Quantitative Proteomics without gels

1
Quantitative Proteomics without gels (iTRAQ)
technology Bill Simon and Toni Slabas

Proteomics Facility, School of Biological and
Biomedical Sciences, University of Durham, UK
2
Why do we need quantitative proteomic analyses

• comparison between tissue types.
• biomarker discovery - compare healthy and
  diseased states.
• response to drug or pathogen treatment.
• study stress responses.

High throughput
3
Is 2D gel technology dead ?
4
MudPIT Multidimensional Protein Identification
Technology Shotgun analyses driving force
speed and accuracy of mass spec.
The problem is with the complexity of the mixture
and duty time in the mass spectrometer Advantage
s ?
5
Advantages

• Avoid potential problems with gels.
• Diversity of proteins analysed- hydrophobic,
  acidic and basic proteins.
• Wide dynamic range transcription factors
  storage proteins.
• Membrane protein analyses.
• Smaller amounts of starting material 100mg
  whereas 1mg for gels.
• Quantification made at the peptide level.
• Can handle complex mixtures.

6
What is the Instrumentation and Technology
? Chromatography Mass spectrometry
MudPIT
Buffers and Salt
7
Instrumentation and technology required for
successful quantitative proteomics via mass
spectrometry
mass spectrometer (Qstar tandem MS)
nano-flow HPLC (MDLC system)
8
nanoflow HPLC system (MDLC) high throughput flow
path
Buffers and Salt
LC is either desalting samples or running
gradient what is mass spec doing ?
gradient elution and MS analysis
C18 trap column
C18 capillary analytical column Zorbax 300SB-C
18, 3,5um, 15X 0.075mm
high voltage nano-spray needle
nano flow rate (nl)
nanoflow HPLC system / mass spectrometer Duty
Time
9
QStar triple quad TOF mass spectrometer
high voltage nano-spray needle

• Precursor ion scan (peptide mass analyses) -
  Ions are focused in Q0, filtered according to
• mass range (300-3000) in Q1 and then released
  into the TOF detector for mass measurement.
• Product ion scan (peptide sequence analyses)
  ions are focused in Q0, filtered according to
• individual peptide ion mass in Q1, fragmented
  in Q2 and fragment ions released into the TOF
• detector for mass measurement.

10
How Much Can This Machine Do ?
peptides continuously eluting from nano-LC
column for 120 minutes
Several peptide eluting at the same time
detect peptide ions
mass spectrometer continuously acquiring data
fragment ion 1 fragment ion 2 fragment ion 3
cycle time 10 seconds
Duty Time
11
What does QStar Data Look Like ?
total ion chromatogram
Several peptide ions eluting
1st most abundant
10 second cycle time
2nd most abundant
3rd most abundant
12
Tandem Mass Spectrometry and Peptide Sequence
A B C D E F
F E D C B A
b6 b5 b4 b3 b2 b1
y6 y5 y4 y3 y2 y1
single peptide
Fragment ions are designated b or y series
depending on which end of the peptide they are
derived from (N and C-terminus respectively)
13
MudPIT Allows Identification But Not
Quantification
Solved salt and duty time problem
14
Strategies available for mass spectrometer based
proteomic quantification

• SILAC Stable Isotope Labelling with Amino
  Acids in Cell Culture cells grown in
• the presence of light
  (control) and heavy (treated) amino acids
  measuring
• ratios at the peptide level via MS data .
• ICAT - Isotope Coded Affinity Tags - two
  reagents heavy and light
• labelling done at the protein level
  cysteine containing peptides
• affinity selection quantification via MS
  data.
• iTRAQ Isotope Tagging Reagents for Relative
  and Absolute Quantificaton -
• four reagents multiplexing - labelling done
  at the peptide level
• all peptides are labelled quantification
  done via MS-MS data.
• multi-dimensional LC-MS approach.

Ross et al., (2004) Molecular and Cellular
Proteomics 31154-1169
15
iTRAQ Isotope Tagging Reagents for Relative and
Absolute Protein Quantification
amine specific reagents peptide N-terminus and
lysine side chains are labelled
16
iTRAQ and
Labelling
collision induced fragmentation
Peptide Identification sequence ions combined
from all four peptides
Quantification
114
31
115
30
116
29
117
28
17

4 accurately quantified protein samples
reduce alkylate and digest with trypsin to
produce peptide pools
label each peptide pool with an iTRAQ
reagent (114 115 116 117)
pool peptide samples
iTRAQ Workflow
cation exchange fractionation usually around 20 -
30 fractions
nanoLC-MS-MS 3 hours for each cation fraction
identification
bioinformatics ProQuant ProGroup processing
software
20,000 MS-MS spectra
18
Thats the theory what about practice ? Four
labels allows you to look at four samples
multiplexing If two of the samples are
identical then label ratio should be 11
19
Use of iTRAQ reagents in a quantitative heat
shock analyses of wild type Synechocystis and a
thermally tolerant knockout strain lacking the
histidine kinase 34 gene.

• Synechocystis PCC 6803 is a unicellular
  photosynthetic prokaryotic algae used extensively
  in stress response studies using both proteomic
  and genomic technologies.
• Complete genome sequence available (3168 genes).
• Extensive library of gene knockouts available.
• cDNA arrays of almost the complete genome
  available

20

• Background
• Radio-labelling data shows that major proteomic
  changes occur within 1
• hour of heat shock.
• Transcriptome and DIGE data have identified
  changes in classical heat
• shock responsive proteins along with proteins
  involved in other
• metabolic processes.
• The Hik 34 knockout strain shows increased
  thermal tolerance and has
• elevated levels of heat shock proteins under
  both the control temperature
• and heat shock conditions.
• Hik34, or a down stream component in the
  signalling pathway of which
• it is a component is a negative regulator of
  heat shock responsive genes.

Suzuki et al (2005) Plant Physiol. 1381409-1421,
Slabas et al., (2006) Proteomics 6(3) 845-864.
21
What did we label and why ?
100 mg soluble proteins from each treatment.

• WT at 320C - iTRAQ 114 reagent.
• WT at 42oC - iTRAQ 115 reagent.
• Hik 34 at 32oC - iTRAQ 116 reagent.
• Hik 34 at 42oC - iTRAQ 117 reagent.
• Allows for more information
• Three independent biological experiments

22
Cyanobacterial (Synechocystis sp. PCC6803) system
complexity
3000 genes 1000 membrane proteins - 2000 soluble
proteins 50 expressed in this analysis 1000
proteins 20 - 30 tryptic peptides from each
protein 20000-30000 peptides 25 cation
exchange fractions gt 1000 peptides in each
fraction 120 minute nano-LC gradient gt 10
peptides a minute into the MS
23
First HPLC- cation exchange offline
not peptide resolution - simplification of
complexity
resulted in 24 x 200ml fractions
24
Second HPLC- RP nano LC online with the Qstar
mass spectrometer
Only one third of each cation exchange fraction
loaded
Fraction 5
Fraction 7
Fraction 6
Peptides were eluted from a 75mm C18 capillary
column with a linear gradient 0-42 ACN, 0.1
formic at 0.4min at a LC-MS flow rate of approx
200nl min.
25

• Analyses of Peptide Complexity in RP-HPLC
• At any time point more than one peptide elutes
• Peptides at any single time point are identified
  and quantified
• Each time point contains multiple peptides up
  to 3 can be sequential sequenced.
• How many peptides can you identify ?

26
Total ion count spectra for the three MS-MS
experiments from a single RP separation of one
cation fraction
1st most abundant
2nd most abundant
3rd most abundant
Each cation fraction typically resulted in
between 1000 - 1200 MS-MS spectra
27
Fragment ion spectra of a single iTRAQ labelled
peptide
iTraq reporter ions
Sequence ions
28
iTRAQ reporter ion region of fragmentation
spectra used for quantification
hik 32oC
Wt 32oC
hik 42oC
Wt 42oC
29
Bioinformatics / data processing
each experiment (4 treatments)
24 fractions x 2 hours MS data 3 x MS-MS spectra
every 10 seconds
identification and quantification on all MS-MS
spectra ProQuant
consolidation of data from all fractions ProGroup
ability to mine data
comparison of 3 independent experiments
30
Data identification and quantification using
ProQuant software.
Focuses into the 114 117 region of each of the
MS-MS fragmentation spectra and quantifies the
peak areas for all four tags. Determines amino
acid sequence data from the fragment ion data in
each of the MS-MS spectra. Uses a novel
sequence matching algorithm in a database search
to match the peptide amino acid sequence and
produce a protein identification. Tabulates all
of the peptide identification data with the
quantification data and produces an output with a
statistical measure of both the
identification and quantification data.
31
Data
Simply too much to handle Human computer
insufficient
32
ProQuant data output from a single fraction from
1 experiment
protein level
33
ProQuant data output from a single fraction from
1 experiment
peptide level
34
Link to MS-MS spectra for peptide sequence
evidence
sequence evidence
35
Link to MS-MS spectra for peptide quantification
evidence
quantification evidence
36
ProGroup data refinement and data mining

• Pulls together all of the ProQuant data from all
  of the fractions in a single experiment.
• Removes data redundancy caused by multiple hits
  for similar proteins.
• Generates best hit protein ID with a statistical
  confidence value ProtScore
• generated from the individual confidence value
  generated for each peptide match.

A protein with 3 unique peptide IDs 2 at 99
and 1 at 95 would have a ProtScore of 5.3.

• Generates a protein quantification value for all
  proteins identified in
• each treatment with the statistical confidence
  value (p-value) for each
• quantification measurement reported.
• Generates a coloured mapped output of the
  protein ID confidence and the
• quantification data.

37
Colour mapped output of the protein ID confidence
and the accumulate quantification data from all
fractions of a single experiment.
38
iTRAQ data format in ProGroup
WT 42oC v WT 32oC
Hik 32oC v WT 32oC
Hik 42oC v WT 32oC
protein ID confidence score (a single peptide at
99 confidence scores 2)
P value and error factor () for quantification
ratio average of all peptides measured
iTRAQ 114 reagent - WT at 320C iTRAQ 115 reagent
- WT at 420C iTRAQ 116 reagent - Hik 34 at 320C
iTRAQ 117 reagent - Hik 34 at 420C
39
ProGroup report statistics for protein
identification from the 3 individual
Synechocystis iTRAQ experiments
experiment 1
experiment 2
experiment 3
40
Summarised protein identification data from the 3
Individual experiments
62
approximately 67 of proteins identified were
found in all three experiments.
41
Analyses of quantification data from each of the
three experiments for a number of heat shock
responsive proteins
42
Summary

• Introduced gel free shotgun proteomic analyses
  using multidimensional nano LC-MS
• approaches.
• Demonstrated the use of these approaches combined
  with iTRAQ labelling in an analyses of
• the heat shock response of the unicellular
  cyanobacteria Synechocystis PC6803.
• This allowed gt 500 proteins to be identified
  from the soluble proteome of the organism
• at gt 99 confidence level from each of three
  independent experiments.
• A total of 654 unique Synechocystis proteins
  identified at this confidence level.
• Quantification and protein expression data
  available for all proteins identified.
• Demonstrated the MDLC approach together with
  iTRAQ labelling as a
• powerful tool for quantitative proteomic
  analyses.

43
What next ?
membranes