v.5 ..
6m"
can. . .
“1.84:".ch ,
53”
fig. ..
in. .
. in?» .
a.» fl;

..
In,

 

339;“..u
.1 E»

in. iv.L.i..-:4
9 ”Wells“ l

: than“: ‘ ‘ .
. Lil: . V. 5:1»...1....MénudJQ4ln7smrL .SNAflt‘.
x . VHS...

@ffi

 

“42813

cl

LIB

”IIIIHIHIIMI‘I’I’H‘HW:IJI'I‘I‘IITI'I'I'I’IInmmI ”

3 1293 01770 5090
LIBRARY

Michigan State
University

This is to certify that the

dissertation entitled

SYSTEMATIC STUDIES OF CHARGED DERIVATIVES OF
PEPTIDES ANALYZED BY MASS SPECTROMETRY

presented by

Kenneth David William Roth

has been accepted towards fulfillment
of the requirements for

Ph . D . Chemistry

degree in

M

Date 111 [6% /77? fl Majorpmrm,

 

 

MS U i: an Affirmative Action/Equal Opportunity Institution 0-12771

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

 

 

 

 

 

DATE DUE DATE DUE DATE DUE
m 1 Q MAR 2 a 2004

L053 * 0 1

sap 192004

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1m elm-mu

 

SYSTEMATIC STUDIES OF CHARGED DERIVATIVES OF PEPTIDES
AN ALYZED BY MASS SPECTROMETRY

By

Kenneth David William Roth

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1 999

ABSTRACT

SYSTEMATIC STUDIES OF CHARGED DERIVATIVES OF PEPT IDES
ANALYZED BY MASS SPECTROMETRY

By

Kenneth David William Roth

Mass spectrometry is a powerful analytical technique for sequence analysis of
peptides and proteins. However, the fragmentation patterns of peptides are often
complex and difficult to interpret. Charge derivatization was developed as an approach
to simplify the fragmentation patterns of peptides and to increase the sequence
information available from analysis by mass spectrometry. This dissertation provides an
introduction to several types of mass spectrometers, the use of mass spectrometry for
peptide sequencing, and common types of peptide fragment ions. A review of charge-
derivatization techniques is also included with special attention given to the tris(2,4,6-
trimethoxyphenyl)-phosphonium-acetyl (T MPP+-Ac) derivatization technique.

The fragmentation of TMPPI-Ac-derivatized peptides is described in detail with
special attention paid to fragmentation during analysis by FAB-CAD-MS/MS. The
preparation of a weighable quantity of a purified TMPP+-Ac-derivatized peptide is also
described. Based on the UV absorbance of the derivatized and underivatized peptide,
reaction yields as high as 98.5% were observed. The selectivity of the TMPP+-Ac
reagent for the N—terminus was studied in the presence of lysine, cysteine, and tyrosine.
The derivatization reaction was studied in an unbuffered solution containing 4-dimethyl-

arninopyridine and in a pH 8.3 Tris-HCl buffer solution. The reagent reacted with lysine,

Kenneth David William Roth

cysteine, and tyrosine to some extent in both the buffered and unbuffered solution.
Derivatization in the buffered solution did not provide enhanced selectivity and resulted
in a decrease in the desired N-terminal-derivatized product. Derivatization of the
cysteine side-chain was more complete than derivatization of the N-terminus.
Derivatization of lysine and tyrosine was less complete than derivatization of the
N-terrninus. No signal enhancement was observed for a TMPP+-Ac-derivatized peptide
analyzed by MALDI-MS compared to the signal obtained from an equal amount of the
underivatized peptide. However, when a mixture of the derivatized and underivatized
peptide was analyzed by MALDI-MS, the signal intensity of the underivatized peptide
was suppressed.

The use of mass spectrometry and charge derivatization to distinguish between
normal and isomeric forms of aspartate (Asp) and glutamate (Glu) was studied. Small
underivatized peptides containing the isomeric residues B-Asp or 'y-Glu produced more
intense y ions and less intense [MW-H20] fragment ions than analogous peptides
containing normal Asp or Glu. The fragmentation differences were more dramatic for
pairs of charge-derivatized peptides. Charge-derivatized peptides containing B-Asp or
y—Glu produced *bn and *b...1+H20 fragment ions that were more intense, and *dn
fragment ions that were less intense, than those produced from analogous charge-
derivatized peptides containing normal Asp or Glu. Fragmentation mechanisms were

proposed to rationalize these differences in fragment ion intensities.

Copyright by
Kenneth David William Roth
1 999

ACKNOWLEDGMENTS

I want to thank my research advisor, Professor J. Throck Watson for his guidance
and support during my graduate career. I want to thank Professor John Allison for his
constructive criticism of, and suggestions for, my dissertation. I also want to thank the
students in Prof. Watson’s lab and the staff of the Mass Spectrometry Facility for their
assistance and friendship during my time at Michigan State.

Most importantly,I want thank my wonderful wife, Jennifer, for her love and
support. She has worked and sacrificed just as much as I have in order to complete my
studies. Even in the difficult times, she never stopped believing in me, loving me, and
encouraging me. I couldn’t have done it without you, Jennifer. I also want to thank our
children, Nathan and Maggie, for sacrificing time with their Dad, so that he could finish
school.

I want to thank my parents, Professor R. Waldo Roth and Mrs. Marlene Roth for
teaching me the importance of an education and for giving me a love for learning. Mom
and Dad, thank you for modeling unconditional love and for teaching me that character is
more important than status or degrees. Thank you also to my other parents, Monty and
Kathy Riffer for treating me like a son and for their support. In addition, I want to thank
the other members of my family and my friends who helped support our family during
the long process of completing my studies. I am humbled by, and grateful for, your

support.

TABLE OF CONTENTS

LIST OF TABLES ........................................................................................................... viii
LIST OF FIGURES ............................................................................................................ ix
CHAPTER 1: INTRODUCTION ...................................................................................... 1
I. Introduction ............................................................................................................. 1
II. Ionization Techniques ............................................................................................. 2

A. Fast Atom Bombardment (FAB) and Liquid Secondary-Ionization
Mass Spectrometry (LSIMS) ........................................................................... 2
B. Matrix-assisted Laser Desorption/Ionization (MALDI) .................................. 3
C. Electrospray Ionization (ESI) ........................................................................... 4
III. Mass Analyzers ....................................................................................................... 7
A. Sector Analyzers .............................................................................................. 7
B. Time-of-Flight (TOF) Analyzers ................................................................... 10
C. Quadrupole Analyzers .................................................................................... 12
D. Ion Trap Analyzers ......................................................................................... 15
IV. Peptide Sequencing by Mass Spectrometry .......................................................... 20
- V. Peptide Fragmentation ........................................................................................... 21
VI N -Terminal-Charged Derivatives .......................................................................... 30
A. Quaternary Ammonium Derivatives .............................................................. 30
B. Quaternary Phosphonium Derivatives ........................................................... 35
VII. N-Terminal Derivatives with High Proton Affinities ............................................ 40
VIII. C-Terminal-Charged Derivatives .......................................................................... 42
D(. Side-Chain Derivatives .......................................................................................... 46
X. Negatively—Charged Derivatives ........................................................................... 48
XI. Concluding Commentary ...................................................................................... 50
XII. Summary ............................................................................................................... 53
XIH. References ............................................................................................................. 55

CHAPTER 2: STUDIES OF TRIS(2,4,6-TRIMETHOXYPHENYL)PHOSPHONIUM-
ACETYL (TMPP+-AC) DERIVATIZATION FOR PEPTIDE SEQUENCING BY MASS

SPECTROMETRY ........................................................................................................... 63
I. Introduction ........................................................................................................... 63
II. Experimental ......................................................................................................... 63
A. Preparation of TMPP+-Ac-Succinimide Bromide ([TMPP+-Ac-OSu][Br-]). 63

B. Preparation of TMPP+-Ac-S-Pentafluorophenyl Bromide
([TMPP+-Ac-SPf][Br-]) ......................................................................... 65
C. TMPP+-Ac Derivatization of Peptides in Unbuffered Solution ..................... 67
D. TMPP+-Ac Derivatization of Peptides in Buffered Solution ......................... 67
E. HPLC Separation ........................................................................................... 67
F. Analysis .......................................................................................................... 69
Fast Atom Bombardment (FAB) .................................................................... 69
MALDI ........................................................................................................... 70

vi

ESI—MS on an Ion Trap Instrument ................................................................ 70

ESI-MS on a Triple Quadrupole Instrument .................................................. 71
ESI-In-Source Fragmentation (ISF)-MS ........................................................ 71
III. Fragmentation of TMPP+-Ac-derivatized Peptides ............................................... 72
A. Fragmentation of theTMPP+-Ac Group ......................................................... 72
B. Sequence-specific Fragmentation of TMPP+-Ac-derivatized Peptides by
FAB-CAD-MS/MS ........................................................................................ 77
C. Sequence-specific Fragmentation of TMPP+-Ac-derivatized Peptides by
MALDI-PSD, ESI-CAD-MS/MS, and ESI-ISF-MS ..................................... 85
IV. Preparation of a Weighable TMPP+-Ac Derivative Standard ............................... 86
V. Reaction Yields and Selectivity ............................................................................ 88
A. Reaction Yields in Buffered and Unbuffered Solutions ................................ 88
B. TMPP+-Ac Derivatization of Lysine .............................................................. 89
C. TMPP+-Ac Derivatization of Cysteine ........................................................... 93
D. TMPPI-Ac Derivatization of Tyrosine .......................................................... 97
VI. Ionization Efficiency and Signal Supression ....................................................... 101
VII. References ........................................................................................................... 104
CHAPTER 3: FRAGMENTATION BEHAVIOR OF PEPTIDES AND CHARGE-
DERIVATIZED PEPTIDES CONTAINING ISOMERIC ACIDIC RESIDUES .......... 106
I. Introduction ......................................................................................................... 106
A. Isomeric Residues ........................................................................................ 106
B. Formation of B-Aspartate and y—Glutamate ................................................. 107
C. Conventional Analysis of B-Aspartate and 'y-Glutamate .............................. 111
D. Mass Spectrometric Analysis of B-Aspartate and y-Glutamate ................... 113
11. Experimental ....................................................................................................... 116
A. Reagents ....................................................................................................... 116
B. Charge Derivatization .................................................................................. 117
C. Analysis ........................................................................................................ 118
III. Results and Discussion ........................................................................................ 120
A. Underivatized Peptides ................................................................................. 120
y Ions ............. _ ............................................................................................... 122
[MW-H20] and [MH+-NH3] Ions ................................................................ l 3 1
w and v Ions ................................................................................................. 135
Cm], bnq'l'Hzo, and bu IOI‘IS .......................................................................... 135
Internal Ions ................................................................................................. 136
B. Charge-derivatized Peptides ......................................................................... 136
I"bu Ions ........................................................................................................ 138
I"(1.,I Ions ........................................................................................ 145
*b...1+H20 Ions ............................................................................................ 150
*c...1 Ions ...................................................................................................... 153
*an Ions ........................................................................................................ 155
[c-Hzor and [C+H-H20]2+ ......................................................................... 157
IV. Conclusions ......................................................................................................... 157
V. References ........................................................................................................... 160

vii

Table 2.1

Table 2.2

Table 3.1

Table 3.2

Table 3.3

Table 3.4

Table 3.5

Table 3.6

LIST OF TABLES

Common TMPP+-Ac fragment ions observed in MS/MS spectra ................. 74

Intensities of fragment ions from TMPP+-Ac-derivatized peptides as a
function of amino acid residue ....................................................................... 78

Abundance ratios for underivatized peptides analyzed by FAB-CAD-MS/MS
...................................................................................................................... 123

Abundance ratios for underivatized peptides analyzed by MALDI-PSD 124

Abundance ratios for underivatized peptides analyzed by ESI-CAD-MS/MS ..

...................................................................................................................... 125
Branching ratios for derivatized peptides analyzed by FAB-CAD-MS/MS ......
...................................................................................................................... 139
Branching ratios for derivatized peptides analyzed by MALDI-PSD ......... 140

Branching ratios for derivatized peptides analyzed by ESI-CAD-MS/MS . 141

viii

Figure 1.1
Figure 1.2
Figure 1.3
Figure 1.4
Figure 1.5

Figure 1.6

Figure 1.7

Figure 1.8

Figure 1.9

Figure 1.10
Figure 1.11
Figure 1.12
Figure 1.13
Figure 1.14
Figure 1.15
Figure 1.16
Figure 1.17
Figure 1.18
Figure 1.19
Figure 1.20

Figure 1.21

LIST OF FIGURES

Diagram of an electrospray source ................................................................ 5
Diagram of a quadrupole ion trap analyzer ................................................. 16

The effect of charge localization on fragmentation by high-energy CAD .. 22

Proposed structures of peptide fragment ions .............................................. 24
Alternative structures of peptide fragment ions ........................................... 25
Proposed structures of fragment ions produced from charge-derivatized

peptides ........................................................................................................ 27
Methyl iodide derivatization scheme ........................................................... 31
Trialkylammonium-acetyl derivatization scheme ....................................... 31
Thiocholine iodide derivatization scheme ................................................... 31
“CSQ” derivatization scheme ...................................................................... 31
N-terminal triphenylphosphonium derivatization scheme ........................... 36
TMPP+-Ac derivatization scheme ............................................................... 36
Derivatization schemes of Renner and Spiteller .......................................... 36

N-hydroxysuccinimide-Z-(3-pyridyl)acetate (SPA) derivatization scheme 41

C-terminal pyridinium derivatization scheme ............................................. 41
C-terminal a'cyl urea derivatization scheme ................................................ 41
Carboxypeptidase derivatization scheme .................................................... 41
C-terminal N,N-dimethylethanolamine derivatization scheme ................... 45
C-terminal triphenylphosphonium derivatization scheme ........................... 45

2,4,6-trimethylpyridinium derivatization scheme for lysine side-chains 45

Conversion of lysine side-chains to homoarginine ...................................... 47

ix

Figure 1.22
Figure 1.23
Figure 2.1
Figure 2.2
Figure 2.3

Figure 2.4

Figure 2.5

Figure 2.6

Figure 2.7

Figure 2.8

Figure 2.9

Figure 2.10

Hymln

fiwmlu

Figure 2.13

Figure 2.14

Figure 2.15

Pentafluorobenzoyl fluoride derivatization scheme .................................... 47

C-terminal 4-aminonaphthalenesulphonic acid (AN SA) derivatization ...... 47

Synthesis of TMPP+-Ac reagents ................................................................ 64
FAB-MS spectra of TMPP+-AC-OSu (A) and TMPP+-AC-SPf (B) ............. 66
TMPP+-Ac derivatization scheme ............................................................... 68

FAB-CAD-MS/MS spectra of TMPP+-Ac—OSu (A) and TMPP+-Ac-SPf (B)
..................................................................................................................... 73

Low mass fragments of TMPPI-Ac-derivatized peptides analyzed by FAB-
CAD-MS/MS (A) and MALDI-PSD (B) .................................................... 75

Low mass fragments of TMPP+-Ac-derivatized peptides analyzed by ESI-
CAD-MS/MS on an ion trap (A) and ESI-ISF-MS (B) ............................... 76

Formation of *an ions involving transfer of hydrogen from the B-carbon .. 80

FAB-CAD—MS/MS spectra of TMPP+-Ac-osteocalcin fragment (45-49),
C+ = 1154, showing *an-14, *an-30, and *an-46 peaks ............................... 82

Possible rationalization for losses of 14, 30, and 46 Da from TMPP+-Ac
derivatives .................................................................................................... 84

HPLC chromatographs of TMPP+-Ac-OSu + buccalin reaction mixtures
performed in unbuffered (A) and buffered (B) solutions ............................ 87

HPLC chromatographs of TMPP+-Ac—OSu + A-‘y-EKAA reaction mixtures
performed in unbuffered (A) and buffered (B) solutions ............................ 90

MALDI-PSD spectra of N-terminal (A), side-chain (B), and doubly-
derivatized (C) A-y-EKAA .......................................................................... 92

HPLC chromatographs of TMPP+-Ac-OSu + y—ECG reaction mixtures
performed in unbuffered (A) and buffered (B) solutions ............................ 94

ESI-CAD-MS/MS spectra of N-terminal (A), side-chain (B), and doubly-
derivatized (C) y-ECG ................................................................................. 96

HPLC chromatographs of 'IMPP+-Ac-OSu + VQAAIDYING-NH; reaction
mixtures performed in unbuffered (A) and buffered (B) solutions ............. 98

Figure 2.16

Figure 2.17

Figure 3.1
Figure 3.2

Figure 3.3

Figure 3.4

Figure 3.5
Figure 3.6
Figure 3.7
Figure 3.8

Figure 3.9

Figure 3.10

Figure 3.11

Figure 3.12

Figure 3.13
Figure 3.14

Figure 3.15

Figure 3.16

MALDI-PSD spectra of N-terminal (A) and side-chain (B) derivatized
VQAAIDYING-NHZ, and the ESI-CAD-MS/MS spectrum of the doubly-

derivatized (C) product ................................................................................ 99
MALDI-MS spectra of buccalin (A), TMPP+-Ac-buccalin (B), and an

equimolar mixture of the two (C). ............................................................. 102
Structures of aspartate, B-aspartate, glutamate, and y-glutamate ............... 108
Formation of B-aspartate in solution .......................................................... 109

ESI-CAD-MS/MS spectra of Glu-Cys-Gly (A) and y-Glu-Cys-Gly (B) at a
relative collision (energy of 14.9% ............................................................. 126

ESI-CAD-MS/MS spectra of Asp-Phe-OMe (A) and B-Asp-Phe-OMe (B) at

a relative collision energy of 17.4% .......................................................... 127
y ion formation by glutamyl (A) and 'y-glutamyl (B) peptides .................. 128
y ion formation by aspartyl (A) and B-aspartyl (B) peptides ..................... 130

[MW-H20] formation by glutamyl (A) and 'y-glutamyl (B) peptides ....... 133
[MH‘i-HZO] formation by aspartyl (A) and B-aspartyl (B) peptides .......... 134

ESI-ISF-MS spectra of AEKAA (A) and A-y-EKAA (B) at a cone voltage
of 55 V ....................................................................................................... 137

MALDI-PSD spectra of TMPP+-Ac-Asp-Phe-0Me (A) and
TMPPI-Ac-B-Asp-Phe-OMe (B) ............................................................... 142

*b.. formation by derivatized aspartyl (A) and B-aspartyl (B) peptides 144

MALDI-PSD spectra of TMPP+-Ac-Asp-Phe-OMe (A) and methylated
TMPP+-Ac-Asp-Phe-0Me (B) .................................................................. 146

MALDI-PSD spectra of derivatized AEKAA (A) and A-y-EKAA (B) 148
*dn formation by derivatized aspartyl (A) and B-aspartyl (B) peptides 149

*b...1+HzO formation by derivatized aspartyl (A) and B-aspartyl (B)
peptides ...................................................................................................... 152

*cM formation by derivatized aspartyl (A) and B-aspartyl (B) peptides .. 154

xi

Figure 3.17 *cM formation by derivatized aspartyl (A) and glutamyl (B) peptides
without transfer of the acidic proton .......................................................... 156

Figure 3.18 ESI—CAD-MS/MS spectra of the C+ (A) and [C+H]2+ (B) precursor ions of
TMPP+-Ac-AEKAA at collision energies of 72.5 eV and 32.5 eV,

respectively ................................................................................................ l 58

xii

CHAPTER 1

INTRODUCTION

1. Introduction

Mass spectrometry is a powerful analytical technique based on the principle that
the motion of charged particles (ions) in a vacuum can be controlled by the use of electric
and magnetic fields. The behavior of these ions is also related to the mass-to-charge
ratio. Therefore, by subjecting ions to an electric or magnetic field in a vacuum, it is
possible to determine the mass—to-charge ratio (m/z) of those particles. If the charge is
known, then the mass of the ion can be determined. When fragmentation of the analyte is
induced, the m/z values of the fragment ions can provide clues to the structure of the
original analyte.

This chapter briefly discusses some of the techniques for forming ions from the
condensed phase, some of the commonly used approaches for the determination of m/z,
and some methods for inducing fragmentation. In addition, the use of mass spectrometry
for the analysis of peptides is introduced, along with a discussion of commonly observed
peptide fragment ions. Finally, the use of charge-derivatization to aid the sequence
determination of peptides is discussed, and the scientific literature involving charge-

derivatization is reviewed.

II. Ionization Techniques
A. Fast Atom Bombardment (FAB) and Liquid Secondary-Ionization Mass
Spectrometry (LSIMS)

Although numerous desorption/ionization techniques can be used to generate gas-
phase ions for analysis by mass spectrometry, only three are discussed here: fast atom
bombardment (FAB), matrix-assisted laser desorption/ionization (MALDI), and
electrospray ionization (ESI). Fast atom bombardment was introduced by Barber et al.
(1). During the desorption/ionization process, a beam of Ar or Xe atoms with several
keV of translational energy strikes a solution of analyte dissolved in a liquid matrix.
Commonly used matrices include glycerol, thioglycerol, and 3-nitrobenzyl alcohol. The
translational energy from the primary ion beam is transferred to the analyte and matrix,
resulting in desorption of the analyte and the matrix into the gas phase. Ionization of the
analyte has been proposed to occur in the solution prior to desorption or by proton
transfer in the gas phase as a result of collisions with ionized matrix molecules in the high
pressure selvedge region (2, 3).

FAB is closely related to two other desorption/ionization techniques: secondary
ionization mass spectrometry (SIMS) and liquid secondary ionization mass spectrometry
(LSIMS). LSIMS differs from FAB in that the primary beam is composed of ions
(usually Cs+) instead of neutral atoms. Desorption/ionization by LSIMS has been
reported to give more efficient ionization of analytes and lower detection limits than FAB
as a result of better focusing of the primary beam (4). The mass spectra obtained by FAB
and LSIMS are very similar (5); as a result, the terms FAB and LSIMS are sometimes

used interchangeably. SIMS was developed prior to the development of FAB and differs

from FAB and LSIMS in that the primary beam strikes an analyte in the solid phase
without any matrix. The extent of fragmentation during desorption/ionization by SIMS is
much greater than that during desorption/ionization by FAB or LSIMS. As a result,
SIMS is usually limited to the atomic analysis of surfaces.

In FAB or LSIMS, a primary atom or ion with a kinetic energy in the keV range
strikes the matrix, and the energy is dispersed by collisional processes among the matrix
and analyte molecules that occupy about the top 80 A of the matrix surface (6). Some of
this energy is transferred into kinetic energy for the matrix and analyte, allowing
desorption into the gas phase. Another portion of this energy is transferred into internal
energy, producing ions with internal energies in the range of 8—20 eV (7). Once
desorption/ionization has occurred, the ions are accelerated toward the analyzer by a

strong potential gradient.

B. Matrix-assisted Laser Desorption/Ionization (MALDI)

The use of lasers for desorption/ionization of organic compounds for subsequent
analysis by mass spectrometry dates back to the 19705 (8). However, the use of laser
desorption was limited by the extensive fragmentation induced during the
desorption/ionization process. The introduction of a crystalline matrix to aid the
desorption/ionization process resulted in a technique called matrix-assisted laser
desorption/ionization (MALDI) (9).

In preparation for analysis by MALDI—MS, the analyte is co—deposited with a large
excess of organic matrix. Some of the most commonly used matrices are a-cyano-4-

hydroxycinnamic acid, 2,5-dihydroxybenzoic acid, and 3,5-dimethoxy-4-hydroxy-

cinnamic acid (sinapinic acid). These matrices have strong UV absorption in the 320-350
nm range to allow efficient energy transfer from a nitrogen laser (337 nm) or a frequency-
doubled Nd—YAG laser (355 nm). These matrices also crystallize readily, have low vapor
pressures, low energies of sublimation, and can readily donate protons to the analyte
during the desorption/ionization process (10). A fraction of the energy absorbed by the
matrix is transferred to the analyte, apparently by the coupling of the lattice vibrations of
the matrix to the internal vibrations of the analyte (11). The ionization mechanism is not
clearly understood, although several possible mechanisms have been proposed ( 12).
Although the mechanisms of desorption and ionization are not well understood,
the use of MALDI as a desorption/ionization technique has grown tremendously during
recent years. This growth is due to the ability to desorb biomolecules with masses of

hundreds of kilodaltons and the ability to detect analytes in the picomole range.

C. Electrospray Ionization (ESI)

Fenn and coworkers (13) developed electrospray as an ionization source for mass
spectrometry, building on the work of ' Dole et al. (14). Common features of the
electrospray source are shown in Figure 1.1. Electrospray ionization may be a misnomer
because the technique primarily involves the transfer of ions in solution to the gas-phase
rather than the ionization of neutral species. During the electrospray process, a solution
of analyte is passed through the electrospray capillary held at high potential. As the
solution exits the electrospray capillary, the high electric field generates a spray of highly-

charged droplets at atmospheric pressure. As the droplets pass through a potential and

850w >mamo=oofi cm Co Emcmmfi. .: 939".

ES 1: u C E: F u n:
E::om> Scan. 30
:9... 330m 9.35 N2

23me
05530.5(

 

.mNAII>II._mc<

W

t

.oEEEw 980 $2865
ENNoz .2500

. . . . amino
>meamocfim_m
9.00
mean. 5:3...
Famozomm

 

fl ....... . .. . . A5 23.2..

pressure gradient toward the analyzer, the solvent is removed from the analyte until only
individual gas-phase ions remain. One important feature of the electrospray process is
that the charge on the ions produced is similar to the charge of the ions in solution. As a
result, multiply—charged ions are readily produced by this mechanism.

When the analyte solution reaches the electrospray capillary held at a high
positive potential, positive ions in the solution accumulate at the surface of the solution.
The electric field draws the charged solution into a cone shape, called a Taylor cone. At
the tip of the cone, small droplets (diameter z 1 um) are drawn into the gas phase to
produce the spray (15). Most commercial instruments also use a nebulization gas to
assist in the formation of the spray, especially at high flow rates (>50 [IL/min).

After droplets have been sprayed, the size of the droplet decreases by solvent
evaporation until the Rayleigh limit is reached. Solvent evaporation is assisted in many
instruments by the use of a drying gas or heating of the source. The Rayleigh limit is the
point at which the charge repulsion becomes sufficient to overcome the droplet surface
tension and is given by the following equation:

QR2 = 641C280'YRR3 (Equation 1.1)
where QR is the charge in coulombs, so is the permitivity of vacuum, 7 is the surface
tension, and RR is the radius in meters. When the Rayleigh limit is reached, the droplet
undergoes Coulombic explosion resulting in the formation of numerous smaller droplets.
The processes of solvent evaporation and Coulombic explosion are repeated until very
small droplets (< 10 nm in diameter) are formed. There are two theories that have been
proposed to describe the transition from very small droplets to single ions. The single ion

droplet theory (16) proposes that Coulombic explosion eventually produces very small

droplets containing only a single ion; and that solvent evaporation of that droplet
produces a gas-phase ion. The ion evaporation theory (17, 18) holds that Coulombic
explosion produces small droplets containing multiple ions; and that individual ions
evaporate from the droplet into the gas phase.

Before reaching the analyzer, the ions pass through a nozzle, also called a cone,
and then through a skimmer. In the first differentially-pumped region (pressure z 1 torr),
collisional fragmentation can be introduced by raising the voltage gradient in the region
and giving the ions enough energy to fragment during collisions. This type of
fragmentation is called in-source fragmentation (ISF), nozzle-skimmer dissociation, or
cone voltage fragmentation. More detailed information on electrospray ionization can be

found in several recently published reviews (19-22).

111. Mass Analyzers
A. Sector Analyzers

Sector analyzers use magnetic fields, electric fields, or both to determine the m/z
ratios of ions generated in the source. The kinetic energies of ions leaving the source
region are

KE = v. mv2 = zeV (Equation 1.2)

where KE is kinetic energy, m is mass, v is velocity, 2 is the number of elementary
charges on the ion, e is the electronic charge, and V is the acceleration potential. When
an ion enters a magnetic field, B, the ion follows a circular path that is perpendicular to
the direction of the magnetic field. The radius, Rm, of the circular path is given by:

Rm = mv/Bze or (m/z)v = RmBe. (Equation 1.3)

Combining Equations 1.2 and 1.3 gives the equation for separation of ions by a magnetic
sector analyzer:
m/z = Rm2B2/2eV. (Equation 1.4)

Therefore, by using a fixed value of Rm and scanning B, ions of different m/z values can
be made to pass through a detector slit to acquire a mass spectrum. Equation 1.4 assumes
that the kinetic energy of the ions is determined only by the m/z value and the
acceleration potential. In reality, ions with the same m/z value can have different kinetic
energies and velocities as a result of the ionization process and thermal energy. This
difference in velocity for ions of the same m/z can alter the flight path of the ions, as
shown in Equation 1.3, resulting in a loss of resolution.

An electric sector analyzer with an electric field, E, deflects ions in a circular path
with a radius, Re:

R, = mv2/zeE = KE/22eE. (Equation 1.5)

As a result, an electric sector analyzer can be used in conjunction with a magnetic sector
analyzer to compensate for the initial kinetic energy spread. Isomass ions with large
kinetic energies are deflected by a larger radius and than those with small kinetic
energies. By carefully selecting the electric field, the difference in the R, of isomass ions
is used to compensate for the difference in Rm caused by differences in kinetic energy.
This allows isomass ions to reach the detector at the same time, resulting in an increase in
resolving power.

Sector instruments that utilize both an electric and magnetic sector (double
focusing instruments) can be used to generate MS/MS data by a technique called linked

scanning. When fragmentation occurs in the field-free region between the source and the

first sector analyzer, the precursor ion with mass m; produces a product ion with mass
mg. The fragment ion will have a velocity that is nearly equal to the velocity of the
precursor ion v1, but it will have a smaller kinetic energy. Based in Equation 1.5, the
conditions that allow the precursor ion and product ion to pass through the electric sector
are

eEl = m1v12/Rcz (Equation 1.6)

eEz = mzvlszcz (Equation 1.7)
where E1 is the electric field that passes the precursor ion and E2 is the electric field that
passes the product ion. Based on Equation 1.3, the conditions that allow the precursor
and product ions to pass through the magnetic sector are

CE} = mm/Rmz (Equation 1.8)

eBz = m2v1/Rmz. (Equation 1.9)
Combining Equations 1.6-1.9 results in the following relationship:

BI/El = BZ/Ez = RJme = constant (Equation 1.10)

Therefore, by holding the electric and magnetic fields at a constant ratio, only ions arising
from the precursor ion (those with a velocity of VI) will reach the detector. By scanning
the electric and magnetic fields at a constant ratio, a product ion mass spectrum can be
acquired. Linked scanning can be used to detect metastable ions, or following the
introduction of a collision gas into the first field-free region, to detect ions generated by

high-energy collisionally-activated dissociation (CAD) from a precursor ion.

B. Time-of-Flight (TOF) Analyzers

Time-of—flight (TOF) analyzers are simple and inexpensive analyzers that are
well-suited for pulsed ionization techniques, such as MALDI. A time-of—flight analyzer
works by subjecting a group of ions to an accelerating potential, V, in the ion source,
allowing the ions to drift down a field—free flight tube of length, L, and measuring the
time, t, required to drift down the tube to a detector. Time-of-flight analysis is based on
the relationship between kinetic energy, velocity, potential, and mass-to-charge ratio,
defined previously in Equation 1.2. Solving Equation 1.2 for velocity yields

v = (2eV)"S (In/z)”. (Equation 1.11)
Equation 1.11 shows that when a group of ions is placed in the same electric field, the
velocity of each ion is dependent on the m/z value. Since the time required to traverse the
flight tube is UV, the time-of-flight is:
TOF = t = Lao-V)”: (m/z)"“ . (Equation 1.12)
By measuring the time-of—flight, the m/z of each ion can be determined. Unlike, scanning
analyzers, time~of—flight analyzers allow analysis of all the ions present in the source at a
given time. As a result, TOF analyzers usually are more sensitive than scanning
analyzers. Also, TOF analyzers have no theoretical upper mass limit.

The resolving power of TOF analyzers is usually lower than that achieved by
other analyzers. In the ion source, ions of the same m/z value can have differences in
their kinetic energies and they can be moving in different directions. These differences
affect the time-of-flight and produce peak broadening. Also, not all ions are the same
distance from the analyzer when the accelerating potential is applied. This difference

affects the flight distance of the ions, resulting in loss of resolution. The limited

10

resolving power of TOF analyzers has been addressed by the introduction of reflectron
TOF and by the introduction of delayed extraction. Reflectron TOF was introduced by
Mamyrin et al. (23), and it uses a reflectron or ion mirror to compensate for the initial
kinetic energy spread of the analyte. At the end of the flight tube, the ions are reflected
by a potential field to a detector located part of the way back down the flight tube. Ions
with a higher kinetic energy will travel farther into the ion mirror than ions with lower
kinetic energy, allowing ions with smaller kinetic energies to reach the detector at the
same time as those with greater kinetic energies. As a result, the peak broadening from
differences in kinetic energy is minimized.

Delayed extraction is based on the principle of time-lag focusing developed by
Wiley and McLaren (24). Delayed extraction uses a delay between the formation of ions
and the extraction of those ions from the source in order to compensate for differences in
the initial velocities of the ions. After the ions are formed, they move in a variety
directions and speeds. After the delay time has passed, a two-field acceleration potential
is applied that gives a greater kinetic energy to the ions farther from the detector than to
the ions closer to the detector. As a result, the ions that were moving away from the
detector most rapidly receive the biggest increase in velocity and those moving most
rapidly toward the detector receive the smallest increase in velocity. Therefore, the ions
farther from the detector are able to reach it at about the same time as those closest to the
detector at the time of acceleration. The delay time that gives the highest resolving power
is mass dependent, so the benefits from delayed extraction decrease as the m/z of the

analyte gets farther from the m/z for which the delay time was optimized.

ll

MS/MS analysis is possible for reflectron TOF analyzers by a technique called
post-source decay (PSD). PSD was introduced by Spengler et al. in 1992 (25) to allow
MS/MS analysis of peptides ionized by MALDI. As shown in Equation 1.1], the velocity
of an ion leaving the acceleration region is dependent on its m/z value. If the ion
fragments in the flight tube, the fragments will have the same velocity as the precursor
ion, but they will have smaller kinetic energies than the precursor. In a linear TOF
analyzer, the fragment ions will reach the detector at the same time as the intact ions, but
if the fragment ions are reflected by an ion mirror, they will be separated on the basis of
kinetic energy and therefore on the basis of mass

m = 2KE/v2. (Equation 1.13)
PSD analysis uses an ion gate to deflect all the ions except those with a velocity equal to
that of the desired precursor. The reflectron then separates the fragment ions on the basis
of mass and sends them to the detector to be recorded as a PSD spectrum. In practice,
PSD spectra are usually collected in several segments so that the reflectron voltages can
be adjusted to give the best resolution for the fragment ions. The individual segments are
then “stitched” together by computer to give a complete PSD spectrum. PSD spectra
have been observed to resemble low-energy CAD spectra, but they have traits of both

low-energy and high-energy CAD spectra (26).

C. Quadrupole Analyzers
The quadrupole mass analyzer consists of four parallel rods, ideally of hyperbolic
cross-sections that lie along the z-axis. Two rods lie in the x-z plane, and two rods lie in

the y-z plane. The radius from the center axis of the quadrupole to the surface of a

12

hyperbolic rod is r0. Most commercial instruments use cylindrical rods instead of
hyperbolic rods because cylindrical rods are easier to manufacture. When cylindrical rods
are separated by a radius of 1.148%, they can generate a good approximation of the ideal
hyperbolic field (27).
For the analysis of positive ions, the potential applied to the rods in the x-z plane
is identical and consists of a positive direct current (dc) component with an amplitude U
and an alternating current (ac) component with an amplitude V and a frequency f. The
frequency of the ac potential lies within the radiofrequency (rf) range. The rods in the y-z
plane have an rf potential with the same amplitude, V, and frequency, f, as that applied to
the x-z rods, but the two waveforms are 180° out of phase (opposite in sign). Also, the dc
potential on the y-z rods is the negative (-U) of that applied to the x-z rods. The potential
distribution ((1)) within the quadrupole at any given time, t, can be described by the
expression
(1) = [U + Vcos(o)t)][(x2 — y2) / 2r02] (Equation 1.14)
where to is the angular frequency (2111). The force, F, exerted on a charged particle along
each axis can be determined by taking the partial derivative of Equation 1.14 with respect

to each axis and multiplying by the charge on the particle (ze):

Fx = -[U + Vcos(o)t)](zex/r02) (Equation 1.15)
F, = [U + Vcos((ot)](zey/r02) (Equation 1.16)
Fz = 0. (Equation 1.17)

Equation 1.17 shows that the quadrupole analyzer does not affect the motion of ions in
the z direction. Since F = ma, ions with a small mass undergo greater acceleration than

ions with a larger mass. As a result, low-mass ions oscillate in response to the changing

13

rf potential, whereas high-mass ions are relatively unaffected by the rf potential. The
force exerted by the dc potential U along the x-axis acts to focus ions toward the center of
the quadrupole. However, ions with m/z values below a certain level will collide with the
quadrupole rods and be neutralized as a result of the oscillations induced by the rf
potential. The dc potential along the y-axis attracts ions to the rods in the y-z plane.
Since high-mass ions are relatively unaffected by the rf potential, ions with a m/z value
greater than a cutoff value will collide with the rods in the y-z plane. Therefore, the x-z
rods act as high-pass filters and the y-z rods act as low-pass filters. Only ions with Hill
values that are stable in both the x and the y directions are able to reach the detector.
Usually the width of this stability region is less than one dalton. By ramping the
potentials U and V at a fixed ratio, the m/z values of ions that are stable in the quadrupole
are scanned, producing a mass spectrum after the ions are detected.

Based on the relationship F = ma and the Equations 1.15-1.17, the trajectories of
ions in the x and y directions can be described by the following equations:

dzx/dt2 + (zex/mr02)[U + Vcos((ot)] = 0 (Equation 1.18)
dzy/dt2 - (zey/mr02)[U + Voos(cot)] = 0 (Equation 1.19)

These equations are related to the Mathieu differential equations that describe the motion
of ions in a quadrupole field. The Mathieu equations are discussed in more detail in the
section on ion trap analyzers.

Tandem mass spectrometric analysis (MS/MS) can be accomplished using a triple
quadrupole instrument. The first and third quadrupoles use a combination of dc and rf
potentials to allow ions with selected m/z values to pass, as described above. In the

second quadrupole, precursor ions selected by the first quadrupole collide with a collision

l4

gas to form product ions by low-energy CAD. The second quadrupole uses only an rf
potential to focus the product ions formed and to transmit them to the third quadrupole.
The third quadrupole scans the product ions formed to generate a mass spectrum. A
product ion scan is accomplished by setting the first quadrupole to pass only ions of a
given m/z value, and scanning the product ions formed. A precursor ion scan is
accomplished by setting the third quadrupole to pass only product ions of a given m/z
value and scanning the first quadrupole to determine which precursor ions fragment to
from the desired product. A constant neutral loss scan is accomplished by scanning the
first and third quadrupoles with a fixed difference in the m/z values passed by each. In
this way, only precursor ions that fragment to form product ions resulting from a given

neutral loss can reach the detector.

D. Ion Trap Analyzers

The quadrupole ion trap can function as both an instrument to store gas-phase ions
and as a mass (m/z) analyzer. Figure 1.2 shows the major components of one commercial
ion trap analyzer. The ion trap has three electrodes that generate the potential wells for
storing and analyzing ions. In the center of the instrument is a ring electrode that is
shaped like a doughnut with a hyperbolic interior. At the two open ends of the ring
electrode are end-cap electrodes, also with hyperbolic faces. The potential well that traps
ions is generated when a radio-frequency (rf) potential is applied to the ring electrode and
the end-cap electrodes are grounded.

The motion of ions in a quadrupole field can be described mathematically by the

15

.cozmcoeoo «mos—8.59:. E9». co_mm_:.:oa 53> “695.com
.._oN>_mcm an: :2 283653 a .6 89920 .N... .239”.

ouchoofi
amoucm mocmbcm

     

ouoboom
9E
209300

ouoboofi oocmbcm

 

AIII

Alli
5 mac.

88¢qu
9 So go.

 

 

 

16

solutions to the Mathieu equation (28):

dzu/dfi2 + (au — 2quc082§)u = 0 (Equation 1.20)
where u represents the coordinate axes (x, y, and 2), § is a dimensionless parameter equal
to tot/2, (t) is the radial frequency (in radians/second) of the rf potential applied to the ring
electrode, t is time, and an and qu are additional dimensionless parameters known as
trapping parameters. Equation 1.20 can be rearranged to describe the force on an ion:

Fu = ma = m(d2u/dt2) = -(m(t)2/4)(au — 2qucostot)u (Equation 1.21)
where m is mass and a is acceleration. The force on an ion can also be expressed as the
partial derivative of the potential with respect to the coordinate axis, u:

Fu = -e(B(I>/au) (Equation 1.22)
where e is the electronic charge and (I) is the potential at any point within the field. The
quadrupole potential can be expressed as:

(I) = (Do/r02 (xx?- + cry2 +712) (Equation 1.23)
where (Do is the applied electric potential on the ring electrode, 2., G, and y are weighing
constants for the x, y, and z coordinates, and r0 is the internal radius of the ring electrode,
as shown in Figure 1.2. The applied electric potential on the ring electrode is expressed
as:

(Do = U + Vcos(o)t) (Equation 1.24)
where U is the amplitude of the direct current (dc) potential, and V is the amplitude of the
rf potential. For an ion trap, Equations 1.21-1.24 can be simplified to find the trapping
parameters, a and q (29):

ar = 8eU / m(r02 + 2202)(02; qr = -4eV / m(r02 + 2202)(02 (Equations 1.25)

17

a2 = -l6eU / m(ro2 + 2202)w2; q2 = 8eV / m(r02 + 2202x102 (Equations 1.26)
where ar and qr are trapping parameters in the radial direction, a2 and q2 are the trapping
parameters in the z direction, r0 is the internal radius of the ring electrode, and Zn is V2 the
distance between the end-cap electrodes, as shown in Figure 1.2. It should be noted that
the trapping parameters a and q are inversely proportional to m/e, however these
equations were derived based on the assumption of a charge of 1, see Equation 1.22.
Actually, the term e should be replaced by ze where z is the number of charges on an ion.
Having made that correction, it can be seen that the trapping parameters are inversely
proportional to Hill.

The solutions to Mathieu's equation describe regions in a, q space where ions are
stable. Ions in a trap are stable when the values of another trapping parameter, [5“, are
between 0 and 1 in both the r and 2 directions. 13., is approximately equal to (au + qu2/2)”.
A more detailed discussion of ion motion in an ion trap can be found in the tutorial by
March (28). The trapping parameters, au and qu, are functions of U, V, (o, and (m/z)'1.
Therefore, for given values of U, V, and 0.), only m/z values above a certain threshold are
stable. Ions with m/z values below that threshold will have 13,, values > 1, resulting in
ejection from the ion trap (usually in the z direction). By increasing the amplitude of the
rf potential V, the m/z threshold increases resulting in the ejection of more ions from the
trap. By placing a detector behind a hole in the endcap electrode and gradually raising the
rf potential, it is possible to successively eject and detect ions with a series m/z values,
producing of a mass spectrum.

Tandem mass spectrometry can be accomplished by a technique called resonant

excitation. Equations 1.24 and 1.25 imply a relationship between the frequency of the

18

potential on the electrodes and the stability of ions of a particular m/z. Resonant
excitation uses a small oscillating potential at the end-cap electrodes to excite ions of a
particular m/z value along the z axis. This oscillating potential is in addition to the rf
potential applied to the ring electrode. Ions with m/z values less than that of the desired
precursor ion can be ejected by ramping the rf potential, V, as described above. Ions with
m/z values greater than that of the desired precursor ion are ejected by resonant
excitation. During resonant excitation, ions with selected m/z values (those greater than
the desired precursor) are excited to the extent that they are no longer stable in the ion
trap. When only the precursor ions remain in the trap, they are excited by resonant
excitation. However, this time resonant excitation is used to increase the kinetic energy
of the ions, but not to eject them. The kinetically excited precursor ions collide with the
helium buffer gas (pressure == 1 mTorr) present in the trap and fragment by collisionally-
activated dissociation. The kinetic energy of the precursor ions (and therefore the extent
of fragmentation) can be adjusted by changing the amplitude of the potential used for
resonant excitation. For this reason, the collision energies used in an ion trap are
sometimes expressed as a percentage of the maximum potential that can be applied before
the precursor ions are ejected from the trap. Once fragmentation occurs, the m/z values
of the fragment ions can be determined by scanning the rf voltage as described above.
The CAD-MS/MS spectra obtained on an ion trap may differ from those obtained on
other instruments because the precursor ions spend more time in the collision region and

experience more collisions than the precursor ions analyzed by other techniques.

19

 

 

1). Pi

rein
sequt'
appl:
mils:
Still

den

pl.

IV. Peptide Sequencing by Mass Spectrometry

During the past few decades, mass spectrometry has arisen as a powerful
technique for the analysis of peptides and proteins (30) to now complement chemical
sequencing (31) and gene sequencing (32, 33). Although mass spectrometry had been
applied to the analysis of peptides as early as 1959 (34), analysis by electron ionization
mass spectrometry (El-MS) made only a limited contribution to the area of peptide
sequence determination. For the E1 methods, micromolar quantities of peptides had to be
derivatized prior to analysis by direct probe (35) or gas chromatography mass
spectrometry (GC-MS) (36). The direct probe method allowed sequencing of peptides
with ten or fewer amino acid residues, and the GC-MS method was typically used for
peptides that contain six or fewer residues.

The development of desorption/ionization techniques, e.g., field desorption (37),
plasma desorption (38), and fast atom bombardment (FAB) (I), allowed the analysis of
underivatized peptides by mass spectrometry. FAB and LSIMS gained widespread use as
alternative approaches for peptide analysis. The utility of FAB and LSIMS was
strengthened by the development of tandem mass spectrometry (MS/MS), which allows
fragmentation to be induced by collisionally activated dissociation (CAD), also called
collision-induced dissociation (CID), to produce a variety of structurally significant
fragment ions (39—41). In addition, MS/MS allows fragment ions from the analyte to be
separated from background ions during analysis of the sample, thereby simplifying the
spectra and allowing analyses of simple mixtures.

The introduction of these new ionization techniques and MS/MS revolutionized

the mass spectral analysis of peptides; however two problems remained: lack of

20

 

 

 

senslm'll

 

develop
comm
11111151
increu

desoq

VHF

sensitivity and difficulty in data interpretation. Several types of derivatives were
developed to improve the detection limits of peptide analytes. Some of these derivatives
contained nonpolar functionalities to increase the surface activities of peptides during
analysis by FAB and LSIMS (42-44). Other derivatives generated preformed ions, which
increased ionization efficiencies of many analytes during analysis by

desorption/ionization techniques (45).

V. Peptide Fragmentation

Peptide mass spectra can be complicated due to the many types of fragment ions
that can arise during desorption/ionization and MS/MS (41). The observed peptide
fragments can be amino-terminal fragments (an, bn, cn, and (In ions), carboxy-terminal
fragments (xn, yn, 2“, v,,, and Wu ions), or internal fragments. The extent of charge
localization influences the types of fragment ions produced (46). The effect of charge
localization on fragmentation by high-energy CAD is summarized by Figure 1.3.
Peptides with an acetylated amino-terminus (N-terminus) are protonated on one of the
amide nitrogens, and fragment primarily by charge-driven mechanisms (40) to produce
primarily b., and ya ions. Peptides with basic sites have some degree of charge-
localization, and may produce fragments by a mixture of charge-directed and charge-
remote mechanisms, as indicated by the middle region in Figure 1.3. The extent of
charge localization and charge-directed fragmentation depend on the basicity of the site,
with the arginine (Arg) side-chain being more basic than histidine (His) or lysine (Lys)

side-chains, which in turn are more basic than the free N-terminus of the peptide (47).

21

.n_<o 3.05.59; 3 cos—ScoEmg co cozmufioo. 0920 Lo Echo on... .mé 939....

 

2.2
.5 w a? ...> 93. 22:58”. no «on»... 32
a» a an
._o A >
at w an 259: E58
monsoon. 9< a: o of moan osso mean 8:5
59:00 53:8 2me o: 2me 0... £980
noN=m>tou
omega as as 53:8 as as noosaoa
monsoon oogaoa oogaoa 8558-2
oaasoo A 982

 

5:33.33 «9.25 .0 «Emma

22

 

 

 

Charge-

“ 1111 [hi

peptii
form
‘1?“
great

1110‘

F1

11;
1111:

3m

Charge-derivatized peptides have complete charge-localization and produce fragments
with the charge retained at the site of derivatization.

Charged derivatives were developed to simplify and direct the fragmentation of
peptides to facilitate interpretation of their mass spectra. Charged peptide derivatives
form fragment ions through charge-remote mechanisms (48, 49). As a result, specific
types of fragment ions are observed. Charge-remote fragmentation mechanisms require
greater internal energies than charge-directed fragmentation mechanisms (50), and are
most readily formed by high-energy CAD. The structures of peptide fragment ions and
the mechanisms that produce those ions are the subject of debate (40, 51-59). Figure 1.4
shows the originally proposed structures for the generic fragment ions of underivatized
peptides (40). In this review, we use the nomenclature recommended by Johnson et al.
(54); this is a refinement of the nomenclature originally suggested by Roepstorff and
Fohlman (60). Alternative generic structures for selected peptide fragment ions are
shown in Figure 1.5. The alternative an (56), cn (53), and ya (40) structures were
proposed for fragment ions that formed by charge-remote mechanisms. The all, and yn
ions are proposed to form by charge-driven mechanisms when basic residues (Arg, Lys,
or His) are absent and by charge-remote mechanisms when basic residues are present.
The Cu ions are proposed to form primarily through charge-remote mechanisms and are
related to the presence of certain amino acid residues in the n+1 position (53). The
alternative bn structure (57, 58) was based on ab initio calculations that showed that it
was a more stable structure than that originally proposed for bn ions. The bn+HzO ion
has a mass that is one dalton higher than that of the ca ion, and arises from the loss of one

amino acid residue from the peptide carboxy-terminus (61-63). Because the mass of the

23

mco_ EoEmm: 0033 no 00.3026 nomoaoi .vé 059“.

f saga: a. .5

Egoiwiz 01:n__0
_ . I .1.

 

+z

_:oh301:u_0112a_001:01:z
m

 

+I

Pic

10130151126015?

1 cm

5E00$wi2+o0iznwlf2+
m cm

:oEoolzwlzzvoiolxzoIP
m

 

0:0. 358.36

".5

f saga: o_ as
Iwizifimoiwlzzvz
21.0 m

fznoolzwixzhxglzwlzz:
.m m

+0101:01:2I:i.m001:.401:z*z
a. a

T:
zwflazqaoolxmlszI

am

 

0:0. REE—3.2

24

0:2 EoEmg 02300 no 00.30350 025592 .m... 050E

:ohxoolzwlxzvoolzwflzz
m cm

 

a:

0:0. Esp—:36

 

fzoiszhfiooixwizz v_.:
cm m

 

I:

Iooolzoizhfioiznwizz Y:
_ .._ .. _

 

+1

0
_ wiwizhaooizwlxzrz
cm 2 Em m

f saga: o_ .5

a: T:
. > :wizloimizqz

 

0
=m \ /.m .
+I

 

 

0:0. 358.6%:

”oft...

25

 

 

 

 

bfllhl
SlflJClUl
$111112“)

thngt

cone
dent
addc
hag
itn

den

ha
the
dfi

hr

an

10

de

bn+HzO ion is so close to that of the ca ion, it is frequently misidentified. The bn+HzO
structure shown in Figure 1.4 was proposed by Thome, Ballard, & Gaskell (63), but other
structures have been proposed (55, 62). The bn+HzO ion is proposed to arise from
charge-remote fragmentation (55, 62, 63).

The nomenclature used here for the fragment ions of charge-derivatized peptides
corresponds to the nomenclature of protonated peptides when the mass shift (due to
derivatization) in (singly-charged) precursor ions is taken into account, but an asterisk is
added to denote the presence of the derivative moiety. For example, the N-terminal
fragment ion that results from cleavage of the CH(R.,)-—CO bond is called an an ion when
it arises from a protonated peptide, and is called an *an ion when it arises from a charged
derivative.

The structures of fragment ions from charged derivatives and the associated
fragmentation mechanisms are also debated (56, 64—67). Proposed structures of some of
these fragment ions are shown in Figure 1.6. The two proposed structures for *an ions
differ on which hydrogen atom is lost. The first structure results from the loss of a
hydrogen atom from the B—carbon during analysis by FAB-high—energy CAD-MS/MS (64,
67). The second structure results from the loss of a hydrogen atom from the neighboring
amide nitrogen (65) during analysis by MALDl-PSD. The uppermost proposed structure
for the *bn ion in Figure 1.6 (56) was proposed to arise from charged peptide derivatives
with any amino acid present at residue n. An alternative structure (65) was proposed for
the *bn ion in the special case when aspartic acid is present at residue n; in this case,
derivatized peptides produce abundant *bn ions that correspond to fragmentation on the

C-terminal side of aspartic acid (65, 66). Lin and Glish (68) observed *bn+HzO fragment

26

G) = fixed charge moiety

N-terminal Ions

 

*a : Rn Rn
I‘I R \C":/
G—(NH—éH—CO)? NH—CH
(R' is usually H)
*a o

n' R Rn
G—(NH-éH—CO-hTMéH

*bn:

R '3"
@—(NH—éH—CO)F NH—C=C=O

*bn: (when Asp is residue 11)
O

G-(NH-EH—CO); 109—g

mun-120: 0

R Rn
(ta—(NH—éH—CQFNH—éH-CO—OH

* u
on.

R R.
Q—(NH-éH—CQFNH—CH—CO—NHZ,

*da

nl

R pH—R'
G—(‘NH—(‘SH—CO); NH—CH

(R' is usually H)

C-terminal Ions

 

o=c= NH—éH—CO-(NH—éH—CW

*yn-Z:
R" R
NH=0-CO-(-NH—éH—CO-)F®
”V":
R, R
NH,—CH—co-(NH—cH—co-)F@
*zn-I-1 :
R, R
-éH—CO—(-NH—éH—CO-)F®
*vn:
R
NH=CH—CO-(-NH—(|:H—CO-)-nT®
*wn:
R—cH R
éH—CO—(NH—éH—COfiT-(B
(R' is usually H)

Figure 1.6. Proposed structures of fragment ions
produced from charge-derivatized peptides.

27

ions from a charge-derivatized peptide, but referred to them as bn+OH ions. Because no
structures have been proposed for *bn-i-HzO ions, the bn+HzO structure proposed for
underivatized peptides by Thome, Ballard, & Gaskell (63) was adapted. The *cn and ’"du
structures shown are those proposed by Stults et al. (66). The *xn structure was proposed
by Watson et al. (67). The *yn and *yn-Z structures were proposed by Johnson (64) and
by Watson et al. (67). Charged derivatives fragmented by high-energy CAD produce a
combination of "‘yll and *yn-Z ions; the *yn-Z ions are the more common (67). The *zn
and *z..+1 ions are both formed from charged derivatives, but a structure has only been
proposed for the *z,,+1 ion (56). The structures of the *vn and *wn ions were proposed by
Johnson (64). The structure of the *w., ion was also proposed by Watson et al. (67).

Charge derivatization limits the types of fragment ions generated from peptides.
Derivatization at the N-terminus causes N-terminal charge-remote fragments to be
observed, and derivatization at the carboxy-terrninus (C-terminus) causes C-terminal
fragment ions to be observed. High-energy CAD of N-terminal-charged derivatives
produces primarily "‘an and *dn ions plus a few *bn and *c,. ions (69), whereas C-terminal
charged derivatives produce primarily *yn-2, *yn, "‘VI, and "'WD ions (67, 70). This review
will focus on derivatives that localize a charge on the peptide to control its fragmentation
during analysis by MS or MS/MS.

The vast majority of charged derivatives were developed for use with FAB or
LSIMS desorption/ionization techniques. Most of these papers were published between
the mid-19803 and the mid-19905. Many instruments during this period made use of
high-energy CAD, which can provide the internal energies necessary to induce charge-

remote fragmentation. Derivatization to increase analyte signal intensity was also more

28

important when FAB or LSIMS was used, because the response of many peptides could
be increased by increasing their relative hydrophobicity (42-44).

The use of charged derivatives with plasma desorption (PD) has been quite
limited (71, 72), mainly because PD was never as widely used as FAB or LSIMS.
However, the high internal energies of ions generated by PD produce abundant fragment
ions without the use of CAD-MS/MS techniques.

The analysis of charged derivatives by electrospray ionization (ESI) has received
limited attention (50, 73-77). It is difficult to form charge-remote fragment ions during
analysis by ESI because ions are produced with lower internal energies than those formed
by LSIMS (50), and because the majority of electrospray instruments use low-energy
CAD to form fragment ions for analysis by tandem mass spectrometry. Also, ESI
provides peptide detection limits that are orders of magnitude lower than those obtained
by FAB or LSIMS, so there has been little need to pursue derivatization approaches to
increase signal intensities of the analytes.

Only recently (65, 78-85) have charged derivatives been applied to the analysis of
peptides by matrix-assisted laser desorption/ionization (MALDI). MALDI has detection
limits for peptides and proteins that are several orders of magnitude better than those
obtained with FAB or LSIMS; hence, there was little incentive to develop derivatization
approaches to increase signal intensities of the analytes. Before the development of post-
source decay (PSD) (25), most MALDI instruments were incapable of generating and
observing fragment ions of peptides. Although the spectra of underivatized peptides
obtained by MALDI—PSD have been observed to resemble those obtained by LSIMS low-

energy CAD (26), PSD is an effective technique to form fragment ions of charged

29

derivatives, and now charge derivatization to control fragmentation is an attractive

opnon.

VI. N-Terminal-Charged Derivatives
A. Quaternary Ammonium Derivatives

In 1984, Kidwell and coworkers reported the development of several
derivatization approaches. The first approach (86, 87) to N—terminal derivatization used
methyl iodide to form a trimethylammonium derivative as shown in Figure 1.7.
Because this derivatization approach had low yields, another approach (86, 87) that
involved successive derivatization of the peptide N-terminus with chloroacetyl chloride
followed by reaction with triethylamine to give a triethylammonium derivative was
proposed, as shown in Figure 1.8A. Both of these derivatization approaches will also
derivatize unprotected lysine side-chains. The derivatized peptides were randomly
cleaved with acid or enzymes to generate a series of derivatives that differ by the number
of residues present. The cleavage products were esterified and acetylated to suppress
ionization of the underivatized fragments. These mixtures were analyzed by secondary
ionization mass spectrometry (SIMS), which was the precursor of FAB and LSIMS. The
derivatives had detection limits in the low nanogram (picomole) range; detection limits
were not given for the underivatized peptides. Data were presented only for peptides that
contain three or fewer residues.

Stults, Halualani, & Wetzel presented a selective procedure for attaching a
quaternary ammonium group to the peptide N-terminus (88). Peptide (1-30 nmol) was

reacted with iodoacetic anhydride followed by thiocholine iodide, each at a controlled pH,

30

111,1

Figt

NH

 

 

 

 

c 1 1
MHz-Peptide-COOH —Hi-> :N-PeptldeCOOl-I

Figure 1.7. Methyl iodide derivatization scheme.

R, 0
narrows-coon —LL> Ra—tEI-CHzg-NH-Pepfldo-COOH
R2
. 11 . n
Reagents: 1= Hal-CH2 -R1 1i= Rz-N-R4

A) HaI=CI, R, =c1, R2: R,=R,=Et
13) Ha|=Cl, R1 =01, R,=R,=R,=CH,
c1 Hal = 1, R, = 00(0101121, R, = R3: C11,, R,= 0H,, c,H,,, or c,,H17

Figure 1.8. Trialkylammonium-acetyl derivatization scheme.

0

NHrPepflde-COOH A, Jill-CHZCHZS CHzg-NH-Peptlde-COOH

Reagents: i=(ICH2CO)20, 11= $41—0me

Figure 1.9. Thiocholine iodide derivatization scheme.

0
I O ‘1' I /U\
NHz-PoptHO-COOH ——> -—lil— (CH2)5 NH-PapfldO-COOH

Figure 1.10. "C50" derivatization scheme.

31

as shown in Figure 1.9. The pH control allowed for the selective derivatization of the
N-terminus by exploiting differences between the pKa of the N-terminus and the pKa of
any basic amino acid side-chain. However, protection of free cysteines was required to
prevent derivatization (89). Although the reactions were complete in a few hours, the
products had to be purified by HPLC prior to analysis. Although analysis by FAB or
LSIMS followed by CAD-MS/MS gave typical N-terminal charge-remote fragment ions,
the spectrum was complicated by the loss of trimethylamine from several ions.

Vath and Biemann proposed another procedure for attaching a trimethyl-
ammonium group to the N-terminus of a peptide (90). The procedure involved reaction
of approximately one nmol of peptide with chloroacetyl chloride followed by
trimethylamine, as shown in Figure 1.8B. This two-step procedure involved a total
reaction time of three hours. The signal intensity of the derivative was approximately
one-half that generated from an equal amount of the underivatized peptide during analysis
by FAB-MS. The derivative gave common charge-remote fragmentation products when
analyzed by FAB-CAD-MS/MS.

Stults et al. developed a similar reaction scheme for attaching a
dimethylalkylammonium (DMAA) group at the peptide N-terminus (65). The
derivatization involved reaction of at least 100 pmol of peptide with iodoacetic anhydride
followed by reaction with a dimethylalkylamine, as shown in Figure 1.8C. The
derivatization required a total reaction time of about two hours and gave overall yields of
60-80%. Although the pH control of the iodoacetylation gave 70-90% selectivity for the
N-terminus over the e-amino group of lysine, cysteine residues had to be protected to

prevent derivatization. The presence of a nonvolatile buffer and excess reagents

32

necessitated HPLC cleanup prior to analysis by mass spectrometry. The derivatives were
analyzed by FAB-high-energy CAD-MS/MS or by LSIMS—high-energy CAD-MS/MS.
The dimethylhexylammonium and dimethyloctylammonium derivatives of a hexapeptide
yielded signal intensities that were higher by a factor of 2-5 compared to the
underivatized peptide. The increase in signal intensity is presumably the result of an
increased surface activity; however, such enhancements were not observed during
analyses of larger peptides (a l4-mer and a 19-mer). The trimethylammonium derivative
gave a slight decrease in signal intensity compared to that for the underivatized peptide.
The derivative gives abundant N-terminal charge-remote fragments and a small loss of
the hexyl group from the dimethylalkylammonium functionality.

Stults has applied the N-terminal derivatization procedure to prepare peptide
derivatives with a dimethylhexylammonium group for analysis by E81 (76). The signal
from the derivative was 2-5 times less intense than that from the underivatized peptide
when analyzed by ESI—MS. The doubly-charged ions of the underivatized peptide
[M + 2H]2+ and the derivative [C + H]2+ were used as precursors for analysis by ESI—low-
energy CAD-MS/MS. In contrast to the product-ion spectrum of the underivatized
peptide, the product-ion spectrum of the derivative gave primarily *an and *bn ion peaks;
no C-terminal fragment ion peaks were observed. The presence of abundant *bn ions and
the presence of only one *d., is in contrast to the fragmentation patterns observed by high-
energy CAD (65, 68). The mechanism of fragment ion formation is not clear. The
precursor ion has a charged quaternary ammonium moiety at the N-terminus and a proton
located on the peptide. Fragmentation could occur by charge-remote or charge-directed

mechanisms, or a mixture of the two. Charge-directed mechanisms might by expected to

33

produce doubly-charged fragment ions (when the second charge is retained on the
N-terminal fragment) or abundant C-terminal fragments, such as yn ions (when the
second charge is transferred to the C—terminal fragment); however, neither doubly-
charged fragment ions, nor C-terminal fragment ions were reported. This derivatization
technique was also used by Dongré et al. (49) to prepare the trimethylammonium acetyl
derivative of leucine enkephalin for analysis by electrospray surface-induced dissociation
tandem mass spectrometry (ESI-SlD-MS/MS).

Bartlet-Jones et al. (77) developed another derivatization approach (Figure 1.10)
to attach a trimethylammonium functionality to the peptide N-terminus. This derivative
has been named “CSQ”, presumably because it consists of a five-carbon chain and a
quaternary ammonium group. The CSQ derivatization reaction was complete in 10
minutes and was performed on 20-50 fmol of peptide; 50 fmol of derivatized peptide was
readily observed by MALDI. This derivatization approach was studied in detail in a later
paper (83). MALDI—PSD analysis of the derivatized peptide with a C-terminal arginine
residue generated several yn ions. Modification of Lys and Arg side-chains was
recommended to reduce their basicity. The chemistry used to modify lysine side-chains
also results in cleavage of the N-terminal residue. The modification of lysine,
derivatization of the N-terminus, and modification of arginine required four reaction
steps, more than four hours, 5-10 pmol of peptide, and HPLC purification. When the
peptide was charge-derivatized and basic side-chains were modified, only N-terminal
fragment ions were produced, primarily *an and "‘bn ions. Abundant *bn fragment ions
have not usually been observed for other charge-derivatized peptides analyzed by high-

energy CAD (68) or MALDI-PSD (64, 84). However, *bn ions have been observed in the

34

MALDI-PSD spectra of charge-derivatized peptides when very high laser powers were
used for desorption (64). When Hines et al. analyzed peptides derivatized with the CSQ
reagent (78), several *3" fragment ion peaks and one *bn fragment ion peak were
observed in the MALDI-PSD spectrum; however, the most abundant fragments were
immonium ions and ions that presumably resulted from a combination of backbone

cleavage and the loss of trimethylamine from the derivative ([*a.,-59] and [*bn-59] ions).

B. Quaternary Phosphonium Derivatives

Wagner et al. presented an approach for attaching a triphenylphosphonium group
to either the N-terminus or the C-terminus of a peptide (69). For the N—terminal
derivative, approximately one nmol of peptide was reacted with 2-bromoethyl-triphenyl-
phosphonium bromide at pH 9 for 3 hours at 37°C, as shown in Figure 1.11A. The
derivatization reactions gave yields greater than 75% (66), but with some difficulty in
reproducibility. These derivatives provided charge localization and an enhanced surface
activity relative to the underivatized peptides on the FAB matrix. As a result of
derivatization, the fragmentation was simplified and the signal intensity increased relative
to those obtained from the underivatized peptides during analysis by FAB. The detection
limit of some peptide derivatives was less than 5 picomoles. Both the N-terminal and
C-terminal derivatives produced abundant charge-remote fragments that arose from the
derivatized terminus. However, when a peptide containing a disulfide bond was
derivatized at the N-terminus, the mass spectrum provided less structural information
than the spectrum of the native peptide (91). The derivatization reaction used pH control

to prevent reaction with basic side chains, but it required removal of the buffer salts prior

35

©P—R

MHz-Pepflde-COOH ——> ©iP—CH2CH2-NH-Popudo-coon

A) R = CHZCHzBr
B) R = CH=CH2

Figure 1.11. N-terminal triphenylphosphoniun derivatization scheme.

Br-
11
MHz-Peptide-COOH —> TMPP+-CH2-C—NH-Peplide-COOH

Br" (II) F F
Reagent: TMPP+-CHz-C —S«@—F
F F

OCH3 ,
TMPP+= [c H‘s—H— Q 15"—
OCH3 3

Figure 1.12. TMPP+-Ac derivatization scheme.

Ar-802Cl
NH2-Peptlde-COOH ——> Ar-SOz-NH-Peptldo-COOH
\
AlAr= / . B)Ar= Q

Br

Figure 1.13. Derivatization schemes of Renner and Spiteller.

36

to analysis; this procedure was later modified to eliminate the need for nonvolatile buffers
(92). With the new procedure, the peptide was reacted with vinyltriphenylphosphonium
bromide in a mixture of acetonitrile and pyridine, as shown in Figure 1.11B. Although
this new procedure eliminated the need for sample cleanup, the pH control was uncertain
and allowed the possibility of reaction with basic side-chains.

Bunk and Macfarlane (70) reported the N-terminal derivatization of bradykinin
with vinyltriphenylphosphonium bromide, as shown in Figure 1.11B. The reagents were
combined in a pH 9.0 buffer solution and allowed to react at room temperature overnight.
The derivatives underwent metastable decay to produce N-terminal fragments, with *an
ions dominating, when analyzed by plasma desorption mass spectrometry (PD-MS). A
peak for the protonated derivative [C + H]2+ and a few C-terminal fragments were also
observed. Although an increase in N-terminal fragmentation was observed, the data
obtained by PD-MS analysis of the derivatives gave unsatisfactory results especially
when larger peptides were derivatized and analyzed (71). The triphenylphosphonium
group was readily cleaved through a reverse Michael-type reaction; this cleavage led to a
decrease in sequence-specific fragment ions. Also, the derivative formed primarily
metastable fragments that produced broader peaks than the fragment ions obtained from
underivatized peptides.

Liao and Allison reported that charge derivatization of a hexapeptide led to signal
enhancement when analyzed by MALDI (81). They used the procedure of Wagner et al.
(69) to prepare the N- and C-terminal ethyltriphenylphosphonium derivatives. A one-
pmol sample of the underivatized peptide gave no signal when analyzed by MALDI, but

the N - and C-terminal derivatives both produced strong signals at the one-pmol level.

37

Huang et al. introduced a derivatization procedure that attaches a tris(2,4,6-
trimethoxyphenyl)phosphonium (TMPP+-Ac) group to the peptide N-terminus (80). At
least 10 pmol of peptide were reacted with S-pentafluorothiophenyl [tris(2,4,6-
trimethoxyphenyl)phosphoniumjacetate bromide and 4-(dimethylamino)pyridine
(DMAP) for 15 minutes at room temperature, as shown in Figure 1.12. Although, the
reaction gave yields in excess of 90% for most peptides, a few peptides have yields as
low as 64%. The reaction mixture usually does not require purification prior to analysis
by FAB or MALDI. In addition, peptides composed primarily of amino acid residues
with nonpolar side-chains show significant signal enhancement upon derivatization and
analysis by FAB or MALDI. We have observed recently that lysine, cysteine, and
tyrosine will form a TMPP+-Ac derivative along with TMPP+-Ac derivatization of the
N-terminus when the reaction is performed in an unbuffered system. The spectra of the
derivative obtained by FAB-high-energy CAD-MS/MS show a strong series of charge-
directed fragments. A variety of N-terminal and C-terminal fragment ions were observed
in the MS/MS spectrum of the underivatized peptide. There was no complete series of
ions among the fragment types produced from the underivatized peptide. In contrast, the
N-terminal derivative shows a complete series of readily discemable *an ions. It is also
noteworthy that the signal-to-background ratios of fragment peaks produced from the N-
terrninal derivative were better than the signal-to-background ratios of fragment peaks
produced from the underivatized peptide. When TMPP‘lAc derivatized peptides were
analyzed by LSIMS-CAD-MSn with a quadrupole ion trap instrument, abundant

*bn+HzO ions were produced in addition to a series of *an ions (67).

38

The TMPP+-Ac derivative has also been applied to peptide analysis by
electrospray with an elevated cone voltage to promote in-source fragmentation (ISF) (75).
With the cone set at a typical potential of 37 V, the peptide derivatives did not fragment
significantly, and the derivatives were observed as the doubly-charged derivative [C +
H]2+. When the cone voltage was raised to 100 V, extensive fragmentation occurred and
the doubly-charged derivative was not observed. The fragmentation produced primarily
*an and *bn ions, with some "‘cn and *d,, ions produced. This pattern is similar to the
fragmentation of the dimethylalkylammonium acetyl derivative when analyzed by ESI-
low-energy CAD-MS/MS (76).

The TMPP+-Ac derivative (Figure 1.12) has been used to direct the fragmentation
of peptides during analysis by MALDI-PSD (64, 79, 80). The spectra consist of a strong
series of *an ions plus a few *bn, *cn, and *dn ions. It is interesting to note that the *dn
ion peaks in the MALDI-PSD spectra are less intense than those in the FAB-high-energy
CAD-MS/MS spectra (64). When an aspartic acid residue occurred at residue n,
abundant *bn, *dn, and ”‘an fragment peaks were observed instead of the *an fragment
peak. At very high laser powers, derivatized peptides produced a variety of *bn, *cn, and
*d,, fragments, that were not produced at lower laser powers. The TMPP+-Ac reagent has
been used to derivatize protein digest mixtures at the 25 pmol level (80). This reagent
was also used to derivatize peptides located in an electrophoresis gel or a transfer
membrane (84). The modified peptides were analyzed by MALDI and MALDI-PSD

directly from the gel or membrane on a modified sample plate.

39

VII. N -Terminal Derivatives with High Proton Affinities

In addition to derivatives with a fixed charge, several derivatives with high proton
affinities have been developed. Although the basic groups provide some degree of charge
localization, there is potential for protonation at other sites or for proton transfer to other
sites. These derivatives are analogous to peptides with basic residues (Arg, His, or Lys)
at the N-terminus. As a result, the nomenclature used for these derivatives will be the
same as that used for underivatized peptides.

Renner and Spiteller (93) reported N-terminal derivatization schemes that use
dansyl chloride, as shown in Figure 1.13A, and 2-bromo-5-(dimethylamino)-benzene-
sulfonyl chloride, as shown in Figure 1.13B. Although these derivatives did not have a
fixed charge, the aromatic amine was preferentially protonated during analysis by FAB.
The derivatives had greater signal intensities and an increase in fragmentation (without
CAD) compared to the underivatized peptides. Although the FAB-MS analysis of the
dansyl derivative produced only N-terminal fragments (an and b..), the second derivative
produced a.., b.., and yll ions. The all and bll ions that arise from the second derivative
produced two peaks for each ion, as a result of the 79Br and 8'Br isotopes. These isotope
peaks allowed the N-terminal fragments be distinguished from the C-terminal fragments.
The chemistry employed would also likely derivatize basic side-chains as well as the N-
terrninus of the peptide.

Sherman et al. (94) have used N-hydroxysuccinimide-2-(3-pyridyl)acetate (SPA)
to derivatize the peptide N-terminus, as shown in Figure 1.14, in preparation for analysis
by ESI-MS; this chemistry was optimized by Cardenas et al. (72). Although the

N-pyridylacetyl functionality is not charged, it does have a slightly increased proton

4O

0
@Jt
0"; {0301
NHz-Pepflde-COOH -—->0 NH-Peptlde-COOH

Figure 1.14. N-Hydroxysuccinimide-Z-(3-pyridyl)acetate (SPA)
derivatization scheme.

NH20H2CH2—+N©

MHz-Peptlde-COOH 755’ MHz-Pepflde-CGNHCHzcwiNQ

Figure 1.15. C-terminal pyridinium derivatization scheme.

0

1
MHz-Peptldo-COOH —> NHz-Popflde-CO—N—CIII-NFICHchZCHz—N<

ii 11 l
—> MHz-Peptide-CO—N—C—NHCHchZCl-Iz—TL

Reagents: i= >N—c1-1ch2c1-12N=c=1~1c1-12c113
11=CH,1

Figure 1.16. C-terminal acyl urea derivatization scheme.

1, 11 1
MHz-Pepflde-CO-OCH3 ——> MHz-Peptide-CO-NH—tltIL

Reagents: i = Carboxypeptidase, NHz-N(CH3)2, ii = CH3I

Figure 1.17. Carboxypeptidase derivatization scheme.

41

affinity compared to that of the peptide N-terminus. The derivatives were analyzed by
ESI-MS/MS on a triple quadrupole instrument (93) or by ESI—MS“ on a quadrupole ion
trap instrument (72). The spectra of the derivative contained N-terminal fragment (bn and
an) peaks and C-terminal fragment (yn) peaks. However, the bn fragment ion peaks of the
derivative were more intense than the corresponding bn fragment ion peaks of the
underivatized peptides. The optimized reaction occurs in four minutes, and gives yields
in the range of 80—100% for peptides that do not contain lysine or tyrosine. The reagent
reacts with the side—chain of lysine and reacts to a lesser extent with the side-chain of
tyrosine. Reactions with the side-chains produce derivatives that have multiple
N-pyridylacetyl functionalities. This derivatization has been used on peptides with as
many as 15 residues.

Naven et al. (82) used MALDI-PSD to study peptides derivatized with SPA
(Figure 1.14). The derivatization was performed on-target with as little 50 fmol of
peptide. The procedure was also applied to tryptic peptides electroblotted onto a PVDF
membrane. The reaction was essentially quantitative at the N-terminus and the lysine
side-chains. The PSD spectra contained peaks for a series of abundant bn ions and some

an and yn ions.

VIII. C-Terminal-Charged Derivatives

In addition to the N-terminal derivatives discussed previously, Kidwell and
coworkers developed several C-terminal derivatization approaches (86). Coupling of the
peptide C-terminus to a charged pyridinium salt with 1-(3-dimethylaminopropyl)-3-

ethylcarbodiimide (EDC) was used the generate a derivative with a charged pyridinium

42

functionality, as shown in Figure 1.15. Alternatively, the peptide C-terminus could be
reacted with excess EDC to form a basic acyl urea derivative. This derivative could be
analyzed directly or could be reacted further with methyl iodide to quatemize the terminal
nitrogen, as shown in Figure 1.16. The reagents from Figures 1.15-1.16 will also
derivatize unprotected acidic side-chains. To avoid derivatization of the side-chains, an
enzymatic approach to C-terminal derivatization was also reported, as represented in
Figure 1.17. The peptide methyl ester was formed and reacted with 1,1-dimethyl-
hydrazine, using carboxypeptidase Y to catalyze the reaction. This reaction was followed
by reaction with methyl iodide to form a quaternary ammonium functionality. Like the
N-terminal derivatives discussed above, these derivatized peptides were randomly
cleaved, and the cleavage products were esterified and acetylated to suppress ionization
of the underivatized fragments during analysis by SIMS.

Figure 1.18 shows a selective C-terminal derivatization procedure reported by
Bennett and Day (95). Peptides were reacted with acetic anhydride and N,N-dimethyl-
ethanolamine to generate a C-terminal tertiary amine. The tertiary amine was reacted
with methyl iodide to generate a C-terminal quaternary amine. Although the procedure
was selective, it was very time-consuming. The procedure was only applied to
dipeptides, and required large samples (approximately 100 nmol for the derivatization
procedure and 10 nmol for the analysis by FAB). This chemistry was adapted by Vath,
Zollinger, & Biemann (96) for gas-phase derivatization. The adapted procedure allowed
derivatization and FAB-MS analysis of 360 pmol of a pentapeptide; no signal

enhancement was observed as a result of derivatization.

43

Hirayama et al. used a similar approach to generate the tertiary amine derivative
(97). The procedure was the same as that shown in Figure 1.18, except that the
quatemization with methyl iodide was omitted. The dimethylamine functionality is
reportedly preferentially protonated during analysis by FAB-MS. However, the FAB-
CAD-MS/MS spectrum of the derivative contained peaks for ions that arise from charge
retention at the N-terminus (bn ions) and the C-terminus (yn ions). The precursor-ion
spectrum of m/z 72 (a fragment of N,N-dimethylethanolamine) contained only peaks for
ions that arise from charge retention at the C-terminus. Although sequence information
was generated, this approach was not very effective at localizing the charge, and probably
should be avoided for peptide sequencing.

A C-terminal triphenylphosphonium derivatization procedure was reported by
Wagner et al. (69). The peptide was combined with 2-aminoethyltriphenylphosphonium
bromide and dicyclohexylcarbodiimide (DCC) at pH 5 to form the C-terminal derivative,
as shown in Figure 1.19. Like the N-terminal derivative reported previously, the
derivative increased signal intensity during analysis by FAB-MS and simplified
fragmentation by FAB-CAD-MS/MS. Although the derivatization procedure did not
require removal of buffer, it had the potential to derivatize acidic side chains. However,
reaction products from acidic side chains were not observed. The MS/MS spectra of C-
terminal derivatives contained only C-terminal fragments. Using the same derivatization
approach, Liao and Allison (81) observed an increase in signal intensity for a peptide

derivative analyzed by MALDI.

NHfieptlde-COOH —"i'—> CHgtl'i—NH-Pepflde-CO—OCHchz—N<

CHal ll l+
———> CH3C—NH-Poptlde-CO—OCH2CH2—tlil—

Reagents: l '-'-' A620, ll = HOCHZCHZN(CH3)2

Figure 1.18. C-terminal N, N-dimethylethanolamine derivatization scheme.

NHZCHZCHZECg—Q
+
MHz-Pepflde-CCDH TD? MHz-Popflde-CO- NHCHZCHz—Pr-Q

Figure 1.19. C-terminal triphenylphosphonium derivatization scheme.

4.
(CH2).— “”2 (39‘ (CHzlti'Q
___, |

NHz-Pepflde-COOH NHz-Peptlde-COOH

Figure 1.20. 2,4,6-Trimethylpyridinitm derivatization
scheme for lysine side-chains.

45

IX. Side-Chain Derivatives

Johnson (39, 63) reported the conversion of a lysine side-chain to a charged 2,4,6-
trimethyl pyridinium group, as shown in Figure 1.20. Peptide (20 nmol) was reacted with
2,4,6-trimethylpyrylium tetrafluoroborate for 18-24 hours at room temperature. Excess
reagent and buffer salts were removed by solid phase extraction and HPLC. No
underivatized peptide was observed during analysis by FAB-MS. Upon analysis by FAB-
CAD-MS/MS, primarily charge-remote fragments were detected. If the derivatized lysine
was at the N-terminus, then N—terminal fragments were formed. If the derivatized lysine
was at the C-terminus, then C-terminal fragments were formed. If an internal lysine was
derivatized, then a mixture of N- and C-terminal fragments was formed.

In response to the unsatisfactory results obtained from plasma desorption analysis
of triphenylphosphonium derivatives (70), Bunk and Macfarlane developed another
derivatization procedure (71); the conversion of lysine residues to homoarginine used
O-methylisourea under basic conditions, as shown in Figure 1.21. The reaction mixture
was purified by HPLC prior to analysis. The homoarginine residue does not possess a
fixed charge, but has a high proton affinity similar to that for arginine. This derivative
localizes the charge on the homoarginine residue. During analysis by plasma desorption,
sequence-specific fragmentation was increased for some derivatives. This derivative also
provides a means for distinguishing between lysine and glutamine residues. The most
obvious limitation of this approach is that it is only useful for peptides that contain lysine.
The site of derivatization is not known unless the location of the lysine residue is known.
Also, there is potential for the charge to be spread over several sites if either an arginine

or additional lysine residues are present in the peptide.

46

NHz
NH:
(CH,),— NH.2 0130an (0H,),— NH’gNH
NHz-Pe fldo—COOH ——> MHz-PerLtIdO-COOH

Figure 1.21. Conversion of lysine side-chains to homoarginine.

‘1?
FQC—F
F F 0

F F ll
NHl-PepfldO-COOH W F@ C — NH-PODMG—COOH
F F

Figure 1.22. Pentafluorobenzoyl fluoride derivatization scheme.

Ra‘s" o
R-NH-PeptIde-COOH W R-NH-Peptide-CO—NH 803'
R = H orAc

Figure 1.23. C-terminal 4-aminonaphthalenesulphonic acid (ANSA) derivatization.

47

X. Negatively-Charged Derivatives

Kiplinger et al. (98) used pentafluorobenzoyl fluoride to derivatize peptides at the
N-terminus, as shown in Figure 1.22. The derivatization was useful for analyses by
LSIMS in the positive and negative modes. In negative LSIMS, the pentafluorobenzoyl
group was preferentially ionized to produce a species with a negative charge at the N-
terminus. The derivative produced a peak that was 2-40 times more intense than the peak
produced by the underivatized peptide. However, in the negative mode no significant
fragmentation was observed. The response of the derivative to negative LSIMS-CAD-
MS/MS was not examined. In positive LSIMS, no significant enhancement of the signal
intensity was observed, but extensive fragmentation was produced without the use of
MS/MS. However, in the positive mode, the location of the charge was not fixed, and N-
terrninal and C-terminal fragments were both observed. The derivatization reaction
occurred rapidly, and the derivative could be analyzed without any sample cleanup.
However, the reaction was performed on a large scale (approximately 100 nmol), and the
reagent reacted with the side-chains of lysine and tyrosine.

Lindh et al. reported a C-terminal derivatization approach for use with negative
LSIMS (99). The peptide or acetylated peptide was reacted with 4-amino-naphthalene-
sulphonic acid (ANSA) and a carbodiimide coupling agent (EDC) to form the negatively-
charged derivative, as shown in Figure 1.23. Three small peptides with 3-5 residues were
derivatized and analyzed. The reaction was performed on 10-100 nmol of peptide at
25°C for 2 hours to produce yields of 10%-100%, depending on the peptide. After
purification by HPLC, the derivatives were analyzed by negative LSIMS. When the

peptides were acetylated and derivatized with ANSA, the detection limits obtained upon

48

o“

 

analysis by negative LSIMS were a factor of 10-40 times improved compared to those of
the underivatized peptide. Analysis by negative LSIMS-CAD-MS/MS produced a variety
of C-terminal charge-remote fragments and a few N-terminal fragments. Derivatization
with ANSA can occur at the side—chains of aspartic acid and glutamic acid residues in
addition to derivatization at the C-terminus. Other side reactions, such as internal
cyclization, have been observed during derivatization.

Griffiths et al. analyzed ANSA-derivatized peptides using ESI—MS, ESI-ISF-MS,
and ESI—high-energy CAD-MS/MS on a sector instrument equipped with an array
detector (73). Detection limits in the range of one frnol were obtained from the
derivatized peptides during analysis by negative ESI-MS, a factor of 5-50 better than the
detection limits of the underivatized peptides. Negative ESI-ISF-MS data were obtained
from 10-30 pmol of derivatized peptide, and negative ESI—high-energy CAD-MS/MS data
were obtained from as little as 50 fmol of derivatized peptide. Derivatized peptides
analyzed by ESI-ISF-MS and by ESI-high-energy CID produced many of the same
fragment ions; however, some unique fragment ions were generated by each
fragmentation method. Both fragmentation methods produce multiple types of fragments,
complicating interpretation.

In a related study, ANSA-derivatized peptides were analyzed by negative ESI-
low-energy CAD-MS/MS on a hybrid magnetic sector-orthogonal time-of-flight
instrument (74). MS/MS data could be obtained from a few picomoles of derivatized
peptide. Primarily C-terminal charge-remote fragments were observed; however, a few c“
ions were observed. The appearance of the low-energy CAD-MS/MS spectra were

dependant on the mass of the collision gas used. When helium was used as the collision

49

gas, the fragments generated were primarily the result of backbone cleavage. When xenon
was used as the collision gas, fragments from side-chain cleavage were observed in
addition to those from backbone cleavage. Many of the fragment ion peaks in the ESI-
low-energy-CAD-MS/MS spectra were the same as those in the ESI—ISF—MS and ESI-
high-energy CAD-MS/MS spectra (73); however, the spectra produced by each of these

methods were different.

XI. Concluding Commentary

Theideal charged derivatization approach would have several attributes. The
chemistry should occur rapidly with high product yield without any unwanted
derivatization of peptide side-chains. The derivatives should not require purification
prior to analysis. The derivative should be compatible with all mass spectrometry
ionization and fragmentation techniques. The ideal derivative would have a better
detection limit than the underivatized peptide. The derivative should have a fixed charge,
and produce a series of charge-remote fragments during CAD. The added chemical
derivative group should not fragment during analysis by tandem mass spectrometry
because that fragmentation might complicate the mass spectrum and reduce the intensity
of the other fragment ion peaks.

Three of the derivatization approaches discussed in this review come closer to the
ideal than the others. Those derivatization approaches produce the dimethylalkyl-
ammonium acetyl (DMAA) derivatives reported by Stults et al. (65) (Figure 1.8C), the
CSQ derivatives of Spengler et al. (83) (Figure 1.10), and the [tris(2,4,6-trimethoxy-

phenyl)-phosphonium] acetate (TMPP+-Ac) derivatives reported by Huang et al. (80)

50

(Figure 1.12). These approaches can be used to attach a fixed charge to the N-terminus of
the peptide in high yields.

The DMAA derivatization is performed with commercially available reagents.
Although the C5Q and TMPP+-Ac derivatization reagents must be synthesized, the
TMPP+-Ac reagent, and possibly the C5Q reagent, are stable for months. The TMPPI-Ac
derivatization occurs in a fifteen-minute step, the DMAA derivatization occurs in two
steps that require a total of two hours, and the CSQ derivatization requires between
twenty minutes (without side-chain derivatization) and four hours (with derivatization of
Lys and Arg). The amount of peptide required for the derivatization reactions has been
reported as 100 pmol for DMAA and the low pmol range for the TMPP+-Ac and C5Q
derivatizations. However, the CSQ derivatization procedure can be performed on sub-
picmole quantities of peptide, if neither Lys nor Arg needs derivatization. All three
reactions occur with high yields and are selective for the N-terminus when the pH of the
reaction mixture is controlled. The DMAA and C5Q derivatizations usually require
purification by HPLC to remove excess reagent. Although such purification usually is
not required for the TMPP+-Ac derivatization when the reaction mixture is analyzed by
FAB or MALDI, it is required in analyses by ESI.

The TMPP+-Ac derivative has been analyzed successfully with FAB, E81, and
MALDI techniques. Although the DMAA derivatives have only been analyzed by FAB
and E81, they are probably compatible with MALDI. Likewise, although the CSQ reagent
has only been analyzed by MALDI, it should be compatible with FAB and E81.

DMAA derivatives give a small increase in signal intensity in some conditions

and a small decrease under others. No loss of signal intensity was observed for CSQ

51

derivatization; however, the relative signal intensities of underivatized and derivatized
peptides were not reported. For most peptides, the TMPP+-Ac derivative has only a small
effect on signal intensity; however, some hydrophobic peptides have shown a significant
signal enhancement upon derivatization. All of the derivatives have detection limits in
the low picomole to sub-picomole range.

Because the derivative moieties have fixed charges, the fragment ions produced
from these derivatives are proposed to form by charge-remote mechanisms. FAB-high-
energy CAD-MS/MS analysis of the DMAA and TMPP+-Ac derivatives produces
primarily "‘an and "‘dn plus a few *bn and *cn ions. In addition, the DMAA derivative
loses the alkyl group to generate a peak that can complicate the high-mass end of the
spectrum. When DMAA (76) and TMPP+-Ac (75) derivatives are analyzed by ESI,
primarily "‘an and "‘bn ions are generated plus a few *cn and *dn ions. Because the
precursor ions have a double-charge, it is not clear whether the fragment ions arise by
charge-remote fragmentation, charge-directed fragmentation, or a combination of the two.

When TMPP+-Ac derivatives are analyzed by MALDI-PSD (64, 79, 84), a
dominant series of *an ions is produced and a few *bn, *cn and "'dn ions are produced.
Peptides derivatized with CSQ (including modification of Lys or Arg) are reported to
form primarily *an and *bn ions, however "‘cn and "'dn ions also form to a lesser extent.
In a preliminary study by Hines et al. (78), a CSQ derivative was reported to form
primarily *an-59 and *bn-59 ions. The differences in these fragmentation patterns are
difficult to explain. Perhaps the laser power (or some other experimental parameter)
causes the differences in fragmentation, as has been documented in one case of

fragmentation of charged derivatives during analysis by MALDI-PSD (64). The CSQ

52

derivative moiety has a smaller head group than the TMPP+-Ac moiety, and the CSQ
moiety has a flexible five-carbon chain. These differences might allow the formation of
secondary structures in which the charge can interact with the peptide backbone to
influence fragmentation behavior. All three of these derivatization techniques show
promise; however they have yet to be proven in applications of peptide sequencing by

mass spectrometry of biological samples.

XII. Summary

Major advantages of charged derivatives are that they simplify the fragmentation
of peptides and direct the fragmentation to one end of the peptide. Simplification of the
mass spectrum is particularly helpful for interpreting data from peptides with a
completely unknown sequence. Derivatives with a fixed charge are the more effective in
directing fragmentation than derivatives with elevated proton affinities. N-terminal
derivatization can be made selective by exploiting the differences in pKa of the N-
terminus and e-NHz group on lysine; however, the pKa values for the C-terminus and the
acidic side chains are too similar to allow selective derivatization. Because N-terminal
charged derivatives produce fewer types of fragment ions than C-terminal derivatives, the
spectra of N-terminal derivatives are easier to interpret than those of the C—tenninal
derivatives.
Charged derivatives of peptides are especially useful in analyses based on FAB and
LSIMS, because some derivatives can provide signal enhancement and simplify
fragmentation. The utility of MALDI-PSD for the analysis of charged derivatives has

been demonstrated in recent years, although different derivatives appear to produce

53

different fragmentation patterns. The analysis of charged derivatives by ESI has been the
subject of preliminary studies; however, a careful and extensive study has yet to be
published. In addition, it is not clear whether these derivatives fragment by charge-

remote mechanisms, charge-driven mechanisms, or some combination of the two.

54

XIII. References

l.

10.

ll.

12.

13.

Barber, M.; Bordoli, R. S.; Sedgwick, R. D.; Tyler, A. N. J. Chem. Soc., Chem.
Commun. 1981, 325-327.

Rouse, J. C.; Allison, J. J. Am. Soc. Mass Spectrom. 1993, 4, 259-269.

Sunner, J.; Marales, A.; Kebarle, P. Anal. Chem. 1987, 59, 1378-1383.

Stoll, R. G.; Harian, D. J.; Hass, J. R. Int. J. Mass Spectrom. Ion Processes 1984,
61, 71-79.

Aberth, W.; Straub, K. M.; Burlingame, A. L. Anal. Chem. 1982, 54, 2029-2034.
Standing, K. G.; Chait, B. T.; Ens, W.; McIntosh, G.; Beavis, R. Nucl. Instrum.
Methods 1982, 198, 33-38.

Takayama, M. J. Am. Soc. Mass Spectrom. 1995, 6, 114-119.

Posthumus, M. A.; Kistemaker, P. G.; Meuzelaar, H. L. C.; Ten Noever de Brauw,
M. C. Anal. Chem. 1978, 50, 985-991.

Karas, M.; Bachmann, D. B.; lBahr, U.; Hillenkamp, F. Int. J. Mass Spectrom. Ion
Processes 1987, 78, 53-68.

Juhasz, P.; Costello, C. E.; Biemann, K. J. Am. Soc. Mass Spectrom. 1993, 4, 399-
409.

Vertes, A.; Gijbels, R.; Levine, R. D. Rapid Commun. Mass Spectrom. 1990, 4,
228-233.

Liao, P.-C.; Allison, J. J. Mass Spectrom. 1995, 30, 408-423.

Whitehouse, C. M.; Dreyer, R. N.; Yamashita, M.; Fenn, J. B. Anal. Chem. 1985,

57, 675-679.

55

14.

15.

l6.

l7.

l8.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

Dole, M.; Mack, L. L.; Hines, R. L.; Mobley, R. C.; Ferguson, L. D.; Alice, M. B. J.
Chem. Phys. 1968, 49, 2240-2249.

Wilm, M. S.; Mann, M. Int. J. Mass Spectrom. Ion Processes 1994, 136, 167-180.
Schmelzeisen-Redecker, G.; Buttering, L.; Rollgen, F. W. Int. J. Mass Spectrom. Ion
Processes 1989, 90, 139-150.

Iribame, J. V.; Thompson, E. A. J. Chem. Phys. 1976, 64, 2287-2294.

Thompson, B. A.; Iribame, J. V. J. Chem. Phys. 1979, 71, 4451-4463.

Fenn, J. 8.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Mass
Spectrom. Rev. 1990, 9, 37-70.

Gaskell, S. J. J. Mass Spectrom. 1997, 32, 677-688.

Kebarle, P.; Tang, L. Anal. Chem. 1993, 65, 972A-986A.

Smith, R. D.; Loo, J. A.; Loo, O.; Busman, M.; Udseth, H. R. Mass Spectrom. Rev.
1991, 10, 359-451.

Mamyrin, B. A.; Karataev, V. I.; Shmikk, D. V.; Zagulin, V. A. Sov. Phys. JEPT
1973, 37, 45-48.

Wiley, W. C. and McLaren, I. H. Rev. Sci. Instr. 1955, 26, 1 150-1157.

Spengler, B.; Kirsch, D.; Kaufmann, R.; J aeger, E. Rapid. Commun. Mass Spectrom.
1992, 6, 105-108.

Rouse, J. C.; Yu, W.; Martin, S. A. J. Am. Soc. Mass Spectrom. 1995, 6, 822-835.
Miller, P. E.; Denton, M. B. J. Chem. Ed. 1986, 63, 617-622.

Mathieu, E. J. Math. Pure Appl. 1868, 13, 137-203.

March, R. E. J. Mass Spectrom. 1997, 32, 351-369.

56

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

Yates, J. R. J. Mass Spectrom. 1998, 33, 1-19.

Edman, P. Acta Chem. Scand. 1950, 4, 283-293.

Maxam, A. M.; Gilbert, W. Proc. Natl. Acad. Sci. USA. 1977, 74, 560-564.
Sanger, F.; Nicklen, S.; Coulson, A. R. Proc. Natl. Acad. Sci. USA. 1977, 74, 5463-
5467.

Biemann, K.; Gapp, F.; Siebl, J. J. Am. Chem. Soc. 1959, 81, 2274-2275.

Vilkas, E.; Lederer, E. Tetrahedron Lett. 1968, 26, 3089-3092.

Carr, S. A.; Herlihy, W. C.; Biemann, K. Biomed. Mass. Spectrom. 1981, 8, 51-61.
Beckey, H. D. Principles of Field Ionization and Field Desorption Mass
Spectrometry. Pergamon, Oxford, 1977.

Macfarlane, R. D. Anal. Chem. 1983, 55, 1247A-1264A.

Biemann, K.; Scoble, H. A. Science 1987, 237, 992-998.

Johnson, R. S.; Martin, S. A; Biemann, K. Int. J. Mass Spectrom. Ion Processes.
1988, 86, 137-154.

Papayannopoulos, I. A. Mass Spectrom. Rev. 1995, 14, 49-73.

Falick, A. M.; Maltby, D. A. Anal. Biochem. 1989, 182, 165-169.

Ligon, W. V. Anal. Chem. 1986, 58, 485-487.

Naylor, S.; Findeis, A. F., Gibson, B. W.; Williams, D. H. J. Am. Chem. Soc. 1986,
108, 6359-6363.

Busch, K. L.; Unger, S. E.; Vincze, A.; Cooks, R. G.; Keough, T. J. Am. Chem. Soc.
1982, 104, 1507-1511.

Renner, D.; Spiteller, G. Biomed. Environ. Mass Spectrom. 1986, 13, 405-410.

57

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

Harrison, A. G. Mass Spectrom. Rev. 1997, 16, 201—217.

Jensen, N. J .; Tomer, K. 8.; Gross, M. L. J. Am. Chem. Soc. 1985, 107, 1863-1868.
Gross, M. L. Int. J. Mass Spectrom. Ion Processes 1992, I 18/119, 137-165.

Dongré, A. R.; Jones, J. L.; Somogyi, A; Wysocki, V. H. J. Am. Chem. Soc. 1996,
118, 8365-8374.

Biemann, K.; Martin, S. A. Mass Spectrom. Rev. 1987, 6, 1-76.

Cordero, M. M.; Houser, J. J.; Wesdemiotis, C. Anal. Chem. 1993, 65, 1594-1601.
Downard, K. M.; Biemann, K. J. Am. Soc. Mass Spectrom. 1993, 4, 874-881.
Johnson, R. S.; Martin, S. A.; Biemann, K.; Stults, J. T.; Watson, J. T. Anal. Chem.
1987, 59, 2621-2625.

Rouse, J. C. Ph.D. Dissertation, Michigan State University, East Lansing, MI,
1993.

Wagner, D. S. Ph.D. Dissertation, Michigan State University, East Lansing, MI,
1992.

Yalcin, T.; Khouw, C.; Csizmadia, I. G.; Peterson, M. R.; Harrison, A. G. J. Am.
Soc. Mass Spectrom. 1995, 6, 1165-1174.

Yalcin, T.; Csizmadia, I. G.; Peterson, M. R.; Harrison, A. G. J. Am. Soc. Mass
Spectrom. 1996, 7, 233-242.

Yu, W.; Vath, J. E.; Huberty, M. C.; Martin, S. A. Anal. Chem. 1993, 65, 3015-
3023.

Roepstorff, P.; Fohlman, J. Biomed. Mass Spectrom. 1984, I I, 601.

58

61.

62.

63.

65.

66.

67.

68.

69.

70.

71.

72.

Deery, M. J.; Summerfield, S. G.; Buzy, A.; Jennings, K. R. J. Am. Soc. Mass
Spectrom. 1997, 8, 253-261.

Gonzalez, J .; Besada, V.; Garay, H.; Reyes, 0.; Padron, G.; Tambara, Y.; Takao, T.;
Shimonishi, Y. J. Mass Spectrom. 1996, 31, 150-158.

Thome, G. C.; Ballard, K. D.; Gaskell, S. J. J. Am. Soc. Mass Spectrom. 1990, I,
249—257.

Johnson, R. S. Ph.D. Dissertation, Massachusetts Institute of Technology,
Cambridge, MA, 1988.

Liao, P.-C.; Huang, Z.-H.; Allison, J. J. Am. Soc. Mass Spectrom. 1997, 8, 501-509.
Stults, J. T.; Lai, J.; McCune, S.; Wetzel, R. Anal. Chem. 1993, 65, 1703-1708.
Watson, J. T.; Wagner, D. S.; Chang, Y. S.; Strahler, J. R.; Hanash, S. M.; Gage, D.
A. Int. J. Mass Spectrom. Ion Processes. 1991, 111, 191-209.

Lin, T.; Glish, G. L. Proceedings of the 45'” ASMS Conference on Mass
Spectrometry and Allied Topics, Palm Springs, California, June 1-5, 1997.

Zaia, J.; Biemann, K. J. Am. Soc. Mass Spectrom. 1995, 6, 428-436.

Wagner, D. S.; Salari, A.; Gage, D. A.; Leykam, J.; Fetter, J.; Hollingsworth, R.;
Watson, J. T. Biol. Mass Spectrom. 1991, 20, 419-425.

Bunk, D. M.; Macfarlane, R. D. Int. J. Mass Spectrom. Ion Processes. 1991, 111,
55-75.

Bunk, D. M.; Macfarlane, R. D. Int. J. Mass Spectrom. Ion Processes. 1993, 126,

123-136.

59

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

Cardenas, M. S.; van der Heeft, E.; de Jong, A. P. J. M. Rapid. Commun. Mass
Spectrom. 1997, II, 1271-1278.

Griffiths, W. J .; Lindh, I.; Bergman, T.; Sjovall, J. Rapid Commun. Mass Spectrom.
1995, 9, 667-676.

Lindh, 1.; Griffiths, W. J.; Bergman, T.; Sjovall, J. Int. J. Mass Spectrom. Ion
Processes 1997, I 64, 71-79.

Sadagopan, N.; Wu, J .; Yang, Y.; Huang, Z.-H.; Watson, J. T. Proceedings of the
45m ASMS Conference on Mass Spectrometry and Allied Topics, Palm Springs,
California, June 1-5, 1997.

Stults, J. T. Proceedings of the 40m ASMS Conference on Mass Spectrometry and
Allied Topics, Washington, DC, May 31-June 5, 1992.

Bartlet-Jones, M.; Jeffery, W. A.; Hansen, H. F.; Pappin, D. J. C. Rapid Commun.
Mass Spectrom. 1994, 8, 737-742.

Hines, W.; Peltier, J.; Hsieh, F.; Martin, S.A. Proceedings of the 43” ASMS
Conference on Mass Spectrometry and Allied Topics, Atlanta, Georgia, May 21-26,
1995.

Huang, Z.-H.; Wu, J.; Gage, D. A.; Watson, J. T. Proceedings of the 45‘” ASMS
Conference on Mass Spectrometry and Allied Topics, Palm Springs, California, June
1-5, 1997.

Huang, Z.-H.; Wu, J.; Roth, K. D. W.; Yang, Y.; Gage, D. A.; Watson, J. T. Anal.
Chem. 1997, 69, 137-144.

Liao, P.-C.; Allison, J. J. Mass Spectrom. 1995, 30, 511-512.

60

83.

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

Naven, T. J. P.; Jeffery, W. A.; Bartlet-Jones, M.; Rahman, D.; Pappin, D. J. C.
Proceedings of the 45m ASMS Conference on Mass Spectrometry and Allied Topics,
Palm Springs, California, June 1-5, 1997.

Spengler, B.; Luetzenkirchen, F.; Metzger, S.; Chaurand, P.; Kaufmann, R.; Jeffery,
W.; Bartlet-Jones, M.; Pappin, D. J. C. Int. J. Mass Spectrom. Ion Processes. 1997,
169/170, 127-140.

Strahler, J. R.; Smelyanskiy, Y.; Lavine, G.; Allison, J. Int. J. Mass Spectrom. Ion
Processes 1997, 169/170, 111-126.

Kidwell, D. A.; Ross, M. M.; Colton, R. J. J. Am. Chem. Soc. 1984, 106, 2219-2220.
Kidwell, D. A.; Ross, M. M.; Colton, R. J. Springer Ser. Chem. Phys: SIMS 4,
1984, 36, 412-414.

Stults, J. T.; Halualani, R.; Wetzel, R. Proceedings of the 3 7m ASMS Conference on
Mass Spectrometry and Allied Topics, Miami Beach, Florida, May 21-26, 1989.
Wetzel, R.; Halualani, R.; Stults, J. T.; Quan, C. Bioconjugate Chem. 1990, I, 114-
122.

Vath, J. E.; Biemann, K. Int. J. Mass Spectrom. Ion Processes. 1990, 100, 287-299.
Chang, Y. S.; Gage, D. A.; Watson, J. T. Biol. Mass Spectrom. 1993, 22, 176-180.
Wagner, D. S.; Nieuwenhuis, T. J .; Chang, Y. S.; Gage, D. A.; Watson, J. T.
Proceedings of the 40112 ASMS Conference on Mass Spectrometry and Allied Topics,
Washington, DC, May 31-June 5, 1992.

Renner, D.; Spiteller, G. Angew. Chem. Int. Ed. Engl. 1985, 24, 408-409.

61

94.

95.

96.

97.

98.

99.

Sherman, N. E.; Yates, N. A.; Shabanowitz, J.; Hunt, D. F.; Jeffery, W.; Bartlet-
Jones, M.; Pappin, D. J. C. Proceedings of the 43rd ASMS Conference on Mass
Spectrometry and Allied Topics, Atlanta, Georgia, May 21-26, 1995.

Bennett, B. D.; Day, R. A. Proceedings of the 35m ASMS Conference on Mass
Spectrometry and Allied Topics, Denver, Colorado, May 24-29, 1987.

Vath, J. E.; Zollinger, M.; Biemann, K. Fresenius Z. Anal. Chem. 1988, 331, 248-
252.

Hirayama, K.; Akashi, S.; Yuji, R.; Niitsu, U.; Fujimoto, Y. Organic Mass
Spectrom. 1993, 28, 1516-1524.

Kiplinger, J. P.; Contillo, L.; Hendrick, W. L.; Grodski, A. Rapid Commun. Mass
Spectrom. 1992, 6, 747-752.

Lindh, I.; Griffiths, W. J .; Bergman, T.; Sjovall, J. Rapid Commun. Mass Spectrom.

1994, 8, 797-803.

62

CHAPTER 2
STUDIES OF TRIS(2,4,6-TRIMETHOXYPHENYL)PHOSPHONIUM-ACETYL
(TMPP+-AC) DERIVATIZATION FOR PEPTIDE SEQUENCING BY

MASS SPECTROMETRY

I. Introduction

The tris(2,4,6—trimethoxyphenyl)phosphonium (TMPPI-Ac) reagents were
developed to overcome some of the limitations of the triphenylphosphonium (TPP)
reagents. The structures of the TMPP+-Ac reagents are shown in Figure 2.13. The
TMPP head group is more polar than the TPP head group. As a result, the TMPP+-Ac
reagents have a greater water solubility than the TPP reagents. This allows reactions to
be performed in aqueous solutions. Generally, peptides have much higher solubility in
water than in the nonaqueous solvents sometimes used for TPP derivatization reactions.
Derivatization in aqueous solution also allows control of the reaction pH to provide more
selective reactions. The general fragmentation characteristics of TMPP+-Ac-derivatized
peptides are described in this chapter. Also, this chapter explores the reaction yields and
selectivity of the TMPP+-Ac reagents. Finally, the ionization efficiency and signal

suppression effects of TMPP+-Ac derivatives are discussed.

11. Experimental
A. Preparation of TMPP+-Ac-Succinimide Bromide ([TMPP+-Ac-OSu][Br'])
The TMPP+-Ac-succinimide reagent was prepared using a two-step synthesis as

shown in Figure 2.1. Preparation of the bromacetic acid ester is shown in Figure 2.1A.

63

O
. 11
A) Br-CHz-C-Br + H-Y
O

i Y = S'CGF5 OI‘ O-N

C1’
Br—CHz-C-Y + H-Br

[_ __

OCH 0

3
11
B) CH3O@——P + Br-CHz-C-Y

_ OCng 3

 

 

OCH3 Br- c")
CH3O‘— —P+—CH2C-Y
OCH3

— —3

 

 

Figure 2.1. Synthesis of TMPP+-Ac reagents.

64

N-hydroxysuccinimide (3.0 g, 26 mmol) was dissolved in 20 mL warm benzene.
Bromoacetyl bromide (2.5 mL, 29 mmol) was added dropwise, and the mixture was
refluxed 20 hours in a sand bath at 90 °C with a drying tube attached to exclude moisture.
Upon cooling, a white ppt formed. The reaction mixture was filtered and the ppt was
recrystallized from benzene. The yield of the bromoacetic acid succinimide ester
intermediate was 0.71 g (3.0 mmol, 12%).

Reaction of the intermediate with TMPP to form the TMPP+-Ac reagent is shown
in Figure 2.1B. The bromoacetic acid succinimide ester intermediate (1.0 g, 4.4 mmol)
was dissolved in 10 mL benzene. A solution of tris(2,4,6-trimethoxyphenyl)phosphine
(2.0 g, 3.8 mmol) in 25 mL benzene was added dropwise with stirring. A fine white
precipitate formed immediately. A 35 mL aliquot of hexane was added to force
remaining product out of solution. The supernatant was removed, and the precipitate was
washed twice with 40 mL benzene/hexane (1:3, vzv) and once with 30 mL hexane. After
drying overnight in a vacuum dessicator over phosphorous pentoxide and NaOH pellets,
the pure product (1.0 g, 34%) was obtained as a fine, white powder. The product was

analyzed by positive FAB-MS as shown in Figure 2.2A.

B. Preparation of TMPP+-Ac-S-Pentafluorophenyl Bromide ([TMPP+-Ac-SPf][Br'])
TMPP+-Ac—S-pentafluorophenyl ([TMPP+-Ac-SPf][Br']) was prepared by a two-
step synthesis analogous to that described above (1). The FAB-MS spectrum of the

TMPP+-Ac-SPf reagent is shown in Figure 2.2B.

65

 

 

 

 

 

 

 

 

 

 

 

 

[TMPP‘LAc-OSuHBr']
a};
c -
s
s
0
.2
E .
CD
a: ~1-
, *1” [TMPP-I-Hl“ TMPP :f-OSu
533
* * 623 688
1 -- -1 \ 321.1._l_
m/z '
_ 181
B TMPP+-Ac-8Pf
c-I-
- 773
g- [TMPP*-Ac-SPI‘][Br']
2 _
.9
5
<1)
g _ ['l'MPP-i-Hl"
% 533
‘r 167
. 121 \ 197 559 591
365 /57.'y 731
‘ 35Q Iii/81 (I 633 l 1
mlz l I r

Figure 2.2. FAB-MS spectra of TMPP+-Ac-OSu (A) and TMPP*-Ac-SPf (B).

66

C. TMPP+-Ac Derivatization of Peptides in Unbuffered Solution

The TMPP+-Ac derivatization scheme is shown in Figure 2.3. Peptide derivatives
were prepared using the following procedure. 2 nmol of peptide (in 2 [IL 10%
acetonitrile), 2 1.1L of 10 mM 4-dimethylaminopyridine in 50% acetonitrile (v/v), and 1
11L of 10 mM [TMPP+-Ac-succinimide][Br'] in 50% acetonitrile were combined,
vortexed for 1 minute, and allowed to stand at room temperature for 30 minutes. For
analysis by FAB or MALDI, the reaction mixture was analyzed without any purification.
Prior to analysis by ESI, the derivatized peptide was purified by HPLC as described

below.

D. TMPP+-Ac Derivatization of Peptides in a Buffered Solution

Peptide derivatives were prepared using the following procedure. Peptide (0.5-10
nmol in 10 11L 10% acetonitrile (v/v)), 2 1.1L 50 mM TRIZMA (Tris-HCl) buffer (pH 8.3),
and 2 11L of 25 mM [TMPP+-Ac-succinimide][Br'] in 10% acetonitrile were combined.
The pH of the reaction mixture was checked and, if necessary, the pH of the reaction
mixture was adjusted to pH 8.3 with NaOH solution. The reaction mixture was then
vortexed for 1 minute and allowed to stand at room temperature for 30 minutes. No
purification was performed prior to analysis by FAB or MALDI. However, purification

by HPLC was required prior to analysis by ESI.

E. HPLC Separation
Prior to analysis by ESI-MS or ESI-CAD-MS/MS, the reaction mixture was

purified by HPLC. The HPLC system consisted of two Waters Model 6000A pumps

67

.0E0co0 533302.00 0<.+mn=2._. .md 059“.

9:980 ethos: m a

210: + oosaoatzziwazodal

 

 

O O M
>
£5 om .QEQ E02
ZO<BNI
000m E0mm0m 0<i+n_n=>_._.
m a
mosaoatmzz + 2.. On__0~101wn_l
m
0 O 1

 

:00
O
:00

l

 

:00
O
100

[

 

68

controlled by a Waters Millenium data system. Purification was performed on a
Phenomenex Jupiter C18 reverse-phase column (S-ttm particle size, 4.6 x 250 mm) with
the eluant monitored at 215 nm. The mobile phase contained the following: (A) 0.1%
TFA, (B) acetonitrile / 0.1% TFA (90:10, v/v). The fraction containing the derivatized
peptide was collected and concentrated using a Speed-Vac (Savant Instruments,

Farmingdale, NY).

F. Analysis
Fast Atom Bombardment (FAB)

Fast atom bombardment and FAB-CAD-MS/MS spectra were obtained on a
double-focusing mass spectrometer (JEOL HX-l 10, JEOL, Tokyo, Japan). The
instrument utilizes a beam of 6-keV xenon atoms and a lO-kV acceleration voltage in the
source. The FAB matrix used was thioglycerol/2-hydroxyethyl disulfide (1:1, v/v). One
11L of underivatized peptide solution (0.5-2 nmol/1.1L in 50% acetonitrile) or one 1.1L of
derivatized peptide solution (100-500 pmol/11L in 50% acetonitrile) was combined with
the matrix for each analysis. For CAD-MS/MS experiments, helium was used as the
collision gas, the abundance of the precursor ions was attenuated to 50% for CAD
experiments, and 10 scans were averaged to obtain each spectrum. Ions generated by
FAB from Ultramark 1621 (PCR Inc., Gainesville, FL) were used for instrument

calibration. The data were acquired with a J EOL MS-MP8020D data system.

69

MALDI

MALDI mass spectra were obtained on a Voyager Elite reflectron time-of-flight
mass spectrometer (PerSeptive Biosystems, Inc., Framingham, MA) equipped with a
nitrogen laser (337 nm, 3-ns pulse) and delayed extraction. The acceleration voltage used
was 20-24 kV and the delay time was 50-100 ns. Data were acquired with the data
system provided and based on a transient recorder with 2-ns resolution. The matrix used
was or-cyano-4-hydroxycinnamic acid, dissolved in water/acetonitrile (1:1, v/v) to give a
saturated solution at room temperature. To prepare the sample, 1 [LL of the solution
containing the underivatized peptide (at a concentration of 5-10 pmol/1.1L) or the peptide
derivative (at a concentration of 5-10 pmol/11L) was added to 1 [IL of the matrix solution
and applied to a gold or stainless steel sample plate. The resulting mixture from the
derivatization was used directly. The mixture was then allowed to air dry before being
introduced into the mass spectrometer. Each spectrum was produced by accumulating
data from 256 laser shots. PSD spectra were generated by obtaining several spectra, each
optimized for a different range of m/z for the fragment ions. Selected regions of each are
stitched together to yield a single PSD mass spectrum. The time-to-mass conversion for
PSD was achieved by calibrating with the peak representing the C+ cation of the charged

derivative to be analyzed and the peak at m/z 379 representing the matrix dimer.

ESI-MS on an Ion Trap Instrument
ESI-CAD-MS/MS analysis was carried out on a quadrupole ion trap instrument
(LCQ, Finnigan, San José, CA). Spectra were acquired at a capillary voltage of 4.25 kV,

a cone voltage of 28 V, and a capillary temperature of 220 °C. Samples were analyzed by

70

direct infusion at a flow rate of 3 uL/min, with 20 scans averaged to generate each
spectrum. Peptides were diluted to a concentration of 5 pmol/11L in acetonitrile/0.1%
formic acid (1:1, v/v). TMPP+-Ac-derivatized peptides were diluted to a concentration of
5 pmol/1.1L in acetonitrile/water (50:50). The singly-charged precursor ions were selected

for analysis by MS/MS using a selection window of 3 daltons. Helium was used as the

collision gas, and MS/MS spectra were collected over a range of collision energies.

ESI-MS on a Triple Quadrupole Instrument

ESI mass spectra were obtained on a Sciex API 2000 mass spectrometer.
Samples were analyzed by direct infusion at a flow rate of 10 111/min, and 20 scans were
averaged to generate each spectrum. Peptides were diluted to a concentration of 50
pmol/1.11 in acetonitrile/water/formic acid (50:50:01). T MPP+-Ac-derivatized peptides
were diluted to a concentration of approximately 50 pmol/11.1 in acetonitrile/water (50:50).

Nitrogen was used as the collision gas for low-energy CAD experiments.

ESI-In-Source Fragmentation (ISF)-MS

ESI-ISF—MS mass spectra were obtained on a Micromass Platform mass
spectrometer. In—source fragmentation, also called nozzle-skimmer dissociation or cone
voltage fragmentation, is achieved by raising the cone voltage in order to increase the
kinetic energies of analytes. The energetic ions fragment during collisions in the source
region. Samples were analyzed by direct infusion at a flow rate of 20 til/min, and 7 scans

were averaged to generate each spectrum. TMPP+-Ac-derivatized peptides were diluted

71

to a concentration of approximately 5 pmol/11.1 in acetonitrile/water (50:50).

Fragmentation was induced by raising the voltage on the cone to the range of 55-100 V.

III. Fragmentation of TMPP+-Ac-derivatized Peptides
A. Fragmentation of the TMPP+-Ac Group

The FAB-CAD-MS/MS spectra of [TMPP+-Ac-OSu][Br'] (A) and [TMPP-Ac-
SPf][Br'] (B) are shown in Figure 2.4. The MS/MS spectra contain a number of fragment
ions including peaks at m/z 121, 151, 167, 181, 197, 351, 365, 381, 501, 515, 543, 559,
573, and 591. Also, losses of 15, 31, 32, and 46 Da from the molecular cation were
observed. Possible assignments of the major peaks in these spectra are shown in Table
2.]. Many of the fragments produced by these reagents are similar to the fragments
produced by alkyl triphenylphosphonium salts (2).

TMPP+-Ac-derivatized peptides produce many of the same fragments when
analyzed by FAB-CAD-MS/MS, as shown in Figure 2.5A. TMPP+-Ac-derivatized
peptides analyzed by MALDI-PSD produce fragments at m/z 181, 351, 365, 381, 501,
559, and 573, as shown in Figure 2.5B. Derivatized peptides analyzed by ESI-CAD-
MS/MS in an ion trap produced TMPP+-Ac related peaks at m/z 351, 365, 381, 399, 501,
527, 559, and 573 as shown in Figure 2.6A. Derivatized peptides analyzed by ESI-CAD-
MS/MS on a triple quadrupole instrument produced peaks at m/z 559, 570, and 573. The
TMPP+-Ac-related peaks produced by derivatized peptides analyzed by ESI-in-source
fragmentation-MS were observed at m/z 181, 197, 351, 365, 381, 501, 559, 570, and 573,

as shown in Figure 2.6B.

72

c+

q 573 6188

. TMPP+-Ac-OSu

656

a-
2 673
E /
Q) d 642 /l
.2
E
d)
a: _

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ci-
, 573 773
‘ TMPP+-Ac-SPf
.é‘l 181
(I)
C
% 742
3-
.g
g 758
0 559 11/
m 543 727 / /
A!
-151 365 515 //
167197 501 /
121 \ H 351481 \l l U591 711
-4443 w L
H I I m/z I

Figure 2.4. FAB-CAD-MSIMS spectra of TMPP"-Ac-OSu (A)
and TMPP+-Ac-SPf (B).

73

Table 2.1. Common TMPP"-Ac fragment ions

_m_/;
121
136
151
157
181
197
351
365
381
501
533
543
559
573
590
591

0+ - 62

C“ - 46

C” - 32

o*-31

0* - 15

observed in MS/MS spectra.

Possible Structure

 

[(CH30)2Ph+' 'CH3]
(CH3O)2Ph+'
[(C"|30)3F’h+ '0]
(CH30)3F’|‘1+
(CH30)3F’h-CH2+
(CH30)3Ph-OCH2+
[(CH30)3Ph]P+[(CH3O)2Ph-OH]
[((3"|30)3Ph]2':’+
{[(CH30)3Ph]2P "' O}”
[TM PP - OCH3]+'
[TMPP + H]+
[TMPP+-CH=C=O - 2(CH3)]
TMPP+-CH=CH2
TMPP+-CH=C=O
TMPP+-CH2-CO-NH2
TMPP+-CH2-CO-OH
[C " 2(OCH3)]+

[o - ooH3 -CH3]‘”

[C - CH30H]+

[C - OCH3]+°

[C ' CH3]+'

74

Relative Intensity

Relative Intensity

 

181

FAB-CAD-MSIMS

559
573

515

 

 

 

 

 

 

 

 

573
B
MALDI-PSD
1 81 559
365 501

197 351 381 \
u _ w WL/ u-:\_:L L / A: .JW;}a-—MJJM

r r mI/z r T

Figure 2.5. Low mass fragments of TMPP*-Ac-derivatized peptides
analyzed by FAB-CAD-MSIMS (A) and MALDI-PSD (B).

75

 

 

 

573

 

 

 

 

ESI-CAD-MSIMS
.3."
(0
5
E x§\
0 I
.2 I
E l
G)
0! 559 E
351 381 :
I
\ 385 399 501 527 :
I
[A 1. 4 Jul +T .L—
300 400 500
mlz
181 B
ESI-lSF-MS 573
.é‘
(D
C
.9
s
d)
:3 570
_g
d)
0‘ 197 365
351
381 501 559 L
260 360 ' 460 560
mlz

Figure 2.6. Low mass fragments of TMPP”-Ac-derivatized peptides analyzed
by ESl-CAD-MSIMS on an ion trap (A) and ESI-lSF-MS (B).

76

B. St

'1)

B. Sequence-specific Fragmentation of TMPP+-Ac-derivatized Peptides by
FAB-CAD-MS/MS

Twenty peptides were derivatized using [TMPP+-Ac-OSu][Br°] in unbuffered
solutions and analyzed by FAB-CAD-MS/MS as described above. The effects of the
amino acid residue at the nth position on the intensity of the *3“, *bn, *c...1, and *dn ion
abundances were studied. The nomenclature used here for the fragment ions of charge-
derivatized peptides corresponds to the nomenclature of protonated peptides when the
mass shift (due to derivatization) in (singly-charged) precursor ions is taken into account,
but an asterisk is added to denote the presence of the derivative moiety. The possible
structures of these ions are described in Chapter 1. The average signal-to-background
ratios (S/B) of the *a.,, *b... *c...1, and *dn peaks for each type of amino acid residue are
shown in Table 2.2. The average S/B ratios are grouped into five different categories.
An average S/B of S l, was designated with “-” to denote that the ion was not observed.
If the ion had an average S/B of > 1 and S 2, it was denoted with a “vw” to show that the
average abundance was very weak. Although it can be argued that peaks with very weak
abundances are the product of random noise, this category was included because the
fragment ions were sometimes clearly observed and sometimes not observed. Fragment
peaks with an average S/B of > 2 and S 4 were designated “w” for weak; those with an
average S/B of > 4 and S 10 were designated “m” for medium; and those with an average
S/B of > 10 were designated “3” for strong.

The "'an ions were observed for every type of amino acid residue. Residues with
aromatic side-chains (His, Phe, Trp, and Tyr) produced strong "‘an ions, whereas glycine

residues produced weak "‘an ions. The lower abundance of *3“ ions produced from

77

Table 2.2. Intensities of fragment ions from TMPP-Ac-
derivatized peptides as a function of amino acid residue.

 

Residue

al-
N
:1

*bn

*cn-1

*dn

 

 

Ala
Arg
Asn
Asp
Cys
Gln
Glu
Gly
His
lle
Leu
Lys
Met
Phe
Pro
Ser
Thr
Trp
Tyr
Val

mmm333m3333m53333333

333333

m(d). W(d')

3

m
m
VW

m
m(d). W(d')

VW
VW
m

 

(s = SIB > 10, m = SIB 4-10, w = SIB 2-4, vw = S/B1-2)

78

 

glycine residues suggests that *an ions may form most readily when there is a hydrogen
atom on the B-carbon available for transfer, as shown in Figure 2.7. According to this
mechanism, the formation of "‘an fragments at glycine could not occur because glycine
has no B—carbon atom on its side-chain. As a result, formation of "‘aa ions at glycine
would have to go through a less favorable mechanism, resulting in a decrease in the "‘an
ion abundance. As a result of deuterium—labeling experiments on charged derivatives,
Wagner (3) also concluded that the preferred mechanism for "‘an ion formation during
FAB-CAD-MS/MS involves transfer of a hydrogen from the B-carbon. However,
TMPP+-Ac derivatives analyzed by MALDI-PSD were found to generate abundant "‘an
ions for all amino acid residues except for proline, which lacks an amide hydrogen (4).
This suggests that different fragmentation mechanisms may predominate for charged
derivatives analyzed by FAB-CAD-MS/MS and those analyzed by MALDI-PSD.

The only *bn ions observed occurred when an aspartate residue was present at the
nth position. The proposed mechanism for ”1).. formation from derivatized aspartyl
peptides is discussed in detail in Chapter 3. The formation of *cn.1 ions was only
observed when selected residues were present at the nth position. The most abundant
*cn.1 fragment ions were observed from fragmentation preceeding serine or threonine.
Less abundant *c...1 fragments were observed when arginine, asparagine, aspartatate,
cysteine, or lysine were present at the n‘h position. These findings agree well with
sequence dependence observed by Downard and Biemann (5) for c.” fragment ions
produced from underivatized peptides.

The formation of *dn fragment ions was observed for most amino acid residues.

However, no "‘dn fragments were observed for residues lacking a B-carbon on the

79

RI"
RFC-{1} H

| I
WNH-Cfi-CIF/N-S
C+ 0
RI RII
\ /
i?
WNH-CH + H2+ O=C=N-5

*
an

Scheme 2.7. Formation of *an ions involving
transfer of hydrogen from the B-carbon.

80

side-chain (alanine, glycine, and proline), and the average S/B ratios for residues with
aromatic side-chains (histidine, phenylalanine, tryptophan, and tyrosine) were very weak
or not observed at all. Likewise, underivatized peptides generally do not produce dn
fragment ions resulting from the loss of aromatic side-chains (6, 7). The low abundance
of tin and "‘dn fragment ions probably results from the stability of the B-y bond for
peptides with aromatic side-chains. It is interesting to note that the aromatic residues
form "‘an ions with increased SIB ratios and ”Mn ions with decreased SIB ratios.
Apparently, the reduced ability to form *dn ions allows the formation of more abundant
*3“ ions.

In addition to the fragment ions described above, TMPP+-Ac-derivatized peptides
were observed to form fragment ions of 14, 30, and 46 daltons less than the masses of "‘a,I
ions, as shown in Figure 2.8. Note the fragment ions peaks near the *al and *3; peaks.
For simplicity, these ions will be called *an-14, a"an-30, and *an-46 ions. These fragment
ions were most commonly observed near "‘an ions produced from cleavage at an aromatic
residue. In order to investigate the origin of these fragment ions, the peptide osteocalcin
fragment (45—49) (FYGPV) was derivatized and incubated in D20 overnight to allow
deuteration of the exchangeable hydrogen atoms. The FAB-CAD-MS/MS spectrum of
the deuterated peptide contained the same *an-14, *an-30, and *an-46 ions, although the
masses were shifted by the presence of deuterium atoms. This result demonstrates that
the exchangeable hydrogen atoms on the peptide backbone are not lost in the formation
of *an-14, *an-30, and *an-46 ions. Osteocalcin fragment (45-49) was then derivatized
using vinyltriphenylphosphonium according the procedure of Wagner et al. (8) to

generate the N-terminal triphenylphosphonium-acetyl (TPP-Ac) derivative. The *an—14,

81

Relative Intensity

*a1

692

*az
855

TMPP+-Ac-FYGPV

 

 

 

 

 

 

 

 

 

 

 

800 700 800 900 1000 11'00
m/z

Figure 2.8. FAB-CAD-MS/MS spectra of TMPP+-Ac-osteocalcin fragment
(45-49), C" = 1154, showing *an-14, *an-30, and *an-46 peaks.

82

*an-30, and *an-46 fragment ions were not observed in the FAB-CAD-MS/MS spectrum
of this derivative. This result suggests that the formation of *an-l4, *an—30, and *an-46
fragment ions is related to losses from the TMPP+ head group. The abundances of these
fragment ions are roughly proportional to the abundance of the neighboring *an ion. As a
result, they are more readily observed in the vicinity of aromatic "'an ions because
aromatic residues produce the most abundant "'an ions, as discussed above.

Figure 2.9 provides a possible rationalization for the formation of I"an-14, *an-30,
and *an-46 ions. The *an-14 ion could possibly be formed by the net loss of 14 daltons
from the TMPP head group combined with "‘all fragmentation. However, the formation
of this fragment must involve a rearrangement of some type, because the loss of 8 CH2
neutral is not reasonable. The *an-3O ion could possibly be formed by the loss of an
O=CH2 neutral from the TMPP head group combined with *an fragmentation. The “‘81..-
46 ion could be formed by loss of CH3-O-CH3 from the head group combined with *an
fragmentation. FAB-CAD—MS/MS spectra of the "'au fragment ions produced in the
FAB-MS spectrum of TMPP+-Ac-osteocalcin fragment (45-49) contain abundant *an-46
ions and less abundant *an-14 and *an-3O ions. This observation suggests that I"am-46
ions may form directly from "‘an ions, but that *an-14 and *an-BO ions may from other
precursors. The formation of *an-l4, *an-30, and *an-46 ions has not been observed in
the spectra of TMPP+-Ac-derivatized peptides analyzed by MALDI-PSD, ESI—CAD-
MS/MS, or ESI-ISF-MS. The formation of these fragment ions appears to be limited to

high-energy CAD.

83

.mm>:m>:mu o<..+n_n=2._. So: mo 9. new .9... .3 Co momwo. .2 cozmuficozfi o_n_mmon_ .mN 239“.

m m

zoo zoo
foo foo a 0an
o :00
mo_aon_-o<l+n__l.@10fo All mn_ama-o<l+n__l1|©om:o
m foo m 8100 3.38.83
m foo a £00
_ _
8_amn_-o<1....n__|©: AIII 8_amn_-o<dn__|©omzo
m m
m m
m :00 m 100
_ _

mn_amn_-o<.un_ll@10I All mn_amn_-o<.nn__|©uomzo
_

m 8:00 m ”:00 3.338..

84

C. Sequence-specific Fragmentation of TMPP+-Ac-derivatized Peptides by
MALDI-PSD, ESI-CAD-MS/MS, and ESI-ISF-MS

Although the sequence-specific fragmentation of TMPP+-Ac-derivatized peptides
during analysis by MALDI-PSD, ESI-CAD-MS/MS, and ESI—ISF-MS was not studied as
systematically as fragmentation during analysis by FAB-CAD-MSIMS, several trends
were observed. Similar to observations in FAB-CAD-MS/MS, the most prominent
fragment ions produced during analysis by MALDI-PSD are the "‘an ions. However, in
contrast to FAB-CAD-MS/MS, the *dn and *c..-1 fragments ions are not usually observed
during MALDI-PSD analysis. Notable exceptions to these trends occur when the
derivatized peptide contains an aspartate residue. When an aspartate residue is present at
the nth position, no "'an ions are observed. Instead, *cn.1, *d... and "‘bu fragment ions are
observed. The formation of these fragmentation ions is discussed further in Chapter 3.
In addition, *b..+HzO fragment ions arising from the penultimate residue are frequently
observed in the MALDI-PSD spectra of TMPP+-Ac-derivatized peptides when the C-
terrninal residue is basic (arginine or lysine).

Both "'an and “b, ion series are observed in the ESI-CAD-MS/MS and ESI-ISF-
MS spectra of charge-derivatized peptides. In addition, losses of H20 and NH3 from "‘an
and *bn ions are commonly observed. Similar to MALDI-PSD spectra, the presence of
aspartate at the n‘h position of a TMPP+-Ac-derivatized peptide produces *c...1, *d.., and
more intense *bn ions. Also, *dn ions are sometimes observed when glutamate is present
at the nth position. The presence of *b..+Hzo ions at the penultimate position is observed
when a basic residue is present at the C-terminal position. In addition, a series of

*bn+H20 fragment ions is observed in the ESI-CAD-MS/MS spectra of TMPP-Ac-

85

 

derivatized peptides analyzed on an ion trap instrument. These fragment ions are

proposed to form by sequential loss of the C-terminal residue (9).

IV. Preparation of a Weighable TMPP+-Ac Derivative Standard

In the past, determination of reaction yields of the TMPP+-Ac derivative were
based on HPLC peak areas or the intensities of mass spectral peaks because no standard
was available for the TMPP+-Ac derivative. As a result, it was necessary to assume that
the response of the TMPP+-Ac-derivatized peptide was similar to that of the
underivatized peptide in order to estimate yields. In order to avoid this assumption, a
weighable amount of purified derivative was prepared. This standard allows the reaction
yields to be corrected for differences in the response of the underivatized peptide and the
TMPP+-Ac-derivatized peptide. Also, the use of a standard allows the study of ionization
efficiency and signal suppression for the underivatized and derivatized peptide.

To prepare the standard, 700 nanomoles of buccalin (GMDSLAFSGGL-NHZ,
MW = 1053.3 g/mol) were charge-derivatized in an unbuffered solution by the procedure
described above. The reaction mixture was separated by HPLC using the following
gradient: 25-40% B in 12 minutes, hold at 40% B for 18 minutes, 40-60% B in 15
minutes, hold at 60% B for 10 minutes. The resulting HPLC chromatogram is shown in
Figure 2.10A. Note the buccalin peak at a retention time of 9.9 minutes and the TMPP+-
Ac-buccalin peak at a retention time of 34.3 minutes. The TMPP+-Ac-bucca1in fractions
were collected, combined, and dried on a Speed-Vac. A total of 0.72 mg of the TMPPI-
Ac-buccalin standard was obtained. The molecular weight of the TMPP+-Ac-bucca1in

cation is 1626.8 g/mol. Assuming the product is present in the form of the

86

_9’
a:

 

 

 

28.3 34,3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I 5 mlz 623
g A 95 I ‘g 53.8
E ‘5 =
> + CD
a. a g,
E O D. <
C .E 2 I
to E I- E.
a 7“; ‘2L
‘8' g 1-
g; g mlz 585 ml
0 2
§ 4 242 849 TMPP
3 32.8 45'5
< \
Buccalin 23'5
9.9 J
L I A ._ W L
0 1'0 2'0 30 4'0 80 80
Time (minutes)
3.5 c
29.3 =
I §
0 “.3
‘5 2
E 19.3 I" ii
In “- a
‘- (D D.
N -: >3 2 mlz 623
*8 '7 *- *' 54 3
8 g 35.5 °
8 ’6.
g 02- mlz
«4 - 1- 649 TMPP
2 3‘33"" mlz 565 33.0 45,7
. 24.3 W W
0 1'0 2'0 3'0 40 5b 80

Time (minutes)

Figure 2.10. HPLC chromatograms of TMPP*-Ac-OSu + buccalin reaction
mixtures performed in unbuffered (A) and buffered (B) solutions.

87

trifluoroacetate salt, the formula weight of the salt would be 1739.8 g/mol and the

product would correspond to 410 nmoles of the derivative.

V. Reaction Yields and Selectivity
Reaction Yields in Buffered and Unbuffered Solutions

Any chemistry for the derivatization of the peptide N-terminus has the potential to
derivatize the side—chains of lysine residues and possibly the side-chains of other
residues. The unwanted derivatization of lysine side-chains has been reported by several
authors (IO-13). The derivatization of cysteine side-chains has been observed by Stults et
al. (14) and Wetzel et al. (15). Tyrosine side-chain derivatization was reported by
Cardenas er al. (10) and Kiplinger et al. (13). Careful regulation of the pH of the reaction
mixture has been reported as a means for selective derivatization of the N-terminus in the
presence of lysine (14, 15).

The chemistry originally proposed for the TMPP+-Ac derivatization used
4-dimethylaminopyridine as a basic catalyst (1). This reaction mixture is unbuffered, but
measurements with a pH electrode give a reading of approximately 9. The use of pH 8.2
phosphate buffer or Tris-HCl buffer has been reported as a means to selectively
derivatize the peptide N-terminus (16). The charge-derivatization of buccalin
(GMDSLAFSGGL-NHz) was studied in an unbuffered reaction mixture and in a reaction
mixture buffered to pH 8.3 with a Tris-HCl buffer in order to determine reaction yields.
The HPLC separation of the reaction mixtures was performed using the gradient

described in the previous section. The HPLC fractions were collected and analyzed by

88

MALDI-MS. Underivatized buccalin and the TMPP+-Ac-buccalin standard were
analyzed separately to determine the response of each species at 215 nm.

The HPLC chromatograms for the unbuffered and buffered reaction mixtures are
presented in Figure 2.10. Note the peak areas of the buccalin peaks (retention time = 9.9
minutes) and the TMPP+-Ac-buccalin peaks (retention times = 34.3 and 35.5 minutes).
Based on peak areas, the unbuffered reaction (A) had a yield of approximately 99.5%.
The derivatized peptide had a chromatographic peak area that was approximately 3 times
as large as an equal amount of the underivatized peptide. Based on the standards, the
unbuffered reaction had a yield of approximately 98.5% of the derivative. The buffered
reaction (B) had a yield of approximately 93% based on peak areas and a yield of 80%

based on the standards.

B. TMPP-Ac Derivatization of Lysine

It should be possible to selectively derivatize the peptide N-terminus in the
presence of the lysine by exploiting the pKa differences between the N—terminus (about
8.0) and the lysine side-chain (about 10.5) (17, 18). The extent of charge-derivatization
of lysine side-chains was studied using the peptide Ala-D—y-Glu-Lys-D-Ala—D-Ala. The
peptide was derivatized using [TMPP+-Ac-0Su][Br'] in an unbuffered solution and in a
pH 8.3 Tris-HCl buffer. The reaction mixture was separated by HPLC using the
following method: 1-25% B in 10 minutes, 25-45% B in 13.3 minutes, hold at 45% B for
6.7 minutes, 45-75% B in 40 minutes. The HPLC fractions were collected and analyzed
by MALDI-MS. The chromatograms and peak identities are shown in Figure 2.11. Note

the peak areas of the underivatized peptide (RT = 8.9 min), the N-terminal derivative

89

 

 

 

 

 

m w L
Z
w
W 4.
1 m u
<<xm+io<rnast~ 1
2.
3. 6.2.8.8? 98 as M U
3
3 8m ~21 1M
8

 

8. o:_u:>aoc_Em_>5¢E5-v

5. 259535855 8M\
55826255.; 8

9

 

<<xm$<9 8\

5'0

40

Time (minutes)

I

30

20

 

A

 

 

 

 

E: m K we oocmeomn<

 

TMPP

4
3 .
Sum. < 85. aasEm 4/

3 N 2.
m. 10.2.8829 98 E 5

Mm_._._..o<.+n.n__2._.1 3 m

393558858 5. .rlM\

 

«Saazhoixwi m

<§m+< m

 

 

 

17

4'0 5'0 8'0

30
Time (minutes)

20

10

 

 

E: m E am mocmEomn<

Figure 2.11. HPLC chromatograms of TMPP*-Ac-OSu + A-y-EKAA reaction

mixtures performed in unbuffered (A) and buffered (B) solutions.

90

(RT = 25.8 min), the lysine side-chain derivative (RT = 26.6 min), and the doubly-
derivatized peak (RT = 41.2 min). The 4-dimethylaminopyridine used in the unbuffered
solution elutes at a retention time of 9.8 minutes, overlapping with the peak from the
underivatized peptide, as shown in Figure 2.11A. This peak was collected and separated
with a different gradient to separate the two components. The MALDI-PSD spectra of
the purified peptide derivatives are shown in Figure 2.12. Note the absence of *al, *bl,
*b1+H20, *az, *b2, and *c2 peaks in the spectrum of the side-chain derivative (B). Also,
there is a *y3 peak in the spectrum representing cleavage at the N-terminal side of the
derivatized lysine residue.

Based on the HPLC peak areas of the purified products, the unbuffered reaction
mixture contained approximately 1% of the unreacted peptide, 61% of the N-terminal
derivative, 5% of the lysine side-chain derivative, and 33% of the doubly-derivatized
product, as shown in Figure 2.11A. When the derivatization was performed in a pH 8.3
Tris-HCl buffer, the reaction mixture contained approximately 9% of the unreacted
peptide, 54% of the N-terminal derivative, 19% of the lysine side-chain derivative, and
18% of the doubly-derivatized product, as shown in Figure 2.11B. The actual molar
percentage of the underivatized peptide should be greater than that determined by HPLC
peak areas because the underivatized peptide should have a lower molar absorptivity than
the derivatized peptide due to the absence of the TMPP+-Ac group. Based on the peak
area of a standard, the molar percentage of the underivatized peptide was 5% for the
unbuffered reaction mixture and 37% for the buffered reaction mixture, as shown in
Figure 2.11B. Also, the molar percentage of the doubly-derivatized product should be

less than that determined by HPLC because the doubly-derivatized product should have a

91

Relative Intensity

 

 

 

 

   

 

 

 

*b1+H2° *b +H 0 C“
882 ,b ‘9902 1081
2
A 773 N-terminal
*a +2
1617
ta1 *a2 ;a444
818 745 *62 . *b
573 “D1 790 ”3 9742
844 901
800 700 800 900 1000
mlz
*b4+H20 c+
90 1081
B Side-chain *a5+2 ‘
1017
*y *34
TMPP*-Ac-NH-(CH2)4-CH=NH 8631 944
873 *8, *b4
972
57 901
L . /.
800 700 800 900 1000
m/z

TMPP+-Ac-A-'y-EKAA
1 061

   
  
  

Relative Intensity

 

 

 

C 2(TMPP+-Ac)-A-vy-EKAA
1075
c2+ -H +
573 l 1633]
1029
800 800 1000 1200 1400 1600
mlz

Figure 2.12. MALDI-PSD spectra of N-terminal (A), side-chain (B),
and doubly-derivatized (C) A-y-EKAA.

92

larger molar absorptivity due to the presence of a second TMPP+-Ac group. Based on
this data, the only advantage of the buffered solution was a reduction in the amount of the

doubly-derivatized product.

C. TMPP+-Ac Derivatization of Cysteine

The extent of cysteine side-chain derivatization was studied using the peptide
glutathione ('y-Glu-Cys-Gly). The peptide was derivatized using [TMPP+-Ac-OSu][Br']
in an unbuffered solution and in a pH 8.3 Tris-HCl buffer. The reaction mixture was
separated by HPLC using the following gradient: 25-45% B in 13.3 minutes, hold at
45% B for 6.7 minutes, 45-75% B in 40 minutes. The HPLC fractions were collected and
analyzed by MALDI-MS and ESI-MS. The chromatogram and peak identities are shown
in Figure 2.13A. Note the peak areas of the N-terminal derivative (RT z 16.6 min), the
cysteine side—chain derivative (RT = 17.8 min), and the doubly-derivatized product (RT z
35.4 min). The peak representing the N-terminal derivative (RT —- 16.6 min) was broad,
presumably as a result of the free cysteine group. Based on the HPLC peak areas, the
unbuffered reaction mixture produced approximately 10% of the N-terminal derivative,
24% of the cysteine side-chain derivative, and 66% of the doubly-derivatized product, as
shown in Figure 2.13A. The amount of unreacted peptide was not determined because
the peptide eluted with the solvent front. The molar percent composition of the doubly-
derivatized product should be less than that determined by peak areas because it would be
expected to have a larger molar absorptivity than the single-derivatized products. When
the derivatization was performed in a pH 8.3 Tris—HCl buffer, the product had the

following composition based on peak areas: 1% of the N-terminal derivative, 85% of the

93

 

 

 

 

 

 

 

 

n 7 R

w e.

w

w
4.
a. m m
% Swine/4.882%
8 10.28.8558 NE 7. 21L
4mm
2
new NE. 2A

8
1 79883918.»

A e \
68.3585 8.

6 mEUt>QOEEm—>505_DIV

 

 

3. w

E: mew E 85983.

I

I

20

 

30 40 50 60

Time (minutes)

10

 

 

mlz 623

P6.

08.194.88.25
80.2.8839 1.
3

 

3

10.91.82» 58 NE 7.
«Nu 81m 8:.
7 83.2.8829

8

9
1 .47.. 98858558.»

 

24.8
ll

 

11.1 L31. -11..-.

 

08.3585 M

40

I

30
Time (minutes)

20

 

 

l

 

 

 

E: m K .m 8cm€omn<

Figure 2.13. HPLC chromatograms of TMPP"-Ac-OSu + y-ECG reaction

mixtures performed in unbuffered (A) and buffered (B) solutions.

94

cysteine side-chain derivative, and 14% of the doubly-derivatized product, as shown in
Figure 2.13B. Rather than limit derivatization of the cysteine side-chain, the use of Tris-
HCl buffer appears to limit N-terminal derivatization, thereby increasing the relative
amount of side-chain derivatization. It appears that blocking of the cysteine side-chain
(19) prior to charge-derivatization is necessary to achieve selective derivatization of the
N-terminus.

The ESI-CAD-MS/MS spectra of the derivatized products obtained on an ion trap
instrument are shown in Figure 2.14. The MS/MS spectra of the N-terminal derivative
(A) and the side-chain derivative (B) are surprisingly similar. The MS/MS spectrum of
the N—terminal derivative (A) has peaks at m/z 607 (TMPP+-Ac-SH) and mlz 751 (*y;)
that appear to arise from TMPP+-Ac derivatization of the cysteine side-chain. These
peaks probably result from incomplete separation of the N-terminal derivative from the
cysteine side-chain derivative. The major differences in the MS/MS spectra of the N-
terrninal derivative (A) and the side-chain derivative (B) is the presence of peaks at mlz
591 (*bo+H20) and 702 (*bl) in the spectrum of the N-terminal derivative and the
absence of those peaks the in spectrum of the side—chain derivative. Both of these peaks
are indicative of ’y-glutamate at the n = 1 position of a charge-derivatized peptide, as
discussed in Chapter 3. The MS/MS spectrum of the C2+ precursor ion of the doubly-
derivatized product has peaks that are indicative of derivatization at the N-terminus and
at the cysteine side-chain, as shown in Figure 2.14C. It is interesting to note that the
spectrum produced during analysis of the doubly-derivatized product by MALDI-MS has
no discernible C2+ peak, a very low intensity [C2+ - H]+ peak, and an intense

[C2+ - TMPP+-Ac]+ peak (data not shown). Therefore, analysis of the doubly-derivatized

95

Relative Intensity

TM PP+-Ac-SH

 

 

 

 

 

Relative Intensity

 

 

 

 

 

Relative Intensity

 

 

 

 

 

A 573 807
N-terminal c+
880
. C*-NH
b°+H2° 8833
// 702 751 i
V J. L L - i . 1 L fin. ._L Bf . r
m/z
TMPP+-AcsH
B 573 807
Side-chain
c-I-
4 c+’NH3880
883
*y, 835
751 I
. 1 .- l - I - .
mlz
02+
_ 726-5 TMPP"-Ac-ry-ECG
*b, 880
1 TMPP‘*-Ac-SH 702
C 807 *yz 848
573 751
1 °”” ’8.
591
['l'MPP-I-H] 898
1 533 L R l I
-_.L-_.L/._ _ 7 AA A LL . L Afik L m
mlz

Figure 2.14. ESl-CAD-MS/MS spectra of N-terminal (A),
side-chain (B), and doubly-derivatized (C) y—ECG.

96

product by MALDI-MS produces a spectrum that is very similar to that produced from

the single-derivatized products.

D. TMPP+-Ac Derivatization of Tyrosine

The extent of charge-derivatization of the tyrosine side-chain was studied for the
peptide acyl carrier protein fragment (65-74) amide (VQAAIDYING-NHZ). The peptide
was derivatized using [TMPP+-Ac-OSu][Br'] in an unbuffered solution and in a pH 8.3
Tris-HCl buffer. The reaction mixtures were separated by HPLC using the same gradient
used to separate the glutathione reaction mixture. The chromatograms and peak
identities are shown in Figure 2.15A. Note the peak areas of the tyrosine side-chain
derivative (RT z 19.8 min) and the doubly-derivatized product (RT —- 41.5 min). The
peak with a retention time of approximately 24 minutes represents a mixture of the
N-terminal derivatization product and the TMPP+-Ac-OH peak. This peak was collected
and separated by HPLC to determine the relative contribution of each component to the
total peak area.

The MALDI-PSD and ESI-CAD-MS/MS spectra of the purified peptide
derivatives are shown in Figure 2.16. Note that the N-terminal derivative (A) produced
only N-terminal fragments. The side-chain derivative (B) produced a mixture of
N -terminal and C-terminal fragments. All of the fragments produced include the tyrosine
residue. Also, an internal peak representing most of the derivatized tyrosine residue was
observed at mlz 708. The ESI—CAD-MS/MS spectrum of the doubly-derivatized product
has peaks found in both the N-terminal and C-terminal spectra. In addition, peaks

representing doubly-charged fragment ions were present at mlz 953, 1095.5, 1104. The

97

 

 

 

 

 

 

 

 

 

m ~ 4. 11
am >: w w
A
N:zéz_3.5.3558.n.55M
mas: m
3
7.
98 as m
m $9322.88? \ 6.2.88?
2 6 W
mom NF: 1.

.. \

N:2.sz_A..n_n__\fi.e<.:6_<<o> M

A

 

5.1
3

60

40 50

30
Time (minutes)

20

 

 

 

E: m N am 02.52036. >3

5.
6
4

7.
0
3

3.
4
2
3.
8
1

5.
3

m

 

 

 

 

 

 

 

 

 

 

 

«8 us
«282.524.023.882? MW
4
2
ans: .8. m
60.2.8.8?
N11
8
5.8 as E. W
332993.288? \ 10.91883
8. A
.2. 88 NE kw
8.5.2.882» .

NIzez5555528623> M

 

B m. e

E: m FN um 8cm€omn< >3

80

50

40

30
Time (minutes)

 

Figure 2.15. HPLC chromatograms of TMPP”-Ac-OSu + VQAAIDYING-NHZ

reaction mixtures performed in unbuffered (A) and buffered (B) solutions.

98

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

644 .32 A
772 N terminal
9% *c5 ' C“
5 1072 *6, 1634
:5 *a3 a 1098
a: * 5 *c
g *c 843 912 1027 *bs 1463 *8:
— 1 1170 1532
6‘5 688 \1‘3
L 1| 1 1305 1418 9L
mlz
W4
1037 +
TMPP*-Ac-O-CBH4-CH2-CH=NH B 1 6634
5'; 708 Side-chain
.8 /
s *3
g 693 Was 1407 *a, 1553
E 680
§ \
.V4 1 17° Doubly-derivatized
g 1037 c2+-NH
g £553
E r6, + TMPP+-Ac]2+ ./ [62+ . TMPP-Aer
.42: 953 1634
E *
é, a, 32 1633 1 01 8 .ys oy.’ 1.187 "y9
64:1 772 843 i L 1265 1336 1535
A 1 1' L . L .1 1‘1 . A AL 44 “A
600 800 1000 ml 1200 1400 1600
2

Figure 2.16. MALDI-PSD spectra of N-terminal (A) and side-chain (B)
derivatized VQAAIDYING-NHZ, and the ESl-CAD-MSIMS

spectrum of the doubly-derivatized (C) product.

99

MALDI-MS spectrum of the doubly-derivatized product (Figure 2.16C) contains a strong
peak resulting from the loss of a TMPP+-Ac group (mlz 1634), a very weak peak
resulting from loss of a proton (mlz 2207), and no C2+ peak (m/z 1104). As a result, it
was not possible to obtain a MALDI-PSD spectrum from a doubly-derivatized precursor
ion.

Based on the HPLC peak areas of the purified products, the unbuffered reaction
produced approximately 87% of the N-terminal derivative, 1% of the tyrosine side—chain
derivative, and 12% of the doubly-derivatized product, as shown in Figure 2.15A. The
amount of unreacted peptide was not determined because the peptide eluted with the
solvent front. When the derivatization was performed in a pH 8.3 Tris-HCl buffer, the
reaction mixture contained approximately 75% of the N-terminal derivative, 15% of the
tyrosine side-chain derivative, and 10% of the doubly-derivatized product, as shown in
Figure 2.15B. Although the amount of underivatized peptide was not determined, the
small peak areas for the derivatives are indicative of low yields in the Tris-HCl buffer.
Again, no advantage was observed for derivatization in the buffer except for a decrease in
the amount of doubly-derivatized product.

In summary, TMPP+-Ac-derivatization of peptides in a pH 8.3 Tris-HCl buffer
resulted in lower reaction yields than derivatization in an unbuffered solution. Also, no
increase in selectivity for the N-terminus was observed for peptides containing lysine,

cysteine, or tyrosine when derivatization was performed in a pH 8.3 Tris-HCl buffer.

100

VI. Ionization Efficiency and Signal Supression

Samples of underivatized and TMPP+-Ac-derivatized standards of Buccalin
(GMDSLAFSGGL—NHZ) were analyzed by MALDI-MS to study the ionization
efficiency and signal supression effects of the underivatized and derivatized peptides.
Three spots were prepared containing 5 pmol of the underivatized peptide; three spots
were prepared containing 5 pmol of the derivatized peptide; and 3 spots were prepared
containing 5 pmol each of the underivatized and derivatized peptide. Each spot was
analyzed by MALDI-MS with 256 laser shots. The intensities of the MH+, [M+Na]+, and
[M+K]+ peaks were added for the underivatized peptide, and the intensities of the C+ and
[C-H+Na]+ peaks were added for the derivatized peptide. The MALDI-MS spectra
produced are shown in Figure 2.17.

The average intensity of the TMPP+-Ac-buccalin peaks (3700 :1: 1000 counts,
Figure 2.17A) was smaller than the average intensity of the underivatized buccalin peaks
(5200 i 900 counts, Figure 2.178), but the difference in average intensities was not
statistically significant at the 95% confidence level (as determined by a two-sample t-
test). However, it is clear that the ionization efficiency of buccalin was not increased by
charge-derivatization. In contrast, charge-derivatization of another peptide has been
observed to produce a dramatic increase in ionization efficiency (20). Apparently,
charge-derivatization provides an increase in ionization efficiency for peptides that
normally do not respond well to MALDI-MS, but derivatization does not appear to offer
any further increase in ionization efficiency for peptides that respond well without

derivatization.

101

4000-‘[Buccalin + H?

 

 

 

 

 

 

 

*3 1053 A
5 300m ,
‘E Buccalin
— +
E 2000. figggcalin 4' Na]
U) .
"' uccahn + ‘*
(I) 1000- [B 1091 K]
O- . , _
1000 1200 mlz 1400 1600
TMPP*-Ac-Buccalin
4000‘ 1625
9%
g 3000- B
E TMPP*-Ac-Buccalin
“<6 2000‘
.5» ret +1 mar
"’ 1000‘ 1647
O...
1000 1200 m [2 1400 1600
TMPP*-Ac-Buccalin
1625
4000:1
.é‘ - +
g 3000 [306164151511 + H] _C
E [Buccalin + Na]“ Mlxture
- 2000‘ 1075
2 + [0* -H +Nal“
.9 uccalin +
w 10001 i / U91 K] 177
0- u —r—_——1———r———r——_-r—Lk~—:fi
1000 1200 mlz 1400 1600

Figure 2.17. MALDI-MS spectra of buccalin (A), TMPP*-Ac-buccalin (B),
and an equimolar mixture of the two (C).

102

When an equimolar mixture of buccalin and TMPP+-Ac-buccalin was analyzed in
three separate spots by MALDI-MS, the signal from the buccalin was suppressed, as
shown in Figure 2.17C. The average signal intensity of the buccalin peaks was reduced
from 5200 :1: 900 to 1680 :t 700, whereas the average signal intensity of the TMPP+-Ac-
buccalin (4900 :1: 2800) was not reduced. The ratio of the TMPP+-Ac-buccalin intensity
divided by the buccalin intensity had an average value of 2.8 :1: 0.7. Statistical analysis
confirmed that the buccalin intensity was significantly reduced at the 95% confidence
level (two-sample t-test), and that the ratio of TMPP+-Ac-buccalin intensity to the
buccalin intensity was significantly greater than 1.0 at the 95% confidence level (one-
sample t-test).

The signal suppression of the underivatized peptide by the TMPP+-Ac-derivatized
peptide suggests that MALDI-MS analysis is not a reliable method for estimating
reaction yields for TMPPI-Ac derivatization reactions. The suppression of underivatized
peptide means that the reaction yield estimates of TMPP+-Ac derivatization reactions

would be overestimated when determined by MALDI-MS.

103

VII. References

10.

11.

12.

Huang, Z.-H.; Wu J.; Roth, K. D. W.; Yang, Y.; Gage, D. A.; Watson, J. T. Anal.
Chem. 1997, 69, 137—144.

Claereboudt, J .; Baeten, W.; Geise, H.; Claeys, M. Org. Mass Spectrom. 1993, 28,
71-82.

Wagner, D. S. Ph.D. Dissertation, Michigan State University, East Lansing, MI,
1992.

Liao, P.-C.; Huang, Z.-H.; Allison, J. J. Am. Soc. Mass Spectrom. 1997, 8, 501-509.
Downard, K. M.; Biemann, K. J. Am. Soc. Mass Spectrom. 1993, 4, 874-881.
Johnson, R. S.; Martin, S., A.; Biemann, K. Int. J. Mass Spectrom. Ion Processes
1988,86, 137-154.

Papayannopoulos, I. A. Mass Spectrom. Rev. 1995, 14, 49-73.

Wagner, D. S.; Nieuwenhuis, T. J.; Chang, Y. S.; Gage, D. A.; Watson, J. T.
Proceedings of the 40m ASMS Conference on Mass Spectrometry and Allied Topics,
Washington, DC, May 31-June 5, 1992.

Lin, T.; Glish, G. L. Proceedings of the 45m ASMS Conference on Mass
Spectrometry and Allied Topics, Palm Springs, California, June 1-5, 1997.

Cardenas, M. S.; van der Heeft, E.; de Jong, A. P. J. M. Rapid. Commun. Mass
Spectrom. 1997, II, 1271-1278.

Kidwell, D. A.; Ross, M. M.; Colton, R. J. J. Am. Chem. Soc. 1984, 106, 2219-2220.
Kidwell, D. A.; Ross, M. M.; Colton, R. J. Springer Ser. Chem. Phys: SIMS 4,

1984, 36, 412-414.

104

l3.

14.

15.

l6.

17.

18.

19.

20.

Kiplinger, J. P.; Contillo, L.; Hendrick, W. L.; Grodski, A. Rapid Commun. Mass
Spectrom. 1992, 6, 747-752.

Stults, J. T.; Lai, J .; McCune, S.; Wetzel, R. Anal. Chem. 1993, 65, 1703-1708.
Wetzel, R.; Halualani, R.; Stults, J. T.; Quan, C. Bioconjugate Chem. 1990, I, 114-
122.

Huang, Z.-H.; Shen T.; Wu, J.; Gage, D. A.; Watson, J. T. submitted to Anal.
Biochem. 1999.

Greenstein, J. P.; Winitz, M. Chemistry of the Amino Acids; Wiley: New York,
1961; pp 486-488.

Means, G. E.; Feeney, R. B. Chemical Modification of Proteins; Holden-Day: San
Francisco, 1971; pp 14-15, 68-71.

Coligan, J. E.; Dunn, B. M.; Ploegh, H. L.; Speicher, D. W.; Wingfield, P. T.
Current Protocols in Protein Science, Vols. 1-2; Wiley: New York, 1998; pp 15.1.1-
15.1.18.

Liao, P.—C.; Allison, J. J. Mass Spectrom. 1995, 30, 511-512.

105

CHAPTER 3
FRAGMENTATION BEHAVIOR OF PEPT IDES AND CHARGE-

DERIVATIZED PEPTIDES CONTAINING ISOMERIC ACIDIC RESIDUES

I. Introduction
A. Isomeric Residues

Mass spectrometry is well suited to the determination of the amino acid sequence
of peptides. However, distinguishing between amino acid residues that differ in mass by
less than one mass unit is difficult. The monoisotopic masses of lysine and glutamine are
12809496 and 128.05856 mass units, respectively, a difference of 0.0364 mass units.
These residues can be distinguished by analysis on a high-resolution mass spectrometer
or by reaction with phenylisothiocyanate (1). The glutamine side-chain does not react
with phenylisothiocyanate, but the side-chain of lysine reacts to form a phenylthio-
carbamoyl derivative that is 135 daltons heavier than the mass of the lysine side chain.

Leucine and isoleucine residues are isomers with a monoisotopic mass of
11308406 mass units. These residues can be distinguished by the way they fragment
under high-energy collisionally-activated dissociation (CAD) conditions (2, 3). The (1..
ion produced by cleavage at the site of leucine is 42 daltons lighter than the
corresponding an ion, whereas two different dll ions result from similar cleavage at
isoleucine; the (1,, ions from isoleucine have masses that are 14 and 28 daltons less than
the corresponding a,l ion. Similarly, the mass of the Wn ion produced from leucine is 59

daltons less than the corresponding yn ion, and the WI] ions produced from isoleucine are

106

31 and 45 daltons less than the corresponding yu ion. Neither the (in ions nor the wn ions
described are produced during low-energy CAD.

Another challenge for mass spectrometry is identification of the isomeric forms of
aspartic acid and glutamic acid residues shown in Figure 3.1. The isomeric form of the
aspartic acid residue is called B-aspartate (IS-Asp) or isoaspartate. The side-chain of the
B-Asp residue is shorter by one methylene unit, but the peptide backbone has an
additional methylene unit between the Ot-carbon and the next amide toward the C-
terrninus; likewise, the side-chain of the y—glutamate residue (y—Glu) is shorter by two
methylene units and the backbone is two methylene units longer between the a-carbon
and the next amide bond.

The y—glutamate residue is present in a few naturally-occuring peptides. These 7-
glutamyl peptides have been found in brain tissues (4, 5), human urine (6-8), flaxseed (9),
the cell walls of E. coli (10), in folic acid cofactors of E. coli (11), and in the biomolecule
glutathione. Glutathione is a very important biopeptide that is involved in the transport
of amino acids across cell wells (12, 13), decomposition of hydrogen peroxide (14), DNA

and protein synthesis ( 15), and cellular regulation (16) among other functions.

B. Formation of B-Aspartate and 'y-Glutamate
Isoaspartate residues occur naturally in human urine (6, 8), and they can form by
the rearrangement of aspartic acid, asparagine, or aspartic acid esters (17). Isoaspartate
also forms in tryptic digests of serine proteases (18). The formation of isoaspartate from
asparagine, aspartate, or aspartic acid esters at mildly acidic, neutral, or basic pH is

shown in Figure 3.2. Aspartic acid, asparagine, or aspartic acid esters can undergo

107

OH

NH

O
Aspartate

0 OH
O

&NH

B—Aspartate

A

HO O

NH
O

Glutamate

O OH

NH
O

y-Glutamate

Figure 3.1. Structures of aspartate, B-aspartate, glutamate,
and y—glutamate.

c0328 5 2mtmamm-n 06 sewage“. .N.m 9:9“.

mm mam -
o e t < a
IZ
\c/112 48
N— O
-O O
506/.
00
N Hm mchao ac: I O moo/co
- rye
oEEEBSw .10

m 0
mi N I Z
2 V\ /x 4|
0 >
O XI

08:398.
m 0
J1... ._....
o -o
0
«66512.10;
m 0
J1... If.
0 x

O

109

rearrangement to form a succinimide structure. The formation of the succinimide
structure requires the loss of hydrogen from the amide nitrogen to form a reactive anion.
The succinimide ring opens to form a mixture of aspartate and B-aspartate at a ratio of
approximately (1 :3). Opening of the succinimide can be accompanied by a small amount
of racemization to form D-aspartate and D—isoaspartate residues (19). Analogous
reactions occur for glutamine, glutamate, and glutamic acid esters, but at much slower
rates (20).

The formation of the succinimide is the rate-limiting step in isoaspartate
formation. Fromation of the succinimide from Asn requires deprotonation of the peptide-
bond nitrogen; whereas succinimide formation from Asp requires both deprotonation of
the peptide-bond nitrogen and protonation of the side-chain carboxyl to form a hydroxyl
leaving group (21, 22). At 37°C and pH 7.4, the half-lives of aspartic acid methyl esters
in peptides are on the order of minutes, the half-lives of asparagine residues are on the
order of days, and the half-lives of aspartate residues are on the order of weeks (22).

The rate of succinimide formation is affected by the residue following the Asx
residue (where Asx denotes Asn, Asp, or aspartic acid ester). Succinimide formation
occurs most rapidly in peptides where Asx is followed by glycine or serine (22, 23).
Glycine has minimal steric hindrance. The —OH side-chain of serine is relatively small
and may stabilize the nitrogen anion intermediate necessary to form the succinimide.
Asn—Gly and Asn-Ser sequences are common sites for deamidation in proteins, although
the stability of the Asx residues is dependent on the three-dimensional structure of the
protein (24). The reaction depicted in Figure 3.2 does not occur when Asx is followed by

proline, because the proline nitrogen cannot form a reactive anion (19). In proteins,

110

isoaspartate most commonly forms from deamidation of asparagine, but isomerization of
aspartyl residues also occurs under mild conditions (25-27).

Asparagine residues are most stable in the pH range of 3 to 4 (28). As the pH
increases, the rate of succinimide formation increases, presumably because the extent of
deprotonation of the peptide-bond nitrogen increases at higher pH. At a pH range of l to
2, asparagine undergoes direct hydrolysis to form Asp without formation of B-Asp.

Deamidation and isoaspartate formation are of interest to the biotechnology
industry because of the importance of the stability and purity of protein-based
pharmaceutical products (29). Deamidation/isoaspartate formation, oxidation of
methionine, and oxidation of cysteine are considered the principal pathways of
degradation of native protein structures under mild conditions (30). Deamidation of Asn
and isomerization of Asp in proteins can cause loss of biological activity in proteins when
the change occurs in the region of the active site (25, 31-33). In some cases, more stable
variants of the proteins can be produced by replacing the unstable Asx residues with Gln,

Glu, or other residues (25, 32).

C. Conventional Analysis of B-Aspartate and y-Glutamate

A variety of techniques have been used to separate and identify B-Asp containing
proteins and peptides. As a general rule, it is easier to detect deamidation (conversion of
asparagine to a mixture of aspartate and B-aspartate) products than to distinguish between
the two forms of aspartate. Deamidated proteins can be separated from native proteins
using polyacrylamide gel electrophoresis (34), isoelectric focusing (35), and ion-

exchange chromatography (36, 37). In a few cases, hydrophobic interaction

111

chromatography (37) and affinity chromatography (38) have been successfully used to
separate the isoaspartyl and aspartyl forms of an intact protein; however, separation of B-
aspartyl variants from aspartyl variants usually requires preliminary digestion of the
protein. Isoaspartyl peptides can be separated from analogous aspartate, asparagine, or
succinimide-containing peptides by reverse-phase HPLC (39) or capillary electrophoresis
(40). When separated by reverse-phase chromatography, isoaspartyl peptides elute
before aspartyl peptides because the isoaspartyl side-chain ionizes to a greater degree
than the aspartyl side-chain. The determination of the relative acidities of the Asp and [3-
Asp side-chains are based on the acid dissociation constants of the 01- and B-carboxyl
groups of aspartic acid. The pKa of the B-carboxyl (analogous to the Asp side-chain) is
3.65, and the Ot-carboxyl (analogous to the B-Asp side-chain) pKa is 1.88. Similarly, the
pKa of the y—carboxyl of glutamic acid (analogous to the Glu side-chain) is 4.25, whereas
the a-carboxyl pKa (analogous to the 'y-Glu side-chain) is 2.19 (41).

Approaches to detecting deamidation include crystallographic methods (42),
release of ammonia (43), and cleavage with hydroxylamine (44). The most widely used
techniques for the identification of isoaspartate-containing proteins and peptides are
Edman degradation (45) and methylation with protein L—isoaspartyl O-methyltransferase
(PIMT) (46). Other approaches include 2-D NMR (47), tritium incorporation (48),
binding with monoclonal antibodies (49), and changes in the activity of certain proteases
(50). However, none of the techniques listed above is without limitations. Edman
degradation provides only indirect evidence of isoaspartate in the form of low yields for
the expected PTH-amino acids beginning at the site of the isoaspartate residue, and these

results can be inconclusive. The low yields are the result of inefficient cleavage of the

112

extended isoaspartyl backbone. PIMT is not commercially available, and the assay
requires micromolar quantities of PIMT for efficient methylation; also, the assay requires

the use of radiolabeled reagents for detection.

D. Mass Spectrometric Analysis of B-Aspartate and y-Glutamate

Mass spectrometry has also been used to detect deamidation and isoaspartate
formation. Deamidation of the asparagine residue of a peptide to form the analogous
aspartyl or isoaspartyl peptide can be recognized by the addition of one dalton to the
mass of the peptide (51). Likewise, the succinimide intermediate can be identified by a
loss of NH3 from the asparaginyl peptide or the loss of H20 from the aspartyl peptides
(52). When fragmentation occurs, mass spectrometry can verify the location of the newly
formed Asp or B-Asp residue based on the residue mass (53). However, this only
identifies the site of deamidation; it does not distinguish between aspartyl and isoaspartyl
residues. Mass spectrometry can also be used as an alternative to radiolabeling to detect
methylation of isoaspartyl residues by PIMT (54).

The presence of normal aspartate residues in peptides has been observed to
significantly alter fragmentation during analysis by mass spectrometry. Preferential
cleavage on the C-terminal side of Asp to form bu or yll ions has been observed for
peptides analyzed by LSIMS (55), ESI (56-58), and MALDI (55, 59). In addition, the
presence of aspartate at the nth residue of a charged-derivatized peptide has been
observed to produce abundant *c...1, *d.., and ”'13:. fragment ions during analysis by

MALDI-PSD (60).

113

Several researchers have attempted to use mass spectrometry to identify
isoaspartyl residues. Okada and Kawase (61) found differences in the fragmentation of
derivatized dipeptides with an N-terminal Glu or y—Glu residue when analyzed by EI-MS,
but the differences were not significant for larger peptides. Dipeptides with either Glu or
y—Glu at the N—terminus were analyzed by chemical ionization (62). The glutamyl
dipeptides produced abundant [MH-HZOTr and [MH-2(H20)]+ fragments whereas the
y-glutamyl peptides produced abundant fragment ions corresponding to protonated
pyroglutamic acid (m/z 130) and y1 fragment ions. Small peptides (24 residues) with an
acidic residue (Asp, B-Asp, Glu, or y—Glu) at the N-terrninus were analyzed by LSIMS-
high-energy CAD-MS/MS (63, 64). The spectra of the 'y-glutamyl peptides had a strong
yn peak corresponding to cleavage at the C-terminal side of the 'y-Glu residue; this peak
was absent in the spectra of the glutamyl peptides, but a strong peak at [MH-HZO]+ was
present instead. The relative intensities of the ya and [MH-H20]+ fragments differed for
aspartyl and B-aspartyl peptides, but the differences were not significant enough to allow
identification of the peptides unless both forms of the peptide are analyzed. Qin (65)
observed a stronger [MH-H20]+ peak for aspartame (Asp-Phe-OCH3) than for
B-aspartame (B-Asp-Phe-OCH3) when the samples were analyzed by atmospheric
pressure chemical ionization low-energy CAD-MS/MS.

Lloyd and co-workers found differences in the relative intensities of an and bn
ions generated from aspartyl, B-aspartyl, glutamyl, and y—glutamyl peptides during
analysis by FAB-CAD-MS/MS (66). They found that the ratio of the intensity of the hu
peak to the intensity of the an peak [I(bn)/I(an)] was S 1 for residues with a—linkages

(aspartate and glutamate) and that I(bn)/I(an) was 2 10 for residues with [3- or y—linkages.

114

However, the intensities of the an and bn peaks were dependent on the collision gas
pressure. Under similar conditions, Papayannopoulos and Biemann (52) observed
abundant an and bn ions for an aspartyl peptide, but not for the analogous B-aspartyl
peptide.

Violand et al. compared the fragmentation of an asparagine-containing peptide
(VFI‘NSLVFGTSD) and the analogous isoaspartyl peptide (VFTQ-QSLVFGTSD) during
analysis by FAB—CAD-MS/MS (67). They observed that the yg peak from the isoaspartyl
peptide was much more intense than the y9 peak from the asparagine containing peptide.
Unfortunately, the analogous aspartyl peptide was not analyzed, so it is unclear if the
fragmentation of the 0t- and B-aspartyl peptides would differ.

Carr et al. analyzed aspartyl and isoaspartyl forms of a tryptic peptide by FAB-
high-energy CAD-MS/MS (53, 68). The w ion produced from the aspartyl peptide was
more abundant than that produced by the isoaspartyl peptide. When the side chains of
Asp and B-Asp were methylated, the B-aspartyl peptide produced an abundant v6
fragment, but no W6 fragment, whereas the aspartyl peptide produced both v6 and W6
fragments. However, this approach is limited to high-energy CAD because v5 and W6
ions are not produced during low-energy CAD.

Isoaspartyl peptides with a basic residue near the N-terminus produced a bu +
H20 ion corresponding to cleavage at the N-terminal side of the isoaspartate residue (52).
N 0 corresponding fragment was produced from analogous aspartyl peptides; instead, a Cu
ion was formed. Bean and Carr (69) also observed a bn + HzO fragment in the high-

energy CAD-MS/MS spectrum of a B—aspartyl peptide, and in the spectra of charge-

derivatized B-aspartyl peptides. Schindler et al. (70) observed the formation of a ho +

115

H20 fragment ion in the ESI-low-energy CAD-MS/MS spectrum of an isoaspartate-
containing peptide.

Although several researchers have investigated the use of mass spectrometry to
identify B-aspartyl peptides, the studies have generally examined only a small number of
peptides. Also, very little work has been done using matrix-assisted laser desorption/
ionization with post-source decay (MALDI-PSD) or electrospray ionization collisionally-
activated dissociation tandem mass spectrometry (ESI-CAD-MS/MS) to identify
B-aspartate or y—glutamate residues.

This chapter presents the results of a study of six pairs of peptides,
differing only by the presence of a normal or isomeric acidic residue, by fast atom
bombardment collisionally-activated dissociation tandem mass spectrometry (FAB-CAD-
MS/MS), MALDI-PSD, and ESI-CAD-MS/MS. In addition, the results of parallel
experiments with the charged derivatives of each of these peptides are described. The
fragmentation of these peptides is evaluated for diagnostic purposes, and possible

fragmentation mechanisms are proposed.

II. Experimental
A. Reagents

The peptides Asp-Gly, B-Asp-Gly, aspartame, B-aspartame, reduced glutathione,
A1a-D-y-Glu-Lys-D-Ala-D-Ala, angiotensin II, and [B-Aspr-angiotensin II were obtained
from Sigma Chemical Co. (St. Louis, MO) and were used without further purification.
TRIZMA buffer was also obtained from Sigma. Buccalin and acyl carrier protein

fragment (65-47) amide were obtained from Bachem Bioscience Inc. (King of Prussia,

116

PA) and were used without further purification. Glu-Cys-Gly and Ala-Glu-Lys-Ala-Ala
were prepared in house using an Applied Biosystems automated synthesizer and were
purified by preparative scale reverse-phase HPLC. [B-Asp]3-buccalin and [B-Asp]6-acyl
carrier protein fragment (65-74) amide were prepared from buccalin and acyl carrier
protein fragment (65-74) amide, respectively, using the method of McFadden and Clarke
(71). The [tris-(2,4,6-trimethoxyphenyl)phosphonium] acetyl (TMPP+-Ac) derivatives
were made with the charged group attached to the peptide N-terminus. The synthetic
aspects are discussed in detail elsewhere (72). The reagent used is [TMPP+-Ac-
succinimide][Br']. It is prepared by first reacting bromoacetyl bromide with
N-hydroxysuccinimide and then adding tris-(2,4,6-trimethoxyphenyl)phosphine to

generate a salt.

B. Charge Derivatization

Peptide derivatives were prepared using the following procedure. 0.5-2 nmol of
peptide (in 2 1.1L 10% acetonitrile), 2 1.1L of 10 mM 4—dimethylaminopyridine in 50%
acetonitrile (v/v), and 1 11L of 10 mM [TMPP+-Ac-succinin1ide][Br'] in 50% acetonitrile
were combined, vortexed for 1 minute, and allowed to stand at room temperature for 30
minutes. FAB-CAD-MS/MS and MALDI-PSD analyses were performed without further
purification. Prior to analysis by ESI-CAD-MS/MS, the derivatized peptide was purified
by HPLC. The HPLC system consisted of two Waters Model 6000A pumps controlled
by a Waters Millenium data system. Purification was performed on a Phenomenex
Jupiter C18 reverse-phase column (S-ttm particle size, 4.6 x 250 mm) with the eluant

monitored at 215 nm. The mobile phase contained the following: (A) 0.1% TFA, (B)

117

acetonitrile / 0.1% TFA (90: 10, v/v). The gradient elution was carried out at a flow rate
of lmIJmin with gradient 25-45% B in 13 min, hold at 45% B for 7 min, 45-75% B in 40
min. The fraction containing the derivatized peptide was collected and concentrated

using a Speed-Vac (Savant Instruments, Farmingdale, NY).

C. Analysis

Fast atom bombardment-collisionally activated dissociation-tandem mass
spectrometry (FAB-CAD-MS/MS) spectra were obtained on a double-focusing mass
spectrometer (JEOL HX-l 10, JEOL, Tokyo, Japan). The instrument utilizes a beam of
6-keV xenon atoms and a lO-kV acceleration voltage in the source. The FAB matrix
used was thioglycerol/Z-hydroxyethyl disulfide (1:1, v/v). One pl. of underivatized
peptide solution (0.5-2 nmol/11L in 50% acetonitrile) or one 11L of derivatized peptide
solution (100-500 pmol/11L in 50% acetonitrile) was combined with the matrix for each
analysis. The pressure of the helium collision gas was adjusted to produce 50%
attenuation of the precursor ion abundance. Ten scans were averaged to obtain each
spectrum. Although the collision gas pressures used allow for multiple collisions, better
signal-to-background ratios were obtained at 50% attenuation (multiple-collision
conditions) than at 10% attenuation (single-collision conditions). The differences in the
CAD-MS/MS spectra of peptide pairs analyzed at 50% attenuation were similar to those
of peptide pairs analyzed at 10% attenuation. Ions generated by FAB from Ultramark
1621 (PCR Inc., Gainesville, FL) were used for instrument calibration. The data were

acquired with a J EOL MS-MP8020D data system.

118

MALDI-PSD mass spectra were obtained on a Voyager Elite reflectron time-of-
flight mass spectrometer (PerSeptive Biosystems, Inc., Framingham, MA) equipped with
a nitrogen laser (337 nm, 3-ns pulse) and delayed extraction. The acceleration voltage
used was 20 kV and the delay time was 100 ns. Data were acquired with the data system
provided and based on a transient recorder with 2-ns resolution. The matrix used was
Ot-cyano-4-hydroxycinnamic acid, dissolved in water/acetonitrile (1:1, v/v) to give a
saturated solution at room temperature. To prepare the sample, 1 1.1L of the solution
containing the underivatized peptide (at a concentratiOn of 50-100 pmol/1.1L) or the
peptide derivative (at a concentration of 10 pmol/11L) was added to 1 ILL ofthe matrix
solution and applied to a gold or stainless steel sample plate. The resulting mixture from
the derivatization was used directly. The mixture was then allowed to air dry before
being introduced into the mass spectrometer. Each spectrum was produced by
accumulating data from 256 laser shots. PSD spectra were obtained at a laser power 25%
greater than the threshold laser power. The instrument generates PSD data by obtaining
several spectra, each optimized for a different range of m/z for the fragment ions.
Selected regions of each are stitched together to yield a single PSD mass spectrum. The
time-to-mass conversion was achieved by calibrating with the peak representing the C”
cation of the charged derivative to be analyzed and the peak at m/z 379 representing the
matrix dimer.

ESI—CAD-MS/MS analysis was carried out on a quadrupole ion trap instrument
(LCQ, Finnigan, San José, CA). Spectra were acquired at a capillary voltage of 4.25 kV,
a cone voltage of 28 V, and a capillary temperature of 220 °C. Samples were analyzed by

direct infusion at a flow rate of 3 uL/min, with 20 scans averaged to generate each

119

spectrum. Peptides were diluted to a concentration of 5 pmol/11L in acetonitrile/0.1%
formic acid (1 :1, v/v). TMPP+- derivatized peptides were diluted to a concentration of 5
pmol/11L in acetonitrile/water (50:50). The singly-charged precursor ions were selected
for analysis by MS/MS using a selection window of 3 daltons. Helium was used as the
collision gas, and MS/MS spectra were collected over a range of collision energies.

Some ESI-CAD-MS/MS were obtained on a Sciex API 2000 mass spectrometer.
Samples were analyzed by direct infusion at a flow rate of 10 til/min, and 20 scans were
averaged to generate each spectrum. Peptides were diluted to a concentration of 50
pmol/uL in acetonitrile/water/formic acid (50:50:01). TMPP+-Ac-derivatized peptides
were diluted to a concentration of approximately 50 pmol/11L in acetonitrile/water
(50:50). Nitrogen was used as the collision gas for low-energy CAD experiments.

ESI mass spectra were also obtained on a Micromass Platform mass spectrometer.
Samples were analyzed by direct infusion at a flow rate of 20 til/min, and 7 scans were
averaged to generate each spectrum. Peptides were diluted to a concentration of 50
pmol/11L in acetonitrile/water/formic acid (50:50:01). TMPP+-Ac-derivatized peptides
were diluted to a concentration of approximately 5 pmol/11L in acetonitrile/water (50:50).

Fragmentation was induced by raising the voltage on the cone.

III. Results and Discussion
A. Underivatized Peptides

Previous studies on the fragmentation acidic peptides suggested that the presence
of a B-aspartate or 'y-glutamate residue at the nth position may alter the relative

abundances of [MH+-H20], cm, bn.l+HzO, an, and bn fragments; in addition, the

120

abundances of y, w, and v fragments resulting from cleavage near the acidic residue may
be altered (52, 53, 59, 61-70). Also, the MS/MS spectra collected for this study
suggested that the abundances of [MH+-NH3], (In and internal acyl fragments may be
altered. In order to evaluate the relative intensities of these fragment peaks, spectra were
collected in triplicate on each instrument and the percent contribution of each fragment
peak was calculated using the following equation:

A(x) = 100 * I(x) / [ I(MH+-HzO) + I(MH+-NH3) + I(cn.1) + I(b...1+H20) + I (11,.) + I(an) +

I(bn) + I(y) + I(w) + I(v) + I(intemal)]

where A(x) is the abundance ratio of a particular fragment ion peak, I(x) is the intensity
of a particular fragment ion peak, I(y) is the intensity of the y fragment resulting from
cleavage at the C-terminal side of the acidic residue, and I(intemal) is the total intensity
of internal acyl peaks representing fragments with an acidic residue at the N-terminal side
of the fragment. This equation has the same form as equations used to determine
branching ratios (73); however the A(x) values cannot be considered branching ratios
because not all of the fragment ions form from a common precursor. The abundance
ratio, A(x), is a number between 0 and 100 that represents the percent contribution of a
particular ion to the total ion abundance of the fragment ions considered.

To determine abundance ratios for the ESI-CAD-MS/MS experiments, the
relative collision energy was raised until the most intense fragment peak was
approximately equal to the intensity of the precursor ion peak. The relative collision
energy was in the range of 15-45% (of the maximum collision energy that can be
generated by the instrument) and the precursor ion attenuation was between 55% and

83%. Collisionally-activated dissociation in an ion trap occurs under multiple-collision

121

conditions because of the relatively high collision gas pressure (~1 mTorr) used. The
abundance ratios obtained by FAB-CAD-MS/MS, MALDI-PSD, and ESI-CAD-MS/MS
are presented in Tables 3.1-3.3. The standard deviations of these abundance ratios are

S 5 units.

y Ions

Most of the peptide pairs analyzed had large differences in the relative intensities
of the y ion peaks formed by cleavage at the C-terminal side of the acidic residue. Other
researchers have observed a similar trend for peptide pairs analyzed by chemical
ionization mass spectrometry (CI-MS) (62) and by liquid secondary-ionization-tandem
mass spectrometry (LSIMS/MS) (63, 64). The y ion abundance ratios obtained by FAB-
CAD-MS/MS and ESI-CAD-MS/MS were larger for B-aspartyl or 'y-glutamyl peptides
than for the analogous normal peptide for all peptide pairs except for the pair of
angiotensin II and [B-Asp]'-angiotensin II. For this pair, the trend was reversed. The
differences in y ion abundance ratios were even larger between glutamyl and 'y-glutamyl
peptides than the differences between pairs of aspartyl and B-aspartyl peptides, as
illustrated in Figures 3.3 and 3.4. Note that no y; peak is present in Figure 3.3A, but an
intense y; peak is present in Figure 3.3B. The spectra in Figure 3.4A and 3.4B both have
yl fragment peaks, but the y; peak is more intense in the B-aspartyl spectrum (Figure
3.43).

Possible mechanisms to rationalize the differences in y-ion abundance ratios are
shown in Figure 3.5. Peptides with a normal glutamate residue at the N-terminus (A)

form y ions accompanied by the loss of a neutral with a strained aziridinone ring (74). If

122

 

3- 3 3- 3 no no. 3- to. 3 a «3. ea :8:

3 no 3- 3. we. no. o.o o.o 2 2. no. :5

o.~ no 3m 3. Na 3 od 3. 3. R we N12.62.3523,?

m...” I «.2 od 3 3 ad od 3 2 o.» «12.02.5253

2. 3- we 3- 3- 3- 2. od 2 3- to. ._.—5

o.o 3 o.~ 3 3 3 3. 9o 9. 3. 3 412-589.5830

o.o ed 3. 3 3 3 od 3 8 om 33 «12-408385

3. o.o hm. No 2. od 3. 4.- t 2- m ea

3” od 3 ad ad ad 2 t 2 on t Max:518

2. O... 3. No 2. 9o 3 2 8 on 8 “Erika

3.. 3 3 3 we 3 o. 3. m- m 3- Eu

2: 3 so. 3 we no 3 we 2 «N 3 54me

we 3 3.2 on 2.. ea c v.» 3 8 m <§m<

ed to me. 3 o.o o.o od o.o mu. 8 8. :5

od 2. 3 3 o.o ad ad 9o 3 9 3 com»

0.0 so 3 3 2. 9o 2. o.o : K a wow

90 3. no 3 od o.o o.o od 3 mm 5- :5

od 3 to Na od 3. ed 2. Z a. No 6.2018

ad to be 3. ad ad o.o od 31 Mm E 22045
€525? 134. 3? 3.? Sites? Are? 32 E< 3:31.27. 6f..:§< $< 6888

.wémzdéei 3 68228 8288 86.2.85 6.. 8:9 8:853 .3. 6.an

123

 

mm- 3. 3. 3 no to o.o o.o no o.~ Q? :5 =8:

2- to. ma 3. 3 5 9o 3. no 3- m. :5

mm 3 2 3 N." 3 2. 9o 3 3. om «12.02.5259

2 o.» 9 3 2. 3 3 od 3 3 2 N:z.wz_>n=§d>

ad ad ad ad ad ad ad S. R 2- 3- :5

o.o ad ad ad ad ad od 2. 3 8 aw fzJoomufimonzw

od 3 9o o.o ad ad od 2. 8 t 2 «12-603585

ad ad ed 3 9o o.o o.o o.o : 5 I- :5

od 3. 2. ad ad ad ad o.o 2 3 t nag/Eon

9o 3 o.o od 3 ad ad S. R 3 8 "Erika

2- od 3 o 3. 2. 9o 2. 3, 2 o :5

3. o... S 2 9o Qo 3. od 3 mm 2 Sam?

2 od 3 2 3. ad ad ad 32 8 2 <<xm<

o.o o.o od 3. 3. od 3. S. o 2. od :5

ad ad o.o ad ad o.o od 0... mm 2 o.o com»

2. ad ad ad ad ad ad ad an 8 o.o mom

2. od o.o ad ad ad 2. 9o 3 2 t- :5

9o o.o ad ad ad ad ad ad 3 2 8 220.32.

2. 3. 9o 3 od 2. S. 2. Law 8 8 22un
$525? 3?. 37. 3% 6f+2£< is? if E< 1:2-.122 627:2? S< ~38“.

.omn.-.o._<_2 3 233.2 825$ 856252: .8 83mm 8:853. 3‘. 9:9“.

124

 

 

no. o.o 0.0 0.0 3 od No no 3 on mm- :5 cum:

n? o.o 3- 3. o.o 0.0 to to 2 2- .8- so

3 o.o 3. No o.o od 3. o.o 3 3 3 fzéz_>on_<<a>

to ad 5 ad ad ad to to 8 3 to «12.02.5253

3. 3 od 3. ad ad ad 3 mm mm. 3- ta

2. 9o o.o ad od ad 2. od 2. 8 3 «12--.835830

od 3. 9o o.o ad ad ad ad 82 o to «Iz-wawufimozo

2.? o.o ad ad ed 3 9o 3 2- 2- 3 :5

to S. o.o ad ad ad 3 3 8 on a "Excsmon

3 ad ad ad o.o od 3 3 8 cm 2. “31:53

5- od 3 o.o od 3 3 9o 3 S 8- so

3. ed 3 o.o o.o 9o o.o o.o 3 mm 5 42me

3 od 3 ad ad ad 3 S. «.2 2 2 <<xm<

3 od 3. od od 2. od od 3- 8 8- :5

9o o.o 3 od o.o od o.o 9o 3 m 8 00w»

2. od 3 ad ad o.o o.o od ed as o 00m

od o.o 9o 3. ad 3 9o 3. 3 mm X..- :5

S. .3 ad ad od 3 3 9o 3 E S «29.8

2. 3. od ad ad ad o.o 9o 92 8 «- mic-mo
€525? 3.2 E? 1.? aft-s:- AE3< $7. E< 2124.2? 6f-.:s=< S4. mason.

$236543 3 8%...» $258 $32.52: .9 8%.. 83.822 .3” was»

125

Normal

Relat've Inten 'ty
11111111ulluuluuluulu11]unlunluuln1111111111111

 

lsomer

Relative Intensity

D1
130

 

z2
1 62
1

ha
233

 

mlz

L l.

MH+-H20

290

 

MH+-H20
290

r-l

 

l_
I l I

Figure 3.3. ESI-CAD-MS/MS spectra of Glu-Cys-Gly (A) and
y-GIu-Cys-Cly (B) at a relative collision energy of 14.9%.

126

MH+-H20

 

 

 

 

 

 

 

 

 

 

277
3 A
3 Normal
g.
m-
c-
92
Si
2: b2
.3. 263
75: a2
'13 y- 235 MH+
_ 180 1245 Jl 295
l . . A ll - L
mu MH+
_ b2 295
: B 263
g»; lsomer
83 a
(D- 2
g: 235
.f‘>_’i 249 MH+-H20
E: 277
a)- Y1 /
‘1: 180 245
mlz

Figure 3.4. ESI-CAD-MS/MS spectra of Asp-Phe-OMe (A) and
B-Asp-Phe-OMe (B) at a relative collision energy of 17.4%.

127

A0 COOH
| C")

Glu: (EH
2 ,-<":NH
CHz—CH)\>
N— H

l

o
l
Hooo-on-oH2<II + +NH3-CH-CHS-NH-s
J.

R O
z-CH-C- NH-S

B) on
7-6.”: HOOC- CH CH2

5"ch= O R C")
H [NHz—CH-(ui-NH-s

l

HOOCQ R o
+ + I ll
s-N NH3-CH-C-NH-5

0

Figure 3.5. y ion formation by glutamyl (A)
and y-glutamyl (B) peptides.

128

the glutamate residue is not at the N -terminus, a more stable diketopiperazine neutral can
be formed. y—Glutamyl peptides (B) can form y ions accompanied by a stable
S-membered pyrrolidone ring (64). The pyrrolidone neutral formed by y—glutamyl
peptides is more stable than the aziridinone neutral formed by N-terminal glutamyl
peptides, and the MS/MS data suggest that the pyrrolidone neutral may be more stable
than the diketopiperazine neutral.

The larger y ion abundance ratios of B-aspartyl peptides can be explained by the
mechanisms in Figure 3.6 (55). Both aspartyl (A) and isoaspartyl (B) peptides can
possibly transfer a proton from the side-chain to a neighboring amide nitrogen. Although
this transfer produces additional charged sites on the ion, the intermediate could be
stabilized by ionic interactions (represented by dotted lines on Figure 3.6). The
carboxylate oxygen can then attack the amide carboxyl group to form a succinic
anhydride neutral and a y fragment ion. Based on these mechanisms, the relative
abundances of the y ions would be dependent on the ease of proton transfer from the
acidic side-chain.

The B-aspartyl side-chain is more acidic in solution than the aspartyl side-chain.
This assertion is based on the acid dissociation constants of the or- and B-carboxyl groups
of aspartic acid in solution. The pKa of the B-carboxyl (analogous to the Asp side-chain)
is 3.65, and the (x-carboxyl (analogous to the B-Asp side-chain) pKa is 1.88. Similarly,
the pKal of the y—carboxyl of glutamic acid (analogous to the Glu side-chain) is 4.25,
whereas the or-carboxyl pKa (analogous to the 'y-Glu side-chain) is 2.19 (41). Although
losing a proton in solution is quite different than losing a proton in the gas phase, it is

possible that the B-aspartyl side-chain loses a proton more readily than the aspartyl

129

A) Asp:

o o
H o H (I; 0‘
(I: H+ l ‘. H+
are i It") —> Mr» W
s-NH-CH-p-NH-CH-cuz-s S—NH-CH-C-Kle-CH-fiz-s
o o o o
o l W
s—NH Ff
o + NHz-CH-("3-s
o
o
B) B-Asp:

O + O
H n

H+
n
c—o—H R c—o-~~~ R
I V l -—> i NR II I
s—NH-CH-CHz-CHD-NH-CH-le-S S-NH-CH-CHz-CHZ-NHz-CH-C-S

O

i +

O H
s—NH II? I
o + NHz-CH-(fi-s

O

0

Figure 3.6. y ion formation by aspartyl (A)
and B-aspartyl peptides (B).

130

side-chain in the gas phase. If so, the B-aspartyl peptides would be expected to produce
more abundant y ions than the normal aspartyl peptides, as was observed from the
abundance ratios. However, the differences in the y-ion abundances from pairs of larger
isomeric peptides (10-11 residues) were much smaller than those observed for pairs of
small isomeric peptides.

It is not clear why the y ion abundance ratio of angiotensin II was greater than that
of [B-Asp]l-angiotensin H. Ionic salt bridges have been proposed to form between
aspartate residues and neighboring arginine residues (75). The formation of salt bridges
can significantly alter the fragmentation of peptides (76). The formation of this salt
bridge may stabilize the ionized B-aspartate side-chain and limit formation of the succinic

anhydride.

[MH+-H20] and [MW-NH3] Ions

The [MH+-H20] abundance ratios from pairs of small acidic peptides (2-5
residues) were noticably different for MS/MS spectra generated by FAB or ESI. For
these peptide pairs, the [MH+-H20] abundance ratios were greater for normal peptides
than for the isomeric peptides. The differences between glutamyl and ’y-glutamyl
peptides were greater than the differences observed between aspartyl and B-aspartyl
peptides, as demonstrated by Figures 3.3 and 3.4. Similar results have been observed by
other researchers using CI-MS (62), LSIMS-CAD-MS/MS (63, 64), and by APCI-CAD-
MS/MS (65). For larger peptide pairs (8-11 residues), the abundance ratios did not

follow this trend.

131

The differences in the [MH+—H20] abundance ratios of glutamyl and y—glutamyl
peptides can be explained by the mechanisms shown in Figure 3.7. The glutamyl
peptides can readily lose water through the formation of a pyroglutamate structure (64); a
similar reaction can occur in strongly acidic solutions (17). Peptides with an N-terminal
y—glutamate residue cannot lose water easily through the analogous mechanism because
the product would have a strained 3-membered ring. When y—glutamate is present at any
position other than the N-terminal position, the carboxy side chain can be attacked by a
different amide nitrogen to form a product with a more stable 6-membered ring.

During mass spectral analysis, aspartyl and B-aspartyl peptides probably
dehydrate through the formation of a succinimide structure, as shown in Figure 3.8. This
mechanism is analogous to the dehydration in solution presented in Figure 3.2. An
alternative mechanism for the dehydration of aspartyl and isoaspartyl peptides was
proposed by Qin (65) and is based on the high temperature dehydration of dry aspartame
(77). It is not clear why dehydration occurred to a greater extent for aspartyl peptides
than for B-aspartyl peptides. The mechanisms proposed in Figure 3.8 do not predict
significant differences in the [MH+-H20] peak intensities of aspartyl and B-aspartyl
peptides. The most likely explanation is that the competing formation of y ions limits the
dehydration of B-aspartyl peptides. Since the aspartyl and B-aspartyl peptides can
fragment by a similar mechanism, it is not surprising that the differences in their
dehydration peak intensities would be less significant for larger peptides.

Differences in the [MH+-NH3] abundance ratios were observed for several pairs

of peptides. However, these differences do not follow any discernible pattern. Rather,

132

‘3'": H02 Z CH2 R

 

o o f
H+
l-H2o

”O R
5-N ("3- --NH CH-C-S

L o <">|

 

 

 

H.
B) _?
y-Glu: "IQ/('3 R
H— N- CH- CH 2-CH 2-c-NH-(':H- C-5
l o c")
H.
H 20
R
lob—CH2 -CH 2-c-NH- CH-C-S
C") ,
H+O

Figure 3.7. [MH+-H20] formation by
glutamyl (A) and y-glutamyl (B) peptides.

133

8258 av _Bamma 21$ 3:88 3 c2558. _o~z-+..=2_ .3” 2:9".
+1 +1

0 o w 2%
To-:o-z-o-~_._o-:o-xz..
$-10: TNI
. ._.

 

 

IZ.~

 

 

_ _ _
of m 1%0

o a

+1

 

 

n__u IZ.~
T 0-IM_VIZ

 

134

the abundance ratio differences seem to be the result of large changes in the intensities of

peaks representing [MH+-H20] and the y ions.

w and v Ions

Differences in the w and v ion abundances of aspartyl and B-aspartyl peptides
have been reported by Carr et al. (53), but no significant differences were observed for
the peptides in our study. The peptide pairs analyzed produced only low-intensity w and
v peaks, most likely because they lacked charge localization near the C-terminus.
Although no significant differences in w and v abundance ratios were observed for these
pairs of peptides, significant differences might be observed for those with a basic residue

near the C-terminus (such as tryptic peptides), when fragmented by high-energy CAD.

c...“ b...1+H20, and bn Ions

Although differences have been reported in the c“ (52), b...1+H20 (52, 69, 70)
and bn (66) peak intensities of peptide pairs with an acidic residue at the nth position, no
significant differences in the c“, bn.1+H20, or bll abundance ratios were observed for the
peptide pairs analyzed here. This may be because the c“, b...1+H20, and l)ll ions arising
from the peptides in our study were of low abundance. It is possible that significant
differences would be observed for peptide pairs with sequences that allow the formation

of abundant c“, b...1+H20, or bll ions.

135

Internal Fragment Ions

Whenever internal acyl fragment ions were produced with the acidic residue at the
N-terminus of the fragment, the abundance ratio was greater for the peptide containing
B-aspartate or y—glutamate than for the analogous peptide containing normal aspartate or
glutamate, as demonstrated by the ESI-ISF-MS spectra in Figure 3.9. Note that the
internal acyl ion at m/z 258 (EK) produced from the aspartyl peptide (A) is less abundant
than the corresponding fragment produced from the B-aspartyl peptide (B).
Unfortunately, the peak intensities of these fragments were too small for the differences
in abundance ratios to be statistically significant, but it is possible that the presence of

B—aspartate or y—glutamate enhances the formation of internal acyl ions.

B. Charge-derivatized Peptides

Charge-derivatization of peptides localizes the charge away from the peptide
backbone, and minimizes or eliminates fragmentation by charge-driven mechanisms (60).
Charge-derivatization of the peptides in this study produced striking changes in
fragmentation patterns during analysis by MS/MS. Examination of the MS/MS spectra
of charged derivatives suggested that the presence of a B-aspartate or 'y-glutamate residue
at the nth position may alter the relative intensities of *cM, *b...1+H20, *dn, *an, and *bn
fragment peaks. The nomenclature used here for the fragment ions of charge-derivatized
peptides will be presented in a review of charge-derivatization techniques (78) and
corresponds to the nomenclature of protonated peptides when the mass shift (due to
derivatization) in (singly-charged) precursor ions is taken into account. An asterisk is

added to denote the presence of the charged derivative moiety. It is necessary to

136

Relative Intensity

Relative Intensity

 

4.

 

 

 

 

 

 

 

Normal
b2
201
b2+Hzo
EK 347 D4
EK-H2o 253 400
240 ’4
b: 418
I ’ 1 286 329 /
_l “ ‘ ~
mlz
lsomer
Y3
289
-EK-H o
y 240’
82 253 329 b
4
fl“ 400 Y4
418
l. 1.1.. ‘ ._k....l.._..-.. _..L_L_
mlz

Figure 3.9. ESI-ISF-MS spectra of AEKAA (A)
and A-(y-E-KAA (B) at a cone voltage of 55 V.

137

MH"
489

 

 

Ml-I"
489

 

distinguish between the fragment ions of underivatized and charge—derivatized peptides
because the structures and fragmentation mechanisms may be different (78). In order to
evaluate the relative intensities of these fragment peaks, MS/MS spectra for the
derivatized peptides were obtained in triplicate on each instrument and branching ratios
for each of the fragment peaks were calculated using the following equation:
B(x) = 100 * I(x) / [ I(*cn.1) + I(*b.,-1+H20) + I(*d.,) + I(*a..) + I(*b,.)]

where B(x) is the branching ratio of a fragment ion peak and I(x) is the intensity of that
particular fragment ion peak. Unlike the similar equation used to compare the relative
abundances of fragment ions from underivatized peptides, the values determined by this
equation represent true branching ratios because all of the fragment ions are believed to
come from the same precursor, an N-terminal derivatized peptide with the charge located
on the derivative moiety.

To determine branching ratios for the ESI-CAD-MS/MS experiments, the relative
collision energy was raised until the most intense fragment peak was approximately equal
to the intensity of the precursor ion peak. For singly-charged precursor ions, the relative
collision energy was in the range of 37-60% (of the maximum collision energy that can
be generated by the instrument). The branching ratios obtained by FAB-CAD-MS/MS,

MALDI-PSD, and ESI-CAD-MS/MS are presented in Tables 3.4-3.6. The standard

deviations for these ratios were S 9 units.
*bnlons

The differences in the “b, branching ratios of charge-derivatized peptides were

quite large. Most derivatized peptides with B-aspartyl or y—glutamyl residues at the nm

138

Table 3.4. Branching ratios for derivatized peptides
analyzed by FAB-CAD-MSIMS.

 

Peptide B(*c.,.1) B(*b...1+H20) B(*d..) B(*a..) B(*b..)
DF-OMe 9.9 7.7 36 39 8
BDF-OMe 7.9 13.3 21 24 35
Diff 2.0 -5.6 15 15 ~27
ECG 2.0 13 47 34 3
yECG 10.7 31 24 1 1 24
Diff -8.7 -18 23 24 ~21
AEKAA 14 10 63 7.7 6
AyEKAA 12 35 22 1 1 .6 20
Diff 2 -25 41 -3.9 -14
DRWIHPF 1 1 3 31 23 31
BDRWIHPF 15 14 14 12 44
Diff -4 -1 1 17 1 1 -13
GMDSLAFSGGL-NH; 16 10 39 24 10
GMBDSLAFSGGL—NHZ 21 23 19 17 19
Diff -5 -13 20 7 -9
VQAAIDYING-NH; 19 10 28 21 23
VQAAIBDYING—NHZ 16 1 1 21 29 23
Diff 3 -1 7 -8 0
Mean Diff -1.5 -12 20 8.2 -15

139

Table 3.5. Branching ratios for derivatized peptides
analyzed by MALDI-PSD.

 

Peptide B(*c.,.1) B(*b.,.1+H20) B(*d..) B(*a.,) B(*b..)

DF-OMe 28 14 37 6.5 15
BDF-OMe 16 29 5 1.7 49
Diff 12 -15 32 4.8 -34

ECG 0.0 0.0 32 63 5

yECG 0.0 0.0 5 14 81
Diff 0.0 0.0 27 49 -76

AEKAA 0.0 3 95 1 1
AyEKAA 0.0 74 1 1 1 13
Diff 0.0 -71 94 -10 -12
DRWIHPF 18 10 16 2.4 53
BDRWIHPF 24 20 8 1.3 46

Diff -6 -10 8 1 .1 7
GMDSLAFSGGL-NHz 18 16 41 1.9 23
GMBDSLAFSGGL-NH; 1 1 36 18 1.6 33
Diff 7 -20 23 0.3 -10
VQAAIDYING-NH; 18 17 34 2.1 29
VQAAIBDYING-NH; 18 38 15 2.2 26
Diff 0 -21 19 -0.1 3
Mean Diff 2.2 -23 34 7.5 -21

140

Table 3.6. Branching ratios for derivatized peptides
analyzed by ESI-CAD-MS/MS.

 

Peptide Bren-1) B(*bn-1+H20) B(*dn) Evan) Bvbn)
DF-OMe 11.2 5.3 21 0.4 62
BDF-OMe 2.3 7.8 1 0.0 89
Diff 8.9 -2.5 20 0.4 -27
ECG 0.0 0.0 0.0 0.0 0.0
YECG 0.0 0.0 0.0 0.0 0.0
Diff 0.0 0.0 0.0 0.0 0.0
AEKAA 0.9 13 81 1 4
A‘YEKAA 0.8 27 0 22 50
Diff 0.1 -14 81 -21 ~46
DRWIHPF 5.5 0.5 8.1 0.0 86
BDRVYIHPF 10.0 5.7 8.9 1.0 74
Diff -4.5 -5.2 -0.8 -1.0 12
GMDSLAFSGGL-NHz 2.8 4.7 18 0.1 75
GMBDSLAFSGGL—NH; 2.8 5.2 18 0.1 74
Diff 0.0 -0.5 0 0.0 1
VQAAIDYING-NH: 0.6 1.2 0.6 0.1 97
VQAAIBDYING-NHg 4.0 2.3 2.6 0.1 91
Diff -3.4 -1.1 -2.0 0.0 6
Mean Diff 0.2 -4.6 20 -4.3 -11

141

Relative Intensity

 

 

 

 

*bo+H20
591

*c

 

*d1

Relative Intensity

 

*a1-H20

l 616 Mai

642

 

 

A
Normal

821

 

lsomer

m/z

C4-
867

C4»
867

 

Figure 3.10. MALDI-PSD spectra of TMPP+-Ac-Asp-Phe-0Me (A)
and TMPP+-Ac-B-Asp-Phe-0Me (B).

142

position had larger B(*bn) ratios than analogous derivatized peptides with normal
aspartyl or glutamyl residues. An example of this difference is provided in Figure 3.10.
Exceptions to the trend were observed for larger derivatized peptides (8 or more residues)
analyzed by MALDI-PSD and ESI-CAD-MS/MS. Also, the ESI-CAD-MS/MS spectra
of TMPP+-Ac-Glu-Cys-Gly and TMPP+-Ac-y-Glu-Cys—Gly were dominated by a
rearrangement peak at m/z 607 (TMPP+-Ac-SH). As a result, it was not possible to
calculate reliable branching ratios for these derivatives when analyzed by ESI—CAD-
MS/MS. For the larger derivatized peptides, the differences in the B(*bn) ratios were not
statistically significant based on a two-sample t—test at a 95% confidence level. Lloyd
and coworkers (66) also observed more intense bn fragment peaks in the FAB-high
energy CAD-MS/MS spectra of underivatized peptides with B-Asp or 'y-Glu at residue n
than for analogous peptides with Asp or Glu at residue 11.

The differences in the B(*bn) ratios can be rationalized by the mechanisms
presented in Figure 3.11. Mechanism A is an adaptation from that presented by Yu et al.
(55) to explain the enhanced bu formation from underivatized peptides with an aspartate
residue at position 11. A similar mechanism was proposed by Summerfield, Whiting, and
Gaskell (75). According to the mechanism, the acidic proton of the aspartate side—chain
can be transferred to the neighboring amide nitrogen. The intermediate can be stabilized
by ionic interactions (represented by dotted lines in Figure 3.11). The ionized side-chain
can then attack the amide carbonyl to form a *bn ion with a stable succinic anhydride
structure. Figure 3.113 shows a similar mechanism for the formation of *bn ions from
derivatized B-aspartyl peptides. The side-chain of the B-aspartate residue is more acidic

in solution than the aspartate side-chain (41) and could possibly be more acidic in the gas

143

9

c—o\H
A) R CH2
ASP: WH-cH-co NH-cH-co NH-s
122
CH2 J a
WH ('3H-co NH-ch (3-rqo NH2-s
R CH2 [\C \O
MH-cH-co NH-CH\C + NH2-5
9 o—H
B) R C/

- ; I I
B As” @VVLNH-CH-CO-NH-CH-CHz-CO-NH-S

9
WNH-cH-co NH-CH-CHz-CD NHz-s

l

9 0‘9/O\c-o
MH-CH-CO-NH-CH\092 + NH2-5

Figure 3.11. *b'1 formation by derivatized
aspartyl (A) and B-aspartyl (B) peptides.

144

phase. If so, the initial proton transfer would proceed more readily for B—aspartyl
peptides, resulting in greater *bn ion formation.

The *bn fragmentation mechanisms were investigated further by preparing methyl
esters of selected TMPP+-Ac-derivatized peptides to eliminate the acidic proton on the
side-chains. When the esterified TMPP+-Ac-derivatized peptides were analyzed by
MALDI-PSD, the *bn fragment peaks were absent, as shown in Figure 3.12. This
verifies that the transfer of a proton from the acidic side-chain is necessary for the
formation of "‘b,I fragment ions in agreement with the mechanism shown in Figure 3.11.

Derivatized glutamyl and ‘y-glutamyl peptides could form *bn fragments by
mechanisms similar to those in Figure 3.11, producing "‘bn ions with a six-membered ring
instead of the five-membered ring produced by aspartyl and B-aspartyl peptides. If the
relative solution phase acidities hold in the gas phase, the derivatized ‘y—glutamyl peptides
would be expected to form more intense "‘bll peaks than the derivatized glutamyl peptides
because of the difference in the side-chain acidities. This is consistent with the data as
summarized by the *bn branching ratios in Tables 3.4-3.6.

The similarity of the branching ratios for larger acidic peptides analyzed by
MALDI-PSD and ESI-CAD-MS/MS is probably the result of competition from the

greater number of other fragmentation pathways available.

*dn Ions
The branching ratios for the *d.. fragment ions also had large differences for

analogous pairs of peptides. For purposes of this study, the "‘dll fragment of derivatized

peptides containing B-Asp or y—Glu at residue n is defined as the fragment with the same

145

*d1 A C*-15

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

616 852
Normal 1
*a1 821
’5‘; .c 660
c o *b
91590 1
s 642 688
d)
.2
E
0)
cc W
w
600
Gag-g6
*a1
674 B
Methylated

.é‘
(I)
C
93
s
:2: *co 642
g 590 2
6: 91

m)

NJ
600 T 760 ' 800

 

Figure 3.12. MALDI-PSD Spectra of TMPP+-Ac-Asp-Phe-0Me (A)
and Methylated TMPP+-Ac-Asp-Phe-0Me (B).

146

m/z value as the *dn fragment of an analogous derivatized peptide containing normal Asp
or Glu. For B-aspartyl or y—glutamyl peptide derivatives, these "‘dn fragments do not
result from cleavage between the B- and y— carbons of the side-chain, but these fragments
may have the same structures as the *dn fragments of normal aspartyl or glutamyl peptide
derivatives, as illustrated by Figure 3.14.

For all the derivatized peptides analyzed by FAB-CAD-MS/MS and MALDI-
PSD, the derivatives containing normal Asp or Glu produced "'dn peaks that were more
intense (as determined by branching ratios) than the "‘dll peaks produced from the
analogous derivatized peptide containing B-Asp or y—Glu. An example of these
differences is shown in Figure 3.13. The same trend was observed for TMPP+-Ac-
aspartame/1‘ MPP+-Ac-B-aspartame and for TMPP+-Ac-Ala-Glu-Lys-Ala-Ala/
TMPP+-Ac-Ala-D-‘y-Glu-Lys-D-Ala-D-Ala as analyzed by ESI-CAD-MS/MS. The ESI-
CAD-MS/MS spectra of TMPP+-Ac-Glu-Cys-Gly and TMPP+-Ac-y-Glu-Cys-Gly were
dominated by a rearrangement peak at m/z 607 (TMPP+-Ac-SH), so none of the expected
peaks was observed. For larger derivatized peptides (8 or more residues), no significant
differences in the *dn branching ratios, B(*dn), were observed. These fragmentation
differences provide clues to the fragmentation mechanisms for the formation of *dn
fragment ions. These fragments probably do not form by the same mechanism as the (ln
ions of underivatized peptides (3) because (Iu fragments are generally not observed during
low-energy CAD (79).

Possible mechanisms for the formation of *dn fragment ions are shown in Figure
3.14. The top mechanism was proposed by Liao, Huang, and Allison (60) and the lower

mechanism is an adaptation of it. Both mechanisms involve a proton transfer from the

147

 

    

 

 

 

 

 

 

 

*31 c+
616 A 1061
.é‘
(D
C
.9
E
Q)
.2
E
a)
0:
fi— i—gi_k~- *7? _r-i—‘fi’
600 700 800 m/z 900 1000
*b1+H20 0+
662 B 1061
a. Isomer
.7) a1
2:, 616
3; *b,+H2o
g *d 990
E 682 *a2 773 W * °+'C°2
a: 7 745 944 b4 1017
(I 790 972 J
600 700 800 m/ 960 1090
Z

Figure 3.13. MALDI-PSD spectra of derivatized
AEKAA (A) and A-y-EKAA (B).

148

Asp' | V
@WbNH-CH-CHS-NH-CH-CHZ-S
O 0
F5
@VVMNH-CH=CH2 + CO2 + H-(E-NH-CH-CIE-5
O O
B) 9
B-Asp: 919/— H [7 F5
@WbNH-CmCH2-("3-NH-CH-ICI-6
O O
5*
@VVMNH-CH=CH2 + CO2 + H-CHS-NH-CH-CHS-5
O 0

Scheme 3.14. *dn formation by derivatized
aspartyl (A) and B-aspartyl (B) peptides.

149

acidic side-chain to the C-terminal fragment. It is likely that the side-chain is involved in
fragmentation because *du fragments are not observed for most residues during analysis
by MALDI-PSD or low-energy CAD. The formation of (in fragment ions from
underivatized peptides during analysis by low-energy CAD has been observed by other
researchers (75, 80) for peptides with an acidic residue at the 11‘11 position and a basic
residue near the N-terminus. Selected peptide derivatives were methylated to eliminate
the acidic proton. After methylation, the "'dn peaks were less intense, but were still
present, as shown in Figure 3.12. This result suggests that transfer of an acidic proton is
involved in the preferred fragmentation mechanism, but *dn fragments can also form
through other mechanisms.

It is not clear why the '"dl2 branching ratios are larger for derivatized aspartyl and
glutamyl peptides than for the analogous derivatized B-aspartyl and 'y-glutamyl peptides.
The mechanisms in Figure 3.14 suggest that both aspartyl and B-aspartyl peptide
derivatives should be able to form *dn fragment ions with comparable efficiency. The
difference in ‘l‘dn branching ratios may be the result of competition from "‘bII and

*bn.1+H20 formation which are favored by B-aspartyl and 'y-glutamyl peptide derivatives.

*b...1+I-I20 Ions

For the derivatized peptides analyzed with an acidic residue at the nth position, the
*b...1+H20 branching ratio was larger for the derivatized B-aspartyl or y—glutamyl peptide
of a given pair than the branching ratio for the corresponding derivatized aspartyl or
glutamyl peptide. However, not all of the differences in the branching ratios were

statistically significant as determined by a two-sample t-test. When the derivatized

150

peptides studied had the acidic residue at the N-terminal position, a *bo-I-HzO ion
(TMPP+-Ac-OH) was formed with an m/z value of 591. The formation of *bn.1+H20
ions from charge-derivatized B-aspartyl peptides was observed by Bean and Carr (69)
during analysis by FAB-CAD-MS/MS. Similarly, the formation of b...1+H20 fragments
from underivatized B-aspartyl peptides has been observed by several researchers (52, 69,
70). The formation of analogous fragments from underivatized and charge-derivatized
peptides suggests that the both bn.1+H20 and *bn.1+H20 fragments may form by charge-
remote mechanisms.

The differences in *bn.1+H20 branching ratios can be explained by the
mechanisms presented in Figure 3.15. The bottom mechanism is adapted from that
proposed by Papayannopoulos and Biemann (52) to explain the formation of bn.1+H20
ions from underivatized peptides with B-aspartate at the nth position. Aspartyl peptides
cannot form
*bn.1+H20 fragment ions by this mechanism (A) because one of the neutral fragments
produced would be unstable. Derivatized 'y-glutamyl peptides could form *bn.1+H20
fragments by a mechanism similar to the lower mechanism in Figure 3.15. Derivatized
peptides with a normal glutamyl residue cannot readily form *b...1+H20 fragments by the
same mechanism because one of the neutral fragments formed would be unstable, just as
for the derivatized aspartyl peptides.

The intensity of the *bn.1+H20 peak produced from a charge-derivatized
B-aspartyl peptide with an esterified side-chain was much lower than the intensity of the
same derivatized peptide with an unmodified side-chain (the *bn.1+H20 branching ratio

decreased from 24 to 6 as a result of esterification). This indicates that a proton transfer

151

A) Ff G) CH2

l
Asp; WH-CH-C-NH-CH-CO-NH—s

Fr 9 9H2
@WbNH-CH-C-OH + co + NH=Cc>H-co-NH-s

B) H/O‘E
B-Asp: 9 (91) i

WH-CH-C—NH-CH-CHz-CO-NH-s
/

Ho o— J
9 i/U 7
@WLNH-CH-q/V
NH—CH

i ‘CHz-co-NH-s

O

R
@VVbNH-CIDH-C-OH + co + NH=CH-CH2-CO-NH—S

Figure 3.15. *bn_1+H20 formation by derivatized
aspartyl (A) and B-aspartyl (B) peptides.

152

is involved in the formation of *b...1+H20 fragments by derivatized B-aspartyl peptides,

and it is consistent with the mechanism presented in Figure 3.15.

*cn.1 Ions

The differences in the *c...1 branching ratios of normal derivatized peptides and
isomeric derivatized peptides were small and did not follow any clear trend. However,
for many of the derivatized aspartyl peptides, the *c...1 peak was more intense than the
adjacent *b...1+H20 peak as shown in the top portion of Figure 3.10 (see peaks at mlz
590 and 591). Papayannopoulos and Biemann (52) reported a similar observation for
underivatized peptides. The greater intensity of the *c...1 peaks than the *bn.1+H20 peaks
of derivatized aspartyl peptides is probably a result of a less favorable pathway for
derivatized aspartyl peptides to form *b...1+H20 fragment ions, as depicted in Figure
3.15A. At least some of the intensity attributed to the *b...1+H20 peaks actually arises
from the isotope peak of the *c...1 ions, so the differences between the *c...1 and
*bn.1+H20 peak intensities should be greater than that shown in the branching ratios. It
is interesting that for many of the derivatized glutamyl peptides, the *bn.1+H20 peak is

more intense than the ”‘an peak.

Liao, Huang, and Allison (60) proposed a mechanism for the formation of *cm
fragments from charge-derivatized peptides with an aspartate at residue 11. This
mechanism is shown at the top of Figure 3.16. The bottom of Figure 3.16 shows that
derivatized B-aspartyl peptides are less likely to form *cM fragment ions by this

mechanism because the neutral fragment produced would be unstable. However, when

153

9

A :
Sp R \ (6H2 R
WcH-R-NH- CH-R-NH-CH-R-‘S
("3 ('3 ("3
Fr (2
WCH-R-NHZ + 002 + CH2=CH-R--NH-CH--R-S
o o o
B)
0
.As - / \
B p R 1,H c=o 9
WcH-R-NH-c NCH 2-R-NH-CH-R"
('3 ("3 ('3
R G R
WRH R-N H2 + 602 + CH: CH 2-R-NH ('3H-R-S
("3 ("3 ("3

Figure 3.16. *cn,1 formation by derivatized
aspartyl (A) and B-aspartyl (B) peptides.

154

the aspartyl side-chain of a derivatized peptide was converted to a methyl ester, *c...1
fragment ions were still readily formed, as shown in Figure 3.12. This result contradicts
the mechanism of Figure 3.16, because it involves the transfer of an acidic proton from
the aspartyl side-chain for *c.“ ion formation. The mechanism of Downard and Biemann
(81) can be adapted to describe the formation of *c.“ fragment ions from charge-
derivatized aspartyl and glutamyl peptides without the transfer of a proton from the
carboxylic acid group, as shown in Figure 3.17. In the top mechanism, the hydrogen
atom on the methylene group is activated because it is adjacent to a carbonyl. B-Aspartyl
peptides cannot form *c“ fragment ions by this mechanism because there is no
methylene group on the B-aspartyl side-chain. The formation of ”‘an ions from
derivatized glutamyl peptides is not favored because the neutral formed would be
unstable. It is not clear whether *c...1 fragment ions form exclusively through the
mechanisms in Figure 3.17 or whether they form by a combination of the mechanisms

shown in Figure 3.16 and 3.17.

"‘an Ions

The differences in the *an branching ratios are small. For many derivatized
peptide pairs, the "‘all branching ratio was larger for the derivatized peptide containing a
normal aspartyl or glutamyl residue than for the derivatized peptide containing a
B-aspartyl or y—glutamyl residue. This difference in branching ratios is probably a default
result of significant changes in the branching ratios of some of the other fragmentation

products as described earlier. Since the differences in the *an branching ratios are minor,

155

 

‘0 6—0H

Asp: R (H; CI3 H R
WcH-R-NchH-R-NH-RHR4
O O Q
5‘ ' F3
WCH‘9'NH2 + HO-le-CH=CH-("3-NH-CH-C"3‘5
O o (3 o
B) (“3
Glu: ('3—OH
H—RH
F: 9H2 1
WCH-R-NHZCH-R-NH-CH-R-s
O O 0
i
9 6) (-3 F}
WCH-R-NH2+HO-R-CH-CH2=CH-R-NH-CH-R-s
O O 0

Figure 3.17. *cn_1 formation by derivatized aspartyl (A) and
glutamyl (B) peptides without tranfer of the acidic proton.

156

it is likely that the "‘an fragmentation mechanism does not involve the acidic side-chain.

A reasonable fragmentation mechanism was proposed by Liao, Huang and Allison (60).

[C-Hzol” and [C+H-H20]2+ Ions

In contrast to the spectra of underivatized peptides, derivatized peptides analyzed
by FAB-CAD-MS/MS, MALDI-PSD, ESI-triple quadrupole-MS/MS, or ESI-ISF-MS
produced extremely weak [C-H20]+ peaks. However, ESI-triple quadrupole-MS/MS
spectra acquired from the doubly-charged precursor, [C+H]2+, of TMPP+-Ac-AEKAA
produced a significant [C+H-H20]2+ fragment ion peak, as shown in Figure 3.18. The
low intensities of the [C-H20]+ peaks suggests that a mobile proton may be necessary to
produce intense dehydration peaks from acidic peptides during analysis by these
techniques. When dehydration did occur, no difference in the relative intensities of the

[C+H-H20]2+ peaks of the normal or isomeric peptide were observed.

IV. Conclusions

There are differences in the fragmentation patterns of pairs of underivatized
aspartyl or glutamyl peptides and their analogous B-aspartyl or y—glutamyl peptides; the
relative abundances of fragment ions are compared on the basis of branching ratios. For
the peptides in this study, B-aspartyl or y-glutamyl peptides produced more abundant y
ions and less abundant [MW-H20] ions than those produced by the corresponding
normal aspartyl or glutamyl peptides. However, these differences are less apparent for
larger peptides (8-11 residues). Fragmentation mechanisms are proposed that rationalize

the differences observed. Differences in the abundances of internal acyl ions were

157

A A

. Relative Intensity

 

c+

 

 

 

 

 

 

 

Relative Intensity _

 

 

 

 

 

 

 

 

1061
[TMPP-CH=CO]*
573
A $1.
C-I-
l
l
l *d“
687
* 559-
11,241 *34 *b4‘l'H20
*b 732 944 99°
1 . *a '30 *b
644 a2-30 843 \ 4
. Ill 1 71.5 2 L 1 1,972] i
mlz
* 1
687”‘b2-41
B 732
1 + 2+
[C+H]2+ [C + H]
531
'32-30
'31 715
616
*83'30
Hem-112013: 843
i 522 *b
1
. *6 +H 0
‘ £73 644 a‘ 9902
. / *c1 94: *b4 / C"
559 l 661 l l h 972 1061
L . ‘ *3 v l‘ 1 . 4 ‘ TA . H ‘ L. . ‘
m/z

Figure 3.18. ESI-CAD-MS/MS spectra of the C“ (A) and
[C+H]2+ (B) precursor ions of TMPP*-Ac-AEKAA at collision
energies of 72.5 eV and 32.5 eV, respectively.

158

observed, but the intensities of the corresponding peaks were too small for the differences
to be statistically significant. No differences in the abundances of v, W, cm], b...l+H20,
or bu ions were observed for the peptide pairs in this study, but differences might be
observed for peptides that produce these fragment ions in large abundance.

The fragmentation patterns of pairs of charge-derivatized peptides also had
several differences. Charge—derivatized B-aspartyl or y—glutamyl peptides produced more
abundant *bn and *bn-1+H20 ions, and less abundant "‘dII ions than the analogous charge-
derivatized aspartyl or glutamyl peptides. Mechanisms are presented that rationalize the
greater abundance of *bn and *bn.1+H20 ions for derivatized B-aspartyl and y—glutamyl
peptides. The decrease in the "‘dll branching ratios could not be rationalized
mechanistically, and it appears to be a default result of competition by the *bn and *b...
1+H20 fragmentation pathways. Methylation of the acidic side—chains showed that
transfer of an acidic proton from the side-chain plays a role in the formation of *b.., *d..,
and *bn.1+H20 ions. As with the underivatized peptides, the differences in the branching
ratios were smaller for pairs of larger (8-11 residues) charge-derivatized peptides.

Although differences in the fragmentation patterns of pairs of underivatized
peptides and pairs of charge-derivatized peptides were observed, these differences appear
to be analytically useful only for small peptides (2-5 residues). For larger peptides (8-11
residues), these differences become smaller as a result of competition from the greater

number of available fragmentation pathways.

159

V. References

10.

ll.

12.

Edman, P. Acta Chem. Scand. 1950, 4, 238-293.

Johnson, R. S.; Martin, S. A.; Biemann, K.; Stults, J. T.; Watson, J. T. Anal. Chem.
1987, 59, 2621-2625.

Johnson, R. S.; Martin, S. A.; Biemann, K. Int. J. Mass Spectrom. Ion Processes.
1988, 86, 137-154.

Nakamura, K.; Higashiura, K.; Nishimura, N .; Yamamoto, A.; Tooyama, 1.; Kimura,
H.; Ienaga, K. J. Neurochem. 1990, 55, 1064-1066.

Reichelt, K. L. J. Neurochem. 1970, 17, 19-25.

Buchanan, D. L.; Haley, E. E.; Markiw, R. T. Biochemistry 1962, I, 612-620.
Martin, S. A.; Karnovsky, M. L.; Krueger, J. M.; Pappenheimer, J. R.; Biemann, K.
J. Biol. Chem. 1984, 20, 12652-12658.

Lou, M. F. Biochemistry 1975, 14, 3503-3508.

Klosterrnan, H. J .; Lamoureux, G. L.; Parsons, J. L. Biochemistry, 1967, 6, 170-177.
Van Heijenoort, J.; Elbaz, L.; Dezélée, P.; Petit, J.-F.; Bricas, E.; Ghuysen, J.-M.
Biochemistry 1969, 8, 207-213.

Ferone, R.; Hanlon, M. H.; Singer, S. C.; l-lunt, D. F. J. Biol. Chem. 1986, 261,
16356-16362.

Meister, A. In Functions of Glutathione: Biochemical, Physiological,
Toxicological, and Clinical Aspects; Larsson, A., Orrenius, S., Holmgren, A.,

Mannervik, B., Eds.; Raven Press: New York, 1983; pp 1-22.

160

13.

14.

15.

l6.

17.

18.

19.

20.

21.

22.

Vifia, J .; Vina, J. R. In Functions of Glutathione: Biochemical, Physiological,
Toxicological, and Clinical Aspects; Larsson, A., Orrenius, S., Holmgren, A.,
Mannervik, B., Eds.; Raven Press: New York, 1983; pp 23-30.

Jones, D. P.; Kennedy, F. G. In Functions of Glutathione: Biochemical,
Physiological, Toxicological, and Clinical Aspects; Larsson, A., Orrenius, S.,
Holmgren, A., Mannervik, B., Eds.; Raven Press: New York, 1983; pp 109-116.
Holmgren, A. In Functions of Glutathione: Biochemical, Physiological,
Toxicological, and Clinical Aspects; Larsson, A., Orrenius, S., Holmgren, A.,
Mannervik, B., Eds.; Raven Press: New York, 1983; pp 199-204.

Orrenius, S; Jewell, S. A.; Bellomo, G.; Thor, H.; Jones, D. P.; Smith, M. T. In
Functions of Glutathione: Biochemical, Physiological, Toxicological, and Clinical
Aspects; Larsson, A., Orrenius, S., Holmgren, A., Mannervik, B., Eds.; Raven Press:
New York, 1983; pp 261-272.

Barany, G.; Merrifield, R. B. In The Peptides, Volume 2; Gross, E.; Meinhofer, J.,
Eds.; Academic Press: New York, 1979; pp 1-284.

Naughton, M. A.; Sanger, F.; Hartley, B. S.; Shaw, D. C. Biochem. J. 1960, 77,
149-163.

Geiger, T.; Clarke, S. J. Biol. Chem. 1987, 262, 785-794.

Clarke, S.; Stephenson, R. C.; Lowenson, J. L. In Stability of Protein
Pharmaceuticals, Part A, Chemical and Physical Pathways of Protein Degradation;
Ahem, T. J ., Manning, M. C., Eds.; Plenum Press: New York, 1992; pp 1-29.
Brennan, T. V.; Clarke, S. Protein Sci. 1993, 2, 331-338.

Stephenson, R. C.; Clarke, S. J. Biol. Chem. 1989, 264, 6164-6170.

161

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

Tyler-Cross, R.; Schirch, V. J. Biol. Chem. 1991, 266, 22549-22556.

Clarke, S. Int. J. Peptide Protein Res. 1987, 30, 808-821.

George-Nascimento, C.; Lowenson, J .; Borissenko, M.; Calderon, M.; Medina-
Selby, A.; Kuo, J .; Clarke, S.; Randolph, A. Biochemistry 1990, 29, 9584-9591.
Potter, S. M.; Henzel, W. J .; Aswad, D. W. Protein Sci. 1993, 2, 1648-1663.
Shahrokh, Z.; Eberlein, G.; Buckley, D.; Paranandi, M. V.; Aswad, D. W.; Stratton,
P.; Mischak, R.; Wang, Y. J. Pharmaceut. Res. 1994, 11, 936-944.

Patel, K.; Borchardt, R. T. Pharmaceut. Res. 1990, 7, 703-711.

Liu, D. T. Trends Biotechnol. 1992, I 0, 364-369.

Aswad, D. W. In Deamidation and Isoaspartate F onnation in Peptides and Proteins;
Aswad, D. W, Ed.; CRC Press: Boca Raton, 1995; pp 1-6.

Daumy, G. 0.; Wilder, C. L.; Merenda, J. M.; McColl, A. S.; Georghegan, K. F.;
Ottemess, I. G. FEBS Lett. 1991, 278, 98-102.

Friedman, A. R.; Ichhpurani, A. K.; Brown, D. M.; Hillman, R. M.; Krabill, L. F.;
Martin, R. A.; Aurcher-Neely, H. A.; Guido, D. M. Int. J. Peptide Protein Res.
1991, 37, 14-20.

Johnson, B. A.; Langmack, E. L.; Aswad, D. W. J. Biol. Chem. 1987, 262, 12283-
12287.

Sun, A.-Q.; Yiiksel, K. 13.; Gracy, R. W. Arch. Biochem. Biophys. 1992, 293, 382-
390.

Zale, S. E.; Klibanov, A. M. Biochemistry, 1986, 25, 5432-5444.

Artigues, A.; Birkett, A.; Schirch, V. J. Biol. Chem. 1990, 265, 4853-4858.

162

37.

38.

39.

40.

41.

42.

43.

45.

46.

47.

48.

49.

Di Donato, A.; Ciardiello, M. A.; de Nigris, M.; Piccoli, R.; Mazzarella, L.;
D’Alessio, G. J. Biol. Chem. 1993, 268, 4745-4751.

Cacia, J.; Quan, C. P.; Vasser, M.; Sliwkowski, M. B.; Frenz, J. J. Chromatogr.
1993, 634, 229-239.

Purcell, A. W.; Aguilar, M. I.; Hearn, M. T. W. J. Chromatogr. 1989, 476, 113-123.
Stevenson, C. L.; Anderegg, R. J.; Borchardt, R. T. J. Pharm. Biomed. Anal. 1993,
l I, 367-373.

Lide, D. R., Ed. CRC Handbook of Chemistry and Physics, 76th Edition; CRC Press:
Boca Raton, 1995; p 7-1.

Kossiakoff, A. A. Science 1988, 240, 191-194.

Galletti, P.; Jardino, P.; Ingrosso, D.; Manna, C.; Zappia, V. Int. J. Peptide Protein
Res. 1989, 33, 397-402.

Blodgett, J. K.; Loudon, G. M.; Collins, K. D. J. Am. Chem. Soc. 1985, 107, 4305-
4313.

Smyth, D. G.; Stein, W. H.; Moore, S. J. Biol. Chem. 1963, 238, 227-234.
Lowenson, J. D.; Clarke, S. In Deamidation and Isoaspartate Formation in Peptides
and Proteins; Aswad, D. W, Ed.; CRC Press: Boca Raton, 1995; pp 47-64.

Chazin, W. J.; Kordel, J.; Thulin, E.; Hofmann, T.; Drakenberg, T.; Forsen, S.
Biochemistry 1989, 28, 8646-8653.

Di Donato, A.; Galletti, P.; D’Alessio, G. Biochemistry, 1986, 25, 8361-8368.
Lehrman, S. R.; Hamlin, D. M.; Lund, M. E.; Walker, G. A. J. Protein Chem. 1992,

l I, 657-663.

163

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

Aswad, D. W.; Guzzetta, A. W. In Deamidation and Isoaspartate Formation in
Peptides and Proteins; Aswad, D. W, Ed.; CRC Press: Boca Raton, 1995; p 18.
Hassell, A. M.; Johanson, K. 0.; Goodhart, P.; Young, P. R.; Holskin, B. P.; Carr, S.
A.; Roberts, G. D.; Simon, P. L.; Chen, M.-J.; Lewis, M. J. Biol. Chem. 1989, 264,
4948-4952.

Papayannopoulos, I. A.; Biemann, K. Peptide Research 1992, 5, 83-90.

Carr, S. A.; Bean, M. F.; Hemling, M. E.; Roberts, G. D. In Biological Mass
Spectrometry; Burlingame, A. L., McCloskey J. A., Eds.; Elsevier: Amsterdam,
1990, pp 621-652.

Galletti, P.; Ingrosso, D.; Manna, C.; Sica, F.; Capasso, S.; Pucci, P.; Marino, G.
Eur. J. Biochem. 1988, 177, 233-239.

Yu, W.; Vath, J. E.; Huberty, M. C.; Martin, S. A. Anal. Chem. 1993, 65, 3015-
3023.

Goodlett, D. R. Proceedings of the 45m ASMS Conference on Mass Spectrometry
and Allied Topics, Palm Springs, CA, June 1-5, 1997.

Katta, V.; Chow, D. T. Proceedings of the 45'” ASMS Conference on Mass
Spectrometry and Allied Topics, Palm Springs, CA, June 1-5, 1997 .

Tsaprailis, G.; Nair, H.; Somogyi, A.; Wysocki, V. H.; Zhong, W.; Futrell, J. H.;
Summerfield, S. G.; Gaskell, S. J. Proceedings of the 46th ASMS Conference on
Mass Spectrometry and Allied Topics, Orlando, FL, May 31-June 4, 1998.

Qin, J.; Chait, B. T. J. Am. Chem. Soc. 1995, 117, 5411—5412.

Liao, P.-C.; Huang, Z.-H.; Allison, J. J. Am. Soc. Mass Spectrom. 1997, 8, 501-509.

Okada, K.; Kawase, M. Chem. Pharm. Bull. 1977, 25, 1497-1508.

164

62.

63.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

Nagasawa, H. T.; Magnan, S. D. J.; Foltz, R. L. Biomed Mass Spectrom., 1982, 9,
252-256.

Ienaga, K.; Nakamura, K.; Higashiura, K.; Toyomaki, Y.; Kimura, H. Chem. Pharm.
Bull. 1988, 36, 2796-2801.

Nakamura, K.; Higashiura, K.; Ienaga, K. Chem. Pharm. Bull. 1989, 37, 73-76.
Qin, X.-Z. J. Mass Spectrom. 1997, 33, 55-63.

Lloyd, J. R.; Cotter, M. L.; Ohori, D.; Doyle, D. L. Biomed. Env. Mass Spectrom.
1988, 15, 399-402.

Violand, B. N .; Schlittler, M. R.; Toren, P. C.; Siegel, N. R. J. Protein Chem. 1990,
9, 109-117.

Carr, S. A.; Hemling, M. E.; Bean, M. F.; Roberts, G. D. Anal. Chem. 1991, 63,
2802-2824.

Bean, M. F.; Carr, S. A. Proceedings of the 40’” ASMS Conference on Mass
Spectrometry and Allied Topics, Washington, DC, May 31-June 5, 1992.

Schindler, P.; Muller, D.; Méirki, W.; Grossenbacher, H.; Richter, W. J. J. Mass
Spectrom. 1996, 24, 967-974.

McFadden, P. N.; Clarke, S. J. Biol. Chem. 1986, 25, 11503-11511.

Huang, Z.—H.; Wu, J.; Roth, K. D. W.; Yang, Y.; Gage, D. A.; Watson, J. T. Anal.
Chem. 1997, 69, 137-144.

Guettler, R. D.; Jones, G. C.; Posey, L. A.; Kirchner, N. J.; Keller, B. A.; Zare, R. N.
J. Chem. Phys. 1994, 101, 3763-3771.

Cordero, M. M.; Houser, J. J.; Wesdemiotis, C. Anal. Chem. 1993, 65, 1594-1601.

165

75.

76.

77.

78.

79.

80.

81.

Summerfield, S. G.; Whiting, A.; Gaskell, S. J. Int. J. Mass Spectrom. Ion Processes
1997, 162, 149-161.

Schnier, P. D.; Price, W. D.; Jockusch, R. A.; Williams, E. R. J. Am. Chem. Soc.
1996, 118, 7178-7189.

Graves, D. J.; Luo, S. J. Agric. Food Chem. 1987, 35, 439—442.

Roth, K. D. W.; Huang, Z.-H.; Sadagopan, N.; Watson, J. T. Mass Spectrom. Rev.,
1998, 17, in press.

Papayannopoulos,1. A. Mass Spectrom. Rev. 1995, 14, 49-73.

Summerfield, S. G.; Cox, K. A.; Gaskell, S. J. J. Am. Soc. Mass Spectrom. 1997, 8,
25-31.

Downard, K. M.; Biemann, K. J. Am. Soc. Mass Spectrom. 1993, 4, 874-881.

166

 

"Illlllllllllllllllllllli