1
A novel titanium dioxide-polydimethylsiloxane plate for phosphopeptide
2
enrichment and mass spectrometry analysis
3
Chao-Jung Chen,1,3* Chien-Chen Lai,5 Mei-Chun Tseng,6 Yu-Ching Liu,3
4
Yu-Huei Liu,1,4 and Fuu-Jen Tsai2,4
5
6
1
Graduate Institute of Integrated Medicine and 2College of Chinese Medicine, China Medical
7
University, Taichung 40402, Taiwan
8
3
9
Genetics and Medical Research, China Medical University Hospital, Taichung 40402, Taiwan
Proteomics Core Laboratory, Department of Medical Research and 4Department of Medical
10
5
11
Taichung 40402, Taiwan
12
6
Institute of Molecular Biology, National Chung Hsing University,
Institute of Chemistry, Academia Sinica, Taipei 115, Taiwan
13
14
15
*Corresponding author:
16
Chao-Jung Chen, Tel: +886 4 22052121 ext 2761; E-mail:cjchen@mail.cmu.edu.tw; Fax:
17
886-4-22037690
1
1
Abstract
2
The phosphorylation of proteins is a major post-translational modification that is required for the
3
regulation of many cellular processes and activities. Mass spectrometry signals of
4
low-abundance phosphorylated peptides are commonly suppressed by the presence of abundant
5
non-phosphorylated peptides. Therefore, one of the major challenges in the detection of
6
low-abundance phosphopeptides is their enrichment from complex peptide mixtures. Titanium
7
dioxide (TiO2) has been proven to be a highly efficient approach for phosphopeptide enrichment
8
and is widely applied. In this study, a novel TiO2 plate was developed by coating TiO2 particles
9
onto polydimethylsiloxane(PDMS)-coated MALDI plates, glass, or plastic substrates. The
10
TiO2-PDMS plate (TP plate) could be used for on-target MALDI-TOF analysis, or as a
11
purification plate on which phosphopeptides were eluted out and subjected to MALDI-TOF or
12
nanoLC-MS/MS analysis. The detection limit of the TP plate was ~10-folds lower than that of a
13
TiO2-packed tip approach. The capacity of the ~2.5 mm diameter TiO2 spots was estimated to be
14
~10 µg of β-casein. Following TiO2 plate enrichment of SCC4 cell lysate digests and
15
nanoLC-MS/MS analysis, ~82% of the detected proteins were phosphorylated, illustrating the
16
sensitivity and effectiveness of the TP plate for phosphoproteomic study.
17
2
1
Keywords:
2
Phosphopeptide; Titanium dioxide; Polydimethylsiloxane; Plate; Mass
3
spectrometry; Phosphoproteomics
4
ABBREVIATIONS
5
PDMS, polydimethylsiloxane; MALDI-TOF MS, matrix-assisted laser desorption
6
ionization time-of-flight mass spectrometry; DHB, 2,5-dihydroxybenzoic acid; PA,
7
phosphoric acid; TFA, trifluoroacetic acid; FA, formic acid.
8
3
1
1. Introduction
2
The reversible phosphorylation of proteins is a major post-translational modification that is
3
required for the regulation of a wide range of biological process. Analyses of the dynamics of
4
protein phosphorylation cascades can help reveal significant signaling events that are correlated
5
with the regulation of specific cellular functions. MS-based approaches have proven to be
6
powerful tools for phosphoprotein characterization and phosphorylation site mapping.1,
7
However, the MS signals of phosphorylated proteins in low abundance are commonly suppressed
8
by
9
phosphopeptides have a higher degree of acidity than unmodified peptides, the positive-ion
10
formation of phosphopeptides is less efficient, resulting in the greater difficulty in identifying
11
phosphopeptides by MS. Therefore, the purification of phosphopeptides from unmodified
12
peptides prior to MS analysis is crucial. To improve the detection sensitivity and enrichment
13
specificity for phosphopeptides, novel approaches have been continuously reported.3-5 One
14
commonly used approach is immobilized metal ion affinity chromatography (IMAC), which uses
15
metal ions such as Fe3+ and Ga3+ to chelate with the phosphate group.6, 7 However, because the
16
IMAC approach is based on ionic interaction, the selectivity is usually limited by the nonspecific
17
binding of acidic non-phosphorylated peptides.8 Because metal oxides are stable in a wide pH
18
range, allowing the use of acidic buffers to prevent nonspecific binding, metal oxide affinity
the
presence
of
abundant
non-phosphorylated
4
peptides.
Furthermore,
2
because
1
chromatography (MOAC) shows better specificity for enriching phosphopeptides. Owing to its
2
high stability and specific affinity toward phosphopeptides, titanium dioxide (TiO2) is gaining
3
more use in MOAC approaches and it is fabricated as columns, pipette tips, or coating layers on
4
magnetic nanoparticles for purification purposes. In the TiO2-LC column system, a large quantity
5
of sample is often used, with the requirement of a high-pressure pumping device. A pressurized
6
system or repeated aspirate-eject cycles of pipetting is operated for packed micro-tips (e.g.,
7
StageTips).9 Complicated synthesis process of magnetic TiO2 particles could increase the
8
experimental cost and difficulty. However, these off-probe methods are often time-consuming
9
and laborious, and suffer sample loss during multistep sample handling.10 In addition, the
10
purification reproducibility may be affected by the beads-packing quality, elution flow rate, and
11
manual operation. In contrast, the direct enrichment and detection of phosphopeptides on a
12
MALDI target is a high-throughput process, and could be more sensitive owing to the fewer
13
steps in sample handling with minimal sample loss.11 One simple approach to immobilize TiO2
14
on MALDI plates is to directly sinter the TiO2 particles onto it.12 However, a drawback of this
15
approach was presented as the lack of control and full characterization of the structure of the
16
TiO2 spots. TiO2 films with a more controllable thickness can by fabricated by the pulsed laser
17
deposition method.13 However, because TiO2 particles or layers are not feasible for reuse and are
18
not easily removed from the plate, these on-probe approaches consume expensive commercial
5
1
MALDI targets. Disposable TiO2 plates have been fabricated by using aluminum foil14 and glass
2
slides15, 16 instead of commercial plates. Printing, spraying, or deposition/calcination10 processes
3
were used to fabricate the TiO2 film on the substrates. However, these disposable TiO2 plates
4
have to be equipped with modified MALDI targets or homemade sample stages.
5
Herein, a simple approach was developed to fabricate removable TiO2 spot arrays on
6
substrates for phosphopeptide purification or on-target MALDI-TOF analysis. By first coating a
7
polydimethylsiloxane (PDMS) film on the MALDI target or other substrates, TiO2 particles can
8
then be adhered onto the PDMS-coated plates, followed by on-probe MALDI-TOF detection or
9
phosphopeptide purification. The sensitivity, selectivity, reproducibility, and sample capacity
10
toward phosphopeptides of this novel design and its application in phosphoproteomics were
11
demonstrated.
12
13
2. Materials and Methods
14
PDMS (polydimethylsiloxane) prepolymer was purchased from Dow Corning (Sylgard®
15
184, Midland, MI, USA). Acetonitrile (ACN), formic acid (FA), and NH4OH were purchased
16
from J. T. Baker (Phillipsburg, NJ, USA). Dithiothreitol (DTT), iodoacetamide (IAA),
17
N,N,N′,N′-tetramethylethylenediamine (TEMED), phosphoric acid (PA), TFA, α-casein, β-casein,
6
1
ovalbumin, myoglobin, cytochrome C, angiotensin II, adrenocorticotropic hormone (ACTH)
2
fragment 18-39 (RPVKVYPNGAEDESAEAFPLEF, 2465.2 m/z), and TiO2 nanoparticles (<150
3
nm, part no. 700347) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Tryptic BSA
4
and 2,5-dihydroxybenzoic acid (DHB) were purchased from Bruker Daltonics (Germany).
5
Trypsin (modified, sequencing grade) was from Promega (Madison, WI, USA). A mixture of 4
6
phosphopeptides (part no. 186003285) was purchased from Waters (Milford, MA, USA). TiO2
7
particles (5 µm) were purchased from GL Sciences (Tokyo, Japan). BSA, ammonium
8
bicarbonate (ABC), Tris-HCl, and SDS were purchased from Bio Basic Inc. (Toronto, Canada).
9
Deionized water (Milli-Q, Millipore, USA) was used to prepare the sample and matrix solutions.
10
2.1 Protein digestion
11
Ovalbumin, α-casein, β-casein, BSA, myoglobin, or cytochrome C protein was individually
12
dissolved in 50 mM ABC and heated to 90°C for 20 min. The denatured proteins were reduced
13
with 10 mM DTT for 20 min at 56°C, followed by the addition of 55 mM IAA for 30 min in the
14
dark at 25°C. Trypsin was added to the protein solution at an enzyme-to-substrate ratio of 1:50
15
(w/w) for 12 h at 37°C. The peptide solution was dried in a centrifugal concentrator (miVac Duo
16
Concentrator; Genevac, Stone Ridge, NY, USA).
17
2.2 Phosphopeptide enrichment by TiO2-PDMS plate (TP plate)
7
1
The flow chart for TP plate fabrication and phosphopeptide enrichment is shown in Fig. 1.
2
In step 1, a PDMS-coated plate was first fabricated. Prepolymer (reagent A) was mixed in an
3
Eppendorf tube with its curing agent (reagent B) in a 10:1 volume ratio and spun down. An
4
aliquot (3-5 µL) of the mixture was smeared on a glass/plastic plate, or a stainless steel MALDI
5
plate (ground target; Bruker Daltonics). A roller (part no. 165-1279; Bio-Rad) was used to flatten
6
the PDMS prepolymer to make a thin layer on the plate, which was then incubated in an oven for
7
polymerization at 80°C for 1 h. Once the PDMS film had been formed, the plate was washed
8
with a 0.1% FA solution. In step 2, aliquots (5 µL) of a TiO2 particle solution (1mg in 500 μl,
9
70% ACN) were deposited on a PDMS-coated plate to make the spot arrays. The plate was then
10
incubated in an oven at 80°C for 10 min, after which it was flushed with water to remove the
11
excess TiO2. Such TiO2-PDMS plate (hereafter named TP plate) was then ready for use.
12
The TiO2 spots on the TP plate were first washed with an 80% ACN/2% TFA solution to
13
remove contaminants and nonspecifically adsorbed compounds. Protein digest samples dissolved
14
in loading buffer (80% ACN, 2% TFA, and 20-200 mg mL-1 of DHB) were loaded onto the TiO2
15
spots and incubated for a certain time. The spots were then washed with the 80% ACN/2% TFA
16
solution to remove non-phosphorylated peptides. For on-probe MALDI-TOF analysis,
17
phosphopeptides were eluted with 3-5 µL of 0.05% NH4OH. After the spots had dried, 2 µL of a
8
1
DHB/PA matrix solution (2 mg mL-1 DHB in 25% ACN and 1% PA) was applied to the TiO2
2
spots, and after DHB/analyte co-crystallization, the TP plate was sent into the MALDI ionization
3
source for MALDI-TOF analysis. For PDMS-coated plate/MALDI-TOF analysis, the
4
NH4OH-eluted phosphopeptide solution on TiO2 spots was collected and deposited on the
5
PDMS-coated MALDI plate, followed by matrix deposition and MALDI-TOF analysis. For cell
6
lysate analysis, the NH4OH-eluted phosphopeptide solution on TiO2 spots was collected and
7
dried in a centrifugal concentrator, followed by resuspension in 0.1% FA solution and
8
nanoLC-MS/MS analysis.
9
2.3 Phosphopeptide enrichment by TiO2-packed tips
10
TiO2-packed tips were fabricated according to a previously published method.17 In this
11
study, TiO2 beads with a packed length of ~2 mm were adopted. In the detection limit test,
12
β-casein in an 80% ACN/2% TFA/20 mg mL-1 DHB solution was loaded into the TiO2-packed
13
tip. The tip was then washed with 25 µL of an 80% ACN/2% TFA solution to remove
14
non-phosphorylated peptides. The bound phosphopeptides were eluted with 20 µL of 0.05%
15
NH4OH (pH 10.5). Water washing and another elution with 10 µL of a 50% ACN/0.1% FA
16
solution were used to elute phosphopeptides that may be adsorbed on the C8 disk. The sample
17
loading, washing, and elution steps were all operated by centrifugation. The solutions eluted
9
1
from 0.05% NH4OH and 50% ACN/0.1% FA were collected separately and dried in a centrifugal
2
concentrator, without combination to avoid salts formation. The dried samples were dissolved
3
with 2 µL of a DHB/PA matrix solution (2 mg mL-1 DHB in 25% ACN and 1% PA) and
4
combined, followed by analysis with MALDI-TOF on a PDMS-coated MALDI target.
5
2.4 Cell culture, lysis, and digestion
6
The SCC4 oral epithelial cell line was grown in DMEM/F12 supplemented with 10% fetal
7
bovine serum and 1% penicillin G at 37°C in a 5% CO2 atmosphere. The cells were harvested,
8
washed 3 times with PBS, and then lysed in lysis buffer (0.25 M Tris-HCl, pH 6.8, 1% SDS).
9
The proteins from the cell lysate were subjected to gel-assisted digestion.18, 19 Solutions of 18.5
10
μL of acrylamide (40%, 29:1), 2.5 μL of 10% ammonium persulfate, and 1 μL of TEMED were
11
sequentially added to the 400-µg protein solution (50 μL). The tube-gel was cut into small pieces
12
and washed 4 times with 25 mM ABC (pH 8.2) containing 50% ACN for 15 min each time,
13
followed by dehydration with pure ACN. The gel pieces were reduced with 200 μL of 10 mM
14
DTT for 30 min at 56°C, followed by alkylation with 200 μL of 55 mM IAA for 30 min at 25°C
15
in the dark. After alkylation, the gel pieces were washed with 500 µL of 25 mM ABC/50% ACN
16
buffer, dehydrated with 100% ACN, and digested with trypsin (1:25 (w/w) trypsin-to-protein
17
ratio) in 25 mM ABC at 37°C overnight. After digestion, the tryptic peptides were extracted from
10
1
the gel using 0.1% FA in 50% ACN, and 100% ACN sequentially. All of the extracted solutions
2
were combined and applied to TP plate for phosphopeptide purification and nanoLC-MS/MS
3
analysis.
4
2.5 MALDI-TOF MS Analysis
5
Peptides and phosphopeptides were analyzed by MALDI-TOF/TOF-MS (Ultraflex III
6
TOF/TOF; Bruker Daltonics) equipped with a smartbeam laser. Peptide mass calibration for
7
MALDI-TOF was performed with a peptide calibration standard kit (Bruker Daltonics) to
8
calibrate an instrument mass range of 800–4000 Da. Spectra were acquired in reflector mode,
9
using a 25-kV accelerating voltage, 26.3-kV reflector voltage.
10
2.6 NanoLC-MS/MS analysis and database search
11
NanoLC-MS/MS was performed with a nanoflow UPLC system (UltiMate 3000 RSLCnano
12
system; Dionex, The Netherlands) coupled with a hybrid Q-TOF mass spectrometer (maXis
13
impact; Bruker). The sample was injected into a homemade tunnel-frit trap column (5 μm C18
14
particles, 180 μm i.d. x 20 mm)20 with a flow rate of 8 μL min-1 and a duration of 4 min. The
15
trapped analytes were separated on a commercial analytical column (Acclaim PepMap C18, 2
16
μm, 100 Å, 75 μm i.d. × 250 mm; Thermo Scientific, USA) with a flow rate of 300 nL min-1. An
11
1
acetonitrile/water gradient of 1-50% within 40 min was used for separation of the
2
phosphorylated peptides. For MS/MS detection, peptides with a charge of 2+, 3+, or 4+ were
3
selected for data-dependent acquisition.
4
The LC-MS/MS data were searched against the SwissProt (release 51.0) database using
5
the MASCOT search algorithm (version 2.2.07). The search parameters of MASCOT for
6
peptides and MS/MS mass tolerance were 0.05 Da. The search parameters selected were as
7
follows. Taxonomy: human; enzyme: trypsin; fixed modification: carbamidomethyl (cysteine);
8
and variable modification: oxidation (methionine) and phosphorylation (serine, threonine, and
9
tyrosine); missed cleavage: 2. Peptides and phosphopeptides were considered to be identified if
10
their ion score was higher than 25. False discovery rates (FDRs) on MASCOT were estimated to
11
be < 0.3%.
12
3. Results and Discussion
13
3.1 TP plate fabrication
14
PDMS, an elastic, transparent, chemically inert, and hydrophobic polymer, is broadly
15
applied in microfluidic systems. Recently, we used PDMS as the coating material to fabricate
16
hydrophobic MALDI plates for improving the detection sensitivity of biomolecules.21 In this
12
1
study, we found that by applying an aqueous solution of TiO2 particles onto a PDMS polymer
2
surface, followed by air-drying or heating in an oven, the TiO2 particles were immobilized on the
3
PDMS membrane surface after water flushing. This special phenomenon may be attributed to the
4
formation of noncovalent bonding or oxygen-bridged heterometallic Si-O-Ti bonds between the
5
TiO2 and the siloxane group of PDMS.22, 23 As shown in Fig. 1, to fabricate TiO2 spots, aliquots
6
(5 μL) of TiO2 particle solution (70% ACN) were deposited on the PDMS-coated plate. After
7
flushing the plate surface with water, excess non-adhered TiO2 was flushed out from the PDMS
8
surface, leaving behind a compact, thin, and circular layer of TiO2 spot arrays with diameters of
9
~2.5 mm2 (Fig. 2). The adhesion strength between the TiO2 particles and PDMS membrane could
10
tolerate repeated pipetting without the particles peeling off. The TP plate can be fabricated on
11
plastic substrates or on a MALDI target plate for on-probe MALDI-TOF analysis. After use, the
12
TiO2-PDMS film can easily be manually peeled off and refabricated with the steps shown in Fig.
13
1, to avoid any sample memory effect.
14
3.2 Phosphopeptide enrichment of TP plate with micro- and nano-TiO2 particles
15
Because the acidic residues of glutamic acid and aspartic acid can cause nonspecific binding
16
between the peptides and TiO2, nonspecific binding of non-phosphorylated peptides is often
17
encountered on the TiO2 particles. In a recent study, DHB was added to the sample loading
13
1
solution in order to alleviate electrostatic interactions between acidic residues and TiO2 while
2
retaining the high binding affinity of TiO2 for phosphopeptides.17 Therefore, in this study, DHB
3
was added to the sample loading solution and applied to the TP plate to alleviate nonspecific
4
binding of non-phosphorylated peptides. The enrichment efficiency of the TP plate was first
5
evaluated using 1 μg of ovalbumin digests. Without enrichment, non-phosphorylated peptides of
6
the ovalbumin digest presented as dominant peaks in the MS spectrum, whereas the ion signal of
7
the phosphopeptide (2088.89 m/z, EVVGSAEAGVDAASVSEEFR) was relative low
8
(Supplemental Fig. S-1a). After on-probe analysis with the TP plate containing 5-µm-sized TiO2
9
particles, the 2088.89 m/z peak signal was greatly enhanced as the major peak, and no
10
non-phosphopeptides peaks appeared (Supplemental Fig. S-1b). The metastable ion (2002.8 m/z)
11
appeared due to the loss of an 86 m/z fragment of the 2088.89 m/z phosphopeptide peak rather
12
than the loss of phosphoric acid as H3PO4 (98 Da) or HPO3 (80 Da). Metastable phosphopeptide
13
ions formed from the loss of 86 Da fragments have also been reported by Vener et al.
14
Bykova et al..25
24
and
15
Owing to the high surface area-to-volume ratio of nanoparticles, TiO2 nanoparticles could
16
be expected to have a higher sample capacity for phosphopeptide enrichment and more
17
applicability in a plate system as compared with column or tip systems. Therefore, the
14
1
performance of the TP plate made of nano-TiO2 particles (<150 nm) was also evaluated. As
2
shown in Supplemental Fig. S-1c, the 2088.89 m/z peak presented as the highest peak; however,
3
non-phosphopeptide peaks were also significantly observed. The nonspecific binding of the
4
nanoTP plate system could be attributed to the more abundant active sites of nanoparticles
5
available for binding phospho- and non-phosphopeptides, so that the nonspecific binding force
6
could not be effectively eliminated, even by using saturated DHB. Therefore, the 5-µm-sized
7
TiO2 particles were chosen to be more optimal for fabricating the TP plate in this study, and this
8
was also the size widely used in tip and column systems.
9
3.3 Phosphopeptide enrichment from complex sample by TP plate
10
The enrichment performance of TP plate in dealing with trace amount of complex sample
11
was also evaluated by purifying phosphopeptides from a protein mixture containing 3
12
non-phosphorylated proteins (BSA, myoglobin, and cytochrome C) and 3 phosphorylated
13
proteins (α-casein, β-casein, and ovalbumin). As shown in Fig. 3a, without TP plate enrichment,
14
non-phosphorylated peptides were the major peaks in the spectrum; after TP plate enrichment,
15
the phosphopeptide peaks presented as the major peaks, whereas the non-phosphorylated peptide
16
signals were nearly unobservable (Fig. 3b). Table 1 summarizes the identified phosphopeptides
17
and their average signals from 4 TiO2-spot measurements (1 spectrum acquired from each TiO2
15
1
spot). The selectivity of the TP plate was further evaluated by capturing phosphorylated peptides
2
from complex mixtures of 3 phosphorylated protein digests (α-casein, β-casein, and ovalbumin)
3
and 3 non-phosphorylated protein digests (BSA, myoglobin, and cytochrome C) with ratios of
4
1:100 (mol/mol).
5
major peaks. (Supplemental Fig. S-2) These results indicated the high selectivity of the TP plate
6
in enriching phosphopeptides from a complicated protein mixture, thereby improving the
7
phosphopeptide signals.
8
3.4 Recovery and detection limit
After TP plate enrichment, the phosphopeptide peaks still presented as the
9
To evaluate the phosphopeptide recovery of the TP plate, the 0.05% NH4OH-eluted
10
phosphopeptides (VNQIG(pT)LSESIK, 1368.678 m/z; and VNQIGTL(pS)E(pS)IK, 1448.644
11
m/z) were deposited on a PDMS-coated plate with a mix of an internal standard (2 fm
12
angiotensin II, theoretical mass of 1046.542 Da). Before the TP plate enrichment, 2
13
phosphopeptides (1368.68 m/z and 1448.64 m/z) and 1 peptide (angiotension II, 1046.54 m/z)
14
were observed, with the peak height ratios of 1368.68/1046.54 m/z and 1448.64/1046.54 m/z
15
being ~0.37 (STD, 0.05) and ~0.13 (STD, 0.04), respectively. After TP plate enrichment, these
16
peak height ratios were slightly decreased to ~0.33 (standard deviation (STD), 0.05) and ~0.12
17
(STD, 0.03), respectively (Fig 4). Therefore, the recovery for the two 20 fm phosphopeptides
16
1
was estimated to be ~90% and ~92%, respectively, illustrating a very low sample loss in dealing
2
with such trace sample amounts.
3
The detection limit for TP plate-enriched phosphopeptides was also evaluated and compared
4
with that from TiO2-packed tips. One spectrum was acquired from each TiO2 spot or tip
5
replicated measurements. On-probe MALDI-TOF analysis of the 5 fm β-casein digests showed
6
that the average S/N (STD) signals of the phosphopeptide (2061.8 m/z) was ~12.0 (3.6) from 3
7
TiO2-spots measurements (Fig. 5a). The eluted phosphopeptides from the TP plate were also
8
applied to a PDMS-coated plate for improving its sensitivity, on which 2 fm β-casein digests
9
were observable with an average S/N (STD) of ~14.7 (4.7) from 4 TiO2-spots measurements (Fig
10
5b). Using TiO2-packed tips, the 2061.8 m/z phosphopeptide peak was only detectable on a
11
PDMS-coated plate when applying 20 fm β-casein digests with an average S/N (STD) of ~12.6
12
(10.1) from 3 TiO2-packed tips measurements (Fig. 5c). The higher sensitivity of the TP plate
13
with 4 folds (on-probe MALDI-TOF analysis) and 10 folds (with TP plate enrichment and
14
PDMS-coated plate detection) enhancement compared with TiO2-packed tips may be attributed
15
to the available use of lower TiO2 particle quantities on each spot than on the tips, resulting in
16
lower sample loss from the TiO2 beads. Furthermore, the additional centrifugal concentration
17
steps to reduce the eluted sample solution in the TiO2-packed tips carry the risk of sample loss.
17
1
3.5 On-spot concentration
2
The TP plate showed an additional advantage, having the configuration of hydrophilic TiO2
3
spots surrounded by a hydrophobic PDMS coating, which can enlarge the contact angles of the
4
sample droplets, allowing for applying larger sample volumes on the TiO2 spot in order to
5
concentrate low-abundance samples. Increasing the analyte concentration or the sample volume
6
usually results in improved MS detection sensitivity. In most cases, it is more desirable to apply a
7
large sample volume, because it allows for the direct use of diluted samples, which avoids
8
sample loss or contamination in concentration preparations. The large volume concentration on
9
the TP plate was also investigated. Samples of β-casein tryptic digests with a loading volume of
10
1μL (20 nM), 10 μL (2 nM), 20 μL (1 nM) and 50 μL (0.4 nM) were deposited on ~2.5 mm i.d.
11
TiO2 spots, and their corresponding concentration results were evaluated by MALDI-TOF
12
analysis with an internal standard peak of a 4 fmol ACTH peptide. Supplemental Fig. S-3a shows
13
the peak ratio curve of the phosphopeptide peak (2061.8 m/z) to the internal standard peak
14
(2465.2 m/z) in the various diluted samples. The 2 nM β-casein digests of 10 μL volume were
15
completely concentrated onto the spot, their peak ratio being similar to that of 1 μL (20 nM)
16
β-casein digests. However, significant sample loss of ~36% and ~57% occurred in the use of the
17
20 μL and 50 μL diluted sample volumes, respectively, due to their loaded area exceeding that of
18
1
the TiO2 spots (~2.5 mm2) (Supplemental Fig. S-3b). The maximum applicable sample volume
2
for on TiO2-spot concentration can be increased on a larger size of TiO2 spot, which can easily be
3
made by applying a larger volume of TiO2 particle solution on the PDMS-coated plate. Therefore,
4
the TP plate could also be a feasible approach for preconcentrating low-concentration
5
phosphopeptides.
6
3.6 Sample capacity
7
Sample capacity is an important issue for sample enrichment approaches. With the
8
restriction of the plate format, the sample capacity could be too low to meet with the analysis of
9
cell lysate, and the sample capacity of previous studies is rarely reported. H. Wang et al.10 have
10
reported their photocatalytically patterned TiO2 plate having sample capacity of 1 pmol (~24 ng)
11
for β-casein on 1 mm spots, which is comparable with the result reported with IMAC spots. J. D.
12
Dun et al.26 also reported their phosphopeptide enrichment plate with polymer brushes having ~6
13
ng of phosphoangiotensin on 1 mm spots. To evaluate the sample capacity of the TP plate,
14
increasing amounts of β-casein digests from 0.5 to 100 µg were applied to the TP plate, on which
15
TiO2 spot arrays of ~2.5 mm diameter had been made. The eluted phosphopeptides from the
16
increasing concentration of samples were separately mixed with the same increased-folds
17
amount of ACTH peptide (18-39 clip) (ACTH fragment, RPVKVYPNGAEDESAEAFPLEF,
19
1
2465.2 m/z) that acted as the internal standard signal. For example, 100 fm of ACTH peptide was
2
mixed with the eluted phosphopeptide of 0.5 µg β-casein, and 1 pm, 2 pm, 10 pm, and 20 pm
3
ACTH were likewise separately mixed with the eluted β-casein samples of 5, 10, 50, and 100 µg
4
prior to analysis on the PDMS-coated plate. The peak ratio at each sample applied amount was
5
determined with 3 TiO2-spot replicated measurements (2 spectra were acquired for each TiO2
6
spot). Supplemental Fig. S-4 shows a similar peak signal ratio (2061.8/2465.2 m/z) of ~15 for
7
0.5 and 5 µg of β-casein, indicating that this protein amount was still below the beads capacity. A
8
slightly decreased ratio of ~13.6 was observed for the 10 µg sample, and the ratios dropped
9
significantly for the 50 µg and 100 µg sample applications. Therefore, the maximum sample
10
capacity of the ~2.5 mm diameter TiO2 spots was estimated to be ~10 µg of β-casein and was
11
further calculated as ~1.6 µg on 1 mm spots. To our knowledge, our TP plate provides the
12
highest sample capacity when compared to previous studies and the reason could be attributed to
13
the spherical binding surface of the TiO2 micro-particles.
14
3.7 Phosphoprotein analysis of SCC4 cell lysate
15
NanoLC-MS/MS
has
been
proven
to
be
a
highly
reliable
strategy
for
16
protein/phosphoprotein identification in complex protein samples.27-29 A high protein amount is
17
usually injected for nanoLC-MS/MS, to increase the identified protein numbers. Because of the
20
1
low stoichiometry of phosphorylation, a higher initial protein amount (100 µg to 1 mg) is usually
2
needed to obtain a higher amount of phosphopeptides after TiO2 enrichment. To deal with a large
3
amount of initial protein sample, TiO2-packed tips or TiO2-packed LC columns with a larger
4
sample capacity is the first selection. However, with the continuously improving sensitivity of
5
LC-MS techniques, the applied sample amount is gradually decreasing, allowing for the analysis
6
of trace and valuable samples. The high sample recovery of plate-based enrichment approaches
7
are attractive alternatives in dealing with trace samples. To evaluate the performance of the TP
8
plate for phosphoproteomics analysis, the SCC4 cell line, a malignant keratinocyte line from
9
squamous carcinomas of the human tongue, was lysed and purified on the TP plate, followed by
10
nanoLC-MS/MS analysis. Given that the sample capacity of the ~2.5 mm i.d. TiO 2 spot is ~10
11
µg, TiO2 spots with ~5 mm i.d. were used to purify the ~40 µg protein samples without
12
exceeding their sample capacity. Results showed that ~460 peptides and ~287 proteins
13
(averaging 4 TiO2-spots replicates) were identified (Supplemental Table S-1); ~80% (RSD,
14
~5.6%) of peptides and ~82% (RSD, ~3.8%) of proteins were phosphorylated. The high
15
selectivity in the analysis of complicated cell lysate is comparable to the reported results of
16
packed-tip
17
(http://www.piercenet.com/product/magnetic-phosphopeptide-enrichment-kit). Herein, a protein–
18
protein interaction network of the identified phosphoproteins was constructed by ingenuity
methods20,30
and
magnetic
21
TiO2
particles
1
pathways analysis (IPA Ingenuity System Inc, USA), where the networks related to reported
2
functions were generated from the identified phosphoproteins (Supplemental Fig. S-5(a-d) and
3
Table S-2). One of the networks including the most proteins was designated as “cell death and
4
survival, cellular assembly and organization, cell morphology” and contained 72 proteins. (Fig.6
5
and Fig. S-4a) A total of 45 proteins were identified by our strategy and marked with a red color
6
in the network. Some proteins in the network were of great interest in cancer development,
7
metastasis, and apoptosis. For example, epidermal growth factor receptor (EGFR), a cell-surface
8
receptor, is activated by binding of its specific ligands of the EGF family. EGFR has been
9
recognized as an important target for cancer therapy, and various kinds of EGFR inhibitors are
10
currently being used in the treatment of human cancers.31 Different phosphorylation patterns of
11
EGFR result in opposite effects of several cancer-related downstream signaling activations.32, 33
12
Phosphorylation at Thr-693 on EGFR was detected in this study. CD44 is a cell-surface
13
glycoprotein that participates in a wide variety of cellular functions, including cell–cell
14
interactions, and cell adhesion and migration, as well as tumor metastasis.34 The detection of
15
phosphorylation at CD44 Ser-706 may represent its active state for further protein kinase C
16
active signaling. Jun (transcription factor AP-1) is one of the oncogenic transcription factors. The
17
phosphorylation of Jun on the sites of Ser-63 and Ser-73 by Jun N-terminal kinases causes
18
increased transcriptional activity.35 The activation and repression of Jun-dependent pathways
22
1
have been a hot issue in studies of tumor development.36 Phosphorylation at Ser-63 on Jun was
2
detected in this study, indicating the tumor characteristic of the SCC4 cell line. Tyrosine
3
phosphorylation was also identified on cyclin-dependent kinase 1 (CDK1) and Bcl-2-associated
4
transcription factor 1 (BCLF1), both related to apoptosis. Knowing the phosphoprotein profiles
5
and their pivotal roles at a more global level will be helpful for the thorough investigation of the
6
signaling transduction pathways responding to, for example, kinase inhibitor pharmacology.
7
4. Conclusions
8
This study devised a simple and novel approach for phosphopeptides enrichment and
9
on-probe MALDI-TOF analysis. By applying a TiO2 particle solution onto a PDMS-coated plate,
10
TiO2 spot arrays were formed after drying. The TP plate demonstrated the effective enrichment
11
of phosphopeptides from complex protein samples, with good sample recovery and a low
12
detection limit. Compared with previously reported approaches, the TP plate has the advantages
13
of simpler fabrication procedure, lower sample loss, higher analytical reproducibility, high
14
throughput, and ease in manual operation. The high percentage of phosphorylated proteins
15
detected shows that the TP plate coupled with nanoLC-MS/MS analysis is an effective and
16
sensitive approach to analyze phosphoproteins from cell lysates. This platform can be further
17
applied for profiling dynamic phosphorylation with quantitative proteomics, such as
23
1
isotope-labeled (iTRAQ, isobaric tags for relative and absolute quantitation)37 and label-free
2
quantitative approaches.38
3
Acknowledgements
4
This
5
101-2113-M-039-002-MY2,
6
(BM102021124), diabetes (BM102010130), and stroke biosignature (BM102021169) projects
7
from Academia Sinica, Taiwan.
work
was
supported
by grants
from
the
National
NSC101-2632-B-039-001-MY3),
8
9
24
Science
Council
Urothelial
(NSC
Carcinoma
1
REFERENCES
2
[1]
J.A. Jebanathirajah, J.L. Pittman, B.A. Thomson, B.A. Budnik, P. Kaur, M. Rape, M.
3
Kirschner, C.E. Costello, P.B. O'Connor, J. Am. Soc. Mass Spectrom. 16 (2005)
4
1985-1999.
5
[2]
C. T. Chen, Y. C. Chen, Anal. Chem. 77 (2005) 5912-5919.
6
[3]
A.B. Iliuk, V. A. Martin, B.M. Alicie, R.L. Geahlen, W.A. Tao, Mol. Cell. Proteomics 9
7
8
(2010) 2162-2172.
[4]
9
10
X. Zhao, Q. Wang, S. Wang, X. Zou, M. An, X. Zhang, J. Ji, J. Proteome Res. 12 (2013)
2467-2476.
[5]
11
Y. J. Tan, D. Sui, W. H. Wang, M. H. Kuo, G. E. Reid, M. L. Bruening, Anal. Chem. 85
(2013) 5699-5706.
12
[6]
A. Stensballe, S. Andersen, O. N. Jensen, Proteomics 1 (2001) 207-222.
13
[7]
M.C. Posewitz, P. Tempst, Anal. Chem. 71 (1999) 2883-2892.
14
[8]
J.D. Dunn, G.E. Reid, M.L. Bruening, Mass Spectrom. Rev. 29 (2010) 29-54.
15
[9]
J. Rappsilber, M. Mann, Y. Ishihama, Nat. Protoc. 2 (2007) 1896-1906.
25
1
[10]
H. Wang, J. Duan, Q. Cheng, Anal. Chem. 83 (2011) 1624-1631.
2
[11]
Y. Zhang, L. Li, P. Yang, H. Lu, Anal. Methods 4 (2012) 2622-2631.
3
[12]
L. Qiao, C. Roussel, J. Wan, P. Yang, H.H. Girault, B. Liu, J. Proteome Res. 6 (2007)
4
5
4763-4769.
[13]
6
7
1932-1942.
[14]
8
9
[15]
[16]
A. Eriksson, J. Bergquist, K. Edwards, A. Hagfeldt, D. Malmstrom, V. Agmo Hernandez,
Anal. Chem. 82 (2010) 4577-4583.
[17]
14
15
M.L. Niklew, U. Hochkirch, A. Melikyan, T. Moritz, S. Kurzawski, H. Schluter, I. Ebner,
M.W. Linscheid, Anal. Chem. 82 (2010) 1047-1053.
12
13
H. Bi, L. Qiao, J.M. Busnel, V. Devaud, B. Liu, H.H. Girault, Anal. Chem. 81 (2009)
1177-1183.
10
11
F. Torta, M. Fusi, C.S. Casari, C.E. Bottani, A. Bachi, J. Proteome Res. 8 (2009)
M.R. Larsen, T.E. Thingholm, O.N. Jensen, P. Roepstorff, T.J. Jorgensen, Mol. Cell.
Proteomics 4 (2005) 873-886.
[18]
X. Lu, H. Zhu, Mol. Cell. Proteomics 4 (2005) 1948-1958.
26
1
[19]
2
Y.T. Lu, C.L. Han, C.L. Wu, T.M. Yu, C.W. Chien, C.L. Liu, Y.J. Chen, Proteomics Clin.
Appl. 2 (2008) 1208-1222.
3
[20]
C.J. Chen, W.Y. Chen, M.C. Tseng, Y.R. Chen, Anal. Chem. 84 (2012) 297-303.
4
[21]
C.J. Chen, C.C. Lai, M.C. Tseng, Y.C. Liu, S.Y. Lin, F.J. Tsai, Anal. Chim. Acta 783
5
(2013) 31-38.
6
[22]
D. Hoebbel, T. Reinert, H. Schmidt, J. Sol. Gel. Sci. Technol. 6 (1996) 139-149.
7
[23]
D. Hoebbel, T. Reinert, H. Schmidt, J. Sol. Gel. Sci. Technol. 7 (1996) 217-224.
8
[24]
A.V. Vener, A. Harms, M.R. Sussman, R.D. Vierstra, J. Biol. Chem. 276 (2001)
9
10
959-6966.
[25]
11
12
(2003) 26021-26030.
[26] J.D. Dunn, E.A. Igrisan, A.M. Palumbo, G.E. Reid, M. L. Bruening, Anal. Chem. 80
13
14
15
N.V. Bykova, A. Stensballe, H. Egsgaard, O.N. Jensen, I.M. Moller, J. Biol. Chem. 278
(2008), 5727–5735.
[27]
H.Y. Wu, V.S.M. Tseng, L.C. Chen, Y.C. Chang, P.P. Ping, C.C. Liao, Y.G,Tsay, J.S. Yu, P.
C. Liao, Anal. Chem. 81 (2009) 7778-7787.
27
1
[28]
2
3
Chen, J. Proteome Res. 7 (2008) 4058-4069.
[29]
4
5
[30]
Y. Kyono, N. Sugiyama, K. Imami, M. Tomita, Y. Ishihama J. Proteome Res. 7 (2008)
4585–4593.
[31]
8
9
F. Wu, P. Wang, J. Zhang, L.C. Young, R. Lai, L. Li, Mol. Cell. Proteomics 9 (2010)
1616-1632.
6
7
C.F. Tsai, Y.T. Wang, Y.R. Chen, C.Y. Lai, P.Y. Lin, K.T. Pan, J.Y. Chen, K.H. Khoo, Y.J.
Y. Hiraishi, T. Wada, K. Nakatani, I. Tojyo, T. Matsumoto, N. Kiga, K. Negoro, S. Fujita,
Pathol. Oncol. Res. 14 (2008) 39-43.
[32]
10
J.S. Biscardi, M.C. Maa, D.A. Tice, M.E. Cox, T.H. Leu, S.J. Parsons, J. Biol. Chem. 274
(1999) 8335-8343.
11
[33]
J.S. Biscardi, D.A. Tice, S.J. Parsons, Adv. Cancer Res. 76 (1999) 61-119.
12
[34]
H. Ponta, L. Sherman, P.A. Herrlich, Nat. Rev. Mol. Cell Biol. 4 (2003) 33-45.
13
[35]
R.J. Davis, Cell 103 (2000) 239-252.
14
[36]
P.K. Vogt, Nat. Rev. Cancer 2 (2002), 465-469.
28
1
[37]
2
3
4
D. Ghosh, Z. Li, X.F. Tan, T.K. Lim, Y. Mao, Q. Lin, Mol. Cell. Proteomics. 7 (2013)
1865-1880.
[38]
S. Nahnsen, C. Bielow, K. Reinert, O. Kohlbacher, Mol. Cell. Proteomics 12 (2013)
549-556.
5
29
1
FIGURE CAPTIONS
2
Fig. 1. Schematic representation of the TP plate fabrication and phosphopeptide enrichment
3
procedures.
4
Fig. 2. TiO2 spot arrays (~2.5 mm i.d.) on a MALDI-TOF target. The image was obtained using
5
an optical scanner.
6
Fig. 3. MALDI mass spectra of a 6-protein digests mixture (20 fm of BSA, myoglobin,
7
cytochrome C, α-casein, β-casein, and ovalbumin) (a) before and (b) after TP plate enrichment.
8
●α-S1-casein, ◆ α-S2-casein, ★ β-casein, ▇ ovalbumin, ○ metastable ion after loss of 86 Da of
9
the phosphopeptide peak.
10
Fig. 4. Phosphopeptide recovery test of the TP plate. The two 20-fm phosphopeptide standard
11
with and without TP plate enrichment were separately mixed with the internal standard of
12
angiotensin II (2 fm, 1046.54 m/z) and deposited on a PDMS-coated plate for MALDI-TOF
13
analysis. The sequence of the 2 phosphopeptides are VNQIG(pT)LSESIK (1368.68 m/z) and
14
VNQIGTL(pS)E(pS)IK (1448.64 m/z). The comparison results were obtained from averaging
15
the signals of 3 TiO2-spot measurements (3 spectra acquired from each TiO2 spot).
16
Fig. 5. Detection limit test of the TP plate and TiO2-packed tips by MALDI-TOF. (a)
30
1
Phosphopeptide from 5 fm β-casein digests was purified and detected on the TP plate. For
2
off-probe analysis, β-casein digests were captured by (b) the TP plate and (c) TiO2-packed tips,
3
and the eluents were deposited on PDMS-coated plates for MALDI-TOF analysis.
4
Fig. 6. The protein–protein interaction network of “Cell Death and Survival, Cell Cycle, and
5
Organismal Survival”, generated by ingenuity pathways analysis. Identified proteins in this study
6
were marked with a red color in 4 networks.
7
8
9
10
11
12
13
TABLES
14
15
16
17
18
19
20
21
Table 1. Identified phosphopeptides from digests of ovalbumin, α-casein S1 and S2, and β-casein.
Their signals were compared before and after TP-plate enrichment.
22
23
24
31
1
2
3
Fig. 1
32
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Fig. 2
20
21
22
23
24
25
26
27
33
1
2
3
4
5
6
7
8
9
10
11
Fig. 3
12
13
14
15
16
17
18
19
34
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Fig. 4
15
16
17
18
35
1
2
3
4
5
6
7
8
9
10
11
Fig. 5
36
1
2
3
4
5
6
7
8
Fig.6
37
1
2
3
4
5
6
7
8
9
10
Table 1
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
38