Overview - University of Alabama at Birmingham 01-13-09-2.pdf · PT X-S NH-PTX NHS/EEDQ DMF MAb PTXMAb Ahmad Safavy ... 1346.0 1354.4 1362.8 1371.2 1379.6 1388.0 Mass (m/z) 0 0 20

1

S. Barnes-UAB 1/13/09

Application of massApplication of massspectrometry to the analysisspectrometry to the analysis

and identification of peptides,and identification of peptides,proteins and other biologicalproteins and other biological

moleculesmolecules

Stephen Barnes, PhD4-7117

[email protected]


Overview• MALDI-TOF MS

– Protein modifications– Peptide mass fingerprinting

• Electrospray MS– Analysis of intact proteins– Molecular weight calculations– Max Entropy for MW estimation

• Peptide analysis– Purity - ESI-MS is a revelation

• Integration of MS with LC and CE– Multidimensional LC of peptides

• Tandem MS– Identifying peptide amino acid sequences

2


Matrix-Assisted LaserDesorption Ionization (MALDI)

Flight tube and drift region tomeasure the time-of-flight (TOF)

Accelerating pulse

Short laserpulse

detector


Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry

(MALDI-TOF MS)

• Advantages of MALDI-TOF– More tolerant to common buffers than ESI– High degree of sensitivity, moderate mass accuracy,

and mass resolution– High mass compounds, i.e. proteins, PEG…

• Common Applications of MALDI-TOF– Masses of large proteins and other compounds– Enzymatic digestion profiles of proteins to establish

their identity– Peptide sequencing (TOF-TOF)– In situ protein/peptide imaging

3


Factors from conventional experimentsthat impact MALDI-TOF analysis

• Tolerance of buffers/chemicals used insample preparation– NaCl up to 150 mM– Urea up to 2-3 M (carbamoylation can occur!)– Guanidinium-HCl up to 2 M

• Tolerance of detergents– SDS up to 0.05%

• Staining Protocols– Whole proteins form adducts with Coomassie– Silver staining modifies selected peptides


Benefit of removing salt from tryptic digest

900 1520 2140 2760 3380 4000Mass (m/z)

0

8477.0

01020

304050

60708090100

% In

tens

ity

1929.87

2209.971993.96

1868.97 2148.002897.04

1804.902225.00 Salt Contamination

899.0 1519.2 2139.4 2759.6 3379.8 4000.0Mass (m/z)

0

2.7E+4

0102030405060708090

100

% In

tens

ity

1804.96

2164.04

1994.97

2148.001928.96

1475.782290.101868.94

2706.121439.742383.97

1179.65 2502.28 3024.06

ZipTip (desalting)

Trypsin autolysis peak

4


Chemically modifying an antibodyChemically modifying an antibodyto bind radioactive metal ionsto bind radioactive metal ions

aaa

Scheme 1

PTX-2'-OH + Succinic Anhydride PTX-2'-O2CCH2CH2CO2H

PTX-SX

aaa

Scheme 2

+NH2 PTX-SX NH-PTXNHS/EEDQ

DMF

MAb PTXMAb

Ahmad Safavy


Structure of modified antibodyStructure of modified antibody

Ahmad Safavy

5


Modification of an antibody byMALDI-TOF

Ahmad Safavy60000 87982 115964 143946 171928 199910

Mass (m/z)

0

670.1

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity

C 155148

BSA D 77246

Reacted Ab

[M+H]+

[M+2H]2+

60000.0 87975.4 115950.8 143926.2 171901.6 199877.0Mass (m/z)

0

917.6

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity

BSA

B 152223

A 76337

Unreacted Ab

[M+2H]2+[M+H]+

Δ = 2925 Da


900 1520 2140 2760 3380 4000

Mass (m/z)

0

8401.3

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity

2452.171366.82

1494.90

2646.331267.76 2164.05

2517.291388.78 2186.02 3141.451544.76

3163.46

A MALDI-TOF mass spectrum ofpeptides

6


1346.0 1354.4 1362.8 1371.2 1379.6 1388.0

Mass (m/z)

00

20

40

60

80

100%

Inte

nsity

1366.82

1367.82

1368.82

1369.831

900 1520 2140 2760 3380 4000Mass (m/z)

0

8401.3

0

20

40

60

80

100

% In

tens

ity

1366.82 2452.17

Isotope profile of an individual peptide ion

Why?


Stable isotopes of the most abundantelements found in peptides

*Varies according to its source

Element Mass AbundanceH

C

N

O

1.00782.0141

12.000013.003414.003115.000115.994916.999117.9992

99.985%0.015%

99.89%*1.11%*

99.64%*0.36%*

99.76%*0.04%*0.20%*

S 31.9721 94.93%*32.9715 0.76%*33.967935.9671

4.29%*0.02%*

7


m/z

4360.45All 12C

4361.45One 12C

Isotope pattern for a larger peptide (207 C-atoms)Two 12C

Three 12CFour 12C

Five 12C

Six 12C

Seven 12C


How to represent the mass ofcompound?

• At high resolution where the isotopicpeaks are fully resolved, then we candetermine the monoisotopic mass foreach one

• At low mass resolution (where the isotopepeaks cannot be resolved) what isobserved is the average mass

8


60000.0 87975.4 115950.8 143926.2 171901.6 199877.0Mass (m/z)

0

917.6

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity

BSA

B 152223

A 76337

Average massFor MALDI-TOF spectra above m/z 5000, it becomesdifficult/impossible to resolve the isotope profile. Instead, theaverage mass is calculated.


Mass (m/z)1000 8800 16600 24400 32200 40000

100

% In

tens

ity

12521.80

12680.12

12837.86

6260.9725207.30

12365.22

[M+2H] 2+

[2M+H]+

[M+H] +

CONTROL

PROTEIN + ONE 4HNE

PROTEIN + TWO 4HNE

PROTEIN +THREE 4HNE

-

Modification of cytochrome C by thereactive aldehyde 4HNE

MALDI spectra usually contain only the molecular ion [M+H]+.The mass accuracy is approximately +1-2 Da. Hard to pick out the12C isotope peak. The 4HNE Michael adduct should be 156 Da.

9


Increased resolution and sensitivity in reflector vs. linear mode

1560 1566 1572 1578 1584 1590Mass (m/z)

0

6.3E+4

0

10

20

30

40

50

60

70

80

90

100

% In

tens

ity

1570.841571.82

1572.82

1573.78

1574.63

1575.66

Reflector Mode

1560 1566 1572 1578 1584 1590Mass (m/z)

0

5959.3

0102030405060708090100

% In

tens

ity

1572.98

1571.97

1573.98

1575.05Linear Mode


Peptide mass fingerprinting• This method was developed because of the

availability of predicted protein sequences fromgenome sequencing

• Proteins did not have to have been previouslysequenced - only that the open reading frame inthe gene is known - the rest is a virtual exercisein the hands of statisticians, bioinformaticistsand computers

• However, remember the matching is only as goodas the database content - this can change

10


Protein analysis by MALDI 2009

Peptideextraction

DesaltingZiptip

WaterBath37oC

Incubateovernight

destain

Speed-Vac

trypsin

1:20

Eppendorftube

MALDI plate

899.0 1219.2 1539.4 1859.6 2179.8 2500.0Mass (m/z)

0

102030405060708090

100

% In

tens

ity

1774.86

1729.77

1618.84 1827.90 2184.02

1535.75

1361.711756.85

1585.73 2255.061551.73956.53 2303.15

1512.751396.69

1407.67


Proteolytic enzymes used tohydrolyze proteins

The choice of enzyme largely depends on the nature of the aminoacid sequence and the specific issue that is being addressed

• Trypsin - cleaves at arginine and lysine residues

• Chymotrypsin - cleaves hydrophobic residues

• Arg-C - cleaves at arginine residues

• Glu-C - cleaves at aspartate/glutamic acid residues

• Lys-C - cleaves at lysine residues

• V8-protease - cleaves at glutamic acid residues

• Pepsin - cleaves randomly, but at acid pH

See http://www.abrf.org/JBT/1998/September98/sep98m_r.html

11


Searching databases with peptidemasses to identify proteins

Best site is at www.matrixscience.com

The program (MASCOT) can search the OWL orNCBI databases using a set of tryptic peptidemasses, or the fragment ions (specified orunspecified) of peptides

Presents the expected set of tryptic peptides foreach matched protein


Choice of peptidase• Analogous to DNA restriction enzymes

• Tryptic peptide fingerprinting may identify, notone, but several highly related protein candidates(e.g., actins)

• Inspection of the sequences may reveal that thereis a difference at one residue that distinguishesbetween two candidates.

• If for instance it is a glutamate, then use of Glu-Cor V8-protease may enable the two proteins to becorrectly identified

• INSPECT sequences carefully

12


Sequence of β-lactoglobulin

MKCLLLALAL TCGAQALIVT QTMKGLDIQKVAGTWYSLAM AASDISLLDA QSAPLRVYVEELKPTPEGDL EILLQKWENG ECAQKKIIAEKTKIPAVFKI DALNENKVLV LDTDYKKYLLFCMENSAEPE QSLACQCLVR TPEVDDEALEKFDKALKALP MHIRLSFNPT QLEEQCHI


MK

CLLLALALTCGAQALIVTQTMKGLDIQK

VAGTWYSLAMAASDISLLDAQSAPLR

VYVEELK

PTPEGDLEILLQKWENGECAQK

K

IIAEK

TK

IPAVFK

IDALNENK

VLVLDTDYK

KYLLFCMENSAEPEQSLACQCLVR

TPEVDDEALEK

FDK

ALK

ALPMHIR

LSFNPTQLEEQCHI

13


MKCLLLALALTCGAQALIVTQTMKGLD

IQKVAGTWYSLAMAASD ISLLD AQSAPLRVYVE

E LKPTPE GD LE ILLQKWE NGE CAQKKIIAE

KTKIPAVFKID ALNE NKVLVLD TD YKKYLLFCME

NSAE PE QSLACQCLVRTPE VD D E ALE KFD

KALKALPMHIRLSFNPTQLE E QCHI

Peptides from digestion with Glu-C


Amino acid residue massesAlanine 71.037 Leucine 113.084Arginine 156.101 Lysine 128.094Asparagine 114.043 Methionine 131.040Aspartic acid 115.027 Phenylalanine 147.068Cysteine 103.009 Proline 97.053Glutamic acid 129.043 Serine 87.032Glutamine 128.058 Threonine 101.048Glycine 57.021 Tryptophan 186.079Histidine 137.059 Tyrosine 163.063Isoleucine 113.084 Valine 99.068

The m/z value of a peptide [M+H]+ is the sum of the residuemasses plus 18.015 for H2O plus 1.008. So, what is it forISLLD?

113.084 + 87.032 + 113.084 +113.084 + 115.027 + 18.015 + 1.008 = 560.334

14


837.4764 916.47341064.44661065.58271245.58451658.78432275.25862313.25882647.20232707.3760

800.4876 929.54551003.56051232.66341259.77221337.66321447.70321811.89962307.30062819.5265

Expected peptides from trypsin andGlu-C digestion of bovine β-

lactoglobulin

Assumes all cuts are complete, there is no oxidation ofMet residues, and Cys residues are unmodified


Enter the ions here in this box

15



Protein records provided by MASCOT search

16


Comparison of observed andpredicted tryptic peptides


Search against SwissProt database

17


Things to consider when doingpeptide mass fingerprinting

• Proteins can be oxidized both biologically (real data)and during the workup

• Treat the protein or the peptide digest with a reagentthat reacts with Cys sulhydryl groups - e.g.,iodoacetamide, iodoacetic acid, N-ethylmaleimide or4-vinylpyridine. Cysteines may also have reacted withacrylamide in the gel.

• Set the options in the fixed or variable modificationboxes before searching

• Allow for at least one missed cleavage - trypsin doesnot cut when Lys or Arg are followed by a Pro residue


Other web sites for peptideanalysis

• http://prowl.rockefeller.edu/– Choose ProFound

• http://prospector.ucsf.edu/– Choose MS-fit

18


Further information on identifiedprotein

• Take the protein identifier number:– For bovine β-lactoglobulin it is gi|520– Go to http://www.ncbi.nlm.nih.gov– Under protein, paste in the gi number– A link to the protein will appear– Click on Blink - this is similar to BLAST, but better– Go to Structure: enter beta lactoglobulin bos taurus

• This generates 21 different crystal structure records -select #18 - 1CJ5

– To view a 3D-image of the protein, first download Cn3Dfrom the NCBI site or RasMol

• Bring a picture of beta-lactoglobulin to the next class


Examples for homework (due Jan 20)

• Identify the following proteins from these MALDI ions(corrected for isotope effects):

– 2280.11, 1700.89, 1393.71, 1385.74, 1164.60, 1087.49, 1007.54,969.44, 916.50, 865.44, 813.40 (human)

– 2582.21, 2511.17, 1677.96, 1647.77, 1581.81, 1068.48, 1018.49,1011.58, 1001.45 (rat)

– 1993.90, 1964.91, 1802.88, 1756.90, 1738.67, 1716.91, 1668.68,1581.82, 1510.72, 1497.76, 1428.66, 1386.68, 1383.69, 1250.62,1241.69, 1209.60, 1172.60, 1161.69, 1140.56, 1131.65, 1128.62,1107.49, 1093.52, 1027.46, 1026.51, 1006.50 (mouse)

• Set the number of missed tryptic cuts to 0 and tryvarying the mass accuracy from 0.02 to 1.0 Da. Howdoes this alter the MOWSE score?

• Do any of these have crystal structures?

19


Take home question (due Jan 20)

1. What is the monoisotopic mass of humancalgranulin A?

– Hint: workout the empirical formula of humancalgranulin A - its sequence can be obtainedfrom record P05109 athttp://www.ncbi.nlm.nih.gov

2. What is the molecular weight of the mostabundant species of calgranulin A ?

– Hint: assume that the abundance of 13C is1.11% of total carbon atoms - do not worryabout 2H or other isotopes


Blurring of protein space• Identification using MALDI-TOF

with MASCOT depends on:– Number of peptides recognized as

being part of the protein– The mass accuracy of the peptides

that are recognized– Pre-2000, an accuracy of better than

0.05 Da in a 1000 Da peptide (i.e., 50ppm) was sufficient to distinguish theunknown protein from the otherproteins in the databases at that time

– Now, the protein information space hasbecome more dense and MALDI-TOFis no longer adequate

– Previously identified proteins may notbe correct

20


Electrospray ionization• ESI-MS is very sensitive to the presence of

electrolyte species -– these ionize more easily than solutes and may

also form adducts with solutes• In ESI-MS, multiple charge states are

possible– These lead to more accurate MWs

• This is a softer ionization than MALDIwhere the UV laser at 337 nm alters thechemistry of modifications such as Tyr-NO2 and Cys-SNO


ESI-MS and purity of peptides

Guarantees of purity based on observation of “a single peak byreverse-phase HPLC” and by “it gave the correct sequence whenanalyzed by Edman degradation” are hollow.The lower spectrumwas of a “pure” HPLC peak. The method of purification wasamended and the upper spectrum was obtained

21


Ionizing proteins and peptides

• +H3NCHR1CO(NHCHRnCO)nNHCHR2COOH is theion that’s found in dilute acid solution

• If there are internal basic residues, then the ionswill be of the form [M+nH]n+, where n = 1, 2, etc.

• A tryptic peptide will have a N-terminal aminogroup and an amino group from Arg or Lys– If the peptide has a mol. wt. of 1000 Da, then the singly

charged ion will have a m/z of 1001, whereas the doublycharged ion has a m/z of 501


ESI mass spectrum of ribonuclease

Peak (m/z) Intensity Charge (est.) Mol. Wt. (Est.)

978.00 7,778 14.00000 13,677.891,053.00 18,532 13.02656 13,675.901,141.00 59,087 11.95446 13,679.911,245.00 33,275 10.96146 13,683.911,369.00 32,390 10.03219 13,679.921,521.00 35,668 8.99995 13,679.931,711.00 16,624 7.99996 13,679.941,956.00 3,333 6.97955 13,684.94

Cumulative MWestimate = 13,680.29

SD = 2.94

22


Calculation of molecular weightsand ion states

• For two ions in a series for a peptide of molecular weight M,the lower m/z value (x) will be for the n+1 ion state and thelarger m/z value (y) will be for the n+ ion state.

- (1) (M+n)/n = y- (2) (M+n+1)/(n+1) = x

• Hence- (3) M+n = ny and M = ny-n- (4) M+n+1 = (n+1)x and M = (n+1)x-(n+1)

• Hence- ny-n = (n+1)x - (n+1)- ny-n-xn+n = x-1- n(y-x) = x-1- n = (x-1)/(y-x)

• The value of n can then be substituted in equation (1) toobtain the molecular weight of the peptide


Deconvolution of MS data

• When several proteins are present,then their multiply charged ionclusters overlap

• Can this be overcome? - yes, use theMaxEntropy program provided byMicromass

23


800 900 1000 1100 1200 1300 m/z0

100

%

0

100

%

483

933.433

816.759 1088.845

553931.102

814.8361086.109

1102.455

Oxidized Aprotinin ESI mass spectrum

Control Aprotinin ESI mass spectrum

6400 6500 6600 6700

mass

0

100

%

0

100

%

6.21e3

6527

6511

6543

1.21e36510

Deconvoluted mass spectra

Oxidized

Control

Junlong Shao


1000 1100 1200 1300 1400 1500 1600 1700m/z

0

100

%

1361225.144

1148.588

1081.069

1022.515

1312.467

1413.476

1530.979

18300 18400 18500 18600 mass0

100

%

1.39e318362

18345

18329

184261837818393 18458

1848918505

1852218540 18575

Deconvolution of oxidized forms of β-lactoglobulin

Junlong Shao

24

S. Barnes-UAB 1/13/09700 80 0 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000

m/z0

100

%

TOF MS ES+ 552

A: 57195.23±12.89B: 50011.25±18.34C: 15157.60±0.94

A521101.1

A48/B421192.70

A451272.08

A411396.16

A56/B491022.47

C18843.12 Each ion is [M+nH]n+

For 50+ charge state ofa 50 kDa protein,

m/z = [50,000 + 50]/50 = 1,001

ESI spectrum of bacterially expressed protein

Courtesy of Mindan Sfakianos


15000 20000 25000 30000 35000 40000 45000 50000 55000 60000 65000mass0

100

%

15,157.6 + 0.9 Da

50,011.3 + 18.3 Da6xHis-tag BAT

57,195.2 + 12.9 DaE. coli GRoEL

MaxEnt deconvolution of MWs

Courtesy of Mindan Sfakianos

25


12356NativeCytochrome C 12512

Cytochrome C + One 4HNEMichael Addition (+156)

12494

CytochromeC+One 4HNESchiff Base(+138)

12668

Cytochrome C + Two 4HNEMichael Addition (+312)

12823

Cytochrome C+Three 4HNEMichael Addition (+467)

12300 12400 12600 12700 12800Mass

0

100

%

12500

ESI-MS of 4HNE-Modified Cytochrome C

Courtesy of Amanda Isom (d. 2005)


Summary of determining MW by ESI

• The multiple charge states of a proteinallow:– Mol Wt of large proteins to be estimated– It’s a super SDS-PAGE gel

• Important to remember that the proteinsample must be free of salt– Typically, a sample is cleaned up on a short

reverse-phase column prior to electrospray– Alternative, use ammonium acetate as buffer

26


Studying high molecularweight complexes by ESI

• Most instrument ESI interfaces have alimited m/z range - up to 3,000

• In protein complexes water, and hence H+ions, is “squeezed” out, therebysubstantially increasing observed m/zvalues

• Interfaces that pass ions with m/z valuesabove 10,000 have been designed


nanoESI-MS of HMW complexes of small heat shock proteins

Sobbott et al., J BiolChem 277:38921

Note the large m/z values(6,000-7,000) for theobserved ions

The ESI data weredeconvoluted to revealthe distribution of themasses of the complexes

27


Use of FT-MS in ESI of proteins

• The very high resolving power of FT-MSenables a direct measure of charge stateof an individual ion since each peptide orpolypeptide will have several/manyisotope peaks

• The distance in Da between successiveisotope peaks of a multiply charged ion isthe reciprocal of the number of charges


Bovine Serum Albumin (66 kDa)4.7 T Actively Shielded Magnet

1510.6 1510.8 1511.0 m/z

1600 1800 m/z

44+ESI: BSA

Bruker Daltonics1400

44+

28


LC-MS of peptide mixtures

Q1 Q2

Collision gas

Electrostaticreflector

TOF detector

Analytical reverse phase column75 µm i.d. x 15 cm

Flow rate 200 nl/min

Acetonitrile gradient

pre-columnfor desalting

waste

Loadsample


The MUDPIT approach

A B

CD E

MUMUlti-lti-DDimensional imensional PProtein rotein IIdentification dentification TTechnologyechnology

Digest with trypsinwithout any proteinseparation

20-50 trypticpeptides per protein(100-250 peptides inthis example)

Can they be resolved?

29


• Hydrolyze everything!

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

• Can be fractionated, but still10,000-20,000 to differentiate

• Enormous bioinformaticsproblem

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

NH4Cl (1-200 mM step gradient)

200 mM

Cation exchangecolumn (H+)

John Yates


Cation exchange of peptides

NH3 +-R.COOH

-O3SNH

3 +-R.COOH

-O3S

Stepwise elution with NH4+

NH3+-R1.COOH

Electrostatic capture ontoresin bead in H+ form

NH3+-R2.COOH

NH3+-R3.COOH

30


Connecting CE and LC toMALDI analysis

CE analysis

nanoLC analysis

Parallel capture of effluents of 8 nanoLC separations on Mylar - can be scanned simultaneously by fast laser

Creates 20 mm wide tracks thatcan be scanned by MALDI laser

for MS analysis


Pros/Cons of laying down LC orPros/Cons of laying down LC orEC separations on matrix plateEC separations on matrix plate

• Allows off-line analysis both in real time and thenin a retrospective mode

• MALDI-TOF analysis is very fast

• Can also do TOF-TOF MS-MS analysis

• BUT what happens chemically on the acidicenvironment on the surface of the plate duringstorage?

• Also, can the laser beam cause chemicalchanges?

31


Sequencing of peptides

• Using tandem mass spectrometry in a triplequadrupole, Q-tof, or ion trap instrument, theparent ion is first selected in the first quadrupole

• The parent ion is collided with argon gas and itbreaks into fragments (daughter ions)

• By identifying the daughter ions, the peptideamino acid sequence is inferred


Q1 Q2 Q3 Detector

-++--

-- --- -- -

Collision gasN2

GasSample solution 5 KV

Tandem mass spectrometry on atriple quadrupole instrument

• Daughter ion spectra

• Parent ion spectra

• Multiple reaction ionscanning

32


Q1 Q2 Q3 Detector

-++--

-- --- -- -

Collision gasN2

GasSample solution 5 KV

Triple quad versus Q-tof and sensitivity

The quadrupole analyzer (Q3) is slowand insensitive - it’s a filter - thusthrows away large amounts of data

Q1 Q2

Collision gas

Electrostaticreflector

TOF detector

TOF detector collects all ions generated andyields fmol rather than pmol sensitivity

Also gives far greater mass accuracy -from 1000 ppm on the triple quad to <20 ppmon the Q-tof

Crucially important for automatedinterpretation of MS-MS spectra to yieldamino acid sequence


Fragmentation of a peptide

100 200 300 400 500 600 700 800 900 1000 1100 1200 1300mass0

100

%

y101100.61

y9987.52

b3375.21

b2 262.12

234.13

y1 175.12

y2 274.19

y8 916.49

parent681.45

y4 487.32

y5 602.32

y6 730.43

b8875.42

y111247.69

y7 859.45

b4446.23

Documents

Overview - University of Alabama at Birmingham 01-13-09-2.pdf · PT X-S NH-PTX NHS/EEDQ DMF MAb PTXMAb Ahmad Safavy ... 1346.0 1354.4 1362.8 1371.2 1379.6 1388.0 Mass (m/z) 0 0 20