THE INFLUENCE OF CATION AND “ALTERNATIVE” AMINO ACIDS ON
THE FRAGMENTATION PATHWAYS OF METAL CATIONIZED AND
PROTONATED PEPTIDES
A Thesis by
SILA OCHOLA
B.ED, Egerton University, Nairobi , Kenya, 1998.
Submitted to the Department of Chemistry
and the faculty of the Graduate School of
Wichita State University
in partial fulfillment of
the requirements for the degree of
Master of Science
MAY 2008
©Copyright 2008 by Sila Oduor Ochola
All Rights Reserved
THE INFLUENCE OF CATION AND “ALTERNATIVE” AMINO ACIDS ON
THE FRAGMENTATION PATHWAYS OF METAL CATIONIZED AND
PROTONATED PEPTIDES
I have examined the copy of this thesis for form and content and recommend that it be
accepted in partial fulfillment of the requirements for the Degree of Master of Science,
with a major in Chemistry.
Michael J. Van Stipdonk, Committee chair.
We have read this thesis and recommend its acceptance.
Erach R. Talaty, committee member.
Francis D’Souza, committee member.
Paul Rillema, committee member.
William Parcell, committee member.
iii
ACKNOWLEDGEMENT
I wish to express my sincere and heartfelt gratitude to my faculty advisor Dr.
Michael J. VanStipdonk for his kindness and being able to guide me academically. I have
benefited a lot in working in his research laboratory and my knowledge of peptide
chemistry has really improved.
I am very grateful to Dr. Erach Talaty for his guidance especially with the
reaction mechanisms. He encouraged me to pursue my dreams such as applying for Ph.D
studies in Chemistry. I do sincerely thank him for being a member of my committee. My
sincere gratitude is with Dr. Francis D’souza for his guidance and being a great teacher.
From him I have acquired a great deal of Analytical Chemistry knowledge and most of
all I thank him for being a member of my committee. My sincere thanks to Dr. Paul
Rillema, first for accepting to be a member of my committee. I do thank him for listening
to my requests and for his teaching excellence. I have always benefited from his wise
counsel. I am thankful to Dr William Parcell for kindly accepting to be in my committee
and for his valuable input.
I am very grateful to my research group members for their encouragement and
constructive input in my thesis work. I wish to thank all the faculty and staff of
Chemistry Department.
The work described in this thesis was supported by a grant from the National
Science Foundation (CAREER-0239800) and a Faculty Scholar Award from the Kansas
Biomedical Research Infrastructure Network. Funds for the purchase of the ESIinstrumentation were provided by the Kansas NSF-EPSCoR Program and Wichita State
University.
iv
ABSTRACT
Tandem mass spectrometry and collision-induced dissociation (CID) are the
“workhorse” methods for protein identification in proteomics investigations. Recent
studies have demonstrated significant differences in the CID spectra of Li+, Na+ and
silver cationized peptides, particularly with respect to the preferred product ions. For
example, the former produce primarily (bn+17+Li)+ while the latter preferentially
generates (bn-1+Ag)+ species. To improve our understanding of peptide fragmentation in
general, three separate studies were initiated. The objective of the first study was to
determine the CID patterns for thallium(I) cationized peptides and compare them to those
from Ag, Na, and protonated analogues. The goal was to determine whether thallium,
which represents a monovalent cation of relative hardness that differs from that of the
group I metals, would demonstrate reaction pathways similar to group(I) cations or Ag(I).
CID results show that the tendency to produce (bn+17+Tl)+ or (bn-1+Tl)+ depended
significantly on the peptide sequence. Also, the multi-stage CID of Tl+ cationized
peptides fails in the determination of the peptide sequence.
The second objective of this research was to determine the influence of a 4aminomethylbenzoic acid (4AMBz) residue on the relative intensities of (b3-1+cat)+ and
(b3+17+cat)+ fragment ions using tetrapeptides of the general formula A(4AMBz)AX and
A(4AMBz)GX (where X = G, A, V). For Li+ and Na+ cationized versions of the peptides
there was a significant increase in the intensity of (b3-1+cat)+ for the peptides that contain the
4AMBz residue, and in some cases the complete elimination of the (b3+17+cat)+ pathway.
The influence of the 4AMBz residue is attributed to generation of a highly-conjugated
oxazolinone species as (b3-1+cat)+, which increases the stability of this product relative to the
rival (b3+17+cat)+ ion. This conclusion is supported by dissociation profiles, which suggest
v
that the energetic requirements for generation of (b3-1+cat)+ are significantly lower when the
4AMBz residue is positioned such that it should enhance formation of the conjugated
oxazolinone. The objective of the third study was to determine the effect of the same
residues on the formation of (b3-1+cat)+ products from metal (Li+, Na+ and Ag+)
cationized peptides. The larger amino acids suppress formation of b3+ from protonated
peptides with general sequence AAXG (where X=β-alanine, γ-aminobutyric acid or εaminocaproic acid), presumably due to the prohibitive effect of larger cyclic
intermediates in the “oxazolone” pathway. However, abundant (b3-1+cat)+ products are
generated from metal cationized versions of AAXG. Using a group of deuterium-labeled
and exchanged peptides, we found that formation of (b3-1+cat)+ involves transfer of
either amide or α-carbon position H atoms, and the tendency to transfer the atom from
the α-carbon position increases with the size of the amino acid in position X. To account
for the transfer of the H atom, a mechanism involving formation of a ketene product as
(b3-1+cat)+ is proposed.
vi
TABLE OF CONTENTS
CHAPTER
PAGE
CHAPTER I
INTRODUCTION………………………………………………...1
CHAPTER II
EXPERIMENTAL METHODS
Mass Spectrometry……………………………………………….12
Electrospray Ionization…………………………………………..14
Quadrupole Ion trap Analyzer…………………………………...17
Tandem Mass Spectrometry……………………………………..20
Procedure………………………………………………………...24
CHAPTER III
Collision Induced Dissociation of Tl+-cationized peptides……...28
CHAPTER IV
Influence of a 4-aminomethylbenzoic acid residue on the
competitive fragmentation pathways during CID of metal
cationized peptides……………………………………………….48
CHAPTER V
Formation of (b3-1+cat)+ Ions from Metal-cationized Tetrapeptides
containing β-Alanine, γ-Aminobutyric Acid or ε-Aminocaproic
Acid Residues……………………………………………………70
CHAPTER VI
Conclusion……………………………………………………....97
REFERENCES
………………………………………………….........................100
vii
LIST OF FIGURES
FIGURE
PAGE
Figure 1.1
Biemann nomenclature for fragment ions bn and yn ions………………...3
Figure 1.2
Representative structures of protonated b and y ions……………………3
Figure 1.3
CID spectra of protonated AAXG……………………………………….11
Figure 2.1
The components of mass spectrometer………………………………….13
Figure 2.2
Principle of electrospray ionization……………………………………...16
Figure 2.3
Basic components of quadrupole ion trap analyzer……………………...19
Figure 2.4
Collision induced dissociation of precursor ions………………………...22
Figure 2.5
Principle of Tandem Mass Spectrometry………………………………...23
Figure 2.6
Multiple reaction vessel peptide synthesis apparatus……………………27
Figure 3.1
CID spectra of protonated and metal cationized VAAF…………………35
Figure 3.2
CID spectra of protonated and metal cationized YGGFL……………….36
Figure 3.3
CID spectra of protonated and metal cationized VGVAPG……………..37
Figure 3.4
Multi-stage CID of (LGGFL + Tl)+ product ions………………………..40
Figure 3.5
Multi-stage CID of (FGGLL + Tl)+ product ions………………………..41
Figure 3.6
Multi-stage CID of (FGGGL + Tl)+ product ions………………………..43
Figure 3.7
CID of (AGGFL + Tl)+ product ions…………………………………….44
Figure 4.1
Structures of the group of peptides analysed…………………………….62
Figure 4.2
CID spectra of protonated and metal cationized GGGA………………..63
Figure 4.3
CID spectra of protonated and metal cationized AAAG………………...64
Figure 4.4
CID spectra of protonated and metal cationized A(4AMBz)AG………...65
viii
LIST OF FIGURES (CONTINUED)
FIGURE
PAGE
Figure 4.5
CID spectra of Na+ cationized A(4AMBz)AX and A(4AMBz)GX……..66
Figure 4.6
CID spectra of Na+ cationized AA(4AMBz)G and (4AMBz)AAG……..67
Figure 4.7
CID profile of Na+ cationized AXGV (X= A or 4AMBz)……………….69
Figure 5.1
Sequence/structures of model peptides used to study the influence of βA,
γAbu and Cap on formation of (b3-1+cat)+ from metal-cationized
AAXG……………………………………………………………………86
Figure 5.2
CID (MS/MS) spectra generated from AA(γAbu)G series cationized by:
(a) H+, (b) Li+, (c) Na+ and (d) 107Ag+ and 109Ag+……………………….87
Figure 5.3
Spectra generated by CID (MS/MS) of D+ cationized, fully deuteriumexchanged: (a) AAAG, (b) AA(βA)G, (c) AA(γAbu)G and
(d) AA(Cap)G……………………………………………………………88
Figure 5.4
Spectra generated by CID (MS/MS) of Li+ cationized, fully deuteriumexchanged: (a) AAAG, (b) AA(βA)G, (c) AA(γAbu)G and
(d)AA(Cap)G…………………………………………………………….89
Figure 5.5
Spectra generated by CID (MS/MS) of 109Ag+ cationized, fully deuteriumexchanged: (a) AAAG, (b) AA(βA)G, (c) AA(γAbu)G and
(d) AA(Cap)G……………………………………………………………90
Figure 5.6
CID (MS/MS) spectra of Li+ cationized (a) AA(γAbu)G and
(b)AA(α-d2-γAbu)G……………………………………………………..91
Figure 5.7
CID (MS/MS) spectra of Ag+ cationized (a) AA(γAbu)G and
(b) AA(α-d2-γAbu)G…………………………………………………….92
Figure 5.8
CID (MS3) spectra of (a) (b3-1+Li)+Ag+ and (b) (b3-1+Ag)+ derived from
AA(γAbu)G………………………………………………………………93
ix
LIST OF SCHEMES
SCHEME
PAGE
Scheme 1
Pathway to (b3)+ (1a) and (b3+17+cat)+ (1b) from AAAG through 5membered cyclic intermediates ….............................................................10
Scheme 2
Multi-stage CID scheme of thallium cationized LGGFL………………..40
Scheme 3
Multi-stage CID scheme of thallium cationized FGGLL………………..42
Scheme 4
Multi-stage CID scheme of thallium cationized FGGGL………………..42
Scheme 5
Multi-stage CID scheme of thallium cationized AGGFL………………..45
Scheme 6
Multi-stage CID scheme of sodium cationized AGGFL………………...45
Scheme 7
Multi-stage CID scheme of protonated AGGFL………………………...46
Scheme 8
Multi-stage CID scheme of silver cationized AGGFL…………………..47
Scheme 9
Reaction mechanism for formation of (b3)+ from A(4AMBz)AG …........68
Scheme 10
Pathway to (b3-1+cat)+ from metal cationized AA(α-d2-γAbu)G through 7membered cyclic intermediate, with transfer of amide position H atom...94
Scheme 11
Potential pathway to (b3-1+cat)+ from metal cationized AA(α-d2-γAbu)G
through 7-membered cyclic intermediate, with transfer of α-carbon
position H atom………………………………………………………….95
Scheme 12
Pathway to (b3-1+cat)+ from metal cationized AA(α-d2-γAbu)G through
alternative ketene mechanism……………………………………………96
x
LIST OF TABLES
TABLE
PAGE
Table 1
Representative product ion distribution of protonated and metal cationized
YGGFL, VGVAPG, WHWLQL peptides……………………………….38
Table 2
Representative product ion distributions of protonated and metal
cationized GPA, FGG and GHG peptides……………………………….39
xi
CHAPTER I
Introduction
Tandem mass spectrometry combined with soft ionization methods such as
electrospray ionization and matrix assisted laser desorption ionization (MALDI) provides
an effective approach for peptide characterization. The identity of a peptide can be
determined by isolating cationized precursor ions using one mass spectrometry (MS)
stage, ion-activation using methods such as collision induced dissociation (CID) or
surface induced dissociation in a second MS stage, and then analysis of the fragmentation
product ions using a third MS stage. The fragmentation or MS/MS spectrum is used to
deduce the amino acid sequence of the original peptide.
Protonated peptide may dissociate into a wide variety of fragment ions including
the an, and bn ions which are the N-terminal fragments, and yn ions that are C-terminal
fragments [1, 2] (figure 1.1). The reaction pathways leading to bn and yn ions have been
explained qualitatively using the mobile “proton model” [3-16] and more generally by the
“pathways in the competition” model [17]. The main tenet of the mobile proton model is
that protons undergo intramolecular migration from the most basic group on a peptide to
the site of cleavage, where they can weaken the C-N bond and make the carbonyl C atom
more susceptible to intermolecular nucleophilic attack. The recently introduced pathways
in competition fragmentation model provides a more general framework, taking into
account additional features such as proton transfer and peptide isomerization, structures
and transition state energies in the dissociation step, and the thermodynamics associated
with the separation of ion-molecules into fragmentation products.
1
Based on several experimental and theoretical studies [18-26], it is now generally
agreed upon that formation of b-and y ions involves the attack of the carbonyl carbon by
the carbonyl oxygen of the amino acid on the N-terminal side, thus forming a fivemembered ring intermediate. C-N bond cleavage without proton transfer results in the b
ion formation, while one additional proton transfer step leads to generation of a y ion.
According to this mechanism b-ions have a cyclic structure whereas y ions are linear
(figure 1.2). Because it is difficult to predict the type of fragment ions that will be formed
for a given peptide, and identification of products that contain the N or C terminus is
difficult without a priori knowledge of the sequence of the peptides, de novo sequencing
of protonated peptides based on interpretation of fragmentation patterns alone has been a
challenge.
CID of alkali metal-cationized peptides has been studied as an alternative
approach to peptide sequencing. There have been several studies reported in the literature
on the CID of alkali metal-cationized peptides [27-37]. The dissociation of (M+cat)+ ions,
where M=peptide and cat= alkali metal, is known to reveal the metal binding sites and
also produce specific fragmentations that are different from those of (M+H)+ ions. The
major difference between the CID of alkali metal-cationized peptides and protonated
peptides is that the former predominantly form (bn+17+cat)+ while the latter forms bn and
yn ions (figure 1.2).
2
b1
a1
H2N
O
CH
b2
a2
C
NH
C
CH
CH
NH
b3
O
NH
C
R
O
CH
C
OH
R
R
x3
x2
y3
Figure 1.1
R
O
a3
x1
y2
y1
Biemann nomenclature for fragment ions where bn type ions contain
N-terminus and yn type ions contain the C-terminus.
R2
O
O
H2N
N
H
R1
N
H
O
R4
+
+
R3
H2N
H
OH
O
bn +
Figure 1.2
O
H
N
R5
yn+
representative structures of protonated b and y ions.
The mechanism for formation of (bn+17+cat)+ involves the transfer of the
hydroxyl group from the C-terminus of the peptide to the adjacent amino acid and
subsequent loss of the residue mass of the C-terminal amino acid leading ultimately to the
product of a new alkali cationized peptide lacking the original C-terminal residue
(scheme1b). The rearrangement reaction can be particularly well suited for sequencing
because the metal ion is retained in the N-terminal fragment ion, which can undergo the
same fragmentation reaction multiple times in a multiple-stage CID experiment.
Formation of (bn-1+cat)+ ion is proposed to take place via a nucleophilic attack on
the carbonyl carbon at the cleavage site by the oxygen atom on the adjacent carbonyl
group to the N-terminal side of the cleavage site( as shown for b3+ from AAAG in
3
scheme 1a). Several previous studies have shown remarkable differences between Ag+
and alkali metal cationized peptides [27, 35]. For example, the most prominent species
observed during the multistage CID of alkali metal cationized leucine enkephalin are the
(bn+17+cat)+ ions. At higher CID stages however, dissociation of the YG (b2+17+cat)+
results in the production of (a2-1+cat)+ species. In the investigation of YGGFL it was
found that (b4-1+Ag)+ was favored over both the (b4+17+Ag)+ and (a4-1+Ag)+ ions at
activation amplitude ranging from 20-30% of the excitation voltages accessible on the
LCQ-DecaTM and an activation time of 30ms.
The multiple-stage CID of Ag+ cationized leucine enkephalin can be initiated with
either the (bn-1+Ag)+ or (bn+17+Ag)+ produced at the MS/MS stage. Multiple stage CID
of (bn-1+Ag)+ proceeds through sequential elimination of CO to produce (an-1+Ag)+ ions
and then subsequent loss of imine of the C-terminal amino acid to generate the next
(b2-1+Ag)+ species. Similar to the alkali cationized peptides, CID of (b2-1+Ag)+ produces
prominent (a2-1+Ag)+ ions. It has been noted that (b1) ions are not normally observed in
the CID of protonated peptides and even with metal cationized peptides, formation of
(b1-1+cat)+ is not the favored process. The lack of (b1-1+cat)+ is attributed to the fact that
the mechanism for the generation of (bn) products involves the intervention of a five
membered ring structure. This particular mechanism requires cleavage of a carbonyl
moiety to the N-terminal side of the amide bond in the dissociation reaction. In the case
of the (b2-1+cat)+ or (b2+17+cat)+ species, such a carbonyl group is not present and CID
of these precursor ions leads primarily to the production of the cationized imines
(a2-1+cat)+.
4
Recently, studies have shown that CID of Ag+-cationized peptides is more
informative in terms of peptide sequencing than either alkali metal cationized peptides or
protonated peptides [37]. Research studies by Siu and co-workers have demonstrated that
the CID of argentinated peptides results in the production of (bn-1+cat)+ and
(an-1+cat)+ ions which are similar to the CID of protonated peptides, while the production
of (bn+17+cat)+ ions are in line with the CID of alkali metal cationized peptides [38]. The
group of Siu and co-workers have found that the specific advantage of studying the CID
of argentinated peptides is the fact that a prominent triplet of product ions consisting of
(an-1+cat)+, (bn-1+cat)+ and (bn+17+cat)+ ions is consistently observed in the CID
spectrum. This observation leads to direct determination of C-terminal amino acid.
Despite nearly two decades of effort to resolve the mechanisms behind peptide
dissociation, many details remain unclear. For this reason, three specific and focused
studies were designed and carried out. These independent studies are related in that they
are all designed to provide new information, or analytical approaches to generate new
information, about the dissociation of gas-phase protonated and metal-cationized
peptides.
Results from these studies demonstrate that there are significant differences in the
CID spectra of Ag+ and alkali (Li+ and Na+) cationized peptides. The former primarily
form (bn-1+cat)+ and the later forms (bn+17+cat)+ as the major product ions respectively.
For identical peptides, the product ions of (M+Ag)+ are more intense than those of
(M+Na)+, and relative intensities of the (bn+OH+Ag)+ ions are typically lower than those
of the (bn+OH+Na)+ ions. While the differences may be due to differences in hardness of
the cations, testing of this hypothesis is difficult because of the lack of monovalent
5
transition metal ions to compare to Ag+. The objective of the first study was to investigate
the CID patterns of Tl+ cationized peptide, and in particular whether thallium would
demonstrate reaction pathway similar to group(I) cations or Ag+. The first set of
experiments involved CID of protonated and metal cationized peptides followed by the
multi-stage CID of the major product ions.
Apart from the identity of cation (metal or hydrogen), the C-terminal amino acid
can influence the fragmentation pathway. This phenomenon was illustrated by the earlier
studies by Glish and co-workers who investigated the influence of C-terminal residue in
the formation and abundance of (bn+17+cat)+ ions using a quadrupole ion trap mass
spectrometer [34]. After acquiring a large number of MS/MS spectra of alkali cationized
peptides, two general categories of dissociation behavior were determined based on
which of the 20 common amino acids were at the C-terminus. Fifteen amino acids are in
category I and five in category II. Peptides with C-terminated category I amino acids
were found to form primarily the (bn-1+Na +OH)+ product ions. The category I amino
acids include alanine, valine, leucine, isoleucine, methionine, proline, phenylalanine,
tyrosine, lysine, histidine, cystine, arginine, glutamic acid, glycine and aspartic acid.
Although peptides with C-terminated category II amino acids (tryptophan, serine,
threonine, asparagine and glutamine) generally do not produce (bn-1+Na+OH)+ ion as the
most abundant ion following CID, they do provide other alkali cationized analogs, such
as bn-1 and characteristic fragments. These dissociation patterns combined with the
capability of ion trapping instruments to induce multiple CID stages, provide
opportunities for a new method in peptide sequencing.
6
In an attempt to enhance the understanding of how cation and sequence influence
the peptide fragmentation, our group recently investigated the dissociation of metal
cationized, model N-acetylated tetrapeptides, with the general sequence AcFGGX, that
featured C-termini designed to allow transfer of the –OH required to generate the
(b3+17+cat)+ product ion, but not necessarily as the most favored pathway [29]. The
amino acid placed at position X either required a larger cylclic intermediate than the fivemembered ring presumably formed with α-amino acids (β-alanine, γ-aminobutyric acid
and є-amino-n-caproic acid to generate six-, seven- or nine-membered rings,
respectively) or prohibited cyclization because of the inclusion of a rigid ring (para- and
meta-aminobenzoic acid). For Ag+, Li+, and Na+- cationized AcFGGX, formation of
(b3+17+cat)+ was suppressed when the amino acids requiring adoption of larger ring
intermediates were used, while amino acids that prohibit cyclization eliminated the
reaction pathway completely.
To build upon the ealier studies of metal cationized peptides, and to determine the
extent to which changes of the size of the cyclic intermediate may influence the tendency
to form sequence ions, our group recently investigated and reported on the incorporation
of “alternative” amino acids such as β-alanine (βA), γ-aminobutyric acid (γABu), εcaproic acid (Cap), and 4-aminomethylbenzoic acid (4AMBz) into the sequence of model
protonated peptides, and the effect(s) of these residues on relative product ion intensities
[39] (figure 1.3). For protonated peptides, the position of the “alternative” amino acids in
XAAG, AXAG, and AAXG (where X represents the position of the “alternative” amino
acid) had a significant influence on the CID spectrum by inhibiting or completely
suppressing the formation of specific bn+ and yn+ ions. This effect was attributed to the
7
prohibitive effect of forcing cyclization and intramolecular nucleophilic attack to
progress through larger cyclic intermediates, which would be kinetically slower to form
and entropically less favored, when amino acids such as βA, γAbu or Cap were used.
Cyclization was prohibited when the 4AMBz residue was used because the rigid aromatic
ring separates the nucleophile from the electrophilic site of attack.
While investigating the CID of protonated XAAG, AXAG and AAXG peptides
we found that while the 4AMBz residue tended to eliminate specific bn+ and yn+ ions
because the aromatic ring precluded cyclization and intramolecular nucleophililc attack,
it also enhanced the formation of other bn+ ions. For example, increased intensities of b3+
and b2+ ions were observed for the peptides A(4AMBz)AG and (4AMBz)AAG,
respectively, suggesting that 4AMBz enhances formation of b type product ions that arise
via cleavage of the amide bond one sequence position to the C-terminal side of the
residue. The positive effect was explained by formation of a highly conjugated
oxazolinone species, with the aromatic ring as a substituent.
Therefore, the objective of the second study was to extend our earlier experiments
and determine the influence of 4AMBz on the formation of the rival (b3-1+cat)+ and
(b3+17+cat)+ products from protonated and metal cationized tetrapeptides. For this study
peptides with general sequence A(4AMBz)AX and A(4AMBz)GX, where X = G, A, and
V, were synthesized. The CID of protonated and metal cationized (Li+, Na+ and Ag+)
forms of these model peptides was then examined by multiple-stage ion trap tandem mass
spectrometry. The specific hypothesis tested was that the presence and position of the
4AMBz residue would enhance formation of the (b3-1+cat)+ ion, primarily through
8
generation of the stable, highly conjugated oxazolinone, at the expense of the normally
favored (b3+17+cat)+ species.
The objective of the third group of experiments was to determine the effect on
fragmentation patterns of changing the size of the putative cyclic intermediate formed
during the nucleophilic attack. In this third study, the influence of βA, γAbu and Cap on
the tendency to form (b3-1+cat)+ products from Li+, Na+ and Ag+ cationized AAXG was
investigated. The initial hypothesis was that the potential prohibitively large cyclic
intermediates would suppress formation of (b3-1+cat)+, as was observed for the
protonated versions of the peptide. However, the metal cations may coordinate with the
peptide through interactions with multiple amide carbonyl O atoms, and thus kinetically
assist formation of a reactive configuration from which nulceophilic attack occurs. As we
show here, (b3-1+cat)+ is a prominent, if not dominant, reaction product from the metal
cationized AAXG.
9
Scheme 1a
O
H
N
H+
H2N
O
H
N
N
H
O
Scheme 1b
OH
O
H2N
N
H
O
H2N
OH
O
O
H
O
N
H
O
+
N
O
H2N
OH
O
H
N
N
H
O
H
O
+
O
cat+
O
N
N+
H
H
O
O
H2N
cat+
H
H
N+
H
O
OH
O
O
H
N
H2N
O
O
H
N
O
H+
N
H
cat+
O
H
N
N
H
O
Proton transfer
H
N
H
N
H2N
O
O
O
OH
H
N
H2N
H2N
O
cat+
OH
O
N
H
O
O
N
H
N+
CO, NH=CH2
H
O
O
H
N
b3+
H
N
H2N
O
cat+
O
N
H
OH
O
(b3+17+cat)+
Scheme 1
Pathway to (b3)+ (1a) and (b3+17+cat)+ (1b) from AAAG through 5membered cyclic intermediates.
10
100
R. I. (%)
AA(βA)G
(y2)
80
60
40
(b3)
20
0
100
100
125
150
200
225
275
300
AA(γAbu)G
60
40
(y3)
20
0
100
100
125
150
80
R. I. (%)
250
(y2)
80
R. I. (%)
175
(y3)
(y2)
175
200
225
250
275
300
AA(Cap)G
60
(y3)
40
20
0
125
100
R. I. (%)
80
60
150
175
200
225
250
275
300
(y3)
(y2)
325
350
AA(4AMBz)G
(b2)
40
(b3)
20
0
120
140
160
180
200
220
240
260
280
300
m/z
(used with permission from reference 39).
Figure 1.3
CID spectra of protonated AAXG.
11
320
340
360
CHAPTER II
Experimental Methods
Mass Spectrometry
Wilhelm Wien laid the foundation for the development of mass spectrometry in
1898, when he discovered that charged particles could be deflected by a magnetic field.
In experiments conducted between 1907 and 1913, J.J. Thomson studied the deflection of
a beam of positively charged ions using combined electrostatic and magnetic fields. The
two fields were oriented in such a way that the ions were deflected through small angles
in two perpendicular directions: ion trajectories under the influence of the respective
fields produced a series of parabolic curves on photographic plates. Each parabola
corresponded to ions of a particular mass-to-charge ratio and specific position of each ion
depended on its velocity [40].
Modern mass spectrometry is an analytical technique used to measure the massto-charge ratio (m/z) of ions. Mass spectrometry is useful for quantification of atoms and
molecules. Because molecules usually have distinctive fragmentation patterns that
provide structural information, mass spectrometry is also used for determining chemical
and structural information about molecules [41]. In an extension of the work by Wien and
Thomson, m/z values are determined by monitoring signal at a detector as a function of
some applied electric or magnetic field. The applied field causes a separation that is
dependent on m/z value.
The ion separation power of a mass spectrometer is described by its resolving
power, which is defined by the equation:
R = m/∆m
12
where m is the ion mass and ∆m is the difference in mass between two resolvable peaks
in a mass spectrum. A typical mass spectrometer comprises of three parts: an ion source,
a mass analyzer and a detector as shown in figure 2.1.
In this study the focus was on use of Electrospray ionization (ESI) and
Quadrupole Ion Trap (QIT) as ionization and mass analysis methods, respectively.
Ion source
Analyzer
Ion Detector
Inlet system
Sample introduction.
Data system
Figure 2.1
The components of a mass spectrometer.
13
Electrospray Ionisation
Electrospray ionization ( ESI ) is a method used in mass spectrometry that is
especially useful in producing ions from macromolecules because it overcomes the
tendency of these molecules to fragment when ionized. The phenomenon of electrospray
has been known for over hundred years. Chapman in the 1930’s carried out experiments
using electrospray ionization. Some 30 years later (1960), the pioneering work of
Malcolm Dole et al demonstrated the use of electrospray to ionize intact chemical
species. A further 20 years elapsed until work in the laboratory of John Fenn
demonstrated for the first time the use of ESI for the ionization of high mass biologically
important compounds and their subsequent study by mass spectrometry. This work was
to win John Fenn a share of the 2002 Nobel Prize for chemistry. Fenn and co-workers in
1980’s successfully demonstrated the basic principles and methodologies of the ESI
technique, including soft ionization of involatile and thermally labile compounds,
multiple charging of proteins and intact ionization of complexes. ESI-MS is now a basic
tool used in probably every biological chemistry in the world [42, 43].
In electrospray ionization, the analyte is introduced to the source in solution from
syringe pump. Flow rates are characteristically in the order of 1µl/min to 5µl/min. The
liquid is pushed through a very small charged metal, capillary. The potential difference
with respect to the counter electrode is in the range of 2.5 to 5kV. This liquid contains the
substance which is to be studied, the analyte, dissolved in a solvent, which is generally
more volatile than the analyte. Volatile acids or buffers are always added to this solution.
The analyte exists as an ion in solution either in protonated form or as an anion. As like
charges repel, the liquid pushes itself out of the capillary and forms a mist of aerosol of
14
small droplets about 10µm in diameter. This jet of aerosol droplets is sometimes
produced by a process involving the formation of a Taylor cone and a jet from the tip of
this cone. A neutral carrier gas, such as N2 is sometimes used to desolvate the liquid that
is to assist in evaporation of the neutral solvent in the small droplets. As the evaporation
occurs, the droplets shrink until they reach the point that the surface tension can no longer
sustain the charge, this is referred to as the Rayleigh limit. This is the limiting droplet size
at which self-fragmentation occurs with static charge droplets generated in ESI or other
ionization processes. At this limit the proximity of the molecules becomes unstable as the
similarly charged molecules come close together and the droplets once again explode.
This explosion is referred to as coulombic fission because it is the repulsive coulombic
forces between charged analyte that influence it. This process repeats itself until the
analyte is free of solvent and is a lone ion. The lone ion then continues along to the mass
analyzer of a mass spectrometer. In most mass spectrometers, the ESI process is carried
at atmospheric pressure.
In electrospray process the ions observed are ionized by the addition of a proton
(hydrogen ion) to give [M+H]+ (M= analyte molecule, H= hydrogen ion). In other cases a
metal cation such as sodium ion is used [M+Na]+ or the molecule can be deprotonated
[M–H]-. Also in electrospray, multiply charged ion such as [M+2H]2+ are often observed.
Electrospray ionization is a very soft ionization as very little residual energy is
retained by the analyte upon ionization. This is why ESI-MS is such a significant
technique in studying non-covalent gas-phase interactions. The electrospray process is
capable of transferring liquid phase non-covalent complexes into the gas-phase without
disrupting their non-covalent interactions.
15
Spray needle tip
Rayleigh limit at
this point
++++++
++++++
+++++
Solvent
evaporation
+++
Coulombic
explosion.
Multiply charged
droplet
+ve
Power supply
-ve
+
+
+
Analyte ions
Counter electrodes
Figure 2.2
Principle of electrospray ionization.
16
Quadrupole Ion-trap Analyzer
A quadrupole ion-trap is a sensitive and versatile component of the mass
spectrometer. The quadrupole ion-trap (QIT) was developed by the third Nobel Prize
winning mass spectrometry pioneer, Wolfgang Paul. His work in the early 1950’s led to
the development of the basic parameters of today’s benchtop instruments [44]. However
it took breakthroughs in design at the Finnigan MAT in the 1980’s to make the
quadrupole mass spectrometer, the simple to use instrument it is today.
The quadrupole ion-trap mass analyzer used in this study consists of three
hyperbolic electrodes: the ring electrode and the entrance and exit end-cap electrodes.
Various voltages are applied to these electrodes which results in the formation of a cavity
in which the ions are trapped. Ions produced by electrospray ionization are focused using
an octupole transmission system into the ion-trap. Ions may be channeled into the trap
through the use of pulsing lens or through a combination of rf potentials applied to the
ring electrode. The pulsed transmission of ions into the trap distinguishes ion-traps from
“beam” instruments such as quadrupole where ions continually enter the mass analyzer.
The time during which ions are allowed into the trap termed the “ionization period”, is set
to maximize signal while minimizing space-charge effects. The space-charge effects
arises due to too many ions in the trap which distort the electric fields, leading to
considerably impaired performance. The ion-trap is kept at a pressure due to the presence
of helium. The usual pressure inside the trap is between 10-3 – 10-6 torr. Collision with
helium gas present in the ion trap reduce the kinetic energy of the ions and serve to
quickly focus trajectories toward the center of the ion-trap, enabling the trapping of the
injected ions.
17
Trapped ions are further focused toward the center of the trap through the use of
an oscillating voltage, called the fundamental rf , applied to the ring electrode. An ion
will be stably trapped depending upon the values for the mass (m) and charge of the ion
(e), the radial size of the ion trap (r), the oscillating frequency of the fundamental rf (ω),
and the amplitude of the voltage on the ring electrode (v). The reliance of ion motion on
these parameters is described by the dimensionless parameter qz
qz= 4eV/mr2ω2
It is important to note that quadrupole ion trap is concerned with the criteria that
govern the stability (or instability) of the trajectory of an ion within the field, that is the
experimental conditions that determine whether an ion is stored within the device or is
ejected from the device[45, 46]. Ions can be stored in the trap provided that their
trajectories fall within a region of stability within the quadupolar field as defined by the
Mathieu parameters az and qz. Depending upon the amplitude of the voltage placed on the
ring electrode, an ion of a given m/z will have az and qz values that are within the
boundaries of the stability diagram and the ion will be trapped. If az and qz values at that
voltage fall outside the boundaries of the stability region, the ion will hit the electrode
and will be lost.
18
Fundamental rf.
~
Ring electrode
Endcap electrodes
v
Auxiliary ac.
Figure 2.3
Basic components of quadrupole ion trap analyzer.
19
Tandem Mass Spectrometry
Tandem mass spectrometry MS/MS involves various steps of mass selection or
analysis generally accompanied by some form of fragmentation. Tandem mass
spectrometry can be effected in space by using different mass analyzer. For example one
mass analyzer can isolate the peptide for investigation. A second mass analyzer then
stabilizes the peptide ions while they collide with a neutral gas such as helium, causing
them to fragment. A third mass analyzer then catalogs the fragments produced from the
peptides. Tandem MS can also be done in a single mass analyzer over time as in
quadrupole ion trap. This involves the use of one mass analyzer multiple times to perform
the task of isolation and fragmentation of the precursor and product ions.
There are various methods for fragmenting molecules for tandem MS, including
collision induced dissociation (CID), electron capture dissociation (ECD), electron
transfer dissociation (ETD), Infrared multi photon dissociation (IRMPD) and Black body
infrared radiation (BIRD).
In this study collision induced dissociation was used to investigate the metal
cationized peptides. All CID processes occurring regularly can be separated into one of
two categories based mostly on the translational energy of the precursor ion. For organic
ions of moderate mass (several hundred Daltons), low energy collisions, common in
quadrupole and ion trap instruments occur in the 1-100eV range of collision energy and
high energy collision seen in sector and TOF/TOF instruments are in kiloelectrovolt
range.
In a quadrupole ion trap, the precursor ions are isolated and accelerated by “onresonance” excitation causing collision to occur and product ions are detected by
20
consequent ejection from the trap. In on-resonance excitation, the isolated precursor ion
is excited by applying a small a.c potential across the endcap electrodes, corresponding to
the secular frequency of the ion. Ion activation times the order of tens of milliseconds can
be used without considerable ion loses, therefore collisions can occur during the
excitation period. Because of the time scale, this excitation technique falls in the category
of so called “slow heating” processes. However, excitation in an ion trap is still fairly
fast, due to the high pressure of the helium present in the trap (≈ 10-3 Torr). For low
collision energies, excitation is mostly vibrational since the interaction time coincides
with a bond’s vibration period [47].
Multiple stages of CID can be performed in ion trap instrument, although the
experiment is restricted to the product ion scan. In product ion scan, the mass analyzer is
used to select the user-specified sample ions arising from a particular component, usually
the molecular ion ([M+H]+, [M-H]- or [M + cat]+). These chosen ions collide with neutral
gas molecules in this case helium which cause fragment ions to be formed (figure 2.4).
These product ions are isolated and fragmentation can be performed resulting in MS3
spectrum (figure 2.5). This process can be repeated a number of times, resulting in a
series of MSn spectra where ‘n’ represents the number of times the isolationfragmentation cycle has been carried out. All the fragment ions arise directly from the
precursor ions specified in the experiment, and this produces a fingerprint pattern specific
to the compound under investigation. This is useful for the elucidation of fragmentation
pathways and for the identification of molecular structures of the ions.
21
Ion source
Mass analyzer
MS
I
Detector
Ion source
Mass analyzer
m/z
MS
. ..
..........
.......
I
m/z
Detector
Collision cell
MS/MS
I
m/z
Figure 2.4
Collision induced dissociation of precursor ions.
22
Intensity
[M+X]+
Product ions
Isolate and fragment
intensity
Isolate and fragment
m/z
Figure 2.5
Principle of Tandem mass spectrometry
23
Sample Preparation Procedures
Peptides VAAF, FGGFL, AGGFL, LGGFL,FGGGL,YGGFL,VGVPAG,
FGGAL, GGGA, AAAG, (4AMBz)AAG, A(4AMBz)AX and A(4AMBz)GX used in
this study were generated by solid-phase synthesis methods using Wang resin,
conventional coupling procedures [53] and Fmoc-protected amino acids in a custombuilt, multiple reaction vessel peptide synthesis apparatus(figure 2.6). Metal nitrates
LiNO3, NaNO3, AgNO3, TlNO3 and Glycine (G), alanine (A), and valine (V)-loaded
Wang resins, Fmoc-alanine (Ala), β-alanine [NH2CH2CH2CO2H, βA], γ-aminobutyric
acid [NH2CH2CH2CH2CO2H, γAbu], γ-aminobutyric acid labeled with deuterium at the
α-carbon position [NH2CH2CH2CD2CO2H, α-d2-γAbu], ε-aminocaproic acid
[NH2CH2CH2CH2CH2CH2CO2H, Cap] and 4-aminomethylbenzoic acid
[NH2CH2C6H4CO2H, 4AMBz] were purchased from Sigma-Aldrich (St. Louis, MO) and
used as received. Peptides, once cleaved from the resin using trifluoroacetic acid, were
used without subsequent purification in the CID studies.
Solutions of each peptide were prepared by dissolving the appropriate amount of
solid material in 1:1(v/v) mixture of HPLC grade MeOH and de-ionized H2O. Equimolar
metal nitrate solutions were prepared in deionized H2O to produce the required final
concentration. To investigate the dissociation of AAXG peptides for which exchangeable
H atoms were replaced with D, the peptide was incubated in a mixture (50:50 by volume)
of D2O and CH3OD (Aldrich Chemical, St. Louis MO) overnight prior to analysis.
Solutions for ESI experiments were prepared by combining the peptide and metal
nitrate solutions in 1:1 ratio. ESI mass spectra were collected using a Finnigan LCQDecaTM ion–trap mass spectrometer (Thermoquest Corporation, San Jose, CA, USA)
24
Mixtures (1:1 v/v) of metal nitrate and peptide, prepared by combining respective
stock solutions were infused into the ESI–MS using the incorporated syringe pump at
flow rate of 5µl/min. The atmospheric pressure ionization stack settings for the LCQ
(lens voltages , quadrupole and octapole voltage offsets) were optimized for maximum
(M + cat)+ ion transmission to the ion-trap mass analyzer using the auto-tune routine
within the LCQ tune program. Following the instrument tune, the spray needle voltage
was maintained at +5kV, the N2 sheath gas flow rate at 25 units was used. This is set
arbitrary for the LCQ Deca system. The capillary temperature for this experiment was set
at 2200c. This served as desolvation temperature to convert ions from solution phase to
gas phase. The ion-trap analyzer was operated at ≈ 1x 10-6 Torr. Helium gas, admitted
directly into the ion-trap, was used as the bath/buffer gas to improve the trapping
efficiency and as the collision gas for CID experiments. The metal cationized peptides
were isolated using a width of 1.5 m/z units. To induce collision activation, the activation
amplitude for CID was set between 10 to 30% of 5V. The activation amplitude defines
the amplitude of the RF energy applied to the end caps electrodes to effect dissociation.
The Q setting used to adjust the frequency of the RF excitation voltage was set at 0.3.
The activation time employed at each CID stage was set at 30ms. Spectra were generated
by plotting relative intensity to the most abundant ion as a function of m/z.
Energy resolved CID profiles generated by scanning the activation amplitude
were corrected using the following equations:
(AA/30) x ( 0.4 + ( 0.002 x Precursor ion mass )
CAA= (AA/30) x ( 0.4 + ( 0.002 x Precursor ion mass )/(3N-6).
25
Where CAA is the corrected activation amplitude, AA is the normalized collision energy
and N is the degrees of freedom.
The corrected activation amplitude scale was arbitrarily chosen to facilitate the
comparison and measurement of the relative product ion abundances with minimal bias
that might originate due to the differences in precursor ion mass of peptides examined
here. This allowed a comparison of the relative collision energies/voltages necessary to
induce formation of (b3-1+cat)+ and (b3+17+cat)+ for peptides with and without the
4AMBz residue.
For ion isolation experiments to probe D for H back exchange, precursor ions
were isolated using an isolation width of 0.9 – 1.1 m/z units, which was sufficient to
isolate a single isotopic peak. The normalized collision energy was set at 0% (i.e. no
imposed collisional activation), the activation Q at 0.30, and the isolation time altered
from 10 msec to 1 second. During the isolation period the ions may collide with H2O,
which is present as a contaminant in the ion trap and because of its use as solvent for ESI.
The check for D for H back exchange was done because such a process could effect the
splitting of (b3-1+cat)+ products into distinct isotopic peaks (vide infra), which was the
principal feature used to identify and measure the extent of transfer of α-carbon position
H or D atoms during fragmentation reactions.
26
Figure 2.6
Multiple reaction vessel peptide synthesis apparatus.
27
CHAPTER III
Collision Induced Dissociation of Tl+-cationized peptides
Introduction
As noted in the introductory chapter, tandem mass spectrometry of cationized
peptides and proteins in the gas phase is commonly used to determine the sequence of
peptides and identity of the proteins. This information is dependent on the number and
identities of the cationizing species and their interactions with the peptide or protein. The
fragmentation behavior of protonated peptides has been extensively examined. However,
complementary fragmentation and or specific sequence information may be obtained by
performing collision induced dissociation experiments on metal cationized peptides. For
example, collision induced dissociation (CID) of sodiated peptide leads to fragmentation
adjacent to the C-terminal residue, while protonated peptide usually fragments at various
places along the peptide backbone [27].
Recent studies have demonstrated significant differences in the CID spectra of
protonated and metal cationized peptides [27-38]. An important distinction between the
CID spectra generated from protonated peptides and those derived from metal cationized
peptides is the preferred formation in the latter of an N-terminus containing (b-type)
rearrangement ion labeled (bn+17+cat)+ or (bn+OH+cat)+. While for the protonated
peptides, sequence specific b- and y- ions are often the dominant fragment ions in the low
energy CID spectra.
Studies in our laboratory have shown distinct differences between Ag+ and alkali
metal (Li+, Na+) cationized peptides [27], including the fact that the most prominent
species observed during the CID of alkali metal(Li+, Na+) cationized leucine enkephalin
28
are the (bn+17+cat)+ ions, while Ag+ cationized peptides favor formation of (bn-1+Ag)+
over both the (bn+17+Ag)+ and (an-1+Ag)+ ions. Similar to the alkali cationized peptides,
multi-stage CID of (b2-1+Ag)+ produces prominent (a2-1+Ag)+ ions.
While the differences between Ag+ and alkali-metal (Li+, Na+) ions may be due,
in part to differences in polarizing power and hardness of the cations, testing of this
hypothesis is difficult because of the lack of monovalent transition metal ions to compare
to Ag(I).
In this set of experiments the multi-stage CID of Tl+ cationized peptides was
investigated and compared to those of Ag+, Na+, and protonated analogues. The main
goal was to determine whether Tl+ cation would demonstrate reaction pathway similar to
alkali metal cations or Ag+.
Result and Discussion
CID of protonated and metal cationized VAAF, YGGFL and VGVAPG
For the sake of illustration and comparison, figure 3.1 shows the CID spectra of
protonated, sodiated, argentinated and thallium cationized VAAF. The major peaks
observed in spectra of protonated VAAF were b3+ (242) and y2+ (237) product ions. The
difference in m/z values between the (M+H)+ and b3+ is 165u (407-242) thus identifying
phenylalanine (147u) and H2O (18u). The ∆m/z value resulting in the formation of y2+
(237) was 170u which is consistent with the loss of N-terminus residues valine (99u) and
alanine (71u). For Na+ cationized VAAF, the peak with the largest m/z value was
(b3+17+Na)+ (282u). The amino acid cleaved from the C-terminus was identified as
phenylalanine (147u), since (M+Na)+ had an m/z value of 429u. The Ag+ cationized
29
version generated (b3-1+Ag)+ (348,350) as the major product ion more than (a3-1 + Ag)+
(320,322) and (b3 + 17 + Ag)+ (366,368) species. Silver ion has two stable isotopes 107Ag
and 109Ag of almost equal abundance and the product ion spectra shown are those of the
two isotopes. In this survey, the most prominent product ion resulting from the CID of
Tl+ cationized VAAF was (b3+17+Tl)+ having a mass unit of 462u and 464u due to two
isotopes of thallium (205Tl and 207Tl). Other product ions observed were, (b3-1+Tl)+ (444,
446) and (b2-1+Tl)+ (373, 375). One striking observation is that, among the metal
cationized peptide (VAAF) investigated Tl+ cationized peptide generated greater
abundance of (b2-1+Cat)+ ions.
The most prominent product ion generated from the CID of protonated YGGFL
figure 3.2, was b4+ (425u) ion due to the neutral loss of leucine (113). The other product
ion observed was a4+ (397u). In contrast to CID of protonated peptide, CID of Na+
cationized YGGFL generated (b4+17+Na)+ in abundance. Other minor products obtained
were (b4-1+Na)+ and (a4-1+Na)+ ions. The remarkable differences between the protonated
peptides and Na+ cationized versions can be attributed to the difference in the binding
sites within the peptide. The dissociation of alkali cationized peptides is proposed to
involve the binding of the alkali metal ion to amide carbonyl atoms, rather than amide
nitrogen atoms [38].
In accord with earlier studies, triplet of product ions, consisting of (b4-1+Ag)+,
(a4-1+Ag)+ and (b4+17+Ag)+ were observed in the CID spectra of Ag+ cationized
YGGFL. It is interesting that the yield of (b4+17+Ag)+ ion from YGGFL was
approximately 35%(figure 3.2c), significantly higher than the corresponding ions
30
(b3+17+Ag)+, generated from VAAF (figure 3.1c). This may be due to the beneficial
interaction between the metal ion and the aromatic ring at the second amino acid, in
combination with the size of the peptide, to make the C-terminal rearrangement reaction
favorable.
Similar to Na+ cationized versions, the most prominent product ion resulting from
the CID of Tl+ cationized leucine enkephalin was (b4+17+Tl)+ ion (figure 3.2d). This is in
contrast to the major product ion observed in the CID of (Tl + VGVAPG)+ (figure 3.3d),
in which (b4-1+Tl)+ had the highest peak intensity. This trend in which the major product
ion was either (bn+17+Tl)+ or (bn-1+Tl)+ was also observed with other Tl+ cationized
peptides investigated in this study.
Figure 3.3 shows the CID spectra of VGVAPG in (a) protonated, (b) Na+, (c) Ag+
and (d) Tl+ cationized forms. For this peptide, each cation demonstrated a unique pattern
of fragmentation. One interesting observation was that there was high abundance of
(a4-1+Tl)+ compared to (a4-1+Ag)+ among the two metal cationized peptides. The spectra
shown in figures 3.1-3.3 are representative, of the spectra generated from majority of the
peptides included in this study.
The study revealed that unlike the Ag+ cationized versions, CID of the alkali
cationized (Li+, Na+) versions of the peptides did not generate prominent (bn-1+cat)+ ions.
The different behaviors of Ag+ and Na+ metal ions are strikingly consistent with their
differences in hardness as lewis acids, Na+ prefers to interact with the harder O (oxygen)
donors, whereas Ag+ prefers to bind to softer N (nitrogen) donors. It is interesting then,
that generation of (bn-1+cat)+ which involves transfer of proton ultimately to an N- atom
is the preferred pathway for Ag+ cationized peptides. The loss of H2O, which involves
31
transfer of a proton instead to an O- atom is the preferred pathway for the alkali
cationized (Li+, Na+) peptides.
The CID of Tl+ cationized peptides (MS/MS) resulted in the formation of either
(bn+17+Tl)+ or (bn-1+Tl)+ (figures 3.1-3.3) product ions . It was found that the tendency
to produce (bn-1+Tl)+ versus (bn+17+Tl)+ products depended significantly on peptide
sequence (Tables 1 and 2). With several peptides, the product ion distribution generated
from Tl+ cationized species were most similar to those generated from group I metals,
while for others the distribution was more similar to those observed for protonated and
Ag+ cationized peptides. These observations suggest that a straight forward relationship
between cation hardness and reaction pathway, will be difficult to identify.
Multiple-stage CID of Tl+ cationized product ions
MSn experiments were also conducted to elucidate the fragmentation pathways for
the dissociation of (bn+17+Tl)+ or (bn-1+Tl)+ product ions, (figures 3.4-3.7). The MS3
product ion spectra (3.4(b)-3.7(b)) were generated by isolating and dissociating the most
abundant peak in the MS/MS spectra (3.4(a)-3.7(a)). Subsequent MSn spectra were then
generated by isolating and dissociating the most abundant peaks resulting from the prior
CID stage (MSn-1). The collision activation of (b4+17+Tl)+ (597) from (LGGFL+Tl)+
generated a new product ion (b3+17+Tl)+ (450) via the loss of 147u which is consistent
with the neutral loss of F (phenylalanine) from the C-terminus. Subsequent CID of the
(b3+17+Tl)+ (450) product ion (MS4) resulted in the loss of 75u and 245u consistent with
the loss of (glycine+ H2O) and elimination of the metal ion to form (b2-1+Tl)+ (374.9)
and Tl+ (205) ions as the major product ions which is in contrast to the behavior of Ag+ or
32
alkali cationized versions. This reaction pattern was also observed with the multi stage
CID of (FGGLL+Tl)+ peptides (figure 3.5). As illustrated in figure 3.6, the dissociation
pattern of (FGGGL+Tl)+ was unique in that the CID of (b4+17+Tl)+ having mass unit of
541 resulted in the production of (b3-1+Tl)+ (466) and Tl+ (205) as the major product
ions. In most of the peptides investigated, the multi stage CID of the thallium product
ions resulted in the high yield of the bare metal cation (Tl+) signifying a weak binding
between the metal and the peptide.
Inspection of the data showed a striking influence of the sequence of the peptide
on the relative intensities of the (bn+17+cat)+ and (bn-1+cat)+ product ions. For instance,
at the same corrected activation amplitude,(spectra not shown), CID of Ag+ cationized
GGG and GGA produced (bn+17+Ag)+ product intensities of 8% and 45% respectively,
indicating an increase in intensity as the side group of the C-terminal amino acid become
bulky. This is in accord with earlier studies by Glish and coworkers. However, in contrast
to their studies, the influence of the N-terminal amino acid was also observed. For
example, FGG and YGG produced higher relative intensities of (bn+17+cat)+ and
(bn-1+cat)+ product ions than GGG. It is apparent that the N-terminal amino acid lies on
neither the five-membered ring during the C-terminal rearrangement nor that formed
during the N-terminal rearrangement, and thus is not expected to play a significant role in
the stability of the rings or influence the probability of their formation. In this case the
bulky group at the N-terminus influences the binding of the metal ion, and in particular
increases the probability that the metal ion can migrate to the cleavage site to polarize the
carbonyl bond. Indeed, the yield of the (b2+17+cat)+ and (b2 -1+cat)+ product
33
ions are significantly higher for FGG and YGG, with the aromatic side groups capable of
coordinating the metal ion and altering the metal ion binding position on the peptide.
Summary
In this study, the CID results showed that the tendency to produce (bn+17+Tl)+ or
(bn-1+Tl)+ depended significantly on the peptide sequence. The survey demonstrated that
MSn of the protonated versions of the peptides failed to give the accurate sequence
information. The sequencing failed because of the tendency to eliminate amino acids
from either the N- or C-terminus, leading to an ambiguity in sequence determination. It
was observed that multistage CID (MSn) of Ag+ and alkali metal cationized peptides can
lead to unambiguous determination of the sequence of peptides (Scheme 6 and 8),
whereas Tl+ cationized versions are not preferable. This is because at higher CID stages
of the latter, the reaction channel leading to the elimination of the metal cation is more
preferred (scheme 2-5) than the formation of (bn) or (an) sequence ions which is observed
with former species. One interesting observation was that among the metal cations
investigated, Tl+ cationized peptides tended to produce the highest yield of (yn-1+cat)+
species. The appearance of the prominent (yn-1+cat)+ ion may be attributable to the metal
ion affinity for this product ion and is being explored further in our research work.
34
100
R. I. (%)
75
(a)
237
y2+
(H+)
VAAF
242
b3+
50
b2+
171
y1+ 166
25
0
125
150
y3+
308
197
175
200
225
250
275
300
-H 2O
389
361
325
350
375
400
425
100
25
0
125
150
175
200
264
236
193 211
225
282
(b3+17+Na)+
(b3-1+Na)+
50
(a3-1+Na)+
(b2-1+Na)+
R. I. (%)
(b2+17+Na)+
(b) (Na+)
75
250
275
300
325
350
375
400
425
450
100
R. I. (%)
75
(c) (Ag+)
(b2-1+Ag)+
50
(a3-1+Ag)+
320,322
(a2-1+Ag)+
249,251
0
100
125
150
175
200
225
250
275
300
100
R. I. (%)
(b3+17+Ag)+
366,368
277,279
25
75
(b3-1+Ag)+
348,350
(d) (Tl+ )
(b2-1+Tl)+
373,375
325
350
(b3-1+Tl)+
444,446
375
400
425
450
475
500
525
600
625
462,464
(b3+17+Tl)+
50
25
0
175
200
225
250
275
300
325
350
375
400
425
450
475
500
525
550
575
m /z
Figure 3.1
CID spectra of protonated and metal cationized VAAF
35
100
(a)
R. I. (%)
75
YGGFL
425 b4+
(H+)
a4+
397
50
538
y3+
336
25
510
0
225
275
300
325
350
375
400
425
450
475
50
(b3+17+Na)+
318
25
(b4-1+Na)+
(Na+)
500
525
550
575
600
550
575
600
625
650
675
775
800
465
(b4+17+Na)+
(b)
75
R. I. (%)
250
(a4-1+Na)+
100
419 447
0
225
100
R. I. (%)
75
250
275
300
325
350
375
400
(c)
425
450
475
500
525
531,533
(b4-1+Ag)+
(Ag+)
50
503,505
(a4-1+Ag)+
(b3-1+Ag)+
384,386
25
549,551
(b4+17+Ag)+
0
350
375
400
425
450
475
(d)
(Tl+)
50
(b3-1+Tl)+
480,482
25
500
(a4-1+Tl)+
R. I. (%)
75
325
(b3+17+Tl)+
498,500
525
(b4-1+Tl)+
300
100
550
575
600
645,647
(b4+17+Tl)+
627,
599,
629
601
0
400
425
450
475
500
525
550
575
600
625
650
675
700
725
m /z
Figure 3.2
CID spectra of protonated and metal cationized YGGFL.
36
750
100
327
b4+
(a)
R. I. (%)
75
(H+)
a4+
299
50
a3+
228
211
173
25
VGVAPG
b3+
256 282
b5+
424
-H2O
481
0
250
(a4-1+Na)+
100
(b)
R. I. (%)
75
(Na+)
50
275
321
25
300
325
350
375
400
425
349 367
450
475
500
525
(b5+17+Na)+
225
(b5-1+Na)+
200
(b4+17+Na)+
175
(b4-1+Na)+
150
464
-H2O
503
446
0
275
300
325
350
375
(Ag+)
50
25
405,407
225
250
275
300
325
350
375
400
500
525
550
(M+Ag)+
433,435
605,607
(a3-1+Tl)+
50
Tl+
203,205
450
475
500
501,503 529,531
(b3-1+Tl)+
(Tl+)
425
(a4-1+Tl)+
(d)
R. I. ( %)
475
458,
430, 460
432
525
550
(b5+17+Tl)+
200
100
25
450
362,364
0
75
425
(b4-1+Tl)+
R. I. (%)
(b3-1+Ag)+
(c)
75
400
(b4-1+Ag)+
250
(a4-1+Ag)+
225
100
575
600
625
-H2O
683,685
644,
646
0
200 225 250 275 300 325 350 375 400 425 450 475 500 525 550 575 600 625 650 675 700 725
m /z
Figure 3.3
CID spectra of protonated and metal cationized
VGVAPG.
37
YGGFL
Product
ions
(% R.I)
M-H2O
b4+17
b4
b3+17
b3
Tl
8
100
34
2
4
H
22
0
100
0
2
Ag
1
27
100
0
2
Na
0
100
6
0
0
Product
ions
(% R.I)
M-H2O
b5+17
b5
b4+17
b4
Tl
56
19
12
11
100
H
2
0
3
0
100
Ag
5
0
6
0
100
Na
12
100
14
18
30
WHWLQL
Product
ions
(% R.I)
M-H2O
b5+17
b5
b4
y4
Tl
21
100
61
17
14
H
42
0
100
73
36
Ag
7
8
100
26
2
Na
8
100
16
2
3
Table 1
representative product ion distributions.
VGVAPG
38
GPA
Product
ions
(% R.I)
M-H2O
b2+17
b2
a2
y1
Tl
100
15
47
0
39
H
18
0
100
8
0
Ag
55
83
30
8
74
Na
100
57
21
0
5
FGG
Product
ions
(% R.I)
M-H2O
b2+17
b2
a2
a1
Tl
21
7
34
0
79
H
1
0
100
0
0
Ag
9
21
100
37
0
Na
60
52
100
0
0
GHG
Product
ions
(% R.I)
M-H2O
b2+17
b2
b1
-35u
Tl
100
0
34
0
0
H
83
0
81
26
100
Ag
2
0
100
0
0
Na
56
0
100
0
0
Table 2
representative product ion distributions of protonated and metal
cationized GPA, FGG and GHG peptides.
39
100
597
(b4+17+Tl)+
(a)
80
(b4-1+Tl)+
60
579
40
(a4
CID of 710
-1+Tl)+
(-H2O)
692
551
20
0
(%)
0
100
Relative intensity
100
200
300
400
500
600
700
800
(b3+17+Tl)+
450.1
(b)
80
CID of 710_597
60
40
Tl+
20
205
(b3-1+Tl)+
432
(b4+17+Tl)+
596.8
0
0
100
100
300
400
(Tl+)
205
(c)
80
200
500
600
700
800
(b2-1+Tl)+
374.9
CID of 710_597_450
60
(b3+17+Tl)+
450
40
20
0
0
100
200
300
400
500
600
700
800
m/z
Figure 3.4
Multi-stage CID of (LGGFL+Tl)+ product ions.
(M+Tl)+
710
692
MS/MS
579
-113 (L)
(b4-1+Tl)+
597(b4+17+Tl)+
MS3
-147 (F)
450.1(b3+17+Tl)+
MS4
-75 (G+H2O)
205(Tl)+
Scheme 2
374.9(b2-1+Tl)+
Multi-stage CID scheme of thallium cationized LGGFL.
40
100
(b4-1+Tl)+
(b4+17+Tl)+
597
80
60
(an-1+Tl)+ 579
551
(a)
40
CID of 710
20
0
Relative intensity (%)
0
100
200
300
400
500
600
700
800
(b3+17+Tl)+
484
100
80
60
CID of 597
(b)
Tl+
40
205
20
0
0
100
200
Tl+
205
100
80
60
300
400
500
600
700
800
(b2-1+Tl)+
324
409
CID of 484
(b3+17+Tl)+
466
(c)
40
20
0
0
100
200
300
400
500
600
700
m/z
Figure 3.6
Multi-stage CID of (FGGLL+Tl)+ product ions.
41
800
(M+Tl)+
710
551
(a4-1+Tl)+
MS/MS
579 (b4-1+Tl)+
-133(L)
597 (b4+17+Tl)+
MS3
-113 (L)
205 (Tl)+
484 (b3+17+Tl)+
MS4
409 (b2-1+Tl)+
205 (Tl)+
Scheme 3
-75 (G+H2O)
MS4
Multi-stage CID scheme of thallium cationized FGGLL.
(M+Tl)+
654
MS/MS
523
(b4-1+Tl)+
-113 (L)
541 (b4+17+Tl)+
522.9 (b4-1+Tl)+
MS3
-75 (G+H2O)
205 (Tl)+
466
323.9
205 (Tl)+
Scheme 4
Multi-stage CID scheme of thallium cationized FGGGL
42
(b4+17+Tl)+
100
80
541
(a)
CID 654
60
(b4-1+Tl)+
40
523
20
0
0
Relative intensity (%)
100
100
200
300
Tl+
205
(b)
400
600
700
800
(b3-1+Tl)+
80
466
324
60
500
(b4-1+Tl)+
522.9
CID 654_541
40
20
0
0
100
200
300
400
500
600
700
800
323.9
100
(c)
80
CID 654_541_466
60
40
Tl+
437.8
20
205
0
0
100
200
300
400
500
600
700
m/z
Figure 3.6
Multi-stage CID of (FGGGL+Tl)+ product ions.
43
800
(b4+17+Tl)+
100
(a)
80
(b4-1+Tl)+
60
555
537
CID 668
40
20
668
0
0
100
200
300
400
100
80
600
700
800
408
(b)
CID 668_555
60
40
Tl+
20
537
205
0
0
100
100
80
200
Tl+
205
(c)
300
400
279.9
60
40
500
600
700
800
(b3-1+Tl)+
333
(b2-1+Tl)+
Relative intensity (%)‫‏‬
500
(b3+17+Tl)+
390
CID 668_555_408
20
0
0
100
200
300
400
500
600
m/z
Figure 3.7
Multi-stage CID of (AGGFL+Tl)+ product ions.
44
700
800
(M+Tl)+
668
MS/MS
537 (b4-1+Tl)+
-113 (L)
555 ( b +17+Tl)+
4
MS3
205 (Tl)+
-147 (F)
408 (b3+17+Tl)+
MS4
-18 (H2O)
-75
205 (Tl)+
390 (b3-1+Tl)+
333
(b2-1+Tl)+
Scheme 5
Multi-stage CID scheme of thallium cationized AGGFL.
(M+Na)+
487
MS/MS
-114 (L)
373 (b4+17+Na)+
MS3
226
-147 (F)
(b3+17+Na)+
168.9 (b2+17+Na)+
MS4
-18 (H2O)
150.8
+
(b2-1+Na)
208 (b3-1+Na)+
Scheme 6
Multi-stage CID scheme of sodium cationized AGGFL
45
(M+H)+
464
MS2
-18 (H2O)
446.1
-131
332.9 (b4)+
-28
MS3
FL
305 (a4)+
MS4
-74
231
-28
MS5
203 (AGG+H)+
MS6
-71 (A)
132 (GG+H)+
Scheme 7
Multi-stage CID scheme of protonated AGGFL
46
(M+Ag)+
572
-131
MS/MS
-113 (L)
459 (b4+17+Ag)+
441 (b4-1+Ag)+
MS3
-28
412.9 (a4-1+Ag)+
F
-119
MS4
293.9 (b3+17+Ag)+
MS5
-28
265.9 (a3-1+Ag)+
MS6
G
-29
236.9 (b2-1+Ag)+
Scheme 8
Multi-stage CID scheme of silver cationized AGGFL
47
CHAPTER IV
Influence of a 4-aminomethylbenzoic acid residue on competitive fragmentation
pathways during CID of metal cationized peptides
Introduction
As noted in earlier sections, a prominent, if not the preferred dissociation pathway
for Li+ and Na+ cationized peptides is formation of (bn+17+cat)+ ions [33,34,48]. Except
for a relatively few cases in which peptides contain arginine residues to localize the
added proton [49,50], the protonated analog (bn+17+H)+ is rarely observed. The currently
accepted mechanism for generation of (bn+17+cat)+ ions is cyclization and intramolecular
nucleophilic attack with proton transfer. As in the case of the bn+/(bn-1+cat)+ pathway, the
electrophilic site of attack in the (bn+17+cat)+ pathway is the carbonyl carbon of the
amide group being cleaved. In this case, however, the nucleophile is the carbonyl oxygen
of the C-terminal carboxylic acid [31,35,51,52] (reaction b in scheme 1). For metal
cationized peptides in particular, the (bn-1+cat)+ and (bn+17+cat)+ ions are the products of
competing reaction pathways that involve cleavage of the same amide bond.
Incorporation of “alternative” amino acids such as β-alanine (βA), γ-aminobutyric
acid (γABu), ε-aminocaproic acid (Cap), and 4-aminomethylbenzoic acid (4AMBz) into
the sequence of a model peptide is known to have a substantial effect on relative product
ion intensities [39]. For protonated peptides, the position of the “alternative” amino acids
in XAAG, AXAG, and AAXG (where X represents the position of the “alternative”
amino acid) [39] inhibits or completely suppresses formation of specific bn+ and yn+ ions.
This observation was attributed to the prohibitive effect of forcing cyclization and
intramolecular nucleophilic attack to progress through larger cyclic intermediates, which
48
would be kinetically slower to form and entropically less favored, when amino acids such
as βA, γAbu or Cap were used. Cyclization was prohibited when the 4AMBz residue was
used because the rigid aromatic ring separates the nucleophile from the electrophilic site
of attack. For the metal (Li+, Na+ and Ag+) cationized peptides [29] the “alternative”
amino acids, when placed at the C-terminus of the model peptide acetyl-FGGX, inhibited
the formation of the (b3+17+cat)+, because these amino acids either require larger cyclic
intermediates to transfer the required O and H atoms, or prohibited cyclization because of
the inclusion of a rigid aromatic ring.
During the initial CID study of protonated XAAG, AXAG and AAXG it was
discovered that while the 4AMBz residue can cause the disappearance of specific bn+ and
yn+ ions because the aromatic ring precluded cyclization and intramolecular nucleophililc
attack, formation of other bn+ ions appeared to be enhanced. For example, increased
intensities of b3+ and b2+ ions were observed for A(4AMBz)AG and (4AMBz)AAG,
respectively, suggesting that 4AMBz enhances formation of b type product ions that arise
via cleavage of the amide bond one sequence position to the C-terminal side of the
residue. The positive effect was attributed to the fact that the oxazolinone product formed
in the dissociation reaction would be highly-conjugated (and thus stable) and feature an
aromatic ring substituent. This is shown for protonated A(4AMBz)AG in scheme 9.
The overall aim of the experiments described in this chapter was to extend earlier
experiments and investigate the influence of 4AMBz on the formation of the rival
(b3-1+cat)+ and (b3+17+cat)+ products from metal cationized tetrapeptides. As noted
above, these specific ions are generated by competing reactions that ultimately involve
cleavage of the same amide bond. For this study, model peptides with general sequence
49
A(4AMBz)AX and A(4AMBz)GX, where X = G, A, and V, were synthesized. The CID
of protonated and metal cationized (Li+, Na+ and Ag+) forms of these model peptides was
then examined by multiple-stage ion trap tandem mass spectrometry and by probing the
change in product ion intensities with applied collision voltage (so called CID or
dissociation profiles). The specific hypothesis tested was that the presence and position of
the 4AMBz residue would enhance formation of the (b3-1+cat)+ ion, primarily through
generation of the stable, highly conjugated oxazolinone, at the expense of the normally
favored (b3+17+cat)+ species.
Results
CID of cationized GGGA and AAAG
The structures of the group of peptides synthesized and examined by ion-trap CID
are provided in figure 4.1. The first goal of this study was to probe for a general effect of
sequence on the formation of bn+, (bn-1+cat)+ and (bn+17+cat)+ ions. To this end the CID
patterns for cationized GGGA and AAAG were examined. Figure 4.2 shows the CID
(MS/MS stage) spectra derived from GGGA cationized by attachment of H+ (4.2a), Li+
(4.2b), Na+ (4.2c) and Ag+ (4.2d). CID of protonated GGGA produced a prominent peak
at m/z 243 via a loss H2O. One possible mechanism by which the m/z 243 product may be
generated is the conventional “oxazolone” pathway [17] operative in formation of bn+ and
yn+ ions. For generation of the [M-H2O+H]+ ion from GGGA (m/z 243), this pathway
would involve cyclization, nucleophilic attack and elimination of H2O from the Cterminal acid group [54]. An alternative route would be a retro-Koch or retro-Ritter
50
type reaction, originally proposed by O’Hair and coworkers, which would involve
elimination of an amide carbonyl oxygen atom [28,55,56]. Subsequent CID (MS3 stage,
spectrum not shown) of the species at m/z 243 showed the loss of 28 and 29 mass units
(u) with the former pathway more prominent than the latter. Loss of 28 as CO is the
characteristic decomposition pathway for bn+ ions that feature the oxazolinone ring
[19,20]. Loss of 29 has been observed in our laboratory for species in which the H2O
molecule is eliminated from an amide position, and in the present case reflects the loss of
HN=CH2 from the N-terminal end of the m/z 243 species.
The dominant product ion in the CID spectra derived from the alkali metal
cationized versions of GGGA was (b3+17+cat)+ at m/z 196 and 212 for Li+ (4.2b) and
Na+ (4.2c), respectively. In each case, the rival (b3-1+cat)+ ion appeared at relative
intensities of only ~5% (m/z 178 and 194 for Li+ and Na+, respectively). Li+ cationized
product ions in particular are prone to the formation of adducts through association
reactions with adventitious H2O in the ion trap, and an H2O adduct to the (b3-1+Li)+
species derived from AAAG would be isobaric with (b3+17+Li)+. To test for generation
of an H2O adduct, (b3-1+Li)+ was selectively isolated and stored in the ion trap, without
imposed collisional activation, for 30 msec after its formation by CID of Li+ cationized
AAAG (spectrum not shown): all species other than the one selected for isolation were
resonantly ejected from the ion trap. The relative intensity of a peak 18 u higher,
indicative of formation of an H2O adduct, was only ca. 5%, relative to (b3-1+Li)+, thus
indicating that the species as m/z 196 in figure 4.2b is primarily attributable to formation
of (b3+17+Li)+. No H2O adducts to (b3-1+Na)+ were observed when the same test was
applied to the Na+ cationized version of AAAG.
51
CID of Ag+ cationized GGGA (4.2d) generated (b3+17+Ag)+ (m/z 296 and 298),
(b3-1+Ag)+ (m/z 278 and 280) and (a3-1+Ag)+ (m/z 250 and 252). For the Ag+ cationized
version, the (b3+17+Ag)+ ion was observed at a relative intensity of ca. 25%, while the
rival (b3-1+Ag)+ product was the base peak. As for the Li+ and Na+ cationized versions of
the peptide, isolation of the (b3-1+Ag)+, without imposed collision activation,
demonstrated that formation of H2O adducts was not a significant source of error while
determining relative (b3+17+Ag)+ peak intensity.
Figure 4.3 shows the CID spectra derived from AAAG cationized by attachment
of H+ (4.3a), Li+ (4.3b), Na+ (4.3c) and Ag+ (4.3d). For the protonated version of the
peptide, b3+ at m/z 214 was the dominant sequence ion generated, with b2+, y2+ and a3* at
m/z 143, 147 and 169, respectively, also observed. A shift in preference from b4+ for
GGGA to b3+ for AAAG demonstrates that the formation of bn+ ions is, in general, less
favored when the cyclization and nucleophilic attack steps occur at or “across” a glycine
residue.
As shown in figures 4.3b and 4.3c, the change in peptide sequence also influenced
significantly the product ion distribution for the Li+ and Na+ cationized versions of
AAAG, respectively. While (b3+17+Li)+ at m/z 238 was the most abundant product ion
generated from the Li+ cationized version of AAAG, the (b3-1+Li)+ and (a3-1+Li)+
products at m/z 220 and 192 were generated at relative intensities significantly higher
than from GGGA. For the Na+ cationized version of AAAG, the (b3-1+Na)+ species at
m/z 236 was the most abundant product. For Ag+ cationized AAAG (4.3d), (b3-1+Ag)+
species at m/z 320 and 322 was the dominant product ion, and the rival (b3+17+cat)+
product, which would have m/z values of 338 and 340, was not observed. As with the
52
GGGA peptides, isolation and storage of (b3-1+cat)+ derived from AAAG, without
imposed collision activation, showed that H2O adducts did not contribute significantly to
the apparent (b3+17+cat)+ peak intensities.
CID (MS3 stage, spectra not shown) of (b3+17+Li)+ and (b3+17+Na)+ generated
(b2+17+Li)+ and (b2+17+Na)+, respectively, as the dominant product ions. Loss of H2O
from (b3+17+Li)+ or (b3+17+Na)+ to generate the (b3-1+cat)+ species was not a prominent
fragmentation pathway. This observation further confirms that the (b3+17+cat)+ and
(b3-1+cat)+ species represent the products of competing dissociation pathways (from the
same precursor ion) that involves cleavage of the same amide bond, rather than two steps
along a sequential dissociation pathway.
CID of cationized A(4AMBz)AG
Comparison of the CID of cationized GGGA and AAAG demonstrate that relative
product ion intensities are influenced by the position of the G residue(s) in the peptide
sequence, and that formation of product ions that involve cyclization and nucleophilic
attack across a glycine residue, such as formation of (b3+17+cat)+ from AAAG or b3+
from protonated GGGA, are less favored compared to a case where the same processes
instead involve an alanine residue. The high yield of both (b3+17+cat)+ and (b3-1+cat)+
from Li+ or Na+ cationized versions of AAAG made this system particularly well suited
for an examination of the influence of the 4AMBz residue on competition between the
two dissociation pathways.
Figure 4.4 shows CID spectra derived from A(4AMBz)AG cationized by
attachment of H+ (4.4a), Li+ (4.4b), Na+ (4.4d) and Ag+ (4.4e). The sole product ion
53
generated from protonated A(4AMBz)AG was b3+ at m/z 276. Both b2+ and y2+ were
generated in relatively high abundance from protonated AAAG, but were not observed in
the CID spectrum of protonated A(4AMBz)AG. Elimination of the pathways leading to
b2+ and y2+ which would have m/z values of 205 and 147, respectively, is attributed to the
fact that the cyclization and nucleophilic attack steps necessary to generate the products
are inhibited by the rigid aromatic ring of the 4AMBz residue [39].
Two major product ions were observed following CID of Li+ cationized
A(4AMBz)AG: one at m/z 282 and another at m/z 300. The species at m/z 282 is formed
via a neutral loss of 75 u and is attributed to (b3-1+Li)+. The species at m/z 300 was
initially attributed to (b3+17+Li)+. As noted above, Li+ cationized product ions are prone
to the formation of adducts through association reactions, and the H2O adduct to the
(b3-1+Li)+ species derived from A(4AMBz)AG would be isobaric with (b3+17+Li)+. The
(b3-1+Li)+ species was selectively isolated and stored in the ion trap, without imposed
collisional activation, for 30 msec after its formation by CID of Li+ cationized
A(4AMBz)AG. As shown in figure 4.4c, isolation and storage of (b3-1+Li)+ lead to
formation of an abundant ion at m/z 300, which identifies the latter species as an H2O
adduct. The ion at m/z 300 also displayed characteristics of loosely-bound adducts in ion
trap mass spectrometers [57-59], including a pronounced tail to the low-mass side of the
peak and a chemical mass shift. We therefore conclude that the peak at m/z 300 observed
in figure 4.4b is not (b3+17+Li)+, but instead an H2O adduct to (b3-1+Li)+.
As shown in figures 4.4d and 4.4e, (b3-1+cat)+ at m/z 298 and 383 was the
dominant product ion generated by CID of the Na+ and 109Ag+ cationized versions of
A(4AMBz)AG, respectively. Because of lower hydration energy, H2O adducts to Na+
54
cationized products are rarely, if ever, observed in our ion trap under the experimental
conditions used here. To eliminate any ambiguity in the evaluation of CID product ion
yields, the (b3-1+Na)+ and (b3-1+Ag)+ product ions were isolated and stored, without
imposed collision activation, as described above for the Li+ cationized version. No
formation of H2O adducts was observed. The (b3+17+cat)+ product ion at m/z 316 was
generated at a relative intensity of ~3% for Na+ cationized A(4AMBz)AG, and was not
observed for the Ag+ cationized version of the peptide. Comparison of the spectra shown
in figure 4.4 to those in figure 4.3 demonstrates the significant influence of the 4AMBz
residue on the product ion intensities, and in particular, the competition between
(b3-1+cat)+ and (b3+17+cat)+ from Li+ and Na+ cationized versions of the peptides.
CID of Na+ cationized A(4AMBz)AX and A(4AMBz)GX
Comparison of the CID spectra in figures 4.2 and 4.3 reveals the sensitivity of
relative product ion intensities to the sequence of the model peptides, particularly with
respect to the specific positions of G and A residues. To examine further the competition
between the (b3-1+cat)+ and (b3+17+cat)+ product ions, the CID of A(4AMBz)AX and
A(4AMBz)GX, where X= A or V was investigated. In these experiments, the hypothesis
was that the use of larger amino acids at the C-terminus would enhance formation of
(b3+17+cat)+, as was observed in the comparison of ion intensities for metal cationized
GGGA and AAAG (figures 4.2 and 4.3). For the A(4AMBz)GX series, any enhancement
to the yield of (b3+17+cat)+ due to the large C-terminal amino acid could also be
amplified by the fact that the G residue adjacent to the C-terminus may make formation
of (b3-1+cat)+ less favorable. For these experiments, CID of the [M+Na]+ ions
55
of A(4AMBz)AX and A(4AMBz)GX was examined because the Na+ cationized product
ions are not susceptible to the gas-phase H2O addition reactions that plague experiments
with Li+ cationized versions.
Figure 4.5 shows CID spectra derived from Na+ cationized A(4AMBz)AA (4.5a),
A(4AMBz)AV (4.5b), A(4AMBz)GA (4.5c) and A(4AMBz)GV (4.5d). No appreciable
increase in the intensity of (b3+17+Na)+ was observed when the C-terminal G residue of
A(4AMBz)AG was replaced with A or V. As is apparent in figures 4.5a and 4.5b, the
(b3+17+Na)+ ion was generated at relative intensities below 5%, and was nearly identical
to the yield of the same species from Na+ cationized A(4AMBz)AG. The relative
intensity of (b3+17+Na)+ was markedly higher for the A(4AMBz)GX peptides (figures
4.5c and 4.5d, respectively), and appeared at ca. 10% and 40% for A(4AMBz)GA and
A(4AMBz)GV, respectively.
We attribute the increase in (b3+17+Na)+ intensity relative to (b3-1+Na)+ observed
for the A(4AMBz)GA and A(4AMBz)GV peptides to the fact that the reaction to produce
the latter product involves cyclization of a G residue, thus making the pathway less
favorable. However, regardless of the sequence and identity of the C-terminal amino
acids, for no peptide containing the 4AMBz residue was the intensity of the (b3+17+Na)+
product greater than (b3-1+Na)+.
CID of Na+ cationized AA(4AMBz)G and (4AMBz)AAG
The spectra shown in figures (4.4, 4.5) demonstrate a significant influence by the
4AMBz residue on the competition between formation of (b3-1+cat)+ and (b3+17+cat)+
from the Li+ and Na+ cationized forms of the model peptides. An important question was
56
whether the influence was due to the specific position of 4AMBz, or instead the mere
presence of the residue within the peptide sequence. In a previous CID study of model
peptides with “alternative” amino acids at the C-terminal position [39], we observed the
complete elimination of the (bn+17+cat)+ pathway for peptides with an aminobenzoic
acid residue. To test for an influence by the position of the aromatic residue on
dissociation patterns in the present set of experiments, CID of cationzied AA(4AMBz)G
and (4AMBz)AAG was examined. For the sake of brevity, only the CID spectra of Na+
cationized AA(4AMBz)G and (4AMBz)AAG are shown in figure 4.6.
For Na+ cationized AA(4AMBz)G, the dominant product ion generated was
(b3+17+Na)+ at m/z 316. The (b3-1+Na)+ and (a3-1+Na)+ product ions, which would have
m/z values of 298 and 270, respectively, were not observed. Similar results were obtained
for the Li+ cationized peptide (spectrum not shown). The absence of the (b3-1+cat)+ and
(a3-1+cat)+ peaks is consistent with the inhibition, by the 4AMBz residue, of the
cyclization step and nucleophilic attack necessary to produce the species along the
“oxazolone” pathway. A similar effect was noted in our earlier study of the CID of
protonated peptides. For AA(4AMBz)G, the 4AMBz residue clearly influences the
competition between the rival pathways by eliminating the (b3-1+cat)+ pathway.
For Na+ cationized (4AMBz)AAG, the (b3+17+Na)+, (b3-1+Na)+ and the
(a3-1+Na)+ ions at m/z 311, 298 and 270, respectively, were observed with relative
abundances similar to those observed in figure 4.3c for Na+ cationized AAAG. For this
particular peptide, the 4AMBz residue is located at the N-terminus of the peptide and
should not affect formation of (b3+17+Na)+ or (b3-1+Na)+ by impeding the cyclization or
nucleophilic attack steps required for generation of these product ions. We can therefore
57
conclude that it is not the mere presence of the 4AMBz that is responsible for an effect on
the competition between (b3+17+Na)+ and (b3-1+Na)+, but instead the specific position of
the residue in the peptides.
Dissociation profiles generated using AAAG and A(4AMBz)AG
The observations made using the CID spectra derived from protonated and metal
cationzied A(4AMBz)AG clearly demonstrate that the 4AMBz residue influences
product ion intensities, and especially the competition between the competing (b3-1+cat)+
and (b3+17+cat)+ products from Li+ and Na+ cationized peptides. The next question that
arose was whether the effect of the 4AMBz residue is one of promoting the rate of
formation of (b3-1+cat)+, presumably because of the generation of the stable, highly
conjugated oxazolinone structure, or instead the suppression or inhibition of the
(b3+17+cat)+ pathway. For precursor ions of a given internal energy, the relative peak
heights for competing dissociation pathways are dependent on their respective reaction
rates [60,61]. To probe in a crude way the response of product ion intensities to changes
in internal energy, CID profiles, which show changes in precursor and product ion
intensities as a function of the voltage applied to the end-cap electrodes of the ion trap
voltage to induce dissociation, were examined. For these specific experiments, the
[M+Na]+ ions of AAGV and A(4AMBz)GV were subjected to CID because a prominent
(b3+17+Na)+ product was generated from the latter peptide despite the presence of the
aromatic amino acid in the sequence.
In general, higher collision voltages applied to induce dissociation should lead to
higher precursor ion internal energies, but it has been noted that the correlation between
58
collision energies and ion internal energies and their distributions is not necessarily
straightforward [61]. Because of uncertainties in ion kinetic energies in ion traps and in
the ultimate conversion of collision energy to internal energy, and the need to tune
precisely the activation R.F. voltage and frequency, the quadrupole ion trap is less well
suited for energy-resolved CID measurements than, for example, guided ion beam
instruments. However, the intent in collecting and examining the CID profiles here was
not to determine and report the appearance energies or thresholds for the respective
dissociation pathways, but instead to probe for gross changes in the relative intensities of
(b3-1+cat)+ and (b3+17+cat)+ as a function of the voltage applied to affect CID when the
4AMBz residue is in a sequence position to influence a fragmentation pattern.
Figure 4.7 shows the CID profile for Na+ cationized AAGV and A(4AMBz)GV,
collected by scanning the normalized collision energy from 10 through 35% (roughly
0.75 → 1.2 V applied). As noted in the experimental section, the normalized collision
energy was converted to applied voltage in the laboratory frame of reference, and then
divided by the vibrational degrees of freedom (DOF) of the precursor peptide. Using this
approach, the rise in product ion intensities can be compared at the similar applied
collision voltage per DOF to minimize any erroneous observations and conclusions due
to “degrees of freedom” effect on the internal energy needed to induce a fragmentation
reaction [61].
For Na+ cationized AAGV, the CID profiles demonstrate that the (b3+17+Na)+ ion
is the dominant product at all applied CID voltages beyond which dissociation is
observed, and reached a fraction of ion abundance of 0.1 at ca. 0.00960 V/DOF. The rival
(b3-1+Na)+ species failed to reach the same fractional abundance of 0.1 even at the
59
highest applied collision voltages. The CID profiles for Na+ cationized AAGV are
therefore consistent with a lower critical energy, and thus higher formation rate at a given
precursor ion internal energy, for generation of (b3+17+Na)+ compared to (b3-1+Na)+.
For Na+ cationzied A(4AMBz)GV, the profiles for both (b3+17+Na)+ and
(b3-1+Na)+ are shifted to lower relative collision voltages per DOF primarily because of
the greater number of DOF due to the inclusion of the larger 4AMBz residue. The
important observation, however, is that for Na+ cationized A(4AMBz)GV, it is the
intensity of the (b3-1+Na)+ species that begins to rise at lower relative collision voltages,
reaching a fraction of the ion abundance of 0.1 at ca. 0.00617 V. The (b3+17+Na)+
product instead reaches a fractional abundance of 0.1 at 0.00665 V. The decrease of
(b3-1+Na)+ ion intensities at higher V/DOF (i.e. above ~.0067) is due to dissociation of
(b3-1+Na)+ into (a3-1+Na)+ via the loss of CO. The curve showing the increase in
(a3-1+Na)+ intensity as a function of applied CID voltage was not included in figure 4.7
for the sake of clarity. In any case, the CID profile for Na+ cationized A(4AMBz)GV
suggests that the critical energy for generation of (b3-1+Na)+ is decreased relative to that
of (b3+17+Na)+ when the aromatic amino acid is located in a position such that it would
lead to the formation of the highly conjugated oxazolinone, thus increasing the formation
rate of the former species compared to the latter.
60
Summary
The purpose of this study was to determine the influence of
4-Aminomethylbenzoic acid on the formation of the rival (b3-1+cat)+ and (b3+17+cat)+
species. Generation of these product ions involves cleavage of the same amide bond, but
through entirely different mechanisms. For model peptides with general formula
A(4Ambz)AX and A(4Ambz)GX, our study shows that formation of the (bn-1+cat)+
species is favored over the (bn+17+cat)+ species when the amide bond cleaved is one
residue to the C-terminal side of the 4Ambz residue. We attribute this effect to the
potential formation of a highly conjugated oxazolinone species as (b3-1+cat)+.
Energy-resolved CID, which measures the collision energies necessary to
generate product ions from a particular precursor ion, shows that formation of (b3-1+cat)+
occurs at significantly lower collision energies when the 4Ambz residue is present to
affect the reaction pathway. This result is suggestive of an influence of the aromatic
residue on the activation energy of the reaction to produce (b3-1+cat)+. Placing residues
with bulky side groups at the C-terminus, which are known to enhance formation of
(b3+17+cat)+, caused a modest increase in relative abundance of the product. However,
generation of (b3-1+cat)+ remained the dominant dissociation pathway.
61
H
N
H2N
O
N
H
O
H
N
H2N
H
N
O
H
N
N
H
H2N
GGGA
OH
AAAG
OH
A(AMBz)AA
OH
A(AMBz)AV
OH
A(AMBz)AG
OH
A(AMBz)VG
O
O
O
H
N
OH
O
O
N
H
O
H
N
O
O
O
O
N
H
H
N
H2N
H
N
O
O
O
O
N
H
H
N
H2N
H
N
O
O
O
O
N
H
H
N
H2N
H
N
O
O
O
H2N
H
N
O
H
N
H2N
O
O
N
H
H
N
O
OH
(AMBz)AAG
O
O
N
H
H
N
AA(AMBz)G
O
OH
O
Figure 4.1
Sequence/structures of model peptides used to study the influence of
4AMBz on the formation of (b3-1+cat)+ and (b3+17+cat)+.
62
100
R. I. (%)
80
(a)
y2
60
a4
+
+
215
20
125
150
175
(b)
196
200
225
250
275
300
+
(b3+17+Li)
60
40
+
(b3-1+Li)
20
178
0
150
100
80
R. I. (%)
+
b4 /(M-H2O+H)
+
172
40
80
175
200
(c)
225
212
250
275
250
275
+
(b3+17+Na)
60
40
+
(b3-1+Na)
20
194
0
150
100
80
R. I. (%)
b3
147
0
100
100
R. I. (%)
243
+
175
(d)
200
225
+
(b3-1+Ag)
278,280
60
+
(b3+17+Ag)
40
20
0
200
296,298
+
(a3-1+Ag)
250,252
225
250
275
300
325
350
375
m/z
Figure 4.2
CID (MS/MS) spectra generated from cationized GGGA: (a) H+
cationized, (b) Li+ cationized, (c) Na+ cationized and (d) Ag+
cationized.
63
400
100
R. I. (%)
80
(a)
b2
40
125
271
175
200
225
250
275
300
238
192
220
+
(a3+17+Li)
60
+
(b3+17+Li)
+
(b3-1+Li)
40
20
80
R. I. (%)
169
150
(b)
260
0
150
100
175
200
(c)
225
250
275
300
325
236
+
(b3-1+Na)
+
(b3+17+Na)
254
60
+
(a3-1+Na) 208
40
20
293
0
150
100
80
R. I. (%)
+
+
143
20
80
b3
147
60
0
100
100
R. I. (%)
y2
214
+
175
200
225
(d)
250
+
(b3-1+Ag)
275
300
325
350
320,322
60
40
+
(a3-1+Ag)
20
0
200
292,294
220
240
260
280
300
320
340
360
380
400
420
m/z
Figure 4.3
CID (MS/MS) spectra generated from cationized AAAG: (a) H+
cationized, (b) Li+ cationized, (c) Na+ cationized and (d) Ag+
cationized.
64
100
(a)
R. I. (%)
80
40
20
R. I. (%)
80
250
275
+
(b3-1+Li)
300
325
350
375
400
325
350
375
400
325
350
375
400
325
350
375
400
282
60
300
40
20
80
R. I. (%)
225
(b)
0
200
100
225
250
(c)
275
+
(b3-1+Li)
300
282
300
60
40
20
0
200
100
80
R. I. (%)
276
60
0
200
100
225
250
275
(d)
300
+
(b3-1+Na)
298
60
40
20
0
200
100
80
R. I. (%)
+
b3
225
250
275
300
(e)
+
(b3-1+Ag)
383
60
40
20
0
225
355
250
275
300
325
350
375
400
425
450
475
m/z
Figure 4.4
CID (MS/MS) spectra generated from cationized A(4AMBz)AG: (a)
H+ cationized, (b) Li+ cationized, (c) Li+ cationized, isolation and
storage of (b3-1+Li)+, (d) Na+ cationized and (e) Ag+ cationized.
65
100
R. I. (%)
80
(a)
+
(b3-1+Na)
60
40
+
(b3+17+Na)
20
0
200
100
R. I. (%)
80
225
250
275
300
325
375
400
350
375
400
350
375
400
350
375
400
+
(b3-1+Na)
40
+
(b3+17+Na)
20
316
225
250
(c)
275
300
325
284
+
(b3-1+Na)
60
40
+
(b3+17+Na)
20
302
0
200
100
80
350
298
60
80
R. I. (%)
316
(b)
0
200
100
R. I. (%)
298
225
250
(d)
275
300
325
284
+
(b3-1+Na)
60
302
+
(b3+17+Na)
40
20
0
200
225
250
275
300
325
m/z
m/z
Figure 4.5
CID (MS/MS) spectra generated Na+ cationized peptides:
(a)A(4AMBz)AA, (b) A(4AMBz)AV, (c) A(4AMBz)GA and
(d) A(4AMBz)GV.
66
100
(a)
316
+
(b3+17+Na)
R. I. (%)
80
60
40
+
(M+Na)
373
-CO2 -H O, -H2O
2
329 NH3
355
338
20
0
200
100
225
250
275
300
325
350
375
400
298
(b)
+
(b3-1+Na)
+
(b3+17+Na)
R. I. (%)
80
316
60
+
(a3-1+Na)
270
40
+
-H2O
355
20
0
200
225
250
275
300
325
350
(M+Na)
373
375
400
m/z
Figure 4.6
CID (MS/MS) spectra generated Na+ cationized peptides: (a)
AA(4AMBz)G, (b) (4AMBz)AAG.
67
O
N
H
H
N
H+
H2 N
O
H
N
O
OH
O
Proton transfer
O
N
H
H
N
H2 N
H
N
H+
O
OH
O
O
O H
N
H+
O
H
N
H 2N
N+
H
O
O
O
N+
H
H
N
H 2N
O
Scheme 9
O
H2 N
OH
b3 +
Reaction mechanism for formation of (b3)+ fromA(4AMBz)AG.
68
O
OH
1.0
+
(b3-1+Na) , AAGV
+
(b3+17+Na) , AAGV
Fraction of ion abundance
0.8
+
(b3-1+Na) , A(4AMBz)GV
+
(b3+17+Na) , A(4AMBz)GV
0.6
0.4
0.2
0.0
0.005
0.006
0.007
0.008
Applied CID voltage/DOF
Figure 4.7
CID profiles for Na+ cationized AAGV and A(4AMBz)GV).
69
0.009
CHAPTER V
Formation of (b3-1+cat)+ Ions from Metal-cationized Tetrapeptides Containing βAlanine, γ-Aminobutyric Acid or ε-Aminocaproic Acid Residues
Introduction
Formation of a 5-member cyclic intermediate is thought to be an integral
part of the reactions to produce the bn+ and yn+ sequence ions from protonated peptides
[19,20,23-26], and analogous (bn-1+cat)+ products from metal-cationized peptides
through the “oxazolone” mechanism (shown for b3+ from AAAG in scheme 1a). The
(bn+17+cat)+ species are prominent, if not the preferred, dissociation products for Li+ and
Na+ cationized peptides [33,34,36,48]. The mechanism for the (bn+17+cat)+ pathways is
also thought to involve cyclization and intramolecular nucleophilic attack with proton
transfer (scheme 1b) [31,32,35,52]. The pathway depicted in scheme 1b includes
concerted opening of the cyclic intermediate and elimination of CO and an imine.
Gronert, Lebrilla and coworkers have proposed instead a plausible route that involves an
anhydride intermediate after ring opening [52]. In any case, the salient feature is that, as
in the case of the bn+/(bn-1+cat)+ pathway, the electrophilic site of attack in the
(bn+17+cat)+ pathway is the carbonyl carbon of the amide group being cleaved. In this
case, however, the nucleophile is the carbonyl oxygen of the C-terminal carboxylic acid.
Recently we investigated and reported on the incorporation of “alternative” amino
acids such as β-alanine (βA), γ-aminobutyric acid (γABu), ε-aminocaproic acid (Cap),
and 4-aminomethylbenzoic acid (4AMBz) into the sequence of a model peptide, and the
effect(s) of these residues on relative product ion intensities [29, 39]. The goal of this
third study was to determine the effect on fragmentation patterns of changing the size of
the putative cyclic intermediate formed during the nucleophilic attack. For protonated
70
peptides, the presence of βA, γAbu or Cap in XAAG, AXAG, and AAXG (where X
represents the position of the “alternative” amino acid) inhibited or completely
suppressed formation of specific bn+ and yn+ ions [39]. This observation was attributed to
the prohibitive effect of forcing cyclization and intramolecular nucleophilic attack to
progress through larger cyclic intermediates, which should mainly be kinetically slower
to form and entropically less favored, when the larger amino acids were used. Cyclization
is prohibited when the residue is 4AMBz because the rigid aromatic ring separates the
nucleophile from the electrophilic site of attack. For metal (Li+, Na+ and Ag+) cationized
peptides, similar suppression and inhibition of the formation of (b3+17+cat)+ was
observed when the “alternative” amino acids were placed at the C-terminus of the model
peptide acetyl-FGGX [29].
In this third study, the influence of βA, γAbu and Cap on the tendency to form
(b3-1+cat)+ products from Li+, Na+ and Ag+ cationized AAXG was investigated. The
initial hypothesis was that the potential prohibitively large cyclic intermediates would
suppress formation of (b3-1+cat)+, as was observed for the protonated versions of the
peptide. However, the metal cations may coordinate with the peptide through interactions
with multiple amide carbonyl O atoms, and thus kinetically assist formation of a reactive
configuration from which nulceophilic attack occurs. As we show here, (b3-1+cat)+ is a
prominent, if not dominant, reaction product from the metal cationized AAXG. More
importantly, isotope labeling demonstrates that formation of the product involves the
transfer of an H atom that originates from an α-carbon position of the amino acid in
position X.
71
Results
CID of AA(γAbu)G
The sequences and structures of the model peptides included in this study are
shown in figure 5.1. Figure 5.2 shows CID (MS/MS stage) spectra for AA(γAbu)G
cationized by: H+ (2a), Li+ (2b), Na+ (5.2c) and Ag+ (both the 107Ag+ and 109Ag+ isotopic
peaks, (5.2d). For the protonated version of AA(γAbu)G, elimination of the full Cterminal G amino acid (75 mass units, u) to generate b3+ would furnish an ion peak at m/z
228. However, as in our earlier investigation [39], b3+ is not observed for AA(γAbu)G
and the dominant product ion generated was instead y2+ at m/z 161. Suppressed formation
of b3+ for the peptide containing γAbu is consistent with, at least in part, potential
changes to the rates of the putative ring-closure step leading to nucleophilic attack in the
“oxazolone” mechanism: the 7-member ring for γAbu should be kinetically slower to
form than the conventional 5-member rings that are postulated as intermediates for αamino acids.
For Li+ and Na+ cationized versions of AA(γAbu)G (figures 5.2b and c,
respectively), the (b3+17+cat)+ product, generated by neutral loss of 57 mass units (u),
was prominent. Other product ions observed included (M-H2O+cat)+, (y3-1+cat)+,
(a3-1+cat)+ and (y2-1+cat)+. A more important observation, however, was the appearance
of the (b3-1+cat)+ product. The relative abundance of (b3-1+cat)+ is comparable to that of
(b3+17+cat)+ for the Li+ and Na+ cationized versions of the peptides, and is the dominant
product ion generated from the Ag+ cationized version of the peptide. Thus, despite the
potential prohibitive effect of the larger cyclic intermediates that presumably suppress
formation of b3+ for the protonated peptides, the analogous (b3-1+cat)+ ion is prominent
72
in the spectra derived from the metal cationized peptides. Formation of (b3-1+cat)+ was
not observed for the metal cationized versions of AA(4AMBz)G (spectra not shown),
consistent with the hypothesis that prohibiting cyclization altogether blocks formation of
the bn+/(bn-1+cat)+ products.
CID of deuterium exchanged and labeled AAXG
Figure 5.3 shows the CID spectra generated from (M+D)+ of AAXG peptides for
which exchangeable H atoms were replaced with D by incubation of the peptides in a
mixture of D2O and CH3OD. For D+-cationized AAAG (figure 5.3a), the b3+ ion appears
as multiple isotopic peaks. Isolation and storage of the m/z 218 isotopic peak, without
imposed collisional activation (spectra not shown) suggested that the m/z 217, 216 and
215 isotopic peaks are instead generated by D for H back-exchange through collisions
with adventitious H2O in the ion trap. This was confirmed by monitoring exchange of D
for H with increasing isolation times. Investigation of the cause for the rapid D for H
exchange for b3+ derived from AAAG is beyond the scope of this report, and is being
studied in greater detail using a more extensive set of model peptides.
As shown in figure 5.3b, the b3+ ion at m/z of 218, via loss of 78 u, was also
observed following CID of deuterium exchanged AA(βA)G, consistent with loss of fully
deuterium exchanged glycine from the C-terminus. D for H back exchange was minimal,
and became significant only at isolation times longer than 1 second (data not shown). For
AA(βA)G, AA(γAbu)G and AA(Cap)G (figures 5.3b through d, respectively), formation
of y2+ following CID of (M+D)+ occurred through the net elimination of 144, consistent
with loss of the two N-terminal A residues, each with a single D atom. Isolation and
73
storage of y2+, without imposed collisional activation, showed that the tendency for D for
H back exchange was negligible.
The CID spectra generated from Li+ and Na+ cationized versions of the
deuterium-exchanged forms of AAXG were qualitatively similar. For the sake of brevity,
the results from the Li+ versions are presented here as representative of both systems. As
shown in figure 5.4 for Li+ cationized, deuterium-exchanged AAAG, the (b3+17+Li)+ and
(b3-1+Li)+ products were observed at m/z 243 and 223, respectively. These m/z values
correspond to product ions generated by elimination of 58 and 78 u, respectively. The
formation pathways proposed for both the (bn-1+cat)+ and (bn+17+cat)+ products involve
transfer of H atoms from exchangeable sites, specifically amide and acid positions for the
former and latter, respectively. The elimination of 58 and 78 u during formation of
(b3+17+Li)+ and (b3-1+Li)+ from deuterium-exchanged AAAG is therefore consistent
with the respective proposed mechanistic pathways, and demonstrates that generation of
the products does not include significant scrambling of isotope labels or transfer of H
atoms from positions other than the amino terminus, amide positions, or the C-terminal
acid group. Both the (b3+17+Li)+ and (b3-1+Li)+ were isolated and stored in the ion trap,
without imposed collisional activation, for periods ranging from 10 msec to 1 sec (spectra
not shown): no appreciable D for H back exchange for the ions was observed.
For Li+ cationized, deuterium-exchanged AA(βA)G (5.4b), AA(γAbu)G (5.4c)
and AA(Cap)G (5.4d), formation of (b3+17+Li)+ continued to involve elimination of 58
u, consistent with transfer of the C-terminal acid-position D atom during the dissociation
reaction (as shown with a C-terminal H atom in scheme 1b). However, the (b3-1+Li)+ ion
appeared as two isotopic peaks (m/z 223, 224; 237, 238 and 265, 266 for X = βA, γAbu
74
and Cap, respectively) that arise through neutral losses of 77 or 78 u. Elimination of 78 u
is consistent with transfer of an amide-position D atom to the C-terminal residue, and loss
of fully deuterium exchanged glycine to furnish (b3-1+Li)+. The neutral loss of 77 u is
noteworthy, because it signals transfer to the C-terminal G residue of an H atom, rather
than a D atom, in the dissociation pathway. Because the peptides were fully D exchanged
(amino, amide and acid positions labeled with D), the transfer of an H atom implicates
one of the methylene groups within the residue at position X. As is evident in the
comparison of the spectra in figure 5.4, the tendency to lose 77 u rather than 78 u
increases following the trend βA < γAbu < Cap, i.e, as the size of the amino acid at which
cyclization and intramolecular nucleophilic attack should occur to produce (b3-1+cat)+
increases. Similar trends were observed for the Na+ cationized versions of the deuteriumexchanged peptides. As in the case of the AAAG peptide, both the (b3+17+Li)+ and
(b3-1+Li)+ derived from Li+ and Na+ cationized AA(βA)G, AA(γAbu)G, and AA(Cap)G
were isolated and stored in the ion trap, without imposed collisional activation, for
periods ranging from 10 msec to 1 sec (spectra not shown). No appreciable D for H back
exchange for the respective ions was observed.
The (a3-1+Li)+ ion at m/z 195 was a prominent product in the spectrum derived
from Li+ cationized, fully deuterium exchanged AAAG. CID (MS3 stage, spectrum not
shown) of (b3-1+Li)+ from the deuterium exchanged AAAG led primarily to (a3-1+Li)+
via elimination of CO. The (a3-1+Li)+ product is absent in the CID spectra generated
from Li+ cationized, fully deuterium exchanged AA(βA)G (5.4b), AA(γAbu)G (5.4c) and
AA(Cap)G (5.4d), nor was the ion generated by CID of (b3-1+Li)+ derived from the
respective peptides, as discussed in more detail in a later section.
75
Figure 5.5 shows the CID spectra for Ag+ cationized AAAG (5.5a), AA(βA)G
(5.5b), AA(γAbu)G (5.5c) and AA(Cap)G (5.5d). For the sake of clarity, the spectra
shown in figure 5.5 were generated using CID of only the 109Ag+ adduct of the peptides.
The (b3-1+109Ag)+ ion is the dominant product generated from each of the peptides, and
(b3+17+109Ag)+ appeared at relative intensities < 5% for the AA(βA)G, AA(γAbu)G and
AA(Cap)G peptides. As in the case of the Li+ and Na+ cationized versions, formation of
(b3-1+109Ag)+ generated from deuterium-exchanged AAAG involved primarily
elimination of 78 u, which represents fully deuterium exchanged glycine. However, loss
of 77 u was observed from the deuterium-exchanged AA(βA)G, AA(γAbu)G and
AA(Cap)G, and the fraction of product ion derived from this pathway increased
systematically with the size of the “alternative” amino acid. No D for H back exchange
was observed when either (b3+17+109Ag)+ or (b3-1+109Ag)+ derived from Ag+ cationized
AA(βA)G, AA(γAbu)G, and AA(Cap)G were isolated and stored in the ion trap, without
imposed collisional activation, for periods ranging from 10 msec to 1 sec (spectra not
shown). As in the case of the Li+ (and Na+) cationized versions of deuterium exchanged
AA(βA)G, AA(γAbu)G and AA(Cap)G, the (a3-1+109Ag)+ product was not generated by
CID of (b3-1+109Ag)+.
The CID spectra shown in figures 5.4 and 5.5 demonstrate that formation of
(b3-1+cat)+ from metal cationized AA(βA)G, AA(γAbu)G and AA(Cap)G involves, in
part, transfer of an H atom that does not originate in an exchangeable position. To
determine the specific position from which the H atom is transferred during formation of
(b3-1+cat)+, the CID of metal cationized AA(α−d2-γAbu)G was examined. This peptide
was chosen for study because α-d2 labeled γAbu is commercially available, and because
76
our initial hypothesis was that the α-C position is the most likely origin of the H atom. In
figures 5.6 and 5.7, high-resolution ZoomScan spectra collected after CID of native
AA(γAbu)G and AA(α−d2-γAbu)G are compared for the Li+ and 109Ag+ cationized forms
of the peptides, respectively. For AA(α−d2-γAbu)G (figures 5.6b and 5.7b), the
(b3-1+cat)+ ion appears as two isotopic peaks, which are shifted higher by 1 and 2 mass
units compared to the unlabeled version of the peptide. The higher mass isotopic peaks
(m/z 236 and 338 for Li+ and Ag+, respectively) are generated by the net elimination of
75 u and formation of (b3-1+cat)+ via a reaction that involves transfer of an H atom from
an amide position to the departing, C-terminal G residue. The lower mass peaks (m/z 235
and m/z 337 for the Li+ and Ag+ cationized versions, respectively) correspond to
formation of (b3-1+cat)+ by transfer of a D atom, thus supporting the hypothesis that the
α-C position is the source of the non-exchangeable H atom in the reaction pathway.
Regardless of the metal cation, exclusive loss of 78 u was observed following CID of
AA(α−d2-γAbu)G that was incubated in D2O/CH3OD to induce H/D exchange (spectra
not shown). This result suggests that only H (or D) atoms from amide N positions or the
α-carbon position of residue X are transferred in the reaction to produce (b3-1+cat)+.
Proposed mechanism for (b3-1+cat)+ with transfer of α-position H atom
The results shown thus far demonstrate that placement of larger amino acids such
as βA, γAbu or Cap in position X of AAXG does not appreciably inhibit formation of
(b3-1+cat)+ from the model peptide, unlike the results obtained for b3+ from protonated
versions of the same peptides [39]. The pronounced transfer of an H atom from an αcarbon position suggests that formation of (b3-1+cat)+ occurs, at least in part, through a
77
mechanistic pathway distinct from that proposed for peptides with conventional α-amino
acids. Two reasonable general pathways can be envisioned for (b3-1+cat)+ from AAXG
where X= βA, γAbu or Cap: one involving formation of cyclic intermediate larger than
the 5-member ring proposed for α-amino acids, and a second pathway that instead
involves tautomerization and generation of a metal-cationized ketene.
Using the peptide AA(γAbu)G for discussion, a mechanism involving
intramolecular nucleophilic attack and formation of a cyclic intermediate is shown in
scheme 10. Once the cyclic intermediate is formed, proton transfer from the amide N
position of the γAbu residue to the amide N atom of the C-terminal glycine residue must
occur. Cleavage of the C-terminal amide bond would then liberate glycine as a neutral
and furnish 3H-tetrahydro-1,3-oxazepine, a 7-membered ring analogue to the
oxazolinone (a 5-membered ring) generated by α-amino acids. In our earlier study of
protonated peptides, residues such as γAbu and Cap were found to suppress formation of
bn+ ions when the residues were positioned such that they influence formation of the
cyclic intermediate. The larger cyclic intermediates were thought to be kinetically and
entropically less favored, thus leading to diminished ion yields. In the present case,
formation of (b3-1+cat)+ for which amide position H (or D) atoms are transferred to the
departing neutral by the mechanism shown in scheme 10 suggests that the metal ion may
facilitate nucleophilic attack by coordinating with the proper amide carbonyl oxygen
atoms to assist formation of the reactive configuration, thus kinetically assisting the
reaction pathway.
78
Though the reaction shown in scheme 10 may account for generation of
(b3-1+cat)+ with specific transfer of an amide-position H (or D) atom, it cannot explain
the formation of the same species when an H atom from an α-carbon position is
transferred. Assuming formation of the same intermediate after nucleophilic attack,
transfer of the α-position D atom to the O atom would have to be accompanied by
opening of the ring as shown in scheme 11. However, the cyclic intermediate depicted in
scheme 11 has no properly located keto group to participate in keto-enol tautomerization,
or otherwise render the α-deuterium atoms sufficiently acid; so it is difficult to
mechanistically explain the transfer of deuterium from the α-carbon position necessary to
complete the reaction to produce (b3-1+cat)+ and account for the splitting of the product
ion as observed in figures 5.6 and 5.7
A rational alternative to a mechanism that involves cyclization is formation of a
metal-cationized ketene as shown in scheme 12. The appearance of (b3-1+cat)+ in the
CID spectra of metal-cationized AAXG, when X = γAbu or Cap, but not b3+ from
protonated analogues, and that generation of (b3-1+cat)+ from AAXG involves transfer of
an H atom from the α-position of X suggests that the acidity of this H atom has been
enhanced by the intervention of the metal ion. The normal pKa of this H atom is about
25-30. However, as suggested below, this H atom acquires the character of allylic to an
iminium ion prior to fragmentation. The pKa of H in an allylic position to
electronegatively-substituted imines such as oxime tosylates cannot be much higher than
16-18 because its removal by a base such as ethoxide (-OEt) promotes the Neber
rearrangement via an azirine [62]. Moreover, condensation of a CH3 group, adjacent to an
iminium ion, with orthoformate to form a cyanine dye in acetate buffer [63] indicates that
79
its pKa must be significantly lower than 16-18. Any process that interferes with the
production of carbocation A’ (scheme 12) would disfavor loss of glycine. Such a process
is cyclization involving nucleophilic attack by the oxygen of the penultimate alanine
residue of AAXG (iminium ion A, scheme 11). When X = Cap, this ring is ninemembered, whereas X = γAbu, βA or Ala leads to seven-, six- and five-membered rings,
respectively. It is well-known that formation of medium-sized rings (sizes 7-9) is least
favored based on both thermodynamic and kinetic factors. Thus, the acyl ion
condensation affords these rings in lower yields than both bigger or somewhat smaller
ones [64], and the Dieckmann condensation yields virtually no product corresponding to
medium-sized rings [65]. Therefore, we expect the loss of C-terminal glycine to be
greatest with X = Cap because of least competition with formation of A’, and diminish in
the order γAbu > βA > Ala.
If the (b3-1+cat)+ ion is indeed a metal-cationized ketene one would not expect it
to lose CO upon further dissociation. Ketenes (as odd-electron species) by themselves do
lose CO upon electron impact to produce electron deficient carbenes. However,
formation of such carbenes in the present case would be discouraged by the proximity of
electron deficient metal ions. Figure 5.8 shows spectra generated by CID (MS3 stage) of
(b3-1+Li)+ (5.8a) and (b3-1+109Ag)+ (5.8b) derived from the respective metal-cationized
versions of AA(γAbu)G. For neither species is the loss of CO a prominent pathway. The
dominant product fragmentation pathway for (b3-1+Li)+ involves either elimination of 18
u, presumably as H2O, or loss of 44 u. Elimination of 44 u was observed by our group for
dissociation of (b3-1+cat)+ ions generated from metal-cationized AXGG, where X=βA,
γAbu or Cap, and attributed to elimination of CO2 to furnish a metal-cationized azirine
80
[66]. In the prior study, the (b3-1+cat)+ ions were all 5-member ring oxazolinones, and
decomposition of the azirine led to formation of a metal-cationized nitrile product. What
specific atoms make up the neutral(s) eliminated at 44 u in the present study is not known
and will be determined using a more extensive series of isotope labeled peptides. Other
dissociation pathways observed were formation of (b2-1+Li)+ and (a2-1+Li)+ via
elimination of 85 u (residue mass of γAbu) and 113 u (residue mass of γAbu and CO),
respectively. For (b3-1+109Ag)+ , similar dissociation pathways were observed. However,
formation of (b2-1+109Ag)+ and (a2-1+109Ag)+ was more favored compared to loss of H2O
or 44 u.
It is clear from the spectra in figure 5.8, and the lack of (a3-1+cat)+ products in the
CID spectra of Li+ and Ag+ cationized AA(γAbu)G and AA(Cap)G (figures 5.3 and 5.4),
that (b3-1+cat)+ with the larger “alternative” amino acids does not dissociate via the
elimination of CO as do conventional bn+/(bn-1+cat)+ ions. Loss of CO from the
(bn-1+cat)+ and analogous bn+ species generated from AAAG, and related peptides
composed of α-amino acids, is assisted by concomitant facile formation of an imine
product (scheme 1a). For AA(βA)G, AA(γAbu)G and AA(Cap)G, cyclic (b3-1+cat)+
products formed via the pathway shown in scheme 10 would not be expected to form the
associated (a3-1+cat)+ species because opening of the ring, with concomitant elimination
of CO, can not lead to formation of an imine product as depicted in scheme 1 without
significant rearrangement. As noted above, metallated ketenes likely also would not
eliminate CO.
Loss of H2O from (b3-1+Li)+ and (b3-1+Na)+ probably occurs through a retroKoch or retro-Ritter type pathway and elimination of an amide O atom, as proposed first
81
by Reid and coworkers [55] for protonated peptides and supported by subsequent
experiments by our group with metal-cationized peptides [28, 35, 56]. The formation of
(b2-1+cat)+ and (a2-1+cat)+ from (b3-1+cat)+ can instead be attributed to the conventional
“oxazolone” pathway, with attack by the carbonyl O atom of the alanine residue adjacent
to the N-terminus upon either a C atom within the large ring formed in scheme 10, or the
amide group between alanine and the γAbu-ketene to the peptide generated in scheme 12.
Regardless of the precursor ion structure, such a nucleophilic attack would furnish a
metal-cationized oxazolinone as (b2-1+cat)+ which could decompose further, through loss
of CO, to produce (a2-1+cat)+. CID (MS4 stage, spectra not shown) of (b2-1+Li)+ and
(b2-1+Ag)+ derived from AA(γAbu)G or AA(Cap)G produces the (a2-1+cat)+ species via
the loss of CO.
Apparent influence of X on yield of other product ions
Another interesting observation was the apparent influence of X in AAXG on the
relative abundance of the (M-H2O+cat)+ and (b3+17+cat)+ products relative to
(b3-1+cat)+. As noted above, an investigation of the elimination of H2O from model,
metal cationized peptide-esters showed that the preferred pathway likely involves a retroKoch type-process that involve loss of an amide-position carbonyl O atom [32, 33]. This
is opposed to loss of the C-terminal -OH group through a pathway similar to that
operative for generation of the (bn-1+cat)+ and b3+ ions shown in scheme 1a. In the
previous study, loss of H2O was shown to involve loss of both amide N position and αcarbon position H atoms [56], consistent in the present study with the significant fraction
of the (M-H2O+Li)+ generated from the fully deuterium exchanged form of AA(βA)G
82
and AA(γAbu)G via the loss of HDO rather than D2O. The reason for the enhanced loss
of H2O from AA(βA)G and AA(γAbu)G is not clear, but we have found that the intensity
of (M-H2O+cat)+ increases when the favored (bn+17+cat)+ pathway is inhibited or
suppressed [29].
Noting the fact that the (b3-1+Li)+ species appears as two isotopic peaks, the ratio
of the abundance of (b3-1+Li)+ relative to (b3+17+Li)+ increases from ca. 0.75 (measured
using integrated peak areas) for AAAG to ca. 1.5 for AA(βA)G and AA(γAbu)G, and
further to nearly 5 for AA(Cap)G. The change in relative abundance is surprising given
the fact that the (b3+17+Li)+ is often the preferred, if not dominant, product generated
from metal-cationized peptides. We recently investigated the competition between
formation of (b3-1+Li)+ and (b3+17+Li)+ from A(4AMBz)AG and found that there was a
significant increase in the intensity of (b3-1+Li)+ for the peptides that contain the 4AMBz
residue, and in some cases the complete elimination of the (b3+17+Li)+ pathway [67].
The influence of the 4AMBz residue was attributed to the fact that (b3-1+Li)+ would be a
highly-conjugated species containing an aromatic ring substituent. Comparison of CID
profiles suggested that the competition may be influenced by a decrease of the critical
energy for generation of (b3-1+cat)+ relative to that of (b3+17+cat)+.
However, in the present study the initial assumption was that the larger amino
acids would suppress formation of (b3-1+cat)+ because of the larger cyclic intermediates,
and in doing so presumably increase the competitiveness of the (b3+17+cat)+ product. As
shown in scheme 1b, formation of (b3+17+cat)+ requires attack on an amide C atom by
the C-terminal carboxyl group to generate an oxazolidinone intermediate, whereas
formation of (b3-1+cat)+ via the ketene pathway would involve attack by the adjacent
83
nitrogen atom on the same carbocation (scheme 12), to produce an iminium ion.
Preferred formation of (b3-1+cat)+ over (b3+17+cat)+ as the size of the amino acid in
position X increases may be rationalized by considering that the nitrogen atom may be
more nucleophilic than the carbonyl group that would attack in the pathway shown in
scheme 1b, and thus make formation of the former product more favorable.
The (y3-1+cat)+ ion was also more prominent in the CID spectra generated from
Li+ and Na+ cationized AA(γAbu)G and AA(Cap)G when compared to AAAG.
Computational studies by Paizs and Suhai provide compelling evidence for an integrated
pathway that involves simultaneous cleavage of the C-C bond of the N-terminal amino
acid residue, and the N-terminal amide C-N bond to generate y3+, a1+ and CO [68]. Such a
pathway was found, for protonated glycine-glycine, to be energetically more favorable
than one involving formation of an aziridinone product via cyclization and attack by the
N-terminal amino group. In the present study, it is not clear what role the metal ion may
play in enhancing the formation of (y3-1+cat)+, and a more detailed investigation of the
enhanced formation of the species is currently underway using a larger group of model
peptides.
84
Summary
In this study we examined the effect of β-alanine, γ-aminobutyric acid or εaminocaproic acid residue on the formation of (b3-1+Cat)+ products from metal(Li+, Na+
and Ag+) cationized peptides. Although, the “alternative” amino acids suppress the
formation of b3+ from protonated peptides with the sequence AAXG due to the
prohibitively large cyclic intermediate in the “oxazolone” pathway, the results above
show that abundant (b3-1+Cat)+ are generated from metal cationized versions of AAXG.
Using a group of D-labeled and exchanged peptides, we found that the formation
of (b3-1+cat)+ involves transfer of either amide or α-carbon position H atoms. An
important observation was that the tendency to transfer the α-carbon position H atoms
increase with the size of the alternative amino acid in position X.
85
A
H
N
H2N
O
N
H
O
N
H
H
N
H2 N
H
N
O
Figure 5.1
O
H
N
O
H
N
O
OH
O
γAbu
O
O
H
N
N
H
O
H2 N
OH
βA
O
H2N
O
H
N
OH
O
Cap
O
H
N
N
H
O
OH
O
Sequence/structures of model peptides used to study the influence of
βA, γAbu and Cap on formation of (b3-1+cat)+ from metal-cationized
AAXG.
86
R. I. (%)
100
161
a
80
+
y2
60
+
b3
40
+
y3
144
20
232
0
130 140 150 160 170 180 190 200 210 220 230 240 250 260 270
R. I. (%)
100
80
+
[b3-1+Li]
b
+
[b3+17+Li]
234
60
274
252
+
40
[y3-1+Li]
+
[y2-1+Li]
20
238
167
0
140 150 160 170 180 190 200 210 220 230 240 250 260 270 280
R. I. (%)
100
80
+
[b3-1+Na]
c
[y3-1+Na]
60
+
40
[y2-1+Na]
20
[a3-1+Na]
250
268
+
[b3+17+Na]
+
290
+
254
222
183
0
150 160 170 180 190 200 210 220 230 240 250 260 270 280 290 300
R. I. (%)
100
338
d
80
336
+
[b3-1+Ag]
60
[b3+17+Ag]
+
40
[y2-1+Ag]
20
267 269
352
+
354
0
240 250 260 270 280 290 300 310 320 330 340 350 360 370 380
m/z
Figure 5.2.
CID (MS/MS) spectra generated from AA(γAbu)G series cationized
by: (a) H+, (b) Li+, (c) Na+ and (d) 107Ag+ and 109Ag+.
87
R. I. (%)
100
a
152
+
y2
+
b2
80
60
146
40
217
+
b3
20
218
215
a3*
151
216
170,171
0
140
R. I. (%)
100
80
b
150
160
170
180
190
200
210
220
230
152
60
40
20
218
0
140
R. I. (%)
100
80
c
150
160
170
180
190
200
210
220
230
166
60
40
20
0
150
R. I. (%)
100
80
d
160
170
180
190
200
210
220
230
240
194
60
40
20
0
180
190
200
210
220
230
240
250
260
270
m/z
Figure 5.3.
Spectra generated by CID (MS/MS) of D+ cationized, fully deuteriumexchanged: (a) AAAG, (b) AA(βA)G, (c) AA(γAbu)G and (d)
AA(Cap)G.
88
R. I. (%)
100
a
80
+
+
[b3+17+Li]
223
[a3-1+Li]
60
243
+
[b3-1+Li]
195
+
[y3-1+Li]
40
20
229
0
180
R. I. (%)
100
80
190
200
210
220
230
b
240
250
243
270
263 262
223
60
260
224
40
229
20
0
180
R. I. (%)
100
80
190
200
210
220
230
240
250
260
257
c
270
275
276
238
237
60
243
40
20
0
190
R. I. (%)
100
80
200
210
220
230
240
250
260
270
280
266
d
60
40
265
303 304
285
271
20
0
220
230
240
250
260
270
280
290
300
310
m/z
Figure 5.4.
Spectra generated by CID (MS/MS) of Li+ cationized, fully
deuterium-exchanged: (a) AAAG, (b) AA(βA)G, (c) AA(γAbu)G and
(d) AA(Cap)G
89
R. I. (%)
100
a
80
325
+
[b3-1+Ag]
60
+
[a3-1+Ag]
40
326
297
20
0
290
R. I. (%)
100
300
310
320
b
80
330
340
350
360
350
360
325
60
326
40
20
345
0
290
R. I. (%)
100
300
310
320
c
80
330
339
340
340
60
40
359
20
0
300
R. I. (%)
100
80
310
320
330
340
d
350
360
370
368
60
367
40
20
387
0
330
340
350
360
370
380
390
400
m/z
Figure 5.5.
Spectra generated by CID (MS/MS) of 109Ag+ cationized, fully
deuterium-exchanged: (a) AAAG, (b) AA(βA)G, (c) AA(γAbu)G and
(d) AA(Cap)G.
90
100
(a)
234
(b3-1+Li)+
80
R. I. (%)
(b3+17+Li)+
252
60
(y3-1+Li)+
40
238
20
0
230
100
234
236
238
240
242
244
246
(b)
80
R. I. (%)
232
60
248
250
252
254
(b3+17+Li)+
254
(b3-1+Li)+
235
236
(y3-1+Li)+
40
240
20
0
232
234
236
238
240
242
244
246
248
250
252
m/z
Figure 5.6
CID (MS/MS) spectra of Li+ cationized (a) AA(γAbu)G and
(b) AA(α-d2-γAbu)G.
91
254
256
100
(a)
(b3-1+Ag)+
80
R. I. (%)
336
60
40
(b3+17+Ag)+
354
20
0
332
100
336
(b)
80
R. I. (%)
334
337
338
340
342
344
346
348
350
352
354
356
338
(b3-1+Ag)+
(b3+17+Ag)+
60
356
40
20
0
334
336
338
340
342
344
346
348
350
352
354
m/z
Figure 5.7.
CID (MS/MS) spectra of Ag+ cationized (a) AA(γAbu)G and
(b)AA(α-d2-γAbu)G.
92
356
358
100
216
(a)
190
80
R. I. (%)
-H2O
-44
60
-71
40
-113
121
20
234
149
0
100
-85
(b3-1+Li)+
163
125
150
175
200
225
250
251
(b)
-85
R. I. (%)
80
-113
223
60
-43
40
293
-H2O
336
318
20
0
(b3-1+Ag)+
225
250
275
300
325
350
m/z
Figure 5.8.
CID (MS3) spectra of (a) (b3-1+Li)+Ag+ and (b) (b3-1+Ag)+ derived
from AA(γAbu)G.
93
AA(α-d2-γAbu)G+cat+
H
N
H2N
O
D
N
H
O
D
H
N
O
cat+
OH
O
H2N
H N
+
D
D
NH
H
O
:
NH
cat+ :O:
N
O
H
O
H2N
cat+
:
-
O
H :O:
N
O
H
+
D
N
D
NH
O
O
NH
O
N
N
:O:
O
H
+
N
H
OH
D
D
H2N
O
H2N
O
OH
(b3-1+cat)+
NH
O
OH
N
Scheme 10
cat+
O
N
H
D
D
O
cat+
:
O
OH
OH
H2N
H2N
O
O
cat+
D
D
Pathway to (b3-1+cat)+ from metal cationized AA(α-d2-γAbu)G
through 7-membered cyclic intermediate, with transfer of amide
position H atom
94
AA(α-d2-γAbu)G+cat+
O
H
N
H2N
D
N
H
O
D
cat+
O
H
N
OH
O
H2N
:
cat+ :O:
NH
N
O
H
H2N
X
D
O
NH
O
O
O
H N
O
N
H
+
cat+
OH
D
H
N
H
O
Scheme 11.
N
D
O
D
N
H
H
N+
D
O
OH
:O:
cat+
OH
D
N
H
O
-
:
O
O
H2N
OH
D
D
H N
H2N
H
N
A
O
cat+
C
O
(b3-1+cat)+
Potential pathway to (b3-1+cat)+ from metal cationized AA(α-d2γAbu)G through 7-membered cyclic intermediate, with transfer of αcarbon position H atom.
95
AA(α-d2-γAbu)G+cat+
O
H
N
H2N
N
H
O
:
A’
:O:
O
H
N
-
D
N
H
H
N+
O
OH
D
:O:
-
:
O
cat+
O
H
N
H2N
OH
D
cat+
H2N
O
H
+ N
:
O
OH
O
D
N
H
cat+
O
H
N
D
O
H
N
H2N
D
D
N
H
O
OH
:O:
-
:
O
H
N+
D
cat+
O
H
B
H
N
H2N
O
D
N
H
O
OH
N
D
cat+
C
O
(b3-1+cat)+
Scheme 12.
Pathway to (b3-1+cat)+ from metal cationized AA(α-d2-γAbu)G
through alternative ketene mechanism.
96
CHAPTER VI
CONCLUSIONS
The objective of the first set of experiments in this study was to investigate the
CID patterns for thallium(I) cationized peptides and compare them to those from Ag, Na
and protonated versions. In comparison to H+, Na+, and Ag+, we found that lower
desolvation temperatures and milder ionization conditions were required to produce
abundant (M+Tl)+ ions, an observation that suggested weak binding of Tl+ by the entire
suite of peptides. This conclusion was supported by appearance of prominent Tl+ signals
in the CID spectra of each peptide. In addition, the activation amplitudes required to
induce fragmentation (ca. 20-25%, 25-30% and 30-35% for Tl, Ag, and Na respectively)
of Tl cationized peptides were comparable to those for Ag cationized versions and lower
than Na cationized analogues. As a result of tendency to eliminate Tl+ during CID, use of
MS/MS experiments to determine sequence from the C-terminus was not possible.
The objectives of the second set of experiments was to investigate the influence of
4-Aminomethylbenzoic acid on the formation of the rival (b3-1+cat)+ and (b3+17+cat)+
species. Generation of these product ions involves cleavage of the same amide bond, but
through entirely different mechanisms. For model peptides with general formula
A(4Ambz)AX and A(4Ambz)GX, our study shows that formation of the (bn-1+cat)+
species is favored over the (bn+17+cat)+ species when the amide bond cleaved is one
residue to the C-terminal side of the 4Ambz residue. As proposed in our earlier study of
the effect of “alternative” amino acids on the CID of protonated peptides, it is likely that
the influence of the 4AMBz residue observed here with the metal cationized peptides can
97
be attributed to generation of a highly-conjugated oxazolinone species as (b3-1+cat)+,
which would increase the stability of this product relative to the rival (b3+17+cat)+ ion.
Energy-resolved CID, which measures the collision energies necessary to generate
product ions from a particular precursor ion, shows that formation of (b3-1+cat)+ occurs
at significantly lower collision energies when the 4Ambz residue is present to affect the
reaction pathway. This result is suggestive of an influence of the aromatic residue on the
activation energy of the reaction to produce (b3-1+cat)+. Placing residues with bulky side
groups at the C-terminus, which are known to enhance formation of (b3+17+cat)+, caused
a modest increase in relative abundance of the product. However, generation of
(b3-1+cat)+ remained the dominant dissociation pathway. The relative energies of
precursor and product ions, and of transition states and relevant intermediates have yet to
be determined, and would benefit from a comprehensive computational study.
In the last set of experiments, the effect of the “alternative” βA, γAbu and Cap
residues on the tendency to form (b3-1+cat)+ ion from metal cationized peptides was
examined. For protonated versions of these peptides, the larger amino acids suppress
formation of b3+ from peptides with general sequence AAXG, presumably due to the
prohibitive effect of larger cyclic intermediates in the “oxazolone” pathway. However,
we found that abundant (b3-1+cat)+ products are generated from Li+, Na+ and Ag+
cationized versions of AAXG, despite the potential negative effects of the large cyclic
intermediates. Using a group of D-labeled peptides, we found that formation of
(b3-1+cat)+ from the AAXG peptides involves transfer of both amide position H atoms, as
well as those attached to C atoms. The most likely source of the H atom transferred is
identified as the α-carbon position using AA(γAbu)G for which the α-C position of the
98
γAbu residue was labeled with D. To account for the unusual transfer of the α-C position
H atom, and formation of (b3-1+cat)+ from the AAXG peptides, a mechanism was
proposed that involves tautomerization at the X-G amide group and bond cleavage to
furnish a metal-cationized ketene. Unlike the “oxazolone” mechanism used to explain
generation of conventional bn+ and (bn-1+cat)+ ions, formation of the ketene does not
involve cyclization and nucleophilic attack by a carbonyl O atom, and thus is less likely
to be hindered by the large cyclic intermediates necessary for formation of oxazolinone
products.
99
REFERENCES
100
LIST OF REFERENCES
1. Biemann K, Biomed Environ. Mass Spectrom. 1988; 16: 99-111
2. Roepstorff P, Fohlman J. Biomed Environ. Mass Spectrom. 1984; 11: 601.
3. Jones J L, Dongre A R, Somogyi A, Wysocki V H. J. Am. Chem. Soc. 1994; 116:
8368-8369.
4. Dongre A L, Jones J L, Somogyi A, Wysocki V H, J. Am. Chem. Soc. 1996; 118:
8368-8374.
5. Tsaprailis G, Nair H, Somogyi A, Wysocki V H, Zhong W, Futrell J H,
Summerfield S G, Gaskell S J. J. Am. Chem. Soc. 1999; 121: 5142-5154.
6. Wysocki V H, Tsaprailis G, Smith L L, Breci L A. J. Mass Spectrom. 2000; 35:
1399-1406.
7. Tsang C W, Harrison A G. J. Am. Chem. Soc. 1976; 98: 1301-1308.
8. Harrison A G, Yalcin T. Int. J. Mass Spectrom. Ion Processes 1997; 165: 339347.
9. Burlet O, Yang C Y, Gaskell S J. J. Am. Chem. Soc. Mass Spectrom. 1992; 3:
337-344.
10. Cox K A, Gaskell S J, Morris M, Whiting A. J. Am. Soc. Mass Spectrom. 1996; 7:
522-531.
11. Summerfield S G, Whiting A, Gaskell S J. Int. J. Mass Spectrom. Ion Process.
1997; 162: 149-161.
12. Summerfield S G, Cox K A, Gaskell S J. J. Am. Soc. Mass Spectrom. 1997; 8: 2531.
13. Tang X, Boyd R K. Rapid Comm. Mass Spectrom. 1992; 6: 651-657.
14. Tang X, Thibault P, Boyd R K. Anal. Chem. 1993; 65: 2824-2843.
101
15. Csonka I P, Paizs B, Lendvay G, Suhai S. Rapid Commun. Mass Spectrom. 2000;
14: 417-427.
16. Paizs B, Csonka I P, Lendvay G, Suhai S. Rapid Comm. Mass Spectrom. 2001;
15: 637-647.
17. Paizs B, Suhai S. Mass Spectrom. Rev. 2004; 24: 508-548.
18. Papayannopoulos I A. Mass Spectrom. Rev. 1995; 14: 49-73.
19. Yalcin T, Khouw C, Csizmadia I G, Peterson M R, Harrison A G. J. Am. Soc.
Mass Spectrom. 1995; 6: 1165-1174.
20. Yalcin T, Csizmadia I G, Peterson M R, Harrison A G. J. Am. Soc. Mass
Spectrom. 1996; 7: 233-242.
21. Nold M J, Wesdemiotis C, Yalcin T, Harrison A G. Int. J. Mass Spectrom. Ion
Processes.1997; 164: 137-153.
22. Paizs B, Lendvay G, Vekey K, Suhai S. Rapid Comm. Mass Spectrom. 1999; 13:
525-533.
23. Harrison A G, Csizmadia I G, Tang T H. J. Am. Soc. Mass Spectrom. 2000; 11:
427-436.
24. Paizs B, Suhai S. Rapid Comm. Mass Spectrom. 2002; 16: 375-389.
25. Paizs B, Suhai S. J. Am. Soc Mass Spectrom. 2004; 15: 103-112.
26. Polfer N C, Oomens J, Suhai S, Paizs B. J. Am. Chem Soc. 2005; 127: 1715417155.
27. Barr J M, Van Stipdonk M J. Rapid Comm. Mass Spectrom. 2002; 16: 566-578.
28. Anbalagan V, Patel J N, Niyakom G, Vanstipdonk M J. Rapid Comm. Mass
Spectrom. 2003; 17: 291-300.
29. Anbalagan V, Silva A T, Rajagopalachary S, Bulleigh K, Talaty E R, Van
Stipdonk M J. J. Mass Spectrom. 2004; 39: 495-504.
102
30. Anbalagan V, Bulleigh K, Angela H, Do T, Qun Wu, Van Stipdonk M J. Rapid
Comm. Mass Spectrom. 2006; 20: 227-232.
31. Tang X, Ens W, Standing K G, Westmore J B. Anal. Chem. 1998; 60: 1791.
32. Renner D, Spiteller G. Biol. Environ. Mass Spectrom. 1988; 15: 75.
33. Grese R P, Cerny R L, Gross M J. J. Am. Chem. Soc. 1989; 111: 2835.
34. Glish L T, Payne A H, Glish G L. J. Am. Soc Mass Spectrom. 2001; 12: 497-504.
35. Anbalagan V, Perera B A, Silva A T, Gallardo A L, Barber M, Barr J M, Terkali
S M, Talaty E R, Vanstipdonk M J. J. Mass Spectrom. 2002; 37: 910.
36. Teesch L M, Orlando R C, Adams J. J. Am. Chem Soc. 1991; 113: 3668-3675.
37. Chu I K, Gou X, Lau T C, Siu K W M. Anal. Chem . 1999; 71: 2364.
38. Grese R P, Gross M L. J. Am. Chem Soc. 1990; 112: 5098.
39. Talaty E R, Travis J C, Sandra O, Van Stipdonk M J. Rapid Comm. Mass
Spectrom. 2006; 20: 3443-3455.
40. Mass Spectrometry Resource, University of Bristol. www. Chm.bris.uk.
41. Mclafferty F W, Turecek F, Inter. Mass Spectra 4th Edition ( May 1993).
42. Dole M, et al. J. Chem Phys. 1968; 49: 2240.
43. Fenn J. B, et al. J. Phys. Chem. 1984;88: 4451.
44. Paul W Agewandte chemie, International edition, 29, 1990, 739
45. Raymond E M. J. Mass Spectrom. 1997; 32: 351-369.
46. Randall E P. Anal. QMS Extrel Quad. Mass Spectrom. 575 epsilon.
47. Lekha S, Volmer D A. J. Mass Spectrom. 2004; 39: 1091-1112.
103
48. Lin T, Glish G L. Anal. Chem. 1998; 70: 5162-5165.
49. Thorne G C, Ballard K D, Gaskell S J. J. Am. Soc. Mass Spectrom.1990; 1: 249251.
50. Gonzalez J, Besada V, Garay H, Reyes O, Padron G, Tambara T, Takao T,
Shimonishi Y. J. Mass Spectrom. 1996; 31: 150-156.
51. Chu I K, Shoieb T, Guo X, Rodriquez C F, Lau T C, Hopkinson C F, Lau T C,
Hopkinson A C, Siu K W M. J. Am. Soc Mass Spectrom. 2001; 12: 163-175.
52. Feng W Y, Gronert S, Fletcher K A, Warres A, Lebrilla C B. Int. J. Mass
Spectrom. 2003; 222: 117-134.
53. Fmoc Solid Phase Peptide Synthesis- A practical Approach, Chan W C, White P
D. Eds. Oxford University Press Inc, New York, 2000.
54. Balta B, Aviyente V, Lifshitz C. J. Am. Soc Mass Spectrom. 2003; 14: 1192-1203.
55. Reid G E, Simpson R J, O’Hair R A. J. Am. Soc. Mass Spectrom. 1998; 9: 945956.
56. Talaty E R, Cooper T J, Piland D L, Bateman D J, Syed A, Stevenson W, Van
Stipdonk M J. Rapid Comm. Mass Spectrom. 2006; 20: 3007-3017.
57. McClellan J E, Murphy III J P, Mulholland J J, Yost R A. Anal. Chem. 2002; 74:
402-412.
58. Murphy J P III, Yost R A. Rapid Comm. Mass Spectrom. 2000; 14: 270-273.
59. Vachet R W, Hartmann J A R, Callahan J H. J. Mass Spectrom. 1998; 33: 12091225.
60. Gabelica V, De Pauw E. Mass Spectrom. Rev. 2005; 24: 566-587.
61. Vekey K. J. Mass Spectrom. 1996; 31: 445-463.
62. Parcell, RF. Chem. Ind. (London), 1963, 1396
104
63. Cao H, Xiong Y, Wang T, Chen B, Squier TC, Mayer MU. J. Am. Chem. Soc.
2007, 129, 8672- 8673.
64. Finley KT. Chem. Rev. 1964; 64: 573-589.
65. Schaefer JP, Bloomfield JJ. Org. React. 1967; 15: 1-203.
66. Cooper TJ, Talaty ER, Van Stipdonk MJ. J. Am. Soc. Mass Spectrom. 2005; 16:
1305-1310.
67. Osburn S, Ochola S, Talaty E, Van Stipdonk MJ, Rapid Comm. Mass Spectrom.
2007; 21: 3409-3419.
68. Paizs B, Suhai S, Rapid Comm. Mass Spectrom. 2001; 15: 651-662.
105