic drug, for example, at m/z 547.2206. This ion can be used to selectively recognize DM1 conjugated peptides in the LC–MS

164 LC•GC Europe March 2018 BIOPHARMACEUTICAL PERSPECTIVES chromatogram. For that, one can (peak 3) are observed. Of best known are columns packed operate the QTOF-MS system in the particular interest is the with sub-2-μm porous particles or all-ions MS/MS mode, which means detection of isomeric partially sub-3-μm superficially particles. that the quadrupole is operated in the conjugated heavy chain peptide Micro-pillar array columns are another RF only mode, thereby transferring THTCPPCPAPELLGGPSVFLFPPKPK novel development offering highly all peptides to the collision cell where (peaks a and b), in which only one efficient separations. In this article, the CID takes place. As illustrated in of the two cysteine residues is use of the micro-pillar array column in Figure 11, when extracting from the conjugated. mAb and ADC peptide mapping has data the specific fragment ion at m/z been illustrated. In combination with 547.2206 at high mass accuracy, all Conclusion high resolution mass spectrometry, conjugated peptides are revealed In the 21st century, numerous high sequence coverage is obtained in the chromatogram of Kadcyla novel developments were made and post-translational modifications compared to Herceptin. The latter in LC column technology and the such as glycosylation, deamidation, chromatogram is virtually empty, illustrating the selectivity that is offered by this all-ions MS/MS functionality. Figure 12 shows the extracted ion chromatograms associated with a selection of identified conjugated peptides. A striking observation is the appearance of isomeric DM1 conjugated peptides, which can be explained by the existence of two stereochemical configurations (diastereomers) of the antibody–drug linkage through maleimide. Peptide Mapping of the Antibody– Drug Conjugate Adcetris: A similar strategy was applied to reveal the conjugation sites on Adcetris (brentuximab-vedotin). Approved in 2011, this ADC is directed to CD30, a major marker of Hodgkin lymphoma and systemic anaplastic large cell lymphoma (ALCL). Adcetris combines the antibody brentuximab with the antimitotic drug monomethylauristatin E (MMAE) conjugated to interchain cysteine residues via a cathepsin cleavable valine-citrulline linker. With only eight residues (four on each half antibody) available for conjugation, Adcetris is much simpler compared to Kadcyla. Figure 13 shows the LC–MS total compound chromatogram and the all-ions MS/MS chromatogram of an Adcetris tryptic digest, respectively. In analogy to DM1 conjugated peptides, MMAE conjugated peptides also contain specific fragment ions originating from the cytotoxic molecule, that is, at m/z 718.5113. When extracting the latter ion from the all-ions MS/MS data, the conjugated peptides are revealed. Three intense peaks explaining the full conjugation of MMAE at the four interchain cysteine residues in light chain peptide GEC (peak 2), heavy chain peptides SCDK (peak 1), and THTCPPCPAPELLGGPSVFLFPPKPK www.chromatographyonline.com 165 BIOPHARMACEUTICAL PERSPECTIVES

Biopharmaceutical Analysis” supplement Figure 12: Extracted ion chromatograms of selected peptides showing the appearance to LCGC Europe 28(s10), 16–23 (2015). of isomeric conjugated peptides. The location of the conjugation site in the mAb (8) A. Beck and J.M. Reichert, mAbs 5, structure is shown as well. 621–623 (2013). (9) S. Panowski, S. Bhakta, H. Raab, P. Polakis, and J.R. Junutula, mAbs 6, 34–45 (2014). (10) A. Wakankar, Y. Chen, Y. Gokarn, and F.S. Jacobson, mAbs 3, 161–172 (2011). (11) A. Beck and J.M. Reichert, mAbs 6, 15–17 (2014). (12) A. Beck, L. Goetsch, C. Dumontet, and N. Corvaia, Nat. Rev. Drug Discov. 16, 315–337 (2017). (13) A. Beck, E. Wagner-Rousset, D. Ayoub, A. Van Dorsselaer, and S. Sanglier- Cianférani, Anal. Chem. 85, 715–736 (2013). (14) K. Sandra, I. Vandenheede, and P. Sandra, J. Chromatogr. A 1335, 81–103 (2014). (15) S. Fekete, D. Guillarme, P. Sandra, and K. Sandra, Anal. Chem. 88, 480–507 (2016). (16) D.J. Kroon, A. Baldwin-Ferro, and P. Lalan, Pharm. Res. 9, 1386–1393 (1992). (17) B. He, N. Tait, and F. Regnier, Anal. Chem. 70, 3790–3797 (1998). (18) F. Regnier, J. High Resol. Chromatogr. 23, 19–26 (2000). (19) J.H. Knox, J. Chromatogr. A 960, 7–18 (2002). (20) P. Gzil, N. Vervoort, G. Baron, and G. Desmet, Anal. Chem. 75, 6244–6250 (2003). LC–MS peptide map of Adcetris. (a) Total compound chromatogram and Figure 13: (21) W. De Malsche, H. Gardeniers, and G. (b) all-ions MS/MS chromatogram. The ion at m/z 718.5113 was extracted at high mass Desmet, Anal. Chem. 80, 5391–5400 accuracy. The location of the conjugation site in the mAb structure is shown as well. (2008). (22) W. De Malsche, J. Op de Beeck, S. De Bruyne, H. Gardeniers, and G. Desmet, Anal. Chem. 84, 1214–1219 (2012).

Koen Sandra is the editor of “Biopharmaceutical Perspectives”. He is the Scientific Director at the Research Institute for Chromatography (RIC, Kortrijk, Belgium) and R&D Director at anaRIC biologics (Ghent, Belgium). Jonathan Vandenbussche is LC–MS Technician at RIC. Isabel Vandenheede is Project Manager Biologics at RIC. Bo Claerebout is R&D Engineer at PharmaFluidics, Ghent, Belgium. Jeff Op de Beeck is Application Development Manager at PharmaFluidics. isomerization, oxidation, N-terminal References Paul Jacobs is Chief Operating cyclization, and C-terminal lysine (1) G. Köhler and C. Milstein, Nature 256, Officer (COO) at and co-founder of 495–497 (1975). truncation can be elucidated. Its use (2) N.A.P.S. Buss, S.J. Henderson, M. PharmaFluidics. in assessing comparability between McFarlane, J.M. Shenton, and L. de Wim De Malsche is Associate an originator and biosimilar mAb Haan, Curr. Opin. Pharmacol. 8, Professor at the Department of 620–626 (2012). has furthermore been demonstrated. (3) J.G. Elvin, R.G. Couston, and C.F. van Chemical Engineering at the Vrije The resolving power offered by a der Walle, Int. J. Pharm. 440, 83–98 Universiteit Brussel, Belgium and micro-pillar array column allows an (2013). co-founder of PharmaFluidics. in-depth study of ADC conjugation sites (4) D.M. Ecker, S.D. Jones, and H.L. Levine, Gert Desmet is a Full Professor in mAbs 7, 9–14 (2015). and accommodates the separation (5) G. Walsh, Nat. Biotechnol. 32, 992–1000 Chemical Engineering at the Vrije of isomeric conjugated peptides. In (2014). Universiteit Brussel, Belgium and a combination with all-ions MS/MS, (6) H. Kaplon and J.M. Reichert, mAbs 10, co-founder of PharmaFluidics. 183–203 (2018). conjugated peptides can selectively be (7) K. Sandra, I. Vandenheede, E. Pat Sandra is President of RIC and recognized to assist data interpretation. Dumont, and P. Sandra, “Advances in Emeritus Professor at Ghent University.

166 LC•GC Europe March 2018 PRODUCTS MALS detector HPLC kits The μDAWN is, Sciencix’s HPLC performance according to the maintenance (PM) kits are company, the world’s designed to be equivalent to first multi-angle light the corresponding OEM kits. scattering (MALS) According to the company, the detector that can be kits contain essential parts to coupled to any UHPLC keep HPLC systems performing system to determine at an optimal level. absolute molecular weights and sizes of polymers, peptides, www.sciencix.com and proteins or other biopolymers directly, without resorting Sciencix, Burnsville, Minnesota, USA. to column calibration or reference standards. The WyattQELS Dynamic Light Scattering (DLS) module, which measures hydrodynamic radii “on-the-fly”, reportedly expands the versatility of the μDAWN. www.wyatt.com Wyatt Technology, Santa Barbara, California, USA.

Preparative system Extraction technique AECS-QuikPrep Ltd, UK has announced an upgraded range of Isolute Hydro DME+ provides Quattro CCC and CPC, plus Partitron extremely efficient removal CPC. According the the company, of matrix components and the scope of applications ranges from hydrolysis enzymes from milligrams to recently released Partitron urine samples, according to CPC of 40 litres, capable of tonnes the company, using a simple pass through workflow. per annum preparations. AECS’s CPC Dual-mode extraction (DME) provides effective removal of can be utilized at process scales in matrix components using a combination of liquid partitioning chromatographic extraction, plus for 2018, “on-column” synthesis and scavenging modes. This novel extraction technology reactor and extractor operations. AECS reports that unfavoured removes urinary components such as pigments, salts, urea, synthetic reactions can be driven to very high efficiencies, 97%+, creatinine, and residual hydrolysis enzyme. when using on-column synthesis, allowing unique ionic liquids and http://www.biotage.com/product-page/isolute-hydro- other compounds to be economically obtained. dme-dual-mode-extraction-products www.quattroprep.com Biotage AB, Uppsala, Sweden. AECS-QuikPrep LTD, Cornwall, UK.

GC column oven GPC/SEC/IPC software The fluidless column oven PSS has launched WinGPC (FCO) can easily be installed UniChrom 8.3 SR1, GPC/SEC/ in any GC or GC–MS system IPC software with full 21CFR11 to speed analyses up to ten compliance also for (multi-angle) times faster than conventional light scattering or heparin ovens, according to the company. The FCO works by placing analysis. The integrated column a steady-state gradient temperature profile on the analytical database tracks injections and column, eliminating normal temperature cycling, which allows monitors column performance. one sample to be analyzed directly after another. All results are reported with measurement uncertainty. www.gcovens.com PSS WinGPC UniChrom can be used with all LC hardware GC Ovens Inc., Nevada, USA. and can control more than 200 modules from different vendors. www.pss-polymer.com PSS Polymer Standards Service GmbH, Mainz, Germany.

www.chromatographyonline.com 167 PRODUCTS

Chiral columns Reference materials YMC has launched a 3-μm version of Spex CertiPrep has introduced the recently introduced immobilized a new pesticide mix to address polysaccharide Chiral Art Cellulose-SJ the European Commission’s column which is available in a 5 μm Regulation 2017/170. The particle size format. Both use an Commission is amending immobilized chiral selector. Chiral Art Annexes II, III, and V to Cellulose-SJ columns are very robust Regulation (EC) No 396/2005 and can be used in reversed-phase, of the European Parliament normal-phase, and SFC mode with a and of the Council as regards to maximum residue levels wide range of solvents, according to the company. for bifenthrin, carbetamide, cinidon-ethyl, fenpropimorph, https://ymc.de/chiral-columns.html and triflusulfuron in or on certain products. YMC Europe GmbH, Dinslaken, Germany. www.spexeurope.com Spex Europe, Dalston Gardens, Stanmore, UK.

HILIC phases GC instruments ACE HILIC Method Development Kits group together the acidic character The Pegasus BT 4D offers ACE HILIC-A phase, the basic enhanced sensitivity by coupling character ACE HILIC-B phase, and the benchtop Pegasus BT with the neutral character ACE HILIC-N the high performance GC×GC phase—each offering alternative thermal modulation system. This selectivity—to provide users with a combination gives the Pegasus powerful screening platform for the BT 4D the ability to interrogate development of a successful and challenging samples where robust HILIC separation, according to the company. They are the best sensitivity is needed, according to the company. available in 1.7 μm, 3 μm, and 5 μm and a variety of column Software and hardware features simplify quantitation, dimensions. while also making GC×GC easier to use and understand, www.ace-hplc.com according to the company. Advanced Chromatography Technologies Limited, ww.leco.com Scotland, UK. Leco Corporation, Berlin, Germany.

FFF system HPLC columns Postnova Analytics has The core–shell product range announced the launch of the of Macherey-Nagel has been EAF2000 - a new simultaneous extended by Nucleoshell Electrical and Asymmetrical Bluebird RP 18. As a result of Flow Field Flow Fractionation polar endcapping and (EAF4) system designed to core–shell technology, the enhance separation and characterization of biopharmaceuticals, HPLC columns Bluebird RP environmental, and nanomaterials. 18 can be used for fast analyses under highly aqueous http://www.postnova.com/overview_759.html conditions, especially for very polar analytes, such as Postnova Analytics GmbH, Landsberg, Germany. organic acids, water -soluble vitamins, pesticides, and antibiotics. www.mn-net.com Macherey-Nagel GmbH & Co. KG, Düren, Germany.

168 LC•GC Europe March 2018 THE APPLICATIONS BOOK

March 2018 www.chromatographyonline.com CONTENTS THE APPLICATIONS BOOK

Food and Beverage

171 The Grass Isn’t Always Greener: Removal of Purple Pigmentation Photo Credit: victoriaKh/Shuttestock.com Credit: Photo from Cannabis Using QuEChERS Extraction and Chlorofiltr® dSPE Cleanup Danielle Mackowsky, UCT, LLC

172 E ffects of Irradiation on Sodium Alginate Wyatt Technology

Medical/Biological

173 Kinase Fragments Dimerize Without Oligomerization Domains

Photo Credit: Fotokostic/Shutterstock.com Thomas Huber, University of Zurich, Department of Biochemistry

Pharmaceutical/Drug Discovery

174 Extended Dynamic Range ELSD for Impurity Profiling and Purification SEDERE SAS

176 Analysis of Residual Solvents in Drug Products Using Nexis GC-2030 Combined With HS-20 Headspace Sampler—USP <467> Residual Solvents Procedure A

Photo Credit: kentoh/Shutterstock.com Shimadzu Europa GmbH

Cover Photography: Shutterstock.com Photo Credit: Komar art/Shutterstock.com Photo Credit: Jezper/Shutterstock.com

170 THE APPLICATIONS BOOK – MARCH 2018 FOOD AND BEVERAGE

The Grass Isn’t Always Greener: Removal of Purple Pigmentation from Cannabis Using QuEChERS Extraction and Chlorofi ltr® dSPE Cleanup Danielle Mackowsky, UCT, LLC

Cannabis testing laboratories have the challenge of removing a variety Results of unwanted matrix components from plant material prior to running extracts on their LC–MS/MS or GC–MS systems. The complexity of the cannabis plant presents additional analytical challenges that do not need to be accounted for in other agricultural products. While novel methods have been developed for the removal of chlorophyll, few sample preparation methods, if any, have been devoted to removal of other coloured pigments from other popular cannabis strains.

Table 1: Extraction and analytical materials 50-mL centrifuge tubes and mylar pouches ECMSSC50CT-MP containing 4 g MgSO4 and 1 g NaCl SpinFiltr™ dSPE cleanup tube 150 mg Figure 1: Cannabis strains used (Clockwise from top left): Agent ECQUSF54CT MgSO4, 50 mg PSA, 50 mg C18 and 50 mg Orange, Tahoe OG, Blue Skunk, Grand Daddy, and Grape Drink. ChloroFiltr® Selectra® Aqueous C18 HPLC column SLAQC18100ID21-3UM 100 × 2.1 mm, 3-μm Selectra® Aqueous C18 guard column SLAQC18GDC20-3UM 10 × 2.1 mm, 3-μm

Procedure Sample Extraction + a) Add 1 g of cannabis 10 mL H2O. b) Vortex briefl y and hydrate for 10 min. c) Add 10 mL acetonitrile + 2% formic acid + the QuEChERS extraction salts from the Mylar pouch (ECMSSC50CT-MP) to the 50-mL tube, and shake vigorously for 1 min manually or using a Figure 2: Cannabis samples following hydration. Left: Grape Drink strain, right: Agent Orange strain. Spex 2010 Geno-Grinder at 1000 strokes/min. d) Centrifuge at ≥3000 rcf for 5 min. Conclusion ® e) Transfer 1 mL supernatant to SpinFiltr™ dSPE cleanup tube. A blend of MgSO4, C18, PSA, and Chlorofi ltr allowed for the most f) Vortex for 30 s and centrifuge at ≥3000 rcf for 2 min. effective sample cleanup, without loss of pesticides and mycotoxins, g) Transfer purifi ed and fi ltered extract into autosampler vial for for all cannabis samples tested. Average recovery of the 48 analysis. pesticides and four mycotoxins using the selected dSPE blend was 75.6%, while the average recovery when including GCB instead Instrumental of Chlorofi ltr® was 67.6%. Regardless of the sample’s original LC–MS/MS: Thermo Scientifi c™ Dionex™ Ultimate™ 3000 UHPLC pigmentation, this blend successfully removed both and TSQ Vantage™ (MS/MS) chlorophyll and purple hues from all strains tested. Column: 100 × 2.1 mm, 3-μm UCT Selectra® Aqueous C18 HPLC The other six dSPE blends evaluated were unable to Guard Column: 10 × 2.1 mm, 3-μm UCT Selectra® Aqueous C18 provide the sample cleanup needed or had previously Injection Volume: 5 μL demonstrated to be detrimental to the recovery of Mobile Phase A: + + D.I. H2O 5 mM NH4HCO2 0.1% formic acid pesticides routinely analyzed for in cannabis. Mobile Phase B: + + Methanol 5 mM NH4HCO2 0.1% formic acid Column Flow Rate: 0.30 mL/min UCT, LLC 2731 Bartram Road, Bristol, Pennsylvania 19007, USA Tel: (800) 385 3153 E-mail: [email protected] Website: www.unitedchem.com

THE APPLICATIONS BOOK – MARCH 2018 171 FOOD AND BEVERAGE

Effects of Irradiation on Sodium Alginate Wyatt Technology

Sodium alginate is a food thickening agent that may be irradiated for sterilization. Changes in biopolymer molar mass and conformation are analyzed by size-exclusion chromatography multi-angle light scattering (SEC-MALS).

Introduction Sodium alginates are widely used as a food additive or as a sterile wound dressing. The molecular weight and conformation properties of this polysaccharide contribute directly to their end-use performance. When used as a thickening agent, the higher the molecular weight, the better the gel properties will be. There has been considerable interest in recent years in the development of suitable test methods to characterize foods that have been irradiated, as part of a continuing effort to prolong shelf life. This note describes work undertaken on a grade of sodium alginate used as a food thickening agent, employing a DAWN multi-angle light scattering detector and Optilab differential refractometer (both from Figure 1: Inset: Differential molecular weight plots for sodium Wyatt Technology) in conjunction with SEC, to determine the effect of alginate before and after irradiation. Main: RMS Radius versus MW gamma irradiation on the biopolymer. plot showing the change in conformation that occurs after irradiation.

Experimental Conditions A 7.6 mm × 300 mm HEMA Bio Linear column (PSS) was installed The radiation dose given to the sodium alginate caused the molecule in a Waters 600 chromatograph (Waters Corp.). The fl ow rate was to change from a fairly dense, sphere-like, cross-linked structure to 1.0 mL/min of pure HPLC-grade water. The alginate’s specifi c a rod-like conformation. This suggests that the cross-linking bonds refractive index increment, dn/dc, was measured off-line in an Optilab, are damaged or destroyed by the irradiation, causing the molecule which operates at the same wavelength as the DAWN. Absolute molar to extend or expand, and hence modifying its performance as a food mass (MW) and size (rms radius, Rg) at each elution volume were thickener or wound dressing. determined by analysis of the light scattering and refractive index signals in ASTRA software (Wyatt Technology), which converted the Conclusions data to differential MW distributions and conformation plots. SEC-MALS provides detailed information on the molecular weight and conformational changes induced in sodium alginate and similar Results polysaccharides by irradiation. The molecular-level changes may The differential molecular weight distributions presented in the then be correlated to macroscopic changes in end-use performance inset of Figure 1 clearly indicate the degradation in molecular for a deeper understanding of this phenomenon and the viability of weight (MW) that occurs after irradiation, with peak MW dropping such treatment. by up to 70%. Of much greater interest, however, was the change in molecular conformation, shown in the main fi gure. Multi-angle light scattering shows these changes by determining both the size and MW, independently, at each elution volume and plotting the log (radius) as a function of the log (MW). The resulting slopes reveal whether the molecule is a sphere (slope of about 0.33), random coil (slope of 0.5–0.6), or a rod (slope of 1.0). The conformation plots prove that the natural, un-irradiated alginate Wyatt Technology Corp. has a compact, sphere-like structure. In the irradiated sample, by 6330 Hollister Avenue, Santa Barbara, California 93117, USA contrast, their slope of about 0.88 indicates that the molecules Tel.: +1 (805) 681 9009 • Fax: +1(805)681 0123 apparently have “opened” to a nearly rod-like structure. Website: www.wyatt.com • E-mail: [email protected]

172 THE APPLICATIONS BOOK – MARCH 2018 MEDICAL/BIOLOGICAL

Kinase Fragments Dimerize Without Oligomerization Domains Thomas Huber, University of Zurich, Department of Biochemistry

Self-association is crucial for regulation of certain kinase proteins. However, a kinase fragment that lacks the oligomerization domain is still dimeric in solution, as determined by size-exclusion chromatography multi-angle light scattering (SEC-MALS).

Introduction Determination of oligomeric states is an important issue in protein chemistry. For example, self-assembly via oligomerization domains is crucial for the regulation of several protein kinases. Determination of the oligomeric state of fragments of these kinases is a means of verifying the involvement of each domain in self-assembly. Analytical size-exclusion chromatography (SEC) is widely used for determining molecular weight and oligomeric state of proteins in solution, but it exhibits some important limitations. For example, interactions of proteins with column material can lead to delayed elution and hence erroneous results when relying on column calibration. Since even ideal elution occurs according to hydrodynamic size rather than true molecular weight, there are no appropriate molar mass gel ﬁltration standards for proteins, fragments, or complexes of nonglobular structure that present a different size or molecular weight dependence than globular proteins. The use of size-exclusion chromatography in combination with Figure 1: Molar mass, as determined by multi-angle light multi-angle light scattering (SEC-MALS) determines molecular weights scattering, versus elution volume of kinase fragment (red), independently of elution time and conformation, overcoming the need bovine serum albumin (BSA, blue), and alcohol dehydrogenase (ADH, green). Molar masses deduced from the elution volumes for standards and the errors inherent in analytical SEC. of kinase fragment and ADH are shown to be misleading when This note describes the analysis of a kinase fragment lacking compared with absolute molar masses from SEC-MALS. its association domain to determine its oligomeric state in solution. SEC-MALS revealed that the kinase moiety clearly remains dimeric in ADH tetramer (green trace, 150 kDa) eluted between the monomer solution, even in the absence of its purported oligomerization domain. and dimer of BSA, possibly because of ADH-column interactions that caused it to elute late relative to its size. Comparison of the fragment’s Experimental Conditions elution volume to ADH monomer and tetramer would mislead the A HP-SEC column was calibrated using bovine serum albumin investigator to assume a tetrameric state, even further removed from (BSA) monomer and dimer. The kinase fragment and alcohol the truth than comparison to BSA. dehydrogenase (ADH, 38 kDa) were each run on the column, and the elution times compared to those of BSA monomer (66.4 kDa) Conclusions and dimer (133 kDa). Absolute molar mass (MW) of the proteins SEC-MALS provides true solution molecular weight for proteins, at each elution volume were determined by analysis of signals overcoming the inherent errors produced by reliance on column from the multi-angle light scattering and refractive index detectors calibration. Here we have shown that kinase fragments are dimeric, (DAWN and Optilab, respectively, Wyatt Technology) in ASTRA even without the purported oligomerization domain; but they are software (Wyatt Technology). Chromatograms were overlaid with not trimeric or tetrameric as might have been deduced via column molecular weight values calculated for each elution time along the calibration. peaks, as seen in Figure 1.

Results The monomeric kinase fragment has a sequence molar mass of 53.5 kDa. The fragment (red trace) eluted at nearly the same volume as the BSA dimer (blue trace), suggesting that its molar mass is approximately 140 kDa, or trimeric. However, MALS determines an absolute molar mass in solution of 108 kDa, revealing that Wyatt Technology Corp. the protein is actually a dimer. The molecular weight is absolutely 6330 Hollister Avenue, Santa Barbara, California 93117, USA uniform across the peak, indicating a high degree of homogeneity. Tel.: +1 (805) 681 9009 • Fax: +1(805)681 0123 Such early elution is indicative of a nonglobular conformation. Website: www.wyatt.com • E-mail: [email protected]

THE APPLICATIONS BOOK – MARCH 2018 173 PHARMACEUTICAL/DRUG DISCOVERY

Extended Dynamic Range ELSD for Impurity Profi ling and Purifi cation SEDERE SAS

When running impurity profi le analysis, dynamic range of the detector is one of the most important parameters. Current norms in the pharmaceutical industry require at least three orders of magnitude to be able to quantify 0.1% impurities. Running low dynamic range detector analysis often requires a multi-injection sequence with external calibration, since it is impossible to get a single injection chromatogram with a nonsaturated major peak along with minor peaks simultaneously. Evaporative light scattering detection (ELSD) is a nearly universal technique that should be considered as an advantageous alternative to ultraviolet (UV) detection in impurity profiling, because the response factor is generally tighter than with UV detectors, providing a more accurate picture of impurity profile. However, in some cases, ELSD does not provide enough dynamic range to assess impurity profile within one chromatogram. A new TM generation ELSD equipped with a patented feature (SAGA ) is Figure 1: SEDEX LT-ELSD Model LC, new generation shown to offer an efficient way to override this limitation, providing SAGATM-enabled ELSD. more than four orders of magnitude of dynamic range within a single setting, thus allowing the impurity profiling in one injection (Figure 1). This process is called SAGATM (SEDEX Automated Gain Adjustment). Thanks to an innovative design, this new ELS detector can adapt in real time the gain setting to avoid any signal saturation, and can provide a quantitative view of the chromatogram with just a single setting, thus drastically enhancing the dynamic range of the detector (Figure 2). The use of dedicated drivers for data exchange between the SAGATM-enhanced ELSD and the chromatography data software (CDS) is mandatory to avoid the limitation induced by the use of A/D converters. Several drivers are available for major CDS. This new ELSD feature can also be advantageously used in purification, where higher dynamic range allows better monitoring of the purification process, thus avoiding saturation of the detector without compromising sensitivity of minor compounds. Figure 2: Improvement of dynamic range with SAGATM: Application the same injection is replicated on an ELSD with or without SAGATM. The SAGATM feature allows quantitation without Hydrocortisone was selected as a standard compound to investigate signal saturation because of extended dynamic range. the use of SAGATM in a pharmaceutical analysis environment. This work was only focused on a quantitative analysis of the main Results compound and other impurities described in the pharmacopoeias In our method, the first step was to find the nominal were not considered. Prednisolone was also selected as a qualifier concentration C0 of hydrocortisone that allows the quantitation compound for specificity of the method during the tests. of a peak at 0.1% of C0 concentration (that is, over the limit A simple reverse phase UHPLC method was selected to decrease of quantification [LOQ]) as defined in pharmacopoeias and cycle time. The SAGATM-enabled ELSD involved in this work was a regulation texts. The second step was to verify that the response SEDEX LT-ELSD LC. This detector is optimized for standard HPLC curve of hydrocortisone was mathematically correlated with performed in routine analyses and in regulated environments. A the concentration at levels up to 120% C0, providing both the dedicated SAGA-compatible driver for Openlab was used. method capability of assay determination and related substances

174 THE APPLICATIONS BOOK – MARCH 2018 PHARMACEUTICAL/DRUG DISCOVERY

quantification within the same nominal concentration injection of hydrocortisone. Sensitivity (1) was assessed by a visual determination of the peak at limit of detection (LOD) and signal-to-noise ratio (S/N) at LOQ. Hydrocortisone SNR = 13.2 The peak at 0.05% C0 was visible, which confirmed that LOD of Signal (mV) Prednisolone the method is below 0.05%. On LOQ peak at 0.1% C0 the S/N was found to be higher than 10 (see the Figure 3), which confirmed that the sensitivity of the method is acceptable for impurity assessment. In addition, %RSD (n = 6) at LOQ was 6.3%, which is below the standard maximum limits of %RSD at LOQ found in the literature (10% to 20%) (2). 0 0,5 1 1,5 2 2,5 3

Linearity was checked in a continuous domain from LOQ (0.1% Time (min)

C0) up to 120% C0 (Figure 4). The results confirmed that the method = is linear with r² 0.9997, using a direct linear correlation; however, Figure 3: Chromatogram example of hydrocortisone the confidence interval on the origin (a/Sa) is remote from the and prednisolone mix at 0.1% C0. Student coefficient for this set of data. This is explained by a small nonlinear domain in the lowest concentrations. For assay method validation, the linearity study can be considered with a shorter set of concentrations (typically from 10% to 120%

C0). In that domain the linearity is demonstrated (Figure 5) with a correlation coefficient r² = 0.9996, the a/Sa factor is 0.556 (below the Student coefficient for this set of data). The linear relationship is a first step for assay validation, the second point to confirm is that the method is able to provide reproducible results. The %RSD (n = 6) at C0 was 1.0% (equal or below the limits found in the literature), which means that the method can be used for hydrocortisone assay. For the lower range (used for the determination of related substances), three options are possible: • The use of limit tests where injections of impurity standards at the Figure 4: Complete 13-levels linearity data on the specification limit are injected before the sample to be qualified, so = range 0.1% C0 to 120% C0 (r² 0.9997). that direct comparison of areas may give a “compliant to specifications” or an “out of specification” answer. The excellent and demonstrated • In purification, where the concentration of the sample is generally homogeneity of response coefficient of ELSD (3) may offer the option unknown and for which a signal saturation risk exists and may to use a single injection of the major compound at the specification result in a sample loss. limit as a reference standard for limit test, other impurities are • In pharmaceutical analysis, where quality assessment of the considered as related substances to that major compound. products requires an important dynamic range. • The use of nonlinear quantification for low range concentrations: most current chromatography data software provides the References capability to work with a nonlinear relationship (using a (1) International Conference on Harmonization, ICH Quality Guidelines: polynomial or power function). Validation of Analytical Procedures: Text and Methodology (ICH, Geneva, • Taking advantage of the wide dynamic range, it is possible to shift Switerland, 1994). the domain to higher concentrations. Increasing the reference (2) J. Ermer and J.H. McB. Miller, Eds., Method Validation in Pharmaceutical

concentration C0 slightly should allow the peak of 0.1%C0 to be Analysis. A Guide to Best Practice (John Wiley & Sons, found in the linear region of response of the method, allowing a 2006), Chapter 2.6.5. true direct linear relationship to be used for quantitative studies. (3) Liling Fang et al., J. Comb. Chem. 2, 254–257 (2000).

Conclusion A prevalidation study was performed on hydrocortisone using an ELS detector featuring a novel extended dynamic range technology. This new feature allows the ELSD to acquire data without limitation SEDERE SAS in the higher signal range, so that sensitivity is always at the highest 841, Boulevard Duhamel du Monceau, 45160 Olivet - France setting (no more need for attenuation or gain parameters). This Tel: +33 (0)2 38 66 84 47 Fax: +33 (0)2 38 56 46 85 feature also brings a relevant solution: Email: [email protected]

THE APPLICATIONS BOOK – MARCH 2018 175 PHARMACEUTICAL/DRUG DISCOVERY

Analysis of Residual Solvents in Drug Products Using Nexis GC-2030 Combined With HS-20 Headspace Sampler—USP <467> Residual Solvents Procedure A Shimadzu Europa GmbH

Residual solvents in pharmaceuticals are defi ned as volatile organic Table 2: HS-20 method for USP 467 procedure A compounds used in or generated from the manufacture of drug substances, pharmaceutical additives, or drug products. They are Oven temperature 80 ºC strictly controlled according to risk classifi cations from Class 1 to Sample line temperature 110 ºC Class 3, which are based on the risk to human health. Headspace Transfer line temperature 120 ºC GC methods specifi ed in theUSP (US Pharmacopeia), General Chapters <467> Residual Solvents, are commonly used for analysis Vial stirring Off of residual solvents. This application note presents data obtained Vial volume 20 mL using the Shimadzu HS-20 Headspace Sampler and Nexis GC-2030 Gas Chromatograph (Figure 1), from Class 1 and Class 2 standard Vial heat-retention time 60 min solutions, in accordance with Water-Soluble Articles, Procedure A, Vial pressurization time 1 min in USP <467> Residual Solvents. Vial pressure 75 kPa

Class 1 Loading time 1 min Figure 2 shows the Class 1 standard solution chromatogram. Needle fl ush time 5 min Procedure A requires that the signal-to-noise (S/N) ratio obtained for 1,1,1-Trichloroethane in this chromatogram be 5 or higher. As shown, the S/N ratio was 220. Even for carbon tetrachloride, which had the lowest sensitivity level, the S/N was 20. Table 3 indicates the S/N ratio of each peak and the repeatability of the peak area (n = 6).

Instruments and Analytical Conditions

Table 1: GC method for USP 467 procedure A

Model Nexis GC-2030

Detector FID-2030 Figure 1: Nexis GC-2030 with HS-20.

Headspace sampler HS-20

Column SH-Rxi-624 Sil MS 30 m, ID 0.32 mm, df 1.8 m 40 ºC (20 min) - 10 ºC/min - 240 ºC (20 min) Column temperature Total 60 min 1 Injection mode Split 1:5

Carrier gas controller Constant linear velocity (He) 2 4 Linear velocity 35 cm/s

Detector temperature 250 ºC 5 FID H2 fl ow rate 40 mL/min 3

FID make up fl ow rate 30 mL/min (He) 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 Time (min) FID air fl ow rate 400 mL/min

Injection volume 1 mL Figure 2: Chromatogram of water-soluble articles Class 1 standard solution by procedure A.

176 THE APPLICATIONS BOOK – MARCH 2018 PHARMACEUTICAL/DRUG DISCOVERY

12 13 14 1 5 8 3 4 56 7 8 11 15 1: Methanol 7 1: Hexane 2: Acetonitrile 2: Nitromethane 3: Methylene chloride (DCM) 3: Chloroform 4: trans-1,2-Dichloroethylene 4: 1,2-Dimethoxyethane 5: cis-1,2-Dichloroethylene 5: Trichloroethane 1 6: Tetrahydrofuran 3 6: Pyridine 2 7: Cyclohexane 7: Methylbutylketone 6 8: Methylcyclohexane 8: Tetraline 9 9: 1,4-Dioxane 2 4 0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 25.0 27.5 30.0 32.5 10: Toluene Time (min) 11: Chlorobenzene 5.0 10.0 15.0 20.0 25.0 30.0 35.0 Resolution 2.4 3 4 1 12: Ethylbenzene 13: m,p-Xylene Time (min) 2 14: o-Xylene 15: Cumene

2.0 3.0 4.0 5.0 6.0 Figure 4: Chromatogram of water-soluble articles Time (min) Class 2B standard solution by procedure A. Figure 3: Chromatogram of water-soluble articles Class 2A standard solution by procedure A. Class 2 Table 3: S/N ratio and repeatability of Class 1 Due to the large number of components in the Class 2 standard solution, it was separated into two mixtures: A and B. Respective No. Compounds S/N ratio %RSD (n = 6) measurement results are shown in Figure 3 and Figure 4. Procedure A requires that the resolution for acetonitrile and methylene chloride 1 1,1-Dichloroethane 320 2.8 in the Class 2 standard solution Mixture A chromatogram be 1.0 or 2 1,1,1-Trichloroethane 220 2.3 greater. Figure 3 shows that, using the Restek SH-Rxi-624SilMS low-bleed 3 Carbon tetrachloride 20 2.9 column, the specified peaks are completely separated, with a resolution of 2.4. 4 Benzene 170 2.5

5 1,2-Dichloroethane 60 3.4

Shimadzu Europa GmbH Albert-Hahn-Str. 6–10, D-47269 Duisburg, Germany Tel.: +49 203 76 87 0 Fax: +49 203 76 66 25 E-Mail: [email protected] Website: www.shimadzu.eu

THE APPLICATIONS BOOK – MARCH 2018 177 EVENT NEWS

Analytica 2018 10 May 2018 From Cradle to Grave: The Chromatographic (Analytical) Analytica will take place from Method Life Cycle 10–13 April 2018 at the Messe Piccadilly, London, UK München, in Munich, Germany. E-mail: [email protected] The Analytica conference is a Website: www.chromsoc.com/chrom- highlight of Analytica and will socevents take place at the International Congress Center (ICM) from 13–18 May 2018 10–12 April 2018. 42nd International Symposium on The Analytica conference Capillary Chromatography (ISCC) covers cutting-edge research and and 15th GC×GC Symposium applications using modern chemical and bioanalytical technologies. The Palazzo dei Congressi, Riva del increasing digitization in the laboratory and the handling of the flood of Garda, Italy results are a major focus this year. What does having to process thousands E-mail: [email protected] of samples every hour mean for laboratory management? How can this Website: http://www.chromaleont.it/iscc flood of data be reliably evaluated and managed? There are full-day symposia on chromatography and spectroscopy, 3–7 June 2018 and new developments in instrumental analytics remain at the core of the 66th Conference on Mass Analytica conference. In biosciences, “multi-omics” will be covered in Spectrometry and Allied Topics detail. Cutting-edge techniques for the analysis of proteomes, genomes, San Diego, California, USA and metabolomes will also be a focus in the conference sessions. The E-mail: [email protected] session “Big Data Tools for Omics”, chaired by systems biologist Lennart Website: www.asms.org/conferences/ Martens from the University of Ghent (Belguim), offers insights into data annual-conference/annual-conference- handling for “omics” applications, and a series of lectures on microbiome homepage analytics will focus on big data and bioinformatics. “Trends in Analytical Toxicology” will be of interest to forensic analysts 26–27 June 2018 attending the symposium and includes presentations on the detection of The 2nd Copenhagen Symposium on drugs in the dental material of deceased persons and urine screening by Separation Sciences (CSSS 2018) means of paper-spray mass spectrometry (MS). DGI-Byen Hotel, Copenhagen, Denmark The conference also offers a varied programme for environmental E-mail: [email protected] analysts. One dedicated lecture series will address perfluorinated Website: https://cphsss.org compounds. These stable chemicals are used in many products, from 29 July–2 August 2018 fire-extinguishing agents to paper cups, and contaminate large areas 47th International Symposium on worldwide. The detection of perfluorinated compounds is challenging High Performance Liquid Phase because there are no analytical standards for many analytes in this large Separations and Related Techniques class of compunds. Marriott Wardman Park, Washington, One challenge unites all analysts involved in analytical science. DC, USA Decreasing detection limits and increasingly powerful measuring methods, E-mail: [email protected] along with increasing automation and growing sample throughput rates Website: www.hplc2018.org have caused data volumes to swell. The chemometrics symposium provides information on mathematical and statistical methods that help to 23–27 September 2018 analyze vast analytical datasets. 32nd International Symposium on The Analytica conference invites you to think outside the box and allow Chromatography yourself to be inspired by colleagues from other analytical disciplines, Cannes-Mandelieu, France according to the organizers. E-mail: [email protected] This year’s programme has been compiled once more by the Website: http://isc2018.fr Association of German Chemists (Gesellschaft Deutscher Chemiker, GDCh), the Society for Biochemistry and Molecular Biology (Gesellschaft 17–19 October 2018 für Biochemie und Molekularbiologie, GBM), and the German Society for 12th International Conference on Clinical Chemistry and Laboratory Medicine (Deutsche Gesellschaft für Packed Column SFC (SFC 2018) Klinische Chemie und Laboratoriumsmedizin, DGKL). Strasbourg, France The Analytica exhibition is a major global event with than 32,000 visitors E-mail: register@greenchemistrygroup. and over 1150 exhibitors. For futher information go to www.analytica.de org Website: www.greenchemistrygroup.org

Please send any upcoming event information to lewis.botcherby@

ubm.com Photo Credit: Vlada Photo/Shutterstock.com

178 LC•GC Europe March 2018 F3 U2 TU RE4

NE3 TW O4 RK

GL 03 B2 AL2

Highlights in 2018: The World’s No. 1 Live Labs and The largest laboratory trade fair in the world features the entire range of products and solutions for industry Digital Trans- and research laboratories. formation The first-rate scientific analytica conference, world premieres, the latest product developments, unique Live Labs, special shows, forums and Focus Days await you!

April 10–13, 2018 I analytica exhibition April 10–12, 2018 I analytica conference 26th International Trade Fair for Laboratory Technology, Analysis, Biotechnology and analytica conference www.analytica.de Metabolomics, Flavor and Fragrance, Lab Food Safety, Pharmaceutical, Environ- mental… the list is long. In every ﬁeld, laboratories must deliver improvements in throughput and performance that are Growth only possible with efﬁcient automation.

The MultiPurpose Sampler MPS performs sample preparation, analyte concentration and introduction to GC/MS or LC/MS.

The MPS even generates your standards, adds them to the samples and runs the calibration. Highly precise and efﬁcient automation for accurate results.

NEW: Sequence by barcode Place your samples in the MPS. Click-Start the system. Method and Sequence are automatically set up.

How can we support your growth?

Sequenceq byy Barcode