.45.... L. «m... .. x .
fikfififi?

.hw...

h.
‘(cu‘

6.5:».

.$§?
wafinjvuhhwflfim
nu, flaw... hf.
vnfimamfig
....

0

53.
11.51., a. ..
43.32.. V .3“:
. h ltui.
fit) .. till. 3|! .5
7.1.1.311... Sinnflw.

,3 a. .2:
mm. 2

7
"-'\.-I ‘I
...... is!!!
1...! ........!... ,

‘5 :I'I! \§:.!’X“
é. 3(-

.. l.

.l
3 VXtfiixfit‘

D-I\1tmmnslz

ltl. 5.
Sibel} Vt. :Iggt"
.ixl cl.-v..l 9:01le :11. ‘01.“.
11.15:!) D.%I..cc.ll.. I...
I»...

. I .3331! I!!.Ca.|)!:l.
a ‘$
1'... \ulzi:
3x
vlllo~\I.tln

.. a“;

. t..|q||.|‘v.
LII...!11.430 I!
. ; 1:

 

 

1mmm ‘IJBRARY

.‘Q' 3 Mlfis "9“” State
W ‘ mmmmw'

 

 

 

This is to certify that the
thesis entitled

FUNCTIONALIZED SURFACES FOR SELECTIVE CAPTURE
OF BIOMOLECULES

presented by

Elizabeth A. lgrisan

has been accepted towards fulfillment
of the requirements for the

MS. degree in Chemistry

 

 

”flux; Km

Major Erofessor’sggpature
Ogjallzoaq
Date

MSU is an Affirmative ActiorVEquaI Opportunity Employer

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K2/Proj/Aoc8PresICIRC/Dateomindd

 

-%

FUNCTIONALIZED SURFACES FOR SELECTIVE CAPTURE OF
BIOMOLECULES

By

Elizabeth A. lgrisan

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Chemistry

2009

ABSTRACT

FUNCTIONALIZED SURFACES FOR SELECTIVE CAPTURE OF
BIOMOLECULES

By
Elizabeth A. lgrisan

This thesis describes the synthesis of several functional surfaces for isolation of
phosphopeptides, glutathione—S-transferase (GST), and glycopeptides. Gold surfaces
modified with Zr02/p01y(styrene sulfonate) films selectively capture phosphorylated
peptides from an unpurified protein digest. Substrates prepared by heating an array of
Ti02 nanoparticles also enrich phosphopeptides, and these plates selectively recover
~70% of a synthetic H5 phosphopeptide (125 fmol) in the presence of 1 pmol of
nonphosphopeptide mixture. On-plate capture is attractive for high recoveries prior to
analyses by matrix-assisted laser desorption/ionization mass spectrometry (MALDI-MS).

In efforts to create surfaces that bind GST-tagged proteins, gold substrates
modified with poly(acrylic acid) or poly(2-hydroxyethyl methacrylate) (PHEMA)
brushes were derivatized with reduced glutathione in a variety of methods. These
polymer brushes should have much higher binding capacities than monolayer films.
Unfortunately the immobilized glutathione did not bind GST, perhaps because of residual
glutathione in the sample. PHEMA brushes were also functionalized with 3-
aminophenylboronic acid (APBA) to create surfaces that capture glycopeptides. The
APBA-modified films bind small amounts of simple cis-diol-containing carbohydrates,
and initial studies show that glycopeptide enrichment on these polymer brushes improves
the MALDI-MS signal-to-noise ratio for some glycopeptides present in a digest of

horseradish peroxidase. However, signals are weak and nonspecific binding occurs.

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................................. vi
LIST OF FIGURES ........................................................................................................... vii
LIST OF ABBREVIATIONS ........................................................................................... xiii
Chapter One: Introduction ............................................................................................... 1
1.1 Outline ......................................................................................................................... l
1.2 Surface Modification.......................................................' ........................................... 2
1.2.1 Polymer Grafting ..................................................................................................... 2
1.2.2 Polyelectrolyte Multilayers ...................................................................................... 4
1.3 Post—Translational Modifications of Proteins ............................................................. 5
1.3.1 Protein Phosphorylation ........................................................................................... 5
1.3.2 Protein Glycosylation ............................................................................................... 8
1.4 Mass Spectrometry for Characterization of Phosphoproteins/Phosphopeptides

and G1ycoproteins/Glycopeptides ............................................................................. 1 1
1.4.1 Ionization Sources .................................................................................................. 12
1.4.1.1 Matrix-Assisted Laser Desorption/Ionization ..................................................... 13
1.4.1.2 Electrospray Ionization ....................................................................................... 15
1.4.2 Mass Analyzers Compatible with MALDI ............................................................ 16
1.4.2.1 Time-of-Flight Mass Analyzer ........................................................................... 16
1.4.2.2 Linear Ion Trap Mass Analyzer .......................................................................... 17
1.4.3 Challenges in Phosphopeptide and Glycopeptide Detection by Mass

Spectrometry .......................................................................................................... 21

1.5 Enrichment Techniques for Phosphopeptides and Glycopeptides Prior to

MS Analysis .............................................................................................................. 22
1.5.1 Enrichment Methods for Phosphorylated Peptides Prior to MS Analysis ............. 22
1.5.1.1 Immobilized Metal Affinity Chromatography .................................................... 23
1.5.1.2 Metal Oxide Affinity Chromatography .............................................................. 25
1.5.2 Enrichment Methods for Glycoproteins/Glycopeptides Prior to MS Analysis ..... 29
1.5.2.1 Lectin Affinity Chromatography ........................................................................ 30
1.5.2.2 Covalent Binding Enrichment Methods .............................................................. 31
1.6 Research Overview ................................................................................................... 37
1 .7 References ................................................................................................................. 39
Chapter Two: Metal Oxide-Modified Plates for Analysis of Phosphopeptides
by MALDI-MS ................................................................................................................. 46
2. 1 Introduction ............................................................................................................... 46
2.2 Experimental ............................................................................................................. 48
2.2.1 Materials ................................................................................................................ 48
2.2.2 Protein Digestion ................................................................................................... 50
2.2.3 Labeled Phosphopeptide Synthesis ........................................................................ 51
2.2.4 Fabrication of Polyelectrolyte/Metal Oxide Films ................................................ 52
2.2.5 Preparation of TiOz-modified MALDI Plate ......................................................... 55

iii

2.2.6 Enrichment Protocol for Plates Modified with Polyelectrolyte/Metal Oxide

Films ...................................................................................................................... 56
2.2.7 Protocol for Enrichment of H5 Peptide Using Plates Modified with TiOz

by Heating .............................................................................................................. 57
2.2.8 Characterization and Instrumentation .................................................................... 58
2.3 Results and Discussion ............................................................................................. 59
2.3.1 Fabrication and Characterization of Polyelectrolyte/Metal Oxide-Modified

Plates ...................................................................................................................... 59
2.3.2 Analysis of Ovalbumin Digests Using MALDI-MS ............................................. 61
2.3.3 Calibration of H5 and D5 Signals for Recovery Analysis ...................................... 66
2.3.4 Enrichment of H5 Peptide from Peptide Mixtures Using TiOz-modified Plate ..... 67
2.4 Conclusions ............................................................................................................... 7 1
2.5 References ................................................................................................................. 72
Chapter Three: Binding of Glutathione-S-Transferase to Glutathione-
F unctionalized Polymer Brushes for Protein Enrichment ........................................... 74
3. 1 Introduction ............................................................................................................... 74
3.2 Experimental ............................................................................................................. 76
3.2.1 Reagents and Materials .......................................................................................... 76
3.2.2 Fabrication of Glutathione-Functionalized PAA Films ......................................... 77
3.2.3 Fabrication of Glutathione—Functionalized PHEMA Brushes Using Succinic

Anhydride and NHS/EDC ...................................................................................... 81
3.2.4 Binding of Glutathione to Brominated-PHEMA Brushes ..................................... 84
3.2.5 Binding of Glutathione to Maleimide-Derivatized PHEMA Brushes ................... 85
3.2.6 Dialysis of Glutathione-S-Transferase ................................................................... 86
3.2.7 Purification of Glutathione-S-Transferase Using ZipTipmg Pipette Tips .............. 88
3.2.8 Characterization and Instrumentation .................................................................... 88
3.3 Results and Discussion ............................................................................................. 89
3.3.1 Fabrication and Characterization of Glutathione-Functionalized PAA Films ....... 89
3.3.2 Fabrication and Characterization of Glutathione-Functionalized PHEMA

Brushes ................................................................................................................... 91
3.3.3 Immobilization of Glutathione on Brominated-PHEMA Brushes ......................... 92
3.3.4 Fabrication and Characterization of Glutathione Immobilized on Maleimide-

PI-IEMA Brushes .................................................................................................... 94
3.3.5 Attempts to Purify Glutathione-S-Transferase to Increase Binding ...................... 97
3.3.6 Binding Purified GST to GSH-Maleimide-PHEMA Brushes ............................. 100
3.4 Conclusions ............................................................................................................. 101
3.5 References ............................................................................................................... 103
Chapter Four: Enrichment of Glycopeptides Using Aminophenylboronic Acid-
Derivatized Polymer Brushes ........................................................................................ 105
4. 1 Introduction ............................................................................................................. 105
4.2 Experimental ........................................................................................................... 108
4.2.1 Materials .............................................................................................................. 108
4.2.2 Protein Digestion ................................................................................................. 109
4.2.3 Fabrication of APBA-PHEMA Brushes .............................................................. 110

iv

4.2.4 Binding of Fructose to APBA-PHEMA Brushes ................................................. 111

4.2.5 Protocol for Enrichment of Glycopeptides Using APBA-PHEMA Brushes ....... 112
4.2.6 Characterization and Instrumentation .................................................................. 112
4.3 Results and Discussion ........................................................................................... 115
4.3.1 Characterization of APBA-PHEMA Brushes ...................................................... 115
4.3.2 Characterization of Fructose Bound to APBA-PHEMA Brushes ....................... 117
4.3.3 Analysis of Horseradish Peroxidase Digests Using MALDI-MS ........................ 119
4.4 Conclusions ............................................................................................................. 126
4.5 References ............................................................................................................... 127
Chapter Five: Conclusions and Future Work ............................................................. 129

LIST OF TABLES

Chapter Two: Metal Oxide-Modified Plates for Analysis of Phosphopeptides by
MALDI-MS

Table 2.1: Comparison of the ability of several commercially available enrichment
materials and the TiOz-modified plates to recover 125 fmol of H5 peptide from 1 pmol of
nonphosphorylated digest mixture ..................................................................................... 70

Chapter Four: Enrichment of Glycopeptides Using Aminophenylboronic Acid-
Derivatized Polymer Brushes

Table 4.1: N-linked glycopeptides present in a tryptic HRP digest. The oligosaccharide
structure of each glycopepdide is shown, along with its [M+H]+ m/z value. N# represents
the N-linked glycosylation site on asparagines residues. Hex stands for hexose, HexNAc
for N-acetylhexosamine, dHex for deoxyhexose, and Pent for pentose .......................... 121

vi

LIST OF FIGURES

Chapter One: Introduction

Figure 1.1: Schematic diagram of the enrichment process where a protein digest is
spotted on a functionalized surface. Nonmodified peptides and digest reagents are rinsed
away. Bound modified peptides are eluted and crystallized upon addition of matrix. The
analyte can be analyzed directly on the plate using MALDI-MS ........................................ 2

Figure 1.2: Schematic diagram comparing “grafting to” versus “grafting from” methods
for growing films on surfaces .............................................................................................. 3

Figure 1.3: Scheme of the mechanism for transition metal-catalyzed ATRP, where km,
kdcact, kp, and kt represent the rate constants for activation, deactivation, polymerization,
and termination, respectively ............................................................................................... 4

Figure 1.4: Schematic diagram of a negatively charged substrate before (a) and after
adsorption of a polycation and rinsing with H20 (b), and after subsequent adsorption of a
polyanion and a second rinsing with H20 (0). Repetition of the process yields multilayer
films ..................................................................................................................................... 5

Figure 1.5: Kinase-catalyzed phosphorylation of serine by ATP ........................................ 6

Figure 1.6: Cartoon showing reversible protein phosphorylation and the conformational
change of the protein upon addition of the phosphate group ............................................... 7

Figure 1.7: Schematic diagram showing the synthesis of the core oligosaccharide of a
glycoprotein, where the oligosaccharide moiety is built up by the successive addition of
monosaccharide units, (a) and (b), which occurs on the cyctosolic face of the ER. The
oligosaccharide is translocated across the ER membrane and into the lumen (c) and
additional monosaccharide units can be added while inside the lumen (not shown here).
The core oligosaccharide is transferred from the dolichol phosphate to an Asn residue of
the protein ((1), forming an N-linked glycoprotein (e). The fully translated protein is
released from the mRNA (f) and the dolichol phosphate is translocated (g) and a
phosphate is removed (h). The glycoprotein is further modified within the ER and in the

Golgi complex ...................................................................................................................... 9
Figure 1.8: Structures of a) N-acetylgalactosamine 0-linked to the hydroxyl group of
serine and b) N-acetylglucosamine N—linked to the amide nitrogen of asparagines ........... 10
Figure 1.9: Schematic diagram of a mass spectrometer .................................................... 12

Figure 1.10: Schematic diagram of the ionization of a MALDI sample, where the organic
matrix is present in a much higher quantity than the analyte. After co-crystallization of
the biomolecule and matrix, a pulsed laser is fired at the sample and the matrix and
analer are both desorbed and ionized ............................................................................... 15

vii

Figure 1.11: Schematic diagram of the Theme vMALDI LTQ XL instrument with a
linear ion trap ..................................................................................................................... 18

Figure 1.12: Schematic diagram of a Therrno linear ion trap. Ions are trapped in the
center section of the 2D LIT, and the front and back sections are necessary to minimize
distortion of the electric field of the center section. This improves trapping efficiency.
Ions are radially ejected through two slots (30 x 0.25 mm) in the exit rods in the center
section ................................................................................................................................ 19

Figure 1.13: In the LIT, ions have stable trajectories when q is less than 0.908. Circles
represent ions in the stable region and the size of the circles corresponds to the relative
size of m/z values ............................................................................................................... 20

Figure 1.14: Schematic diagram showing the general procedure for enrichment of
phosphopeptides using IMAC or MOAC on a column. The column packing is different
for each method, but both techniques involve loading the sample onto a column, rinsing
to remove any contaminants, and elution of enriched phosphopeptides. The analyte can
then be analyzed with MS .................................................................................................. 23

Figure 1.15: Structures of a) iminodiacetate (IDA), and b) nitrilotriacetate (NTA) .......... 24

Figure 1.16: Schematic diagram comparing the chelating bidentate structure of salicylic
acid on TiOz with the bridging bidentate complex formed between phosphate and
titanium dioxide ................................................................................................................. 27

Figure 1.17: Schematic diagram showing the immobilization of a thin film of N-[3-
(trimethoxysilyl)propyl]ethylenediamine (TMSPED) on glass, attachment of gold
nanoparticles (NPs) to the film, and coating of the nanoparticles with TiOz. In the TiOz
coating step, a solution of titanium isopropoxide is spin-coated on the surface, followed
by annealing to give TiOz nanoparticles ............................................................................ 28

Figure 1.18: Schematic diagram showing the oxidation of cis-diol-containing
glycoprotein with sodium periodate, coupling with hydrazide-functionalized beads, and
then tryptic digestion of the protein, yielding glycopeptides covalently bound to the
magnetic particles. Glycopeptides bound to the resin are isotopically labeled and cleaved
using PNGase F, resulting in formerly N-linked isotopically labeled peptides, followed by
analysis using either ESI or MALDI ................................................................................. 33

Figure 1.19: Relevant equilibria for the formation of a boronate ester from phenylboronic
acid and a cis-diol. The boronate ester may exist in both trigonal and tetrahedral forms.
Km-g and Km are the equilibrium constants for the formation of the trigonal and tetrahedral
forms of phenylboronic acid, respectively ......................................................................... 34

viii

Figure 1 .20: Schematic diagram of a) reaction between (3-
glycidyloxypropyl)trimethoxysilane (GLYMO) and 3-aminophenylboronic acid (APBA)
to form GLYMO-APBA, followed by attachment to mesoporous silica FDU-l2; and b)
grafting of GLYMO-APBA to mesoporous silica, then enrichment of glycopeptides
within the silica pore .......................................................................................................... 37

Chapter Two: Metal Oxide-Modified Plates for Analysis of Phosphopeptides by
MALDI-MS

Figure 2.1: Schematic diagram showing the modification of a gold substrate with a) a
SAM of MPA, followed by the deposition of b) the polycation PAH, c) the polyanion
PSS, and finally (1) the nanoparticles, either Ti02 or Zr02. Additional bilayers can be
formed by alternating steps c) and d). Inset shows the structures of the polyelectrolytes
used .................................................................................................................................... 54

Figure 2.2: Photograph of a) Ti02-modified magnetic MALDI plate with machined wells,
and b) the MALDI sample plate holder, which contains magnets ..................................... 56

Figure 2.3: Schematic of the enrichment process used to determine the recovery of the H5
peptide from a protein digest mixture using a T102 modified plate. The D5 peptide served
as an internal standard ........................................................................................................ 58

Figure 2.4: Reflectance FTIR spectra of a) a self-assembled monolayer of MPA on a gold
substrate, and the same film after b) deposition of the polycation, PAH, and finally c)
deposition of the polyanion, PSS ....................................................................................... 60

Figure 2.5: SEM images of a) Ti02 adsorbed on an MPA-PAH-PSS modified gold
substrate and b) a magnified image of the same film. Ti02 was adsorbed from a 1 mg/mL
suspension .......................................................................................................................... 61

Figure 2.6: Positive ion MALDI mass spectra of 2 pmol of ovalbumin digest analyzed
using a) conventional MALDI-MS, and b) a ZrO2-PSS-PAH-MPA-modified gold plate
with 0.1% TFA loading solution, rinsing with 66 mg/mL DHB (80% ACN/O.1% TFA)
solution and 80% ACN/0.1% TFA solution, and addition of matrix prior to MALDI-MS.
Asterisks (*) represent phosphorylated peptides ............................................................... 63

Figure 2.7: Amino acid sequence for chicken egg ovalbumin (Swiss-Pro: P01012).
Phosphorylation sites are designated by bold and italic type and a (p) label. Tryptic
cleavage sites are labeled in bold ....................................................................................... 64

Figure 2.8: Multistage tandem mass spectrometry of a monophosphorylated peptide ion,
m/z 2089, isolated from 2 pmol of ovalbumin digest enriched using a Zr02-PSS-PAH-
MFA-modified gold plate: a) CID MS/MS of the m/z 2089 isolated [M+H]+ precursor
ion, and b) CID MS3 of m/z 1973 isolated from the MS/MS product ion spectrum of m/z
2089 from spectrum a). In b) several yn type ions were identified .................................... 65

ix

Figure 2.9: Sequence and m/z values of labeled synthetic phosphorylated peptides, a) H5
peptide and b) D5 peptide ................................................................................................... 66

Figure 2.10: Calibration curve comparing the ratio of peak intensities (Ins/IDS) of
synthetic phosphopeptides, H5 and D5, from MALDI-MS spectra as a function of the
amount of H5 peptide. 125 fmol of D5 peptide was present in each sample as an internal
standard .............................................................................................................................. 67

Figure 2.11: a) Positive ion MALDI mass spectrum of 125 fmol of H5 peptide enriched
from 1 pmol of a protein digest using a Ti02-modified plate, and b) an enlarged region of
the spectrum around peaks due to H5 and D5 peptides. 125 fmol of D5 was added as an
internal standard. The recovery of H5 peptide from the mixture is 69 1 7%. The
modified plate was prepared by heating Ti02 nanoparticles on a magnetic plate ............. 69

Chapter Three: Binding of GIutathione-S-Transferase to Glutathione-
Functionalized Polymer Brushes for Protein Enrichment

Figure 3.1: Schematic diagram showing the general procedure for separation of GST-
tagged proteins using a column. The sample is loaded onto a column packed with
glutathione-modified beads, rinsed to remove any non-GST-tagged proteins and
contaminants, and finally the GST-tagged proteins bound to the resin are eluted and
collected for further analysis .............................................................................................. 75

Figure 3.2: Structure of reduced L-glutathione, which is composed of glycine, cysteine
and y-glutamic acid residues .............................................................................................. 76

Figure 3.3: Schematic diagram showing a gold-coated substrate after a) formation of an
MUA SAM, b) activation of the MUA SAM, c) attachment of PTBA, d) hydrolysis of
PTBA to give immobilized PAA, e) activation of PAA with NHS/EDC, and f)
immobilization of glutathione. In some cases, the PAA of step d was activated with
ethyl chloroformate, and steps c, d, and e were repeated prior to immobilization of
glutathione .......................................................................................................................... 80

Figure 3.4: Schematic diagram of a gold-coated substrate after a) formation of an MUD
SAM, b) initiator attachment to the MUD SAM, c) polymerization of HEMA, d)
derivatization of PHEMA with SA, e) activation with NHS, and f) immobilization of
glutathione onto PHEMA-SA films ................................................................................... 83

Figure 3.5: Schematic diagram the modification of a gold-coated substrate with a)
PHEMA, b) brominated-PHEMA, c) glutathione immobilized on the PHEMA film ....... 85

Figure 3.6: Schematic diagram of the modification of a gold-coated substrate after a)
formation of a PHEMA-SA film activated with NHS/EDC, b) attachment of maleimide,
and c) immobilization of glutathione on the PHEMA film... ........................................... 86

Figure 3.7: Reflectance FTIR spectra of a gold-coated Si wafer after a) deposition of a
PAA film, b) activation of the PAA with NHS, c) reaction of glutathione with the
activated PAA bilayer, and (1) exposure of the glutathione-containing film to GST. The
initial PAA film was deposited using two activation, PTBA deposition, and hydrolysis
steps .................................................................................................................................... 90

Figure 3.8: Reflectance FTIR spectra of a) a PHEMA brush immobilized on a gold-
coated substrate, b) functionalization of PHEMA with succinic anhydride, c) activation of
PHEMA-SA with NHS, (1) immobilization of glutathione onto PHEMA-SA, and e)
binding of GST to glutathione-immobilized polymer brushes .......................................... 92

Figure 3.9: Reflectance FTIR of a) PHEMA immobilized on a gold-substrate, b) PHEMA
brominated with bromoacetyl chloride, and c) immobilization of glutathione on PHEMA
film ..................................................................................................................................... 93

Figure 3.10: Reflectance FTIR spectra of PHEMA-SA a) before and after b)
derivatization of with N-(2-aminoethyl)maleimide trifluoroacetate, c) immobilization of
reduced glutathione through the maleimide, and (1) exposure of GST to the glutathione-
functionalized film ............................................................................................................. 95

Figure 3.11: Spectrum resulting from subtraction of the PHEMA-SA spectrum (Figure
3.10a) from the PHEMA-SA-maleimide spectrum (Figure 3.10b) showing the
immobilization of the maleimide to the film ..................................................................... 96

Figure 3.12: Spectrum resulting from subtraction of the PHEMA—SA-maleimide-GSH
spectrum (Figure 3.10c) from the PHEMA-SA-maleimide-GSH spectrum after exposure
to GST (Figure 3.10d) showing that there is essentially no GST bound to the film ......... 97

Figure 3.13: Positive-ion conventional MALDI mass spectra of 20 pmol of GST that was
purified using dialysis. The signal for the [M+H]+ ion of glutathione (GSH) has an m/z
value of 308. The inset shows an expanded region of the mass spectrum ........................ 98

Figure 3.14: CID MS/MS spectrum of glutathione (m/z 308) isolated from 20 pmol of
GST purified by dialysis. Losses of water as well as glycine (G) and glutamic acid (E)
residues are labeled ............................................................................................................ 99

Figure 3.15: Positive-ion MALDI mass spectra, obtained by conventional analysis, of 12
pmol of GST purified according to the second dialysis protocol and by adsorption and
elution from ZipTipmg pipette tips. Reduced glutathione is present at m/z 308. The inset
shows an enlarged region of the mass spectrum .............................................................. 100

Figure 3.16: Reflectance FT IR spectra of PHEMA-maleimide-glutathione films after

immersion in a) as received GST and b) GST purified by dialysis with PBS and desalting
with ZipTipsTM. The inset shows the subtraction of these spectra .................................. 101

xi

Chapter Four: Enrichment of Glycopeptides Using Aminophenylboronic Acid-
Derivatized Polymer Brushes

Figure 4.1: Comparison of the binding capacity of a thin monolayer film and a thicker
polymer brush. Spheres represent either a protein or peptide ......................................... 108

Figure 4.2: Schematic diagram showing modified gold-coated substrates after a)
adsorption of an MUD SAM, b) attachment of initiator to MUD, c) polymerization of
HEMA, d) derivatization of PHEMA with succinic anhydride (SA), e) activation of
PHEMA-SA with NHS, and f) derivatization of PHEMA-SA with 3-aminophenylboronic
acid (APBA) ..................................................................................................................... 114

Figure 4.3: Reflectance FTIR spectra of a PHEMA brush immobilized on a gold-coated
substrate before a) and after b) functionalization with succinic anhydride, c) activation of
PHEMA-SA with NHS, and d) attachment of APBA to the PHEMA-SA film .............. 116

Figure 4.4: Schematic diagram showing fructose binding, under basic conditions, to a
gold-coated substrate modified with an APBA-PHEMA film ........................................ 117

Figure 4.5: Reflectance FT IR spectra of a) a gold-coated substrate modified with APBA-
PHEMA (the substrate was immersed in a pH 10.6 solution for 15 min prior to taking the
spectrum), and b) fructose bound to the same APBA-PHEMA-modified substrate in pH
10.6 solution. The inset shows the subtraction of the spectrum of APBA-PHEMA film
(a) from the spectrum of the fructose-bound film (b) ...................................................... 119

Figure 4.6: Amino acid sequence of horseradish peroxidase. N—glycosylation sites are
labeled with bold and italic type and with a pound (#) sign. Cysteine residues involved in
disulfide bonds are underlined and tryptic cleavage sites are in bold .............................. 120

Figure 4.7: Positive ion MALDI mass spectra of 5 pmol of HRP digest analyzed using a)
conventional MALDI-MS, and b) an APBA-PHEMA-modified plate with H20 as the
loading solution and rinsing with ethanol. In b), glycopeptides were eluted with 1 11L of
0.1% TFA, followed by addition of matrix prior to analysis. Pound signs (#) represent
signals due to glycopeptides ............................................................................................ 122

Figure 4.8: Positive ion MALDI mass spectra of 2 pmol of HRP digest in 0.1 M
ammonium bicarbonate obtained using a) conventional analysis, and b) binding to
APBA-PHEMA-modified gold-coated plates that were rinsed with 0.1 M ABC, followed
by ethanol, prior to elution of glycopeptides by deposition of 1 uL 0.1% TFA and
addition of matrix solution ............................................................................................... 125

xii

LIST OF ABBREVIATIONS

ACN ............................................................... acetonitrile

ADP ............................................................... adenosine diphosphate

APBA ............................................................. 3-aminopheny1boronic acid

ATP ................................................................ adenosine triphosphate

ATRP ............................................................. atom transfer radical polymerization

BAC ............................................................... bromoacetyl chloride

BIBB .............................................................. bromoisobutyryl bromide

bpy .................................................................. 2,2’-bipyridine

BSA ................................................................ bovine serum albumin

a-CHCA ......................................................... a-cyano-4-hydroxycinnamic acid

CID ................................................................. collision-induced dissociation

Con A ............................................................. concanvalin A

DHB ............................................................... 2,5-dihydroxybenzoic acid

D/I .................................................................. desorption/ionization

DMAP ............................................................ 4-dimethylaminopyridine

DMF ............................................................... N,N’ -dimethylformamide

DTT ................................................................ dithiothreitol

EDC ............................................................... N—(3-dimethylaminopropyl)-N ’-
ethylcarbodiimide hydrochloride

ER .................................................................. endoplasmic reticulum

ESI ................................................................. electrospray ionization

FAB ................................................................ fast atom bombardment

FD .................................................................. field desorption

xiii

FT IR ............................................................... Fourier transform infrared

GalNAc .......................................................... N—galactosamine

GlcNAc .......................................................... N—glucosamine

GSH ............................................................... reduced glutathione

GST ................................................................ Glutathione-S-Transferase

IDA ................................................................ iminodiacetate

IMAC ............................................................. immobilized metal affinity chromatography
LbL ................................................................. layer-by-layer

LD .................................................................. laser desorption

LIT ................................................................. linear ion trap

MALDI .......................................................... matrix-assisted laser desorption/ionization
MOAC ........................................................... metal oxide affinity chromatography
MPA ............................................................... mercaptopropionic acid

MS .................................................................. mass spectrometry

MS/MS ........................................................... tandem mass spectrometry

MUA .............................................................. mercaptoundecanoic acid

MUD .............................................................. mercaptoundecanol

m/z .................................................................. mass-to-charge ratio

ND .................................................................. nanodiamond

NHS ............................................................... N—hydroxysuccinimide

NP .................................................................. nanoparticle

NTA ............................................................... nitrilotriacetic acid

PAA ............................................................... poly(acrylic acid)

xiv

PAH ............................................................... poly(allylamine hydrochloride)

PDMS ............................................................. polydimethylsiloxane

PEM ............................................................... polyelectrolyte multilayer
PHEMA ......................................................... poly(2-hydroxyethyl methacrylate)
phos b ............................................................. phosphorylase b

PNGase F ....................................................... protein-N—glycanase F

PSS ................................................................. poly(sodium 4-styrene sulfonate)
PTBA ............................................................. poly(tert—butyl acrylate)

PTM ............................................................... post-translational modification
QIT ................................................................. quadrupole ion trap

RNase B ......................................................... ribonuclease B

SA .................................................................. succinic anhydride

SALDI ............................................................ surface-assisted laser/desorption ionization
SAM ............................................................... self-assembled monolayer

SEM ............................................................... scanning electron microscopy
S/N ................................................................. signal-to-noise ratio

tBOC .............................................................. t—butyl-dicarbonate

TEA ................................................................ triethylamine

TFA ................................................................ trifluoroacetic acid

THAP ............................................................. 2’,4’,6’-trihydroxyacetophenone
TOF ................................................................ time-of-flight

WGA .............................................................. wheat germ agglutinin

XV

Chapter One: Introduction

1.1 Outline

The focus of this thesis is the development of functional surfaces for selective
capture and enrichment of analytes, specifically digested proteins with post-translational
modifications (PTMs), prior to their analysis by matrix-assisted laser
desorption/ionization mass spectrometry (MALDI-MS). Figure 1.1 demonstrates the
enrichment process. To put the research in perspective, this chapter first discusses
surface modifications and some of their applications (section 1.2). Subsequent sections
describe the significance of two PTMs of proteins, phosphorylation and glycosylation, as
well as the primary method for their characterization, MS (sections 1.3 and 1.4).
Although MS is the leading technique for phosphoprotein and glycoprotein analysis,
challenges in the detection of proteins or peptides with PT Ms have led to the
development of methods for enriching these species prior to analysis. Section 1.5
provides an overview of several enrichment techniques for phosphopeptides and
glycopeptides. The research discussed in this thesis primarily focuses on the
modification of MALDI sample plates to provide rapid, high—throughput on-plate
enrichment with minimal sample loss and sample handling. Section 1.6 gives an

overview of this research.

 

0 Modified Peptide

O Nonmodified Peptide
O Digest Reagents

 

 

 

Rinse Elute and Matrix Crystals
Apply Digest and Dry Add Matrix

—» ——»m

Functionalized
Surface

Figure 1.1: Schematic diagram of the enrichment of modified peptides from a protein
digest spotted on a functionalized surface. Nonmodified peptides and digest reagents are
rinsed away, and bound modified peptides are eluted and allowed to crystallize with a
matrix. The analyte can be analyzed directly on the plate using MALDI-MS. Figure
adapted from Dunn et a1.1

1.2 Surface Modification
1.2.1 Polymer Grafting

Modification of solid supports with thin organic films allows tailoring of surfaces
for a variety of applications, including protein isolation. One common modification
involves grafting polymer brushes on a solid substrate, and either chemisorption or
physisorption can link the brushes to the surface. The polymer films described in this
thesis contain a covalent linkage to an alkanethiol self-assembled monolayer (SAM),
which is adsorbed to a gold-coated substrate. In general, polymer films can be “grafted
to” or “grafted from” a surface (Figure 1.2), and this thesis includes both forms of
grafting. Grafting of poly(acrylic acid) (PAA) films to a gold-coated surface (Chapter 3)
occurs by covalently linking polymer chains to a surface. However, the “grafting to”
method typically yields thinner films with lower chain densities than the “grafting from”
strategy. This is due to limited access to reactive sites after grafting even a low density of
chains to the surface. The chains lie predominantly parallel to the substrate, thus

covering sites for further grafting. As an alternative, grafting of poly(2-hydroxyethyl

methacrylate) (PHEMA) can occur by growth of the polymer chains from initiators
immobilized on a gold-coated surface (Chapters 3 and 4). This grafting method allows
thicker films to form because small monomers can continue to diffuse to the reactive sites
as the polymer grows. In “good” solvents, the growing polymer chains also swell, which
facilitates diffusion of small molecules in the film. Therefore, polymer films can be
easily derivatized for a variety of applications including the rapid capture of

biomolecules.

m = monomer

Eire i

”grafting to” ”grafting from”

 

Figure 1.2: Schematic diagram comparing “grafting to” versus “grafting from” methods
for growing films on surfaces. Figure adapted from Bruening et al.2

Many polymerization techniques are applicable to growth of polymers from
surfaces, including free radical polymerization, anionic polymerization, cationic
polymerization, controlled radical polymerization, and atom transfer radical
polymerization (ATRP). This thesis focuses on the use of surface-initiated ATRP to
rapidly grow PHEMA films from gold-coated substrates. Derivatization of the brushes

with reduced glutathione or aminophenylboronic acid (APBA) aims at developing

surfaces for the capture of glutathione-s-transferase (GST)-tagged proteins or
glycopeptides, respectively.

Matyjaszewski and Sawamoto introduced ATRP in 1995.” This technique is a
particularly appealing form of polymerization due to its controlled growth of polymers
that results in a narrow molecular weight distribution.5 Surface initiated ATRP often
generates uniform brushes, and the use of commercially available initiators and transition

6 The controlled growth of polymer brushes in ATRP

metal catalysts is also attractive.
occurs because the rate of activation of dormant halogen-terminated chain ends by the
oxidation of copper (I) is significantly slower than the reverse reaction (Figure 1.3). This

results in a low radical concentration that favors polymerization over termination and

gives a fairly constant rate of polymerization.5

 

kact
RX + Cu(l)X/ bpy A —‘ R . + Cu(ll)X2 / bpy
Initiator kdeaCt kt
kp
Termination
Monomer

Figure 1.3: Scheme of the mechanism for transition metal-catalyzed ATRP, adapted from

Matyjaszewski.5 km, kdeact, kp, and kt represent the rate constants for activation,
deactivation, polymerization, and termination, respectively.

1.2.2 Polyelectrolyte Multilayers

Solid supports can also be modified by layer-by-layer (LbL) adsorption,7'8’9 which
in one of its simplest forms involves exposing a substrate to alternating polyanion and
polycation solutions, with rinsing with H20 after adsorption of each “layer” (Figure 1.4).

The application of many adsorption steps yields thicker films, and virtually any charged

substrate can be modified in this fashion. The ionic strength, pH, and polyion
concentration in the solution also affect the thickness and stability of the film,7 and the
polyelectrolyte multilayer (PEM) films can terminate with either a polycation or
polyanion, including charged nanoparticles. Because of the versatility of LbL adsorption,

a number of studies attempted to employ PEM films in applications such as catalysis,

  

 

drug delivery, and sensing.2’7"0’ll
1. Polycation 1. Polyanion
2. H20 Rinse 2. H20 Rinse
all ----- I —-> b) ----- ——> c) -----

 

 

 

Figure 1.4: Schematic diagram of a negatively charged substrate before (a) and after
adsorption of a polycation and rinsing with H20 (b), and after subsequent adsorption of a
polyanion and a second rinsing with H20 (0). Repetition of the process yields multilayer
films. Figure adapted from Bruening et a1.2

1.3 Post-Translational Modifications of Proteins
1.3.1 Protein Phosphorylation

Protein phosphorylation is a reversible post-translational modification (PTM)
regulated by kinases and phosphatases in eukaryotic cells. Although phosphorylation is
just one of hundreds of PTMs, it is a key regulatory mechanism and plays an essential
role in several cellular functions, including membrane transport, gene expression, and
apoptosis.12’13”"“5"6"7’18 In 1955, Fischer and Krebs first recognized the significance of
protein phosphorylation and dephosphorylation as a regulatory mechanism in the cell,19
and many studies focused on this PTM since the early 19905 when Liu and coworkers
determined that the drug Cylcosporin specifically inhibits the protein phosphatase PP2B

to make organ transplantation possible.20

In eukaryotic cells, protein phosphorylation can occur at the hydroxyl groups of
the amino acid residues serine (S), threonine (T), and tyrosine (Y) when adenosine
triphosphate (ATP) donates its gamma phosphate to the amino acid residue and becomes
adenosine diphosphate (ADP) (Figure 1.5).” Reversible protein phosphorylation is an
enzymatically regulated process, where kinases catalyze the addition of a phosphate
group and phosphatases catalyze dephosphorylation (Figure 1.6). Upon phosphorylation,
the protein may undergo conformational changes that change its activity, so
phosphorylation of a protein can, in a sense, “turn on” a regulatory pathway, which can

then be “turned off” by dephosphorylation.22

NH2
/N \N
R “Y [3 on < I 2
if i ii “ ..
o o\H + °o—I|=—o—Fl>—o—r|>—o 0
R. 0.6 .09 09 ATP
"’2 HOH HH
Mg +
kInase ””2
N \
</ I j‘
/
”N/R ii I N N
0 0-2-09 + °°"i"°"i‘° .,
(La 0, o, ADP
R' H H
OH H

Figure 1.5: Kinase—catalyzed phosphorylation of serine by ATP, adapted from Walsh.21

Phosphorylation

ATP ADP
Kinase

Phosphatase

0

Dephosphorylation

Figure 1.6: Cartoon showing reversible protein phosphorylation and the conformational
change of the protein upon addition of the phosphate group.

Because phosphorylation plays a major role in several regulatory mechanisms,
abnormal phosphorylation of proteins either leads to or results from numerous types of
diseases, including diabetes, muscular dystrophy, and various forms of cancer.23’24
Therefore, the identification and analysis of phosphorylation processes is essential for
understanding cellular regulation and for the advancement of pharmaceutical targets.25
However, the amount of a particular phosphorylated protein in the cell is low relative to
the amount of nonphosphorylated proteins, and a protein often contains multiple serine,
threonine, and tyrosine residues. Both of these facts make identification of

phosphorylation sites challenging. Even so, such identifications can sometimes be made

using mass spectrometry, which is discussed in section 1.4.

1.3.2 Protein Glycosylation

Protein glycosylation is a common post-translational modification that occurs in
eukaryotic cells. Processing of the core oligosaccharides of the glycan moiety of the
glycoprotein begins in the endoplasmic reticulum (ER) and further trimming and
modification of the glycoprotein occurs once moved to the Golgi complex?“27 Figure
1.7, adapted from Nelson,22 shows a simplified scheme of the synthesis of the core
oligosaccharide on an N-linked glycoprotein. Synthesis begins on the cytosolic surface
of the ER, where monosaccharide units are added individually to dolichol phosphate.
After addition of two N-acetylglucosamines (GlcNac, step a) and five mannose units
(step b), the oligosaccharide in the figure is translocated across the ER membrane and
into the lumen (step c), where more monosaccharide units can be added (not shown here).
The core oligosaccharide is then transferred from dolichol phosphate to an asparagine
(Asn) residue on the growing polypeptide chain with the assistance of the
oligosaccharyltransferase enzyme complex (step d),26 yielding an N—linked glycoprotein
(e). The fully translated protein is then released from the mRNA (step f) and the dolichol
phosphate is translocated back to the cytosolic surface of the ER (step g) and a phosphate
group is released (step h). The glycoprotein is further modified within the ER and in the
Golgi complex, with the assistance of several enzymes. Processing of the glycan moiety
within the ER produces relatively uniform glycoproteins, as any alterations administered
in the ER are common among all glycoproteins. Most structural diversity between

glycoproteins is not introduced until the biomolecules are in the Golgi complex.26

@ UMPGD

P SGDP SGDP-Man P +UDP 2UDP-GICNAC

P P P

@

   

Endoplasmic Reticulum

 

 

 

 

 

 

Cytosol
5’
mRNA E1 3
A GlcNAc H30 CH3
0 Mannose _ (II
Dolichol Phosphate H CH32 O—IID—OH
n OH
Dolichol Phosphate

 

 

Figure 1.7: Schematic diagram showing the synthesis of the core oligosaccharide of a
glycoprotein, where the oligosaccharide moiety is built up by the successive addition of
monosaccharide units, (a) and (b), which occurs on the cyctosolic face of the ER. The
oligosaccharide is translocated across the ER membrane and into the lumen (c) and
additional monosaccharide units can be added while inside the lumen (not shown here).
The core oligosaccharide is transferred from the dolichol phosphate to an Asn residue of
the protein ((1), forming an N—linked glycoprotein (e). The fully translated protein is
released from the mRNA (1) and the dolichol phosphate is translocated (g) and a
phosphate is removed (h). The glycoprotein is further modified within the ER and in the
Golgi complex. Figure adapted from Nelson.22

The glycan moiety of a glycoprotein attaches to the polypeptide chain through
either an N— or O-linkage (Figure 1.8). An oligosaccharide can attach its anomeric carbon

to the hydroxyl group on either a serine or threonine residue, forming an O-glycosidic

\D

linkage. Figure 1.8a shows an 0-linkage between N-acetylgalactosamine (GalNAc) and
the hydroxyl group on serine. An oligosaccharide can alternatively attach its anomeric
carbon to the amide nitrogen on an asparagine or glutamine residue, forming an N-
glycosidic linkage}:2 Figure 1.8b shows an N—linkage between N—acetylglucosamine
(GlcNAc) and the amide nitrogen on asparagine. The asparagine, however, must be two
amino acid residues away from either serine or threonine to form an N—linkage,
AsnXSer/I‘ hr, where X is any amino acid residue except proline, which because of its
ring structure inhibits the formation of a glycosidic linkage. The presence of either serine
or threonine in the third position is necessary for the formation of a hydrogen bond
between the carbonyl group on either Asn/Gln and the hydroxyl group on Ser/T hr. This
facilitates the formation of the glycosidic linkage by decreasing the dissociation constant

of the amide nitrogen.28

(a) O-linked (b) N-linked
\
9 :H
CH2 CH2 0 c=o
ll H2 |
OH 0 H H 0 HN-C-C '(l3H
H C=0 H
H H H2 1 H NH
O O-C 'CH OH
H I o H
NH /
H NH H NH
$=O % (3:0
CH3 CH3
GalNAc Ser GlcNAc Asn

Figure 1.8: Structures of a) N-acetylgalactosamine 0-linked to the hydroxyl group of
serine and b) N—acetylglucosamine N—linked to the amide nitrogen of asparagine. Figure
adapted from Nelson et a1.22

10

The glycan moiety of the glycoprotein assists in the folding and conformational
stability of the individual protein, whereas the entire glycoprotein plays a significant role

2627.29.30.31 Protein

in protein recognition, the immune system, and cell-to-cell recognition.
glycosylation is enzymatically controlled, as discussed above, which means that the
monosaccharide units comprising the oligosaccharide can be easily modified if the
enzyme or reactant concentration is altered. This variation in oligosaccharide structure
may alter the function and activity of the glycoprotein, which can potentially lead to
disease?“27 Thus, it is not surprising that abnormal glycosylation of proteins plays a key
role in several diseases, including diabetes, liver disease, and various forms of
cancer.32’33'34’35 The identification and analysis of protein glycosylation is, therefore,
essential for understanding cellular regulation and these diseases. Unfortunately,
determination of glycosylation sites presents a challenge similar to that associated with
the identification of phosphorylation sites, namely the low abundance of glycoproteins
and peptides relative to that of nonglycoproteins and peptides. Additionally, the
oligosaccharides attached to the protein can contain various monosaccharide units, and

this diversity complicates analysis of protein glycosylation. Nevertheless, mass

spectrometry (Section 1.4) is a useful tool for glycoprotein analysis.

1.4 Mass Spectrometry for Characterization of Phosphoproteins/Phosphopeptides
and Glycoproteins/Glycopeptides

Throughout the recent decades, mass spectrometry (MS) has become the premier
method for protein sequencing and for the identification of phosphorylation and
glycosylation sites. First introduced in 1897 by Sir Joseph J. Thomson,36 MS has

developed into a tool that is useful for a wide range of applications. Although there are

11

several types of mass spectrometers, all instruments have the general design described in
Figure 1.9. Ions analyzed by MS must be in the gas phase; however, the analyte may be
introduced into the instrument in the liquid, gas, or solid phase. The ionization source
converts the analyte into gas-phase ions, and the mass analyzer separates the ions based
on their mass-to-charge ratios (m/z). The detector then “counts” these ions, and a
computer processes the data to give a mass spectrum.

Sam le .~—.~-.1h Ionization ..-- Mass a
p :. . e.__.__ ; . Detector ,2, ,.
lnlet 2! Source Analyzer r ,1.

.._~

Computer

Figure 1.9: Schematic diagram of a mass spectrometer.

1.4.1 Ionization Sources

The ionization source and mass analyzer are the two most important components
of the mass spectrometer. Initially designed by Arthur J. Dempster in 1918, electron
ionization (E1) was one of the first ionization methods developed and remains the most
common form of ionization for organic molecules. While the development of the El
source was a major advancement for mass spectrometry, this ionization method limits
samples to those that are volatile and thermally stable. This ionization method is not
conducive to the analysis of large biomolecules, such as proteins and peptides, because
vaporization of these molecules is challenging without extensive thermal degradation.
These challenges were overcome in the late 19703 and early 19805 with the development
of desorption/ionization (D/I) techniques that first included field desorption (FD) and
laser desorption (LD), which are able to produce gas-phase ions from condensed-phase
samples. While these D/I methods greatly expanded the range of compounds open to
analysis by mass spectrometry, field desorption and laser desorption have limited use

12

today.37 FD involves heating of the sample on the probe with an electric field as high as
108 V/cm and requires an experienced operator.37 LD involves a pulsed laser focused on
a solid target, which rapidly heats the analyte causing it to desorb and ionize.
Unfortunately, the use of LD is limited to compounds with molecular weights of
approximately 500 Daltons (Da), as any analyte over this mass will fragment.37 This
mass limitation was overcome with the development of fast atom bombardment (FAB) by

Barber and Bordoli in the early 19805.38’39

in which a high energy beam of neutral atoms
or molecules is focused on the sample.37 Unfortunately FAB is an insensitive ionization
method and has been superseded by matrix-assisted laser desorption/ionization (MALDI)
and electrospray ionization (ESI). MALDI is related to laser desorption in that a pulsed
laser is used to desorb and ionize the analyte, but MALDI incorporates a matrix to aid in
this process. Although ESI is not a desorption/ionization method, it is capable of ionizing
non-volatile analytes and has overcome mass and sensitivity limitations. A5 soft

ionization techniques, MALDI and E81 have revolutionized mass spectrometry and its

applications to biomolecule analysis. These techniques are further discussed below.

1.4.1.1 Matrix-Assisted Laser Desorption/Ionization
The principle of MALDI was developed in the late 19805 by Hillenkamp and
Karas, who discovered that the energy required to ionize alanine decreases 10-fold in the

0

presence of tryptophan.4 This concept was adapted from Tanaka, who originally

demonstrated that an inorganic cobalt nanopowder, mixed with glycerol, provided rapid

I and has since been applied to

heating necessary to desorb and ionize the analyte,4
MALDI, where the compound of interest is mixed with a matrix prior to analysis. The

matrix consists of a small organic molecule that strongly absorbs light at the wavelength

13

of the laser. Common matrices include 2,5-dihydroxybenzoic acid (DHB), a-cyano-4-
hydroxycinnamic acid (a-CHCA), and trihydroxyacetophenone (THAP). A solution with
a 1:1000 molar ratio of analyte to matrix crystallizes on a MALDI sample plate, which is
typically made of stainless steel and occasionally coated with gold. After evaporation of
the solvent and crystallization of the matrix-analyte mixture, a pulsed laser, usually a
nitrogen laser at 337 nm, is fired at the sample, as depicted in Figure 1.10, and the matrix
sublimes due to rapid heating by the laser. The energy absorbed by the matrix transfers
to the analyte molecule (M), causing the analyte molecule to desorb and ionize with the
matrix. MALDI is a soft ionization technique, so singly-charged intact ions form with
minimal fragmentation, as described in Karas’ “lucky survivors” model,42 and peaks
present in a MALDI mass spectrum typically correspond to a singly protonated species,
[M+H]+. Signals due to sodium adducts, [M+Na]+, also commonly appear in MALDI
mass spectra. MALDI is a useful ionization technique for large, non-volatile, and
thermally labile species, including proteins and peptides, polymers, carbohydrates, and
inorganic compounds.37 Additionally, sample preparation is simple and MALDI exhibits
a relatively high tolerance to contaminants such as salts and buffers. MALDI is the

ionization technique used for all MS experiments described in this thesis.

14

  
 

Organic Matrix: 0
Biomolecule: .

  

 

 

Figure 1.10: Schematic diagram of the ionization of a MALDI sample, where the organic
matrix is present in a much higher quantity than the analyte. After co-crystallization of
the biomolecule and matrix, a pulsed laser is fired at the sample and the matrix and
analyte are both desorbed and ionized.
1.4.1.2 Electrospray Ionization

Another technique that has greatly expanded biomolecule analysis is electrospray
ionization (ESI). Developed in the late 19805, ESI is unique in that ionization occurs
under atmospheric pressure and ions often form as multiply-charged species. Ions form
due to the application of an electric field (3-6 kV) to a solution passing through a

7 This electric field causes charge to

capillary tube with a flux of 1—10 ILL/min.3
accumulate at the surface of the solution at the end of the capillary, and highly charged
droplets form as the solution breaks free. These charged droplets pass through an inert
gas, typically nitrogen, which removes any remaining solvent. Charge accumulates on
the surface of the droplet and Coulombic explosions result due to charge repulsion.
Secondary droplets continue to break down until only individual ions remain, many of

which may be multiply-charged. ESI is widely used for analysis of proteins and synthetic

polymers, and even small polar molecules.37

15

1.4.2 Mass Analyzers Compatible with MALDI
1.4.2.1 Time-of-Flight Mass Analyzer

MALDI is a pulsed ionization method, producing ions in bundles, and the time—
of-flight (TOF) mass analyzer analyzes packets of ions at a time. Hence, the TOF
analyzer is highly compatible with a MALDI source. The TOF analyzer separates ions
based on the time required to travel through a field-free region from the ion source to the
detector. Bundles of ions are ejected from the source and a potential is applied across an
electrode and extraction grid, through which the ions pass, causing them to accelerate
toward the flight tube. Accelerated ions with the same charge achieve the same kinetic
energy, and because kinetic energy (ER) is proportional to mass (m) and the square of
velocity (v) (Equation 1.1), ions with different masses have different characteristic
velocities. Therefore, in TOF analyzers ions are separated based on velocity, and mass-
to-charge ratios are determined by measuring the time required for the ions to pass from
the source to the detector. This is described by Equation 1.3, which states that the mass-
to-charge ratio (m/z) is directly proportional to the product of the charge on an electron (e
= 1.67 x 10'19 C), the accelerating voltage (V), and the square of the time (t), divided by
the square of the distance traveled by the ion (d). The length of the flight tube is typically
1 to 2 m. As described by Equation 1.3, the ion with the smallest m/z value will have the

smallest flight time and will, therefore, reach the detector first.

16

 

Equation 1.1: Ek = émv2 = zeV

 

d
Equation 1.2: v = 7
2
Equation 1.3: fl = Zth
z d 2

 

 

 

Since its first use as a mass analyzer in the 19505 by Wiley and McLaren, several
developments such as the use of an electrostatic reflector and delayed pulsed extraction
greatly improved the resolution of TOF analyzers. In addition to improvements in mass
resolution, TOF measurements are attractive for analysis because mass spectra can be
collected quite rapidly, making TOF a high-throughput technique. The mass range of
TOF is, in principle, unlimited. Karas and coworkers showed that even a protein as large

as 300 kDa can be examined.43

1.4.2.2 Linear Ion Trap Mass Analyzer

All mass spectrometry experiments described in this thesis were conducted using
a Therrno VMALDI LTQ XL instrument that contains a linear ion trap (Figure 1.11). The
vMALDI source, where v stands for vacuum (170 mTorr), houses the sample plate, a
fiber optic cable connected to the N2 laser, a camera, and a modified quadrupole (000).
In this particular instrument, the N2 laser (337 nm) irradiates the sample at an incident
angle of ~30° and the diameter of the laser spot is about 80-120 pm. The RF quadrupole
(00) and octapole (01) are located after the source and guide ions into the linear ion trap.
Before the RF quadrupole and octapole is the 000 quadrupole, which increases the

translational kinetic energy of the ions. The RF quadrupole creates a downhill potential

17

gradient to guide the ions and then the octapole also increases the translation kinetic
energy of the ions. Helium is used asthe bath gas, at 10'3 Torr, in the linear ion trap to
collisionally cool the ions, as well as for collision-induced dissociation (CID) for tandem
mass spectrometry experiments. Using the parameters discussed above, the upper m/z

value for this instrument is 4000.

 

O H
Laser Beam g g
/ RF quadrupole "' RF octopole -'
I
.fi/ 000 ) Q0 I Q1

 

 

 

 

B = Linearlon Trap
\ 7 l | r 2
Sample Skimmer x
Plate
Figure 1.11: Schematic diagram of the Therrno vMALDI LTQ XL instrument with a
linear ion trap. Figure is redrawn from Garrett and coworkers.44

 

 

 

The quadrupole ion trap (QIT), also known as the Paul trap, was developed by
Wolfgang Paul in the 19605. In independent studies in 1993, Schwartz and Jonscher first
coupled the MALDI source and quadrupole ion trap to one another.45’46 Two-
dimensional (2D) quadrupole ion trap mass analzyers, or linear ion traps (LIT), are
similar to three-dimensional (3D) QITs, except 2D analyzers confine ions in the radial
dimension (xy plane) of the quadrupole via application of an RF potential, while a DC
potential confines ions in the axial dimension at the ends of the trap.37 Briefly, the
Therrno 2D ion trap consists of four parallel hyperbolic rods, which are divided into three
sections (Figure 1.12). Upon entering the LIT along the z-axis, ions are collisionally
cooled with an inert gas (He) and oscillate in the xy plane due to the RF potential applied
to the rods, while moving along the z-axis in between the end electrodes. A DC voltage
is applied to the ends of the quadrupole to prevent the ions from exiting the LIT.37 Ions

18

have stable trajectories when q is less than 0.908 in the LIT for a given VRF range (Figure
1.13), which is described in Equation 1.4, where Q = 2m, which is the applied RF
frequency (v = 1.2 MHz), VRF is the magnitude of RF voltage (ramped from zero to 10
kV), and r0 is the distance from the center of the trap to the end (r0 = 4.12 mrn).37’47 Ions
below the mass range are lost because they have unstable trajectories.37 The major
advantage of the linear ion trap over the Paul ion trap is that the LIT has a much higher
trapping efficiency. This 2D trap has an efficiency of 55-70%, while the Paul trap only
has 5% trapping efficiency.37 The LIT is capable of storing more ions because space
charge effects are minimal compared with the Paul ion trap since ions are focused along

the xy plane instead of in all three directions.

Back
Section

Center

  
 
   

Front
Section

2 (axial)
x(radial)

Figure 1.12: Schematic diagram of a Therrno linear ion trap. Ions are trapped in the
center section of the 2D LIT, and the front and back sections are necessary to minimize
distortion of the electric field of the center section. This improves trapping efficiency.
Ions are radially ejected through two slots (30 x 0.25 mm) in the exit rods in the center
section. Figure adapted from de Hoffmann et al.37

19

 

 

 

IOI lle IOI H

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.908

qx

Figure 1.13: In the LIT, ions have stable trajectories when q is less than 0.908. Circles
represent ions in the stable region and the size of the circles corresponds to the relative
size of m/z values. For the a-q stability diagram, see either de Hoffmann et a]. or
Douglass et al.37’47

 

_ 4zr3VRF

Equation1.4: _ 2 2
erQ

Equation 1.5: fl = [a + ((12 /2)]y2

Equation 1.6: f = ,6V / 2

 

 

 

Ions are expelled radially from the trap through two slots (30 x 0.25 mm) in the
exit rods using resonance ejection (Figure 1.12). An ion in the linear ion trap oscillates at
a secular frequency specific to its m/z value. When an RF voltage at a particular
frequency is applied along the z-axis, the oscillations of ions in the trap with the
corresponding secular frequency will begin to increase. The secular frequency f of an ion
is related to the applied voltage through the ,6 term in Equations 1.5 and 1.6, where a is
zero in the LIT. Therefore, f is directly proportional to VRF. Equation 1.5 is applicable
only for q < 0.4. If the amplitude of the applied voltage is large enough, the oscillations
of the ions will be so great that they will be destabilized and ejected from the trap.37 Two
detectors are used in the Therrno LIT to ensure that all ions are expelled from the trap and

are detected.

20

Tandem-in-time mass spectrometry (MS/MS), including multistage tandem mass
spectrometry (MSn, where n>2), is easily performed in the linear ion trap. First, ions with
a specific m/z value are selected and all other ions are ejected from the trap (isolation).
These isolated precursor ions are then fragmented via the application of a low amplitude
resonance excitation voltage which increases the kinetic energy of the ions and causes
them to collide with the helium gas present in the trap (where q is 0.25). The collisions
convert kinetic energy to internal energy and cause the ions to fragment. Finally, the

product ions are expelled from the linear ion trap by resonance ejection.37

1.4.3 Challenges in Phosphopeptide and Glycopeptide Detection by Mass
Spectrometry

Although mass spectrometry is now the premier tool for analysis of both
phophopeptides and glycopeptides, there are several challenges in detecting these species.
In most eukaryotic cells, the amount of phosphorylated protein is low relative to the

48,49,50

amount of nonphosphorylated protein, and the same challenge exists for

glycosylated protein.51 The structural diversity of glycoproteins also makes it difficult to

determine the monosaccharide units that make up a glycan.52

Phosphorylation may
decrease the ionization efficiency of some species in the presence of
nonphosphopeptides, and this becomes increasingly challenging as the number of

53.54

phosphorylation sites increases. However, recent studies suggest that

phosphopeptides have similar ionization efficiencies as nonphosphopeptides.55

21

1.5 Enrichment Techniques for Phosphopeptides and Glycopeptides Prior to MS
Analysis

As discussed above, phosphopeptides and glycopeptides are present in low
amounts relative to other peptides present in a sample, making analysis by mass
spectrometry challenging. A number of recently developed techniques aid in the
detection of phosphorylated and glycosylated peptides by selectively capturing these
species to overcome their low abundance. The sections below provide a brief summary
of the most widely used enrichment methods for both phosphopeptides and,
glycopeptides. For more thorough descriptions of enrichment techniques, refer to recent

56.57

reviews by Dunn et a1 and Thingholm et al for phosphorylated peptides and Ito et al,

Liu et al, and Mechref et al for glycosylated peptides.58’59’60

1.5.1 Enrichment Methods for Phosphorylated Peptides Prior to MS Analysis
Phosphopeptide enrichment techniques can be broadly grouped into
immunoprecipitation, affinity chromatography, and chemical derivatization or covalent
binding. Irnmunoprecitation is limited to tyrosine phosphorylation and certain motifs for
serine and threonine phosphopeptides,61 so covalent binding and affinity chromatography
are more general techniques. Although covalent binding may allow higher enrichment
specificities than affinity chromatography, the majority of the covalent procedures
require multiple steps.‘r’2’63’64 Hence, affinity chromatography is the most widespread
method for phosphopeptide enrichment. Typical forms of affinity methods include
immobilized metal affinity chromatography (IMAC) and, more recently, metal oxide
affinity chromatography (MOAC). These two techniques are similar in that

phosphopeptides selectively bind to a resin, while rinsing removes unbound species,

22

including impurities. Finally, elution yields concentrated analyte, and the
phosphopeptides can be analyzed via mass spectrometry. Figure 1.14 shows this general
process. Although IMAC and MOAC may not provide the specificity of covalent
binding, they are relatively simple, rapid enrichment techniques and are useful for

numerous applications.

 

 

 

 

Protein %
° 0
DIgestOg:N 42 Mr MI Phosphopeptide
“gowiéQO W Nonphosphopeptide
<3 W? O Digest Reagent

000. 0,20 9 000'
o. o , v0.4.0 0' o
lMACor.... RInse .. 52—. Elute ....
MOAC ..OOO “’ '30...) “* ....O
.6 t“
Resin 0 O . ”V .63 O .
0.. .9." o ‘0'.
O. .0 O. .0 O. .0
00.0 a... a...

 

 

 

 

 

 

 

42 na sis
2%)?“ :WS -_) AByLZlS
ggggx Enriched

Phosphopeptides

 

 

 

Figure 1.14: Schematic diagram showing the general procedure for enrichment of
phosphopeptides using IMAC or MOAC in a column format. The column packing is
different for each method, but both techniques involve loading the sample onto a column,
rinsing to remove any contaminants, and elution of enriched phosphopeptides. The
analyte can then be analyzed with MS. Figure adapted from Dunn et al.56

1.5.1.1 Immobilized Metal Affinity Chromatography
Immobilized metal affinity chromatography (IMAC) was introduced in 1975 by

Porathfi5 and relies on the interaction between immobilized metal-ligand complexes and a

23

specific functional group, in this case, phosphate. The two most common metal-binding
ligands in IMAC are iminodiacetate (DA) and nitrilotriacetate (NTA) (Figure 1.15).
These ligands strongly complex metal ions such as Fe(III), Ga(III), Zr(IV), and Al(III).
Hochuli and coworkers first applied NTA to IMAC in 1987 because of its strong
complexing ability.66 This tetradentate ligand should bind metal ions more strongly than

IDA, which is a tridentate ligand. Therefore, NTA more effectively prevents metal

leaching than IDA.
O
O
_ _ o" o"
a) o o b)
A N

N O

O \H
'o 0

Figure 1.15: Structures of a) iminodiacetate (IDA), and b) nitrilotriacetate (NTA).

Unfortunately, binding of acidic residues to metal ion complexes results in
significant adsorption of nonphosphopeptides. Methyl esterification of the carboxyl
groups on glutamic and aspartic acid can sometimes overcome this challenge,67 but this
introduces an additional step in the enrichment process. Additionally, because
derivatization is usually less than 100%, esterification complicates mass spectra. The
most critical challenge with IMAC is the potential sample loss due to multiple rinsing and
elution steps on the column (Figure 1.14), and some commercial methods even require
additional rinsing steps. To overcome this challenge, several groups developed

techniques to enrich sample directly on a MALDI plate.

24

In 1993, Hutchins and coworkers introduced surface-enhanced affinity capture in
on-plate capture of a glycoprotein, which was analyzed directly with MALDI-TOF-MS.68
Two years later, Brockman and coworkers derivatized a MALDI target with a self-
assembled monolayer (SAM) for the specific capture of biomolecules.69 After incubation
of the sample, the plate was rinsed, and matrix was added directly to the plate prior to
analysis by MALDI-T0F-MS."9 These works, along with many others, lead to the
development of on-plate enrichment using IMAC in 2005. Using gold MALDI targets as
substrates, Shen and coworkers derivativized SAMs with NTA and then formed the
Ga(III)-ligand complex.70 The Ga(III)-NTA-SAM-derivatized MALDI plates allowed
successful enrichment of synthetic phosphopeptides. Our group recently used MALDI
plates coated with poly(2-hydroxyethyl methacrylate) (PHEMA) brushes derivatized with
Fe(III)-NTA to enrich phosphopeptides prior to MS analysis.71 The high capacity of

polymer brushes relative to monolayers greatly enhances enrichment efficiency.

1.5.1.2 Metal Oxide Affinity Chromatography
Metal oxide affinity chromatography (MOAC) has recently become one of the
most successful methods for enriching phosphopeptides. Sano, Pinske, and Larsen were

among the first to introduce titanium dioxide resins for enrichment of phosphorylated

72,73.74,75

peptides. Although titanium dioxide is probably the most widely used metal

oxide, zirconium dioxide is also capable of enriching phosphorylated peptides and it has

been reported that it has a higher binding affinity toward phosphate than carboxylate

76.77

anions. The isoelectric points of Ti02 and ZrO2 are approximately 6 and 7,

78.79

respectively, so at low pH both of these metal oxides are positively charged and are

able to selectively adsorb phosphopeptides. Other metal oxides, including aluminum

25

oxide (A1203), niobium oxide (Nb205), and aluminum hydroxide (Al(0H)3), have also
been used for the selective enrichment of phosphorylated peptides.80‘8"82

In MOAC, proteolytic digests are typically loaded onto a column packed with a
metal oxide resin, such as Ti02, and the loading solution usually contains either acetic
acid or trifluoroacetic acid to minimize unwanted binding of acidic peptides. Pinske and
coworkers loaded their samples in 0.1 M acetic acid, rinsed the column with 0.1 M acetic
acid in 80% acetonitrile, and eluted the phosphopeptides by increasing the pH to 9.0 with
ammonium bicarbonate.73 However, some residual nonspecific binding of
nonphosphopeptides containing acidic residues occurred. Derivatization of the acidic
residues using O-methyl esterification, resulted in little nonspecific affinity for the
Ti02.73

Larsen and coworkers used a 0.1% TFA loading buffer and a rinsing solution
containing 2,5-dihydroxybenzoic acid (DHB) to minimize nonspecific adsorption.72 TFA
is much more acidic than acetic acid and this loading solution has a pH of 1.9, compared
with 2.7 for acetic acid. The lower pH in the loading solution protonates acidic residues
more effectively and therefore aids in preventing nonspecific adsorption to the Ti02
resin. These studies also used an ammonium hydroxide solution at pH 10.5 to elute
phosphopeptides, which resulted in higher recoveries compared with a pH 9.0 elution.
This method allowed identification of 20 phosphopeptides in a MALDI-MS spectrum of
500 fmol of a-casein digest along with essentially no signals from nonphosphopeptides.72
Larsen and coworkers also examined the ability of other acids to reduce binding of

nonphosphorylated peptides to Ti02 resins. They determined that 2,5-DHB, salicylic

acid and phthalic acid exhibit the greatest ability to inhibit binding of nonphosphorylated

26

peptides, followed by benzoic acid, cyclohexane carboxylic acid, phosphoric acid, TFA,
and acetic acid.72 When binding to the Ti02 surface, salicyclic acid creates a chelating
bidentate structure, compared with the bridging bidentate complex formed between
phosphate and Ti02 (Figure 1.16). As a result of these differences, Larsen suggests that
DHB, which is similar to salicylic acid, competes for binding sites with

nonphosphopeptides and not with phosphopeptides.72

 

 

0 0V“
O/ \O
O O
Ti02 Ti02 Ti02
Chelating Bridging
Bidentate Bidentate

Figure 1.16: Schematic diagram comparing the chelating bidentate structure of salicylic
acid on Ti02 with the bridging bidentate complex formed between phosphate and
titanium dioxide. Figure adapted from Larsen et a1.”

As with IMAC, work with metal oxides has primarily occurred on columns where
potential sample loss is an issue due to multiple rinsing and elution steps. On-plate
techniques have also been applied to enrichment methods using metal oxides. Lin and
coworkers immobilized Ti02-coated gold nanoparticles on a glass slide for on—plate
enrichment of phosphopeptides and analysis by MALDI-TOF-MS (Figure 1.17).83 More
recently, EkstrElm and coworkers prepared polymer MALDI plates with channels packed
with Ti02 for the selective enrichment of one pmol to 100 fmol of B-casein digest.84 The
wash, rinse, elution, and matrix solutions were pulled through the channels using a

vacuum and the sample was flipped over for MALDI-MS analysis as the matrix

27

crystallized on the rear side of the plate. Other titanium dioxide on-plate enrichment
techniques include work by Qiao et al, where an array of Ti02 nanoparticles was
prepared on a stainless steel MALDI target by heating an array of 2 ILL drops of a 100
mg/mL T102 suspension on the plate at 400 °C for 1 h.85 Tan et a1 affixed TiO2-coated
magnetic nanoparticles to a stainless steel MALDI target by holding a magnet to the rear

side of the plate while the sample was loaded, incubated, rinsed, and mixed with matrix.86

H

   

gN
N N
< 0 g
78. . ,SK 7 .
o o Spln- o oo o
coatIng

Glass —-> Glass

   

OH OH OHOH

TMSPED

Figure 1.17: Schematic diagram showing the immobilization of a thin film of MB-
(trimethoxysilyl)propyllethylenediamine (TMSPED) on glass, attachment of gold
nanoparticles (NPs) to the film, and coating of the nanoparticles with Ti02. In the Ti02
coating step, a solution of titanium isopropoxide is spin—coated on the surface, followed
by annealing to give Ti02 nanoparticles. The figure is adapted from Lin et al.8

 

A number of studies also examined the use of magnetic beads for capturing
phosphopeptides in solution. Magnetic beads are attractive for enrichment because they
can simply be collected in an external magnetic field, and they have a high surface area to
volume ratio that gives a high binding capacity. In 2005, magnetic nanoparticles were
first coated with metal oxides for phosphopeptide enrichment.87'88'89 Typically, coating
of the F6304 beads with SiO2 occurs using either sodium silicate or tetraethyl
orthosilicate, and subsequent formation of the metal oxide employs titanium butoxide,

28

zirconium butoxide, or aluminum isopropoxide for either titania, zirconia, or alumina-
coated beads, respectively. To enrich a phosphopeptide sample, a proteolytic digest in
TFA is mixed with magnetic beads (~25 pg) in a microcentrifuge tube and incubated for
as little as 30 s. The beads are typically rinsed with an acetonitrile/T FA solution, and a
magnet is used to hold the beads to the wall of the tube while the solution is decanted.
Matrix solution that contains phosphoric acid to elute the phosphopeptides can be mixed
with the beads and then spotted to the MALDI target for direct MS analysis, or beads can
be directly spotted on the MALDI plate without addition of matrix, which is attractive as
no elution step is necessary. Without matrix, surface-assisted laser desorption/ionization
(SALDI) MS is used,87 but this technique generally yields lower signals than MALDI. In
addition to metal oxide coatings, magnetic nanoparticles can also be modified with
IMAC materials. Xu and coworkers functionalized magnetic nanoparticles with Fe(III)-
IDA in 2006 and used these beads for phosphopeptide enrichment prior to MALDI-
TOF/TOF—MS analysis.90 The development of on-plate enrichment methods and the use
of modified magnetic particles has the potential to improve the analysis of

phosphopeptides by reducing sample handling and sample 1055.

1.5.2 Enrichment Methods for Glycoproteins/Glycopeptides Prior to MS Analysis
Lectin affinity chromatography and reversible covalent binding are the two main
methods for enrichment of glycoproteins and glycopeptides prior to analysis by mass
spectrometry, with lectin affinity chromatography being more common. While both of
these techniques provide selective enrichment of glycopeptides, lectin chromatography is

0

a highly specific form of separation.9 These two types of enrichment are briefly

described below.

29

1.5.2.1 Lectin Affinity Chromatography

Lectins are proteins that have a high affinity for certain carbohydrates, and these
proteins are categorized based on the monosaccharide to which they bind most strongly.
Hence, lectin affinity chromatography separates proteins based on their glycan moiety.
Typically, serum glycoproteins are digested using trypsin, loaded onto a column
containing one or more immobilized lectins bound to a support, and rinsed. Finally,
bound peptides are eluted and deglycosylated using protein-N-glycanase F (PNGase F).9|
Deglycosylation is often necessary because the glycan moiety typically has a very large
molecular weight that precludes MS detection in some cases because the m/z value is past
the upper limit of many mass analyzers.

Although a vast amount of work has been accomplished using lectins

. . . . . 3.94
Immobilized 1n columns,929

multiple rinsing and elution steps increase the possibility
of sample loss, as with IMAC and MOAC columns for phophopetide isolation (Figure
1.14). Even Top Tips (Glygen), pipet tips loaded with lectins such as concanavalin A on
agarose, require several washings and multiple elutions prior to analysis by MALDI-
TOF-MS.95 Recent methods address these issues and aim to minimize sample handling
and 1055. One of these techniques that involves the establishment of lectin microarrays
was recently reviewed by Hu et al.96 In these methods, lectins are immobilized on a solid
support, such as gold or polydimethylsiloxane (PDMS), through either covalent bonding
or physical adsorption, and the glycopeptide sample can be directly applied to the
microarray. Wong and coworkers covalently immobilized lectins in an array on a PDMS

substrate and spotted glycoproteins directly on the microarray. Following sample

incubation, matrix was added to the microarray, which was then affixed to a MALDI

30

target for analysis by MALDI-TOF-MS.97 Microarrays are advantageous because
different lectins can be immobilized on each spot of the plate to simultaneously enrich
different glycoproteins from the same sample.

Magnetic particles can also be modified for glycoprotein analysis. Sparbier and
coworkers demonstrated that magnetic beads functionalized with concanavalin A (Con
A) and wheat germ agglutinin (WGA), the two most commonly employed lectins, are

capable of enriching glycosylated proteins.98’99

Glycoprotein samples were incubated for
1 h, washed, and eluted under acidic conditions, and the particles were held in place using
an external magnetic field, while the supernatant was decanted. Proteins were digested
with trypsin following elution and were then analyzed using MALDI-TOF-MS. The
authors show that these modified beads specifically bind glycoproteins based on the
functionality of the particles.99 These magnetic particles significantly decrease sample

handling and potential sample loss, which are challenges with column lectin affinity

chromatography.

1.5.2.2 Covalent Binding Enrichment Methods

Several covalent binding techniques exist for glycosylated peptide enrichment,
and some of these methods have been applied to modified magnetic beads. Two common
modifications of particles for binding glycosylated peptides include functionalization
with hydrazide and boronic acid groups. In 2003, Zhang and coworkers used hydrazide
functionalized particles to enrich N—linked glycosylated peptidesloo In this work, a
glycoprotein sample was oxidized using sodium periodate, which converts cis—diol
groups on the glycan moiety to aldehydes (Figure 1.18). The oxidized monosaccharide

was then coupled with hydrazide groups immobilized on a resin, forming a covalent

31

hydrazone bond, and nonglycoproteins were removed by rinsing. The covalently bound
glycoprotein was then digested with trypsin, leaving glycosylated peptides bound to the
resin whereas nonglycosylated peptides were rinsed away. The a-arnino groups on the
immobilized glycopeptides were isotopically labeled while bound to the resin, and then
the formerly N-linked glycopeptides were cleaved from the oligosaccharide chain using
PNGase F (Figure 1.18). The released peptides were identified and quantified using
either microcapillary high-performance liquid chromatography electrospray ionization
(uLC-ESI) MS/MS or by uLC separation followed by MALDI MS/MS analysis. This
enrichment technique, however, is only applicable toward N—linked glycoproteins as
PNGase F cleaves the bond between the innermost GlcNAc and asparagine residues of
the glycan group from N—linked glycoproteins,'°' and there is no comparable enzyme for
0-linked glycoproteins. Zhang and coworkers demonstrate that hydrazide-functionalized
beads are capable of selective capture of N—linked glycoproteins, as essentially all peaks
present in the mass spectra were due to expected N-linked glycopeptides.100 Since this
work, separation of glycopeptides using hydrazide chemistry has become increasingly
popular. Hydrazide chemistry is an effective enrichment technique for the identification
and quantification of glycopeptides, both by itself, as well as coupled to other separation

methods such as lectin affinity chromatography.'02'1()3"04'”)5'l06

32

CH20H CHZOH CHZOH Tryptic CHng

O Oxidation O Coupling O Digestion
OH O ——-> ———> '_>
O O O | a) O R 0 fl)

0
l I
(:)Iigosaccharide o (géligosaccharide IN I Oligosaccharide IN N Oligosaccharide
fi l

Protein Protein NH NHProtein H NHPeptide

NHz NHz

NH NH o O
Isotopic
LabeHng

Labeled
“LC-ESI Peptides CHZOH PNGase F 01"ng

. O
\ SN?" + Cleavage 2

Intensity

Intensity

2. MALDI-MS

. N N Oligosacchande N N Oligosaccharide
"“"C sepa'am” 8H NH NH NH*Labeled Peptide

00

Figure 1.18: Schematic diagram showing the oxidation of cis-diol-containing
glycoprotein with sodium periodate, coupling with hydrazide-functionalized beads, and
then tryptic digestion of the protein, yielding glycopeptides covalently bound to the
magnetic particles. Glycopeptides bound to the resin are isotopically labeled and then
cleaved using PNGase F to give formerly N-linked isotopically labeled peptides prior to
analysis by either ESI or MALDI. Figure adapted from Zhang et al.100

In addition to hydrazide-functionalized supports, boronic acid-derivatized resins
can selectively and rapidly enrich glycosylated peptides prior to MS analysis. The
affinity of boronic acid groups for cis-diol-containing compounds was discovered in the
late 19405,107"08'109 and this chemistry is applicable to glycopeptide enrichment because
essentially all carbohydrate groups in glycopeptides contain one, if not multiple, cis-diol
groups. The covalent interaction between boronic acid and cis-diols is relevant to both
N—linked and 0-linked glycopeptides, so this method is essentially applicable to all
glycoproteins. Figure 1.19 shows the equilibria for the formation of the boronate ester

from phenylboronic acid and a general cis-diol.

33

 

i“ i“
9 /OH
B _ +
\OH +H20, H \OH
R2 R1 R2 R1
-2 H O -2 H O
2 2
K - K
trig tet
HO OH HO OH
R2
0
B\ R1 e\B/
0 +H20, H \O
____.. R1

 

Figure 1.19: Relevant equilibria for the formation of a boronate ester from phenylboronic
acid and a cis-diol. The boronate ester may exist in both trigonal and tetrahedral forms.

ng and Ktet are the equilibrium constants for the formation esters from the trigonal and
tetrahedral forms of phenylboronic acid, respectively. Figure adapted from Yan et a1.110

While a cis-diol-containing compound can bind to phenylboronic acid in either
the tetrahedral or trigonal forms, the equilibrium constant is higher for formation of the
tetrahedral form, Kw, > Km} 10’1” The pKa of the boronic acid and of the boronate ester
formed, as well as the pH of the solution and the type of buffer used and its
concentration, will all affect the binding between boronic acid and the cis-diol-containing
compound.”0‘Ill The optimal pH for binding can be estimated by simply averaging the
ngacid and pKa,d,ol.' '0 While this predicted pH may be helpful, the actual optimal pH will
depend on the reaction conditions. In general, compounds containing cis-diols most
effectively bind to boronic acids under basic conditions and are released by lowering the

pH.
34

In recent years, the interaction between boronic acid and cis-diol groups has been
exploited for the enrichment of glycosylated proteins and peptides. In addition to their
work in 2005 and 2006 involving the immobilization of lectins on beads, Sparbier and
coworkers also used commercially available phenylboronic acid-functionalized beads for
glycoprotein enrichment.98'99 Glycoprotein samples were incubated under slightly basic
conditions (pH 8.5) for 1 h at room temperature under gentle shaking, and bound
glycoproteins were eluted under acidic conditions and analyzed using MALDI-
TOF/TOF-MS. Sparbier et al demonstrated that these phenylboronic acid-functionalized
beads are capable of selective enrichment of glycopeptides. The authors also showed that
while beads containing immobilized phenylboronic acid, Con A, and WGA are all able to
capture glycopeptides, each material has its individual binding profile.98’99

In a similar method, Zhou and coworkers synthesized aminophenylboronic acid-
functionalized magnetic nanoparticles and incubated them with tryptically digested
proteins under slightly alkaline conditions (pH 8.5) for 90 min.112 The nanoparticles
were rinsed, and bound peptides were eluted under acidic conditions and analyzed by
MALDI-QIT-TOF MS. The authors showed that signals (and signal to noise) due to
glycopeptides significantly increased after enrichment compared with conventional
MALDI analysis.112 Yeap and coworkers functionalized nanodiamond (ND) with
succinic anhydride and subsequent reaction with aminophenylboronic acid and then
enriched intact proteins, including ovalbumin and ribonuclease B (RNase B).113
Glycoproteins were incubated with the ND in phosphate buffer for 3 h at either pH 7.4 or

9, and the ND powder was separated by centrifugation. After removal of the supernatant,

matrix solution was mixed with the ND prior to spotting on the sample plate, and these

35

samples were air-dried and analyzed by MALDI-TOF MS. The glycosylated proteins
ovalbumin and RNase B were identified in separate experiments and bovine serum
albumin (BSA), a nonglycosylated protein, was not present in mass spectra when succinic
anhydride was used as a “spacer” between the ND particle and aminophenylboronic
acid.“3

Most recently, Xu and coworkers used mesoporous silica for selective enrichment
of glycopeptides.114 In this work, aminophenylboronic acid was immobilized within the
pores of the silica (Figure 1.20). A glycopeptide sample in ammonium bicarbonate was
added to a suspension of boronic acid-functionalized silica and was incubated for just 15
min with shaking. The supernatant was decanted after centrifugation, and the beads were
washed with ammonium bicarbonate, which was again decanted after centrifugation. A
solution containing TFA and acetonitrile was used to elute the glycopeptides (30 min)
from the silica. After centrifugation, the eluent and matrix were spotted onto a MALDI
target for analysis by MALDI-QIT MS. This boronic acid-functionalized material was
capable of enriching 23 fmol of tryptically digested horseradish peroxidase (HRP) with

little nonspecific binding.I '4

36

a) OH

H CQ 0 APBA H3CO FDU 12
H3CO',Si’\’O\/Q 3 H3 CO CSSi’\’O\)\/NO/OH 0H “—9 Si 0 :giNOWNO/B‘OH
H3CO H3Cd
GLYMO GLYMO- APBA FDU12- GLYMO- APBA
b) ' FDU-12 " " FDU12-GLYMO-APBA
' \x ,( " Glycopeptide X ,(
Graftin ' —( ~ Enrichment 3» —<‘ .
____g Y ___a . O . Y

,(«K -,(-J\

f }

 

—( GLYMO-APBA

u Glycopeptides

 

 

 

Figure 1 .20: Schematic diagram of a) reaction between (3-
g1ycidyloxypropyl)trimethoxysilane (GLYMO) and 3-aminophenylboronic acid (APBA)
to form GLYMO- APBA, followed by attachment to mesoporous silica FDU- 12; and b)
grafting of GLYMO- APBA to mesoporous silica, then enrichment of glycopeptides
within the silica pore. Figure adapted from Xu. ”4

1.6 Research Overview

This thesis focuses on the design and fabrication of various modified MALDI
plates for the direct enrichment of biomolecules prior to MALDI-MS analysis. On-plate
enrichment techniques can potentially reduce sample handling, preparation time, and
sample 1055. Chapter Two describes deposition of Ti02 nanoparticles on a gold-coated
substrate modified with a bilayer of polyelectrolytes and on-plate enrichment of
phosphorylated peptides using the modified substrate. The chapter includes a comparison
of the enrichment performance of these plates with substrates containing Ti02
nanoparticles immobilized by simple heating. The MALDI plates with TiO2 particles
immobilized by heating showed a phosphopeptide recovery of 69%. Chapter Three
presents methods for the synthesis of polymer brushes derivatized with reduced

glutathione, with the intention of specifically binding GST-tagged proteins. Chapter Four

37

describes the derivatization of thicker poly(2—hydroxyethyl methacrylate) brushes with
aminophenylboronic acid. These polymer-modified plates are examined for the specific
on-plate enrichment of glycosylated peptides from an unpurified glycoprotein digest.
Finally, Chapter Five summarizes the conclusions of this work and suggests future

directions.

38

1.7 References

1 Dunn, J. D.; lgrisan, E. A.; Palumbo, A. M.; Reid, G. E.; Bruening, M. L. “Use of High-
Capacity Polymer Brushes Immobilized on MALDI Plates for the Analysis of
Phosphopeptides by MS,” Poster Presented at the 55th ASMS Conference, June 4-7,
2007, Indianapolis, IN.

2 Bruening, M. L.; Dotzauer, D. M.; Jain, P.; Ouyang, L.; Baker, G. L. Langmuir 2008,
24, 7663-7673.

3 Kamigaito, M.; Sawamoto, M.; Higashimura, T. Macromolecules 1995, 28, 5671-5675.

4 Wang, S.; Matyjaszewski, K. Journal of the American Chemical Society 1995, 117,
5614-5615.

5 Matyjaszewski, 1c; Xia, J. Chemical Reviews 2001, 101, 2921-2990.
6 Matyjaszewski, K.; Spanswick, J. Materials Today 2005, 8, 26-33.
7 Ai, H.; Jones, 5. A.; Lvov, Y. M. Cell Biochemistry and Biophysics 2003, 39, 23-43.

8 Bertrand, P.; Jonas, A.; Laschewsky, A.; Legras, R. Macromolecular Rapid
Communications 2000, 21, 319-348.

9 Decher, G. Science 1997, 277, 1232-1237.
'0 Hammond, P. T. Advanced Materials 2004, 16, 1271—1293.

‘1 Johnston, A. P. R.; Cortez, C.; Angelatos, A. S.; Caruso, F. Current Opinion in Colloid
& Interface Science 2006, 11, 203-209.

'2 Hunter, T. Cell. 2000, 100, 113-127.
‘3 Duronio, v. Biochemical Journal. 2008.415, 333-344.

‘4 Palfrey, H. C.; Alper, S. L.; Greengard, P. Journal of Experimental Biology. 1980, 89,
103-1 15.

‘5 Sombroek, D.; Hofmann, T. G. Cell Death and Difieremiarion. 2009, 16, 187-194.
'6 Stamm, 5. Journal ofBiological Chemistry. 2008, 283, 1223-1227.

17 Turner, K. M.; Burgoyne, R. D.; Morgan, A. Trends in Neurosciences. 1999, 22, 459-
464.

39

 

‘8 Whitmarsh, A. J.; Davis, R. J. Cellular and Molecular Life Sciences 2000, 57, 1172-
1 183.

‘9 Fischer, E. H.; Krebs, E. G. Journal of Biological Chemistry 1955, 216, 121-132.

20 Liu, J.; Farmer, J. D.; Lane, W. S.; Friedman, J.; Weissman, 1.; Schreiber, S. L. Cell
1991, 66, 807-815.

2' Walsh, C. T. Posttranslational Modification of Proteins; Roberts and Company:
Englewood, CO, 2006, 35-36.

22 Nelson, D. L.; Cox, M. M. Principles of Biochemistry; W. H. Freeman and Company:
New York, NY, 2005, 228-230, 1069-1071.

23 Cohen, P. European Journal of Biochemistry 2001, 268, 5001-5010.

24 Yu, L.; Issaq, H. J.; Veenstra, T. D. Proteomic Clinical Applications 2007, 1, 1042-
1057.

” McConnell, 1. L.; Wadzinski, B. E. Molecular Pharmacology 2009, 75, 1249-1261.
26 Helenius, A.; Aebi, M. Carbohydrates and Glycobiology 2001, 291, 2364-2369.

27 Rudd, P. M.; Dwek, R. A. Critical Reviews in Biochemistry and Molecular Biology
1997, 32, 1-100.

2“ Rao, V.S.R.; Qasba, P.K.; Balaji, P.V.; Chandrasekaran, R. Conformation of
Carbohydrates; Harwood Academic Publisher: Amsterdam, Netherlands, 1998, 131.

29 Durand, G.; Seta, N. Clinical Chemistry 2000,46, 795-805.

30 Paulson, J. C. Trends in Biochemical Sciences 1989, 14, 272-276.

3' Varki, A. Glycobiology 1993, 3, 97-130.

32 Coffman, F. D. Critical Reviews in Clinical Laboratory Sciences 2008, 45, 531-562.
33 Dias, W. B.; Hart, G. W. Molecular Biosystems 2007, 3, 766-772.

34 Saldova, R.; Wormald, M. R.; Dwek, R. A.; Rudd, P. M. Disease Markers 2008, 25,
219-232.

35 Turner, G. A. Clinica Chimica Acta 1992, 208, 149-171.

36 Griffiths, I. W. Rapids Communications in Mass Spectrometry 1997, 11, 2-16.

40

 

37 de Hoffmann, E.; Stroobant, V. Mass Spectrometry: Principles and Applications; John
Wiley & Sons: West Sussex, 2007 , 29-40, 43-55.

3" Barber, M.; Bordoli, R. 5.; Sedgewick, R. D.; Tyler, A. N. Journal of the Chemical
Society, Chemical Communications 1981, 325-327.

39 Barber, M.; Bordoli, R. 5.; Sedgewick, R. D.; Tyler, A. N. Nature 1981, 293, 270-275.
4" Karas, M.; Bachmann, D.; Hillenkamp, F. Analytical Chemistry 1985, 57, 2935-2939.

4' Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshida, Y.; Yoshida, T. Rapid
Communications in Mass Spectrometry 1988, 2, 151-153.

42 Karas, M.; Gliickmann, M.; Schafer, J. Journal of Mass Spectrometry 2000, 35, 1-12.

43 Karas, M.; Ingendoh, A.; Bahr, U.; Hillenkamp, F. Biomedical and Environmental
Mass Spectrometry 1989, 18, 841.

4“ Garrett, T. J.; Yost, R. A. Analytical Chemistry 2006, 78, 2465-2469.

45 Schwartz, J. C.; Bier, M. E. Rapid Communications in Mass Spectrometry 1993, 7, 27-
32.

4" Jonscher, K.; Currie, G.; McCormack, A. L.; Yates, J. R. Rapid Communications in
Mass Spectrometry 1993, 7, 20-26.

47 Douglass, D. J .; Frank, A. J .; Mao, D. M. Mass Spectrometry Reviews 2005, 24, 1-29.

48 Kratzer, R.; Eckerstrom, C.; Karas, M.; Lottspeich, F. Electrophoresis 1998, I9, 1910-
1919.

49 Stensbalk, A.; Anderson, S.; Jensen, 0. N. Proteomics 2001, 1, 207-222.
5° Zeller, M.; Koenig, 5. Analytical and Bioanalytical Chemistry 2004, 378, 898—909.

5' Ullmer, R.; Plematl, A.; Rizzi, A. Rapic Communications in Mass Spectrometry 2006,
20, 1469-1479.

52 Medzihradszky, K. F. Methods in Enzymology 2005, 405, 116-138.

53 Liao, P. C. Leykam, J.; Andrews, P. C.; Gage, D. A.; Allison, J. Analytical
Biochemistry 1994, 219, 9-20.

54 Craig, A. G.; Hoeger, C. A.; Miller, C. L. Goedken, T.; Rivier, J. E.; Fischer, W. H.
Biological Mass Spectrometry 1994, 23, 519-528.

41

 

55 Steen, H.; Jebanathirajah, J. A.; Rush, J.; Monice, N.; Kirschner, M. w. Molecular &
Cellular Proteomics 2006, 5, 172-181.

56 Dunn, J. D.; Reid, G. E.; Bruening, M. L. Mass Spectrometry Reviews 2009 DOI
10.1002/mas.20219.

57 Thingholm, T. E.; Jensen, 0. N.; Larsen, M. R. Proteomics 2009.9, 1-18.

58 Ito, 8.; Hayama, K. Hirabayashi, J. Methods in Molecular Biology; Humana Press:
Clifton, NJ. 2009, 534, 195-203.

59 Liu, x.; Ma, L.; Li, J. Analytical Letters 2008, 41, 268-277.

(’0 Mechref, Y.; Madera, M.; Novotny, M. V. Methods in Molecular Biology; Humana
Press: Clifton, NJ. 2008, 424, 373-396.

61 Schmidt, S. R.; Schweikart, F.; Andersson, M. E. Journal of Chromatography B 2007,
849, 154-162.

62 Bodenmiller, B.; Mueller, L. N.; Pedrioli, P. G.; Pflienger, D. Molecular Biosystems
2007, 3, 275-286.

‘53 Tao, w. A.; Wollscheid, B.; O’Brien, R.; Eng, J. K. Nature Methods 2005, 2, 591-598.
64 Zhou, H.; Watts, J. D.; Aebersold, R. Nature Biotechnology 2001, 19, 375-378.

65 Porath, J .; Carlsson, J .; Olsson, I.; Belfrage, G. Nature 1975, 258, 598-599.

66 Hochuli, E.; Dobeli, H.; Schacher, A. Journal of Chromatography 1987, 411, 177-184.

67 Ficarro, S. B.; McCleland, M. L.; Stukenberg, P. T.; Burke, D. J.; Ross, M. M.;
Shabanowitz, J .; Hunt, D. F.; White, F. M. Nature Biotechnology 2002, 20, 301-305.

68 Hutchins, T. W.; Yip, T. Rapid Communications in Mass Spectrometry 1993, 7, 576-
580.

69 Brockmann, A. H.; Orlando, R. Analytical Chemistry 1995, 67, 4581-4585.

7" Shen, J. W.; Ahmed, T.; Vogt, A.; Wang, J. Y.; Severin, J.; Smith, R.; Dorwin, 5.;
Johnson, R.; Harlan, J .; Holzman, T. Analytical Biochemistry 2005, 345, 258-269.

7‘ Dunn, J. D.; lgrisan, E. A.; Palumbo, A. M.; Reid, G. E.; Bruening, M. L. Analytical
Chemistry 2008, 80, 5727-5735.

42

 

72 Larsen, M. R.; Thingholm, T. E.; Jensen, 0. N.; Roepstorff, P.; Jorgensen, T. J. D.
Molecular & Cellular Proteomics 2005, 4, 873-886.

73 Pinske, M. W. H.; Uitto, P. M.; Hilhorst, M. J.; Ooms, B.; Heck, A. J. R. Analytical
Chemistry 2004, 76, 3935-3943.

74 Sano, A.; Nakamura, H. Analytical Sciences 2004, 20, 565-566.

75 Sano, A.; Nakamura, H. Analytical Sciences 2004, 20, 861-864.

76 Blackwell, J. A.; Carr, P. W. Journal of Chromatography 1991, 549, 43-57.
77 Blackwell, J. A.; Carr, P. w. Journal of Chromatography 1991, 549, 59-75.
78 Vassileva, E.; Proinova, 1.; Hadjiivanov, K. Analyst 1996, 121, 607-612.

79 Renger, C.; Kuschel, P.; Kristoffersson, A.; Clauss, B.; Opperrnann, W.; Sigmund, W.
Journal of Ceramic Processing Research 2006, 7, 106-112.

8° Wang, Y. B.; Chen, w. Wu, J. 5.; Guo, Y. L; Xia, x. H. Journal ofAmerican society
of Mass Spectrometry 2007, 18, 1387-1395.

8' Ficarro, s. B.; Parikh, J. R.; Blank, N. C.; Marto J. A. Analytical Chemistry 2008,80,
4606-4613.

82 Wolschin, F.; Wienkoop, 5.; Weckwerth, w. Proteomics 2005, 5, 4389-4397.
83 Lin, H. Y.; Chen, C. T.; Chen, Y. c. Analytical Chemistry 2006, 78, 6873-6878.

8" EkstrElm, 5.; Wallman, L.; Helldin, G.; Nilsson, J.; Marko-Varga, G.; Laurell, T.
Journal of Mass Spectrometry 2007, 42, 1445-1452.

85 Qiao, L.; Roussel, C.; Wan, J .; Yang, P.; Girault, H. H.; Lin, B. Journal of Proteome
Research 2007, 6, 4763-4769.

86 Tan, F.; Zhang, Y.; Wang, J.; Wei, J.; Qin, P.; Cai, Y.; Qian, X. Rapid
Communications in Mass Spectrometry, 2007, 21, 2407-2414.

87 Chen, c. T.; Chen, Y. C. Analytical Chemistry 2005, 77, 5912—5919.

88 Chen, C. T.; Chen, W. Y.; Tsai, P. J.; Chien, K. Y.; Yu, J. S.; Chen, Y. C. Journal of
Proteome Research 2007 , 6, 316-325.

89 Lo, C. Y.; Chen, W. Y.; Chen, C. T.; Chen, Y. C. Journal of Proteome Research 2007,
6, 887-893.

43

 

90 Xu, S. Y.; Zhou, H. J.; Pan, C. S.; Fu, Y.; Zhang, Y.; Li, X.; Ye, M. L.; Zou, H. F.
Rapid Communications in Mass Spectrometry 2006, 20, 1769-1775.

91 Drake, R. R.; Schwegler, E. E.; Malik, G.; Diaz, G.; Block, T.; Mehta, A.; Semmes, 0.
J. Molecular & Cellular Proteomics 2006, 5, 1957-1967.

92 Geng, M.; Zhang, X.; Bina, M.; Regnier, F. Journal of Chromatography B: Biomedical
Sciences and Applications 2001, 752, 293-306.

93 Kaji, H.; Saito, H.; Yamauchi, Y.; Shinkawa, T.; Taoka, M.; Hirabayashi, J .; Kasai, K.;
Takahashi, N.; Isobe, T. Nature Biotechnology 2003, 21, 667-672.

94 Xiong, L.; Regnier, F. E. Journal of Chromatography B: Analytical Technologies in
the Biomedical and Life Sciences 2002, 782, 405-318.

95 Kubota, K.; Sato, Y.; Suzuki, Y.; Goto-Inoue, N.; Toda, T.; Suzuki, M.; Hisanaga, S.;
Suzuki, A.; Endo, T. Analytical Chemistry 2008, 80, 3693-3698.

96 Hu, 5.; Wong, D. T. Proteomics: Clinical Applications 2009, 3, 148-154.
97 Wong, D.; Hu, 5. WIPO Patent WO/2007/082057.

98 Sparbier, K.; Koch, S.; Kessler, 1.; Wenzel, T.; Kostrzewa, M. Journal of Biomolecular
Techniques 2005, 16, 407-413.

99 Sparbier, K.; Wenzel, T.; Kostrzewa, M. Journal of Chromatography B 2006, 840, 29-
36.

‘0” Zhang, H.; Li, x.; Martin, D.; Aebersold, R. Nature Biotechnology 2003, 21, 660-666.

‘01 Maley, F.; Trimble, R. B.; Tarentino, A. L.; Plummer, T. H. Analytical Biochemistry
1989, 180, 195-204.

102 Cao, J.; Shen, C.; Wang, H.; Shen, H.; Chen, Y.; Nie, A.; Yan, G.; Lu, H.; Liu, Y.;
Yang, P. Journal of Proteome Research 2009, 8, 662-672.

‘03 Chen, R.; Jiang, x.; Sun, D.; Han, G.; Wang, F. Ye, M.; Wang, L.; Zou, H. Journal of
Proteome Research 2009, 8, 651-661.

‘04 McDonald, c. A.; Yang, J. Y.; Marathe, v.; Yen, T.; Macher, B. A. Molecular &
Cellular Proteomics 2009, 8, 287-301.

'05 Mirzaei, H.; Baena, B. Barbas, C. Regnier, F. Proteomics 2008, 8, 1516-1527.

44

 

'06 Zhou, L.; Beuerman, R. W.; Chew, A. P.; Koh, S. K.; Cafao, T. A.; Urrets-Zavalia, E.
A.; Urrrets-Zavalia, J. A.; Li, S. F. Y.; Serra, H. M. Journal of Proteome Research 2009,
8, 1992-2003.

‘07 Boeseken, J. Advanced Carbohydrate Chemistry 1949, 4, 189-210.

'08 Sugihara, J. M.; Bowman, C. M. Journal of the American Chemical Society 1958, 80,
2443-2246.

'09 Lorand, J. R; Edwards, J. 0. Journal of organic Chemistry 1959, 24, 769-774.
”0 Yan, J .; Springsteen, G.; Deeter, S.; Wang, B. Tetrahedron 2004, 60, 11205-11209.
I” James, T. D. Topics in Current Chemistry 2007, 277, 107-152.

”2 Zhou, W.; Yao, N.; Yao, G.; Deng, C.; Zhang, X.; Yang, P. Chemical
Communications 2008, 5577-5579.

“3 Yeap, w. 5.; Tan, Y. Y.; Loh, K. P. Analytical Chemistry 2008, 80, 4659-4665.

“4 Xu, Y.; Wu, Z.; Zhang, L.; Lu, H.; Yang, P.; Webley, P.; Zhao, D. Analytical
Chemistry 2009, 81, 503-508.

45

Chapter 2: Metal Oxide-Modified Plates for Analysis of Phosphopeptides by

MALDI-MS

2.1 Introduction
Reversible phosphorylation of proteins within cells of both prokaryotic and

' Although it is just one of

eukaryotic organisms is an essential regulatory mechanism.
hundreds of post-translational modifications (PTMs), protein phosphorylation is
responsible for regulating numerous cellular functions, including membrane transport,
gene expression, and apoptosis.2’3‘4'5'6’7 Abnormal phosphorylation has been identified as
either the cause or consequence of several diseases, including diabetes, muscular
dystrophy, and various forms of cancer.8'9 Therefore, the identification of
phosphorylation sites, as well as the quantification of phosphorylated species, is
necessary to understand these biochemical processes and diseases. Mass spectrometry,
particularly with electrospray ionization (E81) and matrix-assisted laser
desorption/ionization (MALDI), has become the premier tool for the identification and
quantification of phosphorylated proteins and peptides. Both of these forms of mass
spectrometry use soft ionization techniques that allow minimal fragmentation and
therefore the identification of the intact biomolecule.

Despite the advantages of mass spectrometry, the low abundance of
phosphorylated peptides and proteins relative to nonphosphorylated species makes
detection challenging.'0’”’12 However, a number of recently developed techniques that
isolate phosphorylated peptides from nonphosphorylated peptides make detection by

mass spectrometry more feasible. Some of these methods include immobilized metal

affinity chromatography (IMAC), reversible covalent binding, and metal oxide affinity

46

chromatography (MOAC).'3’l4 While IMAC is one of the most commonly used
phosphOpeptide enrichment techniques, acidic amino acid residues, including aspartic
acid and glutamic acid, also have some affinity for IMAC resins, which results in the
non-specific binding of nonphosphorylated species.15 Recently developed metal oxide
resins exhibit strong affinities for phosphorylated species, and under appropriate
conditions, the non-specific binding of acidic residues to metal oxides is minimal. In
most enrichment methods with either metal oxides or IMAC, enrichment occurs in a

microcolumn.16'”’18

Unfortunately, the column-based methods often decrease
throughput, as the sample must be loaded onto the column, rinsed to remove any unbound
species, and then eluted to collect the analyte. The column eluate is finally either
analyzed by ESI-MS or mixed with matrix and spotted on a target and analyzed by
MALDI-MS.

This chapter describes two on-plate enrichment techniques where samples are
spotted on a modified MALDI plate and rinsed to remove impurities prior to addition of
matrix and subsequent analysis. Both methods utilize metal oxide-modified plates for
selective capture of phosphorylated peptides. These on-plate enrichment techniques
allow for essentially no sample loss as the enriched phosphopeptides are analyzed
directly on the modified substrate and do not need to be transferred to a conventional
MALDI target, as is necessary with rnicrocolumns.

The first on-plate enrichment technique utilizes MALDI substrates modified by
the electrostatic layer-by-layer (LbL) adsorption of polyelectrolytes and metal oxides,

either Ti02 or ZrO2 nanoparticles (Figure 2.1). LbL deposition of polyelectrolytes and

nanoparticles was used previously for a variety of applications and is a simple method for

47

'9’20’21 Gold-coated substrates were modified with these

substrate modification.
polyelectrolyte/metal oxide layers, and an unpurified protein digest was added to the
plate and allowed to incubate for 1 h. The positively charged nanoparticles directly
interact with the phosphorylated peptides.

The second method is similar to work by Qiao and coworkers,22 in which an array
of sintered Ti02 nanoparticles is prepared on a modified MALDI target by heating the
TiO2-covered plate to 400 °C. Here, we add 1 to 2 IIL of a T102 suspension to wells of a
modified MALDI plate and heat the plate by ramping from room temperature to 400 °C
in a nitrogen-filled furnace. With these plates, we examine the recovery of a synthetic
phosphorylated peptide from an excess of nonphosphorylated peptide and compare these
results with recoveries from commercially available enrichment materials. The recovery

of this synthetic peptide using the Ti02-modified MALDI plates is comparable to and

even higher than several commercially available IMAC and metal oxide materials.

2.2 Experimental
2.2.1 Materials

Chicken egg ovalbumin, bovine serum albumin (BSA), rabbit phosphorylase b
(phos b), and pig esterase were purchased from Sigma and digested using sequencing
grade modified trypsin from Promega. Other digest reagents include: Tris-HCl
(lnvitrogen), urea (J. T. Baker), 1,4-dithio-DL-threitol (BioChemika), ammonium
bicarbonate (Columbus Chemical), and iodoacetamide (Sigma). Gold-coated silicon
wafers (Silicon Quest International) were prepared by sputter coating the Si wafers with
20 nm of chromium, followed by 200 nm of gold. The sputter coating was performed by

Lance Goddard Associates, Foster City, CA. The reagents used for the fabrication of

48

polyelectrolyte/metal oxide modified plates include 3-mercaptopropionic acid (Aldrich),
poly(allylamine hydrochloride) (Mw 70,000 Da, Sigma), poly(sodium 4-styrene
sulfonate) (Mw 70,000 Da, Sigma), titanium (IV) oxide (particle size, ~100 nm, Aldrich),
and zirconium (IV) oxide (Aldrich). Titanium (IV) oxide used for modifying the MALDI
plate by heating was received from Evonik Degussa (Piscataway, NJ) as a gift. Reagents
for enrichment include trifluoroacetic acid (Aldrich), phosphoric acid (Aldrich), HPLC
grade acetonitrile (EMD), glacial acetic acid (Spectrum) and 2,5-dihydroxybenzoic acid
(Aldrich). Deionized water was obtained using a Millipore purification system (Milli-Q,
18 MQcm). HPLC grade methanol (Sigma), isopropyl alcohol (EMD), and ammonium
hydroxide (Columbus Chemical) were used to clean the conventional stainless steel
MALDI plate according to Therrno Scientific’s deep cleaning procedure in which the
plate was first thoroughly rinsed with isopropanol and methanol, and then sonicated for
30 min in a solution consisting of 451 mL of acetonitrile, 451 mL of water, and 108 mL
of ammonium hydroxide. After sonication, the plate was rinsed with water and methanol,
and then dried under a stream of N2 gas. Finally, HPLC grade hexane (J. T. Baker) was
rubbed over the plate using a cotton swab, followed by rinsing with hexane to create a
hydrophobic surface. The plate was dried under a stream of N2 gas and stored in
nitrogen-filled glove bag while not in use.

For synthesis of H5 and D5 peptides, sequenal grade trifluoroacetic acid (TFA)
was purchased from Pierce (Rockford, IL) and N-a-Fmoc-protected amino acids and
Wang resins that were derivatized with Fmoc-protected amino acides (100-200 mesh)
were obtained from EMD Biosciences (San Diego, CA). DIG—propionic anhydride was

purchased from CDN Isotopes (Pointe-Claire, Quebec, Canada) and propionic anhydride

49

was obtained from Sigma Aldrich (St. Louis, MO). Reagent grade N,N’-
dimethylforrnamide (DMF), purchased from Spectrum Chemicals (Gardena, CA), was

dried with 4-A molecular sieves.

2.2.2 Protein Digestion

For the tryptic digestion of protein samples, twenty 100 pg samples of each
protein were separately dissolved in 20 pL of 6 M urea containing 50 mM tris-HCl. To
measure out 100 pg of a protein, 2 mg of protein was first dissolved in 1 mL of deionized
water, and twenty 100 pL aliquots were prepared and dried separately using a SpeedVac.
For the digestion of a nonphosphorylated protein mixture, 100 pg of each protein, BSA,
phos b, and esterase, were combined in one Eppendorf tube and dissolved in 20 pL of the
urea/tris-HCl solution. To reduce any disulfide linkages, 5 pL of 10 mM 1,4-dithio-DL-
threitol (DTT) was added to each sample, and the protein solutions were heated in a water
bath at ~65°C for 1 h. After cooling the sample to room temperature, 160 pL of 50 mM
ammonium bicarbonate and 10 pL of 100 mM iodoacetamide were added to each protein
solution, and the samples were placed in the dark for 1 h. Finally, 10 pL of 0.5 pg/pL
modified trypsin was added to each sample prior to incubation for ~16 hours at 37 0C.
For the digestion of the nonphosphorylated protein mix, the volumes of ammonium
bicarbonate, iodoacetamide, and trypsin solutions were tripled. The digestion reaction
was finally quenched with the addition of 11 pL of glacial acetic acid. Samples were
dispensed into Eppendorf tubes in 22 pL aliquots and stored in a -70 °C freezer until

further use.

50

2.2.3 Labeled Phosphopeptide Synthesis

The labeled phosphopeptide synthesis protocol is the same as described
previously.23 Briefly, the synthetic peptides, CH3CH2CO—LFTGHPEpSLEK (H5 peptide)
and CD3CD2CO—LF'I‘GHPEpSLEK (D5 peptide), were prepared using manual stepwise
Fmoc-based solid-phase peptide synthesis on Fmoc-Lys(boc)-Wang resins (0.05 mol).
Fmoc amino acids (0.25 mmol) were preactivated by thorough mixing with 0-
(benzotriazol-1-yl)-N,N,N’,N’,—tetramethyluronium tetrafluoroborate (0.25 mol), 1-
hydroxybenzotriazole hydrate (0.25 mmol), and NJV-diisopropylethylamine (0.38 mmol)
in DMF, then coupled with the peptidyl resin for 15 min. For phosphoamino acid
incorporation, Fmoc-Ser(PO(Ole)OH-OH) was used and was coupled as described
above, except a 3-fold excess of NN-diisopropylethylamine was used. Fmoc removal
was carried out using a 30% piperidine solution in DMF for 25 min with shaking. N-
terminal acetylation was achieved after the final Fmoc deprotection step by the addition
of 0.25 mmol of NW-diisopropylethylamine and either 0.17 mmol of dlo-propionic
anhydride or 0.17 mmol of propionic anhydride in DMF in the resin while shaking for 15
min. The side-chain protecting groups and resin were cleaved from the peptide with
2.5% triisopropylsilane and 2.5% water in trifluoroacetic acid for 2 h. The resulting
peptides were precipitated in diethyl ether and redissolved in 25% aqueous acetic acid,
then lypholized, and purified by Reverse Phase-High Performance Liquid
Chromatography (RP-HPLC) using an Aquapore RP-300 column (4.6 mm; Perkin Elmer,
Wellesley, MA) and a linear gradient elution with a flow rate of 1 mL/min from 0-100%
B. Here solvent A was 0.1% TFA aqueous solution and solvent B was 0.089% TFA/60%

acetonitrile in water. This synthesis was conducted by Amanda Palumbo at Michigan

51

State University. The Genomics Technology Support Facility at Michigan State
University determined the concentrations of aqueous stock solutions of the H5 and D5
peptides by amino acid analysis. Aliquots of the stock solutions were concentrated to

100 pmol and stored at -20° C until further use.

2.2.4 Fabrication of Polyelectrolyte/Metal Oxide Films

Gold-coated silicon wafers (1.1 x 2.4 cm) were UV/ozone-cleaned for 15 min and
immersed in a 1 mM solution of 3-mercaptopropionic acid (MPA) in ethanol (0.87 pL
MPA in 10 mL ethanol) for l h to give a self-assembled monolayer (SAM) of MPA on
the gold surface. Wafers were rinsed with deionized water, followed by ethanol and
dried in a stream of N2 gas. The monolayer-modified wafers were then immersed for 2
min in a pH-4.5, 0.02 M poly(allylamine hydrochloride) (PAH, molarity is given with
respect to the repeating unit) solution containing 0.5 M NaCl, (pH was adjusted using
HCl). The films were rinsed with deionized water for l min, then dried in astream of N2
gas. Poly(4-styrene sulfonate) (PSS) was deposited on the films by immersing the wafers
for 2 min in 3 mg/mL PSS containing 0.5 M NaCl (the pH of this solution was adjusted
to 2.2 with HCl). Wafers were rinsed with water for 1 min and dried in a stream of N2
gas. Aqueous suspensions of ZrO2 were prepared with various concentrations, including
0.1 mg/mL, 0.2 mg/mL, and 1 mg/mL. The pH of the ZrO2 suspension was lowered to
~l .5 using HCl 50 that the nanoparticles were positively charged, and the suspension was
sonicated to ensure even distribution of the nanoparticles. Wafers were immersed in the
ZrO2 suspension for up to 20 min (Figure 2.1). They were then washed by placing them
in a dilute HCl solution, pH ~l.5, for 1 min to remove any loosely bound ZrO2. The

wafers were then dried completely in a stream of N2 gas. For thicker films, additional

52

bilayers of PSS/ZrO2 can be deposited. However, only one bilayer of PSS/ZrO2 was
typically employed, as discussed below. Films containing Ti02 were prepared in the
same fashion, except PSS-terminated films were immersed in a 1 mg/mL T102
suspension with a pH of ~1.5 for 20 min. Wafers were again washed in dilute HCl, pH
~1.5 for 1 min and dried in a stream of N2 gas. Wafers were originally cut to fit the
reflectance FTIR spectrophotometer sample holder (1.] x 2.4 cm). To fit in the modified
stainless steel MALDI plate, which was machined to hold standard microscope slides (1.7
x 7.7 x 0.1 cm), the wafers were cut to fit the width of the modified plate (1.7 cm).
Sample wells were created on the modified wafers by lightly scratching circles (2 mm
diameter) onto the wafer using a tungsten carbide-tipped pen. Typically six wells were
scratched per wafer and each well is able to hold 2-3 pL of aqueous solution. Wafers

were attached to the modified MALDI plate using double-sided tape.

53

a)

O

m
CD CD (I)

:= :=o :=0
0,, 0° 0,,

Polycation
Deposition

 

Polycation

Ge

Polyanion

b)
TiOZ or ZrOZ

Nanoparticle

 

 

 

m

a) m m

g: 0%: O: O
00 00 00

Polyanion
Deposition

 

C)

  
 
  

M .

CH2
I e e
NH3C|

m
(D (I) (1)

:=0 :=O :=0
00 O" 0

e o
Poly(allylamine 803 Na

Nanoparticle hydr‘gREride) Poly(sodium
Deposition 4-styrene sulfonate)
PSS

 

 

 

d)

3.4

Figure 2.1: Schematic diagram showing the modification of a gold substrate with a) a
SAM of MPA, followed by the deposition of b) the polycation PAH, c) the polyanion
PSS, and finally (1) the nanoparticles, either T102 or ZrO2. Additional bilayers can be
formed by alternating steps c) and d). Inset shows the structures of the polyelectrolytes
used.

54

2.2.5 Preparation of T102-modified MALDI Plate

A second method of preparation of a T102-modified MALDI plate was adapted
from Qiao, et al.22 In this case, 1 g of T102 nanoparticles was first heated at 300 °C for 2
h and the nanoparticles were separated by grinding them for 2 h using a mortar and
pestle. Throughout these two hours, a total of 1 mL of 10% acetic acid was added to the
nanoparticles to keep them wet. A 100 mg/mL suspension was then prepared in 90%
ethanol and sonicated for 1 h, and a 4 mg/mL suspension was prepared by diluting the
100 mg/mL T102 suspension with water and sonicating for 15 min. Sample wells were
machined into a magnetic plate (1.65 x 7.65 cm) which fits precisely into a stainless steel
MALDI plate holder (this plate is the same size as a standard microscope slide and
attaches to the MALDI plate holder which contains magnets) as shown in Figure 2.2.
Both the plate and the holder were available through Therrno for the vMALDI LTQ XL
mass spectrometer. The wells are 0.23 mm deep and have a diameter of either 2.4 mm or
3.2 mm. Either I or 2 pL of the 4 mg/mL T 102 suspension was spotted into the machined
wells and allowed to dry in the air at room temperature. The plate was heated under
nitrogen in a furnace which was ramped from room temperature to 400 °C in ~40 min and
held at 400 °C for 1 h. The plate was then allowed to cool to room temperature and

stored in a desiccator until further use.

55

Modified
Plate

    

b)

.

 

Figure 2.2: Photograph of a) T102—modified magnetic MALDI plate with machined wells,
and b) the MALDI sample plate holder, which contains magnets.

2.2.6 Enrichment Protocol for Plates Modified with Polyelectrolyte/Metal Oxide
Films

For analysis of protein digests, 1 pL of digest solution was spotted in the 2 mm
diameter wells that were scratched in the polyelectrolyte/metal oxide-modified gold
wafers, as described above. Either 0.1 M acetic acid or 0.1% trifluoroacetic acid (TFA)
was used as the solution from which digests were loaded on the plate. Samples were
incubated for 1 h, and additional loading solution (0.5 or 1 pL without digest) was added
to the wells as the sample solution evaporated throughout the incubation time. After 1 h,
a modified wafer with six sample wells was rinsed with 5—10 mL of 66 mg/mL 2,5-
dihydroxybenzoic acid (DHB) in 80% acetonitrile/0.1% TFA, followed by 5-10 mL of
80% acetonitrile/0. 1% TFA solution, then dried under a stream of N2 gas. After drying, 1

pL of 0.1% TFA was added to each well, followed immediately by 0.25 pL of 40 mg/mL

56

DHB solution (1:4 acetonitrile: 1% phosphoric acid). After crystallization of the DHB
matrix, the wafer was affixed to the modified MALDI target using double-sided tape.

The same protocol was followed for both T102 and ZrO2-modified plates.

2.2.7 Protocol for Enrichment of H5 Peptide Using Plates Modified with T102 by
Heating

Stock solutions containing 100 pmol of either the H5 or D5 synthetic peptides
were prepared in 200 pL of deionized water, and solutions containing 125, 62, 31, 16, 8
fmol/pL were prepared from the original stock solutions by serial dilutions with water. A
calibration curve obtained through conventional MALDI analysis was prepared prior to
enrichment to compare the signals due to the H5 and D5 peptides. The matrix solution
used for conventional analysis was 1 pL of 10 mg/mL DHB solution (1:1 acetonitrile: 1%
H3PO4). For the enrichment of the synthetic H5 phosphopeptide using the plates modified
with T102 (Figure 2.3), 1 pmol of peptide mix, consisting of BSA, phosphorylase b, and
esterase, in 0.1% TFA was spotted in the sample wells, followed immediately by 125
fmol of H5 peptide. The H5 peptide was incubated for 30 min and 0.1% TFA solution
was periodically added as the solvent evaporated throughout the incubation time. The
plate containing four wells was rinsed with 10 mL of 20 mg/mL DHB solution (1:1
acetonitrile: 0.2% TFA), followed by 10 mL of 1:1 acetonitrile: 0.2% TFA solution, then
dried under a stream of N2 gas. Phosphopeptides were desorbed using 1 pL of 1%
phosphoric acid, and 125 fmol of the D5 peptide was added as an internal standard just
prior to addition of 0.25 pL of matrix solution, which contained 40 mg/mL DHB in 1:1
acetonitrile: 1% H3P04). Upon crystallization of the matrix, the plate was easily fixed to

the modified MALDI plate via the magnets in the plate.

57

 

0 H5 Peptide
O Other Peptides
O Digest Reagents

TiO2-modified
a D5 Peptide

 

 

Add 1 pmol

Digest Mixture
_ Add 125 fmol ’
H5 peptide

 
  
 
 
 
 

 

 

Matrix Crystals

Figure 2.3: Schematic of the enrichment process used to determine the recovery of the H5
peptide from a protein digest mixture using a T102 modified plate. The D5 peptide served
as an internal standard.

2.2.8 Characterization and Instrumentation

The polyelectrolyte layers were characterized by reflectance FTIR spectrometry
using a Nicolet Magna 560 spectrophotometer with a Pike grazing angle (80°) accessory.
The thicknesses of the polyelectrolyte layers were determined using a rotating analyzer
spectroscopic ellipsometer (J. A. Woollam, M-44), assumingla film refractive index of
1.5. Scanning Electron Microscopy (SEM) was used to characterize the T102
nanoparticles deposited on PSS-terminated films. All SEM images shown here were
taken by David Dotzauer of Michigan State University using a Hitachi S—4700-II field-
emission electron microscope. Mass spectra were obtained using a MALDI linear ion
trap mass spectrometer (Therrno vMALDI LTQ XL), and tandem MS was carried out
using collision-induced dissociation (CID). For most samples, MS/MS, wideband
MS/MS, and MS3 were used for the identification of phosphopeptides. All spectra were

obtained in positive ion mode.

58

2.3 Results and Discussion
2.3.1 Fabrication and Characterization of Polyelectrolyte/Metal Oxide-Modified
Plates

The fabrication of the polyelectrolyte/metal oxide-modified plates (Figure 2.1)
was characterized using reflectance FTIR spectroscopy, and the film thickness after each
step was determined using ellipsometry. The reflectance FT IR spectra in Figure 2.4
confirm the deposition of the polyelectrolyte layers. The peak near 1730 cm'1 in
spectrum a) is due to the acid carbonyl group of the MPA immobilized on the gold
substrate. Spectrum b) which was taken after adsorption of positively charged PAH on
the negatively charged MPA SAM contains a peak near 1570 cm”, which is probably due
to the primary amine of PAH and the formation of the acid salt in the monolayer.
Finally, spectrum c) in Figure 2.4 confirms the adsorption of the polyanion, PSS, onto the
MPA-PAH film. The two sharp peaks at 1035 cm'1 and 1010 cm'1 are characteristic of
the sulfonate group in PSS, as are the peaks at 1220 and 1174 cm". These MPA-PAH-

PSS films are approximately 30 A thick, as determined by ellipsometry.

59

 

 

sulfonate
\
0.0010
8 C)
I: .
(U
'9 primary amine
8 b) \I
.D
<
acid carbonyl \
a)

 

 

4000 3500 3000 2500 2000 1500 1000
Wavenumbers (cm'1)

Figure 2.4: Reflectance FTIR spectra of a) a self-assembled monolayer of MPA on a gold
substrate, and the same film after b) adsorption of the polycation, PAH, and finally c)
adsorption of the polyanion, PSS.

Although reflectance FT IR spectroscopy and ellipsometry were used to
characterize and confirm the growth of each of the polyelectrolyte layers, the deposition
of either T102 or ZrO2 onto PSS could not be confirmed using these analytical tools.
These metal oxides interfered with the IR signal, as well as the ellipsometric data, most
likely due to a high surface roughness. Therefore the deposition of nanoparticles was
examined using SEM. The image in Figure 2.5a shows a widespread distribution of T102
nanoparticles on a PSS-terminated film, but the nanoparticles have aggregated (Figure

2.5b). These TiO2 nanoparticle aggregates are approximately 2-3 pm in diameter.

60

Despite this aggregation, the metal oxide nanoparticles were still able to interact with and

enrich the phosphopeptides from the protein digest, as discussed below.

 

Figure 2.5: SEM images of a) T102 adsorbed on an MPA-PAH-PSS modified gold
substrate and b) a magnified image of the same film. T102 was adsorbed from a 1 mg/mL
suspensron. '
2.3.2 Analysis of Ovalbumin Digests Using MALDI-MS

Initial studies of the ability of polyelectrolyte/metal oxide-modified MALDI
targets to enrich phosphopeptides focused on tryptic digests of ovalbumin, which has two
phosphorylation sites and contains a disulfide bond (Figure 2.7). DTT and iodoacetamide
were added to the ovalbumin digests for the cleavage of the disulfide bond and the
carbamidomethylation of the resulting cysteine residues, respectively. Conventional
MALDI-MS analysis of an ovalbumin digest (Figure 2.63) shows the presence of three
monophosphorylated peptides, which have [‘M+H]+ m/z values of 2089
(EVVGpSAEAGVDAASVSEEFR), 2512 (LPGFGDpSIEAQCGTSVNVHSSLR), and
2903 (FDKLPGFGDpSIEAQCGTSVNVHSSLR). This third phosphopeptide at 2903 is
the result of a miscleavage at lysine (K62). Although the conventional mass spectrum

shows signals due to all three phosphorylated peptides, the use of ZrO2-modified gold
61

plates to selectively enrich phosphopeptides greatly simplifies the mass spectrum by
essentially eliminating signals due to nonphosphorylated peptides (Figure 2.6b). After
enrichment on a gold plate modified with a PAH/PSS/T 102 film, the monophosphorylated
peptide at m/z 2089 gives the dominant signal in the mass spectrum. However, signals
due to the phosphorylated peptides at m/z 2512 and‘2903 are much smaller compared
with that at 2089; the signal at m/z 2903 is especially difficult to identify. This unusually
high signal of the peptide with m/z 2089 may be due to the additional acidic residues,
glutamic (E) and aspartic (D) acid, that enhance enrichment. This phosphopeptide
contains five acidic residues, whereas the phosphopeptide with m/z 2512 contains two
acidic residues and the third phosphopeptide with m/z 2903 contains three. The peptide
with m/z 2089 may also ionize particularly well, as suggested by its relatively strong
signal in the conventional MALDI mass spectrum. While there is minimal non-specific
binding of nonphosphorylated peptides, the signal at m/z 1774 is likely due to the
nonphosphorylated peptide ISQAVHAAHAEINEAGR, which contains two histidine and
two glutamic acid residues. One study noted that these amino acid residues may have
some affinity for metal oxides.17 Glutamic acid, aspartic acid, and cysteine residues can
be derivatized by O—methyl esterification to minimize non-specific binding,15 but this
additional step would take away from the advantages of rapid analysis using these
modified plates. Incomplete esterification also complicates mass spectra. There is also a
peak present at m/z 2132, which is 43 m/z units higher than the signal at 2089. This peak
may be due to the carbamylation of the phosphopeptide at m/z 2089, which can result
from heating protein digests containing urea. The peaks at m/z 697 and 1045 could not

be identified, and neither peak was present in the conventional MALDI mass spectrum.

62

 

 

 

 

 

 

 

6‘) 100%=16200
*2089
3 *2512
'7, l *2903
C
2 /
E
g _
-— b o =
E ) *2089 1004. 22400
G)
o:
*
697 2132 2512 *2903
1045 1774 / /
A . I: .l [I . .. 1.: .L 'I. I I
500 1000 1500 2000 2500 3000 3500 4000

Figure 2.6: Positive ion MALDI mass spectra of 2 pmol of ovalbumin digest analyzed
using a) conventional MALDI-MS, and b) a ZrOz-PSS-PAH-MPA-modified gold plate
with 0.1% TFA loading solution, rinsing with 66 mg/mL DHB (80% ACN/0.1% TFA)
solution and 80% ACN/0.1% TFA solution, and addition of matrix prior to MALDI-MS.

m/z

Asterisks (*) represent phosphorylated peptides.

63

 

1 MGSIGAASME FCFDVFKELK VHI-IANENIFY CPIAIMSALA MVYLGAKDST RTQINKWRF 6O
61 DKLPGFGDpSI EAQCGTSVNV HSSLRDILNQ ITKPNDVYSF SLASRLYAEE RYPILPEYLQ 120
121 CVKELYRGGL EPINFQTAAD QARELINSWV ESQTNGIIRN VLQPSSVDSQ TAMVLVNAIV 180
181 FKGLWEKAFK DEDTQAMPFR VTEQESKPVQ MMYQIGLFRV ASMASEKMKI LELPFASGTM 240
241 SMLVLLPDEV SGLEQLESII NFEKLTEWTS SNVMEERKIK VYLPRMKMEE KYNLTSVLMA 300
301 MGITDVFSSS ANLSGISSAE SLKISQAVHA AHAEINEAGR EWGpSAEAGV DAASVSEEFR 360
361 ADHPFLFCIK HIATNAVLFF GRCVSP

Figure 2.7: Amino acid sequence for chicken egg ovalbumin (Swiss-Pro: P01012).
Phosphorylation sites are designated by bold and italic type and a (p) label. Tryptic
cleavage sites are labeled in bold.

Tandem mass spectrometry (MS/MS) is a useful tool for accurate identification of
phosphopeptides present in mass spectra. Without MS/MS, it is possible to incorrectly
assign m/z values to phosphorylated species, especially at larger m/z values. All MS/MS
analysis was conducted using a vMALDI linear ion trap mass spectrometer with low
energy collision-induced dissociation (CID). In Figure 2.8 the precursor ion selected was
the monophosphorylated peptide EVVGpSAEAGVDAASVSEEFR, [M+H]+ m/z 2089.
The loss of water (18 Da) from this precursor ion, [M+H-H20]+, yields the dominant
peak in Figure 2.8a. Also present, although with much lesser intensity, is the signal due
to loss of 98 Da from the precursor ion, at m/z 1991, and may be due to the loss of either
phosphoric acid, [M+H-H3PO4]+, or the loss of water and monophosphate, [M+H—H20-
HPO3]+.24 The signal at m/z 1973 is due to loss of both water and phosphoric acid (116
Da), [M+H-H20-H3PO4]+. To confirm the identification of the [M+H-H20-H3PO4]+ peak
at m/z 1973, MS3 was used to identify the amino acid residues present in this peak
(Figure 2.8b). In general, bng) and yn type ions are most commonly formed when a
protonated peptide is fragmented using CID. Signals from several yn type ions were
present in the MS3 spectrum in Figure 2.7b, confirming that the peak at m/z 2089 is the

phosphorylated peptide.

64

 

 

 

 

 

 

 

 

 

 

a) 2071
[M+H-H20]+ __,
;:;:; Igggim ;>~°;;I>m;;
IIGSIIIWAI IIII
Q)
E
Q)
.2
E
Q)
at
[M+H-H3Po,]+ [M+H].
[M+H-H20-H3PO4]+ \
- .1973 \n .
600 1000 1400 1800 2200
m/z
b) Y9
3‘
E
OJ
E
G)
.2
13
g V9‘H20
\ y13 1973
Y
Y5 Y5 Vs ylo 11 V14 V15 Y15 V17 y13
t l H “2 t l
600 950 1300 1650 2000
m/z

Figure 2.8: Multistage tandem mass spectrometry of a monophosphorylated peptide ion,
m/z 2089, isolated from 2 pmol of ovalbumin digest enriched using a ZrOz-PSS- PAH-
MPA- modified gold plate. a) CID MS/MS of the m/z 2089 isolated [M+H] precursor
ion, and b) CID MS3 of m/z 1973 isolated from the MS/MS product lon spectrum of m/z

2089 from spectrum a). In b) several yn type ions were identified.

65

2.3.3 Calibration of H5 and D5 Signals for Recovery Analysis

A calibration curve for the synthetic H5 and D5 phosphopeptides, where the D5
peptide serves as an internal standard, allowed quantitation of the enrichment efficiency
for the H5 peptide. Since these synthetic peptides differ only in the deuterated label on
the D5 peptide (Figure 2.9), an equimolar mixture of the H5 and D5 peptides should result
in approximately equal ion intensities in the MALDI mass spectrum. The calibration
curve was created by varying the amount of H5 peptide (125, 62, 31, 16, and 8 fmol) in
samples containing 125 fmol of D5 peptide. These H5 amounts were chosen in the
MALDI-MS linear dynamic range for these synthetic phosphorylated peptides. Figure
2.10 shows the ratio of the peak intensities of the H5 and D5 peptides (1H5/Ips) versus the
amount of H5 peptide in the sample. The plot is linear, and the desorption and ionization
efficiencies of the two peptides are similar, although signals from the H5 peptide are as

much as 20% less than expected for equal sensitivities for the two peptides.

O

O .
H C k D3C /u\
3 \fi LFTGHPEpSLEK \8 LFTGHPEpSLEK
2

2
3) H5 peptide (m/z 1393.6) b) D5 Peptide (m/2 1393-5)

Figure 2.9: Sequence and m/z values of labeled synthetic phosphorylated peptides, a) H5
peptide and b) D5 peptide.

66

 

 

 

 

 

1.0
y = 0.0065x + 0.011
0.3 -
R2=0.9994

m 0.5 -
_O
E
- 0.4 -

0.2 -

0.0 I I T I

o 25 50 75 100 125

Amount H5 (fmol)

Figure 2.10: Calibration curve comparing the ratio of peak intensities (lug/105) of
synthetic phosphopeptides, H5 and D5, from MALDI—MS spectra as a function of the
amount of H5 peptide. 125 fmol of D5 peptide was present in each sample as an internal
standard.

2.3.4 Enrichment of H5 Peptide from Peptide Mixtures Using Plate Modified by
Heating of Nanoparticles.

Initial studies of the ability of the T i02-modified plates to enrich phosphopeptides
examined the recovery of the H5 peptide from a mixture containing digested BSA,
esterase, and phosphorylase b. These studies utilized magnetic plates modified by simple
heating of Ti02 nanoparticles. One pmol of the digest mixture in 0.1% TFA was spotted
within a TiOz-modified well, followed by addition of 125 fmol of H5 peptide in 0.1%
TFA, and the mixture was incubated for 30 min. After rinsing with 10 mL of 20 mg/mL
DHB (1/1 ACN/0.1% TFA), followed by 10 mL of 1/1 ACN/0.1% TFA, the plate was

dried under N2 gas. To desorb the H5 phosphopeptide, 1 uL of 1% phosphoric acid was

67

added to the well, followed immediately by 125 fmol of the D5 peptide as an internal
standard. Finally, 0.25 11L of matrix solution (40 mg/mL DHB in 1/1 ACN/ 1% H3PO4)
was added to the well. Figure 2.11a shows the mass spectrum of the H5 peptide that was
enriched from the peptide mixture, and Figure 2.11b shows an expanded region around

the m/z values of the H5 and D5 peptides, 1393.6 and 1398.6, respectively.

68

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

05 peptide 100% = 15500

>.

.t’

m -

QC) H5 peptide

4.:

E

G)

.2

H
L9

G)

a:

1000 1250 1500 1750 2000
m/z
D5 peptide

>

.4: Recovery

3 69 i 7%

0)

*5 H5 peptide

(D

>

'4:

L5

(D

I

1390 1395 1400 1405

m/z

Figure 2.11: a) Positive ion MALDI mass spectrum of 125 fmol of H5 peptide enriched
from 1 pmol of a protein digest using a T102-modified plate, and b) an enlarged region of
the spectrum around peaks due to H5 and D5 peptides. 125 fmol of D5 was added as an
internal standard. The recovery of H5 peptide from the mixture is 69 :1: 7%. The
modified plate was prepared by heating TiOz nanoparticles on a magnetic plate.

69

Although the full MALDI mass spectrum (Figure 2.1 la) has a significant amount
of noise, the recovery of the synthetic phosphopeptide, H5, from the peptide mix
containing digested BSA, esterase, and phosphorylase b, is 69 i 7%, as determined from
the ratio of the H5 and D5 signals and the calibration curve in Figure 2.9. Much of the
noise present in Figure 2.11a is likely due to impurities found on the plate, due to the
machining process. The plate was cleaned prior to modifying with Ti02; however, this
may not have removed all of the impurities from the machining process. Nevertheless,
the 70% recovery is comparable to and even higher than for several commercially
available IMAC and metal oxide materials, as described in Table 2.1. When analyzing
the recovery of 125 fmol of this same H5 peptide from 1 pmol of the same peptide
mixture using the ZipTipMC pipette tips containing Fe (HI) complexes, our group
previously determined that this commercially available IMAC material has a recovery of
only 12 :t 2%. Pipet tips with TiOz or ZrOz embedded into the tip walls (Glygen)
exhibited recoveries of 22 i 8% and 68 -_1- 5% for the ZrOz and TiOz-containing tips,

respectively.23

Table 2.1: Comparison of the ability of several commercially available enrichment
materials and the TiOz-modified plates to recover 125 fmol of H5 peptide from 1 pmol of
nonphosphorylated digest mixture.23

 

 

 

 

 

 

 

 

 

 

 

. . Percent (%) Recove of H5
Enrichment Material Pguide from Digest rMixture
Fe(III)-NTA Monolayer 9 i 2

Millipore ZipTipsMc 12 i- 2
Qiagen IMAC Chip 13 -I_- 3
Glygen Zr02 N uTips 22 i 8
Glygen Ti02 NuTips 68 i 5
TiOz Plates by Heating 69 :l: 7
Fe(III)—NTA-PHEMA 73 i 12

 

7O

2.4 Conclusions

Both types of metal oxide-modified plates allow separation and detection of
phosphopeptides from protein digests without any additional purification prior to
enrichment. Specifically, the PAH/PSS/ZrOZ plates significantly improved the signal due
to the monophosphorylated peptide with m/z 2089 from an ovalbumin digest. The other
two phosphopeptides in the digest can also be detected using this enrichment method, but
their signals are low. The use of this modified plate removed virtually all signal due to
nonphosphorylated peptides initially present in the protein digest. While these
PAH/PSS/ZrOz-modified plates selectively captured phosphopeptides from an ovalbumin
digest, the ZrOz nanoparticles aggregated on the surface (as seen by the SEM images),
and it was often difficult to create a uniform surface. The second on-plate enrichment
technique employing an array of TiOz nanoparticles on a MALDI plate by simple heating
resulted in 69 i 7% recovery of the synthetic H5 peptide, which is comparable to, or even
significantly better than, the recovery observed from various commercially available
resins. Despite the reasonably high recovery using these plates, there was a significant
amount of noise present in the mass spectra due to the inability to completely clean the
surface of the plate. However, both of these modified plates allow for rapid, simple on-
plate enrichment techniques, which allow the detection and analysis of phosphopeptides

from protein digests by MALDI—MS.

71

2.5 References

1 Nelson, D. L.; Cox, M. M. Lehninger Principles of Biochemistry. W. H. Freeman and
Company: New York, NY, 2005.

2 Hunter, T. Cell. 2001 100, 113-127.
3 Duronio, v. Biochemical Journal 2008, 415, 333-344.

4 Palfrey, H. C.; Alper, S. L.; Greengard, P. Journal of Experimental Biology 1980, 89,
103-115. ‘

5 Sombroek, 1).; Hofmann, T. G. Cell Death and Differentiation 2009, 16, 187-194.
6 Stamm, S. Journal of Biological Chemistry 2008, 283, 1223-1227.

7 Turner, K. M.; Burgoyne, R. D.; Morgan, A. Trends in Neurosciences 1999, 22, 459-
464.

8 Cohen, P. European Journal of Biochemistry 2001, 268, 5001-5010.

9 Yu, L.; Issaq, H. J .; Veenstra, T. D. Proteomic Clinical Applications 2007, 1, 1042-
1057.

‘0 Kratzer, R.; Eckerstrom, C.; Karas, M.; Lottspeich, F. Electrophoresis 1998, 19, 1910-
1919.

“ Stensbalk, A.; Anderson, 3.; Jensen, o. N. Proteomics 2001, 1, 207-222.
'2 Zeller, M.; Koenig, S. Analytical and Bioanalytical Chemistry 2004, 378, 898-909.

'3 Dunn, J. 1).; Reid, G. E.; Bruening, M. L. Mass Spectrometry Reviews 2009, nor
10.1002/mas.20219.

‘4 Thingholm, T. E.; Jensen, 0. N .; Larsen, M. R. Proteomics 2009, 9, 1451-1468.

15 Ficarro, S. B.; McCleland, M. L.; Stukenberg, P. T.; Burke, D. J.; Ross, M. M.;
Shabanowitz, J .; Hunt, D. F.; White, F. M. Nature Biotechnology 2002, 20, 301-305.

‘6 Larsen, M. R.; Thingholm, T. E.; Jensen, 0. N.; Roepstorff, P.; Jorgensen, T. J. D.
Molecular & Cellular Proteomics 2005, 4, 873-886.

‘7 Pinske, M. w. H.; Uitto, P. M.; Hilhorst, M. J.; Ooms, 13.; Heck, A. J. R. Analytical
Chemistry 2004, 76, 3935-3943.

72

 

‘8 Wilson-Grady, J. T.; Villen, J .; Gygi, S. P. Journal of Proteome Research 2008, 7,
1088-1097.

'9 Dotzauer, D; Dai, J .; Sun, L.; Bruening, M. L. Nanoletters 2006, 6, 2268-2272.

20 Kommireddy, D. S.; Patel, A. A.; Shutava, T. G.; Mills, D. K.; Lvov, T. M. Journal of
Nanoscience and Nanotechnology 2005, 5, 1081-1087.

2' Burghard, Z.; Tucic, A.; Jeurgens, L. P. H.; Hoffmann, R. C.; Bill, J .; Aldinger, F.
Advanced Materials 2007, 19, 970-974.

22 Qiao, L.; Roussel, C.; Wan, J .; Yang, P; Girault, H. H.; Liu, B. Journal of Proteome
Research 2007 , 6, 4763-4769.

23 Dunn, J. D.; lgrisan, E. A.; Palumbo, A. M.; Reid, G. E.; Bruening, M. L.; Analytical
Chemistry 2008, 80, 5727-5735.

24 Palumbo, A. M.; Reid, G. E. Analytical Chemistry 2008, 80, 9735-9747.

73

Chapter Three: Binding of Glutathione-S-Transferase to Glutathione-

Functionalized Polymer Brushes for Protein Enrichment

3.1 Introduction

Glutathione-S-Transferase (GST) is a detoxification enzyme that defends the cell
against reactive oxygen species.“3 This enzyme can also serve as a protein tag in fusion
proteins,4 which are created by joining two or more peptides or proteins together. In this
case, a peptide or protein is fused to the C-terminus of GST, creating the recombinant
GST-tagged protein. These fusion proteins are used in a variety of areas of research
including vaccine development, immunodetecion, and analysis of protein-protein
interactions.5‘6’7 Before any vaccination or immunological studies can be conducted, the
GST-tagged protein must be purified from contaminants, including bacterial proteins.
One domain of GST has a strong affinity for glutathione (GSH), specifically the y-
glutamic acid residue of glutathione.8 Therefore, affinity chromatography using
glutathione is one of the most commonly used purification techniques for GST-tagged
proteins. Glutathione can be covalently immobilized on beads, such as agarose, and

9,10,11,1213 After loading of the desired

packed into a column to bind GST-tagged proteins.
tagged protein, the beads are washed to remove any unbound species and contaminants.
To elute the bound GST-tagged proteins, the column can be washed with a low pH buffer
or free glutathione, which competitively displaces the GST-tagged proteins from the
immobilized glutathione. The proteins can also be cleaved from the GST-tagged using a
protease and then removed from the column. ”"5

While GST-tagged proteins are typically purified using glutathione-affinity

chromatography on a column, this method is relatively time consuming because the

74

analyte must be loaded onto the column, rinsed to remove unbound species, eluted to
remove bound species, and then analyzed (Figure 3.1). This chapter discusses attempts to
prepared glutathione-modified MALDI plates modified for rapid, on-probe capture of
GST—tagged proteins. In principle, GST-tagged proteins can be bound to the glutathione

immobilized on the modified-surface and directly analyzed on the plate, increasing

 

 

 

 

 

 

 

 

throughput.
Protein 3’ go
Mixture 03$? W M! GST-tagged Protein
W so Mr Non GST-tagged Protein
5 04“ O Other Contaminants
1v
,-
Resin
Immobilized
with
Glutathione
-
42. Analysis
$351!; _> of GST-tagged
Proteins
Purifed GST-

 

tagged Proteins
Figure 3.1: Schematic diagram showing the general procedure for separation of GST-
tagged proteins using a column. The sample is loaded onto a column packed with
glutathione-modified beads, rinsed to remove any non-GST-tagged proteins and
contaminants, and finally the GST—tagged proteins bound to the resin are eluted and
collected for further analysis. Figure adapted from Dunn et a].16

Reduced glutathione is a tripeptide consisting of glycine, cysteine, and y—glutamic
acid residues, where glutamic acid is bound to cysteine through its carboxyl group
(Figure 3.2). This chapter describes immobilization of reduced glutathione on various

polymer—modified films via reaction of the thiol group on the cysteine residue with

75

poly(acrylic acid) (PAA) and poly(hydroxyethyl methacrylate) (PHEMA) films.
Additionally, initial studies of the binding of free GST to these plates were conducted to

examine the feasibility of potentially binding GST fusion proteins.

 

SH
0 O O

H
)0»
HO . OH

NH;

IZ

 

 

 

Figure 3.2: Structure of reduced L-glutathione, which consists of glycine, cysteine and y-
glutamic acid residues.

3.2 Experimental
3.2.1 Reagents and Materials

Silicon (100) wafers (Silicon Quest International) were sputter coated with 20 nm
of chromium, followed by 200 nm of gold by Lance Goddard Associates (Foster City,
CA). The reagents used for modifying the gold-coated plates include 11-
mercaptoundecanoic acid (Aldrich), ethyl chloroformate (Aldrich), 4-methylmorpholine
(Aldrich), methanesulfonic acid (Mallinckrodt), dichloromethane (Mallinckrodt), 11-
mercaptoundecanol (Aldrich), 2-bromoisobutyryl bromide (Aldrich), triethylamine (Jade
Scientific), 2-hydroxyethyl methacrylate (Aldrich), cupric bromide (Aldrich), cuprous
chloride (Aldrich), 2,2’-bipyridine (Aldrich), succinic anhydride (J .T. Baker), 4-
dimethylaminopyridine (Sigma), N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide
hydrochloride (Sigma), N—hydroxysuccinimde (Aldrich), bromoacetyl chloride (Aldrich),
pyridine (Jade Scientific), N-(2-aminoethyl)maleimide trifluoroacetate (Fluka),
tetrahydrofuran (Mallinckrodt), and reduced L—glutathione (GSH, Sigma). Glutathione-s-

transferase from equine liver (Sigma) was used in studies of binding to substrates

76

modified with immobilized glutathione. Dimethylforrnamide (Spectrum) was dried using
3-A molecular sieves (Spectrum), and deionized water was obtained from a Millipore
purification system (Milli-Q, 18 MQcm). HPLC grade methanol (Sigma), isopropyl
alcohol (EMD), and ammonium hydroxide (Columbus Chemical Industries) were used to
clean the conventional stainless steel MALDI plate according to Therrno Scientific’s
deep cleaning procedure, as described in Chapter 2. The plate was stored in a nitrogen-
filled glove bag while not in use. The matrix used for all MALDI-MS experiments in this
chapter was 1 11L of a 10 mg/mL 2,5-dihydroxybenzoic acid (DHB, Aldrich) in 1:1
HPLC grade acetonitrile (EMD): deionized water. This solution was applied directly to
the protein sample on the conventional stainless steel MALDI plate. Slide-A-Lyzer
Dialysis cassettes (Pierce Biotechnology Inc.) were used for the dialysis of glutathione-s-
transferase. The reagents used for the purification of this protein through dialysis include
glacial acetic acid (Spectrum), sodium acetate (Sigma), sodium phosphate dibasic
(Spectrum), and potassium phosphate monobasic (Spectrum). Trifluoroacetic acid
(Aldrich) was employed for desalting of Glutathione-S-Transferase using ZipTipmg

pipette tips (Millipore).

3.2.2 Fabrication of Glutathione-Functionalized PAA Films

Gold-coated silicon wafers (1.1 x 2.4 cm) were UV/ozone-cleaned for 15 min and
immersed in a 1 mM solution of ll-mercaptoundecanoic acid (MUA) in ethanol (2.2 mg
MUA in 10 mL ethanol) for 1 h to form a self-assembled monolayer (SAM). Wafers
were then rinsed with deionized water, followed by ethanol and dried under a stream of
nitrogen gas. The MUA SAM was activated by immersing the wafers in 100 mM ethyl

chloroformate, 90 mM 4-methylmorpholine in MN -dimethylforrnamide (DMF) (100 uL

77

of ethyl chloroformate and 100 11L of 4-methylmoropholine in 10 mL of DMF) for 10
min until the solution turned yellow. Activated films were rinsed with ethyl acetate and
dried under a stream of N2. Polymer was attached to the films by immersing wafers in a
solution of amino-terminated poly(tert-butyl acrylate) (PTBA) (0.4 g PTBA in 10 mL of
DMF) for 1 h. Amino-terminated PT BA was synthesized as described previously.17
Polymer films were then rinsed with ethanol and dried with nitrogen gas. PT BA films
were hydrolyzed to form poly(acrylic acid) (PAA) by immersing wafers in a solution of
0.15 M methanesulfonic acid in dichloromethane (100 11L of MeSO3H in 10 mL of

CH2C12) for 10 min. Hydrolyzed films were rinsed with ethanol followed by deionized
water and dried under a stream of nitrogen. To create thicker films, a second layer of
PTBA was attached to the film by first activating the initial layer of PAA with ethyl
chloroformate and 4-methylmoropholine in DMF as before. The second layer of polymer
was then attached by immersing activated films in a solution of PTBA in DMF and
hydrolyzing the polymer to form a bilayer of PAA. For glutathione attachment, this
bilayer of PAA was first activated with 50 mM N—(3-dimethylaminopropyl)-N’-
ethylcarbodiimide hydrochloride (EDC, 0.096 g) and 50 mM N-hydroxysuccinimide
(NHS, 0.058 g) in 10 mL water for 30 min. The activated bilayer of PAA was rinsed
with deionized water, followed by ethanol and dried under a stream of N2 gas. Reduced
glutathione was immobilized on the activated films (Figure 3.3) by immersing wafers in
33 mM glutathione (GSH) in deionized water (0.1 g glutathione in 10 mL water) for ~24
h at 37 °C. Wafers were finally rinsed with deionized water and dried under N2. In an
attempt to bind glutathione-S-transferase (GST) to the glutathione-functionalized films

prior to GST purification by dialysis and desalting by ZipTipsTM, 100 uL of 1 mg/mL of

78

an aqueous undialyzed GST solution was dispensed on the polymer brush so that the
entire gold-coated wafer was covered with the GST solution "for l. h. The film was rinsed

with ethanol and dried under a stream of nitrogen gas.

79

immobilization of glutathione.

ml} Mia.

lEthyl Chloroformate
1-4 Methylmorpholine

bgswiim.

NilPTBAJhr
C)
SM ”:TR’NHZ
MeSOgHH
CHZCIOZ
d)
SNL TIER/NW
lNHS
leoc
e)H_S N91“ RTR R/NH2
GSH,pH7 0 Cl)

37°C, 24h 0WD
0 OH
0
” M R NH
I / 2
E—s 5 u If)? NH2
‘ o

O S
NH

HN O

OH

Figure 3.3: Schematic diagram showing a gold-coated substrate after a) formation of an
MUA SAM, b) activation of the MUA SAM, c) attachment of PT BA, (1) hydrolysis of
PTBA to give immobilized PAA, e) activation of PAA with NHS/EDC, and f)
In some cases, the PAA of step d was activated with
ethyl chloroformate, and steps 0, d, and e were repeated prior to immobilization of
glutathione.

80

3.2.3 Fabrication of Glutathione-functionalized PHEMA Brushes Using Succinic
Anhydride and N HS/EDC

Gold-coated silicon wafers (1.1 x 2.4 cm) were UV/ozone-cleaned for 15 min and
immersed in a 1 mM solution of mercaptoundecanol (MUD) in ethanol (4.1 mg MUD in
20 mL ethanol) for ~16 h to form an MUD SAM. Wafers were rinsed with deionized
water followed by ethanol and dried under a stream of N2 gas. Gold wafers (typically 8
at a time) were arranged in a crystallizing dish and placed in a nitrogen-filled glove bag.
Then 0.12 M triethylamine (TEA) (0.33 mL of TEA in 20 mL dry DMF) was added to
the crystallizing dish, followed by 0.1 M 2—bromoisobutyryl bromide (BIBB, 0.25 mL in
20 mL DMF) that was added dropwise with swirling over the duration of 10 min. Wafers
were removed from the solution and rinsed with DMF. After drying in the glove bag for
10 min, they were removed and rinsed with ethyl acetate, deionized water, and ethanol,
and then dried under a stream of N2 gas. Reflectance FTIR was used to identify the
presence of the ester carbonyl peak (~1730 cm") resulting from the attachment of the
initiator to the MUD SAM. Using freeze-pump-thaw cycling, 30 mL of 2-hydroxyethyl
methacrylate (HEMA) and 30 mL of deionized water were degassed in a Schlenk flask.
During the third cycle, 165 mg CuCl, 108 mg CuBr2, and 640 mg 2,2’-bipyridine were
added to the frozen mixture of HEMA and deionized water. The catalyst dissolved in
mixture of HEMA and deionized water as it thawed. The freeze-pump-thaw cycle was
completed, followed by two additional cycles. The flask was transferred to the nitrogen-
filled glove bag, and the HEMA solution was distributed equally among four 20-mL
scintillation vials, each containing two wafers with initiator attached to the MUD SAM.

Wafers were typically immersed in the HEMA solution for 2 h, giving 45-50 nm thick

81

PHEMA films. The PHEMA films were removed from the vials and rinsed with DMF,
deionized water, and acetone and were then characterized using reflectance FTIR
spectroscopy. To convert the hydroxyl groups of the PHEMA brushes to carboxylic acid
groups, the films were then immersed in a 10 mL of DMF containing 0.1 g succinic
anhydride (SA) and 0.2 g 4-dimethylaminopyridine (DMAP) and heated at 55 °C for 3 h.
These films were rinsed with DMF, deionized water, and ethanol and were dried under a
stream of N2 gas. The carboxylic acid groups were then activated using 50 mM N-(3-
dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride (EDC, 0.096 g) and 50 mM
N-hydroxysuccinimide (NHS, 0.058 g) in 10 mL water for 30 min. Activated films were
rinsed with water and ethanol, and then dried under N2 gas. Reduced glutathione was
immobilized onto these polymer brushes (Figure 3.4) by immersing the films in a pH 6.7,
33 mM glutathione solution containing 10 mM Na2HPO4, 0.15 M NaCl, and 1 mM
EDTA (0.1 g glutathione in 10 mL of buffer). Films were incubated for ~16 h at room
temperature and were rinsed with buffer solution, followed by ethanol and dried under
nitrogen. In an attempt to bind glutathione-S-transferase (GST) to the glutathione-
functionalized films prior to GST purification by dialysis or desalting by ZipTipsTM, 100
uL of 1 mg/mL undialyzed GST in phosphate buffer solution (PBS, 80 mM Na2HPO4, 2
mM KH2POe, 140 NaCl, 10 mM KCl, pH 7.4) was dispensed on the polymer brush so
that the entire gold-coated wafer was covered with the GST solution for l h. The wafer

was rinsed with buffer solution, followed by ethanol and dried under a stream of N2 gas.

82

3% Ma.

lBIBB, TEA
lin DMF, 10 min

b)E%—si“\vh?\o’fl\+/B

HEMA H2O
CuCl, CuBr2, bpy
r.t., 2h

55°C 3 h

Cl) Br
n O

O”C\O/\/O\"/\/“\OH

o
01 NHS, EDC
n3

6) 30 min
O”C\ /\/O\"/\/“\O_ NPOH
GSH, pH 6. 7
rt. ~16 h
f) Br
n O NH2
OI/C\O/\/O\"/\/”\S O
O
HN O
O
OH

Figure 3.4: Schematic diagram of a gold-coated substrate after a) formation of an MUD
SAM, b) initiator attachment to the MUD SAM, c) polymerization of HEMA, d)
derivatization of PHEMA with SA, e) activation with NHS, and f) immobilization of
glutathione onto PHEMA-SA films.

83

3.2.4 Binding of Glutathione to Brominated-PHEMA Brushes

PHEMA films were prepared on gold-coated wafers as described above. Instead
of functionalizing these brushes with succinic anhydride and activating them with EDC
and NHS, the films were brominated by immersing films for 1 h under N2 in a 10 mL
N,N’-dimethylformamide solution containing 100 uL of bromoacetyl chloride (BAC) and
100 11L of pyridine. The wafers were rinsed with DMF followed by ethyl acetate and
dried under a stream of N2 gas. Glutathione was bound to the polymer brushes (Figure
3.5) by immersing the brominated films in a pH 7.8, 33 mM aqueous glutathione solution
for ~16 h. Wafers were rinsed with deionized water and dried under a stream of N2 gas.
Finally, 100 uL of 1 mg/mL GST in phosphate buffer solution (80 mM Na2HPO4, 2 mM
KH2PO4, 140 NaCl, 10 mM KCl, pH 7.4) was dispensed on the glutathione-
functionalized polymer brushes, completely covering the gold-coated substrate with the
solution, in an attempt to bind GST to the glutathione-derivatized films prior to GST
purification by dialysis or desalting by ZipTipsmg. This solution was allowed to incubate
for l h, and the wafer was then rinsed with PBS followed by ethanol and dried under

nitrogen gas.

84

b) Br
,C O
O/ \O/\/ \n/\Br

o
GSH, pH 7.8 NH2
~16 h
B o

n
o”C\o/\/O\n/\S/INH
O

HN O
O

OH

Figure 3.5: Schematic diagram of the modification of a gold-coated substrate with a)
PHEMA, b) brominated-PHEMA, and c) glutathione immobilized on the PHEMA film.

3.2.5 Binding of Glutathione to Maleimide-Derivatized PHEMA Brushes

PHEMA brushes were prepared on gold-coated wafers and were derivatized with
succinic anhydride and activated with EDC and NHS, as described above. Activated
PHEMA-SA films were derivatized with a maleimide by immersing the films in a 10
mg/mL solution of N—(2-aminoethyl)maleimide trifluoroacetate in tetrahydrofuran (THF)
(0.05 g maleimide in 5 mL of THF) solution, containing 100 uL of triethylamine (TEA)
for ~21 h. Wafers were sonicated in DMF, then rinsed with DMF followed by ethyl
acetate and dried under a stream of N2 gas. Glutathione was immobilized on the film
(Figure 3.6) by immersing the brushes in 100 mM of reduced glutathione in phosphate
buffer solution (0.3 g glutathione in 10 mL of buffer solution) for 24 h. The PBS

contained 80 mM Na2HPOa, 2 mM KH2PO4, 140 mM NaCl, and 10 mM KCl, at a pH of

85

7.4. Wafers were rinsed with PBS, followed by ethanol and dried with N2. GST solution
was dispensed onto the glutathione-immobilized PHEMA brushes as described above, in
an attempt to bind GST to the film prior to GST purification by dialysis or desalting by
ZipTipsTM. After incubating for 1 h, the wafer was rinsed with buffer solution followed

by ethanol and was dried under a stream of nitrogen gas.

3) Br
W" O 0
C O
0” \0/\/ NO-N
_ O
O

O N O TEA in THF
~12h
b) Br H O
n NH2 0
c o N
0” \O/\/ NuA/ol
O
GSH in PBS
Cl Br lrwt 24h
n
’/C\ New /\/N
O 0 NM
HN O
O
OH

Figure 3.6: Schematic diagram of the modification of a gold-coated substrate after a)
formation of a PHEMA-SA film activated with NHS/EDC, b) attachment of maleimide,
and c) immobilization of glutathione on the PHEMA film.

3.2.6 Dialysis of Glutathione-S-Transferase

A Slide-A-Lyzer Dialysis cassette (Pierce Biotechnology Inc., MWCO 10 kDa)
was used to remove impurities, specifically reduced glutathione, from Glutathione—S-
Transferase according to the procedure provide by Pierce. The presence of free

glutathione could interfere with the binding of GST to the glutathione immobilized on the

86

polymer brushes. Briefly, 500 mL of buffer solution containing 100 mM acetic acid and
100 mM sodium acetate in deionized water (2.85 mL of acetic acid, 0.038 g of sodium
acetate) was prepared. The dialysis cassette was immersed in the buffer for 30 Sec to
hydrate the membrane of the cassette. Using a syringe, 1 mL of deionized water was
carefully injected into the cassette to ensure that the membrane was fully sealed and that
no water leaked out. The water was removed, and 1 mL of 0.5 mg/mL GST in 5% acetic
acid was injected into the cassette. Any air remaining in between the membranes was
removed to maximize the surface area exposed to the sample buffer solutions. A buoy
was slipped onto the Slide-A-Lyzer cassette so that it would float in the buffer solution.
The acetic acid/sodium acetate buffer solution was replaced with fresh buffer solution
every 2 h for the first 6 h, and the protein was dialyzed for ~16 h. A syringe was used to
remove the GST solution from the cassette, being careful not to puncture the membrane
with the syringe. The protein sample was dried down and analyzed by MALDI-MS.

A second GST sample was dialyzed using the Slide-A-Lyzer Dialysis cassette
with a different buffer. solution, which was previously used by Chen and coworkers.18 In
this case, a phosphate buffer solution (PBS) containing 10 mM Na2HPO4, 1.8 mM
KH2PO4, 140 mM NaCl, and 2.7 mM KCl (5.4 g Na2HPO4, 0.49 g KH2P04, 16.4 g NaCl,
0.40 g KC] in 2 L water, pH 7.4) was employed in place of the acetic acid/sodium acetate
buffer. The membrane was hydrated by immersing it in the PBS for 30 sec, and the
absence of leaks was ensured by injecting 1 mL of deionized water into the cassette prior
to sample injection. A syringe was used to inject 1 mL of GST solution into the cassette,
which was then placed in the PBS. The PBS was replaced with fresh solution every 2 h

for the first 6 h and was then replaced with deionized water, which was replaced only

87

once after 2 h. The sample was dialyzed for a total of 20 h. The GST sample was
removed from the cassette using a syringe and was dried down and analyzed by MALDI-

MS.

3.2.7 Purification of Glutathione-S-Transferase Using ZipTipc1s Pipette Tips

The dried down GST sample from the second dialysis (~50 ug) was resuspended
in 167 11L of deionized water. Two 11L of this 0.3 mg/mL GST solution were desalted
using 10 11L ZipTipc13 pipette tips (Millipore). Briefly, the tips were equilibrated by
wetting with acetonitrile twice, followed by aspirating 0.1% trifluoroacetic acid (TFA)
and dispensing the solution two times. The sample was bound to the ZipTip pipette tip
by aspirating and dispensing the GST solution 10 times. The 0.1% TFA wash solution
was aspirated and dispensed three times to remove any unbound species. The protein
was finally eluted by aspirating with 0.1% TFA/50% ACN and dispensing the solution
three times. The eluent was spotted onto a conventional stainless steel MALDI plate, and
l uL of 10 mg/mL 2,5—dihydroxybenzoic acid (DHB) in 1:1 ACN:H2O was added
directly to the sample on the MALDI plate as the matrix. In addition to desalting this
purified GST sample, an unpurified GST sample was also desalted in the same fashion

and was then analyzed by MALDI-MS.

3.2.8 Characterization and Instrumentation

Polymer brushes were characterized by reflectance FTIR spectroscopy using a
Nicolet Magna 560 spectrophotometer with a Pike grazing angle (80°) accessory. The
thicknesses of the functionalized polymer brushes were determined using a rotating
analyzer spectroscopic ellipsometer (J. A. Woollam, M-44), assuming a film refractive

index of 1.5. Mass spectra were obtained using a MALDI linear ion trap mass

88

spectrometer (Therrno vMALDI LTQ XL), and tandem MS was carried out using low
energy collision-induced dissociation (CID). All spectra were obtained in positive ion
mode. The mass spectra in this chapter were collected by Jamie Dunn of Michigan State

University.

3.3 Results and Discussion
3.3.1 Fabrication and Characterization of Glutathione-Functionalized PAA Films
The fabrication of glutathione—derivatized PAA films (Figure 3.3) was
characterized using reflectance FTIR spectroscopy, and film thicknesses after each step
were determined using ellipsometry. The reflectance FTIR spectra in Figure 3.7 confirm
the functionalization of the PAA brushes. The peak at 1730 cm'1 in spectrum a) of Figure
3.7 is due to the acid carbonyl group of the PAA immobilized on the gold-coated
substrate. These PAA films, which are prepared using two PT BA deposition and
hydrolysis steps, have a thickness of approximately 50 A. After activation of PAA with
NHS, the IR spectrum of the film contains succinimide ester peaks at 1790 and 1760 cm'1
(Figure 3.7b). The asymmetric stretch due to succinimide at 1743 cm’1 overlaps with the
carbonyl stretch (1730 cm") of the previously formed acid carbonyl, forming a broad
peak with an absorbance more than double that of the acid carbonyl of PAA. Upon
immobilization of glutathione on the PAA films, a broad peak appears around 3300 cm'1
probably due to the presence of two amine groups in GSH (Figure 3.7c). Additionally,
the very strong shoulder at 1690 cm'1 is due to the amide I band, which partly overlaps
with the carbonyl stretch, resulting in a much broader peak. The second amide band
appears at 1520 cm'l. Figure 3.7d shows the same glutathione-functionalized film after

exposure to GST for 1 h. The absorbance of the amide I band decreases, as does the

89

carbonyl stretch, perhaps indicating that some of the PAA desorbed from the surface.
While it is appears that glutathione was immobilized on the PAA, it is not clear that GST
was bound to the immobilized glutathione. There is no peak present in the GST spectrum

that positively identifies the binding of this protein to the glutathione-functionalized PAA

 

  
    
   

 

 

 

  
 

 

 

 

 

 

films.
I 0.010
W __
. Amidel
Amine Carbonyl stretch / I Band
stretch Amide ll
\ ' ( Band
W J
K Activated
EsterCarbonyl
Succinimide\
Ester
\
w
Macaw
M
3800 3300 2800 2300 1800 1300 800
cm'1

Figure 3.7: Reflectance FTIR spectra of a gold-coated Si wafer after a) deposition of a
PAA film, b) activation of the PAA with NHS, 0) reaction of glutathione with the
activated PAA bilayer, and (1) exposure of the glutathione-containing film to GST. The
initial PAA film was deposited using two activation, PT BA deposition, and hydrolysis
steps.

90

3.3.2 Fabrication and Characterization of Glutathione-Functionalized PHEMA
Brushes

PHEMA brushes grown from immobilized initators are much thicker than grafted
PAA films, so derivatized PHEMA may bind more GST than derivatized PAA. Figure
3.4 shows the procedure for preparing PHEMA films and one method for derivatizing
them with glutathione to bind GST. The reflectance FTIR spectra in Figure 3.8 confirm
the synthesis and derivatization of PHEMA. The strong ester carbonyl absorbance at
1730 cm‘1 (spectrum a) and the hydroxyl stretch at 3650 — 3100 cm'1 (not shown here) are
characteristic of PHEMA films. After reaction of PHEMA with succinic anhydride (SA),
the absorbance of the ester carbonyl doubles due to an additional ester group formed on
each repeating unit of the polymer chain (spectrum b, Figure 3.8). The hydroxyl stretch
disappears (not shown here) as a result of the complete derivatization of the polymer
chains. The reaction with succinic anhydride results in an increase in film thickness from
50 nm to approximately 75 nm. Spectrum c) is consistent with the activation of the
polymer brush with NHS, as it shows the succinimide ester peaks at 1813 and 1784 cm'1
and the asymmetric stretch of succinimide at 1750 cm'l, which overlaps with the carbonyl
stretch. Figure 3.8d presents the spectra of the film after reaction with glutathione. The
peak at about 1580 cm'1 may be due to the amide II band, althouth the absence of the
amide I band suggests that immobilization may not have been successful. The peak at
1580 cm”1 could also be due to deprotonated carboxylic acid groups. Finally, spectrum e)
in Figure 3.8 shows the spectrum of the glutathione-derivatized film after exposure to
“impure” GST. Spectra d) and e) are very similar, and thus it appears that GST did not

bind to the immobilized glutathione.

91

 

 

 

 

0.10
Carbonyl stretch
e)
K Amide II Band
d
\ Activated ester carbonyl
Succinimide
Ester
\\ PHEMA-SA ester carbonyl
‘ C
b)
a) MEMA ester carbonyl
2000 1800 1600 1400 1200 1000 800

cm'1

Figure 3.8: Reflectance FTIR spectra of a) a PHEMA brush immobilized on a gold-
coated substrate, b) functionalization of PHEMA with succinic anhydride, c) activation of
PHEMA-SA with NHS, (1) immobilization of glutathione onto PHEMA-SA, and e)
binding of GST to glutathione-immobilized polymer brushes.
3.3.3 Immobilization of Glutathione on Brominated-PHEMA Brushes

To determine whether the glutathione immobilization method might be limiting
GST binding, we also immobilized glutathione via reaction with PHEMA brushes that
were activated with bromoacetyl chloride (Figure 3.5). When immersed in a solution of
glutathione, the reactive bromine is replaced via the formation of the thiol ether from the
reduced glutathione. This derivatization was again characterized using reflectance FTIR

spectroscopy (Figure 3.9). Spectrum 3) in the figure shows the presence of stretches

typical of PHEMA, and the increase in absorbance of the peaks present at 1270 and 1200

92

cm"l (spectrum b) may be due to the presence of the alkyl bromide. The absence of a
hydroxyl peak in spectrum b suggests that the reaction of PHEMA with the acid chloride
is complete. The peak due to the ester carbonyl stretch shows a high energy shoulder due
to the electronegativity of bromine, which withdraws electrons from the newly formed
carbonyl ester. Attempts to immobilize reduced glutathione on the brominated-PHEMA
film involved immersing the gold-coated substrate in an aqueous glutathione solution.
However, this reaction was not successful as the spectrum of the film did not change

significantly after immersion in the glutathione (compare spectra 3% and 3.9c).

 

| 0.020

 

 

Ester carbonyl

bromide

 

 

b) A

PHEMA hyd roxyl stretch

 

 

 

 

 

/ PHEMA ester carbonyl
W
3800 3300 2800 2300 1800 1300 800
cm'1

Figure 3.9: Reflectance FT IR of a) PHEMA immobilized on a gold-substrate, b) PHEMA
brominated with bromoacetyl chloride, and c) immobilization of glutathione on PHEMA
film.

93

3.3.4 Fabrication and Characterization of Glutathione Immobilized on Maleimide-
PHEMA Brushes

In another method for immobilizing glutathione, PHEMA brushes were first
derivatized with a maleimide (Figure 3.6). For optimal GST binding, not only does
glutathione need to be attached to the modified substrate, it should be immobilized
through its thiol group and not through the amine. Immobilized maleimides have been
used to anchor thiol—terminated molecules in a variety of applications,19’2°‘21 and this
reaction should ensure the covalent attachment of glutathione to the polymer film through
its thiol group. Functionalization of the PHEMA brush with maleimide occurred via
reaction of the succinimidyl ester of PHEMA with N—(2-aminoethyl)maleimide
trifluoroacetate. Comparison of spectra a) and b) in figure 3.10 confirms the attachment
of the maleimide. The peak at 1714 cm"1 is due to the carbonyl stretch of the maleimide
immobilized on a PHEMA-SA film. Figure 3.11 is the subtraction spectrum of PHEMA-
SA (spectrum a in Figure 3.10) from PHEMA-SA-maleimide (spectrum b in Figure
3.10). This subtraction more clearly shows the asymmetric carbonyl stretching of the
maleimide at 1714 cm‘1 and the amine stretch between 3410 and 3240 cm]. The
asymmetric aromatic C-C stretching is present at 1525 cm'l, as well as the symmetric

stretching at 1404 cm’1 due to C-N-C in the maleimide.”

After reaction of glutathione
with maleimide-functionalized brushes (Figure 3.10c), the amine stretch broadens and
amide bands appear near 1660 cm'I and1530 cm".

Unfortunately, the spectrum of the film after exposure to GST (spectrum d in

Figure 3.10) is very similar to the spectrum after immobilization of glutathione.

Moreover the ellipsometric thickness of the film (~90 nm) changed by < 2 nm after

94

exposure to GST. Since GST is a large protein (Mw = 45-50 kDa), the thickness of the
film should have increased more than 20 nm if the protein was actually bound throughout
the polymer brush. Figure 3.12 shows the subtraction of the PHEMA-SA-maleimide-
GSH spectrum (Figure 3.10c) from the PHEMA-SA-maleimide-GSH after exposure to
GST (Figure 3.10d). This subtraction results in a spectrum with negative absorbances

and no clear evidence of GST binding.

 

   
   
   
  
          

 

      

 

  
 

 

 

 

 

0.10
0')
Amide I .

Band Amide II

1 Band
C) \PHEMA-SA-maleimide

Carbonyl Stretching
Maleimide
/C-N-C Stretch
b)
\PHEMA-SA ester carbonyl

2200 2000 1800 1600 1400 1200 1000 800
cm'1

Figure 3.10: Reflectance FTIR spectra of PHEMA-SA a) before and after b)
derivatization of with N-(2-aminoethyl)maleimide trifluoroacetate, c) immobilization of
reduced glutathione through the maleimide, and d) exposure of GST to the glutathione-
functionalized film.

95

 

I 0'020 Asymmetric /

Carbonyl Stretching

Aromatic
C-C

       
 

Symmetric
C-N-C Stretch

 

Amine Stretch

/

 

 

 

 

 

M
3750 3250 2750 2250 1750 1250 750

cm'1

Figure 3.11: Spectrum resulting from subtraction of the PHEMA-SA spectrum (Figure
3.10a) from the PHEMA-SA-maleimide spectrum (Figure 3.10b) showing the
immobilization of the maleimide to the film.

96

 

\ 1770 /

 

 

 

 

 

1404 1045
1713\it
3750 3250 2750 2250 1750 1250 750

cm'1

Figure 3.12: Spectrum resulting from subtraction of the PHEMA-SA-maleimide-GSH
spectrum (Figure 3.10c) from the PHEMA-SA-maleimide-GSH spectrum after exposure
to GST (Figure 3.10d) showing that there is essentially no GST bound to the film.

3.3.5 Attempts to Purify Glutathione-S-Transferase to Increase Binding

According to the manufacturer, the GST used in binding experiments has
glutathione present as an impurity, which may inhibit the binding of GST to the
immobilized glutathione. Both dialysis and ZipTipCtg pipette tips were employed in
efforts to purify the GST and remove unwanted glutathione. However, matrix-assisted
laser desorption/ionization-mass spectrometry (MALDI-MS) shows that these methods
do not completely remove glutathione (Figure 3.13). The GST sample analyzed here was
dialyzed for 16 h in the acetic acid/sodium acetate buffer using a Slide-A-Lyzer cassette,

and the buffer solution was replaced with fresh solution every two hours for the first six

97

hours. Reduced glutathione has an [M + H]+ m/z value of 308.3, and the conventional
MALDI mass spectrum of the Glutathione-S-Transferase sample after dialysis suggests
that glutathione is still present (Figure 3.13). Tandem mass spectrometry (MS/MS)
confirmed the assignment of glutathione to the peak with an m/z value of 308 (Figure
3.14). Reduced glutathione is a tripeptide, containing glycine (G), cysteine (C), and
glutamic acid (E) residues (Figure 3.2). The MS/MS spectrum of this molecule (Figure
3.14) shows the loss of either water or one of these residues, confirming the assignment
of the glutathione to the peak present at [M+H]+ m/z 308. The peak found at [M+H]+ m/z
of 262 (Figure 3.14) which is a loss of 46 W2 units and may be due to the loss of water

and the loss of a carbonyl group (CO).

 

 

 

 

 

 

 

 

 

100% = 39200
100% = 1510
\[GSH + H]+

.5

0')

C
3
5

OJ
.2
4.:

(U
a:
O:

[GSH + H]"
- n. ._. 1 J, I?
250 275 300 325 350 375 400
m/z

Figure 3.13: Positive-ion conventional MALDI mass spectra of 20 pmol of GST that was
purified using dialysis. The signal for the [M+H]+ ion of glutathione (GSH) has an m/z
value of 308. The inset shows an expanded region of the mass spectrum.

98

 

 

 

 

 

 

 

 

SH
1 Ln 0 0 [M+H-H20]+ \5
HO N :1ka 290
1 1° 1 ”“2 l
r l r
Glycine Cystine Y-Glutamic acid
5 (G) (C) (E)
g [M+H—E]+
+5 *" 179
Q)
'5 [M+H—H20 - c0]+
CD
76 262
°‘ [M+H—
[M+H—E—H20]+ +
162 “M
[M+H-G]+ 272
2 3
100 150 200 250 300
m/z

Figure 3.14: CID MS/MS spectrum of glutathione (m/z 308) isolated from 20 pmol of
GST purified by dialysis. Losses of water as well as glycine (G) and glutamic acid (E)
residues are labeled, as well as the loss of a carbonyl group (CO).

Upon discovering that glutathione was still present in the protein sample, a second
GST sample was dialyzed in a phosphate buffer solution using a Slide-A-Lyzer cassette,
as described in Section 3.2.6. The PBS was replaced with fresh solution every two hours
for the first six hours and was then replaced'with deionized water, which was replaced
once after two hours; the GST sample was dialyzed for 20 h in total. This same GST
sample was also purified using ZipTipc18 pipette tips to remove any unwanted impurities.
MALDI-MS was used to analyze the purified GST and to determine whether any
glutathione remained in the sample. Figure 3.15 shows the conventional MALDI mass
spectrum of GST, again revealing the presence of a peak at m/z 308, which suggests that

glutathione is still present in the protein sample. MS/MS was used to confirm the identity

99

of this signal and several product ions generated by CID tandem mass spectrometry are
the same as those identified previously (Figure 3.14). Unfortunately, reduced glutathione

is still present in the GST sample, despite multiple purification methods.

 

 

100% = 18700
100% = 6060

 

 

 

 

 

 

[GSH + H]+

a?

U?

C

B

E

g 300 310 320 330 340 350
‘5 ml:

1T.)

:1:

[GSH + H]+
250 275 300 325 350 375 400

m/z

Figure 3.15: Positive-ion MALDI mass spectra, obtained by conventional analysis, of 12
pmol of GST purified according to the dialysis with phosphate buffer and by adsorption
and elution from ZipTipmg pipette tips. Reduced glutathione is present at m/z 308. The
inset shows an enlarged region of the mass spectrum.

3.3.6 Binding Purified GST to GSH-Maleimide-PHEMA Films

Although reduced glutathione was present in the GST sample after dialysis and
desalting, a wafer with glutathione immobilized on a maleimide-functionalized PHEMA
film was immersed in an aqueous solution of the “purified” GST for 1 h and rinsed with
buffer solution. Figure 3.16 compares the reflectance FI‘ IR spectrum of films exposed to
(a) as-received and the (b) purified GST. The inset in Figure 3.16 shows the subtraction

of these spectra. While there is a difference between the spectra, most of the differences

100

are due to the underlying films. The two spectra were obtained from different wafers,
which had slightly different thicknesses, and the subtraction spectrum largely resembles a
spectrum of PHEMA-maleimide. From these data, the dialysis and desalting of GST do
not appear to have enabled this protein to bind to the glutathione-modified surface. It is
also possible that the GST sample was not eluted from the ZipTipc.g column following

desalting, which would also inhibit its binding.

 

 

 

 

 

 

 

 

 

 

 

0.050
3750 2750 1750 750
cm'1
W
a)
3750 3250 2750 2250 1750 1250 750
cm'1

Figure 3.16: Reflectance FTIR spectra of PHEMA-maleimide-glutathione films after
immersion in a) as received GST and b) GST purified by dialysis with PBS and desalting
with ZipTipsTM. The inset shows the subtraction of these spectra.

3.4 Conclusions
Reduced glutathione was immobilized on both poly(acrylic acid) and

poly(hydroxyethyl methacrylate) films. While several methods of attaching glutathione

101

to these polymer brushes were explored, PHEMA films derivatized with a maleimide are
perhaps most promising because glutathione should be immobilized through the thiol
group of the cysteine residue, instead of through the amine group. Immobilization
through the amine group of glutathione may decrease the ability to bind GST.
Unfortunately, the question of whether immobilized glutathione binds Glutathione-S-
Transferase is still open. Glutathione is present as an impurity in the GST sample even
after dialysis and desalting using ZipTipag pipette tips. The presence of glutathione in
the protein may inhibit binding of GST to the immobilized glutathione. The reflectance
FTIR spectra of films exposed to GST show no signs of protein binding. Perhaps GST
was not eluted from the ZipTipcrg packing material, also preventing binding to the film.
The GST sample analyzed here (Sigma) consists of only 60% protein. Other
Glutathione-S-Transferases have a significantly higher purity, which may allow more
' efficient binding to immobilized glutathione. Further studies of GST binding in the
absence of glutathione are needed to show whether immobilized glutathione could be

useful in binding GST-tagged proteins.

102

3.5 References

‘ Mezzetti, A. 1).; Ilio, C.; Calafiore, A. M.; Aceto, A.; Marzio, L.; Frederici, G.;
Cuccurullo, F. Journal of Molecular and Cellular Cardiology, 1990, 22, 935-938.

2 Oesch, F.; Gath, I.; Igarashi, T.; Glatt, H.; Thomas, H. NATO ASI Series, Series A:
Lifesciences, 1991, 202, 447-461.

3 Rietjens, I. M.; Lemmink, H. H.; Alink, G. M.; van Bladeren, P. J. Chemico-Biological
Interactions, 1987, 62, 3-14.

4 Kaplan, W.; Husler, P.; Klump, H.; Erhardt, J.; Sluis—Cremer, N.; Dirr, H. Protein
Science, 1997, 6, 399-406.

5 Kursula, I.; Heape, A. M.; Kursula, P. Protein and Peptide Letters, 2005, 12, 709-712.
6 Smith, D. B. Methods in Enzymology, 2000, 326, 254-270.

7 Smyth, D. R.; Mrozkiewicz, M. K.; McGrath, W. J.; Listwan, P.; Kobe, B. Protein
Science, 2003, 12, 1313-1322.

8 Wilce, M. C. J .; Parker, M. W. Biochimica et‘ Biophysica Acta, 1994, 1205, l-18.

9 Lipin, D. I.; Lua, L. H. L.; Middelberg, A. P. J. Journal of Chromatography A, 2008,
1190, 204-214. .

10 Murray, A. M.; Kelly, C. D.; Nussey, S. S.; Johnstone, A. P. Journal of Immunological
Methods, 1998, 218, 133-139.

11 Sadilkova, L.; Osicka, R.; Sulc, M.; Linhartova, I.; Novak, P.; Sebo, P. Protein
Science, 2008, 17, 1834-1843.

‘2 Scheich, c.; Sievert, v.; Biissow, K. BMc Biotechnology, 2003, 3, 1-8.
‘3 Smith, D. B.; Johnson, K. s. Gene, 1988,67, 31-40.
‘4 Simons, P. c.; Vander Jagt, D. L. Analytical Biochemistry, 1977, 82, 334-341.

'5 Thermo Fisher Scientific Inc. “Reusable Glutathione Agarose Resin for Purifying GST
Fusion Proteins,” 2009, http://www.piercenet.com/products/

‘6 Dunn, J. D.; Reid, G. E.; Bruening, M. L. Mass Spectrometry Reviews 2009, DOI
10.1002/mas.20219.

'7 Bruening, M. L.; Zhou, Y. F.; Aguilar, G.; Agee, R.; Bergbreiter, D. E.; Crooks, R. M.
Langmuir, 1997, 13, 770-778.

103

 

'3 Chen, H.; Luo, 8.; Chen, R.; Lii, c. Journal of Chromatography A, 1999, 852, 151-
159.

‘9 Houseman, B. T.; Gawalt, E. s.; Mrksich, M. Langmuir, 2003, 19, 1522-1531.
20 Jin, L.; Horgan, A.; Levicky, R. Langmuir, 2003, 19, 6968-6975.

2' Xia, B.; Xiao, S.; Guo, D.; Wang, J .; Chao, J .; Liu, H.; Pei, J .; Chen, Y.; Tang, Y.; Liu,
J. Journal of Materials Chemistry, 2006, 16, 570-578.

104

Chapter Four: Enrichment of Glycopeptides Using Aminophenylboronic Acid-

Derivatized Polymer Brushes

4.1 Introduction

Protein glycosylation is one of the most common, as well as complex, post-
translational modifications (PTMs) in the cell. The core oligosaccharides of the glycan
moiety are attached to the polypeptide chain and processed by enzymatic reactions in the
endoplasmic reticulum (ER), and the glycoprotein is subsequently transported to the
Golgi complex for further trimming and processing."2 The carbohydrate moiety of the
glycoprotein aids in the folding and conformational stability of the protein, and the entire
glycoprotein plays a key role in protein recognition, the immune system, and cell-to-cell

l'2'3‘4‘5 Unlike the polypeptide chain, whose composition is controlled

recognition.
genetically, the carbohydrate group on a glycoprotein is controlled by enzymatic
reactions. Thus, the glycan moiety may change if the enzyme or reactant concentration
varies, and these glycan variants sometimes alter protein activity and function, which can
lead to disease.“2 In fact, abnormal glycosylation of proteins is either the cause or

consequence of numerous hereditary and acquired diseasesf”7

Abnormal glycosylation
patterns were first associated with cancer in the late 1970s,8 and subsequent studies
showed that changes in glycosylation and in levels of glycosylated proteins play a role in
liver disease, diabetes, and various forms of cancerf’m’n‘l2 Therefore, the identification
of glycosylation sites, as well as the quantification of glycosylated species, is necessary
to understand and identify biochemical processes and diseases.

Due to the development of electrospray ionization (ESI) and matrix-assisted laser

desorption/ionization (MALDI) over the last two decades, mass spectrometry (MS) has

105

become the premier tool for the analysis of carbohydrates and glycoproteins.13’14’15'l‘5’l7

Despite the ability of these soft ionization MS techniques to identify intact biomolecules,
however, the low abundance of glycosylated peptides and proteins makes their detection
challenging.18 Glycosylation sites are also diverse and may contain a range of different
carbohydrates at each site, which also makes their analysis more challenging.
Nevertheless, a number of recently developed techniques that isolate glycosylated
peptides from nonglycosylated species make detection by mass spectrometry more
feasible. The most common method to separate glycoproteins and glycopeptides from a
complex mixture is lectin affinity chromatography. Lectins are proteins, such as
concanavalin A (Con A) and wheat—germ agglutinin (W GA), that have a high affinity for

19’20’2'32 Lectin affinity chromatography separates glycoproteins

specific carbohydrates.
based on their glycan moiety, and the lectins are typically bound to beads, such as
agarose, that are packed into a column. Unfortunately, column-based methods often
decrease throughput, as the sample must be loaded onto the column, rinsed to remove any
unbound species, and then eluted to collect the analyte. The column eluate is finally
either analyzed by ESI-MS or mixed with matrix and spotted on a target and analyzed by
MALDI-MS.

More recently, hydrazide- and boronic acid-functionalized magnetic beads have
been used for enrichment of glycoproteins and peptides.23’24’25’26’27 These beads
covalently bind glycosylated proteins and peptides and have the ability to enrich a wider
range of glycoproteins and peptides than lectins, which may be necessary, depending on

the aim of the separation. Typically the magnetic beads must be loaded with analer and

then collected prior to rinsing to remove contaminants. Finally, the analyte is eluted, the

106

beads are collected, and the eluate is analyzed by ESI-MS or MALDI-MS (after addition
of matrix in the latter case). This is a tedious process, as collection of beads can be
difficult, and sample loss is also a concern using this enrichment technique.

This chapter describes attempts to develop an on-plate enrichment technique,
where an unpurified glycopeptide sample is spotted on a modified MALDI plate that
covalently binds glycopeptides. After rinsing to remove any impurities and
nonglycopeptides, matrix is added prior to analysis by MALDI-MS. This on-plate
enrichment technique should, in principle, exhibit minimal sample loss because the
enriched glycoopeptides are analyzed directly on the modified plate with no sample
handling steps.

Specifically, the on-plate enrichment technique employs gold-coated substrates
modified with polymer brushes that are functionalized with 3-aminophenylboronic acid
(APBA). APBA covalently binds cis-diol groups, which are presentin essentially all
carbohydrate groups. The derivatized brushes are significantly thicker than the
monolayer films that were used to cover the magnetic beads discussed above. Because of
their greater thickness, polymer brushes have a higher binding capacity than monolayer
films and should, therefore, have a greater capability to enrich glycopeptides, as

illustrated in Figure 4.1.

107

 

Monolayer Film Polymer Film

Figure 4.1: Comparison of the binding capacity of a thin monolayer film and a thicker
polymer brush. Spheres represent either a protein or peptide.
4.2 Experimental
4.2.1 Materials

Peroxidase from Horseradish (HRP) was purchased from Sigma and digested
using sequencing grade modified trypsin from Promega. Other digest reagents include
Tris-HCl (lnvitrogen), urea (J. T. Baker), 1,4-dithio-DL-threitol (BioChemika),
ammonium bicarbonate (Columbus Chemical Industries), and iodoacetamide (Sigma).
Silicon wafers (Addison Engineering) were sputter coated with 20 nm of chromium,
followed by 200 nm of gold (Lance Goddard Associates, Santa Clara, CA). The reagents
used for the preparation of surface-modified gold-coated plates include 11-
mercaptoundecanol (Aldrich), 2-bromoisobutyryl bromide (Aldrich), triethylamine (Jade
Scientific), 2-hydroxyethyl methacrylate (Aldrich), cupric bromide (Aldrich), cuprous
chloride (Aldrich), 2,2’-bipyridine (Aldrich), succinic anhydride (J.T. Baker), 4-
dimethylaminopyridine (Sigma), N—(3-dimethylaminopropyl)-N’-ethylcarbodiimide
hydrochloride (Sigma), N-hydroxysuccinimde (Aldrich), and 3—aminopheny1boronic acid
monohydrate (Aldrich). Fructose (Aldrich) was bound to the surface-modified plates.
Dimethylformamide (Spectrum) was dried using 3-A molecular sieves (Spectrum), and
deionized water was obtained through a Millipore purification system (Milli-Q, 18

108

MQcm). HPLC grade methanol (Sigma), isopropyl alcohol (EMD), and ammonium
hydroxide (Columbus Chemical Industries) were used to clean the conventional stainless
steel MALDI plate according to Therrno Scientific’s deep cleaning procedure as
described in Chapter 2. The plate was in a nitrogen-filled glove bag while not in use.
Ammonium bicarbonate (Columbus Chemical Industries) was used as the loading
solution for some protein samples. The matrix used for all MALDI-MS experiments in
this chapter was 0.25 uL of a 40 mg/mL solution of 2,5-dihydroxybenzoic acid (DHB,
Aldrich) in 1:1 HPLC grade acetonitrile (EMD): 0.1% trifluoroacetic acid (Aldrich),
which was applied directly to the protein sample on the conventional stainless steel

MALDI plate or within a well scratched on the modified gold-coated MALDI plates.

4.2.2 Protein Digestion

For the tryptic digestion of horseradish peroxidase samples (MW ~44 kDa),
twenty 100 pg samples of protein were separately dissolved in 20 uL of 6 M urea
containing 50 mM tris-HCl as discussed in Chapter 2. To reduce any disulfide linkages
present, 5 [IL of 10 mM 1,4-dithio-DL-threitol (DTT) was added to each sample, and the
protein solutions were heated in a water bath at ~65°C for 1 h. After cooling the sample
to room temperature, 160 11L of 50 mM ammonium bicarbonate and 10 11L of 100 mM
iodoacetamide were added to each protein solution, and the samples were placed in the
dark for 1 h. Finally, 10 uL of 0.5 ug/uL modified trypsin was added to each sample, and
they were incubated for ~16 hours at 37 0C. The digestion reaction was quenched with
the addition of 11 11L of glacial acetic acid, which lowered the pH to ~3. Samples were
dispensed into Eppendorf tubes in 22 uL aliquots and stored in a -70 °C freezer until

further use.

109

4.2.3 Fabrication of APBA-PHEMA Brushes

Gold-coated silicon wafers (1.1 x 2.4 cm) were UV/ozone-cleaned for 15 min and
immersed in a 1 mM solution of mercaptoundecanol (MUD) in ethanol (4.1 mg in 20 mL
ethanol) for ~16 h to form a self-assembled monolayer (SAM). Wafers were rinsed with
deionized water, followed by ethanol and dried under a stream of N2 gas. Gold wafers
(typically 8 at a time) were arranged in a crystallizing dish and placed in a nitrogen-filled
glove bag. Then 0.12 M triethylamine (TEA) (0.33 mL of TEA in 20 mL dry DMF) was
added to the crystallizing dish, followed by dropwise addition of 0.1 M 2-
bromoisobutyryl bromide (BIBB, 0.25 mL in 20 mL DMF) with swirling over a 10 min
period. Wafers were removed from the crystallizing dish and rinsed with DMF. After
drying in the glove bag for 10 min, they were removed and rinsed with ethyl acetate,
deionized water, and ethanol, and then dried under a stream of N2 gas. Reflectance FTIR
spectroscopy was used to identify the presence of the ester carbonyl peak (1730 cm") due
to the attachment of the initiator to the MUD SAM. Using freeze-pump-thaw cycling, 30
mL of 2-hydroxyethyl methacrylate (HEMA) and 30 mL of deionized water were
degassed in a Schlenk flask. During the third cycle, 165 mg CuCl, 108 mg CuBr2, and
640 mg 2,2’-bipyridine were added to the frozen mixture of HEMA and deionized water.
The catalyst dissolved once the mixture of HEMA and deionized water thawed. The
freeze-pump-thaw cycle was completed, followed by two additional cycles. The flask
was transferred to the nitrogen-filled glove bag and was distributed equally among four
20-mL scintillation vials, each containing two wafers modified with initiators attached to
the MUD SAM. Typically wafers were immersed in the HEMA solution for 2 h, which

gave film thicknesses of 45-50 nm. Thicknesses of the PHEMA films were varied by

110

either shortening or lengthening the polymerization time. The PHEMA films were
removed from the vials and rinsed with DMF, deionized water, and acetone, and were
characterized using reflectance FTIR spectroscopy. To convert the hydroxyl groups of
the PHEMA brushes to carboxylic acid groups, the films were immersed in 10 mL of
DMF containing 0.1 g succinic anhydride (SA) and 0.2 g 4-dimethylaminopyridine
(DMAP) and heated at 55 °C for 3 h. These films were rinsed with DMF, water, and
ethanol and dried under a stream of N2 gas. The carboxylic acid groups were then
activated using 50 mM N-(3-dimethylaminopropyl)-N’-ethylcarbodiimide hydrochloride
(EDC, 0.096 g) and 50 mM N-hydroxysuccinimide (NHS, 0.058 g) in 10 mL water for 30
min. Activated films were rinsed with water and ethanol, and then dried under N2 gas.
Finally, the PHEMA brushes were functionalized with 3-aminophenylboronic acid
(APBA) (Figure 4.2) by immersing films for ~16 h in 10 mL of DMF containing 0.031 g
APBA and 0.06 g DMAP. These APBA-PHEMA brushes were rinsed with DMF,
followed by ethanol and dried using N2 gas. All steps of the fabrication process were
characterized by reflectance FTIR spectroscopy, and wafers were initially cut to fit the
FT IR sample holder (1.1 x 2.4 cm). To fit the modified stainless steel MALDI plate,
wafers were cut to a width of 1.7 cm. Circular sample wells, with a diameter of 2 mm,
were scratched onto the wafers using a tungsten carbide-tipped pen. Typically six wells
were created on each wafer. The modified wafers were then secured to the modified

stainless steel MALDI plate using double-sided tape.

4.2.4 Binding of Fructose to APBA-PHEMA Brushes
Fructose was bound to APBA-derivatized PHEMA brushes (Figure 4.4) by

immersing wafers for 15 or 30 min in a 0.01 M aqueous solution of fructose (0.018 g

111

fructose in 10 mL deionized water) at pH 10.6 (pH was adjusted with 0.125 M NaOH,
final NaOH concentration was ~lmM). Wafers were briefly rinsed with ethanol (~10 s)
and dried under a stream of nitrogen gas. Films with bound fructose were characterized

using reflectance FTIR spectroscopy and ellipsometry.

4.2.5 Protocol for Enrichment of Glycopeptides Using APBA-PHEMA Brushes

For analysis of tryptic protein digests, 1 uL of digest solution was spotted in the 2
mm diameter wells that were scratched in the APBA-PHEMA-modified gold wafers.
Protein digest stock solutions were diluted in loading solution, either deionized water or
0.1 M ammonium bicarbonate, and samples were then spotted directly on the modified
wafer. Samples were incubated for 15 min, and additional loading solution (0.5 or 1 uL
of loading solution without digest) was added to the wells as the sample solution
evaporated throughout the incubation time. After 15 min, samples in deionized water
were rinsed with ~5 mL of ethanol and samples in 0.1 M NH4HCO3 were rinsed with ~5
mL of loading solution, followed by ~5 mL of ethanol, and then dried under a stream of
N2 gas. After drying, 1 11L of 0.1% TFA was added to each well to elute the
glycopeptides, followed immediately by addition of 0.25 uL of 40 mg/mL DHB solution
(1:1 acetonitrile: 0.1% trifluoroacetic acid) to deposit the matrix. After crystallization of
the solution, the gold-coated wafer was secured to the modified MALDI target using

double-sided tape.

4.2.6 Characterization and Instrumentation

The polymer brushes were characterized by reflectance FTIR spectrometry using
a Nicolet Magna 560 spectrophotometer with a Pike grazing angle (80°) accessory. The
thicknesses of the polymer films were determined using a rotating analyzer spectroscopic

112

ellipsometer (J. A. Woollam, M-44), assuming a film refractive index of 1.5. Mass
spectra were obtained with a MALDI linear ion trap mass spectrometer (Therrno

vMALDI LTQ XL) in positive ion mode.

113

2E. M...

lBIBB, TEA
lin DMF, 10 min

”E. “*5“ka

HEMA, H20
CuCl, CuBr2, bpy
r..t, 2hrO

clgsmowm

SA, DMAP 0
55°C, 3 hr

d) Br
n O

C O
O” \O/\/ NOH

ClrlNHs, EDC O

O/\/OH

30 min

APBA, DMAP
B ~16 hr

“ ° <1
C O OH
O” \O/\/ NN ?/
o H OH

Figure 4.2: Schematic diagram showing modified gold-coated substrates after a)
adsorption of an MUD SAM, b) attachment of initiator to MUD, c) polymerization of
HEMA, d) derivatization of PHEMA with succinic anhydride (SA), e) activation of

PHEMA-SA with NHS, and f) derivatization of PHEMA-SA with 3-aminophenylboronic
acid (APBA).

1’)

114

4.3 Results and Discussion
4.3.1 Characterization of PHEMA-APBA Brushes

The fabrication and derivatization of PHEMA brushes (Figure 4.2) were
characterized using reflectance FTIR spectroscopy, and film thicknesses were determined
after each step using ellipsometry. The reflectance FT IR spectra in Figure 4.3 confirm
the growth and derivatization of PHEMA on a gold-coated surface. After growth of
PHEMA, the spectrum of the 50 nm-thick films contains the expected ester carbonyl
(1730 cm], spectrum 4.3a) and hydroxyl stretches (3650 — 3100 cm”, not shown).
Functionalization of PHEMA with succinic anhydride results in a doubling of the ester
carbonyl absorbance (compare spectra 4.3a and 4.3b) due to an additional ester group in
each repeating unit of the polymer chain (Figure 4.1). The hydroxyl stretch also
disappears as a result of the essentially complete derivatization of the hydroxyl groups.
After functionalization with succinic anhydride, the polymer brushes increase in
thickness to approximately 75 nm. Spectrum 4.3c shows that the activation of the
polymer brush with NHS gives rise to succinimide ester peaks at 1813 and 1784 cm'1 and
an asymmetric stretch of succinimide at 1750 cm'1 that overlaps with the carbonyl stretch.
Attachment of APBA to the PHEMA-SA film occurs through the amine group (Figure
4.2). After derivatization with APBA, the FTIR spectrum of the film contains a hydroxyl
stretch at 3550 — 3250 cm'1 that may overlap with the N—H stretch of the amide. The
amide I band exists as a shoulder around 1664 cm], and vibrations due to the phenyl
group of APBA appear at 1600, 1575, 1540, and 1483 cm]. The amide II band should
also contribute to the absorbance around 1550. Vibrations due to the boronic acid occur

around 1362 and 1342 cm". After derivatization with 3-aminophenylboronic acid, the

115

polymer film thickness increases to approximately 100 nm, suggesting widespread

 

 

 

 

 

 

    

 

derivatization.
I 0.050 Amide | Boronic
Band .Phenvl acid
PHEMA-APBA carbonyl X V'bl rat'Ons J l
d) ~
\ Activated ester carbonyl
Succinimide
Ester \\ l I’ PHEMA-SA ester carbon I
C)
‘KPHEMA ester carbony
2200 2000 1800 1600 1400 1200 1000 800
cm'1

Figure 4.3: Reflectance FTIR spectra of a PHEMA brush immobilized on a gold-coated
substrate before a) and after b) functionalization with succinic anhydride, c) activation of
PHEMA-SA with NHS, and d) attachment of APBA to the PHEMA-SA film.

116

Br
n O O OH
/C 0 e
O/ \O/\/ \'l/\/U\N /B< .
o H HO OH

Fructose
Br pH 10.6
. «I
C O 9 O
0” \O/\/ NM /B< O

HO OH

Figure 4.4: Schematic diagram showing fructose binding, under basic conditions, to a
gold-coated substrate modified with an APBA-PHEMA film.

4.3.2 Characterization of Fructose Bound to APBA-PHEMA Brushes

Before enriching any glycopeptides using the APBA-PHEMA brushes, fructose, a
simple carbohydrate, was bound to the APBA—derivatized polymer film under basic
conditions (Figure 4.4) to ensure that these modified plates were able to bind a compound
containing cis-diols. Substrates modified with APBA-PHEMA brushes were first
immersed in a pH 10.6 solution (pH adjusted with 0.125 M NaOH, where the final NaOH
concentration is ~1mM) without fructose for 15 min (Figure 4.5a), then rinsed with
ethanol and dried under a stream of N2 prior to obtaining a reflectance FTIR spectrum
and an ellipsometric thickness of the film. Subsequently, the films were immersed in a
0.01 M fructose solution, pH 10.6 for 15 min (Figure 4.5b) and rinsed with ethanol before
again obtaining an IR spectrum and ellipsometric thickness. This procedure was
employed to ensure that the reflectance FTIR spectra and film thicknesses before and
after fructose binding were being compared under the same conditions. The inset of

Figure 4.5 shows the FTIR difference spectrum between APBA-PHEMA films before

117

and after exposure to fructose. There is an increased absorbance in the region from 3550
— 3000 cm", which may be due to the increase in hydroxyl groups present in fructose.
Sharp peaks in the subtraction spectrum at 1174 and 1066 cm’1 may also stem from C-O-
C stretches in fructose. The thickness of the APBA-PHEMA film, initially 82 nm,
increased 5 nm after immersion in the fructose solution. Taken together, these results
provide evidence for fructose binding, but the amount of binding is not extensive. The
high density of the polymer brushes may limit the ability of fructose to diffuse to boronic
acid sites, thus decreasing the binding of the cis-diol-containing carbohydrate. Longer

incubation times did not appear to increase fructose binding.

118

 

 

c-o-c j
Hyd roxyl Stretches
/ Stretch

   
   

0.002 0.020

 

 

 

 

3750 2750 1750 750

 

 

 

 

a)

 

3750 3250 2750 2250 1750 1250 750
cm'1

Figure 4.5: Reflectance FTIR spectra of a) a gold-coated substrate modified with APBA-
PHEMA (the substrate was immersed in a pH 10.6 solution for 15 min prior to taking the
spectrum), and b) fructose bound to the same APBA-PHEMA-modified substrate in pH
10.6 solution. The inset shows the subtraction of the spectrum of APBA-PHEMA film
(a) from the spectrum of the fructose-bound film (b).

4.3.3 Analysis of Horseradish Peroxidase Digests Using MALDI-MS

Initial attempts to enrich glycopeptides on APBA-PHEMA-modified MALDI
plates focused on tryptic digests of horseradish peroxidase (HRP), which contains nine N-
linked glycosylation sites (Figure 4.6). However, only eight glycopeptides result from
tryptic digestion because one peptide contains two glycosylation sites. This protein also
contains four disulfide bonds, which were reduced using DTT and iodoacetamide. Table
4.1 lists the sequences of the eight N—linked glycopeptides, the composition of their

oligosaccharide chains, and their experimental m/z values, [M+H]+, as reported

119

previously by Wuhrer et al.28 Unfortunately, two glycopeptides (69-92 and 214-236)
cannot be detected using the vMALDI LTQ XL because their [M+H]+ m/z values (4058.4
and 4986.2, respectively) exceed the upper mass limit of the mass analyzer of this
instrument (4000). However, this should not affect the enrichment or analysis of the

other glycopeptides using the APBA-PHEMA-modified MALDI plates.

3 l QLTPTFYDNS EPMVSNIVRD TIVNELRSDP 6 0

6 1 RIAAS I LRLH FHDSFVNGED AS I LLDN”TTS 9 O

9 1 FRTEKDAFGN AN SARGFPVI DRMKAAVE SA 1 2 O
1 2 1 EPRTVSEADL LT IAAQQSVT LAGGPSWRVP 1 5 O
1 5 1 LGRRDSLQAF LDLANANLPA PFFTLPQLKD l 8 0
1 8 1 SFRNV GLN’ RS SDLVALSGGH TFGKNQQRFI 2 l O
2 1 1 MDRLYN” FSNT GLPDPTLMTT YLQTLRGL_C_ P 2 4 O
2 4 1 WGMLSALW FDLRTPTI FD NKYYVNLEEQ 2 7 0
271 KGLIQSDQEL FSSPMATDTI PLVRSFAMST 3 00
3 0 1 QTFFNAFVEA MDRMGN’ ITPL TGTQGQI RLN 3 3 0

33 1 QRVVNSNS
Figure 4.6: Amino acid sequence of horseradish peroxidase.28 N-glycosylation sites are

labeled with bold and italic type and with a pound (#) sign. Cysteine residues involved in
disulfide bonds are underlined and tryptic cleavage sites are in bold.

The conventional MALDI mass spectrum (Figure 4.78) of 5 pmol of HRP digest
in H2O shows the presence of five glycosylated peptides with [M+H]+ m/z values of
1843, 3355, 3607, 3673, and 3896. While not listed in Table 4.1, the peak at m/z 3896 is
due to the peptide sequence LHFHDCFVNGCDASILLDNH‘TSFR, but the glycan has

28 Although the conventional mass spectrum

only three hexose units instead of four.
shows signals due to several of the glycopeptides in the HRP digest, enrichment of the
HRP digest using an APBA-PHEMA-modified MALDI plate aids in increasing signal-to-
noise ratio (S/N) of these glycopeptides, as well as decreasing the presence of some

nonglycosylated peptides and other contaminants (Figure 4.7b). Unfortunately, the mass

spectrum in Figure 4.7b has a relatively low signal-to-noise ratio and several signals due

120

to nonglycopeptides are still present. Relative to the conventional MALDI mass
spectrum, the S/N increased for the glycopeptides at m/z of 3355, 3673, and 3896, and
remained the same for the other two glycopeptides identified in the mass spectrum, at m/z
values of 1843 and 3607. Figure 4.7 gives the S/N values for these five glycopeptides.
The relatively short HRP incubation time (15 min) may account for the absence of the
other glycopeptides described in Table 4.1, as well as the low S/N ratio. Following
incubation, the modified MALDI plates were briefly rinsed with ethanol and then dried
before eluting with 0.1% TFA and addition of the matrix. Perhaps a more thorough
rinsing would result in the removal of the nonglycosylated species seen in Figure 4.6b. It
should also be noted that conventional MALDI analysis of 1 pmol of digested HRP in
H2O yielded a mass spectrum with an S/N of 31.4 for the most intense peak; whereas the
most intense peak in the conventional MALDI mass spectrum of 5 pmol of HRP in H2O
(Figure 4.7) has an SIN of 24. This suggests that the increased amount of digest reagents
and salt present in a 5 pmol of HRP digest sample may decrease the S/N of the mass

spectra, as well as interfere with the enrichment of glycopeptides from the digest.

Table 4.1: N-linked glycopeptides present in a tryptic HRP digest. The oligosaccharide
structure of each glycopepdide is shown, along with its [M+H]+ m/z value. N“ represents
the N-linked glycosylation site on asparagine residues. Hex stands for hexose, HexNAc
for N-acetylhexosamine, dHex for deoxyhexose, and Pent for pentose.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

$133331? Peptide Sequence Glycan Structure28 m/z
31-49 QLTPTFYDNSCPN4'VSNIVR HeX3HexNAc2dHex1Pent1 3323
69-92 LHFHDCFVNGCDASILLDN’HTSFR HeX4HexNAC2dHex1Pent1 4058

184-189 NVGLN#R HeX3HexNAC2dHex1Pent1 1844
214-236 LYWFSNTGLPDPrLNitTYLQTLR HeX3HexNAC2dHex1Pentt 4986
237-254 GLCPLNGNi'LSALVDI-‘DLR HeX3HexNAC2dHex1Pentt 3607
272-294 GLIQSDQELFSSPNfiATDTIPLVR HeX3HexNAC2dHex1Pent1 3674
295-313 SFAN#STQTFFNAFVEAMDR HeX3HexNAc2dHextPent1 3355
314-328 MGNirrPLTGTQGQIR HeX3HexNAc2dHex1Pent1 2612

 

121

 

a) 100% = 5590

it

3673 ,
S/N = 1.6 3:95

 

 

 

 

 

 

“1843 “
S/N = 51 g 3607 1 6
3. 3355 S/N = 1.8 '
(7, S/N = 2.0
c \\
a)
4.:
.E
a)
3 b) 100% = 2310
4.1
L“
a)
a:
”3896
”1843 “3673 5”“ =
S/N = 5.0 S/N = 4.5 4-8
”3355
S/N =
1.4
h

 

 

500 1000 1500 2000 2500 3000 3500 4000
m/z
Figure 4.7: Positive ion MALDI mass spectra of 5 pmol of HRP digest analyzed using a)
conventional MALDI-MS, and b) an APBA-PHEMA-modified plate with H2O as the
loading solution and rinsing with ethanol. In b), glycopeptides were eluted with 1 11L of

0.1% TFA, followed by addition of matrix prior to analysis. Pound signs (#) represent
signals due to glycopeptides.

122

Most studies of enrichment of glycopeptides using phenylboronic acid have been
performed under basic conditions. Under alkaline conditions phenylboronic acid exists in
a tetrahedral form, instead of its trigonal form, and while cis-diol-containing compounds
may bind to phenylboronic acid in either of its forms, they are more likely to bind to

2930 Figure 4.8a shows the conventional

phenylboronic acid in the tetrahedral form.
MALDI mass spectrum of 2 pmol of HRP digest in 0.1 M ammonium bicarbonate, where
43 11L of digest stock solution (pH ~3) was diluted in 184 uL of 0.1 M NH4HC03 to raise
the pH. Figure 4.8b displays the mass spectrum of the same digest/NH4HC03 solution
that was enriched on the APBA-PHEMA plate. While the higher pH of the NH4HCO3
solution should aid in the covalent binding of glycopeptides to aminophenylboronic acid-
functionalized polymer brushes, these basic conditions result in no signals due to
glycopeptides. Conventional MALDI analysis of 2 pmol of HRP in 0.1 M ammonium
bicarbonate shows significantly greater noise than the conventional mass spectrum of
HRP diluted in deionized water (Figure 4.7a), and this noise may mask the signals due to
several glycosylated and nonglycosylated peptides. The high salt content likely
suppresses signals.31 Unfortunately, Figure 4.8b suggests that the tryptic HRP
glycopeptides were not enriched from the ammonium bicarbonate solution using APBA-
PHEMA brushes, as no signals from any of the peptides listed in Table 4.1 can be
identified. Perhaps a lower concentration of ammonium bicarbonate in the loading
solution would decrease noise present in the conventional analysis and not interfere as
much with the enrichment process. Xu and coworkers used a 50 mM NH4HCO3 solution

for enrichment of glycopeptides using their APBA-functionalized mesoporous silica

beads.24 Other works use a phosphate buffer as the loading solution at a pH of either 7.4

123

or 9,27 which could be better because the binding conditions would be less harsh than in a
basic solution. Recently, La§tovickova and coworkers determined that when a 1 mg/mL
solution of ribonuclease B, a common glycoprotein, is diluted in a 20 mM ammonium
bicarbonate solution and is analyzed by MALDI-MS, the intensity and resolution of this
protein peak decreases significantly.32 This decrease in intensity and resolution occurs
with a number of different matrices: 2,5-dihydroxybenzoic acid (DHB), 2,4,6-
trihydroxyacetophenoe (THAP), a-cyano-4-hydroxycinnamic acid (a-CHCA), or
sinapinic acid as the matrix. However, the authors show that the use of binary matrices
helps to alleviate this problem, specifically the combination of DHB/or—CHCA,
DHB/T HAP, and DHB/sinapinic acid.32 It would be worthwhile to compare the mass
spectra of the enrichment of HRP in NH4HCO3 with DHB as the matrix, as shown above,

with the same enrichment, only using one of these binary matrices.

124

 

a) 100% = 1320

 

 

 

 

 

 

 

 

'01 100% = 2890

Relative Intensity

 

 

 

 

 

 

500 1000 1500 2000 2500 3000 3500 4000

m/z

Figure 4.8: Positive ion MALDI mass spectra of 2 pmol of HRP digest in 0.1 M
ammonium bicarbonate obtained using a) conventional analysis, and b) binding to
APBA-PHEMA-modified gold-coated plates that were rinsed with 0.1 M NH4HC03,
followed by ethanol, prior to elution of glycopeptides by deposition of 1 ILL 0.1% TFA,
addition of matrix solution, and MALDI-MS.

125

4.4 Conclusions

Poly(2-hydroxyethyl methacrylate) brushes were successfully derivatized with 3-
aminophenylboronic acid to yield APBA-PHEMA-modified gold-coated substrates.
These polymer films bind at least small amounts of fructose, a simple cis-diol-containing
carbohydrate. When used as MALDI substrates for on-plate enrichment of glycopeptides
from a protein digest, the APBA-PHEMA films marginally improved the signal-to-noise
ratios of some of the glycopeptides from an HRP digest in water, although a large number
of peaks due to nonglycosylated species are still present in the MALDI mass spectra.
Even though phenylboronic acids bind cis-diol-containing compounds more completely
under basic conditions, enrichment of glycopeptides from an HRP digest in ammonium
bicarbonate resulted in no signals due to glycosylated peptides. If the optimal binding
and matrix conditions can be determined, APBA-PHEMA-modified MALDI plates have

the potential for rapid and simple on-plate enrichment of glycopeptides.

126

4.5 References
l Helenius, A.; Aebi, M. Carbohydrates and Glycobiology, 2001, 291, 2364-2369.

2 Rudd, P. M.; Dwek, R. A. Critical Reviews in Biochemistry and Molecular Biology,
1997, 32, 1-100.

3 Durand, G.; Seta, N. Clinical Chemistry, 2000,46, 795-805.
4 Paulson, J. C. Trends in Biochemical Sciences, 1989, 14, 272-276.
5 Varki, A. Glycobiology, 1993, 3, 97-130.

6 Jaeken, J .; Matthijs, G. Annual Review of Genomics and Human Genetics, 2001, 2, 129-
151.

7 Freeze, H. H. Encyclopedia of Biological Chemistry, 2004, 2, 302-307.

8 Rostenberg, I.; Guizar-Vazquez, J .; Suarez, P.; Rico, R.; Nungaray, L.; Dominguez, C.
Journal of the National Cancer Institute, 1978, 60, 83—87.

9 Coffman, F. D. Critical Reviews in Clinical Laboratory Sciences, 2008, 45, 531-562.
‘0 Dias, W. B.; Hart, G. W. Molecular Biosystems, 2007, 3, 766-772.

“ Saldova, R.; Wormald, M. R.; Dwek, R. A.; Rudd, P. M. Disease Markers, 2008, 25,
219-232.

'2 Turner, G. A. Clinica Chimica Acta, 1992, 208, 149-171.
‘3 Dell, A.; Morris, H. R. Carbohydrates and Glycobiology, 2001, 291, 2351-2356.
'4 Hagglund, P.; Larsen, M. R. Spectral Techniques in Proteomics, 2007 , 81-100.

'5 Hemandez-Borges, J .; Neusiiess, C.; Cifuentes, A.; Pelzing, M. Electrophoresis, 2004,
25, 2257-2281.

‘6 Kurogochi, M.; Nishimura, S. Analytical Chemistry, 2004, 76, 6097-6101.
'7 Zaia, J. Mass Spectrometry Reviews, 2004, 23, 161-227.

‘8 Ullmer, R.; Plematl, A.; Rizzi,'A. Rapic Communications in Mass Spectrometry, 2006,
20, 1469-1479.

‘9 Endo, T. Journal of Chromatography A, 1996, 720, 251-261.

127

 

2° Hirabayashi, J. Journal ofBiochemistry, 2008, 144, 139-147.

21 Monzo, A.; Bonn, G. K.; Guttman, A. Trends in Analytical Chemistry, 2007, 26, 423-
432.

22 Satish, P. R.; Surolia, A. Journal of Biochemical and Biophysical Methods, 2001, 49,
625-640.

23 Sparbier, K.; Wenzel, T.; Kostrzewa, M. Journal of Chromatography B, 2006, 840, 29-
36.

24 Xu, Y.; Wu. Z.; Zhang, L.; Lu, H.; Yang, P.; Webley, P. A.; Zhao, D. Analytical
Chemistry, 2009, 81, 503-508.

25 Yeap, w. 3.; Tan, Y. Y.; Loh, K. P. Analytical Chemistry, 2008, 80, 4659-4665.

26 Zhang, H.; Li, x.; Martin, D. 13.; Aebersold, R. Nature Biotechnology, 2003, 21, 660-
666.

27 Zhou, W.; Yao, N.; Yao, G.; Deng, C.; Zhang, X.; Yang, P. Chemical
Communications, 2008, 5577-5579.

2" Wuhrer, M.; Hokke, C. H.; Deelder, A. M. Rapid Communications in Mass
Spectrometry 2004, 18, 1741-1748.

29 James, T. D. Topics in Current Chemistry 2007, 277, 107-152.
30 Yan, J .; Springsteen, G.; Deeter, S.; Wang, B. Tetrahedron 2004, 60, 11205-11209.
_ 3‘ Xu, Y.; Bruening, M. L.; Watson, J. T. Mass Spectrometry Reviews 2003, 22, 429-440.

32 La§toviékové, M.; Chmelik, J.; Bobalova, J. International Journal of Mass
Spectrometry 2009, 281, 82-88.

128

Chapter Five: Conclusions and Future Work

The use of ZrO2-PSS-PAH-modified plates for phosphopeptide enrichment and
subsequent identification of phosphorylation sites shows great potential. These metal
oxide-modified plates can separate phosphorylated peptides from an unpurified protein
digest and are fully capable of phosphopeptide enrichment at the picomolar level. After
enrichment using ZrO2-PSS-PAH-modified plates, the intensity of the phosphorylated
peptide peaks from a 2 pmol ovalbumin digest significantly improved, compared with the
intensity of these peaks from the conventional MALDI mass spectrum. Additionally,
there was little nonspecific binding when using these plates modified with metal oxides.
Plates prepared by heating an array of TiO2 nanoparticles also appear to be promising for
enrichment and analysis of phosphopeptides and their phosphorylation sites. These TiO2-
modified plates selectively enriched 125 fmol of a synthetic H5 phosphopeptide in the
presence of 1 pmol of a nonphosphopeptide mixture with a recovery of 69%. Both
techniques provide rapid and selective on-plate enrichment of phosphorylated peptides.
Further work regarding the ZrO2-PSS-PAH-modified plates should include examining
whether they can enrich phosphopeptides at the low fmol level. It would also be
interesting to compare these ZrO2 plates with other metal oxides such as aluminum oxide.
The only samples studied here were synthetic peptides and simple protein digests. For
modified plates to have true significance in the area of cancer research or in the study of
cellular regulatory mechanisms, they must be capable of enriching phosphopeptides from
biological samples, and this merits further investigation.

In an effort to expand the utility of on-plate enrichment, gold substrates modified

with poly(acrylic acid) or poly(2-hydroxyethyl methacrylate) films were derivatized with

129

reduced glutathione in a variety of methods. Unfortunately these glutathione-
immobilized films did not bind free GST, perhaps because of the presence of glutathione
in the GST sample. Despite desalting and purification of free GST by dialysis, MALDI
mass spectra confirmed that some glutathione remained in the sample. The GST initially
contained only 60% protein, and purer samples are needed to determine if the presence of
glutathione is truly preventing GST from binding to the films. Much work still needs to
be completed in this area.

Poly(2-hydroxyethyl methacrylate) brushes were also functionalized with 3-
aminophenylboronic acid to provide gold-coated substrates for on-plate enrichment of
glycopeptides prior to analysis by MALDI-MS. These modified films bind small
amounts of simple cis-diol-containing carbohydrates, specifically fructose. Initial studies
of the utility of these APBA-PHEMA—modified surfaces involved the enrichment of
either 2 or 5 pmol of glycoproteins from digested HRP. While these derivatized polymer
brushes improved the signal-to-noise ratio for some glycopeptides in an HRP digest in
water, a significant amount of nonspecific binding of nonglycosylated species was
apparent in the MALDI mass spectra. Even though phenylboronic acids bind cis-diol-
containing compounds more completely under alkaline conditions, enrichment of
glycopeptides from an HRP digest in ammonium bicarbonate resulted in no detectable
signals due to glycosylated peptides. Loading and rinsing solutions as well as incubation
times need to be optimized for the covalent binding between cis-diol-containing
glycopeptides and APBA-functionalized brushes. The use of binary matrices may
improve the intensity‘and resolution of glycopeptides peaks in the presence of salts and

other contaminants. With optimal binding conditions, the APBA-PHEMA-modified

130

MALDI plates might provide a useful method for rapid and simple on-plate

glycopeptides enrichment.

131

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII