A pseudo-targeted MS method for the sensitive analysis of protein

Subscriber access provided by UNIV OF DURHAM

A pseudo-targeted MS method for the sensitive analysis of protein phosphorylation in protein complexes Jiawen Lyu, Yan Wang, Jiawei Mao, Yating Yao, Shujuan Wang, Yong Zheng, and Mingliang Ye Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b00749 • Publication Date (Web): 16 Apr 2018 Downloaded from http://pubs.acs.org on April 16, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 30 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

1

A pseudo-targeted MS method for the sensitive analysis of

2

protein phosphorylation in protein complexes

3

4

Jiawen Lyu1,2, Yan Wang1,2, Jiawei Mao1,2, Yating Yao1,2, Shujuan Wang3, Yong Zheng3, Mingliang

5

Ye1,2*

6 7

1

8

R&A Center, Dalian Institute of Chemical Physics, Chinese Academy of Sciences (CAS), Dalian

9

116023, China;

CAS Key Laboratory of Separation Sciences for Analytical Chemistry, National Chromatographic

10

2

University of Chinese Academy of Sciences, Beijing 100049, China

11

3

State Key Laboratory of Proteomics, Beijing Proteome Research Center, National Center for

12

Protein Sciences (Beijing), Beijing Institute of Lifeomics, Beijing 102206, China

13

*

14

411-84379620. E-mail: [email protected].

15

16

17

To whom correspondence should be addressed: (M.L. Ye) Phone: +86-411-84379610. Fax: +86-

18

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

1

2

Abstract

3

In this study, we presented an enrichment-free approach for the sensitive analysis

4

of protein phosphorylation in minute amounts of samples, such as purified protein

5

complexes. This method takes advantage of the high sensitivity of parallel reaction

6

monitoring (PRM). Specifically, low confident phosphopeptides identified from the

7

data dependent acquisition (DDA) dataset were used to build a pseudo-targeted list for

8

PRM analysis to allow the identification of additional phosphopeptides with high

9

confidence. The development of this targeted approach is very easy as the same

10

sample and the same LC-system were used for the discovery and the targeted analysis

11

phases. No sample fractionation or enrichment was required for the discovery phase

12

which allowed this method to analyze minute amount of sample. We applied this

13

pseudo-targeted MS method to quantitatively examine phosphopeptides in affinity

14

purified endogenous Shc1 protein complexes at four temporal stages of EGF signaling

15

and identified 82 phospho-sites. To our knowledge, this is the highest number of

16

phospho-sites identified from the protein complexes. This pseudo-targeted MS

17

method is highly sensitive in the identification of low abundance phosphopeptides,

18

and could be a powerful tool to study phosphorylation-regulated assembly of protein

19

complex.

20


Page 2 of 30

Page 3 of 30 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

2

1. Introduction

3

Proteins barely perform biological functions on their own but interact with each

4

other to execute an obligation together1. Protein complex is a group of interacted

5

proteins that work together to execute a specific biologic function2. Protein

6

phosphorylation plays a crucial role in regulating the assembling of protein complex

7

3,4

8

complex. Direct analysis of protein digest by shotgun proteomics for protein

9

phosphorylation is very poor in sensitivity because the phosphopeptides are of low

10

abundance in the sample and their ionization is seriously suppressed by the coexisted

11

non-phosphopeptides5. To

12

phosphopeptides, specific enrichment of phosphopeptides was often performed prior

13

to LC-MS/MS analysis6,7, which could identify and quantify over 10,000 of

14

phosphopeptides in a typical phosphoproteome analysis8,9. Though the enrichment

15

method performs very well for phosphoproteomics analysis, it might be not fitted to

16

analyze trace amount of sample, especially in the protein complex sample which is

17

often as low as a few microgram level. This is because huge sample loss may occur

18

when enrichment was performed for trace amount of sample. Alternatively, the sample

19

complexity can also be reduced by fractionation of the peptide sample prior to LC-

20

MS/MS analysis, which allowed identification of low abundant peptides10. To our

21

knowledge, this multidimensional separation scheme was never applied to analyze

22

protein complex probably because it also require large amount of sample.

. It is important to analyze the dynamical change of phosphorylation in the protein

reduce

the

interference


of

high

abundant

non-


1

Shotgun proteomics with mass spectrometer operated in data dependent

2

acquisition (DDA) mode is a powerful tool for the discovery of new proteins but faces

3

serious issues with reproducibility and sensitivity11. More importantly, this data

4

acquisition method is difficult to identify low abundant peptides as they have less

5

chance to be delivered to MS2 for fragmentation12. Targeted proteomics using either

6

parallel reaction monitoring (PRM) or multiple reaction monitoring (MRM) has

7

gained popular recent years due to its high sensitivity, high reproducibility and high

8

accuracy in quantification

9

the analysis of protein phosphorylation in protein complex. A classical targeted MS

10

workflow includes two phases. First, in a pilot experiment for discovery, a large

11

quantity of sample is subjected to either 2D-LC-MS/MS analysis or phosphopeptide

12

enrichment to maximize the identifications. Second, in a targeted analysis phase, the

13

identifications of interest can be specifically monitored by 1D-LC-MS with MRM or

14

PRM analysis to quantify across different samples. The sample amount for the pilot

15

experiment was always over 100 ug for phosphorylation analysis15-17. This amount is

16

clearly too much for the development of targeted approach for the analysis of protein

17

complex. In this study, we aim to develop a sensitive method for the identification and

18

quantification of p-sites in the protein complex.

13,14

. It is of interest to develop targeted MS approach for

19

The recent developed quadrupole-Orbitrap offers specific trapping capacities to

20

enhance the analysis of low abundance peptides18,19. Compared with the MRM

21

method in triple quadrupole MS, PRM in quadrupole-Orbitrap has higher resolution

22

and selectivity20. In addition to quantitative proteomics, PRM can also be applied in


Page 4 of 30

Page 5 of 30 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

targeted peptide identification as it has all fragment ion information of the pre-

2

selected precursor instead of giving only 3-5 pre-selected transitions in MRM21.

3

Taking the advantage of PRM’s high sensitivity and the feature of available full

4

fragment ions in tandem spectra for peptide identification, we proposed a pseudo-

5

targeted MS method for the sensitive analysis of protein phosphorylation in protein

6

complex. The development of this targeted approach is very easy as the same sample

7

and the same LC-system with different acquisition modes, i.e. DDA and PRM were

8

used for the two phases. No sample fractionation or enrichment was required for the

9

discovery phase which allowed this method to analyze minute amount of sample. This

10

method was applied to analyze the phosphorylation in endogenous Shc1 complexes.

11

With three runs of DDA in discovery phase and three runs of PRM in targeted

12

analysis, totally about 3 µg was sufficient for the whole workflow and the

13

identification of phospho-sites increased by around 50%. Take its advantage of

14

accurate quantification, the PRM based method enabled comprehensive mapping the

15

dynamic change of the Shc-1 complexes upon EGF stimulation.

16

17

2. Method

18

Chemicals.

19

DMEM medium was from Gibco (Thermo Fisher Scientific). Anti-Flag M2 agarose

20

beads were from Sigma. The Flag-Shc1 HeLa cell line was a gift from Beijing

21

Proteome Research Center (BPRC). The wild type Hela cell was purchased from

22

National Infrastructure of Cell Line Resource and was used as negative control. The



1

following antibodies were used in western blotting: anti-EGFR pY1068, anti-ERBB2,

2

anti-ERBB3 from Cell Signal Technology, anti-SHC and anti-GRB2 from Abcam,

3

anti-PIK3C2B from Proteintech and anti-Flag from Sigma Aldrich. All the other

4

reagents were purchased from Sigma Aldrich.

5

Cell culture and western blotting verification.

6

The cell line was seeded at 15cm dish (Corning) and cultured at 37℃ in 5% CO2

7

with DMEM plus 10% FBS. To generate phosphopeptide samples for method

8

comparisons, cells were cultured in serum-free medium for 12 h followed with the

9

stimulation with EGF (100 ng/ml) for 0, 2, 5, 20 min. Western blotting verification

10

was performed with anti-Flag and anti-EGFR pY-1068 antibodies.

11

Immunoprecipitation and digestion of Shc-1 complexes.

12

The immunoprecipitation of Shc-1 complexes were performed as previously

13

described22. Briefly, the cells stimulated with EGF were immediately washed three

14

times with pre-cold PBS to quench the cell signal transduction. Then each dish of

15

cells was lysed on ice by NP40 lysis buffer (50 mM HEPES-NaOH, pH 8.0, 150 mM

16

NaCl, 0.5% NP40, 2.5 mM MgCl2, 1 mM DTT, 2% protease inhibitor cocktail, 10%

17

glycerol, 1 mM NaF, 1 mM β-glycerolphosphate, 10 mM Na4P2O7, 1 mM NaVO4).

18

Then the total cell lysate was centrifuged at 20800 g and 4℃ for 20 min and the

19

supernatant was transferred. The protein concentration of the lysate was determined

20

by Bio-Rad protein assay. Flag-Shc1 was pulled down by adding 6 µL (bed volume)

21

anti-Flag M2 agarose beads for 4h or overnight at 4℃. The beads were spun down at

22

5000 g and then washed 3 times with lysis buffer and 2 times with 20 mM ammonium

23

bicarbonate (ABC). The cleaned beads were suspended in 20 µL 20 mM ABC. To


Page 6 of 30

Page 7 of 30 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

perform an on-bead digestion, 400 ng trypsin (from Sigma) was added to the

2

suspension before incubating the suspension at 37℃ overnight with shaking at 800

3

rpm. The digestion was stopped by adding 3% formic acid to the reaction and the

4

supernatant was transferred into a clean tube and dried down.

5

Enrichment of phosphorylated peptides by Ti-IMAC.

6

The sample pretreatment process was described in previous publications6. Briefly,

7

Ti-IMAC was added to the digest according to a ratio of 1:20 (peptides: beads). After

8

30min incubation, the beads were wash by buffer 1 (50% ACN, 6% TFA, 200 mM

9

NaCl) to remove the non-specific peptides, followed with another wash step with

10

buffer 2 (30% ACN, 0.1% TFA) to desalt. Finally, the phosphopeptides were eluted

11

from beads by adding 10% ammonium hydroxide and dried down.

12

Liquid Chromatography.

13

The dried samples were redissolved in 0.1% FA (1% FA only for Ti-IMAC

14

enrichment sample) and loaded onto a 209-µm inner diameter 3-cm trap column

15

(packed in-house with 5 µm, 120 Å, C-18 resins, from Sunchrom) using a flowrate of

16

5 µL/min of mobile phase A and washed for 10 min. H2O containing 0.1% formic acid

17

was used as mobile phase A, while acetonitrile containing 0.1% formic acid was used

18

as mobile phase B. Then the peptides were eluted from the trap column and the

19

reversed-phase separation was accomplished using a 150 -µm inner diameter 20-cm

20

analytical column with a pulled tip (packed in-house with ReproSil-Pur C18-AQ 1.9

21

µm resin). The 130 min separation gradient was set as follow: the flowrate is 600

22

nL/min, 2% mobile phase B from 0 to 10 min, 7% B at 11 min, 27% B at 75 min, 45%



1

B at 105 min, 90% B at 108 min, 90% B at 118 min, 2% B at 120 min, 2% B at 130

2

min.

3

Mass spectrometry.

4

All data were acquired by Q-Exactive HF hybrid quadrupole-Orbitrap mass

5

spectrometer (Thermo, San Jose, CA) except the comparison with iDDA where Q-

6

Exactive hybrid quadrupole-Orbitrap mass spectrometer was used (Refer to

7

supporting information). The DDA runs were conducted by a TopN method in which a

8

high resolution (resolution of 60000 at 200 m/z) full MS acquisition was followed by

9

20 fast dd-MS2 acquisitions (resolution of 15000). The FTMS acquisition of full MS

10

was set as followed: Automatic Gain Control (AGC) target, 3e6; Maximum injection

11

time (Max IT), 20 ms; scan range, from 350 to 2000 m/z. The parameter of dd-MS2

12

acquisition was set as following: AGC target, 5e5; Max IT, 50 ms; isolation window,

13

1.6 m/z; normalized collision energy, 27%; Centriod mode. The ions that carrying

14

charge lower than +1 and higher than +8 were excluded, and dynamic exclusion was

15

set as 20 s.

16

The PRM runs were conducted under a Full MS and PRM tandem method with

17

inclusion mode on. The isolation list for the identification of phosphopeptides was

18

generated by sorting the potential phosphopeptides that identified by previous DDA,

19

the process of which is described in the next section. The MS was run on positive

20

mode. For full MS scanning, the resolution was set at 60000, AGC target 3e6, Max.IT

21

20 ms and a profile mode was used for spectrum data. For the setting of MS2, which

22

refers to PRM in this method, resolution was set at 30000, AGC target 5e5, Max.IT


Page 8 of 30

Page 9 of 30 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

503ms, isolation window 1.6 m/z, NCE 27, and the profile mode was also used for

2

spectrum data of MS2. For details of PRM quantification, please refer to supporting

3

information.

4

Building of pseudo-targeted library.

5

The human proteome was obtained from UniProt (http://www.uniprot.org/) and

6

included 20157 sequences of proteins . All of the *.raw files that generated from DDA

7

method were converted to *.MGF files by Thermo Proteome Discoverer (version 1.4).

8

Mascot was used to search the *.MGF files against the database of human with

9

parameter set as follow: precursor-ion mass tolerance, 10 ppm; fragment-ion mass

10

tolerance, 0.05 Da; protease, trypsin with 2 missed cleavages. Variable modifications

11

were set: oxidation on methionine (M, +15.9949 Da), phosphorylation on

12

serine/threonine/tyrosine (S/T/Y, +79.9663 Da). After database searching, the proteins

13

were filtered with < 1% FDR. The proteins identified in each sample were used to

14

generate a new focused database for re-analysis of the *.MGF files to identify

15

potential phosphopeptides. For each matched phosphopeptides, the top-scored

16

spectrum was extracted from the original searching DAT file via an in-house Java

17

language package. Thus, the retention time, m/z, peptide sequence, peptide score and

18

protein accessory of the phosphopeptides were available for the subsequent formation

19

of PRM inclusion list. The phosphopeptides that didn’t belong to the protein

20

complexes were removed and phosphopeptides that scored higher than 30 were also

21

discarded as they were already confidently identified. The phosphopeptide

22

identifications that conformed to the above criteria would be considered as targets for



1

subsequent PRM acquisition. The isolation windows were set to ±3min around the

2

extracted retention time of the top-scored spectrum.

3

Data analysis.

4

All phosphorylation sites, which will be abbreviated as phospho-site or p-site,

5

were identified by MaxQuant23 as it has a module to evaluate the p-site localization

6

confidence. Mascot does not has this function, but its data is easy for us to extract the

7

information on the retention time, m/z, peptide sequence that need for the generation

8

of pseudo-targeted library . The reviewed human proteome mentioned above was also

9

used as a database in the identification of phospho-site. Trypsin was set as the specific

10

enzyme and max miss cleavage was set at 2. Besides phosphorylation on S/T/Y,

11

oxidation on methionine and acetylation on protein N-term were also set as variable

12

modifications. For other items, the default settings were retained. The exported p-sites

13

that localization probability less than 0.75 and score less than 40 were discarded.

14

Quantitative data were processed with Skyline24 and MS2 ion chromatograms of

15

interested peptides were used as surrogates for the quantification of proteins or

16

phospho-sites. For details of label free quantification (LFQ) and PRM quantification,

17

please refer to supporting information. The raw data were uploaded onto JPOST

18

Repository. The accession numbers are PXD008923 for ProteomeXchange and

19

JPST000385 for jPOST.

20

3. Results and discussion

21

Protein phosphorylation plays an important role in the dynamical assembling of

22

protein complex. It is well known that the EGF receptor tyrosine kinase (EGFR) is


Page 10 of 30

Page 11 of 30 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

autophosphorylated upon EGF stimulation, which provide the binding sites for

2

scaffold protein Shc122.

3

phosphorylated at multiple sites to provide docking sites for cytoplasmic targets.

4

Therefore, Shc1 will recruit many proteins to form protein complexes upon EGF

5

stimulation. Using EGF-dependent Shc1 complex as a test example, we first applied

6

the conventional method to analyze its protein components and their phosphorylation

7

after 2 min EGF stimulation. A special HeLa cell line expressing Flag tagged Shc1

8

was used for this study. The presence of Flag tagged Shc1 and the activation of EGFR

9

were verified by western blotting (Figure S1). When the Shc1 was pulled down by

10

immunoprecipitation using anti-Flag antibody, its bona fide interaction proteins as

11

well as a large number of nonspecific bound proteins were also pulled down25. To

12

distinguish these non-specific bound proteins, the Hela cells also expressing Flag tag,

13

e.g. Flag-GFP, would be an ideal control. However, such a cell line is not available in

14

our lab. Instead, the wild type Hela cells were used as the negative control in this

15

study. After stimulated with EGF for 2min, the cells from these two cell lines were

16

lysed and the Shc1 complex were immunoprecipitated under the same conditions.

17

The obtained proteins were then digested and analyzed by LC-MS/MS in DDA mode

18

as in the standard shotgun proteomics workflow. The identified proteins from these

19

two samples were compared by label-free quantification using MaxLFQ26 and the

20

proteins quantified with significant difference (p < 0.05) was obtained by t test

21

(permutation-based FDR control, 1%) using Perseus27. Only the proteins that

22

exhibited significant high expression in the Flag-Shc1 HeLa cell line were considered

Once Shc1 is associated with EGFR, it will be



Page 12 of 30

1

as the potential components of the protein complexes (Figure S2). This resulted in the

2

identification of 29 potential protein components. Since Shc1 recruits proteins to form

3

protein complexes upon EGF stimulation, the bona fide complex components will

4

increase in abundance in the pulled sample with EGF stimulation compared with the

5

unstimulated one. Because of its high sensitivity, PRM was applied to monitor the

6

dynamical change of the potential components in the Shc1 complex between the cells

7

treated with EGF for 2min or not. In addition to the 29 candidate proteins identified

8

above, 7 additional proteins, PPP1R12A, PEAK1, AP2A2, PIK3CB, ERRFI1, ASAP3,

9

PPP1CB, that were reported to be related to EGF-dependent Shc1 complexes22 were

10

also added into the potential component list for PRM analysis (see supplementary

11

results). After analysis, we compared the abundances of above 36 proteins in the IP

12

samples derived from Flag-Shc1 HeLa cells stimulated with 2min or not. It was

13

found that 33 proteins were up-regulated over 50% (Figure S3a). These proteins

14

together with Shc1, totally 34, were confirmed as the components of the EGF-

15

dependent Shc1 complex. Among these 34 components, 20 were reported

16

increases in abundance for five proteins, i.e. PIK3C2B, ERBB3, ERBB2, SHC, and

17

GRB2, after EGF stimulation were verified by Western Blotting (Figure S4). After the

18

component proteins were determined, we then identified the phosphorylation sites on

19

these proteins using DDA data acquired from the IP sample with 2 min EGF

20

stimulation. We searched the DDA data against human proteome database using

21

MaxQuant by setting variable phosphorylation modification on residues of

22

Ser/Thr/Tyr. A total of 2166 unique peptides with score > 40 were identified from 592


22

. The

Page 13 of 30 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

proteins, including 94 unique phosphopeptides. It was found 52 unique

2

phosphopeptides were derived from the 34 component proteins, which resulted in the

3

identification of 42 p-sites with the localization probability higher than 0.75.

4

Using conventional approach, we identified 42 p-sites from the Shc1 complex

5

upon EGF stimulation for 2min. We then proposed a pseudo-targeted MS method as

6

shown in Figure 1 to identify more p-sites from the protein complex. It is well known

7

that targeted analysis using PRM or MRM is more sensitive than conventional

8

shotgun proteomics using DDA .Unlike MRM, PRM can provide full tandem spectra

9

for peptide identification. PRM was usually used in targeted proteomics, but not in

10

discovery proteomics, as the peptides to be analyzed must be set before LC-MS/MS

11

analysis. Then how can we take the advantage of its high sensitivity to identify

12

additional phosphopeptides with low abundance? The key issue is to build a library of

13

potential phosphopeptides to be monitored. We reason that the potential list could be

14

obtained from the DDA data. DDA is known to bias to identify high abundant

15

peptides28. It has poor sensitivity to identify low abundant phosphopeptides especially

16

when the unphosphorylated peptides and phosphorylated peptides were loaded

17

together for LC-MS/MS analysis. However, we believe that some low abundant

18

phosphopeptides are also fragmented in DDA mode but failed to be identified using

19

the strict filtering criteria. To discover these potential low abundant phosphopeptides,

20

we used a focused database containing only the proteins identified in the sample for

21

database searching of the DDA data. For the case of Shc1 complex sample, we firstly

22

identified the proteins presented in the 2min IP sample through searching the DDA



1

data against human proteome database by MASCOT. These identified proteins were

2

then used to construct a focused database, which was used to search the same DDA

3

data to identify the potential phosphopeptides. No filtering criteria were applied for

4

this step so that we can maximize the number of potential phosphopeptide

5

identifications. From this step, we identified 1757 potential phosphopeptides. Among

6

these, 222 phosphopeptides were derived from the component proteins of Shc1

7

complex. The high scored phosphopeptide identifications were removed as they were

8

already confidently identified. After this step, 152 phosphopeptide identifications

9

scored less than 30 were left. These identifications were of low confidence, but some

10

true positive phosphopeptide identifications should present in this list as the peptide

11

mass and a few fragment ions were matched. Their masses, charges and retention

12

times were used to build an inclusion list for PRM analysis. Considering the protein

13

complex sample is not as complex as the proteome sample, relative wide retention

14

time isolation window of 6 min was set to trap the potential phosphopeptides as much

15

as possible. After the targeted analysis of these peptides by PRM, the resulting tandem

16

spectra were searched against the human database. In this step, we used MaxQuant for

17

database searching to identify phosphopeptides and localize p-sites. It should be noted

18

two search engines, Mascot and MaxQuant, were used in this study. Mascot was used

19

to generate the potential phosphopeptide library for PRM analysis because the

20

retention time was more easy to be extracted by our in house written Java script while

21

MaxQuant was used to identify phosphopeptides and phospho-sites in the final results

22

because it has a module to evaluate the confidence of identified p-sites. After filtered


Page 14 of 30

Page 15 of 30 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

with the same strict criteria as in DDA data, 167 unique non-phosphopeptides and 43

2

unique phosphopeptides were confidently identified from the PRM dataset. Because

3

the presences of false matched phosphopeptides in the low-scored phosphopeptide

4

target list, the non-phosphopeptides or phosphopeptides were also identified from

5

proteins other than those presented in the protein complex. These identifications were

6

removed and only the 37 unique phosphopeptides derived from Shc1 protein complex

7

were kept. These phosphopeptides leaded to the mapping of 34 p-sites on 13

8

component proteins. Among the 34 p-sites, 19 were the same with the 42 p-sites

9

identified by DDA. Therefore, 15 high confident p-sites were newly identified by this

10

method. Thus, the total number of p-sites identified from 2 min stimulated sample was

11

increased to 57. We named this method as the pseudo-targeted MS method because

12

we do not know if the targets we set are true positive target peptides or not. Among

13

152 potential phosphopeptides from the Shc1 complex that were targeted during PRM

14

analysis, only 37 unique phosphopeptides were confidently identified from the protein

15

complex after searching PRM spectra. Clearly most of them are not true positive

16

targets. However, it should be noted that the scores for 49 peptides were improved

17

(above the identical line) after PRM analysis (Figure 2a). One example is given in

18

Figure 2b. A very poor MS/MS spectrum was generated in DDA, while a fragment

19

rich spectrum was generated in PRM mode for the same phosphopeptide. The DDA

20

spectrum cannot yield a confident phosphopeptide identification (Mascot score of 6)

21

while the PRM spectrum yielded a high confident phosphopeptide identification with

22

score of 34. Clearly the improved spectra quality in PRM enabled confident



1

identification of extra p-sites in Shc1 protein complex.

2

In above experiments, three replicate runs were performed and searched together

3

for the analysis of the 2min stimulated IP sample in both the DDA and PRM modes.

4

Due to the high complexity of the proteome sample and the random sampling in DDA

5

mode, running LC-MS/MS for multiple times is an effective way to improve the

6

analysis coverage in shot-gun proteomics29. We investigated if this is the case for the

7

analysis of p-sites in protein complex. For easy comparison, the MS data for the three

8

replicate runs were searched separately. As shown in Figure 3a, the three replicate

9

DDA runs identified 35, 33 and 36 p-sites on the Shc1 complex, respectively.

10

Combining the results of these three runs leaded to the identification of 43 p-sites

11

(Figure 3a) . The number did not increase significantly as most of these identifications

12

were the same for different runs (Figure 3b). This is probably because the IP sample is

13

much simpler than the common complex proteome sample and so that the mass

14

spectrometer can reproducibly fragment these phosphopeptides. For the PRM analysis,

15

the three replicate runs leaded to the identification of 30, 32 and 32 p-sites from the

16

Shc1 protein complex (Figure 3c). Among them, 12, 12 and 14 p-sites were newly

17

identified from the Shc1 protein complex compared with the p-sites already identified

18

by DDA data. Combining these three PRM runs, the newly identified p-sites on Shc1

19

complex increased to 17. The increase for the three replicate PRM runs was even

20

lower than the DDA data as the reproducibility for the targeted analysis was better

21

(Figure 3d). Clearly, replicate runs, either in DDA or PRM, does not contribute many

22

new identifications. However, compared with the 42 sites identified by DDA, the


Page 16 of 30

Page 17 of 30 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

PRM analysis operated in pseudo-targeted strategy leaded to 15 new site

2

identifications, which improved by 35.7%. Clearly the pseudo-targeted strategy is

3

able to identify low abundant phosphopeptides that cannot be achieved by

4

conventional DDA method.

5

Inclusion list can also used in DDA mode, termed as iDDA, to enhance the

6

detection sensitivity of the peptides of interest30. We then compared the performance

7

of iDDA with PRM using the 2min stimulated IP sample. Filling time of 503 ms was

8

set in PRM as it yielded maximum number of phospopeptides during our initial

9

optimization of PRM (data was not shown). For fair comparation, filling time of 503

10

ms was also set in iDDA rather than 50ms in conventional DDA we mentioned above.

11

Basically, the comparison experiments were performed in parallel with identical

12

conditions, i.e. the same sample, the same inclusion list, the same machine (QE rather

13

than QE HF was used here) with the same parameter setting. It can be seen from

14

Figure S5.a that both iDDA and PRM yielded much more p-site identifications than

15

conventional DDA did. However, the p-sites identified by PRM were 37.5% more

16

than those identified by iDDA. Clearly PRM has higher sensitivity. This could be

17

partly attributed to their different trigger mechanism. In iDDA, MS2 were triggered

18

only when the presence of the peaks in the full MS corresponding to m/z values set in

19

the inclusion list. While in PRM, MS2 were triggered by pre-set retention times and

20

m/z values independent on MS scan. The low abundance phosphopeptides may not

21

yield strong enough peaks in MS1 to trigger MS2 in iDDA, however these MS2

22

spectra can still be available in PRM mode. This explained why more MS2 spectra



Page 18 of 30

1

from the protein complex were collected in PRM mode compared with in iDDA mode

2

(Figure S5.b).

3

This PRM based pseudo-targeted method was further applied to analyze the Shc1

4

complexes with other EGF stimulation times. Obvious increases in the number of the

5

identified p-sites were observed in all the samples as shown in Table 1. Especially 21

6

more p-sites were identified by PRM in the 20 min sample, which accounted for 50%

7

of the p-sites identified by DDA. From this study, we identified 10, 57, 63 and 63 p-

8

sites in Shc1 complexes at 4 states of EGF signaling pathway (rest stage/0 min, early

9

stage/2 min, medium stage/5 min, late stage/20 min), respectively. Accumulatively, a

10

total of 82 unique p-sites were identified in this study. The identified p-sites were

11

listed in Table S2. It was found that over 96% of these p-sites were included in

12

PhosphoSitePlus,

13

modifications31. To our knowledge, this is the highest number of p-sites identified

14

from the EGF-dependent Shc1 complexes. In a previous work22, 22 p-sites were

15

identified on the Shc1 complexes. With the benefits of high sensitivity and high

16

resolution of PRM, the identification of p-sites in the protein complex was

17

dramatically enhanced by the pseudo-targeted strategy.

a

database

of

experimentally

observed

post-translational

18

In conventional phosphoproteomics analysis workflow, the phosphopeptides were

19

specifically enriched from protein digest prior to LC-MS/MS run. Because the

20

interferences

21

phosphoproteomics coverage could be dramatically improved. We also tested the

22

performance of this approach for the analysis of the p-sites in protein complex. The

from

the

non-phosphopeptides

were


almost

eliminated,

the

Page 19 of 30 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

digest of the 20min EGF stimulated Shc1 complex IP sample, which has been

2

identified most p-sites in both acquisition modes, was subjected to Ti-IMAC

3

enrichment followed by LC-MS/MS analysis. It was found only 5 p-sites on the Shc1

4

complex were identified, which is much fewer than the identification in sample

5

without Ti-IMAC treatment (see Table.1). This is not difficult to understand. The

6

pulled down complex is about a few µg, which is much fewer than the common initial

7

amount for Ti-IMAC enrichment of proteome sample (typically >100 µg15). The

8

enrichment step generated huge sample loss due to the minute amount of sample,

9

which resulted in poor identification. Alternatively, the method presented in this study

10

without enrichment yielded better results.

11

Due to its advantage of accurate quantification, the PRM based method enabled

12

comprehensive mapping the dynamic change of the shc-1 complex upon EGF

13

stimulation (Figure 4). After EGF stimulation, a series of proteins has been recruited

14

to the complex. A total of 33 proteins were up-regulated after EGF stimulating, among

15

which 12/ 15/ 6 proteins achieved peak point at 2min/ 5min/ 20min, respectively

16

(Figure S3b). Some proteins were down-regulated in later stages, as the intensity

17

decreased by 50% comparing to the peak value but still maintained at a remarkable

18

high level. Similarly, dynamical change in protein phosphorylation on the protein

19

complexes was also observed in different stages of EGF stimulation. The numbers of

20

p-sites identified on each protein were labeled on Figure 4. With the assembling of

21

protein complex and signal transduction, the total number of p-sites reached

22

maximum at late stage (20 min after EGF stimulation). About 42.7% p-sites were



1

identified in all EGF stimulated states (2min/5min/20min) (Figure S6), suggesting

2

these sites probably started to be activated immediately after the cells were stimulated

3

and most of p-sites maintained activated from early stage to late stage. The peptides

4

that carrying the p-sites were monitored by PRM and the dynamic change of 45 p-

5

sites across the stimulation was observed (Table S1 and Figure S7.a). Changes in

6

levels of phosphopeptides were further normalized to changes in protein expression to

7

derive changing tendency in occupancy of phosphorylation sites (Figure S7.b). The

8

normalized fold-change indicates the change of phospho-site occupancy. It was found

9

the occupancy for many sites altered during the EGF stimulation, suggesting

10

phosphorylation-mediated dynamic regulation of protein complex. For example, the

11

quantitative tendencies of Y1172 and Y1197 on EGFR were observed to reach peaks

12

after 5 min stimulation, which is similar to the result of previous work that the two p-

13

sites reached peaks at early stage and then decreased22. After normalized by protein

14

change, the occupancy of these two sites started to increase at 2min and were down-

15

regulated at 5 min. The occupancies that seemed high at 20 min may be the result of

16

down-regulation in protein level. The dynamic occupancy changes on these two sites

17

suggested that the sites started to be activated immediately after stimulation and didn’t

18

increase as the same proportion as the increasing of protein level. Bring together, the

19

above results indicated that the PRM based method could be a powerful tool to reveal

20

the process for the dynamical assembly of protein complex.

21

22

23


Page 20 of 30

Page 21 of 30 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

4. Conclusion

2

In this study, we presented a pseudo-targeted MS method to identify low

3

abundance phosphorylation in minute amount of sample. The development of this

4

targeted approach is very easy as the same sample and the same LC-system were

5

used for the two phases. No sample fractionation or enrichment was required for the

6

discovery phase which allowed this method to analyze minute amount of sample. We

7

have demonstrated that this method has higher sensitivity to identify phosphorylation

8

sites on endogenous Shc1 protein complexes. PRM is typically used in targeted

9

proteomics to quantify the peptides of interest, while the pseudo-targeted MS method

10

presented in this study allowed PRM to identify new peptides. In this new method,

11

low confident peptides identified from the DDA dataset are used to build a pseudo-

12

targeted library for PRM analysis which enabled the identification of new high

13

confident peptides. Thus this strategy is not limited to analyze protein

14

phosphorylation. It is also applicable to analyze other low abundant peptides in a

15

sample with trace amount.

16

Acknowledgments

17

This work was supported, in part, by funds from the China State Key Basic

18

Research Program Grants (2016YFA0501402), the National Natural Science

19

Foundation of China (21605140, 21235006, 21535008, 81600046). MY is a recipient

20

of the National Science Fund of China for Distinguished Young Scholars (21525524).

21

Competing financial interests: The authors declare no competing financial

22

interest.



1 2 3

Supporting information available Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

4 5

Reference

6

(1) Gavin, A. C.; Bosche, M.; Krause, R.; Grandi, P.; Marzioch, M.; Bauer, A.; Schultz, J.; Rick, J. M.;

7

Michon, A. M.; Cruciat, C. M.; Remor, M.; Hofert, C.; Schelder, M.; Brajenovic, M.; Ruffner, H.; Merino,

8

A.; Klein, K.; Hudak, M.; Dickson, D.; Rudi, T., et al. Nature 2002, 415, 141-147.

9

(2) Guruharsha, K. G.; Rual, J. F.; Zhai, B.; Mintseris, J.; Vaidya, P.; Vaidya, N.; Beekman, C.; Wong, C.;

10

Rhee, D. Y.; Cenaj, O.; McKillip, E.; Shah, S.; Stapleton, M.; Wan, K. H.; Yu, C.; Parsa, B.; Carlson, J. W.;

11

Chen, X.; Kapadia, B.; VijayRaghavan, K., et al. Cell 2011, 147, 690-703.

12

(3) Vazquez, F.; Grossman, S. R.; Takahashi, Y.; Rokas, M. V.; Nakamura, N.; Sellers, W. R. J. Biol. Chem.

13

2001, 276, 48627-48630.

14

(4) van Attikum, H.; Fritsch, O.; Hohn, B.; Gasser, S. M. Cell 2004, 119, 777-788.

15

(5) Lawrence, R. T.; Searle, B. C.; Llovet, A.; Villen, J. Nature methods 2016, 13, 431-434.

16

(6) Zhou, H.; Ye, M.; Dong, J.; Corradini, E.; Cristobal, A.; Heck, A. J.; Zou, H.; Mohammed, S. Nature

17

protocols 2013, 8, 461-480.

18

(7) Kang, T.; Kim, J. H.; Hong, I.; Park, N.; Heinsen, H.; Lee, J.-Y.; Ravid, R.; Ferrer, I.; Yoo, J. S.; Kwon, K.-

19

H.; Park, Y. M. Anal. Bioanal. Chem. 2014, 406, 5433-5446.

20

(8) Roumeliotis, T. I.; Williams, S. P.; Goncalves, E.; Alsinet, C.; Del Castillo Velasco-Herrera, M.; Aben,

21

N.; Ghavidel, F. Z.; Michaut, M.; Schubert, M.; Price, S.; Wright, J. C.; Yu, L.; Yang, M.; Dienstmann, R.;

22

Guinney, J.; Beltrao, P.; Brazma, A.; Pardo, M.; Stegle, O.; Adams, D. J., et al. Cell reports 2017, 20,

23

2201-2214.

24

(9) Bekker-Jensen, D. B.; Kelstrup, C. D.; Batth, T. S.; Larsen, S. C.; Haldrup, C.; Bramsen, J. B.; Sorensen,

25

K. D.; Hoyer, S.; Orntoft, T. F.; Andersen, C. L.; Nielsen, M. L.; Olsen, J. V. Cell systems 2017, 4, 587-599

26

e584.

27

(10) Yang, F.; Shen, Y.; Camp, D. G., 2nd; Smith, R. D. Expert review of proteomics 2012, 9, 129-134.

28

(11) Gillet, L. C.; Navarro, P.; Tate, S.; Roest, H.; Selevsek, N.; Reiter, L.; Bonner, R.; Aebersold, R.

29

Molecular & Cellular Proteomics 2012, 11.

30

(12) Nahnsen, S.; Kohlbacher, O. BMC Bioinformatics 2012, 13.

31

(13) Pan, S.; Aebersold, R.; Chen, R.; Rush, J.; Goodlett, D. R.; McIntosh, M. W.; Zhang, J.; Brentnall, T.

32

A. J. Proteome Res. 2009, 8, 787-797.

33

(14) Elschenbroich, S.; Kislinger, T. Molecular bioSystems 2011, 7, 292-303.

34

(15) de Graaf, E. L.; Kaplon, J.; Mohammed, S.; Vereijken, L. A.; Duarte, D. P.; Redondo Gallego, L.;

35

Heck, A. J.; Peeper, D. S.; Altelaar, A. F. J Proteome Res 2015, 14, 2906-2914.

36

(16) Soste, M.; Hrabakova, R.; Wanka, S.; Melnik, A.; Boersema, P.; Maiolica, A.; Wernas, T.; Tognetti,

37

M.; von Mering, C.; Picotti, P. Nat Methods 2014, 11, 1045-1048.

38

(17) Degryse, S.; de Bock, C. E.; Demeyer, S.; Govaerts, I.; Bornschein, S.; Verbeke, D.; Jacobs, K.; Binos,

39

S.; Skerrett-Byrne, D. A.; Murray, H. C.; Verrills, N. M.; Van Vlierberghe, P.; Cools, J.; Dun, M. D.


Page 22 of 30

Page 23 of 30 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

Leukemia 2017.

2

(18) Gallien, S.; Domon, B. Bioanalysis 2014, 6, 2159-2170.

3

(19) Bourmaud, A.; Gallien, S.; Domon, B. Proteomics 2016, 16, 2146-2159.

4

(20) Sherman, J.; McKay, M. J.; Ashman, K.; Molloy, M. P. Proteomics 2009, 9, 1120-1123.

5

(21) Peterson, A. C.; Russell, J. D.; Bailey, D. J.; Westphall, M. S.; Coon, J. J. Molecular & Cellular

6

Proteomics 2012, 11, 1475-1488.

7

(22) Zheng, Y.; Zhang, C.; Croucher, D. R.; Soliman, M. A.; St-Denis, N.; Pasculescu, A.; Taylor, L.; Tate,

8

S. A.; Hardy, W. R.; Colwill, K.; Dai, A. Y.; Bagshaw, R.; Dennis, J. W.; Gingras, A. C.; Daly, R. J.; Pawson,

9

T. Nature 2013, 499, 166-171.

10

(23) Cox, J.; Mann, M. Nat. Biotechnol. 2008, 26, 1367-1372.

11

(24) MacLean, B.; Tomazela, D. M.; Shulman, N.; Chambers, M.; Finney, G. L.; Frewen, B.; Kern, R.;

12

Tabb, D. L.; Liebler, D. C.; MacCoss, M. J. Bioinformatics 2010, 26, 966-968.

13

(25) Mellacheruvu, D.; Wright, Z.; Couzens, A. L.; Lambert, J. P.; St-Denis, N. A.; Li, T.; Miteva, Y. V.;

14

Hauri, S.; Sardiu, M. E.; Low, T. Y.; Halim, V. A.; Bagshaw, R. D.; Hubner, N. C.; Al-Hakim, A.; Bouchard,

15

A.; Faubert, D.; Fermin, D.; Dunham, W. H.; Goudreault, M.; Lin, Z. Y., et al. Nature methods 2013, 10,

16

730-736.

17

(26) Cox, J.; Hein, M. Y.; Luber, C. A.; Paron, I.; Nagaraj, N.; Mann, M. Molecular & Cellular Proteomics

18

2014, 13, 2513-2526.

19

(27) Tyanova, S.; Temu, T.; Sinitcyn, P.; Carlson, A.; Hein, M. Y.; Geiger, T.; Mann, M.; Cox, J. Nature

20

methods 2016, 13, 731-740.

21

(28) Lesur, A.; Domon, B. Proteomics 2015, 15, 880-890.

22

(29) Liu, H. B.; Sadygov, R. G.; Yates, J. R. Anal. Chem. 2004, 76, 4193-4201.

23

(30) Jaffe, J. D.; Keshishian, H.; Chang, B.; Addona, T. A.; Gillette, M. A.; Carr, S. A. Molecular & Cellular

24

Proteomics 2008, 7, 1952-1962.

25

(31) Hornbeck, P. V.; Kornhauser, J. M.; Tkachev, S.; Zhang, B.; Skrzypek, E.; Murray, B.; Latham, V.;

26 27

Sullivan, M. Nucleic Acids Res. 2012, 40, D261-D270.

28



1

2

Table1.

3

4 5

Page 24 of 30

Table 1. P-sites identified in the Shc1 protein complexes by DDA and PRM

6 EGF stimulation 0min

2min

5min

20min

time DDA

6

42

46

42

PRM

+4

+15

+17

+21

total

10

57

63

63


Page 25 of 30 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

2

Figure legends

3 4

Figure 1. The pseudo-targeted strategy for the sensitive analysis of protein phosphorylation in

5

protein complex. The low confident phosphopeptide identifications from DDA data were used to

6

build the pseudo-targeted library for PRM analysis, which enabled the identification of additional

7

high confident phosphopeptides.

8

Figure 2. The pseudo-targeted strategy improves the spectra quality. (a) Comparison of peptide

9

matching scores in DDA and PRM for the same ions. Scores of 49 peptides were improved after

10

PRM, locating above the identify line. (b) The matching score for the identification of a

11

phosphopeptide, SPFGSPSAEAVSSR, was 6 in DDA, while it was improved to 34 after PRM.

12

Figure 3. Multiple runs of DDA or PRM analysis do not significantly improve the coverage. The

13

numbers of p-sites identified by the three replicate (a) DDA runs and their combination, (c) PRM

14

runs and their combination; The overlap of p-site identifications for the three replicate (b) DDA

15

and (d) PRM runs. The Shc1 complex immunoprecipitated from cells stimulated with EGF for 2min

16

was used as the sample for all the analysis. The p-sites shown were all identified from the Shc1

17

complex. The numbers of p-sites for the combined results were slightly different with those in the

18

Table 1 because the MS data for each run were searched separately here.

19

Figure 4. The PRM based method enabled comprehensive mapping the dynamic change of the

20

Shc-1 complex upon EGF stimulation. The binding partners of Shc1 at rest phase were labeled in

21

blue. The proteins recruited after EGF stimulation was shown in orange. The proteins down-

22

regulated over 50% compared with their peak values were labeled in white. The number in the

23

circle indicates the number of p-sites identified from that protein.

24



1

Figure 1

2

3 4

Page 26 of 30

5


Page 27 of 30

1

Figure 2

2

a

Phosphopeptide Non-phosphopeptide Unmatched ion

120

100

80

PRM score

1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


60

40

20

0 0

20

40

60

80

100

120

DDA score

b

Score: 6

Score: 34

3 4

5



1

Page 28 of 30

Figure 3

2

a

b

50 40

43

35

33

36

30

20 10 0

2 min DDA1 2 min DDA2 2 min DDA3 3 DDA runs Combined

c

d

40 30 20 10

12

12

14

18

20

18

17

19

0

2 min PRM1 2 min PRM2 2 min PRM3 3 PRM runs Combined

3 4

Newly identified P-sites in PRM

P-sites identified in DDA

5


Page 29 of 30 1 2 3 4 5 6 7 8 9 10 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 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


1

Figure 4.

2

3

4



1

TOC graphic

2 3

4

Page 30 of 30


A pseudo-targeted MS method for the sensitive analysis of protein

Recommend Documents