1

Introduction to Proteomics

Åsa Wheelock, Ph.D.Division of Respiratory Medicine &

Karolinska Biomics Centerasa.wheelock@ki.se

In: Systems Biology and the Omics Cascade, Karolinska Institutet, June 9-13, 2008

2

Focus of course: Tools for data analysis

Your analysis is no better than data you have collected...

The goals of this proteomics overview:• Understand possibilities & limitations • Pros and cons of different method• Sources of variance in proteomics• Take advantage of proteomics core facilities • Perform proteomics collaborations • Write a short research proposal in

proteomics

3

Proteomics publications in Pubmed

0

500

1000

1500

2000

2500

3000

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

First ”proteomics”publication

4

Why Proteins?!?• Business end of the cell• Detailed information with limited efforts

– As compared to metabolomics

TRANSCRIPTOMICS

PROTEOME

METABOLOME

Limited info from mRNA

Detailed infoRobust technology

Techincally challenging

• Relatively robust methods available

5

Proteomics Methodology• No ”protein PCR”

– 4 nucleotides vs 20+ amino acids– Post-translational modifications (PTM)

3 MAIN PROTEOMICS PLATFORMS• Gel based methods • Shotgun methods (mass spec-based)

”chromatography-based”, ”gel-free”• Array based (antibody based)

6Klose, J. 1975. Humangenetic 26, 231-43O’Farrel, P. 1975. J. Biol.Chem, 250, 4007-21

Gel based: 2-Dimensional Electrophoresis

Isoelectric point (pI)

Net

cha

rge

3 4 5 6 7 8 9 10 11 pH

+2+1

0-1-2

log

Mw

mobility

Molecular weight

SEPARATION VISUALISATION

QUANTIFICATION

Stoichiometric protein stain => 3rd dimension

7

Image acquisition

Image acquisition using fluorescent scanner

8

Quantitative image analysis

1. Detect spots

2. Match spots across gels

3. Quantify spot volumes

Pixel intensity => 3rd dimensionSpot volume = protein quantity

9

Protein identification

⇒ Mass spectrometry(MALDI-TOF/TOF)

Trypsin digestion

⇒ DATABASE SEARCH => IDENTIFICATION(Swiss-prot, EnSemble) (statistical probability)

10

Protein Identification

Protein Trypsine digestion Peptides EXPERIMENTAL Peptide masses

Protein database

DTHKSEIAHRFKDLGEEHFKGLVLIAFSQYLQQCPF DEHVKLVNELTEFAKTCVADESHAGCEKSLHTLFGDELCKVASLRET

Virtual digest

MS analysis

EXPECTED Peptide masses

Statistical matching

11

Peptide mass mapping

LHTLFGDR

MS/MS analysis=> sequence information

MS analysis=> peptide masses

Statistical matching

DLGEEHFK database search

DTHKSEIAHRFKDLGEEHFKGLVLIAFSQYLQQCPF DEHVKLVNELTEFAKTCVADESHAGEKSLHTLFGRELCKVASLRET

Homology search Validate statistical hit

12

Shotgun vs. Gel-based proteomics

Adapted from Patterson and Aebersold, Nature Genetics 2003, 33:311-23. Fig. 3

extract

Separate Quantify

Separate(LC)

Digest

Digest

MS/MS

MS/MS

(2DE)

ID

Quantify

13

Semi-quantitative proteomics

Both 2DE and MS-based methods NOT quantitative by nature

Co-separation: 2 samples => ratiosTags => Semi-quantitative proteomics

14

Shotgun isotope-tagging :Isotope coded affinity tag (ICAT)

15

Multiplexing in 2DE: DIGEMultiplexing in 2DE: DIGE-- Differential Gel ElectrophoresisDifferential Gel Electrophoresis

Protein abundance = 532/633 signal ratio

Co-separation by 2DE

Direct ratiometric normalization

Spot quantification

Statistical analysis

Protein visualization (super-imposable images)

λex 633nm ⇒ Cy5λex 532nm ⇒ Cy3

Treated sample: Covalently labeled (Cy3)

Control sample: Covalently labeled (Cy5)

16

Semi-quantitative proteomics

Both 2DE and MS-based methods NOT quantitative by nature

Co-separation: 2 samples => ratiosTags => Semi-quantitative proteomics

Pooled internal standard + 2-3 samples => Relative quantification

17

Internal Standards in 2DE: DIGEInternal Standards in 2DE: DIGE

Protein abundance = relative to IS

Co-separation by 2DE

Direct ratiometric normalization

Spot quantification

Statistical analysis

Protein visualization (super-imposable images)

λex 633nm ⇒ Cy5λex 532nm ⇒ Cy3

Treated Sample: Covalently labeled (Cy3)

Control Sample:Covalently labeled (Cy5)

λex 488nm ⇒ Cy2

Pooled Internal standard (Cy2)

18

Proteomics in Pubmed

0

500

1000

1500

2000

2500

3000

1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006

proteomics 2DE

First ”proteomics”publication

First DIGE-method

published

19

Sample A

Trypsin digestion (peptides)

31114

PRG

30

115

PRG

29

116

PRG

28

117

PRG

Multiplexing in MS: iTRAQ- isobaric Tag for Relative and Absolute Quantitation

Sample B Sample C Sample D

a b c diTRAQ labelling

MS/MS m/z

ID

Quantification

20

Differential labelling opens up new possibilities

• Cysteine oxidative states

• Identify peptides on plasma membrane surface

• Cellular re-localization

21

2D or not 2D?2D or not 2D?Gel-based methods: 2-D electrophoresis

+ soluble proteins

+ post-translational modifications

22

Post-translational modifications”Spot trains”

Intact proteins

23

2D or not 2D?2D or not 2D?Gel-based methods: 2-D electrophoresis

+ soluble proteins

+ post-translational modifications

- technical variance, time consuming

MS-based (Gel-free) methods: ICAT, iTRAQ

+ membrane proteins

+ low abundance proteins

24

Extremes of physiochemical properties: Peptides

• Charge

- pI range from 3-12

• Size

- Mw range of 5 – 500,000 kDa

• Hydrophobicity

- membrane proteins

25

2D or not 2D?2D or not 2D?Gel-based methods: 2-D electrophoresis

+ soluble proteins

+ post-translational modifications

- technical variance, time consuming

MS-based (Gel-free) methods: ICAT, iTRAQ

+ membrane proteins

+ low abundance proteins

- expensive, data intense

26

Shotgun approcahes and gel-based approaches complementary

No ”true” proteomics technique yet

SHOTGUNGEL-BASED

27

日本Japan

103

100 101 102 103 104 105 106 107 108 109 1010 1011 1012

COPIES of each

PROTEIN

Osaka 大阪

Kyoto 京都

Kobe 神戸

104

1012

DYNAMIC RANGE

28

- 400 reported PTMs

Protein copy numbers in blood (log10)0 1 2 3 4 5 6 7 8 9 10 11 12

Phos

phor

ylat

ion

site

s

7

6

5

4

3

2

1

Post-translational Modifications (PTMs)

29

Variance in 2DE• BIOLOGICAL VARIANCE• Experimental variance

– Pre-fractionation, isolation & labelling of proteins – Protein staining

• Technical variance– Gel-to-gel variation in 2DE– Image acquisition (scanner)

• Post-experimental variance– Software-induced variance– User dependant variance

30

Variance in 2DE• BIOLOGICAL VARIANCE•• Experimental varianceExperimental variance

– Pre-fractionation, isolation & labelling of proteins – Protein staining

• Technical variance– Gel-to-gel variation in 2DE– Image acquisition (scanner)

• Post-experimental variance– Software-induced variance– User dependant variance

31

Remember that variance adds up:Multiple-step method is not your friend...

10% 40%

32

Variance in 2DE• BIOLOGICAL VARIANCE• Experimental variance

– Pre-fractionation, isolation & labelling of proteins – Protein staining

• Technical variance– Gel-to-gel variation in 2DE– Image acquisition (scanner)

• Post-experimental variance– Software-induced variance– User dependant variance

33

Technical variance in 2DE

Solubilization Rehydration Isoelectric focusing

SDS-PAGELoad IPG strip Protein visualization

Recovery?

Recovery?

Inhomogenouselectric field?

Gel-to-gel variations?

-SH modifications?

Staining artifacts?

Backgroundfluorescence?

Biological variance

34

Tools to reduce variance

Technical variance• Internal standard:

– DIGE

• Software algorithms:– Background subtraction– Normalization

35

Dynamic range of scanner

16 bit pixel resolution (216 ∼ 65,000 ~ 105)Make sure you are using the entire range!

36

Variance in 2DE• BIOLOGICAL VARIANCE• Experimental variance

– Pre-fractionation, isolation & labelling of proteins – Protein staining

• Technical variance– Gel-to-gel variation in 2DE– Image acquisition (scanner)

• Post-experimental variance– Software-induced variance– User dependant variance

37

2DE analysis software

• Main purpose: match and quantify spots• Normalization: reduce gel-to-gel variation• Background subtraction:

– Reduce background noise– Increase signal/noise ratio– Increase sensitivity

38

39

40

Global Background SubtractionGlobal Background Subtraction

PDQuestPDQuest: Floating/Rolling Ball: Floating/Rolling Ball

41

Software induced variance Software induced variance

Expression; No background subtraction

0

25

50

75

100

1 51 101 151 201 251 301 351 401 451

CV

(%) n

=5

Expression; No background subtraction

0

25

50

75

100

1 51 101 151 201 251 301 351 401 451

CV

(%) n

=5

Average CV=4.6%Average CV=10%

PDQuest PG200PD Quest; Power mean

0

25

50

75

100

1 51 101 151 201 251 301 351

CV

(%) n

=5

PD Quest; Power mean

0

25

50

75

100

1 51 101 151 201 251 301 351

CV

(%) n

=5

Software variance up to 30% of technical variance

42

Applications of proteomicsApplications of proteomicsBIOMARKER DISCOVERY• Biomarker of disease & susceptibility

CLINICAL APPLICATIONS• Pharmaceutical target identification• Improved diagnostics

MECHANISTIC STUDIES• Protein-protein interactions• Protein adduction /Altered protein expression• Hypothesis generation: avoid local ”maxima”• Systems Biology

43

Proteomics in the future

• Improved sensitivity– Currently: scratching the surface– laser capture microdissection

• Protein microarrays– Antibody arrays (e.g. for cytokines)– Tissue microarrays (Peter Nilsson, Friday)

• In vivo subcellular localization assays• Protein amplicifation method?

– i.e. ”protein-PCR”

44

Focus on INTERPRETINGINTERPRETING data,not on ACQUIRING ACQUIRING data.

Pathway Analysis• Integrate data from omics cascade• Integrate heatmap with biological pathways

Proteomics in the NEAR future...

45

Preview of coming attractionsPreview of coming attractions……KEGG & KEGG & KegArrayKegArray

46

Take home messages...

...keep your variance downand your dynamic range up!

...keep your false positives down,and your power up!