. m, .9: ,
3.1%.?
h a...
. u

 

)1 E. In“...
Tut
, .

n, Hutu“:

.3..,.u.“.,ru.l|P5 kw. u v

 

. ‘ . . i. ‘ ‘ . ‘ I
, 5).. ‘ ‘ . . w, .ﬂ. _. u ..,_f.u........rxt an
» V $3.3m; V ‘ .. ‘ ‘ .t .

)3 .. ..‘ . ‘1‘..V:...... . l u.
3:}. :3_.,.5Iu.a..r1Wﬂ.i dwmw£.r.3hu. 46 .aﬁfymuuwx. Jamm— , m . . . ‘ ,

 

 

This is to certify that the
dissertation entitled

DESIGN AND IMPLEMENTATION OF PATTERNED SURFACES
FOR ON-PROBE CLEANUP AND CONCENTRATION OF
PROTEINS, PROTEIN DIGESTS, AND DNA PRIOR TO ANALYSIS
BY MALDI-TOF-MS

presented by

Yingda Xu

has been accepted towards fulﬁllment
of the requirements for the

Doctor degree in Chemistry

 

 

Major Professor’ 3 Signature

0

Date

MSU is an Afﬁrmative Action/Equal Opportunity Institution

 

LIBRARY
Michigan State
University

 

 

 

 

 

PLACE IN REfURN 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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 a/a‘RCIDatootnwd-ms

 

DESIGN AND IMPLEMENTATION OF PATTERNED SURFACES FOR
ON-PROBE CLEANUP AND CONCENTRATION OF PROTEINS, PROTEIN
DIGESTS, AND DNA PRIOR TO ANALYSIS BY MALDI-TOF-MS

By

Yingda Xu

A DISSERTATION

Submitted to
Michigan State University
In partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Chemistry

2004

ABSTRACT

DESIGN AND IMPLEMENTATION OF PATTERNED SURFACES FOR
ON-PROBE CLEANUP AND CONCENTRATION OF PROTEINS, PROTEIN
DIGESTS, AND DNA PRIOR TO ANALYSIS BY MALDI-TOF-MS

By
Yingda Xu

This dissertation describes a surface science/mass spectrometry effort to
develop and characterize patterned surfaces that serve as matrix-assisted laser
desorption/ionization (MALDI) sample platforms capable of concentrating and
purifying proteins and DNA. The use of these patterned surfaces can also enhance
the detectability of peptides, especially phosphopeptides, from protein proteolytic
digest mixtures. Using micro-contact printing, small (ZOO-um diameter) hydrophilic
spots of bare gold, chemically anchored poly(acrylic acid) (PAA), or immobilized
polyethylenimine (PEI) are patterned at S-mm intervals in a hydrophobic ﬁeld
consisting of a self-assembled monolayer of hexadecanethiol. Dilute or
salt-contaminated protein, DNA, or protein proteolytic digest samples are applied
onto the hydrophilic spots, dried, and then rinsed with water to remove water-soluble
contaminants, simplify a digestion mixture, or separate non-phosphopeptides from
phosphopeptides. One of the key features in this work is the combination of a
functionalized surface with a small spot to afford both concentration of analyte via
evaporation to a small spot size and puriﬁcation by selective adsorption. The

polymeric anchors bind the analytes during the water-rinsing step, and the

subsequently added acidic matrix solution releases analytes for their incorporation
into the matrix crystals.

Use of these patterned surfaces decreases the detection limit for the analysis
of dilute protein samples by MALDI-MS. For example, 1-5 fmol of insulin chain A,
insulin chain B, insulin, and ribonulcease A can be routinely observed with patterned
surfaces, while conventional stainless steel probes allow only 50-100 fmol detection
limits. The detection limits for salt-containing samples decrease by at least one
order of magnitude for use of patterned surfaces compared with use of non-patterned
on-probe decontamination methods.

The patterned surfaces also allow the detection of more tryptic peptides in
protein digestion mixtures. For example, use of a patterned PAA surface revealed
22 peptides as compared to only 11 peptides observed with a SS plate. Thus, the
PAA surface allows a much higher conﬁdence level for protein identiﬁcation during
myoglobin peptide mapping. Patterned surfaces also allowed 13% higher sequence
coverage for a larger protein, bovine serum albumin.

Modiﬁed probes containing small spots modiﬁed with polycations show great
promise for selectively enriching phosphorylated peptides directly on the probe prior
to MALDI-MS analysis. The positively charged anchors selectively bind the
negatively charged phosphopeptides, while the water-rinse removes other signal

suppressing contaminants or non-phosphopeptides.

To my loving and supportive parents

iv

ACKNOWLEDGEMENTS

I thank the M.S.U. Department of Chemistry for a very challenging and rewarding
graduate program. I owe much of my gratitude to my advisors J. Throck Watson and
Merlin L. Bruening. Their remarkable abilities to explaining complicated material in
a simple and concise manner are very contagious and have helped me to become a
better presenter. I would also like to thank Professor Doug Gage for serving my
committee even after he went to Pﬁzer. My family has always supported me and
showed pride in all of my endeavors. Without the values that they instilled upon me
at a young age, I never would have made it this far. I must say thanks to the mass
spectrometry facility for allowing me to use their instrument, even after I broke the
instrument twice. I also thank some of my friends that I met at M.S.U. such as
Guangming Wang, Rong Yang, Wei Wu, and Jianfeng Qi. I cherish the golden time
we enjoyed together. I would also like to thank Wenxi Huang, Jinhua Dai and all of
the old and current Watson and Bruening group members that I worked with at

Michigan State.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................... vii
LIST OF FIGURES ............................................................................. viii
LIST OF ABBREVIATIONS .................................................................. xii
Chapter One: Introduction

I. Introduction to MALDI ........................................................................ 2
II. On-Probe Sample Puriﬁcation ................................................................. 8
III. Dissertation Outline ........................................................................ 28
References ........................................................................................ 29

Chapter Two: Patterned Monolayer/Polymer Modiﬁed Metal Surfaces as Sample
Platforms for Analysis of Dilute or Salt-Contaminated Protein Samples by
MALDI-MS

I. Introduction .................................................................................... 35
11. Experimental Section ........................................................................ 38
III. Results and Discussion ..................................................................... 42
IV. Conclusions .................................................................................. 57
References ........................................................................................ 59

Chapter Three: Use of Polymer-Modiﬁed MALDI-MS Probes to Improve Analyses of
Protein Digests and DNA

1. Introduction .................................................................................... 61
11. Experimental Section ........................................................................ 63
III. Results and Discussion ..................................................................... 68
IV. Conclusions .................................................................................. 81
References ........................................................................................ 82

Chapter Four: Use of Polymer-Modiﬁed MALDI-MS Probes to Enhance Detection of
Phosphopeptides in Phosphoprotein Digest

1. Introduction .................................................................................... 83
11. Experimental Section ........................................................................ 87
III. Results and Discussion ..................................................................... 89
IV. Conclusions .................................................................................. 97
References ........................................................................................ 98

Chapter Five: Conclusions and Future Work

1. Conclusions ................................................................................ 102
11. Future Work ............................................................................... , 103
References ..................................................................................... 106

vi

LIST OF TABLES

Table 2.1 MALDI—MS signal intensities for replicate samples on patterned HDT
SAM/Au surface ................................................................................. 47

Table 3.1 Properties of peptides detected with MALDI-MS analysis of a myoglobin
tryptic digest using different surfaces: (a) Stainless Steel; (b) Hydrophobic HDT SAM
on Au; (c) HDT SAM/PAA-PEI on Au; ((1) HDT SAM/PAA on Au .................. 69-72

vii

 

LIST OF FIGURES

Figure 1.1 Scheme of Matrix-Assisted Laser Desorption/Ionization Time of Flight
Mass Spectrometer (MALDI—TOF-MS). Adapted from Introduction to Mass
Spectrometry 3“! edition J. Throck Watson p279. ............................................ 4

Figure 1.2 MALDI mass spectra of: (a) 0.5 pmol Insulin, no salt; (b) 5 pmol Insulin, 1
M NaOAc .......................................................................................... 6

Figure 1.3 MALDI TOF mass spectra of a 21-mer of single-stranded DNA showing
the effect of added NaCI: (a) control sample with no NaCl added; (b) 5 mM NaCl; (0)
25 mM NaCl; ((1) 250 mM NaCl. 50 nmol of DNA 21-mer were used in each case.
Adapted from T. A. Shaler et a1. Anal. Chem. 1996, 68, 576-579 ......................... 7

Figure 1.4 Scheme showing on-probe decontamination with a commercial membrane
attached to a conventional MALDI sample plate... ...................................... 10

Figure 1.5 Mass spectra showing the efficacy of a PE ﬁlm for adsorbing bovine serum
albumin (BSA) from a solution containing excess SDS. Control spectrum (top): pure
BSA applied to a PE membrane; Contaminated spectrum (middle): BSA in 0.73%
SDS applied to a PE membrane; Decontaminated spectrum (bottom): BSA in 0.73%
SDS applied to the PE membrane and subsequently vortexed in 50% methanol for 30 5.
All spectra were acquired at a protein loading of l pmol/mmz, and sinapinic acid was
added as matrix. Adapted from Blackledge et a1. Anal. Chem. 1995 67, 843-848.. . . 14

Figure 1.6 Schematic structure of a PU membrane showing the hard and soﬁ domains.
Adapted from McComb et al. Rapid Commun. Mass Spectrom. 1997 11,
1716-1722 ....................................................................................... 16

Figure 1.7 MALDI-TOF mass spectra of 200 pmol of myoglobin in the presence of
200 nmol of NaCl on a PU membrane: (a) original unwashed sample, (b) washed once
with water, (c) washed twice and (d) washed three times. Figures adapted from
McComb et a1. Rapid Commun. Mass Spectrom. 1997 11, 1716-1722 .................. 17

Figure 1.8 Conceptual schematic diagram showing desalting of a sample on a MALDI
probe surface modiﬁed with a thin layer of matrix crystals ................................ 19

Figure 1.9 Conceptual illustration of SAM formation and subsequent mechanism of

protein adsorption.Adapted from Brockman et al. Anal. Chem. 1997 69,
4716-4720 ......................................................................................... 24

viii

 

Figure 1.10 Ion-pairing solid phase extraction MALDI/MS spectrum of insulin (2
pmol/uL) from a solution of the peptide saturated with sodium acetate. These spectra
were acquired using (A) a strong cation-pairing surface

(3-mercapto—l -propanesulfonic acid) and (B) a strong anion pairing surface
(2-aminoethanethiol hydrochloride). Figure was adapted from Warren et a1. Anal.
Chem. 1998 70, 3757-3761 ..................................................................... 26

Figure 2.1 Comparison of sample deposition on conventional (top) and prestructured
(bottom) sample supports ...................................................................... 36

Figure 2.2 Schematic ﬂowchart for the preparation of a patterned HDT-SAM/PAA
sample probe and‘subsequent on-probe cleaning and concentrating of a
salt-contaminated protein sample for analysis by MALDI-MS ........................... 41

Figure 2.3 MALDI mass spectra of RNase A obtained using different sample probes.

(A) 600 fmol RNase A using a conventional 2-mm diameter sample well (70 mM DHB
as matrix); (B) 1 frnol RNase A using a 200-mm diameter Au spot in a patterned HDT
SAM (10 mM DHB as matrix) ................................................................ 44

Figure 2.4 Mass spectra of (A) 3 fmol insulin chain B and (B) 2.5 frnol insulin.
Spectra were obtained from 200-mm diameter Au spots in HDT SAMs (10 mM DHB
as matrix) ......................................................................................... 45

Figure 2.5 MALDI mass spectrum of insulin adsorbed from 0.25 mL of solution
containing 25 frnol insulin and 1 M NaOAc. The sample was applied to a 200-mm
diameter PAA spot in a HDT SAM and rinsed with water. DHB (10 mM) was added as
matrix ............................................................................................. 52

Figure 2.6 Reﬂectance FTIR spectra of PAA ﬁlms after exposure to a l-mM RNase A
solution for 30 min at different pH values: (A) pH 5.5; (B) pH 1.6; (C) pH 12.2.

Films were rinsed with water prior to measurements. In these studies, substrates were
not patterned, but completely coated with PAA to provide enough capacity to achieve a
sufﬁcient signal-to-noise ratio in the IR spectrum .......................................... 53

Figure 2.7 Amide absorbances in reﬂectance F TIR spectra and the protonated protein
signals obtained using MALDI-MS after exposure of sample plates to RNase A
solutions (1 mM, no salt) at different pH values, followed by water rinsing. Mass
spectra were taken on patterned HDT SAM/PAA plates (10 mM DHB as matrix) and
FTIR spectra were measured on plates coated only with PAA ............................ 54

Figure 2.8 Amide absorbances in reﬂectance FTIR spectra and the protonated protein
signals obtained using MALDI-MS after exposure of sample plates to RNase A
solutions at different concentrations (pH 7, no salt), and water rinsing. Mass spectra
were taken on patterned HDT SAM/PAA plates (10 mM DHB as matrix) and FTIR

spectra were measured on plates coated only with PAA .................................... 55

Figure 3.1 MALDI mass spectra of a tryptic digest of myoglobin (500 fmol) deposited
on different probe surfaces: (a) conventional stainless steel (b) hydrophilic anionic
spots in a hydrophobic ﬁeld, HDT SAM/ PAA; (c) a hydrophobic surface, HDT SAM;
(d) hydrophilic cationic spots in a hydrophobic ﬁeld, HDT SAM/ PAA-PEI. In
(b)-(d), the samples were deposited on the polymer-modiﬁed surface and rinsed with
water prior to addition of a-CHCA. Stars (*) indicate peaks that can be assigned to
tryptic fragments of myoglobin ............................................................... 74

Figure 3.2 MALDI mass spectrum of a BSA (1 pmol) tryptic digests obtained on a SS
probe. Stars (*) indicate peaks that can be assigned to tryptic fragments of
BSA ............................................................................................... 77

Figure 3.3 MALDI mass spectra of 60 frnol of a DNA 24-mer in 800 mM NaOAc
deposited on different probes: (a) SS without rinsing; (b) HDT SAM/ PAA-PEI with
rinsing after sample drying. A comatrix consisting of 3-HPA and ammonium citrate
was used in both cases .......................................................................... 78

Figure 3.4 MALDI mass spectrum of a DNA mixture contaminated with 1 M NaOAc.
The sample was deposited on a modiﬁed MU A-PEI surface, air dried, rinsed with
water, and subsequently matrix was applied ................................................ 80

Figure 4.1 MALDI mass spectra of: (a) 5 pmol Angiotensin (A); (b) 5 pmol
Phosphorylated Angiotensin (AP); ( c) 2.5 pmol of A and AP. a-CHCA was used as
matrix in every case ............................................................................. 85

Figure 4.2 MALDI mass spectrum of: (a) Equimolar mixture of angiotensin (A) and
phosphoangiotensin (AP), 2.5 pmol each, analyzed using stainless steel probe; (b) the
same sample as in (a), applied to a PAA-Fe3+ modiﬁed probe, dried, and rinsed with
water. Saturated a-CHCA was used as matrix in both cases ............................. 92

Figure 4.3 MALDI mass spectra of a peptide mixture prepared by mixing two
phosphopeptides, “AP" and “UP”, and a non-phosphopeptide, “A”, 5 mM each, with a
tryptic digest of myoglobin. (a)The sample was analyzed on a SS probe; (b) The
sample was analyzed with a patterned, PAA-PEI-modiﬁed probe. Saturated a-CHCA
was used as matrix in both cases .............................................................. 93

Figure 4.4 Mass spectra of a tryptic digest of ovalbumin (1 pmol) obtained using
different MALDI probes: (a) conventional SS; (b) PAA-Fe“; (c) PAA-PEI; (d) after
phosphatase treatment and analyzed on SS; (6) after phosphatase treatment and
analyzed on PAA-Fe“. In (b), (c) and (6), samples were deposited on the probe and
rinsed with water prior to addition of matrix. Peaks labeled with “P" represent
phosphorylated peptides and peaks labeled with “P-80” represent corresponding

 

nonphosphorylated peptides ................................................................... 95

Figure 4.5 Mass spectra of a tryptic digest of b-casein (100 fmol) obtained using
different MALDI probes: (a) conventional SS; (b) HDT SAM/ MUA-PEI. In (b), the
sample was deposited on the probe and rinsed with water prior to addition of matrix.
The peaks labeled with “P” represent phosphopeptides ................................... 96

xi

LIST OF ABBREVIATIONS

A ............................................. Angiotensin

AC ........................................... Ammonium Citrate

AP ............................................ phophorylated Angiotensin

B&B .......................................... Bull & Breese

BSA .......................................... Bovine Serum Albumin

or-CHCA .................................... a-cyano-4-hydroxy cinnamic acid

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

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

DTT .......................................... Dithiolthreitol

ESI ........................................... Electrospray Ionization

HDT .......................................... Hexadecanethiol

3-H PA ........................................ 3-Hydroxypicolinic Acid

HPLC ........................................ High Performance Liquid Chromatography
1 MAC ........................................ Immobilized Metal Affinity Chromatography
MALDI ...................................... Matrix Assisted Laser/Desorption Ionization
MLK3 ....................................... Mixed Linkage Kinase 3

MOWSE .................................... Molecular Weight Search

MS ............................................ Mass Spectrometry

MUA .......................................... Mercaptoundecanoic Acid

NaOAc ....................................... Sodium Acetate

xii

NTA ........................................... Nitrilotriacetic Acid

OM ............................................ Octadecyl Mercaptan
PAA ........................................... Poly(acrylic acid)

PDMS ......................................... Polydimethylsiloxane

PE ............................................. Polyethylene

PEI ............................................. Polyethylenimine

PP ............................................. Polypropylene

pS .............................................. phosphorylated Serine

pT .............................................. phosphorylated Threonine
PTBA .......................................... Poly(tert-butylacrylate)
pY .............................................. phosphorylated Tyrosine
PU ............................................. Polyurethane

PVDF .......................................... Polyvinylidenediﬂuoride
RMS ........................................... Root Mean Square

RNase A ....................................... Ribonuclease A

SAM ........................................... Self-assembled Monolayer
SS .............................................. Stainless Steel

Teﬂon .......................................... Poly(tetraﬂuoroethylene)
Zitax ............................................ Poly(tetraﬂuoroethylene)

xiii

Chapter One: Introduction

Mass spectrometry has been an important tool for structural analysis of small
molecules for a long time, but historically it has not been very useful for biochemical
analyses because the ionization of non-volatile macromolecules, such as proteins and
DNA, via traditional methods is difﬁcult, if not impossible. However, the
introduction of two new ionization techniques, electrospray ionization (E81)1 and
matrix-assisted laser/desorption ionization (MALDI)2, changed this situation.
Proteins and DNA with molecular weights as high as one million Daltons can now be
ionized and detected via mass spectrometry. These ionization techniques triggered
the explosive development of biochemical mass spectrometry in the last decade, and
two seminal contributors were recognized with the 2002 Nobel Prize.

Although both E81 and MALDI are capable of ionizing macromolecules, these
methods are often complementary. Usually multiply charged ions are observed in
ESI-MS, while predominantly singly charged ions are detected in MALDI-MS. ESI
is very sensitive and able to handle complicated mixtures when combined with high
performance liquid chromatography (HPLC), but on the other hand, MALDI is
attractive due to its simplicity in sample preparation, ease of instrument operation, and
high speed of data acquisition.

This dissertation describes my efforts to simplify the procedure for preparing
dilute or salt-contaminated samples for MALDI-MS. The use of patterned, functional
surfaces allows concentration and puriﬁcation of samples directly on the modiﬁed
sample probe. To place this work in context, this chapter contains a brief introduction
to MALDI, including sections on the matrix, sample preparation, instrumentation, and
problems that result from contamination. As my research focuses on on-probe

decontamination of samples, I also present an extensive review of surface

modiﬁcations previously used for on-probe decontamination for MALDI-MS. Finally,

this chapter contains a brief outline of the dissertation.

I. Introduction to MALDI

A. Roles of the Matrix in MALDI

Before the appearance of MALDI, several studies demonstrated laser
desorption/ionization of nonvolatile samples.3'5 However, the high laser intensity
required to desorb macromolecules leads to strong fragmentation and fast sample
consumption, thus limiting the application of this method. The addition of a matrix of
small organic molecules, such as nicotinic acidz, derivatives of cinnamic acid", or 2,5-
dihydroxybenzoic acid7, to macromolecular samples allows the use of much lower
laser intensities to desorb nonvolatile samples. The matrix molecule to analyte ratio is
very high, in the range of 10,000:1 to 100,000:1, so the analyte molecules are well
separated from one another. Absorption of a pulse of laser radiation (often from an
N2 laser, wavelength 337 nm) rapidly heats the surface of the matrix crystal, leading
to sudden vaporization of the top layers of the dried sample. It is under debate
whether ionization of the macromolecular analytes occurs in the solid phase (primary
ion formation) or during the desorption process (secondary ion formation)8. The
presence of contaminants, such as salts or surfactants, in the sample may interfere
with the co-crystalization process between the matrix and analytes by displacing
analytes from matrix crystal lattices, and thus should be kept to a minimum. The
matrix plays a central role in the ionization procedure, and new compounds for this

purpose continue to be studied.2’9'l7

B. Sample Preparation for MALDI-MS

For any MALDI process, the sample must be mixed with the matrix molecules.
The sample preparation can be quite simple: a drop (0.5 to 1 uL) of analyte solution
and a drop of matrix solution (same volume) are mixed and then deposited and dried
on a stainless steel (SS) or Au sample plate (dried droplet method)2. In some cases,
acetone is used as the solvent for the matrix solution to provide “fast evaporation” of
the sample on the surface”. This results in small, homogeneous polycrystals on the
plate. A “two-layer” method'7 can also yield uniform samples. In this case, a thin
layer of seed matrix crystals is ﬁrst deposited, followed by the deposition of
matrix/analyte solution. These modiﬁed sample preparation methods offer better

shot-to-shot signal reproducibility than the traditional dried droplet method.

C. Instrumentation (MALDI-TOF MS)

After the sample dries, the sample plate is inserted into a MALDI-time of ﬂight
(TOF) mass spectrometer (Figure 1.1). Ions formed upon laser irradiation are
accelerated with a high voltage (20-25 kV) before traveling in a ﬁeld-free ﬂight tube
(time-of-ﬂight tube). Ions with the same number of charges have the same kinetic
energy, but due to differences in molecular mass, they have different velocities and
are thus separated by the different times they spend in the time of ﬂight tube before

hitting the detector.

D. Salt Contamination
In spite of the many successes with MALDI-MS, practical problems still exist
in the application of this popular technique. One challenge is that contaminants

typically found in biological extracts often degrade or eliminate analyte signals. For

 

Beam Splitter v

 

 

 

 

 

 

 

~ a ~~ Laser
T' —~—‘ f f - , Transient
“gger "'c liegrder;
Sample I II F
PrObe All—ll? ++ +++ «If:
. i . T.
. F11 t Tube
1 gh Detector
Ion ~
Source _
W ‘ Protein
7 I; it .
' ' Matrix
Probe Tip

Figure 1.1 Scheme of Matrix-Assisted Laser Desorption/Ionization Time of
Flight Mass Spectrometer (MALDI-TOF-MS). Adapted from Introduction to
Mass Spectrometry 3rd edition I. Throck Watson p279

 

example, denaturants such as urea or guanidine-HCI are used to prevent aggregation
and precipitation of hydrophobic proteins,‘8 but residual amounts of these denaturants
decrease or eliminate signals in MALDI-MS. The presence of stabilizing salts or
surfactants also degrade the quality of mass spectra,18 although a low concentration of

- . 9. 0
these contaminants can be tolerated.‘2 l 2

As reported by Kallweit et al., the quality
of mass spectra usually decays when the salt concentration reaches 100 mM.19 When
the concentration of an interfering salt is around 1 M, analyte signals are usually
undetectable. This is clearly shown in Figure 1.2, where a strong signal was obtained
from as little as 0.5 pmol pure insulin (Figure 1.2 a); but no signal was detectable
from a 5-pmol sample contaminated with l M sodium acetate (NaOAc), even though
10 times more insulin was present (Figure 1.2 b).

Salt contamination is also problematic in the analysis of DNA samples by
MALDI-MS. As reported by Shaler et al.,21 when salt concentrations reach 0.1 M, the
peak intensity and resolution for DNA oligomers decrease due to formation of
multiple cation adducts. This is shown in Figure 1.3, where the quality of the mass
spectra decreases with increases in the concentration of NaCl.

To solve this salt-contamination problem, some analysts resort to HPLC to
purify samples prior to MALDI-MS analysis.22 However, HPLC is time-consuming,
and puriﬁed analytes are usually present at low concentrations because of the
relatively large volume of mobile phase employed for elution. Sample loss due to
adsorption on the inside walls of containers during evaporation of solvent to

concentrate analytes can become a serious secondary problem. Recently, micro-

columns have been designed to minimize sample loss and shorten puriﬁcation times.

 

 

 

 

é a
a L . L
g :
1:”
b
2000 M. f 16000

Figure 1.2 MALDI mass spectra of: (a) 0.5 pmol Insulin, no salt; (b) 5 pmol
Insulin, 1 M NaOAc.

Wiriﬁuﬁﬁ? Zita-QC '._'.I 0. «man: mat-m1.

 

 

 

 

 

 

._ E .._.... ..__._..___....._._..__. ,_..._....___. _
a (a) r’i UK“.
.5 "% ,Ancx -A_»~Mmu.Aw.-.dmn./-h-ﬁ'ﬁ' \“V‘ . v-A"'\"v-"-v .~ . Ann-«Mb».-.
C ' i
m \
H
E (b) j “my, .
0 ‘H~‘\-.\~...sn v-4.“,\,_a~,Ah~Mn\av‘-A'/v ‘P‘.\ V h-r\,w-..~ a; ,d ._ -, .‘_’.~ 7 .7 .A ’
>
H
(U
E (C) .II \ v
m r144 “o wavu‘m Mm ’ "W “I 4‘ ~ ~ ,. v '-
~ «-
\x
(d) ,
“2:32?" ":"~"‘ “: *Mr'“’""1"":".'_. ._.._..-- .___._ " ‘ ’ ""”'”.*.";;.‘;:;~.-~a
5000 6000 7000 8000
m/z

Figure 1.3 MALDI TOF mass spectra of a 21-mer of single-stranded DNA
showing the effect of added NaCl: (a) control sample with no NaCl added; (b)
5 mM NaCI; (c) 25 mM NaCl; (d) 250 mM NaCl. 50 nmol of DNA 21-mer
were used in each case. Adapted from T. A. Shaler et al. Anal. Chem. 1996,
68, 576-579.

 

—- -. .J-I—"lnlu z-

Commercially available Ziptip23 (Millipore) micro-columns contain a ﬁxed small
volume (0.2 - 0.6 uL) of chromatography stationary phase at the end of IO-uL pipette
tips. Analytes are concentrated on the stationary phase and eluted with a small
volume of MALDI matrix solution. A similar approach using reversed-phase nano-

columns was developed by Gobom et al.24

After loading of the sample onto the
nano-column and subsequent washing to remove salts, the analyte was eluted directly
on to the MALDI-MS target with 50-100 nL of matrix solution. Micro-extraction

5In

chips were also used for sample clean-up and enrichment of trace peptides.2
comparison to conventional HPLC separations, miniaturized off-probe puriﬁcation
techniques have greatly simpliﬁed sample preparation. However, these micro-

partitioning operations can still be time-consuming, and they add a signiﬁcant cost to

analyses.

[1. On—Probe Sample Purification

The main focus of my research is development of new sample puriﬁcation
methods that are performed directly on the MALDI sample probe. This work builds
on previous methods for on-probe sample puriﬁcation, and this section summarizes
prior work in this area. The basic strategy employed in all on-probe puriﬁcation
methods is to deposit sample solutions onto modiﬁed probes that allow selective
adsorption of a speciﬁc class of analyte molecules. Subsequent rinsing removes
contaminants such as salts, while leaving behind the analyte of interest, e.g., proteins
or DNA. Finally, matrix is added to the sample prior to analysis by MALDI-MS (See
Figure 1.4). This strategy eliminates the need for chromatographic separation prior to
sample deposition, and should increase the sample throughput and reduce the cost of

analyses.

Essentially, three types of desalting stages have been incorporated into MALDI
sample platforms: ﬁlms of commercial polymers, thin layers of matrix crystals, and
self-assembled monolayers (SAMs)/ ultrathin polymer ﬁlms. Most of these systems
allow separation of contaminants from adsorbed analytes with a simple rinsing

I .2o3 - - - - 4 - .
2 3 or even wrthout a rrnsrng step In some cases.3 Mechanisms for selective

step,
binding to the sample plate range from afﬁnity interactions to simple hydrophobic
adsorption, and here we focus on methods that utilize non-speciﬁc hydrophobic and
ionic interactions. Although such interactions do not allow highly selective
discrimination among macromolecules, they are capable of separating contaminants
such as salts and surfactants from macromolecules. This affords a simple and general
method for purifying samples on the probe for analyses by MALDI-MS. Below I

discuss relevant literature on decontamination with commercial polymers, thin layers

of matrix crystals, and self-assembled monolayers (SAMs)/ ultrathin polyrrrer ﬁlms.

A. Decontamination Using Films of Commercial Polymers
Many types of commercially available polymer membranes have been used to
selectively adsorb proteins from salty solutions as well as to interface mass
spectrometry with separation techniques, such as gel electrophoresis. This section is

divided into subsections according to the material employed for decontamination.

i. Polyvinylidenediﬂuoride and Nitrocellulose

The ﬁrst type of polymer membrane utilized for protein adsorption onto a MALDI
probe was polyvinylidenediﬂuoride (PVDF). Hefta and coworkers35 initially
deposited sample directly onto an unmodiﬁed, stainless steel MALDI probe and

rinsed it with 0.1% TFA in a 30% aqueous acetonitrile solution to remove

Polymer film

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l y W .va
. I Sample , WWW evevv
I ' I deposition
Metal plate Water
rinse

Add 0 O 0 O O O O 0
matrix
0 Protein
V Salt

Matrix crystals

Figure 1.4 Scheme showing on-probe decontamination with a commercial
membrane attached to a conventional MALDI sample plate.

contaminants. Addition of matrix allowed detection of MALDI signals for
cytochrome P450, but cytochrome P450D gave no detectable signal. To improve this
method, they utilized a piece of PVDF membrane as the adsorbing medium on the
MALDI probe. This yielded signal for cytochrome P450D, but charging of the
insulating PVDF membrane degraded the achievable mass accuracy.

To overcome the membrane charging problem, Mock et al.'2’27’3°
electrosprayed a thin layer of nitrocellulose onto the probe surface, rather than using a
thick piece of nitrocellulose membrane physically attached to the probe.
Nitrocellulose was originally used to decontaminate samples prior to plasma
desorption mass spectrometry, but in that case, a piece of nitrocellulose membrane
was employed.37 The success of desalting with electrosprayed nitrocellulose depends
greatly on the experimental procedure. For example, wetting of the rough,
hydrophobic nitrocellulose surface with 15% aqueous methanol facilitates good
contact of the surface with the salt-contaminated sample solution during loading;
drying of the sample prior to rinsing with water is also important for achieving
optimal signals, especially in the case of dilute sample solutions. A high acetonitrile
content (>50%) in the matrix solvent and elongated drying times were necessary to
facilitate desorption of protein from nitrocellulose for incorporation into the matrix
crystals.

Both Hillenkamp38 and F enselau3 ("3 9 extended the utility of PVDF membranes
by using them as substrates for electroblotting prior to MALDI-MS. Membranes with
higher surface areas yielded spectra with higher mass resolution,38 presumably
because the more porous structure allowed incorporation of more matrix molecules to
achieve an optimum matrix—to-analyte ratio. Vestling and F enselau4O performed on-

membrane digestion of proteins, and a subsequent water rinse proved sufﬁcient for

11

 

removal of contaminants introduced by the digestion buffer solution. Digestion of
cytochrome c on the membrane yielded more detectable peptides than a similar
procedure performed in solution with subsequent transfer of the digest to the MALDI
sample plate. This result was ascribed to different accessibilities to cleavage sites in
solution and on the membrane; however, adsorptive loss on vial walls and pipette tips
during sample transfer could also be part of the problem with the solution digestion

procedure.

ii. Nylon
Mass spectra from proteins adsorbed to polymer surfaces appear to depend on

both the protein and the polymer composition. For example, Zaluzec et al.32

reported
only weak signals from proteins adsorbed on either nitrocellulose or PVDF; however,
they showed that both ribonuclease A and trypsinogen give strong MALDI signals
when adsorbed to either nylon-66 or charge-modiﬁed nylon (Zetabind). Nylon-66
also allowed analysis of a DNA-binding regulatory protein from a solution containing
6 M guanidine-HCI. (The guanidine-HCI is needed to solubilize the protein.) No
signal could be observed for the analyte from the 6 M guanidine-HCI solution without
puriﬁcation by adsorption of the analyte on the nylon-66.

Decontamination using zetabind (Life Science Products) substrates also results
in a decrease in the number of cation adducts observed in some MALDI mass spectra.
When a 10-pmol sample of bovine insulin was spiked with 10-fold excess AgNO3 and
analyzed on a conventional MALDI plate, four Ag-ion adduct peaks were observed in

the mass spectrum. Deposition of the same sample on zetabind followed by a water

rinse yielded only the mono Ag-ion adduct.

12

 

iii. N aﬂon
Naﬁon (Dupont) is a perﬂuorinated polymer that contains sulfonate groups
that serve as ion-exchange sites. Unlike the desalting procedures employed with
PVDF, nitrocellulose or nylon, Naﬁon can bind interfering cations, and thus, in some

. . . . . 4.4]
cases, no washing step rs necessary prror to adding matnx.3

The ion-exchange
capacity of Naﬁon is limited, however, and thus high concentrations of salt cannot be
tolerated. Additionally, no desalting effect was observed when sample solution and
matrix were premixed, possibly because the presence of an overwhelming number of
protons competed with cations for binding-sites, or the surface was less effective in

binding cations than protons. Naﬁon proved particularly effective for analysis of real

biological mixtures, such as milk or egg whites, by MALDI-MS.

iv. Polyethylene (PE) and Polypropylene (PP)

Polyethylene (PE)42‘43 and polypropylene (PP)43 provide hydrophobic surfaces
for sample cleanup. Commercial membranes prepared from these polymers allowed
more reproducible mass spectra from contaminated solutions than commercial PVDF
and nitrocellulose membranes or C8 and C18 extraction disks. The reason for
improved reproducibility with PE and PP appears to be the small uniform pores in
these membranes, which allow formation of small, homogeneous matrix crystals.
Usually, the residue from an aqueous solution containing as little as 0.1% SDS will
eliminate MALDI-MS signals.44 However, using a PE-modiﬁed MALDI probe as a
desalting stage, Blackledge and Alexander obtained a mass spectrum of bovine serum
albumin from a sample containing 0.73% SDS after vortexing the sample-coated
probe in 50% aqueous methanol for 305. (See Figure 1.5) Similarly, Woods et al.‘43

observed signals from a 0.5-pmol-protein sample doped with 500 mM NaCl, 5%

13

I

bontrol

 

 

 

 

 

_.: Salty
0:“
l
L # \Pesalted
0‘ ' '2oo'oo i 4'00‘00‘ - 60606 F 80600' '

m/z

Figure 1.5 Mass spectra showing the efﬁcacy of a PE ﬁlm for adsorbing bovine
serum albumin (BSA) from a solution containing excess SDS. Control
spectrum (top): pure BSA applied to a PE membrane; Contaminated spectrum
(middle): BSA in 0.73% SDS applied to a PE membrane; Decontaminated
spectrum (bottom): BSA in 0.73% SDS applied to the PE membrane and
subsequently vortexed in 50% methanol for 30 5. All spectra were acquired at a
protein loading of 1 pmol/mmz, and sinapinic acid was added as matrix.
Adapted from Blackledge et al. Anal. Chem. 1995 67, 843-848.

14

glycerol, and 1% Triton X—100. In the latter case, the sample was spotted onto the
membrane surface, allowed to dry, and then rinsed three times with 70% aqueous
methanol. No explanation was provided for the mechanism of protein binding to

thepolyrners, but presumably, hydrophobic interactions played an important role.

v. Polyurethane (PU)

McComb et al.45‘47 employed PU (Stevens Elastomerics) as a platform to
desalt protein samples. This work was based on the study of Oleschuk and Chow“,
who showed that neutral species adsorb more strongly to PU than do charged species.
Ether-type PU contains moderately polar, hard urethane domains and non-polar, soft
polyether domains; proteins may interact with this material via hydrophobic
interactions with the polyether domain or H-bonding with the urethane domain (See
Figure 1.6). SEM images of the sample (not shown) showed that water rinsing
removed NaCl crystals from the PU surface; peak shape and resolution in mass
spectra of NaCl-containing, 200 pmol of myoglobin improved with increasing

numbers of washes. (See Figure 1.7)

vi. Paraffin and Teﬂon

In the late 90$, Guo’s group34'49

utilized parafﬁn wax ﬁlms and
poly(tetraﬂuoroethylene) (Teﬂon) for sample cleanup. In the case of paraffin wax-
modiﬁed probes for analysis of DNA samples, no rinsing step was incorporated in
sample preparation, but salt tolerance for successful analysis by MALDI” increased
from 5 mM to 100 mM of NaCl. The authors proposed that the hydrophobic parafﬁn

ﬁlms enhanced the rate of matrix/DNA crystallization relative to crystallization of

highly polar salts.

15

Amorphous

Soft Domain
Crystalline
Hard Domain

0
+O-(CH2)4,,C-O " C-N-Q—CHz-Q-N-C— (CH )4- -0+.
H

Soft Segment Hard Segment Soft Segment
Figure 1.6 Schematic structure of a PU membrane showing the hard and soft

domains. Adapted from McComb et al. Rapid Commun. Mass Spectrom. 1997
11,1716-1722.

Intensity

 

 

 

 

 

 

 

 

 

1+
. [M+H]
2+ a
[M+ 2H]
‘. b
.4
C
I
d
8000 10000 12000 14000 16000

m/z

Figure 1.7 MALDI-TOF mass spectra of 200 pmol of myoglobin in the presence
of 200 nmol of NaCl on a PU membrane: (a) original unwashed sample, (b)
washed once with water, (c) washed twice and (d) washed three times. Figures
adapted from McComb et al. Rapid Commun. Mass Spectrom. 1997 11, 1716-
1722

 

Guo also studied the use of Teﬂon as a sample loading and washing platform.
A 70% aqueous methanol solution was used to rinse off salt (1 M NaOAc), and
hydrophobic interactions were proposed as being responsible for retaining the protein
while salts were rinsed away. Polytetraﬂuoroethylene (Zitex) has also been used for

blotting of proteins,50 but in that case, no puriﬁcation procedure was employed.

vi. Ion-exchange Materials

Salt contamination leads to the formation of multiple cation adducts with
polyanionic DNA, which results in broad peaks in mass spectra. Nordhoff et al.5 I
added a few ammonium-loaded cation exchange beads to a sample droplet to replace
DNA-bound sodium or potassium ions with ammonium ions, and subsequent loss of
ammonia during desorption resulted in greatly simpliﬁed mass spectra. The polymer
beads interfered with neither the crystallization of matrix nor the laser ablation
process, and both signal intensity and mass resolution improved. Smirnov and

coworkers5 2

later coated MALDI probes with either polyethylenimine or
polyvinylpyrollidone and used these materials to adsorb DNA from salt-containing
solutions. These MALDI plates allowed both puriﬁcation and concentration of DNA

samples.

B. Decontamination Using Thin Layers of Matrix Crystals

In these methods, a thin layer of a relatively water-insoluble matrix, e.g., sinapinic
acid, serves as the medium that selectively binds proteins in the presence of salt. Fast
evaporation of solvent or crushing of raw matrix crystals produces a thin layer of

micro-crystals on the MALDI probe, and exposure of the micro-crystal-coated metal

18

 

‘A A ’A‘A A I. Applicatiog

 

 

 

 

of sample

 

 

 

solution
Water
rinse
-- - - -
i - o... -I . é --
. gaser desorption - Aﬁ‘ I :A g - _:"
l
.._.....-_...—..-.MJ~‘"J Lbﬁuanmn .--....._.~_,_

 

 

 

A Salt

0 protein

0 Matrix crystals

Figure 1.8 Conceptual schematic diagram showing desalting of a sample on a
MALDI probe surface modified with a thin layer of matrix crystals.

19

_.—._ — rev-- wv . Mun-u-

surface to sample followed by rinsing with water yields a puriﬁed matrix-analyte ﬁlm.
(See Figure 1.8) Beavis and Chait53 ﬁrst demonstrated this concept by immersing
analyte/matrix crystals into cold, distilled water to remove water-soluble
contaminants. Signal intensities for the analyte did not decrease after rinsing,
suggesting that the amount of protein lost from the crystals during washing was
negligible. Removal of interfering salts greatly improved mass resolution,
presumably because of a decrease in the concentration of cation adducts.

Below, I ﬁrst review a study of the mechanism by which proteins adsorb to matrix
crystals because binding to the matrix is the heart of the puriﬁcation process. The
subsequent section discusses the various methods for preparing MALDI probes
modiﬁed with matrix crystals. Procedures employed for probe modiﬁcation strongly

affect the quality of mass spectra.

i. Mechanism of Adsorption to Matrix Crystals

Beavis and Bridson studied the mechanism of protein adsorption to matrix crystals
using X-ray crystallography and staining of proteins with Coomassie Brilliant blue.54
X-ray structures showed that the planar trans-sinapinic acid molecules hydrogen
bonded to each other in extended sheets in the crystal lattice, while staining patterns
demonstrated that proteins contacted only the crystal faces parallel to these extended
sheets. The crystal plane that interacts with the protein contains no H-bond donor or
acceptor atoms, suggesting that only hydrophobic interactions occur between the
protein and the crystal face.

Crystallographic and staining studies also suggested a mechanism by which SDS
interferes with the MALDI process. A low concentration of SDS did not affect

crystallization of the matrix, but it did abolish staining of the crystal. Previous studies

20

 

showed that the hydrophobic tails of SDS molecules can bind to the hydrophobic
portion of proteins to form rod-like particles55 and change the amphiphilic nature of
the protein. The hydrophobic portions of the protein would then no longer be
available to bind to the hydrophobic crystal surface. The lack of binding between
proteins and matrix in the presence of SDS suggests that a single layer of matrix

crystals will not be effective for purifying samples contaminated with surfactants.

ii. Preparation of Desalting Matrix Crystals
1. Thin Layer of Matrix Crystals

Xiang and Beavisll

utilized crushed crystals as “seeds” to facilitate the
formation of a polycrystalline ﬁlm at the base of a salt-contaminated sample droplet.
In this work, they ﬁrst applied a drop of matrix solution (without analyte) to the probe
surface and allowed it to dry. They then crushed this compact deposit using a glass
slide and brushed the surface with a tissue, leaving behind only a trace of micro-
crystals. A solution containing matrix, analyte, and contaminants was then applied to
the probe surface. Within several seconds, an opaque ﬁlm formed at the base of the
droplet, and after 1 minute, the plate was immersed in water to remove contaminants.
The matrix ﬁlm was stable under these rinsing conditions. This procedure yielded a
strong MALDI-MS signal for 1 pmol horse skeletal muscle myoglobin even when the
solution contained 6 M urea.

Cadene and Chait56 used a modiﬁed thin-layer method to analyze membrane
proteins in the presence of non-ionic detergents. They covered the MALDI probe
with a small amount of matrix solution, and wiped it dry just prior to complete

evaporation of the solvent to produce a thin layer of matrix crystals. Using this

surface, deposition of samples containing protein, matrix, and 0.5 mM surfactant

21

 

followed by rinsing with water allowed successful analysis of several different
membrane proteins by MALDI-MS. However, concentrations of surfactants higher
than 0.5 mM do decrease analyte signals. Detergent tolerance is especially important
in the analysis of membrane proteins because surfactants are required to prevent
precipitation of these macromolecules.

Vorrn et al.57 deposited a thin layer of matrix crystals by fast evaporation of
solvent to improve mass accuracy and achieve high sensitivity in MALDI-MS,
presumably due to the capacity for removing salts and the formation of smaller
matrix/protein crystals. They chose acetone as a matrix solvent so that the
matrix/protein solution would spread quickly on the probe to produce a thin and
homogeneous layer of micro crystals. The choice of matrix was limited to the less
water-soluble matrices to facilitate a washing step. This method resulted in a
resolving power of 5700 for medium-sized (m/z~3000) peptides. Zhang et al.17
modiﬁed this procedure slightly by ﬁrst depositing a thin ﬁlm of matrix and then
applying a sample droplet that contained both analyte and matrix. A discernible
MALDI signal for 250 frnol bovine serum albumin in a sample contaminated with 1

% SDS was obtained successfully with this method.

2. Hydrophilic Spots of Matrix in Hydrophobic Polymer Layers
Using a small spot of matrix crystals in a pre-structured sample probe, Gobom
et a1.5 8’5 9 improved the detection limit for analytes in salt-contaminated samples
during analysis by MALDI-MS. They ﬁrst coated the probe with a thin, hydrophobic
polymer layer that contained an array of tiny holes (400 pm in diameter). Growth of a
thin layer of matrix crystals only in the holes provided localized adsorption sites for

the protein or peptides, while keeping the surrounding region hydrophobic. Because

22

min-pedal:- .r'v'ﬁ"ﬂ'¢’?3s. 4152- '5, |s_-.-_l .

of the difference in hydrophobicity between the matrix crystals and the surrounding
hydrophobic polymer region, the sample droplet evaporated to the size deﬁned by the
localized spot of crystals, concentrating the analyte molecules. As with other
methods, a rinsing procedure washes away contaminants and leaves protein behind.
The small spot sizes resulted in 10-20 frnol detection limits, even for the analyte in

salt-contaminated samples.

C. Decontamination using SAMs and Ultrathin Polymer Films

The use of SAMs and ultrathin polymer ﬁlms for modifying MALDI probes
allows tailoring of this interface for puriﬁcation of protein samples. Additionally, the
minimal thickness of these coatings should alleviate problems with charging of the
modiﬁed probes. Several studies demonstrated the attachment of a monolayer of
antibodies to a MALDI probe to speciﬁcally bind an antigen in the presence of
contaminantsf’m’z Those studies fall into the category of sample cleanup with afﬁnity
interactions, and thus are not in the scope of this project. Here I discuss the
fabrication of monolayer/ultrathin polymer ﬁlms that can adsorb proteins from
contaminated solutions via hydrophobic or electrostatic interactions.

i. Decontamination using Hydrophobic Interactions

SAMs of alkanethiols provide a well-characterized surface"3

for implementing
hydrophobic interactions in the puriﬁcation of analytes from sample solutions prior to
analysis by MALDI-MS. Brockman et al."4 modiﬁed MALDI probes with a
monolayer of octadecyl mercaptan (OM) by immersing a clean gold MALDI probe
into an etlranolic OM solution. (See Figure 1.9)

Analysis of samples using these probes involved the standard procedure described

above: deposition of sample, rinsing, deposition of matrix solution, evaporation of

23

 

Octadecyl Mercaptan

HWAM

Go|d HW
surface Hs/WVWW\/\

HS/\/WW\/\/\/\
lSample deposition

W
lWater rinse

W.

 

 

 

 

Figure 1.9 Conceptual illustration of SAM formation and subsequent mechanism
of protein adsorption.Adapted from Brockman et al. Anal. Chem. 1997 69, 4716-
4720.

24

solvents, and analysis of the crystalline residue by MALDI-MS. Using this
procedure, analyte signals are rather difﬁcult to obtain, and the analyst must search all
over the modiﬁed sample probe for a few spots that yield reasonable single-shot
spectra. Brockman et al. speculate that this occurs because a limited contact area
between the sample droplet and the probe surface (high contact angle) results in
limited adsorption of the analyte. Improved reproducibility results from immersing
the SAM-modiﬁed probe into the salt-containing protein solution overnight (8 b).

However, the longer immersion time greatly decreases sample throughput, and
additional analyte solution is necessary to cover the probe. A subsequent study
showed that the amount of protein binding to the SAM is independent of analyte

concentration in solution, but dependent upon immersion time."5

This happens
presumably because the SAM has a limited binding capacity and, thus, increasing the
protein concentration may not increase the amount of protein adsorbed to the surface.
However, after the formation of the ﬁrst protein layer, some loose layers of protein
may form by attaching themselves to underlying proteins via H-bonding, electrostatic

interactions, or hydrophobic interactions. The formation of these “loose” layers may

be time-dependent, and thus, longer immersion times lead to higher adsorption levels.

ii. Decontamination Using Electrostatic Interactions
Because of the slow protein adsorption on hydrophobic monolayers, Orlando
and coworkers began investigating the utility of ionic monolayers for extracting
proteins and peptides from salty solutions prior to analysis by MALDI-MS.29
Remarkably, ionic monolayers successfully extracted proteins from solutions
contaminated with 20% Triton X-100, 8 M urea, and saturated sodium acetate. (See

Figure 1.10)

25

 

Insulin (M-I-H)+ \

Intensity

 

I. A A L

6100

l

 

 

51 00 5300 5500 5700 5900
m/z

Figure 1.10 Ion-pairing solid phase extraction MALDI/MS spectrum of insulin (2
pmol/mL) from a solution of the peptide saturated with sodium acetate. These
spectra were acquired using (A) a strong cation-pairing surface (3-mercapto-1-
propanesulfonic acid) and (B) a strong anion pairing surface (2-aminoethanethiol
hydrochloride). Figure was adapted from Warren et al. Anal. Chem. 1998 70,

3757-3761 .

26

Results from these two experiments suggest that the mechanism for protein
extraction by these monolayers is ion pairing and not hydrophobic interactions or
other forces. First, washing the dried salty sample with organic solvent rather than
deionized water has no signiﬁcant effect upon analyte signal intensities. If
hydrophobic interactions dominated protein extraction, rinsing with organic solvent
should eliminate, or at least decrease, protein or peptide signals. Second, the pH of
rinsing solutions affects the amount of protein bound to monolayers containing weak
acid groups. Rinsing with pH 2.3 solutions removes peptides bound to carboxylate-
terminated surfaces, but has little effect on peptide adsorption at sulfonate-terminated
surfaces. At pH 2.3, carboxylate groups should be protonated (neutral), while
strongly acidic sulfonate groups will be deprotonated (negatively charged). Thus,
results from this experiment suggest that ion pairing is the main driving force for
peptide adsorption. Cations in contaminated solutions may compete with
proteins/peptides for adsorption sites, but the peptides/proteins have multiple binding
sites, and hence, achieve a stronger overall interaction with the charged surface.

Figure 1.10 shows that both positively (mercaptoethylamine) and negatively
(mercaptopropanesulfonic acid) charged surfaces are capable of extracting insulin
from highly contaminated solutions. This result raises the question as to how
electrostatic interactions can occur between a single protein and both positively and
negatively charged monolayers. Additionally, insulin contains only two positively
charged residues and would not be expected to bind strongly to a sulfonated surface in
the presence of a large excess of cations. These conditions suggest that binding of
proteins to charged monolayers is not driven solely by electrostatics; entropy changes

due to release of bound water upon protein binding may also favor adsorption.

27

Jain)“. -:‘T mm

 

One of the main problems in using SAMs for decontamination of samples
prior to analysis by MALDI—MS is their limited binding capacity, which results in
weak signal intensities for the analyte. To address this problem, Zhang and Orlando
immobilized polylysine chains onto a gold surface via a succinimide-containing

33 Signal intensities in MALDI mass spectra of proteins adsorbed to this

monolayer.
surface increased with the molecular weight of the polylysine. This result probably
occurs because an increase in polylysine molar mass translates into an increase in

immobilized amine groups on the probe, thereby providing more binding sites for

protein adsorption.

III. Dissertation Outline:

Above 1 reviewed the wide variety of on-probe decontamination methods used
prior to analysis by MALDI-MS. Below I will describe the approach I employed to
achieve both puriﬁcation and concentration of proteins (chapter two), peptide
mixtures (chapter three), and DNA (chapter three) on the MALDI probe. Chapter 4
will present my most recent work that aims at enhancing the detectability of

phosphopeptides via selective, on-probe adsorption.

28

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

(12)

(13)

Reference:

Fenn, J. 8.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M.
Electrospray ionization for maqlss spectrometry of large biomolecules
Science 1989, 246, 64-71.

Karas, M.; Hillenkamp, F. Laser desorption ionization of proteins with
molecular masses exceeding 10,000 daltons Anal. Chem. 1988, 60, 2299-2301.

Balasanmugam, K.; Viswanadham, S. K.; Hercules, D. M. Characterization of
polycyclic aromatic hydrocarbons by laser mass spectrometry Anal. Chem.
1986, 58, 1102-1108.

Li, L.; Lubman, D. M. Resonant two-photon ionization of enkephalins and
related peptides volatilized by using pulsed laser desorption in supersonic
beam mass spectrometry Anal. Chem. 1988, 60, 1409-1415.

Cotter, R. J. Laser mass spectrometry: an overview of techniques, instruments
and applications Anal. Chim. Acta 1987, 195, 45-59.

Beavis, R. C.; Chait, B. T. Cinnamic acid derivatives as matrices for
ultraviolet laser desorption mass spectrometry of proteins Rapid Commun.
Mass Spectrom. 1989, 3, 432-435. :.

Strupat, K.; Karas, M.; Hillenkamp, F. 2,5-Dihydroxybenzoic acid: a new
matrix for laser desorption-ionization mass spectrometry Int. J. Mass
Spectrom. 1991, 111, 89-102.

Zenobi, R.; Knochenmuss, R. Ion formation in MALDI mass spectrometry
Mass Spectrom. Rev. 1999, 17, 337-366.

Bai, J.; Liu, Y.-H.; Cain, T. C.; Lubman, D. M. Matrix-Assisted Laser
Desorption/Ionization Using an Active Perﬂuorosulfonated Ionomer Film
Substrate Anal. Chem. 1994, 66, 3423-3430.

Weinberger, S. R.; Boemsen, K. 0.; F inchy, J. W.; Robertson, V.; Musselman,
B. D. An evaluation of crystallization for matrix-assisted, laser desorption
/ionization of proteins Proceedings of the 41st ASMS Conference on Mass
Spectrometry and Allied Topics 1993, p775a-b.

Xiang, F .; Beavis, R. C. Growing protein-doped sinapic acid crystals for laser
desorption: an alternative preparation method for difﬁcult samples Org. Mass
Spectrom. 1993, 28, 1424-1429.

Mock, K. K.; Sutton, C. W.; Cottrell, J. S. Sample immobilization protocols
for matrix-assisted laser-desorption mass spectrometry Rapid Commun. Mass
Spectrom. 1992, 6, 233-238.

Kochling, H. J.; Biemann, K. Novel sample preparation for MALDI
Proceedings of the 43rd ASMS Conference on Mass Spectrometry and Allied
Topics 1995, p1225.

29

(14)

(15)

(16)

(17)

(18)

(19)

(20)

(21)

(22)

(23)

(24)

Vorm, 0.; Roepstorff, P.; Mann, M. Improved resolution and very high
sensitivity in MALDI TOF of matrix surfaces made by fast evaporation Anal.
Chem. 1994, 66, 3281-3287.

Li, L.; Golding, R. E.; Whittal, R. M. Analysis of single mammalian cell
lysates by mass spectrometry J. Am. Chem. Soc. 1996, 118, 11662-11663.

Kussmann, M.; Nordhoff, E.; Rahbek-Nielsen, H.; Haebel, S.; Rossel-Larsen,
M.; Jakobsen, L.; Gobom, J .; Mirgorodskaya, E.; Kroll-Kristensen, A.; Palm,
L.; Roepstorff, P. Matrix-assisted laser desorption/ionization mass
spectrometry sample preparation techniques designed for various peptide and
protein analytes J. Mass Spectrom. 1997, 32, 593-601.

Zhang, N.; Doucette, A.; Li, L. Two-layer sample preparation method for
MALDI mass spectrometric analysis of protein and peptide samples
containing sodium dodecyl sulfate Anal. Chem. 2001, 73, 2968-2975.

Ikeguchi, M. Protein denaturation and roles of denaturants Seibutsu Butsuri
2002, 42, 72-74.

Kallweit, U.; Boemsen, K. 0.; Kresbach, G. M.; Widmer, H. M. Matrix
compatible buffers for analysis of proteins with matrix-assisted laser

desorption/ionization mass spectrometry Rapid Commun. Mass Spectrom.
1996, 10, 845-849.

Bomsen, K. 0. Inﬂuence of salts, buffers, detergents, solvents, and matrices
on MALDI-MS protein analysis in complex mixtures Method. Mol. Biol.
2000, 146, 387-404.

Shaler, T. A.; Wickham, J. N.; Sannes, K. A.; Wu, K. J .; Becker, C. H. Effect
of impurities on the matrix-assisted laser desorption mass spectra of single-
stranded oligodeoxynucleotides Anal. Chem. 1996, 68, 576-579.

Qi, J .; Isupov, M. N.; Littlechild, J. A.; Anderson, L. E. Chloroplast
glyceraldehyde-3-phosphate dehydrogenase contains a single disulﬁde bond
located in the C-terminal extension to the B subunit .1. Biol. Chem. 2001, 276,
35247-35252.

Bagshaw, R. D.; Callahan, J. W.; Mahuran, D. J. Desalting of in-gel-digested
protein sample with mini-C18 columns for matrix-assisted laser desorption
ionization time of ﬂight peptide mass ﬁngerprinting. Anal. Biochem. 2000,
284, 432-435.

Gobom, J .; Nordhoff, E.; Mirgorodskaya, E.; Ekman, R.; Roepstorff, P.
Sample puriﬁcation and preparation technique based on nano-scale reversed-
phase columns for the sensitive analysis of complex peptide mixtures by
matrix-assisted laser desorption/ ionization mass spectrometry J. Mass
Spectrom. 1999, 34, 105-116.

30

 

Q
.‘
Ia
IL.
t
i
F
l

(25)

(26)

(27)

(28)

(29)

(30)

(31)

(32)

(33)

(34)

(35)

(36)

Ekstrom, S.; Malmstrom, J.; Wallman, L.; Lofgren, M.; Nilsson, J.; Laurel],
T.; Marko-Varga, G. On-chip microextraction for proteomic sample
preparation of in-gel digests Proteomics 2002, 2, 413-421.

Fan, X.; Beavis, R. C. A method to increase contaminant tolerance in protein
matrix—assisted laser desorption/ionization by the fabrication of thin protein-
doped polycrystalline ﬁhns Rapid Commun. Mass Spectrom. 1994, 8, 199-
204.

Mock, K. K.; Sutton, C. W.; Keane, A.; Cottrell, J. S. Sample immobilization
protocols for matrix assisted laser desorption mass spectrometry Tech. Protein
Chem. IV, [Pap Protein Soc. Symp.], 6th 1993, 65-72.

Rouse, J. C.; Vath, J. E. On-the-probe sample cleanup strategies for
glycoprotein-released carbohydrates prior to matrix-assisted laser desorption-
ionization time-of-ﬂight mass spectrometry Anal. Biochem. 1996, 238, 82-92.

-"cm.u..- In 18%.

Warren, M. E.; Brockman, A. H.; Orlando, R. On-probe solid-phase
extraction/MALDI-MS using ion-pairing interactions for the cleanup of
peptides and proteins Anal. Chem. 1998, 70, 3757-3761.

 

Xiang, F .; Beavis, R. C. A method to increase contaminant tolerance in protein
matrix-assisted laser desorption/ionization by the fabrication of thin protein-
doped polycrystalline ﬁhns Rapid Commun. Mass Spectrom. 1994, 8, 199-
204.

Xu, Y.; Watson, J. T.; Bruening, M. L. Patterned monolayer/polymer ﬁlms for
analysis of dilute or salt-contaminated protein samples by MALDI-MS Anal.
Chem. 2003, 75, 185-190.

Zaluzec, E. J .; Gage, D. A.; Allison, J .; Watson, J. T. Direct matrix-assisted
laser desorption ionization mass spectrometric analysis of proteins
immobilized on nylon-based membranes J. Am. Soc. Mass Spectrom. 1994, 5,
230-237.

Zhang, L.; Orlando, R. Solid-Phase Extraction/MALDI-MS: Extended ion-
pairing surfaces for the on-target cleanup of protein samples Anal. Chem.
1999, 71 , 4753-4757.

Hung, K. C.; Ding, H.; Guo, B. Use of poly(tetraﬂuoroethylene)s as a sample
support for the MALDI-TOF analysis of DNA and proteins Anal. Chem. 1999,
71, 518-521.

Hefta, S. A.; Stahl, D. C.; Mahrenholz, A. M.; Martino, P. A.; Rutherford, S.
M.; Lee, T. D. Analysis of highly glycosylated or hydrophobic membrane
proteins by laser desorption time-of-ﬂight mass spectrometry Proceedings of
the 39th Conference on Mass Spectrometry and Allied Topics 1991, pl4l6.

Vestling, M. M.; F enselau, C. Surfaces for interfacing protein gel

electrophoresis directly with mass spectrometry Mass Spectrom. Rev. 1995,
14, 169-178.

31

(37)

(38)

(39)

(40)

(41)

(42)

(43)

(44)

(45)

(46)

(47)

Jonsson, G. P.; Hedin, A. B.; Hakansson, P. L.; Sundqvist, B. U. R.; Saeve, B.
G. S.; Nielsen, P. F.; Roepstorff, P.; Johansson, K. E.; Kamensky, I.;
Lindberg, M. S. L. Plasma desorption mass spectrometry of peptides and
proteins adsorbed on nitrocellulose Anal. Chem. 1986, 58, 1084-1087.

Strupat, K.; Karas, M.; Hillenkamp, F.; Eckerskorn, C.; Lottspeich, F. Matrix-
assisted laser desorption ionization mass spectrometry of proteins
electroblotted after polyacrylamide gel electrophoresis Anal. Chem. 1994, 66,
464-470.

Vestling, M. M.; Fenselau, C. Polyvinylidene diﬂuoride (PVDF): an interface
for gel electrophoresis and matrix-assisted laser desorption/ionization mass
spectrometry Biochem. Soc. T. 1994, 22, 547-551.

Vestling, M. M.; Fenselau, C. Poly(vinylidene diﬂuoride) membranes as the
interface between laser desorption mass spectrometry, gel electrophoresis, and
in situ proteolysis Anal. Chem. 1994, 66, 471-477.

Bai, J.; Liu, Y. H.; Lubman, D. M. On-probe puriﬁcation for matrix-assisted
laser desorption/ionization mass spectrometry(MALDI/MS) Proceedings of
the 42nd ASMS Conference on Mass Spectrometry 1994, p976.

Blackledge, J. A.; Alexander, A. J. Polyethylene membrane as a sample
support for direct matrix-assisted laser desorption/ionization mass
spectrometric analysis of high mass proteins Anal. Chem. 1995, 6 7, 843-848.

Worrall, T. A.; Cotter, R. J.; Woods, A. S. Puriﬁcation of contaminated
peptides and proteins on synthetic membrane surfaces for matrix-assisted laser
desorption/ionization mass spectrometry Anal. Chem. 1998, 70, 750-756.
Puchades, M.; Westrnan, A.; Blennow, K.; Davidsson, P. Removal of sodium
dodecyl sulfate from protein samples prior to matrix-assisted laser

desorption/ionization mass spectrometry Rapid Commun. Mass Spectrom.
1999, 13, 344-349.

McComb, M. E.; Oleschuk, R. D.; Chow, A.; Ens, W.; Standing, K. G.;
Perreault, H.; Smith, M. Characterization of hemoglobin variants by MALDI-

TOF MS using a polyurethane membrane as the sample support Anal. Chem.
1998, 70, 5142-5149.

McComb, M. E.; Oleschuk, R. D.; Chow, A.; Perreault, H.; Dworschak, R. G.;
Znamirowski, M.; Ens, W.; Standing, K. G.; Preston, K. R. Application of
nonporous polyurethane (PU) membranes and porous PU thin ﬁlms as sample
supports for MALDI-MS of wheat proteins Can. J. Chem. 2001, 79, 437-447.

McComb, M. E.; Oleschuk, R. D.; Manley, D. M.; Donald, L.; Chow, A.;
O'Neil, J. D. J.; Ens, W.; Standing, K. G.; Perreault, H. Use of a non-porous
polyurethane membrane as a sample support for matrix-assisted laser
desorption/ionization time-of-ﬂight mass spectrometry of peptides and
proteins Rapid Commun. Mass Spectrom. 1997, 11, 1716-1722.

32

 

(48)

(49)

(50)

(51)

(52)

(53)

(54)

(55)

(56)

(57)

(58)

(59)

Oleschuk, R D.; Chow, A. The separation and isolation of gold by selective
extraction and transport through a polyurethane ether-type membrane T alanta
1996, 43, 1545—1554.

Hung, K. C.; Rashidzadeh, H.; Wang, Y.; Guo, B. Use of parafﬁn wax ﬁlm in
MALDI-TOF analysis of DNA Anal. Chem. 1998, 70, 3088-3093.

Bodnar, W. M.; Anderegg, R. J.; Moyer, M. B. Analysis of blotted peptides
and proteins by MALDI TOF mass spectrometry Proceedings of the 42nd
ASMS Conference on Mass Spectrometry 1994, p683.

Nordhoff, E.; Ingendoh, A.; Cramer, R.; Overberg, A.; Stahl, B.; Karas, M.;
Hillenkamp, F.; Crain, P. F. Matrix-assisted laser desorption/ionization mass
spectrometry of nucleic acids with wavelengths in the ultraviolet and infrared
Rapid Commun. Mass Spectrom. 1992, 6, 771-776.

Smirnov, I. R; Hall, L. R.; Ross, P. L.; Haff, L. A. Application of DNA-
binding polymers for preparation of DNA for analysis by matrix-assisted laser

desorption/ionization mass spectrometry Rapid Commun. Mass Spectrom.
2001, 15, 1427-1432.

Beavis, R. C.; Chait, B. T. High-accuracy molecular mass determination of
proteins using matrix-assisted laser desorption mass spectrometry Anal. Chem.
1990, 62, 1836-1840.

Beavis, R. C.; Bridson, J. N. Epitaxial protein inclusion in sinapic acid crystals
J. Phys. D: Appl. Phys. 1993, 26, 442-447.

Reynolds, J. A.; Tanford, C. The gross conformation of protein-sodium
dodecyl sulfate complexes J. Biol. Chem. 1970, 245, 5161-5165.

Cadene, M.; Chait, B. T. A robust, detergent-friendly method for mass

spectrometric analysis of integral membrane proteins Anal. Chem. 2000, 72,
5655-5658.

Vorm, 0.; Mann, M. Improved mass accuracy in matrix-assisted laser
desorption/ionization time-of-ﬂight mass spectrometry of peptides J. Am. Soc.
Mass Spectrom. 1994, 5, 955-958.

Gobom, J.; Schuerenberg, M.; Mueller, M.; Theiss, D.; Lehrach, H.; Nordhoff,
E. Alpha-Cyano-4-hydroxycinnamic acid afﬁnity sample preparation. A
protocol for MALDI-MS peptide analysis in proteomics Anal. Chem. 2001,
73, 434-43 8.

Schuerenberg, M.; Luebbert, C.; Eickhoff, H.; Kalkum, M.; Lehrach, H.;

Nordhoff, E. Prestructured MALDI-MS sample supports Anal. Chem. 2000,
72, 3436-3442.

33

(60)

(61)

(62)

(63)

(64)

(65)

Nedelkov, D.; Nelson, R. W. Design and use of multi-afﬁnity surfaces in
biomolecular interaction analysis-mass spectrometry (BIA/MS): A step toward
the design of SPR/MS arrays J. Mol. Recognit. 2003, 16, 15-19

Nelson, R. W.; Nedelkov, D. Bioactive chip mass spectrometry US. Patent #
6569383, 2003.

Walker, A. K.; Wu, Y.; Tirnmons, R. B.; Kinsel, G. R.; Nelson, K. D. Effects
of protein-surface interactions on protein ion signals in MALDI mass
spectrometry Anal. Chem. 1999, 71 , 268-272.

Ulman, A. An Introduction to ultrathin organic ﬁlms; Academic Press:
Boston, MA. 1991.

Brockman, A. H.; Dodd, B. S.; Orlando, R A desalting approach for MALDI-
MS using on-probe hydrophobic self-assembled monolayers Anal. Chem.
1997, 69, 4716-4720.

Brockman, A. H.; Shah, N. N.; Orlando, R. Optimization of a hydrophobic

solid-phase extraction interface for matrix-assisted laser desorption/ionization
J. Mass Spectrom. 1998, 33, 1141-1147.

34

4x154"

 

Chapter Two: Patterned Monolayer/Polymer Modified Metal Surfaces as
Sample Platforms for Analysis of Dilute or Salt-Contaminated Protein
Samples by MALDI-MS

I. Introduction
This project was partially inspired by Nordhoff and coworkers’ research on

prestructured MALDI sample supports,l where ZOO-um diameter gold spots on
hydrophobic Teﬂon provide anchors for dilute samples during analyses by
MALDI-MS. Due to the water-repellant nature of hydrophobic Teﬂon, the
sample droplet stays preferentially on the hydrophilic gold. With the evaporation
of solvent, analytes are concentrated in the small area deﬁned by the ZOO-pm gold
spot. Compared to the use of a conventional metal MALDI plate (stainless steel
or gold), where the sample droplet spreads out in the 2-mm diameter sample well
(as illustrated in Figure 2.1), use of the prestructured sample support can decrease
the detection limit for an analyte by more than an order of magnitude.

However, only pure samples can be analyzed using this modiﬁed probe, as the
contaminants would be concentrated with the analyte and would cause an even
more serious problem during the desorption/ionization process. I report here the
modiﬁcation of a conventional MALDI plate with a patterned self-assembled
monolayer (SAM) of hexadecanethiol (HDT) prepared by micro-contact printing.2
In one manifestation of these patterned probes, small (ZOO-um diameter) spots of
hydrophilic bare gold patterned in the hydrophobic HDT SAM provide anchors for
droplets of the sample solution and allow subsequent concentration of samples
during solvent evaporation (in a similar fashion to use of a prestructured sample
support). This process yields both decreased detection limits and increased signal

reproducibility, as seen in other work with use of small sample spots,“4

35

sample

 

 

 

droplet
5 reads
Conventional drying
gold sample
plate I“ "1
ZOOOum
sample
droplet beads
/ UP
drying
Prestructured _> [
sample +1 H-
SUPPOWS 200nm

EMMA-j

Teﬂon

 

g cojting ,

 

 

Figure 2.1 Comparison of sample deposition on conventional (top) and

prestructured (bottom) sample supports.

36

but these patterned substrates can be conveniently prepared in nearly any

laboratory.

A second, more versatile manifestation of the patterned gold surface allows
protein puriﬁcation to be combined with concentration of the solute to a small spot.
This method achieves low detection limits even with contaminated samples. In
this manifestation, using a method developed by Crooks and coworkers, I graft
hydrophilic poly(acrylic acid) ﬁhns to the bare gold spots in the otherwise
hydrophobic surface presented by the SAM of hexadecanethiol.5’6 Although
specially modiﬁed sample plates with patterned hydrophilic spots are already
commercially available (Anchorchip, Bruker,
http://www. daltonics. bruker. com/appliedtions/shared/maldiapp1 070anchor WEB. p
dj), the use of PAA allows decontamination of protein samples along with lower
detection limits and increased reproducibility. Thus, the PAA-patterned sample
supports combine the advantages of patterned substratesl and sample
decontamination.7 Additionally, PAA provides a versatile support surface
because it can be derivatized with a wide variety of functional groups, as will be
seen in chapter 3.5

I also report reﬂectance FTIR spectroscopy studies of the mechanism of
protein binding to PAA ﬁlms. Infrared spectroscopy provides a powerful tool for
measuring the amount of adsorbed protein in the ﬁlm as a function of sample

solution conditions such as pH. These studies show that protein binding to PAA

37

 

occurs primarily due to electrostatic interactions, and thus is dependent upon pH
and the concentrations of proteins and salts. Additionally, IR spectra show that
the acidic matrix solution removes proteins from PAA ﬁlms, allowing
co-crystallization with the matrix. These studies also demonstrate that IR spectra
of protein adsorbed on the polymer-modiﬁed surface correlate well with
MALDI-MS signal intensities obtained from corresponding HDT SAM/PAA

surfaces after addition of matrix.

11. Experimental Section
A. Materials and Solutions

Dilute (200 amol/ttL to 18 pmol/ILL) protein solutions were prepared in
deionized H20 (milli-Q, 18 MQcm), 8 M NaOAc, or 1 M NaOAc within 1 h of the
corresponding experiment. For solutions in deionized water, pH values were
adjusted by addition of 1 M HCl or IM NaOH. All proteins and chemicals were
obtained from Sigma or Aldrich unless noted otherwise. Gold-coated substrates
(200 nm of gold sputtered on 20 nm of Cr on Si(100) wafers ) were prepared by

Lance Goddard Associates (Foster City, CA).

B. Fabrication of Patterned Substrates

Poly(dimethylsiloxane) (PDMS) stamps were prepared according to the

2

methods developed by Whitesides and coworkers. Brieﬂy, the desired pattern

38

 

(ZOO-um spots separated by 5 mm) was printed on an ordinary transparency sheet
using a laser printer. A “pre-baked” (30 min, 60 °C) photoresist (AZ P4210,
Clariant) spin-coated on a Si wafer was then selectively exposed to UV light (Hg
lamp) using the transparency as a mask, and the photoresist was immersed in
developer (AZ 421K, Clariant) for 1 min to yield ZOO-um diameter protrusions.
After “post-baking” (30 min, 60 °C), a PDMS elastomer solution (Sylgard 184,
Dow Corning) was poured onto the photoresist template and cured overnight at
60°C. The PDMS stamp, consisting of ZOO-um diameter indentations, was then
peeled off the photoresist. The stamp can be used more than 500 times without
losing the patterning resolution.

Prior to the printing of patterns, gold-coated substrates were ﬁrst cleaned
for 15 min in a UV/ozone cleaner (Boekel 135500), rinsed with deionized water for
~10 s, and dried with N2. PDMS stamps were “inked” with HDT by swabbing the
stamp with a ~10 mM solution of HDT in ethanol, and then dried by a stream of
N2. The stamp, still moistened with residual HDT, was then gently contacted to
the gold substrate (as suggested early in Figure 2.2), and air bubbles were removed
by lightly pressing on the stamp. After ~30 s, the stamp was removed, leaving a
thin layer of HDT. Excess HDT was rinsed away with ethanol and water to yield
a pattern of gold spots in the HDT SAM (HDT SAM/Au).

To further modify the patterned HDT SAM surface (step 2 in Figure 2.2),

we chemically grafted PAA onto the bare gold spots according to the procedure of

39

B . I

Crooks and coworkers.6 Brieﬂy, we dipped the gold-coated wafer into 1 mM
mercaptoundecanoic acid (MUA) in ethanol for 60 s, and then rinsed it with water
and ethanol. After activating the MUA using ethyl chloroforrnate, we chemically
attached amino-terminated poly(tert-butyl acrylate) to the MUA linker. Finally,
we hydrolyzed the tert-butyl ester groups using methanesulfonic acid in

dichloromethane.

C. Sample Preparation Prior to Analysis by MALDI-TOF-MS

For analysis of dilute, pure protein samples on HDT SAM/Au-patterned
surfaces, 0.25-0.5 uL of the sample solution was applied to a bare gold spot, after
which an equal volume of matrix solution (10 mM 2,5-dihydroxybenzoic acid
(DHB) in deionized water) was added and the mixture was allowed to dry. When
using HDT SAM/PAA surfaces, 0.25 pL of salty protein solution was deposited on
the PAA spot and the droplet was allowed to stand for 3-5 min. Before the drop
dried, the surface was rinsed with copious amounts (~10 mL) of MilliQ water and
dried with N2. Subsequently, 0.25 uL of matrix solution was added and allowed
to dry. The gold wafers were attached to a disposable MALDI plate using
double-sided tape or superglue. When the sample preparation process required

rinsing with water, the wafers were always attached to the plate

40

 

. .s n * . LEFT TIM—r .".:T'““. '1‘

PDMS stamp Negatively

HDTS AM charged PAA
Remove / \ Deposition

stamp £3553 £3335
f I"’1 of PAA
200 um
Gold plate HDT Addition of salt-
contaminated
sample solution

M t' - l t
a "X ana Y e Sample droplet

   
 
 

 

1) (l COCO/513' Positively charged
I l} 9 analyte

100 i '
V \‘1

Mg: Rinse and Solvent

 

evaporation

add matrix

Figure 2.2 Schematic flowchart for the preparation of a patterned HDT-
SAM/PAA sample probe and subsequent on-probe cleaning and
concentrating of a salt-contaminated protein sample for analysis by MALDI-
MS.

41

after drying with N2. as otherwise it took a long time (~30 min in extreme cases) to evacuate

the sample inlet system prior to analysis by MALDI-MS.

D. Instrumentation

Mass spectra were obtained with a PE Biosystems Voyager Elite or STR
MALDI TOF mass spectrometer using an accelerating voltage of 25 kV, a 95%
grid voltage, 0.05 % guidewire voltage, and an extraction delay time of 100-150
nsec to accumulate ion current associated with 30-50 laser pulses. Ellipsometric
experiments were performed using a rotating analyzer spectroscopic ellipsometer
(J.A. Woollam) and assuming a ﬁlm refractive index of 1.5. Reﬂectance FT IR
spectra were obtained with a Nicolet Magna 560 spectrophotometer with a Pike
grazing angle (80°) accessory. The spectrophotometer was housed in a glove box

to minimize interference from water vapor.

III. Results and Discussion

A. Patterned HDT SAM/Au surfaces as sample plates

The hydrophilic bare gold spots in HDT SAMs provide anchors for the
sample/matrix droplet, allowing concentration of a sample by deposition of the
solute to a smaller surface area during the evaporation of solvent. This improves
both the detection limit and signal reproducibility in MALDI-MS. The detection

limit (signal-to-background ratio 2 3) for ribonuclease A (RNase A) in

42

conventional MALDI (2-mm diameter sample wells) is approximately 100 frnol,
while the detection limit using a patterned HDT SAM/Au surface is about 1 frnol
as illustrated by the mass spectra in Figure 2.3.

Detection limits also decrease for insulin, insulin chain A, and insulin
chain B (0.3-3 frnol, see Figure 2.4, for representative spectra). From the spectra
in Figure 2.4, one may have the impression that the detection limit can be further
lowered as the signal-to-noise ratio is much higher than 3: 1. Indeed, signals from
0.2 fmol insulin and other small peptides have been observed, but not on a routine
basis; this may be due to the fact that at very low concentrations, most of the
peptides are adsorbed to the pipette tip and, thus, are not available for analysis in
MALDI-MS. A similar phenomenon was observed in another study of a
prestructured sample probe.l In this chapter, I provide only the spectra that can be
obtained routinely. The surface area in a 2—mm sample well is 100 times that of a
patterned 0.2-mm diameter gold spot, so in principle, the sample concentration on
the patterned surface should be 100 times higher than that on a conventional
MALDI plate when equal amounts of analytes are deposited. This surface area
ratio may explain the improvement in detection limit for RNase A when using the
patterned HDT SAM/Au surface. However, detection limits for smaller proteins
decrease by a factor of only ~10 when using the patterned plate, indicating that
surface area is not the only variable affecting sensitivity. Previous studies with

small sample spots also showed a ~10-fold decrease in detection

43

W I.- I 547 Mini TA.‘ WW. ~79}?!

 

 

Intensity

 

 

 

l l I
13000 14000 15000 16000

m/z
Figure 2.3 MALDI mass spectra of RNase A obtained using different sample
probes. (A) 600 fmol RNase A using a conventional 2-mm diameter sample

well (70 mM DHB as matrix); (B) 1 fmol RNase A using a ZOO-mm diameter
Au spot in a patterned HDT SAM (10 mM DHB as matrix).

44

R.I.

 

VWN Wu.» .—.-_.

 

, .1 ll

1
«WWW—w «p. “MM-WW» -v- “u-

 

 

 

 

l
2200

3200

m/z

4200

5200

6200

Figure 2.4 Mass spectra of (A) 3 fmol insulin chain B and (B) 2.5 frnol
insulin. Spectra were obtained from ZOO-mm diameter Au spots in

HDT SAMs (10 mM DHB as matrix).

45

limits.l

Another important experimental factor is the matrix solution because both
matrix and analyte molecules are concentrated when solvent evaporates. If the
initial matrix concentration is too high, the size of the solid residue is usually larger
than the dimensions of the underlying gold spot, and it is difﬁcult to obtain
reproducible signals. The matrix solvent is also an issue. If the solvent evaporates
too quickly, as can happen with acetone for example, the size of the solid residue is
often greater than the size of the bare gold spot. Usually, matrix solutions made
with water, 70% acetonitrile and 30% water, or 50% acetone and 50% water give
reasonable sample drying rates and small sizes of solid residues.

As the sample spot size approaches the area illuminated by the laser, signal
reproducibility also improves compared to that achieved from a conventional
MALDI sample plate. For replicate measurements with RNase A, insulin, insulin
chain A, and chain B made at several different Au spots on a patterned probe, the
relative standard deviation in signal intensity is less than 30% (four samples
analyzed in each case, see Table 2.1). In these experiments, the operator ensured
that the laser beam illuminated matrix crystals, and signal was observed in every
case. As the sample size is about 3 or 4 times larger than the cross section of the
laser beam (i.e., 25-33% of the sample is illuminated on each laser shot), usually 3
or 4 spectra (a spectrum consists of the averaged signal following 30 to 50 laser

shots) can be obtained within one sample well, and we can get usable signal in

46

Table 2.1 MALDI-MS signal intensities for replicate samples on patterned
HDT SAM/Au surface

 

Relative
Standard Standard
Protein Intensity Intensity Intensity Intensity deviation Average Deviation
Insulin
chainA
(16fmol) 3720 3811 4764 4769 579 4266 14
Insulin
chain B
(16fmol) 23000 21000 16000 21000 2986 20250 15
Insulin
(32 fmol) 27000 15000 16000 22000 5598 20000 28
RNaseA
(50 fmol) 2206 4067 3564 2789 823 3156.5 26

 

 

most cases (>75%). Excluding spots where no measurable signal is obtained
(<25%), the relative standard deviation of measurements at different spots within
the same ZOO-um diameter sample well is less than 30%. We have not been able
to obtain similar reproducibility with the same dried-droplet-method sample
preparation procedure on a stainless steel probes. In the latter cases, variation in
signal intensities can be as high as several orders of magnitude; very often no
signals can be detected in speciﬁc sample locations (>95% in worst cases, i.e.,
signal might be detected from as few as 1 position out of 20). The variation in
signal intensities from the patterned substrates is comparable to that obtained
previously using probes modiﬁed with 300-um diameter matrix/nitrocellulose

spots, in which case samples were deposited with a piezoelectric pipette.3

47

The ideal diameter of the solid residue should be even smaller than 200 um
so that the laser beam would illuminate the whole sample area. (The laser beam
that we use has a diameter of approximately 100 pm, and the intensity of the laser
is less at the edge of the beam than in the center.) Unfortunately, the sample
deposition process provides an effective minimum limit on the size of the
hydrophilic gold spot. To detach the sample from the pipette tip, there must be a
sufﬁciently strong attraction between the surface and the droplet, and this
attraction is too weak to deposit samples on hydrophilic gold spots with diameters
<100 pm. A similar phenomenon has been reported in the use of prestructured
supports.7 We have experimented brieﬂy with 500-, 200-, and 100-um Au spot
diameters, and the ZOO-pm diameter Au spots yielded the most reproducible

signals and lowest detection limits.

B. Patterned HDT SAM/PAA Surfaces as Sample Plates

Although patterned HDT SAM/Au substrates give low detection limits, they
are not capable of decontarninating protein solutions. In fact, contaminants will
be concentrated with this system. To prepare patterned surfaces capable of both
concentrating and desalting solutions, we grew PAA ﬁlms in the bare Au holes of
patterned HDT SAMs (see Figure 2.2). In the decontamination procedure, we
deposited ~0.25 uL of a salt-containing solution on the PAA spot, waited for 3-5

minutes to allow partial evaporation of solvent, rinsed the spot with water, dried it

48

with N2, and subsequently added matrix solution to the sample residue. The sample
size, after evaporation of matrix solvent, is about the size of the underlying PAA
spot. Using these ﬁlms, the detection limit for 1 M NaOAc-contaminated insulin
is approximately 25 frnol (Figure 2.5); this value is 20- to 100-fold lower than
detection limits reported with the use of non-patterned polylysine surfaces8 or
SAM surfaces.9

Patterned HDT SAM/PAA sample plates also allow low (20-50 frnol)
detection limits for insulin chain A, insulin chain B, and RNase A in the presence
of 1 M NaOAc. Both conventional MALDI sample wells and patterned HDT
SAM/Au plates (with or without rinsing) give little or no signal with application of
as much as 1 pmol of RNase A in the presence of 1 M NaOAc. Detection limits
obtained using patterned HDT SAM/PAA sample plates for salty solutions are even
lower than uncontaminated-solution detection limits obtained using conventional
MALDI plates. This is due, again, to concentration of the protein on the small
sample spot. Although contaminants in the sample solution are also concentrated
during evaporation of the solvent, they can be rinsed off easily with water while the

analyte molecules remain bound to PAA.

C. Mechanism of Desalting and Incorporation of Proteins into the
MALDI Matrix

In the decontamination and concentration process, positively charged

49

proteins most likely bind to the surface due to electrostatic interactions with PAA,
which contains many carboxylate groups at pH values above its pKa (54.5).lo
Because many proteins can contain multiple positive charges, they adsorb to the
negatively charged surface more strongly than simple salts, and thus salts can be
removed preferentially by rinsing with water. Subsequent addition of an acidic
MALDI matrix should remove the protein from the ﬁlm by protonating PAA,
thereby eliminating electrostatic interactions. Reﬂectance FTIR spectroscopy
studies conﬁrm this mechanism of protein binding and incorporation of analyte
into the MALDI matrix.

If electrostatic interactions are essential for decontamination, protein
absorbances in IR spectra and signal intensities in mass spectra should depend on
the pH of the sample solution. At pH values where the protein is highly positively
charged and the ﬁlm is negatively charged, maximum adsorption should occur.
Thus, the solution pH should be below the pI of the protein (9.7 for RNase A,
calculated by GPMAW software), but above the pKa of PAA (4.5). Figures 2.6
and 2.7 show that maximum amide absorbances occur at an adsorption pH of ~5.5.
(Amide absorbances should be approximately proportional to the amount of
adsorbed protein.) The absorbance maximum at pH 5.5 represents a compromise
between achieving the highest number of positive charges on the protein and the
highest negative charge density in the ﬁlm. These IR absorbance data are

consistent with mass spectral data, which also show maximum signal intensity

50

when the protein solution has a pH of 5.5 (Figure 2.7). The pH for maximum
adsorption will, of course, vary somewhat from protein to protein, depending on p1
values.

Protein adsorption should also depend on the concentration of both
protein and salt in sample solutions. Figure 2.8 shows a plot of amide absorbance
as a function of the concentration of RNase A in the solution to which the sample
plate was exposed (solution pH of 7.0). The plot suggests that adsorption increases
linearly as RNase A concentrations increase from 0.1 to 10 pmol/uL. Compared
with the acid carbonyl absorbance (~0.003 for a protonated PAA ﬁhn), the amide
absorbance (0.05) for a sample plate exposed to 20 pmol/uL RNase A solution
suggests that multilayers of protein are forming in highly extended PAA chains.
Ellipsometric measurements conﬁrm multilayer formation as ﬁlm thickness

increases 10-fold (from 25 A to 250 A) after exposure

51

 

 

 

Intensity

l
l:
I
II

, “ ’1 - .
V'VW'yMTVM'QI’MMMUW \M’JW VVIJWNW-Jxm

M f-M A.v’ \

5000 5500 6000 6500

m/z
Figure 2.5 MALDI mass spectrum of insulin adsorbed from 0.25 uL of
solution containing 25 fmol insulin and 1 M NaOAc. The sample was
applied to a ZOO-pm diameter PAA spot in a HDT SAM and rinsed with
water. DHB (10 mM) was added as matrix.

 

 

 

52

 

Amide Absorbance

 

 

 

 

2000 1750 1500 1250 1000 l
Wavenumbers (cm'1)

 

Figure 2.6 Reﬂectance FTIR spectra of PAA ﬁlms after exposure
to a 1-pM RNase A solution for 30 min at different pH values: (A)
pH 5.5; (B) pH 1.6; (C) pH 12.2. Films were rinsed with water prior
to measurements. In these studies, substrates were not patterned,
but completely coated with PAA to provide enough capacity to
achieve a sufﬁcient signal-to—noise ratio in the IR spectrum.

53

 

 

 

 

 

 

 

 

0.016
. A MALDI Signal Intensity E]
0'0“ Cl Amide Absorbarce A :02”
8 2
5 0012- 9,
C
E 0010- 13 g
8 ::
<5) 0008- g ‘ 2%“
U _
'g 0006- 9
< ‘3 <
0.0041 ‘ D 2
A
0.002 5 - . .
0 2 4 6 8 10 12 14
pH

Figure 2.7 Amide absorbances in reflectance FTIR spectra and the protonated
protein signals obtained using MALDI-MS after exposure of sample plates to
RNase A solutions (1 11M, no salt) at different pH values, followed by water rinsing.
Mass spectra were taken on patterned HDT SAM/PAA plates (10 mM DHB as
matrix) and FTIR spectra were measured on plates coated only with PAA.

54

 

 

 

 

 

 

 

0.06.... ---L.,ﬁ,
0.05 ~ A 2:
8 . 'g
g 0.04 1 . 1 g
2 0.03 « A E
f, 0.02 - A .09;
:9 ._
<EE 0.01 « 8 3
0,00 - ‘3 Amide Absorbance A g
MALDlSignallntensity O ’

 

0 5000 10000 15000 20000 25000
Concentration fmol/pL

Figure 2.8 Amide absorbances in reﬂectance FTIR spectra and the protonated
protein signals obtained using MALDI-MS after exposure of sample plates to RNase
A solutions at different concentrations (pH 7, no salt), and water rinsing. Mass
spectra were taken on patterned HDT SAM/PAA plates (10 mM DHB as matrix) and
FTIR spectra were measured on plates coated only with PAA.

55

of PAA to 20 pmol/uL of protein. Evidence for the formation of multilayers helps
to explain why patterned surfaces are so successful in concentrating protein
analytes from salty samples.

MALDI signals should be proportional to the amount of protein adsorbed
on the surface. However, MALDI signal intensities increased more rapidly with
concentration than did amide absorbances. This may be due to the fact that in the
MALDI experiments, the solution is concentrated on the small spot, while in the IR
experiments, no concentrating of protein occurs (the non-pattemed PAA-coated
sample plate was immersed in a protein solution); on the other hand, MALDI does
not necessarily correlate well with quantity of analyte in the absence of an internal
standard.

Using reﬂectance FT IR spectroscopy, we also examined the effect of salt
on protein adsorption to PAA. The presence of 1 M NaOAc decreases amide
absorbances due to RNase A adsorption by 90% (adsorption pH of 8.5). At high
concentrations, salts are evidently able to compete with RNase A for binding sites.
This is consistent with a mechanism of protein adsorption based on electrostatic
forces.

To examine whether proteins are removed from PAA ﬁlms upon exposure
to matrix solutions, we ﬁrst immersed a PAA-ﬁlm-covered gold wafer in a
lpmol/pL RNase A solution (pH=7) for ~12 h, and then rinsed the wafer with

water, dried it with N2, and measured the IR spectrum of the ﬁlm. Next, we

56

I
q

mamas-n Jiam“. IPA-m
4

covered the entire surface with 5 pL of 20 mg/mL DHB in water. After
evaporation of the matrix solution solvent, we rinsed the surface with water, dried
it with N2, and measured a second IR spectrum. The diminution in amide
absorbance (>90% decrease) following matrix deposition and rinsing with water
showed that most of the protein molecules were removed from the ﬁlm. These
results suggest that the acidic matrix protonates the carboxylate groups in the PAA
ﬁlm, thereby breaking the ionic interactions between PAA and the positively
charged proteins, and allowing protein incorporation into the matrix. This is
consistent with the pH-dependent adsorption discussed above.

One of the drawbacks of the PAA system described herein is that only
positively charged proteins can be effectively captured from salt-containing
solutions. However, we can easily derivatize the PAA substrate to produce an
amine-terminated polymer ﬁlm,11 which would be positively charged at neutral pH
(chapter 3). Using amidation or esteriﬁcation, one could also modify the surface
with speciﬁc bioreactive groups, such as antibodies, that have speciﬁc afﬁnity for

antigens in salt-containing mixtures.12

IV. Conclusions
Micro-contact printing of HDT SAMs affords rapid formation of patterned
MALDI probes that decrease the detection limit and increase reproducibility in

MALDI-MS. Graﬁing of PAA into bare gold spots of patterned SAMs increases

57

A_—-‘.—.—-—A M

 

the versatility of this system and allows desalting of sample solutions prior to
MALDI. The combination of a patterned surface and a polymer with affinity for
charged proteins yields low-frnol detection limits even in the presence of 1 M
NaOAc. Reﬂectance FTIR spectra conﬁrm that protein binding occurs due to
electrostatic interactions. These spectra indicate that protein binding is dependent

upon the pH of sample solutions as well as on protein and salt concentration.

58

References:

( 1) Schuerenberg, M.; Luebbert, C.; Eickhoff, H.; Kalkum, M.; Lehrach, H.;
Nordhoff, E. Prestructured MALDI-MS sample supports Anal. Chem. 2000,
72, 3436-3442.

(2) Xia, Y.; Whitesides, G M. Soft lithography Angew. Chem. Int. Edit. 1998,
37, 550-575.

(3) Miliotis, T.; Kjellstrom, S.; Nilsson, J.; Laurell, T.; Edholm, L.-E.;
Marko-Varga, G Ready-made matrix-assisted laser desorption/ionization
target plates coated with thin matrix layer for automated sample deposition
in high-density array format Rapid Commun. Mass Spectrom. 2001, 16,

l 1 7- 126.

(4) Little, D. P.; Cornish, T. J .; O'Donnell, M. J .; Braun, A.; Cotter, R. J .;
Koester, H. MALDI on a chip: analysis of arrays of low-femtomole to
subfemtomole quantities of synthetic oligonucleotides and DNA diagnostic
products dispensed by a piezoelectric pipet Anal. Chem. 1997, 69,
4540-4546.

 

(5) Bruening, M. L.; Zhou, Y.; Aguilar, G; Agee, R.; Bergbreiter, D. E.; Crooks,
R. M. Synthesis and characterization of surface-grafted, hyperbranched
polymer ﬁlms containing ﬂuorescent, hydrophobic, ion-binding,
biocompatible, and electroactive groups Langmuir 1997, 13, 770-778.

(6) Lackowski, W. M.; Ghosh, P.; Crooks, R. M. Micron-scale patterning of
hyperbranched polymer ﬁlms by micro-contact printing J. Am. Chem. Soc.
1999, 121, 1419-1420.

(7) Gobom, J .; Schuerenberg, M.; Mueller, M.; Theiss, D.; Lehrach, H.;
Nordhoff, E. Alpha-cyano-4-hydroxycinnamic acid afﬁnity sample
preparation. A protocol for MALDI-MS peptide analysis in proteomics
Anal. Chem. 2001, 73, 434-438.

(8) Zhang, L.; Orlando, R. Solid-Phase Extraction/MALDI-MS: Extended
ion-pairing surfaces for the on-target cleanup of protein samples Anal.

Chem. 1999, 71, 4753-4757.

(9) Warren, M. E.; Brockman, A. H.; Orlando, R. On-probe solid-phase
extraction/MALDI-MS using ion-pairing interactions for the cleanup of

59

peptides and proteins Anal. Chem. 1998, 70, 3757-3761.

(10) Peez, R. F.; Dermody, D. L.; Franchina, J. G; Jones, S. J .; Bruening, M. L.;
Bergbreiter, D. E.; Crooks, R. M. Aqueous solvation and functionalization
of weak-acid polyelectrolyte thin ﬁlms Langmuir 1998, 14, 4232-4237.

(11) Xiao, K. P.; Kim, B. Y.; Bruening, M. L. Detection of protamine and
heparin using electrodes modiﬁed with poly(acrylic acid) and its amine

derivative Electroanalysis 2001, 13, 1447-1453.

(12) Nelson, R. W. the use of bioreactive probes in protein characterization
Mass Spectrom. Rev. 1997, 16, 353-376.

60

Chapter Three: Use of Polymer-Modiﬁed MALDI-MS Probes to Improve
Analyses of Protein Digests and DNA

1. Introduction:

Several recent reports demonstrate the effectiveness of on-probe sample
puriﬁcation prior to analysis by MALDI-MS."l7 In these puriﬁcation procedures,
surface-modiﬁed MALDI probes bind speciﬁc analytes, while allowing contaminants
to be washed away. By reducing problems of signal suppression or adduct formation,
such techniques facilitate successful detection of proteins and DNA during analysis of
complex mixtures or solutions contaminated with salts, surfactants, or urea. As
mentioned in Chapter 1, successful methods for MALDI sample probe modiﬁcation

3,12

include adsorption of self-assembled monolayers (SAMs) and deposition of

polymers (e.g., polyvinylidenediﬂuoride (PVDF),”‘18‘2‘ naﬁon,22‘23 and Teﬂon.5’24) or
water-insoluble ﬁlms of matrix.5’24’28

Chapter 2 described preparation of polymer-modiﬁed surfaces for MALDI probes
using a combination of micro-contact printing and polymer grafting.29 Deposition of
poly(acrylic acid) (PAA) in ZOO-um diameter gold spots in a hydrophobic monolayer
allows simultaneous concentration and puriﬁcation of protein from salt-contaminated
solutions. Self-centering of sample droplets onto the hydrophilic PAA spots affords
sample concentration on the small (ZOO-um diameter) area during solvent
evaporation,30 while the carboxylate groups of PAA selectively bind proteins having

net positive charges, even in the presence of 8 M NaOAc. In this chapter, I describe

enhancements to the utility of MALDI probes patterned with PAA through further

61

chemical (modiﬁcation) elaboration of these polymeric surfaces. Derivatization of
PAA can be controlled to produce one of several different surfaces on the MALDI

probe to improve the analysis of protein digests and to remove adducts from DNA.

The availability of a variety of probe surfaces can be especially useful in analyzing
protein digests. In conventional MALDI-MS of protein digests, many proteolytic
peptides cannot be detected because of signal suppression by salt and other
contaminants such as urea or guanidine hydrochloride. This leads to low sequence
coverage that sometimes makes protein identiﬁcation difﬁcult. To solve this problem,
researchers often resort to HPLC separation of the digest mixture, but this approach
sacriﬁces the high throughput potential usually available in MALDI-MS. To guide
development of a more rapid technique that still enhances the number of proteolytic
fragments detected during analysis of a protein digest by MALDI-MS, we evaluated
sample probe surfaces modiﬁed with cationic, anionic, or hydrophobic groups. Each
of these surface types yielded a different mass spectrum with the deposition of a given
protein digest, followed by rinsing and subsequent addition of matrix. For example,
after the analysis of partially digested myoglobin by MALDI-MS, combining data
from a patterned PAA surface and a conventional stainless steel (SS) probe yielded 2.4
times more identiﬁable proteolytic fragments than did results of analysis using only the
SS probe. The effect was less dramatic with a larger protein, bovine serum albumin
(BSA), but sequence coverage still increased from 61.3 to 74.5% when including data

obtained from individual use of several different surface-modiﬁed probes.

Use of derivatized PAA ﬁlms that present polycationic surfaces is also effective

62

for on—probe decontamination of DNA samples prior to analysis by MALDI-MS. The
salt-tolerance level for analysis of DNA samples by MALDI-MS is usually around 100
mM.3 ' At higher salt concentrations, both the peak intensity and resolution decrease
due to formation of DNA adducts consisting of multiple cations (see chapter 1, Figure
1.3). The addition of ammonium citrate (AC) to the matrix solution helps to solve
this problem because the ammonium ions displace the Na+ or K+ bound to DNA; the
bound NH4+ is lost subsequently as neutral ammonia at the ionization/desorption
stage.”34 However, puriﬁcation of DNA samples that contain salt concentrations
>100 mM is necessary even when AC is added.3 ' Smirnov et al. previously showed
that pure polyethylenimine (PEI) or poly(pyrolidone) ﬁlms are suitable for on-probe
puriﬁcation of DNA samples prior to analysis by MALDI-MS.lo Here, we show that
derivatization of patterned PAA ﬁlms with PEI also yields polycationic surfaces that
adsorb DNA. Simple rinsing of the dried sample on the PAA-PEI surface removes
salts (800 mM NaOAc), but not DNA; this procedure leads to regeneration of signals

from protonated DNA.

II. Experimental Section
A. Materials and solutions

All proteins and chemicals were obtained from Sigrna-Aldrich unless noted
otherwise. A 24-mer (TTI CAC CCC TCT ATG ACC GCT ACC), 20-mer (AAC
CTT GGA ACC TI‘G GAA CC), 7-mer (TTT TTT T), and 10-mer (AAC CTT GGA A)

of DNA were prepared in the Macromolecular Structure Facility at Michigan State

63

University. A DNA l4-mer (TTG GCC AAT TCC GG) and a DNA 12-mer (CCG
GAA TTG GCC) were obtained from Gene Link. Protein or DNA solutions were
prepared in deionized H20 (milli-Q, 18 Macm) or 800 mM NaOAc. Gold-coated
substrates (200 nm of gold sputtered on 20 nm of Cr on Si(100) wafers) were prepared

by Lance Goddard Associates (Foster City, CA).

B. Protein Digestion

Horse heart myoglobin, BSA, and ribonuclease A (RNase A) were digested by
trypsin (Promega) according to the procedures provided by the supplier. Brieﬂy,
~20ug of protein were ﬁrst denatured either by 50 uL 6 M urea in Tris buffer (diluted
to 0.6 M urea before addition of trypsin, ﬁnal volume ~500 uL) or by heating at 65 °C
in 50 uL water for 30 min (ﬁnal volume ~ 500 uL by addition of 50 mM ammonium
bicarbonate). If the protein contained disulﬁde bonds (BSA and RNase A ), it was
then reduced by 5 pL of 10 mM dithiotheitol (DTT) and alkylated with 20 11L of 100
mM iodoacetamide. All proteins were ﬁnally incubated with 0.5-1 pg trypsin at 37°C
overnight. The digestion was stopped by addition of acetic acid to achieve a pH of ~3

or storage of the digestion mixture in a -80 °C freezer.

C. Fabrication of patterned substrates

The procedure for preparing patterned hexadecanethiol (HDT) SAMs on gold was

29,35

reported previously. Brieﬂy, we ﬁrst prepared a polydimethylsiloxane (PDMS)

36

stamp according to the methods developed by Whitesides and coworkers. Wetting

64

of the PDMS stamp with ~10 mM HDT in ethanol, followed by pressing the stamp
onto an ozone-cleaned gold wafer for 30 sec transferred the pattern of the stamp onto
the gold surface. After rinsing with ethanol and water, immersion of the patterned
wafer in a l-mM ethanolic solution of mercaptoundecanoic acid (MUA) for 60 8
resulted in a patterned MUA/HDT SAM. Activation of the carboxylic acid groups of
MUA with ethylchloroformate, grafting of amino-terminated poly(tert-butyl acrylate)
(PTBA) to these groups, and subsequent hydrolysis in methanesulfonic acid solution
yielded hydrophilic PAA37 spots surrounded by hydrophobic regions of HDT. To
modify the patterned surfaces, we activated the PAA with ethylchloroformate again
and allowed this surface to react with 20 mg/mL PEI in N,N-dimethylformamide
(DMF) for 1 h. Rinsing with ethanol and water removed physisorbed PEI, and
subsequent rinsing of the surface with water or dilute HCl protonated the amine groups
in PEI, thereby creating hydrophilic, positive spots in the hydrophobic HDT SAM.
PEI can also be attached directly to an ethylchloroformate-activated MUA/ HDT SAM
to produce a MU A-PEI-modiﬁed probe. In an alternative chemical elaboration
procedure, we immersed the patterned HDT SAM/ PAA ﬁlm into an aqueous solution
of 100 mM Fe(N03)3 or Fe(C104)3 for 15 min to form PAA-Fe“, complexes and

subsequently rinsed the ﬁlm with water.

D. Sample Preparation prior to Analysis by MALDI-TOF-MS
For analysis of protein digests, four types of surfaces were used: conventional SS,

gold modiﬁed with hydrophobic HDT SAMs, and gold patterned with HDT SAM/

65

PAA or HDT SAM/ PAA-PEI ﬁlms. A droplet of protein digest solution (0.25 uL)
was applied to each modiﬁed surface and allowed to air dry. Subsequently, the
sample area was rinsed with water (~10 mL, from a wash bottle) for about 5 s (to
remove interfering contaminants) and dried with N2. A matrix solution (0.25 to 0.5
uL of 50:50:01 v:v:v acetonitrile:water:triﬂuoroacetic acid saturated with
or-cyano-4-hydroxycinnamic acid, or-CHCA) was then added to the spots and allowed
to air dry. In the case of the conventional SS surface, no water rinsing was included,
and matrix was added before the sample dried. For the analysis of salt-contaminated
DNA solutions, a 0.25-uL sample was applied to a PAA-PEI or MUA-PEI spot, air
dried, and then rinsed with water from a wash bottle for ~5 3. Subsequently, a 10 to
50 mM AC solution saturated with 3-hydroxypicolinic acid (3-HPA) was used as a

comatrix.

E. HPLC Separation of Protein Digests

Aliquots (10 to 50 pL) of myoglobin tryptic digests were injected into an HPLC
and subjected to reverse-phase separation with a 5% to 95% solvent B gradient elution
in 60 minutes. (Solvent A: 95% H20, 5% acetonitrile, 0.1% triﬂuoroacetic acid;
Solvent B: 5% H20, 95% acetonitrile, 0.1% triﬂuoroacetic acid) Each elution

fraction (25 to 80uL) was collected, evaporated to 5 to 10 1.1L with a Speed Vac, and

analyzed by MALDI-MS.

G. Instrumentation and Data Analysis

66

Mass spectra were obtained with a PE Biosystems Voyager STR MALDI TOF
mass spectrometer using an accelerating voltage of 20 kV, a 94% grid voltage, a 0.05%
guidewire voltage, and an extraction delay time of 100-150 nsec to accumulate ion
current associated with 50-250 laser pulses. A 2-point m/z calibration for protein
digests was usually performed using intense peaks corresponding to previously
identiﬁed peptides. Peaks were selected according to the following protocol: the root
mean square (RMS) noise level as a percentage of the base peak was ﬁrst calculated
for the whole spectrum, and this value was multiplied by 20 to obtain the percent peak
area threshold for peak detection. Signals were assigned to speciﬁc peptides using the
Mascot program (http://www.matrixscience.com) with a mass tolerance of 1 Da.
Peptides corresponding to signals consisting of a small shoulder or a peak with height
<3 times the local noise were not counted as peptide matches. (This was the case for
fewer than 6% of identiﬁed peaks. Molecular weight search (MOWSE) scores were
calculated without removing low-intensity peaks.) Ellipsometric determinations of
polymer thickness were performed with a rotating analyzer spectroscopic ellipsometer
(J .A. Woollam, M—44), assuming a ﬁlm refractive index of 1.5. Reﬂectance FTIR
spectra were obtained with a Nicolet Magna 560 spectrophotometer using a Pike
grazing angle (80°) accessory; the spectrophotometer was housed in a glove box to

minimize interference from water vapor.

67

III. Results and Discussion
A. Analysis of Protein Digests by MALDI-MS

Mass spectra of a tryptic digest of myoglobin show that use of different sample
probe surfaces permits the detection of different, and sometimes complementary,
arrays of proteolytic fragments. Figure 3.1a shows a MALDI mass spectrum of a
tryptic digest of myoglobin that was denatured with 6 M urea and analyzed using a
conventional SS probe. The signal at m/z 1608 is particularly strong, and small
signals from ten other protonated proteolytic fragments are also visible. The
dominating signal at m/z 1608 may arise from selective tryptic cleavage that leads to a
high concentration of this peptide, or this fragment may have an especially high
ionization or detection efficiency. To gain further insight into the MALDI process,
we used HPLC to quantitatively assess the distribution of proteolytic fragments in the
tryptic digest of myoglobin. The chromatographic peak (data not shown)
representing the tryptic fragment responsible for the mass spectral peak at m/z 1608
was not particularly intense (compared with other peaks), indicating that the
concentration of this peptide in the digest mixture was not unusually abundant. This
observation suggests that this fragment (m/z 1608, residues 17-31, see Table 3.1 for
composition) has a higher ionization or detection efﬁciency than other components in
the digestion mixture. Ionization of this peptide (m/z 1608) may be especially

tolerant to the presence of urea (see below).

68

.oEEoe wEeeamotoo 2: co wows—Lu mo Bass: .89 2: m_ :owbﬁU: Kozaomcomv 0338 354,75 3 35:28
298 Eoﬁozaoﬁé a m_ :xer mumm: ”2338.2: use :99852: 5953 3585.6 05 2 12 «zoo: 6253 96:83:00 2: L8 38an
33> NE. 33285 65 m_ L283)? 5585 m L0 $95 32: 3 38833 628on $9: 5452 05 E @3830 32an wEecoamotoQ 2:

Lo 2:? NE. 2: m_ 209852.. ”8:268 :BEwQAE 2: E 598:: 2632 Box oEEm mEneo new $5.83 2: ﬂ :ncméﬁm: 638 £5 E

 

 

69

OOqumM>M<<EZmn3mq< o cmmm- emmd- Gawmm www.mwmm mm _ - em _
MH<ImO<Am¥Am<mIIO¥¥ m oxen owed- v.2 _N Redo—N co - an
¥P<ImO<AmMAm<MIIO¥ V cm? mmd- 803— «9:3. 8 - E.
MmIA>E:<QmEmI; _ ovoe- .24.? 2.2%— 332: w: - mo—
MH<ImO<AmMAm<MIIO m oewv md- e302: Summw— ea - ow
M03>ZA>OOBmOquO _- co mmd- e _ 2 Sam _ M: e _ -—
MQn—Mmqhmmmprnﬁ o owa mmd- mm. 33 3.83 we - mm
m5>mOOIOSQ<m> T o: 86 meom: 5.80— _m .. E
MFZ<OO<Q<OLZOAE o 34mm 85 vcdom _ 8.82 mm _ - a:
MA_OOA<HA>>HOI m oﬂm- n no- woe? _ mmwnm _ P - we
MmgthIOHn—q c 3.3. cod- 34. :2 wm. Km— ? - mm
oocoscom ember—U New.“ :2 BED 333:2 Cedar—2 new - :Bm

 

saw 48.2% a

.moomtsm E23425 wEm: “mowE use»: :EoﬁoxE a mo mix—mew m2404<2 5:5 388% .8252. Lo moEomoi _.m 29¢

 

 

 

m3>moomo<5<m>mo3>zd>ooamoomae N- o2 mg- 3.3% 43.9% 2;
Me2<00<Q<0moom§mma>m_~<om55> a 8:- w _ .o- wSmm $.32 an:
MESSEQEBEEE: 4 8%- 8a 2.83 2.88 w: - 3
Mm5>m=<omzm§ _ 8%- 8.0 2.2“: News: w : - a:
x0e:—
ooeoscom owcmno maem. :2 SEQ 3.832 998qu can - tam

 

3.. .8 2% E: 38:22: .4.

70

 

 

 

EE<OO<Q<OmDOa§m24>::<QmEm—1;v=& N 30v- and- 2:33 Nedmwm m2 - mo
ME>mOOIO<E<m>MO>9>2.3002500qu m- omw two- 2.53m Nméovm _ m-—
¥H2<OO<Q<OmQOmIMm3432—385Ema». _ 83. $6.. mdomm Edemm m2 - m2
v:_OOA<._.1_>>FOIMv_u_Dmm<¥2m<mHMAI m cow. 8.? w.w_~m haw—mm E. - we
Mmmq>::<om_mmq>v=m§$ v 8%. 8.0 2.88 3.80m x: - 3
Mm=4>m=<amﬁmq>v=ﬁ N ammo- mmd wemmm mosmmm w: - mm
OOmOumM>v~<<EZMmqmq< o ommm- and $.mwmm aqmwmm m2 - em—
MmIA>I=<QmEmA> _ 94%. Ed 36%: 3.2:: w: - mo—
¥03>Zq>003m00m40 T co mod 05— woe—M: 2 - _
MMEOOA<EA>>HOI m Chem- Ed mweom _ 3.82 mm - we
v_.__OOx_<.5>>HOI N oﬂm- med weﬁnm— Zuwnm— R - we
38?: EM
coco—zoom emcenu mﬁm :2 830 333.2 99832 - team

 

=< eo man—#35 E<m Ham .9

71

 

 

 

§2<oo<m<8aom§m§>m§am~mmimé N 221 :3- SANNN 38% N2 - a
m5>moumc<5<m>m03>zg>ooamuamqw N- CNN 8.? 3.3% :3: 3 - _
m~2<oo<a<0mawmmmm5>52amEma» _ 83- No. mac? 9%? N2 - 2:
mmdcgﬁgogqammézmégé V cm. 3- No.32 $.32 NS?
xdwwiﬁgogqammﬁzmépxé N 8v. Nod- 3sz $sz R - NV
xm5>m=<amzmu>v§§m v 8%. Ned N_.N8N ENSN w: - 8
Mmmqéinmamﬁmé N $8- 86 MENNN $.38 w: - 8
mqmﬁmanﬁm—EEB N 83.. e .o- ZEN 3.28 8 - NN
cohoqmgmiﬁzmim? o oNNN- So- GQNN oQNN N2 - v:
Edooqﬁgogqam? N omNT 2d- NSNNN NQNNN wEm
mdooqﬁqgogqam? N 2:- So- mnoﬂN 3%: NEW
mugamm§2m<mbimum N 82 ”to 3.2: SEN No-8
ﬁémoﬁmﬁmﬁmmwﬁ m 8% mod v.2 _N mi: N 3 - NN
aqammﬁzmﬁhﬁim N ON: mNd NNQON $.88 No-3
Mgmmoﬁaﬁmémmox v onm 86 NNNNE aNNwe 8 - E
Mmmammmﬁmmmotq _ oaN- 8d NNNNE 3.22 S - NN
mm5>§<omam§ _ 8%- o 2.38 2.3: w: - m2
E<mmo§§qm<mmmo N oEN N _ .o 8.5: 2.32 8 - 8
595815;? N SNN- :.o 8.22 I .22 2 - 3
m3>mocmo<5<m> 7 o: 33 N88 3.82 a - :
MmdooiSEo: N EN. :3 3.82 £82 M: - we
v353<5>>5m N 8:- Nod N35 25 R - we
#89: a: mam
85:ch ems—5 mﬁm SEQ A2832 ©9852 - tam

 

:< .3 <4»— Em (HQ: 6

72

Figure 3.1 also presents the MALDI mass spectra of the same tryptic digest of
myoglobin deposited on a HDT SAM/ PAA ﬁlm, a HDT SAM, and a HDT SAM/
PAA-PEI ﬁlm. The digest mixture was applied to these modiﬁed surfaces, allowed to
dry, and rinsed with water before adding matrix. The spectra clearly show that the
relative signal intensity at m/z 1608 decreases dramatically when using the modiﬁed
surfaces with the rinsing step. Additionally, signals from a number of previously
undetected peptides appear. The diminution of the signal at m/z 1608 possibly occurs
because the water rinse removes a large fraction of this relatively hydrophilic peptide
(see table 3.1 for peptide sequence). Other peptides that are hydrophobic or capable
of electrostatically interacting with a modified surface are less likely to be removed by
rinsing.

Table 3.1 lists the tryptic fragments of myoglobin detected from various probes,
their amino acid sequences, their net charges at pH 5.3 (~pH of the rinsing solution),
and their Bull and Breese (B & B) hydrophobicity indices. Three of the four
fragments detected using the hydrophobic HDT SAM have highly negative B & B
indices, suggesting that these peptides are primarily retained via hydrophobic
interactions. Peptides detected on PAA surfaces generally have either negative B & B
indices or they are positively charged, suggesting that both hydrophobic and
electrostatic interactions contribute in the binding process. For the PAA-PEI surface,
most of the detected peptides have negative B & B indices, and the two peptides with
positive B & B indices were negatively charged, suggesting that both hydrophobic and

electrostatic interactions were responsible for peptide adsorption.

73

 

 

 

 

 

E

%V

l:"
1‘—

:1

:ﬁ—'

l

L,

{—3

Ti’

if

' C-

 

lntensi

 

lwulh ill. 1.. :

p.
l’
y

- . d
iL" - 'l

i ‘ ‘7" l " I“. '

Hill A. .. v L l A ‘ l
12'00 19'00 26'00 33'00 4000

 

 

 

 

mlz

Figure 3.1 MALDI mass spectra of a tryptic digest of myoglobin
(500 fmol) deposited on different probe surfaces: (a) conventional
stainless steel (b) hydrophilic anionic spots in a hydrophobic ﬁeld,
HDT SAM/ PAA; (c) a hydrophobic surface, HDT SAM; (d)
hydrophilic cationic spots in a hydrophobic ﬁeld, HDT SAM/ PAA-
PEI. In (b)-(d), the samples were deposited on the polymer-
modiﬁed surface and rinsed with water prior to addition of a-
CHCA. Stars (*) indicate peaks that can be assigned to tryptic
fragments of myoglobin.

74

Overall, the MALDI mass spectrum obtained using the HDT SAM/
PAA-modiﬁed surface (Figure 3.1b) contains signals due to 22 proteolytic fragments
of myoglobin, and only seven of these signals were detectable when using the SS
sample probe, from which only 11 total tryptic fragments were detected. This
increase in the number of tryptic fragments with the PAA-modiﬁed surface
signiﬁcantly increased the conﬁdence level in identifying myoglobin. For example,
when using the Mascot program to identify myoglobin with data obtained using the SS
probe, the probability-based MOWSE score was 82. In contrast, protein
identiﬁcation using data accumulated with the HDT SAM/ PAA-modiﬁed surface gave
a MOWSE score of 215. Mass spectral data obtained by using the PAA-modiﬁed
probe greatly decreased the probability (from 0.25% to 10"3'9%) that the protein
identiﬁcation occurred due to a random match. Interestingly, although analyses from
HDT SAM and HDT SAM/ PAA-PEI surfaces yielded 11 proteolytic fragments
(Figures 3.1c and 3.1d), all of these signals were also present in the MALDI mass
spectra obtained using either the PAA-modiﬁed sample surface or the SS probe.

The data in Figure 3.1 were obtained from myoglobin that was denatured with
6 M urea. When myoglobin is denatured at 65 °C before digestion by trypsin, the
mass spectrum of the digest from a conventional SS sample plate shows about the
same number of tryptic fragment signals as does the spectrum taken using a
PAA-modiﬁed surface, although relative intensities are quite different. In the case of
the SS sample probe, addition of urea (0.6 M) to the heat-denatured digest decreases
the number of identiﬁable proteolytic fragments to the level seen in the spectrum of the
digest prepared using preliminary urea denaturation. This suggests that the main
advantage of using the polymer-modiﬁed surfaces on the sample probe lies in allowing

the removal of urea. (Although thermal denaturation avoids urea-contamination of a

75

 

protein sample destined for analysis by MALDI-MS, it breaks electrostatic bonds only
reversibly and is rarely capable of disrupting the hydrophobic core of most proteins.
Thus, 6 M urea continues to be the chaotropic agent of choice in the preliminary
preparation of proteins for proteolytic digestion.‘)

Sequence coverage is another ﬁgure of merit in protein mass mapping by
MALDI-MS. Usually, ~90 to 100% sequence coverage can be achieved for smaller
proteins like myoglobin; however, for bigger proteins, such as BSA, sequence
coverages are usually low.2 The conventional MALDI mass spectrum (See Figure
3.2) of an incomplete tryptic digest of BSA (urea denaturation) contains signals from
numerous tryptic fragments, resulting in a sequence coverage of 61.3%, or 372 amino
acid residues out of 607. With additional MS data obtained using each of the three
modiﬁed probes (data not shown), sequence coverage increases by 13.2 percent to
include another 80 amino acid residues. B. DNA Analysis Using Positively Charged
MALDI Probes Patterned HDT SAM! PAA-PEI-modiﬁed surfaces are also useful for
on-probe decontamination of DNA samples prior to analysis by MALDI-MS. As a
previous study shows,3 the salt-tolerance level for successful analysis of DNA samples
by MALDI-MS is around 100 mM. At higher salt concentration, both mass
spectrometric signal intensity and resolution degrade due to formation of
multiple-cation adducts of DNA. This problem is apparent in Figure 3.4a, where a
broad, low-intensity signal occurs when a 60-fmol sample of a DNA 24-mer in 800
mM NaOAc is analyzed by MALDI-MS from a SS sample probe. However, a sharp
peak from the protonated DNA 24-mer (Figure 3.4b) results with use of a PAA-PEI

—modified sample probe and a water rinse before analysis by MALDI-MS.

76

Intensity

?+

2500 3250 4000
(mlz)

 

1000 1750

Figure 3.2 MALDI mass spectrum of a BSA (1 pmol) tryptic
digests obtained on a SS probe. Stars (*) indicate peaks that
can be assigned to tryptic fragments of BSA.

77

 

 

100%=400

iv-
Il‘lL.

>‘ 3' H Ili'lff; I

t 1‘ ‘g‘ll AAA; “it!!! wﬁi‘ldllplll“ gill? “’ '1 «ii’llil‘yk 1M1)”;L a

(D I! 1 W” ‘l'iilllli W

: ,Miwaem ,rr.__

0:

E l

_ l 100%=1000
l b
1:1..1'

6000 7000 8000

mlz

Figure 3.3 MALDI mass spectra of 60 frnol of a DNA 24-mer in 800
mM NaOAc deposited on different probes: (a) SS without rinsing; (b)
HDT SAM/ PAA-PEI with rinsing after sample drying. A comatrix
consisting of 3-HPA and ammonium citrate was used in both cases.

78

 

We note that the DNA sample should be rinsed with H20 only after it dries; otherwise,
the water rinse removes DNA molecules as well as contaminating salts. In contrast,
when decontarninating salty protein solutions prior to analysis by MALDI, strong
signals appear regardless of whether rinsing takes place before or after drying of the
sample.4 Presumably, the difference in behavior between proteins and DNA is related
to their different solubilities. Proteins have limited solubility in water, and they tend
to bind to each other via hydrophobic interactions. In contrast, DNA molecules are
highly soluble in water, so rinsing of a wet sample may break the limited electrostatic
bonds between DNA and the PEI surface. However, during the time in which DNA
samples are allowed to dry on the probe, the DNA may rearrange around the positive
groups of PEI to establish multiple binding interactions that can better withstand water
rinsing. The addition of ammonium citrate to the matrix is also necessary in order to
observe a MALDI signal from the analyte. The citrate anions probably compete with
the bound DNA for the positive sites in PEI, helping to desorb DNA for incorporation
into matrix crystals. The ammonium ions should also displace the remaining Na+ or
K+ adducts, and subsequent elimination of ammonium adducts as NH3 leaves only the
protonated DNA species, which is represented by a well-resolved peak in the MALDI
mass spectrum (Figure 3.3b). Similar results can be obtained with a mixture of
several DNA strands in 1 M NaOAc. With a mixture of a 7-mer, an 8-mer, a IO-mer,
and a 12-mer in l M NaOAc, signals are observed for all DNA components in
MALDI-MS using MUA-PEI- or PAA-PEI-modified sample probes. (see Figure 3.4,

for an example)

79

 

 

>.100 ‘/
3: 80
(D l
C 60
d’ 1 3 10
E 40 ; r 8-mer -mer 12-mer ‘
—- l 1‘
20 t l L, l
1;"); m, V n . l. r J
2000 2300 2600 2900 3200 3500
mlz

Figure 3.4 MALDI mass spectrum of a DNA mixture contaminated with 1 M NaOAc.
The sample was deposited on a modified MUA-PEI surface, air dried, rinsed with
water, and subsequently matrix was applied.

80

IV. Conclusions

Parallel analyses of tryptic digests of proteins by MALDI-MS from sample
probes consisting of polymer-modiﬁed surfaces containing hydrophobic, cationic, or
anionic sites greatly increase the number of peptides that can be detected. The
enhancement is likely due to removal of urea after deposition of the sample, and prior
to deposition of the MALDI matrix. Polycationic surfaces on MALDI sample probes
are also useﬁil in preventing formation of multiple-ion adducts of DNA during analysis
of salt-contaminated DNA samples by MALDI-MS, thereby yielding well-resolved

peaks for protonated DNA segments.

81

.ﬂn‘l.

 

(1)

(2)

(3)

(4)

References:

Stone, K. L.; LoPresti, M. B.; Williams, N. D.; Crawford, J. M.; DeAngelis, R.;
Williams, K. R. Tech. Protein Chem. 1989, 377-91.

Kadlcik, V.; Strohalm, M.; Kodicek, M. Biochem. and Biophys. Res. Commun.
2003, 305, 1091-1093.

Shaler, T. A.; Wickham, J. N.; Sannes, K. A.; Wu, K. J .; Becker, C. H. Anal.
Chem. 1996, 68, 576-9.

Xu, Y.; Watson, J. T.; Bruening, M. L. Anal. Chem. 2003, 75, 185-190.

82

 

Chapter Four: Use of Polymer-Modiﬁed MALDI-MS Probes to Enhance
Detection of Phosphopeptides in Phosphoprotein Digest

1. Introduction:
Characterization of protein phosphorylation is essential in proteomics because this
posttranslational modiﬁcation is one of the most important mechanisms for regulating

"2 It is

signal transduction, gene expression, and protein synthesis in eukaryotic cells.
estimated that approximately one-third of all proteins in eukaryotic cells are
phosphorylated at any given time.3 However, the analysis of these phosphoproteins is
not straightforward due to challenges such as substoichiometn'c levels of
phosphorylation, variation of phosphorylation sites with cell state, and different
degrees of phosphorylation at different sites for the same phosphoprotein.4’5
Traditionally, protein phosphorylation was studied by incorporation of 32F into the
cellular proteins via treatment with radio-labeled ATP.“7 However, this method
suffers from complications due to problems with handling radioactive materials and
limited 32F incorporation.3’9

In recent years, mass spectrometry (MS) has become an indispensable tool for

3 -S.7.9-24

studying phosphoproteins. The detection of phosphopeptides by MS can

involve a number of different techniques, including a combination of peptide mapping

13,24 10.16.25 11,21,23

and phosphatase treatment, post—source decay, precursor ion seaming,

- 12.19
and neutral loss scanning. '

However, the detection of phosphoproteins or
phosphopeptides by MS is challenging because signals from these species are often

suppressed by non-phosphopeptides in the digest mixture, especially when

83

 

phosphorylation occurs at a substoichiometric level. This is clearly shown in Figure
4.1a-c. When pure angiotensin (A, sequence DRVYIHPF, Figure 4.1a) or pure
phosphoangiotensin (AP, sequence DRVpYIHPF, Figure 4.1b) was analyzed
individually by MALDI-MS, a very strong signal was observed in either case.
However, when an equimolar mixture of A and AP was analyzed using MALDI-MS,
only signal due to protonated A was observed (Figure 4.1c). Because ionization of
phosphopeptides is comparitively weak, several protocols have been designed to
enrich phosphopeptides (remove non-phosphorylated peptides) before analysis by MS.
For example, antibodies that bind speciﬁcally to phosphotyrosine have been used to
immunoprecipitate phosphotyrosine(pY)- containing peptides.26 However, this
method is limited to pY—containing peptides or proteins, and cannot be applied to
phosphoserine-(pS) or phosphothreonine(pT)-containing peptides. All of the
antibodies for pS- and pT-containing peptides found thus far are sequence dependent.
Considering the 1800:200:1 ratio of pS:pT:pY found in eukaryotic cells“, this
pY-speciﬁc antibody method is not particularly useful. A second protocol for
enriching phosphopeptides employs reverse phase HPLC to separate
non-phosphopeptides from phosphopeptides along with concentration of isolated

Phophopeptides by solvent evaporation.”27

However, multiply-phosphorylated
peptides can be quite hydrophilic and co-elute with salts, while less hydrophilic,
mono-phosphorylated peptides may co-elute with the corresponding

non-phosphorylated peptides, making HPLC separation futile. Adsorption of

peptides to vial walls during evaporation and handling can be a serious secondary

84

 

Z

 

 

 

 

bi _‘ ”"4 F- A - "'“_—' "‘1"- — w ‘— *‘T
-- % AP 2.
a \ :
_=. r A . + - - — ~ 4
\‘5 !
3 AP not detectable 1
1000 1100 1200

mlz

Figure 4.1 MALDI mass spectra of: (a) 5 pmol Angiotensin (A); (b) 5 pmol
Phosphorylated Angiotensin (AP); ( c) 2.5 pmol of A and AP. a-CHCA was

used as matrix in every case.

85

 

problem that reduces sensitivity.

Recently, immobilized metal afﬁnity chromatography (IMAC) has become
popular for the enrichment of phosphopeptides prior to MS analysis. This technique
utilizes the complexation of phosphate groups by metal ions to retain phosphopeptides,

while other peptides are removed during washing.”30

However, nonspeciﬁc binding
of peptides containing multiple acidic residues and difﬁculty in eluting multiply
phosphorylated peptides can be problematic in IMAC. Thus, proper use of this
technique requires that the analyst has a thorough understanding of the operating
principles of the separation system. IMAC beads carrying adsorbed phosphopeptides
also have been analyzed directly on MALDI probes following multiple rinsing with

- - - - 9.17.29
dilute acetic acrd solutions.

Compared with the traditional IMAC technique, this
method avoids the time-consuming HPLC process. Deposition of these beads
followed by addition of matrix solution is sufﬁcient to observe signals due to
mono-phosphopeptides. However, in order to desorb multi-phosphorylated peptides
from beads, addition of IOOmM ammonium dihydrogenphosphate to these heads is
necessary.29

In this chapter, I discuss the use of patterned cationic surfaces on MALDI probes
to selectively bind phosphopeptides to enhance their detectabilities. During such
experiments, peptide mixtures containing both phospho- and nonphospho-peptides are
applied to the cationic surfaces, dried, and then rinsed with water to remove most of

the nonphosphopeptides. Surface bound phosphopeptides are released by the

subsequently added strongly acidic matrix solutions for their incorporation into matrix

86

 

crystals. This technique offers potential advantages over HPLC and traditional IMAC

in that it allows for rapid sample preparation and avoids loss of peptide on vial walls.

II. Experimental Section
A. Materials and solutions

All proteins and chemicals were obtained from Sigma unless noted otherwise.
Protein solutions were prepared in deionized H20 (milli-Q, 18 MQcm). Three
peptides, angiotensin (A), phosphoangiotensin (AP) and another phosphopeptide with
unknown sequence (UP, m/z 1858) were kindly provided by Rhonda Husain from the
Mass Spectrometry Facility at Michigan State University. Mixed linkage kinase 3
(MLK3) was kindly provided by Professor Gallo of Michigan State University and
prepared by Karen Alexandra Schachter. A peptide mixture simulating a
phosphoprotein digest was prepared by mixing A, AP, and UP (~5 11M ﬁnal
concentration of each species) with a myoglobin (~l 11M) tryptic digest. Gold-coated
substrates (200 nm of gold sputtered on 20 nm of Cr on Si(100) wafers ) were prepared

by Lance Goddard Associates (Foster City, CA).

B. Protein Digestion and Peptide Dephosphorylation

Horse heart myoglobin, B-casein and ovalbumin were digested by trypsin
(promega) according to the procedures provided by the supplier. Brieﬂy, ~5-20 pg of
protein was ﬁrst denatured either by 50 11L of 6 M urea in Tris buffer (diluted to 0.6 M

urea before addition of trypsin, ﬁnal volume ~500 1.1L) or by heating at 65 °C in 50 11L

87

 

water for 30 min (ﬁnal volume ~ 500 pL by addition of 50 mM ammonium
bicarbonate). If the protein contained disulﬁde bonds (B-casein and ovalbumin), it
was then reduced by 5 11L of 10 mM dithiothreitol (DTT) and alkylated with 10 uL of
100 mM iodoacetamide. All proteins were ﬁnally incubated with 0.5-1 pg trypsin at
37 °C overnight. In most cases, the digestion was stopped by addition of acetic acid
to achieve a pH of ~3. Digested ovalbumin was also incubated with 10 U (enzyme
activity unit) phosphatase (New England Biolabs) at 37 °C for 3 hours in a pH 7.5 tris
buffer to remove phosphate groups.
C. Fabrication of Patterned Substrates

The procedure for preparing patterned PAA-PEI or MUA-PEI was discussed in
chapters 2 and 3. I report here the preparation of another patterned cationic surface,
PAA-Fe“, from patterned anionic PAA. This surface was designed to be somewhat
similar to stationary phases used for IMAC. A slide coated with patterned PAA was
immersed into a 100 mM FeCl3 aqueous solution or a 100 mM Fe(ClO4)3 ethanolic
solution for 30 minutes, followed by water or ethanol rinsing to remove unattached

F e3 +.
D. Sample Preparation prior to Analysis by MALDI-TOF-MS

Sample solution (0.25 1.1L) was applied to the PAA-PEI or PAA-Fe3+ surface,

allowed to dry, and then rinsed with water (pH ~5) or dilute acetic acid (100 mM,
pH~4). Matrix solution (0.25 11L saturated a-CHCA, in 50:50:01

acetonitrile:water:triacetic acid) was applied aﬁer drying the water-rinsed surface with

88

P m_e "in

 

‘5]...5 a,

N2. Finally, the wafer carrying the sample was attached to a disposable MALDI plate

using superglue, and the sample was analyzed by MALDI-TOF-MS.

E. Instrumentation and Data Analysis

Mass spectra were obtained with a PE Biosystems Voyager STR MALDI TOF
mass spectrometer using an accelerating voltage of 20 kV, a 94% grid voltage, a 0.05
% guidewire voltage, and an extraction delay time of 100-150 nsec to accumulate ion
current associated with 50-250 laser pulses. A 2-point calibration for protein digests
was usually performed using intense peaks corresponding to previously identiﬁed
peptides. Peaks were assigned to protonated phosphopeptides by the presence of a
m/z — 80 peak (loss of HPO3), or the disappearance of a particular signal after
treatment of the same sample with phosphatase. Ellipsometric determinations of
polymer thickness were performed with a rotating analyzer spectroscopic ellipsometer
(J .A. Woollam, M—44), assuming a ﬁlm refractive index of 1.5. Reﬂectance FTIR
spectra were obtained with a Nicolet Magna 560 spectrophotometer using a Pike
grazing angle (80°) accessory; the spectrophotometer was housed in a glove box to

minimize interference from water vapor.

III. Results and Discussion
A. Simple Mixture of Phosphopeptides

To study the effectiveness of patterned, cationic surfaces for the analysis of
phosphopeptides, I ﬁrst obtained the MALDI mass spectrum of a simple mixture of

equimolar A and AP. The results are shown in Figure 4.2. When this sample was

89

analyzed using a conventional stainless steel plate, only signal due to A was observed,
even though the same amounts of A and AP were present in the mixture. However,
when analyzing the same sample using a PAA-Fe3+-modiﬁed surface and a water rinse
after sample deposition, signal due to AP appeared. This result was exciting, as it
showed that a simple water rinse helped to solve the suppression problem; but it was
not as ideal as we had hoped it would be. Ideally, the water rinse should remove most
of the nonphosphopeptide while the phosphopeptide remains bound to the surface via
electrostatic interactions. In the mass spectrum in Figure 4.2b, the peak due to the
nonphosphopeptide is still the dominant peak, suggesting that a good portion of this
peptide remains on the surface after the water rinse. These results suggest that the
modiﬁed cationic surface has higher affinity to phosphopeptides than to
non-phosphopeptides, but the selectivity still needs to be improved in order to totally
separate nonphosphopeptides from phosphopeptides, and thus completely eliminate

the suppression problem.

B. Simulated Phosphoprotein Digest

In order to obtain a more complicated phosphopeptide-containing mixture that
represents a typical tryptic digest of a phosphoprotein, I intentionally mixed three
peptides, A, AP, and UP, with a tryptic digest of myoglobin. As shown in Figure 4.3,
when the mixture was analyzed on a SS probe, signals due to the two phosphopeptides,
AP and UP, were hardly observable. However, when the same sample was applied to

a PAA-PEI-modiﬁed probe and rinsed with water, signals due to AP and UP were

90

greatly enhanced, and signals due to the tryptic peptides from myoglobin decreased
dramatically. Interestingly, the peak that is due to A is still the dominant peak,
suggesting that this peptide has a stronger afﬁnity to PAA-PEI than other

non-phosphopeptides in the myoglobin tryptic digest mixture.

C. 'Ii'yptic Digest of Phosphoproteins

Analysis of ovalbumin by MALDI-MS further illustrates the challenge in detecting
phosphorylated fragments using a SS probe. Figure 4.4a shows a MALDI mass
spectrum of a tryptic digest of ovalbumin loaded onto a conventional SS sample probe.
Some of the peaks that represent phosphorylated fragments (peaks labeled with “P”,
m/z 2091, EWGpSAEAGVDAASVSEEF R, m/z 2513,
LPGFGDpSIEAQCGTSVNVHSSLR, and m/z 2903,
FDKLPGFGDpSIEAQCGTSVNVHSSLR) show low signal intensities. I assigned
these peaks to phosphorylated tryptic fragments based on the sequence of ovalbumin
and the presence of accompanying peaks corresponding to MH+ — 80 Da (loss of
HP03). My attempt to use LC-MS to identify phosphorylated peptides failed,
presumably because other peptides co-eluted with the phosphorylated proteolytic
fragments. Use of both PAA-Fe3+ and PAA-PEI probes (Figure 4.4b and 0) resulted in
a greatly increased intensity for the phosphorylated fragment at m/z 2091. For
PAA-PEI, we attribute these results to binding of anionic, phosphorylated peptides to

the polycationic surface of the modiﬁed MALDI probe; while in the case of patterned

91

 

 

 

 

 

 

 

 

A\‘ T 5 ss 1
3‘1 1 AP not detectable 1
'35 1’ ‘ \ a 1
c __

31 A PAA F 3+1
_=. 1 \ ‘ ‘ e l

i 1

% AP 1

_. - - \* - bi

1000 1100 12
mlz

Figure 4.2 MALDI mass spectrum of: (a) Equimolar mixture of
angiotensin (A) and phosphoangiotensin (AP), 2.5 pmol each, analyzed
using stainless steel probe; (b) the same sample as in (a), applied to a
PAA-Fe3+ modiﬁed probe, dried, and rinsed with water. Saturated a-
CHCA was used as matrix in both cases.

92

'0 ' .-unbﬂ

 

 

Intensity
E

/UP

1 b

1000 1400 1800 2200 2600 3000
mlz

Figure 4.3 MALDI mass spectra of a peptide mixture prepared by mixing
two phosphopeptides, “AP" and “UP", and a non-phosphopeptide, “A", 5 11M
each, with a tryptic digest of myoglobin. (a)The sample was analyzed on a
SS probe; (b) The sample was analyzed with a patterned, PAA-PEl-modiﬁed
probe. Saturated a-CHCA was used as matrix in both cases.

93

 

PAA-Fe“, the phosphate groups might reversibly form complexes with Fe“. The
IMAC technique (including use of IMAC beads directly on the probe) also relies on
binding through metal ion/ phosphate complexes,22‘28'3 ' but that technique is not as
simple as use of polymer-modiﬁed MALDI sample probes.

To verify that the labeled peaks (p) in Figure 4.4 represented phosphorylated
fragments, I incubated an aliquot of the same tryptic digest of ovalbumin with
phosphatase, an enzyme that selectively removes phosphate groups. After 3 hours of
incubation at 37 °C, 1 analyzed this phosphatase-treated digestion mixture using a SS
probe and the cationic PAA-Fe” surface. The MALDI mass spectra (Figure 4.4d and
4.4e) showed no detectable peaks for the phosphorylated peptides seen in Figures 4.4b
and c, and corresponding nonphosphorylated peptide peak intensities increased.
Similar results were obtained for a B-casein digest, where signals due to two
phosphopeptides (m/z 2063.0, FQpSEEQQQTEDELQDK and m/z 3123.9,
RELEELNVPGEIVEpSLpSpSpSEESITR) were selectively enhanced as shown in
Figure 4.5. It should also be noted that both of these peptides are highly negatively

charged, and the acidic amino acid residues may also contribute to the binding process.

94

 

100%=2.400 1

p
P P \ a
1\ ii .LLz\L1 r. A L41. . l ‘_
P ,
\

100%=1o.oooi

 

 

g PAA-Fe3+

   

; p .
PAA-PEI \ 100%=15,ooo1

Intensity

 

 

100%=5.ooo 1

P- o 1
pa \ d .
i 100%=4.000 1
2 P' "\° 1
m--.n§1__nl.l.n -L1lx .. a

1 000 1500 2000 2500 3000 3500

mlz

Figure 4.4 Mass spectra of a tryptic digest of ovalbumin (1 pmol) obtained
using different MALDI probes: (a) conventional 88; (b) PAA-Fe3+; (c) PAA-
PEI; (d) after phosphatase treatment and analyzed on 88; (e) after
phosphatase treatment and analyzed on PAA-Fe3+. In (b), (c) and (e),
samples were deposited on the probe and rinsed with water prior to addition
of matrix. Peaks labeled with “P” represent phosphorylated peptides and
peaks labeled with “P-80” represent corresponding nonphosphorylated
pepﬁdes.

 

 

95

 

 

 

 

 

 

1 1000/0:20.000
l
l
9 '1 1? P
2 Multidi'u.1};-u11 .. 1| .1 111' “111.1111: \' .. . _ a
a:
a P
E \ p 100%=4,000
\
1 1 1
11.110.11.10...11.4119?.1. ~11; 111. Wt“ we. J A -1 -991 ' . b
1000 1600 2200 2000 3400 4000
mlz

 

Figure 4.5. Mass spectra of a tryptic digest of b-casein (100 frnol) obtained
using different MALDI probes: (a) conventional SS; (b) HDT SAM/ MUA-PEI.
In (b), the sample was deposited on the probe and rinsed with water prior to
addition of matrix. The peaks labeled with “P” represent phosphopeptides.

96

IV Conclusion

Detection of phosphorylated peptides during MALDI-MS of digests deposited on
polycationic surfaces consisting of PAA-PEI or PAA-Fe3+ is particularly attractive
because these analytes are often undetectable when using conventional SS sample
probes. The enhanced detectabilities for phosphopeptides are probably due to the

removal of interfering non-phosphopeptides, which reduces suppression.

97

 

(1)
(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

Reference

Hunter, T. Signaling--2000 and beyond Cell 2000, 100, 113-127.

Marks, F. Protein phosphorylation, 1996.

Zolnierowicz, S.; Bollen, M. Protein phosphorylation and protein phosphatases.
De panne, belgium, september 19-24, 1999 Embo. J. 2000, 19, 483-488.

Kinumi, T. Study of phosphoprotein and phosphopeptide by mass spectrometry
J. Mass Spectrom. Soc. Japan 2003, 51, 324-329.

McLachlin, D. T.; Chait, B. T. Analysis of phosphorylated proteins and
peptides by mass spectrometry Curr: Opin. Chem. Biol. 2001, 5, 591-602.

Haystead, T. A. J .; Garrison, J. C. Study of protein phosphorylation in intact
cells Protein Phosphorylation (2nd Edition) 1999, 1-31.

Yan, J. K; Packer, N. H.; Gooley, A. A.; Williams, K. L. Protein
phosphorylation: Technologies for the identiﬁcation of phosphoamino acids J.
Chromatogr. A 1998, 808, 23-41.

Fadden, P.; Haystead, T. A. Quantitative and selective ﬂuorophore labeling of
phosphoserine on peptides and proteins: Characterization at the attomole level

by capillary electrophoresis and laser-induced ﬂuorescence Anal. Biochem.
1995, 225, 81-88.

Zhou, W.; Merrick, B. A.; Khaledi, M. G; Tomer, K. B. Detection and
sequencing of phosphopeptides affinity bound to immobilized metal ion beads
by matrix-assisted laser desorption/ionization mass spectrometry J. Am. Soc.

Mass Spectrom. 2000, 11 , 273-282.

Annan, R. S.; Carr, S. A. Phosphopeptide analysis by matrix-assisted laser

98

 

3413-3421.

(11)

(12)

(13)

(14)

(15)

(16)

(17)

(18)

(19)

(20)

Carr, S. A.; Huddleston, M. J .; Annan, R S. Selective detection and sequencing
of phosphopeptides at the femtomole level by mass spectrometry Anal.
Biochem. 1996, 239, 180-192.

Hunter, A. P.; Games, D. E. Chromatographic and mass spectrometric methods
for the identiﬁcation of phosphorylation sites in phosphoproteins Rapid
Commun. Mass Spectrom. 1994, 8, 559-570.

Larsen, M. R.; Sorensen, G L.; Fey, S. J .; Larsen, P. M.; Roepstorff, P.
Phospho-proteomics: Evaluation of the use of enzymatic de-phosphorylation 1.
and differential mass spectrometric peptide mass mapping for site speciﬁc "'

phosphorylation assignment in proteins separated by gel electrophoresis
Proteomics 2001, I, 223-238.

Mann, M.; Ong, S.-E.; Gronborg, M.; Steen, H.; Jensen, 0. N.; Pandey, A.
Analysis of protein phosphorylation using mass spectrometry: Deciphering the
phosphoproteome Trends Biotechnol. 2002, 20, 261-268.

 

“_ _-I _

Marcus, K.; Immler, D.; Stemberger, J .; Meyer, H. E. Identiﬁcation of platelet
proteins separated by two-dimensional gel electrophoresis and analyzed by
matrix assisted laser desorption/ionization-time of ﬂight-mass spectrometry
and detection of tyrosine-phosphorylated proteins Electrophoresis 2000, 21,
2622-2636.

Metzger, S.; Hoffmann, R. Studies on the dephosphorylation of
phosphotyrosine-containing peptides during post-source decay in

matrix-assisted laser desorption/ionization J. Mass Spectrom. 2000, 35,
1165-1177.

Raska, C. S.; Parker, C. E.; Dominski, Z.; Marzluff, W. F.; Glish, G L.; Pope, R.
M.; Borchers, C. H. Direct maldi-ms/ms of phosphopeptides affmity-bound to

immobilized metal ion affmity chromatography beads Anal. Chem. 2002, 74,
3429-3433.

Resing, K. A.; Ahn, N. G Protein phosphorylation analysis by electrospray
ionization-mass spectrometry Methods Enzymol 1997, 283, 29-44.

Schlosser, A.; Pipkom, R.; Bossemeyer, D.; Lehmann, W. D. Analysis of
protein phosphorylation by a combination of elastase digestion and neutral loss

tandem mass spectrometry Anal. Chem. 2001, 73, 170-176.

Shou, W.; Verma, R.; Arman, R. S.; Huddleston, M. J .; Chen, S. L.; Carr, S. A.;

99

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

(29)

(30)

Deshaies, R. J. Mapping phosphorylation sites in proteins by mass
spectrometry Method. Enzymol. 2002, 351, 279-296.

Steen, H.; Kuester, B.; Fernandez, M.; Pandey, A.; Mann, M. Detection of
tyrosine phosphorylated peptides by precursor ion scanning quadrupole tof
mass spectrometry in positive ion mode Anal. Chem. 2001, 73, 1440-1448.

Thompson, A. J .; Hart, S. R.; Franz, C.; Bamouin, K.; Ridley, A.; Cramer, R.
Characterization of protein phosphorylation by mass spectrometry using
immobilized metal ion affinity chromatography with on-resin b-elimination
and michael addition Anal. Chem. 2003, 75, 3232-3243.

Wilm, M.; Neubauer, G; Mann, M. Parent ion scans of unseparated peptide
mixtures Anal. Chem. 1996, 68, 527-533.

Zhang, X.; Herring, C. J .; Romano, P. R.; Szczepanowska, J .; Brzeska, H.;
Hinnebusch, A. G; Qin, J. Identiﬁcation of phosphorylation sites in proteins
separated by polyacrylamide gel electrophoresis Anal. Chem. 1998, 70,
2050-2059.

Schnolzer, M.; D. Lehmann, W. Identiﬁcation of modiﬁed peptides by
metastable fragmentation in maldi mass spectrometry Int. J. Mass Spectrom.
1997, 169/170, 263-271.

Stancato, L. F.; Petricoin, E. E, 3rd Fingerprinting of signal transduction
pathways using a combination of anti-phosphotyrosine immunoprecipitations
and two-dimensional polyacrylamide gel electrophoresis Electrophoresis 2001,
22, 2120-2124.

Vacratsis Panayiotis, 0.; Phinney Brett, 8.; Gage Douglas, A.; Gallo Kathleen,
A. Identiﬁcation of in vivo phosphorylation sites of mlk3 by mass spectrometry
and phosphopeptide mapping Biochemistry 2002, 41, 5613-5624.

Andersson, L.; Porath, J. Isolation of phosphoproteins by immobilized metal
(fe3+) affinity chromatography Anal. Biochem. 1986, 154, 250-254.

Hart, S. R.; Waterﬁeld, M. D.; Burlingame, A. L.; Cramer, R. Factors
governing the solubilization of phosphopeptides retained on ferric nta imac
beads and their analysis by maldi toﬁns J. Am. Soc. Mass Spectrom. 2002, 13,
1042-1051.

Posewitz, M. C.; Tempst, P. Immobilized gallium(iii) affinity chromatography
of phosphopeptides Anal. Chem. 1999, 71 , 2883-2892.

100

 

(31) Papac, D. I.; Hoyes, J .; Tomer, K. B. Direct analysis of afﬁnity-bound analytes
by maldi/tof ms Curr: Biol. 1994, 66, 2609-2613.

101

 

Chapter Five: Conclusions and Future Work

I. Conclusions:

Microcontact printing of HDT SAMs affords rapid formation of patterned MALDI
probes that decrease detection limits and increase reproducibility in MALDI-MS.
Graﬁing of PAA into bare gold spots of patterned SAMs increases the versatility of
this system and allows desalting of sample solutions prior to analyses by MALDI.
The combination of patterning and surface modiﬁcation with a polymer that binds
charged proteins yields low-femtomole detection limits, even for samples containing
1M NaAc. Reﬂectance FT-IR spectra conﬁrm that protein binding occurs due to
electrostatic interactions and, thus, is highly dependent upon the pH of sample
solutions. Protein binding to the polymer-modiﬁed surface also depends on protein
and salt concentration.

Patterned surfaces also show great promise for the analysis of more complicated
mixtures, such as protein digests. Parallel analyses of tryptic digests of proteins by
MALDI-MS from sample probes consisting of polymer-modiﬁed surfaces containing
hydrophobic, cationic, or anionic sites greatly increase the number of peptides that can
be detected. The enhancement is likely due to removal of urea after deposition of the
sample, and prior to deposition of the MALDI matrix. Detection of phosphorylated
peptides during MALDI-MS of digests deposited on polycationic surfaces consisting
of PAA-PEI or PAA-Fe3+ is particularly attractive because these analytes are often
undetectable when using conventional SS sample probes. Polycationic surfaces on

MALDI sample probes are also useful in preventing formation of multiple-ion adducts

102

 

of DNA during analysis of salt-contaminated DNA samples by MALDI-MS, thereby
yielding well-resolved peaks for protonated DNA segments.

These examples illustrate the utility of patterned surfaces as MALDI probes, but
there are many other areas in which patterned probes may prove useful. Below, I
describe one of these areas along with possible improvements in the design and use of

patterned probes.

11. Future Work
1. Catch and Release of Free Cysteine-Containing Peptides Using Au

Shotgun proteomics uses trypsin to digest the whole proteome and
multi-dimensional liquid chromatography to separate the resulting complicated peptide
mixture to reduce ionization suppression problems during subsequent analyses by
MALDI-MS. 1'7 However, due to the complexity of the mixture, it is rarely possible to
totally separate all the peptides from one another; current approaches aim at
simplifying the mixture rather than total separation. We propose to use bare Au
surfaces to catch and release free cysteine-containing peptides to simplify complex
peptide mixtures in those cases where interest is focused on this speciﬁc class of
peptides.

The formation of SAMs of alkanethiols on gold has been thoroughly studied} but
the study of cysteinyl peptide adsorption to Au is fairly new.9 Kirk and Bohn recently
used a short peptide containing an N-terrninal cysteine as a model for cysteinyl peptide

adsorption. Matrix was added to the Au colloids carrying the adsorbed peptides, but

103

 

only weak signals due to the peptide were observed in the subsequent MALDI
experiment. Weak signals probably occurred because the MALDI matrix is not
strong enough to remove (cleave) the majority of the adsorbed peptides from Au.
Preliminary experiments show that dilute 12 solutions can remove alkanethiol SAMs
from the Au surface.10 Thus, addition of 12 to the matrix solution may allow stronger

signals to be observed for similar experiments with peptides adsorbed to Au.

2. PAA-nitrilotriacetic acid-Fe3+ Surfaces for Enhanced Capture of
Phosphopeptides

Use of PAA-Fe3+ surfaces to selectively adsorb phosphopeptides on MALDI
probes shows some promise in increasing the detectability of phosphopeptides.
However, the non-speciﬁc adsorption of acidic amino acid residues containing
carboxylic acid groups decreases the selectivity of this surface. Use of more stringent
washing with a low pH solution may increase selectivity for phosphopeptide
adsorption, but Fe” ions may also be removed from the surface during washing
because PAA is not a very strong chelator of Fe“. In contrast, due to its tetradentate
nature, nitrilotriacetic acid (NT A) is a much stronger chelating agent for Fe“ and thus
should withstand stringent rinsing. Derivatization of PAA with NTA followed by the
formation of a PAA-NTA-Fe3+ complex, should result in ﬁlms that more selectively

adsorb phosphopeptides.

104

 

3. Patterned Polymer Brushes

A low binding capacity due to the limited thickness of PAA ﬁlms may lead to the
low signal intensities in MALDI-MS experiments using these polymer-modiﬁed
sample probes. Thus, it may be necessary to grow thicker ﬁlms to improve binding
capacities of these surfaces. Our current methodology uses only one cycle of PAA
deposition to obtain a ﬁlm thickness of ~20 A. More cycles of derivatization with
PAA may create thicker ﬁlms, but the hydrophobic/ hydrophilic pattern may be
compromised due to physisorption of PTBA to the hydrophobic HDT SAMs.
Recently, several methods have been developed for creating thicker, patterned ﬁlms by

”'14 By combining our expertise in MALDI sample probe

growth of polymer bruses.
modiﬁcation with established techniques for growing polymer brushes, we should be

able to increase the binding capacity of modiﬁed surfaces and observe stronger signal

intensities in the corresponding mass spectra.

105

 

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

(11)

References

Lee, C.-1.; Hsiao, H.-h.; Lin, C.-w.; Wu, S.-p.; Huang, S.-y.; Wu, C.-y.; Wang, A.
H. J .; Khoo, K.-h. Strategic shotgun proteomics approach for efﬁcient

construction of an expression map of targeted protein families in hepatoma cell
lines Proteomics 2003, 3, 2472-2486.

McDonald, W. H.; Yates, J. R., 111 Shotgun proteomics: Integrating
technologies to answer biological questions Curr: Opin. Mol. Then 2003, 5,
302-309.

McDonald, W. H.; Yates, J. R., 111 Shotgun proteomics and biomarker
discovery Dis. Markers 2002, 18, 99-105.

Saraf, A.; Yates, J. R., 111 Shotgun proteomics and its applications to the yeast
proteome Protein Arrays, Biochips, and Proteomics 2003, 233-253.

Wolters, D. A.; Washbum, M. R; Yates, J. R., 111 An automated
multidimensional protein identiﬁcation technology for shotgun proteomics
Anal. Chem. 2001, 73, 5683-5690.

Wu, C. C.; MacCoss, M. J. Shotgun proteomics: Tools for the analysis of
complex biological systems Curr 0pin.Mol. Ther 2002, 4, 242-250.

Zhu, W.; Reich, C. 1.; Olsen, G J .; Giometti, C. S.; Yates, J. R., IH Shotgun
proteomics of methanococcus jannaschii and insights into methanogenesis J.
Proteome Res. 2004, 3, 538-548.

Ulman, A. An introduction to ultrathin organic ﬁlms; academic press: Boston,
Ma. 1991.

Kirk, J. S.; Bohn, P. W. Surface adsorption and transfer of organomercaptans to
colloidal gold and direct identiﬁcation by matrix assisted laser
desorption/ionization mass spectrometryJ. Am. Chem. Soc. 2004, 126,
5920-5926.

Templeton, A. C.; Hostetler, M. J .; Kraft, C. T.; Murray, R. W. Reactivity of
monolayer—protected gold cluster molecules: Steric effects J. Am. Chem. Soc.
1998, 120, 1906-1911.

Schmelmer, U.; Jordan, R.; Paul, A.; Eck, W.; Golzhauser, A.; Grunze, M.;
Ulman, A. Synthesis of patterned polymer brushes via surface-initiated

polymerization on self-assembled monolayers Polym. Prepr 2003, 44,
562-563.

106

ﬂ griﬂvﬁnaule-uﬁ

(12) Schmelmer, U.; Jordan, R.; Geyer, W.; Eck, W.; Golzhauser, A.; Grunze, M.;
Ulman, A. Surface-initiated polymerization on self-assembled monolayers:
Ampliﬁcation of patterns on the micrometer and nanometer scale Angew. Chem.
Int. Ed. 2003, 42, 559-563.

(13) Park, J .-W.; Thomas, E. L. Nano-pattemed polymer brushes from a surface
reactive rod-coil diblock copolymer Polym. Mater. Sci. Eng. 2001, 85,
167-168.

(14) Benoit, D.; Husemann, M.; Mecerreyes, D.; Hawker, C. J.; Hedrick, J. L.; Shah,
R.; Abbott, N. L. Novel strategies for the preparation of patterned polymer
brushes using surface-initiated polymerization; 217th ACS national meeting,
Anaheim, CA, March 21-25, 1999.

107

 

   

11111111111111.11111111.111