Quantification in MassSpectrometry: the who, thewhat and the how much…

Stephen Barnes, PhDTargeted Metabolomics and Proteomics

Laboratory

Synopsis• Quantification during discovery

– Spectral counting– ICAT (isotope-coded affinity tag)– iTRAQ– SILAC

• Discovery starting points• Multiple reaction ion monitoring (MRM)

– Advantages and limitations– Importance of bioinformatics and

computational support

• Hydrolyze everything!

• For a cell expressing 5,000 proteins,this leads to >100,000 peptides

• Can be fractionated, but still 10,000-20,000 to differentiate

• Enormous bioinformatics problem

MUDPIT - MUDPIT - MUMUlti-lti-DDimensionalimensionalPProtein rotein IIdentification dentification TTechnologyechnology

MS-MS analysison Qqtof

Massive computing

nanoLC

0-40% MeCNgradient

10 mM

20 mM

40 mM

60 mM

80 mM

100 mM

NH4Ac (1-200 mM step gradient)

200 mM

Cation exchangecolumn (H+)

John Yates

Spectral counting

• Spectral counting depends on how many peptidesfrom a protein are detected in a tandem MSexperiment

• A completely digested protein (1 pmol) shouldproduce 1 pmol of each of its constituent peptides

• However, the molar ionization intensity for thesepeptides varies at least 100-fold

• A low abundance protein will have only a fewrecognizable peptides

• It’s a robust method that doesn’t require specialreagents and is part of a discovery experiment

Isotope-coded affinity technology

This reagent reacts with cysteine-containing proteins (80-85% of proteome)

Labeling can be replacement of hydrogens (X) with deuterium, or better toexchange 12C with 13C in the linker region (this avoids chromatography issues)

iTRAQ™ Reagent Design

Amine specific

PRG

Peptide Reactive Group

Charged Neutral loss

Isobaric Tag(Total mass = 145)

Reporter Balance PRG

• Gives strong signature ion inMS/MS

• Gives good b- and y-ion series• Maintains charge state• Maintains ionization efficiency of

peptide• Signature ion masses lie in quiet

region

Reporter(Mass = 114 thru 117)

• Balance changes inconcert withreporter mass tomaintain total massof 145

• Neutral loss inMS/MS

Balance(Mass = 31 thru 28)

Isobaric Tag(Total mass = 145)

= MS/MS Fragmentation SiteIsobaric Tag

Total mass = 145

Reporter Group mass114 –117 (Retains Charge)

Balance GroupMass 31-28 (Neutral loss)

Amine specific peptidereactive group (NHS)

N

N

O

ON

O

O

Slide provided byApplied Biosystems

SILAC and proteinquantification

• Stable isotope labeling with amino acids in cellculture (SILAC)

• See http://www.silac.org• Amino acids such as leucine, lysine or tyrosine

are added as light (all 12C) or heavy forms (all 13C)– e.g., 13C6-lysine, 13C9-tyrosine

• These amino acids are incorporated into proteinsand increase masses of peptides

• Samples are mixed and digested with proteases• The intensities of the peptide pairs (M, M+6)

compared

Pulsed SILAC

Proteomics 2009, 9, 205-209

Limitations of quantitative methods• Spectral counting: depends on the number of

peptides produced from a protein which arerecognizable -– however, tandem MS is a stochastic process– can measure 2-fold differences

• ICAT: the proteins that have cysteines and onesthat are free and not oxidized are limited– the reagent is reacted with the intact proteins

• iTRAQ: reacts with lysines (more abundant thancysteines)– the large number of steps that are involved leads to

variable losses– Also, the label is added after the hydrolysis

• SILAC: requires good chromatography– best for cells - too expensive for animal studies

On the way to absolutequantification

• All forms of proteomics and mass spectrometryare a compromise

• At best only a small part of the proteome can beobserved due to sample solubility, complexityand concentration (dynamic range of 9-10 ordersof magnitude), as well as instrument detectionbias

• If we “know” what we want to measure, then wecan focus our attention on a limited, but ifcarefully selected, important group of proteinsand determine them at much higher sensitivity

Protein/peptideseparation + mass

spectrometry

Bioinformaticsanalysis

Signaling andprotein complexes

analysis

Quantitative MRManalysis

Gel-LC for proteinseparation

Workflow for generation of proteomics data

Microarrayanalysis

Biological andexperimentalknowledge

2D-DIGE andother

electrophoresis

microRNAanalysis

Multiple Reaction Monitoring

Q0 Q1 Q2 Q3

LIT

Q0 Q1 Q2 Q3

LIT

FragmentPeptide Y or B ion

– Highest specificity and sensitivity for detectingcomponents in a complex mixture

– Largest linear dynamic range for quantitation– Well accepted as the MS technique for quantification

(small molecule world)– Triple quadrupole or Q TRAP® system mass

spectrometers

Quantification of peptides• MRM-MS collects data points to describe a peak• Enough ions can be collected in 10-20 msec,

enabling a switch to a different parention/fragment ion combination (“channel”) for thenext 10-20 msec, etc.

• When the list of peptide measurements has beencompleted, then we go back to the top of the listfor the next cycle (multiple reaction ionmonitoring - MRM)

• The data points (10 per peak) are used to recreatethe peak and hence measure the area underneathit.

Oxidized tryptic peptide aa57-69 from hαB-crystallin

A P S W F D T G L S E M R

200 400 600 800 1000 1200 1400m/z0

10

20

30

40

50

60

70

80

90

100b6

736.2

b3256.1

b4474.2 b5

621.3

b7837.3

b91007.5

b111223.5

b121370.5

O O O

169.0 256.1 474.2 621.3 736.3 837.3 894.4 1007.5 1094.5 1223.5 1370.6 b2-b13

y4538.3

y2322.2

y3451.2

y5651.4

y6708.3

y7809.4 y8

924.5 y91071.5

y111376.5

1473.6 1376.6 1289.6 1071.5 924.5 809.4 708.3 651.4 538.3 451.2 322.2 175.1

y1-y13

756.8/793.457APSW*FDTGLSEMR69

764.8/793.457APSW**FDTGLSEMR69

nanoLC-MRM-MS analysis reveals that UVA lightoxidation of tryptophan occurs at several sites

Addition of one Oleads to fivepossible isomers

Addition of two Oleads to tenpossible isomers

756.8/809.457APSWFDTGLSEM*R69

Single isomer

[O]

Max. 5166.7 cps.

5 10 15 20 25 30 35 40 45 50 55 60 65 70Time, min

0200400600800

1000120014001600180020002200240026002800300032003400360038004000420044004600480050005167

Intensity, cps

33.57

LAPAVSPEK MRM Parent ion: 456.26++ m/z Daughter Ion: 559.3 m/z (y5)

Cystin peptide detected in whole kidney digest by MRM-MS

Choosing the precursor andfragment ions for MRM

• Without preliminary data– Use the Sciex software program MIDAS - it

interrogates a library of known peptides andselects the predicted most intense ions

– Our experience is that it is a poor predictor• With preliminary data

– Inspect the peptide MSMS spectra that wehave previously obtained to identify proteins

– These spectra are “guaranteed” to work sincethey are obtained on the same instrument;however, the Qtrap and the triple quadcollision parameters are not the same

DIGE of two cell protein extracts

Red: cell 1; Green: cell 2

Partial list of protein identities of spots indicated to be significantlydifferent between cell 1 and cell 2 (rescued with Polaris)

Spot # Protein Identification MOWSE Accession mw (obs) mw (pred) pI (obs) pI (pred) T-test ratio R/Cscore number kDa kDa

83 No identificaiton N/A N/A 250.00 N/A 5.75 N/A 1.60E-03 -2.03

211 gelsolin, cytosolic 49 gi|90508 101.00 81.00 5.85 5.53 2.16E-05 -2.02

212 gelsolin, cytosolic 203 gi|90508 100.00 81.00 6.01 5.53 7.98E-05 -2.05

404 Keratin 8 502 gi|76779293 54.00 55.00 5.75 5.71 7.30E-07 -2.52

406 Keratin 8 995 gi|76779293 50.00 55.00 5.75 5.71 4.17E-06 -5.5

412 Keratin 8 988 gi|76779293 52.00 55.00 5.83 5.71 6.02E-07 -3.77

468 Keratin 18 320 gi|6754482 43.00 47.00 5.31 5.31 1.10E-04 -2.47Keratin 20 269 gi|21592285 49.00 5.31

476 hypothetical protein LOC100037084 678 gi|147899980 45.00 47.00 5.32 5.22 4.69E-06 -3.24(Keratin 18)

477 hypothetical protein LOC100037084 1251 gi|147899980 47.00 48.00 5.40 5.22 1.65E-08 -4.44(Keratin 18)

490 eukaryotic translation initiation factor 4A isoform 1 471 gi|4503529 43.00 46.00 5.60 3.40E-02 -3.47hypothetical protein LOC100037084 (Keratin 18) 384 gi|147899980 48.00 5.22Keratin 20 344 gi|21592285 49.00 5.31

Being selective - pragmaticallychoosing the MRM ions from existing

tandem spectra

5 10 15 20 25 30 35 40 45 50 55 60 65Time, min

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

1920

1- Phosphoglycerate kinase2- Gelsolin-like Capping Protein3- Keratin 184- Keratin 205- Aldolase A6- ERp61 7- Enolase 1, alpha8- Keratin 199- (3)-Phosphoglycerate dehydrogenase10-Glyceraldehyde-3-phosphate dehydrogenase

11- Elongation Factor Tu12- Pyruvate kinase M13- Beta actin14- Alpha actin15- Tubulin, beta 516- Keratin 817- ATPase beta subunit18- Alpha tubulin19- Heterogeneous nuclear ribonucleo protein K20- Keratin 7

Combination of allthe MRM channels

Proteins indicated to be of interest by 2D DIGE detected inone MRM run, with forty channels set, two for each peptide

21 days 50 days

100 days

Decreasing amount ofbeta B1 crystallin in thelens with the age of therat - note the BSApeptide as the internalstandard

Day 21 Day 50

Day 100

Increasing amount ofbeta B2 crystallin in thelens with the age of therat - note the BSApeptide as the internalstandard

Linearity of the area response for the peptide from astarting protein (alphaA-crystallin)

Selecting an internalstandard

• A protein that is not present in the sample of interest canbe– e.g., in rat lens, add BSA and measure a unique peptide

• A synthetic 13C-labeled peptide for each protein of interest– Very elegant , but expensive ($1,000 per peptide)– Does not assess hydrolysis efficacy

• Concatenation (joining the peptides as their oligonucleotideanalogs, creating a macro cDNA and expressing the hybridprotein)– Very elegant, sophisticated use of molecular biology

• SILAP labeling - creates proteome-wide, stable isotopelabeled proteins

Concatenation - making 13C-labeled peptide internal standards

K

K

K

K

K

K

Convert peptidesequences tooligo DNAsequences

Splice together theindividual oligo DNAs toform a composite cDNA

Insert cDNA into aplasmid andrecombinantly expressin bacteria in thepresence of Lys-13C6

15N2

NH2- COOH

K*

K*K*

K*

K*

K*

Treat with trypsin

Anderson & Hunter, 2005

SILAC use as an internal standard

Journal of Proteome Research 2009, 8, 1565-1576

Stable isotope amino acids http://www.isotope.com

Amino acid Amount (g) Price ($)• L-Arg 13C6 0.1, 0.5 650, 1880• L-Arg 13C6 15N4 0.1, 0.5 855, 2445• L-Lys 13C6 0.1, 0.25 443, 1129• L-Lys 13C6 15N2 0.1, 0.25 345, 580• L-Tyr 13C9 0.1, 0.25 635, 1240• L-Tyr 13C9

15N 0.1, 0.25 275, 410

These are all 97-99% pure. Prices for 99% orbetter are not listed, but will be higher.

Summary

• Quantitative proteomics has arrived• Specific hypotheses can now be tested

with high precision• Limitations lie in having enough

bioinformatics resources to “boil”approaches down

• Other limitations are in the basicbiochemistry of proteolysis

Many thanks to• Landon Wilson• Ray Moore• Sai Sai Dong• Richie Herring

• David Stella• Michael Heaven• Kyle Floyd• Elizabeth Gulsby

• Chiquito Crasto, PhD• Helen Kim, PhD• Matt Renfrow, PhD• Om Srivastava, PhD

UAB Centers support• Acute Kidney Injury Center• Botanicals• CNGI• Polycystic Kidney Disease• Skin Disease Research• (Cancer Center)

Federal support• R21 AT004621• S10 RR17261• S10 RR19231