Case Center for Proteomics and Bioinformatics Workshop Series March 5: Masaru Miyagi and Chao Yuan Will Present: “Quantitative proteomic analysis using stable isotopic labeling” April 2: Elizabeth Yohannes Will Present: “2D-DIGE workflow: An applied quantitative proteomics for molecular signature identification for diabetic bladder and erectile dysfunction” We will continue the workshop series in the Fall with the following topics: Structural Mass Spectrometry Posttranslational modifications of proteins and peptides Interaction Proteomics Bioinformatics CWRU has a license for Ingenuity software. The software conducts pathways analysis from your supplied targets. If you are interested in utilizing this software, send an e-mail request to [email protected]
79
Embed
Case Center for Proteomics and Bioinformatics Workshop Seriesproteomics.case.edu/pdfs/proteomicsworkshop_feb.pdf · Case Center for Proteomics and Bioinformatics Workshop Series ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Case Center for Proteomics and Bioinformatics Workshop SeriesMarch 5: Masaru Miyagi and Chao Yuan Will Present:
“Quantitative proteomic analysis using stable isotopic labeling”
April 2: Elizabeth Yohannes Will Present:“2D-DIGE workflow: An applied quantitative proteomics for molecular signature identification for diabetic bladder and erectile dysfunction”
We will continue the workshop series in the Fall with the following topics:
Structural Mass SpectrometryPosttranslational modifications of proteins and peptidesInteraction ProteomicsBioinformatics
CWRU has a license for Ingenuity software. The software conducts pathways analysis from your supplied targets. If you are interested in utilizing this software, send an e-mail request to [email protected]
Section 2 - Gel Based AnalysisExample of Tissue Analysis in Animal Models
Cardiovascular
Quantitative ProteomicsDetermine changes in a protein’s abundance and/or expression due to a biological challenge (disease,drug)
Better understand biological function and pathophysiology of diseaseBiomarker Discovery
Disease preventionTreatment
Quantitative Proteomics – Analysis Platforms
‘Top Down’ ProteomicsQuantify in tact proteins
2-Dimensional gel electrophoresis2D-DIGE, SELDI-TOF
‘Bottom Up’ ProteomicsQuantify peptides via enzymatic cleavage of one or many proteins
Isotope Coding – iTRAQ, SILAC, O18
Direct Quantification – Label free expressionTargeted Quantification – MS Western, Isotope Dilution Tandem Mass Spectrometry
Label Free ExpressionAdvantages
Can accommodate complex experimental designsCan provide increase proteome coverageEasy transition to a ‘bottom up’ validation analysisAmendable to low sample concentration (600 nanogramsfor LC/MS/MS)
DisadvantagesDifficulty quantifying post translational modificationsHigh degree of sample complexity for the mass spectrometer
Anderson, N. L. (2002) Mol. Cell. Proteomics 1: 845-867
Proteins Measured Clinically in Plasma Span > 10 orders of Magnitude in Abundance
Sample Preparation and Fractionation Strategies
DepletionPlasma, Serum and CSF
Affinity depletionAlbumin only7 or 14 most abundant plasma proteins
FractionationStrong cation exchange chromatography (peptide)One dimensional gel (protein)
Label Free Analysis WorkflowC T
C TBiomarker Discovery
Stats.Analysis
Diffset N
(… = Mass, Time, Intensity and Sequence)
…T…C
Digestion
Raw Data AcquisitionLC/MS/MS Data Processing
NetworkAnalysis
Diffset 1
450 500 550 600 650 700 750 800 850 90m/z
0
5
10
15
20
25
30
35
40
45
50
55
60
Rel
ativ
e A
bund
ance
440.54470.27
586.82
682.29
821.37509.27
660.30754.55
522.26497.93 655.97620.76
801.78548.51 703.39581.79
787.36 845.94 890.7
732.312 Mo DM
1 Mo DM
3 Day DM
C:\FT_MS Raw Data\...\Dani15_LF_1 8/7/2007 10:28:51 PM
NL:5.81E6Base Peak F: FTMS + p NSI Full ms [400.00-1600.00] MS Dani15_LF_1
NL:4.14E6Base Peak F: FTMS + p NSI Full ms [400.00-1600.00] MS dani15_lf_11
NL:5.71E6Base Peak F: FTMS + p NSI Full ms [400.00-1600.00] MS dani15_lf_6
Retention Time
Label Free Mass Spectrometry Data
Quantification of 1,000 of peptides
Under Accurate Mass Conditions there are 160,000 units on the Y axis
m/z 400 - 1,200 (+ / - 0.005)
< 5 ppm mass error
Can be thought of an ‘array’technology
Data AnalysisProcess Raw Data
Intensities vs. SamplesPeak list generation (Mass)
Chromatographic Peaks Peptide sequence assignment
Data QCFilter poor quality peptides
Replace missing valuesNormalization
Statistical AnalysisANOVA
PCAFDR
Raw Imputed
NormalizedpeptideN
(… = Mass, Time, Intensity and Sequence)
…SN…S1
peptidet1
Jean eudes Dazard, Rob Ewing
Bioinformatic AnalysisData mining to enable biological interpretation of high dimensional datasets
Molecular network relevant to disease pathway understudy
Software tools which we use:Ingenuity Pathway Analysis (IPA)
Curated literature and protein interaction databasesPathway Studio
Automated literature search and protein interaction databases
MetaCoreProtein interaction databases
Example 1 – DiabetesMark Chance Ph.d – Center for Proteomics
Urine Biomarker for diabetic complications (CAD and DM nephropathy)
20 million people in the US have Type I or II diabetes; 150 million worldwideEnd Stage Renal Disease (ESRD) and Coronary Artery Disease are important disease complications
DM is leading cause of kidney failure Approximately 60,000 with ESRD die yearly
CAD leading cause of mortality in the US with a 3-10 fold increase risk in those with diabetes
Biomarkers of DM Complications-Limitations
Increased urinary albumin excretion is a current biomarker for both diabetic nephropathy and coronary artery diseaseMeasurement can be non linear over time and high incidence of remission is observed, questions as to sensitivity
Perkins et al, Current Diabetes Reports 5:455-463, ‘05
Frequency of micro-albuminuria remission and progression
Uro-genital ComplicationsChanges in GFRDefects in proximal tube re-absorptionDefects in lysosomal processingChanges in protein expression in relevant tissues
Urinary Proteomics Discovery in STZ induced Diabetes Complications Rat Model
3 day 10 day 1 month 2 monthControl 1 3 2 3
STZ Treated 3 4 4 3
1 2 3 4 6 8
Development Progression of DiseaseEnd Organ Alterations
# Months after STZ treatment
Digested Peptides from Urine: LC/MS Chromatograms
INTENSITY
450 500 550 600 650 700 750 800 850 90m/z
0
5
10
15
20
25
30
35
40
45
50
55
60
Rel
ativ
e A
bund
ance
440.54470.27
586.82
682.29
821.37509.27
660.30754.55
522.26497.93 655.97620.76
801.78548.51 703.39581.79
787.36 845.94 890.7
732.312 Mo DM
1 Mo DM
3 Day DM
C:\FT_MS Raw Data\...\Dani15_LF_1 8/7/2007 10:28:51 PM
NL:5.81E6Base Peak F: FTMS + p NSI Full ms [400.00-1600.00] MS Dani15_LF_1
NL:4.14E6Base Peak F: FTMS + p NSI Full ms [400.00-1600.00] MS dani15_lf_11
NL:5.71E6Base Peak F: FTMS + p NSI Full ms [400.00-1600.00] MS dani15_lf_6
Retention Time
Summary Statistics for Diabetic ComplicationProteomics Discovery
2868 chromatographic peaks were generated812 of these received a peptide assignment 494 peptides passed QC filter and were analyzed by ANOVAOut of the 494 peptides, 64 of these had p-values of <0.05 for interaction (time & treatment)These peptides represent approximately 30 different proteins that:showed progression upon treatment
Specific Biomarker Candidates
Down –Regulated Protein (Hypoinsulinemia)
Up-Regulated Proteins (Hyperglycemia)
Major Urinary Protein Precursor Collagen alpha 2(1) chain precursor
Biomarker Discovery – Pilot Human Study
Goal – To transfer rat urine proteomic analysis protocols to human urine and to discriminate between sample types within pilot study
Coronary Artery Calcification in Type I Diabetes Study (CACTI)
Optimized protocol for 4 mL human urine3 controls, 3 Diabetics, 3 Diabetics with macro-albuminuriaAge matched (33-55 years)
7446 diffsets were generated1165 of these received a peptide assignment 826 peptides passed QC filter and were analyzed by ANOVA
Contrast Significant Peptides (up or down regulated)
Adj P-value <0.05
Control vs DM 27
DM vs DM -ALB 377
Control vs DM -ALB 435
PCA of Label Free CACTI Analysis
DM-Alb
Control
DM
Changes in Protein Levels
Albumin Transferrin
CD 59 Glycoprotein PrecursorUromodulin
ConclusionsDeveloped and optimized an label free expression protocol for urine samplesBiologically significant changes in urinary proteins are observed in rat model across treatment and timePilot study shows differences in human data consistent with rat modelValidation study to to define predictors of CAD and ESRD underway
Example 2 – CardiovascularDaniel Simon MD – Cardiovascular Medicine
Pathway discovery for understanding the molecular basis of restenosis and drug eluting stents
Restonsis is a healing response to mechanical injury
Inflammation plays a critical role in neointimal thickening
Inflammatory cells contribute to the bulk within the intimaIncrease growth and chemotactic factorsProduction of enzymes which can facilitate cell migrationLocal inflammatory signals release the bone marrow derived stem cells that participate in the healing response of the blood
SMC and endothelial cells proliferation
Simon, DI et al. Circulation. 2005;111:2257-2273
Simon, DI et al. Circulation. 2007;115:548-549
Restenosis(DysregulatedHealing)
▪ Smooth Muscle Cell Proliferation▪ Matrix Growth
Neointimal Thickening
Reendothelialization(Normal Healing)
Inflammation▪ Activation and Cell-to-cell Interaction of Leukocytes,
Contrast Significant Peptides (up or down regulated) Adj P-value <0.10
Drug stent vs bare metal 48 hours 280 (56 Proteins)
15701 diffsets were generated1370 of these received a peptide assignment 1064 peptides passed QC filter and were analyzed by ANOVA
Drug/Bare Metal at 48 hours
Green
Red
Follow Up
Two animal models to evaluate importance of CRP and complement pathway in recruitment of stem cell progenitors to vessel healing
Femoral hind limb artery ischemiaCRP knock in mice – increase recruitmentComplement knock out mice – decrease recruitment
Example 3 - CancerKenneth Cooke – Pediatric Oncology/Hematology
Biomarker discovery to understand the immunologic mechanisms that contribute to the development of idiopathic pneumonia syndrome (IPS) following allogeneic stem cell transplantation (SCT).
IPS is a major complication of SCT
3-15% of all SCT recipients60-80% mortalityEvidence from rodent models suggest lung is a target immunologically mediated damage
Inflammtory effectors of TNFαDonor derived T cell effectors
Cooke,KR et al. Bone Marrow Transplantation. 2004;34:753-765
Cytokine Plasma Measurements in BMT Patients
TNFα
Cont NL
Cont BMT
Pre-BMT
IPS Dx
0
50
100
150
**
*****
**
**
pg/m
L
IL-6
Cont NL
Cont BMT
0
5
10
500
1000
1500 ***
**
******
Pre-BMT
IPS Dx
pg/m
L
Cooke and Yanik et al. Blood. 2008;112 (8):3073-81
Proteomic Analysis
Goal -To identify proteins that are involved in the development of IPS in SCT patients
24 plasma samples were analysed from SCT patients at BMT and post (14 days) BMT or Dx of IPS
6 Non-progressors6 IPS patients
Sample Prep Depletion – 7 most abundant plasma proteinsFractionation – SCX of pooled digests
PCA
IPS dx
IPS d0
C d 0
C d 14
Label Free Analysis – Summary Statistics
14523 diffsets were generated1253 of these received a peptide assignment 1088 peptides passed QC filter and were analyzed by ANOVA
Contrast Significant Peptides (up or down regulated)
Secreted protein involved in the initial immune cell clustering during inflammatory responses and may regulate the chemotactic activity of chemokines.
LPS sensing by the Toll like receptor (TLR4) and innate immune responses are regulated by lumican
Lumican knock out mice are hypo responsive to LPS induced septic shock – poor induction of TNF-alpha, IL –beta and IL6 in serum
Individual Peptide Intensities for Lumican
Lumican Abundance at Day 0
0
2,000,000
4,000,000
6,000,000
8,000,000
10,000,000
12,000,000
14,000,000
16,000,000
1_0 2_0 3_0 4_0 5_0 7_0 8_0 9_0 10_0 11_0 12_0
IPS and Non IPS at Day 0
Ave
rage
Inte
nsity
Acr
oss
All L
umic
an
Pept
ides
Proteins – Early BiomarkersAHSG Alpha-2-HS-glycoprotein ORM1 Alpha-1-acid glycoprotein 1
A positive acute phase protein – TNF-alpha and IL-6 may up regulate the expression in the liver of this protein
A negative acute phase protein – TNF-alpha and IL-6 may down regulate the expression in the liver of this protein
Footprint of IL-6 and TNF-α signalingInteraction effect – disease over timeGreen Red
Follow Up
Validation of Lumican by WesternInvestigate knock out of rodent model of disease
Additional network analysis of using Pathway Studio for further data miningAnalysis including day 7 samples to detect potential early biomarkers id’d in initial analysis
Inte
nsity
MS/MSRP Column
Concentration
Biomarker Validation Strategy: Isotope Dilution Tandem Mass Spectrometry
High sensitivity, high specificity, absolute quantification
Quantification of Pro Alpha (2) 1 Collagen Peptide (GEPGSVGAQGPPGPSGEEGK)
Will briefly discuss mass spectrometry and data analysis
Consistency
Animal Preparation
Control (n=4)
Buffer
Isolate Hearts
Diabetic (n=4)
STZ
Isolate Hearts
Insulin (n=4)
STZ
Insulin
Isolate Hearts
Week 0
Week 4
Week 6
PI: Andrea Romani (Physiology, CWRU)
Protein ExtractionUnlike body fluids, tissue has significant amounts of structural and membrane proteins which are difficult to be extracted.In addition to mechanical homogenization and sonication, detergents are often used to facilitate the extraction of protein.Good detergents must
Have high/consistent extraction efficiencyBe compatible with digestion & LC/MS
Micro-dissected breast carcinoma tissue
Benefits of 1D-Gel
Compatible with most extraction buffers.
RemoveDetergentsAbundant protein speciesLow molecular weight contaminates
Protein fractionation
Fractionation
Divide whole proteome into multiple fractionsGoal: to increase proteome coverage
Two concerns regarding in-gel-digestion:-Digestion efficiency-Extraction efficiencyUse an improved in-gel-digestion protocolPulverize gel pieces -> increased exposure to trypsinExtract 5 times with shaking/sonication-> max. recovery
Next: Miss-cleavage
Miss-Cleavage
Maximum peptide recovery does not mean maximum digestion efficiency.Protein might have partial 2nd structure, or be protected from gel matrix that prevent full tryptic digestion.Some chemical properties of peptides also induce miss-cleavageThese larger peptides with miss-cleavages can be successfully extracted because of our vigorous extraction method.Miss-cleavage is usually not a problem for protein ID.But it poses serious challenges for protein quantification as peptides mixtures with different miss-cleavage rates can hardly be compared, and exclusion of these peptides will dramatically reduce the number of quantifiable peptides.Therefore, we used a two step digestion protocol.
Comparison of two digestion protocols
Band #6 (Pool)
In-gel-digestion
extraction
2nd digestion
LC/MS
Peptide/Protein ID
Calculate miss-cleavage rate
One-Step Two-Step
Total unique peptides: 806 (two step); 846 (one step)
Comparison of Miss-cleavage Rates
50%
Miss-cleavage Rates-Band 6 of 12 Samples
Average CV2 miss cleavages 1.6% 16%
1 miss cleavage 10.4% 14%
Total miss-cleaved 12.0% 13%
Next: properties of miss-cleavage sites
Nature of Miss-cleavage
R/K E/D or phosphorylation
Distribution of Miss-cleavages
Further reduction of miss cleavage rates using an bacterial trypsin
1.7% of total
Miss cleaved peptides from all band 6
A Promising Bacterial Trypsin (SET)•Streptomyces griseus trypsin (SET) was known to be resistant to autolysis.•We tested SET for digestion efficiency.•Porcine trypsin and SET were used to digest equal amounts of BSA for 5 hours.
Pep
tide
coun
t
Next: LC/MS analysis
71
LC/MS
Mass SpectrometerRP Column
HPLC
Care was taken to ensure the reproducible performance of LC/MS.Duration: 8 days; RT: < 2min; Mass Accuracy: < 2 ppm
72
MS of intact peptide (precursor ions)
m/z
Intensity
73
MS of fragment ions
P E P T I D E
P E P T I D Eb5 y3
y4
b3
b2
b1
y2
y3
y5y6b4
b5
R
RHe
………………………………………………………………
CID
Label Free MS Data Analysis
Peptide M
Intensity
Sample NAll Bands…
Sample 1All Bands
Peptide 1
LC/MS & Data Analysis
4106 unique peptides and 820 unique proteins were were identified with high confidence.Data are subjected to label free data analysis using a commercial software.Some changes can be observed even at chromatographic level.
Control 1
Diabetic 1
Insulin 1
Control 2
Diabetic 2
Insulin 2
Control 3
Diabetic 3
Insulin 3
Control 4
Diabetic 4
Insulin 4
Summary
1D GelProtein extraction “bottom-up” proteomics
•Compatible with most detergents•High Reproducibility•Economical and efficient fractionation tool•Optimized in-gel-digestion protocol
Next Workshop:
Quantitative proteomic analysis using stable isotopic labelingMasaru Miyagi and Chao Yuan
March 5, 2009
Key to label free MS analysis: Reproducibility
AcknowledgementsMark Chance Ph.D Director Center for Bioinformatics and Proteomics
CollaboratorsDaniel Simon M.D.Ken Cooke M.D.CACTI
Marian Rewers, David Maahsand Janet Snell-Bergeon
Andrea Romani Ph.D
All Center Staff and Faculty
Jean Eudes Dazard Ph.DBiostatistics and experimental design