MASS SPECTROMETRY ANALYSIS OF PROTEINS AND PROTEIN
PHOSPHORYLATION USING FUNCTIONALIZED POLYMER MATERIALS
By
Weihan Wang

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Chemistry
2011

ABSTRACT

MASS SPECTROMETRY ANALYSIS OF PROTEINS AND PROTEIN
PHOSPHORYLATION USING FUNCTIONALIZED POLYMER MATERIALS
By
Weihan Wang
Posttranslational protein phosphorylation helps control many cellular processes, so
identification of stimuli- induced phosphorylation events is vital to understanding cell regulation.
Characterization of phosphorylated proteins by mass spectrometry (MS) typically employs small
phosphopeptides (<3000 Da) generated by complete protein digestion.

However, efficient

phosphopeptide ionization for MS and tandem MS (MS/MS) analysis requires enrichment of
phosphopeptides from complex protein digests. Alternatively, digestion of proteins to larger
peptides (middle-down proteolysis) can improve protein sequence coverage in MS analysis and
reveal a wider range of phosphorylation sites than complete digestion. Initially, this dissertation
reviews recent studies on phosphopeptide enrichment and middle-down proteolysis. Based on
these studies, the research aims to (1) improve enrichment of small phosphopeptides for MS
analysis of target phosphoproteins, and (2) generate large tryptic peptides for middle-down
proteomics.
In studies on enrichment, polymer brushes derivatized with nitrilotriacetate-Fe(III)
complexes (NTA-Fe) or oxotitanium (oxoTi) capture phosphopeptides from protein digests.
Specifically, polymer microspots derivatized with NTA-Fe and surrounded by hydrophobic
poly(dimethylsiloxane) facilitate pre-concentration and enrichment of dilute phosphopeptides
from only 1 L of digest. Matrix-assisted laser desorption/ionization (MALDI)-MS directly on
the 250-m spots leads to a sub-fmol detection limit for a -casein phosphopeptide. Compared to

films derivatized with NTA-Fe, polymer-oxoTi brushes improve the capture and detection of
mono-phosphopeptides in mixtures of model protein digests. Moreover, MALDI collisioninduced dissociation (CID)-MS/MS on polymer-oxoTi identified unknown phosphopeptides
from several p65-associated phosphoproteins. These phosphoproteins immunoprecipitated
together with p65 from the nuclear extracts of human acute monocytic leukemia cells.
Bifunctionalization of polymer brushes with both oxoTi and NTA-Fe groups leads to
simultaneous low-bias enrichment of mono- and multi-phosphorylated peptides from moderately
complex model protein digests. For larger samples (tens of Ls), stacking two membranes
modified with polymer-oxoTi and polymer-NTA brushes, respectively, in an HPLC fitting
enhances phosphopeptide enrichment. The two membra nes sequentially capture mono- and
multi-phosphorylated peptides from 3 phosphoprotein digests (~1 g total mass) in the presence
of ~70 g of non-phosphoprotein digest.
To control digestion for middle-down MS analysis of proteins, pumping protein solutions
through nylon membranes containing immobilized trypsin affords digestion times ranging from a
few milliseconds to seconds. In modified membranes with 0.45 m pores, a 50- ms digestion of
apomyoglobin leads to peptides with MWs primarily between 2.5 and 6 kDa. Five of these
peptides cover the entire protein sequence.

A 10- ms -casein residence time in modified

membranes with 1.2-m pores yields peptides with masses primarily between 3 and 20 kDa.
Two large peptides constitute the entire protein sequence, and one of them covers all the five
phosphorylation sites. These simple enrichment methods and new digestion techniques for
middle-down proteomics should facilitate future studies of protein phosphorylation.

I dedicate this dissertation to my parents, Jiaxiang Wang and Baozhu Wang, and my wife, Jia
Liu, for their love and support.

iv

Acknowledgement
I would like to express my sincere thanks to my advisor, Prof. Merlin Bruening, for his
guidance and support.

I am lucky enough to join Prof. Brue ning‟s group and receive

comprehensive trainings on instrumental analysis and material synthesis. His patience and
encouragement helped me overcome difficulties at the early stage of my Ph.D study; and his trust
gave me confidence and freedom in exploring new areas in the last two years. I will never forget
those exciting moments when we got interesting results.
I am grateful to my second reader, Prof. Gavin Reid, for his strict training on mass
spectrometry and helpful discussions on several research projects. Some of my manuscripts
could not have been published smoothly without his commenting and revisions.
I also thank Profs. Jetze Tepe, Dan Jones, Gregory Baker, Gary Blanchard and Scott
McLuckey for their kind help and suggestions.

Thanks to Prof. Jetze Tepe for providing

precious samples to test my brush weapons! Thanks to Prof. McLuckey from Purdue University
for his collaboration in middle-down protein analysis.
I would like to acknowledge the collaborations and discussions from many researchers in
and outside Bruening group. Many thanks to Yujing Tan for our long time collaboration and
friendship built from those unforgettable numbers “2061, 2089, 3122…”. I will always
remember Somnath Bhattacharjee and Yong Guan for their insightful discussio ns on organic
synthesis and countless “gossip” of the organic world, and Fei Xu for her dedication on the
project of membrane digestion. Because of the collaborations with Amanda Palumbo, as well as
with Ken Chanthamontri from Purdue University, I was able to extend my skills to address
important biological problems. I also would like to thank my friend and former labmate, Mingxi

v

Zhang, who shared his insights toward life and chemistry with me every week via internet. The
bifunctional polymer brushes actually began growing from our MSN dialog window.
Most importantly, none of the work would have been possible without the love from my
family. I owe so much to my parents, Jiaxiang Wang and Baozhu Wang, who brought their only
child up, but could only hear his voice on the phone once a week in the last five years. Thanks to
my wife, Jia Liu, for her constant support and understanding. Each piece of my work is soaked
with her love and encouragement!

vi

TABLE OF CONTENTS
LIST OF TABLES ...........................................................................................................................x
LIST OF FIGURES ........................................................................................................................ xi
LIST OF SCHEMES..................................................................................................................... xvi
LIST OF ABBREVIATIONS ...................................................................................................... xvii
Chapter 1. Introduction ................................................................................................................... 1
1.1 Protein phosphorylation ........................................................................................................ 1
1.2 Overview of strategies for MS-based protein identification and characterization of
phosphorylation ........................................................................................................................... 4
1.3 Phosphopeptide enrichment strategies .................................................................................. 9
1.3.1 History of enrichment methods for phosphoprotein analysis ......................................... 9
1.3.2 Improvement of phosphopeptide enrichment with conventional IMAC-Fe3+ beads and
TiO2 particles ......................................................................................................................... 10
1.3.3 Novel phosphopeptide enrichment materials................................................................ 16
1.3.4 Ion exchange for phosphopeptide fractionation ........................................................... 21
1.3.5 Phosphopeptide enrichment directly on modified MALDI plates................................ 25
1.4 Middle-down protein analysis by MS ................................................................................. 28
1.4.1 Advantages of middle-down protein analysis............................................................... 28
1.4.2 Middle-down proteolysis methods ............................................................................... 30
1.5 Tandem MS and data analysis for peptide sequencing, protein identification and
phosphorylation localization ..................................................................................................... 38
1.6 MS-based quantification of proteins and protein phosphorylation ..................................... 40
1.7 Research aims...................................................................................................................... 43
REFERENCES.......................................................................................................................... 46
Chapter 2. Phosphopeptide Enrichment on Functionalized Polymer Microspots for MALDI-MS
Analysis......................................................................................................................................... 56
2.1 Introduction ......................................................................................................................... 56
2.2 Experimental ....................................................................................................................... 59
2.2.1 Chemicals and materials ............................................................................................... 59
2.2.2 Fabrication of functional polymer brushes on Si wafers .............................................. 60
2.2.3 Polymer brush characterization .................................................................................... 61
2.2.4 Sample preparation and MS analysis............................................................................ 61
2.3 Results and discussion......................................................................................................... 63
2.3.1 Fabrication and characterization of PHEMA-NTA-Fe(III) microspots surrounded by a
hydrophobic film ................................................................................................................... 63
2.3.2 MALDI-MS analysis of phosphoprotein digests enriched on PHEMA-NTA-Fe(III)
microspots .............................................................................................................................. 68
2.3.3 Enrichment of phosphopeptides from mixtures of protein digests ............................... 72

vii

2.3.4 Mechanism of the enrichment of phosphopeptides on PHEMA-NTA-Fe(III)
microspots .............................................................................................................................. 74
2.4 Conclusions ......................................................................................................................... 76
REFERENCES.......................................................................................................................... 79
Chapter 3. Identification of p65-associated Phosphoproteins by Mass Spectrometry after Onplate Phosphopeptide Enrichment Using Polymer-oxotitanium Films......................................... 82
3.1 Introduction ......................................................................................................................... 82
3.2 Experimental ....................................................................................................................... 85
3.2.1 Chemicals and materials ............................................................................................... 85
3.2.2 Preparation of polymer-oxotitanium Films on Au Wafers (Au-P-oxoTi) .................... 87
3.2.3 Characterization of Au-P-oxoTi films .......................................................................... 87
3.2.4 Immunoprecipitation (IP) of p65 and associated proteins from THP-1 nuclear extracts
............................................................................................................................................... 90
3.2.5 Preparation of tryptic digests ........................................................................................ 90
3.2.6 Phosphopeptide enrichment .......................................................................................... 91
3.2.7 MALDI-MS and CID-MS/MS analysis........................................................................ 94
3.3 Results and discussion......................................................................................................... 95
3.3.1 Preparation and characterization of polymer-oxotitanium (P-oxoTi) hybrid materials 95
3.3.2 Enrichment of model phosphopeptides on Au-P-oxoTi plates ..................................... 98
3.3.3 Comparison with other enrichment methods .............................................................. 109
3.3.4 Analysis of immunoprecipitated p65-associated phosphoproteins............................. 112
3.3.5 Semi-quantitative analysis of p65-associated phosphoproteins upon TNF- treatment
............................................................................................................................................. 122
3.3.6 Relevance of the identified proteins co-immunoprecipitated with p65 ...................... 123
3.4 Conclusion......................................................................................................................... 125
REFERENCES........................................................................................................................ 128
Chapter 4. Low-bias Enrichment of Mono- and Multi-phosphorylated Peptides Using
Bifunctional Polymer Brushes .................................................................................................... 133
4.1 Introduction ....................................................................................................................... 133
4.2 Experimental ..................................................................................................................... 135
4.2.1 Chemicals and materials ............................................................................................. 135
4.2.2 Synthesis of bifunctionalized polymer brushes on Au wafers.................................... 136
4.2.3 Synthesis of functionalized polymer brushes in nylon membranes ............................ 140
4.2.4 Characterization of polymer brushes .......................................................................... 142
4.2.5 Phosphopeptide enrichment ........................................................................................ 143
4.2.6 MS analysis of phosphopeptides and data analysis .................................................... 143
4.3 Results and discussion....................................................................................................... 144
4.3.1 Polymer synthesis on Au wafers and in nylon membranes ........................................ 144
4.3.2 On-plate phosphopeptide enrichment using oxoTi-PGMA-Fe brushes ..................... 149
4.3.3 Enrichment of phosphopeptides using functional membranes ................................... 161
REFERENCES........................................................................................................................ 168
Chapter 5. Controlled Proteolysis with Trypsin-modified Membranes to Obtain Large Peptides
for Mass Spectrometry ................................................................................................................ 170
viii

5.1 Introduction ....................................................................................................................... 170
5.2 Experimental ..................................................................................................................... 180
5.2.1 Chemicals and materials ............................................................................................. 180
5.2.2 Trypsin immobilization in nylon membranes ............................................................. 181
5.2.3 Protein solutions for tryptic digestion......................................................................... 181
5.2.4 Protein digestion ......................................................................................................... 183
5.2.5 Gel electrophoresis ..................................................................................................... 184
5.2.6 ESI-MS analysis and peptide identification ............................................................... 184
5.3 Results and discussion....................................................................................................... 186
5.3.1 Short residence times in trypsin- modified membranes yield large tryptic peptides... 186
5.3.2 Larger membrane pores allow briefer tryptic digestion ............................................. 204
5.3.3 Influence of denaturation on tryptic digestion of lysozyme ....................................... 207
5.3.4 Protein identification by peptide mass fingerprinting (PMF) ..................................... 215
5.4 Conclusion......................................................................................................................... 220
REFERENCES........................................................................................................................ 222
Chapter 6. Summary and Future Directions ............................................................................... 224
6.1 Research summary ............................................................................................................ 224
6.2 Future work ....................................................................................................................... 226
6.2.1 Improving on-plate phosphopeptide enrichment by decreasing sample complexity.. 226
6.2.2 Large-scale phosphopeptide enrichment using tandem membranes .......................... 227
6.2.3 Quantitative evaluation of phosphopeptide recovery ................................................. 227
6.2.4 Tuning flowrate, membrane pore size, trypsin local “concentration”, and substrate
protein concentration to control membrane-based digestion ............................................... 228
6.2.5 Immobilization of pepsin in nylon membranes for protein proteolysis in acidic
conditions............................................................................................................................. 230
6.3 Potential impact ................................................................................................................. 231
REFERENCES........................................................................................................................ 234

ix

LIST OF TABLES
Table 2.1. Phosphopeptides that result from tryptic digestion of -casein and ovalbumin. The
amino acids in parentheses represent cleavage sites and are not part of the peptide sequence. ... 69
Table 3.1. N-Propionylated standard phosphopeptides employed for determination of enrichment
recoveries. ..................................................................................................................................... 86
Table 3.2. Phosphopeptides observed in the positive- ion mode MALDI-MS spectra of tryptic casein, -casein, and ovalbumin digests. The phosphopeptide sequence (if identified), number of
+

phosphoryl groups, and monoisotopic m/z of [M+H] are listed. .............................................. 101
Table 3.3. Peptides identified from the Au-P-oxoTi enriched digests of the nuclear extracts from
TNF-α-treated THP-1 cells. ........................................................................................................ 114
Table 4.1. The amount of Fe and Ti in functional polymer brushes. .......................................... 147
Table 4.2. Phosphopeptides detected in positive- ion mode MALDI-MS spectra. The list contains
the phosphopeptide sequence (if identified), the range of amino acids, the number of phosphoryl
+

groups, and the monoisotopic m/z of [M+H] ............................................................................ 153
Table 5.1. Apomyoglobin (average MW 16951 Da) tryptic peptides detected by ESI-MS. ...... 195
Table 5.2. -casein (average MW 23983 Da) tryptic peptides detected by ESI-MS.................. 204
Table 5.3. Tryptic peptides of lysozyme after Cys carbamidomethylation (average MW 14770
Da after carbamidomethylation) detected by ESI-MS. ............................................................... 218
Table 5.4. Protein identification based on MASCOT PMF searching using ions genera ted by
post-ion/ion reaction ESI-MS. .................................................................................................... 219

x

LIST OF FIGURES
Figure 1.1. Protein phosphorylation at Ser/Thr (top) and Tyr (bottom) residues. .......................... 2
Figure 1.2. “Bottom- up” and “top-down” strategies for MS-based protein identification and
characterization. .............................................................................................................................. 5
Figure 1.3. Three typical modes of phosphopeptide enrichment .................................................. 11
Figure 1.4. Schematic overview of a vented column system with a triple stage precolumn. ....... 14
4+

Figure 1.5. Synthesis of –PGMA-RPO 3-Zr brushes in a capillary column using ATRP of
glycidyl methacrylate and subsequent derivatization ................................................................... 19
Figure 1.6. SCX-IMAC strategy for large-scale identification and characterization of
phosphorylation sites in mouse liver proteins............................................................................... 23
Figure 1.7. F-doped SnO 2 film for phosphopeptide enrichment .................................................. 27
Figure 1.8. Mechanisms of the acid cleavage at Asp-X (path A) and X-Asp (path B) bonds (X
represents any amino acid residue). .............................................................................................. 32
Figure 1.9. Molecular structure of the pea Lhcb1 antenna protein (a) and liquid chromatographyESI MS of the Lhcb antenna proteins after brief typtic digestion (b)........................................... 36
Figure 2.1. Si wafer coated with residual PDMS (light gray) and microspots of PHEMA-NTAFe(III) brushes (dark gray).. .......................................................................................................... 64
Figure 2.2. Contact angles measured 0, 2, or 5 min after placing water droplets (0.5 L) on
different surfaces........................................................................................................................... 65
Figure 2.3. SEM image and EDS mapping of a microspot of PHEMA-NTA-Fe(III) brushes
surrounded by a residual PDMS layer on Si. ................................................................................ 66
Figure 2.4. Transmission FTIR absorbance spectrum of a double-side polished Si wafer coated
with PHEMA brushes on both sides before and after derivatizations .......................................... 68
Figure 2.5. MALDI mass spectra of β-casein digests with and without phosphopeptide
enrichment..................................................................................................................................... 70
Figure 2.6. MALDI mass spectra of ovalbumin digests after enrichment on PHEMA-NTA-Fe(III)
brushes. ......................................................................................................................................... 71
Figure 2.7. MALDI mass spectra of mixtures of protein digests with and without phosphopeptide
enrichment..................................................................................................................................... 73
xi

Figure 2.8. Charge-coupled device images of sample enriched on (a) a 2-mm spot and (b) a
microspot....................................................................................................................................... 74
Figure 3.1. (a) Schematic drawing of the formation of the P-oxoTi hybrid structure, and the
binding and elution of phosphopeptides in this material. (b) Procedures for on-plate
phosphopeptide enrichment using Au-P-oxoTi. ........................................................................... 88
Figure 3.2. Reflectance FT-IR spectra of a PHEMA film before (a) and after reaction with
succinic anhydride (b), and subsequent immersion in titanium isopropoxide (c). ....................... 96
Figure 3.3. MALDI mass spectra of model protein digests without phosphopeptide enrichment (a)
and after enrichment on Au-P-oxoTi MALDI plates (b-d)........................................................... 99
Figure 3.4. MALDI mass spectra of model protein digests that were enriched on Au-P-oxoTi
wafers. ......................................................................................................................................... 104
Figure 3.5. MALDI mass spectra of -casein digests that were enriched on Au-P-oxoTi wafers.
..................................................................................................................................................... 106
Figure 3.6. Recoveries of CH3CH2CO-phosphopeptides from solutions with different TFA
concentrations and various amounts of BSA digest, urea, and ammonium salts ....................... 108
Figure 3.7. MALDI mass spectra of protein digest mixtures after on-plate IMAC enrichment
using Au-P-NTA-Fe (a, b) and in-solution enrichment on TiO 2 particles (c, d). ....................... 110
Figure 3.8. p65 Western blotting (using C-22B4 rabbit polyclonal antibody) of 25 µ of THP-1
g
cell nuclear extract (NE) before (right lane) and after (left lane) IP. .......................................... 113
Figure 3.9. MALDI mass spectrum of a Au-P-oxoTi-enriched trypsin-digest of p65immunoprecipitated nuclear extract from TNF-α-treated THP-1 cells....................................... 113
Figure 3.10. CID-MS/MS product ion spectra of m/z 2061 (a), 2109 (b), 2125 (c), 2166 (d), 1629
(e), 1677 (f), 1693 (g), 1734 (h), and 1367 (i) (from the MALDI mass spectrum shown in Figure
3.9). ............................................................................................................................................. 116
Figure 3.11. MALDI mass spectra of the samples resulting from (a) the trypsin-digested,
ZipTipC18-desalted, p65- immunoprecipitated nuclear extracts from TNF-α-treated THP-1 cells,
and (b) the Au-P-oxoTi enriched trypsin digest of the p65-IP beads exposed to a PBS blank. . 121
Figure 3.12. MALDI mass spectra of the (a) TNF-α-treated THP-1 cell and (b) untreated THP-1
cell samples resulting from nuclear extraction, p65 IP, trypsin digestion, phosphopeptide
enrichment, and addition of 100 amol of an internal standard phosphopeptide ......................... 123
Figure 4.1. Scheme for orthogonal modifications of PGMA brushes to obtain the bifunctional
oxoTi-PGMA-Fe. ........................................................................................................................ 137
xii

Figure 4.2. Structures of control polymer brushes for phosphopeptide enrichment. .................. 139
Figure 4.3. Reflectance FTIR spectra of a PGMA film on a Au-coated wafer before (a) and after
the following sequential steps (b- f)............................................................................................. 145
Figure 4.4. ATR-IR spectra of nylon membranes before (a) and after (b) modification with
PMES, (c) activation with NHS/1-ethyl-3-(3-dimethylaminopropyl)carbodiimide and (d) reaction
with aminobutyl NTA (pH 10). .................................................................................................. 149
Figure 4.5. MALDI mass spectra after on-plate phosphopeptide enrichment from a 1-L 0.1%
TFA solution containing 200 fmol each of three phosphoprotein digests using 60 nm-thick (prior
to addition of affinity groups) polymer films ............................................................................. 151
Figure 4.6. MALDI mass spectra after on-plate phosphopeptide enrichment from a 1-L 0.1%
TFA solution containing 200 fmol each of three phosphoprotein digests using 60 nm-thick (prior
to addition of affinity groups) polymer films ............................................................................. 152
Figure 4.7. MALDI mass spectra after on-plate phosphopeptide enrichment from 1-L of 0.1%
TFA containing 200 fmol each of three phosphoprotein digests using 120 nm-thick (prior to
addition of affinity groups) polymer films.................................................................................. 155
Figure 4.8. MALDI mass spectra after on-plate phosphopeptide enrichment from 1-L of 1%
TFA containing 200 fmol each of three phosphoprotein digests using 120 nm-thick (prior to
addition of affinity groups) polymer films.................................................................................. 157
Figure 4.9. MALDI mass spectra after on-plate phosphopeptide enrichment from 1 L of 1%
TFA containing 200 fmol each of three phosphoprote in digests using 220 nm-thick (prior to
addition of affinity groups) polymer films.................................................................................. 158
Figure 4.10. MALDI mass spectra after on-plate phosphopeptide enrichment from a 1 L of 1%
TFA containing 50 fmol each of three phosphopro tein digests using 220 nm-thick (prior to
addition of affinity groups) polymer films.................................................................................. 159
Figure 4.11. MALDI mass spectrum after phosphopeptide enrichment from 1 L of 1% TFA
containing 200 fmol each of three phosphoprotein d igests and 2 pmol of nonphosphoprotein
digest using 220 nm-thick (prior to addition of affinity groups) oxoTi-PGMA-Fe. .................. 160
Figure 4.12. MALDI mass spectra of phosphopeptides enriched from 200 L of 0.1% TFA, 50%
ACN containing three phosphoprotein digests ........................................................................... 162
Figure 4.13. (a-c) MALDI mass spectra after phosphopeptide enrichment from 200-L of 1%
TFA, 50% ACN containing 10 pmol each of three phosphoprotein digests (-casein, -casein
and ovalbumin), and 1 nmol of BSA digest. Figure (d) is a mixture of 1 L of eluate from (a) and
1 L of eluate from (b)................................................................................................................ 164
Figure 5.1. Scheme of trypsin-catalyzed proteolysis. ................................................................. 172
xiii

Figure 5.2. Enzyme-catalyzed protein digestion pathways: (a) one-by-one scheme and (b) zipper
scheme......................................................................................................................................... 175
Figure 5.3. Scheme for (a) protein digestion using a nylon membrane containing immobilized
trypsin and (b) the assembled setup that performs protein digestion.......................................... 179
Figure 5.4. SDS-PAGE of intact -casein (lane 1) and -casein digests (lanes 2 and 3) .......... 187
Figure 5.5. ESI-Q-TOF mass spectrum of apomyoglobin digested for 2 h in solution. ............. 188
Figure 5.6. ESI-Q-TOF mass spectrum of apomyoglobin digested for 30 min in solution. ....... 189
Figure 5.7. Comparison of typical experimental (bottom) and simulated (top) spectra of
identified apomyoglobin peptides or protein. ............................................................................. 190
Figure 5.8. ESI-Q-TOF mass spectrum of apomyoglobin digest generated in a trypsin- modified
membrane (50- ms residence time, 0.45 m pore size). .............................................................. 193
Figure 5.9. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of
apomyoglobin digested in a trypsin- modified membrane (50 ms residence time, 0.45 m pore
size). ............................................................................................................................................ 194
Figure 5.10. ESI-Q-TOF mass spectrum of a -casein tryptic digest generated by 16- h insolution digestion. ....................................................................................................................... 197
Figure 5.11. Comparison of typical experimental (bottom) and simulated (top) spectra of
identified -casein peptide ions.. ................................................................................................ 198
Figure 5.12. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of -casein
digested in solution for 5 min. .................................................................................................... 200
Figure 5.13. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of -casein
digested in solution for 1 min. .................................................................................................... 201
Figure 5.14. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of -casein
digested in solution for 20 s. ....................................................................................................... 202
Figure 5.15. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of -casein
digested in a trypsin- membrane membrane (2 ms residence time, 0.45 m pore size) .............. 203
Figure 5.16. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of -casein
digested in a trypsin- membrane membrane (10 ms residence time, 1.2 m pore size). ............. 205

xiv

Figure 5.17. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of
apomyoblobin digested in a trypsin- membrane membrane (50 ms residence time, 1.2 m pore
size). ............................................................................................................................................ 206
Figure 5.18. ESI-Q-TOF mass spectrum of lysozyme before (top view) and after (bottom view)
carbamidomethylation of Cys residues. ...................................................................................... 208
Figure 5.19. ESI-Q-TOF mass spectrum of alkylated lysozyme (without dialysis) digested for 16
h in solution................................................................................................................................. 209
Figure 5.20. Comparison of typical experimental (bottom) and simulated (top) spectra of
identified lysozyme peptides (or protein). .................................................................................. 210
Figure 5.21. ESI-Q-TOF mass spectrum of alkylated lysozyme (without dialysis) digested for 20
s in solution. ................................................................................................................................ 211
Figure 5.22. ESI-Q-TOF mass spectrum of alkylated lysozyme (without dialysis) digested in a
trypsin- modified membrane (2 ms residence time, 0.45 m pore size)...................................... 212
Figure 5.23. Comparison of typical experimental (bottom) and simulated (top) spectra. .......... 213
Figure 5.24. ESI-Q-TOF mass spectrum of alkylated, dialyzed lysozyme after digestion in a
trypsin- modified membrane (0.32-s residence time, 0.45 m pore size). .................................. 214
Figure 5.25. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of alkylated,
dialyzed lysozyme digested in a trypsin- membrane membrane (0.32-s residence time, 0.45 m
pore size). .................................................................................................................................... 216
Figure 5.26. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of alkylated,
dialyzed lysozyme digested in a trypsin- membrane membrane (1.3 s residence time, 0.45 m
pore size). .................................................................................................................................... 217

xv

LIST OF SCHEMES
Scheme 2.1. Fabrication of micro-arrays of PHEMA-NTA-Fe(III) brushes surrounded by
hydrophobic poly(dimethylsiloxane) (PDMS), and on-plate enrichment of phosphopeptides prior
to MALDI-MS analysis ................................................................................................................ 58

xvi

LIST OF ABBREVIATIONS
PTM

Posttranslational modification

ATP

Adenosine triphosphate

Ser

Serine

Thr

Threonine

Tyr

Tyrosine

Lys

Lysine

Arg

Arginine

Glu

Glutamic acid

Asp

Aspartic acid

Asn

Asparagine

His

Histidine

Trp

Tryptophan

Met

Methionine

Cys

Cysteine

TNF

Tumor necrosis factor

NF-κB

Nuclear factor kappa-light-chain-enhancer of activated B cells

THP-1

Human acute monocytic leukemia cell

BSA

Bovine serum albumin

MW

Molecular weight

PMF

Peptide mass fingerprinting

MS

Mass spectrometry

MALDI

Matrix-assisted laser desorption/ionization
xvii

ESI

Electrospray ionization

CID

Collision-induced dissociation

ETD

Electron-transfer dissociation

TOF

Time-of-flight

HPLC

High performance liquid chromatography

FTIR

Fourier transform infrared

SEM

Scanning electron microscopy

ICP-OES

Inductively coupled plasma-optical emission spectrometry

%SCT

100% sequence coverage peak intensity threshold

IMAC

Immobilized metal ion affinity chromatography

SCX

Strong cation exchange

SAX

Strong anion exchange

IP

Immunoprecipitation

ACN

Acetonitrile

DMF

Dimethylformamide

THF

Tetrahydrofuran

TFA

Trifluoroacetic acid

FA

Formic acid

ATRP

Atom-transfer radical polymerization

PDMS

Poly(dimethylsiloxane)

PHEMA

Poly(2-hydroxyethyl methacrylate)

PGMA

Poly(glycidyl methacrylate)

PMES

Poly(2-(methacryloyloxy)ethyl succinate)

xviii

PSS

Poly(sodium 4-styrenesulfonate)

oxoTi

oxotitanium

EDTA

Ethylenediaminetetraacetic acid

SA

Succinic anhydride

DMAP

4-Dimethylaminopyridine

EDC

N-(3-Dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride

NHS

N-hydroxysuccinimide

NTA

Nitrilotriacetate

DHB

2,5-Dihydroxybenzoic acid

DTT

Dithiothreitol

IAA

Iodoacetamide

xix

Chapter 1. Introduction
This dissertation describes research on polymer-based techniques for phosphopeptide
enrichment and size-controlled proteolysis prior to mass spectrometry (MS) analysis. To put this
work in perspective, this introduction first describes protein phosphorylation and gives an
overview of current strategies for MS-based

identification and

characterization of

phosphorylation. This chapter then reviews and discusses in detail progress in phosphopeptide
enrichment techniques and “middle-down” MS analysis of proteins, which both relate closely to
the research in this dissertation. Subsequently, the chapter briefly introduces tandem MS and
quantitative MS analysis, two rapidly growing areas that promise qualitative and quantitative
protein characterization, respectively.

The final section of this introduction presents a brief

outline of the dissertation.

1.1 Protein phosphorylation
Posttranslational modification (PTM) of proteins by phosphorylation is an ubiquitous,
highly evolved regulatory mechanism for creating diversity in both eukaryotes and prokaryotes.
Phosphorylation of specific proteins provides an important control mechanism for cellular
processes such as metabolic pathways, kinase cascade activation, gene transcription, membrane
transport and motor mechanisms.[1-3]
Protein phosphorylation occurs most commonly at Ser, Thr and Tyr residues, which all
have a side chain with a terminal –OH group that can be phosphorylated via the transfer of the PO3

2-

from adenosine triphosphate (ATP) (Figure 1.1).[1] Protein kinases catalyze this process,

and around 30% of cellular proteins are phosphorylated on at least one residue at a certain time
1

2-

Figure 1.1. Protein phosphorylation at Ser/Thr (top) and Tyr (bottom) residues. The –OH groups of Ser and Thr attack -PO3 from
2-

ATP (catalyzed by Ser/Thr specific kinases) to form pSer and pThr. A Tyr specific kinase transfers the -PO3 to Tyr to form pTyr.
(Redrawn from a figure in reference [1]).

2

during their life cycle.[4,5] Although for a wide range of species the abundances of Ser, Thr and
Tyr in proteins are around 8.5%, 5.7% and 3.0% (2.8: 1.9: 1), respectively,[6] recent global mass
spectrometry (MS)-based studies of human protein phosphorylation suggest that the distribution
of phosphorylated Ser, Thr and Tyr (pSer, pThr and pTyr) is ~21: 4.4: 1.[7,8] The differences in
the ratios between the phosphorylated residues and their non-phosphorylated forms stem from
protein kinases that exhibit dramatically different specificities toward Ser, Thr and Tyr. Even
though protein kinases typically share a common fold in their structure, they differ greatly in the
charge and hydrophobicity of their surface residues.[9] The two main groups of protein kinases
show either Ser/Thr (~80% of kinases) or Tyr specificity,[10] with the latter group having a
deeper catalytic cleft that distinguishes the specificities of the two classes.[11,12]
The number of sites (Ser/Thr or Tyr) that a kinase can phosphorylate on different
substrate proteins ranges from several to hundreds.[9] Catalysis of phosphorylation normally
occurs after the kinase recognizes the consensus residues (amino acids situated immediately Nterminally and C-terminally to the phosphorylation site) of the substrate protein via charge,
hydrogen bonding or hydrophobic interactions. For example, the consensus sequence of the
substrate proteins of PKA (protein kinase A or cAMP-dependent protein kinase) is R-R-X-S/T ( is a hydrophobic residue and X is any amino acid residue). When PKA comes close to its
substrate protein, two Glu residues near the catalytic region of PKA recognize the two Arg
residues of the consensus sequence via electrostatic interactions, and the hydrophobic pocket of
PKA favors binding of the hydrophobic residue ().[5]
Phosphorylation controls protein function in several ways. For example, the introduction
2-

of –OPO 3

(pKa2 ~6) to Ser14 of glycogen phosphorylase causes intra-subunit contact of the

phosphate group with Arg69 and Arg43 via charge pairing, and consequently activates the
3

catalytic function of this protein in cleaving glycogen to glucose-1-phosphate.[2,13] In addition
to local conformational changes, many independently folding protein domains show affinity for
short phosphorylated peptide sequences of certain other proteins.[14-16] These recognition
domains are key elements in signal transduction in cells.
In signaling pathways, proteins play different roles in sequential steps during which they
can be phosphorylated by specific kinases. Some of the kinases act in cascades: each kinase
phosphorylates hundreds to thousands of proteins at each step, giving rise to dramatic signal
amplifications.[17] Along with hundreds of protein kinases that work in eukaryotic cells to turn
“on” signaling, there are also many protein phosphatases that turn “off” signaling pathways. The
phosphatases either activate water in a bimetallic center for hydrolytic attack on pSer or pThr, or
covalently release the phosphoryl group from pTyr using a nucleophilic thiolate side chain.[18]
The spatial and temporal regulation of both protein kinases and phosphatases ensure an
appropriate balance of protein phosphorylation in signaling pathways.[19] Due to the crucial
roles of phosphorylated proteins in cells, a comprehensive analysis of protein phosphorylation is
necessary for understanding biological functions at a molecular level.

1.2 Overview of strategies for MS-based protein identification and
characterization of phosphorylation
A comprehensive phosphoprotein analysis involves identification of phosphoproteins,
localization of phosphorylation sites and quantitation of phosphorylation.[21] Current MS-based
protein analysis mainly relies on digestion of proteins into small peptides (“bottom-up” approach,
Figure 1.2) using proteases (also called proteolytic enzymes) such as trypsin. Prior to MS, the

4

Figure 1.2. “Bottom- up” and “top-down” strategies for MS-based protein identification and
characterization (redrawn according to a figure in reference [20]).
bottom- up approach typically includes multiple separation strategies, e.g. reversed phase and
size-exclusion chromatography, isoelectric focusing, affinity purification, and ion exchange
chromatography at the protein or peptide level.[20,22] For a sample with several proteins of
specific interest, the “sort-then-break”[20] workflow (Figure 1.2, path A1) is common. In this
approach, proteins are fractionated prior to proteolytic digestion of each fraction, and the
resulting peptides are either directly analyzed for protein identification via “peptide mass
fingerprinting” (PMF),[23] or further separated before MS/MS for protein identification and/or
PTM characterization.[24,25] In recent years, a “break-then-sort” approach, sometimes known as
“shotgun proteomics” (Figure 1.2, path A2), has gained popularity.[20,26] In this strategy,
5

protein digestion occurs without any prefractionation, and the resulting peptides are separated
using multiple separation modes (normally liquid chromatography, LC, is the final step) prior to
MS/MS for peptide sequencing with the assistance of database searching. This approach is
attractive for large-scale and high-throughput analysis of complex samples, especially for PTM
characterization on the proteome-wide scale.[27]
Despite its simple concept, shotgun proteomics requires high MS sensitivity and efficient
peptide separation because of the greatly increased sample complexity (e.g. digested whole cell
lysate) compared to the “sort-then-break” approach. Because proteins are expressed at a wide
range of concentrations, signals for low- level proteins are often undetectable after the conversion
of intact proteins into a complex mixture of peptides.[20,28] Moreover, peptides with extreme
hydrophilicity, hydrophobicity, isoelectric point (pI) values or unexpected modifications (either
PTM or chemical modification during sample preparation) are also easily lost during peptide
separation/enrichment and MS analysis.[27,28] Nevertheless, in the case of global analysis of
phosphoproteins from whole cell lysates or subcellular fractions, most studies employ a “breakthen-sort” approach to identify hundreds to thousands of phosphopeptides that come from several
tens to

hundreds of phosphoproteins.[29] For

targeted

characterization of protein

phosphorylation, recombinant proteins or proteins isolated using antibodies are also often
digested prior to MS analysis of the phosphopeptides.[30,31]
Due to the low level of phosphopeptides relative to non-phosphopeptides, directly
subjecting a protein digest to LC-electrospray ionization (ESI) MS/MS results in the loss of most
phosphopeptide signals, mainly due to the low retention of phosphopeptides on C18 columns and
ion

suppression

of

phosphopeptides

by

non-phosphopeptides.

Thus,

MS-based

phosphoproteomics studies in recent years inevitably included enrichment of phosphopeptides

6

prior to LC-MS analysis.[32] For example, using the combination of strong cation exchange
(SCX) chromatography and immobilized metal ion affinity chromatography (IMAC) to separate
and enrich phosphopeptides, Gygi and coworkers employed LC-MS/MS and database searching
to characterize over 5500 and 13000 phosphorylation events from several mg of proteins from
mouse liver and Drosophila embryos, respectively, in two days.[33]

In analysis of

phosphorylation of targeted proteins, enrichment of phosphopeptides from less complex digests
(e.g. recombinant or gel electrophoresis- fractionated proteins) led to greatly improved
phosphopeptide signals for MALDI-MS/MS or ESI-MS/MS characterization.[30] Section 1.3
reviews in detail the various phosphopeptide enrichment strategies for different types of samples.
An alternative to the bottom- up approach is “top-down” analysis (Figure 1.2). Instead of
breaking proteins into small peptides (mainly <3 kDa), the top-down strategy introduces intact
proteins into the gas phase for fragmentation inside the mass spectrometer.[34] This approach
aims to examine the entire protein sequence and comprehensively identify PTMs. Because of the
high masses of proteins relative to peptides, top-down methodologies require a much greater MS
resolving power than bottom- up methods. Thus, expensive high resolution mass analyzers such
as Fourier transform ion cyclotron resonance (FT-ICR) with electron-capture dissociation (ECD)
are typically needed to provide accurate mass values and efficient ion fragmentation.[20]
However, with recent advances in instrumentation the top-down approach is becoming more
attractive for MS characterization of modifications of a single protein or a simple protein mixture
when the protein mass is <50 kDa.[20,34] Examples of top-down MS protein characterization
include comprehensive mapping of histone PTMs,[35,36] identification of a family of highly
related genes encoding proteins with multifaceted mass differences,[37] analysis of membrane

7

proteins,[38] and revealing the alterations of phosphorylation between the truncated and fulllength forms of cardiac myosin binding protein C (cMyBP-C).[31]
In spite of successes with targeted studies of single proteins, the top-down method suffers
from several issues. Challenges in coupling proteome fractionation with MS analysis restrict the
throughput of top-down proteome analysis,[39] and thus far only a few studies reported analysis
of several hundred proteins in a short period of time. Another challenge for top-down protein
characterization is the less efficient MS/MS for whole proteins compared to MS/MS of small
peptides. Currently, direct MS/MS of a protein with a mass >50 kDa typically generates
fragments that give the sequence information primarily for only the two tails of the protein,
probably because of the folded protein confirmation in the gas phase.[31,34]
To overcome the limitations of the top-down approach in protein characterization with
MS, a few recent studies proposed and implemented a so-called “middle-down” method.[20,40]
This strategy produces relatively large peptides (typically 2000-10000 Da compared to <3000 Da
for tryptic peptides) either via enzymes or chemicals that cleave proteins at a single type of
amino acid residue (e.g. Lys, Glu or Asp), or via mild (short-time) tryptic digestion. These long
peptides maintain advantages of the top-down strategy such as high coverage of PTMs on a
single peptide chain and enhanced ionization for phosphopeptides, but the middle-down
approach decreases the resolution requirement for analysis. Section 1.4 enumerates and discusses
some strategies for “middle-down” protein analysis.

8

1.3 Phosphopeptide enrichment strategies
1.3.1 History of enrichment methods for phosphoprotein analysis
Traditionally, phosphoprotein analysis relied heavily on antibody-based purification. In
this strategy, single phosphoproteins are isolated from a cell lysate by immunoprecipitation using
immobilized antibodies with specific affinity toward the target protein, followed by gel
electrophoresis separation and Western blotting to find the correct protein band for MS
analysis.[29] In principle enrichments of phosphoproteins or phosphopeptides using antibodies
against phosphorylated serine, threonine or tyrosine residues are possible. However, only
phospho-Tyr specific antibodies are available for general phosphopeptide isolation, which
hampers routine analysis of proteins containing phospho-Ser and phospho-Thr.[29,41]
An emerging awareness of the importance of pro tein phosphorylation expedited the
development of more general phosphopeptide isolation techniques in recent years. In one of
these techniques, reversible covalent binding, selective reactions convert phosphorylated
residues to other functional groups (e.g. thiols) that can bind to beads modified with coupling
groups (e.g. dithiopyridine). Eluting the beads with small molecules or strong acids releases the
peptides for MS analysis.[42] Although very specific in principle, this approach either enriches
only the peptides with pSer or pThr residues, or suffers from a high sample loss or side
reactions.[42]
The most common current approach to phosphopeptide isolation involves binding
phosphopeptides to supports containing immobilized metal ions or metal oxides that have
specific affinity toward the phosphates on peptides (Figure 1.3). Due to the complexity of
different types of biological samples, recently developed techniques using this concept cover a

9

wide range of innovations, including novel phospho-specific materials (both new affinity groups
and new supporting materials), new phosphopeptide isolation modes, and optimized enrichment
procedures using commercially available columns/beads. Several reviews of phosphopeptide
enrichment techniques recently appeared.[29,32,42-45] However, these reviews mainly describe
new affinity materials and separation methods and do not typically present the drawbacks of
these techniques or specify the type of biological samples for which specific techniques are
effective. Understanding technical limitations is vital because selection of a proper- but not
necessarily the “most powerful”- method for isolating phosphopeptides depends on the sample
source and complexity. With this in mind, this section mainly discusses techniques reported in
the last three years, with an emphasis on how the features of these methods dictate their
applicability to specific biological samples.

1.3.2 Improvement of phosphopeptide enrichment with conventional IMAC3+

Fe beads and TiO2 particles
Since the first separation of phosphorylated proteins with IMAC in 1986,[46] this
technique has been widely used in phosphopeptide enrichment for MS-based characterization of
protein phosphorylation.[32,42-45] Although metal oxides have been used more and more often
for isolating phosphopeptides in recent years, IMAC is still often employed in large-scale
multistage purification of phosphopeptides from complex protein digests.[33,47-50] Among
different forms of IMAC materials, commercial resins modified with iminodiacetic acid (IDA)3+

Fe

3+

or nitrilotriacetate (NTA)-Fe

are most common.[42] Resins are either packed in columns

(Figure 1.3a) or dispersed in solutions (Figure 1.3b) for capturing and releasing phosphopeptides.
3+

The biggest challenge for IMAC-Fe

is the co-enrichment of acidic non-phosphorylated
10

Figure 1.3. Three typical modes of phosphopeptide enrichment: (a) enrichment using a column packed with metal- ion affinity
beads or TiO 2 particles, (b) enrichment in a solution using dispersed micro- or nano-size particles with their surfaces modified with
metal- ion complexes or metal oxides, and (c) enrichment directly on a flat surface (often a MALDI plate) modified with metal- ion
complexes or metal oxides. (For interpretation of the references to color in this and all other figures, the reader is referred to the
electronic version of this dissertation.)
11

peptides, which normally contain multiple Glu or Asp residues. Although methyl esterification of
carboxylic acid residues can overcome such problems,[51] this step may cause sample loss and
decreased peptide solubility, and incomplete or side reactions may result in complicated mass
3+

spectra.[47,52] IMAC-Fe

also suffers from biased enrichment toward multiply phosphorylated
3+

peptides.[42] Because the binding affinity of IMAC-Fe

for mono-phosphopeptides is much

weaker than the affinity for multiply phosphorylated peptides (even in mild acids such as 0.1%
TFA), mono-phosphopeptides are frequently lost in the flow-through and washing solutions
employed during enrichment.[47]
3+

To improve IMAC-Fe

specificity toward phosphopeptides, Tsai et al. systematically

evaluated the influence of pH and buffer concentration on the capture and release of
phosphopeptides during their isolation from cell lysates.[53] Interestingly, increasing the acetic
acid-sodium acetate buffer concentration while keeping the pH around 3 enhances the
phosphopeptide specificity but avoids more acidic conditions that lead to high specificity but low
phosphopeptide recovery. The presence of organic solvent (normally acetonitrile, ACN) also
affects phosphopeptide binding and release. Jensen and co-workers found that increasing the
ACN content in loading and washing solutions enhances the degree of protonation of carboxylic
acids relative to protonation of the phosphate groups.[54] Based on this finding, including 60%
ACN and 100 mM acetic acid in IMAC loading and washing solutions provides a moderately
acidic condition (~pH 3.2) that prevents binding of most of the non-phosphopeptides from a
digest mixture with 1:1000 molar ratio of phosphopeptide-to-nonphosphopeptide.
Metal oxides are beginning to displace metal- ion affinity resins for phosphoproteome
studies. They are either packed in columns (Figure 1.3a) or dispersed in a solution (Figure 1.3b)

12

for enrichment. Due to their phosphopeptide specificity and compatibility with a wide range of
buffers and pH values, TiO 2 particles are especially popular for phosphopeptide enrichment.
Previous studies explored the addition of small organic acids to loading and washing solutions to
suppress the binding of acidic non-phosphorylated peptides to TiO 2 particles. The presence of
2,5-dihydroxybenzoic acid, salicylic acid, phthalic acid and benzoic acid can improve
phosphopeptide specificity, but some reports suggest that these compounds are sometimes
incompatible with LC-ESI and may reduce the number of identified phosphopeptides.[42,55]
Recent work using TiO 2 particles for phosphopeptide enrichment addressed additional
factors that influence large-scale phosphoprotein analysis. Li et al. found that the peptide-toTiO2 ratio affects specificity in phosphopeptide enrichment from a digest of Hela cell lysate
protein.[56]

As the peptide-to-TiO 2 mass ratio decreases from 4 to 0.025, the number of

identified phosphopeptides and their percentage of the total number of identified peptides
initially increase and then decrease. The highest value for both phosphopeptide percentage and
total number of phosphopeptides occurs with a 0.5-0.125 peptide-to-TiO 2 mass ratio.
Interestingly, the use of higher peptide-to-TiO 2 mass ratios leads to detection of more multiply
phosphorylated peptides than the optimized peptide-to-TiO 2 mass ratio. Changing the peptide-toTiO2 mass ratio from 4 to 0.05 decreased the ratio of multiply phosphorylated peptides to total
phosphopeptides from 90% to 5%. Sequential enrichments using low amounts of fresh beads in
the same digest solution identified more and more mono-phosphopeptides, suggesting that
unidentified mono-phosphopeptides in early enrichment fractions were left in the supernatant

13

and not bound to the particles. These findings should be useful for analysis of multiple
phosphorylations along a short fragment of protein sequence, which is often related to ordered
phosphorylation events.[50]

Figure 1.4. Schematic overview of a vented column system with a triple stage precolumn. The
"sandwich" precolumn consists of three individual precolumns of C18, TiO 2, and C18 material,
respectively. During analysis, all peptides are initially trapped during sample loading onto the
first C18 column. After flow splitting, nonphosphorylated peptides are analytically separated and
subsequently detected, while phosphorylated peptides are trapped on the TiO 2 precolumn at a
very low flow rate. This optimal procedure allows superior binding efficiency, witho ut
increasing total analysis time. (Figure is reprinted from reference [55] with permission of
American Chemical Society (ACS).)

Heck and co-workers developed a sensitive online TiO 2-based phosphoproteomics
system (Figure 1.4).[55] The crucial feature of this system is the C18-TiO 2-C18 “sandwich”
configuration, in which the first C18 column (0.2 L column volume) traps all the peptides at a
relatively fast flow rate (3 µ
L/min), followed by a much slower gradient elution (0.1 µ
L/min)

14

that carries the “concentrated” peptides onto the TiO 2 column where the phosphopeptides bind to
TiO2 particles and non-phosphorylated peptides flow through. Injection of a basic buffer into the
TiO2 column elutes phosphopeptides and carries them to the second C18 column for desalting
and subsequent separation followed by ESI-MS/MS analysis. The slow flow rate (0.1 µ
L/min)
that carries the concentrated peptides from the first C18 column to the TiO 2 column yields a
greatly improved phosphopeptide recovery compared to directly injecting several tens of L of
sample into a TiO 2 column at several µ
L/min (a low flow rate with such a large volume would
result in extended loading times), without any compromise on sample analysis time.
In MS analysis, the presence of many non- and mono-phosphopeptides often suppresses
ionization of multiply phosphorylated peptides and hinders their MS/MS analysis.[29] A very
recent strategy, sequential elution from IMAC (SIMAC) aims to separately elute mono- and
multiply phosphorylated peptides during IMAC to increase the recovery of monophosphopeptides and enhance the ionization efficiency of multiply phosphorylated peptides.[47]
3+

This procedure includes incubating the protein digest with IMAC-Fe

beads in a weakly acidic

(0.1% TFA) solution followed by initial elution with 1% TFA. Further enrichment of the
supernatant and initial eluate on TiO 2 recovers mono-phosphopeptides; while a secondary
3+

elution on IMAC-Fe

beads with pH 11.3 ammonium hydroxide recovers multiply

phosphorylated peptides. In SIMAC of digested protein from human mesenchymal stem cells,
more than 96% of the identified multiply phosphorylated peptide s were in the ammonium
hydroxide buffer, and the number of identified mono-phosphopeptides was more than when
using TiO 2 alone for enrichment. Similarly, Simon et al.[57] demonstrated that sequential elution
15

of TiO 2 with pH 8.5 triethylammonium bicarbonate buffer and then pH 11.5 NH4OH solution
releases mono- and multiply phosphorylated methyl esterified peptides, respectively, in the
different eluates. They found that the mono-phosphopeptides recovered from the first elution are
purer than from a single time elution using a pH 11.5 NH4OH. The above results suggest an
3+

improved enrichment efficiency for both IMAC-Fe

and

TiO 2 after fractionating

phosphopeptides.
3+

I should note that reported enrichment conditions or strategies for IMAC-Fe

or TiO 2

from different groups should be checked carefully when applied to a user‟s specific enrichment
3+

materials and samples. As the supporting materials and ligands for IMAC-Fe , as well as the
particle size and surface properties of TiO 2 particles vary from product to product, the optimized
solvent, pH and sample loading capacity need to be adjusted in order to obtain the best result.

1.3.3 Novel phosphopeptide enrichment materials
3+

Despite the success of IMAC-Fe

beads (which typically employ NTA or IDA as

ligands for binding metal ions) and TiO 2 particles in phosphopeptide enrichment, a number of
recent reports describe new enrichment materials with novel ligands, ions, or supporting
structures to target different aspects for improving phosphopeptide enrichment.

Aiming at

enhancing phosphopeptide enrichment specificity, Zou and co-workers prepared a phosphatecontaining polymer monolith.[58] Grinding the monolith into irregular beads followed by
4+

4+

formation of phosphate- Ti complexes (IMAC-OPO3-Ti ) yielded a resin that enriched
phosphopeptides from model phosphoprotein digests containing a 500 molar excess of non16

phosphoprotein digest. In the phosphoproteome analysis of mouse liver, enrichment with IMAC4+

OPO 3-Ti beads led to enhanced identification of phosphopeptides compared to enrichment
3+

using ZrO 2 particles, TiO 2 particles or commercial IMAC-Fe

beads (2-5 fold increase in both

the total number of phosphopeptides and the percentage of detected peptides that are
4+

phosphorylated. The authors attributed the superior enrichment with OPO 3-Ti

to its strong

binding of phosphopeptides under the highly acidic condition (6% TFA) and the flexible spacing
arm from the supporting polymer. Interestingly, although of all the enrichment materials tested
4+

IMAC-OPO 3-Ti
mouse

liver

allowed identification of the largest number of phosphopeptides from a

lysate

digest,

the

corresponding

ratio

of

multi-phosphorylated

to

monosphosphorylated peptides was the least. This is consistent with prior studies that suggest a
combination of enrichment materials enhances the comprehensiveness of phosphopeptide
identification.[27,47,49]
Zou and co-workers also synthesized highly ordered mesoporous silica particles
4+

containing OPO3-Ti

complexes to isolate phosphopeptides from sera of hepatocellular

carcinoma patients and healthy individuals. Such modified silica facilitates the capture of
phosphopeptides from human plasma based on both size exclusion of undigested proteins and the
4+

high affinity of OPO3-Ti

toward phosphopeptides. This material identified mono-

phosphorylated human fibrinogen peptide A (FPA) and its three isoforms, and quantitative
analysis based on either MALDI-MS or iTRAQ (isobaric tag for relative and absolute
quantitation) labeling suggested that the phosphorylated FPA was expressed differently between
the cancer and healthy groups.

17

To overcome the issue of enriching mainly mono-phosphopeptides with Ti(IV)-based
materials, Wan et al.[59] synthesized a macroporous ordered silica foam (MOSF) with the
surface modified with TiOx that is rich in coordinatively unsaturated Ti(IV) species. The ~100
nm pore size in this material allows for rapid mass transfer, and enrichment of multiphosphorylated peptides occurs even with low amounts of protein (e.g. 10- min enrichment using
40 g of Ti-MOSF in a 50 L solution containing 340 fmol -casein digest) or large amounts of
non-phosphopeptides (e.g. 10:1 mass ratio of BSA/-casein). The authors attribute the high
specificity for binding of multi-phosphorylated peptides to the bidentate phosphate chelation of
the 4-coordinate Ti(IV) species that is rich on the material surface. On conventional TiO 2
particle surfaces phosphate binding occurs in a bidentate bridging mode with a 5-coordinated
Ti(IV) moiety. As multiple phosphorylation events often occur sequentially within a short stretch
of protein, sensitive detection of multi-phosphorylated peptides using such materials will help
investigate ordered phosphorylation events on motifs with more than one phosphorylated
residue.[50]
Although enrichment can occur with dispersed particles, column-based methods
dominate the phosphoproteome field because of their convenience and potential in online
analysis. In a recent study, Lin et al. employed liquid-phase deposition to form a thin layer of
TiO2 nanoparticles on the surface of a capillary.[60]

This column selectively concentrated

phosphopeptides from 2 pmol of -casein digest in the presence of a 10-fold molar excess of
non-phosphoprotein digest. A second deposition of TiO 2 on the capillary wall leads to much
improved selectivity and S/N for phosphopeptides. Lin et al. applied this TiO 2- modified
capillary to enrich substrate peptides phosphorylated by recombinant Abelson tyrosine kinase
18

(Abl) in the presence of kinase inhibitors.[61] Addition of an internal standard phosphopeptide
that mimics the substrate phosphopeptide during enrichment allows quantitative characterization
of Abl inhibitors by MALDI-MS. Because the sample is relatively simple, limited binding
capacity is not an issue.

4+

Figure 1.5. (a) synthesis of –PGMA-RPO3-Zr brushes in a capillary column using ATRP of
glycidyl methacrylate and subsequent derivatization. (b, c) SEM images of the capillary before
(b) and after (c) 20 h of polymer grafting. Figures a, b and c were adopted (with modifications)
from reference [62] with permission of ACS.

To increase the binding capacity of capillary columns for large-scale phosphopeptide
enrichment, Qin et al.[62] recently synthesized 3 to 5 m-thick polymer brushes on the inner
19

wall of a capillary (75 m I.D. × 50 cm).

After surface- initiated atom-transfer radical

polymerization (SI-ATRP) of glycidyl methacrylate to form the brush, modification of the
polymer with a spacing arm and terminal zirconium phosphonate gives a stationary phase that
specifically binds phosphopeptides (Figure 1.5). The loading capacity of a synthetic monophosphopeptide in the modified capillary is ~120 pmol (0.054 mol/mL), more than 1 order of
magnitude higher than the capacity of a conventional open tubular capillary column developed
by the same group.[63] This polymer- modified capillary column can enrich 5 pmol of -casein
phosphopeptides in the presence of a 100- fold molar excess of BSA digest under mildly acidic
conditions (0.1% TFA with 80% ACN). Moreover, phosphopeptide isolation with these columns
after SDS-PAGE fractionation and in- gel digestion facilitated the identification of ~500 unique
phosphopeptides from 400 g of human Hepatocellular carcinoma cell (Hep G2) lysate.
Hou et al.[64] prepared an organic-inorganic hybrid silica monolith column with
4+

immobilized –RPO3-Ti

for IMAC enrichment of phosphopeptides. The porous structure

provides a much higher loading capacity for a synthetic phosphopeptide (1.4 mol/mL) than the
capillary column modified with polymer brushes.[62] In addition, the organic- inorganic hybrid
structure is superior over pure silica monoliths and organic monoliths because of its stability over
a wide range of pH (pH 0.5-11.8) and organic solvent content (i.e. 50% ACN). This high
stability permits the use of mobile phase conditions that minimize co-enrichment of nonphosphopeptides or enhance elution efficiency. In the analysis of a mitochondrial-proteinenriched fraction from 2 g of rat liver tissue, enrichment with the monolith column led to
identification of 224 phosphopeptides from 148 phosphoproteins.

However, multi-

phosphorylated peptides accounted for only 7.1% of the total number of phosphopeptides.

20

The above novel materials present different advantages in isolating phosphopeptides. The
dispersed particles may effectively enrich phosphopeptides from moderately complex samples,
such as certain fractions of cell lysate. Rapid mass transfer in swollen polymer-based affinity
materials or monoliths with small pore size may enhance the rate of phosphopeptide capture and
elution relative to isolation with columns containing commercial IMAC beads or TiO 2 particles.
If so, the new materials might expedite high- throughput phosphopeptide analysis, which requires
short loading and elution times. However, careful comparisons with commercial columns must
be performed to make any firm conclusions.

1.3.4 Ion exchange for phosphopeptide fractionation
For complex biological samples, e.g. digests from whole cell lysate or tissue extracts,
direct enrichment of hundreds to thousands of phosphopeptides using metal ions or metal oxides
is not efficient. More comprehensive isolation of phosphopeptides from such samples typically
relies on pre- fractionation of peptides using ion exchange, particularly SCX where positively
charged peptides bind to the negatively charged stationary phase via electrostatic
interactions.[33,49,50] Elution with a salt and pH gradient releases peptides according to their
charge.[65] A theoretical tryptic digestion of proteins in the human protein database revealed that
68% of the predicted peptides have a +2 net charge at pH 2.7.[65] Phosphopeptides generally
elute

earlier

than

non-phosphorylated

peptides

during

gradient

salt

elution,

and

multiphosphorylated phosphopeptides may not bind to SCX columns at all due to the negative
charge of peptides carrying multiple phosphate groups.[65]

As SCX eluates contain both

phosphopeptides and non-phosphorylated peptides, even in the early eluting fractions,

21

performing IMAC for each fraction typically improves phosphopeptide identification in the later
LC-MS/MS analysis.[33,48]
In a phosphoproteome study of 10 mg of mouse liver tissue,[49] Gygi‟s group first
separated tryptic peptides into 15 SCX chromatography fractions (Figure 1.6). Subsequent
3+

phosphopeptide enrichment by IMAC-Fe

for each fraction, along with a phospho- Tyr specific

antibody enrichment for the total lysate, led to identification of 4038 unique phosphopeptides
that stem from 2149 proteins. However, analysis of SCX fractions 4-10 revealed exclusively
peptides with +1 charge states, suggesting a low resolution in SCX. Similarly, in a large-scale
study of phosphoproteins from postsynaptic density samples using SCX-IMAC, Trinidad et al.
noticed a large elution window for all peptides of a given charge state, and overall elution in
order of charge state.[48] In another phosphoproteome analysis,[50] the SCX-IMAC strategy
revealed 13720 non-redundent phosphorylation sites from 2702 proteins, although 70% of the
phosphopeptides were multi-phosphorylated, suggesting that IMAC still suffers from biased
enrichment of multi-phosphorylated peptides, even after pI-based peptide fractionation. Nie et al.
compared several combined enrichment methods, including SCX, strong anion exchange (SAX)
and TiO 2, in phosphopeptide enrichment from a Hela cell lysate digest. Enrichment of
phosphopeptides using TiO 2 alone and combined SCX/TiO2 reveal identical distributions for the
calculated pI values of the enriched peptides and some bias toward basic or monophosphopeptides.[66] Heck and coworkers[55] showed similar pI distributions for peptides
obtained from solely TiO 2 and combined TiO 2/SCX strategies, suggesting that TiO 2 and SCX
provide parallel enrichment selectivity in isolating phosphopeptides. [55,66] These findings again
3+

suggest that separate IMAC-Fe

and TiO 2 phosphopeptide enrichment will increase the
22

Figure 1.6. Strategy used by Gygi et al. for large-scale identification and characterization of
phosphorylation sites in mouse liver proteins. After homogenization and lysis, 10 mg of protein
was tryptically digested and subjected to a two-step phosphopeptide enrichment. SCX
fractionation isolated many phosphopeptides in early eluting fractions, and subsequent IMAC of
each fraction provided additional selective capture. In addition, 80 mg of tryptic peptides were
enriched for pTyr-containing peptides by immunoaffinity purification. Figure is reprinted with
permission from reference [49] with modifications.

comprehensiveness of phosphoproteome studies, even when SCX is utilized for peptide prefractionation.
Compared with SCX, SAX chromatography provides an alternative or even superior
approach to peptide fractionation before phosphopeptide enrichment. A SAX column binds

23

negatively charged peptides at weakly acidic or neutral conditions and then releases these
peptides during elution with salt or pH gradients.[66,67] Due to the negative charge of the
phosphate groups, phosphopeptides typically elute later than non-phosphorylated peptides.[66,67]
In a study of the phosphoproteome in a human liver,[67] both SAX chromatography and IMAC3+

Fe

enriched acidic peptides, but SAX separated phosphopeptides from acidic non3+

phosphorylated peptides while IMAC-Fe

enriched both species simultaneously. In addition,

SAX chromatography had fewer unbound phosphopeptides than separation with SCX.[67] Two
studies found that SAX typically shows better phosphopeptide fractionation resolution (the width
of fractionation window for a given charge state) than SCX and enriches more multiply
phosphorylated peptides than TiO 2 or SCX.[66,67] Nie et al. showed that an online continuous
pH gradient elution in SAX and offline enrichment of any unbound phosphopeptides using TiO 2
leads to a very comprehensive coverage of both mono- and multiply-phosphorylated peptides.[66]
Large-scale phosphopeptide enrichment typically identifies hundreds to thousands of
phosphoproteins, which allows statistical analysis of phosphorylation patterns. For example, the
mouse liver phosphoproteomics study by Gygi[49] revealed both novel and known motifs for
specific Ser/Thr kinases, with more frequent phosphorylation close to the protein C-termini than
in other regions. They also found some cases in which only one of the two specific loci in a
double phosphorylation motif existed as a mono-phosphorylated species, suggesting sequential
phosphorylation of the two sites. In all, phosphopeptide isolation strategies are flexible, and
combinations of different techniques could be employed in different situations.[66,68,69]

24

1.3.5 Phosphopeptide enrichment directly on modified MALDI plates
Enrichment on an affinity surface through incubation of a drop of sample followed by
rinsing is one of the most direct methods for phosphopeptide isolation. After subsequent addition
of a drop of eluent and matrix, such surfaces can serve as MALDI plates for MS or MS/MS
analysis (Figure 1.3c). This “on-plate enrichment” is simple and minimizes consumption and loss
of samples.[42] With recent advances in surface chemistry, a number of affinity interactions
previously employed for on-column enrichment have been implemented in “on-plate” methods
with sensitive and specific enrichment of phosphopeptides from digests of model proteins.[42,70]
Notably, studies published in the last two years addressed enrichment conditions specific for
MALDI sample preparation, and more importantly, began applying this approach to biological
samples.
4+

Hoang et al. immobilized Zr

on spots of phosphonate-terminated self-assembled
4+

monolayers for phosphopeptide enrichment and direct MALDI-MS analysis.[30] The Zr modified spots are surrounded by a hydrophobic monolayer that confines a 5 L sample for
4+

analysis. A comparison study showed that the film with Zr
4+

specifically than films with Ti

and Fe

3+

isolates phosphopeptides more

ions. Although such films can enrich phosphopeptides

in samples containing 100 fmol of phosphopeptide and 1 pmol of nonphosphoprotein digest, a
spot has a binding capacity of only ~650 fmol of -casein phosphopeptides because of the
limited number of affinity groups in the monolayer. On-plate phosphopeptide isolation using this
film, together with MALDI collision- induced dissociation (CID)-MS/MS and Mascot database
searching, identified a phosphopeptide from recombinant human MAP Kianse1, as well as
several phosphoproteins digested from spots in a 2D electrophoretic gel.
25

Blacken et al. fabricated several metal oxide-coated steels by reactive landing of gasphase ions produced by electrospray ionization of group IVB metal alkoxides.[71] In enrichment
and MALDI-MS analysis of -casein/BSA (1:2 molar ratio) digests on several metal oxide
surfaces, ZrO 2 gave higher intensities of mono- and doubly phosphorylated -casein peptides
than HfO 2 and TiO 2 surfaces. The authors attributed the higher intensities on ZrO 2 to the
optimal balance of surface coverage and metal oxide surface area compared to the other two
materials. Unfortunately, the influence of the intrinsic metal oxide properties on phosphopeptide
binding and elution is not clear. The ZrO 2 surface enriched a phosphopeptide from a digest of cSrc tyrosine kinase, which had been dephosphorylated and then autophosphorylated with ATP
for different times. Addition of an internal standard phosphopeptide in the differentially treated
digest samples and subsequent enrichment allowed monitoring of the kinetics of the
autophosphorylation.
Similarly, Niklew et al.[72] fabricated spots (1- mm dia.) of anatase TiO 2 nanoparticle
films on F-doped conductive SnO 2 using a screen-printing technique (Figure 1.7a-c). The
advantage of such films is their high internal surface area compared to flat surfaces, as revealed
by isolation of phosphopeptides from a 1-L digests mixture containing 15 pmol of BSA and 4
pmol each of -casein and -casein. This TiO 2 film also enriched phosphorylated fibrinopeptide
A from a human fibrinogen digest (m/z 1617 in Figure 1.7d). Moreover, dephosphorylation of
bound -casein peptides by phosphatase treatment directly on the TiO 2 surface was possible,
although less efficient for the dephosphorylation of multiply phosphorylated peptides (Figure 1.7
e and f).
26

Figure 1.7. SEM images of F-doped SnO 2 film before (a) and after (b) modification by screen
printing of TiO 2. (c) A profile scan of one spot by interference spectroscopy illustrates the
average layer thickness on that spot. (d) MALDI mass spectra of the thrombin-digested
fibrinogen (60 pmol) after sample enrichment on the TiO 2 film. (e, f) MALDI-TOF mass spectra
of -casein phosphopeptides enriched on the TiO 2 film before (e) and after (f) treatment with
protein phosphatase directly on the TiO 2 film. Figures were adopted (with modifications) from
reference [72] with permission of ACS.

Porous surface structures can potentially maintain fast mass transfer and increase binding
capacity, and a homogeneous surface is crucial for obtaining reliable enrichment results. One of
the limitations of on-plate enrichment is the difficulty of performing multi-step washing or
elution (e.g. SIMAC in column- mode phosphopeptide enrichment[47]), or multi-dimensional

27

enrichment with more than one phospho-affinity material (e.g. tandem enrichment using IMAC
and TiO 2 ), to improve phosphopeptide recovery. Another drawback lies in the low binding
capacity (typically from low pmol to several hundred fmol of phosphopeptides) on flat surfaces
compared to columns. Last, MALDI-MS/MS is usually less rich in information than ESI-MS/MS
due to the low fragmentation efficiency of singly charged peptide ions in the gas phase produced
by MALDI.[30] However, MALDI-MS also has several advantages over ESI-MS for
phosphopeptide analysis. (1) The domination of singly charged ions greatly simplifies mass
spectra compared to those produced by ESI. (2) The pulsed nature of MALDI makes possible a
data-driven, selective MS/MS process. (3) Matrix/sample spots are often stable for several days
and can be investigated multiple times. Considering the above disadvantages and advantages, onplate enrichment is primarily applicable to analysis of limited amounts of target
phosphoprotein(s), rather than large scale identification and characterization of unknown
proteins. Possible applications of on-plate phosphopeptide enrichment include high-throughput
protein kinase assays and analysis of antibody- isolated phosphoprotein(s) or protein digests from
gel electrophoresis bands.

1.4 Middle-down protein analysis by MS
1.4.1 Advantages of middle-down protein analysis
Long proteolytic peptides (typically 2000-15000 Da) have several advantages for
peptide/protein characterization: (1) They are generally more hydrophobic than short peptides
and separate with higher resolution in LC.[73] (2) Compared to small tryptic peptides, which
mainly generate y-ions during CID-MS/MS due to the basic Arg or Lys residues primarily at the
28

C-terminus, large peptides carry higher numbers of Asp or Lys along the peptide chains and thus
give more efficient MS/MS fragmentation.[73,74] (3) For peptide sequence identification by
database searching after MS/MS fragmentation, larger precursor ions have fewer potential
protein candidates in the database than small ones. In addition, when using high mass accuracy
instrumentation such as Fourier transform MS, setting strict precursor ion mass tolerances for
database searching greatly improves the confidence of peptide assignments. In other words,
identification of large peptides generally requires fewer fragment- ion assignments than
identification of small peptides.[74] (4) Large peptides provide higher coverage of PTM sites
than short tryptic peptides, facilitating the study of PTM coordination at nearby regions of the
protein.[20,35] In all, the middle-down strategy provides many advantages of top-down or
bottom- up strategies, but avoids (to some extent) the instrumental challenges of the top-down
method (e.g. limited MS/MS efficiency and the demand for high mass resolution), as well as the
requirement of multidimensional separation for complex mixtures of small peptides in the
bottom-up method.
Several studies showed that the middle-down approach is an important complementary
method to MS characterization of the whole protein.[31,36,75,76] For example, the Kelleher
group[36,76] demonstrated the power of a middle-down approach in localizing multiple PTMs at
histone H3 tails (1-50 residues) using Glu-C digestion followed by ECD MS/MS of the large
peptide. They found that ECD of the whole histone H3 protein generates abundant fragments
mainly from the unmodified part of the protein, while performing ECD of the tails improves
localization of the majority of histone H3 modifications.[36] Similarly, Ge et al. unambiguously
identified all of the phosphorylation sites in the truncated and full- length forms of cardiac
myosin binding protein C (cMyBP-C) (142 kDa) using top-down and middle-down MS.[31] This

29

combination of approaches ensures the full sequence coverage of cMyBP-C. Other middle-down
examples include: analysis of the length and linkages of polyubiquitin chains by combing partial
digestion and LC-MS,[77] high-confidence identifications of ~30% of the highly basic ribosomal
proteome with an average 46% sequence coverage of 70-79 possible parent proteins,[73]
enhanced determination of multiple phosphorylation sites on recombinant nucleotide-binding
proteins in Escherichia coli (E. coli),[78] and strengthened identification of phosphorylation and
glycosylation sites of proteins using large peptides with the combination of CID, ETD, and CID
of an isolated charge-reduced (CRCID) species derived from ETD.[79]

1.4.2 Middle-down proteolysis methods
1.4.2.1 Targeted proteolysis using proteases with single-residue specificity
To increase the size of proteolytic peptides, a number of studies employed proteases that
cleave proteins at only one residue that appears relatively infrequently along the protein chain.
The most common proteases cleave proteins at multiple types of amino acid residues and give
rise to relatively short peptides. These low-specificity enzymes include: pepsin (cleaves peptide
bonds at hydrophobic and preferably aromatic amino acids such as phenylalanine, tryptophan,
and tyrosine)[80], chymotrypsin (cleaves at the carboxyl side of aromatic amino acid residues,
including tyrosine, tryptophan and phenylalanine)[81] and trypsin (cleaves at the carboxyl side
of Lys or Arg except when either is followed by Pro).[82] In contrast digestion with Glu-C
(cleaves at the C-terminus of Glu),[83] Lys-C (cleaves C-terminally at Lys)[84], and Asp-N
(cleaves N-terminally at aspartic and cysteic acid)[85] may lead to somewhat larger proteolytic
peptides. Examinations of the abundance of amino acid residues in human proteins show similar
fractions for Asp, Lys and Glu (~5%).[74,86] However, for a target protein or protein family, a

30

careful examination of the abundance of Glu, Lys or Arg is important for selecting the proper
protease(s). For example, to circumvent the limitations of top-down analysis of multiple PTMs
on histone, several recent studies focused on the characterization of residues 1-50 where most
PTMs occur.[35,36,75,76] Glu-C digestion of histone provides this large PTM-containing
peptide because of the sparsity of Glu in this region.
In some cases separate digestion with several proteases facilitates complete sequence
coverage with large peptides. Ge et al. clearly demonstrated this strategy.[31] In top-down
analysis of a truncated form (C0-C4, 642 residues) of cMyBP-C using high resolution ESI/FTMS
with ECD- and CID-MS/MS, they identified the sequence and PTMs of only the two tails of the
whole protein. To characterize the central region of C0-C4 (covering 53% of the amino acid
residues), the Ge group separately performed middle-down digestion with three different
proteases, Asp-N, Glu-C and Lys-C. They sequenced 54 resulting peptides, mostly with
molecular masses from 2500-10000 Da, and the peptide revealed the entire sequence and all
phosphorylation sites.
Some studies employing Lys-C for protein digestion also show the advantage of large
peptides in characterizing protein PTMs. Lys-C digestion of a set of recombinant nucleotidebinding proteins from E. coli followed by MALDI QTOF MS/MS greatly improved the
identification of multiply phosphorylated peptides compared to LC-ESI-MS/MS after tryptic
digestion.[78] The preserved positive charge from Arg within the peptide c hain may explain the
enhanced ionization of phosphopeptides.[78]

Wu et al. showed similarly improved PTM

characterization for -casein (a 24 kDa model protein with 5 phosphorylation sites) and human
epidermal growth factor receptor (EGFR, a 132 kDa protein with both phosphorylation and
glycosylation).[74] They observed that the Lys-C digestion of the two proteins into big peptides

31

followed by sensitive online LC-LTQ-FTICR resulted in much higher sequence coverage than
tryptic digestion (e.g. 97% vs. 36% for 50 fmol of -casein, and 95% vs. 70% for 1 pmol of
EGFR). In addition, compared to small tryptic peptides, the middle-down strategy improves the
ionization of peptides with PTMs, as well as the b- ion intensity in CID-tandem MS. These
features led to localization of all the five phosphorylation sites on -casein, as well as 10
glycosylation sites and 3 phosphorylation sites on EGFR.
1.4.2.2 Targeted proteolysis using acid cleavage

Figure 1.8. Mechanisms of the acid cleavage at Asp-X (path A) and X-Asp (path B) bonds (X
represents any amino acid residue). The figure was adapted from reference [86].

In 1950, Partridge et al. found that heating proteins in solutions containing weak acids
preferentially releases aspartic acid from the protein chains.[87] Based on this discovery and
subsequent work,[88] several studies employed acid cleavage (e.g. using formic acid or acetic
32

acid) to digest proteins at either the C-terminal or N-terminal sides, or both sides, of Asp (Figure
1.8) prior to MS analysis of these peptides.[86,89,90] High temperature (>108 ° and low pH
C)
(<2.1) are vital for achieving rapid and efficient digestion,[89,90] and the use of microwaveassisted Asp-specific acid cleavage can shorten the digestion time from hours to 0.5-20 min.[91]
In contrast to enzyme-based digestion, acid cleavage can take place in a salt- free system to avoid
the need for sample desalting prior to MS analysis. Thus, in some cases acid digestion leads to a
mass spectrum with less noise than in spectra obtained after enzymatic digestion.[86,92]
Additionally, acidic solutions (~pH 2) readily denature many proteins to increase their solubility
without adding denaturants such as urea.

Finally, acid digestion avoids the production of

unwanted peptides due to protease autolysis.
In human proteins, 4.9% of the amino acid residues are Asp, which is a little lower than
the abundances of Lys (5.7%) and Arg (5.6%).[86] Thus Asp-cleavage on human proteins
should in principle occur at 43% of the frequency of trypsin cleavage (complete digestion) to
give relatively large peptides. Fenselau and coworkers studied microwave-assisted acid digestion
of several types of proteins,[73,89-91,93] including digestion of highly basic ribosomal proteins
for middle-down analysis.[73] In the human ribosomal database, Asp residues account for only
3.8% of total residues, in contrast to 9.3% and 12.4% for Arg and Lys, respectively.[73] In silico
digestion of 84 human ribosomal proteins suggests that ~60% of tryptic peptides have ≤10
residues, and ~75% of Lys-C digested peptides have ≤20 residues. In contrast, both the in silico
and experimental results show that more than two thirds of the acid-digested peptides have >20
residues.[73] These large peptides led to an average 46.2% sequence coverage for 70 ribosomal
proteins after LC-LTQ-orbitrap analysis. The authors attributed the high sequence coverage to (1)
the multiple internal basic residues that readily produce multiply charged precursor ions and

33

product ions (by CID), and (2) reliable identification of long peptides with only limited numbers
of product ions when using high-resolution mass measurements of precursor and product ions.
In another work, the Fenselau group digested bacteriophage MS2 coat protein (13728 Da
with 129 amino acid residues including 4 Asp) into several large peptides (9000-12500 Da) that
have sequence overlaps and cover the whole protein sequence. [89] After examination by
MALDI- TOF MS, these large peptides help distinguish the MS2 strain coat protein from
bacteriophage R17, a protein with only a 1 Da difference from the MS2 strain protein due to a
sequence reversal at two sites and an Asp/Asn replacement. Such a minor mass difference was
not detectable when the whole proteins were examined with MALDI-TOF. Unfortunately, acid
cleavage shows relatively low digestion efficiency (in terms of consumption of intact proteins
without considering the peptide size) compared to tryptic digestion, even with the assistance of
microwave radiation. However, Fenselau suggests that the parallel analysis of digest products
and intact protein masses improves the specificity in protein source (e.g. microorganism)
identification.[90]
Although acid digestion is attractive, it has several disadvantages. First, as noted above
the protein digestion efficiency using small organic acids is generally lower than that with tryptic
digestion.[90,92] Second, due to the uncertainty of the cleavage side at Asp (C-terminal or Nterminal or both sides), the number of product peptides could be higher than expected (especially
when the digestion time is short).[86,93] Moreover, peptides with a single terminal Asp at a
peptide N- or C-terminus have the same mass (e.g. DGEWQQVLNVWGKVEA and
GEWQQVLNVWGKVEAD), complicating peptide sequence identification based on m/z
values.[86,89,90] The third challenge in acid digestion is the appearance of side reactions, such
as dehydration and formylation when the digestion time is beyond 2 h (no microwave), that

34

complicate peptide sequence identification.[86]

Moreover, variable extents of removal of

glycans and phosphates from proteins during acid cleavage can compromise characterization of
PTMs.[91,93]
In addition to acid cleavage, other chemicals were employed to generate long peptides by
cleavage at Trp residues.[94,95] Vestling et al.[95] reported the digestion of model proteins
specifically at Trp residues using (3-bromo-3-methyl-2-(o-nitrophenylsulfeny1)indolenine.
However, many side reactions occurred, including oxidation of Trp, Met and Cys (forming
disulfide bonds) and substitution of bromine for hydrogen (probably on Tyr).
1.4.2.3 Partial proteolysis for generating large peptides
Partial digestion, which can occur at short digestion times or with a low enzyme-tosubstrate ratio, provides another strategy for obtaining relatively large peptides.[77] A recent
MS characterization of polyubiquitin (polyUb) chain structure by Xu et al. provides an excellent
example of exploiting incomplete digestion.[77] Ub is a 76-residue polypeptide with 7 Lys and 4
Arg residues. In polyUb, the N-terminus (GG-NH2) of one Ub is linked with any primary amine
(from either Lys or the N-terminus) of another Ub.

With a 1:3 trypsin- to-substrate ratio,

ubiquitin or polyUb in its native (folded) form cleaves completely only at R74 in the C-terminal
tail (RGG-NH2), even after incubation at 37 ° for 9 h. Such a specific cleavage produces an
C
almost full- length Ub (residues 1-74, UbR74) and a linkage-specific GG-tagged UbR74.
Combining this digestion strategy with LC-MS, Xu et al. analyzed the length and linkage sites of
human and yeast polyUb chains.
Limiting the digestion time can also lead to partial digestion. Walcher et al.[96] reported
that 3-min tryptic digestion of the type I antenna protein of spinach photosystem II (Lhcb 1)

35

Figure 1.9. (a) Molecular structure of the pea Lhcb1 antenna protein showing the highly
conserved transmembrane helices and the more variable amino terminus. (b) Liquid
chromatography-ESI MS of the Lhcb antenna proteins (from a gel-electrophoresis band) after 3min typtic digestion for simultaneous analysis of intact proteins (identified by index I), truncated
proteins (identified by indexes F1-F3), and aminoterminal peptides (identified by indexes P1-P3).
Figure is reprinted with permission from reference [96].
36

under non-denatured conditions releases only a few small peptides near the protein N-terminus.
In contrast, the structure of the majority of the protein remains, probably due to restricted access
of trypsin to the stable 3-D structure of the transmembrane part of the protein (Figure 1.9a). RPHPLC (monolith)-ESI-MS and -MS/MS successfully separated and identified the tryptic
products, including small peptides, truncated proteins and intact membrane proteins. This
strategy enabled the identification and characterization of a new isoform of Lhcb 1 (Figure 1.9b).
Short-time digestion has also been employed when using proteases with higher specificity such
as Glu-C.[31]
Compared to digestions using pure acids or proteases with high cleavage specificity (e.g.
Glu-C and Asp-N), partial tryptic digestion has not been systematically explored. Examples of
deliberate partial digestion employ proteins in their native confirmation,[77,96] perhaps because
proteolysis of unfolded protein by trypsin is uncontrolled. Limited studies suggest that tryptic
digestion is a complicated process: the proteolysis occurs either in a one-by-one scheme (the
protease breaks only one bond in a substrate protein before immediate dissociation) or a zipper
scheme (the protease remains bound to the protein until digestion is complete).[97,98] Although
less employed for middle-down protein analysis than high specificity proteases such as Glu-C,
trypsin is potentially attractive for middle-down studies because it is available in mg quantities at
relatively low cost. This allows for inexpensive immobilization of trypsin on different supports
for applications such as online protein digestion-LC-ESI-MS analysis.[99] Thus, using trypsin
for partially digesting proteins is potentially viable for routine, especially large-scale, middledown protein MS analysis. Careful studies on tryptic digestion pathways will be essential for
controlled, partial proteolysis.

This dissertation describes immobilization of trypsin in

membranes to allow extremely short digestion times that yield large peptides.

37

1.5 Tandem MS and data analysis for peptide sequencing, protein
identification and phosphorylation localization
Tandem mass spectrometry is vital for peptide or protein sequencing and PTM
characterization.[20,41] Collision- induced dissociation (CID), which increases the internal
energy of gas-phase peptide/protein ions relatively slowly (sec-msec) to induce backbone
cleavage through the lowest-energy dissociation pathway, is the most widely used fragmentation
technique for MS/MS.[41] This process breaks peptide chains at the C-N bond to give a series of
b- and y- fragment ions. For phosphopeptides, a prominent neutral loss of H3PO4 (m/z 98), along
with neutral losses of H2O and NH3 are also very common in CID.[20] Although the neutral loss
of 98 facilitates phosphopeptide identification, it hampers backbone fragmentation for peptide
3

sequencing and assignment of the phosphorylation site. Performing MS/MS/MS (MS ) in an ion
trap, through isolation of the neutral loss ion followed by CID and analysis of the fragment ions,
may overcome the lack of backbone fragmentation in phosphopeptides. This method was
employed in phosphoproteome analysis with software-controlled, neutral loss-dependent
3

3

3

MS ,[65] but the MS efficiency is high primarily for abundant ions.[29] Pseudo MS , where
fragmentation of the precursor ion and the ion originating from the neutral- loss occur
simultaneously, can also improve database searching based on fragment ions.[100]
Two alternative fragmentation methods, ECD and electron transfer dissociation (ETD),
show tremendous promise in fragmentation of large peptides and peptides with PTMs.[20,41] In
ECD, capture of a thermal electron by a peptide ion results in a nonergodic dissociation that
produces fragmentation largely independent on the peptide sequence. [101] Due to this advantage,

38

the labile phosphate groups remain on the resulting c and z fragment ions, enabling
phosphorylation site assignment. However, the requirement of a static magnetic field for
producing thermal electrons restricts the use of ECD to FT-ICR. Similarly, ETD employs anions
with low electron affinities to reduce the charge of peptide ions and induce fragmentation via a
nonergodic pathway to produce c and z ions.[102] Due to its implementation in lower-cost
instruments such as quadrupole ion traps, ETD has great potential for global analysis of protein
phosphorylation, especially for peptides with multiple charges (3+) or multiple phosphorylation
sites.[41,102] Moreover, implementation of ETD on a hybrid linear ion trap-orbitrap enables
sequencing large peptides and proteins.[103] Recent studies showed that the combination of CID
and ETD may yield efficient fragmentation for peptides with a wide range of charge
sates.[79,103]
Finally, computational approaches and software tools are vital for rapidly interpreting the
data in MS/MS spectra.[104] Widely used MS/MS data analysis tools include Sequest,
MASCOT, Phenyx, X!Tandem, Sonar and OMSSA, which commonly identify 5–30% of the
MS/MS spectra data set, with an incomplete list of predefined protein modifications for PTM
screening.[105] In a typical MS/MS data analysis using one of the above tools, the proteins in a
database (e.g. Human, chosen according to the sample category) are digested into peptides in
silico according to the cleavage rules of the protease used for sample digestion. Each peptide
generates a theoretical MS/MS spectrum for comparison with the experimental MS/MS data to
generate a similarity score. The highest score corresponds to the most possible peptide sequence
(with modifications if the variable PTMs are pre-set) and the parent protein(s). This approach,
though successful for sequencing unmodified peptides, is less efficient for peptides with PTMs.
The main reasons include complications in spectra due to the modified groups, the greatly

39

increased number of candidate peptides, and difficulties in predicting fragmentation
patterns.[104,105] Some of these issues can be overcome to some extent by narrowing down the
database category, using high resolution mass spectrometers or employing efficient MS/MS
fragmentation methods (e.g. ETD or ECD for phosphopeptides). In addition, employing more
specific tools (e.g. predictors, network/pathway viewers)[45] may improve peptide sequencing
and PTM assignments.

1.6

MS-based

quantification

of

proteins

and

protein

phosphorylation
Protein phosphorylation is a reversible, highly dynamic process, and changes in the
extent of phosphorylation of specific residues in response to various stimuli may play important
roles in cellular signaling pathways.[29,44] Thus, simply exploring the existence of
phosphorylation or assigning phosphorylation sites on proteins is not sufficient to relate
phosphorylation to cell growth, differentiation, invasion and apoptosis. Traditional analysis of
phosphorylated proteins employed incorporation of radioactive

32

32

P (from [ P]ATP or

32

[ P]phosphate) into phosphoproteins either in vivo or in vitro.[106] Although extremely
sensitive, this technique involves the use of high amounts of radioactivity and does not provide
identification or quantitation of specific phosphorylation sites. Other biochemistry approaches,
including flow kinase activity assays, antibody arrays, cytometry, Western blotting and ELISA
facilitate target phosphoprotein quantification.[107,108] However, these methods reveal changes
in protein phosphorylation only at specific sites, leaving a large unexplored protein region devoid
of phosphorylation quantitation. Over the last few years, advances in MS-based quantitative
40

proteomics have led to discovery-mode phosphorylation quantitation as well as quantitation of
targeted phosphorylation sites.[41,108]
In MS, responses of peptide ions vary greatly and depend on chemical structure,
including amino acid sequence and modifications (e.g. phosphorylation and glycosylation),[41]
and the co-existence of other species during ionization.[109] Thus, relative quantitation of
peptides from two samples frequently requires encoding peptides in one or both samples with
2

stable isotopes (e.g. H,

13

C,

15

N, and

18

O). In chemical labeling, incorporation of isotopes

occurs via reaction between a mass tag and peptide amino acid residues containing amino groups,
carboxylic groups, or cysteines.[110] Reaction of two or several different samples with different
isotopes of the mass tag allows comparison of the intensities of the peptide isotopes, which are
specific to a given sample. In principle, this method applies to both proteins and peptides, but
most applications involve labelling proteolytic digests.[41] One related method, iTRAQ[111]
(isobaric tag for relative and absolute quantitation), labels the primary amines of peptides from
different sources (up to 8) with different isotopic tags that each consists of a reporter and balance
group. The derivatized peptides are indistinguishable in MS (due to the same total mass for the
tags containing the same chemical structure but encoded with different isotopes), but exhibit
low- mass MS/MS signature ions (unique reporter ion for each peptide source) that enable
quantitation.
Using iTRAQ and MS on an LTQ-Orbitrap with high energy collision dissociation, Boja
et al. recently assessed relative changes in protein phosphorylation in the mitochondria upon
physiological perturbation.[112] For example, relative quantification of phosphopeptides of the
pyruvate dehydrogenase-E1α subunit from porcine heart revealed dephosphorylation at three
serine sites by pyruvate dehydrogenase phosphatase in response to dichloroacetate or de41

2+

energization, or Ca . The advantage of iTRAQ over other chemical labeling methods
(especially those comparing precursor ions in MS spectra) lies in its ability in simultaneous
comparing up to 8 samples and providing quantitation spanning 2 orders of magnitude. However,
iTRAQ also suffers from drawbacks such as the existence of isotopic impurities and other
potential interferences in mixed MS/MS.[113] A more serious issue that faces chemical labeling
is that the comparison samples must be prepared in precisely the same way before isotope
labeling to achieve accurate relative quantitation in the final MS step. The multiple sample
preparation steps (e.g. protein purification and digestion) before isotope labeling at the peptide
level may result in quantitation errors.[114]
Metabolic labeling provides an alternative approach for relative quantitation by MS
methods.[41] In stable isotope labeling by amino acids in culture (SILAC)[114,115], one
population of cells grows in media containing normal amino acids, whereas the media for a
second population contains amino acids labeled with stable heavy isotopes.[115] The main
advantage of SILAC lies in its early-stage isotope labeling, which allows combination of
different samples immediately after cell lysis to eliminate the risk of quantitation errors during
the subsequent sample preparation steps.[115,116] The reliability of SILAC makes it a powerful
tool for assessing protein phosphorylation. For example, in a SILAC study to understand how
growth factors influence stem-cell differentiation, Mann and co-workers found that although
both epidermal growth factor (EGF) and platelet-derived growth factor induce Tyr
phosphorylation, only EGF strongly enhances differentiation to bone-forming cells.[117]
Thus far the majority of antibody-based and MS-based (e.g. using isotope labeling)
protein quantitation studies compare relative abundances of proteins or their PTMs in different
samples. Absolute quantitation of protein levels and PTM stoichiometry is even more
42

challenging. Inductively coupled plasma (ICP)-MS can reveal the total amount of P in either the
whole protein sample or single proteins after their isolation.[118] However, because it cannot
identify protein/peptide sequences, ICP-MS is only relevant to simple samples, e.g. the specific
activity of a kinase.[118] For global absolute protein quantitation, Malmström et al. recently
developed an approach that combines selected reaction monitoring measurements with a limited
number of internal reference standards, the average MS signal intensities of the most intense
three peptides per protein, and weighted MS/MS spectra counts.[119] Based on this approach,
the authors generated an absolute protein abundance scale for 83% of the MS-detectable
proteome from Leptospira (a human pathogen).
In addition to absolute protein quantities, the PTM stoichiometry of a given protein also
plays a crucial role in cellular signal transduction. In typical relative MS methods, a several fold
increase in a phosphopeptide signal could correspond to the same fold phosphorylation degree
changes in a protein with <1% of its population phosphorylated or 20% of its population
phosphorylated. The former case might be insignificant for inhibiting phosphorylation unless
turnover is rapid.[44] To address this problem, Steen et al. determined relative changes in
phosphorylation stoichiometry by monitoring the changes in the normalized ion currents of the
phosphopeptide(s) of interest.[120]

1.7 Research aims
A full understanding of relationships between phosphorylation and protein function
requires comprehensive mapping of phosphorylation sites. Previous studies suggest that either
efficient isolation of phosphopeptides from complete tryptic digests (bottom- up approach) or

43

cleaving the proteins into large peptides (middle-down approach) facilitate MS characterization
of phosphoproteins. This research aims to create improved methods for MS characterization of
target proteins (especially phosphoproteins) using both of these approaches. More specifically,
the dissertation will cover:
1. Preconcentration and purification of tryptic phosphopeptides using micro-arrays of
functional polymer brushes on MALDI plates (Chapter 2).
2. MALDI-MS and CID-MS/MS identification of p65-associated phosphoproteins after
phosphopeptide enrichment using polymer-oxotitanium films (Chapter 3).
3. Low-bias enrichment of mono- and multiply phosphorylated peptides with bifunctional
polymer brushes and tandem affinity membranes (Chapter 4).
4. Size-controlled proteolysis using trypsin- modified nylon membranes to provide large
peptides for middle-down protein identification and characterization (Chapter 5).
Chapter 6 summarizes the research and presents future directions.

44

REFERENCES

45

REFERENCES
1.
Walsh, C. T., Posttranslational Modification of Proteins: Expending Nature's Inventory.
Roberts and Company Publishers: Englewood, Colorado, USA, 2006.
2.
Johnson, L. N.; Lewis, R. J., Structural Basis for Control by Phosphorylation. Chem. Rev.
2001, 101, 2209-2242.
3.
Lee, M.; Goodbourn, S., Signalling from the Cell Surface to the Nucleus. In Regulation
of Gene Expression, Portland Press Ltd: London, 2001; Vol. 37, pp 71-85.
4.
Cohen, P., The Regulation of Protein Function by Multisite Phosphorylation - a 25 Year
Update. Trends Biochem.Sci. 2000, 25, 596-601.
5.
Pinna, L. A.; Ruzzene, M., How Do Protein Kinases Recognize Their Substrates?
Biochim. Biophys. Acta-Mol. Cell Res. 1996, 1314, 191-225.
6.
Echols, N.; Harrison, P.; Balasubramanian, S.; Luscombe, N. M.; Bertone, P.; Zhang, Z.
L.; Gerstein, M., Comprehensive Analysis of Amino Acid and Nucleotide Composition in
Eukaryotic Genomes, Comparing Genes and Pseudogenes. Nucleic Acids Res. 2002, 30, 25152523.
7.
Hunter, T.; Sefton, B. M., Transforming Gene-Product of Rous-Sarcoma Virus
Phosphorylates Tyrosine. Proc. Natl. Acad. Sci. U. S. A.-Biol. Sci. 1980, 77, 1311-1315.
8.
Olsen, J. V.; Blagoev, B.; Gnad, F.; Macek, B.; Kumar, C.; Mortensen, P.; Mann, M.,
Global, in Vivo, and Site-Specific Phosphorylation Dynamics in Signaling Networks. Cell 2006,
127, 635-648.
9.
Ubersax, J. A.; Ferrell, J. E., Mechanisms of Specificity in Protein Phosphorylation. Nat.
Rev. Mol. Cell Biol. 2007, 8, 530-541.
10.
Manning, G.; Whyte, D. B.; Martinez, R.; Hunter, T.; Sudarsanam, S., The Protein
Kinase Complement of the Human Genome. Science 2002, 298, 1912-1934.
11.
Brown, N. R.; Noble, M. E. M.; Endicott, J. A.; Johnson, L. N., The Structural Basis for
Specificity of Substrate and Recruitment Peptides for Cyclin- Dependent Kinases. Nat. Cell Biol.
1999, 1, 438-443.
12.
Hubbard, S. R., Crystal Structure of the Activated Insulin Receptor Tyrosine Kinase in
Complex with Peptide Substrate and Atp Analog. Embo J. 1997, 16, 5572-5581.
13.
Barford, D.; Hu, S. H.; Johnson, L. N., Structural Mechanism for GlycogenPhosphorylase Control by Phosphorylation and Amp. J. Mol. Biol. 1991, 218, 233-260.

46

14.
Kuriyan, J.; Cowburn, D., Modular Peptide Recognition Domains in Eukaryotic
Signaling. Annu. Rev. Biophys. Biomolec. Struct. 1997, 26, 259-288.
15.
Pawson, T.; Nash, P., Assembly of Cell Regulatory Systems through Protein Interaction
Domains. Science 2003, 300, 445-452.
16.
Yaffe, M. B.; Elia, A. E. H., Phosphoserine/Threonine- Binding Domains. Curr. Opin.
Cell Biol. 2001, 13, 131-138.
17.
Huang, C. Y. F.; Ferrell, J. E., Ultrasensitivity in the Mitogen-Activated Protein Kinase
Cascade. Proc. Natl. Acad. Sci. U. S. A. 1996, 93, 10078-10083.
18.
Kennelly, P. J., Protein Phosphatases - a Phylogenetic Perspective. Chem. Rev. 2001, 101,
2291-2312.
19.
Bauman, A. L.; Scott, J. D., Kinase- and Phosphatase-Anchoring Proteins: Harnessing the
Dynamic Duo. Nat. Cell Biol. 2002, 4, E203-E206.
20.
Han, X. M.; Aslanian, A.; Yates, J. R., Mass Spectrometry for Proteomics. Curr. Opin.
Chem. Biol. 2008, 12, 483-490.
21.
Mann, M.; Ong, S. E.; Gronborg, M.; Steen, H.; Jensen, O. N.; Pandey, A., Analysis of
Protein Phosphorylation Using Mass Spectrometry: Deciphering the Phosphoproteome. Trends
Biotechnol. 2002, 20, 261-268.
22.
Fournier, M. L.; Gilmore, J. M.; Martin- Brown, S. A.; Washburn, M. P.,
Multidimensional Separations-Based Shotgun Proteomics. Chem. Rev. 2007, 107, 3654-3686.
23.
Henzel, W. J.; Billeci, T. M.; Stults, J. T.; Wong, S. C.; Grimley, C.; Watanabe, C.,
Identifying Proteins from 2-Dimensional Gels by Molecular Mass Searching of PeptideFragments in Protein-Sequence Databases. Proc. Natl. Acad. Sci. U. S. A. 1993, 90, 5011-5015.
24.
Loo, R. R. O.; Hayes, R.; Yang, Y. N.; Hung, F.; Ramachandran, P.; Kim, N.; Gunsalus,
R.; Loo, J. A., Top-Down, Bottom- up, and Side-to-Side Proteomics with Virtual 2-D Gels. Int. J.
Mass Spectrom. 2005, 240, 317-325.
25.
Wang, W. H.; Palumbo, A. M.; Tan, Y. L.; Reid, G. E.; Tepe, J. J.; Bruening, M. L.,
Identification of p65-Associated Phosphoproteins by Mass Spectrometry after on-Plate
Phosphopeptide Enrichment Using Polymer-Oxotitanium Films. J. Proteome Res. 2010, 9, 30053015.
26.
McDonald, W. H.; Yates, J. R., Shotgun Proteomics: Integrating Technologies to Answer
Biological Questions. Curr. Opin. Mol. Ther. 2003, 5, 302-309.

47

27.
Bodenmiller, B.; Mueller, L. N.; Mueller, M.; Domon, B.; Aebersold, R., Reproducible
Isolation of Distinct, Overlapping Segments of the Phosphoproteome. Nat. Methods 2007, 4,
231-237.
28.
Delahunty, C. M.; Yates, J. R., Mudpit: Multidimensional Protein Identification
Technology. Biotechniques 2007, 43, 563.
29.
Thingholm, T. E.; Jensen, O. N.; Larsen, M. R., Analytical Strategies for
Phosphoproteomics. Proteomics 2009, 9, 1451-1468.
30.
Hoang, T.; Roth, U.; Kowalewski, K.; Belisle, C.; Steinert, K.; Karas, M., Highly
Specific Capture and Direct MALDI MS Analysis of Phosphopeptides by Zirconium
Phosphonate on Self-Assembled Monolayers. Anal. Chem. 2010, 82, 219-228.
31.
Ge, Y.; Rybakova, I. N.; Xu, Q. G.; Moss, R. L., Top-Down High- Resolution Mass
Spectrometry of Cardiac Myosin Binding Protein C Revealed That Truncation Alters Protein
Phosphorylation State. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 12658-12663.
32.
Zhao, Y. M.; Jensen, O. N., Modification-Specific Proteomics: Strategies for
Characterization of Post-Translational Modifications Using Enrichment Techniques. Proteomics
2009, 9, 4632-4641.
33.
Villen, J.; Gygi, S. P., The Scx/IMAC Enrichment Approach for Global Phosphorylation
Analysis by Mass Spectrometry. Nat. Protoc. 2008, 3, 1630-1638.
34.
McLafferty, F. W.; Breuker, K.; Jin, M.; Han, X. M.; Infusini, G.; Jiang, H.; Kong, X. L.;
Begley, T. P., Top-Down MS, a Powerful Complement to the High Capabilities of Proteolysis
Proteomics. Febs J. 2007, 274, 6256-6268.
35.
Garcia, B. A.; Pesavento, J. J.; Mizzen, C. A.; Kelleher, N. L., Pervasive Combinatorial
Modification of Histone H3 in Human Cells. Nat. Methods 2007, 4, 487-489.
36.
Garcia, B. A.; Siuti, N.; Thomas, C. E.; Mizzen, C. A.; Kelleher, N. L., Characterizat ion
of Neurohistone Variants and Post-Translational Modifications by Electron Capture Dissociation
Mass Spectrometry. Int. J. Mass Spectrom. 2007, 259, 184-196.
37.
Roth, M. J.; Forbes, A. J.; Boyne, M. T.; Kim, Y. B.; Robinson, D. E.; Kelleher, N. L.,
Precise and Parallel Characterization of Coding Polymorphisms, Alternative Splicing, and
Modifications in Human Proteins by Mass Spectrometry. Mol. Cell. Proteomics 2005, 4, 10021008.
38.
Whitelegge, J.; Halgand, F.; Souda, P.; Zabrouskov, V., Top-Down Mass Spectrometry
of Integral Membrane Proteins. Expert Rev. Proteomics 2006, 3, 585-596.
39.
Siuti, N.; Kelleher, N. L., Decoding Protein Modifications Using Top-Down Mass
Spectrometry. Nat. Methods 2007, 4, 817-821.
48

40.
Young, N. L.; Plazas-Mayorca, M. D.; Garcia, B. A., Systems-Wide Proteomic
Characterization of Combinatorial Post-Translational Modification Patterns. Expert Rev.
Proteomics 2010, 7, 79-92.
41.
Macek, B.; Mann, M.; Olsen, J. V., Global and Site-Specific Quantitative
Phosphoproteomics: Principles and Applications. Annu. Rev. Pharmacol. Toxicol. 2009, 49, 199221.
42.
Dunn, J. D.; Reid, G. E.; Bruening, M. L., Techniques for Phosphopeptide Enrichment
Prior to Analysis by Mass Spectrometry. Mass Spectrom. Rev. 2010, 29, 29-54.
43.
Paradela, A.; Albar, J. P., Advances in the Analysis of Protein Phosphorylation. J.
Proteome Res. 2008, 7, 1809-1818.
44.
Ashman, K.; Villar, E. L., Phosphoproteomics and Cancer Research. Clin. Transl. Oncol.
2009, 11, 356-362.
45.
Rogers, L. D.; Foster, L. J., Phosphoproteomics-Finally Fulfilling the Promise? Mol.
Biosyst. 2009, 5, 1122-1129.
46.
Andersson, L.; Porath, J., Isolation of Phosphoproteins by Immobilized Metal (Fe-3+)
Affinity-Chromatography. Anal. Biochem. 1986, 154, 250-254.
47.
Thingholm, T. E.; Jensen, O. N.; Robinson, P. J.; Larsen, M. R., Simac (Sequential
Elution from IMAC), a Phosphoproteomics Strategy for the Rapid Separation of
Monophosphorylated from Multiply Phosphorylated Peptides. Mol. Cell. Proteomics 2008, 7,
661-671.
48.
Trinidad, J. C.; Specht, C. G.; Thalhammer, A.; Schoepfer, R.; Burlingame, A. L.,
Comprehensive Identification of Phosphorylation Sites in Postsynaptic Density Preparations.
Mol. Cell. Proteomics 2006, 5, 914-922.
49.
Villen, J.; Beausoleil, S. A.; Gerber, S. A.; Gygi, S. P., Large-Scale Phosphorylation
Analysis of Mouse Liver. Proc. Natl. Acad. Sci. U. S. A. 2007, 104, 1488-1493.
50.
Zhai, B.; Villen, J.; Beausoleil, S. A.; Mintseris, J.; Gygi, S. P., Phosphoproteome
Analysis of Drosophila Metanogaster Embryos. J. Proteome Res. 2008, 7, 1675-1682.
51.
Ficarro, S. B.; McCleland, M. L.; Stukenberg, P. T.; Burke, D. J.; Ross, M. M.;
Shabanowitz, J.; Hunt, D. F.; White, F. M., Phosphoproteome Analysis by Mass Spectrometry
and Its Application to Saccharomyces Cerevisiae. Nat. Biotechnol. 2002, 20, 301-305.
52.
Kokubu, M.; Ishihama, Y.; Sato, T.; Nagasu, T.; Oda, Y., Specificity of Immobilized
Metal Affinity- Based IMAC/C18 Tip Enrichment of Phosphopeptides for Protein
Phosphorylation Analysis. Anal. Chem. 2005, 77, 5144-5154.
49

53.
Tsai, C. F.; Wang, Y. T.; Chen, Y. R.; Lai, C. Y.; Lin, P. Y.; Pan, K. T.; Chen, J. Y.;
Khoo, K. H.; Chen, Y. J., Immobilized Metal Affinity Chromatography Revisited: Ph/Acid
Control toward High Selectivity in Phosphoproteomics. J. Proteome Res. 2008, 7, 4058-4069.
54.
Ye, J. Y.; Zhang, X. M.; Young, C.; Zhao, X. L.; Hao, Q.; Cheng, L.; Jensen, O. N.,
Optimized IMAC-IMAC Protocol for Phosphopeptide Recovery from Complex Biological
Samples. J. Proteome Res. 2010, 9, 3561-3573.
55.
Pinkse, M. W. H.; Mohammed, S.; Gouw, L. W.; van Breukelen, B.; Vos, H. R.; Heck, A.
J. R., Highly Robust, Automated, and Sensitive on Line Tio2-Based Phosphoproteomics Applied
to Study Endogenous Phosphorylation in Drosophila Melanogaster. J. Proteome Res. 2008, 7,
687-697.
56.
Li, Q. R.; Ning, Z. B.; Tang, J. S.; Nie, S.; Zeng, R., Effect of Peptide-to-Tio2 Beads
Ratio on Phosphopeptide Enrichment Selectivity. J. Proteome Res. 2009, 8, 5375-5381.
57.
Simon, E. S.; Young, M.; Chan, A.; Bao, Z. Q.; Andrews, P. C., Impro ved Enrichment
Strategies for Phosphorylated Peptides on Titanium Dioxide Using Methyl Esterification and Ph
Gradient Elution. Anal. Biochem. 2008, 377, 234-242.
58.
Zhou, H. J.; Ye, M. L.; Dong, J.; Han, G. H.; Jiang, X. N.; Wu, R. N.; Zou, H. F.,
Specific Phosphopeptide Enrichment with Immobilized Titanium Ion Affinity Chromatography
Adsorbent for Phosphoproteome Analysis. J. Proteome Res. 2008, 7, 3957-3967.
59.
Wan, J. J.; Qian, K.; Qiao, L.; Wang, Y. H.; Kong, J. L.; Yang, P. Y.; Liu, B. H.; Yu, C.
Z., Tio2-Modified Macroporous Silica Foams for Advanced Enrichment of MultiPhosphorylated Peptides. Chem.-Eur. J. 2009, 15, 2504-2508.
60.
Lin, B.; Li, T.; Zhao, Y.; Huang, F. K.; Guo, L.; Feng, Y. Q., Preparation of a Tio2
Nanoparticle-Deposited Capillary Column by Liquid Phase Deposition and Its Application in
Phosphopeptide Analysis. J. Chromatogr. A 2008, 1192, 95-102.
61.
Lu, S.; Luo, Q.; Li, X. C.; Wu, J. H.; Liu, J. A.; Xiong, S. X.; Feng, Y. Q.; Wang, F. Y.,
Inhibitor Screening of Protein Kinases Using MALDI- Tof MS Combined with Separation and
Enrichment of Phosphopeptides by Tio2 Nanoparticle Deposited Capillary Column. Analyst
2010, 135, 2858-2863.
62.
Qin, W. J.; Zhang, W. J.; Song, L. N.; Zhang, Y. J.; Qian, X. H., Surface Initiated Atom
Transfer Radical Polymerization: Access to Three Dimensional Wavelike Polymer Structure
Modified Capillary Columns for Online Phosphopeptide Enrichment. Anal. Chem. 2010, 82,
9461-9468.
63.
Xue, Y. F.; Wei, J. Y.; Han, H. H.; Zhao, L. Y.; Cao, D.; Wang, J. L.; Yang, X. M.;
Zhang, Y. J.; Qian, X. H., Application of Open Tubular Capillary Columns Coated with

50

Zirconium Phosphonate for Enrichment of Phosphopeptides. J. Chromatogr. B 2009, 877, 757764.
64.
Hou, C. Y.; Ma, J. F.; Tao, D. Y.; Shan, Y. C.; Liang, Z.; Zhang, L. H.; Zhang, Y. K.,
Organic-Inorganic Hybrid Silica Monolith Based Immobilized Titanium Ion Affinity
Chromatography Column for Analysis of Mitochondrial Phosphoproteome. J. Proteome Res.
2010, 9, 4093-4101.
65.
Beausoleil, S. A.; Jedrychowski, M.; Schwartz, D.; Elias, J. E.; Villen, J.; Li, J. X.; Cohn,
M. A.; Cantley, L. C.; Gygi, S. P., Large-Scale Characterization of Hela Cell Nuclear
Phosphoproteins. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 12130-12135.
66.
Nie, S.; Dai, J.; Ning, Z. B.; Cao, X. J.; Sheng, Q. H.; Zeng, R., Comprehensive Profiling
of Phosphopeptides Based on Anion Exchange Followed by Flow-through Enrichment with
Titanium Dioxide (Afet). J. Proteome Res. 2010, 9, 4585-4594.
67.
Han, G. H.; Ye, M. L.; Zhou, H. J.; Jiang, X. N.; Feng, S.; Jiang, X. G.; Tian, R. J.; Wan,
D. F.; Zou, H. F.; Gu, J. R., Large-Scale Phosphoproteome Analysis of Human Liver Tissue by
Enrichment and Fractionation of Phosphopeptides with Strong Anion Exchange Chromatography.
Proteomics 2008, 8, 1346-1361.
68.
Motoyama, A.; Xu, T.; Ruse, C. I.; Wohlschlegel, J. A.; Yates, J. R., Anion and Cation
Mixed-Bed Ion Exchange for Enhanced Multidimensional Separations of Peptides and
Phosphopeptides. Anal. Chem. 2007, 79, 3623-3634.
69.
Dai, J.; Jin, W. H.; Sheng, Q. H.; Shieh, C. H.; Wu, J. R.; Zeng, R., Protein
Phosphorylation and Expression Profiling by Yin-Yang Multidimensional Liquid
Chromatography (Yin-Yang Mdlc) Mass Spectrometry. J. Proteome Res. 2007, 6, 250-262.
70.
Jain, P.; Baker, G. L.; Bruening, M. L., Applications of Polymer Brushes in Protein
Analysis and Purification. Annu. Rev. Anal. Chem. 2009, 2, 387-408.
71.
Blacken, G. R.; Volny, M.; Diener, M.; Jackson, K. E.; Ranjitkar, P.; Maly, D. J.;
Turecek, F., Reactive Landing of Gas-Phase Ions as a Tool for the Fabrication of Metal Oxide
Surfaces for in Situ Phosphopeptide Enrichment. J. Am. Soc. Mass Spectrom. 2009, 20, 915-926.
72.
Niklew, M. L.; Hochkirch, U.; Melikyan, A.; Moritz, T.; Kurzawski, S.; Schluter, H.;
Ebner, I.; Linscheid, M. W., Phosphopeptide Screening Using Nanocrystalline Titanium Dioxide
Films as Affinity Matrix-Assisted Laser Desorption Ionization Targets in Mass Spectrometry.
Anal. Chem. 2010, 82, 1047-1053.
73.
Cannon, J.; Lohnes, K.; Wynne, C.; Wang, Y.; Edwards, N.; Fenselau, C., HighThroughput Middle-Down Analysis Using an Orbitrap. J. Proteome Res. 2010, 9, 3886-3890.
74.
Wu, S. L.; Kim, J.; Hancock, W. S.; Karger, B., Extended Range Proteomic Analysis
(ERPA): A New and Sensitive LC-MS Platform for High Sequence Coverage of Complex
51

Proteins with Extensive Post-Translational Modifications-Comprehensive Analysis of BetaCasein and Epidermal Growth Factor Receptor (Egfr). J. Proteome Res. 2005, 4, 1155-1170.
75.
Coon, J. J.; Ueberheide, B.; Syka, J. E. P.; Dryhurst, D. D.; Ausio, J.; Shabanowitz, J.;
Hunt, D. F., Protein Identification Using Sequential Ion/Ion Reactions and Tandem Mass
Spectrometry. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 9463-9468.
76.
Garcia, B. A.; Thomas, C. E.; Kelleher, N. L.; Mizzen, C. A., Tissue-Specific Expression
and Post-Translational Modification of Histone H3 Variants. J. Proteome Res. 2008, 7, 42254236.
77.
Xu, P.; Peng, J. M., Characterization of Polyubiquitin Chain Structure by Middle-Down
Mass Spectrometry. Anal. Chem. 2008, 80, 3438-3444.
78.
She, Y. M.; Xu, X. A. H.; Yakunin, A. F.; Dhe-Paganon, S.; Donald, L. J.; Standing, K.
G.; Lee, D. C.; Jia, Z. C.; Cyr, T. D., Mass Spectrometry Following Mild Enzymatic Digestion
Reveals Phosphorylation of Recombinant Proteins in Escherichia Co il through Mechanisms
Involving Direct Nucleotide Binding. J. Proteome Res. 2010, 9, 3311-3318.
79.
Wu, S. L.; Huehmer, A. F. R.; Hao, Z. Q.; Karger, B. L., On-Line LC-MS Approach
Combining Collision-Induced Dissociation (CID), Electron-Transfer Dissociation (Etd), and CID
of an Isolated Charge-Reduced Species for the Trace-Level Characterization of Proteins with
Post-Translational Modifications. J. Proteome Res. 2007, 6, 4230-4244.
80.
Hamuro, Y.; Coales, S. J.; Molnar, K. S.; Tuske, S. J.; Morrow, J. A., Specificity of
Immobilized Porcine Pepsin in H/D Exchange Compatible Conditions. Rapid Commun. Mass
Spectrom. 2008, 22, 1041-1046.
81.
Appel, W., Chymotrypsin - Molecular and Catalytic Properties. Clin. Biochem. 1986, 19,
317-322.
82.
Lee, T. D.; Shively, J. E., Enzymatic and Chemical Digestion of Proteins for MassSpectrometry. Method Enzymol. 1990, 193, 361-374.
83.
Birktoft, J. J.; Breddam, K., Glutamyl Endopeptidases. In Proteolytic Enzymes: Serine
and Cysteine Peptidases, Academic Press Inc: San Diego, 1994; Vol. 244, pp 114-126.
84.
Raijmakers, R.; Neerincx, P.; Mohammed, S.; Heck, A. J. R., Cleavage Specificities of
the Brother and Sister Proteases Lys-C and Lys-N. Chem. Commun. 2010, 46, 8827-8829.
85.
Drapeau, G. R., Substrate-Specificity of a Proteolytic-Enzyme Isolated from a Mutant of
Pseudomonas-Fragi. J. Biol. Chem. 1980, 255, 839-840.
86.
Li, A. Q.; Sowder, R. C.; Henderson, L. E.; Moore, S. P.; Garfinkel, D. J.; Fisher, R. J.,
Chemical Cleavage at Aspartyl Residues for Protein Identification. Anal. Chem. 2001, 73, 53955402.
52

87.
Partridge, S. M.; Davis, H. F., Preferential Release of Aspartic Acid During the
Hydrolysis of Proteins. Nature 1950, 165, 62-63.
88.

Inglis, A. S., Cleavage at Aspartic-Acid. Method Enzymol. 1983, 91, 324-332.

89.
Swatkoski, S.; Russell, S.; Edwards, N.; Fenselau, C., Analysis of a Model Virus Using
Residue-Specific Chemical Cleavage and MALDI-Tof Mass Spectrometry. Anal. Chem. 2007,
79, 654-658.
90.
Swatkoski, S.; Russell, S. C.; Edwards, N.; Fenselau, C., Rapid Chemical Digestion of
Small Acid-Soluble Spore Proteins for Analysis of Bacillus Spores. Anal. Chem. 2006, 78, 181188.
91.
Swatkoski, S.; Gutierrez, P.; Wynne, C.; Petrov, A.; Dinman, J. D.; Edwards, N.;
Fenselau, C., Evaluation of Microwave-Accelerated Residue-Specific Acid Cleavage for
Proteomic Applications. J. Proteome Res. 2008, 7, 579-586.
92.
Hauser, N. J.; Basile, F., Online Microwave D-Cleavage Lc-ESI-MS/MS of Intact
Proteins: Site-Specific Cleavages at Aspartic Acid Residues and Disulfide Bonds. J. Proteome
Res. 2008, 7, 1012-1026.
93.
Li, J. X.; Shefcheck, K.; Callahan, J.; Fenselau, C., Extension of Microwave-Accelerated
Residue-Specific Acid Cleavage to Proteins with Carbohydrate Side Chains and Disulfide
Linkages. Int. J. Mass Spectrom. 2008, 278, 109-113.
94.
Huang, H. V.; Bond, M. W.; Hunkapiller, M. W.; Hood, L. E., Cleavage at Tryptophanyl
Residues with Dimethylsulfoxide Hydrochloric-Acid and Cyanogen-Bromide. Method Enzymol.
1983, 91, 318-324.
95.
Vestling, M. M.; Kelly, M. A.; Fenselau, C., Optimization by Mass-Spectrometry of a
Tryptophan-Specific Protein Cleavage Reaction. Rapid Commun. Mass Spectrom. 1994, 8, 786790.
96.
Walcher, W.; Timperio, A. M.; Zolla, L.; Huber, C. G., Characterization of a Variant of
the Spinach Psii Type I Light-Harvesting Protein Using Kinetically Controlled Digestion and
Rp-Hplc-ESI-MS. Anal. Chem. 2003, 75, 6775-6780.
97.
Bi, H. Y.; Qiao, L.; Busnel, J. M.; Liu, B. H.; Girault, H. H., Kinetics of Proteolytic
Reactions in Nanoporous Materials. J. Proteome Res. 2009, 8, 4685-4692.
98.
Srividhya, J.; Schnell, S., Why Substrate Depletion Has Apparent First-Order Kinetics in
Enzymatic Digestion. Comput. Biol. Chem. 2006, 30, 209-214.

53

99.
Krenkova, J.; Svec, F., Less Common Applications of Monoliths: Iv. Recent
Developments in Immobilized Enzyme Reactors for Proteomics and Biotechnology. J. Sep. Sci.
2009, 32, 706-718.
100. Schroeder, M. J.; Shabanowitz, J.; Schwartz, J. C.; Hunt, D. F.; Coon, J. J., A Neutral
Loss Activation Method for Improved Phosphopeptide Sequence Analysis by Quadrupole Ion
Trap Mass Spectrometry. Anal. Chem. 2004, 76, 3590-3598.
101. Zubarev, R. A.; Kelleher, N. L.; McLafferty, F. W., Electron Capture Dissociation of
Multiply Charged Protein Cations. A Nonergodic Process. J. Am. Chem. Soc. 1998, 120, 32653266.
102. Syka, J. E. P.; Coon, J. J.; Schroeder, M. J.; Shabanowitz, J.; Hunt, D. F., Peptide and
Protein Sequence Analysis by Electron Transfer Dissociation Mass Spectrometry. Proc. Natl.
Acad. Sci. U. S. A. 2004, 101, 9528-9533.
103. McAlister, G. C.; Phanstiel, D.; Good, D. M.; Berggren, W. T.; Coon, J. J.,
Implementation of Electron-Transfer Dissociation on a Hybrid Linear Ion Trap-Orbitrap Mass
Spectrometer. Anal. Chem. 2007, 79, 3525-3534.
104. Nesvizhskii, A. I.; Vitek, O.; Aebersold, R., Analysis and Validation of Proteomic Data
Generated by Tandem Mass Spectrometry. Nat. Methods 2007, 4, 787-797.
105. Ahrne, E.; Muller, M.; Lisacek, F., Unrestricted Identification of Modified Proteins Using
MS/MS. Proteomics 2010, 10, 671-686.
106. Yan, J. X.; Packer, N. H.; Gooley, A. A.; Williams, K. L., Protein Phosphorylation:
Technologies for the Identification of Phosphoamino Acids. J. Chromatogr. A 1998, 808, 23-41.
107. Pelech, S.; Jelinkova, L.; Susor, A.; Zhang, H.; Shi, X. Q.; Pavlok, A.; Kubelka, M.;
Kovarova, H., Antibody Microarray Analyses of Signal Transduction Protein Expression and
Phosphorylation During Porcine Oocyte Maturation. J. Proteome Res. 2008, 7, 2860-2871.
108. White, F. M., Quantitative Phosphoproteomic Analysis of Signaling Network Dynamics.
Curr. Opin. Biotechnol. 2008, 19, 404-409.
109. Annesley, T. M., Ion Suppression in Mass Spectrometry. Clin. Chem. 2003, 49, 10411044.
110. Julka, S.; Regnier, F., Quantification in Proteomics through Stable Iso tope Coding: A
Review. J. Proteome Res. 2004, 3, 350-363.
111. Ross, P. L.; Huang, Y. L. N.; Marchese, J. N.; Williamson, B.; Parker, K.; Hattan, S.;
Khainovski, N.; Pillai, S.; Dey, S.; Daniels, S.; Purkayastha, S.; Juhasz, P.; Martin, S.; BartletJones, M.; He, F.; Jacobson, A.; Pappin, D. J., Multiplexed Protein Quantitation in

54

Saccharomyces Cerevisiae Using Amine-Reactive Isobaric Tagging Reagents. Mol. Cell.
Proteomics 2004, 3, 1154-1169.
112. Boja, E. S.; Phillips, D.; French, S. A.; Harris, R. A.; Balaban, R. S., Quantitative
Mitochondrial Phosphoproteomics Using Itraq on an Ltq-Orbitrap with High Energy Collision
Dissociation. J. Proteome Res. 2009, 8, 4665-4675.
113. Ow, S. Y.; Salim, M.; Noirel, J.; Evans, C.; Rehman, I.; Wright, P. C., Itraq
Underestimation in Simple and Complex Mixtures: "The Good, the Bad and the Ugly". J.
Proteome Res. 2009, 8, 5347-5355.
114. Mann, M., Functional and Quantitative Proteomics Using Silac. Nat. Rev. Mol. Cell Biol.
2006, 7, 952-958.
115. Ong, S. E.; Blagoev, B.; Kratchmarova, I.; Kristensen, D. B.; Steen, H.; Pandey, A.;
Mann, M., Stable Isotope Labeling by Amino Acids in Cell Culture, Silac, as a Simple and
Accurate Approach to Expression Proteomics. Mol. Cell. Proteomics 2002, 1, 376-386.
116. Schreiber, T. B.; Mausbacher, N.; Breitkopf, S. B.; Grundner-Culemann, K.; Daub, H.,
Quantitative Phosphoproteomics - an Emerging Key Technology in Signal- Transduction
Research. Proteomics 2008, 8, 4416-4432.
117. Kratchmarova, I.; Blagoev, B.; Haack-Sorensen, M.; Kassem, M.; Mann, M., Mechanism
of Divergent Growth Factor Effects in Mesenchymal Stem Cell Differentiation. Science 2005,
308, 1472-1477.
118. Navaza, A. P.; Encinar, J. R.; Sanz-Medel, A., Quantitative Protein Phosphorylation
Analysis: The Role of Icp-MS. J. Anal. At. Spectrom. 2007, 22, 1223-1237.
119. Malmstrom, J.; Beck, M.; Schmidt, A.; Lange, V.; Deutsch, E. W.; Aebersold, R.,
Proteome-Wide Cellular Protein Concentrations of the Human Pathogen Leptospira Interrogans.
Nature 2009, 460, 762-U112.
120. Steen, H.; Jebanathirajah, J. A.; Springer, M.; Kirschner, M. W., Stable Isotope-Free
Relative and Absolute Quantitation of Protein Phosphorylation Stoichiometry by MS. Proc. Natl.
Acad. Sci. U. S. A. 2005, 102, 3948-3953.

55

Chapter 2. Phosphopeptide Enrichment on Functionalized
Polymer Microspots for MALDI-MS Analysis
(The results described in Chapter 2 have been published in: W.H. Wang, M.L. Bruening,
Phosphopeptide enrichment on functionalized polymer microspots for MALDI-MS analysis.
Analyst, 2009, 134, 512–518.)

2.1 Introduction
Mass spectrometry (MS) is the preferred technique for detecting phosphorylation sites on
hydroxylated amino acid residues (Ser, Thr, and Tyr).[1,2] However, the stoichiometry of
phosphorylation is relatively low within cells,[3] so phosphopeptide MS signals are often weak
because of ion suppression in the presence of salts and nonphosphopeptides.[4] Selective
enrichment of phosphopeptides in phosphoprotein digests provides one way to enhance
phosphopeptide MS signals, and enrichment strategies include covalent binding of
phosphopeptides to surfaces[5-7] and selective adsorption of phosphopeptides on immob ilized
metal ion complexes[8-10] or metal oxides.[11,12] The latter two approaches were applied
directly on a matrix-assisted laser desorption/ionization (MALDI) plate prior to MALDI-MS
analysis.[8,10,12] These on-plate enrichment strategies integrate binding of phosphopeptides,
rinsing to remove impurities, and matrix/eluent addition on a single plate to simplify sample
handling, minimize sample loss, and facilitate high-throughput analysis.
Recently, our group developed a strategy for polymer brush-based on-plate enrichment of
phosphopeptides.[13] We grow poly(2- hydroxyethyl methacrylate) (PHEMA) brushes from a
sample probe by atom transfer radical polymerization (ATRP ) and then derivatize these brushes

56

with a high density of nitrilotriacetate-Fe(III) (NTA-Fe(III)) complexes that selectively bind
phosphopeptides. The high binding capacity of the brushes decreases the detection limits for onprobe purification with MALDI-MS analysis compared to enrichment on monolayer and thin
polymer films.[13]
This chapter describes a method for forming PHEMA-NTA-Fe(III) microspots to
selectively concentrate phosphopeptides into small areas, thereby further increasing the
sensitivity of phosphopeptide detection. The strategy for sample concentration (right side of
Scheme 1) includes pinning a sample droplet on a 0.25-mm spot of polymer brush, allowing the
droplet to evaporate to the spot size, rinsing to remove unwanted compounds, and addition of
matrix to elute phosphopeptides and allow MALDI-MS analysis. This method restricts the
2

sample size to ~0.05 mm , which approaches the cross-sectional area of the MALDI laser. Thus
a much larger fraction of the sample is available for ionization upon irradiation than when using
2

conventional MALDI wells that have diameters of 2-4 mm (areas of 3-13 mm ). Similar
strategies were employed with MALDI plates modified with hydrophobic surfaces such as
paraffin wax[14,15] and Teflon.[16,17] However, those methods either concentrate all the
species in a small sample droplet on the plate, or simply remove salts from the sample without
any selectivity for enriching a particular class of biomacromolecules.[14-19] Phosphopeptide
enrichment requires small spots that realize both pre-concentration and selective binding.
In previous work on enrichment of peptides and DNA,[20,21] our group used a microcontact printing/film growth technique[22] to create small (~0.2 mm), hydrophilic spots of
poly(acrylic acid) surrounded by a pre-patterned hydrophobic monolayer. When we attempted to
use this method for synthesizing small spots of PHEMA brushes, however, the pre-patterned
monolayer partially lost its hydrophobicity due to the adsorption of PHEMA or monomer, and
57

Scheme 2.1. Fabrication of micro-arrays of PHEMA-NTA-Fe(III) brushes surrounded by hydrophobic poly(dimethylsiloxane)
(PDMS), and on-plate enrichment of phosphopeptides prior to MALDI-MS analysis. DHB stands for 2,5-dihydroxybenzoic acid.
58

this resulted in inefficient sample pre-concentration during enrichment.
This chapter presents a simple technique for transferring a patterned, hydrophobic
poly(dimethylsiloxane) (PDMS) layer onto a Si wafer by heating of a PDMS stamp that is in
contact with the Si (Scheme 1). The patterned stamp contains micropores (~250 µ diameter)
m
that expose Si for growth of thick polymer brushes. After brush formation, removal of the bulk
PDMS stamp leaves behind a pristine PDMS layer that is hydrophobic and capable of repelling
sample droplets during enrichment. Isolation of phosphopeptides in the small area of the
hydrophilic, polymer brush microspots greatly enhances MALDI-MS signals of low levels of
phosphopeptides. Detection limits are typically 5-fold lower than those obtained with enrichment
on a 2 mm-diameter spot of polymer brushes on a non-patterned substrate.

2.2 Experimental
2.2.1 Chemicals and materials
Most chemicals and materials employed in this work were obtained from Sigma-Aldrich.
Other chemicals include succinic anhydride

(Matheson Coleman & Bell), N ,N-

bis(carboxymethyl)-L- lysine hydrate (aminobutyl NTA, Fluka), anhydrous ferric chloride (FeCl3,
Spectrum), Tris-HCl (Invitrogen), urea (J.T.Baker), DL-dithiothreitol (Fluka), acetonitrile (ACN,
Mallinckrodt Baker), ammonium bicarbonate (Columbus Chemical) and sequencing grade
modified porcine trypsin (treated with L-1-tosylamide-2-phenylethyl chloromethyl ketone,
Promega). N,N-Dimethylformamide (DMF, Jade Scientific) was newly distilled before use, and
deionized water was prepared with a Millipore ion-exchange system (Milli-Q, 18 MΩcm).

59

2.2.2 Fabrication of functional polymer brushes on Si wafers
The left of Scheme 1 shows the steps in fabricating microspots of polymer brushes on a
Si wafer. To prepare the thin PDMS stamp, ~0.5 g of PDMS elastomer solution (Sylgard 184,
Dow Corning) with a 10:1 polymer/curing agent ratio was poured onto a flat 8.5 cm-diameter
sterile polystyrene petri dish (Fisherbrand) and dispersed such that it covered the dish uniformly.
This solution was then cured in air at 50 ° for 2 h to form a thin (<0.1 mm) PDMS stamp,
C
which was cut into 1015 mm pieces. Individual pieces were peeled off the petri dish and placed
on 5- mm thick, pre-cured PDMS prior to punching 250 m-diameter holes through the thin
PDMS stamp and the thick PDMS support with a blunt dispensing needle (i.d. 0.23-0.25 mm,
Part No. ZT-5-026-5-X, ZEPHYRTRONICS). The thin stamp was then peeled off the thick
PDMS and gently attached to a 1520 mm piece of Si wafer that was previously sonicated in
acetone and dried with N 2. After curing the PDMS on the Si wafer in air at 75 ° for 1 h and
C
subsequent UV/ozone treatment for 15 min, initiator was attached to exposed Si by immersing
the stamp-covered wafer overnight in an anhydrous DMF solution containing 7.4 mg/mL (16
mM) 11-(2-bromo-2-methyl)propionyloxyundecenyltrichlorosilane, which was synthesized
according to reference literature procedure.[23] Synthesis of PHEMA brushes on the exposed,
initiator-modified Si spots through CuCl/CuBr2/2,2‟-bipyridine catalyzed ATRP (1 h), and
derivatization of the polymer with NTA-Fe(III) were performed as previously reported by our
group.[13] Finally, the thin PDMS stamp was gently peeled off the Si wafer to leave a thin,
residual PDMS coating surrounding the microspots.
For enrichment on a 2-mm polymer spot, the entire surface of a Si wafer was coated with
polymer brushes in the same way as above, but no PDMS stamp was attached to the Si. To

60

restrict sample droplets to a diameter of ~2 mm, sample wells were made by lightly scratching 2
mm-diameter circles onto the wafer using a tungsten carbide-tipped pen and a teflon mask with
arrays of 2 mm-diameter holes.

2.2.3 Polymer brush characterization
The thicknesses of polymer films on Si wafers were determined using a rotating a nalyzer
spectroscopic ellipsometer (J. A. Woollam, M-44), assuming a film refractive index of 1.5.
Transmission FTIR spectra of films on double-side polished Si wafers were collected using a
Mattson Galaxy Series 3000 instrument and an air background, and static water contact angles
were measured using a First Ten Ångstroms 200 contact angle analyzer. Field-emission scanning
electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDS) mapping were
performed with a Hitachi S-4700II microscope using an acceleration voltage of 15 kV.
To quantify Fe(III) ions in polymer brushes, five single-side polished silicon wafers (2.4
× 1.1 cm dimensions) coated with PHEMA-NTA-Fe(III) brushes (average thickness 78 nm) were
sequentially immersed in the same 7.0 mL of 50 mM EDTA (pH 7.2) for 5 h with stirring (one
wafer was taken out before the new one was put in). The absorbance of Fe(III) in the stripping
solution was determined using a Varian Spectra AA-200 atomic absorption spectrophotometer,
and the amount of Fe(III) was calculated from this absorbance using a calibration curve, which
was determined by measuring the absorbance of FeCl3 standard solutions in 50 mM EDTA.

2.2.4 Sample preparation and MS analysis
To digest proteins into peptides, 100 μg of each protein (β-casein, ovalbumin and bovine
serum albumin (BSA)) were dissolved in 20 μL of 6 M urea, 50 mM Tris-HCl. After addition of
5 μL of 10 mM DL-dithiothreitol (for disulfide bond cleavage) the solution was heated in a water
61

bath at 60 °C for 1 h, followed by dilution with 160 μL of 50 mM ammonium bicarbonate and 10
μL of 100 mM iodoacetamide. The solution was then placed in the dark for 1 h for cysteine
carbamidomethylation. Subsequently, the solution was mixed with 10 μL of 0.5 μg/μL trypsin
prior to incubation at 37 ˚C for 16 h. To quench the tryptic digestion, 11 μL of glacial acetic acid
was finally added to each solution. These digests were dried with a SpeedVac and reconstituted
in 5% acetic acid (HOAc). Digests mixtures were prepared by mixing different volumes of each
digest, and the amounts of salts and digest reagents decrease with decreasing total digest
concentration.
For microspot enrichment of protein digests, 0.5 μL of digest was first spo tted onto the
microspot and allowed to evaporate (and shrink) for ~5 min. The Si wafer was then placed in a
covered, water-saturated petri dish for further incubation for 1 h. (The droplet showed minimal
evaporation in the humidified petri dish.) For enrichment on a 2-mm polymer spot, 0.5 μL of
protein digest was spotted on the 2- mm well and incubated for 1 h. As the sample droplet
evaporated quickly on the 2-mm spot, 0.5 μL aliquots of 5% HOAc were periodically added to
the droplet to compensate for evaporation. The micro- and 2- mm spots were then rinsed with ~30
mL of 3:30:67 HOAc:ACN:water (either from a wash bottle or by a 3- min sonication after an
initial wash-bottle rinse) to remove salts and unbound nonphosphorylated peptides, and the plate
was immediately dried with N 2 to remove any remaining rinse solution. After incubation, rinsing,
and drying, a 0.2-μL droplet of 5 mg/mL 2,5-dihydroxybenzoic acid (DHB, Aldrich, prepared in
1:9 ACN:1% H3PO4) was gently added to the microspot to form matrix crystals covering an area
similar in size to the original microspot. For the 2- mm polymer spot, a 1-μL droplet of 1%
phosphoric acid was added to the well immediately followed by 0.25 μL of 40 mg/mL DHB
(prepared in 1:1 ACN:1% phosphoric acid). The spot of matrix crystals on the non-patterned
62

brushes had a diameter of ~2 mm. Using double-sided tape, the Si wafer was then secured to a
modified stainless steel MALDI sample plate for subsequent MALDI-MS analysis. For
conventional MALDI-MS analysis (no enrichment), 0.5 μL of sample was added to the stainless
steel plate, followed by addition of 1 μL of 10 mg/mL DHB in 1:1 ACN:1% phosphoric acid.
Mass spectra were obtained using a MALDI linear ion trap mass spectrometer (Thermo
model vMALDI LTQ XL, San Jose, CA) in positive ion mode. For each experiment, mass
spectra were obtained from at least 3 spots (micro or 2 mm), and the mass spectra given here are
representative examples.

Protein Prospector (http://prospector.ucsf.edu) was used to make

preliminary assignments to intact precursor ion signals in the mass spectra, and assignments were
confirmed using low energy collision- induced dissociation-tandem mass spectrometry (CIDMS/MS) with standard precursor ion isolation and activation conditions.

2.3 Results and discussion
2.3.1 Fabrication and characterization of PHEMA-NTA-Fe(III) microspots
surrounded by a hydrophobic film
Figure 2.1 shows an image of a patterned plate prepared using the procedure in Scheme 1.
In the patterning process, a PDMS template stamp with 0.25 mm-diameter holes is first attached
onto a native Si wafer such that no air bubbles form. The PDMS stamp protects the Si surface, so
only the uncovered regions (microspots and the margin outside the PDMS stamp) are exposed to
UV/ozone for cleaning and to chemicals for initiator attachment, polymerization, and
derivatization. The template stamp should be as thin as possible (but strong enough to be peeled
off from the petri dish) so that the microwell does not affect the UV/ozone cleaning and
63

Figure 2.1. Si wafer coated with residual PDMS (light gray) and microspots of PHEMA-NTAFe(III) brushes (dark gray). The PDMS stamp did not cover the entire wafer, so brushes also
grew outside the light gray area on the wafer.
subsequent chemical reactions. Initially, we did not cure the stamp on the Si wafer at 75 C, but
the wafer became hot after UV/ozone treatment, and this allowed a thin film of PDMS to adhere
to the oxide layer on the Si after removing the template PDMS stamp. Because the chemical
composition of SiO 2 at the surface is somewhat similar to that of PDMS (-Si(CH3)2-OSi(CH3)2-), some low- molecular-weight PDMS molecules might transfer from the PDMS
template surface to the wafer by covalent bonding. To verify the effect of temperature on this
procedure, we heated a PDMS-coated Si wafer at 75 º for 1 h without UV/ozone treatment.
C
After peeling the PDMS stamp off the wafer, a thin hydrophobic layer remains, and the static
water contact angle on this surface is 98º
(Figure 2.2c), only ~10ºless than on bulk PDMS (110º
,
Figure 2.2b). This angle is higher than that after only UV/ozone treatment and removal of the
PDMS stamp (77º and much higher than that on a native Si wafer (24º Figure 2.2a) or on a
),
,
64

Figure 2.2. Contact angles measured 0, 2, or 5 min after placing water droplets (0.5 L) on
different surfaces: (a) native Si wafer; (b) native PDMS; (c) Si wafer modified with PDMS by
curing a PDMS/Si wafer at 75 º for 1 h and peeling off the bulk PDMS; (d) Si wafer with its
C
entire surface coated with PHEMA-NTA-Fe(III) brushes.
wafer with its entire surface coated with polymer brushes (34º Figure 2.2d). The ellipsometric
,
thickness of the residual film on the Si wafer after heating and removal of the PDMS stamp is
~1.8 nm. Longer heating times do not enhance the hydrophobicity of the Si wafer, and too high a
temperature, e.g. 120 º results in permanent bonding of the entire PDMS stamp to the Si wafer.
C,
Adhesion of the stamp to Si is unfavorable for this work because the sample wells are likely too
deep for efficient enrichment and MALDI-MS.
For sample concentration by evaporation, the PDMS should remain hydrophobic when
exposed to water, and the residual PDMS surface shows a relatively high contact angle (66º
,
Figure 2.2c) after 5 min of evaporation. Moreover, the contact area between the drop and the
residual PDMS surface decreases to ~5% of the initial area after 5 min of drop evaporation,
65

which is essential for concentrating samples. In contrast, for the wafer entirely coated with
polymer brushes (Figure 2.2d), the area of the droplet-substrate contact has a diameter of ~1.8
mm after 2 min of water evaporation. The hydrophobicity of the residual PDMS film is also
stable to sonication in toluene, THF, and acetone for five minutes, in spite of the fact that bulk
PDMS stamps swell in these solvents.
Prior to removal of the PDMS stamp from the Si wafer, immobilization of the silane
initiator on the uncovered Si regions occurs from anhydrous DMF. We use DMF rather than
toluene to avoid swelling of the bulk PDMS template. Growth of PHEMA from immobilized
initiators for 1 h yields ~40 nm-thick polymer brushes on the uncovered Si surface. (The
thickness of the brush was determined ellipsometrically on the uncovered area outside of the
PDMS stamp. See Figure 2.1.)

Derivatization of the PHEMA brushes with NTA-Fe(III)

increases the film thickness to ~80 nm. After removing the PDMS stamp, the hydrophobic region
covered with residual PDMS shows the original color (light gray) of the Si, while the uncovered
regions is a dark gray to blue color (Figure 2.1) due to the thickness of the polymer brushes. The
color distinction is important because it allows rapid location of the polymer brush spots.

Figure 2.3. (a) SEM image of a microspot of PHEMA-NTA-Fe(III) brushes surrounded by a
residual PDMS layer on Si, (b) EDS carbon mapping of the spot, and (c) EDS oxygen mapping
of the spot.

66

SEM images (Figure 2.3a) show that the microspots are well-defined with an average
diameter of 280 µ which is close to the inner diameter (250 µ of the blunt dispensing tip
m,
m)
used to form the pores in the PDMS stamp. EDS mapping shows higher distributions of carbon
and oxygen within the microspot, which is due to the relatively thick PHEMA-NTA-Fe(III)
brushes (Figure. 3b, c). Unfortunately, the signal for Fe is too weak for imaging.
The PHEMA synthesis and subsequent chemical modifications were also performed on
the entire surface of Si wafers to compare enrichment on different size spots (see below) and to
allow FTIR spectroscopic characterization of the brush synthesis. The IR of polymer film after
each reaction step is consistent with previous results. Briefly, the ester carbonyl absorbance at
-1

1725 cm

in Figure 2.4a confirms the formation of PHEMA. The doubled carbonyl absorption
-1

near 1740 cm

in Figure 2.4b reveals ester functionality. The pronounced succinimde ester
-1

peaks at 1817 and 1786 cm

in Figure 2.4c confirms the activation of the carboxylic acid with
-1

NHS. The broad peak near 1650 cm is probably due to the mixed adsorption of amide bond and
-1

carboxylate groups of NTA (Figure 2.4d). Lastly, the small absorption at 1590 cm

indicates the

complexation of Fe(III) to NTA. Atomic absorption spectroscopy shows that the amount of
2

Fe(III) chelated in PHEMA-NTA on fully coated wafers is ~14 nmol/cm . This high density of
complexes provides a high capacity for phosphopeptide binding compared to monolayers of
2

Fe(III) complexes. (A monolayer of alkanethiols has a density around 0.8 nmol/cm , and a
monolayer of NTA-Fe(III) on Si will have an even lower surface coverage.)[24]

67

Figure 2.4. Transmission FTIR absorbance spectrum of a double-side polished Si wafer coated
with PHEMA brushes on both sides (a), and absorbance spectra of the same sample after the
following modification steps: (b) reaction with succinic anhydride (SA), (c) activation with 0.1
M
N-hydroxysuccinimide/N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide
hydrochloride
(NHS/EDC), (d) derivatization with aminobutyl NTA and (e) immersion in 0.1 M FeCl3
followed by rinsing with 250 mM HOAc/30% acetonitrile/70% water. Reference 13 provides a
detailed explanation of similar spectra of PHEMA films on Au.

2.3.2 MALDI-MS analysis of phosphoprotein digests enriched on PHEMANTA-Fe(III) microspots
We initially examined phosphopeptide enrichment from digests of -casein and
ovalbumin. Table 2.1 provides the sequences of tryptic phosphopeptides from these two model
phosphoproteins. Without enrichment of phosphopeptides, 1 pmol of β-casein digest deposited
on a steel MALDI plate gives rise to a complex mass spectrum (Figure 2.5a). The three
phosphopeptides,

1P2062,

4P2966,

and

4P3122,

have

weaker

signals

than

other

nonphosphopeptides. After enrichment of the same amount of β-casein digest on a 2- mm
68

PHEMA-NTA-Fe(III) spot, however, 1P2062 dominates the spectrum, and the two tetraphosphopeptides, 4P2966 and 4P3122, also give signals strong enough for MS/MS (Figure 2.5b).
Moreover, many small peaks correspond to phosphopeptides from α-casein impurities and
phosphopeptide-related species.
Table 2.1. Phosphopeptides that result from tryptic digestion of -casein and ovalbumin. The
amino acids in parentheses represent cleavage sites and are not part of the peptide sequence.
+
Phosphory
Amino acid sequence
[M+H]
-lation sites
m/z
2062
1
(K)FQpSEEQQQTEDELQDK
-casein
2966
4
(R)ELEELNVPGEIVEpSLpSpSpSEESITR
-casein
3122
4
RELEELNVPGEIVEpSLpSpSpSEESITR*
-casein
ovalbumin 2089
1
(R)EVVGpSAEAGVDAASVSEEFR
ovalbumin 2511
1
(K)LPGFGDpSIEAQCGTSVNVHSSLR
ovalbumin 2901
1
(R)FDKLPGFGDpSIEAQCGTSVNVHSSLR*
* 4P3122 and 1P2901 are generated by missed cleavage of the phosphoprotein by trypsin.

Phosphoprotein

At protein amounts ≤1 pmol, enriching the sample on the microspot greatly increases the
signals due to the 4P2966 and 4P3122 peptides. As Figure 2.5c shows, after the enrichment of 1
pmol of β-casein digest on a microspot, the tetra-phosphorylated peptides exhibit some of the
strongest signals in the mass spectrum. The signal- to-noise ratios (S/N) for 4P3122 and 4P2966
are ~9-fold and ~3-fold greater with microspot enrichment than with 2- mm-spot enrichment
(compare Figure 2.5 b and c). The reproducibly weak signal for 1P2062 in Figure 2.5c likely
results from saturation of binding sites on the microspot with tetra-phosphopeptides (see below).
With 2- mm-spot enrichment, tetra-phosphopeptide signals dominate the MALDI mass spectrum
only when the amount of β-casein digest is above 7 pmol (data not shown), whereas at 250 fmol
of digest, signals due to 4P2966 and 4P3122 are already at the detection limit (Figure 2.5d). In
contrast, microspot enrichment still yields a high S/N for multiply phosphorylated peptides in
250 fmol of digested protein (Figure 2.5e). Unfortunately, when the amount of β-casein digest is
69

Figure 2.5. MALDI mass spectra of β-casein digests obtained using: (a) a conventional stainless
steel MALDI plate without enrichment, and (b-h) enrichment on wafers coated with PHEMANTA-Fe(III) brushes. The amount of β-casein digest and the size of enrichment spots are noted
70

in each spectrum. Mono- and tetra-phosphorylated peptides are labeled as 1P and 4P,
respectively, with their S/N in parentheses. Other phosphopeptide-related peaks are labeled with
different symbols: ▼ doubly protonated phosphopeptide, ● -S1-casein phosphopeptide,  S2-casein phosphopeptide, ○ phosphopeptide signal after loss of 98 Da (H3PO4). Peak
identification was facilitated by CID-MS/MS.

Figure 2.6. MALDI mass spectra of ovalbumin digests after enrichment on PHEMA-NTA-Fe(III)
brushes. The amount of ovalbumin digest and the size of enrichment spot are noted in each mass
spectrum. The mono-phosphopeptides are labeled as 1P, with their S/N ratios in parentheses.
Phosphopeptides formed by loss of 98 Da (H3PO4) are labeled as ○; doubly protonated
phosphopeptides are labeled as ▼. Peak identification was facilitated by CID-MS/MS.

71

below 100 fmol, enrichment on the microspot does not allow detection of 4P2966 and 4P3122.
The ability of the microspot to enrich mono-phosphopeptides from very small amounts of
digest is especially remarkable. Even 5 fmol of β-casein digest generates a very strong peak for
1P2062 (S/N=65) with microspot enrichment, whereas the corresponding signal with 2- mm-spot
enrichment is at the detection limit (Figure 2.5 f and g). The lowest level for detecting 1P2062
after microspot enrichment is around 0.8 fmol (S/N=9, Figure 2.5h), which is significantly lower
than the approximate detection limits for 2-mm-spot enrichment (5 fmol, Figure 2.5f).
Enrichment on microspots rather than 2- mm spots also enhances the detection of
phosphopeptides in ovalbumin digests (Figure 2.6). With 1 pmol of digest, enrichment on the 2mm spots followed by MALDI-MS reveals the 1P2089 and 1P2511 peptides, whereas
enrichment on the microspots allows detection of a third phosphopeptide, 1P2901. Moreover, the
lowest detection limit for 1P2089 by MALDI-MS after microspot enrichment is ~5 fmol (S/N=6),
which is 5- fold lower than the approximate detection limit after 2-mm-spot enrichment (25 fmol,
S/N=2, see Figure 2.6 e and f).

2.3.3 Enrichment of phosphopeptides from mixtures of protein digests
Enrichment of a simple mixture of β-casein and ovalbumin digests (25 fmol of each) yields
signals of phosphopeptides from both phosphoproteins (Figure 2.7a). However, in many
biological samples, the phosphoprotein represents only a small fraction of the total protein, so we
also tested phosphopeptide enrichment on microspots of PHEMA-NTA-Fe(III) using a mixture
that contains digested β-casein, ovalbumin and BSA with a molar ratio of 1:1:10. (BSA is not a
phosphoprotein.) Because there are many more nonphosphopeptides than phosphopeptides in
the sample, enrichment of this complex mixture on a 2- mm spot yields relatively large
nonphosphopetide signals and only one strong phosphopeptide signal that can be distinguished
72

Figure 2.7. MALDI mass spectra of mixtures of protein digests with (a-d) and without
phosphopeptide enrichment (e). The digest contents, size of enrichment spot (a-d) and method of
rinsing (a-d) are noted in the spectra. Mono-phosphopeptides are labeled as 1P, with their S/N
ratios in parentheses. Other phosphopeptide-related peaks are labeled with different symbols: ▼
doubly protonated phosphopeptide, ● -S1-casein phosphopeptide,  -S2-casein
phosphopeptide, and ○ phosphopeptide signal after loss of 98 Da (H3PO4). Peak identification
was facilitated by CID-MS/MS.
from adjacent nonphosphopeptides (Figure 2.7b). To improve the efficiency of rinsing, we
immerse the enriched samples in a sonicated HOAc/ACN/water solution. This results in smaller
signals for nonphosphopeptides, but 1P2062 is still the only phosphopeptide (generated from β73

casein and ovalbumin) that can be identified on the 2- mm spot (Figure 2.7c). No signals from
multiply phosphorylated peptides of β-casein or mono-phosphopeptides of ovalbumin are
detectable. Using microspot enrichment with a sonication-assisted rinse, however, the S/N of
1P2062 is 5- fold higher than after 2- mm-spot enrichment, and the mono-phosphopeptide 1P2089
from ovalbumin, as well as three more phosphopeptides from -casein, give detectable signals
(Figure 2.7d).

The S/N for 1P2062 and for 1P2089 are also greater than in conventional

MALDI-MS (Figure 2.7e).

2.3.4 Mechanism of the enrichment of phosphopeptides on PHEMA-NTAFe(III) microspots

Figure 2.8. Charge-coupled device images of sample enriched on (a) a 2-mm spot and (b) a
microspot. The images were obtained after addition of DHB matrix. (Image (a) shows only a
portion of the 2-mm spot. The cross is not at the center of the spot.)
Because of differences in sample area, the surface concentration of phosphopeptides on
the microspot should ideally be 64 times higher than on the 2- mm spot. Figure 2.8 shows that the
matrix on the microspot covers a much smaller area and is more homogeneous than on the 2- mm
spot. However, the S/N obtained with the microspot is only 3- to 20-fold greater than the S/N

74

with the 2- mm spot. Other studies also reported that MALDI-MS signals are not directly
proportional to sample surface area.[14-16] When a peptide/matrix solution deposited on
parafilm formed a layer of sample/matrix crystals covering an area ~1/10 of that on a stainless
steel MALDI plate, the S/N was only 3- to 5- fold higher than on the stainless steel plate.[15]
Moreover, in our work, the improvement in S/N for phosphopeptides using the microspot
approach is much greater for low- femtomole amounts of phosphoprotein digest than for
amounts >100 fmol. There are likely two reasons for these findings.
First, there is already some concentration of the sample during matrix crystallization on
2-mm spots. After adding the DHB matrix solution onto 2-mm spots, the phosphopeptides
dissolve and then partition into certain matrix crystals. Mass spectra typically come from “sweet
spot” crystals that likely contain a disproportionate fraction of the total phosphopeptides,
whereas MALDI-MS signals from other regions of the well are very weak. Addition of 6.4- fold
more matrix to achieve the same amount of matrix per area on the 2- mm spot as on the microspot
does not improve the homogeneity of the 2-mm sample and gives a ~25% decrease in S/N.
Second, there is some loss of phosphopeptides in microspot enrichment, especially at
high levels of phosphopeptide because the binding capacity of the microspot is not sufficient to
capture all of the phosphopeptides in a sample. According to our previous work,[13] the binding
capacity of ~60-nm thick PHEMA-NTA-Fe(III) brushes is 0.6 µ of phosphoangiotensin per
g
2

cm . Considering that the thickness of polymer brushes in the present work is ~80 nm, and that
the number of phosphopeptides is 2 per digested β-casein or ovalbumin molecule (the third
phosphopeptide in both cases is due to unspecific or missed cleavage), as a first approximation
the binding capacity of a 2-mm spot should be ~5 pmol for a β-casein or ovalbumin digest. (This
calculation assumes that the binding capacity in µ is independent of molar mass.) The binding
g
75

capacity of a 0.25- mm spot is only ~80 fmol for a β-casein or ovalbumin digest. The low
binding capacity of the microspot likely explains why the overall S/N improvement after
microspot enrichment is more evident when the amount of phosphoprotein digest is below 100
fmol. At phosphopeptide levels above the binding capacity of the polymer microspot, the excess
phosphopeptides will be removed with the nonphosphopeptides in the rinsing step.
When there is a high excess of phosphopeptides on the microspot, multiply
phosphorylated peptides are more likely to bind to the available affinity sites than monophosphopeptides.[25,26] As Figure 2.5c shows, after enrichment of 1 pmol of β-casein digest on
the microspot (80-fmol binding capacity), the sum of the S/N of the tetra-phosphopeptides is 19
fold higher than the S/N of the mono-phosphopeptide. In contrast, enrichment of the same
amount of sample on a 2-mm spot (~5-pmol binding capacity) results in a sum of the S/N of the
tetra-phosphopeptides that is 9- fold lower than the S/N of the mono-phosphopeptide. In samples
where the amount of phosphopeptide is significantly less than the binding capacity, the brushes
bind both tetra- and mono-phosphopeptides, but the signals of tetra-phosphopeptides are lower,
probably because of inefficient elution or ionization of the tetra-phosphopeptides. When trying
to estimate relative levels of phosphopeptides, it is important to spot dilute samples whose
phosphopeptide amount does not exceed the binding capacity.

2.4 Conclusions
This work presents a simple method for fabricating polymer brush microspots (250 µ
m
diameter) surrounded with a hydrophobic PDMS layer. The use of these spots for on-plate
enrichment of femtomole amounts of phosphopeptides allows pre-concentration of the sample
droplet on the plate and gives an improved phosphopeptide MALDI-MS S/N compared to
76

enrichment on larger spots. Moreover, the polymer microspot can enrich low-femtomole
amounts of phosphopeptides even when the sample contains an order of magnitude excess of
non-phosphoprotein digest. Enhancement of the MALDI-MS S/N for mono-phosphopeptides is
especially evident in the enrichment of low-femtomole amounts of phosphoprotein digests
because the amount of phosphopeptide in the sample is less than the binding capacity. Thus, this
microspot enrichment method may be particularly attractive for analysis of samples such as
immunoprecipitates, where a limited amount of material is available. Moreover, high-throughput
MALDI-MS analysis can be realized by fabricating arrays of microspots on a Si wafer. The
small size of the polymer microspot also enhances the detection of multiply phosphorylated
peptides, but these species still have higher detection limits than mono-phosphorylated peptides.
Because the microspot fabrication does not employ any special instruments, it might be a
substitute for microcontact printing in some cases. Moreover, the principle of this method could
be employed to synthesize microspots of other polymers on Si wafers for the enrichment of
different species.

77

REFERENCES

78

REFERENCES
1.
Paradela, A.; Albar, J. P., Advances in the Analysis of Protein Phosphorylation. J.
Proteome Res. 2008, 7, 1809-1818.
2.
Thingholm, T. E.; Jensen, O. N.; Larsen, M. R., Analytical Strategies for
Phosphoproteomics. Proteomics 2009, 9, 1451-1468.
3.
Mann, M.; Ong, S. E.; Gronborg, M.; Steen, H.; Jensen, O. N.; Pandey, A., Analysis of
Protein Phosphorylation Using Mass Spectrometry: Deciphering the Phosphoproteome. Trends
Biotechnol. 2002, 20, 261-268.
4.
Steen, H.; Jebanathirajah, J. A.; Rush, J.; Morrice, N.; Kirschner, M. W., Phosphorylation
Analysis by Mass Spectrometry - Myths, Facts, and the Consequences for Qualitative and
Quantitative Measurements. Mol. Cell. Proteomics 2006, 5, 172-181.
5.
Goshe, M. B.; Conrads, T. P.; Panisko, E. A.; Angell, N. H.; Veenstra, T. D.; Smith, R.
D., Phosphoprotein Isotope-Coded Affinity Tag Approach for Isolating and Quantitating
Phosphopeptides in Proteome-Wide Analyses. Anal. Chem. 2001, 73, 2578-2586.
6.
McLachlin, D. T.; Chait, B. T., Improved Beta-Elimination-Based Affinity Purification
Strategy for Enrichment of Phosphopeptides. Anal. Chem. 2003, 75, 6826-6836.
7.
Warthaka, M.; Karwowska-Desaulniers, P.; Pflum, M. K. H., Phosphopeptide
Modification and Enrichment by Oxidation-Reduction Condensation. ACS Chem. Biol. 2006, 1,
697-701.
8.
Dunn, J. D.; Watson, J. T.; Bruening, M. L., Detection of Phosphopeptides Using Fe(III)Nitrilotriacetate Complexes Immobilized on a MALDI Plate. Anal. Chem. 2006, 78, 1574-1580.
9.
Ficarro, S. B.; McCleland, M. L.; Stukenberg, P. T.; Burke, D. J.; Ross, M. M.;
Shabanowitz, J.; Hunt, D. F.; White, F. M., Phosphoproteome Analysis by Mass Spectrometry
and Its Application to Saccharomyces Cerevisiae. Nature Biotechnol. 2002, 20, 301-305.
10.
Zhou, H. J.; Xu, S. Y.; Ye, M. L.; Feng, S.; Pan, C.; Jiang, X. G.; Li, X.; Han, G. H.; Fu,
Y.; Zou, H., Zirconium Phosphonate-Modified Porous Silicon for Highly Specific Capture of
Phosphopeptides and MALDI-Tof MS Analysis. J. Proteome Res. 2006, 5, 2431-2437.
11.
Larsen, M. R.; Thingholm, T. E.; Jensen, O. N.; Roepstorff, P.; Jorgensen, T. J. D.,
Highly Selective Enrichment of Phosphorylated Peptides from Peptide Mixtures Using Titanium
Dioxide Microcolumns. Mol. Cell. Proteomics 2005, 4, 873-886.
12.
Qiao, L.; Roussel, C.; Wan, J. J.; Yang, P. Y.; Girault, H. H.; Liu, B. H., Specific onPlate Enrichment of Phosphorylated Peptides for Direct MALDI-Tof MS Analysis. J. Proteome
Res. 2007, 6, 4763-4769.
79

13.
Dunn, J. D.; Igrisan, E. A.; Palumbo, A. M.; Reid, G. E.; Bruening, M. L.,
Phosphopeptide Enrichment Using MALDI Plates Modified with High-Capacity Polymer
Brushes. Anal. Chem. 2008, 80, 5727-5735.
14.
Hung, K. C.; Rashidzadeh, H.; Wang, Y.; Guo, B. C., Use of Paraffin Wax Film in
MALDI-Tof Analysis of DNA. Anal. Chem. 1998, 70, 3088-3093.
15.
Wang, J. H.; Chen, R. B.; Ma, M. M.; Li, L. J., MALDI MS Sample Preparation by Using
Paraffin Wax Film: Systematic Study and Application for Peptide Analysis. Anal. Chem. 2008,
80, 491-500.
16.
Hung, K. C.; Ding, H.; Guo, B. C., Use of Poly(Tetrafluoroethylene)S as a Sample
Support for the MALDI-Tof Analysis of DNA and Proteins. Anal. Chem. 1999, 71, 518-521.
17.
Schuerenbeg, M.; Luebbert, C.; Eickhoff, H.; Kalkum, M.; Lehrach, H.; Nordhoff, E.,
Prestructured MALDI-MS Sample Supports. Anal. Chem. 2000, 72, 3436-3442.
18.
Chen, X. X.; Murawski, A.; Kuang, G.; Sexton, D. J.; Galbraith, W., Sample Preparation
for MALDI Mass Spectrometry Using an Elastomeric Device Reversibly Sealed on the MALDI
Target. Anal. Chem. 2006, 78, 6160-6168.
19.
Jia, W. T.; Wu, H. X.; Lui, H. J.; Li, N.; Zhang, Y.; Cai, R. F.; Yang, P. Y., Rapid and
Automatic on-Plate Desalting Protocol for Mald-MS: Using Imprinted Hydrophobic Polymer
Template. Proteomics 2007, 7, 2497-2506.
20.
Xu, Y. D.; Bruening, M. L.; Watson, J. T., Use of Polymer-Modified MALDI-MS Probes
to Improve Analyses of Protein Digests and DNA. Anal. Chem. 2004, 76, 3106-3111.
21.
Xu, Y. D.; Watson, J. T.; Bruening, M. L., Patterned Monolayer/Polymer Films for
Analysis of Dilute or Salt-Contaminated Protein Samples by MALDI-MS. Anal.l Chem. 2003, 75,
185-190.
22.
575.

Xia, Y. N.; Whitesides, G. M., Soft Lithography. Angew. Chem.-Int. Ed. 1998, 37, 551-

23.
Matyjaszewski, K.; Miller, P. J.; Shukla, N.; Immaraporn, B.; Gelman, A.; Luokala, B. B.;
Siclovan, T. M.; Kickelbick, G.; Vallant, T.; Hoffmann, H.; Pak ula, T., Polymers at Interfaces:
Using Atom Transfer Radical Polymerization in the Controlled Growth of Homopolymers and
Block Copolymers from Silicon Surfaces in the Absence of Untethered Sacrificial Initiator.
Macromolecules 1999, 32, 8716-8724.
24.
Ulman, A., Formation and Structure of Self-Assembled Monolayers. Chem. Rev. 1996,
96, 1533-1554.
25.
Li, Q. R.; Ning, Z. B.; Tang, J. S.; Nie, S.; Zeng, R., Effect of Peptide-to-Tio2 Beads
Ratio on Phosphopeptide Enrichment Selectivity. J. Proteome Res. 2009, 8, 5375-5381.

80

26.
Qiao, L. A.; Bi, H. Y.; Busnel, J. M.; Hojeij, M.; Mendez, M.; Liu, B. H.; Girault, H. H.,
Controlling the Specific Enrichment of Multi-Phosphorylated Peptides on Oxide Materials:
Aluminium Foil as a Target Plate for Laser Desorption Ionization Mass Spectrometry. Chem. Sci.
2010, 1, 374-382.

81

Chapter 3. Identification of p65-associated Phosphoproteins
by Mass Spectrometry after On-plate Phosphopeptide
Enrichment Using Polymer-oxotitanium Films
(The work described in this chapter was performed in collaboration with Dr. Amanda M.
Palumbo from Prof. Gavin E. Reid‟s and Prof. Jetze J. Tepe‟s groups at Michigan State
University. This work has been published in: W.H. Wang, A.M. Palumbo, Y.J. Tan, G.E. Reid,
J.J. Tepe, M.L. Bruening, Identification of p65-Associated Phosphoproteins by Mass
Spectrometry after On-Plate Phosphopeptide Enrichment Using Polymer-oxotitanium Films. J.
Proteome Res. 2010, 9, 3005–3015.)

3.1 Introduction
Nuclear factor kappa- light-chain-enhancer of activated B cells (NF-κB) is an ubiquitous
transcription factor that regulates the transcription of genes by binding to discrete DNA
sequences.[1] The NF-κB signaling pathway impacts many aspects of cellular adaptation such as
activation of immune cell function[2] and programmed cell death[3]. Seven NF-κB family
proteins exist in mammalian cells, including RelA (p65), RelB, c-Rel, p50, p105, p52 and
p100,[4] and these proteins form combinations of homo- and heterodimers. The different NF-κB
proteins share an N-terminal NF-κB/Rel homology domain (RHD), which mediates DNA
binding, dimerization, nuclear translocation and interaction with the inhibitory IκB family
proteins.[4] In resting cells, NF-κB is inactive due to association with IκB proteins. Stimulation
of cells with agonists results in phosphorylation, ubiquitination and degradation of IκBs, which

82

releases NF-κB for nuclear translocation and activation of gene transcription.[5] Well-known
stimuli that induce NF-κB activity include cytokines, γ radiation, and DNA damage.[4]
Within the NF-κB protein family, the heterodimmer p50/p65 is the most abundant and
prototypical form, and only the p65 subunit contains a C-terminal transcriptional activation
domain.

Many biochemical and genetic experiments suggest an important role of post-

translational modifications of NF-κB subunits in its function. For example, different protein
kinases phosphorylate p65 at Thr254, Ser276, Ser311, Thr435, Ser468, Ser529 and Ser536
residues.[1,4,6,7] These modifications enhance the transcriptional activity of p65 and its ability
to interact with co-activators.[1] In addition, the phosphorylation of other proteins may control
their interactions with p65 and alter the regulation of p65- mediated transcriptional activity.[8]
Thus, a full understanding of p65 transcriptional regulation requires examination of the posttranslational modifications of both p65 and its associated proteins in response to different stimuli.
A rapid, sensitive analysis technique that requires low sample volumes would facilitate such
studies.
This work describes an analytical approach to elucidate the phosphorylation status of a
target protein and its associated proteins in small samples (g quantities), with a specific
emphasis on proteins associated with immunoprecipitated p65. Currently, profiling of protein
phosphorylation with MS typically requires enrichment of phosphopeptides from proteolytic
digests using immobilized metal ion affinity chromatography (IMAC) with Fe(III) or Ga(III)
complexes or metal oxide affinity chromatography.[9-12] Several recent studies suggest that
TiO2 particles or IMAC-Ti(IV) frequently exhibit greater specificity for phosphopeptide
adsorption than IMAC-Fe(III) or –Ga(III).[13-19] However, most enrichment methods utilize
column-based separations in which the sample volume is more than that needed for MS,
83

especially for MALDI-MS. Moreover, column-based procedures often require extensive sample
handling that decreases throughput and risks sample loss due to adsorption on columns and
container walls.[9]
One strategy to minimize sample consumption and simplify enrichment is capture of the
desired species (~1 L sample volume) directly on a MALDI plate. Compared to electrospray
ionization (ESI), MALDI allows for lower sample consumption, as well as decreased complexity
in the MS spectrum due to the formation of primarily singly protonated ions. Moreover, the
stability of sample/matrix crystals on MALDI plates allows multiple MS or MS/MS experiments
to be performed on the same spot. Several groups reported on-plate enrichment of model
phosphopeptides with TiO 2-coated surfaces[16,20-22] or other substrates.[9] However, most
studies did not employ a chaotropic denaturing agent (e.g. urea)[23] during sample preparation,
which is often necessary for complete protein digestion, especially for cell lysates.[24] High
concentrations of urea (up to 6 M) are often deleterious to enrichment procedures and MALDIMS.[25] Moreover, as far as we know, none of the previous on-plate enrichment methods were
used to examine the phosphorylation status of proteins immunoprecipitated from cell extracts. In
the case of column-based techniques, the majority of research efforts aimed at enrichment of
hundreds of phosphopeptides from whole cell extracts,[11] rather than addressing biological
questions related to one or several target phosphoproteins that may be present at low levels in
cells.
This chapter describes a polymer brush-oxotitanium (P-oxoTi) hybrid material
synthesized on Au wafers for on-plate enrichment of phosphopeptides in digested
immunoprecipitates that contain a number of proteins and high levels of urea and salts. Selective
binding to titania nanoclusters in P-oxoTi brushes results in more MS signals from
84

phosphopeptides than on-plate capture with IMAC or enrichment with commercial TiO 2
particles.

Moreover, enrichment on Au-P-oxoTi facilitates the identification of several

phosphopeptides from digests of p65-associated proteins immunoprecipitated together with p65
from human acute monocytic leukemia cell (THP-1) nuclear extracts. MALDI-MS/MS of the
enriched phosphopeptides and database searching allow the elucidation of phosphopeptide
sequences, phosphorylation sites, and the parent phosphoproteins.

3.2 Experimental
3.2.1 Chemicals and materials
Most reagents for polymer brush synthesis, sample preparation and phosphopeptide
enrichment were purchased from Sigma-Aldrich. Other chemicals and materials used for protein
digestion, enrichment, and MS include: sequencing grade modified porcine trypsin (treated with
L-1-tosylamide-2-phenylethyl chloromethyl ketone (TPCK), Promega), Tris-HCl (Invitrogen),
urea (J.T.Baker), DL-dithiothreitol (Fluka), ammonium bicarbonate (Columb us Chemical),
trifluoroacetic acid (TFA, EMD Chemicals), acetonitrile (ACN, Mallinckrodt Baker), propionic
anhydride-d10 (98%, C/D/N Isotopes), TiO 2 particles (Titansphere, 5 m, GL Science, Japan),
and desalting pipette tips (ZipTipC18, Millipore). Tetrahydrofuran (THF, Mallinckrodt Baker)
was newly distilled before use, and deionized water was prepared with a Millipore ion-exchange
system (Milli-Q, 18 M cm). Au-coated Si wafers (200 nm of sputtered Au on 20 nm of
sputtered Cr on Si wafers) were prepared by LGA Thin Films.

85

Table 3.1. N-Propionylated standard phosphopeptides employed for determination of enrichment
recoveries.*
Peptide sequence

#
–PO3H2
1
1
1
1

#
-NHCOCH2CH3
1
1
1
1

+

m/z MH after
propionylation
950.5
1277.6
1537.7
2765.6

LPQEpTAR
RDpSLGTYSSR
EVQAEQPSSpSSPR
LGPGRPLPTFPpTSE(carbamidometh
yl)TSDVEPDTR
RYpSpSRSR
2
1
1127.5
pSRSPpSSPELNNK
2
2
1587.7
IHRLARpYpYKTK
2
3
1777.0
* The peptides listed in this table are those with complete N-propionylation (propionylderivatization occurs on both the N-terminus and the primary amine of Lys residues).

A mixture of 10 standard phosphopeptides (donated by Sigma Genosys (The Woodlands,
TX) was N-propionylated for quantitative evaluation of phosphopeptide recovery on Au-P-oxoTi
wafers. The peptide mixture (200 pmol of total phosphopeptides) was dissolved in 20 µ of 50
L
mM NH4HCO 3. Half of the solution was mixed with 20 µ of propionic anhydride solution (5
L
µ of propionic anhydride in 45 µ of methanol) and allowed to react for 2 h at room
L
L
temperature. The sample was dried with a SpeedVac, and the resultant phosphopeptides were
dissolved in 10 µ of 0.1% TFA. The other half of the standard mixture solution was derivatized
L
using deuterated propionic anhydride in the same way. With this protocol, all the standard
phosphopeptides were propionylated on at least one primary amine (no unmodified
phosphopeptides were found in MALDI- mass spectra). The majority of phosphopeptides were
completely propionylated on N-terminus and Lys residues, but some peptides were overpropionylated on other residues. Only phosphopeptides with complete propionylation (Table 3.1)
were used for determination of the percent recovery. Procedures for synthesizing the single
standard phosphopeptide, CD3CD 2CO-LFTGHPEpSLEK, which was used as an internal
86

standard for the semi-quantitative analysis of p65-associated phosphoproteins, were reported
before.[26]
THP-1 cells were obtained from ATCC (Manassas, VA), and primary p65 polyclonal
antibodies (C-20, C-20 agarose conjugate, C22B4), protein A/G PLUS agarose, and normal goat
IgG were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Phosphate buffered
saline (PBS), tumor necrosis factor (TNF)-, and cellular growth media along with its
supplements were acquired from Invitrogen (Carlsbad, CA).

3.2.2 Preparation of polymer-oxotitanium Films on Au Wafers (Au-P-oxoTi)
The synthesis of poly(2- hydroxyethyl methacrylate) (PHEMA) brushes on Au-coated
wafers (2.4 × 1.1 cm) by atom transfer radical polymerization, and the subsequent reaction with
succinic anhydride to introduce carboxylic acid groups were described previously.[27] In a
glove bag, such polyacid-modified (polymer thickness of ~50 nm) Au wafers were immersed in
a stirred, 10- mL THF solution containing 50 mM titanium(IV) isopropoxide (Ti(i-PrO)4) for 30
min to form the P-oxoTi hybrid material (Figure 3.1a). The wafer was finally rinsed with THF,
dried with nitrogen, and removed from the glove bag.

3.2.3 Characterization of Au-P-oxoTi films
The thicknesses of polymer brushes on Au wafers were determined using a rotating
analyzer spectroscopic ellipsometer (J. A. Woollam, M-44), assuming a polymer refractive index
of 1.5. Reflectance FT-IR spectra of the modified Au wafers were collected on a Nicolet Magna
560 spectrophotometer with a Pike grazing angle (80˚) accessory.
The amount of Ti on wafer surfaces was quantified by inductively coupled plasma optical
emission spectrometry (ICP-OES) as follows. First a Au/aqua regia stock solution was prepared
87

86

Figure 3.1. (a) Schematic drawing of the formation of the P-oxoTi hybrid structure, and the binding and elution of
phosphopeptides in this material. (b) Procedures for on-plate phosphopeptide enrichment using Au-P-oxoTi.

88

by dissolving Au from 20 pieces of 2.4 × 1.1 cm Au-coated wafers into 80 mL of aqua regia
solution.

This stock solution was added to some Ti solutions to maintain a constant

concentration of Au in all analysis solutions. To test the stability of P-oxoTi films in loading
(0.1% TFA or 2% TFA) and elution solutions (1% H3PO 4), 0.5 mL of the test solution was
spread on the entire surface of a Au-P-oxoTi wafer (2.4 × 1.1 cm) for either 30 min (for TFA
solution) or 5 min (for 1% H3PO4 solution). The solution was then collected, and the wafer was
rinsed with 5 mL of water. The combined eluate and rinsing solutions were added to 4 mL of
Au/aqua regia stock solution and diluted to 50 mL with deionized water. The remaining Ti on
each wafer was removed together with polymer and Au by immersing the entire wafer in 4 mL
of pure aqua regia for 5 min. This solution, combined with a 5- mL washing solution, was diluted
to 50 mL with deionized water. Triplicate tests were carried out for exposure of the wafers to
each type of solution.

A series of Ti standard solutions were prepared by adding specific

volumes of a Ti ICP standard solution (Sigma) to a 50- mL flask containing 4 mL of Au/aqua
regia stock solution, and finally diluting the solution to 50 mL with D.I. water. Using these
standards, a Ti calibration curve was obtained using ICP-OES, and the amount of Ti in different
eluate solutions was determined using the calibration curve. The total amount of Ti on a single
intact Au-P-oxoTi wafer was calculated by averaging the amount of Ti completely removed from
10 Au-P-oxoTi wafers (705 nm polymer-oxotitanium film thickness). The amount of Ti per unit
2

area was calculated by dividing the amount of Ti on a single Au-P-oxoTi wafer by 2.64 cm (the
area of the 2.4 × 1.1 cm Au wafer).

89

3.2.4 Immunoprecipitation (IP) of p65 and associated proteins from THP-1
nuclear extracts
This part was performed by Amanda M. Palumbo from Prof. Gavin E. Reid‟s group.
Briefly, nuclei were isolated from THP-1 cells with and without TNF- treatment (10 ng/mL
TNF- for 30 min at 37 ° with 5% CO 2 and ambient oxygen), respectively, by centrifugation.
C
These nuclei were lysed followed by determination of protein concentration using Bradford
assay (IgG as standards). Samples for IP were prepared with 1 mg of total nuclear protein in 800
L of PBS buffer. After preclearance of endogenous IgGs with normal goat IgG and Protein
A/G PLUS agarose, the nuclear extracts were immunoprecipitated onto agarose cross- linked with
p65 antibody (overnight incubation at 4 ° The agarose was centrifuged to discard supernatant,
C).
washed, drained and finally stored at -20 ° until use. Blank samples (800 μL PBS) were also
C
“immunoprecipitated” as a control. Experimental details can be found in the Ph.D dissertation of
Dr. Amanda M. Palumbo.[28]

3.2.5 Preparation of tryptic digests
The digestion of standard proteins followed a literature procedure. [29] Briefly, 100 μg of
each phosphoprotein (α-casein, β-casein, and ovalbumin) were dissolved in 20 μL of 6 M urea
containing 50 mM Tris-HCl. Five μL of 10 mM DL-dithiothreitol was added to all the protein
solutions followed by heating in a water bath at 60 °C for 1 h. Ammonium bicarbonate (160 μL
of 50 mM solution) and 10 μL of 100 mM iodoacetamide were added to the protein solutions,
which were then placed in the dark for 1 h. Subsequently, 10 μL of 0.5 μg/μL sequencing grade
TPCK modified trypsin was added to each solution prior to incubation at 37 ˚C for 16 h. To
quench the tryptic digestion, 11 μL of glacial acetic acid was finally added to each solution. As a
90

relatively large amount of bovine serum albumin (BSA) was used as an excess
nonphosphoprotein, TPCK-treated bovine pancreas trypsin with an activity 10,000 N -benzoylL-arginine ethyl ester units/mg protein was used for BSA digestion, with the digestion
procedures given above. All the digest samples were dried with a Speed Vac (Savant SC110) and
dissolved in TFA solutions (0.1% - 2% TFA, depending on sample complexity). The standard
protein digests mixtures were prepared by mixing different amount of each protein digest. For
enrichment using TiO 2 particles (Titansphere), 50% ACN was included in the sample solutions.
For tryptic digestion of immunoprecipitated proteins, the IP samples (agarose) described
above were concentrated to dryness with a Speed Vac, denatured, reduced, alkylated and
tryptically digested (using sequencing grade TPCK-modified trypsin) directly on the IP beads
(the same procedures as for 100 g of standard proteins) without protein elution. The digested
samples were then centrifuged for 1 min at maximum speed and the supernatant was collected.
The residual resin was washed with 20 μL water, centrifuged for 1 min, and the wash solution
was collected and combined with the initial supernatant.

The combined supernatant was

concentrated to near dryness and stored at -20 ° Ten L of 0.1% TFA was added before
C.
enrichment.

3.2.6 Phosphopeptide enrichment
3.2.6.1 On-plate enrichment using modified wafers
Arrays of 2-mm circles were scratched on the modified wafer surface to form sample
wells.[26] Figure 3.1b shows the enrichment procedure. One μL of sample solution was added to
each sample well, followed by incubation in a covered, humid petri dish for 30 min. The wafer

91

was then washed with a solution containing 50% ACN and the same concentration of TFA as in
the sample solution, and dried with nitrogen. Phosphopeptides were eluted from the polymer
brush by adding 1.5 μL of 1% H3PO 4 on the sample well and waiting for 5 min. Finally, 0.25 μL
of 40 mg/mL DHB (prepared in 1% H3PO4 in 50% ACN) was directly added on top of the 1%
H3PO4

droplet,

and

the entire solution

was allowed

to

dry.

The

wafer with

phosphopeptide/matrix spots was attached to a modified steel MALDI plate for MS analysis.
To determine the recovery of standard phosphopeptides (Table 3.1) on Au-P-oxoTi films,
0.4 µ of the 10- µ mixture of CH3CH2CO-phosphopeptides, in which the concentration of
L
L
each peptide was 1 µ was spotted on Au-P-oxoTi. The droplet was immediately mixed with 1
M,
µ of a solution containing specific amounts of BSA digest (0, 1, or 10 pmol) in a TFA solution
L
to give a final sample droplet with the desired concentration of TFA (0.1%, 1%, or 2%) and BSA
digest (0, 0.71, or 7.1 µ on the wafer. After incubation, rinsing, and elution, 0.4 µ of the
M)
L
deuterated analog mixture (CD3CD2CO-phosphopeptides) was added on top of the 1% H3PO4
eluate, followed by addition of DHB matrix. MALDI-MS spectra were taken after allowing the
peptide/matrix solution to

dry.

A calibration curve of unenriched

CH3CH2CO-

phosphopeptide/CD3CD2CO-phosphopeptide MS signal ratios as a function of their molar ratios
was obtained by conventional MALDI-MS on a steel plate. The quantitative analysis of each
sample was carried out three times to obtain the average recovery and its standard deviation.

92

3.2.6.2 Phosphopeptide enrichment using commercial TiO2 particles
Phosphopeptide enrichment using commercial TiO 2 followed a literature procedure[30]
with some modification. Briefly, TiO 2 particles (Titansphere, 15 mg) were dispersed in 300 L
of a solution containing 0.1% TFA and 50% ACN (vortex mixing for 3 min) to form a slurry.
Immediately, 60 L of the slurry was transferred to a tube containing 200 L of loading solution
(0.1% or 2% TFA in 50% ACN). After 3 min of vortex mixing and 5 min of centrifugation with
a microcentrifuge, the supernatant was discarded and the TiO 2 particles were redispersed by
vortex mixing in 60 L of loading solution to form the final TiO 2 slurry. Ten L of this slurry
was added to 80-100 L of sample for enrichment. The TiO 2-sample mixture was treated with
vortex mixing for 3 min after every 5- min of incubation, and the total time for enrichment was
30 min. After spinning down the TiO 2 particles, the supernatant was discarded, and 150 µ of
L
loading solution was added to wash the particles (3 min of vortex mixing, 5 min of centrifugation,
and discarding the supernatant). This washing step was performed three times.
For phosphopeptide elution from TiO 2 particles, our initial studies showed that an eluent
containing 1% H3PO4 and 50% ACN results in better matrix crystallization and often higher MS
phosphopeptide signals than basic elution using 1.6% NH4OH in 50% ACN. The decrease in
signal with the basic eluent may occur because the residual acidic washing solution partially
neutralizes the NH4OH, especially when the eluent volume is low. In addition, the formation of
salts due to the addition of the acidic matrix so lution to the basic eluate may also decrease the

93

MS signals. Thus, phosphopeptides were eluted from TiO 2 by adding 50-75 L of 1% H3PO4 in
50% ACN to the tube followed by 3 min of vortex mixing and 5 min of centrifugation. We used
this volume of eluent to keep the concentrations of phosphopeptides at a similar level as in onplate enrichment, assuming the recovery is 100%. A 1-L aliquot of the eluate was deposited on
the steel MALDI plate prior to addition of matrix (0.25 L of 40 mg/mL DHB in 1% H3PO 4, 50%
ACN) for MS analysis.
For

protein

digests

containing

a

high

level

of

nonphosphopeptides

(e.g.

phosphoprotein/nonphosphoprotein ratio of 1/100), we found that introducing DHB in the
sample and washing solutions results in a high loss of phosphopeptides (both the number of
phosphopeptides and S/N of phosphopeptides in mass spectra). Instead, simply increasing the
TFA concentration helps suppress the binding of nonphosphopeptides on TiO 2 particles.

3.2.7 MALDI-MS and CID-MS/MS analysis
For conventional MALDI-MS analysis of protein digests (no enrichment), 1 μL of the
digest sample was spotted on the stainless steel MALDI plate, followed by matrix addition (0.25
L of 40 mg/mL DHB in 1% H3PO 4, 50% ACN).
Mass spectrometry analysis was performed using an LTQ XL ion trap mass spectrometer
equipped with a vMALDI source (Thermo Fisher Scientific, San Jose, CA). Ion trap CID3

MS/MS and MS experiments were performed on mass-selected precursor ions. Isolation and
activation of the precursor ion was limited to either the monoisotopic ion or the monoisotopic ion
and the

13

C isotope (the first heavy isotope), to decrease the amount of signal in the product ion

94

spectrum resulting from activation of extraneous ions with similar m/z. All MS/MS spectra
shown are typically the average of 500-3000 scans (3 microscans/scan).
MS/MS spectra were transformed to DTA files by Bioworks 3.3 (Thermo Fisher
Scientific), and searched using the Mascot algorithm (Matrix Science) against the NCBInr
database with the taxonomy as metazoa (animals) for standard proteins and homo sapiens
(human) for IP proteins. The allowed mass tolerance was 0.5 Da for both precursor ions and
fragment ions, and the peptide charge state was 1+. The experimental mass values were
monoisotopic and the instrument was set as ESI-Trap. The enzyme was set as semi-trypsin with
the allowance of up to three missed cleavages. Initial variable modifications used included :
carbamidomethylation of Cys (+57 Da), oxidation of Met (+16), and phosphorylation of Ser, Thr,
and Tyr (+80). All results were verified manually with the MS/MS data. If no result could be
manually verified, then other variable modifications were considered for searching, including:
sodium adduct (+22), carbamidomethylation of Glu or Asp (+57), carbamylation of N-term or
Lys (+43), dethiomethylation of Met (-48), and pyro-glu from Gln (-17). In some instances,
peptides could not be identified by Mascot, but their MS/MS spectra showed product ions similar
to those from other identified peptides. For these cases, sequences were identified manually.

3.3 Results and discussion
3.3.1 Preparation and characterization of polymer-oxotitanium (P-oxoTi)
hybrid materials
The goal of preparing P-oxoTi hybrid films is to provide immobilized oxotitanium
clusters that present a high surface area for phosphopeptide enrichment. After forming a
95

Figure 3.2. Reflectance FT-IR spectra of a PHEMA film before (a) and after reaction with
succinic anhydride (b), and subsequent immersion in titanium isopropoxide (c).
-1

polyacid film on the Au-coated Si wafer (confirmed by the COOH absorption at ~3300 cm and
the enhanced C=O absorption in Figure 3.2b compared to 3.2a), the carboxylic acid containing
polymer brushes react with titanium isopropoxide to form oxotitanium clusters (Figure 3.1a).
Figure 3.2c confirms the formation of P-oxoTi species. After the reaction, absorption maxima
-1

-1

appear at 1560 cm (asymmetric stretch) and 1443 cm (symmetric stretch) due to the bidentate
bridging or bidentate chelation between carboxylate and Ti (Figure 3.2c).[31,32] The small
difference between the wave numbers of the two stretching modes rules out the presence of
unidentate coordination, which normally gives rise to a much larger peak separation. [33] The
-1

broad –COOH absorption near 3300 cm

also disappears after the reaction with Ti(i-PrO)4.
2

Elemental analysis by ICP-OES yielded a Ti surface coverage of 1.40.2 µ (304 nmol) /cm
g
96

wafer area in a 70 nm-thick P-oxoTi film. Assuming that the molecular weight of the polyacid
3

repeating unit is 230. and the polymer density is 1.0 g/cm in the dry form, the number of
polymer repeating units per Ti (or COOH/Ti ratio) is ~0.7.
Several studies report that polymer-oxotitanium network materials can form by
polymerization of monomer-capped oxotitanium clusters,[34-36] but the preparation of such
clusters typically requires careful control of hydrolysis and several weeks of reaction.[34,37] In
our approach, the P-oxoTi structure forms during simple immersion of the polymer- modified
wafer in a Ti(i-PrO)4 /THF solution for 30 min (with stirring). The carboxylate groups of the
polymer brushes likely replace some of the isopropoxy groups of Ti(i-PrO)4, and a small amount
of water may hydrolyze other isopropoxy groups to promote the formation of oxo clusters.[36]
Although the composition of the carboxylate-coordinated oxotitanium in the final hybrid
material is still not clear, the molar ratio of polymer repeating units (-COOH groups) to Ti atoms
(~0.7) suggests the formation of small carboxylate-containing oxotitanium clusters.

Some

reported carboxylate-substituted (six-coordinate) titanium polyoxoalkoxy clusters also show a
2-

low degree of condensation (0.33 < O /Ti < 1) along with 2~9 Ti atoms per cluster and
2-

carboxylate/Ti < 2.[35] O /Ti ratios less than 2 could favor the enrichment of phosphopeptides
by presenting many labile Ti species on the cluster surface. Although the Ti precursor, Ti(iPrO)4, hydrolyzes rapidly in aqueous solutions, the as-formed polymer-oxotitanium film is stable
in a 50 mM EDTA solution at pH 7 (no change in IR spectra after a 30-min immersion).
Because protein digests are usually acidified prior to phosphopeptide enrichment, we also
tested the short-term stability of the hybrid films in several acidic solutions. Covering the Au-P-

97

oxoTi surface for 30 min with 0.1% TFA removes 74% of the immobilized Ti. Unfortunately,
exposure to 2% TFA for 30 min removes 662% of the Ti from the film, showing that there is a
risk of losing phosphopeptides under this acidic condition. Coverage of the surface with 1%
H3PO4 for 5 min (mimicking the elution process) removes 163% of the Ti from the Au-PoxoTi brush. Thus, the majority of the Ti will remain in the polymer brush during elution.

3.3.2 Enrichment of model phosphopeptides on Au-P-oxoTi plates
We first examined phosphopeptide enrichment on Au-P-oxoTi using tryptic digests of
model phosphoproteins. This on-plate enrichment requires only 1 μL of sample, and rinsing
removes salts and nonphosphopeptides that can otherwise suppress the MS signals of
phosphopeptides. In initial trials, basic elution with pH 11 NH4OH produced MS signals similar
to those from elution with 1% H3PO 4. (Note that the eluate remains on the wafer for subsequent
on-plate analysis.) However, the matrix crystallization is inhomogeneous after basic elution,
probably because ammonium salts form upon mixing the acidic DHB matrix solution (which
contains H3PO 4 to prevent rebinding of phosphopeptides to the modified wafer after neutralizing
the basic eluate) with the NH4OH eluate. Thus, we chose 1% H3PO4 for phosphopeptide elution
in all subsequent work.
After Au-P-oxoTi enrichment of a mixture of digested -casein, -casein, and ovalbumin
(200 fmol of each) in 0.1% TFA, the MALDI mass spectrum of the captured species reveals 24
phosphopeptides and almost no nonphosphopeptides (Figure 3.3b). In contrast, direct MALDIMS of this

mixture without any enrichment yields detectable signals

for

many

nonphosphopeptides and only 5 phosphopeptides (Figure 3.3a). The signal-to-noise ratios (S/N)
98

Figure 3.3. MALDI mass spectra of model protein digests without phosphopeptide enrichment (a)
and after enrichment on Au-P-oxoTi MALDI plates (b-d). The amounts of digested proteins, and
concentrations of urea and ammonium salts in the sample (before enrichment) are given in each
spectrum. The TFA concentrations in the sample and washing (b-d only) solutions are also given.
Note that for spectrum d, the sample contains a 100- fold molar excess of BSA. MS peaks
labeled with , , and o represent phosphopeptides generated from -casein, -casein and
ovalbumin, respectively, and red symbols denote multi-phosphorylated peptides. Table 3.2
presents sequence information for each phosphopeptide. Peptides due to loss of one or more
H3PO4 units or (H2O+HPO 3) during/after ionization are labeled together with the intact
phosphopeptides using multiple arrows.

99

Figure 3.3 cont‟d

100

Table 3.2. Phosphopeptides observed in the positive- ion mode MALDI-MS spectra of tryptic casein, -casein, and ovalbumin digests. The phosphopeptide sequence* (if identified), number
+
of phosphoryl groups, and monoisotopic m/z of [M+H] are listed.
Protein
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

-casein

20
21
22
23
24
25
26

Phosphopeptide sequence**

#
–
PO 3H2
K.TVDMEpSTEVFTK.K
M(cam)-105
S2 153-164 1
K.VPQLEIVPNpSAEER.L
S1 121-134 1
K. VPQLEIVPNpSAEER.L
Carbamyl (N- S1 121-134 1
term)
K.VPQLEIVPNpSAEER.L
E(cam)
S1 121-134 1
YLGEYLIVPNpSAEER
S1***
1
K.DIGpSEpSTEDQAMEDIK.Q
M(cam)-105 S1 58-73
2
K.DIGpSEpSTEDQAMEDIK.Q
S1 58-73
2
sodium adduct, M(cam)-105
K. DIGpSEpSTEDQAMEDIK.Q
S1 58-73
2
Carbamyl (N-term), M(cam)-105
K.DIGpSEpSTEDQAMEDIK.Q
S1 58-73
2
K.DIGpSEpSTEDQAMEDIK.Q
S1 58-73
2
M(cam)-105, D(cam)
K.DIGpSEpSTEDQAMEDIK.Q
M(ox)
S1 58-73
2
K.YKVPQLEIVPNpSAEER.L
S1 119-134 1
K.DIGpSEpSTEDQAMEDIK.Q
Carbamyl S1 58-73
2
(K)
K.DIGpSEpSTEDQAMEDIK.Q
D(cam)
S1 58-73
2
K. YKVPQLEIVPN pSAEER.L Carbamyl (N- S1 119-134 1
term)
N/A
1
N/A
1
N/A
5
K.QMEAEpSIpSpSpSEEIVPNpSVEQK.H
S1 74-94
5
M(cam)-105, pyro-glu from Q
N/A
2
K.QMEAEpSIpSpSpSEEIVPNpSVEQK.H
S1 74-94
5
Pyro-glu from Q
K.QMEAEpSIpSpSpSEEIVPNpSVEQK.H
S1 74-94
5
N/A
4
N/A
3
R.NANEEEYSIGpSpSpSEEpSAEVATEEVK.I S2 61-85
4
N/A
S2
61- 5
****
85
Sequence
number

101

m/z
+
MH
1418.6
1660.8
1703.8
1717.8
1832.8
1879.7
1901.7
1922.7
1927.7
1936.7
1943.7
1952.0
1970.7
1984.7
1995.0
2027
2037
2575
2655.9
2672
2702.9
2720.9
2760
2819
3008.0
3088.0

Table 3.2 cont‟d

ovalb
umin

casein

1
2
3
4
5
6
o1

K.FQpSEEQQQTEDELQDK.I
K.FQpSEEQQQTEDELQDK.I
Sodium
adduct
K. FQpSEEQQQTEDELQDK.I
Carbamyl
(N-term)
K.FQpSEEQQQTEDELQDK.I E(cam)
R.ELEELNVPGEIVEpSLpSpSpSEESITR.I
A.RELEELNVPGEIVEpSLpSpSpSEESITR.I
R.EVVGpSAEAGVDAASVSEEFR.A

48-63
48-63

1
1

2061.8
2083.8

48-63

1

2104.8

48-63
17-40
16-40
341-360

1
4
4
1

2118.8
2966.2
3122.3
2088.9

* Most of these sequences were identified by Mascot, and all of the sequences were manually
verified. Some phosphopeptides share the same sequence, but are differently modified during
protein denaturing, alkylation, or digestion. To treat all digests equally, even though -casein
(composed of S1 and S2 chains) and -casein do not contain disulfide linkages, each protein was
digested separately using the entire digestion procedure. The overalkylation on Asp, Glu, and
Met were reported previously.[38,39]
** Some of the phosphopeptides were not sequenced (labeled as „N/A‟) due to their weak signals
in the MS full scan. However, their phosphorylation was confirmed according to their dominant
losses of 98 Da in tandem MS, and the number of phosphates that can be confirmed is noted.
*** This is a new sequence variant of the -S1 casein in the region 104–119.[15] The sequence
was manually verified.
**** This is probably a penta-phosphorylated peptide.[15]
of the five phosphopeptides also increase 2-20 fold after enrichment. However, with 0.1% TFA
in the protein digest and washing solutions, MS signals of multi-phosphorylated peptides
(labeled in red) are much weaker than signals of mono-phosphorylated peptides (labeled in
black, Figure 3.3b). Increasing the TFA level to 2% in both sample and washing solutions greatly
enhances signals of multi-phosphorylated peptides, especially those with high m/z, but the S/N
for some mono-phosphorylated peptides decrease by up to 80% (Figure 3.3c). The lower signals
of mono-phosphorylated peptides stem from both the weakened binding of phosphopeptides
(especially mono-phosphorylated peptides) and the loss of Ti in 2% TFA, whereas the higher
signals of multi-phosphorylated peptides likely result from higher ionization efficiency in the
presence of fewer mono-phosphorylated peptides.

102

In biological samples, e.g., whole cell extracts, phosphoproteins are generally much less
abundant than nonphosphoproteins. Even after IP, many proteins (including the antibody) are
typically present. Although electrophoresis can further fractionate such mixt ures, the amount of
targeted phosphopeptide obtained after purification will be low due to both the limited amount of
sample loaded for gel electrophoresis and incomplete recovery after in- gel digestion. Digestion
of targeted phosphoproteins directly on antibody beads can potentially increase the amount of
available phosphoprotein, but subsequent enrichment procedures must isolate phosphopeptides
from a sample containing more nonphosphopeptides than would be present following
electrophoresis and in- gel digestion. Moreover, the salt and urea concentrations in the on-bead
digest solution could be high to facilitate digestion of the greater amounts of proteins compared
to in-gel digestion.
To examine the effect of a large excess of nonphosphopeptides on enrichment, we
captured phosphopeptides from digest mixtures containing 200 fmol each of -casein, -casein,
and ovalbumin along with 20 pmol of BSA, a nonphosphoprotein. In this case, the presence of
0.1% TFA in the sample and washing solution is not sufficient to prevent capture of
nonphosphopeptides on the wafer (Figure 3.4a). Several studies suggest that addition of DHB to
sample and washing solutions improves TiO 2-based enrichment specificity.[9,15,16,22] With
the P-oxo-Ti, however, the presence of DHB during the enrichment of a -casein/BSA mixture
(molar ratio of 1/10) increased the number of nonphosphopeptide signals in the mass spectrum of
the enriched sample (Figures 4b and 4c). This suggests that phosphopeptides and DHB may
exhibit different binding modes on P-oxoTi films and TiO 2 particles.
Although the addition of DHB does not enhance enrichment selectivity for
phosphopeptides over nonphosphopeptides, increasing the TFA concentration effectively
103

Figure 3.4. MALDI mass spectra of model protein digests that were enriched on Au-P-oxoTi
wafers. The amount of protein digests and concentrations of urea, ammonium salts, TFA and
DHB (c only) during sample incubation are noted in each spectrum. MS peaks labeled with 
represent phosphopeptides generated from -casein, and red symbols denote multiphosphorylated peptides. Table 3.2 presents sequence information for each phosphopeptide.
Peptides due to loss of one or more H3PO4 units or (H2O+HPO 3) during/after ionization are
labeled together with the intact phosphopeptides using multiple arrows.
104

suppresses the binding of nonphosphopeptides to Au-P-oxoTi. As Figure 3.3d shows, after the
enrichment of phosphopeptides from digest mixtures containing 200 fmol of -casein, -casein,
ovalbumin, 20 pmol of BSA, and 2% TFA, phosphopeptides dominate the mass spectrum, and
only a few weak signals of nonphosphopeptides are present. Despite the removal of around 65%
of the Ti from the wafer in 2% TFA, signals from 25 phosphopeptides appear in the mass
spectrum.

Remarkably, even though the 1-L sample droplet contains only 23 ng of

phosphoprotein digests out of 1.4 g total protein digests, phosphopeptides dominate the mass
spectrum. Moreover, the concentrations of urea and ammonium salts in this sample are 1.6 M
and 0.11 M, respectively. In conventional MALDI-MS, inhibition of matrix crystallization by
such high concentrations of small compounds typically results in minimal peptide signals. After
enrichment from this relatively complex solution (Figure 3.3d), the mass spectrum is similar to
that after enrichment of the same amount of phosphoprotein digests in the absence of BSA digest
(using 2% TFA during incubation, Figure 3.3c). Moreover, for the sample that gives rise to
Figure 3c, the amounts of total protein (in mass), urea, and salts are also 60 times lower than
when the excess BSA digest is present (Figure 3.3d).
Even in digests containing 6 M urea and 0.4 M ammonium salts, enrichment is possible.
After Au-P-oxoTi enrichment, the MS spectrum of a 500-fmol -casein digest containing 6 M
urea, 0.4 M ammonium salts, and 0.1% TFA reveals both mono- and multi-phosphorylated
peptides (Figure 3.5a). The limit for detecting at least one phosphopeptide from -casein digests
containing such high concentrations of urea and ammonium salts is around 50 fmol in 1 µ
L
(S/N= 8, Figure 3.5b). However, the same detection limit for -casein digests containing 14- fold
diluted urea and ammonium decreases to about 15 fmol in 1 µ (S/N= 17, Figure 3.5c).
L
To quantify phosphopeptide capture and elution efficiencies under different sample
105

Figure 3.5. MALDI mass spectra of -casein digests that were enriched on Au-P-oxoTi wafers.
The amount of -casein digest and concentrations of urea, ammonium salts and TFA during
sample incubation are noted in each spectrum. MS peaks labeled with  represent
phosphopeptides generated from -casein, and red symbols denote multi-phosphorylated
peptides. Table 3.2 presents sequence information for each phosphopeptide. Peptides due to loss
of one or more H3PO 4 units or (H2O+HPO 3) during/after ionization are labeled together with the
intact phosphopeptides using multiple arrows.
106

complexities and acidities, we enriched a mixture of N-propionylated standard phosphopeptides
(CH3CH2CO-phosphopeptides, see Table 3.1 for peptide sequences) under several sample
conditions, and added these peptides‟ deuterated analogs, CD3CD2CO-phosphopeptides, to the
elution droplet (prior to matrix addition) as an internal standard. Using a calibration curve, the
percent recovery of each CH3CH2CO-phosphopeptide was calculated from the ratio of the MS
signals of the nondeuterated and deuterated phosphopeptides. Figure 3.6 shows that in a simple
solution of 0.1% TFA, the on-plate enrichment yields a recovery of 32-95% for the seven
phosphopeptides. Adding 1 pmol of BSA digest to the phosphopeptide mixture results in a 10:1
mass ratio of total nonphosphopeptides/total phosphopeptides in the sample and decreases
recovery by no more than 1/3. One mono-phosphorylated peptide and one multi-phosphorylated
peptide show an unexpected increase in recovery upon the addition of BSA digest. In 1% TFA,
all of the peptides show detectable signals, but increasing the TFA concentration to 2% and the
level of nonphosphoprotein (BSA) digest to 10 pmol causes the loss of nearly all the signals of
three mono-phosphorylated peptides.

Compared to mono-phosphorylated peptides, the bi-

phosphorylated peptides in Figure 3.6 show much higher recoveries under very acidic conditions
or in the presences of a large excess of BSA. We note that, the total amount of phosphopeptides
on the 2-mm-diameter enrichment spot is 4 pmol, which may be close to the binding capacity of
polymer brushes.[26]

Thus after the 30-min incubation in 0.1 or 1% TFA, the multi-

phosphorylated peptides may occupy the majority of affinity sites of P-oxoTi and decrease the
binding of mono-phosphorylated peptides. A similar phenomenon also occurred on the IMAC-Fe
polymer micro-spot as shown in Chapter 1. Future studies should examine recoveries as a
function of the amount of phosphopeptides in the sample.

107

Figure 3.6. Recoveries of CH3CH2CO-phosphopeptides from solutions with different TFA
concentrations and various amounts of BSA digest, urea, and ammonium salts. The number of
+
phosphates and the MH m/z value (mono- isotopic) of each CH3CH2CO-phosphopeptide are
noted below the columns. The sample contained 400 fmol of each CH3CH2CO-phosphopeptide.
For the enrichment of samples containing BSA digest, nonphosphopeptide signals were not
detected in the mass spectra. Table 3.1 give phosphopeptide sequences.

As the above results show, the main disadvantage of increasing the TFA concentration
(especially to above 1%) in the sample and washing solutions is the loss of signal from monophosphorylated peptides.

If the drop in signal in 2% TFA is largely due to removal of

oxotitanium clusters from the film, synthesis of polymer brushes that bind the clusters more
strongly should increase recoveries.

Such brushes could contain phosphate groups or

hydroxamic acids, or other hard ligands. Fortunately, 0.1% TFA is sufficient for excluding
nonphosphopeptides from a modestly complex protein digest sample (e.g. ~0.1 µ L
g/µ
nonphosphopeptides). For unknown protein digests, a good enrichment should provide not only a
„purified‟ mass spectrum, but also sufficient phosphopeptide signals for MS/MS sequencing.
108

Thus, for complex samples, we recommend on-plate enrichment with a series of TFA
concentrations from 0.1%-2% to optimize conditions for specific peptides. Because on-plate
enrichment requires only 1 L of sample and the enrichment process can occur in parallel for a
series of samples, this procedure will not greatly increase the analysis time, and the required
amount of sample will still be relatively small.

3.3.3 Comparison with other enrichment methods
IMAC is probably the most widely used technique for isolating phosphopeptides, so we
compared enrichment on the Au-P-oxoTi wafer with a previously developed IMAC on-plate
enrichment technique that uses gold surfaces coated with polymer brushes derivatized with
nitrilotriacetate-Fe(III) complexes, Au-P-NTA-Fe.[26] Synthesis of this IMAC material starts
from the same polymer used for Au-P-oxoTi but ends with functionalization with NTA-Fe(III).
A previous study by our group[26] and Chapter 2 showed that Au-P-NTA-Fe can enrich both
mono- and multi-phosphorylated peptides from a simple -casein tryptic digest. When tested
with a mixture of three model phosphoprotein digests, however, the IMAC wafer reveals 17
phosphopeptides, but all of them are multiply phosphorylated (Figure 3.7a). In contrast, of the
24 phosphopeptides found on Au-P-oxoTi using the same sample and the same enrichment
conditions, 11 are mono-phosphorylated peptides (Figure 3.3b). For a more complex mixture of
three phosphoprotein digests (200 fmol of each) and BSA digest (20 pmol), increasing the TFA
concentration to 1% in the sample and washing solutions can exclude most of the
nonphosphopeptides from the Au-P-NTA-Fe wafer (Figure 3.7b), but nearly all of the detected
phosphopeptides are multiply phosphorylated. The number of detected phosphopeptides under
these conditions, however, is still fewer than with enrichment on Au-P-oxoTi using 2% TFA
(compare Figures 7b and 3d, 19 vs. 25 total phosphopeptides, and 1 vs. 8 mono-phosphorylated
109

Figure 3.7. MALDI mass spectra of protein digest mixtures after on-plate IMAC enrichment
using Au-P-NTA-Fe (a, b) and in-solution enrichment on TiO 2 particles (c, d). The amounts of
digested proteins, sample volumes (TiO 2 enrichment only), and the TFA concentrations in
sample and washing solutions are noted in each spectrum. In TiO 2 enrichment, a 1 L aliquot of
the 1% H3PO 4 eluate (out of 50 L and 75 L total eluate for c and d, respectively) was spotted
on a stainless steel MALDI plate for MS analysis, but the initial solution contained
correspondingly more phosphoprotein digests than the on-plate enrichment to compensate for the
small fractional amount spotted. MS peaks labeled with , , and o represent phosphopeptides
generated from -casein, -casein and ovalbumin, respectively, and red symbols denote multiphosphorylated peptides. Table 3.2 presents sequence information for each phosphopeptide.
Peptides due to loss of one or more H3PO4 units or (H2O+HPO3) during/after ionization are
labeled together with the intact phosphopeptides using multiple arrows.

110

Figure 3.7 cont’d

111

peptides for Au-P-NTA-Fe and Au-P-oxoTi, respectively).
We also tested the enrichment of phosphopeptides with commercial TiO 2 particles
(Titansphere, 5 m particle size). Dispersing the micron-size TiO 2 particles in 80 L of solution
containing 0.1% TFA, 50% ACN, and three digested phosphoproteins (10 pmol of each)
followed by washing and elution allows MALDI-MS detection of 7 mono-phosphorylated
peptides, 1 multi-phosphorylated peptide, and some non-phosphopeptides (Figure 3.7c). For
samples containing higher levels of nonphosphopeptides, increasing the TFA concentration to 2%
without using DHB as a non-phosphopeptide excluder (which may cause the loss of
phosphopeptides), again, gives mass spectra dominated by phosphopeptides (Figure 3.7d).
However, the detected 10 phosphopepitdes are all mono-phosphorylated.

Under similar

conditions, the Au-P-oxoTi enrichment gives signals from 25 phosphopeptides, 17 of which are
multiply phosphorylated (Figure 3.2d).

3.3.4 Analysis of immunoprecipitated p65-associated phosphoproteins.
Building on our success in phosphopeptide enrichment from tryptic digests of standard
proteins, we turned to the study of phosphorylation of p65-associated proteins in the THP-1 cell
nucleus. THP-1 cells activate the NF-κB pathway upon stimulation with TNF-,[40] and
Western blot analysis of p65-specific immunoprecipitates shows that the TNF- treatment
significantly increases the amount of nuclear p65 (data not shown).
After IP of p65 from the nuclear extract (Western blotting shows that the IP effectively
pulls down p65, Figure 3.8), we denatured and digested p65 and associated proteins directly on
the antibody beads, and then enriched the phosphopeptides from these digests on Au-P-oxoTi
plates. Figure 3.9 shows a representative MALDI mass spectrum of a phosphopeptide-enriched
112

Figure 3.8. p65 Western blotting (using C-22B4 rabbit polyclonal antibody) of 25 µ of THP-1
g
cell nuclear extract (NE) before (right lane) and after (left lane) IP.

Figure 3.9. MALDI mass spectrum of a Au-P-oxoTi-enriched trypsin-digest of p65immunoprecipitated nuclear extract from TNF-α-treated THP-1 cells, where * = identified
phosphopeptides; ♦ = identified nonphosphopeptides; ? = unidentified peptides; and m = matrix
ions. The sample contained 6 M urea, 0.4 M ammonium salts, and 0.1% TFA.
113

tryptic digest from the immunoprecipitate of a nuclear extract from TNF-α-treated THP-1 cells.
Under the 0.1% TFA enrichment condition, nearly all of the signals in the mass spectrum stem
from peptides (indicated by CID-MS/MS product ion spectra). Moreover, phosphopeptides
(indicated by the predominant neutral loss of 98 Da in CID-MS/MS[41]) dominate the spectrum.
The only exceptions include ions corresponding to matrix clusters (labeled as “m”), which are
also present in the mass spectra of some standard protein digests (data not shown). Because the
nonphosphopeptide signals are fewer and weaker than phosphopeptide signals in the mass
spectrum, we did not further increase the TFA concentration in enrichment solutions.

Table 3.3. Peptides identified from the Au-P-oxoTi enriched digests of the nuclear extracts from
TNF-α-treated THP-1 cells.*
m/z
(mono)

Peptide sequence

Protein

Mascot
score

1629

KEEpSEEpSDDDM(cam-105)GF

P1 or P2 isoform of the 60S acidic
ribosomal protein

18

1677
1693
1734
2061
2109
2125
2166
1367

KEEpSEEpSDDDMGF
KEEpSEEpSDDDM(ox)GF
KEEpSEEpSDDDM(cam)GF
KEEpSEEpSDDDM(cam105)GFGLFD
KEEpSEEpSDDDMGFGLFD
KEEpSEEpSDDDM(ox)GFGLFD
KEEpSEEpSDDDM(cam)GFGLFD
pSRpSFDYNYR

as above
as above
as above
as above
as above
as above
as above
human translocation liposarcomaassociated serine-arginine protein

28
N/A
N/A
10
11
N/A
N/A
15

1191
1022

HMYHSLYLK
AAIDWFDGK

ribosomal protein L19
translocation liposarcoma protein

36
39

1515
IWHHTFYNELR
actin (β isoform)
73
* Some of these sequences were identified by Mascot, and all of the sequences were manually
verified. M(cam) corresponds to S-carbamidomethylmethionine, M(cam-105) corresponds to the

114

neutral loss of S-(methylthio)acetamide from M(cam), and M(ox) corresponds to methionine
sulfoxide.

Table 3.3 lists the sequences of the peptides that we identified using the information
obtained by Figure 3.9, CID-MS/MS product ion spectra (Figure 3.10), guided Mascot database
search, and manual spectra interpretation. Among the three identified phosphopeptides
(presented in different modified forms in Table 3.2), KEEpSEEpSDDDMGF is a truncated
version of KEEpSEEpSDDDMGFGLFD. This truncated peptide may result from chymotrypsin
proteolysis during sample preparation, incomplete translation of the C-terminus of the protein, or
proteolytic processing occurring in the cells prior to extraction. It should be noted that, all three
distinct phosphopeptides (and their modified isoforms resulting from sample preparation) are
doubly phosphorylated, and two of those phosphopeptides are highly acidic. Thus these peptides
should have especially high affinity for the oxo-Ti clusters. Other phosphopeptides, particularly
mono-phosphorylated peptides, might be present in the sample but not enriched. However,
under the same salt and TFA conditions, enrichment of phosphopeptides from a -casein digest
revealed both mono- and multi-phosphorylated peptides with similar MS intensities (Figure 3.5a).
We also identified several non-phosphopeptides. Capture of these peptides on Au-P-oxoTi likely
stemmed from non-specific adsorption, but the corresponding parent proteins are probably
specifically associated with p65.
Because the proteins identified

in the immunoprecipitate have other k nown

phosphorylation sites, and the phosphorylation of p65 may also occur, we searched for other
tryptic phosphopeptides of these proteins using the ProteinProspector MS-Digest Program.
However, we could not identify additional phosphopeptides from this IP digest sample. Possible
explanations for the absence of other phosphopeptides include insufficient levels of

115

Figure 3.10. CID-MS/MS product ion spectra of m/z 2061 (a), 2109 (b), 2125 (c), 2166 (d), 1629 (e), 1677 (f), 1693 (g), 1734
(h), and 1367 (i) (from the MALDI mass spectrum shown in Figure 3.9). These peptides were identified as various isoforms of
KEEpSEEpSDDDMGFGLFD (a-d), KEEpSEEpSDDDMGF (e-h), and pSRpSFDYNYR (i). Δ = –98 Da (–H3PO4 or –
(H2O+HPO 3)); □ = –80 Da (–HPO3); o = –18 Da (–H2O); ‡ = –64 Da (–CH3SOH or methane sulfenic acid).

116

Figure 3.10 cont‟d

117

Figure 3.10 cont‟d

118

Figure 3.10 cont‟d

119

Figure 3.10 cont‟d

120

phosphorylation, ineffective enrichment, elution, or ionization, and masses above or below the
user-defined mass range (m/z 800-4000).

The latter reason likely explains the lack of

phosphopeptides from p65 itself. Among the p65 tryptic peptides that contain experimentally
verified phosphorylation sites (we generated digested peptides using the ProteinProspector MSDigest Program with no missed cleavages), only one has a mass between 800-4000 Da. Ryo et
al. showed that this peptide is phosphorylated at Thr254 using immunoblotting.[6]

Figure 3.11. MALDI mass spectra of the samples resulting from (a) the trypsin-digested,
ZipTipC18-desalted, p65-immunoprecipitated nuclear extracts from TNF-α-treated THP-1 cells
(20 L of the IP digest sample was used for desalting, and 2/5 of the tip eluate was spotted on
MALDI steel plate for conventional analysis), and (b) the Au-P-oxoTi enriched trypsin digest of
the p65-IP beads exposed to a PBS blank. None of the observable strong signals in (a) and (b)
showed dominant neutral loss of 98 upon CID-MS/MS (spectra not shown). Signals due to
matrix clusters are noted as “m”.

121

In a control experiment, we desalted the TNF-α-treated IP sample using ZipTipC18
pipette tips and performed conventional MALDI-MS (Figure 3.11a, no phosphopeptide
enrichment). As expected, this approach revealed no phosphopeptides. A separate control
experiment carried out by performing IP from a PBS blank solution also gave no significant
peptide signals in the mass spectrum after on-bead digestion and Au-P-oxoTi enrichment (Figure
3.11b).

3.3.5 Semi-quantitative analysis of p65-associated phosphoproteins upon TNF treatment
For comparison with the TNF-α-treated cell sample, we performed the entire sample
preparation and purification process with untreated THP-1 cells.

Since p65 is present in

significantly lower abundance in the nucleus without TNF-α treatment, any protein that has
specific interactions with p65 (rather than nonspecific interactions with the p65 antibody) should
be present in lower abundance in the nucleus of an untreated sample than in the treated sample,
assuming similar enrichment efficiencies. After parallel IP, digestion, on-plate enrichment, and
elution of both TNF-α-treated and untreated cells, 100 amol (in 0.5 L) of an internal standard
phosphopeptide was added to each sample spot followed by addition of matrix. Figure 3.12
presents the resulting MS spectra for the two samples and shows that the relative signal
intensities (the S/N of phosphopeptide divided by the S/N of internal standard) of the
phosphopeptide ions, m/z 2061 and 2109, identified from the TNF-α-treated sample are 13 and 4
times higher, respectively, than in the untreated sample. Thus, the proteins from which these
phosphopeptides come are most likely specifically associated with p65 in the cell.

122

Figure 3.12. MALDI mass spectra of the (a) TNF-α-treated THP-1 cell and (b) untreated THP-1
cell samples resulting from nuclear extraction, p65 IP, trypsin digestion, phosphopeptide
enrichment, and addition of 100 amol of an internal standard phosphopeptide (CD3CD2COLFTGHPEpSLEK), where * = identified phosphopeptides, and m = matrix ions.

3.3.6 Relevance of the identified proteins co-immunoprecipitated with p65
Table 3.3 lists the proteins we identified from the p65 immunoprecipitate. Ribosomal
proteins are typically involved in cytoplasmic translational processes, but a number of these
proteins have “extra-ribosomal” functions, including regulation of gene transcription, mRNA
translation outside of the ribosome, and DNA repair. [42,43] This work identified 60S acidic
ribosomal protein from the P1 or P2 isoforms as a phosphoprotein associated with p65. In a
previous study, Bouwmeester and coworkers co-purified nine ribosomal proteins and four other
proteins involved in ribosomal function with p65.[44] The large subunit (60S) of eukaryotic
ribosomes contains the P0, P1, and P2 isoforms of the 60S acidic ribosomal protein (collectively
123

known as the “P-proteins”), whose sequences are highly conserved across species.[45] Both the
P1 and P2 proteins participate in peptide synthesis and interact with elongation factors 1 and 2,
which facilitate translational elongation.[46-48] Importantly, the P2 isoform of the 60S acidic
ribosomal protein co- immunoprecipitated with key players of the NF-κB pathway, including IκB
kinase (γ subunit), p52, NF-κB-inducing kinase (NIK), and RelB, indicating that it too plays a
potentially important role in this pathway.[44] However, no reports described the association of
the P-proteins directly with p65. Consequently, the association between the P1 or P2 isoform of
the 60S acidic ribosomal protein and p65 demonstrated here substantiates a function for these
proteins in the NF-κB pathway. RPL19, another identified protein in this study, is part of the
eukaryotic 60S ribosomal subunit and its exact function has yet to be elucidated. As far as we
are aware, this study provides the first evidence for interaction between RPL19 and p65. Similar
to the association between the ribosomal P-protein and p65, this interaction could be cytoplasmic
or nuclear.
A recent study shows that human translocation liposarcoma (TLS)-associated serinearginine protein (TASR) enhances NF-κB mediated transactivation induced by TNF-α,
interleukin-1β, and overexpression of NIK by its interaction with the p65 subunit.[49] This is
consistent with the TASR and TLS co- immunoprecipitation with nuclear p65 that we observe.
TASR serves as a splicing factor, processing pre- mRNA, and interacts with TLS.[50,51]
Specifically, the C-terminal region (residues 266 to the C-terminal residue 526) of TLS links
gene transcription and RNA splicing by recruitment of TASR to RNA polymerase II.[50,51]
However, when TLS constitutes part of an oncogenic fusion protein, where its C-terminal region
is replaced with another protein, it cannot recruit TASR to RNA-polymerase II.[50,52] In our
study, the TLS peptide constitutes residues 349-357, indicating that TLS is likely present in its

124

wild-type form. Collectively these results suggest that the TASR protein co- immunoprecipitated
with p65 may result from a higher order complex of p65, TLS, TASR, and RNA polymerase II.
The last identified protein is actin, which is a highly conserved major component of the
cytoskeleton and plays a critical role in a variety of essential cellular processes.[53,54]
Consistent with our results, several studies demonstrated interactions between p65 and actin.[5557]

However, since actin is also a nuclear protein that interacts with transcriptional

machinery,[58] and the samples prepared here are a nuclear fraction, the p65-actin interaction
proposed in this study may suggest a nuclear role for p65 regulation.

3.4 Conclusion
IP of target proteins from a cell nuclear extract, followed by tryptic digestion directly on
antibody beads and enrichment of phosphopeptides using Au-P-oxoTi reveals the association of
a number of proteins with p65. The on-plate phosphopeptide enrichment strategy consumes only
1 L of protein digest for MALDI-MS analysis, and provides signals from both mono- and
multi-phosphorylated peptides of complex model protein digests. Overall, the Au-P-oxoTi wafer
successfully enriched nine phosphopeptides from trypsin-digested, p65- immunoprecipitated
nuclear extracts from TNF--treated THP-1 cells. CID-MS/MS enabled the identification of the
sequences and possible phosphorylation sites of the nine phosphopeptides along with the
sequences of three nonphosphopeptides. The phosphopeptides result from five p65-associated
proteins, two of which (RPL19 and the 60S acidic ribosomal protein from the P1 or P2 isoforms)
were not previously known to interact with p65. This study also confirms previous results that
implicate a role for TASR, TLS, and actin, in the NF-κB pathway. Additionally, the use of

125

internal standards allows for the comparative analysis of the extent of protein association with
p65 in treated and untreated cells. Collectively, these results broaden the perspective on the
processes involved in the NF-κB pathway.

126

REFERENCES

127

REFERENCES
1.
Perkins, N. D., Integrating Cell-Signalling Pathways with Nf-Kappa B and Ikk Function.
Nat. Rev. Mol. Cell Biol. 2007, 8, 49-62.
2.
Baeuerle, P. A.; Henkel, T., Function Amd Activation of Nf-Kappa-B in the ImmuneSystem. Annu. Rev. Immunol. 1994, 12, 141-179.
3.
Wang, C. Y.; Mayo, M. W.; Baldwin, A. S., Tnf- and Cancer Therapy-Induced Apoptosis:
Potentiation by Inhibition of Nf-Kappa B. Science 1996, 274, 784-787.
4.
Schmitz, M. L.; Mattioli, I.; Buss, H.; Kracht, M., Nf-Kappa B: A Multifaceted
Transcription Factor Regulated at Several Levels. ChemBioChem 2004, 5, 1348-1358.
5.
Scherer, D. C.; Brockman, J. A.; Chen, Z. J.; Maniatis, T.; Ballard, D. W., SignalInduced Degradation of Iκbα Requires Site-Specific Ubiquitination. Proc. Natl. Acad. Sci. U. S.
A. 1995, 92, 11259-11263.
6.
Ryo, A.; Suizu, F.; Yoshida, Y.; Perrem, K.; Liou, Y. C.; Wulf, G.; Rottapel, R.;
Yamaoka, S.; Lu, K. P., Regulation of Nf-Κb Signaling by Pin1-Dependent Prolyl Isomerization
and Ubiquitin-Mediated Proteolysis of P65/Rela. Mol. Cell 2003, 12, 1413-1426.
7.
Yeh, P. Y.; Yeh, K. H.; Chuang, S. E.; Song, Y. C.; Cheng, A. L., Suppression of
Mek/Erk Signaling Pathway Enhances Cisplatin-Induced Nf-Kappa B Activation by Protein
Phosphatase 4-Mediated Nf-Kappa B P65 Thr Dephosphorylation. J. Biol. Chem. 2004, 279,
26143-26148.
8.
Shen, C. H.; Stavnezer, J., Interaction of Stat6 and Nf-Κb: Direct Association and
Synergistic Activation of Interleukin-4-Induced Transcription. Mol. Cell. Biol. 1998, 18, 33953404.
9.
Dunn, J. D.; Reid, G. E.; Bruening, M. L., Techniques for Phosphopeptide Enrichment
Prior to Analysis by Mass Spectrometry. Mass Spectrom. Rev. 2010, 29, 29-54.
10.
Paradela, A.; Albar, J. P., Advances in the Analysis of Protein Phosphorylation. J.
Proteome Res. 2008, 7, 1809-1818.
11.
Rogers, L. D.; Foster, L. J., Phosphoproteomics-Finally Fulfilling the Promise? Mol.
Biosyst. 2009, 5, 1122-1129.
12.
Schmidt, S. R.; Schweikart, F.; Andersson, M. E., Current Methods for Phosphoprotein
Isolation and Enrichment. J. Chromatogr. B 2007, 849, 154-162.

128

13.
Cantin, G. T.; Shock, T. R.; Park, S. K.; Madhani, H. D.; Yates, J. R., Optimizing TiO2Based Phosphopeptide Enrichment for Automated Multidimensional Liquid Chromatography
Coupled to Tandem Mass Spectrometry. Anal. Chem. 2007, 79, 4666-4673.
14.
Kyono, Y.; Sugiyama, N.; Imami, K.; Tomita, M.; Ishihama, Y., Successive and
Selective Release of Phosphorylated Peptides Captured by Hydroxy Acid-Modified Metal Oxide
Chromatography. J. Proteome Res. 2008, 7, 4585-4593.
15.
Larsen, M. R.; Thingholm, T. E.; Jensen, O. N.; Roepstorff, P.; Jorgensen, T. J. D.,
Highly Selective Enrichment of Phosphorylated Peptides from Peptide Mixtures Using Titanium
Dioxide Microcolumns. Mol. Cell. Proteomics 2005, 4, 873-886.
16.
Qiao, L.; Roussel, C.; Wan, J.; Yang, P.; Girault, H. H.; Liu, B., Specific on-Plate
Enrichment of Phosphorylated Peptides for Direct MALDI-Tof MS Analysis. J. Proteome Res.
2007, 6, 4763-4769.
17.
Rainer, M.; Sonderegger, H.; Bakry, R.; Huck, C. W.; Morandell, S.; Huber, L. A.;
Gjerde, D. T.; Bonn, G. K., Analysis of Protein Phosphorylation by Monolithic Extraction
Columns Based on Poly(Divinylbenzene) Containing Embedded Titanium Dioxide and
Zirconium Dioxide Nano-Powders. Proteomics 2008, 8, 4593-4602.
18.
Wan, J. J.; Qian, K.; Qiao, L.; Wang, Y. H.; Kong, J. L.; Yang, P. Y.; Liu, B. H.; Yu, C.
Z., TiO 2-Modified Macroporous Silica Foams for Advanced Enrichment of MultiPhosphorylated Peptides. Chem.-Eur. J. 2009, 15, 2504-2508.
19.
Zhou, H.; Ye, M.; Dong, J.; Han, G.; Jiang, X.; Wu, R.; Zou, H., Specific Phosphopeptide
Enrichment with Immobilized Titanium Ion Affinity Chromatography Adsorbent for
Phosphoproteome Analysis. J. Proteome Res. 2008, 7, 3957-3967.
20.
Bi, H. Y.; Qiao, L.; Busnel, J. M.; Devaud, V.; Liu, B. H.; Girault, H. H., TiO2 Printed
Aluminum Foil: Single-Use Film for a Laser Desorption/Ionization Target Plate. Anal. Chem.
2009, 81, 1177-1183.
21.
Blacken, G. R.; Voln, M.; Diener, M.; Jackson, K. E.; Ranjitkar, P.; Maly, D. J.; Turecek,
F., Reactive Landing of Gas-Phase Ions as a Tool for the Fabrication of Metal Oxide Surfaces for
in Situ Phosphopeptide Enrichment. J. Am. Soc. Mass Spectrom. 2009, 20, 915-926.
22.
Torta, F.; Fusi, M.; Casari, C. S.; Bottani, C. E.; Bachi, A., Titanium Dioxide Coated
MALDI Plate for on Target Analysis of Phosphopeptides. J. Proteome Res. 2009, 8, 1932-1942.
23.
Bennion, B. J.; Daggett, V., The Molecular Basis for the Chemical Denaturat ion of
Proteins by Urea. Proc. Natl. Acad. Sci. U. S. A. 2003, 100, 5142-5147.
24.
Villen, J.; Gygi, S. P., The Scx/IMAC Enrichment Approach for Global Phosphorylation
Analysis by Mass Spectrometry. Nat. Protoc. 2008, 3, 1630-1638.
129

25.
Jensen, S. S.; Larsen, M. R., Evaluation of the Impact of Some Experimental Procedures
on Different Phosphopeptide Enrichment Techniques. Rapid Commun. Mass Spectrom. 2007, 21,
3635-3645.
26.
Dunn, J. D.; Igrisan, E. A.; Palumbo, A. M.; Reid, G. E.; Bruening, M. L.,
Phosphopeptide Enrichment Using MALDI Plates Modified with High-Capacity Polymer
Brushes. Anal. Chem. 2008, 80, 5727-5735.
27.
Sun, L.; Dai, J.; Baker, G. L.; Bruening, M. L., High-Capacity, Protein-Binding
Membranes Based on Polymer Brushes Grown in Porous Substrates. Chem. Mater. 2006, 18,
4033-4039.
28.
Palumbo, A. M. Development and Application of Tandem Mass Spectrometry Methods
for Phosphoprotein Analysis (Ph.D Dissertation). Michigan State University, East Lansing, 2009.
29.
MacCoss, M. J.; McDonald, W. H.; Saraf, A.; Sadygov, R.; Clark, J. M.; Tasto, J. J.;
Gould, K. L.; Wolters, D.; Washburn, M.; Weiss, A.; Clark, J. I.; Yates, J. R., Shotgun
Identification of Protein Modifications from Protein Complexes and Lens Tissue. Proc. Natl.
Acad. Sci. U. S. A. 2002, 99, 7900-7905.
30.
Wilson-Grady, J. T.; Villen, J.; Gygi, S. P., Phosphoproteome Analysis of Fission Yeast.
J. Proteome Res. 2008, 7, 1088-1097.
31.
Doeuff, S.; Dromzee, Y.; Taulelle, F.; Sanchez, C., Synthesis and Solid-State and LiquidState Characterization of a Hexameric Cluster of Titanium(Iv): Ti6(µ -O)2(µ -O)2(µ 2
3
2
Oc4h9)2(Oc4h9)6(Ococh3)8. Inorg. Chem. 1989, 28, 4439-4445.
32.
Sui, R. H.; Thangadurai, V.; Berlinguette, C. P., Simple Protocol for Generating TiO2
Nanofibers in Organic Media. Chem. Mater. 2008, 20, 7022-7030.
33.
Deacon, G. B.; Phillips, R. J., Relationships between the Carbon-Oxygen Stretching
Frequencies of Carboxylato Complexes and the Type of Carboxylate Coordination. Coord. Chem.
Rev. 1980, 33, 227-250.
34.
Moraru, B.; Husing, N.; Kickelbick, G.; Schubert, U.; Fratzl, P.; Peterlik, H., InorganicOrganic Hybrid Polymers by Polymerization of Methacrylate- or Acrylate-Substituted
Oxotitanium Clusters with Methyl Methacrylate or Methacrylic Acid. Chem. Mater. 2002, 14,
2732-2740.
35.
Rozes, L.; Steunou, N.; Fornasieri, G.; Sanchez, C., Titanium-Oxo Clusters, Versatile
Nanobuilding Blocks for the Design of Advanced Hybrid Materials. Mon. Chem. 2006, 137, 501528.

130

36.
Schubert, U., Chemical Modification of Titanium Alkoxides for Sol-Gel Processing. J.
Mater. Chem. 2005, 15, 3701-3715.
37.
Kickelbick, G.; Holzinger, D.; Brick, C.; Trimmel, G.; Moons, E., Hybrid InorganicOrganic Core-Shell Nanoparticles from Surface-Functionalized Titanium, Zirconium, and
Vanadium Oxo Clusters. Chem. Mater. 2002, 14, 4382-4389.
38.
Boja, E. S.; Fales, H. M., Overalkylation of a Protein Digest with Iodoacetamide. Anal.
Chem. 2001, 73, 3576-3582.
39.
Lapko, V. N.; Smith, D. L.; Smith, J. B., Identification of an Artifact in the Mass
Spectrometry of Proteins Derivatized with Iodoacetamide. J. Mass Spectrom. 2000, 35, 572-575.
40.
Lu, Z. G.; Liu, H. S.; Yamaguchi, T.; Miki, Y.; Yoshida, K., Protein Kinase Cδ Activates
Rela/P65 and Nuclear Factor-Κb Signaling in Response to Tumor Necrosis Factor-Α. Cancer
Res. 2009, 69, 5927-5935.
41.
Mann, M.; Ong, S.-E.; Gronborg, M.; Steen, H.; Jensen, O. N.; Pandey, A., Analysis of
Protein Phosphorylation Using Mass Spectrometry: Deciphering the Phosphoproteome. Trends
Biotechnol. 2002, 20, 261-268.
42.
Lindstrom, M. S., Emerging Functions of Ribosomal Proteins in Gene-Specific
Transcription and Translation. Biochem. Biophys. Res. Commun. 2009, 379, 167-170.
43.
Wool, I. G., Extraribosomal Functions of Ribosomal Proteins. Trends Biochem. Sci. 1996,
21, 164-165.
44.
Bouwmeester, T.; Bauch, A.; Ruffner, H.; Angrand, P. O.; Bergamini, G.; Croughton, K.;
Cruciat, C.; Eberhard, D.; Gagneur, J.; Ghidelli, S.; Hopf, C.; Huhse, B.; Mangano, R.; Michon,
A. M.; Schirle, M.; Schlegl, J.; Schwab, M.; Stein, M. A.; Bauer, A.; Casari, G.; Drewes, G.;
Gavin, A. C.; Jackson, D. B.; Joberty, G.; Neubauer, G.; Rick, J.; Kuster, B.; Superti-Furga, G.,
A Physical and Functional Map of the Human Tnf- Α/Nf-Κb Signal Transduction Pathway. Nat.
Cell Biol. 2004, 6, 97-105.
45.
Rich, B. E.; Steitz, J. A., Human Acidic Ribosomal Phosphoproteins P0, P1, and P2:
Analysis of Cdna Clones, in Vitro Synthesis, and Assembly. Mol. Cell. Biol. 1987, 7, 4065-4074.
46.
Gomez-Lorenzo, M. G.; Spahn, C. M. T.; Agrawal, R. K.; Grassucci, R. A.; Penczek, P.;
Chakraburtty, K.; Ballesta, J. P. G.; Lavandera, J. L.; Garcia-Bustos, J. F.; Frank, J., ThreeDimensional Cryo-Electron Microscopy Localization of Ef2 in the Saccharomyces Cerevisiae
80s Ribosome at 17.5 Ǻ Resolution. EMBO J. 2000, 19, 2710-2718.
47.
Macconnell, W. P.; Kaplan, N. O., The Activity of the Acidic Phosphoproteins fro m the
80 S Rat-Liver Ribosome. J. Biol. Chem. 1982, 257, 5359-5366.

131

48.
Sanchezmadrid, F.; Reyes, R.; Conde, P.; Ballesta, J. P. G., Acidic Ribosomal-Proteins
from Eukaryotic Cells - Effect on Ribosomal Functions. Eur. J. Biochem. 1979, 98, 409-416.
49.
Uranishi, H.; Tetsuka, T.; Yamashita, M.; Asamitsu, K.; Shimizu, M.; Itoh, M.; Okamoto,
T., Involvement of the Pro-Oncoprotein Tls (Translocated in Liposarcoma) in Nuclear Factor-Κb
P65-Mediated Transcription as a Coactivator. J. Biol. Chem. 2001, 276, 13395-13401.
50.
Yang, L.; Embree, L. J.; Hickstein, D. D., Tls-Erg Leukemia Fusion Protein Inhibits Rna
Splicing Mediated by Serine-Arginine Proteins. Mol. Cell. Biol. 2000, 20, 3345-3354.
51.
Yang, L.; Embree, L. J.; Tsai, S.; Hickstein, D. D., Oncoprotein Tls Interacts with SerineArginine Proteins Involved in Rna Splicing. J. Biol. Chem. 1998, 273, 27761-27764.
52.
Yang, L.; Chansky, H. A.; Hickstein, D. D., Ews·Fli-1 Fusion Protein Interacts with
Hyperphosphorylated Rna Polymerase Ii and Interferes with Serine-Arginine Protein-Mediated
Rna Splicing. J. Biol. Chem. 2000, 275, 37612-37618.
53.
Carlier, M. F.; Le Clainche, C.; Wiesner, S.; Pantaloni, D., Actin-Based Motility: From
Molecules to Movement. Bioessays 2003, 25, 336-345.
54.
Chen, H.; Bernstein, B. W.; Bamburg, J. R., Regulating Actin-Filament Dynamics in
Vivo. Trends Biochem. Sci. 2000, 25, 19-23.
55.
Are, A. F.; Galkin, V. E.; Pospelova, T. V.; Pinaev, G. P., The P65/Rela Subunit of NfΚb Interacts with Actin-Containing Structures. Exp. Cell Res. 2000, 256, 533-544.
56.
Fazal, F.; Bijli, K. M.; Minhajuddin, M.; Rein, T.; Finkelstein, J. N.; Rahman, A.,
Essential Role of Cofilin-1 in Regulating Thrombin- Induced Rela/P65 Nuclear Translocation and
Intercellular Adhesion Molecule 1 (Icam-1) Expression in Endothelial Cells. J. Biol. Chem. 2009,
284, 21047-21056.
57.
Fazal, F.; Minhajuddin, M.; Bijli, K. M.; McGrath, J. L.; Rahman, A., Evidence for Actin
Cytoskeleton-Dependent and -Independent Pathways for Rela/P65 Nuclear Translocation in
Endothelial Cells. J. Biol. Chem. 2007, 282, 3940-3950.
58.
Zheng, B.; Han, M.; Bernier, M.; Wen, J. K., Nuclear Actin and Actin- Binding Proteins
in the Regulation of Transcription and Gene Expression. FEBS J. 2009, 276, 2669-2685.

132

Chapter 4. Low-bias Enrichment of Mono- and Multiphosphorylated Peptides Using Bifunctional Polymer
Brushes

4.1 Introduction
Although several materials extract phosphopeptides from complex protein digests, these
materials recover unique, partially overlapping sets of phosphopeptides, and no single
enrichment method is comprehensive enough for phosphoproteome analysis.[1] The two most
commmon

phosphopeptide

enrichment

techniques,

immobilized

metal

ion

affinity

chromatography (IMAC) and metal oxide affinity chromatography (MOAC), often show high
specificity for multi-phosphorylated and mono-phosphorylated peptides, respectively.[2] Many
enrichment strategies attempt to simultaneously capture and elute mono- and multiphosphorylated peptides using a single affinity material (e.g. TiO 2 or NTA-Fe(III)),[2] but such
protocols typically lead to biased recoveries for either mono- or multi-phosphorylated peptides in
the following MS analysis.
A few recent studies with MOAC and IMAC show that the binding capacity of the
affinity materials also affects the distribution of recovered phosphopeptides observed by MS. For
example, although previous studies suggest that TiO 2 preferentially enriches monophosphorylated peptides,[3] Li et al. found that a high peptide-to-TiO 2 ratio (insufficient binding
capacity) greatly enhances identification of multi-phosphorylated peptides.[4] These authors
133

recommend pre-examination of samples to find the digest-to-TiO 2 ratio that maximizes
phosphopeptide identifications.[4] Al2O3 foils with a relatively small loading capacity also
specifically capture multi-phosphorylated peptides in on-target phosphopeptide enrichment.[5]
Similarly, Moser et al. increased the capacity of IMAC columns to reduce the bias toward
detection of multi-phosphorylated peptides to some extent.[6] In Chapter 2, we showed that the
limited binding capacity of IMAC-Fe(III) micro-polymer spots leads to preferential enrichment
of tetra-phosphopeptides in -casein digests. Although new elution protocols, e.g. sequential
elution of mono- and multi-phosphorylated peptides from IMAC resins,[7] enhance
identification of different types of phosphopeptides, the intrinsic properties of an affinity
material may still prohibit comprehensive binding of different subsets of phosphopeptides.
Enrichment with both IMAC and TiO 2[8] or sequential capture of phosphopeptides with these
two types of resins can greatly enhance phosphopeptide identification. [9,10]
The studies in this chapter aim to reduce b iases in phosphopeptide enrichment by using
bifunctional polymer brushes as the affinity material. Chapter 3 shows that under mild acidic
conditions (≤ 1% TFA), Au-P-oxoTi shows high specificity toward mono-phosphorylated
peptides, while Au-P-NTA-Fe favors binding of multi-phosphorylated peptides. Considering that
the molar ratio of the polymer repeating unit to Ti in Au-P-oxoTi is close to that of polymer
repeating unit to Fe on Au-P-NTA-Fe (0.7 vs. 1), we modified poly(glycidyl methacrylate)
(PGMA) with equal amounts of NTA-Fe(III) and oxoTi species to simultaneously enrich monoand multi-phosphorylated peptides directly on a modified MALDI plate. Phosphopeptide
enrichment on these bifunctional coatings results in similar signal intensities for mono- and
multi-phosphorylated peptides in a single MALDI mass spectrum. For larger scale

134

phosphopeptide enrichment, we synthesized polymer brushes in the micron-size pores of nylon
membrane discs that serve as enrichment media. Stacking two differentially functionalized
membrane discs in an HPLC fitting allows sequential capture of both mono- and multiphosphorylated

peptides

from samples

that

contain

a

high

non-phosphoprotein-to-

phosphoprotein ratio. A single elution leads to recovery and MS- identification of these
phosphopeptides.

4.2 Experimental
4.2.1 Chemicals and materials
Most chemicals and reagents used in this study were obtained from Sigma-Aldrich. Other
chemicals and

materials include N-hydroxysulfosuccinimide

sodium salt (sulfo-NHS,

ProteoChem), FeCl3 (anhydrous, 98%, Spectrum), dimethylsulfoxide (DMSO, EMD Chemicals),
and acetonitrile (ACN, EMD Chemicals). Tetrahyd rofuran (THF, Mallinckrodt Baker) was
newly distilled before use, and deionized water was obtained from a Millipore ion-exchange
system (Milli-Q, 18 M cm). Au-coated Si wafers (200 nm of sputtered Au on 20 nm of
sputtered Cr on Si) were obtained from LGA Thin Films. Loprodyne® LP nylon membranes
with a 0.45 m pore size were a gift from PALL Corporation. Materials for protein digestion
include: sequencing grade modified porcine trypsin (treated with L-1-tosylamide-2-phenylethyl
chloromethyl ketone, Promega), Tris-HCl (Invitrogen), urea (J.T.Baker), DL-dithiothreitol
(Fluka), ammonium bicarbonate (Columbus Chemical) and trifluoroacetic acid (TFA, EMD
Chemicals).

135

4.2.2 Synthesis of bifunctionalized polymer brushes on Au wafers
Synthesis of bifunctional polymer brushes (oxoTi-PGMA-Fe) on Au-coated wafers
started with self-assembly of a disulfide initiator monolayer during immersion of a cleaned
(UV/O 3 for 15 min) wafer in an ethanol solution containing 1 mM (BrC(CH3)2COO(CH2)10S)2
for 24 h and rinsing with ethanol. (Nishotha Anuraj of our research group prepared the disulfide
initiator following a literature procedure.[11]) ATRP of glycidyl methacrylate (GMA) on the
initiator-modified Au surface followed a modified literature procedure.[12] Typically, a solution
containing glycidyl methacrylate (5 mL, 5.21 g, 36.7 mmol), methanol (4 mL), water (1 mL) and
2,2‟-dipyridyl (141 mg, 0.904 mmol) was degassed with three freeze-pump-thaw cycles. CuCl
(36.4 mg, 0.368 mmol) and CuBr2 (3.9 mg, 0.017 mmol) were added to the frozen solution prior
to three additional freeze-pump-thaw cycles. In a N 2-filled glove bag, the initiator-modified Au
wafers were immersed in the degassed monomer solution for 1 to 4 h to yield PGMA brushes
with different thicknesses on the Au surface. The wafers were finally removed from the
monomer/catalyst solution, rinsed with ethanol, and dried with N 2.
Figure 4.1 shows the PGMA functionalization steps, which were accomplished by
sequentially immersing the Au-PGMA wafer in the following solutions: (1) 1.5 M NaN 3 in
water/DMF (1:1, v/v) at 50 ° for 24 h; (2) 0.1 M succinic anhydride (SA) and 0.16 M 4C
dimethylaminopyridine (DMAP) in anhydrous DMF at 50 ° for 8 h; (3) 0.1 M sulfo-NHS and
C
0.1 M N,N’-diisopropylcarbodiimide (DIC) in DMSO for 12 h; (4) 0.1 M N ,Nbis(carboxymethyl)-L- lysine hydrate (aminobutyl-NTA) in pH 7.4 water/DMSO (1:1, v/v) for 4
h (the solution was prepared by initially adding 0.1 M NaOH until the aminobutyl NTA
136

Figure 4.1. Scheme for orthogonal modifications of PGMA brushes to obtain the bifunctional oxoTi-PGMA-Fe.

137

Figure 4.1 cont‟d

138

Figure 4.2. Structures of control polymer brushes for phosphopeptide enrichment.

139

dissolved. Then 0.1 M HCl was added to lower the solution pH to 7.4 prior to dilution with water
and DMSO); (5) DMF containing 0.1 M propiolic acid, 5 mM CuBr and 8.75 mM 2,2‟-dipyridyl
(the solution was degassed with three freeze-pump-thaw cycles) for 24 h, with modest stirring to
promote mass transfer, followed by rinsing the wafer with 0.1% TFA co ntaining 50% ACN to
protonate the carboxylic acids; (6) 0.1 M pH 7.4 EDTA for 30 min to remove any chelated Cu (I)
or Cu (II) ions in the polymer; (7) 0.1 M Ti(i-PrO)4 in anhydrous THF for 30 min followed by
rinsing with anhydrous THF in the N 2-filled glove bag; (8) 0.1 M FeCl3 for 30 min followed by
rinsing with ethanol. Unless otherwise stated, each of the above steps includes rinsing the wafer
first with the solvent employed for reaction and then with ethanol. The wafer was always dried
with N 2 before the next step.
We also synthesized four control materials (Figure 4.2) on Au wafers for comparison
with the bifunctionalized oxoTi-PGMA-Fe brushes. The PGMA-based control polymers, oxoTiPGMA and PGMA-Fe, were prepared following the bifunctionalization route except omitting the
3+

reaction with Ti(i-PrO)4 or Fe

chelation, respectively. Preparation of PHEMA-based control

polymers, oxoTi-PHEMA and PHEMA-Fe, followed the procedures in Chapters 3.

4.2.3 Synthesis of functionalized polymer brushes in nylon membranes
Synthesis of poly(2-(methacryloyloxy)ethyl succinate) (PMES) brushes in the pores of
nylon membranes followed our previous method.[13] Briefly, a UV/ozone cleaned nylon
membrane (25 mm diameter disc, 0.45 µ pore size) was supported in a home built Teflon cell
m
that allows flow through the membrane.
circulating

a

THF

solution

Immobilization of a silane initiator occurred by
containing

140

16

mM

(11-(2-bromo-2-

methyl)propionyloxy)undecyltrichlorosilane (synthesized as reported [14]) through the membrane
for 2 h followed by rinsing with 20 mL of THF.

A mixture of 2-(methacryloyloxy)ethyl

succinate monomer and 1 M aqueous NaOH (1:1, v/v), and a DMF solution containing 2 mM
CuBr, 1 mM CuBr2, and 6 mM 1,1,4,7,10,10-hexamethyltriethylenetetramine were degassed
separately with three freeze-pump-thaw cycles. In a N 2 bag, the two solutions were mixed at a
10:1 volume ratio, and the resulting solution was circulated through the membrane for 15 min to
accomplish ATRP of MES. Finally, 10 mL of ethanol and 10 mL of water were sequentially
passed through the membrane to remove physisorbed monomer, polymer, or catalyst.
Introduction of oxoTi clusters in the PMES-modified membrane to form Mem-P-oxoTi
occurred by passing ~40 mL of a 50 mM Ti(i-PrO)4/THF solution through a membrane modified
with PMES for 30 min and rinsing with 20 mL of THF. To immobilize NTA groups in the
MES- modified

membrane, 10

mL of an aqueous solution containing 0.1

M N-

hydroxysuccinimide (NHS) and 0.1 M N-(3-Dimethylaminopropyl)-N’-ethylcarbodiimide
hydrochloride was circulated through the membrane for 30 min, followed by 10 mL of 0.1 M
aminobutyl NTA (adjusted to pH 10.0 with 0.1 M NaOH) for 1 h. Ten mL of water was passed
through the membrane for rinsing after each step. Finally, the Fe(III)-NTA complex was formed
by first circulating10 mL of 0.1 M FeCl3 aqueous solution through the membrane for 30 min and
then sequentially passing 10 mL of 0.1% TFA/50% ACN (v/v) and 10 mL of ethanol through the
membrane to remove unbound Fe(III) ions. Both types of functionalized membrane discs were
stored in a desiccator until use. Unless specified, the flow rate in the above procedures was ~10
mL/min except in the ATRP process, where the rate of monomer solution dropped from ~8
mL/min to ~2 mL/min after 15 min.
141

4.2.4 Characterization of polymer brushes
The thicknesses of polymer brushes on Au-coated Si wafers were determined using a
rotating analyzer spectroscopic ellipsometer (J. A. Woollam, M-44), assuming a film refractive
index of 1.5. Reflectance FTIR spectra of the modified Au surfaces were collected on a Nicolet
Magna 560 spectrophotometer with a Pike grazing angle (80° accessory, and polymer growth on
)
nylon membranes was verified by attenuated total reflectance (ATR) FTIR spectroscopy (PerkinElmer Spectrum One Instrument) using an air background. For reflectance IR spectra, a UV/O 3cleaned, Au-coated Si wafer served as the background.
Inductively coupled plasma optical emission spectrometry (ICP-OES) was employed to
determine the amount of Ti and Fe eluted from the bifunctional polymer brushes. For this
determination, each polymer- modified Au-coated Si wafer was immersed in 4 mL of aqua regia
for 5 min to remove the entire polymer along with Au from the Si wafer, and the aqua regia was
diluted to 50 mL with water. To prepare standard solutions (for calibration) with the same
background as the sample solution, 20 pieces of 2.4 × 1.1 cm Au-coated wafers were immersed
in 80 mL of aqua regia to form the Au/aqua regia solution, and 4 mL of this solution was added
to a series of 50-mL volumetric flasks. Certain volumes of 998 ppm Ti and 100 ppm Fe stock
solutions were then added to each flask prior to dilution to 50 mL with water. Us ing these
standards, calibration curves for Ti and Fe were obtained with ICP-OES, and the amounts of Ti
and Fe in different sample solutions were determined using the calibration curve. The surface
coverages of Ti and Fe were calculated by dividing the metal amounts by the wafer area.

142

4.2.5 Phosphopeptide enrichment
Protein digestion and on-plate enrichment of phosphopeptides using modified Au-coated
wafers followed the procedures in Chapter 3. For enrichment using modified membranes, 6- mmdiameter discs were cut from the 25- mm-diameter membranes using a hole punch, and one or
two small discs were inserted in a flangeless fitting system (A-424, Upchurch Scientific) with a
frit between the membrane and the downstream Teflon tubing (I.D. 1/32"). The solution was
loaded in a Hamilton glass syringe and pushed through the membrane disc using a syringe pump
(Fisher Scientific). The dry protein digests were dissolved in 0.1% or 1% TFA containing 50%
ACN. Before enrichment, 30 µ of blank loading solution (the solution used for dissolving
L
samples) was passed through the membrane (100 µ
L/min) for equilibration. Phosphopeptide
isolation occurred by passing the digest solution through the membrane (50 µ
L/min) and
washing with blank loading solution (100 µ
L/min, 200 µ for samples containing only -casein,
L
-casein and ovalbumin digests, and 1 mL for samples containing a mixture of digests of three
phosphoproteins and bovine serum albumin (BSA)). Before elution, ~500 μL of air was loaded
into the syringe and gently pushed through the membrane by hand (~10 s) to remove residual
washing solution. Finally, phosphopeptides were eluted by passing 1% H3PO4 solution
containing 50% ACN through the membrane at 10 µ
L/min, and 10 µ of eluate was collected in
L
an Eppendorf tube. One µ of the eluate was spotted on the steel MALDI plate followed by
L
addition of 2,5-dihydroxybenzoic acid (DHB) for MALDI-MS analysis.

4.2.6 MS analysis of phosphopeptides and data analysis
MS analysis was performed in positive ion mode using an LTQ XL ion trap mass
spectrometer equipped with a vMALDI source (Thermo Fisher Scientific, San Jose, CA). For
143

each experiment, mass spectra were obtained from at least 3 enrichment spots, and the mass
spectra given here are representative examples. Confirmation of phosphopeptide sequences by
CID-MS/MS and database search were the same as in Chapter 3. Table 4.2 gives peptide
sequence information.

4.3 Results and discussion
4.3.1 Polymer synthesis on Au wafers and in nylon membranes
PGMA is attractive as a versatile affinity chromatography material because opening of
the epoxy ring allows facile polymer functionalization. [13-15] Thus,

formation of

bifunctionalized polymer brushes on Au surfaces starts with surface- initiated ATRP of GMA.
Varying the polymerization time from 1 to 4 h yields homogeneous PGMA brushes with
thicknesses that range from 20 to 70 nm, and control over film thickness allows variation of the
phosphopeptide binding capacity of functionalized films.
Reflectance FTIR spectra (Figure 4.3) provide evidence for the synthesis and
derivatization of PGMA to form bifunctional brushes. The ester carbonyl absorbance at 1740
-1

cm

in Figure 2a confirms the formation of PGMA (a, Figure 4.1) on the Au surface.

Subsequent opening of the epoxide by reaction of a with NaN 3 generates adjacent azide and
-1

hydroxyl groups (b, Figure 4.1), as confirmed by the antisymmetric N=N=N stretch at 2100 cm
-1

and the broad hydroxyl stretch at 3100-3650 cm

in Figure 4.3b. Functionalization of the

hydroxyl groups follows the same route used to form PHEMA-SA-NTA in Chapter 2, with some
changes in reaction conditions or reagents. Reaction of b with SA takes longer (8 h) than
144

Figure 4.3. Reflectance FTIR spectra of a PGMA film on a Au-coated wafer before (a) and after
the following sequential steps: (b) reaction with NaN 3; (c) reaction with SA followed by 1- min
deprotonation in 0.01 M NaOH and rinsing with acetone; (d) activation with sulfo-NHS/DIC; (e)
reaction with aminobutyl NTA followed by 1-min deprotonation in 1 mM NaOH and rinsing
with acetone; (f) reaction with propiolic acid. The thickness of the film (dry form) after each step
is noted above each spectrum.

reaction of PHEMA with SA (3 h), probably due to the less swollen structure of PGMA-N3
compared to PHEMA. (We monitored the completeness of this step by occasionally removing
the wafer from the reaction solution, deprotonating the carboxylic acid groups with a 1- min
immersion in 0.01 M NaOH, rinsing with acetone, and examining the intensity of the
-1

antisymmetric COO - stretch at 1590 cm . We stopped the reaction when this absorbance ceased
145

to increase with reaction time.) After activation of the newly formed polyacid (c, Figure 4.1)
-1

with sulfo-NHS/DIC (NHS ester peaks at 1825 cm

-1

and 1793 cm

in Figure 4.3d confirm the

activation), the polymer (d in Figure 4.1) reacts with aminobutyl NTA to yield e (Figure 4.1).
-1

Both the disappearance of the NHS ester peaks and the appearance of a peak at 1685 cm

-1

provide evidence for derivatization with NTA (Figure 4.3e). The absorbance around 1685 cm

and its broad tail likely stem from the newly formed amide bond and the carboxylate groups of
NTA. The introduction of another carboxylic acid group occurs through a Cu(I)-catalyzed (3 + 2)
azide-alkyne cycloaddition, also known as “click chemistry”.[16] We chose this reaction because
of its high efficiency and tolerance to a variety of functional groups.[17] Reaction of 180 nm
thick e (Figure 4.1) with propiolic acid (catalyzed by CuBr/2,2‟-dipyridyl) is nearly quantitative,
-1

as shown by the minimal residual N=N=N stretch at 2100 cm

(Figure 4.3f) after a 48 h “click

reaction”. When e is below 100 nm thick, the N=N=N stretch completely disappears after 24 h of
reaction. A control experiment with “click reaction” conditions in the absence of propiolic acid
for yielded no change in the IR spectrum of polymer e, providing further evidence for the
formation of new carboxylic acid groups in the polymer brushes. Immersing a wafer coated with
f (Figure 4.1) in Ti(i-PrO)4/THF solution leads to formation of oxoTi clusters (see Chapter 3)
coordinated with the carboxylic acid groups attached to the triazole rings (g, Figure 4.1). The
thickness of the polymer increases ~10% after this step. Unfortunately, IR spectra show no
evident changes after this reaction. A control experiment using a Au-PHEMA-SA-NTA wafer
suggested that the NTA groups do not bind a significant amount of Ti(IV) species, as the
polymer thickness does not increase (<1 nm) after exposure to Ti(i-PrO)4. Finally, immersion of

146

polymer g (Figure 4.1) in FeCl3 solution gives the NTA-Fe(III) complex, h (Figure 4.1).
Binding of Fe(III) also does not lead to significant changes in the IR spectrum of the film.
Table 4.1. The amount of Fe and Ti in functional polymer brushes.
oxoTi-PGMA-Fe
Base
polymer*
thickness (nm)
Fe

g/cm2

3
120

4
151

5
193

1.15

1.13

1.73

2.43

3.32

nmol/cm2-nm
Ti

g/cm2

Average 0.28±
0.02
1.45
1.50
2.67

2.76

2.89

nmol/cm2-nm

148

143

2.68

2.90
0.36

0.04

0.56

0.01

2
76.9

PHEMAFe-2***

0.32

1
71.4

PHEMA
-Fe-1**

0.08

Average 0.40±
0.06
* Base polymer is the polymer structure before modification with either Ti(IV) or Fe(III) species.
** The NTA-derivatized PHEMA film was directly immersed in 0.1 M FeCl3.
*** The NTA-derivatized PHEMA film was immersed in Ti(i-PrO)4/THF solution for 30 min
prior to rinsing and immersion in 0.1 M FeCl3.
To demonstrate the presence of Fe and Ti in oxoTi-PGMA-Fe films, we eluted Ti and Fe
from modified wafers using aqua regia and performed ICP-OES to determine Ti and Fe
concentrations in the eluate. Table 4.1 shows that surface coverages of Fe and Ti generally
increase with the thickness of the base polymer film (f). The metal coverages normalized to base
2

2

film thickness are 0.28±
0.02 nmol/(cm -nm) Fe and 0.40±
0.06 nmol/(cm - nm) Ti in oxoTi3

PGMA-Fe films. Assuming the density of the film is 1 g/cm and using a repeating- unit mass of
599.6 g/mol (structure f in Figure 4.1), these metal densities correspond to 1.6± Fe atoms and
0.1
2.4± Ti atoms per repeating unit. Attainment of >1 Fe atom per repeating unit could occur
0.4
because there was some polymerization on the back side of the wafer, which was not coated with

147

Au. (The back side of the wafer appears to change color after polymerization.) In the case of Ti,
there could be several Ti atoms per clusters, and some Ti clusters may also form on the back side
of the wafer.
We also examined the extent of reaction of NTA groups with Ti(i-PrO)4. Immersing an
NTA-derivatized PHEMA brush in Ti(i-PrO)4/THF solution for 30 min followed by Fe(III)
chelation (PHEMA-Fe-2 in Table 4.1) leads to a much smaller loading of Ti in the film
2

compared to in polymer f (0.08 vs. 0.40 nmol Ti atoms/cm -nm). Elemental analysis of a
PHEMA-Fe film (PHEMA-Fe-1 in Table 4.1) that was not exposed to Ti(i-PrO)4 shows 0.1 nmol
2

Ti/cm -nm, which is essentially the experimental uncertainty of the measurement.
To perform larger scale phosphopeptide enrichment, we synthesize and functionalize PMES
-1

brushes in nylon membranes. Although peaks due to the amide groups (1632 cm ) in the base
membrane dominate the ATR-IR spectra of PMES- modified nylon, the carbonyl absorbance at
-1

~1740 cm

(Figure 4b) confirms the formation of PMES on the membrane. The succinimide
-1

-1

ester absorptions at 1820 cm and 1790 cm in Figure 4c show activation of the PMES brushes,
and the disappearance of these peaks in Figure 4d suggests the substitution of NHS by
aminobutyl NTA, although hydrolysis could also occur. The peaks due to the base membrane
-1

mask the amide and NTA absorbances that should appear around 1685 cm

in Figure 4c.

Subsequent passage of FeCl3 solutions through the membrane yields the NTA-Fe(III) complex.
Similar to the synthesis of poly acid-oxoTi on Au wafers (Chapter 3), oxoTi clusters also form
when passing a Ti(i-PrO)4 /THF solution through PMES- modified nylon membranes. Formation
148

of PMES-NTA-Fe in membranes also occurs using the chemistry previously employed on gold
wafers, but solutions are passed through the membrane.

Figure 4.4. ATR-IR spectra of nylon membranes before (a) and after (b) modification with
PMES, (c) activation with NHS/1-ethyl-3-(3-dimethylaminopropyl)carbodiimide and (d) reaction
with aminobutyl NTA (pH 10).

4.3.2 On-plate phosphopeptide enrichment using oxoTi-PGMA-Fe brushes
The previous chapters and a number of literature studies show that preferences for
binding mono- and multi-phosphorylated peptides stem from (1) the intrinsic affinities of metalion complexes or metal oxide materials,[7,18] (2) the binding capacity of the substrates (or
peptide-to-affinity sites ratio),[4,5] and (3) the acidity of the sample solution. [7] With these
considerations in mind, we performed on-plate phosphopeptide enrichment using oxoTi-PGMAFe brushes with different thicknesses. Similar studies with the four control materials in Figure
149

4.2 separately verify the phosphopeptide enrichment specificities of the oxoTi and NTA-Fe.
Among the four control polymers, PHEMA-Fe was used in Chapters 2 and 3, and oxoTiPHEMA was described in Chapter 3.
4.3.2.1 Influence of film thickness on the specificity of affinity groups
We first tested phosphopeptide enrichment from a 1-L mixture of three phosphoprotein
digests (200 fmol each of -casein, -casein and ovalbumin) dissolved in 0.1 % TFA.
Phosphopeptide isolation occurred in 60 nm-thick polymer films functionalized with NTA-Fe(III)
and/or oxoTi complexes (the thickness corresponds to the base polymer prior to modification
with Fe(III) or oxoTi). Enrichment on PGMA-Fe and PHEMA-Fe leads to spectra dominated by
multi-phophorylated peptides (Figures 5a and 6a). In contrast, the two strongest signals in the
mass spectrum of sample enriched on oxoTi-PHEMA stem from mono-phosphorylated peptides
(Figure 4.6b). In the case of oxoTi-PGMA, signals from both types of phosphopeptides appear,
but the spectral pattern (Figure 4.4c) is closer to that of oxoTi-PHEMA than the PHEMA-Fe
control.

We should note that although the four control films have thicknesses of 60 nm,

PHEMA-based brushes typically show 2- to 6-fold greater signal-to- noise ratios (S/N) than
PGMA-based systems. This suggests a lower binding capacity for the PGMA-based control
polymers, which is reasonable because unlike derivatized PHEMA, the functionalized PGMA
contains two branched side chains, and only one of them contains a binding site in the control
polymers.
Films with both oxoTi and Fe(III) complexes enrich a wider range of phosphopeptides
than films containing one of these functionalities. Enrichment on the 60 nm- thick oxoTi-PGMAFe (Figure 4.5c) allows MS identification of 7 more mono-phosphorylated peptides and the same
number of multi-phosphorylated peptides as enrichment on PGMA-Fe (Figure 4.5a), as well as 3
150

Figure 4.5. MALDI mass spectra after on-plate phosphopeptide enrichment from a 1-L 0.1%
TFA solution containing 200 fmol each of three phosphoprotein digests, -casein, -casein and
ovalbumin (3pp). The MALDI plates were coated with 60 nm-thick (prior to addition of affinity
groups) polymer films, whose structures after functionalization are noted in each spectrum. Each
phosphopeptide is labeled as #p#, where the first # denotes the number of phosphates in that
peptide and the second # denotes the peptide number in Table 4.2. Black and red colors represent
mono- and multi-phosphorylated peptides, respectively.
151

Figure 4.6. MALDI mass spectra after on-plate phosphopeptide enrichment from a 1-L 0.1%
TFA solution containing 200 fmol each of three phosphoprotein digests, -casein, -casein and
ovalbumin (3pp). The MALDI plates were coated with 60 nm- thick (prior to addition of affinity
groups) polymer films, whose structures after functionalization are noted in each spectrum. Each
phosphopeptide is labeled as #p#, where the first # denotes the number of phosphates in that
peptide and the second # denotes the peptide number in Table 4.2. Black and red colors represent
mono- and multi-phosphorylated peptides, respectively.

more mono-phosphorylated peptides and 3 more multi-phosphorylated peptides than enrichment
on oxoTi-PGMA (Figure 4.5b). In addition, the MS noise after oxoTi-PGMA-Fe (Figure 4.5c)
enrichment is similar to that from PGMA-Fe (Figure 4.5a) and only 25% of the noise with
oxoTi-PGMA (Figure 4.5b). The presence of two binding functionalities and the resulting
increased binding capacities likely lead to the superior performance of oxoTi-PGMA-Fe relative
to enrichment on oxoTi-PGMA or PGMA-Fe.
152

Table 4.2. Phosphopeptides detected in positive- ion mode MALDI-MS spectra. The list contains
the phosphopeptide sequence* (if identified), the range of amino acids, the number of
+
phosphoryl groups, and the monoisotopic m/z of [M+H] .
Peptide
#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

-casein

20
21
22
23
24
25
26

Phosphopeptide sequence

Sequence
Number**

K.TVDMEpSTEVFTK.K
M(cam)-105
S2 153-164
K.VPQLEIVPNpSAEER.L
S1 121-134
K. VPQLEIVPNpSAEER.L
Carbamyl (N- S1 121-134
term)
K.VPQLEIVPNpSAEER.L
E(cam)
S1 121-134
***
YLGEYLIVPNpSAEER
S1
K.DIGpSEpSTEDQAMEDIK.Q
M(cam)-105 S1 58-73
K.DIGpSEpSTEDQAMEDIK.Q
S1 58-73
sodium adduct, M(cam)-105
K. DIGpSEpSTEDQAMEDIK.Q
S1 58-73
Carbamyl (N-term), M(cam)-105
K.DIGpSEpSTEDQAMEDIK.Q
S1 58-73
K.DIGpSEpSTEDQAMEDIK.Q
S1 58-73
M(cam)-105, D(cam)
K.DIGpSEpSTEDQAMEDIK.Q
M(ox)
S1 58-73
K.YKVPQLEIVPNpSAEER.L
S1 119-134
K.DIGpSEpSTEDQAMEDIK.Q
Carbamyl S1 58-73
(K)
K.DIGpSEpSTEDQAMEDIK.Q
D(cam)
S1 58-73
K. YKVPQLEIVPN pSAEER.L Carbamyl (N- S1 119-134
term)
N/A
N/A
N/A
K.QMEAEpSIpSpSpSEEIVPN pSVEQK.H
S1 74-94
M(cam)-105, pyro-glu from Q
N/A
K.QMEAEpSIpSpSpSEEIVPN pSVEQK.H
S1 74-94
Pyro-glu from Q
K.QMEAEpSIpSpSpSEEIVPN pSVEQK.H
S1 74-94
N/A
N/A
R.NANEEEYSIGpSpSpSEEpSAEVATEEVK.I S2 61-85
N/A
S2
61****
85

153

– Mono.
PO3H2 m/z +
MH
1
1418.6
1
1660.8
1
1703.8
#

1
1

1717.8
1832.8

2
2

1879.7
1901.7

2

1922.7

2
2

1927.7
1936.7

2
1
2

1943.7
1952.0
1970.7

2
1

1984.7
1995.0

1
1
5
5

2027
2037
2575
2655.9

2
5

2672
2702.9

5
4
3
4
5

2720.9
2760
2819
3008.0
3088.0

Ovalbu
-min

-casein

Table 4.2 cont‟d
27
28
29
30
31
32
33

K.FQpSEEQQQTEDELQDK.I
K.FQpSEEQQQTEDELQDK.I Sodium adduct
K. FQpSEEQQQTEDELQDK.I
Carbamyl
(N-term)
K.FQpSEEQQQTEDELQDK.I
E(cam)
R.ELEELNVPGEIVEpSLpSpSpSEESITR.I
A.RELEELNVPGEIVEpSLpSpSpSEESITR.I
R.EVVGpSAEAGVDAASVSEEFR.A

48-63
48-63
48-63

1
1
1

2061.8
2083.8
2104.8

48-63
17-40
16-40
341-360

1
4
4
1

2118.8
2966.2
3122.3
2088.9

* Some phosphopeptides share the same sequence, but are modified differently during protein
denaturing, alkylation, or digestion. “Carbamyl” corresponds to carbamylation, (cam)
corresponds to carbamidomethylation, M(cam-105) corresponds to the neutral loss of S(methylthio)acetamide from M(cam), and M(ox) corresponds to methionine sulfoxide.
** S1 and S2 represent S1- and S2-chains of -casein, respectively.
*** According to Larsen, M. R. et al.,[19] this is a new sequence variant of the -S1 casein in
the region 104–119.
**** According to Larsen, M. R. et al.,[19] this is probably a penta-phosphorylated peptide.

Increasing the thickness of PGMA films from 60 to 120 nm should enhance the
phosphopeptide binding capacity of functionalized films. As comparison of Figure 4.5 and
Figure 4.7 shows, enrichment on the thicker (120 nm) oxoTi-PGMA-Fe and PGMA-based
control materials gives 1.2- to 4- fold higher S/N ratios for phosphopeptides. The high specificity
of oxoTi toward mono-phosphorylated peptides is more evident after the increase in film
thickness (compare Figures 7b and 5b). For specific quantitative comparison of different
enrichments, we chose the four most frequently detected phosphopeptides, 1p2, 1p27, 2p6 and
5p19, which represent a range of degrees of phosphorylation. Signals for all of these
phosphopeptides appear in Figure 4.7c with >44% relative intensity. However, 1p2 and 1p27
show intensities below 37% in Figure 4.7a, and 2p6 and 5p19 show intensities below 23% in
Figure 4.7b. This suggests that oxoTi-PGMA-Fe enriches phosphopeptides with less bias toward
the number of phosphate groups on the peptide than do polymers with only one type of affinity

154

Figure 4.7. MALDI mass spectra after on-plate phosphopeptide enrichment from 1-L of 0.1%
TFA containing 200 fmol each of three phosphoprotein digests, -casein, -casein and
ovalbumin (3pp). The MALDI plates were coated with 120 nm-thick (prior to addition of
affinity groups) polymer films, whose structures after functionalization are noted in each
spectrum. Each phosphopeptide is labeled as #p#, where the first # denotes the number of
phosphates in that peptide and the second # denotes the peptide number in Table 4.2. Black and
red colors represent mono- and multi-phosphorylated peptides, respectively.
155

group. In addition, the MS signal pattern after enrichment on oxoTi-PGMA-Fe remains almost
the same after the polymer thickness doubles (compare Figures 5c and 7c), suggesting that after
surpassing a certain thickness (e.g. 60 nm for 200 fmol of three phosphoprotein digests mixture
in 0.1% TFA) the bifunctionalized polymer maintains its low-bias enrichment of mono- and
multi-phosphorylated peptides.
4.3.2.2 Influence of sample acidity on phosphopeptide enrichment
Sample acidity also affects phosphopeptide enrichment. With PGMA-Fe, increasing the
TFA level in the sample and rinsing solutions from 0.1 to 1% raises the S/N of tetra- and pentaphosphorylated peptides 2- to 3-fold (compare Figures 8a and 7a). Similarly, higher acidity
increases the signals of doubly phosphorylated peptides enriched by oxoTi-PGMA (2.3- to 3.4 fold increase, compare Figures 8b and 7b ). Enrichment of the same sample on the 120-nm thick
oxoTi-PGMA-Fe yields a mass spectrum that is an “average” of the spectra that result from
enrichment on the two control polymers PGMA-Fe and oxoTi-PGMA (compare Figures 8a-c).
Nevertheless, the doubly phosphorylated peptide 2p6 and the penta-phosphopeptide 5p19 now
dominate the spectrum (Figure 4.8c). As mono-phosphorylated peptides generally ionize better
than multi-phosphorylated peptides, such a result suggests a biased enrichment of multiphosphorylated peptides by the 120 nm thick bifunctional polymer film in 1% TFA. Because
binding of mono-phosphorylated peptides to metal affinity sites decreases with higher solution
acidity, the inefficient enrichment of mono-phosphorylated peptides likely stems from
insufficient binding capacity of oxoTi-PGMA-Fe in 1% TFA.
Further increasing the thickness of oxoTi-PGMA-Fe to 220 nm for enrichment of 200
fmol of three phosphoprotein digests in 1% TFA brings the MS pattern (Figure 4.9c) back to that
in Figure 4.7c, which was generated from the same amount of sample using the same polymer
156

Figure 4.8. MALDI mass spectra after on-plate phosphopeptide enrichment from 1-L of 1%
TFA containing 200 fmol each of three phosphoprotein digests, -casein, -casein and
ovalbumin (3pp). The MALDI plates were coated with 120 nm-thick (prior to addition of
affinity groups) polymer films, whose structures after functionalization are noted in each
spectrum. Each phosphopeptide is labeled as #p#, where the first # denotes the number of
phosphates in that peptide and the second # denotes the peptide number in Table 4.2. Black and
red colors represent mono- and multi-phosphorylated peptides, respectively.
157

Figure 4.9. MALDI mass spectra after on-plate phosphopeptide enrichment from 1 L of 1%
TFA containing 200 fmol each of three phosp hoprotein digests, -casein, -casein and
ovalbumin (3pp). The MALDI plates were coated with 220 nm-thick (prior to addition of
affinity groups) polymer films, whose structures after functionalization are noted in each
spectrum. Each phosphopeptide is labeled as #p#, where the first # denotes the number of
phosphates in that peptide and the second # denotes the peptide number in Table 4.2. Black and
red colors represent mono- and multi-phosphorylated peptides, respectively.
158

Figure 4.10. MALDI mass spectra after on-plate phosphopeptide enrichment from a 1 L of 1%
TFA containing 50 fmol each of three phosphoprotein digests, -casein, -casein and ovalbumin
(3pp). The MALDI plates were coated with 220 nm-thick (prior to addition of affinity groups)
polymer films, whose structures after functionalization are noted in each spectrum. Each
phosphopeptide is labeled as #p#, where the first # denotes the number of phosphates in that
peptide and the second # denotes the peptide number in Table 4.2. Black and red colors represent
mono- and multi-phosphorylated peptides, respectively.
159

structure, but with a 120 nm film and weaker acidity for enrichment. In addition, Figure 4.9c
presents ~50% less noise than Figure 4.7c, as well as a nearly perfect combination of the spectra
generated from its PGMA-based control polymers (Figure 4.9 a and b).
Finally we tested samples with lower amounts of phosphoprotein or and excess of nonphosphoprotein. With only 50 fmol each of three phosphoprotein digests in 1% TFA, enrichment
on the 220 nm thick oxoTi-PGMA-Fe yields strong signals for mono- and doubly
phosphorylated peptides due to the combined affinities of oxoTi and NTA-Fe groups (Figure
4.10). Enrichment on this film also leads to strong signals of mono- and multi-phosphorylated
peptides from a mixture containing 200 fmol of three phosphoprotein digests and 2 pmol of BSA
(non-phosphoprotein) digest . In the presence of excess BSA (0.16 mg total protein digest) peak
patterns and S/N are similar for samples with and without BS A (compare Figure 4.11 and Figure
4.9c).

Figure 4.11. MALDI mass spectrum after phosphopeptide enrichment from 1 L of 1% TFA
containing 200 fmol each of three phosphoprotein digests (3pp), -casein, -casein and
ovalbumin, and 2 pmol of BSA digest, The MALDI plates were coated with 220 nm-thick (prior
to addition of affinity groups) oxoTi-PGMA-Fe. Each phosphopeptide is labeled as #p#, where
the first # denotes the number of phosphates in that peptide and the second # denotes the peptide
number in Table 4.2. Black and red colors represent mono- and multi-phosphorylated peptides,
respectively.

160

To summarize these two sections, the dual affinities of high-capacity, bifunctional
polymer brushes allow low-bias enrichment of mono- and multi-phosphorylated peptides from
samples with moderate acidity (0.1-1% TFA) and complexity (e.g. <5-20 g/mL of several
phosphoproteins and < 0.5 mg/mL non-phosphoproteins). The bifunctional film is attractive for
enrichment of target phosphoproteins from low- volume samples such as those that result
immuno-purifications, gel electrophoresis, and high-throughput kinase assays. In addition, the
orthogonal brush derivatization strategy can create bifunctional films for other applications that
require two adjacent functional groups.

4.3.3 Enrichment of phosphopeptides using functional membranes
To enrich dilute phosphopeptides from 10-500 L samples, we modified the micron-size
pores of nylon membranes with PMES-oxoTi (Mem-PMES-oxoTi) or with PMES-NTA-Fe(III)
(Mem-PMES-Fe). In an initial test with Mem-PMES-oxoTi, we enriched phosphopeptides from
a digest mixture containing 3 pmol each of -casein, -casein and ovalbumin in 200 L of 0.1%
TFA in 50% ACN. After passing this mixture through the membrane and eluting with 1%
H3PO4 in 50% ACN, mono-phosphorylated peptide signals dominate the MALDI mass spectrum
of 1 µ of eluate (Figure 4.12a). Increasing the amount of phosphoprotein digests to 10 pmol
L
(without changing the sample volume) does not significantly improve the detection of multiphosphorylated peptides, but the S/N of most of the mono-phosphorylated peptides increases ~2
fold (Figure 4.12b). At these protein amounts, the PMES-oxoTi brushes in the nylon membrane
evidently have sufficient binding capacity to maintain the specificity of oxoTi for monophosphorylated peptides. We should emphasize that only 1/10 of the eluate was spotted on the
MALDI plate in all cases, which leads to lower S/N than might be expected for these relatively

161

Figure 4.12. MALDI mass spectra of phosphopeptides enriched from 200 L of 0.1% TFA, 50%
ACN containing three phosphoprotein digests (3pp), -casein, -casein and ovalbumin. In all
cases, only 1 L of the 10 L eluate was spotted on the sample plate. The membrane
modifications and amount of protein in each sample are noted above the spectra. Each
phosphopeptide is labeled as #p#, where the first # denotes the number of phosphates in that
peptide and the second # denotes the peptide number in Table 4.2. Black and red colors represent
mono- and multi-phosphorylated peptides, respectively.
162

large amounts of protein.
Compared with the Mem-PMES-oxoTi system, enriching the 10-pmol sample with MemPMES-Fe yields a more comprehensive mass spectrum (Figure 4.12c) with strong signals from
both mono- and multi-phosphorylated peptides, including signals from 9 tetra- or pentaphosphopeptides.

The

high

binding

capacity

likely

increases

the

enrichment

of

monophosphopeptides by the NTA-Fe(III) complexes relative to capture in a film on a MALDI
plate.
When enriching a sample containing 10 pmol (each) of -casein, -casein and ovalbumin
digests, and 1 nmol of BSA digest (1:1:1:100 molar ratio), we employed 1% TFA in 50% ACN
to suppress the binding of non-phosphorylated peptides. Under these conditions, however, MemPMES-Fe enriches only

multi-phosphorylated peptides, especially

tetra- and penta-

phosphopeptides (Figure 4.13a). In contrast, Mem-PMES-oxoTi still enriches monophosphorylated peptides (Figure 4.13b) but does not reveal significant signals from tetra- and
penta-phosphorylated species. To realize a more comprehensive enrichment of phosphopeptides
in one sample, we stacked Mem-PMES-oxoTi and Mem-PMES-Fe in series in a single fitting
system.

(The loading solution first encountered the membrane with oxo Ti.) The tandem

membrane enrichment leads to a mass spectrum that essentially combines the signals from the
first two spectra (compare Figure 4.13a-c), although signals are not as strong as after enrichment
with the separate membranes, perhaps because of ion suppression.
We also mixed the eluates generated separately from Mem-PMES-Fe and Mem-PMESoxoTi. MALDI-MS of the mixture shows dominant signals for tetra- and penta-phosphorylated
peptides (Figure 4.13d), suggesting a lower recovery of mono-phosphorylated peptides with
Mem-PMES-oxoTi than for multi-phosphorylated peptides with Mem-PMES-Fe. Separation of

163

Figure 4.13. (a-c) MALDI mass spectra after phosphopeptide enrichment from 200-L of 1%
TFA, 50% ACN containing 10 pmol each of three phosphoprotein digests (-casein, -casein
and ovalbumin), and 1 nmol of BSA digest. The membrane used for enrichment is noted in each
spectrum. Figure (d) is a mixture of 1 L of eluate from (a) and 1 L of eluate from (b). Each
phosphopeptide is labeled as #p#, where the first # denotes the number of phosphates in that
peptide and the second # denotes the peptide number in Table 4.2. Black and red colors represent
mono- and multi-phosphorylated peptides, respectively.

164

Figure 4.13 cont‟d

165

the recovered mono- and multi-phosphorylated peptides prior to MS analysis would probably
improve analysis of the two classes of phosphopeptides. In contrast, the tandem enrichment
leads to more evenly distributed signals for mono- and multi-phosphorylated peptides (Figure
4.13c), probably due to the strong binding of a portion of multi-phosphorylated peptides to MemPMES-oxoTi (the first membrane), which leads to enrichment of only a portion of the multiply
phosphopeptides in Mem-PMES-Fe (the second membrane). Final elution with 1% H3PO4 is
probably sufficient to release multi-phosphorylated peptides from NTA-Fe groups but not from
oxoTi. Considering that NTA-Fe(III) does not capture mono-phosphorylated peptides efficiently
during sample loading and washing with highly acidic solutions(1% TFA), reversing the
membrane sequence in the set-up might improve the tandem enrichment.
In summary, tandem membrane enrichment separates the capture and the elution of
phosphopeptides from the two different materials in space and time. The recovery of mono- and
multi-phosphorylated peptides may change as the flow rate or membrane stacking sequence
changes. Such a tandem strategy should provide a more comprehensive and efficient enrichment
of phosphopeptides from complex protein digests samples than enrichment using a single affinity
material. However, the tandem enrichment strategy will produce a more complex mixture of
phosphopeptides, and separation (e.g. RP-LC) prior to MS may assist analysis.

166

REFERENCES

167

REFERENCES
1.
Bodenmiller, B.; Mueller, L. N.; Mueller, M.; Domon, B.; Aebersold, R., Reproducible
Isolation of Distinct, Overlapping Segments of the Phosphoproteome. Nat. Methods 2007, 4,
231-237.
2.
Dunn, J. D.; Reid, G. E.; Bruening, M. L., Techniques for Phosphopeptide Enrichment
Prior to Analysis by Mass Spectrometry. Mass Spectrom. Rev. 2010, 29, 29-54.
3.
Thingholm, T. E.; Jensen, O. N.; Larsen, M. R., Analytical Strategies for
Phosphoproteomics. Proteomics 2009, 9, 1451-1468.
4.
Li, Q. R.; Ning, Z. B.; Tang, J. S.; Nie, S.; Zeng, R., Effect of Peptide-to-Tio2 Beads
Ratio on Phosphopeptide Enrichment Selectivity. J. Proteome Res. 2009, 8, 5375-5381.
5.
Qiao, L. A.; Bi, H. Y.; Busnel, J. M.; Hojeij, M.; Mendez, M.; Liu, B. H.; Girault, H. H.,
Controlling the Specific Enrichment of Multi-Phosphorylated Peptides on Oxide Materials:
Aluminium Foil as a Target Plate for Laser Desorption Ionization Mass Spectrometry. Chem. Sci.
2010, 1, 374-382.
6.
Moser, K.; White, F. M., Phosphoproteomic Analysis o f Rat Liver by High Capacity
IMAC and LC-MS/MS. J. Proteome Res. 2006, 5, 98-104.
7.
Thingholm, T. E.; Jensen, O. N.; Robinson, P. J.; Larsen, M. R., Simac (Sequential
Elution from IMAC), a Phosphoproteomics Strategy for the Rapid Separation of
Monophosphorylated from Multiply Phosphorylated Peptides. Mol. Cell. Proteomics 2008, 7,
661-671.
8.
Wilson-Grady, J. T.; Villen, J.; Gygi, S. P., Phosphoproteome Analysis of Fission Yeast.
J. Proteome Res. 2008, 7, 1088-1097.
9.
Carrascal, M.; Ovefletro, D.; Casas, V.; Gay, M.; Ablan, J., Phosphorylation Analysis of
Primary Human T Lymphocytes Using Sequential IMAC and Titanium Oxide Enrichment. J.
Proteome Res. 2008, 7, 5167-5176.
10.
Choi, S.; Kim, J.; Cho, K.; Park, G.; Yoon, J. H.; Park, S.; Yoo, J. S.; Ryu, S. H.; Kim, Y.
H.; Kim, J., Sequential Fe3O4/Tio2 Enrichment for Phosphopeptide Analysis by Liquid
Chromatography/Tandem Mass Spectrometry. Rapid Commun. Mass Spectrom. 2010, 24, 14671474.
11.
Shah, R. R.; Merreceyes, D.; Husemann, M.; Rees, I.; Abbott, N. L.; Hawker, C. J.;
Hedrick, J. L., Using Atom Transfer Radical Polymerization to Amplify Monolayers of Initiators
Patterned by Microcontact Printing into Polymer Brushes for Pattern Transfer. Macromolecules
2000, 33, 597-605.

168

12.
Edmondson, S.; Huck, W. T. S., Controlled Growth and Subsequent Chemical
Modification of Poly(Glycidyl Methacrylate) Brushes on Silicon Wafers. J. Mater. Chem. 2004,
14, 730-734.
13.
Slater, M.; Snauko, M.; Svec, F.; Frechet, J. M. J., "Click Chemistry" In the Preparation
of Porous Polymer-Based Particulate Stationary Phases for Mu-Hplc Separation of Peptides and
Proteins. Anal. Chem. 2006, 78, 4969-4975.
14.
Luo, Q. Z.; Zou, H. F.; Xiao, X. Z.; Guo, Z.; Kong, L.; Mao, X. Q., Chromatographic
Separation of Proteins on Metal Immobilized Iminodiacetic Acid- Bound Molded Monolithic
Rods of Macroporous Poly(Glycidyl Methacrylate-Co-Ethylene Dimethacrylate). J. Chromatogr.
A 2001, 926, 255-264.
15.
Roberts, M. W. H.; Ongkudon, C. M.; Forde, G. M.; Danquah, M. K., Versatility of
Polymethacrylate Monoliths for Chromatographic Purification of Biomolecules. J. Sep. Sci. 2009,
32, 2485-2494.
16.
Kolb, H. C.; Finn, M. G.; Sharpless, K. B., Click Chemistry: Diverse Chemical Function
from a Few Good Reactions. Angew. Chem.-Int. Edit. 2001, 40, 2004-2021.
17.
Hein, C. D.; Liu, X. M.; Wang, D., Click Chemistry, a Powerful Tool for Pharmaceutical
Sciences. Pharm. Res. 2008, 25, 2216-2230.
18.
Trost, M.; Bridon, G.; Desjardins, M.; Thibault, P., Subcellular Phosphoproteomics. Mass
Spectrom. Rev. 2010, 29, 962-990.
19.
Larsen, M. R.; Thingholm, T. E.; Jensen, O. N.; Roepstorff, P.; Jorgensen, T. J. D.,
Highly Selective Enrichment of Phosphorylated Peptides from Peptide Mixtures Using Titanium
Dioxide Microcolumns. Mol. Cell. Proteomics 2005, 4, 873-886.

169

Chapter 5. Controlled Proteolysis with Trypsin-modified
Membranes to Obtain Large Peptides for Mass Spectrometry
(The work described in this chapter was performed in collaboration with Ken Chanthamontri
from Prof. Scott McLuckey‟s group in the Department of Chemistry at Purdue University, and
with Dr. Scott Smith from Prof. Reid‟s group in the Department of Chemistry at Michigan State
University.)

5.1 Introduction
Protein analysis using large (2000-15000 Da) proteolytic peptides, also known as middledown proteomics, has several advantages over bottom- up and top-down approaches (see Chapter
1). One of the key challenges in exploiting the assets of middle-down proteomics, however, is
controlled generation of the large peptides. Although acid digestion produces large peptides and
often increases protein solubility, it suffers from low digestion efficiency and removal of protein
modifications.[1,2] Milder and more widely used methods to generate large peptides employ
proteases that cleave proteins at only one residue that appears relatively infrequently along the
protein chain. Such proteases include Glu-C, Lys-C, and Asp-N (see Chapter 1 for details.)
Unfortunately, these high-specificity enzymes are expensive for routine large-scale protein
digestion. Moreover, not all proteins have sequences that will yield large peptides.
Trypsin is a relatively inexpensive protease that cleaves proteins at the carboxylic side of
Lys and Arg, except when either is followed by Pro.[3] Because the total abundance of Lys and
Arg in proteins is relatively high, however, tryptic digestion generates mostly small peptides
(typically below 3000 Da), which are useful for bottom-up protein analysis by MS.[4] However,
170

recent studies suggest that trypsin also generates large peptides with uncleaved Lys or Arg within
the peptide sequences when digestion is properly controlled (see Chapter 1 and below). [5,6] In
addition, the availability of trypsin in mg quantities allows facile immobilization of this enzyme
on a variety of porous materials for applications such as online protein digestion-LC-ESI-MS.[7]
This chapter describes a membrane-based kinetic approach for generating large tryptic peptides
from substrate proteins. Development of such a method requires some understanding of the
tryptic proteolysis mechanism and digestion pathway.
Trypsin is a member of the chymotrypsin clan, [8,9] which is a part of the serine protease
family.[9] Based on their three-dimensional structure, serine proteases can be grouped into
several clans. The proteases of the chymotrypsin, subtilisin and carboxypeptidase C clans all
contain a “catalytic triad” that consists of Ser (nucleophile), Asp (electrophile) and His (base).
These clans exhibit similar spatial arrangements in different proteases to assist hydrolysis of the
substrate protein.[9,10]
Figure 5.1 shows the mechanism of tryptic digestion.

Asp189, Ser195 and His57

constitute the catalytic triad of trypsin (and other chymotrypsin- like proteases), which facilitates
the cleavage of the amide bond of substrate proteins by water. molecules.[8,10,12]

The

specificity of trypsin in cleaving certain sites of a substrate protein stems primarily from Asp189,
which sits at the base of the protease subsite most proximate to the cleavage site.[12,13] During
digestion, the negatively charged Asp189 attracts and stabilizes the positively charged Lys or
Arg of the substrate protein near the catalytic pocket where the hydrolysis occurs.
Although the catalytic role of proteases in cleavage of substrate proteins at specific sites
is well documented,[9,10,12] the protein digestion pathway (how a single protein molecule
gradually degrades to small peptides) is less explored. Nevertheless, controlling the digestion

171

Figure 5.1. Scheme of trypsin-catalyzed proteolysis. Asp189, Ser195 and His57 constitute the catalytic triad of trypsin, and
Asp189 of trypsin stabilizes the substrate protein or peptide in the catalytic pocket. The figure is drawn according to references
[9,10,12].
172

Figure 5.1 cont‟d

173

Figure 5.1 cont‟d

174

Figure 5.2. Enzyme-catalyzed protein digestion pathways: (a) one-by-one scheme and (b) zipper
scheme. E represents a protease, S and S1…Sn represent the substrate protein and its degraded
large peptides, respectively. C1…Cn represent the complexes of protease with the substrate
protein or large peptide, k represents a rate constant at each step, and X1…Xn represent peptides
that either contain no missed cleavage sites (final products), or contain one or more missed
cleave sites that can undergo further proteolysis. The figure is redrawn according to
reference.[11]

pathway is vital to obtaining large peptides for middle-down protein analysis. Previous studies
proposed two possible degradation pathways for enzyme-catalyzed proteolysis: a one-by-one
scheme and a zipper scheme (Figure 5.2).[14-16]

175

In the one-by-one pathway (Figure 5.2a), the protease (E) combines with the substrate
protein (S) to form a complex C1 (e.g. trypsin with the substrate protein stabilized as shown in
Figure 5.1a and b). Proteolysis occurs at a single site on the substrate protein, and the resulting
peptides (S1 and X1) dissociate from the protease. Subsequently the peptide (S1 or possibly X1)
that contains one or more additional cleavage sites (e.g. Lys or Arg for tryptic digestion),
undergoes similar degradation to form smaller peptides (S2 and X2 ). Eventually, all the large
peptides degrade into small peptides that contain no amino acid residues susceptible to
proteolysis.[11]
In the zipper pathway (Figure 5.2b), the complex C formed between protease (E) and
substrate protein (S) undergoes sequential proteolysis at different cleavable sites. This process
releases intermediate products X1, X2,…Xn without dissociating the remaining part of the
protein from the complex. The intermediate peptides could contain missed-cleavage sites.
A first glance at the two digestion pathways suggests that the one-by-one scheme is more
favorable for generating large peptides through quenching of proteolysis at an early stage. In
principle, arresting digestion after a large peptide dissociates from an enzyme and before
formation of a second peptide-enzyme complex will lead to large peptides.

In the zipper

digestion scheme, however, a substrate protein molecule may undergo multiple cleavages before
dissociating from the protease. In a completely zipper mechanism, short-time digestion should
generate mixtures of large and small peptides and intact proteins that did not form complexes
with enzymes. Overall, the zipper mechanism should provide more s mall peptides than the oneby-one mechanism because the interaction between the substrate and the enzyme will lead to a
longer residence time in the membrane and more cleavage for many peptides.
176

Choisnard et al. suggest that at pH 4.5 and 23 ° pepsin-catalyzed hydrolysis follows the
C,
one-by-one scheme for native hemoglobin and the zipper scheme for denatured hemoglobin.[17]
As would be expected, the denatured (unfolded) hemoglobin produces a target peptide twice as
fast as the one-by-one scheme with the native protein. Denaturation also speeds up hydrolysis in
tryptic digestion.[18] The above studies suggest that the rigid, native (folded) proteins may
undergo one-by-one proteolysis , in which only a few cleavable sites of the protein are exposed
to solution and catalytic pockets. In contrast, dena tured proteins have a higher degree of
conformational freedom that leads to easier incorporation into the protease-substrate protein
complex and a higher probability for the zipper digestion pathway, i.e., digestion at multiple sites
without complete dissociation of the enzyme-substrate complex. Although denaturation may
favor generation of smaller peptides, considering that biological samples contain proteins with
varying degrees of folding, digestion after denaturation (unfolding) may yield greater unifor mity
of proteolytic peptides from different proteins. Thus, developing novel methods that “quench”
the zipper scheme by forcing the isolation of briefly digested protein pieces (including both the
released peptides, Xn in Figure 5.2b, and the unreleased residual protein, not shown in Figure
5.2b, bound to protease E) may yield large peptides. However, digestion pathways can also be
the mix of the two schemes.[11]
A few experimental studies examined peptide digestion pathways. In a study of peptide
hydrolysis in human plasma, quantitative MALDI-MS analysis showed a gradual degradation
pathway for fibrinogen peptide A (ADSGEGDFLAEGGGVR). The largest degraded form,
DSGEGDFLAEGGGVR, shows its highest abundance at ~1 h; whereas a smaller peptide,
EGDFLAEGGGVR. shows its highest abundance at ~4 h).[19] Bi et al. noticed similar gradual
degradation of model peptides upon tryptic digestion. [20] Unfortunately, these kinetics studies
177

employed only small peptides and unknown enzymes that may not be relevant to tryptic
digestion of proteins with a MW > 10 kDa.
A recent experimental analysis of peptide degradation in human p lasma suggests an
apparent first-order depletion of the substrate peptide.[19] In a simulation, Srividhya et al.
showed that protein digestion under the one-by-one pathway presents apparent first-order
depletion for the substrate protein regardless of the protease-to-substrate protein ratio.

In

contrast, the zipper pathway allows the apparent first-order protein depletion only when the
protease-to-substrate protein ratio is > 1.[11] This important difference suggests a possible
change of proteolysis pathway upon increasing the protease concentration in the digestion
environment, which may help in the design of novel methods for generating large peptides.
In traditional enzymatic digestion in solution, the protease to substrate protein ratio is low
(1/20 ~1/50) to prevent autolysis. If the digestion follows the zipper scheme (Figure 5.2b), a
small fraction of the substrate protein binds to the protease and undergoes gradual proteolysis.
The rate of protein digestion is low due to the limited ava ilability of trypsin. The problem of
slow digestion could also exist in the one-by-one digestion scheme if the reaction rate constant
k 2 in the first one or two steps is very low. Another limitation of in-solution digestion in
producing large peptides lies in the co-existence of large peptides and intact proteins. Generally,
the large peptides produced from parent proteins are more susceptible to further proteolysis than
the parent proteins. Thus, enzymatic digestion with a limited amount of protease should favor
complete digestion of the pre- formed large peptides instead of producing large peptides from
intact proteins. After quenching of the digestion process at an early stage, the majority of the
protein molecules will still be undigested. However, extending digestion time will preferentially
decrease the size of large peptides rather than forming many new large peptides.
178

Figure 5.3. Scheme for (a) protein digestion using a nylon membrane containing immobilized
trypsin and (b) the assembled setup that performs protein digestion. Both figures are adopted
with permission from figures of reference [21] with modifications.

Porous materials with immobilized proteases normally have a much higher local enzyme
“concentration” than that employed for in-solution digestion. In fact, enzyme immobilization
may provide a protease “concentration” higher than that of the substrate protein. Such a catalytic
environment should increase the proteolysis rate, as most of the substrate protein can undergo
proteolysis simultaneously. This is also more likely to give large peptides and relatively little
intact protein at short times because a large fraction of proteins may be bound to the enzyme.
Whichever pathway (or combination of the two pathways) proteolysis follows, isolation
of the large proteolytic peptides from the enzyme can prevent further degradation into smaller
179

peptides. Our trypsin- modified, microporous nylon membranes provide an ideal platform for
generating large peptides (Figure 5.3) because convection can limit contact between substrates
and proteases. The trypsin concentration in membrane pores is 450 times higher than that in a
typical in-solution digestion, and 9 times higher than the 0.1 mg/mL of substrate proteins passed
though the membrane.[21] Moreover, the micron pore size provides a short diffusion distance
that allows substrate proteins to encounter trypsin molecules in a short time. Previous bottom-up
protein MS analysis showed that -casein and denatured, alkylated BSA undergo complete
tryptic digestion in only 6 s upon passing through the membrane.[21]
This chapter explores proteolysis when the residence time in the membrane is less than a
few hundred ms. Such short exposure times are not feasible with in-solution tryptic digestion
because of the time required for mixing and quenching of digestion. Thus, many of the large
peptides obtained in rapid, membrane-based digestion are not accessible with in-solution
digestion. Moreover, employing ion/ion reactions in the gas phase reduces peptide charge states
to +1 or +2 and greatly simplifies ESI-MS spectra. The combination of the two novel methods
may lead to improved protein identification based on database searching using peptide mass
fingerprinting (PMF).

5.2 Experimental
5.2.1 Chemicals and materials
Chemicals and proteins purchased from Sigma-Aldrich include: poly(sodium 4styrenesulfonate) (PSS) (average molecular weight ~70,000), trypsin (Type I, ~10,000 Nbenzoyl- L-arginine ethyl ester hydrochloride units/mg protein, treated with L-1-tosylamide-2180

phenylethyl chloromethyl ketone, T8802), -casein from bovine milk (70%, C6780), -casein
from bovine milk (98%, C6905), apomyoglobin from horse skeletal muscle (protein sequencing
standard, A8673), lysozyme from chicken egg white (90%, L6876), dl-1,4-dithiothreitol (DTT)
and iodoacetamide (IAA).

Acetonitrile (ACN) was obtained from Mallinckrodt Baker.

Hydrophilic nylon membrane discs (25 mm diameter, 170 m thickness, 0.45 m pore size,
HNWP02500, and 1.2 m pore size, RNWP02500) were obtained from Millipore. Dialysis
cassettes (7 kDa MW cut-off, 0.5 mL capacity, Part No. 66373) were obtained from Thermo
Scientific. Sep-Pak Vac 1cc C18 cartridges (100 mg, 55-92 m beads, Part No. WAT023590)
were purchased from Waters. The low-pressure inline filter system (A-424) and Teflon tubing
(I.D. 1/32" × 1/64") were acquired from Upchurch Scientific.

5.2.2 Trypsin immobilization in nylon membranes
A UV-ozone cleaned (15 min) nylon membrane disc was inserted in an Amicon cell
(Model 8010, 10 mL, Millipore), and 10 mL of 0.02 M PSS, 0.5 M NaCl and 30 mL of water
were sequentially pulled through the membrane at 3 mL/min using a peristaltic pump. Trypsin
immobilization occurred by circulating 3 mL of 0.6 mg/mL trypsin (prepared in 2.7 mM HCl)
through the membrane at 3 mL/min for 1 h followed by rinsing with 30 mL of 1 mM HCl and
drying in air.

5.2.3 Protein solutions for tryptic digestion
-casein (average MW 23694.88 for the S1 chain with 9 reported phosphorylation sites,
and average MW 25308.51 for the S2 chain with 12 reported phosphorylation sites), -casein

181

(average MW 23983.39, including 5 phosphorylation sites) and apomyoglobin (average MW
16951.62) were dissolved in 10 mM NH4HCO 3 at 0.1 mg/mL without further treatment.
Lysozyme (average MW 14305.22) was denatured, reduced and alkylated prior to
digestion. Typically, 3 mg of lysozyme was dissolved in 600 μL of 10 mM NH4HCO 3
containing 6 M urea, followed by addition of 150 μL of 10 mM DTT (prepared in 10 mM
NH4HCO 3). The solution was placed in a 55 ° water bath for 30 min to cleave the disulfide
C
bonds, mixed with 300 μL of 100 mM IAA, and placed in the dark for 1 h to
carbamidomethylate Cys residues.
Part of the above alkylated lysozyme was dialyzed to remove small reagents. In this
procedure, 420 L of the lysozyme solution after alkylation (containing 1.2 mg protein) was
mixed with 30 L of 10 mM NH4HCO 3 and injected into the dialysis cassette. The cassette was
suspended in 300 mL of 10 mM NH4 HCO 3 for 14 h with moderate stirring. The buffer was
changed after 2 and 4 h. The resulting solution containing 2.67 mg/mL (assuming no sample loss)
of purified, alkylated lysozyme was collected from the cassette and stored in an Eppendorf tube
at 4° Fifty g of the purified lysozyme was dried and stored at -20 ° until ESI-MS analysis.
C.
C
The rest of the dialyzed lysozyme was diluted to 0.14 mg/mL with 10 mM NH4HCO 3 for tryptic
digestion.
The undialyzed, alkylated lysozyme (containing urea, DTT and IAA) was diluted to 0.1
mg protein/mL with 10 mM NH4HCO 3 for digestion.

To examine the completeness of

lysozyme alkylation with ESI-MS, 0.5 mL of the 0.1 mg/mL undigested lysozyme was acidified
(0.5% final formic acid (FA) concentration) and desalted using a C18 cartridge. The C18 beads
182

were first washed and equilibrated with acidic solutions by sequentially passing 3 mL of 0.05%
FA in 99.95% ACN and 3 mL of aqueous 0.05% FA through the cartridge. The acidified protein
solution was then forced through the cartridge followed by 0.5 mL of 0.05% FA for washing.
Protein bound to the solid phase was recovered by sequentially passing 0.2 mL of 0.05% FA
buffers containing 40%, 60%, 80% and 99.95% ACN through the cartridge. All the above
procedures were performed manually with a ~0.7 mL/min flowrate. The combined elution
fractions were finally dried with a SpeedVac.

5.2.4 Protein digestion
To perform tryptic digestion using membranes, a 4-mm-diameter membrane was cut from
the 25- mm-diameter membrane modified with trypsin, and inserted in an Upchurch low-pressure
inline filter system. A frit between the membrane and the downstream tubing (Figure 5.3b)
2

served as a support. The effective filtration area was ~2 mm .[21] Prior to digestion 20 L of
10 mM NH4HCO 3 was passed through the membrane at 20 L/min to activate the immobilized
trypsin. The protein solution (0.5 mL) was then passed through the membrane at a specific
flowrate for digestion, and the digest was collected in an Eppendorf tube and dried with a
SpeedVac.
Tryptic digestion of different protein samples was also performed in Eppendorf tubes (insolution digestion) by addition of 0.5 mg/mL trypsin (prepared in 2.7 mM HCl) to the protein
solution at a 1:50 enzyme-to-substrate protein mass ratio. Most of the in-solution digestions were
performed at room temperature, but the 16- h digestion occurred at 37° The digestion was
C.
quenched after a certain period of time by addition of glacial acetic acid to achieve 1% acetic
acid by volume. The acidified digests were then dried with a SpeedVac. For digestion of
183

lysozyme without dialysis, the denatured protein was dissolved at 0.1 mg/mL in 0.5% FA,
desalted with a C18 cartridge as described above, and finally dried with a SpeedVac.

5.2.5 Gel electrophoresis
To evaluate the completeness of membrane-based protein digestion, 0.5 mL of
undigested and digested -casein (in 10 mM NH4HCO 3) was dried and reconstituted in gel
loading buffer at 1 mg/mL.[22] Ten to twenty L of each sample (7 g) was subjected to sodium
dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (gel prepared as a 4% stacking
gel and a 12% separation gel) with Coomassie blue staining to reveal any undigested a-casein
and large peptides.

5.2.6 ESI-MS analysis and peptide identification
Infusion ESI-MS of protein or protein digests was performed on a Q-TOF mass
spectrometer (Agilent G6510) or a modified QqTOF tandem mass spectrometer (QSTAR XL,
AB/SCIEX) that allows ion/ion reactions. The mass of each detectab le ion that did not overlap
with other ions in the spectrum was calculated according to the m/z of the ion and its charge state.
For preliminary sequence assignment, those masses were compared with the theoretical average
masses of the tryptic peptides generated in silico by the ProteinProspector MS-Product program
(m/z ± tolerance). Accession numbers for obtaining sequences of horse myoglobin, bovine 0.5
casein and chick lysozyme in Swiss-Prot are P68082, P02666 and P00698, respectively (signal
peptide or initiator methionine removed). For in silico peptide generation, the protease was set
to trypsin, the maximum missed cleavages was 99, the peptide mass range was 300-30000 Da,
and the minimum peptide length was 1. Additionally, carbamidomethylation at Cys was allowed
as a constant modification of lysozyme, and phosphorylation at Ser and Thr was set as a variable
184

modification of -casein. (Peptides with phosphorylation(s) at sites other than Ser15, Ser17,
Ser18, Ser19 and Ser35 were not included in the theoretical peptide candidates).
For ESI-MS performed with the Agilent G6510 (collaborated with Dr. Scott Smith from
Prof. Gavin Reid‟s group at Michigan State University), the ESI source capillary voltage was 3.5
kV in positive ionization mode. Digests of lysozyme (after desalting with C18 cartridges), casein and apomyoglobin were reconstituted to 30, 5 and 5 M, respectively, in 1% FA in 50%
methanol for infusion ESI-MS analysis at 3 L/min. Because infusion ESI mass spectra of
digests are often complicated by peptide ions with different charge states, some signals from
different peptide ions may overlap and give incorrect m/z values for peptide assignments. To
further confirm the “purity” of each peptide ion in the spectra, the theoretical peptide peak with
specific amino acid composition, charge states and mass spectrometer resolving power was
simulated using IsoPro 3.0. Each experimental peptide peak was compared with the simulated
peak (see examples in Figures 7, 11, 20 and 22), and only those with ±
0.05 m/z accuracy and
similar isotopic peak distributions were confirmed as “identified”.
To simplify ESI-MS spectra, ion/ion reactions were performed in a modified QqTOF
tandem mass spectrometer to reduce the charge state of peptide ions to mainly +1 (with minor
peaks due to +2 species). The experiments and peptide assignment were performed by Ken
Chanthamontri from Prof. Scott McLuckey‟s group in the Department of Chemistry at Purdue
University. Briefly, the ESI spectra of ~20-50 M protein digests containing 50/50/1 (v/v/v) of
methanol/water/acetic acid were directly formed by nano-ESI, followed by ion- ion proton
transfer reactions in Q2 to facilitate the interpretation of spectra. To examine the sequence
coverage of the parent proteins based on the simplified tryptic peptide ions, all the ions were
compared to the in silico generated peptide ions (as described above). Experimental ions were
185

considered as +1 charge states for peptide sequence assignment, with the exception of unresolved
peaks above m/z 5000, where were considered as +1 or +2 charge states.

For protein

identification, all the ions in the post- ion/ion-reaction mass spectra with S/N  5 were considered
as peptides generated from an unknown protein, and subjected to MASCOT PMF searching.
Parameters for database searching included: database (combination of SwissProt, NCBI and
MSDB), protease (trypsin), allowed missed cleavages (9, the maximum available number),
taxonomy (mammals for horse myoglobin and bovine milk -casein, and animals for chick
lysozyme), peptide mass tolerance (1 Da), m/z type (average) and protein mass (provided if
detectable

in the

mass spectrum).

In addition,

fixed

modification was set as

carbamidomethylation at Cys for lysozyme digests, and variable modifications were set as
phosphorylation at Ser/Thr for -casein digests.

5.3 Results and discussion
5.3.1 Short residence times in trypsin-modified membranes yield large tryptic
peptides
In our previous study on -casein digestion within trypsin- modified membranes, gel
electrophoresis showed no intact protein, even with a membrane residence time of only 0.8 sec
(residence times are calculated assuming 50% membrane porosity).[21] Moreover, the resulting
tryptic peptides generally have <10 amino acids. Nevertheless, because the nylon membranes
are only 170 m thick, a simple syringe pump can force the protein solution through the
membrane rapidly enough to achieve a residence time of a few ms. Such a short digestion time

186

together with the resulting immediate isolation of the digest from trypsin may result in large
peptides.

Figure 5.4. SDS-PAGE of intact -casein (lane 1) and -casein digests generated by passing 0.1
mg/mL -casein (in 10 mM NH4 HCO 3) through a trypsin membrane (0.45 m pore size) using
residence times of 1.5 ms (lane 2) and 3.6 ms (lane 3). -S1 and -S2 represent the S1 and S2
chains of -casein, respectively. The loading of protein or protein digests for lanes 1-3 is 7 g.
Gels were stained with Coomassie blue.
Using a trypsin- modified membrane (0.45 m pore size), we performed -casein
digestion at flowrates of 7.2 and 3 mL/min, which give 1.5- ms and 3.6 ms residence times,
respectively, in the membrane. Figure 5.4 suggests that the small difference in residence times
leads to a significant difference in peptide size pattern. With a 3.6 ms residence time, most of
the S-1 and S-2 chains of -casein (both ~25 kDa) hydrolyzed to peptides with a MW of ~5 kDa
(Figure 5.4, lane 3). In contrast, a 1.5- ms residence time generated several peptides with a MW
between 10 and 20 kDa (Figure 5.4, lane 2) without giving rise to an increase in undigested
protein (note the similar amount of residual -casein in lanes 2 and 3 at ~25 kDa). Furthermore,
the strong band at ~5 kDa in lane 2 is a little darker than that in lane 3, suggesting that more
187

41

Sequence coverage 86%
44

26
35

25

5

25
8

15

28
41
21 51
33

44

7
48

8

26

400

500

600

700

m/z

5

2

*

14

52

*

800

41
1
*
900

*
1000

2
*

*
1100

Figure 5.5. ESI-Q-TOF mass spectrum of apomyoglobin digested for 2 h in solution. The black numbers refer to peptides with
no missed cleavage sites, the green numbers refer to peptides with 1 missed cleavage site but MW < 2500 kDa, the red
numbers refer to peptides with 1 missed cleavage site and MW  2500 kDa, and * refers to undigested protein. The inset
presents the peptide pattern, with the horizontal lines showing the possible tryptic cleavage sites. (Peptide number refer to Table
5.1.)

188

8

Sequence coverage 100%
26
44

15

25

25

35
41
45

33

2

51 30
16 41
5
21 28
28

8

48

400

500

42

37
42

33

600

26

44 *
*

700

m/z

*

*

*

*

*
*

37

2

*

*

800

900

1000

1100

Figure 5.6. ESI-Q-TOF mass spectrum of apomyoglobin digested for 30 min in solution. The black numbers refer to peptides
with no missed cleavage site, the green numbers refer to peptides with 1 missed cleavage site but MW < 2500 kDa, the red
numbers refer to peptides with 1 missed cleavage site and MW  2500 kDa, and * refers to undigested protein. The inset
presents the peptide pattern, with the horizontal lines showing the possible tryptic cleavage sites. (Peptide numbers refer to Table
5.1.)

189

a

Theoretical
m/z (av)= 368.48

b

c
Theoretical
m/z (av)= 818.35

[#37+5H]5+
in Figure 6

[#35+2H]2+
in Figure 6

368

369
m/z

Theoretical
m/z (av)=
998.15

370 817

818 m/z 819

820

[P+5H]17+
In Figure 6

998

m/z 998.5

Figure 5.7. Comparison of typical experimental (bottom) and simulated (top) spectra of identified apomyoglobin peptides or
protein. Simulations were performed using peptide (or protein) composition and charge state, with the mass spectrometer
resolving power set as 18000. The peptide number in Table 5.1 (or protein, labeled as P), theoretical m/z, and the complete mass
spectrum for the experimental ion are given in each spectrum.

190

-casein degraded to peptides below 5 kDa (not visible in SDS-PAGE) after a 3.6-ms digestion
compared to a 1.5- ms digestion.

The above results demonstrate that the tryp sin- modified

membrane digests intact -casein very efficiently even when the digestion time is a few ms. In
addition, variation of the protein residence time in the membrane affords some control over
peptide size.
Unfortunately, due to the low purity of the -casein (70%) and the existence of two
chains with similar MWs, infusion ESI-MS gave very complex spectra for -casein digests.
Because many of the large peptides (> 3000 Da) overlapped in the spectra, we could not
determine the exact m/z of each peptide. An accurate m/z value is crucial to determine the
peptide sequence by comparing the experimental m/z values with the theoretical m/z values of
tryptic peptides generated in silico from the known sequence of the parent protein.
We then chose sequencing grade apomyoglobin for further examinations of digestion in
both solution and membranes.

Figure 5.5 shows that 2-h, in-solution digestion extensively

hydrolyzes apomyoglobin (17 kDa) and results in a mass spectrum dominated by small peptides
with few, if any, missed cleavages. The 15 detected peptides cover only 86% of the protein
sequence. Decreasing the in-solution digestion time to 30 min leads to several large peptides,
along with 100% sequence coverage from the combination of peptides (Figure 5.6). However,
the intensity for signals due to small peptides (black numbers) and undigested protein(*‟s) are
quite strong, suggesting a low efficiency in producing large peptides. Figure 5.7 shows the
isotopic distributions of typical apomyoglobin peptide/protein ions that show no evident overlap
with other ions. The high similarity of the experimental and simulated isotope distributions
demonstrates the feasibility of using infusion ESI- mass spectra for assigning apomyoglobin
peptide sequences.

191

In contrast to solution digestion, passing the apomyoglobin solution through a trypsinmodified membrane using a 50- ms residence time cleaves the 17 kDa protein into mainly
peptides with at least one missed cleavage sites (green and red numbers), many of which are
above 2500 Da (red numbers) (Figure 5.8). The combination of only 5 of these large peptides
covers the entire protein sequence, in contrast to the need for 8 peptides to cover the whole
sequence using 30- min in-solution digestion (Figure 5.6). Moreover, compared to in-solution
digestion, proteolysis in the modified membrane gives lower signals for undigested protein.
The above result suggests different apomyoglobin digestion pathways in the trypsinmodified membrane and in solution. For the 30-min in-solution digestion, the combination of
signals from highly abundant ions of small peptides, large peptides and intact protein (Figure 5.6)
suggests either a partial zipper digestion pathway (Figure 5.2b) or increasing rates of digestion as
the sizes of the substrates decrease. Extending the time for in-solution digestion to 2 h depletes
most of the intact apomyoglobin. However, the large peptides also hydrolyzed to small peptides
(Figure 5.5). In contrast, membrane digestion for 50 ms depletes the majority of apomyoglobin,
and the prompt isolation of the large peptides from the digestion system prevents their further
hydrolysis (Figures 8).
Because infusion ESI produces peptide and protein ions with different charge states, ESI
mass spectra are often complicated, and comparison of the relative intensities of different
peptides is difficult. McLuckey‟s group at Purdue University recently developed gas-phase ionion reactions to reduce multiply charged peptide ions to singly or doubly charged ions.[23]
Using this technique, the membrane-digested apomyoglobin shows predominantly peptides with
missed cleavages, and most of the strong signals come from peptides with masses above 3000 Da
(Figure

5.9).

192

2

Sequence coverage 100%
45

37

26

25 8

47
45
37
9
44
51
27
34 25
33
27 33
10
21 46
30
15
36
10

400

500

600

34 47
26

700
m/z

37
34
*
*

800

*

900

*

2

*

1000

1100

Figure 5.8. ESI-Q-TOF mass spectrum of apomyoglobin digest generated in a trypsin- modified membrane (50-ms residence time,
0.45 m pore size). The black numbers refer to peptides with no missed cleavage site, the green number refer to peptides with1
missed cleavage site but MW < 2500 kDa, the red numbers refer to peptides with 1 missed cleavage site and MW  2500 kDa,
and * refers to undigested protein. The inset presents the peptide pattern, with the horizontal lines showing the possible tr yptic
cleavage sites. (Peptide numbers refer to Table 5.1.)

193

2

80
70

37

60
50

8
26

40
30
20
10

45
33
27
49

Peptides above
100%SCT (10%)

30

46 47 9 10
28

34
11

36

*

0
0

0

2000

2000

4000

4000

6000

6000

8000

8000

10000

m/z

10000

12000

12000

14000

14000

16000

16000

Figure 5.9. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of apomyoglobin digested in a trypsinmodified membrane (50 ms residence time, 0.45 m pore size). The black numbers refer to peptides with no missed cleavage
sites, the green numbers refer to peptides with 1 missed cleavage site but MW < 2500 kDa, the red numbers refer to peptides
with 1 missed cleavage site and MW  2500 kDa, and * refers to undigested protein. The inset presents the peptides detected
above the 100% sequence coverage threshold (SCT), with the horizontal lines showing the possible tryptic cleavage sites.
(Peptide numbers refer to Table 5.1)

194

Table 5.1. Apomyoglobin (average MW 16951 Da) tryptic peptides detected by ESI-MS.
Peptide Missed Sequence
#
cleavage #
1

0

1-16

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

1
12
16
0
2
0
2
4
6
7
8
18
1
1
3
5
7
5
8
1
2
3
6
0
1

1-31
1-96
1-133
17-31
17-45
32-42
32-47
32-56
32-63
32-77
32-78
32-153
43-47
48-56
48-63
48-78
48-96
51-79
51-102
57-63
57-77
57-78
63-102
64-77
64-78 or
63-77

Av. m/z
+

Peptide Missed Sequence
#
cleavage #

[MH]
1815.9

27

2

3403.7
10611.6
14677
1606.8
3249
1271.6
1937.0
3004.5
3775.9
5135.8
5263
13557
684.3
1086.5
1857.9
3345
5308
3095
5647
790.4
2150.2
2278.3
4314
1378.8
1506.9

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

2
6
1
5
6
0
4
1
2
3
7
2
6
0
1
0
0
1
2
3
0
2
0
1
0

Av. m/z
+

[MH]
64-79 or 1635.0
63-78
78-96
2110.1
78-133
6176
79-96
1982.0
79-133
6048
79-139
6777
80-96
1853.9
80-133
5920.0
97-102
735.4
97-118
2601.4
97-133
4085.1
97-153
6349.3
99-133
3819.9
99-153
6084
103-118 1885.0
103-133 3368.6
119-133 1502.6
134-139 748.4
134-145 1360.7
134-147 1651.9
134-153 2283.2
140-145 631.3
140-153 1553.7
146-147 310.1
146-153 941.4
148-153 650.3

A good middle-down digestion should hydrolyze the substrate protein to large peptides
and avoid generating large quantities of small peptides. Smaller peptides are generally less
informative in providing sequence and PTM information if their sequences are already covered
by a larger peptide. Moreover, loss of some small peptides during HPLC-ESI-MS results in gaps
in protein sequence coverage if only those small peptides cover that portion of the sequence. To
evaluate middle-down digestion, we define a “100% sequence coverage peak-intensity threshold”
195

(100%SCT) for the post- ion/ion reaction mass spectra and determine the number of peptides
with intensities above this threshhold. At this intensity cutoff, the detected tryptic peptides just
cover the entire protein sequence. For example, in Figure 5.9, the 12 tryptic peptides detected
above the 100%SCT (10% relative MS intensity) cover the entire apomyoglobin sequence.
Setting the selection level above 10% intensity leads to the loss of at least one tryptic peptide and
consequently <100% protein sequence coverage. The fewer tryptic peptides above this threshold,
the more efficient the middle-down digestion is.
Similarly, we performed tryptic digestion of -casein, a 24 kDa protein with 5
phosphorylation sites. A standard 16- h, in-solution digestion produces only 8 small peptides that
are detectable by ESI-Q-TOF (Figure 5.10). (Similar isotopic distributions of the identified
peptide ions and simulated ions, such as those in Figure 5.11, suggest the feasibility of using this
ESI-MS for assigning small -casein peptide sequences.) These small peptides cover only 20%
of the protein sequence, and none of them include the phosphorylation sites. Shortening the insolution digestion time to 5 min generates much larger peptides and no residual protein ions.
However, many of the large peptides with a MW over 10 kDa overlap in the ESI-Q-TOF mass
spectrum and cannot be unambiguously assigned. Thus, we performed ion/ion reactions to
reduce the charge states of these peptides and generate a mass spectrum with only singly charged
peptide ions (Figure 5.12). Although most of the strong signals appear in the 2500-8000 m/z
region, the sequence coverage of the total detected peptides is still 98%. The missing piece is
(R)INK(K), which consists of amino acid residues 26-28 and has a MW of 374 Da. Interestingly,
this small peptide is not detectable in the mass spectrum generated from the 16- h digestion either
(Figure 5.10), suggesting either a low ionization efficiency for this small peptide or further
degradation during digestion.

196

30

Sequence coverage 20%
37

25 33

16
14 22

400

8

500

25

16

33

30

600

700

800

900

1000 1100
m/z

1200 1300 1400 1500 1600 1700

Figure 5.10. ESI-Q-TOF mass spectrum of a -casein tryptic digest generated by 16-h in-solution digestion. The black numbers
refer to peptides with no missed cleavage site, and the green numbers refer to peptides with 1 missed cleavage site but MW <
2500 kDa. The inset presents the peptide pattern, with the horizontal lines showing the possible tryptic cleavage sites, and the
blue P‟s showing the phosphorylation sites. (Peptide numbers refer to Table 5.2)

197

a

Theoretical
m/z (av)= 507.62

b

Theoretical
m/z (av)= 830.97

[#33+H]+
in Figure 10

[#22+2H]+
in Figure 10

507

508
m/z

509

830

831 832
m/z

833

Figure 5.11. Comparison of typical experimental (bottom) and simulated (top) spectra of
identified -casein peptide ions. Simulations were performed using peptide composition and
charge state, with the mass spectrometer resolving power set as 18000. The peptide number,
theoretical m/z, and the complete mass spectrum for the experimental ion are given in each
spectrum.
Further decreasing the in-solution digestion time to 1 min gives 100% sequence coverage,
as the post- ion/ion reaction mass spectrum shows (Figure 5.13). Residues 26-28 now appear as
part of larger peptides 1-28 (peptide #2) and 26-99 (peptide #7). We expected that the shorter
digestion time would yield fewer but larger peptides. However, this brief digestion still generates
14 detectable peptides above 100%SCT, although only 4 peptides can constitute the entire
protein sequence.
The minimum time required to perform in-solution digestion in our hands is 20 s. Even
after this short period, little intact -casein remains in the digest. The peptide pattern moves

198

toward even higher masses (compare Figures 5.13 and 5.14), but the number of tryptic peptides
above 100%SCT drops to 13, and only 3 peptides cover the entire protein sequence.
We then compared the short-time, in-solution digestion with membrane digestion. A 2ms residence time for -casein in the 0.45 m membrane generates a similar peptide pattern as
the 20-s in-solution digestion (compare Figures 5.14 and 5.15), suggesting the same digestion
pathway in both methods. The 2- ms membrane digestion gives 12 peptides above 100%SCT,
among which 4 cover the entire protein sequence.
In addition to expanding sequence coverage, another advantage of middle-down protein analysis
lies in the simultaneous characterization of many PTMs in a single peptide sequence
(“combinatorial” PTM characterization).[24] -casein is phosphorylated at Ser15, Ser17, Ser18,
Ser19 and Ser35.

Previous analysis of -casein phosphorylation based on small tryptic

phosphopeptides inevitably required phosphopeptide enrichment from the complete -casein
digest.[25] Due to the existence of one Arg and three Lys residues between Ser35 and the other
four phosphorylation sites, simultaneous characterization of the five phosphorylation sites on a
single peptide is almost impossible after 16- h, in-solution tryptic digestion. Even with the 20-s,
in-solution digestion and 2- ms membrane digestion, post-ion/ion reaction mass spectra without
prior phosphopeptide enrichment do not yield detectable peptides that extend across the five
phosphorylation sites (Figures 14 and 15). Further increasing the flowrate through the 0.45 m
membrane in an effort to generate such a large peptide was not possible with the syringe pump
that we used.
The similar patterns of predominantly large -casein peptides generated in the membrane
and short-time solution digestion suggest that the proteolysis follows the one-by-one scheme in
both digestion formats. The high efficiency of intact protein depletion in the two methods
199

Sequence coverage 98%

100

100

28 26 23
36

80

80
60

60

9

40

40

20

20

0

0

37
16

0

0

30

1
33

34

11
32
5000

5000

10000

m/z

10000

15000

15000

Figure 5.12. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of -casein digested in solution for 5 min.
The black numbers refer to peptides with no missed cleavage site, and the red numbers refer to peptides with 1 missed cleavage
site and MW  2500 kDa. The inset presents the peptide pattern, with the horizontal lines showing the possible tryptic cleavage
sites, and the blue P‟s showing the phosphorylation sites. (Peptide numbers refer to Table 5.2.)

200

Sequence coverage 100%

120
100
80
60
40
20

120

34

100

Peptides above
100%SCT (10%)

32

80

+2 charged
species

60

40

16 30
3331

20

0

36
2

28

26 23 18
9 10
7 13

0

0

0

5000

5000

10000

m/z

10000

15000

15000

Figure 5.13. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of -casein digested in solution for 1 min.
The black numbers refer to peptides with no missed cleavage site, the green numbers refer to peptides with 1 missed cleavage
site but MW < 2500 kDa, and the red numbers refer to peptides with 1 missed cleavage site and MW  2500 kDa. The inset
presents the detected peptides above 100%SCT, with the horizontal lines showing the possible tryptic cleavage sites, a nd the blue
P‟s showing the phosphorylation sites. (Peptide numbers refer to Table 5.2.)

201

Sequence coverage 100%

100

100

Peptides above
100%SCT (14%)

32

80

80

10
26
2318

60

60

3 34

40

40

36

20

20

0

0

0

0

9

2

28
5000

5000

7

13
19

27
20

29

10000

m/z

10000

24 21
15
15000

15000

Figure 5.14. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of -casein digested in solution for 20 s.
The black numbers refer to peptides with no missed cleavage site, and the red numbers refer to peptides with 1 missed cleavage
site and MW  2500 kDa. The inset presents the detected peptides above 100%SCT, with the horizontal lines showing the
possible tryptic cleavage sites, and the blue P‟s showing the phosphorylation sites. (Peptide numbers refer to Table 5.2.)

202

50

40

30

20

10

Peptides above
100%SCT (22%)

0

0

0

5000

50

50

10000
20000
10000 15000 20000
m/z
32

25000

40

40

Sequence
34
coverage 100%

30

30

26

23

20

20

36

10

10

16

2
1

10
18 12
7 13
9
19

27
24
21

28

3 4

29

0

0

0

0

5000

5000

10000

m/z

10000

15000

15000

Figure 5.15. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of -casein digested in a trypsin- membrane
membrane (2 ms residence time, 0.45 m pore size). The black numbers refer to peptides with no missed cleavage site, and the red
numbers refer to peptides with 1 missed cleavage site and MW  2500 kDa. The inset presents the peptides detected above
100%SCT, with the horizontal lines showing the possible tryptic cleavage sites, and the blue P‟s showing the phosphorylation
sites. (Peptide numbers refer to Table 5.2.)

203

Table 5.2. -casein (average MW 23983 Da) tryptic peptides detected by ESI-MS.
Peptide # Missed
#
cleavages

Sequence

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

1-25
1-28
1-29
1-32
1-169
26-29
26-99
29-32
29-97
29-99
30-97
33-105
33-107
98-105
98-209
100-105
100-107
100-169
100-176

1
2
3
4
11
1
5
1
3
4
2
3
4
1
8
0
1
3
4

m/z
+
[MH]
3122
3477
3605
3976
19518
502
8445
517
7863
8090
7734
8214
8485
873
12680
646
911
7986
8748

Peptide
#

# Missed Sequence
cleavages

20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37

5
7
1
2
6
0
1
5
0
4
0
1
3
0
2
0
1
0

100-183
100-209
106-113
106-169
106-209
108-113
108-169
108-209
114-169
114-209
170-176
170-183
170-209
177-183
177-209
184-202
184-209
203-209

m/z
+
[MH]
9559
12452
1013
7358
11825
748
7093
11559
6363
10829
780
1591
4482
830
3721
2186
2909
742

implies both fast digestion of certain labile sites and rapid dissociation of -casein from trypsin
after each single-site proteolysis.

5.3.2 Larger membrane pores allow briefer tryptic digestion
To circumvent the flux limitation in the modified membrane with 0.45 m pores, we
prepared trypsin- modified membranes with 1.2-m pores. With the larger pore size, a 10-ms
residence time generates several peptides with MWs from 4-20 kDa, and undigested -casein is
also visible. The number of peptides above 100%SCT is only 7, and two large peptides can cover
the entire sequence (Figure 5.16).

Moreover, peptide #5 extends over all the five

phosphorylation sites. Diffusion limitations could result in the significant signals from intact

204

Figure 5.16. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of -casein digested in a trypsin- membrane
membrane (10 ms residence time, 1.2 m pore size). The red numbers refer to peptides with 1 missed cleavage site and MW 
2500 kDa, and * refers to undigested protein. The inset presents the peptides detected above 100%SCT, with the horizontal lines
showing the possible tryptic cleavage sites, and the blue P‟s showing the phosphorylation sites. (Peptide numbers refer to Table
5.2.)

205

Figure 5.17. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of apomyoblobin digested in a trypsinmembrane membrane (50 ms residence time, 1.2 m pore size). The black numbers refer to peptides with no missed cleavage site,
the green numbers refer to peptides with 1 missed cleavage site but MW < 2500 kDa, the red numbers refer to peptides with 1
missed cleavage site and MW  2500 kDa, and * refers to undigested protein. The inset presents the peptides detected above
100%SCT, with the horizontal lines showing the possible tryptic cleavage sites. (Peptide numbers refer to Table 5.1.)
206

protein. In 10 ms, a protein with a diffusion coefficient of 10

-7

2

cm /sec would diffuse on

average around 0.4 m. Thus some protein molecules may not have time to diffuse from the
solution to the enzyme immobilized at the pore walls. This may be especially true for regions
with pore diameters on the higher side of the distribution. Additionally, the surface area to
volume ratio decreases as the pore size increases, so the 1.2 m membrane will also contain a
lower amount of enzyme than the 0.45 m membrane. This could lead to more intact protein and
larger fragments in the 1.2 m system.
Similar to -casein digestion, 50-ms digestion of apomyoglobin in the modified
membrane with 1.2 m pores generates only 6 peptides above 100%SCT (Figure 5.17), in
contrast to 12 peptides for the same protein and residence time using a membrane with 0.45 m
pores (Figure 5.9). Among the large peptides generated with the 1.2 m pores, just two peptides
can cover the entire protein sequence, while digestion in the membrane with 0.45 m pores
requires at least 6 peptides for complete coverage in the post ion-ion reaction mass spectrum.

5.3.3 Influence of denaturation on tryptic digestion of lysozyme
Lysozyme is a 14 kDa protein with 4 disulfide bonds. Direct digestion of this protein in
10 mM NH4 HCO 3 using trypsin- modified membranes (6.5 s residence time) resulted in
identification of only a few small peptides (data not shown) by MALDI-MS. Reduction of the 4
disulfide bonds of lysozyme followed by carbamidomethylation of the Cys residues unfolds the
protein and facilitates protein digestion. The mass spectra in Figure 5.18 show the expected
mass increase for alkylation of all 8 cysteines. (For the 8+ species the inc rease in m/z should be
equivalent

to

the

added

mass

for

one

207

alkylation

or

58.08

m/z

units.)

Original lysozyme

Theoretical MW = 14305.22 Da

Av. m/z ([M+8H]8+) = 1789.13,

Lysozyme reduced by DTT, then carbamidomethylated by IAA
Av. m/z (M+8H]8+)
= 1847.21

Theoretical MW = 14769.71 Da

1790

1800

1810

1820
m/z

1830

1840

1850

1860

All the 4 disulfide bonds are reduced and fully carbamidomethylated at the 8 Cys residues
Figure 5.18. ESI-Q-TOF mass spectrum of lysozyme before (top view) and after (bottom view) carbamidomethylation of Cys
residues.

208

Sequence coverage 84%

12 10

19

43

25
9

22 10

20
28

1
27 41 22

500

600

28
19

12

700

800

24

900
m/z

1000

1100

1200

1300

Figure 5.19. ESI-Q-TOF mass spectrum of alkylated lysozyme (without dialysis) digested for 16 h in solution. The black numbers
refer to peptides with no missed cleavage sites, and the green numbers refer to peptides with 1 missed cleavage site but MW <
2500 kDa. The inset presents the peptide pattern, with the horizontal lines showing the possible tryptic cleavage sites. (Peptide
numbers refer to Table 5.3.)
209

Theoretical
m/z (av)=
1847.21

[alkP+8H]8+
in Figure 18

1846

1847
m/z

Theoretical
m/z (av)=
585.62

[#19+3H]3+
in Figure 19

1848

585

586
m/z

Theoretical
m/z (av)=
1326.53

[#22+H]+
in Figure 19

1325

1327
m/z

1329

Figure 5.20. Comparison of typical experimental (bottom) and simulated (top) spectra of identified lysozyme peptides (or protein).
Simulations were performed using peptide (or protein) composition and charge state, with the mass spectrometer resolving powe r
set as 16000. The peptide number (or protein, labeled as P), theoretical m/z, and the complete mass spectrum for the experimental
ion are given in each spectrum.

210

3

40

Sequence
coverage 91%

3
39 43

41
42

500

34
32

600

33
39

700

42
37

40
28
3

31

34

800
m/z

900

1000

15

1100

Figure 5.21. ESI-Q-TOF mass spectrum of alkylated lysozyme (without dialysis) digested for 20 s in solution. The black numbers
refer to peptides with no missed cleavage site, the green numbers refer to peptides with 1 missed cleavage site but MW < 2500
kDa, and the red numbers refer to peptides with 1 missed cleavage site and MW  2500 kDa.. The inset presents the peptide
pattern, with the horizontal lines showing the possible tryptic cleavage sites. (Peptide numbers refer to Table 5.3.)

211

3
40

Sequence
coverage 91%

3

39

43 41
42

34
32

500

600

33 38 37 42
39

700

28
3

800
m/z

40

31 15
34

900

1000

1100

Figure 5.22. ESI-Q-TOF mass spectrum of alkylated lysozyme (without dialysis) digested in a trypsin- modified membrane (2 ms
residence time, 0.45 m pore size). The black numbers refer to peptides with no missed cleavage site, the green number refer to
peptides with 1 missed cleavage site but MW < 2500 kDa, and the red numbers refer to peptides with 1 missed cleavage site and
MW  2500 kDa. The inset presents the peptide pattern, with the horizontal lines showing the possible tryptic cleavage sites.
(Peptide number refer to Table 5.3.)

212

a

Theoretical
m/z (av)=
511.59

b

Theoretical
m/z (av)=
837.30

[#28+3H]3+
in Figure 23

[#42+3H]3+
in Figure 23

511

512

m/z

837

838
m/z

c

Theoretical
m/z (av)=
1062.23

[#31+6H]6+
in Figure 23

1062
m/z

1062.5

Figure 5.23. Comparison of typical experimental (bottom) and simulated (top) spectra. Simulations were performed using peptide
composition and charge state, with the mass spectrometer resolving power set as 10000. The peptide number, theoretical m/z, and
the complete mass spectrum for the experimental ion are given in each spectrum.

213

Similar to apomyoglobin and -casein, changing the in-solution digestion time for
alkylated lysozyme from 16 h to 20 s increases the protein sequence coverage from 84% to 91%.
Moreover, the tryptic peptides change from predominantly small, completely cleaved peptides to
large peptides with at least one missed cleavage (compare Figures 19 and 21). Figures 20 and 23
compare spectra of several lysozyme peptide/protein ions with their corresponding simulated
spectra. The similar isotopic distributions suggest the feasibility of using these spectra for
peptide sequence assignment.
*
*

*
*
*

*

*
*

300

400

500

600

*

700

800

900

1000

1100

1200

Figure 5.24. ESI-Q-TOF mass spectrum of alkylated, dialyzed lysozyme after digestion in a
trypsin- modified membrane (0.32-s residence time, 0.45 m pore size). * refers to undigested
lysozyme.
Membrane digestion (0.45 m pore size, 2 ms residence time) of the same lysozyme
sample produces a tryptic peptide pattern similar to that after a 20-s in-solution digestion
(compare Figures 21 and 22), but many large peptides with different charge states overlap and
prevent peptide identification. Unfortunately, using ion/ion reactions to reduce the charge state
of the peptides failed, probably due to the presence of small reagents (e.g. urea, DTT and IAA

214

used for protein denaturation, reduction and alkylation) that were not completely removed during
C18 desalting
To circumvent such problems in obtaining post- ion/ion reaction mass spectra, we
dialyzed the alkylated lysozyme to remove the small reagents prior to membrane digestion.
Interestingly, after removing those reagents, digestion of the alkylated lysozyme in the 0.45 mpore size membrane slows dramatically. Even with a 0.32 s residence time, undigested lysozyme
dominates the normal ESI-Q-TOF mass spectrum and interferes with the detection of large
peptides (Figure 5.24). After ion/ion reactions to reduce the charge states of ions to primarily +1,
a number of large peptides appear in the spectrum, although their abundances are much lower
than that of lysozyme (Figure 5.24). The number of tryptic peptides detected above 100%SCT is
only 4, and 2 peptides cover the entire protein sequence. However, one of the 2 peptides
includes almost the entire protein.

Although Figure 5.24 shows a much stronger signal for

undigested lysozyme than for other peptides, the detected peptides are mainly large ones with
MW above 2500 Da. The dramatically different hydrolysis pathway in the digestion samples
with and without dialysis suggests a conformational change of the alkylated lysozyme upon
removal of the small reagents, most notably urea. Interestingly, 4-fold longer digestion in the
same membrane (1.3-s residence time) leads to a complete depletion of lysozyme, and generates
mainly small peptides (Figure 5.25). Thus the peptide lifetime in the membrane must be around
1 s.

5.3.4 Protein identification by peptide mass fingerprinting (PMF)
Large tryptic peptides could potentially prove very useful in protein identification. Thus,
we submitted all the detected ions (not only identified peptides) from different protein digests to
MASCOT for PMF searching, and Table 5.4 shows the results.
215

Membrane digestion of

[P+2H]2+
*

[8+Na]+

6 17 21

Peptides above
100%SCT (12%)

[P+H]+
*
8000

2000
3

4000

22
39 20 32
40
25 42
33
1000

2000

18

7

8000
m/z

6000

10000

12000
m/z

10000

28
29

23

3000

14

[P+3H]3+
*
30
415

3734
4000

m/z

5000

14000

12000

14000

16000

[18+2H]2+
[8+2H]2+
31
16

6000

7000

Figure 5.25. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of alkylated, dialyzed lysozyme digested in a
trypsin- membrane membrane (0.32-s residence time, 0.45 m pore size). The black numbers refer to peptide with no missed
cleavage site, the green numbers refer to peptide with 1 missed cleavage site but MW < 2500 kDa, the red numbers refer to
peptides with 1 missed cleavage site and MW  2500 kDa, and * refers to undigested protein. The inset presents the peptides
detected above 100%SCT, with the horizontal lines showing the possible tryptic cleavage sites. (Peptide numbers refer to Table
5.3.)
216

Sequence coverage 78%
Impurities

41

19

20

22 [22+Na]+
25 35

1000
0

2000

1500
4000

2000
6000

m/z

28 [28+Na]+
29

2500

8000
m/z

3000
10000

23

3500
12000

14000

16000

Figure 5.26. ESI-Q-TOF mass spectrum (after charge reduction by ion/ion reactions) of alkylated, dialyzed lysozyme digested in a
trypsin- membrane membrane (1.3 s residence time, 0.45 m pore size). The black numbers refer to peptides with no missed
cleavage site, the green numbers refer to peptides with 1 missed cleavage site but MW < 2500 kDa, and the red numbers refer to
peptides with 1 missed cleavage site and MW  2500 kDa. The inset presents the peptide pattern, with the horizontal lines
showing the possible tryptic cleavage sites. (Peptide numbers refer to Table 5.3.)

217

Table 5.3. Tryptic peptides of lysozyme after Cys carbamidomethylation (average MW 14770
Da after carbamidomethylation) detected by ESI-MS.
Peptide
#

# Missed
cleavages

Sequence

m/z
+
[MH]
606
1480
1636
5212
6947
8421
or 10912

1
2
3
4
5
6
7

1
2
3
6
7
9
10

8
9
10
11
12
13
14
15
16
17
18
19
20
21
22

15
0
0
1
0
1
2
3
5
6
13
0
0
8
0

1-5
1-13
1-14
1-45
1-61
1-73
1-96
2-97
1-125
2-5
6-13
6-14
15-21
15-33
15-45
15-61
15-73
15-96
15-129
22-33
34-45
34-116
46-61

14284
478
893
1049
874
2181
3590
5325
6802
9293
13151
1325
1428
9474
1754

Peptide
#

# Missed
cleavages

Sequence

23
24
25
26
27
28
29

2
0
1
2
0
0
1

46-73
62-68
62-73
62-96
69-73
74-96
74-97

m/z
+
[MH]
3228
994
1491
3983
517
2509
2636

30
31
32
33
34
35
36
37
38
39
40
41
42
43

3
7
1
2
6
0
1
5
4
1
3
0
2
1

74-114
74-129
97-112
97-114
97-129
98-112
98-114
98-129
113-129
115-125
115-129
117-125
117-129
126-129

4566
6368
1803
2074
3874
1676
1945
3746
2090
1333
1819
1045
1531
505

apomyoglobin “identified” the parent protein with high scores, although the mass of identified
horse myoglobin is a little different from the experimental mass. In-solution digestion of casein for 5 and 1 min failed in protein identification based on the resulting peptides. The 20-s
in-solution digestion gives a little higher score than membrane digestion. Note that the allowed
number of missed cleavage sites is 9. However, the two membrane digestions both generated
large peptides that covered more than 9 missed cleavage sites (Figures 15 and 16). These large
peptides should contribute greatly to PMF searching if a higher number of missed cleavages is
allowed for PMF searching.
218

Brief lysozyme digestion using modified membranes results in accurate protein
identification (exact matched protein MS), while a more complete digestion (0.32 s membrane
digestion) yields a lower score and less accurate protein identification. With longer digestion,
loss of part of the protein sequence likely leads to uncertainty in identifying the parent protein
from a group of proteins where the majority of the sequence identical. Overall, these results
suggest that rapid membrane digestion has the potential to greatly facilitate protein
indentification.

Table 5.4. Protein identification based on MASCOT PMF searching using ions generated by
post-ion/ion reaction ESI-MS.
“Unknown” protein
Name
av. MW
(Da)
Myoglobin 16952
(horse)

-casein
(bovine)

Lysozyme
(chick)

23583*

14770**

Digestion method
Type
Time
Mem
0.45 m
Mem
1.2 m
In-Sln

Mem
0.45 m
Mem
1.2 m
Mem
0.45 m

50 ms

Identified protein
Name
MW
(Da)
Myoglobin (horse)
16940

50 ms
5 min
1 min
20 s
2 ms

Score
***
235
382

N/A
N/A
-casein (bovine)

10 ms

N/A
N/A
23583
*

N/A
N/A
122
91
86

1.3 s Lysozyme, chain L-chicken 14500 90
0.32 s Chain A, structure of N59d 14770 137
hen egg-white lysozyme
**
* This mass does not include increased mass due to phosphorylation.
** This mass includes the increased mass due to full carbamidomethylation.
*** Score= -10Log(P), where P is the probability that the observed match is a random event.
All the scores are the highest scores in each PMF searching, but may be underestimated due to
up to 9 missed cleavages in MASCOT searching.

219

5.4 Conclusion
This chapter presents a novel strategy for the generation of large tryptic peptides for
middle-down MS analysis of proteins. Passing substrate proteins through a trypsin- modified
membrane allows tuning of proteolysis by varying the flow rate or membrane pore size.
Digestion of apomyoglobin using the modified membrane yields mainly large tryptic peptides,
whereas in-solution digestion for 30 min leads to the coexistence of equivalent amounts of small
and large ions and undigested proteins. Passing -casein though a modified 1.2 m membrane
allows a much briefer proteolysis than is available with in-solution digestion for only 20 s and
leads to a few large peptides. As expected, denaturing and alkylation of lysozyme facilitates its
tryptic digestion. Moreover, removal of ~0.1 M urea from a solution of alkylated lysozyme
greatly decreases the rate of proteolysis. Furthermore, the high sequence coverage achieved with
fast protein digestion generally leads to identification of proteins with higher scores and more
accurate assignments for parent protein from protein family members containing similar
sequences.
Control of the size of proteolytic peptides through membrane-based tryptic digestion may
be useful for high-throughput generation of large peptides for protein identification and middledown PTM characterization by MS. This approach is especially attractive when the online
middle-down digestion is coupled with post ion-ion reaction MS.

220

REFERENCES

221

REFERENCES
1.
Li, J. X.; Shefcheck, K.; Callahan, J.; Fenselau, C., Extension of Microwave-Accelerated
Residue-Specific Acid Cleavage to Proteins with Carbohydrate Side Chains and Disulfide
Linkages. Int. J. Mass Spectrom. 2008, 278, 109-113.
2.
Swatkoski, S.; Gutierrez, P.; Wynne, C.; Petrov, A.; Dinman, J. D.; Edwards, N.;
Fenselau, C., Evaluation of Microwave-Accelerated Residue-Specific Acid Cleavage for
Proteomic Applications. J. Proteome Res. 2008, 7, 579-586.
3.
Lee, T. D.; Shively, J. E., Enzymatic and Chemical Digestion of Proteins for MassSpectrometry. Methods Enzymol. 1990, 193, 361-374.
4.
Ahrens, C. H.; Brunner, E.; Qeli, E.; Basler, K.; Aebersold, R., Generating and
Navigating Proteome Maps Using Mass Spectrometry. Nat. Rev. Mol. Cell Biol. 2010, 11, 789801.
5.
Walcher, W.; Timperio, A. M.; Zolla, L.; Huber, C. G., Characterization of a Variant of
the Spinach Psii Type I Light-Harvesting Protein Using Kinetically Controlled Digestion and
Rp-Hplc-ESI-MS. Anal. Chem. 2003, 75, 6775-6780.
6.
Xu, P.; Peng, J. M., Characterization of Polyubiquitin Chain Structure by Middle-Down
Mass Spectrometry. Anal. Chem. 2008, 80, 3438-3444.
7.
Krenkova, J.; Svec, F., Less Common Applications of Monoliths: Iv. Recent
Developments in Immobilized Enzyme Reactors for Proteomics and Biotechnology. J. Sep. Sci.
2009, 32, 706-718.
8.
Huber, R.; Bode, W., Structural Basis of Activation and Action of Trypsin. Accounts
Chem. Res. 1978, 11, 114-122.
9.
Rawlings, N. D.; Barrett, A. J., Families of Serine Peptidases. In Proteolytic Enzymes:
Serine and Cysteine Peptidases, Academic Press Inc: San Diego, 1994; Vol. 244, pp 19-61.
10.
Polgar, L., The Catalytic Triad of Serine Peptidases. Cell. Mol. Life Sci. 2005, 62, 21612172.
11.
Srividhya, J.; Schnell, S., Why Substrate Depletion Has Apparent First-Order Kinetics in
Enzymatic Digestion. Comput. Biol. Chem. 2006, 30, 209-214.
12.
Perona, J. J.; Craik, C. S., Structural Basis of Substrate-Specificity in the Serine Proteases.
Protein Sci. 1995, 4, 337-360.
13.
Schechte.I; Berger, A., On Size of Active Site in Proteases .I. Papain. Biochem. Biophys.
Res. Commun. 1967, 27, 157-&.
222

14.
Bull, H. B.; Currie, B. T., Peptic Hydrolysis of Egg Albumin .1. Kinetic Studies. J. Am.
Chem. Soc. 1949, 71, 2758-2760.
15.
Hollands, T. R.; Fruton, J. S., Kinetics of Hydrolysis of Synthetic Substrates by Pepsin
and by Acetyl-Pepsin. Biochemistry 1968, 7, 2045-&.
16.
Sachdev, G. P.; Fruton, J. S., Kinetics of Action of Pepsin on Fluorescent Peptide
Substrates. Proc. Natl. Acad. Sci. U. S. A. 1975, 72, 3424-3427.
17.
Choisnard, L.; Froidevaux, R.; Nedjar-Arroume, N.; Lignot, B.; Vercaigne-Marko, D.;
Krier, F.; Dhulster, P.; Guillochon, D., Kinetic Study of the Appearance of an Anti-Bacterial
Peptide in the Course of Bovine Haemoglobin Peptic Hydrolysis. Biotechnol. Appl. Biochem.
2002, 36, 187-194.
18.
Pace, C. N.; Barrett, A. J., Kinetics of Tryptic Hydrolysis of the Arginine Valine Bond in
Folded and Unfolded Ribonuclease-T1. Biochem. J. 1984, 219, 411-417.
19.
Yi, J. Z.; Liu, Z. X.; Craft, D.; O'Mullan, P.; Ju, G.; Gelfand, C. A., Intrinsic Peptidase
Activity Causes a Sequential Multi-Step Reaction (Smsr) in Digestion of Human Plasma
Peptides. J. Proteome Res. 2008, 7, 5112-5118.
20.
Bi, H. Y.; Qiao, L.; Busnel, J. M.; Liu, B. H.; Girault, H. H., Kinetics of Proteolytic
Reactions in Nanoporous Materials. J. Proteome Res. 2009, 8, 4685-4692.
21.
Xu, F.; Wang, W. H.; Tan, Y. J.; Bruening, M. L., Facile Trypsin Immobilization in
Polymeric Membranes for Rapid, Efficient Protein Digestion. Anal. Chem. 2010, 82, 1004510051.
22.
http://www.gelifesciences.com/aptrix/upp00919.nsf/Content/elpho _ applications ~elpho
_applications _2d_protein_analysis~elpho_detection~3.+Staining+gels+wit h+Coomassie +
Brilliant+Blue (06/03/2010)
23.
Huang, T. Y.; McLuckey, S. A., Top-Down Protein Characterization Facilitated by
Ion/Ion Reactions on a Quadrupole/Time of Flight Platform. Proteomics 2010, 10, 3577-3588.
24.
Young, N. L.; Plazas-Mayorca, M. D.; Garcia, B. A., Systems-Wide Proteomic
Characterization of Combinatorial Post-Translational Modification Patterns. Expert Rev.
Proteomics 2010, 7, 79-92.
25.
Dunn, J. D.; Reid, G. E.; Bruening, M. L., Techniques for Phosphopeptide Enrichment
Prior to Analysis by Mass Spectrometry. Mass Spectrom. Rev. 2010, 29, 29-54.

223

Chapter 6. Summary and Future Directions

6.1 Research summary
This dissertation describes the development of several polymer-based functional
materials for protein isolation and digestion prior to MS analysis. The studies in chapters 2-4
aim at creating simple methods to enrich phosphopeptides from modestly complex samples such
as immunoprecipitates. Chapter 2 describes a novel method for fabricating microarrays of
functional polymer brushes surrounded by a hydrophobic thin film on Si wafers. A 1-L sample
droplet pinned on a 0.25-mm polymer spot shrinks during evaporation to concentrate peptides
within the droplet. Additionally, during sample incubation, the aminobutyl NTA-Fe(III)
2

complexes in the polymer brushes selectively capture phosphopeptides in an area (~0.05 mm )
that approaches the cross-sectional area of the MALDI laser. Elution of the phosphopeptides
with 1% H3PO4 followed by matrix addition allows sensitive MALDI-MS detection of
phosphopeptides from as little as 1 L of 0.8 nM -casein digest.
The drawbacks of NTA-Fe(III)-derivatized polymer brushes include co-enrichment of
some non-phosphopeptides and biased enrichment toward multi-phosphorylated peptides. In an
effort to overcome these drawbacks, Chapter 3 presents an oxotitanium-derivatized polymer
brush (P-oxoTi) that captures mainly mono-phosphopeptides under weakly acidic conditions (0.1%
TFA). Increasing the TFA concentration to 2% during enrichment on P-oxoTi leads to MS
signals from both mono- and multi-phosphorylated peptides, but with decreased S/N. The PoxoTi polymer brush facilitated the identification of several phosphopeptides from digests of

224

p65-associated proteins immunoprecipitated together with p65 from human acute monocytic
leukemia cell nuclear extracts. MALDI-MS/MS of the enriched phosphopeptides and database
searching elucidated phosphopeptide sequences, phosphorylation sites, and the parent
phosphoproteins.
To further improve phosphopeptide enrichment on polymer- modified wafers, we
orthogonally derivatized poly(glycidyl methacrylate) brushes with o xoTi and NTA-Fe groups
(Chapter 4).

Such bifunctionalized polymer brushes reveal both mono- and multi-

phosphorylated peptides. Thick polymer films (220 nm) with a high binding capacity enrich a
wide variety of phosphopeptides over a range of sample acidity and complexity (0.1-1%
trifluoroacetic acid, 5-20 g/mL of three phosphoproteins and < 0.5 mg/mL of nonphosphoproteins). The above three generations of polymer brushes synthesized on wafers are
attractive for MALDI-MS and -MS/MS analysis of target phosphoproteins in small volumesamples with moderate complexity, e.g., immunoprecipitates, single bands from gel
electrophoresis, and solutions from kinase assays.
Chapter 4 also presents a tandem enrichment approach that captures different subsets of
phosphopeptides sequentially in two nylon membranes derivatized with P-oxoTi and P-NTAFe(III), respectively. Compared to the bifunctional polymer brushes on wafers, this method
provides a higher binding capacity that allows enrichment of phosphopeptides from more
complex samples. The functionalized membranes could in principle facilitate phosphopeptide
sequencing when coupled to HPLC-ESI-MS/MS.
Although enrichment of phosphopeptides from tryptic digests enhances the intensity of
phosphopeptide signals in MS, complete tryptic digestion prior to enrichment often decreases
phosphopeptide signals due to peptide loss in HPLC or inefficient ionization of the

225

phosphorylated species. Moreover, short tryptic peptides do not allow determination of the
“crosstalk” (or correlation) of PTM sites along the entire protein chain. In an effort to overcome
such issues, Chapter 5 presents a proteolysis strategy that employs brief (several ms) digestion in
a trypsin- modified nylon membrane to obtain large proteolytic peptides. For example, limiting
the protein residence time in a modified membrane (1.2 m pores) to 10 ms yields 100%
sequence coverage of -casein (24 kDa, 15 potential trypsin cleavage sites) with as few as 2
peptides. Furthermore, one of the resulting large peptides (19 kDa) covers all the five
phosphorylation sites of -casein. This size-controlled proteolysis enhances the sequence
coverage of target proteins and provides larger PTM coverage on a single peptide than traditional
bottom-up methods.

6.2 Future work
6.2.1 Improving on-plate phosphopeptide enrichment by decreasing sample
complexity
Much of the work presented in this dissertation (Chapters 2-4) focuses on phosphopeptide
enrichment on wafers modified with functional polymer brushes. These materials are specifically
designed to facilitate MALDI-MS analysis of low- volume protein samples. Although sensitive
and specific for detecting phosphopeptides from known model protein digests with moderate
complexity and acidity, this technique still shows limited successes in identifying unknown
phosphorylated proteins.[1] Chapter 3 shows that even after immunoprecipitation, binding of the
target phosphoprotein p65 with many other proteins leads to identification of phosphopeptides
from p65-associated proteins instead of p65 itself. The excess of peptides from the co226

immunoprecipitated proteins might suppress the signals of p65 phosphopeptides and decrease the
efficiency of phosphopeptide enrichment. Thus, further protein purification such as SDS-PAGE
or ion-exchange prior to protein digestion is needed to improve phosphopeptide enrichment.
Moreover, the tryptic digestion of p65 and its associated proteins directly on the antibody beads
is probably not efficient. The high concentration of urea used for protein denaturation pr ior to
digestion might also affect phosphopeptide binding during sample incubation on the Au-P-oxoTi
wafer. Using degradable surfactants[2] for protein elution from the antibody beads followed by
multiple enzymatic digestions (trypsin, Glu-C, Lys-C, etc.) and phosphopeptide enrichment may
give enhanced enrichment.

6.2.2 Large-scale phosphopeptide enrichment using tandem membranes
Chapter 4 presents the possibility of capturing multiple subsets of phosphopeptides using
stacked membranes with different functionalities.

Because polymer brushes grown in

membranes provide higher binding capacities than films on wafers,[3] functionalized membranes
could enrich phosphopeptides from larger volume samples, perhaps with greater complexity.
Before applying the membranes to any complex protein digest, membrane binding capacities
should be evaluated using a group of synthetic mono- and multi-phosphorylated peptides.

6.2.3 Quantitative evaluation of phosphopeptide recovery
In Chapter 4, comparisons of enrichment of mono- and multi-phosphopeptides rely on
MS peak patterns and S/N. Because the ionization efficiencies for mono- and multiphosphorylated peptides are likely very different, a more careful recovery study on d ifferent
subsets of phosphopeptides will require isotope- labeled phosphopeptides for MS quantification
(see Chapter 3 for an example).

Although previous studies described several optimization
227

strategies that enhance phosphopeptide enrichment efficiency in TiO 2 or IMAC columns,[4-8]
none of them used quantitative MS methods for recovery evaluation. The consensus explanation
for biased enrichment of mono-phosphorylated peptides on TiO 2 is strong binding of multiphosphorylated peptides but inefficient elution from the TiO 2 surface.[9] However, the tandem
phosphopeptide enrichment using stacked membranes suggests that multi-phosphopeptides may
not bind well to oxoTi.
Verification of this assumption requires quantitative MS measurements.

More

specifically, these studies should include enrichment of an isotopically labeled protein digest
(with known phosphoproteins) using TiO 2 columns or polymer-oxoTi and collection of the flowthrough, washing, and elution fractions. Mixing each fraction with the same amount of starting
digests with differentially labeled isotopes followed by a second phosphopeptide enrichment
should allow quantitative measurement of different phosphopeptides from each fraction.
Because the recovery of multi-phosphorylated peptides from the Ti-based materials is low,
quantification of multi-phosphorylated peptides using the second-time enrichment will require an
Fe(III)-based affinity material. To avoid the effect of isotopic tags on phosphopeptide binding
and elution, isotope labeling of the flow-through, washing, and eluate fractions performed after
the first enrichment may give better quantification results.

6.2.4 Tuning flowrate, membrane pore size, trypsin local “concentration”, and
substrate protein concentration to control membrane-based digestion
Chapter 5 shows that decreasing digestion times for both in-solution digestion and
membrane-based digestion increases the size of proteolytic peptides. However, apomyoglobin

228

digestion gives rise to very different proteolytic peptides in brief membrane digestion than brief
in-solution digestion. The membrane provides an excess of trypsin that may allow simultaneous,
limited protein cleavage for most apomyoglobin molecules. However, the early apomyoglobin
cleavage in solution is a slow process because there is much less trypsin. In this case, during the
time needed to achieve significant depletion of intact apomyoglobin, the initially formed
peptides degrade further to give a mixture of large peptides, undigested protein, and small
peptides.

Examination of a wider class of standard proteins that digest slowly in solution

(reflected by the existence of a large portion of undigested protein after several minutes of insolution digestion) should provide further support for differences between solution and
membrane-based digestion. The membrane-based digestion should occur at a series of flowrates
because different proteins will present different resistances toward tryptic digestion.

If

hydrolysis of such proteins in trypsin- modified membranes gives more efficient depletion of
intact protein but larger peptides than in-solution digestion, we can conclude that membrane
tryptic digestion provides a different hydrolysis pathway for digestion-resistant proteins.
For more rapidly digested proteins like -casein, short in-solution digestion (<1 min)
already degrades almost all of the protein. Chapter 5 shows that a 20-s, in-solution digestion
depletes most of the -casein and produces large peptides similar to those from a 2- ms
membrane digestion (0.45 m pore size). This suggests a similar digestion pathway for both
methods. Using modified membranes with 1.2 m pores gives larger tryptic -casein peptides,
but a significant portion of the -casein is undigested. Although the intact protein may assist in
protein identification by peptide mass fingerprint database searching, converting a majority of
the proteins to the peptide form is necessary for middle-down MS analysis, especially PTM
characterization via tandem MS. Future work o n such “easily digested” proteins needs fine
229

tuning of membrane parameters to both degrade intact protein and provide peptides with sizes
larger than those produced by the 20-s in-solution digestion. There are several possible ways to
achieve this aim: (1) Minimizing the protein residence in the 0.45 m pores to sub- millisecond
times with a stronger syringe pump. (2) Shortening the trypsin modification time so that less
trypsin will adsorb in the 0.45 m pores after depositing poly(sodium 4-styrenesulfonate) (PSS)
in the membrane. However, this method may lead to binding of positively charged substrate
proteins to the exposed PSS in the membrane pores. (3) Depositing less PSS in the 0.45-m
membrane pores to immobilize less trypsin. To achieve efficient protein depletion, diffusion
distances <1 m promise encounters of each substrate protein with immobilized trypsin.
However, decreasing the local trypsin “concentration” (it should still be close to the substrate
protein concentration) could further minimize proteolysis of the resulting large peptides.
The substrate protein concentration also affects the trypsin-substrate protein ratio. When
the substrate protein concentration is too high, the available trypsin in the membrane pores may
be insufficient for simultaneous digestion of all the substrate molecules. Varying the substrate
concentration over a large range and examining the resulting peptide patterns in mass spectra
may give clues to the digestion pathway.

6.2.5 Immobilization of pepsin in nylon membranes for protein proteolysis in
acidic conditions
Pepsin is another inexpensive protease, but this enzyme is rarely used for traditional
bottom- up protein analysis because its low specificity in protein cleavage leads to short peptides.
Possible protein hydrolysis sites with pepsin include hydrophobic and preferably aromatic amino
acids such as phenylalanine, tryptophan, and tyrosine.[10]

230

Nevertheless, pepsin is also

advantageous for protein digestion in some cases because of its acidic working condition. Acidic
proteolysis with pepsin facilitates digestion of proteins that are insoluble in neutral buffers.[11]
As an example, membrane proteins often require surfactant for solublization prior to tryptic
digestion, but the surfactant complicates downstream peptide separation and MS analysis.[12]
These proteins may, however, dissolve in acidic conditions. Pepsin is also widely used in studies
of protein confirmation by hydrogen/deuterium exchange, where the proteolysis of target
proteins requires an acidic condition to inhibit back exchange of deuterium with hydrogen.[13]
More importantly, fast proteolysis also favors the inhibition of deuterium back exchange.[14]
Previous studies reported immobilization of pepsin for fast proteolysis.[11,14]
Membranes provide a platform for achieving ms residence times, so digestion with pepsinmodified membrane could generate larger peptides than in-solution pepsin digestion.

The

limited digestion time with pepsin may lead to more specific cleavage than longer digestion with
free pepsin in solution. For example, cleavage may occur only at aromatic residues instead of
other hydrophobic residues. Narrowing the digestion specificity for database searching can
improve protein identification.[15] Because of its compatibility with acidic buffers, digestion
with immobilized pepsin may be attractive for integration of protein analysis procedures in an
online system, which could include protein separation, pepsin digestion, peptide separation or
enrichment, and finally ESI-MS.

6.3 Potential impact
The results in this dissertation lay the ground work for potential new devic es that isolate
small phophopeptides or produce large peptides for bottom- up or middle-down MS analysis of
phosphoproteins. The metal-affinity modified polymer brushes grown on wafers greatly simplify
231

phosphopeptide purification processes and facilitate MALDI-MS analysis.

These modified

wafers have potential utility in analysis of target phosphoproteins isolated from complex samples
by up-stream purifications, such as immuno-affinity fractionation and gel electrophoresis.
Extending the growth of the functional polymer brushes to membranes give a higher binding
capacity that allows larger scale phosphopeptide isolation from more complex digests. The
phosphopeptide enrichment platforms may improve the bottom-up

MS analysis of

phosphoproteins. As a complement to small peptide analysis, middle-down analysis requires
generation of large peptides for MS characterization. Immobilization of proteases in membranes
makes possible a kinetic-controlled proteolysis.

Millisecond digestion times produce large

peptides with higher coverages of protein sequences and PTMs than with in-solution digestion.
Further systematic studies with a large group of proteins may give direction for generating
peptides with desired sizes by controlling the flowrate of the substrate protein through the
membrane. Furthermore, the short-time digestion in membranes could be coupled with proteinor peptide-level separation platforms and MS for online protein analysis.

232

REFERENCES

233

REFERENCES
1.
Hoang, T.; Roth, U.; Kowalewski, K.; Belisle, C.; Steinert, K.; Karas, M., Highly
Specific Capture and Direct MALDI MS Analysis of Phosphopeptides by Zirconium
Phosphonate on Self-Assembled Monolayers. Anal. Chem. 2010, 82, 219-228.
2.
Proteasemax™
Surfactant,
Trypsin
Enhancer.
http://www.promega.com/catalog/catalogproducts.aspx?categoryname=productleaf_2163
(03/04/2011)
3.
Jain, P.; Baker, G. L.; Bruening, M. L., Applications of Polymer Brushes in Protein
Analysis and Purification. Annu. Rev. Anal. Chem. 2009, 2, 387-408.
4.
Aryal, U. K.; Ross, A. R. S., Enrichment and Analysis of Phosphopeptides under
Different Experimental Conditions Using Titanium Dioxide Affinity C hromatography and Mass
Spectrometry. Rapid Commun. Mass Spectrom. 2010, 24, 219-231.
5.
Li, Q. R.; Ning, Z. B.; Tang, J. S.; Nie, S.; Zeng, R., Effect of Peptide-to-Tio2 Beads
Ratio on Phosphopeptide Enrichment Selectivity. J. Proteome Res. 2009, 8, 5375-5381.
6.
Simon, E. S.; Young, M.; Chan, A.; Bao, Z. Q.; Andrews, P. C., Improved Enrichment
Strategies for Phosphorylated Peptides on Titanium Dioxide Using Methyl Esterification and Ph
Gradient Elution. Anal. Biochem. 2008, 377, 234-242.
7.
Tsai, C. F.; Wang, Y. T.; Chen, Y. R.; Lai, C. Y.; Lin, P. Y.; Pan, K. T.; Chen, J. Y.;
Khoo, K. H.; Chen, Y. J., Immobilized Metal Affinity Chromatography Revisited: Ph/Acid
Control toward High Selectivity in Phosphoproteomics. J. Proteome Res. 2008, 7, 4058-4069.
8.
Ye, J. Y.; Zhang, X. M.; Young, C.; Zhao, X. L.; Hao, Q.; Cheng, L.; Jensen, O. N.,
Optimized IMAC-IMAC Protocol for Phosphopeptide Recovery from Complex Biological
Samples. J. Proteome Res. 2010, 9, 3561-3573.
9.
Thingholm, T. E.; Jensen, O. N.; Robinson, P. J.; Larsen, M. R., Simac (Sequential
Elution from IMAC), a Phosphoproteomics Strategy for the Rapid Separation of
Monophosphorylated from Multiply Phosphorylated Peptides. Mol. Cell. Proteomics 2008, 7,
661-671.
10.
Hamuro, Y.; Coales, S. J.; Molnar, K. S.; Tuske, S. J.; Morrow, J. A., Specificity of
Immobilized Porcine Pepsin in H/D Exchange Compatible Conditions. Rapid Commun. Mass
Spectrom. 2008, 22, 1041-1046.
11.
Ma, J. F.; Hou, C. Y.; Sun, L. L.; Tao, D. Y.; Zhang, Y. Y.; Shan, Y. C.; Liang, Z.; Zhang,
L. H.; Yang, L.; Zhang, Y. K., Coupling Formic Acid Assisted Solubilization and Online
Immobilized Pepsin Digestion with Strong Cation Exchange and Microflow Reversed-Phase
Liquid Chromatography with Electrospray Ionization Tandem Mass Spectrome try for Integral
Membrane Proteome Analysis. Anal. Chem. 2010, 82, 9622-9625.
234

12.
Lin, Y.; Liu, Y.; Li, J. J.; Zhao, Y.; He, Q. Z.; Han, W. J.; Chen, P.; Wang, X. C.; Liang,
S. P., Evaluation and Optimization of Removal of an Acid-Insoluble Surfactant for Shotgun
Analysis of Membrane Proteome. Electrophoresis 2010, 31, 2705-2713.
13.
Nazabal, A.; Hornemann, S.; Aguzzi, A.; Zenobi, R., Hydrogen/Deuterium Exchange
Mass Spectrometry Identifies Two Highly Protected Regions in Recombinant Full- Length Prion
Protein Amyloid Fibrils. J. Mass Spectrom. 2009, 44, 965-977.
14.
Liuni, P.; Rob, T.; Wilson, D. J., A Microfluidic Reactor for Rapid, Low-Pressure
Proteolysis with on-Chip Electrospray Ionization. Rapid Commun. Mass Spectrom. 2010, 24,
315-320.
15.
Ahrne, E.; Muller, M.; Lisacek, F., Unrestricted Identification of Modified Proteins Using
MS/MS. Proteomics 2010, 10, 671-686.

235