Counting and Label-Free Approaches to Identify PTM Biomarkers
of Cardiovascular Disease in a Mouse Model
Jean L. Spencer, Stephen A. Whelan, Christian F. Heckendorf, Deborah A. Siwik, Wilson Colucci,
Markus M. Bachschmid, Richard A. Cohen, Catherine E. Costello, and Mark E. McComb
Boston University School of Medicine, Boston, MA 02118
Overview
Methods
Results
Results
Results
 Purpose: Counting software and label-free proteomics were
STRAP PTM Analysis
 Software: STRAP PTM (v1.0 beta) freely available at
STRAP PTM Results
STRAP PTM Results
Label-Free Results
PTM Scores (6 modified protein hits)
PTM Map: Albumin
Plasma Proteins vs. Albumin PTMs
MS/MS data
induced by oxidative stress.

Introduction
PeptideProphet
ProteinProphet
CFLQHKDDNP SLPPFERPEA EAMCTSFKEN PTTFMGHYLH EVARRHPYFY
PTM Scoring
APELLYYAEQ YNEILTQCCA EADKESCLTP KLDGVKEKAL VSSVRQRMKC
PTM Score (S): Overall score for a specific PTM (m) on a
SSMQKFGERA FKAWAVARLS QTFPNADFAE ITKLATDLTK VNKECCHGDL
There are many factors that contribute to cardiovascular
specific site (i) of a specific protein (p) based on user-
disease
selectable factors relevant to the system (max = 100)
but
unfavorable
metabolic
conditions
causes. Early detection and monitoring of these adverse effects
Quality
Grouping
plasma proteins are indicators of inflammation and oxidants,
metabolic disease. The appearance and change of posttranslational modifications (PTMs) on circulating proteins may
PTM)
and
label-free
proteomics
PTM Search Space
PTM
Probability
0.2
Q=
0.8
0.8
1.0
using an American diet (high-fat, high-sucrose intake) mouse
model.
Q=
0.2/1
= 0.2
1.0
1.6/2
= 0.8
1.0
Better results
of a specific protein across groups (max = 1)
∆ Mass
Group A
Oxidation*
(M)
[+15.99]
O
Cyano*
(C)
[+25.00]
H(-1) C N
Dioxidation*
(C)
[+31.99]
O(2)
Methylthio
(C)
[+45.99]
H(2) C S
Trioxidation*
(C)
[+47.98]
O(3)
Glutathione
(C)
[+305.07]
10
10
10
10
*Oxidation (M)
30
30
30
HFHS
G = 0/2.1 = 0
Greater
variation
Occupancy (W): Degree of modification of a specific site on
a specific protein with a specific PTM (max = 1)
Sample Processing
 Blood collection at 4 months from each group (n = 4)
 Protein digestion with trypsin to generate peptides
 Plasma
proteomics
HexNAc (S10)
2
W = 2/4 = 0.50
Greater modification
3
Search
PTM Standards
Pipeline
Digested proteins
from
25
62
Uniqueness (U): Rarity of a specific PTM on a specific
protein (max = 1)
*Dioxidation
*Dioxidation(C)
(C)
3030
30
∑
U = 1 - 1/4 = 0.75
U = 1 - 3/4 = 0.25
Less frequent
1.8
17 key sites with PTMs: 15/36 Cys, 2/7 Met
-30
-30
-30
15
15
1010
10
55
Diff0 00
00
ALBU
ALBUC265
C265
-20
-20
-20
*Trioxidation (C)
ALBU C279
ALBU C278
-5-5
-15
-15
1.4
1.2
0.8
Global Sites
PTM
Score
Mod
Mass
Site
46.5
25.00
C392
34
59
25
27
1.000
1.000
0.775
0.600
30.6
47.98
C279
14
27
13
41
12
1.000
0.520
0.774
0.761
30.1
31.99
C265
43
64
21
107
66
0.752
0.840
0.618
0.770
21.8
25.00
C 34
21
34
13
55
1
0.710
0.520
0.982
0.600
20.7
25.00
C487
22
33
11
15
1.000
0.440
0.786
0.600
12.2
25.00
C567
35
48
13
96
0.843
0.520
0.464
0.600
11.7
15.99
M 264
34
43
9
77
96
0.820
0.360
0.445
8.8
47.98
C278
4
13
9
17
36
1.000
0.360
0.321
0.761
6.0
47.98
C514
6
11
5
17
6
0.535
0.200
0.739
0.761
3.5
25.00
C265
12
20
8
32
141
1.000
0.320
0.185
0.600
Control
Control
70
70
Global Sites
-10
-10
Global
GlobalSites
Sites
Total Counts
Control HFHS Diff
1.6
1.0
PTM Scoring: Albumin
ALBU C567
C567
ALBU
ALBU C487
C487
ALBU
-20
-20
-20
10
10
-10
-10
-10
FAQFLDTCCK AADKDTCFST EGPNLVTRCK DALA
188 potentially modified sites/PTMs
Modified Other
Forms Form s
93
55
83
Scoring Factors
Q
G
W
U
Confirmation of Results
MS/MS Fragmentation of Albumin Peptide +3 PTMs
0.890
Peptide: Ions Score 45, Expect 0.002
C392 Glutathione: Ions Score 56
C392 Dioxidation: Ions Score 53
C392 Cyano: Ions Score 81
ALBU C392
C392
ALBU
ALBU C34
C34
ALBU
-10
-10
-10
2020
20
-30
-30
-30
I = total sites
M = total PTMs
N = modified peptide counts
125
Protea Biosciences
depleted mouseAnalysis
plasma
Conc (nM)
Label-Free
Quantitative
Mascot Search
PeptideProphet
MS/MS data
STRAP PTM
ProteinProphet
(v4.1; Nonlinear
Dynamics)
 Progenesis LC-MSEngine
 Files:
Mascot
search engine (v2.3; Matrix Science)
RAW····MGF·······························DAT····PepXML············ProtXML
Global Sites
Cntrl
∑
∑
00
0
ALBU
M264
ALBU M264
APOA1 M218
M218
APOA1
APOA1
APOA1 M216
M216
HFHS
W = 1/4 = 0.25
2
-10
-10
-10
-10
M = total PTMs
N = modified peptide counts
N0 = unmodified peptide counts
∑
NO (C7)
H PO (Y4)
LC-MS/MS
Analysis
Increasing conc
 nanoACQUITY HPLC system (Waters)
O (M4)source (Advion)
 TriVersa NanoMate ESI
 LTQ-Orbitrap
massNOspectrometer
(Thermo Scientific)
OMe
(M11)
(Y4)
 Data-dependent
MS/MS acquisition
Ac (K4)
Database
Trans-Proteomic
Cntrl
EFKAETFTFH SDICTLPEKE KQIKKQTALA ELVKHKPKAT AEQLKTVMDD
10
10
10
Diff 0000
-5
-5
-5-5
*Cyano (C)
20
20
20
555
5
Difference
High-fat, high-sucrose diet
(HFHS)
H(15) C(10)
N(3) O(6) S
Diff = HFHS counts – Control counts
Group B
G = 2.1/2.1 = 1
Composition
Differential Observation of PTMs
σ = std dev of counts
max σ = max σ of all proteins
Control diet
(Control)
Fold Change = HFHS / Control
DYLSAILNRV CLLHEKTPVS EHVTKCCSGS LVERRPCFSA LTVDETYVPK AA
Grouping (G): Variation of a specific PTM on a specific site
Mouse Model
Normalized Abundances: Albumin - PTMs
2.0
P = probability of modified peptides
P0 = probability of unmodified peptides
characterize differential PTMs and potential biomarkers of CVD
Methods
LECADDRAEL AKYMCENQAT ISSKLQTCCD KPLLKKAHCL SEVEHDTMPA
YGFQNAILVR YTQKAPQVST PTLVEAARNL GRVGTKCCTL PEDQRLPCVE
specific PTM on a specific site of a specific protein (max = 1)
to
Retention Time
> 10,000 features
DLPAIAADFV EDQEVCKNYA EAKDVFLGTF LYEYSRRHPD YSVSLLLRLA
Uniqueness
Quality (Q): Goodness of database search results for a
this disease. Here we investigate the power of counting
(STRAP
Occupancy
User-selectable factors
be critical in determining biomarkers and therapeutic targets for
software
KKYEATLEKC CAEANPPACY GTVLAEFQPL VEEPKNLVKT NCDLYEKLGE
on the heart and vasculature are difficult to achieve. Since
nonspecific changes in these components may reflect systemic
Retention Time
> 100,000 features
2
(CVD),
KTCVADESAA NCDKSLHTLF GDKLCAIPNL RENYGELADC CTKQEPERNE
Files: RAW····MGF·······························DAT····PepXML············ProtXML
associated with obesity, diabetes, and hyperlipidemia are major
Albumin
EAHKSEIAHR YNDLGEQHFK GLVLIAFSQY LQKCSYDEHA KLVQEVTDFA
STRAP PTM
1.8
and easily sorted for biologically relevant modifications
Mascot Search
Engine
Plasma
1.6
in differential PTMs were readily detected
Trans-Proteomic
Pipeline
m/z
based on spectral counting and a novel scoring algorithm.
 Results: Trends
Database
Search
m/z
Workflow
1.4

Legend: Control HFHS
Fold Change
model were analyzed using in-house software (STRAP PTM)
http://www.bumc.bu.edu/cardiovascularproteomics/cpctools
1.2
Blood samples from an American diet mouse
1
 Methods:
0.8
biomarkers of cardiovascular disease in a mouse model.
50
50
40
40
Conclusions
30
30
 Select HFHS-induced PTMs were successfully mapped onto
20
20
albumin by counting and label-free approaches.
10
10
00
 STRAP PTM counting provided an easy and quick overview
C392
C392
Global Sites
 Proteins with significant differential PTMs ranked high among
other proteins based on average PTM scores.
 Distributions of differential counts highlighted trends in global
protein sites with PTMs of non-random positive changes.
HFHS
HFHS
60
60
Total Counts
used to characterize differential PTMs and potential
C279
C279
C265
C265
C34
C34
C487
C487
C567
C567
M264
M264
C278
C278
Sites on Albumin
 PTM sites sensitive to HFHS diet were easily identified on
albumin by high PTM scores.
 Top-8
redox-specific sites on albumin were selected and
mapped for further investigation (e.g., C392).
of modified sites on protein sequences.
 Subsequent
quantitative analysis with label-free methods
was facilitated by preliminary analysis with STRAP PTM.
Acknowledgments
 NIH-NHLBI contract HHSN268201000031C
 NIH grants P41 RR10888/GM104603 and S10 RR020946