[ theory ] · 2007-11-21 · drug discovery screening, raw material analysis, impurity profiling,...

A BREAKTHROUGH TECHNOLOGY SHOULDN’T JUST AFFECT YOUR LABORATORY.

IT SHOULD IMPACT THE WAY YOUR ENTIRE ORGANIZATION WORKS.

When technology isn’t holding back scientific

progress, how far will your organization go?

With the revolutionary Waters® ACQUITY

UltraPerformance LC® system, laboratory

scientists have been empowered to develop

methods with greater confidence and efficiency.

ACQUITY UPLC® has given scientists data

and sensitivity far superior to HPLC. As a

result, its impact has been felt in laboratories,

boardrooms, and everywhere science improves

lives. The future of Liquid Chromatography

is the ACQUITY UPLC system. To learn more,

visit www.waters.com/a1.© 2007 Waters Corporation. Waters, The Science of What’s Possible, ACQUITY UltraPerformance LC, UPLC, and ACQUITY UPLC are trademarks of Waters Corporation.

[ theory ]

JUNE 2007 ULTRAPERFORMANCE LC® 3

An Introduction to the ACQUITYUltraPerformance LC® System

I n 2004, Waters introduced a new category of LC technology that changed separationscience for all time.We call it ACQUITY UltraPerformance LC® or UPLC®.Based on a patented 1.8 micron particle and instrumentation designed from the ground

up to maximize the potential of this new particle chemistry, UPLC delivers new levels ofresolution, speed, and sensitivity.

As news of this technology spread and interest quickly caught on, scientists began toembrace its potential to drive new products through the development process faster, pushMS performance to the next level, and open new doors to scientific discovery.

At the request of scientists, we’ve opened the architecture to allow other mass spectrom-etry makers, including ThermoFisher, Bruker Daltonics, and Applied BioSystems/MDSSciex, a means to control the ACQUITY UPLC® system. And we’ve added many newcolumns, detectors, and refinements.

UPLC technology is now mature enough for the most demanding of separations andmeets all the requirements of a highly robust, dependable and reproducible LC product.

As this supplement attests, the applications for UPLC technology are many and thenumber is growing every single day. At last count more than 100 peer-reviewed articleshave appeared in the world’s best-known and most respected scientific journals showcas-ing research obtained in whole or in part with our ACQUITY UPLC system.

Through a connected set of technologies for UPLC (instrumentation, informatics, andchemistries) designed to work together as a total system, we have one goal in mind: to pro-vide you with quality information faster, and make a meaningful impact on your businessand scientific bottom line.

It’s been stated to me that the single-greatest constraint to new scientific discoveries liesin the development of analytical tools. We find ourselves fortunate to be in the criticalpath of scientific progress and behind innovations like UPLC. As we like to think aboutit, it is one example of the way that Waters is making science possible.

Art CaputoExecutive Vice President and President, Waters Business OperationsWaters Corporation

For More Information Circle #

JUNE 2007

Contents6

9

12

16

21

26

Transitioning Existing LC Methods to NewTechnologies: Choosing a Liquid ChromatographyPlatform for 2012Robert Plumb and Jeffrey R. Mazzeo

Transfer of the USP Assay for Simvastatin to UPLC®

Paul Rainville and Robert Plumb

Analysis of Soy Isoflavones from a Dietary Supplement Using UPLC® with PDA andSQ DetectionAntonietta Gledhill

Application of ACQUITYTM TQD for the Analysis ofPesticide Residues in Baby Food James Morphet and Peter Hancock

UPLC®/MS/MS Bioanalytical Method Validation ofAcebutolol and Pindolol Using an AnalogueInternal StandardEd Sprake and Iain Gibb

Peptide Separation Technology: QuantitativeAspects of UPLC® Peptide MappingThomas E. Wheat, Ziling Lu, Beth Gillece-Castro, and Jeffrey R. Mazzeo

LC/MS-based Differential Proteomics of theMitochondria of [PSI+] and [psi-] Saccharomycescerevisiae StrainsJacek Sikora, Michael Dadlez, Chris Hughes, Hans Vissers,Thérèse McKenna, Jim Langridge, and Magdalena Boguta

Removal of Interferences and Easier MetaboliteDetection by Ion Mobility Mass Spectrometry Jose Castro-Perez, Kate Yu, and John Shockcor

ACQUITY UPLC® SystemApplication Notebook – June 2007

29

32

4 ULTRAPERFORMANCE LC®

GIVE LABORATORY SCIENTISTS EASIER ACCESS TO MS/MS.

AND TAKE LABORATORY PRODUCTIVITY TO UNPRECEDENTED LEVELS.

With elevated performance and higher speeds, the Waters® ACQUITY® TQD

gives chromatographers all the benefits of tandem quad MS in a system that’s

easy to operate. Using patented T-Wave™ technology, ACQUITY TQD takes

full advantage of the narrow peak widths and increased sensitivity provided

by UPLC.® This results in high performance quantification, increased

throughput, and the ability to take lab analyses to the next level without compromising

data quality. To learn more, visit www.waters.com/tq1.

© 2007 Waters Corporation. Waters, The Science of What’s Possible, ACQUITY, T-Wave and UPLC are trademarks of Waters Corporation.

PUBS:

American LaboratoryAmerican Drug DiscoveryAmerican Pharmaceutical ReviewBio BusinessBio IT WorldBioTech WorldBio Pharam InternationalChemical & Engineering NewsDrug Discovery and Dev

Drug Discovery NewsFood QualityFood Safety MagazineGenetic Engineering News (GEN)Genome TechnologyGenomics & ProteomicsJour of The American Soc for Mass SpectrometryLab Asia Media GuideLCGC AsiaLCGC N. America

LCGC EuropeMolecular and Cellular Protemomics Pharmaceutical TechnologyPharmaceutical Drug & DiscoveryPharmaceutical Technology EUROPEPharm Form and QualityProteomics JournalScientific Computing & InstrumentationScientific Computing World

Please note this is a COMMON SIZE mechanical file, you will need to center file using the center marks

provided when placing ad in the publication page area. (common size = smallest live/smallest trim / largest bleed)

Job # & CC: WATR7980_A

Job Name: WATR7980_A_TQD_SoWP_DBmec.indd Date: 02-20-07Live: 7"x10" SmallTrim: 7.875"x10.5” LargeTrim: 8.375"x10.875" Bleed: 8.625"x11.375" Page: 1 Rev: 1Stage: mech Release: 02-20-07 DB AD: DP CW: AS PM: DG Scale: 100%Color: 4c Gutter: — Other: — ST: DB

APP

RO

VALS C-LO

MECH PROOF

MK: ______________________ PM: ______________________ AB: ______________________

AD: ______________________ CW: ______________________ RM:______________________

QC: ______________________ CD: ______________________

Initials / Date Initials / Date Initials / Date

Client / Date: ________________________________________________

JUNE 2007

Transitioning Existing LC Methods to New Technologies:Choosing a Liquid Chromatography Platform for 2012

Robert Plumb and Jeffrey R. MazzeoWaters Corporation, Milford, MA

A nalytical laboratories face the continual challenge ofbalancing investment in new technology - to improve

throughput and performance - with the need to run existingmethods and tests to support current production. HighPerformance Liquid Chromatography (HPLC) is one of thesekey analytical technologies that has become a major tool forfood safety testing, life science studies, and pharmaceutical,environmental and chemical analysis.• In the pharmaceutical industry, HPLC is employed

throughout the whole drug analysis process, includingdrug discovery screening, raw material analysis, impurityprofiling, stability studies, pharmacokinetic studies andfinal product testing.

• HPLC has found great application in the environmentaland chemical analysis arena, where it has been employedfor testing water quality, soil samples and product quality.

• Food safety analysis has also benefited from the capabili-ties of HPLC either alone or in combination with massspectrometry (MS), allowing food to be rapidly assessedfor contaminants such as pesticides or veterinary drugs.The popularity of liquid chromatography is due to its

speed of analysis, compatibility with samples to be analyzed,ease of use, simple interfacing with a wide range of detectors(optical, electrochemical, radiochemical, evaporative lightscattering, NMR and MS), wide range of stationary phases,high resolving power and easy scale-up to preparative chro-matography.

Over the last two years, several instrument manufacturershave introduced liquid chromatography systems that aredesigned to take advantage of columns packed with smallerparticles. Compared to traditional LCs, these instruments arecapable of higher pressure operation, have reduced systemvolumes, use faster autosamplers and employ detectors withmuch higher data capture rates.

All of these elements are required to leverage the benefitsof higher resolution, sensitivity and throughput offered bycolumns packed with smaller particles. R&D and methoddevelopment laboratories have extensively adopted these newLC platforms, as they greatly improve the efficiency of prod-uct development.

Methods for products advancing through the pharmaceu-tical pipeline are now being developed with these new LCplatforms. Thus, quality control laboratories must also makesimilar technology investments to enable them to run the

methods that will be required to release new products.Furthermore, many QC labs are in the process of replacingtraditional instruments purchased five to seven years ago thathave fully depreciated.

The convergence of these two trends - new methods beingdeveloped on new LC platforms and the need to replace largenumbers of legacy instruments - leaves QC lab managers at acritical decision point with respect to investing in a platformfor the future. Assuming a depreciation cycle of five years, theplatform chosen today must remain relevant until 2012.

Such a long-term view makes selecting one of the new LCplatforms very attractive. However, given that the chosenplatform will predominantly be running legacy methods dur-ing its early lifecycle, it is absolutely critical that it can effec-tively run these methods. At the same time, it must be fullycapable of running methods being developed today withsmall particle columns - as well as methods developed infuture years that will utilize next-generation columns.

In this paper, we demonstrate that a new LC platform,using UltraPerformance LC® (UPLC®) Technology, can beused to run these legacy methods. We will then demonstratethe stepwise transfer of a legacy method to a short, small par-ticle column for greatly improved throughput with equiva-lent resolution. Finally, we will demonstrate transfer of thelegacy method to a long, small particle column for improvedthroughput with increased resolution.



Figure 1: Omeprazole separation using a 4.6 x 150 mm, 5 mm C8column.

Time (min)

0.45

0.00

AU

Om

epra

zole

- 14

.196

0.00 6.00 12.00 18.00 24.00



DiscussionThe increased chromatographic performance obtained from theWaters® ACQUITY UPLC® system, when used in combinationwith the 1.7 mm stationary phase, is illustrated in the analysis ofOmeprazole. Omeprazole is in a class of drugs called protonpump inhibitors (PPI), which blocks the production of acid bythe stomach.

The first chromatogram, Figure 1, illustrates the chromatog-raphy obtained on a 4.6 x 150 mm, 5 mm C8 column elutedunder isocratic conditions with a phosphate buffer: acetonitrile(3:1) mobile phase at a flow rate of 0.8 mL/min.

We can see from this data that the major peak elutes at aretention time of 14.2 minutes and peak width of nearly 1minute at the base. The small peaks at the front of the chro-matogram are the observed impurities in the sample. The datadisplayed in Figure 2 shows the same separation again performedon the ACQUITY UPLC system, but this time employing anACQUITY UPLC BEH 2.1 x 150 mm, 1.7 mm C8 column,using the same mobile phase conditions. The injection volumeand flow rate were scaled to reflect the change in column inter-nal diameter and particle size, with the flow rate now being0.5 mL/min.

We can see from the data that the retention factors are simi-lar, but the peak width is much reduced, now being only 13 sec-onds wide at the base, giving a plate count of 15,300 comparedto the value of 8704 obtained by the HPLC method. The analy-

sis time is also significantly reduced, from 25 minutes to eightminutes.

Thus this data illustrates that, when scaled correctly and thesame column length is used, the UPLC solution produces supe-rior peak shape and sensitivity. One other major benefits ofincreased performance are the sharper peaks; they not onlyincrease resolution and sensitivity but also simplify the task ofpeak integration, reducing the need for manual reintegration.

One of the major advantages of using UPLC in the field ofproduct release testing is the ability to shorten analysis timewithout reducing peak resolution. This is achieved by scaling theseparation from the existing LC methodology to UPLC by keep-ing the ratio of column length to particle size (L/dp) constant.This increases throughput without compromising analytical per-formance.

Figure 3 shows the separation of the common beta1-selective(cardioselective) adrenoreceptor blocking agent Atenolol, using a3.9 x 300 mm, 5 mm C18 column. The column was eluted at aflow rate of 0.52 mL/min with a mobile phase of 70% phos-phate buffer (pH 3.0) and 30% methanol. Detection was per-formed by UV absorbance at 226 nm.

The inset table in Figure 3 shows the assay reproducibility ofthe separation performance. Here we can see that the retentiontime variation was 0.47% RSD and the peak area deviation cal-culation returned a value of 0.3%.

Figure 2: Omeprazole separation using an ACQUITY UPLC BEH 2.1 x150 mm, 1.7 mm C8 column.

Time (min)0.00

0.90

0.00

AU

Om

epra

zole

- 4.

129

8.806.604.402.20

Figure 3: Atenolol separation using a 3.9 x 300 mm, 5 mm C18 column.

Time (min)0.00

0.000

0.020

AU

Ate

nolo

l - 7

.724

14.007.00

Figure 4: Atenolol separation using an ACQUITY UPLC BEH 2.1 x100 mm, 1.7 mm C18 column.

0.00

0.00

0.06

AU

0.73

1

5.004.003.002.00Time (min)

1.00

Figure 5: Budesonide separation using a 4.6 x 250 mm, 5 mm C18column.

Time (min)

0.000

0.040

AU

4.00 8.00 12.00 16.00 20.00 24.00

JUNE 2007



The assay was repeated on an ACQUITY UPLC BEH 2.1 x100 mm, 1.7 mm C18 column, with the column operated at aflow rate of 0.52 mL/min. In this assay, the particle size wasreduced by a factor of three, from 5 mm to 1.7 mm, while thecolumn length was also reduced by a factor of three, from 300mm to 100 mm, thus keeping the ration of L/dp constant.

The result of scaling this separation to ACQUITY UPLCtechnology can be seen in Figure 4. The analysis time has beensignificantly reduced - from 7.8 minutes to less than one minute- while the assay performance remains unchanged in terms ofplate count.

This magnitude of improvement in throughput and perform-ance can result in a reduced time for product release and/or a reduc-tion in the number of instruments required within the laboratory.

UPLC can both reduce analysis times and increase chromato-graphic performance. One way this can be achieved is by reduc-ing the column length by a factor of two while reducing the par-ticle size by a factor of three, from 5 mm to 1.7 mm. The flatternature of the van Deemter plot for the 1.7 mm material allows afaster mobile phase linear velocity to be employed, furtherspeeding up the analysis.

This approach is illustrated in the analysis of Budesonide, acommon inhaled steroid; the chromatogram shown in Figure 5depicts the separation of the two components on an ACQUITYUPLC system using a 4.6 mm x 250 mm, 5 mm C18 columnwith a mobile phase of 68% 20 mM ammonium formate buffer(pH 3.2) and 32% acetonitrile. Detection was performed byUV absorbance at 240 nm.

In this figure, we can see that there are small low-level impu-rities as well as the R- and S-epimers of the active component.

The assay was transferred to an ACQUITY UPLC BEH 2.1x 100 mm, 1.7 mm C18 column operating with the same mobilephase and a flow rate of 0.6 mL/min. The resulting chro-matogram is displayed in Figure 6.

The analysis time has been reduced from 25 minutes to 10minutes; the chromatographic performance has also improved,producing sharper peaks. The extra chromatographic resolutionand increased sensitivity with the ACQUITY UPLC system hasallowed the detection of more impurities in less time, while theresolution of the two major peaks has been maintained. Usingthis approach, the method was simply transferred to UPLC withboth improved throughput and assay performance.

ConclusionLaboratories making investments in new LC platforms musttake into account their ability to run legacy methods, while alsopreparing themselves for future methods that will employ newcolumn technologies.

UltraPerformance LC is a new category of separation sciencethat builds upon the well-known and established principles ofliquid chromatography. UPLC leverages the performance of sub-2 mm particles to provide increased resolution, sensitivity andthroughput.

The ACQUITY UPLC system has been specifically designedto exploit the capabilities of these new chromatographic station-ary phases, with detailed attention focused on controlling systemvolumes and peak dispersion. Although this system has beendesigned to work with these new small particles, it also providesan excellent platform for easily transferring existing HPLCmethods to the new UPLC methods. Thus, this new technologyis allowing users to confidently move from existing LC methodsto the new technology of UPLC. n

Figure 6: Budesonide separation using an ACQUITY UPLC BEH 2.1 x100 mm, 1.7 mm C18 column.

Time (min)

1

2

3

4

5

6

7

0.00 6.000.00

1.40R-epimer

S-epimer

0.028

0.000

AU

2.20 5.50 7.70 9.90


Transfer of the USP Assay for Simvastatin to UPLC®

Paul Rainville and Robert PlumbWaters Corporation, Milford, MA

S imvastatin (Figure 1) is a lipid-lowering agent that isderived synthetically from a fermentation product of

Aspergillus terreus. After oral ingestion, simvastatin (an inac-tive lactone) is hydrolyzed to the corresponding b-hydroxy-acid form. This is an inhibitor of 3-hydroxy-3-methylglutaryl-coenzyme A (HMG-CoA) reductase. This enzyme catalyzesthe conversion of HMG-CoA to mevalonate, which is an earlyand rate-limiting step in the biosynthesis of cholesterol.

Simvastatin is most commonly administered orally in atablet containing either 5 mg, 10 mg, 20 mg, 40 mg or 80 mgof the active pharmaceutical ingredient (simvastatin) and thefollowing inactive ingredients: cellulose, hydroxypropyl cellu-lose, hydroxypropyl methylcellulose, iron oxides, lactose, mag-nesium stearate, starch, talc, titanium dioxide and other ingre-dients. Butylated hydroxyanisole is added as a preservative.

The Uniformity of Content Assay for simvastatin isaccomplished using high performance liquid chromatogra-phy (HPLC) with ultraviolet detection (UV). The faster thatthis assay can be performed, the faster the finished productcan be released and revenue realized. The current USP 30-NF25 monograph assay for simvastatin tablets calls for the use ofa liquid chromatograph equipped with a 238 nm detectorand a 4.6 mm x 25 cm column containing packing L1 main-tained at a temperature of 45 °C, and a flow rate of about1.5 mL per minute.

The performance of the assay demands a capacity factor,k’, of not less than 3.0. The column efficiency should not beless than 4500 theoretical plates, with a tailing factor of notmore than 2.0 and a relative standard deviation for replicateinjections of not more than 2.0%. The current HPLC-basedassay has an analysis time of 12 minutes, with a retentiontime of 9.28 minutes for the active ingredient simvastatin.

UltraPerformance LC® is a new category of separation sci-ence which builds upon well-established principles of liquidchromatography, using sub-2 mm porous particles. Theseparticles operate at elevated mobile phase linear velocities toproduce rapid separations with increased sensitivity andincreased resolution. Thus, UPLC® technology allows analy-sis times to be dramatically reduced while still meeting assayacceptance criteria based on plate count, resolution and ana-lyte retention. In this application note, we show how theHPLC assay for simvastatin has been transferred to UPLC.The UPLC assay is compared to the USP assay criteria forperformance and quality.

ExperimentalA standard solution of simvastatin was prepared according to theUSP methodology, and then was diluted to 100 mg/mL forchromatographic analysis.

HPLC conditionsLC System: ACQUITY UPLC®

Column: XBridgeTM C184.6 x 250 mm, 5 mm

Mobile Phase: 65% acetonitrile/35% phosphatebuffer, pH 4.5

Flow Rate: 1.5 mL/minInj. Volume: 10 mLTemperature: 45 °CDetection: ACQUITY UPLC PDA @ 238 nm


Figure 1: Chemical structure of simvastatin.

Figure 2: HPLC separation of simvastatin.

0.00

0.00

0.22

AU

9.28

1

20.0010.00Time (min)

JUNE 2007



UPLC conditionsLC System: ACQUITY UPLC Column: ACQUITY UPLC BEH C18

2.1 x 100 mm, 1.7 mm and2.1 x 30 mm, 1.7 mm

Mobile Phase: 65% acetonitrile/35% aqueousFlow Rate: 0.56 mL/minInj. Volume: 0.8 mLTemperature: 45 °CDetection: ACQUITY UPLC PDA @ 238 nm

Results and DiscussionThe HPLC separation of simvastatin is shown in Figure 2. Herewe can see that the analyte was eluted with a retention time of9.28 minutes. The HPLC analysis generated a k' value of 5.5, a4s plate count of 12,112 and a tailing factor of 0.94.

The HPLC method was scaled to UPLC using theACQUITY UPLC Console Calculator. The calculator allowsthe user to input the current HPLC conditions, including col-umn length, mobile phase composition and flow rate. The cal-culator then generates three results: 1) the conditions of EqualEfficiency, 2) the conditions of Maximum Efficiency, and 3)the conditions of Shortest Analysis Time. The HPLC condi-tions for simvastatin were input and the results obtained areshown in Figure 3.

The data generated shows three results relating to threepotential UPLC conditions. The calculator transfers the HPLCcondition to UPLC, taking into account changes in columndiameter and particle size when calculating the flow rate. Theequivalent conditions are calculated by adjusting the flow rate bythe ratio of 2.12/4.62 to account for column geometry, and thenmultiplied by the ratio of 5/1.7 to account for the reduction inparticle size.

UPLC assaysThe injection volume was scaled to account for the change inoverall column volume.

The transferred conditions of Equal Efficiency suggest the useof a 2.1 x 100 mm, 1.7 mm C18 column operating at a flow rateof 766 mL/min and an injection volume of 0.8 mL. As the injec-tion volume suggested was very low, a 2 mL injection loop wasinstalled in the autosampler. The data obtained from the use ofthese conditions is displayed in Figure 4a.

Results indicate an analysis time of 1.41 minutes. TheACQUITY UPLC system produced an efficiency of 12,874, a k’value of 4.88, and a USP tailing factor of 1.10. These values

Figure 4: The UPLC assay of simvastatin using 4a.) the Equal Efficiency conditions, 4b.) the Maximum Efficiency conditions and 4c.) the ShortestAnalysis Time conditions, as suggested by the ACQUITY UPLC Console Calculator.

Time (min)

(a) (b) (c)

AU

1.41

2

1.92

1

0.23

4

0.00 0.00

0.000.00

0.07

0.14

0.21

0.07

0.14

0.21

0.00

0.03

0.06

0.09

0.002.50 2.50 2.50

Figure 3: ACQUITY UPLC Console Calculator.



compare favorably with the original HPLC values of 12,112 forefficiency, 5.5 for k' and a tailing factor of 0.94.

The conditions of Maximum Efficiency required the use of aflow rate of 560 mL/min and an analysis time of 1.92 minutes.The data obtained using these conditions is displayed in Figure4b. Here we can see that the peak retention was slightly greaterthan with the Equal Efficiency conditions. The chromatograph-ic performance calculations revealed that the chromatographicefficiency increased to a value of 17,685, a k' factor of 5, and atailing factor of 1.10.

The conditions of Shortest Analysis Time were performed ona 2.1 x 30 mm, 1.7 mm C18 column, at a flow rate of1.5 mL/min. The data generated is displayed in Figure 4c. Theseconditions resulted in an analysis time of just 0.23 minutes;however, the efficiency count of this separation was below theUSP acceptance criteria for this assay.

L/dp method transferThe method transfer was achieved by keeping L/dp constant; theinitial conditions used a 250 mm column and a 5 mm particle,giving an L/dp value of 50 (250/5). In moving to a 2.1 x 100 mm,1.7 mm C18 column, the L/dp value increased to 59, whichallowed for a more efficient chromatography system. However,when the 2.1 x 30 mm, 1.7 mm C18 column was employed, theL/dp value was reduced to a value of 29; this is why the chromato-graphic efficiency did not meet the USP acceptance criteria.

Assay performanceThe UPLC assay selected for use and evaluation was the oneemploying Equal Efficiency conditions, as it yielded the high-est throughput while meeting the assay acceptance criteria.The assay reproducibility performance was evaluated usingthe conditions developed by the Console Calculator. Theresults generated are shown in Table 1, where we can see thatthe percent RSD for this methodology was determined to be0.883 for six 0.8 mL replicate injections of the simvastatinstandard. These results were well within the USP acceptancecriteria for the methodology.

ConclusionThe USP HPLC-UV methodology for the assay of the choles-terol-lowering drug simvastatin has been successfully trans-ferred to an ACQUITY UPLC method. The final methodol-ogy reduced the analysis time from an analyte retention of9.28 minutes with HPLC to an analyte retention of just 1.41minutes with UPLC (a seven-fold increase in throughput).The ACQUITY UPLC Console Calculator was used to easilyguide the transfer methodology to ACQUITY UPLC usingsub-2 mm particles. The calculator gave three separate optionsfor UPLC methods: Equal Efficiency, Maximum Efficiencyand Shortest Analysis Time conditions. The Equal Efficiencycalculations gave an almost exact transfer of chromatographicperformance when compared to the original HPLC condi-tions. With both the equal and maximum UPLC conditions,the USP assay criteria of efficiency, k’ and tailing, were met orexceeded. The assay reproducibility performance was deter-mined for the Equal Efficiency method and was found to beless than 1% RSD for six replicate injections. Finally, the assaywas transferred to a Shortest Analysis Time method on theACQUITY UPLC system, which gave an analysis time of 0.3minutes with an analyte retention of 0.23 minutes, resultingin an increase in throughput of 40-fold.

We have demonstrated the ease of HPLC to UPLC methodtransferability, and the benefits that can be obtained in anytime-, resource- and/or revenue-conscious laboratory environ-ment where UPLC can significantly increase throughput withquality results. And with a variety of ACQUITY UPLC BEHcolumn dimensions, scientists have the flexibility to tailor theirUPLC separations to the goals at hand.

Injection Peak Name RT Area % Area Height

1 1 Simvastatin 1.374 388681 100 234823

2 2 Simvastatin 1.374 389533 100 235199

3 3 Simvastatin 1.374 383973 100 231690

4 4 Simvastatin 1.375 380276 100 229737

5 5 Simvastatin 1.375 386983 100 233379

6 6 Simvastatin 1.375 386800 100 233178

Mean 1.375 386041

Std. Dev. 0.000 3410

% RSD 0.03 0.833

Table I. Reproducibility of the UPLC simvastatin assay using the Equal Efficiency methodology.

JUNE 2007



Analysis of Soy Isoflavones from a DietarySupplement Using UPLC® with PDA and SQ Detection

Antonietta GledhillWaters Corporation, Manchester, UK

T he consumption of soy products has been linked to manyhealth benefits, as they contain isoflavones. Isoflavones are

commonly known as phytoestrogens and 12 isoflavones foundin soybeans are daidzein (De), glycitein (Gle) and genistein (Ge)and their respective malonyl (6”-O-malonyl-b-glucoside-),acetyl (6”-O-acetyl-b-glucoside-) and glucosyl (b-glucoside-)forms1. Their structures are shown in Figure 1.

Many research studies have indicated that consumption ofisoflavone-containing functional foods are associated with awide variety of health benefits, including prevention of breastand prostate cancers, cardiovascular disease and reduced symp-toms of diabetes and postmenopausal bone loss2-6. These func-tional foods include soy milk and soy flour.

The approval by the U.S. Food and Drug Administration in1999, allowing the food industry to promote soy protein forheart health7, led to an escalation in sales of soy foods as func-tional foods. These foods are also being promoted for theirisoflavones content.

This application note provides a rapid method for usingreversed-phase UPLC® to detect and characterize the isoflavoneglucoside conjugates present in a commercial soy nutritionalsupplement using PDA and MS detection.

ExperimentalThe isoflavones were extracted from the soy capsules using 9:1methanol/water liquid extraction and sonification for 10 min-utes. The extract was filtered through a 0.45 mm filter.

The mass spectrometer was used in two modes: Full scan(m/z: 50 to 550) and single ion recording (SIR) mode.

UPLC conditionsLC System: ACQUITY UPLC®

Column: ACQUITY UPLC BEH C82.1 x 100 mm, 1.7 mm

Temperature: 35 °CFlow Rate: 500 mL/minMobile Phase A: 0.2% formic acid in waterMobile Phase B: MethanolGradient: 25% B for 0.4 min, 25-40%/1.1 min

Hold for 0.8 min

PDA conditionsPDA: ACQUITY UPLC Wavelength Range: 205 to 450 nm

Resolution: 1.2 nmSampling Rate: 20 spectra/s

MS conditionsMS: SQ DetectorIonization Mode: ESI PositiveCapillary Voltage: 2000 V

Desolvation Temp: 400 °CDesolvation Gas: 1000 L/HourSource Temp: 130 °C

Full scan settingsCone Voltage: 37 VAcquisition Range: 50 to 550 m/z

SIR settingsA dwell time of 10ms was used for each SIR and a delay of 5ms.

ResultsThe relative elution order is: glucosyl, malonyl, acetyl and isoflavoneand the retention times are noted in Table II. The ACQUITYUPLC C8 phase was compared to the equivalent C18 column: C8was found to give better separation using this solvent method.

Figure 1: Structures of three soy isoflavones: daidzein, genisteinand glycitein and their conjugates.



In this application note there are two methods described: oneusing full scan and the other an SIR experiment, where ions indica-tive of the target compounds of interest have been selected.

Full scan provides spectral information (Figure 2) from thefragmentation patterns, which can help with structural determi-nation and is useful when identifying unknown compounds.

Examples of daidzein MS spectra using the full scan methodcan be seen in Figure 2. For the conjugated isoflavone systemsthe parent ion of the conjugate ([M+H]+: m/z 417) was presentalong with the parent ion of the isoflavone ([M+H]+: m/z 255).The m/z 439 and 277 may be attributed to [M+Na]+.

Using the full scan data it is possible to extract the ions ofinterest and this procedure has been performed in Figure 3A form/z 255, 271 and 285. The same procedure was performed for260 nm from the PDA detector.

Figure 3B shows the selected ions for genistein and the genis-tein conjugates. The m/z 153 is a product ion from theisoflavone structure (see Discussion).

For quantification experiments, SIR is preferred as it providesmore sensitivity (Figure 4) than the compared extracted ion fullscan data.

DiscussionCurrent interest in soy isoflavones is based on a vast literaturereporting a wide range of biological properties for genistein anddaidzein8-10 and on clinical studies supporting their potentialhealth benefits11, 12.

Studies using a tandem quadrupole MS have described thatthe isoflavones and their glucoside conjugates have a commonproduct ion other than the [M+H]+ from the isoflavone. In thereaction pathway (Figure 5) this product ion may be assigned to[a+1]: for daidzein, genistein and glycitein the m/z values are137, 153, and 167 respectively.

It is possible to see the [a+1] ion using the SIR method, whichallows a separate higher cone voltage energy to be selected for thethree ions from Table I.

Chromatographic data pointsWhen using mass spectrometry, in particular for quantification,it is important to have at least 10 data points across a peak forrepeatable peak integration. For UPLC-based experiments wherethe peak widths are much smaller than comparable HPLCpeaks, MS acquisition rates have to be faster to achieve this.Figure 6 shows the comparison of the data points when the dwelltime is changed in SIR mode.

For the SIR experiment in this application note, a dwell timeof 10 ms was used to achieve the recommended data points forthe compounds analysed.

SIR 1 (Daidzein) SIR 2 (Genistein) SIR 3 (Glycitein)

m/zCone

voltagem/z

Cone voltage

m/zCone

voltage

137 90 153 90 167 90

255 60 271 70 285 70

417 30 433 25 447 25

459 30 475 35 489 25

503 30 519 35 533 45

Table I. SIR settings showing cone voltages used foreach m/z value.

Figure 2: Spectra for daidzein glucoside (daidzin) and daidzein where the parent ions in positive ESI are 417 and 255, respectively.

JUNE 2007



Figure 3: A: Selected wavelength of 260 nm and full scan MS data with m/z ions for daidzein, genistein and glycitein, extracted from the TIC and B: SIRmethod and the respective ions for genistein in the soy supplement.

Figure 4: Comparison of S/N using SIR data (top) and extracted ionfrom full scan data (bottom).

Ret. Time Compound [M+H]

1.59 Daidzein Glucoside 417

1.70 Glycitein Glucoside 447

2.20 Genistein Glucoside 433

2.69Daidzein Malonyl

Glucoside503

2.86Glycitein Malonyl

Glucoside533

3.23Daidzein Acetyl

Glucoside459

3.41Glycitein Acetyl

Glucoside489

3.98 Daidzein 255

4.00Genistein Acetyl

Glucoside475

4.13 Glycitein 285

3.21Genistein Malonyl

Glucoside519

4.72 Genistein 271

Table II. Retention times for the soy isoflavones andtheir conjugates.

Figure 5: Fragmentation pathway of daidzein.



ConclusionIn this application note, a soy supplement has been used to lookat the soy isoflavone content. With the increasing interest infunctional foods and functional ingredients, it is also importantto analyse for these compounds in the functional food and alsotheir bio-availability in the body.

Here, a 5.5 minute method has been described using UV andMS data. For compound structural information a full scanmethod was used, however, if quantification is required, the SIRmethod is recommended as it provides better sensitivity.

References1) P.A. Murphy, T. Song, G. Buseman, K. Barua, G.R. Beecher, D. Trainer, J.

Holden, J. Agric. Food Chem. 47 (1999) 2697.

2) Lamartiniere, C. A. Protection against breast cancer with genistein, a com-ponent of soy. Am. J. Clin. Nutr. 2000, 71, 1705S-1707S.

3) Santiba´n˜ez, J. F.; Navarro, A.; Martinez, J. Genistein inhibits proliferationand in vitro invasive potential of human prostatic cancer cell lines. Anti-cancer Res. 1997, 17, 1199-1204.

4) Anderson, J. W.; Johnstone, B. M.; Cook-Newell, M. E. Meta analysis ofthe effects of soy protein intake on serum lipids. N. Engl. J. Med. 1995, 333, 276-282.

5) Mezei, O.; Banz, W. J.; Steger, R. W.; Peluso, M. R.; Winters, T. A.; Shay,N. Soy isoflavones exert antidiabetic and hypolipidemic effects throughthe PPAR pathways in obese Zucker rats and murine RAW 264.7 cells. J.Nutr. 2003, 133, 1238-1243.

6) Picherit, C.; Coxam, V.; Bennetau-Pelissero, C.; Kati-Coulibaly, S.;Davicco, M.; Lebecque, P.; Barlet, J. Daidzein is more efficient thangenistein in preventing ovariectomy-induced bone loss in rats. J. Nutr.2000, 130, 1675-1681.

7) US Food and Drug Administration. Food labeling, health claims, soy pro-tein and coronary heart disease. Fed Regist 1999;57:699-733.

8) Setchell KDR, Adlercreutz H. Mammalian lignans and phytoestrogens.Recent studies on their formation, metabolism and biological role inhealth and disease. In: Rowland IR, ed. Role of the gut flora in toxicityand cancer. London: Academic Press, 1988:315-34.

9) Adlercreutz H. Phytoestrogens: epidemiology and a possible role in cancerprotection. Environ Health Perspect 1995;7(suppl):103-12.

10) Kim H, Peterson TG, Barnes S. Mechanisms of action of the soyisoflavone genistein: emerging role for its effects via transforming growthfactor signaling pathways. Am J Clin Nutr 1998;68(suppl): 1418S-25S.

11) Setchell KDR. Phytoestrogens: the biochemistry, physiology, and implica-tions for human health of soy isoflavones. Am J Clin Nutr1998;68(suppl):1333S-46S.

12) Setchell KDR, Cassidy A. Dietary isoflavones: biological effects and rele-vance to human health. J Nutr 1999;129(suppl):758S-67S.

Figure 6: Comparison of dwell times and data points.

JUNE 2007



Application of ACQUITY™ TQD for the Analysis ofPesticide Residues in Baby Food

James Morphet and Peter HancockWaters Corporation, Manchester, UK

This application note assesses the suitability of theWaters® ACQUITY™ TQD for tandem quadrupole-based analysis of pesticide residue in baby food.Polarity switching, differing dwell times and ionratio robustness will also be assessed.

T he European Union residue monitoring program, 2005-2007, establishes the need to cover 55 active ingredients in

various foods, including baby foods1,2. Twenty of these pesti-cides are suitable for multi-residue LC/MS analysis; only one hasa negative polarity in electrospray mode, normally requiring twoinjections (one in each polarity ion mode). Consequently com-pounds with negative polarity are often excluded from monitor-ing programs. Ideally, these should be determined in a singleanalysis with polarity switching.

Chemists analyzing pesticide residues are under increasingpressure to broaden the range of pesticides determined in a sin-gle analysis: to improve limits of detection, precision and quan-titation; to increase confidence in the validity of residue data; toprovide faster methods and to reduce usage of hazardous solventswhile maintaining or reducing costs. In order to meet thesedemanding requirements, the scope, sensitivity, efficiency andspeed of multiresidue methods of analysis must be improved.

The introduction of the ACQUITY TQD featuring the TQdetector (Figure 1) allows scientists to perform UPLC® analysisof pesticides while harnessing all the benefits that this new tan-dem quadrupole instrument brings to the laboratory. WithIntelliStart™ technology, the instrument is designed to removethe burden of complicated operation, time-intensive trou-bleshooting and upkeep. Its small footprint will give any labora-tory an advantage as this powerful tool removes the need forlarger instrumentation.

ExperimentalThe sample extraction method has been previously reported3.Extracts of blank baby food matrix in acetonitrile were providedalong with mixtures of the compounds in acetonitrile by theCentral Science Laboratory (CSL), York, UK. Extracts for injec-tion were prepared by spiking the compounds into the babyfood matrix. The supernatant was analyzed on the ACQUITYTQD following dilution with water (1:9) v/v.


Column: ACQUITY UPLC BEH C18 column2.1 x 50 mm, 1.7 mm

Temperature: 40 °CFlow Rate: 600 mL/minMobile Phase A: WaterMobile Phase B: MethanolGradient: Time 0 min 90% A

Time 4 min 100% BTime 5 min 100% B

Total Run Time: 7 minInjection Volume: 20 mL

MS conditionsMS System: ACQUITY TQ detectorIonization Mode: ESI positive and negativeSwitching Time: 0.02 sCapillary Voltage: 2500 VCone Voltage: 35 VDesolvation Gas: Nitrogen, 800 L/Hr, 400 ˚CCone Gas: Nitrogen, 50 L/HrSource Temperature: 140 °CAcquisition Range: Multiple Reaction Monitoring (MRM)Collision Gas: Argon at 4.0 x 10-3 mBar

The ACQUITY TQD was tuned so that the precursor andproduct ions were resolved with a peak width at half height ofless than 0.7 Da.

The list of pesticide residues and the MRM transitions, alongwith the retention times, dwell times, cone voltages and collisionenergies for the method are shown in Appendix 1. Pesticideresidues listed in red were acquired in negative ion mode.

Figure 1: The Waters ACQUITY TQD with the TQ Detector.



Acquisition and processing methodsWaters MassLynx™ software v4.1 was used for acquisition andits TargetLynx™ application manager used for data processing.

Results and DiscussionAll compounds were separated successfully. Figure 2 shows theTotal Ion Chromatogram (TIC) for all compounds.

The polarity switching capability of the ACQUITY TQD isdemonstrated when comparing two co-eluting compounds,lenacil (negative ion compound) and phorate sulfone (positiveion compound). Figure 3 shows the chromatography achievedwith 12 data points over each peak. Both compounds show goodlinearity and give good correlation coefficients (r2), illustrated inFigure 4, while using the 20 millisecond inter-scan delay forpolarity switching.

The ACQUITY TQD has a greater dynamic range than0.0005 to 0.5 mg/mL, a good practical working range for quan-titation that is equivalent to 3 orders of magnitude, as illustrat-ed in Figure 5.

Here, the range is extended from 0.0005 mg/mL to2.5 mg/mL, equivalent to 4.5 orders of magnitude. This widerrange shows the instrument can perform beyond the usual rangeof quantification.

Figure 2: Positive, negative switching TIC for all pesticides.

Time0.00

0

100

%

1.00 2.00 3.00 4.00

Figure 3: Chromatography achieved by two co-eluting compoundsundertaking polarity switching.

Time1.80

100

100

0

0

%

% Lenacil, -ve

Phorate sulfone, +ve

2.00 2.20

2.32

2.30

2.40 2.60 2.80

1.80 2.00 2.20 2.40 2.60 2.80

Figure 5: Calibration curve for lenacil, 0.0005 to 2.5 mg/mL infruit-based baby food.

µg/mL

µg/mL

0.00 0.50 1.00 1.50 2.00 2.50

Resp

onse

Resi

dual

0.0

-10.0

10000

0

Compound name:

Correlation coefficient:

Calibration curve:

Response type:

Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None

Lenacil

r = 0.999623, r2 = 0.999246

6193.91 * x + - 1.82154

External Std, Area

Figure 6: Robustness of ACQUITY TQD over 100 injections ofthree compounds in fruit-based baby food.

Injection number0 20 40 60 80 100

Peak

are

a / c

once

ntra

tion

14000000

12000000

10000000

8000000

6000000

4000000

2000000

0

Azoxystrobin, 4.41% RSDCarbendazim, 4.25% RSDAcetamiprid, 5.71% RSD

Figure 4: Calibration curves for lenacil and phorate sulfone infruit-based baby food while polarity switching is ongoing.

µg/mL0.000

Resp

onse

Compound name:


Calibration curve:

Response type:


Phorate sulfone

r = 0.997579, r2 = 0.995165

192494 * x + 58.2215

External Std, Area

80000

60000

40000

20000

00.100 0.200 0.300 0.400 0.500

µg/mL0.000 0.100 0.200 0.300 0.400 0.500

Resp

onse

3000

2000

1000

0

Compound name:


Calibration curve:

Response type:


Lenacil

r = 0.999385, r2 = 0.998771

6195.2 * x + - 1.79023

External Std, Area

JUNE 2007



The ACQUITY TQD was tested over a 100-injectionsequence to assess its robustness to matrix. This 12-hour batchanalysis contained a series of five matrix-matched standards,

between 0.005 and 0.250 mg/mL, delivering approximately200 mg of matrix. The results for three compounds are illustrat-ed in Figure 6, where the peak area/concentration ratio is plot-ted against injection number.

All three compounds display excellent percent relative stan-dard deviations (%RSD) over the injection sequence indicatinggood instrument robustness.

The ACQUITY TQD's travelling wave collision cell was test-ed for use at short dwell times for the multi residue method.Figure 7 shows methiocarb in fruit-based baby food at a concen-tration of 0.05 mg/mL. The sample was injected seven times insuccession using the same MRM transition, at differing dwelltimes from 5 to 500 milliseconds. The data shows that the sig-nal intensity remains consistent over the range of dwell times. Asthe dwell time decreases the number of points per peak increas-es from 8 points for the dwell time of 500 milliseconds to 220points for the shortest dwell time of 5 milliseconds.

TargetLynx was used to provide automatic quantification andconfirmation with two MRM transitions processed for eachresidue. The browser produced by TargetLynx for thiabendazoleat a spiked concentration of 0.01 mg/mL in fruit-based babyfood is illustrated in Figure 8.

All residues could be screened and confirmed to a concentra-tion of 0.01 mg/mL.

Two MRM transitions were monitored for each compound.The primary transition is used for quantification and the sec-ondary transition is used for confirmation purposes.

Confirmation is achieved by calculating the ion ratio betweenthe primary and secondary transition. All other injections musthave a ratio that lies within 20 percent of the standard for themto be positively confirmed by this technique. Figure 9 shows theion ratios of three example compounds at concentrationsbetween 0.005 and 0.250 mg/mL in baby food over 100 injec-tions. All injections lie within the ±20% boundary required forconfirmation with the average difference shown in the header ofthe graph.

ConclusionA fast and simple UPLC method involving polarity switching ofMRM transitions has been successfully transferred to theACQUITY TQD for the determination of 52 pesticides. Ofthese, 21 pesticides and 7 metabolites are included in the EUresidue monitoring program, 2005-2007.

The ACQUITY TQD was capable of very fast polarityswitching, allowing the analysis of positive and negative com-pounds in a single injection.

The use of very short dwell times of 5 ms was found to haveno effect on signal intensity, indicating that sensitivity can bemaintained as the number of residues is increased.

Robustness of the ACQUITY TQD was proven with a 100-injection sequence after delivery of approximately 200 mg ofmatrix to the instrument.

Confirmation was achieved using a secondary MRM transi-tion over a 100-injection sequence, where the variance was lessthan 20 percent in all cases.

Figure 7: Methiocarb 0.05 mg/mL on the ACQUITY TQD, show-ing that dwell time does not affect signal intensity.

Time2.20

1005 10 20 50 100 200 500 ms

%

02.40 2.60 2.80 3.00

Figure 8: Example TargetLynx browser for thiabendazole infruit-based baby food (0.01 mg/mL).

Figure 9: Plot showing confirmatory ion ratio over 100 injections.

Injection number

Methamidophos, 5.5%

Flufenacet, 4.1%

Butocarboxim Sulfoxide, 5.4%

Ion

rati

o

0

1.4

1.2

1.0

0.8

0.6

0.4

0.2

0.010 20 30 40 50 60 70 80 90 100



Methamidophos 0.43

Acephate 0.53

Omethoate 0.6

Butocarboxim sulfoxide

0.62

Aldicarb sulfoxide 0.68

Butoxycarboxim 0.74

Aldicarb sulfone 0.76

Methomyl 0.88

Oxydemeton-methyl 0.9

Pymetrozine 0.93

Demeton-S-methylsulfone

0.95

Imidacloprid 1.2

Carbendazim 1.5

Methiocarb sulfoxide

1.32

Dimethoate 1.33

Acetamiprid 1.38

Cymoxanil 1.47

Methiocarb sulfone 1.47

Thiacloprid 1.55

Butocarboxim 1.66

Aldicarb 1.69

Thiabendazole 1.74

Carbaryl 2.1

Thiodicarb 2.19

Phorate sulfoxide 2.24

Phorate sulfone 2.28

Lenacil (-ve) 2.31

Azinphos-methyl 2.45

142>94142>125184>143184>125214>183214>155207>75207>132207>89207>132223>106223>166223>86223>76163>88163>106247>169247>109218>105218>79263>169263>121256>209256>175192>160192>132242>185242>168230>125230>171223>126223>56199>128199>111258>122258>107253>126253>90213>75213>156208>116208>89202>175202>131202>145202>127355>88355>108277>97277>143293>97293>115233>151233>107318>160318>261

0.015 28

0.015 22

0.015 26

0.015 23

0.015 22

0.015 23

0.015 29

0.01 21

0.01 26

0.015 31

0.01 32

0.01 28

0.01 27

0.01 28

0.01 23

0.01 33

0.01 23

0.01 28

0.01 34

0.01 30

0.01 13

0.01 46

0.02 24

0.02 21

0.01 24

0.01 24

0.03 50

0.01 20

141381812151261410107127810142817361717162018301324201520158182037203715107725321028161632203024243288

Pesticide RT MRM Transitions Dwell time(s) Cone Voltage (V) Collision Energy (eV)

Appendix 1. ACQUITY TQD MRM parameters

JUNE 2007



AcknowledgementsThe authors would like to thank Richard Fussell at the CentralScience Laboratory (CSL), Sand Hutton, York, UK for kindlysupplying the sample extracts that were analyzed in this project.

References1) AGROW Magazine 468 (2005) 9.

2) C.C. Leandro, P. Hancock, R.J. Fussell, B.J. Keely, J. Chromatogr. A 1144(2007) 161-169. doi: 10.1016/j.chroma.2007.01.030.

3) Determination of Pesticides in Food using UPLC with Polarity SwitchingTandem Quadrupole LC/MS/MS, Waters Application Note720001995EN.

Linuron 2.56

Azoxystrobin 2.58

Methiocarb 2.61

Fludioxinil (-ve) 2.65

Iprovalicarb 2.84

Triadimenol 2.85

Dichlofluanid 2.86

Fenhexamid 2.87

Fenoxycarb 2.87

Flufenacet 2.89

Diflubenzuron (-ve) 2.98

Cyprodinil 3.08

Tolyfluanid 3.09

Zoxamide 3.13

Imazalil 3.15

Phorate 3.19

Hexaflumuron (-ve) 3.37

Fluazinam (-ve) 3.5

Teflubenzuron (-ve) 3.54

Lufenuron (-ve) 3.56

Flucycloxuron (-ve) 3.64

Flufenoxuron (-ve) 3.67

0.01 34

0.01 28

0.01 22

0.03 51

0.01 21

0.01 14

0.01 22

0.01 35

0.01 21

0.01 17

0.03 20

0.01 45

0.01 19

0.01 25

0.01 36

0.01 11

0.02 22

0.02 26

0.02 18

0.02 22

0.02 34

0.02 27

249>160249>182404>372404>329226>169226>121247>180247>126321>119321>203296>70296>99333>123333>224302>97302>55302>88302>116364>152364>194309>156309>289226>93226>108347>137347>238336>187336>159297>159297>69261>75261>97459>276459>175463>416463>398379>196379>339509>175509>326482>156482>462487>156487>329

16151530101928351881016241025352012201011933252810244120201232223921172515402214131622

Pesticide RT MRM Transitions Dwell time(s) Cone Voltage (V) Collision Energy (eV)

Appendix 1. ACQUITY TQD MRM parameters (continued)


UPLC®/MS/MS Bioanalytical MethodValidation of Acebutolol and PindololUsing an Analogue Internal Standard

Ed Sprake and Iain GibbWaters Corporation, Manchester, UK

B eta-blockers are a common class of drugs used to treat con-ditions such as high blood pressure, tachycardia and car-

diac arrhythmia. In this application note, we show the partialvalidation of a bioanalytical method for Acebutolol and Pindololin human plasma using Nadolol as an analogue internal standard(Figure 1). The validation was carried out according to theguidelines in the FDA Guidance for Industry on BioanalyticalMethod Validation.

Through this experiment, we aim to show that the Waters®

ACQUITY UltraPerformance LC® System combined with theQuattro Premier™ XE Mass Spectrometer (UPLC®/MS/MS)operating in MRM mode is an accurate, precise, and robusttechnique which will also yield the benefits of greater speed, sen-sitivity and resolution over HPLC/MS/MS.

ExperimentalDuring this experiment we performed a comparison betweenHPLC and UPLC using a protein precipitation (PPT) samplepreparation method.

Protein Precipitation Method1. 200 µL plasma was spiked with:

• 50 mL IS (1.0 mg/mL in water)• 50 mL spike solution

(from 0.8 ng/mL - 600 ng/mL in water)• When the IS and/or spike solution was not required,

the appropriate volume of water was added 2. 600 mL acetonitrile was added to crash proteins3. Centrifuged at 13,000 rpm for 5 minutes4. 200 mL of supernatant diluted with 800 mL water prior to

injection

Standard curves and QC samples were prepared as described andshown in Table I. Three separately prepared validation batches wereprepared by protein precipitation and run using UPLC/MS/MS. Astandard curve prepared by protein precipitation in human plasmawas run using HPLC/MS/MS for comparison.

A validation batch consisted of the following:• 2 separately prepared calibration curves• 6 individually prepared replicates of each QC

concentration point• A blank and double blank before each curve• 2 carryover blanks after each curve

HPLC conditionsLC System: Waters Alliance® HT SystemColumn: XBridge™ C18

2.1 x 50 mm, 3.5 µm Eluents: A: 2mM ammonium acetate + 0.1% formic

acid in water B: 0.1% formic acid in acetonitrile

Column Temp: 40 ºCSample Temp: 4 ºCFlow Rate: 0.3 mL/minGradient: Time %A %B Curve

0.0 85 15 -1.6 5 95 82.0 85 15 11

Run Time: 3.2 minInjection Volume: 20 mLPressure: 1800 psi


Column: ACQUITY UPLC BEH C182.1 x 50 mm, 1.7 mm

Eluents: A: 2 mM ammonium acetate + 0.1% formic acid in water B: 0.1% formic acid in acetonitrile

Column Temp: 40 ºC


Table I. Spike concentrations and their equivalentconcentrations in human plasma

Sample Type Spike Conc. (ng/mL)Actual Conc. inPlasma (ng/mL)

Standard 0.8 0.2

2 0.5

4 1

20 5

40 10

200 50

320 80

400 100

600 150

QC 0.8 0.2

3 0.75

80 20

300 75

360 90

600 150

JUNE 2007



Table II. Inter-batch statistics for pindolol - 9 calibration standard concentrations over 3 days by UPLC/MS/MS

Conc. Of Pindolol(ng/mL)

Batch A Batch B Batch C Mean SD CV (%) Accuracy (%)

0.2 0.199 0.197 0.198 0.20 0.00 1.16 100.0

0.203 0.203 0.200

0.5 0.473 0.467 0.491 0.50 0.03 6.27 100.3

0.498 0.541 0.537

1 0.967 0.899 0.946 0.99 0.07 7.15 99.1

1.077 1.072 0.982

5 4.893 4.806 4.852 5.05 0.26 5.05 101.1

5.429 5.049 5.294

10 10.134 9.649 9.868 10.17 0.36 3.53 101.7

10.611 10.309 10.429

50 50.858 50.960 52.674 51.84 0.92 1.77 103.7

51.224 52.640 52.708

80 78.223 80.932 77.135 80.03 2.02 2.53 100.0

79.895 82.281 81.736

100 97.434 99.749 102.057 100.59 2.15 2.14 100.6

99.123 102.674 102.530

150 135.588 141.371 134.605 140.38 4.80 3.42 93.6

146.390 144.943 139.356

Gradient 0.031 0.030 0.031 0.031 0.001 2.68 N/A

Correlation 0.997 0.997 0.997

Intercept 0.0001 0.0009 0.0006

Table III. Inter-batch statistics for acebutolol - 9 calibration standard concentrations over 3 days by UPLC/MS/MS

Conc. Of Pindolol(ng/mL)

Batch A Batch B Batch C Mean SD CV (%) Accuracy (%)

0.2 0.189 0.196 0.200 0.20 0.01 4.21 100.4

0.200 0.206 0.214

0.5 0.524 0.496 0.469 0.50 0.03 5.82 100.9

0.545 0.516 0.476

1 0.935 0.885 0.885 0.96 0.08 7.81 96.1

1.071 1.022 0.971

5 4.687 4.971 5.144 5.01 0.19 3.83 100.2

4.939 5.085 5.233

10 9.642 9.912 10.028 9.95 0.22 2.18 99.5

9.783 10.211 10.144

50 51.553 49.417 52.168 51.69 1.50 2.90 103.4

52.211 50.877 53.900

80 81.351 80.481 77.581 80.81 1.70 2.10 101.0

82.154 81.233 82.070

100 100.269 98.684 103.105 102.02 2.64 2.59 102.0

100.689 103.338 106.025

150 139.407 147.795 140.968 144.63 4.51 3.12 96.4

147.382 150.527 141.673

Gradient 0.028 0.027 0.028 0.028 0.0006 2.14 N/A

Correlation 0.997 0.999 0.996

Intercept 0.0003 0.00051 0.0002



UPLC conditions (continued)Sample Temp: 4 ºCFlow Rate: 0.6 mL/minGradient: Time %A %B Curve

0.0 85 15 -0.8 5 95 81.0 85 15 11

Run Time: 1.6 minInjection Volume: 20 mLPressure: 10500 psi

Table IV. Intra- and inter-batch QC statistics for pindolol by UPLC/MS/MS

Intra-Batch Inter-Batch

Conc. Of Pindolol (ng/mL) Batch A n=6 Batch B n=6 Batch C n=6 n=18

0.2 Mean 0.21 0.20 0.19 0.20

SD 0.03 0.01 0.03 0.02

CV (%) 12.21 6.90 14.63 11.53

Accuracy (%) 102.8 99.2 94.1 98.7

0.75 Mean 0.77 0.77 0.73 0.76

SD 0.06 0.05 0.07 0.06

CV (%) 7.36 6.25 9.52 7.82

Accuracy (%) 102.3 103.3 97.0 100.9

20 Mean 19.6 20.6 19.8 20.0

SD 0.80 0.85 0.81 0.88

CV (%) 4.08 4.14 4.07 4.42

Accuracy (%) 98.2 103.0 99.0 100.0

75 Mean 75.5 78.6 76.7 76.9

SD 2.22 2.53 2.18 2.54

CV (%) 2.94 3.22 2.85 3.30

Accuracy (%) 100.6 104.8 102.3 102.6

90 Mean 85.1 88.4 86.6 86.6

SD 1.64 2.75 3.37 2.80

CV (%) 1.93 3.11 3.89 3.24

Accuracy (%) 94.6 98.2 96.2 96.2

150 Mean 140.8 150.7 147.3 146.3

SD 2.53 3.63 3.18 5.16

CV (%) 1.80 2.41 2.16 3.53

Accuracy (%) 93.9 107.7 101.0 97.5

Figure 2: Curve 8 gradient profile.

15.00

95.00

B Composition

Figure 1: Chemical structures of acebutolol, pindolol, andnadolol.

JUNE 200724 ULTRAPERFORMANCE LC®

MS conditionsMS System: Quattro Premier XE tandem quadrupole

mass spectrometerIon Mode: Electrospray positiveCapillary Voltage: 3.00 kVSource Temp: 120 ºCDesolvation Temp: 380 ºCDesolvation Gas: 1000 L/hourCone Gas Flow: 50 L/hourDwell Time: 0.02 secondsInter-scan Delay: 0.01 secondsCollision Gas: Argon (3.45 x 10-3 mbar)Detection Mode: MRM (see below)Compound Transition Cone Voltage(V) Collision Energy(eV)

Acebutolol 337.25>116.00 35 22

Pindolol 249.15>116.00 35 18

Nadolol (IS) 310.30>201.20 25 20

The “Curve” setting in the aforementioned gradient tablesrefers to the gradient profile; adjusting the method to a non-lin-ear curve setting can help separate close running peaks undersome circumstances. A graphical representation of the gradientused for this analysis is shown in Figure 2.

Table V. Intra- and inter-batch QC statistics for acebutolol by UPLC/MS/MSIntra-Batch Inter-Batch

Conc. Of Acebutolol (ng/mL) Batch A n=6 Batch B n=6 Batch C n=6 n=18

0.2 Mean 0.21 0.19 0.19 0.20

SD 0.02 0.02 0.02 0.02

CV (%) 9.02 9.84 10.60 10.98

Accuracy (%) 105.9 93.2 94.5 97.9

0.75 Mean 0.76 0.75 0.76 0.76

SD 0.05 0.05 0.06 0.05

CV (%) 6.05 6.97 7.57 6.52

Accuracy (%) 101.7 99.9 101.2 100.9

20 Mean 19.3 20.3 19.9 19.8

SD 0.85 1.05 0.46 0.87

CV (%) 4.42 5.19 2.30 4.41

Accuracy (%) 96.5 101.3 99.6 99.1

75 Mean 76.8 80.2 78.1 78.3

SD 2.66 3.33 2.59 3.06

CV (%) 3.47 4.16 3.32 3.91

Accuracy (%) 102.4 106.9 104.1 104.4

90 Mean 87.3 92.1 90.9 89.9

SD 2.44 1.90 3.04 3.18

CV (%) 2.80 2.07 3.35 3.53

Accuracy (%) 97.0 102.3 101.0 99.9

150 Mean 148.54 154.47 154.67 152.56

SD 5.95 3.91 4.59 5.45

CV (%) 4.00 2.53 2.97 3.57

Accuracy (%) 99.0 112.0 101.8 101.8

Figure 3: Typical calibration curves for pindolol and acebutololin protein precipitated human plasma by UPLC/MS/MS.

0.0

-10.0

Resp

onse

Compound name: AcebutololCorrelation coefficient: r = 0.999283, r^2 = 0.998567Calibration curve: 0.0273181 * x + 0.000481258Response type: Internal Std (Ref 1), Area * (IS Conc. / IS Area)Curve type: Linear, Origin: Exclude, Weighting: 1/x^2, Axis trans: None

Compound name: PindololCorrelation coefficient: r = 0.998515, r^2 = 0.997033Calibration curve: 0.0298598 * x + 0.00094694Response type: Internal Std (Ref 1), Area * (IS Conc. / IS Area)Curve type: Linear, Origin: Exclude, Weighting: 1/x^2, Axis trans: None

Resp

onse

Resi

dual

Resi

dual

0.0

-5.0

-10.0

4.00

2.00

0.00

4.00

2.00

0.00



Results and DiscussionAll of the calibration standards run by UPLC/MS/MS gener-ated calibration curves with a coefficient of calibration (R2)greater than 0.996. The HPLC/MS/MS run generated cali-bration curves where R2 was greater than 0.997. Typical exam-ples of calibration curves for pindolol and acebutolol (usingUPLC/MS/MS) are shown in Figure 3.

Inter-batch calibration statistics are shown in Tables II and III.The statistics for the standard injections are based on 2 replicateinjections of the 9 calibration points for each of the 3 inter-daybatches. All calibration points show <8% CV with accuracy val-ues between 93.6% - 103.7% for both pindolol and acebutolol.

Statistics for the QC injections, shown in Tables IV and V, arebased on single injections of 6 individually spiked QC solutions ateach concentration, for each of the 3 inter-day batches. Both pin-dolol and acebutolol show <15% CV for the lower limit of quan-titation (LLOQ) with <10% CV for the remainder of the qualitycontrol standards. Inter-batch accuracy values were observedbetween 93.2% - 111.99% for both pindolol and acebutolol.

FDA guidelines recommend that samples at the LLOQshould have less than 20% CV and deviation from the standardcurve. All other unknowns, calibration standards, and QC stan-dards should be within 15%, accuracy values should be within80 - 120% at LLOQ, and 85 - 115% for other standards.

All of the results generated during the validation of this methodcomply with and exceed the guidelines set forth by the FDA.

HPLC vs. UPLCIn Figure 4, we can see that we get a 3.8-fold increase in signal-to-noise by using UPLC vs. HPLC methodology. As well asincreases in signal-to-noise and limit of detection, there is also anincrease in resolution, giving a better chance of separating theanalyte from endogenous peaks. A 2-fold decrease in run timewas also observed, meaning that a validation batch was run inonly 2 hours by UPLC compared to 4 hours when run byHPLC. An example of both an HPLC and a UPLC chro-matogram are shown below for comparison.

ConclusionsWe have successfully produced a validated UPLC/MS/MSmethod for the analysis of Pindolol and Acebutolol in humanplasma over the range of 0.2 - 150 ng/mL. Statistics for accura-cy and precision were within the FDA guidelines for bioanalyti-cal method validation. The data generated by UPLC/MS/MSwere comparable to that generated by HPLC/MS/MS, however,it was shown that by using UPLC, a 4-fold increase in signal-to-noise ratio for the LLOQ, a 2-fold decrease in run time, and anincrease in resolution was achieved. This equates to doubling thethroughput of this method, as well as enabling the acquisition ofmeaningful data for lower sample concentrations. This has sev-eral benefits, for example, as it would allow more accurate meas-urement of the lower part of the PK curve.

Figure 4: Signal-to-noise comparison using the 1 ng/mL calibra-tion standard, HPLC vs. UPLC.

0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60

Time

1.80 2.00 2.20 2.40 2.60 2.80 3.00

2.962.292.181.901.851.731.641.571.43

1.33

1.30

1.26

S/N RMS=112.36

1.051.030.920.23

HPLC

UPLC

0.44 0.54 0.63 0.720.75

3.20

0.700.65

STD1_3PPT_ValidationbatchA_005

1ng/ml StandardXBridgeCurve_0005

100

0

%

100

0

%

0.75 0.80 0.85 0.90

0.94

S/N RMS=425.51MRM of 3 Channels ES+

337.25 > 1168.31e4

MRM of 3 Channels ES+337.3 > 115.8

2.93e3

0.95 1.00 1.05 1.10

Time

1.15 1.20 1.25 1.30 1.35 1.40 1.45 1.50 1.55

Figure 5: Chromatographic comparison, HPLC vs. UPLC.

0.20 0.40 0.60

0.80

0.87

0.98PPT_ValidationbatchA_011 MRM of 3 Channels ES+

TIC1.31e7

150ng/Ml StandardXBridgeCurve_011

MRM of 3 Channels ES+TIC

7.68e5

0.94

HPLC

UPLC

1.06

1.25

0.80 1.00 1.20 1.40 1.60Time

1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20

0.20

100

%

0

100

%

0

0.40 0.60 0.80 1.00 1.20 1.40 1.60Time

1.80 2.00 2.20 2.40 2.60 2.80 3.00 3.20


JUNE 2007



Peptide Separation Technology:Quantitative Aspects of UPLC® Peptide Mapping

Thomas E. Wheat, Ziling Lu, Beth Gillece-Castro, and Jeffrey R. MazzeoWaters Corporation, Milford, MA

T hroughout the development of a biopharmaceutical pro-tein, peptide mapping is used to demonstrate genetic

stability and to confirm the integrity of the protein. Changesin retention time, often in combination with MS or MS/MSdetection, reveal changes in the primary structure of the pro-tein. Modifications such as oxidation, deamidation, dele-tions, sequence clips and glycosylation all affect chromato-graphic behavior. The modified peptides must be separatedfrom the native peptides for a peptide mapping method to beuseful. The presence of a modified peptide in the map of asample reflects the presence of modified protein in the origi-nal sample. In the initial characterization of a protein, it isimportant to develop a peptide mapping method thatresolves modified peptides from native peptides so that allpossible modifications may be detected. As development ofthe biopharmaceutical advances, these peptides must bequantitated. Quantitation is generally expressed as area orheight percent of the native peptide. In this way, the peptidemap can provide information on the mixture of protein formsin each sample so that the safety and efficacy of the prepara-tion may be assured. Methods must, therefore, exhibit excel-lent sensitivity and linearity for quantitative work.

UltraPerformance Liquid Chromatography (UPLC®) hasdemonstrated significant advantages compared to HPLC forpeptide mapping. UPLC gives increased resolution, highersensitivity, excellent peak shapes for glycopeptides, and thepotential to increase throughput1,2. In this application note,we focus on the quantitative aspects of UPLC peptide map-ping with UV detection. The technique is evaluated withrespect to both chromatographic and detection linearity sincethe altered and normal peptides occur at extreme molarratios. Reproducibility of the area measurement at theseextreme ratios is also examined. Results from a mixture ofpeptide standards and from a digest spiked with an amountof peptide are shown.

Materials and MethodsSamples: Waters MassPREP™ Peptide Standard

MixtureWaters MassPREP™ Hemoglobin tryptic digest

Sample Buffer: 95% Buffer A/5% Buffer B

LC System: ACQUITY UPLC®

UPLC Column: Waters Peptide Separation TechnologyACQUITY UPLC BEH 130 C18, 1.7 mm, 2.1 x 100 mm

Flow Rate: 0.2 mL/minuteMobile Phase: A: 0.1% TFA in Milli-Q® water

B: 0.08% TFA in acetonitrileGradient: 0-50% B in 29 minutes (peptide mixture)

0-50% B in 58 minutes (hemoglobin digest)Temperature: 40 °CDetection: UV at 214 nm with 10 mm path length cell

at 10 HzQuantitation: QuanLynx™ application manager within

MassLynx™ v4.1Peptide ID: LCT Premier™ oa-Tof mass

spectrometerIonization Mode: Electrospray, positiveScan: 400 to 1800 m/z at a rate of 2 scans per

second

Results and DiscussionA mixture of peptides, 20 pmoles of each, was analyzed on aPeptide Separation Technology ACQUITY UPLC BEH 130,C18, 1.7 mm, 2.1 x 100 mm column. The separation was mon-itored at 214 nm. Six replicate injections are overlaid in Figure1, demonstrating the reproducibility of UPLC for peptide map-ping. The same sample was subsequently injected at differentlevels to test linearity and sensitivity.

Figure 2 shows the analysis of peptide standards from 250fmol to 100 pmol injected on-column. There is no significantshift in retention or deterioration in peak shape from low tohigh levels. This confirms that the dynamic range of the chro-matographic material and gradient method is sufficient for theanalysis of small amounts of one peptide in the presence of

Figure 1: Six replicate 20 pmol injections of peptide mixture stan-dard were made on a Peptide Separation Technology ACQUITY UPLCBEH 130 C18 1.7 mm, 2.1 x 100 mm column. Retention time, peakheight, and peak area reproducibility were measured for a peptidemapping gradient.

Time

8.00 9.00 10.00 11.00 12.00 13.00 14.00 15.00 16.00 17.00 18.00 19.00

AU

2.0e-2

3.0e-2

4.0e-2

5.0e-2

6.0e-2

7.0e-2

8.0e-2

17.04

13.51

8.51

16.04

18.55



much larger amounts of another. The lower limit is shown inFigure 3 where the six lowest levels are overlaid. The signal-to-noise ratio should be sufficient for quantitation at 250 fmol,and this is consistent with the area reproducibility summarizedin Table I. Peak area response is linear over two-and-a-halforders of magnitude, as shown in Figure 4. Given a typicalinjected amount of 100 pmol of digested protein, these quan-tities would correspond to 0.25% to 10% of a modified pep-tide. These results on standard peptides indicate that UPLCpeptide mapping can be used to quantitate modified peptidesin digests over a wide range of concentration.

A known amount of a specific peptide was added to an actu-al protein digest to test estimates of quantitation. This peptideserves as a surrogate illustrating the behavior of modified pep-tides in the digest. A tryptic digest of hemoglobin (200pmoles) was spiked with a peptide corresponding to a concen-tration range of 0.2 to 2% of the native material. The resultingchromatograms are shown in Figure 5. The whole digest elutesbetween 7 and 53 minutes. Figure 5 focuses on the segment ofquantitative interest and magnifies the elution profile of thesurrogate peptide. The surrogate peptide, marked with a * inFigure 5, elutes just before minute 29. The peptide can be

Figure 2: The peptide mixture was analyzed from 250 fmol to 100 pmol on-column.

Time

8.00 10.00 12.00 14.00 16.00 18.00

AU

2.5e-2

5.0e-2

7.5e-2

1.0e-1

1.25e-1

1.5e-1

1.75e-1

2.0e-1

2.25e-1

2.5e-1

2.75e-1

0.25

1.00.5

105.02.0

5020

100

pmol

Figure 3: Chromatograms of one peptide from 250 fmol to 10 pmol show the high sensitivity region of the standard curve.

Time8.00 8.10 8.20 8.30 8.40 8.50 8.60 8.70 8.80 8.90 9.00

AU

1.4e-2

1.6e-2

1.8e-2

2.0e-2

2.2e-2

2.4e-2

2.6e-2

0.25

1.00.5

105.02.0

pmol

JUNE 2007



detected easily at the 0.2% level. Chromatographic resolutionis maintained at this low level relative to the significantly larg-er peak that elutes 0.25 minutes earlier.

ConclusionsUPLC yields qualitatively and quantitatively reproducible pep-tide maps. Chromatographic resolution and peak shape are con-stant over nearly three orders of magnitude of sample amounton-column. Peak areas may be reliably and reproducibly quanti-tated to sub-picomole levels, as low as 250 fmoles on-column.This level of sensitivity is obtained with UV detection on2.1 mm columns without resorting to special detection strate-gies. The detector response is linear with sample amount. Thissensitivity in combination with robust chromatographic behav-ior enables detection of low-level peptides in a complex digest.

In turn, the technique can be used to measure the specific vari-ants of protein structure that may be observed in the develop-ment of a biopharmaceutical protein.

References1) “Enabling Significant Improvements in Peptide Mapping with UPLC,” J.R.

Mazzeo, T.E. Wheat, B.L. Gillece-Castro, Z. Lu, Waters ApplicationNote 720001339EN.

2) “Next Generation Peptide Mapping with UPLC,” J.R. Mazzeo, T.E. Wheat,B.L. Gillece-Castro, Z. Lu, BioPharm International, January 1, 2006.

Peptide Load(pmoles)

Mean Area

R.S.D. Area

% R.S.D.Area

0.25 18.98 0.70 3.7

0.50 38.35 1.13 2.9

1.00 74.48 1.10 1.5

2.00 151.39 0.79 0.5

5.00 392.83 2.50 0.6

10.00 820.85 12.18 1.5

20.00 1933.30 6.55 0.3

50.00 4368.58 26.18 0.5

100.00 9051.69 117.65 1.3

Table I. Reproductibility of peak area for the peptideeluting at 8.5 minutes. Triplicate injections from 250fmol to 100 pmol were analyzed.

Figure 4: Linearity of UV detector response is shown for triplicateinjections from 250 fmol to 100 pmol of the peptide standard mix-ture. This peptide eluted at 8.5 minutes, as shown in Figures 2 and 3and Table I.

Conc0 10 20 30 40 50 60 70 80 90 100

Resp

onse

0

2000

4000

6000

8000

10000

12000

14000

16000

18000

Figure 5: A tryptic digest of bovine hemoglobin was spiked with thepeptide mixture as a surrogate to represent modified peptides. Thesurrogate peptide concentrations were 0.2 to 2% of the digest con-centration on a molar basis. Panel a) shows the peptide map ofbovine hemoglobin with peptide standards spiked at 2% molar basis.A total injection on-column of 200 pmol of hemoglobin digest wasmade. Panel b) focuses on the elution position of the surrogate.

Time

10.00 12.00 14.00 16.00 18.00 20.00 22.00 24.00 26.00 28.00 30.00 32.00 34.00 36.00

AU

5.0e-2

1.0e-1

1.5e-1

2.0e-1

22.39

15.54

9.7518.00

34.17

29.65

28.60

23.86

24.7127.94

32.84

*

*

Time

b)

a)

28.30 28.40 28.50 28.60 28.70 28.80 28.90 29.00

AU

0.0

1.0e-2

2.0e-2

3.0e-2

4.0e-2

5.0e-2

6.0e-2

7.0e-2

8.0e-2

9.0e-2

1.0e-1

1.1e-1

1.2e-1

1.3e-1

1.4e-1

1.5e-1

1.6e-1

1.7e-1

1.8e-1

1.9e-1



LC/MS-based Differential Proteomics of the Mitochondria of[PSI+] and [psi-] Saccharomyces cerevisiae Strains

Jacek Sikora and Michael DadlezMass Spectrometry Laboratory and Department of Biophysics

Chris Hughes, Hans Vissers, Thérèse McKenna, Jim LangridgeWaters Corporation, Manchester, United Kingdom

Magdalena BogutaDepartment of Genetics, IBB PAS, Warsaw, Poland

P roteomics focuses on the high throughput study of theexpression, structure, interactions, and, to some extent,

function of complex sets of proteins. Differential proteomicsaims at finding differences between two or more multi-proteinsamples, which is imperative for the understanding of many bio-logical problems. [PSI+] is a protein-based heritable phenotype ofthe yeast Saccharomyces cerevisiae, which reflects the prion-likebehavior of the endogenous Sup35 protein release factor.Previous work has shown that the presence of a prion form of thisprotein in the cytosol can cause respiratory deficiency by decreas-ing level of mitochondrially-encoded Cox2 protein1. The maingoal of the presented work is to identify proteins, which are pres-ent at different levels in the mitochondrial fractions of [PSI+] and[psi-] yeast strains. The latter should allow for the identificationof the molecular mechanism of prion-dependent switchingbetween respiratory competence and deficiency.

In this study a label-free LC/MS-based approach was usedwhere data is acquired in an alternating fashion, with low collisionenergy on the gas cell in the first function, switching to elevatedenergy in the alternate scan. In neither scan is a precursor ion iso-lated with the quadrupole, thus providing a parallel approach toion detection and sequencing. The low energy portion of theobtained data sets is typically used for quantification of the pro-teins, whereas the combined low and elevated energy informationare utilized for qualitative, identification purposes. Resultsobtained from yeast mitochondrial fractions allowed differentia-tion of proteins originating from [PSI+] vs. [psi-] strains, leadingto the identification of a significant decrease of Phb1 and Phb2(prohibitins) in mitochondria of the [PSI+] strain. The obtainedresults were confirmed by Western blotting experiments.

ExperimentalSample preparationFractionation of yeast cells grown at 30 ºC on glycerol mediumand purification of mitochondria was performed as previouslydescribed1. Total protein fractions for Western blotting analysiswere prepared using standard NaOH-TCA precipitation.

The mitochondrial fractions were re-solubilized in 50 mMammonium bicarbonate/0.1% RapiGest™ SF solution. The pro-teins were reduced (10 mM DTT) and alkylated (10 mM IAA)prior to enzymatic overnight digestion with trypsin -1:50 (w/w)enzyme: protein ratio at 37 ºC. RapiGest was removed by theaddition of 2 mL conc. HCl, followed by centrifugation, and thesupernatant collected. Samples were diluted with 0.1% formicacid to an appropriate final working concentration prior to analy-sis - corresponding to a 0.3 mg protein digest on-column load.

LC conditions LC/MS quantification experiments were conducted using eithera 1.5 or 3 hr RP gradient at 250 nL/min (5 to 40% acetonitrileover 90 or 180 minutes) on a nanoACQUITY UPLC® Systemutilizing a 1.7 mm BEH C18 NanoEase™ 75 mm x 20 cm col-umn. Each sample was run in triplicate.

MS conditionsThe Q-Tof Premier™ mass spectrometer was programmed toalternate between normal (5 eV) and elevated (25-40 eV) colli-sion energies on the gas cell, using a scan time of 1.5 s per func-tion over 50-1990 m/z (Figure 2).

Data processing and protein identificationData alignment, protein identifications and quantitative analysiswere conducted with the use of dedicated algorithms (WatersProtein Expression Informatics) and searching yeast-specificdatabases. 3D UPLC/MSE visualizations were created with in-house development software.

Western blottingFor immunoblotting, proteins were separated by SDS-PAGE gelsand transferred electrophoretically to nitrocellulose. After incuba-tion with antibody, visualization was made with anti-rabbit oranti-mouse peroxidase-conjugated antibody. Autoradiograms werequantified using ImageQuant (GE Healthcare Life Sciences) withlocal average background correction.

Figure 1: UPLC/MSE alternating scanning principle based onnanoACQUITY UPLC and Q-Tof Premier and subsequent align-ment of the low and elevated energy ions with Protein Expres-sion Informatics software.

JUNE 2007



ResultsAlignmentAs illustrated in Figure 1, alignment of the low energy precursorions and elevated energy fragment ions is performed, to associ-ate them for identification purposes. However, for quantitativepurposes, alignment is also conducted across injections. For a 2-condition comparison this translates into the alignment of atleast 4 to 6 chromatographic data sets - assuming duplicate ortriplicate injections, respectively. Hence, both mass accuracy andchromatographic reproducibility are imperative. The latter isdemonstrated by the results shown in Figure 2, where expandedportions of the low energy BPI chromatogram is shown for threereplicate injections for one of the investigated conditions in thisstudy. A more comprehensive view of the low and elevated ener-gy chromatograms is shown in Figures 3a and 3b. Figure 3ashows the 3D visualization of a low energy UPLC/MSE experi-

ment, with intensity displayed as a function of time and m/z.Figure 3b shows a similar display for the elevated energy infor-mation from the same UPLC/MSE experiment. Shown inset inFigure 3b is a series of elevated energy ions observed from thetime domain direction, which illustrates that high energy frag-ment ions generated can be time and profile aligned to the cor-rect precursor ion from the low energy data, and subsequentlyused for identification purposes.

Relative quantificationA binary comparison of the peptide precursor intensity measure-ments of [PSI+] and [psi-] is discussed in Figure 4. For condi-tions with identical composition and showing no change in con-centration a 45-degree diagonal line with no variation through-out the detected range would be obtained. However, this exam-ple demonstrates significant deviation from a non-regulated typeof distribution, indicating that changes in protein expression

Figure 3a: 3D representation of the low energy chromatogramfrom the UPLC/MSE analysis of the [PSI+] condition, generatedwith in-house software.Figure 3b: 3D representation of the elevated energy chro-matograms of the [PSI+] condition. The inset shows a series ofelevated energy ions with similar profile and elution time fromthe time domain direction, generated with in-house software.

Figure 2: Low energy UPLC/MSE chromatogram details forthree consecutive injections of condition [psi-], illustrating goodquality chromatographic performance and reproducibility.

25.00 27.50

Sample B240206_012 Sm (Mn, 1x2)

240206_011 Sm (Mn, 1x2)

240206_010 Sm (Mn, 1x2)

30.00 32.50 35.00 37.50 40.00

1: TOF MS ES+BPI

3.05e3

1: TOF MS ES+BPI

3.01e3

1: TOF MS ES+BPI

2.91e3

25.00 27.50 30.00 32.50 35.00 37.50 40.00

25.00

25.32

26.89

28.31 29.48

29.83

31.29 33.56

34.2234.68

36.44

39.21

26.33

26.94

28.36 29.48

31.29 32.09 33.56

34.17

36.44

38.36

39.27

38.36

36.44

34.17

33.5131.9931.24

29.4328.25

26.84

26.33

39.21

38.36

100

%

0

100

%

0

100

%

0

27.50 30.00 32.50Time

35.00 37.50 40.00



occurred between the two investigated conditions of interest. The inset displays only those ions that are statistically up- or

down-regulated (p < 0.05 and p > 0.95). For this study, thesepeptides were subsequently searched utilizing both the peptideaccurate precursor mass and accurate mass fragment ion infor-mation to identify the parent protein.

In total, 380 proteins were identified of which approximately45% were regulated or unique to one of the two conditions. Anoverview of the number of identified regulated and non-regulatedproteins is shown in Figure 5. Interrogation of the UPLC/MSE

data identified prohibitin 1 (Phb1) and prohibitin 2 (Phb2),which were down-regulated proteins in the mitochondria of the[PSI+] strain. These proteins have been previously reported asmembrane bound chaperones stabilizing products of mitochondr-ial translation of, for instance, cytochrome c oxidase subunit 2(Cox2), which catalyzes the reduction of oxygen to water.

The regulation of Phb1 and Phb2 was confirmed by means ofWestern blotting. A comparison of the results in terms of relativeamounts is summarized in Figure 6. It can be seen from theseresults, that good agreement was found using both methods.Additionally, Phb1, Phb2 and Cox2 were measured in the totalfraction of the [PSI+] strain, by means of Western blotting. Theseresults showed no significant difference in expression of the target-ed proteins occurred within the cytosol. The level of mitochondri-ally-encoded Cox2 remained unchanged, which might imply thatthe [PSI+] factor can cause delocalization of Phb1 and Phb2.

Conclusions• Significant regulation between the investigated strains -

[PSI+] and [psi-] - have been identified and quantified

• Several proteins of interest, found to be down-regulated inthe nanoscale UPLC/MSE dataset were validated by biological quantification methods

• The Western blot analysis showed excellent correlation withthe UPLC/MSE data

• Further investigation of the up- and down-regulated proteins from this study is being performed

Reference1) A. Chacinska, M. Boguta, J. Krzewska, and S. Rospert.Prion-dependent switching between respiratory competence and deficiency inthe yeast nam9-1 mutant. Mol. Cell. Biol. 2000, 20, 7220-7229.

Figure 4: Log intensity accurate mass/retention time clusters[PSI+] vs. log intensity accurate mass/retention time clusters[psi-]. Shown in inset are the significantly up- and down-regu-lated peptides, which can be subsequently used for identifica-tion by either using PMF strategies or searches using the ele-vated energy fragment ion information.

Log Norm Intensity Sample B

Log

Nor

m In

tens

ity

Sam

ple

A

EMRT Clusters: 6719 out of 34436 Clusters

down regulatedpeptides

up regulatedpeptides

5.253.5

13.77

12.1

10.1

8.1

6.1

4.1

6.05 8.05 10.05 12.05 14.37

Figure 5: Expression distribution of the proteins regulatedand non-regulated mitochondrial proteins for conditions[PSI+] and [psi-].

105

> ± 1.3 (0.30 Inscale) fold

change

> ± 1.62 (0.50 Inscale) fold

change

> ± 2.70 (1.00 Inscale) fold

change

conditionunique

no change total

45

350

300

250

200

150

100

50

0

12

41

234

380

Figure 6: Comparison of Western blotting and nanoscaleUPLC/MSE (left) for the investigated mitochondrial fractions andthe results of Western blotting experiments (right) conductedon the total fractions of yeast strains shows lack of difference inexpression of proteins synthesized in cytosol (Phb1 and Phb2),while mitochondrial encoded Cox2 remains decreased in [PSI+]strains. Relative amounts, i.e. [PSI+]/[psi-], are presented forboth techniques.

mitochondrial fraction

Wes

tern

blotting

LC-M

S

GTP19

7 [psi-

]

GT81-1

D [PSI+

]

GTP19

7 [psi-

]

GT81-1

D [PSI+

]

Wes

tern

blotting

total fraction

0.71

0.39

0.28

0.93

0.69

0.43

0.43

0.91

Phb1

Phb2

Cox2

Mdh1

0.99

0.98

0.32

0.86

JUNE 2007



Removal of Interferences and Easier MetaboliteDetection by Ion Mobility Mass Spectrometry

Jose Castro-Perez, Kate Yu, and John ShockcorWaters Corporation, Milford, MA

T he task of identifying drug metabolites from complexbiological matrices such as bile, plasma, feces, and urine

with traditional techniques can be difficult. One of the typi-cal problems when running in vivo samples is that withoutthe use of radiolabeled compounds, there are no referencepoints to look for xenobiotics. Therefore in the vast majorityof cases, the analyst relies heavily upon personal experienceand customized analytical strategies to detect and identifylow-level metabolites from high endogenous backgrounds.The complexity of this analysis could be significantly reducedby the use of an additional stage of separation, which isorthogonal to the LC and mass spectrometric separations,and occurs on a timescale that is intermediate between thetwo. A technique that possesses this capability is ion mobili-ty spectrometry (IMS). IMS separates ionic species as theydrift through a gas under the influence of an electric field.For any particular ion, the rate of drift depends on its mobil-ity, which in turn is dependent on factors such as mass,charge state and the interaction cross-section of the ion withthe gas. This additional dimension of separation fidelity leadsto improved specificity and sample definition so that moreinformation about the sample can be extracted. The multi-dimensional data produced by the Waters® Synapt™ HighDefinition MS™ (HDMS™) system is visualized and manip-ulated using DriftScope™ Mobility Environment software.

In this work, UPLC®/IMS-TOF-MS (Figure 1) analysis wasconducted on a rat bile sample. By using the DriftScope soft-ware, the metabolites from this complex matrix were easily visu-alized, as the drift time was used to separate background ionsfrom real drug-related metabolites. Further extracted ion chro-matograms were also obtained by selecting specific ions with thesoftware. As a result, the extracted ion chromatogram and MSspectrum for each metabolite were attained without interferencefrom the endogenous compounds.

ExperimentalThe in vivo ketotifen sample was obtained from rat bile. Thedrug was dosed intraperitonally at 10 mg/kg and bile was col-lected from 0 to 3 hours after drug administration. The samplewas diluted 1/10 using water with 0.1 % formic acid prior to theUPLC/MS analysis.

UPLC conditionsLC System: ACQUITY UPLC Column: ACQUITY UPLC BEH C8

2.1 x 100 mm, 1.7 µm Temperature: 45 ˚C

Flow Rate: 600 µL/min Mobile Phase A: Water + 0.1% Formic AcidMobile Phase B: AcetonitrileGradient: 95% A to 20% A in 6.5 min., hold at 0% A

for 0.5 min. before returning to 95% A for re-equilibration

MS conditionsMS System: Synapt HDMS Ionization Mode: ESI PositiveCapillary Voltage: 3200 VCone Voltage: 35 VDesolvation Temp: 400 ˚CDesolvation Gas: 800 L/HrSource Temp: 120 ˚CAcquisition Range: 100-1000 m/zMobility Carrier Gas: Nitrogen at 32 mL/min.

Figure 1: The Waters ACQUITY UPLC system with the Synapt HighDefinition Mass Spectrometry system.



ResultsThe UPLC/IMS-TOF-MS results were reviewed using theDriftScope software. It allowed sample data to be visualizedas mz/dt, or dt/rt, or mz/rt (mz=m/z, dt=drift time, andrt=retention time). The initial review using the dt/rt option

enabled the visualization of all components by using drifttime, which was used to exclude false positives and conse-quently remove interfering ions that masked the detection ofsmaller metabolites. As shown in Figure 2, this was a verycomplex sample. But due to the fact that the drift time infor-mation was available, this data dimension was used to selectthe areas of interest. This can be seen in area A of the figure,where all of the metabolites and the parent drug reside. Asimple lasso around this region within the software allowedefficient export of the corresponding TIC to MassLynx™

software for further data processing. Also in Figure 2, when focusing on areas A and B it was

observed that even though some of the metabolites co-elutedchromatographically with bile salts, they both exhibited differ-ent drift times. This indicated a very selective technique toremove the background ions (D), PEG (B), and bile salt inter-ference (C). Figure 3 shows that once the area of interest hadbeen selected and exported to MassLynx, substantial differenceswere seen in the cleanup of background and bile salt interfer-ences when drift time was used to extract the metabolite infor-mation (B) vs. a normal TIC without the use of IMS (A).

Another interesting observation in Figure 2 was that as onlyarea A was selected, we were consequently above the back-ground noise which was constituted mainly by solvent ions(D) so the noise was zero. Therefore, it becomes even easier todetect very small components in complex samples where thebackground is typically very high.

Figure 2: DriftScope software's dt (msec.)/rt (min.) plots for rat biledosed with ketotifen. Various regions are easily visualized from theplots: (A) metabolites and the parent drug, (B) PEG, (C) bile salts, and(D) solvent ions.

Retention time (mins)

Dri

ft t

ime

(mill

i sec

s)

0

9.6

6.4

3.2

05 10

Figure 3: TICs for ketotifen rat bile sample: (A) TIC without the use of drift time to extract the information, and (B) TIC with the use of drifttime to extract the information..

Time

2.10

2.17

2.382.47

2.77

3.03

0.47

0.99 1.292.10

2.412.49

2.802.88 3.363.54

3.99

4.53

4.674.80

5.495.76 6.02 6.84

7.01

7.46

7.86

0.59

A Raw TIC

B Drift time toextract metabolites

100

%

0

100

%

00.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00

JUNE 2007



Further, we combined the TICs from the non-IMS acquisi-tion vs. those from the IMS-based acquisition from 2 to 3.2minutes where both the metabolites and the parent drug eluted.As seen in the results displayed in Figure 4, it is clear how muchmore visible the metabolites were, showing an overall improve-ment in the corresponding spectra when drift time wasemployed. Even though some of the bile salts had the sameretention times, drift time very effectively removed them fromthe spectra to further aid in the detection of metabolites.

Therefore, from Figure 4 (A) it was deduced that ketotifen atm/z 310 was detected together with all of its major metabolitesin bile corresponding to N-dealkylation m/z 296 (-CH2),hydroxylation m/z 326 (+O), double hydroxylation m/z 342(+2O) and triple hydroxylation m/z 358 (+3O).

Other minor metabolites were also detected, indicatinghydroxylation and reduction to ketone m/z 324 (+O -2H),and double hydroxylation and reduction to ketone m/z 340(+2O - 2H).

ConclusionIon mobility capabilities with the Synapt HDMS system offeran additional fourth dimension of information in the analysisof very complex samples by UPLC/MS. The DriftScopeMobility Environment software is an essential tool to visualizethe data and serves as a very powerful aid to extract the infor-mation required, allowing only relevant drug-related informa-tion to be extracted. IMS technology could significantly accel-erate the identification of metabolite peaks in "cold" in vivometabolism picture earlier on during drug discovery, whereradiolabeled material is usually not available.

AcknowledgementsThe authors would like to thank Laurent Leclercq for the dona-tion of the samples for this study.

Figure 4: Spectra combined over retention times from 2 to 3.2 min. for both TICs: (A) with the use of drift time to extract the information, and(B) without the use of drift time to extract the information.

m/z50

100

A

A

Parent M

M-14

M+16

M+32

M+48

BileSalts

296.11

310.11326.11

342.11

358.11

310.12326.12

327.13382.13

460.25

462.25

480.25

496.24

514.25

515.26

516.29398.13

379.07

B

B

%

0

100

%

0100 150 200 250 300 350 400 450 500 550 600 650 700 750 800

©2007 Waters Corporation. Waters, The Science of What’s Possible, Atlantis,Ultra Performance LC and ACQUITY UPLC are trademarks of Waters Corporation.

WIDER RANGE OF

COMPOUND RETENTION

POLAR RETENTION

SUPERIOR STABILITY AT LOW pH

IMPROVED PEAK SHAPE

100% H2O COMPATIBLE

Atlantis® T3 HPLC and ACQUITY UPLC® HSS T3 columns

are the culmination of more than three decades of

Waters proprietary experience in understanding and

improving retention, selectivity, peak shape and stability

in reversed-phase LC separations. The Waters innovative

and unique T3 bonding and endcapping process provides

superior polar compound retention and aqueous compat-

ibility while enhancing column performance, lifetime,

peak shape and stability.

To learn more, visit www.waters.com/polar

[ S T A B I L I T Y ]

[ theory ] · 2007-11-21 · drug discovery screening, raw material analysis, impurity profiling,...

Documents

Transcript of [ theory ] · 2007-11-21 · drug discovery screening, raw material analysis, impurity profiling,...