note: 345 identification and differential expression ....pdf · identification and differential...
TRANSCRIPT
ApplicationNote: 345
Key Words
• LTQ™
• ADME/Tox
• MALDI TOF
• ProteinIdentification
Proteomics in ADME/Tox Studies: High-ThroughputIdentification and Differential Expression Analysisof Proteins in Mouse Liver Following Drug TreatmentOla Rönn1, Johan Öhman1, Daniel Haid1, Helena Nordvarg1, Lena Hörnsten1, John Flensburg1, Erik Forsberg1, David Fenyö1,Hélèn Bergling1, Gary Woffendin2, and Michaela Scigelova2
1GE Healthcare, Uppsala, Sweden; 2Thermo Fisher Scientific, Hemel Hempstead, UK
Introduction
It is estimated that about 50% of drugs in developmentfail during clinical trials because of deficiencies in theirADME/Tox properties (absorption, distribution, metabo-lism, elimination and toxicity). The cost of these failuresso late in the drug development process is naturally veryhigh. In addition, it has been suggested that over 5% ofhospitalized patients suffer serious adverse reactions todrugs that have successfully completed their developmentand come onto the market. Improved means of gatheringADME/Tox information earlier in the drug developmentprocess should thus benefit both pharmaceutical manu-facturers and, of course, patients.
To understand the processes related to the metabolismand potential toxicity of a candidate drug we chose tocombine the method of 2-D Difference Gel Electrophoresis(DIGE) to highlight proteins which are significantly up-or down-regulated after the administration of the drug,with mass spectrometric analysis to unambiguouslyidentify such proteins.
The DIGE technique in combination with analysesusing DeCyder™ 2D Differential Analysis software allowsfor the concomitant analysis of up to three distinct sam-ples in one gel. This enables a true internal standard,made by pooling all the analyzed samples, to be run alongwith the other two samples, thus eliminating problemsof 2D gel-to-gel reproducibility. Moreover, a gel withinternal standards enables a highly accurate comparisonof expression levels for proteins of interest.
The use of a robotic workstation for spot handling,digestion, and sample target preparation gave the desiredlevel of automation and throughput, enabling a seamlesslink to the mass spectrometric analysis of the digested pro-tein samples. A two-tiered strategy of mass spectrometricanalysis was considered for protein identification: 1) Allsamples were subject to peptide mass fingerprinting, usinga MALDI-TOF mass spectrometer for identification of themajority of gel spots, then 2) tandem mass spectrometry,using nanoelectrospray on an ion trap mass spectrometer,guaranteed unambiguous and exhaustive identificationof those samples not identified by peptide mass finger-printing. The suggested experimental workflow and keytechniques used in the study are illustrated in Figure 1.
Goal
Use a highly automated and reproducible proteomicsapproach to obtain insight into a complex biologicalsystem, combining MALDI TOF and ESI linear ion trapmass spectrometry with 2D-DIGE.
Experimental Conditions
Sample Preparations
The selected candidate drug was administered orally toinbred C57BL/6 mice over five days. Livers from treatedmice and non-treated littermates were removed, dissected,and snap frozen in liquid nitrogen. Pieces of liver fromthree treated mice and three untreated mice (0.5 g each)were rinsed in PBS, and homogenized using a PotterElvehjem homogenizer in 10 volumes (5 mL) lysis buffer(10 mM Tris-HCl pH 8.3, 7 M urea, 2 M thiourea,5 mM Mg(Ac)2, 4% CHAPS) according to the EttanDIGE User Manual1.
2D gelElectrophoresis
Mix
Treated withCandidate
Drug
PooledStandardSample
Label withDIGE Cy2
Label withDIGE Cy5
Untreated
Label withDIGE Cy3
Image Analysis
Ettan SpotHandling
LC-MS/MS
MALDI-ToF
Figure 1: The experimental workflow used for the identificationof differentially expressed proteins.
Page 2 of 6
To remove interfering non-protein material, a 2-DClean-Up Kit was used on a 10×100 µL homogenateaccording to the kit instructions. The 2-D Quant Kitwas used to quantify the prepared samples, following themanufacturer’s protocol.
Samples were aliquoted and pH was adjusted to 8.4by adding 2 µL of 50 mM NaOH to each 100 µL aliquotof homogenate. Treated and un-treated liver samples werelabeled with CyDye DIGE Fluor Cy5 and CyDye DIGEFluor Cy3, respectively, according to the published proce-dure1. In addition, a pooled standard containing all sam-ples included in the experiment was prepared, and labeledwith CyDye DIGE Fluor Cy2, to be used as an internalstandard for the image analysis of the 2D gels. Prior tothe 2D gel analysis the labeled individual samples andthe pooled standard were mixed.
2D Gel Electrophoresis
1st dimension: Cup loading was selected for an analyticalelectrophoresis run. 150 µg of the mixed sample (50 µgof each Cy2-, Cy3-, and Cy5-labeled sample per gel) wasadded to each IPG Strip pH 3-11 NL. For preparativeelectrophoresis, the samples (1 mg) were loaded onto IPGstrips pH 3-11 NL by in-gel rehydration in DeStreak™
Rehydration solution containing 2% IPG buffer pH 3-10. 2nd dimension: Electrophoresis was run overnight
using home-cast gels in Ettan DALT Twelve electro-phoresis system1.
Image Analysis
Each gel was scanned using an imager at 100 µm resolu-tion. Individual images were created to capture differencesin Cy2-, Cy3-, and Cy5-labeled proteins.
Images were analyzed using DeCyder™ software(version 5) in both the DIA and the BVA modules. Forstatistical analysis, the Student T-test value was set to 0.01± 1.5 times intensity. Two-way ANOVA with the setting0-0.001 was used.
The spots were detected using the DIA module, andthen imported to the BVA module where the gels werematched and analyzed further.
Automated Spot Handling
Two preparative gels stained with SYPRO Ruby wereadded to the experimental study, and parallel pick liststhat included all spots of interest (i.e. those that showedmore that 50% change in level following the treatmentwith the candidate drug) were created for subsequentanalysis with peptide mass fingerprinting and LC-MS/MSto determine the proteins’ identity.
Ettan Spot Handling Workstation managed a spotpicking, digestion, and spotting on Ettan MALDI targetslides, the whole procedure being run automaticallyovernight. Gel plugs cut by a 2 mm picking head werewashed twice in 50 mM ammonium bicarbonate in 50%methanol, and once in 75% acetonitrile before drying.For digestion, 10 µL of proteolytic enzyme solution(0.2 µg sequencing grade enzyme, Ettan Chemicals) wereadded before incubation at 37°C for 2 hours. Extraction
was performed in two steps by adding 50% acetonitrile in0.1% trifluoroacetic acid (Ettan Chemicals). A matrixsolution of 5 mg/mL solution of recrystallized α-cyano-4-hydroxycinnamic acid (LaserBio Labs) in extraction liquidwas deposited on the target followed by depositing 10%of the digested sample, saving the rest for analysis byLC-MS/MS, if needed.
Protein Identification using MALDI-ToF and LC-MS/MS
All samples were subject to peptide mass fingerprintingusing an Ettan MALDI ToF Pro system. This methodafforded quick protein identification of most samples.About 30% of analyses were then processed withLC-MS/MS.
The tandem mass spectrometric analysis was per-formed on a Thermo Scientific LTQ linear ion trap fittedwith a Thermo Scientific BioBasic™ 100× 0.1 mm, 5 µmcolumn flowing at 1 uL/min flow rate. A fast gradientprofile enabled a total analysis time of 20 minutes withapproximately 5500 scans acquired per sample using theData Dependent™ “triple play routine”. In this mode pep-tides detected in a survey scan get selected for high resolu-tion ZoomScan™ followed by fragmentation by MS/MS.The built-in features of Data Dependent acquisition andDynamic Exclusion™ maximize the collection of usefuldata, and enable efficient handling of coeluting peptides,ensuring maximum sensitivity and sequence coverage foridentified proteins.
The protein identification was obtained by matchingcollected MS/MS spectra to the peptide sequences in adatabase of mouse proteins. The database search wasdone with BioWorks™ 3.1 software (SEQUEST®) againstmouse.fasta database with added proteolytic enzymesequence. We considered an open enzyme specificity forpotential non specific cleavage sites (those could arisefrom the activity of other proteases present in the liversample), methionine oxidation (a common artifactencountered in protein digests from 2D gels; adding 16to Met residue mass), and a modification of cysteinewith DeStreak (adding 76 to Cys residue mass).
Results and Discussion
The quality of the protein sample is the crucial step toobtain high quality 2D gels and ensure reproducible andreliable results. Therefore, special attention was devotedto sample preparation prior to the DIGE analysis. Ettan2-D Clean-Up Kit was used to efficiently remove non-protein contaminants. The protein content was quantifiedwith Ettan 2-D Quant Kit.
As expected when preparing samples from an organlike the liver, the protein concentration was about 10mg/mL (treated 1, 10.02 mg/mL; treated 2, 9.67 mg/mL;treated 3, 10.23 mg/mL; pooled non-treated control,10.40 mg/mL).
The liver samples were then analyzed by DIGE todetect and quantify the biological differences in thesample. The gel images were scanned and analyzed withimage analysis software, and 1450 spots were matched
Page 3 of 6
between the different images of each of the three gels (oneimage: Cy2 labeled–pooled internal standard; the secondimage: Cy3 labeled–untreated pooled control; and thethird image: Cy5 labeled–treated with candidate drug).
The image analysis highlighted 30 protein spotswhich level changed significantly; greater than 50% up-or down-regulation observed (Figure 2). The differentiallyexpressed proteins were mapped onto the SYPRO Rubystained preparative gel to generate a pick list for the EttanSpot Handling Workstation, which used it to cut out thegel plugs, digest the proteins, and then spot the extractedpeptides onto the MALDI target plate.
The peptide mass fingerprinting on Ettan MALDI ToFconfidently identified 22 of the 30 selected proteins. Figure3 shows a spectrum of formyl transferase as an example.The samples that were not conclusively identified by pep-tide mass fingerprinting technique were then analyzed onthe LTQ linear ion trap. The high speed of data acquisitionon the LTQ enabled us to use a short 20-minute chromato-graphic gradient, thus ensuring a high sample throughput(Figure 4).
The database search with BioWorks 3.1 software(SEQUEST) confidently identified proteins in all samplesanalyzed by LC-MS/MS. Table 1 lists the proteins identi-fied in each of the eight samples with at least two differentpeptides per protein. The high sequence coverage adds fur-ther confidence to the identification. Moreover, it increasesthe chances to locate potential sites of mutations/modifica-
CoverageSample Protein ID (%)
362 Ornithine transcarbamylase 23Methionine adenosyltransferase 8ATP-Citrate lyase 3Proteolytic enzyme 22
938 Serine proteinase inhibitor 36Proteolytic enzyme 17
974 Methionine adenosyltransferase 11Liver carboxyesterase 6Proteolytic enzyme 17
1306 Hydroxymethylglutaryl-CoA synthase 50Elongation factor 1 18Proteolytic enzyme 17
1417 Actin 56Proteolytic enzyme 22
1570 Hepatoma-derived growth factor 8Proteolytic enzyme 22Keratin 4
1805 Thioether S-methyltransferase 22Proteolytic enzyme 17
1824 Isopentenyl-diphosphate delta isomerase 48Proteolytic enzyme 17Keratin 8
Table 1: Proteins identified by LC-MS/MS analysis in eight samples refrac-tive to the identification by peptide mass fingerprinting. Only those proteinsfor which at least two different peptides were identified are listed, togetherwith the sequence coverage observed
Figure 2: The image analysis of the 2D gel by image analysis software
Figure 3: Identification of formyl transferase in sample 552 by peptide mass fingerprinting with Ettan MALDI ToF
Page 4 of 6
tions that might be of special interest. Multiple peptideswith oxidized Met residues were observed (Figure 5A).The sequence coverage detected for proteins in sample1306, and examples of MS/MS spectra with the correctsequence assigned by SEQUEST are given in Figure 5.
RT: 5.8 -12.3 SM: 7G
6.0 6.5 7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 11.0 11.5 12.0Time (min)
05
101520253035404550556065707580859095
100
Rela
tive
Abun
danc
e
1796
1772
2015
1941
18871905
18471736
2049
1692 2076
Figure 4: The fast chromatographic separation of peptide digests enabled by high speed of data acquisition on the LTQ linear ion trap
Figure 5: The sequence coverage and assigned MS/MS spectra for the two murine proteins detected in sample 1306(A: hydroxymethylglutaryl-CoA synthase; B: elongation factor 1)
A)
B)
Page 5 of 6
It is noteworthy, that all samples contained more thanone protein; in most cases three, and in sample 362 evenfour. Two samples were heavily contaminated with ker-atin. The fact that proteins are presented in mixtures mostprobably accounts for the inconclusive results of the pep-tide mass fingerprinting analysis. The summary of theresults from both peptide mass fingerprinting andLC-MS/MS are presented in Table 2.
Sample Change Method Protein ID233 -1.59 PMF Carbamoyl-phosphate synthetase 1267 2.85 PMF Carbamoyl-phosphate synthetase 1275 2.65 PMF Carbamoyl-phosphate synthetase 1340 1.55 PMF Pyruvate carboxylase341 1.74 PMF Pyruvate carboxylase362 -1.54 LC-MS/MS Ornithine transcarbamylase
Methionine adenosyltransferaseATP-Citrate lyase
513 -1.69 PMF Similar to elongation factor 2552 1.81 PMF Formyl transferase563 -1.61 PMF Formyl transferase938 2.75 LC-MS/MS Serine proteinase inhibitor974 -1.84 LC-MS/MS Liver carboxyesterase precursor
Methionine adenosyltransferase1035 -1.68 PMF Liver carboxyesterase precursor1169 1.90 PMF T-complex protein 11171 1.56 PMF Hydroxymethylglutaryl-CoA synthase1230 -2.25 PMF Selenium binding protein 21285 1.59 PMF Hydroxymethylglutaryl-CoA synthase 21289 1.66 PMF Hydroxymethylglutaryl-CoA synthase 21293 1.61 PMF Methionine adenosyltransferase1302 2.02 PMF Methionine adenosyltransferase1306 2.06 LC-MS/MS Hydroxymethylglutaryl-CoA synthase 2
Elongation factor 11417 1.87 LC-MS/MS Actin1464 1.68 PMF Cathepsin D1545 1.64 PMF Farnesyl diphosphate synthase1570 2.59 LC-MS/MS Hepatoma-derived growth factor
Keratin*1746 -2.22 PMF Nudix1804 -2.02 PMF Thioether S-methyltransferase1805 -1.75 LC-MS/MS Thioether S-methyltransferase1824 2.21 LC-MS/MS Isopentenyl-diphosphate delta isomerase
Keratin*1926 -1.58 PMF Major urinary proteins 11 and 81929 -1.54 PMF Major urinary proteins 11 and 8
*A likely contaminant.
Table 2: Mice liver proteins that changed their level significantly (blue for up- and red for down-) following the treatment withthe candidate drug. The method column refers to protein identification by peptide mass fingerprinting (PMF) or capillary LCtandem mass spectrometry (LC-MS/MS)
Conclusion
This study describes the application of an advanced,highly automated proteomics workflow consisting of dif-ferential expression analysis, peptide mass fingerprinting,and tandem mass spectrometry to gain an insight into acomplex biological problem.
The DIGE technology in combination with imageanalysis 2-D software was instrumental for highlightingthe proteins that were significantly up- or down-regulated,while Ettan Spot Handling Workstation streamlined thesample handling for mass spectrometric analysis. Peptidemass fingerprinting with Ettan MALDI ToF providedspeedy identification of the majority of proteins.
The capillary LC-MS/MS technique afforded theidentification of samples containing more than one pro-tein. Such samples cannot be successfully analyzed bypeptide mass fingerprinting analysis, clearly demonstratethe benefit of the MS/MS approach. The speed andsensitivity associated with the LTQ linear ion trap massspectrometer then allow short method times and ensurethe maximum sequence coverage for confident proteinidentification in a high throughput setup.
In addition, a focused study on proteins obtained bysub-cellular fractionation could prove suitable to identifyproteins involved in metabolism and toxicity in the liverfollowing administration of drugs. Greater understandingof these processes should significantly improve the effi-ciency of ADME/Tox screening studies during drugdiscovery campaigns.
References1 Ettan DIGE User Manual 18-1164-40, edition AA; rat liver protocol page 299.
Acknowledgements
Bo Lundgren and Per Garberg (Biovitrum AB, Stockholm Sweden).
AN62549_E 12/07S
Part of Thermo Fisher Scientific
Legal Notices©2007 Thermo Fisher Scientific Inc. All rights reserved. SEQUEST is a registered trademark of the University of Washington. All other trademarks are theproperty of Thermo Fisher Scientific Inc. and its subsidiaries. This information is presented as an example of the capabilities of Thermo Fisher Scientific Inc.products. It is not intended to encourage use of these products in any manners that might infringe the intellectual property rights of others. Specifications,terms and pricing are subject to change. Not all products are available in all countries. Please consult your local sales representative for details.
View additional Thermo Scientific LC/MS application notes at: www.thermo.com/appnotes
In addition to these
offices, Thermo Fisher
Scientific maintains
a network of represen-
tative organizations
throughout the world.
Africa+43 1 333 5034 127Australia+61 2 8844 9500Austria+43 1 333 50340Belgium+32 2 482 30 30Canada+1 800 530 8447China+86 10 5850 3588Denmark+45 70 23 62 60 Europe-Other+43 1 333 5034 127France+33 1 60 92 48 00Germany+49 6103 408 1014India+91 22 6742 9434Italy+39 02 950 591Japan +81 45 453 9100Latin America+1 608 276 5659Middle East+43 1 333 5034 127Netherlands+31 76 587 98 88South Africa+27 11 570 1840Spain +34 914 845 965Sweden / Norway /Finland+46 8 556 468 00Switzerland+41 61 48784 00UK +44 1442 233555USA +1 800 532 4752
www.thermo.com
Thermo Fisher Scientific,San Jose, CA USA is ISO Certified.