Multidimensional protein identification technology-selected reaction monitoring improving detection and quantification for protein biomarker studies

Article abstract of DOI:10.1021/ac2028485

The targeted analysis of proteins in complex biological samples is best achieved using selected reaction monitoring (SRM). To maximize the sensitivity of this approach, sample fractionation or enrichment is still required, particularly to detect less abundant proteins in clinically relevant biofluids. Here, we report the development of multidimensional protein identification technology (MudPIT)-SRM, taking advantage of the robust online strong cation exchange chromatography for tryptic peptide fractionation and combining it with the multiplexed, quantitative attributes of SRM. The classical MudPIT method has been modified with an in-line strategy to introduce reference peptides onto the analytical column to enable quantitation at each salt step. Applying the MudPIT-SRM approach to profile abundant plasma proteins, we demonstrated mean increases in peak areas of almost 90% compared to conventional SRM. MudPIT-SRM analyses of low abundant proteins present in human wound fluid exudates similarly demonstrated increased peak areas and enabled the detection of proteins which were below the lower limit of detection when analyzed by conventional SRM. The MudPIT-SRM method is relatively facile to conduct and offers performance advantages to enhance sensitivity for biomarker studies.

Full text of DOI:10.1021/ac2028485

Article
pubs.acs.org/ac
Multidimensional Protein Identification Technology-Selected
Reaction Monitoring Improving Detection and Quantification for
Protein Biomarker Studies
Christoph Krisp,†,‡ Matthew J. McKay,†,§ Dirk A. Wolters,‡ and Mark P. Molloy*,†,§
†Department of Chemistry and Biomolecular Sciences, Macquarie University, Sydney, Australia
‡Department of Analytical Chemistry, Ruhr-University Bochum, Bochum, Germany
§Australian Proteome Analysis Facility, Macquarie University, Sydney, Australia
S
* Supporting Information
ABSTRACT: The targeted analysis of proteins in complex biological samples is
best achieved using selected reaction monitoring (SRM). To maximize the
sensitivity of this approach, sample fractionation or enrichment is still required,
particularly to detect less abundant proteins in clinically relevant biofluids. Here,
we report the development of multidimensional protein identification
technology (MudPIT)-SRM, taking advantage of the robust online strong
cation exchange chromatography for tryptic peptide fractionation and
combining it with the multiplexed, quantitative attributes of SRM. The classical
MudPIT method has been modified with an in-line strategy to introduce
reference peptides onto the analytical column to enable quantitation at each salt
step. Applying the MudPIT-SRM approach to profile abundant plasma proteins, we demonstrated mean increases in peak areas
of almost 90% compared to conventional SRM. MudPIT-SRM analyses of low abundant proteins present in human wound fluid
exudates similarly demonstrated increased peak areas and enabled the detection of proteins which were below the lower limit of
detection when analyzed by conventional SRM. The MudPIT-SRM method is relatively facile to conduct and offers performance
advantages to enhance sensitivity for biomarker studies.
overcome ion suppression effects and limit interferences of
elected reaction monitoring mass spectrometry (SRM-MS)
Sis a highly selective and sensitive analytical method applied coeluting peptides. Commonly applied methods involve
immunodepletion of the most abundant plasma proteins or
by immunoenrichment to purify an analyte.19,20 Additionally,
strong cation/anion exchange chromatography (SCX/SAX) is
predominantly used as a robust and facile methodology for
charge-based sample fractionation. Interestingly, although
online SCX, exemplified in the multidimensional protein
identification technology (MudPIT),21−23 is a widely dispersed,
robust method that is routinely applied to discovery-based
proteomics, the linking of MudPIT with SRM has not been
reported. The MudPIT strategy combines sample purification
(desalting), prefractionation, and concentration within a simple
chromatographic gradient.24−26 The reduced sample complex-
ity allows for more optimal usage of the mass spectrometer
dynamic range, facilitating sensitivity enhancement.
We anticipate that coupling of online sample fractionation
with SRM will overcome crucial peptide losses, especially for
low abundant proteins. Many of the losses are associated with
the numerous sample manipulation steps associated with offline
methods including, sample concentration, buffer exchanges, or
resuspension. Here, we describe MudPIT-SRM for the first
to biomedical and pharmaceutical research. Over the past few
years, efforts have grown to implement SRM for proteomics
with particular emphasis toward clinical biomarker applica-
tions.1−11 SRM-MS is exceedingly applicable for complex
protein mixtures, enabled by highly sensitive analyte detection
using consecutive mass analysis steps of defined ions. The
selection of defined ions includes the precursor ion and one or
more product ions generated by collision induced dissociation
(CID). The transition of a selected precursor ion to a product
ion provides specificity and high sensitivity detection of a
predefined analyte within the complex background common to
biological samples. SRM is mainly performed on triple-
quadrupole instruments but can also be conducted on Ion-
trap instruments.12,13 The fast ion scanning rates of triple
quadrupole instruments facilitate the simultaneous monitoring
of multiple analytes. Hence, SRM-MS comprises the advantages
of multiplexed, quantitative target analysis within a single run
with reliable analyte identification achieved by monitoring
multiple transitions for each peptide. These advantages,
however, are best exploited with simple protein mixtures or
with highly abundant proteins in complex specimens like blood
plasma.14−18
Received: October 27, 2011
Accepted: January 5, 2012
Published: January 5, 2012
To detect less abundant proteins by SRM, fractionation
strategies are still needed to reduce sample complexity to
© 2012 American Chemical Society
1592
dx.doi.org/10.1021/ac2028485 | Anal. Chem. 2012, 84, 1592−1600

Analytical Chemistry
Article
Figure 1. (a) Flow scheme of sample injection, sample loading, salt step elution, and reference peptide infusion applied in MudPIT-SRM. (b)
Representative elution profile XIC of the reference peptides for CRP and PON1 (three SRMs are used to target each peptide). (c) Proof of
reproducible reference peptide injection using continuous peptide infusion into an infusion loop bypassing the trapping column system. Peptides are
listed on the basis of their elution order.
time and evaluate performance compared to conventional SRM
in terms of sensitivity and quantitation improvements
associated with protein biomarker applications. We also
describe the introduction of labeled reference peptides via an
inline injection loop to facilitate quantitation for MudPIT
analyses. Our study demonstrates that MudPIT-SRM is a
robust method offering improved capacity to resolve and
quantitate candidate biomarkers from biological fluids.
digested with sequencing grade trypsin (Promega, Alexandria,
Australia) at 37 °C in a 1:100 ratio of enzyme to protein in a
30% methanol solution (25 mM NH4CO3; pH 8). Cysteine
residues were first reduced with 10 mM dithiothreitol (DTT)
at 37 °C for 1 h, followed by alkylation at room temperature
using 25 mM iodoacetamide (IAA) for 1 h. Peptides were
vacuum concentration and reconstituted to 2.5 mg/mL in 2%
ACN and 0.1% FA.
LC-MS/MS. Chromatographic separation was performed
using a NanoLC 2D System (Eksigent, Dublin, California,
USA). Twenty five micrograms of peptides (10 μL) were
injected onto the MudPIT trapping system using the integrated
autosampler. The trapping system consists of a self-packed
IntegraFrit fused-silica column with 100 μm inner diameter
(New Objective, Inc., Woburn, MA, USA) containing 5 cm
reversed phase (RP) material (3 μm, 100 Å; Column
Engineering Inc., Ontario, CA, USA) followed by a second
self-packed IntegraFrit column (100 μm inner diameter) filled
with 5 cm strong cation exchange resin (3 μm, 300 Å; PolyLC,
Columbia, MD, USA) and 3 cm RP. It was placed in a 10 port
valve to achieve optimal loading and elution conditions.
Injected peptides retained and purified on the first RP trap
were transferred to the SCX phase using an effective 30 min
(plasma) or 10 min (wound fluid) ACN gradient using an 80%
ACN and 0.1% FA buffer. Ten microliters of autosampler
injections of 15, 30 (300 plasma only), and 1500 mM NH4Ac
(pH 3) was used for sample fractionation based on the binding
strength of the analytes on the SCX resin. Released peptides
were trapped on the second RP trap. ACN gradients (30 min
plasma and 10 min wound fluids), following each salt injection,
enabled analytes to be separated on the 15 cm RP column (3
μm, 100 Å; SGE Analytical Science, Rowville, Victoria,
Australia) in-line with the trapping system. In a second
injection loop, stable isotope labeled reference peptides were
continuously infused using a Harvard model 22 syringe pump
EXPERIMENTAL SECTION
■
Chemicals. HPLC grade acetonitrile (ACN), methanol, and
ammonium acetate (NH4Ac) were purchased from Merck
(Kilsyth, Australia). Formic acid (FA) and ammonium
bicarbonate were obtained from Sigma-Aldrich (Sydney,
Australia).
Sample Collection and Preparation. Human plasma
from a healthy individual was collected in EDTA tubes as
previously described.27 Human wound fluid samples were
delivered from the clinic for plastic surgery and serious burn
victims at the University hospital “Bergmannsheil” in Bochum,
Germany, in a study approved by the Ruhr-University Bochum
ethics committee (Ref. No. 2028). Two types of wound fluids
were collected; nonhealing wounds from foot ulcers of diabetes
mellitus type 2 patients and normal healing burn wounds from
otherwise healthy individuals. Fluids were obtained by applying
open-pored polyvinyl alcohol (PVA) hydro foam to a wound
area to soak excreted material. Soaked foams were removed
from the wound, and the retained proteins were obtained after
several washes with a Complete protease inhibitor cocktail
tablet (Roche, Basel, Switzerland) containing 25 mM
ammonium bicarbonate buffer. Samples were immediately
frozen in liquid Nitrogen and stored at −80 °C.
The protein content in plasma and in the wound fluids was
estimated by BCA assay kit (BCA Protein Assay Reagent,
Pierce, Rockford, Illinois). One milligram of protein was
1593
dx.doi.org/10.1021/ac2028485 | Anal. Chem. 2012, 84, 1592−1600

Analytical Chemistry
Article
Table 1. List of MudPIT and RP-SRM Targeted Peptides in Human Plasma Including Precursor and Product Ion Masses, Mean
Peak Area, and Peptide CVs MudPIT Analyzed Plasma Sample
precursor
mass
(m/z)
mean peak area
(counts) MudPIT-
SRM
product ion
mass (m/z)
mean peak area peak area increase (%) MudPIT-SRM
(counts) RP-SRM MudPIT vs RP-SRM CV (%)
protein
peptide sequence
A1AG2
A1AT
TEDTIFLR
LSITGTYDLK
497.8
556.0
558.8
480.8
511.3
719.3
862.0
634.9
615.8
700.8
694.3
655.0
859.0
606.3
624.8
701.6
466.3
557.8
761.3
885.1
440.7
571.5
656.3
602.3
515.8
714.1
776.0
597.8
642.8
683.9
1183.7
711.8
764.4
797.4
788.4
574.3
565.3
975.5
1111.7
956.6
819.5
1023.5
940.5
975.6
1144.6
695.3
875.5
928.6
689.4
629.4
894.4
1197.6
652.3
650.4
1081.5
1003.5
720.4
813.5
822.4
710.3
771.4
941.5
1313.7
875.4
1.60 × 106
3.43 × 106
2.70 × 105
1.35 × 106
2.85 × 105
2.82 × 105
2.54 × 105
2.01 × 105
3.22 × 107
1.32 × 107
1.04 × 107
4.57 × 105
9.54 × 105
2.08 × 105
1.15 × 106
2.17 × 106
2.51 × 105
1.93 × 106
1.25 × 107
4.78 × 106
8.48 × 105
1.26 × 106
1.90 × 105
2.47 × 106
5.61 × 105
1.15 × 105
4.51 × 104
1.48 × 105
4.78 × 106
5.12 × 105
7.92 × 105
2.37 × 105
1.10 × 105
1.55 × 105
7.25 × 104
2.49 × 105
1.86 × 104
2.28 × 104
1.41 × 104
4.51 × 104
2.37 × 104
3.90 × 106
1.01 × 106
1.04 × 106
3.04 × 104
1.80 × 105
3.75 × 104
1.10 × 105
1.46 × 105
8.10 × 103
1.15 × 105
1.31 × 106
5.08 × 104
8.20 × 104
1.66 × 105
2.59 × 104
2.58 × 105
4.11 × 104
5.66 × 103
9.08 × 102
7.10 × 103
5.04 × 105
1.27 × 105
6.42 × 104
93.09
95.48
73.18
81.47
93.47
88.63
95.02
82.24
72.45
87.90
90.32
92.13
93.35
81.10
82.01
90.44
93.26
96.77
94.06
89.53
98.94
90.34
86.86
86.34
89.57
92.67
95.06
97.99
95.22
89.45
75.11
91.89
89.22 5.26
11.81
5.34
4.62
6.85
2.76
5.78
2.09
0.63
6.96
4.86
8.14
3.46
7.95
5.62
9.23
6.14
3.11
8.21
1.34
3.73
10.79
2.26
1.93
6.54
2.33
7.04
4.19
9.89
2.77
4.46
9.23
2.89
5.40 2.44
A2MG
AACT
AMBP
ANGT
QTVSWAVTPK
ADLSGITGAR
ETLLQDFR
SLDFTELDVAAEK
VLSALQAVQGLLVAQGR
ALQDQLVLVAAK
QGLLPVLESFK
APOA1
DYVSQFEGSALGK
VSFLSALEEYTK
GFEPTLEALFGK
DALSSVQESQVAQQAR
IHWESASLLR
APOB
APOC3
CO3
TELRPGETLNVNFLLR
SSLSVPYVIVPLK
LELSVDGAK
CO4A
VGDTLNLNLR
FIBA
FIBB
GLIDEVNQDFTNR
DNENVVNEYSSELEK
LDGSVDFK
FIBG
HEMO
HPT α
HPT β
KNG1
GGYTLVSGYPK
TEGDGVYTLNDK
VTSIQDWVQK
YFIDFVAR
DIPTNSPELEETLTHTITK
SFFSFLGEAFDGAR
ELLESYIDGR
SAA
THRB
TRFE
TTHY
VTDB
VTNC
mean
EGYYGYTGAFR
GSPAINVAVHVFR
VPTADLEDVLPLAEDITNILSK
FEDGVLDPDYPR
(Harvard Apparatus, Holliston, MA, USA). Reference peptides
(300 fmol) were directly transferred to the analytical column
using the HPLC.
peak smoothing. Within triplicate analyses, retention time shifts
of less than 30 s were accepted as reproducible target retention.
RESULTS AND DISCUSSION
Eluting ions were transferred to the 4000 QTRAP mass
spectrometer (AB SCIEX, Foster City, California, USA) using a
2.4 kV spray voltage and a heating plate temperature of 150 °C.
Precursor ions were isolated in quadrupole 1 (Q1), fragmented
with collision induced dissociation (CID) in Q2 by collision
with nitrogen gas, and their characteristic product ions were
detected in Q3, allowing a dwell time for each transition of 20
ms generating a cycle time of 2−3 s, depending on the number
of transitions. In nonquantitative investigations, eluting target
analytes were validated using SRM triggered MS/MS scans
from isolated precursor ions. The MS/MS spectra were
searched against the human Swiss-Prot database using the
Mascot algorithm with 1.2 Da mass tolerances for precursor
and 0.6 for product ions.
Data Analysis. For each transition, extracted ion chromato-
grams (XIC) were generated and the area under the curve
(AUC) was determined using the imbedded quantitation tool
in the instrument software Analyst (AB SCIEX, Foster City,
California, USA). This tool implements the IntelliQuant
algorithm. Peak integration was performed using a 3 point
■
Flow Path Configuration for MudPIT-SRM. A 2D LC
system with integrated autosampler has been used to achieve
optimal performance, reproducibility, and maximized instru-
ment usage time. The flow path configuration shown in Figure
1a allows high flow (μL/min) sample loading (channel 1) and
nanoflow (nL/min) elution on channel 2. In the loading
position, the analyte is transferred to the MudPIT trapping
system, which consists of a short RP column for online sample
desalting followed by an independent SCX/RP hybrid column
used for SCX sample fractionation, consistent with the classical
MudPIT configuration.27−29 Briefly, retained peptides on the
first RP phase are eluted onto the SCX phase after salts have
been washed away by switching the 10-port valve to the inject
position and starting the ACN elution gradient. Eluting
peptides will bind to the SCX material based on their cationic
strength. Injection of buffer with increasing NH4Ac concen-
tration through the autosampler and 10-port valve in a manner
similar the sample injection allows peptides to be eluted. The
stronger the peptide cationic strength, the higher is the salt
concentration required to elute peptides from the SCX
1594
dx.doi.org/10.1021/ac2028485 | Anal. Chem. 2012, 84, 1592−1600

Analytical Chemistry
Article
Figure 2. LC-SRM analysis targeted the detection of 32 peptides using MudPIT-SRM (Panels A-E) or RP-SRM (Panels F,G). Tryptic peptides from
human plasma were separated using 30 min LC gradients and detected by MudPIT-SRM. (A) Desalting, (B) 15 mM NH4Ac elution, (C) 30 mM
NH4Ac elution, (D) 300 mM NH4Ac elution, and (E) 1.5 M NH4Ac elution. For RP-SRM, peptides were (F) desalted or (G) eluted with 1.5 M
NH4Ac.
material. The whole process only depends on autosampler
directed sample/salt injections and valve switching, thereby
avoiding all offline sample handling after tryptic digestion.
During the entire analysis, three stable isotope labeled reference
peptides (C-reactive protein (CRP); ESDTSYVSLK, labeled
amino acid underlined, and serum paraoxonase/arylesterase 1
1595
dx.doi.org/10.1021/ac2028485 | Anal. Chem. 2012, 84, 1592−1600

Analytical Chemistry
Article
Figure 3. MudPIT-SRM induced peak height increases of (a) complement C3 peptide TELRPGETLNVNFLLR and serum amyloid A protein
peptide SFFSFLGEAFDGAR in human blood plasma and (b) protein S100 peptide LLETEC*PQYIR (* indicates a carbamidomethylation) and
serum amyloid P-component targeted in human wound fluids.
(PON1); peptide 1: YVYIAELLAHK; peptide 2: IQNIL-
TEEPK) are continuously infused into an injection loop using a
syringe pump (placed between valve ports 4 and 7) while the
valve is in the inject position. The reference peptides are
transferred to the analytical column when a sample is loaded or
fractionated by salt injection. Reference peptides, whose
representative XICs are shown in Figure 1b, are retained on
the analytical column prior to target peptide analyte binding
but are eluted collectively with the target peptides, facilitating
relative quantitation within and between sample analyses.
Reproducible injection of the reference peptides is shown in
Figure 1c.
concentration and resuspension after offline SCX fractionation.
Further, offline fractionation strategies are time-consuming and
increase the preanalysis sample preparation complexity.
MudPIT-SRM combines online desalting (cleanup step) and
rapid online SCX fractionation with targeted quantitative mass
spectrometry to facilitate SRM investigations toward clinical
applications in high complex biological material. As MudPIT-
SRM takes advantage of sequential salt fractionation steps,
analysis time is dictated by the degree of sample fractionation
required. For example, a MudPIT-SRM workflow using five
step (1 desalting and 4 salt steps) analysis takes approximately
five times longer than conventional RP-SRM.
The described chromatography setup improves sample
preparation related issues such as sample losses due to
MudPIT-SRM Applied to Human Plasma. To demon-
strate the utility of MudPIT-SRM, we compared it with
1596
dx.doi.org/10.1021/ac2028485 | Anal. Chem. 2012, 84, 1592−1600

Analytical Chemistry
Article
Table 2. Comparison of RP-SRM and MudPIT-SRM Performance in Identifying and Quantifying Targeted Peptides in Human
a
Wound Fluids
protein
peptide sequence
precursor mass (m/z)
product ion mass (m/z)
MudPIT-SRM
d
RP-SRM
peak area changes (%)
n/a
ANXA1
GLGTDEDTLIEILASR
851.9
1015.6
688.4
801.5
895.4
732.4
619.3
732.4
845.5
958.6
1052.5
894.4
995.5
769.4
870.5
1130.6
812.4
755.4
927.5
973.5
787.4
1200.6
933.5
1096.6
862.5
1090.6
833.4
904.5
911.5
1139.7
797.4
1034.5
933.5
734.4
1092.6
934.5
744.4
815.5
1029.6
871.5
968.6
1069.6
887.5
986.6
773.4
1006.5
909.5
503.8
708.4
957.5
871.5
1000.5
825.4
972.5
738.4
551.3
464.3
648.3
849.4
948.5
718.4
ANXA2
ANXA3
SLYYYIQQDTK
GAGTNEDALIEILTTR
GIGTDEFTLNR
711.3
837.4
611.8
876.4
656.4
709.8
630.3
926.5
862.9
851.4
d
q
q
q
d
d
d
n/a
n/a
n/a
GDIR1
GDIR2
MMP2
MMP8
AEEYEFLTPVEEAPK
TLLGDGPVVTDPK
AFQVWSDVTPLR
YYAFDLIAQR
n/a
n/a
q
q
q
q
q
q
q
q
q
−30.3
24.5
9.8
DAFELWSVASPLIFTR
ISQGEADINIAFYQR
QLSLPETGELDSATLK
66.8
n/a
MMP9
AFALWSAVTPLTFTR1
LGLGADVAQVTGALR
841.0
720.9
q
d
q
d
28.3
n/a
PROF1
TFVNITPAEVGVLVGK
SSFYVNGLTLGGQK
STGGAPTFNVTVTK
VGEYSLYIGR
822.5
735.9
690.4
578.8
q
q
q
q
q
q
q
d
60.5
55.3
36.9
n/a
SAMP
AYSLFSYNTQGR
ELPSFLGK
703.8
445.7
624.3
q
d
d
d
d
n/a
n/a
n/a
S100A4
ALDVMVSTFHK
1597
dx.doi.org/10.1021/ac2028485 | Anal. Chem. 2012, 84, 1592−1600

Analytical Chemistry
Article
Table 2. continued
protein
S100A7
peptide sequence
precursor mass (m/z)
product ion mass (m/z)
MudPIT-SRM
q
RP-SRM
q
peak area changes (%)
GTNYLADVFEK
LLETEC*PQYIR
C*IESLIAVFQK
GHFDTLSK
628.8
708.4
821.4
984.5
965.4
1066.5
1196.5
1034.6
818.5
905.5
710.4
563.3
448.3
835.4
1108.5
964.5
417.2
833.4
736.4
−21.2
S100A8
S100A11
S100A12
S100P
711.4
654.4
452.7
679.8
538.3
q
q
d
q
q
q
q
q
49.7
n/a
n/a
YSGSEGSTQTLTK
ELPGFLQSGK
n/a
−63.2
a
Abbreviations used in the table refer to n/a = not applicable, d = detectable, q = quantifiable, and * = carbamidomethylation.
conventional SRM using RP peptide separation using human
plasma. The plasma was immunodepleted of albumin and
immunoglobulin gamma, and 32 peptides from 24 highly
abundant plasma proteins were monitored (Table 1). These
assays of clinical utility were selected on the basis of previous
observations of analytical performance that showed high
quantitative reproducibility using nanoflow LCMS and a 4000
QTRAP with a cohort of cancer patients.30 To maintain similar
reproducibility, the current study utilized RP-SRM with the
MudPIT trapping system still in line using a single high
concentration salt elution step. Therefore, the trapping column
loading capacity and the peptide retention times remained
constant. A sequential 5 step MudPIT-SRM fractionation using
0, 15, 30, 300, and 1500 mM NH4Ac was carried out to achieve
optimal separation without sacrificing excessive analysis time.
An effective organic elution gradient of 30 min follows each salt
fractionation step. Including sample loading and column re-
equilibration, the total analysis time is 7 h per sample.
MudPIT-SRM and RP-SRM detection was compared using
peak intensity and peak area of targeted peptide transitions
(Table 1). Elution profiles of MudPIT-SRM and RP-SRM
analysis (Figure 2) demonstrate improvements in overall peak
intensity and mean peak area for all target peptides using
MudPIT-SRM. Figure 3 depicts examples of intensity increases
associated with MudPIT-SRM for complement C3 (CO3)
peptide TELRPGETLNVNFLLR and the serum amyloid A4
(SAA) peptide SFFSFLGEAFDGAR. From Figure 2, it is
evident that peptides analyzed with MudPIT-SRM showed an
intensity increase of at least 50% compared to conventional RP-
SRM. Determination of AUC for MudPIT- and RP-SRM
(Table 1) further demonstrates the sensitivity enhancement of
noise, the peak areas are able to be accurately summed.
Triplicate MudPIT-SRM analyses of a human plasma sample
reveal excellent reproducibility with mean coefficients of
variation (CV) of 5.40
2.44%, ranging between 0.6 and
11% (Table 1), which is comparable to CVs reported from RP-
SRM studies.11,27,28
Despite the differences of MudPIT- and RP-SRM in the
elution conditions, the two strategies show similarity in the
desalting step. The presence of an APOA1 and a CO3 peptide
in the desalting step of both SRM methods is most likely
contributed to a limitation in binding capacity of peptides on
the SCX material due to the high abundance of these plasma
proteins. Although 5−10% of these peptides were not retained
on the SCX phase, these peptides were still detected and can be
used for quantitation.
The multiple salt step elution increases peptide purity and
reduces the complexity within each elution step improving
known electrospray ionization (ESI) issues such as ion
suppression caused by coelution of analytes. To visualize the
changes in ion suppression, a setup described by Jessome and
Volmer29 has been assessed. Briefly, a mixture of the three
isotope labeled peptides (500 fmol/μL) was continuously
infused (3 μL/h) after the analytical column infused into the
flow path of an elution gradient. Performing a SRM experiment
targeting the three peptides allows for continuous monitoring
of the ionization efficiency. Triplicate analyses of (1) the
background without a complex solution, (2) RP-SRM
separation of human plasma, and (3) MudPIT-SRM separation
of human plasma have been accomplished (Figure S-1,
Supporting Information). Interestingly, this experiment re-
vealed changes in ionization efficiency throughout the entire
gradient, not merely when high amounts of peptides are
coeluting. Monitored background demonstrates reproducible
ionization efficiency of infused peptides within the triplicate
analyses (Figure S-2, Supporting Information). However, both
RP-SRM and MudPIT-SRM showed increased intensities of
CRP and PON1 peptide 2 in the desalting step compared to
the background, which is consistent with improved ionization
in this step. RP-SRM elution of the plasma sample following
the desalting step results in a significant reduction of targeted
MudPIT-SRM as mean peak area increases to 89.2
5.3%
(range of 73−98%) and, therefore, represents almost an order
of magnitude improvement to support improved quantitation
limits. As a consequence of salt step elution methodology,
occasionally, some peptides were spread across more than one
fraction with a high intense peak in one salt fraction and a more
minor component in the next fraction.
However, as long as those peptides are above the lower limit
of quantitation (LLOQ), which we defined as 10 times the
1598
dx.doi.org/10.1021/ac2028485 | Anal. Chem. 2012, 84, 1592−1600

Analytical Chemistry
Article
peptide transition intensities by 70−80%, demonstrating a high
degree of ion suppression. Under the same experimental
conditions, the MudPIT-SRM strategy showed higher ion
counts per second for the desalting step compared to RP-SRM,
which further increased in the 15 mM (10−20% CRP and
PON1 peptide 2; 60% PON1 peptide 1) and 30 mM (50−60%
CRP and PON1 peptide 2; 230% PON1 peptide 1) salt
injection steps before the intensities decreased to comparable
levels as in the RP-SRM elution step when performing the 1.5
M salt elution. Although PON1 peptide 1 showed slightly
different behavior in the ion suppression experiment, which is
most likely related to the 3+ charge state, it followed the overall
trend of the other two peptides.
Targeting Less Concentrated Proteins in Human
Wound Exudates. The MudPIT-SRM method has shown
significant improvements in SRM performance for targeting
highly concentrated plasma proteins; however, a chief purpose
of the online SCX fractionation strategy is not only to increase
quantitation accuracy of highly concentrated proteins but also
to enable the detection and quantitation of less abundant
proteins (lower ng/mL). Additionally, this strategy should
contribute to rapid sample analysis especially for SRM-MS
demands in clinical applications where time can be a key factor.
In order to demonstrate the application of MudPIT-SRM in
this situation, candidate biomarkers from human wound fluids,
which were previously identified in a discovery experiment
comparing fluids of nonhealing wounds to normal healing
wounds from burn victims (manuscript in preparation), have
been analyzed.
A selection of potential discriminating biomarkers that are
commonly known to be of ng/mL concentration in plasma
(e.g., metalloproteinases, S100 proteins) were chosen (Table 2)
and analyzed with MudPIT-SRM using 79 transitions
optimized for SCX elution (Table 2). Corresponding MS/MS
spectra of targeted peptides are shown in Figure S-3,
Supporting Information. In a first instance, peptide elution
gradients were optimized to achieve actual gradient times of
just 10 min. Further improvements might be possible using
faster scanning instruments, since the 10 min gradient is not
limited by the chromatographic separation but the detection of
eluting ions by the MS. Cycle times of 2 s, generated by 20 ms
dwell time, only allow a reduction of the actual gradient time to
10 min. Further gradient reduction would lead to peak width
smaller than 12 s, causing irreproducibility due to inaccurate
data acquisition. Assessment of MudPIT using the highly
abundant plasma proteins (Table 1) showed elution of the
same peptides in the 300 mM and 1.5 M NH4Ac steps; hence,
the 300 mM salt step was removed to optimize reproducibility
and shorten the run. Targeted peptides of low abundant
proteins from a nonhealing wound exudate were analyzed using
the MudPIT-SRM strategy and compared to a RP-SRM
analysis of the same sample.
Triplicate analyses readily demonstrated the advantage of
MudPIT-SRM for detection of less abundant proteins, with
peak area increases of 25−60% (Table 2). Further, the
MudPIT-SRM strategy enabled peptides, which were below
the LLOQ in the RP-SRM approach, to be quantitated. For
example, two different serum amyloid P (SAMP) peptides,
VGEYSLYIGR and AYSLFSYNTQGR, and two Annexin A3
(ANXA3) peptides, GAGTNEDALIEILTTR and GIGT-
DEFTLNR, Rho GDP-dissociation inhibitor (GDIR) 1
(AEEYEFLTPVEEAPK), S100 P peptide YSGSEGSTQTLTK,
S100 A11 (C*IESLIAVF, where * indicates a carbamidome-
thylation of cysteine residues), and the matrix metalloprotei-
nase 9 (MMP9) peptide QLSLPETGELDSATLK were below
the lower limit of detection (LLOD) in conventional RP-SRM
but were quantifiable with MudPIT-SRM. Additionally, the
MudPIT-SRM strategy enabled peptides of ANXA1
(GLGTDEDTLIEILASR) and S100 A4 (ELPSFLGK) to be
detected. These two peptides were below the LLOD using RP-
SRM. However, no improvements were observed for MMP9
peptide LGLGADVAQVTGALR and one S100 A4 peptide
(ALDVMVSTFHK). These peptides remained below the
LLOQ in both chromatographic configurations. Paradoxically,
S100P (ELPGFLQSGK), the S100 A7 peptide GTNYLADV-
FEK, and MMP2 (AFQVWSDVTPLR) revealed significantly
greater peak areas in the RP-SRM strategy. Four of the S100
peptides studied here have previously been analyzed by SRM in
the serum of rheumatoid arthritis patients. In their report, Liao
et al.31used 1 mL of serum, affinity depletion of abundant
proteins, and extensive size exclusion fractionation prior to
SRM. In the current study, we were able to detect each of these
peptides using the MudPIT-SRM configuration following
injection of small volumes of unfractionated wound fluid.
Comparisons of the quantitative characteristics of these two
approaches are difficult to make given differences in samples,
sample preparation strategies, and improvements in mass
spectrometry hardware.
For more than 90% of the targeted low abundant proteins, it
was clearly advantageous to apply MudPIT-SRM over RP-SRM,
allowing peptides to be quantified or detected which were
previously refractory to analyses. Even though MudPIT-SRM
requires longer analysis time compared to RP-SRM, it is the
gains made through sample fractionation that leads to detection
of less abundant analytes. We have shown here that gradient
elution times can be substantially reduced, akin to the time
normally used for RP-SRM, yet tremendous gains in sensitivity
make this approach highly attractive for detection of proteins
that are at low levels or masked in unfractionationed samples
applied to RP-SRM analysis.
CONCLUSION
■
We have described the integration of the widely used MudPIT
approach with SRM and demonstrated its performance
advantages over conventional RP-SRM. MudPIT-SRM reduces
ion-suppression, improves signal by as much as 1 order of
magnitude, and is compatible with rapid chromatography for
analysis of biological fluids. These attributes identify MudPIT-
SRM as a useful method for protein biomarker studies,
particularly suited for the analysis of clinically relevant
biological fluids. One drawback of the MudPIT-SRM approach
is the additional time required to complete the analysis using
multiple salt fractionation steps compared with conventional
RP-SRM. Nonetheless, the superior signal gain achieved with
MudPIT-SRM (∼90%) may overcome this attribute for
analyses of biologically relevant proteins that may otherwise
evade detection using conventional SRM approaches.
ASSOCIATED CONTENT
■
S
* Supporting Information
Ionization efficiency experiments performed on the 4000
QTRAP; real time plots of monitored ion suppression;
acquired SRM-triggered peptide MSMS spectra of targeted
proteins in human wound fluids. This material is available free
of charge via the Internet at http://pubs.acs.org.
1599
dx.doi.org/10.1021/ac2028485 | Anal. Chem. 2012, 84, 1592−1600

Analytical Chemistry
Article
(19) Anderson, N. L.; Anderson, N. G.; Haines, L. R.; Hardie, D. B.;
Olafson, R. W.; Pearson, T. W. J. Proteome Res. 2004, 3, 235−244.
(20) Anderson, N. L.; Jackson, A.; Smith, D.; Hardie, D.; Borchers,
C.; Pearson, T. W. Mol. Cell. Proteomics 2009, 8, 995−1005.
(21) Washburn, M. P.; Wolters, D. A.; Yates, J. R. Nat. Biotechnol.
2001, 19, 242−247.
AUTHOR INFORMATION
■
Corresponding Author
*Address: APAF, Macquarie University, Sydney, Australia,
2109. Tel: +612 9850 6218. Fax: +612 9850 6200. E-mail:
mmolloy@proteome.org.au.
(22) Washburn, M. P.; Ulaszek, R. R.; Yates, J. R. Anal. Chem. 2003,
75, 5054−5061.
ACKNOWLEDGMENTS
(23) Wolters, D. A.; Washburn, M. P.; Yates, J. R. 3rd Anal. Chem.
2001, 73, 5683−5690.
■
C.K. is the recipient of a Macquarie University cotutelle PhD
scholarship. M.P.M. is the recipient of an NHMRC Career
Development Award. Aspects of this research were conducted
at APAF, supported by the Australian Government’s NCRIS/
EIF programs.
(24) Keshishian, H.; Addona, T.; Burgess, M.; Mani, D. R.; Shi, X.;
Kuhn, E.; Sabatine, M. S.; Gerszten, R. E.; Carr, S. A. Mol. Cell.
Proteomics 2009, 8, 2339−2349.
(25) Izrael-Tomasevic, A.; Phu, L.; Phung, Q. T.; Lill, J. R.; Arnott, D.
J. Proteome Res. 2009, 8, 3132−3140.
(26) Addona, T. A.; Shi, X.; Keshishian, H.; Mani, D. R.; Burgess, M.;
Gillette, M. A.; Clauser, K. R.; Shen, D.; Lewis, G. D.; Farrell, L. A.;
Fifer, M. A.; Sabatine, M. S.; Gerszten, R. E.; Carr, S. A. Nat.
Biotechnol. 2011, 29, 635−643.
REFERENCES
■
(1) Anderson, N. L.; Anderson, N. G.; Pearson, T. W.; Borchers, C.
H.; Paulovich, A. G.; Patterson, S. D.; Gillette, M.; Aebersold, R.; Carr,
S. A. Mol. Cell. Proteomics 2009, 8, 883−886.
(2) Lange, V.; Picotti, P.; Domon, B.; Aebersold, R. Mol. Syst. Biol.
2008, 4, 222.
(27) Randall, S. A.; McKay, M. J.; Molloy, M. P. Proteomics 2010, 10,
2050−2056.
(28) Prakash, A.; Rezai, T.; Krastins, B.; Sarracino, D.; Athanas, M.;
Russo, P.; Ross, M. M.; Zhang, H.; Tian, Y.; Kulasingam, V.;
Drabovich, A. P.; Smith, C.; Batruch, I.; Liotta, L.; Petricoin, E.;
Diamandis, E. P.; Chan, D. W.; Lopez, M. F. J. Proteome Res. 2010, 9,
6678−6688.
(3) McIntosh, M.; Fitzgibbon, M. Nat. Biotechnol. 2009, 27, 622−
623.
(4) Rodriguez, H.; Rivers, R.; Kinsinger, C.; Mesri, M.; Hiltke, T.;
Rahbar, A.; Boja, E. Proteomics Clin. Appl. 2010, 4, 904−914.
(5) Schiess, R.; Wollscheid, B.; Aebersold, R. Mol. Oncol. 2009, 3,
33−44.
(6) Wang, P.; Whiteaker, J. R.; Paulovich, A. G. Cancer Biol. Ther.
2009, 8, 1083−1094.
(7) Chiu, C.-L.; Randall, S.; Molloy, M. P. Bioanalysis 2009, 1, 847−
(29) Jessome, L. L.; Volmer, D. A. LCGC North Am. 2006, 24, 498−
510.
(30) McKay, M. J.; Sherman, J.; Laver, M. T.; Baker, M. S.; Clarke, S.
J.; Molloy, M. P. Proteomics Clin. Appl. 2007, 1, 1570−1581.
(31) Liao, H.; Wu, J.; Kuhn, E.; Chin, W.; Chang, B.; Jones, M. D.;
O’Neil, S.; Clauser, K. R.; Karl, J.; Hasler, F.; Roubenoff, R.; Zolg, W.;
Guild, B. C. Arthritis Rheum. 2004, 50, 3792−3803.
855.
(8) Fortin, T.; Salvador, A.; Charrier, J. P.; Lenz, C.; Bettsworth, F.;
Lacoux, X.; Choquet-Kastylevsky, G.; Lemoine, J. Anal. Chem. 2009,
81, 9343−9352.
(9) Gallien, S.; Duriez, E.; Domon, B. J. Mass Spectrom. 2011, 46,
298−312.
(10) Kitteringham, N. R.; Jenkins, R. E.; Lane, C. S.; Elliott, V. L.;
Park, B. K. J. Chromatogr., B 2009, 877, 1229−1239.
(11) Addona, T. A.; Abbatiello, S. E.; Schilling, B.; Skates, S. J.; Mani,
D. R.; Bunk, D. M.; Spiegelman, C. H.; Zimmerman, L. J.; Ham, A.-J.
L.; Keshishian, H.; Hall, S. C.; Allen, S.; Blackman, R. K.; Borchers, C.
H.; Buck, C.; Cardasis, H. L.; Cusack, M. P.; Dodder, N. G.; Gibson, B.
W.; Held, J. M.; Hiltke, T.; Jackson, A.; Johansen, E. B.; Kinsinger, C.
R.; Li, J.; Mesri, M.; Neubert, T. A.; Niles, R. K.; Pulsipher, T. C.;
Ransohoff, D.; Rodriguez, H.; Rudnick, P. A.; Smith, D.; Tabb, D. L.;
Tegeler, T. J.; Variyath, A. M.; Vega-Montoto, L. J.; Wahlander, A.;
Waldemarson, S.; Wang, M.; Whiteaker, J. R.; Zhao, L.; Anderson, N.
L.; Fisher, S. J.; Liebler, D. C.; Paulovich, A. G.; Regnier, F. E.;
Tempst, P.; Carr, S. A. Nat. Biotechnol. 2009, 27, 633−641.
(12) Han, B.; Higgs, R. E. Briefings Funct. Genomics Proteomics 2008,
7, 340−354.
(13) Prakash, A.; Tomazela, D. M.; Frewen, B.; MacLean, B.;
Merrihew, G.; Peterman, S.; MacCoss, M. J. J. Proteome Res. 2009, 8,
2733−2739.
(14) Agger, S. A.; Marney, L. C.; Hoofnagle, A. N. Clin. Chem. 2010,
56, 1804−1813.
(15) Barton, C.; Kay, R. G.; Gentzer, W.; Vitzthum, F.; Pleasance, S.
J. Proteome Res. 2009, 9, 333−340.
(16) Hoofnagle, A. N.; Wu, M.; Gosmanova, A. K.; Becker, J. O.;
Wijsman, E. M.; Brunzell, J. D.; Kahn, S. E.; Knopp, R. H.; Lyons, T. J.;
Heinecke, J. W. Arterioscler., Thromb., Vasc. Biol. 2010, 30, 2528−2534.
(17) Lopez, M. F.; Kuppusamy, R.; Sarracino, D. A.; Prakash, A.;
Athanas, M.; Krastins, B.; Rezai, T.; Sutton, J. N.; Peterman, S.;
Nicolaides, K. J. Proteome Res. 2010, 10, 133−142.
(18) van Winden, A. W. J.; van den Broek, I.; Gast, M.-C. W.;
Engwegen, J. Y. M. N.; Sparidans, R. W.; van Dulken, E. J.; Depla, A.
C. T. M.; Cats, A.; Schellens, J. H. M.; Peeters, P. H. M.; Beijnen, J. H.;
van Gils, C. H. J. Proteome Res. 2010, 9, 3781−3788.
1600
dx.doi.org/10.1021/ac2028485 | Anal. Chem. 2012, 84, 1592−1600