W.M. Keck Foundation Biotechnology Resource Laboratories Erol … · 2019-09-29 · Automated 2D Comparative LC Protein Profiling (Beckman-Coulter PF2D Platform) Partial pI/UV map

W.M. Keck Foundation W.M. Keck Foundation Biotechnology Resource Laboratories Biotechnology Resource Laboratories

Erol GulcicekErol GulcicekChristopher ColangeloChristopher Colangelo

Terence WuTerence Wu

http://info.med.yale.edu/nida-proteomics/

http://info.med.yale.edu/ysm/

http://www.yale.edu/


- Site is currently up with basic information

- Overview of profiling technologies has recently been

added since the url was first distributed

- Ideas for future features are very welcome


Yale/NIDA Neuroproteomics Center Fact Sheet

Title of Center Grant: Yale/NIDA Neuroproteomics Research Center

Theme of Center: "Proteomics of Altered Signaling in Addiction"

Principal Investigator: Kenneth R. Williams, Ph.D.Director, W.M. Keck Foundation Biotechnology Resource LaboratoryProfessor (Adj,) Research Dept. of Molecular Biophysics & BiochemistryYale University School of Medicine

Co-Principal Investigator: Angus C. Nairn, Ph.D.Professor, Department of PsychiatryYale University School of Medicine

#Grants Awarded: 2 Nationally

Total Funding Awarded to Yale:

$7.7 million

Total Budget Period: 9/23/2004 – 5/31/2009

Center Cores: Administrative, Bioinformatics and Biostatistics, Protein Identification, Protein Microarray; andProtein and Lipid Separation and Profiling

Center Investigators: 14 from Yale University, The Veterans Administration Medical Center (West Haven) Rockefeller University

mailto:[email protected]

mailto:[email protected]

Overall Goal:

To substantially increase understanding of the biochemical mechanisms that underlie substance abuse and its treatment.

Specific Aims:

Bring together Yale faculty working at the forefront of such keyneuroscience areas as

Signal transduction, Plasticity, Neuronal morphogenesis, Lipid metabolism in neuronal signaling and Synaptic functionSynaptic response to psychotropic drugs

with experts in

Proteomics, Biostatistics, Bioinformatics.

Proteomics of Altered Signaling in Addiction

Translational Research

Receptor-Mediated Signaling

Transcriptional Regulation

Endocytosis and Exocytosis

Regulation of the Cytoskeleton

Neuronal Development

Neuronal Plasticity

Cell Death

DIGE ICAT 2D-LC Beckman Coulter System MudPIT

Phosphoproteome Profiling

Lipid Profiling

Identification and Characterization of Signaling Proteins Whose Expression is

Altered by Drug Addiction

SerumBiomarkers

Targeted Proteomics

Clinical Studies of Drug Addiction

Yale/NIDA Neuroproteomics ProgramYale/NIDA Neuroproteomics Program

Initial Stage:Initial Stage:- Meet individual PI’s to:

- Better understand the PI’s research area

- Jointly determine best to implement proteomics technologies available

from the Cores.

- Review of current technologies

- Development of new technologies

- Timelines

- Some already know Core personnel and have already initiated studies in which

case please continue working with these staff and additionally, I would be very glad

to talk with you also if you have questions about other technologies.

Erol E. [email protected]

Primary Biological SamplesOrgan, Tissue, cell extracts, biological fluids, etc.

Protein Sample PreparationSub cellular components, fractionation, affinity enrichment, Protein complexes, etc.

Protein Sample Simplification (or Purification):

Top Down MS or MS/MS

Protein Sequencing, and PTM

MudPit MS/MSHT Protein ID &

Quantitation

Digestion

RP LC MS/MSProtein sequence ID, PTM and Quantitation

Peptide Mass Fingerprint (PMF)

Protein ID

Protein Labeling ChemistriesICAT, PhIAT, SILAC,18O proteolysis, etc.

(Non LC) MS/MSProtein

sequence ID, PTM

Digestion

Predominantly ESI FTMS

Predominantly MALDI TOF MS

Predominantly ESI QTOF & IT (3D and 2D), MALDI TOF/TOF MS.

Also FTMS and Triple-Q (incl. Quad. Trap)

Predominantly ESI QTOF& IT (3D and 2D)

Enrichment step and/orPeptide Labeling Chemistries

(SCX, IMAC, iTrAQTM etc.)

1D-SDSPAGE

2D-SDSPAGE DIGE IEF FFEProtein LC

(Q, S, RP, SEX, etc.

Digestion

SELDI orMALDI

TOF MSBiomarker

patterns

Digestion

ProteinsPeptides

LC, M

S and Tandem-M

S

Commonly Used Protein Profiling Workflows in Mass Spectrometry







Quantitation

Digestion



Protein ID



sequence ID, PTM

Digestion








1D-SDSPAGE



Digestion

SELDI orMALDI

TOF MSBiomarker

patterns

Digestion

ProteinsPeptides

LC, M

S and Tandem-M

S







Quantitation

Digestion



Protein ID



sequence ID, PTM

Digestion








1D-SDSPAGE



Digestion

SELDI orMALDI

TOF MSBiomarker

patterns

Digestion

ProteinsPeptides

LC, M

S and Tandem-M

S

Commonly Used Protein Profiling Workflows in Mass Spectrometry

Protein Profiling Overview

Christopher M. Colangelo, PhDDirector of Protein Profiling

300 George StreetRoom G006

(203) 737-2636 (W)[email protected]

Brief Overview of the Keck LaboratoryFounded in 1980:

– 40 Full time staff– 100 Major instrument systems purchased at a cost of >$11

million dollars– 25,000 ft2 of space– 12 individual Resources that provide a wide range of proteomic

and genomic biotechnologiesCompletes 270,000 protein and nucleic acid syntheses and analyses annually for 260 Yale and 640 non-Yale investigators at 300 institutions & companies in 27 countriesNHLBI Proteomics Center (2002)NIH High End Instrumentation Grant (2002): FTICR-MSNational Biodefense Center of Excellence: Proteomics Core (2003)NIH High End Instrumentation Grant: “Center for High Performance Computation In Biomedicine” (2004)NIDA Neuroproteomics Center (2004)

(http://keck.med.yale.edu/)

Disease Based ProteomicsTranslational R

esearch







Lipid Profiling


Up/Down Regulated

SerumBiomarkers

Targeted Proteomics

Clinical Studies of Biomarker Efficacy

Keck MALDI-MS Based Disease Biomarker Discovery Sera from >50 disease patientsSera from >50 normal patients

Automatically desalt each sample (submitted in 96 well plate format) using a Micromass MassPrep robot and C-18 Zip-Tips.

Automated MALDI-MS on Micromass MALDI-L/R instrument: 130,000 m/z vs intensity datapoints acquired from 700 to 28,000 Da.

Optimize marker selection by increasing the size of the training set until the ability to correctly classify a testing set is maximized

Use a Random Forest Algorithm customized by Keck Biostatistics Resource (Wu et al , 2003) to identify 20-40 peptide markers whose intensities can be used to maximally

discriminate all normal from disease samples.

Linearly normalize baseline-corrected intensities in each spectrum to the overall median for all normal + disease spectra in the training set.

Filter out noise and all datapoints which fail to pass peak and unique peptde ion criteria.

Median MALDI-MS Spectra for Sera from 47 Ovarian Cancer (top) vs 42 Control (middle) Patients and the Resulting

Difference Spectrum (bottom)

m/z Versus Intensity Distribution for 129 Samples Around the 2084 Ovarian Cancer Marker

(relative importance = 2.9)

Median Cancer

Median Control

MarkerPosition

Impact of Training Set Size & Number of Biomarkers on Classification of 34 "Unknown" Ovarian Cancer Sera

Training Set Size: 102 Sera#Biomarkers = 25

50%

55%

60%

65%

70%

75%

80%

50 75 100 125 150

#Samples in Training Set

Cor

rect

ly C

lass

ified

Tes

t Ser

a

60%

65%

70%

75%

80%

0 20 40 60 80 100

Number of Biomarkers

%C

orre

ctly

Cla

ssifi

ed T

est S

era

Genomics – The blueprint of lifePCRDNA Central Dogma

Crick (1970) Nature 227, 561-563Crick (1958) Symp. Soc. Exp. Biol. 138

Mullis (1985)

Human Genome SequenceWatson and Crick (1953) Nature 171, 737-734

The International Human Genome Mapping Consortium (2001) Nature 409, 934 - 941 Venter et. al. (2001) Science Feb 16 2001: 1304-1351

What is proteomics?

Genomics – provides us with the words, but doesn’t provide us with any meaning

Proteomics - provides us with the definitions for the genome

The identification, characterization and quantification of all proteins involved in a particular pathway, organelle, cell, tissue, organ or organism that can be studied in concert to provide accurate and comprehensive data about that system.

Protein Profiling – provide us with the grammar and syntax to form a meaningful language

Understanding how proteins interact with each other, their environment, and function with outside molecules

Challenges - proteomics and protein profiling are dynamic languages

How do we bridge the gap from genomics to proteomics?

Genomics

Proteomics

Gene MicroarraysFirst step towards bridging the gap between human genome and function

Keck Human 4.6K cDNA Array Cluster Analysis

mRNA vs. Protein correlation

(R=0.66)

Comparing protein abundance and mRNA expression levels on a genomic scaleGreenbaum D, Colangelo C, Williams K, Gerstein MGenome Biol. 2003; 4(9): 117. Epub 2003 Aug 29.

Reality check for proteomicsAvogadro’s number - 6 x 1023 molecules/moleFor MS proteomics we want 100 fmoles (10-14 moles) = 6.02 x 1010 molecules

cell copy number cells needed for 100 fmoles10 6 billion (6.02 x 109)

100 600 million (6.02 x 108)1000 60 million (6.02 x 107)

10,000 6 million (6.02 x 106)100,000 0.6 million (6.02 x 105)

1g of tissue = 108 cells1000 copy number limit

10cm culture dish = 105 cells100,000 copy number limit

Levels of proteins in plasma

Molecular & Cellular Proteomics 2003 , Anderson and Anderson 2 (1): 50

Protein Profiling


esearch







Lipid Profiling


Up/Down Regulated

SerumBiomarkers

Targeted Proteomics


Cleavable ICAT™ Reagents

Provide 9 Dawhich provides

ratio of heavy/light

peptide conc.

Removes biotin which

improves MS/MS

fragmentation

Reacts with free cysteines

Allows selection and concentration of cysteine-containing peptides

Applied Biosystems

Differential expression analysis on human leukemicprecursor cells vs. differentiated monocytes/neutrophils

16 SCX fractions collected

Cys-containing peptidesaffinity purified/ cleaved with TFA

100ug of protein is labeled with acid cleavable ICATTM

reagent/trypsinized

Online LC-MS/MSanalysis on QStar®

XL System

ICAT-Determined Expression Ratio of Zinc Finger Protein 9 in Two Human Cell Lines

Overall Average Expression Ratio of Protein 9 in Two Human Cell Lines:

3.4 fold decrease(average H/L = 0.297 +0.061)

Web-based query results from Yale Protein Expression Database (YPED)

http://genome2.biology.yale.edu/yp_results/logon.jsp


esearch







Lipid Profiling


Up/Down Regulated

SerumBiomarkers

Targeted Proteomics


Automated 2D Comparative LC Protein Profiling(Beckman-Coulter PF2D Platform)

Partial pI/UV map of a colon cancer cell line before and after drug treatment. The RP-HPLC profiles illustrate differences in the pI 6.0-6.2 fraction which are shown also by the color coded band depictions of these RP-HPLC profiles (from Beckman-Coulter).

Two-dimensional Comparison Plot of

NB4 (Green) and NB4 + RA (Red)

(Neutrophil) from Beckman PF2D

System

| | | | | | | | | | | | | | | | | | | | |1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Fraction

pH8.6 7.9 6.4 5.2 4.3

Ret

entio

n Ti

me

22-

20-

16-

18-

12-

14-

10-

5 mg of total protein

The 2D overlay plot shows the relative differential expression for each of the 21 pH fractions based on overlay comparisons of 214 nm reversed phase chromatograms.

Overlay of NB4 and NB4+RA Reversed Phase Chromatograms from Fraction 7

The differential peaks in the UV chromatogram represent intact proteins that are expressed or modified between control and treated samples.

Retention Time

Abs

orba

nce

(211

nm

)

Life Science Tools is our Business™

APEX FTMSAPEX-Q Industrial Design

Typical Project Milestones

0 1 2 3 4 5 6 7 8 9 10 11 12Weeks

AAA results

Differential Gel

Preparative gel

Mass Spectrometry

Data AnalysisDIGE

AAA results

ICAT Labeling

Cation-exchange

Mass Spectrometry

Data AnalysisICAT

AAA results

2D Chromatography

ICAT Labeling

Mass Spectrometry

Data Analysis

PF2D


esearch







Lipid Profiling


Up/Down Regulated

SerumBiomarkers

Targeted Proteomics


2D Electrophoresis

2D gel electrophoresis a “Gateway” proteomic tool for the separation of complex protein mixtures from many different biological sample types

Pros: -High resolution separation by pI and MW -Possible to resolve several thousand proteins on a single gel-IPG strips feature high reproducibility and resolution (single pH unit)-PTMs, such as phosphorylation and glycosylation are clearly visualized

Cons:- Membrane/ hydrophobic proteins have solubility problems

a) can be addressed by detergents, e.g., ASB-14, Triton X-100b) some proteins are prone to precipitation at their IE pointc) try an alternative technology, such as ICAT

-Gel-to-gel reproducibility and therefore Quantitation is a problem

DIFFERENTIAL FLUORESCENCE GEL ELECTROPHORESISPROTEOME PROFILING (DIGE)

DIGE Detection Limits

# of cells protein copies/cell total # of proteins moles of protein ng of 50kd

1 x 108 1000000 1014 160pmole 8,304(~1 g of tissue) 100000 1013 16pmole 830

10000 1012 1.6pmole 831000 1011 160fmole 8.2100 1010 16fmole 0.8

10 109 1.6fmol 0.08DIGEMS

DIGE method is highly reproducible…but only with good sample preparation!

-use of label compatible buffers7M Urea, 2M Thiourea, 4% CHAPS, in 30 mM TRIS adjusted to pH 8.5

-ensure that the sample preparation is homogenous

-”clean”-up of sample via precipitation, dialysis, nuclease

-low abundance issues are best addressed at the sample preparation stepa)subcellular fractionationb)directed biochemical enrichments/pre-fractionationc)microdissection/punches of specific tissue regionsd)removal of high abundance species

-quantitate samples prior to analysis to normalize, AAA is best

-ensure there is enough sample for downstream studies!

Albumin and IgG depletion from serum

cy 3 image

cy 5 image

spots with a greater or = to 3X increase (Blue) or Decrease (Red)In cy5/cy3 spot volume ratio

Documents

W.M. Keck Foundation Biotechnology Resource Laboratories Erol … · 2019-09-29 · Automated 2D Comparative LC Protein Profiling (Beckman-Coulter PF2D Platform) Partial pI/UV map