.7,
:( ..

... {in

‘6 ‘c .v'\
5.3.. aw
.

r 5 p
.. than...
humus... ......

than... .

ain‘ts‘ a
1:91.) ‘ .1
3|. .1 \ 3.x
1.0.1.3. .
.17 . 1:1.

 

\A..

955:.‘ I...
y. 1:4.
s.

 

 

 

MICHIGANS

IIIIIIIIIIIIII'IIII III III IIIIIIIIIII

1659

II

 

 

 

 

 

 

 

 

 

 

This is to certify than the
dissertation entitled
The characterizatim of thermal degradation of peptides
utilizing K+IJB — K+ ionization of desorbed species
and an investigation of collision—induced dissociation (CID)
of deprotonated peptides

presented by
Huaiqin Wu

has been accepted towards fulfillment
of the requirements for

Ph.D. degree in Chemistry

mfié flax

/ Major professor

Date m 45" 77

 

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

THE CHARACTERIZATION OF THERMAL DEGRADATION OF PEPT IDES
UTILIZING K+IDS - K1“ ION IZATION OF DESORBED SPECIES
AND
AN INVESTIGATION OF COLLISION-INDUCED DISSOCIATION (CID)
OF DEPROTONATED PEPTIDES

By

Huaiqin Wu

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1994

ABSTRACT

THE CHARACTERIZATION OF THERMAL DEGRADATION OF PEPTIDES
UTILIZING K+IDS - K+ IONIZATION OF DESORBED SPECIES
AND
AN INVESTIGATION OF COLLISION -INDUCED DISSOCIATION (CID)
OF DEPROTONATED PEPT IDES

By

Huaiqin Wu

The desorption/ionization technique, K+IDS, was developed in 1984 at MSU. In
this work, the K+IDS technique was further developed and characterized. A new probe
was designed and constructed; a new material was used for the thermionic emitter; local
pressure in the ion source chamber was estimated; the stabilization theory for K+-adducts
was re-evaluated.

Thermal degradation of peptides was characterized utilizing the K+IDS technique.
Charge is not involved in this process; thus it can be treated as the ultimate model for
charge-remote fragmentation, one of the processes proposed for CID of protonated and
deprotonated peptides. Our understanding of thermal degradation of peptides may
provide insights into the C11) process of charged peptides.

The study is based on the use of deuterated, derivatized and specifically-designed
peptides, to trace the source of the shifting hydrogen atom in thermal degradation. It is
concluded that thermal degradation of peptides, at a temperature about 600K, proceeds

via the lowest energy process - cyclization involving their termini.

A new phenomenon - the discrimination in the formation of K+ adducts of
thermal degradation products - was uncovered in this study. It is found that there is a
discrimination against small and cyclic molecules, which are weakly bonded to K+ ions.

In studying fragmentation of charged species, CID of deprotonated peptides was
investigated. Derivatization/charge-localization and deuteration are the major tools in
discovering the sites of deprotonation as well as the fragmentation pathways. The amides
are found to be very important sites — for the first time - in the formation of fragment ions.
Mechanisms are proposed for the formation of the major fragment ions from both regular
and charge-localized peptides. The results obtained are found to be different from those
reported in the literature for dipeptides. The applicability of CID-MS/MS of
deprotonated peptides for peptide sequencing is discussed, as is the low sensitivity of

both MS and MS/MS in negative ion mode.

Dedicated to my mother
and
in memory of my father

Baoshi and Dekui Wu

iv

ACKNOWLEDGMENTS

I would like to express my sincere gratitude to my mentor, Dr. John Allison, for his
guidance and constant support during my graduate study at MSU. especially for the time
and effort in correcting this manuscript and in coaching the seminars. My gratitude also
goes to Dr. Watson for correcting this dissertation, and for his encouragement and support;
to Dr. Leroi and Dr. Reusch for serving on my committee; and to Dr. Hiller for his many

advice about job hunting and for reading this dissertation.

To all the Allison group members, Karen, Jason, Kurt, Kate, Guy, Paochi, Gabi, Tom,
and Garys, thank you all for your help and your friendship during my stay at MSU.

I also owe many thanks to all members of the Mass Spectrometry Facility: Dr. Gage, Dr.
Huang, Mike, Mel, Melinda and Bev. Without their technical support and friendship, my

study here would have been extremely difficult.

And last, but not the least, I would like to thank my wife, Shudong, my mother, brother

and sisters, and to all my relatives and friends for their loving, caring and supporting

during my graduate study at Michigan State University.

 

TABLE OF CONTENTS

List of Tables ........................................................................................ ix
List of Figures ..................................................................................... xiii
Chapter 1 Introduction and ResearCh Objectives ................................................ 1
I. Introduction - A Historical Review of Peptide Sequencing .............................. 1
11. Research Objectives ........................................................................... 6
A. Thermal Degradation of Peptides ..................................................... 7

B. Fragmentation of Deprotonated Peptides ............................................. 9
References ............................................................................................ 10
Chapter2 K+IDS: The Technique and Its Further Characterization ......................... 12
I. Introduction ................................................................................... 12
11. Experimental Section ......................................................................... 13
A. Experiment Setup and K+IDS Probe ................ ‘. ............................... 13

B. The Making of a Thermionic Emitter ................................................ 17

III. K+IDS: Operational Principles .............................................................. 20
A. Thermionic Emission .................................................................. 20

B. Desorption Vs. Degradation .......................................................... 21

C. K"'-Adduct Formation and Stabilization ....... , ..................................... 24

IV. (M+K)'*' Stability Studied by Using a Directly Heated K+IDS Probe ................... 26
A. Characterization of a Novel Direct-Heating K+IDS Probe ........................ 26

vi

B. (M+K)+ Stability Studied by Direct Heating ....................................... 27

References ............................................................................................ 31
Chapter 3 Peptide Analysis by K+IDS Mass Spectrometry ................................... 32
I. Introduction ..................................................................................... 32
II. Experimental Section .......................................................................... 34
III. K+IDS Mass Spectra of Typical Peptides .................................................. 35
A. Nomenclature of K+IDS Fragments ................................................ 35

B Typical K+IDS Spectra of Peptides ................................................. 37,

C. K+IDS Spectra of Polar Peptides and Dehydration Reactions ................... 40

D K+IDS Spectra ofPeptide Derivatives......................................... ..... 48

1. Peptide Amides .................................................................. 48

2. N-Terminally Derivatized Peptides ............................................ 51

3. C-Terminal Esterification ....................................................... 55

4. Peptide Variants ................................................................. 55

E. Internal Ions from Peptides ........................................................... 61

IV. Peptide Sequencing by K+IDS Mass Spectrometry ....................................... 61
References ............................................................................................ 65

Chapter 4 The Characterization of the Primary Thermal Degradation Processes of Peptides

I. Introduction ..................................................................................... 66
II. Experimental Section .......................................................................... 67
III. The Primary Thermal Degradation Processes of Peptides ................................ 68
A. Deuterium-Labelin g Experiments .................................................... 70
B Peptide Derivatization - Evidence for Cyclization .................................. 77
C. Termini,Sidechainsand StericEffects........................... ................... 83
D Thermodynamic Aspects of the Thermal Degradation of Small Peptides ....... 86

vii

IV. Conclusion ...................................................................................... 90

References ............................................................................................ 91
Chapter 5 Response Discrimination in K+-Adduct Formation ............................... 92
I. Introduction ..................................................................................... 92
11. Experimental Section .......................................................................... 93
III. Binding Energies and Stabilities of K+-addums ........................................... 94
IV. Discrimination in the Formation of K+-adducts ............................................ 96

A. Size Discrimination .................................................................... 97

B Structure Discrimination ............................................................ 105
C. Discrimination Studied by Thermal Degradation of Tri- and Tetra-peptidesz. 107
D

Discrimination Against Weak Binding:

Experiments with Mixed Ionic Emitters ........................................... 108

E. Summary (Discrimination in Interpretation of K+IDS Spectra) ................ 112
References .......................................................................................... 113
Chapter 6 Collision-Induced Dissociation of Deprotonated Peptides ...................... 114
I. Introduction ................................................................................... 114
II. Tandem Mass Spectrometry ................................................................ 116
A. Introduction ........................................................................... 116

B. CID-B/E Linked Scans .............................................................. 117

III. Experimental Section ........................................................................ 119
IV. Deprotonation of Peptides .................................................................. 124
A. Deprotonation in the Gas-phase or the Liquid-phase? ........................... 124

1. In Zwitterionic Form or in Neutral Form? ................................. 127

2. Thermodynamic Support for Gas-phase Deprotonation ................. 128

viii

B. The Site of Deprotonation of Peptides ............................................. 131

l. Deprotonation at a—carbon Position? ....................................... 133
2. Deprotonation at Amides or at a—carbon? ................................. 134

3. Why Is Deprotonation of Amides Favored over That of a—carbon?.... 135

4. Charge-localization: Evidence for Amide Deprotonation ................ 137

5. Final Comments on Site of Deprotonation ................................. 139

V. Fragment Ions from CID-MS/MS Spectra of Deprotonated Peptides ................. 139
A. Nomenclature ......................................................................... 139

B. Common Fragment ions ............................................................ 141

C. Ion Structure and Stability .......................................................... 142

VI. Mechanisms for the Fragmentation of Deprotonated Peptides .......................... 146
A. H-Shift: Mechanistic Study by Using Deuterated Peptides .................... 147

B. CID Fragments from Derivatized and Underivatized Peptides ................. 148

VI]. Mechanisms: The Formation of Major Fragment Ions ................................ 151
A. Ions from Amide-Deprotonated Peptides ......................................... 151

B. Ions from C-terminally Deprotonated Peptides ................................... 156

VIII. Fragmentation of Charge-localized Peptides ............................................ 160
A. Charge-localization on Peptides .................................................... 160

B. Fragmentation of Deprotonated Peptides with Charge-localization ............ 163

l. Fragment Ions of C-terminally Charge-localized Peptides ............... 164

2. Fragment Ions of N -terminally Charge-localized Peptides ............... 167

IX. Sequencing with Negative Ion FAB-B/E Linked Scan .................................. 170
A. Peptide Sequencing with CID-MS/MS of negative ions ........................ 170

B. Why is Negative Mode not Favorable? ............................................ 177

X. Conclusions .................................................................................. 184
References .......................................................................................... 186

ix

Appendices ...................................................................................... 188

Appendix I. Local Pressure at the Sample Holder in the Ion Source

and Collision frequency in K+IDS experiments. ............................. 188
Appendix II. Enthalpies of Thermal Degradation of Peptides

from Model Compounds ........................................................ 198

Figure 2.1

Figure 2.2

Figure 2.3
Figure 2.4

Figure 3.1

Figure 3.2
Figure 3.3

Figure 3.4
Figure 3.5

Figure 3.6

Figure 3.7
Figure 3.8

LIST OF FIGURES

a) Schematic of a dual-filament K+IDS probe; ............................... 14
b) Schematic diagram of the K+IDS instrument .............................. 15
c) Schematic of a K+IDS experiment in operation ........................... 16
K+IDS mass spectrum of a mixture of P130200 and PPG725 .............. 18

PEG = poly(ethylene glycol) [H-(OCHzCH2)-OH];

PPG = Poly(propylene glycol) [H-[OCH(CH3)CH2]-OH}.

Temperature dependence of thermal degradation vs. desorption ............ 22
K+IDS spctra of H-YGGFM-NHZ:

a) with the orginal probe, I = 2.70 A;

b) with the new probe. I(sample) = 2.0 A,

direct heating delay = 3 seconds ................................................ 28

1,2 - Elimination mechanism for the CO-NH bond cleavage of peptides. 33
Nomenclature for peptide fragments and ions ................................. 36

K+IDS mass spectra of H-FGGF-OH.

a) Time dependence. b) Average spectrum .................................. 39
K+IDS mass spectra of H-VGVAPG-OH ..................................... 41
K+IDS mass spectra of H-FLEEl-OH ......................................... 44
K+IDS mass spectra of H-PFGK-OH ......................................... 45
K+IDS mass spectra of H-RKEVY-OH ....................................... 47

K+IDS mass spectra of: a) H-YGGFL-OH; b) H—YGGFLNHZ .......... 49

xi

Figure 3.9
Figure 3.10
Figure 3.11

Figure 3.12

Figure 3.13
Figure 3.14

Figure 3.15

Figure 4.1
Figure 4.2

Figure 4.3

Figure 4.4

Figure 4.5
Figure 4.6
Figure 4.7
Figure 4.8
Figure 4.9
Figure 4.10

K+IDS mass spectra of: a) H-GGFL-OH; b) H-GGFLNHz ............ 50

K+IDS mass spectrum of Ac-VGVAPG-OH. ................................ 52
K+IDS spectra of: a) CBZ-GPGGPA-OH; b) CBZ-GPFPL-OH; ....... 53
a) CBZ-GPGGPA-OH; C) t-BOC-PPPP-OH ................................ 54

K+IDS mass spectra of:

a) H-YGGFL—OCH3; b) H-VGVAPG-OCH3 ............................... 56
K+IDS mass spectra of: a) H-VGGFL-OH; b) H-GGVFL-OH ......... 57
K+IDS mass spectra of:

a) H-AGGFL-OH; b) H-(B-A)GGFL-OH. ................................. 58
a) H-AGGFL-OH; c) H—AG(Sar)FL-OH. ................................... 59
K+IDS spectrum of a 22-mer peptide melittin (MW 2846.7). The sequence
of the peptide is H-GIGAVLKVL’I'ITGLPALISWIKRKRQQ-NHZ ..... 64

Schematic diagram of 1,2 - elimination mechanism. ......................... 69
K+IDS spectrum of:

a) H-GP-OH that had been exposed to D20;

b) H-FM-OH that had been exposed to D20 ................................. 71
K+IDS spectra of variants of H-VGVAPG-OH.

a) unlabeled; b) peptide after exposure to D20 .......... : ..................... 75
K+IDS spectra of a-incorporated peptides.

a) H-VGVAadPo-OH; b) H-deodeApo—OH .......................... 76
K+IDS spectrum of CH3CO-VGVAPG-OH .................................. 79
Thermal degradation of peptides ................................................ 80
K+IDS spectra of a) H-AGGFL-OH and b) H-(B-A)GGFL-OH ........... 82
K+IDS spectra of a) H-YGGFL—OH and b) H-VGGFL—OH. .............. 84
K+IDS spectrum of H-GGVFL-OH. .......................................... 85
Thermal degradation pathways of peptides .................................... 87

xii

Figure 5.1
Figure 5.2
Figure 5.3

Figure 5.4

Figure 5.5
Figure 5.6

Figure 5.7

Figure 5.8

Figure 6.1

Figure 6.2

Figure 6.3

Figure 6.4

Figure 6.5

Figure 6.6

Relative responses of five amino acids (a mixture of alanine,

valine, leucine, threonine and phenylalanine) in K+IDS. ................... 98
Relative responses of alanine and phenylalanine in K+IDS. ................ 99
Relative responses of alanine and glycine in K+IDS ........................ 100

Relative responses of alanine, H-AA-OH and H-AAA-OH

‘at m/z 128, 199 and 270. ....................................................... 102

Relative responses of H-GGGG-OH and -H-AAA-OH in K+ IDS ........ 104
Mass chronograms showing the relative responses for molecules

H-GP-OH and cyclo-GP in the K+1DS experiments ........................ 106
Relative responses of K+ and Na+ to a peptide H-VGGFL-OH

and its degradation products .................................................... 110
Relative responses of K+ and Cs“' to a peptide H-VGGFL—OH and its

degradation products. ........................................................... 111

The CID linked scan spectra for H-VGGFL-OH with precursors set at:

a) m/z 490.2, (M-H)‘; b) m/z 489.2. ....................................... 120
a) m/z 490.2, (M-H)‘; c) m/z 491.2 ......................................... 121
Deprotonation of peptides in glycerol matrix with FAB. ................... 125

CID-BIB linked scan spectrum of deprotonated peptides (derivatives):

a) H-VGGFL-AMSA; b) H-VGGFL-ABSA. ............................. 129
Kinetic reasoning for favored deprotonation at amides

over that of a-carbon positions. ............................................... 136
CID-B/E linked scan spectrum of deprotonated peptide

SBA-VGGFL-OH ............................................................... 138
a) Roepstorff nomenclature for peptide fragments;

b) Common fragment ions from deprotonated peptides .................... 140

xiii

Figure 6.7

Figure 6.8

Figure 6.9

Figure 6.10

Figure 6.11
Figure 6.12

Figure 6.13

Figure 6.14
Figure 6.15

Figure 6.16

Figure 6.17

Figure 6.18

Figure 6.19

Figure 6.20

CID-B/E linked scan spectra of deprotonated peptides:

a) H-VGVAPG-OH; b) H-VGVAadPG-OH; ............................. 143
a) H-VGVAPG-OH; c) H—VO‘deadAPG-OH ........................... 144
CID-B/E linked scan spectra of deprotonated peptides:

a) H-VGGFL-OCH3; b) Ac-VGGFL-OCH3 ............................... 150
The formation of Yn-1 and Cn-1 ions via 1,3-elimination mechanisms:

a) charge-induced fragmentation; b) charge-remote fragmentation ...... 152
The formation of Yn and Cn ions via 1,3-elimination mechanisms
(charge-induced fragmentation) ................................................ 153
Cu and Yn ions from displacement reactions ................................. 154
CID-B/E linked scan spectrum of deprotonated peptides:

a) H-YGGFL-OH; b)H-YGGFL-NH2 ...................................... 157
The formation of C ions from C-terminally deprotonated peptide amides.
(H-YGGFL-NHz as an example). ............................................ 158
The formation of Cn ions via a 1,4-elimination mechanism ................ 159
The formation of Y ions from chargeoremote processes

such as E15 reactions shown here ............................................ 161
CID-B/E linked scan spectrum of a deprotonated peptide
H-VHLTEVEK-OH ............................................................. 173
CID-B/E linked scan spectrum of a deproronated peptide
H-RPPGFSPF-OH .............................................................. 174
CID-B/E linked scan spectra of deprotonated peptides:

a) H-FLEEI-OH; b) H-RKDVY-Ol-I ........................................ 176
CID-B/E linked scan spectrum of a deprotonated peptide
H—WAGGDASGE-OH. ........................................................ 178
CID-B/E linked scan spectrum of a deprotonated peptide
H-RVYIHPF-OH ................................................................ 179

xiv

Table 5.1
Table 5.2
Table 6.1
Table 6.2
Table 6.3
Table 6.4

Table 6.5

LIST OF TABLES

Binding energies and dipole moments of various molecules. ............... 94
Relative responses factors, Rrel. for amino acids and small peptides ..... 101
Negative ion FAB responses of a peptide and its derivatives. ............. 126
Gas-phase acidities of various functionalities in peptides ................... 132

Relative intensities of (M-l)‘ and (M-2)’ ions for (deuterated) peptides. . 133
Relative intensities of (M-1)' and (M-2)‘ ions
for (deuterated) peptide esters .................................................. 135

Fragment ions from peptides and their deuterated analogs. ................ 147

XV

Chapter 1

Introduction and Research Objectives

1. Introduction - Peptide Sequencing: Historical Review

The utility of mass spectrometry for peptide sequencing, the elucidation of the
primary structure of peptides (amino acid sequence), was established in the late 19508 [1].
For many years in the following two decades, this sophisticated technique was
demonstrated only in a few highly skilled laboratories, and limited to the analysis of very
small peptides. When fast atom bombardment, FAB [2], was introduced in 1981,
explosive growth in new methods and applications began, because the requirements for
sample volatility were eliminated. Advances in high mass-range instrument development,
especially those made the late 19803, as well as the evolution of electrospray [3] and
matrix-assisted laser desorption/ionization (MALDI) [4], made mass spectrometry
applicable to an ever-increasing number of peptides.

With traditional techniques for peptide sequencing, analysis involves sequential
degradation of the molecule from one end or the other, removing and identifying the amino
acids one at a time. Compared with techniques such as the Edman degradation method [5],
sequencing with mass spectrometry has many potential advantages such as high sensitivity
and high speed. In addition, mass spectrometry can solve many problems that defy Edman
degradation. For example, N-terminally blocked peptides can be sequenced with mass
spectrometry; post-translational modification of protein/peptides can be identified with

mass spectrometry; and peptide mixtures can be analyzed by mass spectrometry.

Furthermore, mass spectrometry can provide precise, accurate confirmation on the
molecular mass of proteins derived from recombinant techniques in molecular biology.

Sequencing a peptide by the recording of a single mass spectrum is a very attractive
proposition which has stimulated considerable research [6]. In fact, as an aid to
microsequencing (automated Edman Degradation), mass spectrometry can improve both the
efficiency and the accuracy of the operation. One of the greatest contributions of mass
spectrometry to protein structure determination has been the ability to obtain precise,
accurate molecular weights of peptides.

Over the last two decades mass spectrometry has advanced to the point where it has
become one of the most broadly applicable tools in analytical chemistry. The rapid pace of
this evolution has been spurred by periodic advances in mass spectrometer hardWare,
computer interfacing, new extended mass spectrometric methods (e.g., tandem mass
spectrometry), interfacing with separation science methods, mass range extension into the
millions of Daltons, and an increasing performance/cost ratio.

In mass spectrometry, the ion formation/ionization process is the starting point for
mass spectrometric analysis. The ionization method used dictates the scope and utility of
the methods in peptide sequencing analysis. The development of various ionization
techniques over the years virtually reflected the advance of mass spectrometry techniques.

The most common ionization technique in mass spectrometry is electron impact
(EI), which is well characterized and understood [7]. This technique has been traditionally
used for the analysis of volatile species or those that can be easily derivatized to some
volatile species. For this reason, the earliest and, for many years, the only way to obtain
the mass spectra of peptides was by making volatile derivatives of small peptides. Two
principal derivatives are often made: polyamino alcohols [8] and N-acetyl-N,O,S-
permethylated peptides [9].

In recent years, desorption/ionization (DI) techniques have become increasingly

popular for analyzing nonvolatile, thermally labile molecules such as peptides. Making

volatile derivatives is undesirable because the process is time consuming, leads to
unwanted products, and usually requires large amounts of sample. Furthermore, there is a
mass limit above which not even derivatized peptides are sufficiently volatile. The variety
of available DI techniques, each of which has its particular advantages for certain
applications, generally eliminates the need for derivatization reactions and opens the door to
direct analysis of even higher mass compounds.

The DI techniques developed in the 1960s and 19705 include field desorption (FD)
[10] plasma desorption (PD) [11], direct chemical ionization (DCI) [12], laser desorption
(LD) [13] and secondary ion mass spectrometry (SIMS) [14]. The development of these
techniques gradually demonstrated that mass spectrometry can be used for the
determination of large, thermally labile biopolymers such as proteins. However, each of
these techniques at that time presented at least one of the following problems such as signal
stability, durability, reproducibility, and difficulty in operation.

Fast atom bombardment (FAB), developed in 1981 by Barber et al. [2] provides for
the first time a technique for ionizing nonvolatile samples that is both simple to use and
gives reproducible results. It has become the dominant ionization technique used for
peptide sequence determination. FAB mass spectra of peptides show MH“, the protonated
molecules, as the predominant ions at the high-mass end of the spectrum, which could be
used as the precursor ion in MS/MS analysis. Although the FAB-MS spectrum usually
displays some sequence ions, they are usually very weak and difficult to be isolated from
matrix peaks. In addition, the mass range of FAB ionization is limited to 3000 Da
(frequently lower than this) for peptides. Techniques with greater mass range need to be
developed.

In the late 19808, the evolution of ionization techniques such as electrospray (ES)
[3] and MALDI [4] greatly extended the mass range. Compounds with mass up to millions

of Daltons have been determined [15]. Most importantly, the electrospray technique also

provides a simple and successful way in coupling separation techniques such as
electrophoresis and liquid-chromatography to mass spectrometry.

It should be pointed out that single stage mass spectrometry usually provides only
molecular weight information, frequently from the mass of the most abundant pseudo-
molecular ions such as protonated peptides (MH+). These ions, formed by commonly
used ionization techniques such as FAB or electrospray, have very little tendency to
fragment.

The lack of fragmentation is very useful for the accurate determination of molecular
weights of peptides, even in quite complex mixtures, but the resulting absence of structural
information is a definite drawback. This problem can be overcome by transferring
additional energy sufficient to force fragmentation of the stable MH+ ions formed in the
initial ionization process. Collision with a neutral gas is the most frequently used method
leading to collision-induced dissociation (CID) [16]. The resulting fragment ions are then
mass analyzed in a second stage of mass spectrometry, hence the term tandem mass
spectrometry (MS/MS). Frequently, the necessary energy can be acquired by multiple
collisions at low energy (eV) or single collisions at high energy (keV).

Tandem mass spectrometry has many advantages over conventional mass
spectrometry. First, it increases the fragmentation of protonated peptides by CID.
Therefore, more structural information can be obtained from a CID spectrum of a peptide.
Secondly, it eliminates peaks that originate from matrix and impurities. Therefore it is
capable of mixture analysis without pre-separation. Thirdly, all ions thus produced are
structurally-related to the precursor ions MH‘“. In addition, tandem mass spectrometry is
capable of differentiating isomeric and isobaric amino acids in peptides in high energy CID
experiments [17-18].

When CID is performed for MS/MS experiments, collisions can be carried out at
different energies. For low-energy collisions, such as those occurring in a triple

quadrupole instrument, the collision energy is usually below 100 eV. With peptides,

cleavage of amide bonds is most frequently observed. With high-energy collisions, such
as those occurring in sector instruments, the collision energy is usually in the range of 1 to
10 keV. This high-energy collision generates more fragments than it does in low-energy
collisions. Side-chain specific fragments - different for different amino acids, are
observed. For peptide sequencing, these fragment ions are extremely important [17-18] in
distinguishing isomeric leucine and isoleucine. However, these ions are usually not
observed in the low-energy CID spectra.

When precursor ions of a peptide are subjected to collision-induced dissociation,
fragment ions indicative of the peptide structure are formed subsequently. Frequently, the
dominant ions and ion series observed are dependent on the amino-acid composition (such
as number of basic amino acids and their positions along the peptide chain) of the peptides
and amino-acid sequence of the peptide. To date, the formation of many ion series have
been proposed - some of them have been experimentally supported. However, questions
like the following are still to be answered: Why does one specific ion series dominate over
another in a CID spectrum of a peptide? For a regular peptide CID-MS/MS spectrum,
what ions are produced by charge-induced processes? What ions are from charge-remote
processes? Which ion series will dominate a spectrum? Why? An better understanding
about the formation of various fragment ions can certainly speed up interpretation of mass
spectra thus peptide sequencing.

In the study of CID fragmentation mechanisms of protonated peptides, various
methods have been utilized. When a hydrogen shift is involved in the formation of a
specific ion series, deuterated peptides were used frequently to trace the source of the
shifting hydrogen atom. Frequently, the CID spectra of partially deuterated peptides (such
as those via HID eXchange in a deuterated solvent), a-carbon deuterated peptides are used.
In tracing the formation of secondary ions derived from specific side-chain cleavages of

regular sequence ions, precursor scans (BZIE linked scan) are frequently used to trace

possible precursors of a fragment ion. Product scans (B/E linked scan) are then performed
to confirm the products from that possible precursor ion.

To date, mass spectrometry has become one of the most important tools in protein
chemistry. Examples are the determination of molecular weight, molecular weight
~ confirmation, identification of post-translational modification, sequencing of N -blocked
peptides. In this dissertation, the research will be focused on fragmentation of peptides in
both neutral and charged (negatively charged form). An example of application of mass

spectrometry for peptide analysis will also be discussed.

'11. Research Objectives

ll. ' '1.” .0 nur' ' 0111 c or: 0' ' u.‘ til: ‘1 3c. The
first one is the primary thermal degradation chemistry of peptides, studied by using
K+IDS. This study will provide a detailed description of thermal degradation of peptides,
based on experiments with peptides and their chemically modified variants such as
derivatized peptides, deuterated peptides and specifically designed peptides. Since thermal
degradation is a charge-free process, the results from this study may provide insight into
charge-remote processes in CID of protonated peptides, as will be explained in more detail
in Chapter 4. The second process is metastable and CID fragmentation of deprotonated
peptides. This process is too complicated to draw much attention until recently, but still
only di- and tri- peptides have been studied so far. Preliminary study with larger peptides
in this research will add much more to our understanding on metastable fragmentation and
CID of deprotonated peptides. At the end of this dissertation, one sample application of

mass spectrometry to peptide analysis will be presented.

A. Thermal degradation of Peptides:

In this research, the primary thermal degradation chemistry of peptides will be
characterized by using a mass spectrometric ionization technique, K+- Ionization of
Desorbed Species (K+IDS). This technique, based on thermionic emission of K+ ions and
subsequent K+ attachment to thermally desorbed neutral analyte molecules and their
thermal degradation products, was introduced in 1984 from this laboratory [19]. The
strength of the K+IDS technique is in the simplicity of both the experiment and the
resulting mass spectra. Since each thermal degradation product gives one peak in the mass
spectrum, K+IDS mass spectra reflect a "molecular weight distribution" of the thermal
degradation products of the analyte molecules, which, in combination with analogous
studies of derivatized and isotope-labeled compounds, can be used for the study of
mmWWfimdanmm However, the simplicity of
the mass spectra is also a complication since only the molecular weight of each degradation
product is directly determined. The approaches presented here are designed to ms];
H... .H .- u _ . ._ l_..h.wth.”.ml I“ .r"”

'.. -.,.. 'm- u . hr! 1.“. 1"! r4... .” . v _ . “my -’
as well as other thermally-labile compounds.

There is a number of reasons why the thermal degradation chemistry of peptides is
of current interest. First, thermal degradation of peptides is a process of rich chemistry
itself. Traditional techniques such as Pyrolysis-Gas Chromatography (Py-GC) are very
difficult to apply to the study of such processes because the degradation products
(pyrolytes) are either too small to provide much useful structure information, too large to
survive GC conditions (undergo further reaction, for example), or too polar to be analyzed
by GC (not volatile enough, for instance). In fact, detailed study on pyrolysis has been
extended only to dipeptides to date. Not much information has been obtained for larger
peptides [20-21]. K+IDS provides a unique way to circumvent this problem and make it

possible to study the thermal degradation of peptides.

A second reason for the interest in thermal degradation is a process in mass
spectrometry (MS) called charge-remote chemistry. When MS is used for structural
analysis of peptides, the analyte molecules undergo desorption and ionization via
techniques such as fast atom bombardment (FAB). When a peptide, M, is converted into a
gas phase cation in the FAB ion source, it appears as a protonated molecule, MH+, which
subsequently fragments (or is induced to fragment), giving structural information. There
are two typesof processes through which protonated peptides can fragment. The first is
that most commonly considered in mass spectrometry for even-electron ions - the chemistry
occurs at the site of protonation, and is W. A second possibility is a charge;
mm; [22-24] mechanism. In the latter, bonds are broken far from the charge site, and are
unaffected by the charge. They are driven by excess energy in the protonated molecule (as
are charge-initiated processes). It is very difficult to distinguish between these two
mechanisms, because, while the precursor to all fragment ions is the protonated peptide
molecule in positive ion FAB, the initial site of protonation is not known and cannot be
controlled. The advantage of the K+IDS technique is that the molecules undergo
fragmentation (thermal degradation) before a charge is placed on the molecule. Thus, for
mass spectrometrists it may represent a model for charge-remote chemistry. Molecular
fragmentation occurs due to energy deposition in the unionized molecule. It is hoped that
the chemistry of bond cleavage in peptides will be sufficiently different for these two cases,
charge-initiated and charge—free, such that the contributions from each in the fragmentation
of protonated peptides can be deduced.

Thirdly, there have been a number of experimentally simple, "direct probe"
approaches for the MS analysis of compounds such as small peptides. In such
experiments, solid samples are introduced into an electron impact (E1) or chemical
ionization (CI) source, and heated. Mass spectra are generated, which are either simple
mass spectra of desorbed molecules, or complex spectra, resulting from the simultaneous

ionization of desorbed molecules and desorbed thermal degradation products [25-26]. The

results of K+IDS experiments may provide direct information on such processes, to assist
in the evaluation of the results of such experiments.

Finally, the average energy obtained by analyte molecules from K+IDS is about the
same as that from FAB experiment. For a peptide of 500 Da (approximately 70 atoms, 3N-
6 = 400 degrees of freedom), assuming 50% of the atoms reach the temperature of 600K
(experimental condition) with starting temperature at 300K (AT = 600-300 = 300K), the
energy gain of the peptide will be roughly about 50% x 400 x RAT/2 = .5x400x8.3x300/2
J = 240 kJ/mole = 60 kcal/mole = 3eV, very close to the average energy deposited in a
high-energy collision process [27]. Therefore, K+IDS spectra may be compared with FAB

spectra under certain circumstances.

B. Fragmentation of Deprotonated Peptides

The second topic in this dissertation is an investigation of fragmentation of
deprotonated peptides induced by collision-induced dissociation. A mechanistic study was
carried out on peptides larger than three amino-acid residues. In this study, extensive
derivatization was performed to vary the site of deprotonation, and the effect of charge site
on fragmentation was evaluated for deprotonated peptides. Deuterated peptides were used
to trace the hydrogen shifts. Based on these experiments, mechanisms for the formation
of the most commonly occurring ions from deprotonated peptides are proposed. Results
from this study may provide valuable insight into CID of protonated peptides.

In mass spectrometry, the effect of charge site for even-electron ions such as
protonated molecules and deprotonated molecules on fragmentation may be very similar.
Therefore, fragmentation of charge-localized peptides in negative mode may produce the
same fragments as those from charge-localized positive ions. Comparison of spectra
obtained from different charge states (positive, neutral and negative mode) may provide

important insights into ion fragmentation processes.

characterization and refinement of the technique will be described.
application of K+IDS for the analysis of peptides will be studied. The applicability of this
technique for peptide sequencing will be discussed. In Chapter 4, the primary thermal
degradation of peptides will be studied by using K+IDS. Mechanisms for thermal
degradation of peptides will be proposed and experimentally supported. In Chapter 5, a
previously un-revealed phenomenon - response discrimination, will be described and the

results will be used to interpret K+IDS spectra.

10

In Chapter 2, the K+IDS technique will be introduced in detail and further

deprotonated peptides will be discussed.

References

1. Biemann, K.; Gapp, F.; Seibl, J. J. Am. Chem. Soc. 1959, 81, 2274.

2. Barber, M.; Bordoli, R. 8.; Sedwick R. D.; Tyler, A. N. J. Chem. Soc. Commun.
1981, 325.

3. Penn, J. 3.; Mann, M.; Meng, C. K.; Wong, 8. F.; Whitehouse, C. M. Science
1989, 246, 64.

4. Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299.

5. Edman, P. Acta Chem. Scand. 1950, 4, 238.

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

7. McLafferty, F. W. Interpretation of Mass Spectra, 3rd Ed, University Science .
Books: Mill Valley, CA, 1980.

8. Biemann K.; Vetter, W. Biochem. Biophys. Res. Commun. 1960, 3, 578-584.

9. Vilkas, E. Lederer, E. Tetrahedron Lett. 1968, 3089.

10. Beckey, H. D. , Int. J. Mass Spectrom. Ion Phys., 1969, 2, 500.

In Chapter 3,

In Chapter 6, fragmentation of

ll.

12.
13.

14.
15.

16.
17.

18.

19.
20.

21.
22.

23.
24.
25.
26.
27.

11

Torgenson, D. F.; Showronski R. P.; Macfarlane, R. D. Biochem. Biophys. Res.
Commun., 1974, 60, 616.

Baldwin, M. A.; McLafferty, F. W. Org. Mass Spectrom. 1973, 7, 1353.
Posthumas, P. G.; Kinstemaker, P. G.; Meuzelaar, H. L. C.; Ten Noever de Brauw,
Anal. Chem. 1978, 50, 985.

Benninghoven, A. Int. J. Mass Spectrom. Ion Phys. 1983, 46, 459.

Chen, R.; Mitchell, D. W.; Chen X.; Wu, Q.; Hofstadlere, S. A.; Bruce, J. E.;
Smith, R. D. The 42nd ASMS Conference on Mass Spectrometry and Allied Topics,
Chicago, (1994), MP203.

McLafferty, F. W., Ed., Tamdem Mass Spectrometry; Wiley: New York, 1983.
Johnson, R. 8.; Martin, S. A.; Biemann K.; Stults, J. T.; Watson, J. T. Anal.
Chem. 1987, 59, 2621

Johnson, R. 8.; Martin, S. A.; Biemann K. Int. J. Mass Spectrom. Ion Processes
1988, 86, 137.

Bombick D.; Pinkston, J. D.; Allison J. Anal. Chem. 1984, 56, 396-402.

Irwin, W. J. Analytical Pyrolysis: A Comprehensive Guide; M. Dekker: New York,
1982; pp 3-44.

Irwin, W. J.; Slack, J. A. Analyst 1978, 103, 673-704.

Gierlich, H. H.; Rollgen, F. W.; Borchers, F.; Levsen, K. Org. Mass Spectrom. 1977,
12, 541-543.

Jensen, N. J.; Tomer, K. B.; Gross, M. L. J. Am. Chem. Soc. 1985, 107, 1863-1868.
Adams, J. Mass Spectrom. Rev. 1990, 9, 141-186.

Svec, H. J.; Junk, G. A. J. Am. Chem. Soc. 1964, 86, 2278-2282.

Biemann, K.; Seibl, J.; Gapp, F. J. Am. Chem. Soc. 1961, 83, 3795-3804.
Wysocki, V. H.; Kentlamaa, H. 1.; Cooks, R. G. Int. J. Mass Spectrom. Ion
Processes, 1987, 75, 189.

Chapter 2
K+IDS: The Technique and Its Further Characterization

I. Introduction

Mass spectrometry has been traditionally used for the analysis of volatile species
and those that can be easily derivatized to volatile compounds. Before the evolution of
fast atom bombardment (FAB), peptide analysis by mass spectrometry was hampered by
the low volatility and thermal instability of peptides. That is, the volatility of a peptide
can not be increased simply by heating, which leads to thermal degradation of the
peptide. In the 19605 and 19703, much effort in mass spectrometry research has been put
into developing new ionization techniques to circumvent the problem. Various so-called
desorption/ionization (D/I) techniques were developed, all based on a simple concept:
Energy deposited on a sample surface in a very short time will induce desorption more
effectively than thermal degradation.

In the search for better techniques at Michigan State University, K'*‘-Ionization of
Desorbed Species, K+IDS, a desorption/ionization technique, was introduced in 1984 by
Bombick and Allison [1] from this laboratory. This technique is based on thermionic
emission of K+ ions and subsequent K+ attachment to thermally desorbed neutral analyte
molecules and their thermal degradation products. Since the introduction of the
technique, K+IDS has been successfully used for the analysis of various compounds such
as saccharides, polymers, peptides and steroids [2-6]. A detailed description on how this

technique works will be presented below.

12

13

11. Experimental

A. Experiment setup and K+IDS probe

The K+IDS experiment is performed on an HP 5985 GC/MS/DS quadrupole mass
spectrometer with a mass range of 10-1000 Da. A schematic of the ion source and
K+IDS apparatus is shown in Figure 2.1. The key device for this K+IDS technique is the
K+IDS probe (Figure 2.1a-c), a modified direct insertion probe. This probe is used to
insert the K+ aluminosilicate bead, the K+ emitter, into the center of the ion source. The
K+ bead filament is made by threading rhenium wire through a two-holed ceramic
leaving a 1-2 mm loop at the top. The ends of the tip are spot-welded to nickel-wire leads
that are fed through the K+IDS probe. The K” bead is made by methods described in part
B of this section [7]. Before use, the bead has to be conditioned inside the ion source of
the mass spectrometer by slowly ramping the current applied to the bead (0-3.0 A, 0.5
A/min.). A low voltage power supply is used to provide to the K+ bead a bias voltage of
0-10 volts with respect to the repeller. K+IDS experiments are generally performed with
a current of 2.2-3.0 A and a bias voltage of O V.

Because the K+IDS signals are usually too short-lived to allow for good
calibration of the instrument with K+IDS adduct ions, the calibration of the instrument
with perfluorotributylamine (PFTBA), especially the mass axis, is performed in the
positive EI mode in the traditional way for GC/MS andlor DIP analyses. The voltages
applied on all the lenses are adjusted for optimum E1 results. After the source is cooled
down, the E1 filament is turned off for K+IDS. The K+ emission is monitored in the
override tuning program and the bead position and direction are adjusted for optimum K+
transmission. Following this K+ bead position adjustment, the calibration was checked

with K+IDS using palmitic acid [CH3(CH2)14COOH] as a benchmark compound to test

the mass accuracy in Karen Light's experiments.

14

K+IDS Probe

    

K+ Emitter /\ Ceramics /Stainless steel tube

  
  

,“a...A‘_ ......

\ ___________ \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\:i$521452"?

/

l\\\\\\\\\\ ‘\ $323,; . .;
Sample holder I

  
   
 
 

 

 

 

I Power Supply J

 

K+ emitter: A solid (potassium aluminosilicate) emits K+ when heated.

made from a mixture of KNO3: SiOZ: A1203 = 2: 2: 1 (molar ratio).

Figure 2.1a Schematic of a dual-filament K+IDS probe.

15

To Quadrupole Mass Analyzer

Lenses \ T Ion Source

K+ Ion . - /

0
En“ “GIN \
m: \§{{{{{{\§{{C{{{{{{{i{ifi'xxiixi’x'iifififi'x‘k’i'x’x’ifii‘x‘ifiii" i?
/

 

 

 

    
    

1| ..

     
  

             
   

 

V

\

: x as» “ ‘1};3-‘202n

....................................

Sample,

Holder - .
/ -[ Bias Voltage control r Power Supply I

Ground V

 

 

 

 

 

 

 

 

 

 

Repeller

 

 

Figure 2.1b Schematic diagram of K+IDS instrumentation.

16

  
        
  

 

 

 

 

K+ Emitter
1200 °C .......... .
I'_K+ (g) ‘— - '{igs‘g‘ fi —:\\ \\\\\\\\\\\\\\\““\“““““‘\“\\\\\\\\\\\\\\\\\\\\\\\\\\ .......
Same Folder , .....
I" l
l Power Supply j
.—> M (g) 4' K+ (g) “'--> (M+K)+ (g) D (g) 4. [Ch (3) n--.) (D+K)+ (g)

 

M: Intact analyte molecule.
D: Thermal degradation products of M.

Figure 2.1c Schematic of a K‘l’IDS experiment in operation.

 

17

It has been noticed, in my experience with K+IDS, that testing calibration with a
single compound such as palmitic acid does not always give reliable results because
calibration in other mass regions could be off even if the mass for the (M+K)+ of palmitic
acid is correct. To ensure better results, a mixture of polymers such as PEG 200
(polyethylene glycol, average molecular mass 200 Da) and PPG 725 (polypropylene
glycol, average molecular mass 725) were used. Both the low-mass region and higher
mass region were considered. The quality of the spectra of the mixture (Figure 2.2), both
in mass accuracy and isotope distribution, was used to judge the quality of a mass
calibration. In the spectrum, peaks at m/z 189, 233, 277, 321, 365 and 409 are (M+K)+
ions of PEGZOO with n = 3, 4, 5, 6, 7 and8, respectively; peaks at m/z 463, 521, 579, 637,
695, 753, 811, 869 and 927 are (M+K)+ ions of PPG 725 with n = 9, 10, 11, 12, 13, 14,
15, 16, 17 and 18 respectively. Because these polymers are even-mass compounds and
K1“ has an odd mass (39 Da), (M+K)+ ions of these polymers should have odd m/z
values. All peaks labeled with odd m/z values indicate a good calibration. Any peak
labeled with even m/z value indicates a bad calibration.

All of the samples used in K+IDS experiments were dissolved or suspended in
methanol to allow transfer of 1-2 pl of sample (approximately 10 pg) onto the sample
holder. The solvent was evaporated before the probe was inserted into the mass

spectrometer.

B. The making of a thermionic emitter

The most difficult part in making a thermionic emitter is making a good glass
mixture and attaching the glass on the rhenium wire. Traditionally, a mixture of KNO3,
Si02 and A1203, in a molar ratio of 2:2:1 is ground and a slurry in acetone is made. A
small drop (Pasteur dropper) of slurry is applied onto the rhenium wire loop, to cover it
evenly. The bead is heated on a Benson burner flame and the solvent is evaporated. The

dried bead is heated in the flame carefully. Gases (N02 and 02) are released and some

18

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

PEGZOO m PPG600
‘ 579 637
- 321
695
‘ 233
— 521
753
365 811

q 189 409 463 869
r‘r‘r‘rl TTLf+-JT1 lit. T‘r—rr lu‘vL-rgrlr‘ iArLr 73% frr? III-I 117‘1 {fir rtrr -rrrrerrr

' l l l I l I I '1 I l I l

 

130 180 230 280 330 360 430 480 530 580 630 680 730 780 830 880 930 980

m/z

Figure 2.2 K+IDS mass spectrum of a mixture of PEG 200 & PPG 600.
PEG = Poly(ethylene glycol) [H-(OCHzCH2)n-OI-I].

PPG = poly(propylene glycol) { H-[OCH(CH3)CH2]n-OH}.

19

binding between different particles develops. This bead is conditioned in the ion source
by ramping the current from 0 to 3.0 A at 0.5 A/min. The probe is then cooled for use.

Although the recipe for making a glass mixture has not changed during the
evolution of K+IDS technique, no one in this group has made a better glass mixture than
that prepared by Daniel Bombick many years ago. The quality of the mixture determines
the easiness for bead making and conditioning. With a bad glass mixture, it is very
difficult to attach the mixture to the wire loop without pop-off after preliminary heating,
or the bead does not survive the conditioning process. Because it is so difficult to make a
good mixture this way, I tried a different way of making the glass mixture.

Since the glass was made from a mixture of KN O3, Si02 and A1203, the result of
heating is the formation of potassium aluminosilicate with the release of N02 and O2
gases at the same time, especially during the conditioning processes. The preliminary
heating removes the glass-carrying solvent such as acetone/methanol, and partly
decomposes the KN03 to K20 which reacts with the other two components to form an
aluminosilicate which binds the mixture together. During the conditioning process, more
aluminosilicates are formed and the beads become more homogeneous with the increase
of temperature. Because of the zeolite structure of the aluminosilicate, in which large
channels are formed, the bead is good for continuous K+ emission when heated.

With this in consideration, ground molecular sieves with K+ as the cation (K+
alurrrinosilicate) were examined as the glass mixture because it contains potassium
aluminosilicate. With the same process, a K+ bead was made (adding a small amount of
Si02 may prevent the bead from becoming too glassy which is bad for heat dissipation).
It was found that ground molecular sieves with K“ as the counter ions is a very good
structural material for the K+IDS bead. There are several advantages for this new bead
material: much shorter grinding time for molecular sieves than that for the traditional

glass mixture, shorter conditioning time because no gas is released during conditioning,

20

higher success probability (less pop-oft) because no gas is released during heating, and

much shorter time is needed to make a quality bead.

III. K+IDS: Operational Principles

There are two basic processes that are important to the K+IDS technique. One is
the use of a thermionic emitter to form gas-phase K+ ions that are used as reagent ions in
a way similar to those used in chemical ionization. The other concept is the use of rapid
heating of a thermally labile compound to promote vaporization of the sample with little
thermal degradation. When these two processes are combined and occur simultaneously
in the ion source of a mass spectrometer, a K+IDS spectrum can be obtained. The K+
ions attach to neutral gas-phase species and are observed as K+ adducts. This K+
attachment induces little or no fragmentation and it samples only the neutrals present in

the gas phase produced by heating the sample.

A. Thermionic emission

The most important aspect of the K+IDS technique is the production of a large
flux of K+ ions in the gas phase. This is accomplished by the use of thermionic emission
materials that are patterned after the aluminosilicate mixtures described by Blewett and
Jones [8] in 1936. They determined that the mixture of A1203, LizO and Si02 in a 1:1:2
ratio, was the best mixture for producing a large flux of Li“ ions. The same mixture
doped with K20 instead of Li20 has proven to be a good source of K+ ions when the
mixture is heated to temperatures between 800 °C and 1200 °C. The emission of K+ ions

is dependent primarily on the work function of the surface and the ionization energy of

21

the potassium atom. Extensive discussion of how these materials are capable of emitting
metal ions can be found in D. B. Bombick’s dissertation [9].

Because K+ was used in K+IDS experiments to "sample" the neutral species in
the gas phase, this technique was previously referred to as K+ chemical ionization (K+CI)
[1] because the K+ ions are simply used as reagent ions. Compared to the commonly
used CI technique, this K+CI was believed to be less dependent on the pressure of the
systems.

It should be pointed out that the thermionic emitter (the I(+ bead) used in these
experiments serves two purposes. First, it is used as the K+ emitter, producing large
fluxes of K+ ions when heated. Second, it also serves to heat the analyte molecules on
the sample holder radiatively, giving rise to thermal degradation and desorption. Because
desorption and K+ emission occurs simultaneously and in very close proximity (1 mm

separation), K+ attachment results.

B. Desorption vs. Degradation

The other process essential to the K+IDS technique is the transfer of the thermally
labile compounds into the gas phase. The use of rapid heating to promote desorption of
intact molecules over competitive decomposition reactions of thermally labile compounds
was described by Friedman and coworkers [10] for the analysis of underivatized peptides.
The enhancement of the volatility of a complex thermally-labile species due to rapid
heating of the sample is based on a classical kinetic analysis of the competitive
decomposition and vaporization processes.

In K‘IDS experiments, with a compound such as glycylglycylglycine on the
surface of the probe, a number of events occur as the temperature of the probe surface
changes from room temperature to a final temperature of about 300 °C. If the surface is
slowly heated, the compound will decompose and the more volatile decomposition

products will vaporize. This is what occurs in pyrolysis-MS analysis [11]. However, on

log k

| Desorption: M(s) or M(l) ---> M(g)
Desorption |

  
     
 
   

Degradation: Decomposition

Degradation
/

 

 

 

Desorption Favored 4— I —> Degradation Favored
l/T

 

+ Lower temperature

R. J. Beuhler, E. Flanigan, L. .l. Greene and L. Friedman, J. Am. Chem. Soc. 96, 3990 (1974)

Figure 2.3 Temperature Dependence of Desorption vs. Degradation

23

rapid heating, temperatures can be achieved above some critical temperature Tc, at which
desorption rates are competitive with, or can exceed, decomposition rates. So it is
possible to raise the temperature above this critical value before a substantial amount of
the analyte is decomposed. This aspect of rapid heating, as shown in Figure 2.3, has been
developed by Friedman [10]. In light of this, the evolution of neutrals in the gas phase
above the emitter varies with time when the sample is heated. At an early stage of rapid
heating, decomposition may dominate until Tc is reached. Then desorption of intact
molecules may dominate the processes. Finally, long-lived species on the surface lead to
further decomposition products. In K+IDS experiments, these two processes are very
competitive and important. One process provides structure information of analyte
molecules and the other one provides molecular weight information.

As discussed above, there is a critical temperature for the desorption of any
specific compound- However, this temperature for a given compound has never been
determined. When K+ IDS experiments were performed in Bombick's work, the
temperatures of both the thermal ionic emitter and the sample surface were the same since
a single-filament probe was used. Due to the rapid heating and high temperature, both
degradation and desorption occur. Therefore, intact molecules and their thermal
degradation products are both detected.

When a two-filament probe was introduced by Daniel Kassel [5], the concept of
rapid heating was challenged because the radiative heating utilized in the experiments
was not expected to generate high heating rate and high final temperature. Karen Light
realized the importance of these temperatures and experiments in the determination of
these data were performed. It was found that, with a heating current of 3.0 A, the final
temperature of the sample surface was around 300 °C (attained in 20 sec) and the heating
rate was about 15 °C/sec. Nevertheless, spectra obtained from this probe were found to

be very similar to those obtained from the single-filament probe.

24

Because the temperature dependence of thermal degradation and desorption are
different, K+IDS spectra acquired at different times into the experiment have different
degradation patterns and relative intensities. In K+IDS -MS, average spectra are
frequently used. Therefore, K+IDS spectra provide information on both low temperature
and higher temperature products, which are all related to the structure of the analyte
molecules.

Since ions in a K+IDS spectrum are K+ adducts of thermal degradation products,
the interpretation of the spectrum must be based on the thermal degradation mechanism
of the analyte. It was assumed that in such a short time in which thermal degradation
could occur, only a few mechanisms may dominate. Of these, 1,2-eliminations were
believed to be the most common [2]: thermal activation of a polar bond, accompanied by
the shifting of a specific hydrogen, resulting in the formation of two stable molecules
from one. This process usually requires little energy. My results, however, suggest that
thermal decomposition with the formation of cyclic products could be a more important
one, especially for peptides. In this case, H-shifts can be more reasonably attributed to
the terminal hydrogens based on the experiments with deuterated peptides, which will be

discussed in Chapter 4.

C. K+-Adduct Formation and Stabilization

It is believed that ions in K+IDS-MS are formed due to thermal degradation and
desorption of neutral species followed by K+ attachment to the gas-phase molecules.
With the commonly used dual-filament probe, this is evidently true because the two
filaments, K+ emitter and sample holder, are spatially separated from each other (the
original single-filament probe was not used for this dissertation work). Previous work
with K+CI [1], in addition to the results of ICR experiments [11] with K+ adducts, suggest

that K" rarely, if ever, induces fragmentation following attachment. Therefore, the peaks

25

in a K+IDS spectrum represent the statistical distribution of neutral thermal degradation
products.

Because energy is released with adduct formation, adduct ions become excited
once formed. Since little fragmentation was observed, adduct ions were believed to be
stabilized somehow. In search for possible sources for such stabilization, Bombick found
that the addition of a neutral collision gas into the ion source produced an increase in
adduct formation for small analyte molecules. When N2 was introduced into the ion
source as a collision gas to a source pressure of approximately one torr, the detection
limit for polyphenylether was lowered from 500 ng to 250 ng at a signal-to-noise ratio of
five. This increase in adduct formation was used as the major evidence of adduct
stabilization by a three-body collision process for nascent K+-adducts [9], a theory
adopted by all of my predecessors.

- However, this small decrease in detection limit by a large amount of collision gas
demonstrated only a weak effect of collision-stabilization. For real K+IDS experiments,
the pressure around the probe tip is calculated to be 10‘3 torr under K+IDS condition,
based on my calculations (see Appendix I for detailed calculations on ion source pressure
and collision frequency). Such low pressure will not allow three-body collisions inside
the ion source. Therefore, three-body collisions are n0t very likely to occur for nascent
K+ adducts.

In considering the stabilization of Li+ adducts, Woodin and Beauchamp [12]
proved that Li+ attachment to an analyte molecule produced an excited complex that
could be stabilized by the emission of a infrared photon. These biomolecular infrared
radiative association reactions are assumed to be dominant stabilization mechanisms at
very low pressures where the time between collisions is greater than 100 ms. This photo
emission mechanism has gained popularity as a possible association mechanism for

interstellar chemistry, but is unlikely to be responsible for the stabilization of the K‘l’

26

adduct ions formed in K+IDS due to the short residence time inside the mass
spectrometer.

As indicated above, K+ adduct ions are excited with adduct formation. However,
the energy release in adduct formation may not be large enough to induce any
degradation, which may lead to K+ detachment. In light of this, the nascent adducts may
not be stabilized at all. Significant adduct ions may have dissociated back to K+ ions
before reaching the detector. The availability of the ions for detection is determined by
the life times of the ions. If the adduct ion is not sufficiently stable, it cannot be detected.
The response discrimination discussed in Chapter 5 - a previously undiscovered
phenomenon in K+IDS - may serve as an evidence that K+ adducts formed in K+IDS are

not stabilized by collisions.

IV. MK+ Stability Studies by Using a Directly Heated K+IDS Probe

A. Characterization of a Novel Direct-Heating K+IDS Probe

In order to make the K+IDS probe more adjustable, a novel probe with direct-
heating capability was designed and made based on the following consideration. (1) For
the original two-filament probe in which samples on a sample holder are heated relatively
slowly (20 'C/sec) , the heating rate varies from time to time, and from one probe to
another due to the uncertainty in bead size, thickness, shape, composition, bead-sample
separation, etc. So, whenever a different bead is made or after a bead is used for a while,
the heating rate will change (2). The final temperature of the sample wire obtainable is
only as low as 300 °C. and the heating rate is not so “rapid” as expected.

With the new probe, the sample holder can be powered to pass electric current.
There are two advantages for this new probe (1). By adjusting the sample heating
current, it is possible to adjust the desorption rate and thermal decomposition rate of the

sample so both molecular weight and structural information can be obtained in a single

27

spectrum. This is especially true for highly thermally-labile compounds such as met-
enkephalin, which produces only a weak molecular ion peak in a K+IDS spectrum (2).
Very polar compounds can produce (M+K)+ ions in K+IDS spectra when the sample is
directly heated. Direct heating of the sample on the new probe can provide real “fast
heating” rates which the normal probe cannot offer to desorb large molecules into the gas
phase. In addition, this probe can also be operated as the original two-filament probe
when the sample filament is not powered.

0n the other hand, this probe also provides one problem. That is, the lifetime of a
sample will become very short when direct heating is provided (much higher heating
rate). With a quadrupole mass spectrometer, this may lead to a very skewed spectrum.
Sometimes, no ions can be detected due to the short sample lifetimes.

Sample spectra of met-enkaphalinamide (Figure 2.4) with two different probes
were obtained. With the original probe the (M+K)+ signal is weak. With the new probe
the direct heating makes [M+K]+ ions much more abundant. However, K+ adducts of
thermal degradation products from the direct heating probe is rare, thus structure
information is lost. Because of this reason, K+IDS spectra of peptides in this dissertation

were obtained by using the original two-filament probe.

B. MK" Stability Studied by Direct Heating

In previous K+IDS papers, it has been suggested that K+ adducts formed are
stabilized by collisions. No fragment ions will be generated from a K+ adduct. For
example, only (M+K)+ ions are observed for many polymers. The rationale for this
speculation is that nascent K+ adduct ions can be collisionally stabilized in the ion source
right after their formation by energy transfer in collision processes. It is assumed that the
ion source pressure in a K+IDS experiment is so high that there are many collisions
happening right after adduct formation so that the energy released during adduct

formation will be transferred to other species, thus the K+-adducts are stabilized.

28

 

 

 

 

 

 

 

a (C3+H+K)+
‘ 333
_ H-YGGFM-NH2
“ (Y3+H+K)+
4 (C2+H+K)+ 391
276
(Y1+H+K)+ I (Y 4+H+K)+ (M+K)"’
‘ 187 228 252 I 300 448 611
%1# A??? I i‘. Jfir‘rii . lr‘rr L. .15? i i i‘.* 1141 Fri turn it“??? Friar]; “i r i i
l l I l l I l T l l I l l
140 180 220 260 300 340 860 420 460 500 540 580 620 660
m/z
b (M+K)"’

H-YGGFM-NHZ 6“

 

 

 

 

‘I
.NH3
(C3-t-H+K)+ 1-
_I 333 I
l
594
T_r l r Ar-r‘ril— ‘1‘ fi?uu‘f I‘r-mr 77‘1““?‘l-‘r-iu77-“L‘7Mr‘ r-r“ Twit-1‘ TrTr I l f]“ Allf—T- r l r r
I I l l 1 hi I l I I l l

140 100 220 260 300 340 380 420 460 500 540 580 620 660

m/z

Figure 2.4 K+IDS spectra of H-YGGFM-NHZ: a) with the original probe, I = 2.70 A;

b) with the new probe. I(sample) = 2.0 A, direct heating delay = 3 seconds.

29

The collision-stabilization theory can be very important if the assumption of high
pressure in the ion source is correct. However, calculations on the source pressure show
that the pressure, generated in a K+IDS ion source with a typical run, falls in about 0.06
Pascal (10’3 to 10‘4 torr), which leads to a conclusion that 90% of K+ ions produced will
not undergo collision even once (see Appendix 1). Consequently, an adduct ion formed
will have much less probability colliding with a neutral species.

Since collision stabilization theory is not very convincing, it is probable that no
stabilization occurs after adduct formation. This is consistent with the observation of
response discrimination that will be discussed in Chapter 5. Discrimination was found
against small and cyclic compounds. The discrimination is a reflection of the stability
difference of different adduct ions. In addition, the results from the study with mixed
ionic emitters in Chapter 5 also confirm that there is significant K+ detachment following
adduct formation. Therefore, K+ adducts are not stable due to K1“ detachment (short
lifetimes). However, the lifetimes for such adducts cannot be estimated at the present
time with confidence.

The stability of K+ adducts should also be examined from the standpoint of
fragmentation of K+ adducts. This process is believed unlikely because the energy
released during adduct formation is too small to induce such processes. The following
experiments with direct heating will provide insights into the stability of K+ adducts.

To study possible fragmentation of K+ adducts (or stability of K+ adducts), the
relative abundance of (M+K)+ and that of a possible fragment (F+K)"' are monitored with
a direct heating K‘l’IDS probe because a large abundance of (M+K)+ can be produced for
study. To circumvent the problem of short sample lifetime, the selected ion monitoring
(SIM) mode is used in these experiments - multiple ions are monitored alternately at very
short time intervals (50 milliseconds for each ion).

Table 2.1 shows an example of temperature dependence of thermal degradation of

a peptide H-YGGFL-OH. For this peptide, the pseudo-molecular ion is at m/z 594, its

30

relative intensity in the mass spectrum is denoted as 1(594); the dominant fragment ion
for this peptide is at m/z 374 - a K+ adduct of (Y3+H) whose relative intensity is denoted
as I(374). Current (A) is the direct heating current. The thermionic emitter was turned on
first (I = 2.65 A) to produce K1” ions. After five seconds of time delay (the radiative
heating may have caused some degradation of the peptide), the power for the direct
heating filament (the sample holder) was turned on to desorb samples on the filament. As

the heating rate is increased, fewer degradation products are formed and more intact

 

 

 

 

 

 

 

peptides are sampled.
Table 2.1 Degradation Vs. Desorption at Various Heating Rate
Current (A) 1(374)/I(594) % degradation* # of scans"
0 20:1 95 120
1.0 1:1 50 8
2.0 1:3 25 4
2.5 1:6 15 2

 

 

 

 

 

*% Degradation = 1374/1594. "'* Number of scans that signal law.

If fragment ions (Y3+H+K)+ are mainly from (M+K)+ ions - K+ adduct ions of
intact molecules, then the relative heating rate will not affect the relative ion abundances
very much. Even at a very high heating rate - much more intact gas-phase molecules are
formed, I(374)/I(594) should stay almost the same. This is obviously not the case, as
indicated by the data in the above table. Therefore, fragment ions are mainly from
attachment of K+ ions to thermal degradation products of neutral peptides.
Fragmentation of K+ adducts, is not significant, if any. Note in the data for experiments
with large heating currents, (Y3+H+K)+ ions partly are from radiative heating in the first

five seconds. Also, direct heating of the peptide can cause degradation to certain extent.

31

Therefore, all of the ion current for (Y3+H+K)+ may be from thermal degradation of

neutral peptides.

References

l.

2
3
4
5
6.
7
8
9

Bombick D.; Pinkston, J. D.; Allison J. Anal. Chem. 1984, 56, 396-402.
Light, K.; Kassel, D. B.; Allison, J. Biomed. Environ. Mass Spectrom. 1989, 18, 177-184.
Bombick, D.; Allison, J. Anal. Chem. 1987, 5 9, 458-466.

Bombick, D; Allison, J. Anal. C him Acta 1988, 208, 99-116

Kassel, D. 8.; Allison, J. Biomed. Environ. Mass Spectrom. 1988, 17, 221.
Rouse, J.; Allison, J. J. Am. Soc. Mass Spectrom. 1993, 4, 259-269.

Light, K., Doctoral Dissertation, Michigan State University, 1990.
Blewett, J. P; Jones, E. J. Phys. Rev. 1936, 50, 464-468.

Bombick, D. Doctoral Dissertation, Michigan State University, 1986.

10. Beuhler, R. J .; Flanigan, E; Greene, L. J .; Friedman L. J. Am. Chem. Soc. 1974, 96, 6.

11. Allison, J.; Ridge, D. P. J. Am. Chem. Soc. 1979, 101, 4998.

Chapter 3

Peptide Analysis by K+IDS Mass Spectrometry

I. Introduction

Since the advent of the K+IDS mass spectrometric technique, K+IDS has been
used for the analysis of various organic compounds. Application of this technique for
peptide analysis was of great interest to the members of Allison group. In the first paper
[1], K+IDS spectra of three peptides, ranging from di- to tetra- peptides, were published.
With the original single-filament probe, the spectra obtained seemed to be complicated,
containing both K+ adducts and surface ionization products. The formation of the
fragment ions were explained by the 1,2-elimination mechanism shown in Figure 3.1.
However, an obvious ion series was not observed for either of the three peptides.

In another paper [2], a sample peptide, hexaglycine, was analyzed and an
excellent mass spectrum was obtained. Dominant fragment ions were K+ adducts of
thermal degradation products from the cleavage of amide bonds (peptide bonds) with the
shift of a hydrogen atom. The cleavages of these peptide bonds were assumed to be a
low-energy 1,2-elimination process. Based on group equivalent values, the energy
required for the process was estimated to be - 5 kcal/mol, suggesting an exothermic
process. This is unreasonably low because such processes are expected to be
endothermic. The cleavages of CHR-NH bonds were also observed in the spectrum, but
with much lower abundance. Cyclization processes were proposed for the formation of

some of the ions.

32

33

 

 

1, 2 - Elimination mechanism

I o t
——> NHZMC=C=O + H-NHWCOOH

NHZMC
PVT Nwacoorr

 

 

Figure 3.1 1, 2 - Elimination mechanism for the CO-NH bond cleavage of peptides.

34

The spectra in these papers demonstrated that K+IDS could be used for the
analysis of peptides. However, obtaining a good spectrum requires a trial-and-error
process because the spectrum of a peptide was highly dependent on the bias voltage set
for the thermionic emitter/sample holder due to the single-filament design of the K+IDS
probe.

In 1988, a two-filament probe was introduced for K+IDS experiments [3]. This
probe, using radiative heating for both degradation and desorption, was reported to be 40
times more sensitive than the original one with one filament. Because the sample is no
longer on the thermionic emitter, changes both on thermal degradation of peptides and
ionization mechanisms might be expected. It was reported that the spectra for many
compounds were almost exactly the same, as indicated in K. Light's dissertation [4].
Similar mechanisms may have been involved. However, no detailed research had been
done on peptide analysis.

In this Chapter, the analysis of peptides with K+IDS, will be examined, the
applicability of this technique to peptide sequencing is evaluated, and the advantages and
limitations of K+IDS for peptide sequencing are discussed. The mechanisms for thermal
degradation of peptides - a very important aspect of this technique, will be investigated

in Chapter 4 in detail.
II. Experimental

Mass spectrometry: The K+IDS technique was performed on an HP 5985
GC/MS/DS quadrupole mass spectrometer. The K+IDS probe was a two-filament direct
insertion probe. One filament is the thermionic emitter, producing K+ ions in the gas
phase. The other is the sample holder, on which sample is radiatively heated and
desorbed. A detailed description of this probe is given in the K‘l‘IDS setup section in
Chapter 2.

35

Chemicals: All peptides were purchased from Sigma Chemical Company, St.
Louis, MO, and used without further purification. The samples were dissolved in
methanol to a concentration of 10 pg/pl, or suspended in methanol. One to two
microliters of solution were transferred to the sample filament. The sample was dried
before being inserted into the mass spectrometer.

Peptide derivatization: N-terminally acetylated peptides such as Ac-VGVAPG-
0H, were formed from the free peptide using acetic anhydride and a small amount of
pyridine as a catalyst, by the procedure described by Knapp [5]. For peptide methyl
esters such as H-VGVAPG-OCH3, a 2N methanolic HCl was prepared first by slowly
(dropwise) dissolving 20 pl of acetyl chloride into 100 pl of methanol. One mg of
peptide was dissolved in 50 pl of the methanolic HCl. Five minutes after the reaction, the
mixture was dried in a Speedvac(trademark) to remove the excess reagent and the residue
re-dissolved in methanol. The mixture is dried because the HCl can corrode the rhenium
wire (sample holder). (For N-acetylation and C-esterification at the same time, a
procedure similar to that for N-acetylation was used. One milligram of a peptide was
dissolved in a mixture of 50 pl of 10% of pyridine and 50 pl of acetic anhydride. The

reaction requires two weeks for completion).

III. K+IDS Mass Spectra of Typical Peptides

A. Nomenclature of K+IDS Fragments

To precisely designate the neutral species and ions formed in K+IDS, a variation
of the original Roepstorff nomenclature [6] is used. For a peptide NH2~~CHRCO-
NH~§~COOH, if the interior amide bond cleaves, two fragments (radicals) will be formed.
The N-terminal fragment, NH2~~CHRCO', is called a Bn fragment and the C-terrninal
fragment, 'NH~~COOH, is called a Yn fragment. The ions formed in the K+IDS

experiment are K+-adducts of even-electron species. Skeletal bond cleavages are

36

A. Roepstorff Nomenclature“ for Peptide Fragments
An 3,, cu

HZNMCHR——CO--NH-- CHR'MCOOH

 

 

 

X. i?— 2..

* P. Roepstorff and J. Fohlman, Biomed. Mass Spectrom. 11, 601(1984)

B. Nomenclature for Neutral Degradation Products

An-H Bn-H Cn+H
_‘ —‘ _.I_ H
HZNMCHR——CO--NH— CHR'MCOOH

H—-. H——.

XII}! fin Zn-H

 

 

 

C. Nomenclature for Ions In K+IDS

(Cn+H+K)"’

(B..-H+K)+
(An-H+K)t———
HzNMCHR—-C0-—NH "' CHR'MCOOH
H-—- (Zn-H+K)+
H——- — (Yn+H+K)+
~— (Xn+H+K)+

 

-——rr

 

 

 

 

 

Figure 3.2 Nomenclature for peptide fragments and ions.

37

accompanied by a H-shift or an OH-shift in thermal degradation processes for peptides.
Therefore, ionic designations such as [Yn+H+K]+, [Bn-H+K]+, [Cn+H+K]+, and
[Bn+OH+ K]"' will be used, where n is the number of amino acid residues (or number of
side chains) contained in that fragment. From this designation scheme, the m/z value of
each ion under discussion is clearly defined. The details of the nomenclature scheme,
indicating the types of ions observed (the types of thermal degradation products
generated), are given in Figure 3.2. The K+ adduct of each intact peptide will be
designated as [M+K]+.

In peptide chemistry, one letter designations are frequently used for amino acid
residues. To designate a peptide completely, the H at the N-terrninus and the OH at the
C-terrninus will be specifically indicated by a dash line (H- and -OH). To avoid
confusion between histidine (H) and hydrogen (H), histidine is underlined as [:1 whenever
a confusion might occur.

With peptides of known amino acid sequence, m/z values of fragment ions can be
calculated by using a commercial program named MacProMass 1.0. This program was
designed for the calculating the m/z values of peptide fragments from protonated
peptides. By making a slight change on the programmable termini, this program can be
used to calculate the m/z values of K+IDS fragments. For example, the (Yn+H+K)+ ions
in K+IDS are 38 mass unit heavier than the corresponding y (Y”) ions from FAB, setting-
up a C terminus to -OK instead of -OH can give m/z values of the C-terminal fragments.
By changing the N terminus from H- to K-, the m/z values of N-terminal fragments for

K+IDS can be calculated.

B. Typical K+IDS Spectra of Peptides
In the K+IDS experiment, microgram quantities of an analyte are placed on a
rhenium wire, and the temperature is increased from room temperature to approximately

300 °C in 10 seconds. The temperature is then held at this value until the sample is

38

depleted. During the experiment, K+ ions attach to desorbed neutral molecules, the mass
spectrometer is repetitively scanned, and a set of mass spectra is collected. For some
samples, there is a time (temperature) dependence of the resulting mass spectra since each
degradation process has a unique activation energy, which leads to the differences in the
relative intensities of mass spectral peaks in different scan files. However, what is
frequently observed is that a number of processes begin to occur at approximately the
same temperature. This might be expected for peptides since, even though each amino
acid residue has a unique side chain, peptides consist of only three types of skeletal
bonds. 1

When the tetrapeptide H-FGGF-OH is rapidly heated, and K+ ions are attached to
molecular species that desorb into the gas phase, typical experimental results are shown
in Figure 3.3a. As desorption and thermal degradation occur, spectra are obtained. The
spectra collected are summed, and a single K+IDS spectrum is reported, which will
provide information on the low-energy therrnal-degradation processes. Note that, as the
thermal degradation chemistry occurs, some of the peptide molecules desorb intact. A
temperature is reached at which desorption and thermal degradation compete. If a
peptide decomposes on the surface to two smaller molecules, they appear to desorb
promptly rather than remain on the surface and decompose further. Thus, the experiment
allows for the primary degradation processes to be investigated. Following such primary
unimolecular processes, prompt desorption of thermal degradation products occurs
because of the temperatures used.

The spectrum shown in Figure 3.3b is a typical K+IDS mass spectrum for a small
peptide. All of the ions detected are K+-adducts. For example, the peak in the spectrum
at m/z 465 represents a K+ (39 Da) adduct of the intact analyte molecule (426 Da). The
peak at m/z 261 represents a K+ adduct of a thermal degradation product with a mass of
222 Daltons. A fragment of the peptide can be generated with this mass by cleaving the

central amide bond, with one additional hydrogen atom, to yield (Y2+H), which can also

Relative Intensity

39

Time Dependence of Spectra: H-FGGF-OH

 

 

‘ (82-H+K)*

., I
|

_ x 10 XF1 I
: (Y2+H+K)+
| 261

- l

(Y1+H+K)+ I
‘ 204 I 243

 

.nl

 

(M440+

b)

(Y3+H-o-K)+
318

 

 

140 160 180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500

m/z

Figure 3.3 K+IDS mass spectra of H-FGGF-OH:
a) Temperature dependence; b) Average spectrum.

40

be written as H-GF-OH. Thus, the ions at m/z 261 are designated as [Y2+H+K]+ or [H-
GFvOH]K+. By correlating K+IDS mass spectra with the su'uCtures of many peptide
analyte molecules, trends in the degradation chemistry become clear, and ionic
designations become apparent. Note that the [Yn+H+K]+ ion series in the spectrum
(n=1-3) provides information on the amino-acid sequence of the peptide and K+-adducts
of the intact, desorbed peptide provide molecular weight information. Thus, K+IDS is a
useful mass spectrometric method as well as a tool for understanding thermal degradation
chemistry.

Figure 3.4 shows the K+IDS mass spectrum of H-VGVAPG-OH as a second
example. Again, the spectrum shows that both the intact molecule and its thermal
degradation products can be formed simultaneously. All of the ions observed are 10"-
adducts of closed-shell molecules that can be related to the parent molecule, formed by
skeletal bond cleavages with accompanying hydrogen atom shifts. For this peptide, the
ion series (Yn+H+K)+ is observed, which can be used to deduce the sequence of the
peptide.

' The spectra shown in Figure 3.3 and Figure 3.4 are typical of the K+IDS results in
the following aspect: the dominant ions are K+ adducts of degradation products that
result from cleavage of amide bonds even though these bonds are the strongest skeletal
bonds. While two neutral products, (Yn+H) and (Bm-H), are generated (simultaneously
and in equal quantities) when an amide bond is broken, the [Yn+H+K]+ ions are more
intense than the corresponding [Bm—H+I(]+ ions. This observation will be discussed in

Chapter 4 .

C. K+1DS Spectra of Polar Peptides and Dehydration Reactions

In K+IDS experiments, peptides with amino-acid side chains containing very

polar functionalities have shown significant dehydration reactions. The major difference

41

 

H-VGVAPG-OH
a:
a ..
B
.53
o-
>
1::
.53 - (AP+K)_.
32 207

1

   

 

140 180 220 260 300

Figure 3.4 K+IDS mass spectrum of H-VGVAPG-OH.

(82-H+K)+ (Y 2+H+K)+ (Y3+H+K)“'
- 195 211 282
\ / /

(C3+H+K)+
3 l l

 

(Y4+H+K)+

381

(Y 5+H+K)"’ (C5+H+I()+
438
\

340 380 420
m/z

(M+K)“'

(M-H20+K)*
519

     

479
/’

460 500

 
  

537

 

 

540 580

the

aria.

42

between peptides containing alkyl, aromatic or H side chains and peptides with other side
chains (more polar) lies in their polarities and chemical reactivities.

The high polarity of some side chains, such as a carboxyl group in glutamic acid,
has direct consequences in the K+IDS experiment. More polar peptides lead to stronger
interactions between peptides, leading to stronger interactions between peptides and the
sample-holder surface, greater vaporization energies and longer heating time before
desorption. The extended heating-time before a peptide molecule can be desorbed while
it acquires energy from heating, can lead to extensive thermal degradation such as
skeletal bond cleavages as well as dehydration. Consequently, there are little or no intact
(M+K)+ ions detected for such peptides. When this occurs, only a few reactions will
dominate and the resulting spectrum can provide only limited sequence information of a
peptide.

For peptides with OH-containing residues such as serine (R = -CH20H) and
threonine [R = -CH(OH)CH3], the side chains are very polar. Peptides containing these
residues do not show much fragmentation (thermal degradation). This result, however,
can not be explained at this time. With -COOH and -NH2 containing residues such as
glutamic acid (R = -CH2CH2COOH), aspartic acid (R = -CH2COOH) and lysine (R =
-CH2CH2CH2CH2NH2), dehydration frequently occurs to a significant extent, both for
intact peptide molecules and for their thermal degradation products.

With these dehydration reactions, ions such as (Yn+H+K-H20)+ and (M-
H20+K)+ are formed. These dehydration ions such as (Yn+H+K-H20)+ ions may be
from dehydration of regular (Yn+H+K)+ ions. They also can be formed directly from
thermal degradation of (M-H20+K)+ ions. It is generally believed that K+ adducts
formed do not produce fragments. Therefore, these ions are most likely the adducts of
their corresponding degradation products. Dehydration probably occurs before K+

attachment. The extent of dehydration varies with amino-acid composition of a peptide.

43

When a glutamic acid/aspartic acid residue is present in a peptide, the extent of
dehydration increases significantly, as is the case in the spectrum of H-FLEEI-OH
(Figure 3.5). This result suggests that carboxylic acid group on the side-chain is involved
in the dehydration process. Frequently, this will generate ions such as (Yn-H20+K)+
ions, (Yn-H20-NH3+I()+ ions, as well as dehydration of the pseudo-molecular ions, (M-
H20+K)+. It has been found that the shifting hydrogen in dehydration is from either the
N-terminus or amide group (see Chapter 4 for details). Many other peptides, such as H—
VGSE-OH, HCO-AGSE-OH and H-RKEVY-OH, show similar dehydration.

As mentioned above, dehydration from side-chain carboxyl groups is much easier
than that from the C—terminus of a peptide. This is probably because these carboxyl
groups can reach farther in various directions due to the longer side chains than the C-
terminal one.

With lysine-containing peptides, such as H-PFGK-OH, significant dehydration
products are also observed as shown in Figure 3.6. A series of dehydration ions such as
(Y2-H20+K)+, (Y3-H20+K)+, (M-H20+K)+ and (M-2H20+K)+ are observed for this
peptide. One plausible explanation for this dehydration process is that the shifting H is

frOm the side-chain NH2. In the dehydration process, intramolecular dehydration occurs

and forms a cyclic amide.

 

I l l I
~~~~coon H-NH~~~~ -----> H20 + ~~~~CONH~~~~

K+IDS spectra of Other lysine-containing peptides such as H-TSK-OH and H-
' KHK-OH also show similar dehydration reactions. However, the loss of the second water
molecule has not been explained. More data are needed to uncover the reason.

When a tyrosine residue is present in a peptide, dehydration generally increases

though the extent varies for different peptides. For the peptides H-YGGFM-OH and H-

 

H-FLEEI-OH

J (Y2+H+K)+
299

(Y2+H-H20+K)+

- (Y1+H+K)+ 281
170

 

130 170 210 250 290 330

Figure 3.5

410

(Y +I-I+K)+

42: ($144910+ (”my

I 541 688

I _36|

:(Y4+H-H20+t<)+ (134-H+K)+ ' l
~H20 l 523 557 : 181
392 |

 

370 410 450

m/z

(Y3+H-H20+K)+

| 670

\/

570 610 650 690 730

 

490 $30

K+IDS mass spectrum of H—FLEEI-OH.

45

 

H-PFGK-OH

(Bz-H-H20+K)+
24$

(Y2+H+1<)+
242
I
(Y2+H-H20+K)+I
224 l
l

. 198
167

 

 

 

130150170190 210 230 250 270 290 310 330 350 370 390 410 430 450 470 490 510

Figure 3.6

(Bz-H+K)“’
283

m/z

(Y3+H-H20+K)+
37 l
I

l 389

K‘TDS mass spectrum of H—PFGK-OH.

I (Y 3+H+K)+

(M-H20+K)“’
468 \

(M-2H20+K)"'
450

 

(M+K)+
486

 

 

46

FGGFM-OH, their K+IDS spectra are almost the same except that the former gives
slightly more dehydration products. This indicates that the dehydration involves the
tyrosine side chain, but is not very significant. For the peptide H-YGG-OH, however,
K+IDS spectrum suggests a significant dehydration of the peptide. Because this peptide
can be desorbed into the gas phase more easily than H-YGGFM-OH, due to its lower
molecular weight (less heat will be absorbed before desorption), the dehydration must be
position dependent. For H-YGG-OH, there is a favorable interaction between the C—
terminal carboxyl group and the phenolic OH on the tyrosine side chain (or N-terminal
NH2) due to the closeness of the two functional groups. This leads to dehydration. For
H-YGGFL-OH, however, these two groups are too far away for a similar reaction to
occur.

Dehydration also occurs for peptide amides even though there are no carboxyl
groups within such molecules. Obviously the reaction proceeds via a different
mechanism. We postulate that the dehydration occurs most probably at the C-terminus

involving the following rearrangement followed by a 1,2-elimination:

NH2~~~CONH2 ----- > {NH2~~~C(OH)=NH} ----- > NH2~~~CN + H20

Other peptide amides such as H-YGGFL-NHz, H-GGFL-NHz, H-YPLG-NHZ and
t-BOC-WMDF—NHz have shown the same dehydration reactions.

When a peptide is too polar, a K+IDS spectrum may not provide information
about either molecular weight or structure of the peptide because the molecules will be
very difficult to desorb into the gas phase. In such a case, some decomposition will occur
and the molecular ions will not be detected. For example, the peptide H-RKEVY-OH
(Figure 3.7) gives only ions from the cleavages of the second amide bond from the N-

terminus, forming (Bz-H) and (Y3+H). No (M-t-K)+ ions are detected.

47

 

(Y3+H+K)+
430

H-RKEVY-OH

(M+K)+
7 1 5

W

130 170 210 250 290 330 370 410 450 490 530 570 610 650 690 730

 

 

 

 

Figure 3.7 K+IDS mass spectrum of H-RKEVY-OH.

sir

dUI

cl:

PCP
H-C

1111.
add:

itpn

mu

48

D. K+IDS Spectra of Peptide Derivatives

1. Peptide Amides

Amidation is one of the most common transformations of peptides [7]. Therefore
analysis of peptide amides is of great importance in biochemistry.

When a peptide is converted into a peptide amide, the K+IDS spectrum changes
dramatically. For a regular peptide H-YGGFL-OH, the dominant ions are (Yn+H+K)+
ions, with (Y3+H+K)+ represented by the base peak of the spectrum. When the
corresponding amide, H-YGGFL-NHz, is analyzed, the dominant ions become
(Cn+H+K)+ ions and (Yn+H+K)+ ions, and the base peak represents (C3+H+I()+ (Figure
3.8).

These observations can be explained by a cyclization mechanism which will be
discussed in Chapter 4. It has been observed that cyclization via the termini of peptides is
the dominant reaction in thermal degradation of peptides.

For H-YGGFL-NHz, cyclizations from both termini are competitive due to
similar steric hindrance. The dominant cyclization from the C -terminus (C3+H+K)+ is
due to the loss of cyclo-(FL) from the N-terminus. The weaker NH-CHR bond probably
is the main reason why the cleavage forming (C3+H+I()+ is more favored than the
cleavage forming (Y3+I-I+K)+.

The change of a peptide Spectrum is not always so dramatic when a regular
peptide is converted into a peptide amide. For example, the spectra of H-GGFL-OH and
H-GGFL-NHz (Figure 3.9) are very similar except that ,(Y1+I-I+K)+ ions of the former
at m/z 170 (weak) is replaced by (C2+H+I()+ ions of the latter at the same m/z value. In
addition, the "apparent" isotopic peak at m/z 317 is not an isotopic peak; it actually
represents a (C3+H+K)+ ion, which can be easily overlooked.

There are several possible reasons for this small change: 1. C1 and C2 are too

small to respond well (see Chapter 5 for size discrimination). 2. Since the N-terminal

49

 

a
‘ (Y3+H+K)+
374

H-YGGFL-OH

 

 

 

 

 

 

 

 

 

— , - +
(Bz-H+K)+ (M HZOK)
I 259 (Y 2+H+K)+ (M-Hzo-NH3+K)+ :
- 317 I
(Y +H+K)+ +
(Y2+H+K)+ I 4 431 IMHO
I 335 |594
170 l
1: rm‘ rrrT 41 fr? Inf} ‘1“: 1114.?1 ' :‘v‘?““r‘n“+fi‘r“n#‘r ”I“: 113*]?11“ ‘ h
I I I I I I I 7 I I I I
130 170 210 250 290 330 370 410 450 490 530 570 610 650
m/z
. b (C3+H+K)+
H-YGGFL-NHz 333
(M+K)+
‘ 593
(Y2+H+K)+
316
_ I
I (Y3+H+K)+
. l 373 'HZO
I
141 {—1 fiTur TA ?‘?1 i ‘rili (“1+3 ‘Tur I ll- Ltut‘t IMMIIAI l‘?A-T‘+AT‘I AInf‘I‘Wh 1;?7 Ii“ I t f i
I I I I I I I I I I I I

 

380

m/z

140 180 220 260 300 340 420 460

Figure 3.8

500

540 580 620

K+IDS mass spectra of: a) H-YGGFL-OH; B) H-YGGFL-NHZ.

 

a (M+K)+
431
_ H -GGFL-OH \
2‘
g .
.5-
0
>
'5
g 4
U
a:
317 374

 

 

 

160 180 200 220 240 260

AA A - gg -4; - _“- A L - .. _ ..
TrleIIYIIUTTrITTIIlTIIIITTITIIIITIYIYYIYITTITIYVITIIUIIIIIIIITTIIT T‘I'FIIT‘I

280 300 320 340 360 380 400 420 440 460

 

 

 

 

 

m/z
b (M+K)+
H-GGFL-NHz .30
‘ X 8
x?
U, .
G
g + X 1
7;- (Y3+H+K)
'1: 373
.3 ‘ (C3+H+K)+
“_ + (Y2+H+I<)+ 317
. 170 281 l
IIIVIIIVIIIIIII'UfiIYYYTI'TIVI'AIYIY‘IIYTI‘ITII'lIIIIITlW—V] TTTIWIT'I‘IVV

150 170 190 210 230 250 270 290 310 330 350 370 390 410 430 450

Figure 3.9

m/z

K‘I’IDs mass spectra of: a) H-GGFL-OH; b) H-GGFL-NHz.

 

 

 

51

glycine residues present such a small steric hindrance for cyclization from the N-terminus

(see Chapter 4 for details), formation of (Yn+H) is still favored.

2. N-terminally derivatized peptides

The structural analysis of N-terminal derivatives of peptides is of great interest to
biochemistry. For example, N-acetylation is a common transformation of peptides for the
N-terminal amines and amines at lysine side chains [7]; CBZ (Carbobenzoxy-:
C6H5CH2-OCO-) and t-BOC (N-tert-butoxycarbonyl: Me3COCO-) are important N-
terrninal protection groups used in peptide synthesis.

When the N-terminus of a peptide such as H-VGVAPG-OH is blocked, by
acetylation for example, the appearance of the spectrum changes dramatically. For the
regular peptide, the dominant ions are (Yn+H+K)+ ion series. With the N-terminal
derivatives of the peptide, these ions become very weak or absent. Instead, the dominant
ion series become (Bn+OH+I()+ ions (Figure 3.10). Because the only possible OH
source is at the C-terminus of the peptide, these ions must be formed via cyclization from
the C-terminus because the regular cyclization reaction from the N-terminus is blocked
by the acetyl group.

N-terminally blocked peptides also produce (Cn+H+K)+ ions, such as
(C3+H+K)+, (C4+H+K)+ and (C5+H+I()+ ions in the above mentioned spectrum. The
formation of these ions will be discussed in Chapter 4.

Similar spectra have been found for other N-terminally blocked peptides such as
CBZ-GPGGPA-OI-I (Figure 11a), CBZ-GPFPL-OH (Figure 11b), t-BOC-PPPP-OH
(Figure 11c), Ac-VGGFL-OH, Ac-YGGFL-OH, HCO-AGSE-OH. In addition to
(Bn+OH+K)+ and (Cn+H+K)+ ions, other ions such as (Yn+H+K)+ ions are also
observed, but with lower abundances.

It should be pointed out that the large N-terminal derivatives show characteristic

neutral losses from the pseudo- molecular ions. Loss of C6H5CH20H (-108) from MK+

Relative Intensity

52

 

 

 

 

 

 

(M+I()+
‘ Ac -VGVAPG-OH (B5+OH+K)+ [79
- 522
1 (C4+H+K)+ (1"1'H20‘q“)+
I /
‘ l
(c3+rr+r<)+ (B3+OH+K)+ l (34+OH+K)+
4 353 354 I 425
. (132+<3H+K>+ \ |
255

130 170 210 250 290 330 370 410 450 490 530 570 610 650

Figure 3.10 K+IDS mass spectrum of Ac-VGVAPG-OH.

S3

 

 

    
 

 

 

a (M+K)+
‘ CBZ-GPFPL-OH 702
+ +
‘ (123'me (B3+0H;91:)\(B4+0H:; $311501on
I (B2+OH+K)+ (YrerHO+
d ' 34 5 51 l “20

 

 

240 280 320 360 400 440 480 520 560 600 640 680 720 760

 

 

 

 

 

 

b (B3+OH+K)+
‘ 402
CBZ-GPGGPA-OH (34+0H+K)+
‘ 459 (M+K)+
(35+0H+K)+ 627
(32+OH+K)+ i
— [cyclo-(GP)+K]"’ i I
193 I I
I
I [cyclo-(PA)-1>K]+ I I 4120
-I I 207 I 365 : 591
I I (Y +K)+ 345 4120 4 -H20
l l :82 326 384 75 556 609
_ 1' I ‘ I

 

160 200 240 280 320 360 400 440 480 520 560 600 640

m/z

Figure 3.11 K+IDS mass spectra of: a) CBZ-GPFPL-OH; b) CBZ-GPGGPA-OH.

 

 

 

(M+I()+

‘ t-BOC-PPPP-OH 545\
d 445

4 I -(Me3COCO-H)
, I -100

[cyclo-(PP)+K]+ I

I 233 I I (B3+OH+K)+

' -CHMe3

- I l 4— (132+oH+r<)+ I 443

I l 293 351 l

263

 

 

210 230 250 270 290 310 330 350 370 390 410 430 450 470 490 510 530 550 570

m/z

Figure 3.11c K+IDS spectrum of t-BOC-PPPP-OH.

55

has been found for CBZ derivatives of peptides; loss of 100 (t-BOC- 1) from the MK+ has
been found for t-BOC derivatives. Therefore, neutral losses of 108 and 100 are indicative

of CBZ and t-BOC derivatives, respectively (Figure Ila-c).

3. Peptides with C-terminal esterification

When the C-terminus of a peptide is blocked with a methyl ester group, the ion
series from K+IDS stay the same as for the underivatized one, with all (Yn+H+K)+ ions
14 mass units higher. This can be seen in the K+IDS mass spectra of H-VGVAPG-
OCH3 and H-YGGFL-OCH3 (Figure 3.12). This is because cyclization from the C-
terminus of a peptide becomes even more difficult than in the case of an underivatized
peptide while cyclization via the N-terminus remains the same. Also found is the
common loss of methanol (-32) from the molecular species which is indicative of a

methyl ester.

4. Peptide variants

H-YGGFL-OH and its derivatives, due to their commercial availability and low
cost, were studied in the early stage of this research. With K+IDS spectra of these
peptide variants, the aromatic side chains (Ry = ~CH2C6H4OH and RF = -CH2C6H5)
were suspected to be the major reason for the formation of the major ions such as
(Y3+H+K)+. To test this speculation, two similar peptides, H-AGGFL—OH and
H-VGGFL—OH, were synthesized and their K‘I'IDS mass spectra were compared with
that of H—YGGFL-OH.

When the first amino acid residue of peptides H-XGGFL-OH (X = Y, A, V) is
changed from tyrosine (Y) to alanine (A) to valine (V), no obvious change in the spectra
is observed (see Figure 3.8, Figure 3.13 and Figure 3.14). This result indicates that the

size of the side chain of the first amino-acid at the N-terminus is not the reason why

 

 

 

 

 

 

56
a (MAW
H-YGGFL-OCH3 608

(M-HzO-NH3+K)+
‘ (Y3+H+K)+ SIB
388 I31
. I
I I-18
.4 I I‘—
(Y1+H+K)+ (Y2+H+K)+ (Y4+H+K)+ I I
4 184 331 445 I590
.‘+3.‘....*..*.. .4.*. ...I #4.: we. ..I. I... .Aniv-va-L-w-T
I I I I I I I I I I I -

140 180 220 260 300 340 380 420 460 $00 540 $80 620 660

 

(M+K)‘*'

b
‘ H-VGVAPG-OCH3 55,

 

 

 

 

 

‘ (Y2+H+K)+ (Y4ggngV
225

I

I
‘ I (Y H K+

5+ + )
B ~H+K +| W3+H+K>+
. ( 2 203) I 296 452 533
l-H-H-I‘I'fi-f'r-IITFI‘. . n7n LIL. 7* 7?: r 113‘ 1M‘F‘1r. L‘awitlr ..*. Lw‘ ‘7"???.
I I I I I I I I

130 170 210 250 290 330 370 410 450 490 530 570 610

Figure 3.12 K+IDS mass spectra of: a) H-YGGFL-OCI-Ig; b) H—VGVAPG—OCH3.

S7

 

 

 

 

 

 

 

 

 

 

a
. (Y3+H+K)+
H -VGGFL-OH 374
3‘
.23 q
:3 X 5 X l
f _ (M+K)+
u? 530
.53 ..
U
M
‘— + (Y2+H+K) +
195 431
Ar‘r“! I if! 1 r‘ 1.11141 r t-r r‘lr LL‘I 1J1! ‘Ir r ‘i‘rTr‘ 1 173‘ L H?! 1‘? ‘
I I I I I I I I I I I
150 190 230 270 310 350 390 430 470 510 550 590
2
b CM+K)+
H-GGVFL-OH 5,,
a?
U)
C .1
B
5 _
D
g (Y3+H+K)+ (Y4+I-I+K)"'
34’ 416 473
-I
130 170 210 250 290 330 370 410 450 490 530 570 610
m/z

Figure 3.13

K‘I’IDS mass spectra of: a) H-VGGFL-OH; b) H—GGVFL-OH.

 

 

 

 
 
 
 

  

 

 

 

 

Figure 3.14 K+IDS mass spectra of: a) H-AGGFL-OH; b) H-(B-A)GGFL-OH.

 

 

 

   

490 530 570

a
. H -AGGFL-OH (Y3+H+K)+
0 374
ha
3% HzN
34 GFL- on
15 X 4
0—
.2
E
SE .
H +
a (Y2+H+K)+ “4+ +K) (M+K)+
("1+H+K)+ (GF K) + 317 431 502
+
170 243
130 170 210 250 290 330 370 410 450 490 530
m/z
4b M+K)+
H-(B-A)GGFL-OH MM“... 5,.
‘ 485
>1 HzN GFL' OH
'5 (Y3+H+K)+
3~ 374
.5
.3 4 (Y4+H+K)+
33 431
or (Y2+H+K)"'
+ 317 467
”VINO 299 339
.J
130 170 210 25) 290 mm/ 31) 410 45)
Z

 

 

59

 

 

  

  
 
 
 

 

 

 

 

 

 

 

Figure 3.14

a
« H -AGGFL-OH (Y3+"*K)+
374
z...
'17}
= .
8
.5. x 4
Q—
.2
I!
§ .
(Y4<1-H+K)+
-* (Y 2+H+K)+ (M+K)+
(Y1+H+K)+ 317 431 502
170 (GF+K)+
243
I
130 170 210 250 290 330 370 410 450 490 530 570
Hill
C
‘ H K "‘
H-AG(Sar)FL-OH “3* + ’
388
.J
5‘
2 .
8
5 _
o
{-3 (M+K)+
% - 516
a: (Y4+H+K)"'
‘ 445
. (Y2+H+K)"' I
317 I
1%? [MATT I“! flurrrAQ‘flA? +1711 TYTAAIAAI‘UIlAIIIIIIA UTIAT +rrr‘r‘ri‘341111
I I I I I I I I I I I I
130 170 210. 250 290 330 370 410 450 490 530 570 610
Z

K'I'IDS specta of: a) H-AGGFL-OH; c) H-AG(Sar)FL-OH.

 

6O

cyclization occurs. It also may be concluded that aromatic side chains have nothing to do
with the dominant cleavages which lead to the formation of the major ions.

Spectra of two peptides (Figure 3.13), H-VGGFL-OH and H-GGVFL-OH, with
the same amino—acid composition but different sequences, were also examined. When the
sequence of a peptide is changed (not the amino acid composition) as in this case,

changes in the relative abundances of the same ion series are observed. For example,

(Y4+H+K)+ is as abundant as (Y3+H+K)"' for H-GGVFL-OH. For H-VGGFL-OH, the
(Y4+H+K)+ is much less abundant than (Y3+H+K)+ because the formation of the later is
more favored. Similar reactions have been observed when the N-terminal amino acid is
glycine. For example, with H-GGFL-OH and H-GPGG-OH, loss of the N-terminal
glycine is also favored. Therefore, the presence of glycine at the N-terminus of a peptide
may affect bond cleavages in thermal degradation of peptides.

Three peptide variants with the same sequence, but with simple modifications
along the peptide chain, were studied. H-AGGFL-OH, H-(B-A)GGFL-OH and H-
AG(Sar)FL-OH are the three peptides studied. The B-alanine residue
(NH2CH2CH2COOH) has its NHz at the B-position while the regular alanine has its
NHz at the a-position. Sarcosine [NH(CH3)CH2COOH] is very similar to glycine
(NH2CH2COOH) except that one of the N-terminal hydrogens is replaced by a methyl
group. When a modification is made to the peptide skeletal bonds, the thermal
degradation pattern may change. For example, when the N-terminal a-alanine is changed
into flalanine, the spectrum changes significantly in terms of relative intensities of all the
ions (Figure 3.14). With a-alanine, (Y3+H+K)+ ions dominate the spectrum, indicating
that there exists one dominant degradation reaction. With B-alanine, all ions formed are
of comparable abundances, suggesting that there is no dominant reaction. Also the
intense peak of (M+K)+ indicates that this peptide is much more stable than the former

due to the difference in amino group positions of the N-terminal amino acid residue.

61

No dramatic change has been noticed when one of the amide hydrogens was
replaced by a methyl group (glycine to sarcosine). The spectra of two peptides, H-
AGGFL-OH and H-AG(Sar)FL-OH, are comparable, with (Y3+H+K)+ dominant. This
suggests that the N-terminus of a peptide affects more than amides on thermal
degradation of peptides. However, the introduction of a methyl group does make the

peptide more stable, and it changes the distribution of degradation products as well.

E. Internal ions from peptides.

In FAB-M/S, an internal ion is formed when two skeletal bonds are cleaved with
the charge on the middle fragment. In K+IDS, internal ions are also observed, especially
for proline-containing peptides. This is the same in K‘I'IDS. The internal ions observed

are shown in the following peptides.

W: W

H—VGGFL—OH lH-YGGFL-OH [cyclo-(GF)]K+ at m/z 243
H-VGVAPG-OH [cyclo-(AP)]K+ at m/z 207
CBZ-GPGGPA-OH [cycle-(GP)]K+ at m/z 195
t-BOC-PPPP-OH [cyclo-(PP)]I(+ at m/z 233
CBZ-GPFPL-OH [cyclo—(PF)]K+ at m/z 283

The presence of these internal ions usually provides only very limited structural

information and only can be used to confirm a partial sequence of a peptide.

IV. Peptide sequencing by K+IDS mass spectrometry

Peptide sequencing is one of the most successful applications of mass

spectrometry. The success of peptide sequencing by mass spectrometry lies in the basic

62

structures of peptides, which contain repeating units of ~NH-CHR-CO- with R different
for different amino acids. Because of this property, peptides contain only three basic
skeletal bonds: NH-CHR, CHR-CO and CO-NH. When consecutive ions from cleavages
of the same type of bonds such as CO-NH bonds, the mass difference of these two ions
reveals the mass of an amino acid residue, from which one amino acid residue is
identified (For isomeric leucine and isoleucine or for isobaric glutamine and lysine, other
ion series or derivatization have to be used for the differentiation). A complete set of
such ions, therefore, can give a complete sequence of the peptide. If one ion series is not
complete, only a partial sequence can be deduced from that ion series. Other ion series
have to be identified to c0mplete the missing sequence by overlapping the partial
sequences deduced from each ion series. In favorable cases, a complete sequence can be
finally determined.

The K+IDS spectrum of a peptide H-YGGFL-OH (Figure 3.8a), for example, can
be used to illustrate this technique. Because (Yn+H+I()+ ions are the dominant ions for
underivatized peptides, they will be used for the purpose of sequence analysis. The mass
difference of the first few peaks from the molecular ion peaks is too small to correspond
to any amino acid residue. The peak at m/z 431 is 163 mass units lower than the peak at
m/z 594, indicating that the first amino acid is a tyrosine residue. The next major peak at
m/z 374 is 57 mass units lower than the peak at m/z 431, indicating a glycine residueby
difference. This process is repeated for peaks at m/z 374, 317 and 170; and a complete
sequence is obtained for this peptide: H-YGGFL/I-OH. Note that isoleucine and leucine
cannot be distinguished with this technique.

For real peptide sequencing, interpretation of a mass spectrum of a peptide is not
always a trivial task. A mass spectrum can contain many incomplete ion series (including
ion series from side chain fragments) and interpretation of a spectrum can be very slow.
Computer programs [8-11] have been introduced to speed up such tedious processes and

the best sequence fit can be obtained within a few minutes.

63

For peptide sequencing with K+IDS, the most important ion series are
(Yn+H+K)+ ion series for underivatized peptides, (Cn+H+K)“' and (Yn+H+K)+ ion
series for peptide amides, and (Bn+OH+K)+ ion series for N-terminally derivatized
peptides.

There are some advantages for the application of this technique for peptide
analysis. It is easy to implement on any quadrupole instrument with a direct insertion
probe. The technique is inexpensive to use. The instrument is simple. Obtaining a
K+IDS spectrum is very fast and the spectra are easy to interpret. All of the ions are from
the analyte molecules since no matrix is used. However, the best application of K+IDS
analysis is still limited to peptide confirmation - both sequence and molecular weight -
due to the limitations shown below.

There are some serious limitations for peptide sequencing by this K‘I’IDS
technique. For di-peptides, no skeletal bond fragment ions can be observed due to fast
desorption without skeletal bond cleavages. Frequently, the only fragment ion is from
dehydration. For tri-peptides, sequencing could be difficult since peptides with low
polarity may not degrade significantly before desorption. The ideal peptides for analysis
by K‘I'IDS are tetra- to hepta- peptides. With larger ones, thermal degradation may
dominate and it can be difficult to obtain molecular weight inforrrration. With large
peptides or proteins, K+IDS spectra can be very complicated, if obtainable. Figure 3.15
shows the spectrum of a peptide of 22 amino acid residues from a mass spectrometer with
a mass range of 1000 Da. It can be seen that most of these peaks are not labeled because

thermal degradation of this large molecule is too complicated.

 

 

 

 

 

 

 

4 277 ‘25 (c5+1r+1()+
‘ (B4-H+K)+ 451
‘ I . (C6+H+K)+
‘ 247 295 I 355 394 563
4 306 536
. 322337
120g” 26 484 509 582
I
200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560 580 600
‘ (87-H+K)+
I 620 677
-I
I 906
4 705 762
651
.I 735 87‘ 029 963
J 794 815 83850 890

 

.1

600 620 640 660 600 700 720 740 760 700 800 820 840 860 000 900 920 940 060 9801000

 

Figure 3.15 K+IDS mass spectrum of a 22mer peptide Melittin (MW 2846.7).

The sequence of the peptide is H-GIGAVLKVL'I'ITGLPALISWIKRKRQQ-NHZ.

65

References

1. Bombick D.; Pinkston, J. D.; Allison J. Anal. Chem. 1984, 56, 396-402.

2. Bombick, D.; Allison, J. Anal. Chem. 1987, 59, 458-466.

3. Kassel, D. B.; Allison, J. Biomed. Environ. Mass Spectrom. 1988, 17, 221.

4. Light, K., Doctoral Dissertation, Michigan State University, 1990.

5. Knapp, D. R. "Chemical Derivatization for Mass Spectrometry" in Methods in
Enzymology, vol. 193, McCloskey, J. A., ed., Academic Press: San Diego, 1990;
pp 314-328.

6. Roepstorff, P.; Fohlmann, J. Biomed. Mass Spectrom. 1984, 11, 601-601.

7. Stults, J. "Peptide Sequencing by Mass Spectrometry", in Biomedical Application of
Mass Spectrometry, vol 34, Suelter, C. H.; J. T. Watson ; ed., Wiley: New York,
1990; PP145-201.

8. Johnson, R. 8.; Biemann, K. Biomed. Environ. Mass Spectrom. 1989, 18, 945.

9. Alberth W.; Straub K. M.; Burlingame A. L.; Anal. Chem. 1982, 54, 2029

10. Siege], M. M.; Basuman, N. Biomed. Environ. Mass Spectrom. 1988, 15, 333.

11. Sakurai, T.; Matsuo, T.; Matsda, H.; Katakuse, I. Biomed. Mass Spectrom. 1984, 11,

396.

Chapter 4

The Characterization of the Primary Thermal Degradation Processes of Peptides

I. Introduction

The pyrolysis/thermal degradation of organic molecules is a process that has been
used both for the production of chemicals and for analysis [1]. In the latter, mass
spectrometry (MS) has been an important tool in characrerizing products of pyrolysis (Py)
of organic compounds, both directly in techniques such as in Py/MS, and with an
additional separation step such as in Py/GC/MS. In pyrolysis studies, temperatures of
600-800°C are frequently used. The characterization of thermal degradation processes
can be challenging since, under some conditions, primary degradation products
decompose further and undergo subsequent chemistry, yielding complex mixtures of
products. To date, little work has been reported on the thermal stability/thermal
degradation of peptides in general; however information is available on the pyrolysis of
small dipeptides and amino acids [2].

In this chapter, the K+IDS technique will be utilized for the characterization of the
primary thermal degradation chemistry of peptides. The strength of the K‘I'IDS technique
is in the simplicity of both the experiment and the resulting mass spectra. Since each
thermal degradation product gives one peak in the mass spectrum, K‘I'IDS mass spectra
reflect a "molecular weight distribution" of the thermal degradation products of the
analyte molecules, which, in combination with analogous studies of derivatized and

isotopically-labeled compounds, can be used for the study of unimolecular, condensed

66

67

phase thermal degradation chemistry. The approaches presented here are designed to
provide insights into the structures of the thermal degradation products, and to
demonstrate a unique and simple tool for characterizing thermal degradation processes for
biomolecules, as well as Other thermally-labile compounds.

Because of the importance of the process, which has been detailed in part II of
Chapter 1, a thorough study has been conducted in this chapter. Deuterated peptides have
been prepared to trace the source of the shifting hydrogen atoms. Derivatized peptides
have been used to determine the roles of peptide termini in thermal degradation.
Specifically designed peptides have been synthesized and used to test the proposed
mechanisms. The effects of varieties of the termini of peptides and steric hindrance have
been considered in thermal degradation. At the end of this chapter, thermodynamics and

kinetics are used to support the proposed mechanisms.

II. Experimental Section

All mass spectrometric analyses were performed on an unmodified HP5985
GC/MS/DS quadruple mass spectrometer with a mass range of 101,000 Daltons
equipped with a modified direct insertion probe, similar to that reported previously by
Rollgen et a1. [3] and by Kassel and Allison [4]. The probe places two wire filaments
into the ion source. One filament serves as the thermionic emitter, generating K+ ions
I when resistively heated. The other filament serves as the sample holder from which the
analyte molecules undergo thermal degradation and desorption. The sample holder is
radiatively- heated from the ionic emitter or resistively heated with a separate power
supply (direct heating). The analyte molecules and their degradation products are
thermally desorbed from the sample holder and form adducts in the gas phase with the

potassium ions emitted from the thermionic emitter.

68

The peptides H-GP-OI—I, H-GGFL-OH, cyclo-GP, H-YGGFDOH, and H-FGGF-
OH were obtained from the Sigma Chemical Co. , St. Louis, Missouri, and used without
further purification. H-AGGFL-OH, H-(B—A)GGFL—OH (B-A is B alanine), H-VGGFL-
OH, H-GGVFL-OH, H-VGVAPG-OH, H-VGVAadPG-OH, H-deodeAPG-OH,
andH-AFdeLA-OH were synthesized in the laboratory of Dr. Phil Andrews at the
University of Michigan. The three partially-deuterated peptides listed above, were
synthesized with the corresponding deuterated amino acids as the starting materials. N-
terminally acetylated peptides such as Ac-VGVAPG-OH, are formed from the free
peptide using acetic anhydride and a small amount of pyridine as a catalyst, by the
procedure described by Knapp [5].

All of the samples used in K+IDS experiments were dissolved or suspended in

methanol to allow transfer of 1-2 Ill of sample onto the tip of the sample holder. The

solvent was evaporated off before the probe was inserted into the mass spectrometer.

III. The Primary Thermal Degradation Processes of Peptides

The thermal degradation chemistry for peptides, involving skeletal bond cleavage
with an accompanying hydrogen atom shift, had been proposed [6—7] to follow a 1,2 -
elimination process, Figure 4.1. However, additional experimental results suggest an
alternative mechanisms should be proposed because it cannot explain many experimental
observations. A 1,2-elimination mechanism would suggest that the probabilities of
cleavage of any of the amide bonds would be similar, since the process involves only
amide bonds and adjacent H atoms. In our experiments, however, the rates for cleaving
the various amide bonds differ dramatically throughout a single peptide, as well as from
one peptide to another. Also, this mechanism cannot explain the formation of [C+H+K]'*'
ions, which are occasionally observed, since there is no hydrogen atom at the required

position (on the carbonyl group) to participate the reaction.

69

 

1,2 - Elimination Mechanism

R O R
l l 1
81szanch —>NH2-NHC=C=O + H-NHMCOOH
H (NHMCOOH

Peptide (Bm'H) (Yn +H)

 

 

 

Figure 4.1 A schematic diagram for the 1,2 - elimination mechanism showing

thermal degradation of peptides.

70

Because the 1,2-elimination mechanism is inconsistent with the experimental
results for peptides, alternative degradation mechanisms were considered. Since H-shifts
are frequently involved in the thermal degradation of peptides, as well as other thermally-
labile molecules [67], determining the source of the shifting H is key to determining the

mechanism. This can be done using deuterium labeling.

A. Deuterium-labeling Experiments

Two deuteration approaches for peptides have been successfully used. The first
method involves deuterium exchange in a deuterated solvent such as CH3OD and D20.
Hydrogen atoms that are bound to the N-terminal nitrogen, amide nitrogens, and the C-
terminal carboxylate group can be exchanged easily at room temperature within seconds.
To demonstrate the utility of HID exchange, consider the dipeptide H-GP—OH. When this
small peptide is subjected to rapid heating, desorption of the intact molecule dominates.
A fraction of the molecules dehydrate, and the (M-H20) molecules desorb as well. The
elimination of water from a peptide poses the same questions as the processes in which
skeletal bonds cleave. To eliminate H20, a C-OH bond cleaves with an accompanying
H-shift. What is the source of the shifting hydrogen atom? Water elimination is
analogous to the process that forms B- and Y-type products. With small dipeptides,
skeletal bond fragmentation is usually a minor process relative to water elimination, thus
they provide an opportunity for characterizing this particular process. Also, small
peptides contain a relatively small number of exchangeable H atoms, and it is easier to
approach 90+% H/D exchange.

If H-GP-OH is dissolved in D20, placed on the K+IDS probe, and the
experiment performed, the spectrum shown in Figure 4.2a is obtained. There are 3
exchangeable H atoms in this dipeptide. Complete exchange of these three H atoms

yields a K+-adduct of the intact peptide at m/z 214. Those molecules undergoing

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

71
a +
(M'I'K)
H -GP-OH d3\ /2r4
— Deuterated by I
E‘ Exchan e H 0 -D20
g; g ( 2 ) I ‘
a- I
3 d2
3;, \
‘2 194
I 195
_ 193\ /
rrwr VflT TT?! TT Vfijl‘fiILeA IITI '1 IILIA I1I I I11 ii! I UYTT
I I I I I I I II I I I
130 140 150 160 170 180 190 200 210 220 230 240
m/z
b 4320
W+K)+
H-FM-OH / D20 I d
w - I \
_._._ I d4 339
g. ‘
g I -DHO
g: - |
O
a, . 319
Q
—I
rT I I I I I I F1 I I I I I I j 1 fiAI “I I I 1‘ In; I T I WW I I‘T IAAIA
I I I I I I I
200 220 240 260 280 300 320 340 360 300
m/z

Figure 4.2 K'I’IDS mass spectra of: a) H-GP-OH that had been exposed to D20;

b) H-FM-OH that had been exposed to D20.

72

incomplete exchange are represented by the peaks at lower m/z values. The distribution
of the [M+K]+ peaks indicates that 90% of the exchangeable H's have been replaced by

deuterium atoms.

The dipeptide H-GP-OH has the elemental composition C7H1 2N 203. The
deuterated form that can be obtained via H/D exchange is C7H9D3N203. In the
spectrum of the deuterated dipeptide, the m/z value of the dominant peak representing the
dehydration product is 194, which corresponds to [C7H9D1N202+K]+- The relationship
of the isotopic variants for this species, relative to that for the [M+K]"’ ions is as follows.

If the loss of H20, in the form of DHO or H20, goes through a 1,2- elimination process,

the degradation will produce the following ions:

m/z 214 <—K+— [ND2-~COOD] 41le _K+_> m/z 195
m/2213 <—K+— [NDHMCOOD] 411912» J$+_> m/z 194
m/z 213 <—K+— [NHDMCOOD] :HQQ» —K-+—> 111/2194
m/z 213 <—Ki— [anmcoom —-HQH—> —K+—> m/z 195

Each expression inside the brackets indicates the two termini of the deuterated peptide
and the m/z value of the corresponding K+-adduct is on the left side of the expression.
The abundance of the K+-adduct of the each deuterated variant reflects its relative
abundance. When a water molecule is lost from the peptide molecule, the adduct formed
will have the m/z value shown on the right side. If exchange of each candidate H atom
occurs to approximately the same extent, and the water elimination occurs via a 1,2-
elimination (involving a H atom close to the C-terminus) the expected ratio of m/z 194 to
195, based on statistical analysis, would be 1: 5, which is NOT in agreement with the
experimental value of 4:1. The expected ratio of m/z 194 to 195 is calculated as follows:
The relative quantities of the completely exchanged peptide H-GP-OH(d3) and the
partially exchanged peptide H-GP-OI—l(d2) are 100 to 35 based on the intensity ratio of

73

their K+ adducts at m/z 214 and 213. Based on the above chart, the expected ratio of m/z
194 to 195 should be (35x1/3): (100+35x2/3) = 1:5. The experimental ratio of m/z 194
and 195 is based on the relative intensities of the peaks at these m/z values.

However, if the loss of H20/HDO/D20 is through a 1,6-elimination in which the
shifting H atom is from the N-terminus (allowing for a D-shift from the opposite end of

the molecule), the elimination would allow for the loss of D20), and would give the

following ions:

m/2214 <49— [ND2~~~COOD] 41291?» —I-<i—> m/z 194
111/2213 <—K—t- [NDHMCOOD] JiQD—> 34> m/z 194

 

m/2213 < K+ [NHD~-CO0D] 4290—) —1(-T—> m/zl93
m/2213 < K+ [anmcoon] 41291:!» —K—*:—> 111/2194

 

The expected ratio of m/z 193 to 194 will be (40 x 1/3) : (100 + 40 x 2/3) = 1: 9, which is
in very good agreement with the experimental data (1:6). Simply, the spectrum suggests
that most of the dipeptide molecules contain 3 deuterium atoms, and the water
elimination involves the loss of D20. This can only happen via a cyclic intermediate.
Therefore, the loss of H20 in the thermal degradation of this dipeptide occurs via a 1,6-
elimination process.

Analyses of another dipeptide H-FM-OH also give the same conclusion, as shown
in Figure 4.2b. For this peptide, a greater exchange rate at the C-terminus CO0H is
suggested based on the following: Because the completely exchanged peptide H-FM-OH
(d4) is more abundant than the partially deuterated peptide H-FM-0H(d3), the dominant
formation of ions at m/z 319 indicates a loss of D20 molecule, which suggests that the
dehydration process is a 1,3- or 1,6—elimination process. The weak ion intensity at m/z
318 thus suggests a dominant loss of DHO from H-FM-0H(d3) with one exchangeable

hydrogen undeuterated. Apparently if the undeuterated hydrogen is on the C-terminus

74

only, then the loss of DHO will be the only reaction for H-FM-0H(d3), which leads to no
ion current at m/z 318. This is almOSt exaCtly true, suggesting that the undeuterated
exchangeable hydrogen is at the C-terminus mostly. Therefore, the exchange rate at the
C—terminus must be the greatest among the four exchangeable sites of the peptide. (Note:

The peptide is assumed to be exchanged completely first. Inside the vacuum lock and the

ion source, back exchanges occur due to the presence of trace amount of H20 vapor.)

The HID exchange in D20 was also undertaken with larger peptides such as the
hexapeptide H-VGVAPG-OH , which has seven exchangeable H atoms. The resulting
K‘I’IDS spectrum is shown in Figure 4.3b. The H/D isotopic variants of the [M+K]+ ion
suggest extensive deuterium'incorporation, with more than 87% of the exchangeable sites
deuterated [For the molecular ions, there are three peaks representing the deuterated
peptide adducts containing seven, six and five deuterium atoms, respectively. The
relative intensities of these deuterated peptides are 90%, 100% and 70% respectively.
Therefore, the percentage deuteration = 100x(90x7+100x6+70x5)/(90x7+100x7+70x7),
in this case]. In this hexapeptide, the dominant ion is the [Y4+H+I(]+ species; resulting
from cleavage of the amide bond with a H shift onto the C-terminal fragment. An
analysis of the isotopic species shows that an exchangeable H shifts in the process, thus a
D-shift is observed in this case as well. It cannot be unambiguously determined whether
the source of this shifting D is the N -terminal amine group, or an amide nitrogen in the N-
terminal fragment. However, it is important to note that the bond cleavage/H shift is not
a 1,2-elimination reaction, so H's that are 3 or more skeletal atoms away must be involved
in the chemistry.

The second way to prepare partially-deuterated peptides is to use partially-
deuterated amino acids in the peptide synthesis. In this way, deuterium atoms can be
placed on carbon atoms in the peptide skeleton as well as in side-chains. The following
partially-deuterated peptides were synthesized for this study: H-VGVAadPG-OH, H-
VGVAB3dPG-0H, H-vadovadAPG-OH, H-VBdGVBdAPG-OH and H-AFBdZLA-OH,

75

 

 

a
H-VGVAPG-OH

a:
E .-
E
o“
>
'13
.25 (Am-K)“
32 207

‘ (B2-H+I()"’ (Y2-1»H+l()+ (Y3+H+K)+

I 195 211 282

/ /

 

 

U4+H+K)+
381

(C3+H+K)"'
31 l

438
\

 

(Y5+H+K)+ (C5+H+K)+

 

(M+K)‘r
537

(114-thong+
519

\

479
/

 

 
  

 

 

 

 

 

 

140 180 220 260 300 340 380 420 460 500 540 580
m/z
b + d6\ /d7
H-VGVAPG-OH “4“”‘0 \ (WW
Deuterated by 385 544
Exchan e (D 0)
£2 g 2 d5\
E
,5-
g I
'6
.2 7 1
0 .
a: I
- (ARK)+ (Y 2+D+K)“ (Y 3+D+K)’r (Y5+D+K)+ i
208 214 286 444 I?
160 200 240 280 320 360ml 400 440 480 520 560 600
2
Figure 4.3 K'I'IDS mass spectra of: a) H-VGVAPG-OH; b) H-VGVAPG-OH exposed

to D20 (hydrogens bonded to N and 0 atoms are exchanged to deuteriums).

76

 

 

 

 

 

 

 
  

 

 

 

a
* H-VGVAadPG-OH fig)"
a: Add = -NHCD(Me)C0-
IS
o—I
>
‘33
2
0‘2 - . (Y4+H+K)+
- (Y2+H+K)+ (Y3+H+K)+ (c3+H+1<)+ 332 (Y5+H+K)+
211 283 311 439
‘I I I‘ \I I II II\I%I“ IlI‘rI WI‘I I I I I I ‘I I4 I‘I I IMIhIw I‘I I I
I I I I I I I I I I
”0 180 220 260 300 340 m[2380 420 460 500 540 580
b (11'4+H+K)+ (M+x)+
H-VOIdGVOIdAPG-OH /382 539
I: Vad = -NHCD(i-C3H7)C0-
3. I
E.
3
'21
Eu
0
a: (Ill-H204“)+
(AP+K)+ (Y2+H+K)+ a3+n+x)+ a5+H+K)+ 52‘
207 211 232
\ / /

 

150

 

190 230

 

270 310 350

  

630

590

390 430 470 510 550
m/z

Figure 4.4 K‘I'IDS mass spectra of: a) H-VGVAadPG-OH; b) H-vadovadAPG-OH.

77

with a and B designating the deuteron position(s). In these peptides, some deuteriums are
present on the backbone of the peptide chain such as - CDCH3 - (in Aad), and others on
the sidechains of the molecules such as -CD3 (in AB3d ) and -CD(CH3)2 (in the VI3d
residue).

The K+1Ds spectra of unlabeled H-VGVAPG-OH, as well as H-VGVAadPG-
0H, and H-VadGVadAPG-OH are shown in Figures 4.3a, Figure 4.4, respectively. The
spectra show that the deuterium incorporation by synthetic methods is complete, and no
reverse-exchange occurs. The results consistently show that, when Y-type ions are
formed, H shifts only occur from nitrogen sites - amide groups and/or the N-terminus.
The source of the shifting H atoms is not those H's that are bound to carbon atoms in the
amino acid residues.

When these experiments are performed, the bond energy of the OH bond cleaved
is also considered. Since the C-BH bond in phenylalanine (R = -CH2C6H5) is one of the
lowest in peptides, the B—deuterated peptide H-AFZBdLA-OH, was synthesized and the
H-shift in this peptide was studied. Consistently, the same result was obtained - no H
shifts from the side-chains occur.

All the experiments with deuterated peptides lead to the same conclusions. Such
deuteration experiments alone, however, do not allow for definite determinations of the
sources of the shifting H that accompanies skeletal bond cleavage to form the dominant
degradation product, (Yn+H). To complete the investigation, other variants of the

peptides were characterized.

B. Peptide Derivatization - Evidence for Cyclization
As demonstrated above, a shifting hydrogen atom is not from a-carbons; it can
only be from either amides or the N-terminus of a peptide. If the N-terminus of a peptide

is blocked, then the only choice will be amides.

78

Insights were gained in the thermal degradation chemistry of small peptides via
the N-acyl derivatives. When a peptide is N -acetylated, and there is no longer an amino
group at the N-terminus, changes in the K+IDS mass spectra are observed. For the
underivatized hexapeptide H-VGVAPG-OH, as shown in Figure 4.3a, the dominant ions
observed in the K'I’IDS spectrum are the [Yn +H +K]+ ions, as has been previously
discussed. When the N-terminus of H-VGVAPG-OH is blocked with an acetyl group
forming CH3C0—VGVAPG-0H, the dominant ions observed are no longer [Yn+H+K]+
ions. Instead, an unexpected series of ions of the type [Bn+0H+K]+ and [Cn+H+K]+
ions become dominant (see Figure 4.5). These observations suggest a scenario which is
summarized in Figure 4.6. First, since there is such a dramatic change in the chemistry
upon acylation when the N-terminal amino group is removed, then the thermal
degradation chemistry for underivatized peptides that yields molecules of the type
(Yn+H) must involve a H-shift from the N-terminus. When this group is removed, no Y—
type products are formed. Second, if the Y-type products are formed via a H-shift from
the N-terminus, then the secondary structure of the decomposing peptide is critical, and
must involve a cyclic intermediate. When the N-terminal amine group that can
participate in H-bonding is removed, products that appear to involve cyclic intermediates
through H-bonding to the C-terminus are observed. It is readily apparent that :QkLshifls
accompanying the cleavage of skeletal bonds many atoms away occur through cyclic
intermediates when the N-terminus is blocked Thus, the peptides appear to react through
cyclic conformations. The H-bonding involving the N-terminus usually dominates.
However, if the N-terminus is blocked, the C-terminus participates in important
secondary interactions. Third, if peptides undergo skeletal bond cleavages and H-shifts
(or 0H shifts) through cyclic intermediates, then one linear product and one cyclic
product should be produced in each case. Note that the dominant (Yn+H) product is
formed by elimination of a cyclic, six-membered ring dipeptide. Thus, the dominant

Cyclic intermediate involves a H-bond via an 8-membered ring. Also, when the N-

79

 

a

1 l ryl

Relative Intensit

 

H-VGVAPG-OH

(Y4+H+K)+
38 l

(MHQ+

(ARK)+
207

211

(C3+H+K)"‘
311

(M-H20+K)+
5 l9

   

180 220 260

(Y 2+H-I-K)+ (Y3+H+K)+
282
\ / /

300 340

 

 

3 80 420

m/z

460

\

(Y 54-11440+ (C5‘1-131+1()+

438 479
\

/

  
  

537

 
  

 

 

580

540

500

 

yJ

l

1

Relative Intensit

 

130

Figure 4.5

Ac -VGVAPG-0H

I
(C3+H+I()+ (B3+0H+K)"'|
354

353
(Bz-I-OH-I-K)+
255

I70 210 250 290 330

370

(C4+H+K)+
424

I425
I

410 450

(35+OH+K)+

522

(B4+OH+K)+

490

 

$30

WNW"

/$79

(M-H20+K)+

561
/

 

 

 

we

570 610 650

K‘I'IDS mass spectra of: a) H-VGVAPG-OH; b) Ac-VGVAPG-OH.

 

R-VGVAPG-OH
(P
Q» ‘23,
//
Q- 300
H 0 MezCH 0 0 A
\ |5+ 5" I I $H3 I5+ 8"
H —C—NH-CH'C-APG-0H cmco-vcv -NH-CH- _?—CII
HN lI'M—Hi“ H .: 5+ C=O
\ O \ /
o—c41 Hc— NH
I \ I
v
H\ IO
HC—o\ M92?!
CH 0 N —CH
Hg NH+ HzN-CH-C-APG-0H ' 3 I I. \
\ CH3CO -VGV-NH-CH —c-0H + o =(~\ C=o
M” I.
H —
O ICHMez Y
H
Cyclo.(VG) H-VAPG-OH CH3C0VGVA-0H Cyclo-(PG)
(82-11) mm) (B4+0H) (Yz-OH)

Figure 4.6 Thermal Degradation of Peptides: (H-VGVAPG-OH and Ac-VGVAPG-OH).

81

terminus is blocked, and cyclic intermediates involving the C-terminus form, the
dominant (Bn+0H) product is formed by, again, elimination of a cyclic dipeptide. Thus,
cyclic interactions that involve 6 and 8-membered rings simultaneously, as shown in

Figure 4.6, form the dominant products.

Similar results from CBZ-derivatives (CBZ = carbobenzoxy- = C6H5-CH20C0-)
such as CBZ-GPGGPA-OH, and t-BOC derivatives (teBOC = tert-butoxycarbonyl =
Me3COC0-)'such as t-BOC-PPPP-OH, were obtained (see Figure 3.11). All the results
indicate that N—terminal derivatives blocks cyclizations from the N-termini and new ion
series, (Bn+0H+K)+ and (Cn+H+K)+, characteristic of the peptide derivatives, are
produced.

To further investigate the possibility of "remote" H-shifts via cyclic
intermediates, the synthesis of peptides with other types of N-terminal modifications, for
subsequent analysis by K+IDS, was pursued. Two forms of the amino acid alanine are
available. In addition to the standard a-amino acid, NH2-CH(CH3)CO0H, B-alanine is
available, NH2-CH2-CH2-CO0H. The presence of B-alanine in a peptide will disrupt the
normal sequence of the skeletal backbone (-NH-CHR-C0-). To investigate the influence
of modifications of the N -terminus on thermal degradation, two peptides were
synthesized, H-AGGFL-OH and H-(B-A)GGFL-0H, and their thermal degradation
chernistries were compared. The resulting K+IDS spectra are shown in Figure 4.7. For
H-AGGFL-OH (Figure 4.7a), formation of (Yn+H) products dominate, with the
[Y3+H+K]"’ ion indicating that elimination of a cyclic dipeptide from the N-terminus is
the dominating thermal degradation pathway. When the N-terminal residue is changed to
B-alanine (Figure 4.7b), the spectrum changes substantially. Amide bond cleavages with
accompanying H-shifts still occur, although they no longer dominate the desorbed

Products. Note that, if amide bond cleavages still involve Heshifts from the N—terminus,
then the formation of (Yn-I-H) products would involve the formation of 4-, 7-, 10-

membered cyclic products. Apparently, none of these are as favored as the process that

82

 

 

 

 

  

 

 

 

Figure 4.7

 

 

 

 

a
. H -AGGFL-0H (Y3+H+K)+
0 374
k.
=5 HzN
g. GFL- OH
E: X 4
o-
.2
E
52’ .
11 K +
m (Y2+H+K)+ «4+ + ) (M+x)+
(Y1+H+K)+ 317 431 502
130 170 210 250 290 330 370 410 450 490 530 570
m/z
«b (Mi-10'"
H-(B-A)GGFL-OH (M-NH3+K)"' 502
" 485
% HZN GFL- 0H
E (Y3+H+K)+
5 374
E -‘ (Y4+H+K)+
g 431
m“ (Y2+H+K)"'
17 467
“I‘m” 2993 339
170
130 170 210 25, 2% mm, 370 410 45) 4% 530 570
z

K+IDS mass spectra of: a) H-AGGFL-OH; b) H-(B—A)GGFL-0H.

83

forms a 6-membered cyclic dipeptide, which occurs for H-AGGFL-OH only. Also now
that the ratio of thermal degradation to desorption of the intact molecule upon heating is
smaller when B-A is at the N-terminus, leading to the [M+K]+ ion as the base peak in
Figure 4.7b. These changes in the spectrum when the N-terminal residue is modified in
this way is consistent with the proposed importance of remote, H-shifts from the N-
terminus, accompanying skeletal bond cleavages throughout the peptide, and the

importance of the formation of a cyclic product.

C. Termini, Sidechains and Steric Effects

It is clear from the above data that the primary, low-energy, thermal degradation
chemistry of peptides involves cyclic intermediates and cyclic products. When the N-
terminus of a peptide is not blocked, cyclization involves the shift of a H atom from the
N-terminus; when the N-terminus is blocked, cyclization involves the C-terminal OH
group. The N-bound H atoms at the termini of a peptide are, in part, most active in their
participation in cyclic interactions because fewer steric obstacles are present, relative to
those encountered by interior amide N-H groups.

If steric effects are important in the thermal degradation of peptides, the sizes of
the sidechain of the amino acid residue at the termini may influence the distribution of
thermal degradation products. Insights into one aspect of steric effects can be realized
through the K+IDS spectra of the following series of peptides with various N-terminal
residues: H-YGGFL-OH, H-VGGFL-OH, and H-AGGFL-OH. These are shown in
Figure 4.8a-b and Figure 4.7a, respectively. When the N-terminal residue is Y, V or A,
the spectra are essentially the same. The mechanism shown in Figure 4.6 for the free
peptide dominates. Note in particular the ratio of the intensities of the ions representing

[Y3+H+K]+ and [Y4+H+K]+, which are formed by eliminating neutrals containing 6-
and 3-membered rings, respectively. Obviously, the former dominates. However, when

the sidechain of the N-terminal residue is smaller, changes are observed. Figure 4%

 

 

 

 

 

 

a
.. (Y3+H+r<)+
H -VGGFL-0H 374
2."
7;. I
53 X5 x1
5 _ I (M+K)+
D
a? 530
.9. -1
D
a:
i + (Y2+H+KY‘ +
. “32““) 317 I (Y4+H+K) 512
195 I 431 I

 

 

 

 

 

 

 

 

150 190 230 270 310 350nV2390 430 470 510 550 590
b
I H-YGGFL-OH (Y2+H+K)+
a 374
d _ +
(32 2:310 (M-HZOK)+
(Y2+H+K)+ (M-Hzo-NH3+K)"'
i ’ 317
(Y2+H+K)'+
- (Y2+H+1<)+ 431 (WK)+
170 335 594
IIII‘TTIT ‘IIIT IIrI JrIIIMIIT‘I III+A7AIIIA+TAIATAAII!IA7 I—IILI‘ITIII IIIT
I I I I I I I I I I I I I
130 ‘70 210 250 290 330 370 410 450 490 530 570 610 650

Figure 4.8

m/z

K+IDS mass spectra of: a) H-VGGFL-OH; b) H-YGGFL-OH.

85

 

 

 

 

 

 

a
I (Y3+H+I()+
H -VGGFL-0H 374
2?
.5 q
i5 x 5 x1
f _, I— (MI-10+
.5 530
:3.
D
a:
— + (Y2+H+K)I' +
. 0324“") 317 I mm“) 512
195 I 431 I I
510 550 590

150 190 230 270 310 350“V 390 430 470
Z

 

b (Mi’K)+
H-GGVFL-OH 5,,
E"
g .
E -
0
g (Y 3+H+K)"’ (Y 4+H+K)+
52 416 473

l

 

 

 

130 170 2l0 250 290 330 370 410 450 490 530 570 610
m/z

Figure4.9 K‘I'IDS mass spectra of: a) H-VGGFL-OH; b) HoGGVFL-OH.

86

shows the K‘I'IDS spectrum of another pentapeptide, H-GGVFL—OH, in which the N-
terrninal glycine contains the smallest sidechain (least steric hindrance). The spectrum is
very different. The relative probabilities of cleavages at various amide bonds change
completely. In this case, the cleavage at the first amide bond is almost competitive with
that at the 2nd amide bond. Apparently, the size of the sidechain in the N-terminal
residue does have an influence on the distribution of the various cyclic intermediates,
with the smallest cyclic intermediates being uncommon, unless the N-terminal residue is
glycine. The same is observed in the spectrum of the tetrapeptide, H-GGFL-OH, where
the N-terminal residue is again 0. More peptide spectra will be useful to determine
whether this effect is always observed for the N-terminal residues, and the precise

secondary interactions that are involved in this "steric effect".

D. Thermodynamic aspects of the thermal degradation of small peptides

To. better understand the mechanisms for the thermal degradation of peptides,’
thermodynamic aspects of the chemistry were considered. The experimental data
presented here suggest that skeletal bond cleavages do not involve 1,2-elimination
processes, but 1,n-elimination processes, where n = 3, 6, 9, etc. Is one energetically
favored over the other?

In all of the thermal degradation processes of peptide molecules that we have
observed, two bonds are formed and two bonds are broken. The estimated reaction
enthalpies for several possible processes are shown in Figure 4.10. To approximate the
reaction enthalpies for some of the degradation processes shown, the dissociation
energies of the bonds formed ( BDE(Bf1) and BDE(Bf2) ) and bonds broken ( BDE(Bb1)
and BDE(Bb2) ) are considered, and the reaction enthalpies are estimated as follows:

AH°nm = BDE(Bb1) + BDE(Bb2) - BDE(Bf1) - BDE(Bf2)

The bond dissociation energies (BDE's) that were used are shown below:

87

2%... .ész
a. a.
LN! ($7 . ®
.1... .4...

E
a- .8
a 3..
2.3. . .. O.
2.2.5....

 

 

“was...
{31.85%
Wm. s,
08%. I

4W n M.

 

 

fir s
ZWWIMU

 

 

. 3.5.8.. 5 81.2.5 5.88.. 3:56 .

8232. .3. Q6359“ 8.5358 .9505. 36 3:9“.

W ® @
\II “It! o olot
.a "8359a.
@Lw
251.21%. . 5H7)
M @ @
2.211.“. I .3?
Lie. a. I.
"coo-55%

88

0
II
NHMCR7—8C 9—4N §JCRI-I --CO0H

9H 93l 10H
H H H

The bond dissociation energies for the three skeletal bonds of peptides were obtained
from small molecules (see Appendix II for details), for which thermodynamic data are
available[8]. For example, BDE(Y-X bond) is actually that for the molecule CH3CO—
NHCH3. Values used for BDE (H-X), where X = NHR, NR2, CHRz, etc. were also
extracted from available data on small molecules. The BDE for a N-H bond in the
terminal amine group is that for a small primary amine.

Reaction enthalpies for 1,2-elimination reactions that are presented again
represent those for smaller analogous systems, where sufficient thermochemical
information is available. For example, the estimated reaction enthalpy for the highest
energy 1,2-elimination, to form {(B+H) and (Y-H)}, represents the enthalpy for the
process:

CH3C0NHCH3 ------ > CH3C0H + NH=CH2

Figure 4.10 first points out that homolytic cleavage, involving the formation of
two radicals, should not be expected since such reactions require much more energy than
reactions in which the number of bonds broken is equal to the number of bonds formed,
such as a 1,2-elimination. Products of simple homolytic cleavages are not observed under
K‘I'IDS conditions. Lower energy 1, 2 - elimination reactions (requiring about 40
kcal/mol) also do not occur, as suggested by deuterium-labeling experiments. The lowest
energy processes are those suggested on the right side of Figure 4.10, in which a cyclic
and a linear product is formed in each case. Like 1,2-elimination reactions, these 1,6-
eliminations involve the cleavage of two bonds and the formation of two bonds.

However, the latter have approximate reaction enthalpies that are less than or equal to 0

89

kcal/mol. The differences in the two processes are fairly clear. In the 1,2-elimination,
one of the bonds "formed" involves the conversion of a single bond to a double bond;
double bonds have BDE‘s less than the sum of two single bond BDE's. Also, in a 1,2-
elimination, the environment from which the H moves is different from that to which it
moves. In contrast, consider the 1,6-elimination process shown in Figure 4.10 that yields
(cyclo-C-H) and (Z+H). 0ne C-N bond is broken while one is formed. Also a CH bond
is formed as a CH bond is broken. Thus, of the mechanisms considered, the 1,6-
elimination reactions represent the lowest energy pathways.

If the conversion of a peptide into two smaller peptides, one linear and one cyclic,
is thermoneutral or exothermic, why are all not observed? Why does cleavage of one
skeletal bond dominate? We suggest that the answer lies in a consideration of the
intermediates through which the 1,6-elimination products are formed. These are also
shown in Figure 4.10, and are arranged to suggest their relative Stabilities. In Figure 4.10,
intermediate IV is expected to be readily accessible. Dipole-dipole interactions will allow
the molecule to align as shown, as a prelude to thermal degradation. In contrast,
intermediate I would allow (cyclo-C—H) and (Z+H) species to be formed, however it is an
unfavored, high-energy transition State. Its formation requires unlikely secondary
interactions according to considerations of the partial charges on the atoms of the peptide
backbone. Thus, the B- and Y-type thermal degradation products are most likely formed
as shown in Figure 4.10 - they evolve from the most stable secondary structure, and the
overall reaction enthalpy is lowest. (In this case, the estimated enthalpy range reflects
differences between N-H bond dissociation energies for H's of either the N-terminal
amino group or from an interior amide group.) The relative Stabilities of the transition

states determine, as usual, the reaction rates.

90

IV. CONCLUSION

K+IDS is shown here to be a useful technique for the mass spectrometric
characterization of the thermal degradation chemistry of small peptides. In this chapter,
one of the most important questions relevant to the technique have been answered, which
allow spectra to be related to underlying thermal degradation processes.

The dominant skeletal bond cleavages in the thermal degradation of peptides
occur most frequently at amide bonds. Cyclo-degradation was found to be the major
process for thermal degradation of peptides. The formation of a cyclic product is favored
because it is the lowest energy process; the dominant cyclization process is the one with
the most stable transition state. For underivatized peptides, cyclization involves the N -
terminus, the shifting H that accompanies skeletal bond cleavage is from the N-terminus.
For N-terminally blocked peptides, cyclization from the N-terminus is hindered, and
cyclization involving the C-terminus can occur. In this case, the shifting H or OH is from
the C—temrinus. I

In this Chapter, we have reached a good understanding on thermal degradation of
peptides. However, the much weaker (Bn-H+l()+ signals than those of (Yn+H+K)+ in
K+IDS spectra of peptides are Still to be understood. In the following Chapter, the
mechanisms for thermal degradation of peptides will be used for this purpose. The cyclic

nature of (Bn-H) will be used in explaining why (Bn-H+K)+ ions are not observed very

often.

91

References

1. Irwin, W. J. Analytical Pyrolysis: A Comprehensive Guide; M. Dekker: New York,
1982; pp 3-44.

2. Irwin, W. J.; Slack, J. A. Analyst 1978, 103, 673-704.

3. Schmelzeisen-Redeker, G.; Giessmann, U.; Rollgen, F. W. Org. Mass Spectrom.
1985, 20, 305-309.

4. Light, K.; Kassel, D. 8.; Allison, J. Biomed. Environ. Mass Spectrom. 1989, 18, 177-184.

5. Knapp, D. R. "Chemical Derivatization for Mass Spectrometry" in Methods in
Enzymology, vol. 193, McCloskey, J. A., ed., Academic Press: San Diego, 1990;
pp 314-328.

6. Bombick, D.; Allison, J. Anal. Chem. 1987, 59, 458-466.

7. Bombick D.; Pinkston, J. D.; Allison J. Anal. Chem. 1984, 56, 396-402.

8. Lias, S. G. ; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.;
Marlard, W. G. J. Phys. Chem. Ref. Data, 1988, 17, Suppl. No. 1.

Chapter 5

Response Discrimination In K+-Adduct Formation

I. Introduction

In K+IDS mass spectrometry, when a peptide molecule undergoes thermal
degradation, two neutral species such as (Yn-t-H) and (Bn-H) are formed and desorbed
into the gas phase. Due to the stoichiometry of the reaction, the quantity of these two
species (or the number density ) are the same inside the ion source. These two gas-phase
neutral species then react with K+ ions generated from the K‘I' thermionic emitter,
forming K+-adducts.

As have been seen in K+IDS mass spectra, however, the peaks representing the
[Bm-H+K]+ ions are much less intense than those for the corresponding [Yn+H+K]+
ions. For example, the thermal degradation of the pentapeptide H-YGGFL-OH will
produce (Y3 + H) and (Bz - H) in a 1:1 ratio. In the spectrum of H-YGGFL-OH,
however, the abundance of [Y3+H+K]+ is 30 times as great as that of [B2-H+K]+.
Why? What are the reasons for this observation? Is that because the (Bm-H) molecules
are unstable? Is that because the volatility of the (Bm-H) molecules is not high enough?
Or is that because of the poor stability of the (Bm-H+K)+ ions?

For ion formation in K+IDS, three processes are involved - degradation,
desorption, and "sampling" (adduct formation with K+ ions). We have investigated the
latter as the most likely source, i.e., that some discrimination exists in the adduct
formation process. The nature of interactions between 10" ions and neutral molecules is

electrostatic, and should be sensitive to molecular size and structure, local dipole

92

93

moments within the molecule, and polarizability. The binding energies are relatively
small, 20-40 kcal/mol [1-2], and adduct ions are presumably stabilized by emission of an
infrared photon, or dissociate after some time period (K+ detachment).

In this Chapter, response discrimination in K+IDS is investigated. The size and
structure dependence of response is studied with peptides of different sizes and structures.
The nature of discrimination will be studied by using mixed thermionic emitters. Results

from this investigation are used to explain the general features of K‘I'IDS spectra.
II. Experimental section

Preparation of mixed ionic emitters: A mixed thermionic emitter produces two or
more different ions such as K‘I' and Na'I‘ ions at the same time from the same emitter. To
make these emitters, a regular K+ ion emitter was made and conditioned first. For a K"
and Na+ ionic emitter, a small drop of Na2C03 aqueous solution was added on the tip,
air dried then put into the ion source for conditioning. The bead/emitter was conditioned
by applying current gradually to 3.0 A, as is for a regular K+ emitter. The emission of
K'I' and Na+ ions was examined to assure that the emission currents for K‘I' and Na+ are
comparable. For a Cs+ and K+ emitter, a similar procedure was used. Instead of using
Na2C03, C520 was dissolved in water and a drop of the solution (CsOH) was applied to
the thermionic emitter. This new bead was conditioned and emission of K+ and Cs+ ions
were examined.

All peptides and amino acids were purchased from Sigma Chemical Co., St.
Louis, M0., and were used without further purification. The samples were dissolved in
water to concentrations of approximately 1p g/ul. For peptide and amino acid mixtures,
all components have roughly the same concentration of l ug/ttl. One to two-I11 aliquots

of solution were transferred to the K+IDS sample-filament.

94

III. Binding Energies and Stabilities of K+-adducts

The interactions between K‘I' ions and neutral molecules has been studied using
high pressure mass spectrometry by Kebarle and his coworkers [1]. In their early study
[1-2] on gas-phase solvation of alkali ions under equilibrium-conditions, binding energies
of K+ ions with various organic and inorganic solvents were determined experimentally.
Some of these data are listed in Table 5.1.

Table 5.1 Binding energies* and dipole moments of various molecules [1-2]

M W -A§°(k§al[mgl) W
Me280 35 25 3.9
DMA 31 24 3.81
DMF 31 23 3.9
Me2C0 26 19 2.88
n-PrNI-Iz 22 14 1.18
Pyridine 21 15 2.2
Me2NH 20 13 1.0
MeNHz 19 13 1.29
Benzene 19 12

NH3 18 12 1.47
H20 17 1 l 1.8

 

* For M-K‘I' ions, binding energy = 7AH°rxn (M + K" ----> MK").

The data in Table 5.1, determined by using high-pressure mass spectrometry, have shown
the following trends:
1. The greater the number of heteroatoms in a molecule, the greater the binding

energy (for molecules with the same size). This clearly indicates that the K+ in an K+

95

adduct ion is not bound to one specific atom. It must interact with many atoms within the
neutral molecule.

2. The larger the size of a molecule, the greater the binding energy.

3. For a small molecule with one heteroatom, its K+ ion affinity is about 10% of
its proton affinity - a very weak binding. Therefore, if the excess energy of the adduct is
large enough to cleave a bond within the adduct, K+ detachment will most probably occur
due to the weak binding.

Classical electrostatic calculations were performed by Clementi and his co-
workers [3-5] for alkali metal adducts; the results were in good agreement with
experimental measurements. Their calculations show that the bonding, particularly for
the larger ions like 10*, is almost purely electrostatic. For instance, the Mulliken
population analysis [3-5] showed that a charge transfer of only 0.018, 0.013, and 0.004
electron from the water molecule to the Li+, Na+, and K‘I' ions, respectively.

From these thermodynamic data, the formation constant for a K+-adduct of a
compound with A0' = -20 kcal/mol, can be estimated to be around 1050 (log K =
~AG'IRT) at 300K. In condensed phases, this means that the K+-adducts formed will
practically stay undissociated once formed, the reverse reaction rate is negligible
compared with the forward reaction rate.

kf
M + K+ <====> MK+
kr
In the gas phase of a high vacuum system, as is the case in a K+1D8 ion source, however,
a large formation constant does not necessarily means great stability because:

1. Formation constants are determined under equilibrium conditions. In a K'I'IDS
ion source of a mass spectrometer, an equilibrium can never be achieved because high
vacuum under K+IDS conditions leads to a very low number of collisions while a large

number of collisions is required for equilibrium.

96

2. A small number of collisions leads to a low formation rate of K+-adducts,
which leads to an even lower collision frequency with other species present in the ion
source. In the formation of K+-adducts, energy is released, which is equal to the binding
energy. Since there is no way for a newly formed K+-adduct to dissipate the energy
released during their formation (through collisions or photon emission), the energy
released may well be used to reverse the process.

3. Since the K+ binding is the weakest "bond" in a K+-adduct -- 20% of a regular
covalent bond, dissociation of K+-adducts will occur due to energy released during K‘I'
binding, and to lack of mechanisms for energy dissipation. Therefore, there is a lifetime
for any K+-adduct formed. If the lifetime of a K+-adduct is shorter than its residence
time inside the mass spectrometer (flight time from ion source to the exit of the
quadrupoles - approximately 50 us in the quadrupole mass spectrometer), the adduct ion
will dissociate and will not be able to reach the detector. (In a solvent system, the
situation is totally different. The energy released in the complexation process can be
taken away very quickly by the large quantity of solvent molecules, and a newly formed
molecule is thus stabilized). Therefore, K’r adducts with long life-times will show greater
response those with shorter life-times.

So the questions are: In K+IDS experiments, what are the most important factors

for a high yield of K+-adducts? What determines the lifetime of K"’-adducts?

IV. Discrimination in the formation of K+-adducts

Since binding energies vary with the type and size of molecules, different thermal
degradation products will have different binding energies. The difference in binding
energy may lead to differences in K+IDS responses. Based on the above considerations,
various compounds were studied for their K+-adduct formation and their relative

responses in K+IDS experiments.

97

For a quadrupole mass spectrometer, ions with higher m/z values usually have
lower transmission efficiencies. This is not considered in this study because the mass
differences of the selected compounds are relatively small.

In this study, the following two important factors were considered: A) The size

of neutral molecules; B) The structures of neutral molecules - linear or cyclic.

A. Size Discrimination

If small compounds such as amino acids are analyzed by K‘I'IDS, desorption can
be made to dominate. Thus, equimolar amounts of two amino acids (A and B) will yield
equal numbers of desorbed molecules for subsequent attachment to available K+ ions. If
the peaks representing [A+K]+ and [B+K]+ are of differing intensities, this would
indicate a response discrimination. The experiments in this section were designed to
determine the extent of discrimination, if any.

Amino acid mixtures are used in this study because they are the basic constituents
of peptides and because they can be the products from thermal degradation of peptides.
They share the same chemical formula NH2CHB_CO0H, with 3, different for different
amino acids. They do not show much chemical reactivity in K+IDS conditions, and the
typical reaction is the direct attachment of a K+ ion.

In the determination of response discrimination, proper quantitation is required.
To do so, relative concentrations of different compounds should be known. For this
purpose, a good solvent should be used and the best solvent for this study is water, in
which many amino acids show good solubilities. Some amino acids were not studied in
this work because of their poor solubilities in water.

Mixtures with two to five different amino-acids were used in this study and the
relative responses were found to be very similar (Figures 5.1 - 5.3). For example, with a

mixture of alanine and phenylalanine, the relative responses of phenylalanine to alanine

98

 

 

 

 

 

 

 

 

 

 

 

 

 

3 Ion counts (full scale) Ph m/z (MI?)

E roost.- 1‘79. e 204

1 100%- 930. 170

_-j Leu

{100%.- 407 153

: Thr

11001— 189 156

: Val

'2 100%- 117 123

_: Ala

‘I A
"'I""'Y"'*'T""1""U'fi'I""I""l’"ff'fjv"v"'1'

1100;- 4452

o 10 20 30 40 so 60 70 so 90 100 110

scan

Figure 5.1 Relative responses of five amino acids (a mixture of alanine, valine,
leucine, threonine and phenylalanine) in K+IDS.

 

 

 

 

 

 

 

 

: 100%: 38784.
j K+ adduct of phenylalanine
.: m/z 204

T I T I ' I V ' I ' fa I ' ' I
.I
_ 10 o 9.. 3 5 oo .
_‘: K+ adduct of alanine
j m/z 204

I I I A ' Ff A'T ‘47 TfiA
: 10 O 7.: 51 2 32 .
. TIC
L
-I
f I a 1 f I I f Y I
o 10 20 so 40 so so
Scan

Figure 5.2 Relative responses of alanine and phenylalanine in K+IDS.

 

 

 

100

 

 

 

 

 

 

 

 

 

 

 

: 1007a: 1495.
_: K+ adduct of alanine
L m/z 128
.4
I F ' 'fi * I ' ' '41 * r ' WT ' ' ' * I ' iff ' ' ffi ‘T_"“""‘
_ 100%. 394.
: K‘I' adduct of glycine
; m/z 114
I ' I ' ' I ' I 'AIA I r r T '

- 100%- 2654. , TIC
q

f 1 V f r 1 f I f I f I I
o 10 20 so 40 so so 70

Sean

Figure 5.3 Relative responses of alanine and glycine in 101138.

101

were found to be 32 to 1 (Figure 5.2). With a mixture of five amino acids - alanine,
threonine, valine, phenylalanine and leucine, the relative responses of phenylalanine to
alanine were calculated to be 30 to 1 (Figure 5.1). This clearly indicates that there is no
competitive processes occurring presumably because there are not many collisions in the
gas-phase, which is consistent with our estimated low pressure shown in Appendix 1.
Note that equal mass of each amino acid was used for the mixtures. Therefore, the
molecular mass of each amino acid should be considered when relative responses are
calculated.

As shown in Figure 5.1, different amino acids have shown different responses to
K‘I' attachment. The smaller ones, such as glycine and alanine, have much lower response
than the large ones, such as phenylalanine. For example, with a mixture of five amino
acids - alanine, phenylalanine, leucine, threonine and valine, their relative responses are
calculated to be alanine(Ala): phenylalaninde(Phe): leucine(Leu): threonine(Thr):
valine(Val) = 1: 32: 18: 8: 7, as shown in Figure 5.1. Glycine(Gly) has the lowest
response to K‘I' adduct formation among all amino acids: Ala: Gly = 6: 1. This is shown
in Figure 5.3. The relative responses of some common amino acids are summarized in

Table 5.2. In the table, the response for alanine is given a value of 1.

Table 5.2. Relative response factors, Rm], for amino acids and small peptides

amino acids
Ala Gly Val Thr Ser Met Leu Phe lle
Rrel 1 1/6 7 8 2 8 18 32 22

small peptides
Alaline H—A-A-OH H-A-A-A-OH H-A-A-A-A-OH
Rrel 1 10 100 200

102

 

 

 

 

 

 

 

 

 

 

E 100%=19264

2‘ m/2270

E 100%=3756

‘J Hill 199

E 100%:439

: Ill/2128

: Y T TL TA

3 1009 =30176

J 0 N/K TIC

o TIT .30 {I ' I30 I'as I'Io III 236 2'25.
scan

Figure 5.4

Responses of alanine, A-A and A-A—A at m/z 128, 199 & 27o.

 

 

 

103

Generally the greater the size and polarizability, the greater the response - the
trend is the same as binding energy trend observed by Kebarle and his coworkers [1-2].
For amino acids with the same size such as valine and threonine, the presence of
heteroatoms can enhance the response due to the increased polarity. The responses of
threonine and valine indicate that threonine is about 15% more sensitive than that of
valine. Dipeptides usually have greater responses than amino acids, tripeptides usually
have greater responses than dipeptides. As the size continues to increase, however, the
response differences gradually level off. The results for a series of small peptides
(Figure 5.4) are also listed in Table 5.2.

The observation of size discrimination can be attributed to the following factors:
1. The thermodynamic stability. The larger a neutral molecule, the more stable its K+
-adduct because there are more atoms to redistribute the energy released during the
adduct formation, and the longer its lifetime. 2. Collision frequency . Since K+ ions
can react with any organic molecules in gas phase, amino acids with larger sizes have
greater collision frequency with K+ ions. In other words, larger molecules have greater
reaction rate, yielding a higher concentration of adduct ions in the gas phase. 3.
Polarizability. The larger a molecule, the greater its polarizability, and the stronger its
interaction with a charged K+. Heteroatoms give greater dipole moments to a molecule,
thus give stronger charge-dipole interactions with K+ ions, thus better stability. 4.
Branching effect. H-AAA-OH and H-GGGG-OH have very similar molecular weight
and size. However, the branched peptide H-AAA-OH (MW = 231) is found to be 4 times
as sensitive as H-GGGG-OH (MW = 246), as shown in Figure 5.5.

Note that the relative response factor should be treated carefully. When an amino
acid residue is added to a peptide molecule, the response may or may not increase
significantly depending on the size of the original molecule. This statement can be
illustrated by the following experimental data for the relative responses of an amino aicd

and five small peptides, assuming the relative response of phenylalanine (H-F—OH) is l:

 

 

 

 

 

 

 

 

 

L 100%: 1002.
j K"' adduct of H-AAA-OH
7. m/2285
l T '1 1 T 1"f'1f1' I IT I Ifi
_I 100%- 8928.
~ K+ adduct of H-GGGG-OH
. rn/2270
' If IfiT I? I 7'1" I' 71' flfvfl V'l
— 100%. 10960.
rfif'l 'I'j"1"ijrfi"Ivvrvjj"r1'v'frfiv"1"'*T
o 25 so 75 100 125 150 175 200 225 250 275
Sean

Figure 5.5

Relative responses of H-GGGG-OH and H-AAA-OH in K‘I'IDS.

 

 

 

105

 

Compound H-F—OH H-FG-OH H-FL-OH H-GGF-OH H-GFL—OH H-AFLA-OH

 

These data indicate that the response factor for the addition of each amino acid
residue is not a constant. For example, from H-F-OH to H-GF-OH, the response
increases by a factor of five; while from H-FL-OH to H-GFL-OH, the increase in
response is 10/7 =1.5; From H-FG-OH to H-GGF-OH, there is no significant change in
response at all. It appears that, as the size of a "large" peptide increases, the response of
the peptides does not change significantly since the additional functionalities can not

interact with K‘I’ ions, thus no further stabilization effect can be observed.

B. Structure Discrimination

In characterizing thermal degradation of peptides, we have found that some of the
neutral species formed are cyclic and others are linear. It is natural to suspect that the
structure of a molecule may affect its binding ability with K+ ions because different
structures can provide different interaction sites and interaction coordination. For
instance, a neutral (Bn-t-H) molecule has a cyclic configuration and not all of its
functional group can interact with K+ efficiently to form a multidentate adduct at the
same time. For a linear molecule, however, it can form such multidentate adducts more
efficiently because of its lack of ring constraint (much more flexible).

This speculation was tested. If equimolar amounts of H-GP-OH and cyclo(-GP)
are analyzed by K+IDS, the resulting spectra only contain the [M+K]+ species of each. If
the intensities of these two peaks are monitored, the results shown in Figure 5.6 are
obtained. The time-dependent response reflects the desorption of the two species, and the

relative responses clearly show that the linear form is "sampled" much more efficiently

106

 

 

 

 

 

 

 

 

 

 

 

 

.3 100%= 26544 (H'GP-OH+K)+

i 111/2211

T 100%: 7192 (Cyclo-GP+K)+

j m/z 193

I r I I ' r :1; a r I T: I A I A I
‘. 100%= 37120 TIC

O 10 Z) 30 4O 50 60 70 a)

Scan Number

Figure 5.6 Mass chronograms showing the relative responses of H-GP-OH and

cyclo-(GP) in the K‘I'IDS experiments.

107

than the cyclic form by K+ ions — the response is 8 times larger for the linear form. It is
reasonable to believe that a linear molecule can form a more stable multidentate complex
with K" than can a constrained cyclic molecule, thus the results of these response studies

are not unexpected.

C. Discrimination Studied By Thermal Degradation 0f Tri- And Tetra-Peptides
The mass spectral data of tripeptides can be used to deduce the relative response
factors of smaller peptides. Note here that skeletal bond cleavages are observed for these
peptides and the formation of (Bz-H) and (Y1+H) is observed.
For example, by comparing the responses from H-GGF-OH and H-FFF-OH, the

response of cyclo4FF) relative to cyclo-(GG) can be deduced. If "I" is used to represent

the relative intensity of ions, then 11(cyclo(FF)):11(F) = 36:3 can be obtained from the
spectrum of H-FFF-OH. From the spectrum of H-GGF-OH, 12(cyclo(GG)):12(F) = 14:29
can be obtained. Note that the subscript “ 1” and “2” indicate the data are from spectra 1
and 2, respectively.

11(cyclo—(FF)):11(F)/12(Cyclo-(GG)):12(F) = 36:3/14:29 = 25

Ir(cycle-(FF)):I1(F)/I2(cyclo-(GG)):I2(F) = I(cyclo-(FF))/I(cycl<>-(GG))

I(cyclo-(FF))/I(cyclo-(GG)) = 25
This means that cyclo-(FF) is 25 times as sensitive as cyclo—(GG)) to K+ attachment. By
the same method, the responses of amino acids can be obtained. For example, by
comparing the spectra of H-GGG-OH and H-GGF—OH, the response of phenylalanine
relative to glycine can be deduced.

The relative responses of linear and cyclic dipeptides can be deduced from a
single tetra-peptide spectrum since the formation of a linear (Y2+H) and a cyclic (Bz-H)
commonly is the dominant reaction. However, loss of H20 sometimes may complicate
the situation (make the response for a linear molecule smaller than it should be). From

the K'I'IDS mass spectra of H-FGGF-OH (Figure 3.3) and H-FGFG-OH, it was found that

108

the linear dipeptide H-FG-OH and H-GF-OH are five times more sensitive than its
corresponding cyclic dipeptide cyclo-(FG) (without considering the loss of H20 from
(Y2+H) to a cyclic (B2-H)). In contrast, from the spectrum of H-PFGK-OH (Figure 3.6), '
the cyclo—(PF) is four times more sensitive than the linear H-GK-OH due to the large size

difference of these two.

D. Discrimination against Weak Binding: Experiments with Mixed Ionic Emitters

From thermodynamic data, we have proposed that discrimination is due to the
small binding energies betWeen K+ ions and different neutral species in the gas-phase.
To demonstrate this in K+IDS, experiments with mixed ionic emitters, which emit more
than one type of alkaline metal ions, have been performed to show the efficiency of alkali
metal ions in sampling the same gas-phase peptide/fragments. To be more specific, we
have used two mixed ionic emitters: one containing Na+ and K+, and the other
containing K+ and Cs+ ions. The results from peptides with these thermionic emitters are

shown below.

1. Na"' and K" ionic emitter

When a peptide such as H-VGGFL-OH is heated in an ion source equipped with a
K‘I' and Na+ mixed emitter, thermal degradation and desorption occur, which is then
followed by alkali metal ion (Na+ and K+) attachment. At any given time, neutral
species are presented to both K+ ions and Na+ ions with equal concentration. Assuming
there is no discrimination for sampling "large" neutral species for both K+ and Na+ ions -
reasonable due to physical trapping, then the peak intensity ratio of (M+K)+ and
(MI-Na)+ will reflect the relative ion currents of K+ and Na+ ions inside the ion source.

If there is no discrimination, for any given fragment F, the peak abundance ratio of

109

(F+K)+ to (F-t-Na)+ would be the same as that of (M+l()+to (M+Na)+. The data shown
below is for the peptide H-VGGFL-OH, which is obtained from Figure 5.7.

I(M+Na+)/I(M+K+) = 13/25 = 1/2

I(Y3+H+Na+)/I(Y3+H+K+) = 55/100 = 1/2

I(B2-H+Na+)/I(Bz-H+K+) = 4/2 = 2/1

At the high-mass region of a spectrum, the large sizes of K+ adducts of both

(Y3+H) and intact analyte molecules make the adducts relatively stable, thus no
significant discrimination is observed. When the size becomes smaller, discrimination
against the formation of K+ adducts shows up. For example, the relative concentration of
K+ ions is twice that of Na+ ions according to the adducts formed at the high mass region
of the spectrum. If there is no difference in the binding energy between K'I‘ ions Na+ ion,
then I(B2-H+Na+)/I(B2-H+K+) should be l/2. In fact, the ratio is found to be 2/1. This
observation indicates that there is a strong discrimination against the formation of K+

adducts or the adduct ions with weaker binding energies.

2. K+ and Cs+ mixed ionic emitter
The same experiment is also performed for K+ and Cs+ mixed ionic emitter, with

Cs+ a weaker binding ion than K+ ion. The data shown below is for H-VGGFL-OH,
which is obtained from Figure 5.8.

I(M+Cs+)/I(M+K+) = 3/1

I(Y3+H+Cs+)/I(Y3+H+K+) = 1/1
In this experiment, the emission of Cs+ ions is about four times that of K‘I' ions. When a
large neutral species such as a pentapeptide is attached to these ions, the large sizes of the
peptides and multiple interaction sites make both the I(+ and Cs+ adducts stable, little
discrimination was seen, as is shown in the ion intensity ratio between Cs+ adducts and
K+ adducts. When a smaller molecule such as the (Y3+H), a peptide, is attached by

these ions, discrimination against the formation of Cs+ adducts is seen. The yield of Cs+

110

 

H-VGGFL-OH ‘jjoiK’I

(Na+ & Ki" emitter)

J.

l

 

 

4 (Y 3+H+Na)+
' 358
(Bz-H+Na)+
I 179 (M+Na)+ (M+K)+
‘ (Bz-H+K)+ 514 530
/l95

 

 

160180 200 220 240 260 280 300 320 340 360 380 400 420 440 460 480 500 520 540 560

m/z

Figure 5.7 Relative responses of K+ and Na+ to a peptide H-VGGFL-OH and its
thermal degradation products.

 

111

 

 

 

 

 

 

 

+

« H-VGGFL-OH (Y3+H+K)+ “gym”
.34— (K+ & Cs+ emitter) 374
8 .
g (M+Cs)+,
o “I 624
.2.
g .
Q)
‘3‘ _ (M+K)+

530
_WWILI I4: IAII Irr‘rrI LILw‘r‘ IIILI LI I‘ rI II
fiI I I I I | I I
480 520 560 600 640 680

160 200 240 280 320 360 400 440

m/z

Figure 5.8 Responses of K+ and Cs‘I' to a peptide H-VGGFL-OH and its

degradation products.

112

adducts is much smaller (only 30%) than that of K+ adducts, as is demonstrated by the

adduct ion abundance ratio.

E. Summary (Discrimination in Interpretation of K+IDS Spectra)

Based on the cyclization mechanism for thermal degradation and the subsequent
discrimination in the K“ adduct formation step against small molecules and against cyclic
molecules, interpretation of K+IDS spectra of peptides becomes straightforward. Many
important features of K+IDS spectra can be explained easily: 1. For underivatized
peptides, (Yn+H+K)+ ions formed are linear while their corresponding (Bn-H+K)+ ions
are cyclic. The discrimination against cyclic molecules leads to the fact that the spectra
of underivatized peptides are dominated by (Yn+H+K)+ ions while their corresponding
(Bn-H+K)+ ions are very weak. 2. For N-terminally blocked molecules such as N-
acetylated peptides, the dominant ions are (Bn+0H+K)+ ions and (Cn+H+l()+ ions
because these ions are linear while their corresponding ions are cyclic. 3. For molecules
such as H-GGFL-OH, (Yn+H+K)+ ion abundance increase with their sizes due to
size/volume effect. 4. The discrimination against peptides with cyclic structures
provided us another evidence for the proposed cyclization mechanism.

In summary, the weak interactions between a K+ ion and a neutral molecule leads
to a finite lifetime for the nascent K+ adducts. The limited lifetime is reflected in the
response discrimination against smaller molecules and cyclic molecules, which give
much weaker signals. The discrimination in K+IDS results in simple, easy-to-interpret
spectra of underivatized peptides, which are dominated by [Yn + H + 10““ ions. The
determination of the relative responses of amino acids and small peptides makes it much
easier to interpret K+IDS spectra, and information regarding the probabilities of bond
cleavages at different sites along a peptide backbone can be evaluated more accurately.
For example, in Figure 3.3, the [Y2+H+K]+ peak is about five times as intense as the

[Y3+H+I(]"' peak, but this does not mean that the rate of bond cleavage at the second

113

amide bond from the N-terminus is five times that for the third amide bond, since there is
a response discrimination against (Y2+H). Since the response for (Y3+H) is about five
times that for (Y2+H), the cleavage at the 2nd amide bond is estimated to occur at a rate
25 times that for the 3rd amide bond. For [Y1+H+K]+, the peak is 80 times less intense
than that for [Y2+H+K]+» but cleavage of the 3rd amide bond from the N-terminus is
only 16 times that of the second amide bond since the response favors (Y2+H). Thus,
relative response studies suggest that the data in Figure 1 reflects the following

distribution of primary, low-energy thermal degradation products, as an example:

 

 

—— 4% > (Cyclo-FGG) + H-F-OH
H-FGGF-OH ——90% > (Cyclo—FG) + H-GF-OH
— 6% > "Cyclo—F" + H-GGF-OH

 

References

l. W. R. Davidson, P. Kebarle, J. Am. Chem. Soc. 1976, 98, 6133-6138.

J. Sunner, P. Kebarle, J. Am. Chem. Soc. 1984, 106, 6135.

Clementi, E.; Popkie, H.; J. Chem. Phys. 1972, 57, 1077.

Kisenmacher, H.; Popkie, H.; Clementi, E. J. Chem. Phys. 1973, 58, 1689.

9'99“!"

Kisenmacher, H.; Popkie, H.; Clementi, E. J. Chem. Phys. 1974, 59, 5892.

Chapter 6

Collision-Induced Dissociation of Deprotonated Peptides

I. Introduction

The determination of the amino acid sequence of peptides, by positive fast atom
bombardment (FAB) mass spectrometry, is one of the best known applications of mass
spectrometry, which has been well studied and well documented [1]. However, the
application of negative ion FAB mass spectrometry to peptide analysis has, with notable
exceptions, been used principally to detect (M-H)‘ ions. Recently, several research groups
have studied the collision-induced dissociation (CID) of deprotonated tri-peptides and di-
peptides [1—9]. It is becoming more clear now that MS/MS data from peptide negative ions
can produce useful structural information, even though the fragmentation behavior of
deprotonated peptides is little understood.

In this chapter, the fragmentation behavior of larger deprotonated peptides, is
studied. In Bowie's early studies on the fragmentation of deprotonated peptides, only
dipeptides and tripeptides were studied. Although their results provided some very useful
information about the fragmentation of deprotonated peptides, especially side-chain specific
fragmentations, the sizes of the peptides studied are not large enough to provide general
rules for the fragmentation of deprotonated peptides. First, smaller deprotonated peptides
usually give relatively more fragmentation because the internal energy deposited from
collisional activation is much greater [10] and the energy can only be redistributed among a

smaller number of atoms, generating various non-sequence ions. Secondly, some

114

115

reactions characteristic of larger molecules may n0t occur for smaller peptides due to
smaller internal energy deposition. For example, C ions have not been frequently observed
for dipeptides and tripeptides [1-10] while these ions are among the two dominant ion
series observed for larger peptides [this work]. These ions provide additional information
about the amino acid sequence of a peptide. Moreover, there is no obvious sequence-ion
series for small peptides because their peptide chains are too short to generate large
numbers of sequence-related ions. One relevant fact is that designations for sequence ions
have not been used in these papers. Finally, there is a pressing need for a better
understanding on the fragmentation of deprotonated peptides since deprotonated peptides,
in some cases, can be more abundant than protonated peptides in an ion source, which can
provide better sensitivities for MS and MS/MS analysis.

In this chapter, several important questions relevant to the fragmentation
mechanisms of deprotonated peptides are addressed in detail. Those are: What is (are) the
structure(s) of the deprotonated peptides from which all the fragments are derived? What
are the dominant fragment ions from CID of deprotonated peptides? How are they most
often formed?

To start with, general concept of tandem mass spectrometry and B/E linked scans
will be introduced in section II. In section IV, the site of deprotonation for peptides will be
investigated with deuterated peptides and derivatized peptides. Deprotonation in the gas-
phase will be proposed and supported. In sections VI and VII, fragmentation of
deprotonated peptides will be investigated. Charge-localized negative ions (peptide
derivatives) are used. Other peptide derivatives and deuterated peptides are also used in
exploring the fragmentation of deprotonated peptides. Mechanisms leading to the
formation of the major ion Series will be proposed. In section VIII, fragmentation of
charge-localized peptides will be discussed. Mechanisms for the formation of fragment

ions will be proposed. Ions formed in negative CID will be compared with ions formed in

116

positive mode. At the end of this chapter, the applicability of negative ion CID of

deprotonated peptides for peptide sequencing will be discussed.
II. Tandem Mass Spectrometry

A. Introduction

When peptides are subjected to fast atom bombardment, protonated peptides are
generated efficiently. These ions are chemically very Stable and have little excess energy.
AS a result, nOt much fragmentation takes place and the signal consists mainly of the
isotope cluster for (M+H)‘*‘. Fragment ion abundances are generally ten times or more
lower and the spectrum is, therefore, dominated by the Signal for protonated peptides.
Thus, such spectra contain little structure information beyond molecular weight.

As a result of lack of fragmentation and the accompanying lack of Structural
information, other methods were developed to induce cleavage of bonds. This can be
achieved by transferring additional energy, sufficient to induce fragmentation of the stable
molecular ion formed in the initial ionization process. Collision with a neutral gas is
presently the most widely used, leading to collision-induced dissociation (CID). The

selection of precursor ions and the subsequent mass analysis are achieved by tandem mass
spectrometry (MS/MS). The precursor ions, generally (M+H)+, in the first mass
spectrometer (MS-1) are fragmented upon collision with inert atoms such as He, or some
other energy transfer process, and the product ions (fragment ions) are recorded in the
second mass spectrometer (MS-2). For MS/MS analysis, the ideal instruments are four-
sector instruments such as JEOL HXl 10/HX110 with EBEB configuration with high

resolution for both precursor ions and product ions , or triple quadrupole instruments

(T SQ).

117

B. B/E Linked Scan

For the JEOL HXl 10 mass spectrometer (EB configuration)- the instrument used
for this dissertation, however, MS/MS was achieved in a different manner named B/E
linked scan. In this process (see ion flow diagram below), precursor ions, along with all
other ions produced in the initial FAB process, collide with He gas in a collision cell
located at the first field-free-region (lFFR) which is right in front of the electric sector,
leading to collision-induced dissociation (CID). When the magnetic field strength (B) and
the electric field strength (E) are scanned at the same time with B/E constant, it is called B/B
linked scan. In a B/E linked scan, only ions with the same velocity as the precursor ions
can pass through the two sectors, and be detected. Therefore, ions detected from a linked

scan are from precursor ions with the same m/z values.

 

Ion source ----> lFFR ----- > E sector ----- > B sector -----> Detector

 

When an ion F1 is fragmented to form F2 during collision with He, a small portion
of the translational energy of F1 (10 keV) is converted into internal energy ( a few electron
volts). Bonds cleave and fragments are formed. All fragments from the same precursor

ions will travel with practically the same velocity as its precursor ions.

For singly charged ions to pass to the detector, they have to follow the equations:
eE = va/Rc
CE = mv/Rm
Where E is the electric field strength, B is the magnetic field strength, Re is the radius of

the electric sector, Rm is the radius of the magnetic sector, v is the velocity of ions, m is the

mass of ions, e is the charge on a Single electron.

118

For a precursor ion F 1 (mass m1) and its product ion F2 (mass m2), both with the
same velocity, to pass the electric sector, they must obey the following equations:
651 = m1V2/Re
e132 = m2v2/Re
Thus 151/E2 = ml/m2. If 131 is set for passing the precursor ion F 1, then the product ion
F2 can pass the electric sector by lowering the electric field strength to E2 in proportion to
its mass reduction from m1 to m2.

For the same ions to pass the second sector (magnetic sector), they must follow the

following equations:
681 = mIV/Rm
682 = mZV/Rm

Hence B1/B2 = m1/m2. By lowering the magnetic field Strength proportionally, a product
ion can pass the magnetic sector being detected.

When B and E are lowered at the same time with the same proportionality, a
fragment ion can pass both detectors being detected and the mass of the fragment ions
(product ions) is directly proportional to the electric and magnetic Strength. This can be
seen by combining the above results: El/Ez = m1/m2 = B1/Bz and B1/E1 = B2/Ez =
constant. By scanning B and E with B/E constant (linked scan), all ions detected are from
the same precursor ions.

For a precursor ion to pass both sectors, the above two equations can be
rearranged into B/E = Rej(va) = K/v. So the B/E ratio is a velocity-related constant for
a given spectrometer. When the instrument is operated in this mode, it can be used as a
velocity selector - only ionS with the same velocity as the precursor ions can pass both
sectors being detected. At the beginning of a B/E linked scan, the B/E ratio is adjusted to
pass selected precursor ions by the computer control system, based on calibration of the

instrument.

119

For B/E linked scans with the JEOL HX 110, the resolution for precursor ion
selection is not very high, 0.5% about a preset mass. Generally ions within a few mass
units of the precursor ions will be selected. However, this inaccuracy in precursor ion
selection does not affect the appearance of a spectrum very much. For example, the (M-
H)‘ for H-VGGFL-OH is at m/z 490.2. The B/E linked scans were performed at m/z at
490.2, 489.2 and 491.2 for this peptide. The spectra (Figure 6.1a-c) were obtained and
compared. It is found that the three spectra are almost exactly the same, except that the one
at (M-H)‘ (490.2) has better Signal/noise ratio and greater totally ion current. This
observation indicates that the CID spectrum for a compound would be acceptable if the

preset m/z values is within a mass unit from the m/z value of MH‘I' or (M-H)‘ ions.

III. Experimental Section

Mass spectrometry: Ions were produced with 6 keV beam of fast Xenon atoms in a
JEOL HX-110 double-focusing mass Spectrometer operated in the negative ion mode. The
accelerating voltage was 10 kV and the resolution was set at 1000 (10% valley). Glycerol
was the liquid matrix chosen to obtain all spectra because it generates spectra with the best
signal-to-background value for all of the peptides Studied. For CID-MS/MS, helium was
used as the collision gas in a cell located in the first field-free region. The helium pressure
was adjusted to reduce the abundance of the precursor ions by 30%. A JEOL DA-5000
data system generated linked scans at constant B/E. With CID experiments at a resolution
of 1000, a main Slit setting at 3000 resolution will produce much better resolved CID
Spectra. This setting is thus recommended and used for CID-B/E linked scans.

Chemicals: The peptides H-VGVAPG-OH, H-YGGFL-OH, H-YGGFL-NHz,
and t-BOC-PPPP-OH were purchased from the Sigma Chemical Co., St. Louis, Missouri,
and used without further purification. Aminomethanesulfonic acid, aminoethanesulfonic

acid and aminobenzenesulfonic acid were purchased from Fluka Chemical Corp.,

120

 

 

 

 

 

R‘a
i j H-VGGFL-OH
: 34
i I
V
e I
8.2 Y -CH4
3 2% - ‘ :2:
3 I C2 C3 Y3 375 C4 446
a 14 172 229 I (23+!)- 334 376 -RL
e , 130 186 261 \ \ 391

c , ‘. — ——- - .I

150 200 250 300 350 400 450
111/2

1
R Ib
e 4
1.. H-VGGFL-OH
i 1
V
:61
A j 474 ,
b
u44 446
3 J
112 .1 172 229 277 331\334 375 375 397
1'1
5 1130 186 261 I \l 391 417433

1 __ _._____ ,, , . I ,- 3.2; - __ __ __, . -

1 200 250 300 350 400 450

 

Figure 6.1 CID-BIB linked scan Spectra for H-VGGFL-OH with precusors set at:

a) rn/z 490.2, (M-H)’; b) m/z 489.2.

121

 

 

 

 

 

 

4
R <3
e 1
1 f H-VGGFL-OH
a
1 3‘
i 3
V
e 1
AZ‘ 4:114
b ‘ Y -
u 1 2727 (Z4-H)' 31:}; RV 474
1 c c
3 . 2 3 Y3 375 c4 446
.14 172 229 I (2340' 334 375 -111,
3 Y1 (Zz-HT 318 331 Y4 417 433
e 130 186 261 \ \ 391
0__ _ ___. ___ _ _ _~-_2..___.3_ .. _ ___‘__‘_
150 200 250 300 350 400 450
111/2
2
4C
R 1
i - H-VGGFL-OH
‘ 4
t 1
i J
.V 4
e 4
1‘ 397 400
A 4 / 474
3 I 375 376 I 435
n 3 446
d 1 172 229 \
331 334
' I 277 391 417
n 1
c 1 130 186 261 \
e ‘I— _. _ _ fl ,_ _--_,,-;V___. ! l _ _‘r — — ___7__7‘___ __
150 200 230 300 350 400 450

Figure 6.1 CID-BIB linked scan spectra for H-VGGFL-OH with precusors at:

a) m/z 490.2, (M-H)’; c) m/z 491.2.

122

Ronkonkoma, New York. H-AGGFL-OH, H-VadGVadAPG-OH, H-VGVAadPG-OH,
H-VGGFL-OH were synthesized in the laboratory at the University of Michigan. All other
derivatization reagents were also purchased from Sigma Chemical Company without
further purification.

Peptide derivatives (C-terminal esterification and N-terminal acetylation): N-
terminally acetylated peptides such as Ac-VGVAPG-OH and Ac-AGGFL-OH, are formed
from the free peptide using acetic anhydride, with a small amount of pyridine as a catalyst.
The C-terminal methyl ester derivatives of peptides, such as H-VGVAPG—OCH3 and H-
AGGFL-OH, were formed from the free peptide using methanolic HCl. Detailed
procedures have been described by Knapp [11]. For peptides with both termini blocked
such as Ac-VGVAPG-OCH3, equal volumes of acetic anhydride and 10% of pyridine
methanol solution was added to the free peptide. The mixture was allowed to stand at room
temperature for 15 minutes (longer time for higher yield). All of the reaction mixtures were
dried on a Speedvac to evaporate all derivatizing reagents. This procedure also helps
eliminate acid-catalyzed hydrolysis if the derivatives are prepared for later use.

Deuterated peptide methyl ester (peptide-OCD3) was prepared as follows. A
procedure very similar to the above was used. However, CD3OD instead of CH3OH, was
used to prepare the hydrogen chloride methanol solution. The reaction mixture was dried
on a Speedvac. During this reaction, exchangeable hydrogen atoms of a peptide (NH2-,
-NH-, -COOH)' will exchange with the -OD deuterium of CD300. To make the derivative
deuterated only on the methyl ester moiety, back exchange with water was used. 50 11.1 of
water was added to the dried vial and the solution was dried on a Speedvac. This back-
exchange procedure was repeated three times and the resulting residueprc-dissolved in water
for use. It is advisable not to use methanol for either back-exchange solvent or solvent for
the product because the transesterification reaction, which occurs easily, can lead to a

mixture of esters.

123

For peptide sulfonic-acid derivatives, the corresponding amino sulfonic acids such
as aminomethanesulfonic acid (AMSA), aminoethanesulfonic acid (AESA) and
aminobenzenesufonic acid (ABSA) were used as the derivatizing reagents. They are
coupled to the C-termini of peptides by using l-ethyl-3-(dimethylaminopropyl)
carbodiimide (EDC) as the coupling reagent in a pyridine-pyridinium chloride buffer
solution of pH 6.

For peptide phosphonic acid derivatives, the corresponding aminophosphonic acids
such as aminomethanephosphonic acid (AMPA), 2-aminoethanephosphonic acid (2-
AEPA), 3—aminopropanephosphonic acid (l-APPA), l-aminopropanephosphonic acid (3-
APPA), 2-aminoethanephosphate (2-AEPO4), 6-aminohexanephosphate (6-AHPO4), were
used as the derivatizing reagents. The same procedure as above was used.

For N-terminal derivatives such as sulfobenzoic acid (SBA = HO3S-C6H4-CO-)
derivatives of peptides, coupling with the above method gives rise to very low yield. The
efficiency of the coupling can be improved by adding N-hydroxylsuccinamide (NHS) at a
higher pH buffer solution (N-morpholine-ethanesulfonic acid buffer at pH 6.5), as
described recently by Sehgal and Vijay [12]. When a strongly acidic derivatizing reagent
such as SBA is used, its pH value should be adjusted before the coupling reaction to assure
proper reaction pH value. Otherwise, the buffer solution can not keep the pH value of the
solution within the desired pH range with the addition of SBA. In our experiment, the
SBA aqueous solution (pH = 2) is neutralized to about pH 6 before it was used. A reaction

time of five hours or longer is usually required to achieve high yield.

124

IV. Deprotonation of Peptides

A. Deprotonation in the gas-phase or the liquid-phase?

When peptides are bombarded with fast atoms, they may be deprotonated in the
liquid-phase or the gas-phase, as shown in Figure 6.2. In the case I, a deprotonated
peptide is formed in the liquid phase due to acid-base equilibrium, this preformed ion is
then desorbed directly from the liquid phase to the gas phase. According to this model,
peptides with high liquid-phase acidity will have much higher efficiency in deprotonation
thus will be more sensitive to MS analysis. In other words, a non-ionic species will have
much lower efficiency of deprotonation or no deprotonation at all. In the second case, a
peptide molecule is desorbed as a neutral into the gas phase; this neutral gas-phase molecule
is then deprotonated by a deprotonated glycerol molecule via gas-phase ion—molecule
reactions.

To test these desorption/ionization models, the responses of H-VGGFL-OH and its
various derivatives were made and tested, each with different ionization efficiencies in the
condensed phase such as glycerol matrix. For these compounds in a polar matrix such as
glycerol, H-VGGFL-OH is considered zwitterionic due to the acid-base properties of the
termini; Ac-VGGFL-OH is slightly acidic due to its C-terminal carboxylic acid group:
SBA-VGGFL-OH is very acidic due to the strong acidic group (sulfonic acid group
~SO3H) at the N-terminus and the carboxylic acid group at the C-terminus; H-VGGFL-
0CH3 is non-ionic. If equal molar amounts of the peptide and their derivatives are
bombarded with fast atoms in a polar matrix such as glycerol, different ion currents of (M-
H)‘ from these compounds will be obtained. If the response, the ion current from (M-H)‘
of H-VGGFL-OH, is defined as 1, the relative response of a derivative can be obtained
from the (M-H)‘ current ratio of the derivative and the peptide. The responses of these

compounds in formation of deprotonated molecules are listed in Table 6.1.

125

 

Deprotonation in the gas-phase (Route 1)

or in the liquid phase (Route 11) ?

(M-H)' (g) 3(2) (M-H)' (g)
A

We) (G-H)‘ (g) Route II

 

 

  

M(l) H (M-H)‘ (l) + H"

 

G = glycerol M = peptide

 

 

Figure 6.2 Deprotonation of peptides in glycerol matrix with FAB.

 

126

Table 6.1 Negative ion FAB responses of a peptide and its derivatives

 

£99114; Chgge state in matrix Relative Response
H-VGGFL-OH zwitterionic 1
Ac-VGGFL—OH Ac-VGGFL-O‘ or neutral 1
SBA-VGGFL-OH 'O3S-C6H4-CO-VGGFL-OH 0.4
H-VGGFL-OCH3 neutral (non-ionic) 0.7

 

All the derivatives give responses smaller than that of H-VGGFL-OH. These
observations lead to at least two conclusions: 1. It is not necessary to have preformed ions
in the liquid-phase to produce gas-phase deprotonated ions. The spectra of H-VGGFL-
OCH3 and H-VGGFL-OH make it very clear. This comparison, however, does suggest
that the C—terrninus of a peptide is an important site of depmtonation. 2. Preformation of
deprotonated peptides in the liquid phase does not enhance the number of gas phase
deprotonated molecules. The responses of the two peptide derivatives, Ac-VGGFL-
0H and SBA-VGGFL-OH, are clear evidences. 3. Therefore, deprotonation in the liquid-
phase is not the mechanism for the formation of gas-phase deprotonated peptides.

Note here there are many factors that could contribute to the change of responses.
For example, derivatization changes the solubility, and surface activity of a peptide (in a
specific matrix) which leads to a change in FAB response/(M-H)' intensity. However, the
effect of these factors can not be fully evaluated at this time. (The important point here is
that, if the liquid-phase mechanism, RCOO' (1) ----> RCOO' (g), does not occur, then we
don't know from where the H+ is lost. It could be -CHR-, -NH-, or -COOH).

127

1. Are peptides desorbed in zwitterionic form or in neutral form?

Peptides in aqueous solution are believed to be in a zwitterionic form. In a very
polar matrix such as the commonly used glycerol, a peptide dissolved in it should also exist
in a zwitterionic form.

When fast atoms (3 keV) bombard peptides in glycerol, it is our belief that only
non-zwitterionic form of peptides are considered to be desorbed and deprotonated. First,
non-ionic forms of peptides have much less interaction with the polar matrix such as
glycerol and requires less energy for desorption, thus are preferentially deprotonated over
the zwitterionic form. In addition, the high stability of (M-H)’ ions from negative FAB-
MS and MI-I+ in positive mode, demonstrated by the intense peaks in the mass spectra,
suggest that peptides are desorbed into the gas phase in a non-zwitterionic form.
Otherwise, deprotonation or protonation alone in the gas phase will release a lot of energy
(a 160 kcal/mol = 7 eV) in addition to the energy obtained from fast atom bombardment.
This energy is far greater than the average internal energy deposition from collision
activation, and is enough to induce high-energy fragmentations such as homolytic
cleavages.

The energy released for the deprotonation of a peptide in zwitterionic form can be
estimated as follows:

H3N+-~COO' ----> H2N-~COO' + H+ (1)

AH“ (l) = -PA (H2N~~~COO' ) = -PA (H2N-C2H5) = -885 kJ/mol

(G-H)‘ + H+ ------ > G (glycerol) (2)

AH' (2) z - AH' acid ( C3H70-H) = -1571 kJ/mol
Combining the above two equations will result in equation (3):
(1) + (2) = (3):
H3N+~-COO' + (G-H)’ ------ > H2N~~~COO' +G (glycerol) (3)
AH' (3) = AH” (1) + AH’ (2)

128

z -1571 - (- 885) = - 686 kJ/mol

= -160 kcal/mol
This means that the deprotonation process is very exothermic if such a process occurs.
However, no significant fragment ions are observed from regular FAB spectra, indicating
that peptides are desorbed into the gas phase in non-zwitterionic forms.

It should be noted that some fragments from relatively high-energy processes are
observed with FAB bombardment. For example, metastable spectra of (M-H)’ of peptides
frequently produces (Zn-H)‘ ions which are supposed to be ionic radicals (negative distonic
ions). The late formation of these ions may suggest that the excess energy obtained from
FAB and from deprotonation is only slightly greater than the bond energy of NH-CHR (~
80 kcal/mol), much less than the number shown above, indicating that the desorbed species

is not in the form of zwitterionic form.

2. Thermodynamic support for gas-phase deprotonation

When the (M-H)‘ of a peptide H-VGGFL-OH is subjected to CID, the ion currents
from fragment ions of both termini are about the same (Figure 6.1a). When AMSA is
coupled to the C-terminus of a peptide such as H-VGGFL-OH, ion current from the C-
terminus is far greater than that from the N-terminus (Figure 6.3a), indicating a significant
charge-localization at the C-terminus of the peptide. When a ABSA derivative is analyzed,
the charge localization effect is even more effective (see Figure 6.3b) because all fragment
ions are C-terminal ions. Is it the change of gas-phase acidity or the change of the liquid-
phase acidity that makes the difference?

Apparently, liquid phase deprotonation can not explain it because both the peptide
and the derivatives are considered as zwitterionic molecules. No preformed (M-H)’ (l) is

present in the polar matrix such as glycerol.

129

 

 

 

 

 

 

 

 

 

 

 

RSI a
‘ 412503
g4; H-VGGFL-AMSA .———
. - Y4
8‘ M- at m/z 583.3
E I ( H) 434 CH4
v _: I -C H
03: (Z4_m-l 3 8
A . 468 |
b2: 7 l I
g 1 (21‘1” 21 I |
1
d 4 207 203 (2240- I
:1; (AMSA-1)“ \ \ 354 (2340' I I
e
t 7 7 _ 7. 7_ 7 7 7 77 J I _ 7 I_ L
1'0 200 300 400 500
111/2
1
R jb
O
1.; H-VGGFL-ABSA .14,
a
g j (M-H)’ at m/z 645.4 “’2
v67: I
’ . <Y2-2m- -ch
Q 1 430 551 I
11 4I (22:91!) (22.-H)- (13°10-
11 415 473 Y4
a *(ABSA-l) Y1 | Y2 Y ,.
a i - , 3
"2.172 w] 235 (2221-!) | 432 I489 (24+!)
° ‘ 226 “5 / | 530
e 1 \I
‘ ___ ____ _____ ___- ______ “777____ _ ____ __ _ ____I1_7_______‘d_ l.
200 300 400 500 600
m/z

Figure 6.3 CID-BIB linked scan spectra of deprotonated peptides (derivatives):
a) H-VGGFL-AMSA; b) H-VGGFL-ABSA.

130

Consider gas-phase deprotonation. Since sulfonic acid is more acidic than
carboxylic acid in the gas phase, thus the former (equation 4) is much more competitive
than carboxylic acid group (equation 5) in deprotonation.

NHz ----- SO3H + (G-H)‘ ----- > NHz ----- SO3' + G (glycerol) (4)

NH2----COOH + (G-H)‘ ----- > NH2---COO‘ + G (glycerol) (5)

Therefore, charge-localization is observed at the sulfonic acid group. Charge-localization is
indicated by the fact that only C-terminal ions are observed for ABSA derivatives.

With gas-phase deprotonation theory, the difference of CID spectra between ABSA
derivatives and AMSA/ABSA derivatives can be explained. The difference lies in their gas
phase acidities. Even though the gas-phase acidity data for these derivatizing reagents are
not available, the following data can provide insights into the effect of aromatic
functionality on gas-phase acidity.

AG'acid (CH3COOH) = 1429 kJ/mol
AG'acid (C5H5C00fl) = 1388 kJ/mol

The introduction of the benzene ring reduces the gas-phase acidity by 40 kJ/mol ==
10 kcal/mol, which makes C6H5COOH more favored in deprotonation than CH3COOH.
If this is the same for the sulfonic acid analog, which should be a fair Statement, then
ABSA with a benzene ring should be favored in deprotonation than AMSA/ABSA.
Therefore, in the gas-phase deproronation reaction, ABSA derivatives are seen
deprotonated more at the SO3H group than AMSA/ABSA derivatives. Because of this,
ABSA has a stronger charge-localization effect than its aliphatic analogs. In summary, the
results from charge-localization experiments support the gas-phase deprotonation

mechanism.

131

B. The Site of Deprotonation of Peptides
In early papers [1-10], the deprotonation site was assumed to be at the C-terminus

of a peptide. The argument for this is based on the following gas-phase acidity data (amide
-NH- was not considered in these paper), with the underlined H considered. Note that the
gas-phase acidity of glycerol is considered because it is the most commonly used matrix for
peptide analysis.

AG'zrcid (peptide CflR) = AG'acid (Cfl3COOH) = 1518 Kcal/mol

AG°acid (peptide COOH) = AG°acid (CH3COOH) = 1429 Kcal/mol

AG°acid (glycerol OH) = AG°acid (C3H70l-l) = 1566 Kcal/mol.
However, the reference data cited are not very convincing because acetic acid is quite
different from amino acids, and because the a-hydrogens in a peptide (except for the N
terminal one) are also dramatically different from the a-hydrogen in acetic acid. The
presence of a amide group ( —CONH- in a peptide) can make the deprotonation of the OL-
hydrogen atom much easier. In addition, with the deprOtonation of an (rt—hydrogen, the
negative charge will be stabilized by two electron withdrawing groups (~NH-‘CR-CO-) and

tWO resonance SITUCIUI‘CSI

o 8)}
a”) ”5 ___ n l
——NH-—CR—C— ‘ —N H—CR=C-—

This argument can be easily supported by comparing the following gas-phase acidity data
with those shown above:
AG'acid(-CONHCHRCONH-) == AG’acid(CH3COOCfl2COCH3) =1440 Kcal/mol
AG'acid (NH2CH2COOH) = 1401 Kcal/mol.
The small difference between these two values suggest there may be significant

deprotonation at a—carbon positions.

132

Table 6.2 compiles a list of gas-phase acidity data for the most common functional

groups in peptides, with the gas-phase acidity value of a commonly used matrix glycerol.

Table 6.2 Gas-phase acidity of various functionality in peptides

W Magic] (kJ/mole) Referenee Compou 11;!
N -terminus 1640 CH 3CH2Nfl2
-COOCH3 1610 HCOOCH3
Glycerol 1550 . CH3CH20fl
CH3CO- 1 520 CH3COOH

amide NH 1480 CH3CONl_-I_2

—CHR— 1 440 CH3COOCHzCOCH3
C-terminus 1400 N H2CH2COO_I;I_

 

If the gas-phase acidity of a functional group is greater than that of glycerol,
energetically it can not be deprotonated by deprotonated glycerol. If the gas-phase acidity
value of a functional group is smaller than that of the glycerol, then it can be deprotonated
by glycerol. Based on the data in Table 6.2, deprotonation can occur at the C-terminus,
the amides and the a—carbons of a peptide because the AG'acid [13] values for these sites
are smaller than that for glycerol.

It should be noted that depretonation at side ehaigs ef seme amine aeids, sueh as

11' 2' -. n ° -Ci -. 0. tin rosin hi iin an. tr 0 ooh: 1 u‘
majfieanfifim’em. However, deprotonation at these positions usually does not generate
significant sequence ions, thus they were n0t studied in this work. The discussion below

will be focused on the deprotonation along peptide backbone structures.

133

1. Does deprotonation occur at a—carbon position?

The N-terminus of a peptide is the least likely site of deprotonation. The high free
energy required for the deprotonation makes it the least favored considering that
deprotonation at many Other sites are much more favored energetically.

Deprotonation at the C—terminus is assumed to be the most probable from other's
work [2-8] due to its high gas-phase acidity. In addition to this site, are there any other
sites of peptides showing significant deprotonation at fast atom bombardment?

To determine the site of deprotonation, deuterium-labeled peptides were used.
Since deuterium exchange reactions are rarely complete, peptides with deuterium
incorporated at a—carbon positions were synthesized and used in our study to assure high
purity of deuterated peptides.

The FAB spectra of Ac-glycine—ordZ (CH3CONHCD2COOH), H-VGVAPG-OH,
H-VGVAadPG-OH and H-vadovadAPG-on, H-AFBdZLA-on were obtained; the
relative intensities of (M-H)‘ and (M-2H)'/(M-D)' ions are listed in Table 6.3. It should be
pointed here that the purity of these deuterated peptides have been examined in these

experiments.

Table 6.3 Relative intensities of (M-l)‘ and (M-2)’ * ions for (deuterated) peptides

Peptide Rel. Int. (M-H)‘ Rel. Int. (M-2H)’/(M-D)'
H-VGVAPG-OH 100 5
H-VGVAadPG-OH 100 5
H-vadovadAPG-OH 100 5
CH3CO-NHCD2COOH 100 5

 

*Note: (M-l)‘ = (M-H)‘; (M-Z)’ = (M-ZH)’ and/or (M-D)‘

The data in the above Table indicate that there is no change in the relative intensity

of (M-H)’ and (M-2H)'/(M-D)' ions. Therefore, there is no noticeable deprotonation at the

134

a—carbon positions (-CHR-) to the carbonyl group. Deprotonation of a—carbon positions

in underivatized peptides does not occur for regular peptides.

2. Does deprotonation occur at amides or at Ot-carbon?

To answer this question, consider the other two potential sites of deprotonation
besides the C-terminus of a peptide. Gas-phase acidity data suggest that deprotonation can
occur at both the amide position and the (rt—carbon position, with the latter favored. In
FAB experiments, which one is dominant?

To determine the deprotonation of amide and a—carbon positions, experiments with
C-terminally blocked and a—deuterated peptides, such as H-VGVAadPG-OCH3 and H-
vadHGvadAPG-ocng, were used, and the data for (M-H)‘ and (M-2H)'/(M-D)' were
compared with that for H-VGVAPG-OCH3 (see Table 6.4). In this process, the (M-2H)’
ion, which has the same m/z value as the (M-D)‘ ion, has been taken into account; the
intensity of (M-2H)‘ ions for the undeuterated peptide is subtracted from that of the (M-2)’
ions [(M-2H)‘ and/or (M-D)' ions] for the deuterated analogs since it appears in the
negative ion FAB-MS spectrum of every compound.

To exclude possible deprotonation at the methyl ester moiety, deuterated methyl
ester, H-VGVAPG-OCD3, was synthesized and examined by mass spectrometry (Table
6.4). It is found that there is no deprotonation at the methyl ester moiety of the derivatized
peptides. Therefore, esterification is an effective method in blocking the C-terminus of a
peptide without introducing new sites of deprotonation.

Based on Table 6.4, it is concluded that there is no removal of a-hydrogens (of
-CHR- groups) for peptides in these experiments. Therefore, deprotonation of amides is
much easier than that of a—carbon positions in FAB-MS experiments. The sites of
deprotonation, along a peptide chain, are the C-terminal carboxyl group and the amide
groups. However, it is NOT clear yet to what extents deprotonation occurs at each of these

positions.

135

Table 6.4 Relative intensities of (M- 1)’ and (M-2)’ ions for (deuterated) peptide esters

 

Peptide Rel. Int. (M-H)‘ Rel. Int. (M-2H)'/(M-D)‘
H-VGVAPG-OCI-l3 100 5
H—VGVAadPG-OCH3 100 5
H—vadovadAPG-ocng, 100 5
H-VGVAPG-OCD3 100 5

 

3. Why is deprotonation of amides favored over that of a-ca rbon
positions? 8

As stated above, thermodynamic data suggest that deprotonation of a—carbon
positions (-CHR-) is favored compared to that of amide hydrogen. However, experimental
results indicate that the expected deproronation of a—hydrogen does not occur. Instead,
deprotonation of amides is favored. Why?

One possible explanation is that the deprotonation kinetics dictates the sites of
deprotonation because the deprotonation process is very short in time. For deprotonation at
these two potential sites, the one with a faster deprotonation rate will dictate the site of
deprotonation while the rates of these reactions are determined by the stability of their
transition states.

To understand the stability of the two transition states in Figure 6.4, consider the
polarity difference between C-H and O-H/N-H bonds. Due to the difference in
electronegativity among C, H, N and O atoms, OH is basically a non-polar bond while 0-
H and N-H are strong polar bonds. For reaction I in Figure 6.4, the strong charge-dipole
(N-H) interaction with a charged oxygen atom of deprotonated glycerol stabilizes the
intermediate/transition state, which leads to the deproronation. In reaction II, however,

there is no such strong interactions (OH is practically a non-polar bond) present, thus the

136

.8868 .8888 we 85 .85 32:8 .3 :ocggenoe 39$ c8 afieaaoc ouoeg We Emmi

 

 

sea 2 8%
828E.
Penguin n: 1 t .. V0
22.54am oz I
0 n3 ow:

. 85m 0 mo

0 :0 Sauce. 2082A a one
:OOUsiA 69.93,.— W

mzlm - - .. o

zo + :0 I a o

oEEoSOxm .M n
«_..—z a Ow?—
N:z mo

$36 + 288330 11%| 2 + -59

 

 

137

intermediate/transition state is not stabilized. Therefore, deprotonation at the (It—carbon

position is not observed. In summary, deprotonation requires a favorable transition state;

and the reaction with stable transition States will dominate.

4. . Charge-localization: Evidence for amide deprotonation

As can be seen from above experiments, deprotonation can occur at amides and the
C-terminal carboxyl group. Is deprotonation at amides significant compared to that at the
C-terminus? If deprotonation occurs only at the C-terminus of a peptide - that is, when
charge is localized at the C—tenninus of a peptide, what ions can be generated from CID of
the charge-localized peptides?

To make C—terminally charge-localized peptides, peptide derivatives were made
with sulfonic acid groups at the C-terrninus. As shown in the spectra of H-VGGFL-OI-I
and its derivatives in Figure 6.13, the dominant ions for the underivatized peptides are C
ions and Y ions, in about equal abundance. The dominant ions from I-I-VGGFL-AMSA
(Figure 6.3a) are C-terminal ions while N-terminal ions are rare and weak. For H-
VGGFL-ABSA (Figure 6.3b), there is no N-terminal ions, and all fragment ions are from
the C-tenninus. These spectra clearly indicate that C-tenninally deprotonated peptides can
generate only C-terminal ions!

Similar experiments were performed for N—terminally charge-localized peptides.
When a sulfobenzoic acid group is coupled to the N-terminus of a peptide such as H-
VGGFL-OH, charge will be localized at the N-terminus in this form, ‘O3S-C6H4-CO-
VGGFL-OH. When this ion was subjected to CID, it was found that all ions generated
were N-terminal ions (Figure 6.5), such as (An-2H)', (Bn-H+OH)‘. For a regular
peptide, these ions are not observed. These results suggest that N -terminally deprotonated
peptides generate only N-terminal ions. It also confirms that N-terminus of a peptide is not

the site of deprotonation.

138

 

 

 

 

1
5 I SBA-VGGFL-OH
184
z 1 (M-H)‘ at 674.3 B4+0H
$5.1 (M'ZH)- 561
e ‘ 515
A I
34‘ (AB-ZHT

‘ (At-2H)‘ 368
3 ‘ (A2-2H)' l

* 254 B3+0H
: «(SBA-HY'ZAO l

. . l l . A
C 7 ,
200 300 400 ' 500 600

Figure 6.5 CID-BIB linked scan mass spectrum of a deprotonated peptide SBA-VGGFL-OH.

139

For a regular peptide, if deprOtonation occurs at the C-terminus, as was stated in
other's work [2-8], then N-terminal ions will not be seen in CID spectrum of that
deprotonated peptide, as suggested by experiments with the above charge—localized
peptides. In fact, both C-terminal and N-terminal ions are observed for CID of regular
deprotonated peptides in about the same abundance. Therefore, there must be significant
deprotonation at amide sites from which both C-terminal and N-terminal ions can be
formed. The total fragment ion current from amide-deprotonated peptides is greater than

that from C-terminally deprotonated peptides.

5. Final comments on the site of deprotonation

In explaining the backbone cleavage of deprotonated di-peptides and tri-peptides,
rearrangement following deprotonation at the C-terminus of a peptide has been proposed
[1] to explain the formation of Y ions. Our experimental results suggest, however, that
there is a significant deprotonation at amides that is responsible for the formation of both
N-terminal ions and C-terminal ions. Deprotonation at amides produces more sequence

ions (C ions at least) than C-terminally deprotonated peptides.

V. Fragment Ions from CID-MS/MS spectra of Deprotonated Peptides.

A. Nomenclature

To designate negative ions precisely, the original Roepstorff nomenclature scheme
[14] will be used, as is shown in Figure 6.6. An, 13“, Cn, Xn, Yn and Zn are used to
represent the fragment formed at specific skeletal bond and terminus, where the subscript n

represents number of side chains contained in a fragment. If a hydrogen atom is removed
from a fragment, it is labeled by "H". For example, (Zn-H)” represents an anion that is

one mass unit short of a regular Zn fragment.

140

 

 

a) Roepstorff nomenclature [14] for peptide fragments

N- terminus All

C- terminus

b) Common fragment ions from CID of deprotonated peptides

Y ions: NHZWCOO’ NHZWCON'WCOOH;
C ions: NHZMCONH' NH2-CON'-CONH2;

(Z-H)‘ ions: “CHRMCOO';

 

Figure 6.6 a) Roepstorff nomenclature for peptide fragments;
b) Common fragment ions from deprotonated peptides.

 

141

When the amide bond of a neutral peptide, NH2~~CO-NI—I~~COOI-I, is cleaved,
two neutral fragments (radicals) are formed. The N-terminal fragment, NH2~~CO-, is
called 8;; fragment, and the C-terminal fragment, -NH~~COOH, is called Yn fragment.
For a Yn anion from CID of deprotonated peptides, the mass of a Yn ion is practically
equal to the mass of the Yn neutral fragment, differing only by the mass of an electron.
Note that Yn ions can be formed in many different ways and their structures can be
different.

It should be noted that the mass of a Yn in negative mode is different from the mass
of a yn in positive mode. In positive mode, a yn is a short-hand expression for a
(Yn+2H)+, which is two mass unit heavier than the Yn anion from negative mode.

Besides the ions from skeletal bond cleavages, there are a few ions from sidechain
cleavages that need to be clarified. The most important ions of this type are Dn ions and W
ions, which are also two mass unit lower than the corresponding dn and Wu ions in
positive mode. Dn ions can be considered formed from their corresponding (An-H)“ ions
[Dn = (An-H)’ - R], and Wu ions can be considered formed from their corresponding (Zn-
H)‘ ions [wn = (Zn-H)‘ - R]. V

Fragmentation in negative mode is more complicated than that in positive mode. In
positive mode, a peak is frequently accompanied by only the isotopic peaks. ”In negative
ions, however, peak clusters are always observed, especially for ions at high mass region
of a spectrum. This complication sometimes makes ion assignment difficult; experiments

with deuterated peptides are also less effective than those for positive ions.

B. Common fragment ion series

i) The most common fragment ions are Yn ions, Cn ions and (Zn-Hr ions (Figure
6.1a - the spectrum of H-VGGFL-OH). The latter two have not been encountered in the

study of smaller peptides such as di- and tri- peptides. Dehydration products of these ions,

142

and fragments of these ions from specific side—chain loss such as loss of CH20 from serine
[1] are also observed.

ii) (An-2H)“, (Bn-ZHY, (Xn-ZH)' ion series are usually not very abundant in
negative FAB spectra, except for those at the high mass end and those with cleavages
around aromatic side chains. However, if more acidic side chains (Asp and Glu) are at or
near the N-terminus of a peptide, (An-Hr and (An-2H)‘ ions can become fairly abundant.
When a charge is localized at the N-terminus of a peptide, (An-H)‘ become one dominant
ion series (Figure 6.5).

iii) (Zn-HY and (An-H)‘ ion series are the only two groups that are not two mass
units lower than their positive ion counterparts.

iv) When a charge is localized at the N-terminus of a peptide, or when a proline

residue is close to the C—terminus of a peptide, (Bn+O)‘ ions/(Bn-H+OH)‘ ions are formed

at the high mass region of a peptide spectrum (Figure 6.7).

C. Ion Structure and Stability

In the fragmentation of deprotonated peptides, Y ions and C ions are the most
frequently observed. To underSIand the formation and their relative abundances, it is
necessary to consider the structures of these ions and structure-stability relationships.

When a peptide is deprotonated at the C-terminus or amides, these deprotonated
peptides will all have resonance structures, which stabilize fragment ions formed in the CID
of deprotonated peptides. Fragments with resonance structures will generate much more
intense peaks in a mass spectrum. In a later part of this work, this assumption is
successfully used to explain the relative intensities of various ion series in deprotonated
peptides.

One of the most important ion series is the Yn ions and the following is the possible
structures of these ions:

‘NHCHR~~~COOH (I)

143

 

 

 

 

 

10
R .a
f 8. H-VGVAPG-OH
? ; (M-H)‘ at rn/z at 497.2
I
:61 -CH4
3 ; -C3H3 ‘
“41 453
n . Y2 c3 (B4+O)- I
:1 j 171 C2 271 Y3 343 (ZS-H)’ (35m)-
n2“ 172 (2340- 341 I 382 440
g j / 226 424
200 250 300 350 400 450
m/z

R l d
:2: H-VGVAa PG-OH
; : (M-H)‘ at m/z 498 CH4
‘v :
e .
A I -C3H8
31.: C2 453
3 j :21 172 (2 H) C3 3 (B40). (85+O) I

. 1 - ‘ - ‘
:1 4 /C2-dl 2327 271 342 3"“ (ZS-H)‘ 441

4 173
g ; / \

g‘ . l

200 250 300 ‘ 350
m/z

Figure 6.7 CID-BIB linked scan mass spectra of deprotonated peptides:

a) H-VGVAPG-OH; b) H-VGVAadPG-OH.

 

 

144

 

O

 

 

 

 

1
R .a
:83 H-VGVAPG-OH
a .
t : (M-H)‘ at m/z at 497.2
I
‘ l
:6 cm!
i: j -C3H8 i
u4~ 453
n 1 Y2 C3 (B4+O)’ I
2 1 171 C2 271 Y3 343 (ZS-HT (85+0)'
n21 /172 (23-H)- 341 I 382 440
g , 226 \ 47A It
0; ‘ . ; J
200 250 300 350 400 450
m/z
R5:c
:4; H-vadeadAPG-on
a J
i ‘ (M-H)’ at m/z 499.2
v3;
3 1
A : Y C3148 :
b ‘ 1 C2 - 455 ‘
24 e l
3 1 ‘7‘ ‘73 (2353;!) 'CH“
d I / (Z3-H)' (:3 Y3 (BM-0)“ (35+O)‘ , I
:14 226 273 342 345 442 i
g : \I l
3‘ ._.. __ _ ___-_'__ __ __________ __ _ ‘ I
200 250 300 350 400 450
m/z

Figure 6.7 CID-BIB linked scan mass spectra of deprotonated peptides:

 

a) H-VGVAPG-OH; c) H-vadovadAPG-ort.

 

145

NHzCHR~~~COO' (II)

NH2-CON'-COOH (III)
From the structures of these ions, it is easy to see that the last two structures are more
stable because they have resonant structures. Therefore these structures are most probably

the structures of Y ions formed in CID experiments.

 

8
to II fa QI A
NHZCHR~~~ C—O NHZCHR~~~ C=o

 

Ntllfe Q8“

H2~~C—-N-COOH v— NH2~~ C=N-COOH

 

Structure I is less stable due to the lack of a resonance structure. However, this structure
can rearrange to a much more Stable form easily via the following reaction (a very

exothermic reaction with 136' z - 60 kcal/mol), especially when the ion is not very large:

'NHCHR-~COOH > NHzCHR-~COO'

 

The other most important ion series from CID of deprotonated peptides is the C-
ions and the following is a list of their possible structures:
NHzCHR~~~CONH' (IV)
NH2~~CON'~~CONH2 (V)
'NHCHR~~~CON H 2 (VI)
Note here that the first two structures are more stable because they have the following

resonance SU’UCIUI'CS.

146

G
Q9/1L
NHZCHR~~~C=N H

o 8
(II {6 RI A
NH2~~C—N~~CONH2 NH2~~ C=N~~CONH2

(Iolfe

N H2CH R~~~C-N H

 

 

 

 

For Z-type of ions, the most commonly seen are (Zn-Hr and (Zn-2H)', with the
former frequently more abundant. The (Zn-H): ions are probably formed via homolytic

cleavage of C-terminally depr0tonated peptides.

R1\ ’R2 R,\ ,R2
IH —-> II + *1
CHCONH-~COO’ CHCONH~~~COO‘
(Zn-HY (Zn-2H)'

VI. Mechanisms for the Fragmentation of Deprotonated Peptides

In the study of ion fragmentation chemistry, two approaches have been frequently
used. One of them is through the use of deuterium-labeled compounds to trace the shift of
a hydrogen atom when it is involved. The other common method is by controlled
derivatization such as derivatizing specific functionality to examine the role of that
functionalities in the fragmentation of a specific compound [11]. In this work, both
deuterated peptides and derivatized peptides will be used to study the CID behavior of

deprotonated peptides.

147

A. H-Shift: Mechanistic Study by Using Deuterated Peptides

Mechanisms regarding the formation of some common ions involving the shift of
hydrogen atoms, from CID of deprotonated di- and tri- peptides, have been proposed in the
literature[1-10]. In these studies, deuterated peptides from deuterium/hydrogen exchange
have been prepared. It has been reported that. for all dipeptides studied [1], the shift of
a-hydrogen was observed.

In this study, partially deuterated larger peptides, with deuteriums at specific or—
carbon positions, were used and the CID spectra of these deprotonated peptides were
obtained. From the CID spectra of H-VGVAPG-OH, H-VGVAadPG-OH and H-
vadHGvadAPG-OH (Table 6.5 and Figure 6.7a-c), it was found that shifting of
deuterium did not occur in the formation of Yn ions. In the formation of C ions, most of
them are formed without a—hydrogen shift being observed. However, an (it-hydrogen
shift does occur to a small extent (30% in the case of H-VGVAadPG-OH), four atoms
away from the deprotonated amide nitrogen. Therefore, the formation of Yn ions does not
involve the shift of a—hydrogens. However, there are at least two ways that C ions can be

formed via one of which there is an a-hydrogen shift.

Table 6.5 Ions from peptides and their deuterated analogs

Peptide m/z(C2) m/z(Y3) m/z(Y4) m/z(Y2)
H-VGVAPG-OH (r) 172 242 398 171
H-vadovadAPo-orr (11) 173 242 399 171
H-VGVAadPG-OH (111) 172/173 243 399 171
Deuterium-shift ? no/ yes no no no

 

Note that the changes in the m/z values of the same ions from different peptides
above (for example Czyof peptides I and II are at m/z 172 and 173, respectively) are not

due to the shift of deuteriums except for C2 of peptide III.

148

This observation carries a very important message. First, the formation of either
Yn ions or Cn ions is not from a 1,2 - elimination process. Secondly, fragmentation of
larger peptides is different from that of dipeptides. In early papers on dipeptides [1], it was
found that the formation of Yn ions involved the shifting of an a—hydrogen atom to the
deprotonated C-terminus. This is clearly not the case in this study.

Since present mechanisms can not explain all the observations in CID spectra of

deprotonated peptides, alternative mechanisms have to be proposed.

B. CID Fragments from Derivatized and Underivatized Peptides
In fragmentation of deprotonated peptides, we believe that the site of deprotonation
plays an important role. To determine the effect of deprotonation sites on fragment ion

series, various peptide derivatives, as listed below, were made with expected deprotonation

 

at different sites:

W mprgtgngnign site Elam
H-V/AGGFL-OH C-termin us and amides Yn & Cn ions
Ac-VGGFL-OH C-terminus and amides Yn & Cn ions
H-VGGFL-ABSA C-terminus C-terminal ions only
H-AGGFL-OCH3 Amides C“ ions and Yn ions
SBA-AGGFL—OH N-terminus (An-2H)' & (Bn+0)‘

 

a) Regular peptides: With these peptides, the most dominant fragment ions are C
ions and Y ions (as shown in Figure 6.1a and 6.7a). There are also some other ions,
however with much lower abundances. With N ~tenninal acetylated peptides, the fragment
ions and their relative abundance are about the same, suggesting that the N-terminus is not

very reactive in fragmentation of deprOtonated peptides.

149

It should be pointed out, for small peptides such as di- and tri-peptides, C ions
were not observed very often from CID of (M-H)’. For larger peptides as was the case in
this study, C ions have been found to be one of the most important ions. This clearly
indicates that the effect of size on fragmentation because the size of a precursor ion affects
the energy deposition and the energy distribution. Note here that the center-of -mass

energy Ecom' maximum amount of energy that can be obtained from collision is
approximately inversely proportional to the mass of the precursor ion. (Ecom =
[MHd(MHc+Mp)]pr, where Ep is the translational energy of (M-H)‘ ion, Mp is the mass
of the (M-H)‘, MHc is the mass of the collision gas He.)

b) Amide-deprotonated peptides: When the C-terminus of a peptide is blocked
methyl ester, the deprotonation sites along the peptide backbone can only be amide
nitrogens. As shown in Figure 6.83-b, both C ion and Y ions are observed for the peptides
H-VGGFL-OCH3 and Ac-VGGFL-OCH3. However, Y ions are suppressed more. This
result clearly indicates that amide deprotonated peptides can produce both C ions and Y
ions.

The ion series from amide deprotonated peptides indicates that amide deprotonated
peptides barely produce Y ions. The suppression of Y ions, in the form of
‘NHCHR~~~COOCH3, is related to the lack of a resonance structure, and thus they are
less stable than their corresponding C ions (see part C of section V).

It should also be noted that the abundances for C ions are not as high as those from
the regular peptides, indicating that the C-temtinus of a peptide contributes to the formation
of C ions via a H-shift from the C -terminus to deprotonation sites.

c) When the C-terminus of a peptide is made to be a very acidic moiety such as
aminobenzenesulfonic acid moiety (ABSA = -NHC6H4SO3H), the C-terminus of the
peptide derivative will be the site of deprotonation and only C-terminal ions, dominantly Y
ions and (Z-H)‘ ions (Figure 6.3), are observed. This observation clearly indicates that C-

terrninally deprotonated peptides can not produce the C ions. Therefore, there must be

150

 

 

 

 

 

 

 
 
 

 

 

20 a
a ‘ C
. . ,3, H-VGGFL-OCH3
1 .
:1? (M-H)- at m/z 504.3
i .
V .
e .
1cr
A 4
b
2 (22-2H)' (ls-21!)“
d 5. 274 3'31
a ' Y3 _
n . C2 I Y (2443)
c < 172 211 242 259' 2921 l 348 339
e 1
0%
150 200 250 300 350 400 450 500
W2
1 b
a I
2.: Ac-VGGFL-OCH3
a 1
I; . (M-H)‘ at m/z 546.3
V
06-: -RF
A 455
b d
M, (2:31 (22+!)-
3 1 X1 \ 275
:24 172 \\ - C4 2 I
c j (24“) 418 I
O 4 Y3 389 I
c , '1 , . - i .
1 0 200 250 300 350 400 450 500
m/z
Figure 6.8 CID-BIB linked scan spectra of deprotonated peptides (derivatives):

 

a) H-VGGFL-OCH3; b) Ac-VGGFL-OCH3.

151

significant deprotonation at the amide sites which are responsible for the formation of C
ions.

When the N-terminus of a peptide is attached to a very acidic moiety such as
sulfobenzoic acid moiety (SBA = HO3S-C6H4-CO-), the deprOtonation site will be at the
N-terminus of the peptide derivative. As shown in Figure 6.5, the dominant ions are N-
terminal ions such as (An-2H)' and (Bn-H+OH)' ions. In the spectra of regular peptides,
this ions are rarely seen, consistent with the gas-phase acidity data which suggest that the

N-terminus of a peptide is not the site of deprotonation for regular peptides.

VII. Mechanisms: The Formation of Major Fragment Ions

A. Ions from Amide-Deprotonated Peptides

1,3 - Elimination mechanisms: When (M-H)‘ ions of n-mer peptides are
subjected to high-energy collisions with He, the an and Cn-1 ions are most frequently.
observed. These ions can be formed only via two mechanisms shown in Figure 6.9. Both
are 1,3 - elimination reactions, with one charge-induced and the other charge-remote.
Fragmentation in Figure 6.9a is more favored than that in Figure 6.9b because charge-
dipole interactions in Figure 6.93 is more favored than those in Figure 6.9b.

In Figure 6.9a, a H-shift occurs from one of the two termini. Based on this
mechanism, it is reasonable to believe that the following fragmentation reaction in Figure
6.10 can also occur. In this mechanism, a H-shift is from an adjacent amide NH. These
reactions can be classified as a 1,3 - elimination reactions.

Displacement - reaction mechanisms: C ions and Y ions can also be formed

via a displacement reaction without a H—shift, as is shown in Figure 6.11. In this

 

152

 

a) Y,” and CM Ions From charge-induced fragmentation

O
G O
H QWCOOH H o
R H

N— R N
.
H. n l

H
e R o R H
O N G)
>\/ ——-> NHzMnu—NH + k0
NH2 4%! o
H

b) Y“ and Cn-l ions from charge-remote fragmentation

O

9
+ H2N "N'M'COOH
‘3 R N

H H

e R R
fNWI‘T) -—> HZNWSMNHz + fico

NH2 0
C Ions
H50

 

Figure 6.9 The formation of Y,” and CM ions via 1,3 - elimination mechanisms.

a) Charge-induced fragmentation; b) Charge-remote fragmentation.

 

153

 

 

The formation of C ions and Y ions

HR H

9/€= e R
HZN'WN) ——> H NmNH +
t’ O 2 «wwwcoon

swoon

HR H

.w o
HzN 3, o ——> HszN + NHWCOOH

p
swoon o

 

Figure 6.10 The formation of Y and C ions via 1,3 - elimination reactions

(Charge - induced fragmentation)

 

 

 

O H
H N 'c':\ I S coon" 0 PI1
2 III/f\fi/ Hsz'c'\e + R /C\'E;NwCOOR"
R' NH
R"OOC

0
Run fiRH
HNmC O HNM 0
2 EN 9 2 C\N

-—> R"OOCWNH +

meme“ Nx 9X”:
H

HR'

 

Figure 6.11 Cn ions and Yu ions from displacement reactions.

 

155

mechanism, only electron-pair-shifts are involved. Because of this, steric hindrance for
such processes are small thus such processes may be significant.

The above reaction mechanisms can easily explain many experimental facts. First,
they can explain the fact that the relative intensities of Y ions and C ions are about the same
for underivatized peptides. The reason is that the formation mechanisms for these ions are
exactly the same. Even though the C ions thus formed are more stable due to their resonant
structures, the following rearrangement reaction (very exothermic reaction!) can proceed to
stabilize the Y ions thus formed, especially when the ions formed are not very large (so the
termini of the ion will more probably interact with each other):

‘NHCH(R3)CO~~~COOH < ------ > NHzCl—I(R3)CO~~~COO'
AG’ == AG°acid (CH3COOB) - AG°acid(C2HsNflz)
= (1400 - 1650) kJ/mol = - 60 kcal/mol

In the gas-phase of a high vacuum, different parts of a large molecule tend to interact with
each other due to charge-dipole, dipole-dipole interactions (and van de Waal force), leading
to folding of ions/molecules in the gas phase. Because of these interactions, this
exothermic reacrion is very likely to occur. 80 even though the Y ions formed initially do
not appear very stable, this rearrangement reaction can give much more stable species.
Secondly, this mechanism can explain the relative intensities of C ions and Y ions
when the C-terminus of a peptide is transformed into a methyl ester. It was observed that
esterifrcation severely suppressed the formation of Y ions. This is because the above
rearrangement reaction can not occur to stabilize the Y ions thus formed. The poor stability
of the Y ions (due to the lack of resonance structures) will suppress their formation. Thus
the C ions formed with resonance structures will be much more abundant than Y ions.
Supporting evidence for the displacement mechanism is the spectral difference
between the negative ion FAB-CID-B/E linked scan spectra of H-YGGFL-OH and H-
YGGFL-NI-Iz. When the C-terminus is a carboxylic acid, the intensity ratio of C2 to C3 is

156

3:1 (Figure 6.12a). However, when the C-terminus is an amide, the ratio become 1:3
(Figure 12b). This change occurs because the displacement reaction via the C-terminus is
sterically the most favored site thus reacrion via the C-terminus is favored. Since 6-
membered ring structure is generally the most stable, loss of a 6-membered neutral species
and the formation of C3 is the most favored (Figure 6.13). Note here that the C-terminus
is certainly not the only site for deprotonation, but fragment ions from deprotonation site
other than the C-terminus should be less abundant because fragmentation of these
deprotonated peptides will present more steric hindrance. Note here this case is almost
exactly the same as when it occurred in thermal degradation of peptides and peptide amides
[15] (also see Chapter 3 for detail).

1,4 - Elimination: As seen in the experiments with deuterated peptides, part of
the C ions formed does involve shift of (rt—hydrogen atoms, four atoms away from the
cleavage site. To explain the formation of these ions, a six-membered ring intermediate is
proposed and a tat—hydrogen is transferred to the deprotonated amide nitrogen atom (Figure
6.14). This mechanism suggests that a deprotonated amide nitrogen atom can deprotonate
an tat-carbon position (~CHR-) when enough energy is supplied (AG' = AG'acid(-CONH-)
- AG'acid(-CHR-) as = 10 kcal/mol, see section III). Another possibility is that the shifting
hydrogen goes to a non-deprotonated amide, forming a C ions such as NH2~~~CON'

-~NH2, by the same mechanism.

B. Ions from C-terminally deprotonated peptides

If the C-terminus of a peptide is the site of deprotonation, then the charge is fixed at
the C-terminus and the situation will be the same as the sulfonic acid derivatives discussed
earlier in this article. As can be seen in Figure 6.3b, all ions from the C-terminally

deprotonated peptides are C-terminal ions. N-terminal ions can not be produced from the

157

 

H

.a
: H-YGGFL-OH :31 W
‘ (M-H)‘ at m/z 554.3

co
. 4

O)
.11.

 

445

"f
n
N
Q
"3
N
5.

  
 
 

C3 —C3H3

N
l A A A

 

005000.366? O4H—rtarwwan

 

 

no
t»
O

 

 

1 0 200 7 300 400 500
111/2
1
R 1b
3 . H-YGGFL-NHZ
:81 (M-I-I)‘ at m/z 553.3
1 1
Ve
e , C3
A j 293 Y4
Sr (23- ' 390
n
d
a
n
C
O

 

 

 

 

 

Figure 6.12 CID-BIB linked scan mass spectra of (M-H)’ ions of:

a) H-YGGFL-OH; b) H-YGGFL-NHZ.

938on 5 ma «IZ-An_OO>-EV

.moEEa acumen Bacouoaocéacgorv 80¢ 32 U do cou~§8 2:. 2.0 83mm”—

 

 

O O

"—
m
m m 28— U
,zg/Tm I szloowlm

4m z + N:zlogl I: All
2 ,2

w. m a. E o m a.
a _ _ _ i
\U/\ m Z/ \ :U \U/ _ \ I
o o 2:! oo>l : :z xofl/o zmIoQT:
m u 09 up

 

 

 

159

 

 

 

 

Figure 6.14 The formation of C“ ions via a 1,4 - elimination mechamism.

160

C-terminally deprotonated peptides. So only Y and (Z-H)‘ ions are derived from these
charged species.

To form a Yn ion from a C-terminally deprotonated peptide, shifting of a hydrogen
atom from the N-terminal side of the cleavage site must have occurred because C-terminally
fixed—charge peptides showed abundant Y ions. This process may proceed via a 1,2-
elimination mechanism or via cyclization from the N-terminus or amides. Since the charge
is fixed at the C terminus of the peptide, this will be a charge-remote process and steric
hindrance at related sites should be very important, as is demonstrated in the thermal
degradation of peptides. However, this process may take more time to occur since the

reaction is a charge-remote process.

From the CID Spectra of H-VGVAPG-OH, H-VGVAadPG-OH/ocng. and H-
VadHGVadAPG-OH/OCH3, shifting of a—hydrogens are not observed. This indicates
that the 1,2 - elimination process does n0t occur on the time scale of the experiment.
Therefore, cyclization is the fragmentation mechanism (see Figure 6.15). 15 the shifting
hydrogen from the N-terminus? This is most probable for the loss of small cyclic neutral
molecules. To form larger neutral cyclic molecules, a H-shift from amides is probably the
best choice. Note in this process, cyclization is not charge induced, it is a remote-site
process. In the gas phase, the spontaneous folding in minimizing the surface tension may

help speed up these reactions.
VIII. Fragmentation of charge-localized peptides

A. Charge-localization on peptides

In mass spectrometry, fragment ions are believed to be from fragmentation of the
corresponding molecular ions. In electron impact (El) ionization, odd-electron molecular
ions are formed because an analyte molecule either gains or loses one electron.

Fragmentation of these ions are frequently driven by the radical. With softer ionization

161

 

 

R‘ 1|?»
l
H\ CH H\ ’CH‘ go
’ ‘ —N W‘XX) N C
If gg H ___). l I + HZNWCOO
/C\ ,N_ /C\ N

 

Figm'e 6.15 The formation of Y ions from charge-remme processes

such as 1315 reactions shown here.

 

162

techniques such as FAB, even electron ions are formed. For example, molecular ions
formed in FAB are MH+ or (M-H)’, by gaining or losing a proton. When these ions are
forced/induced to fragment, fragmentation is considered to be driven by the charge -
namely charge-induced fragmentation. Following the protonation or deprotonation
process, not only the protonation site and its adjacent atoms have the greatest changes in
electron density, but also it is the deprotonation site that is affected first and most by the
energy release during a protonation process {Deprotonation of peptides for example, AG“ ==
AG°acid (CH3COOfl) - AG'acid(glycerol-OH) = (1400 ~1550) kJ/mol = 35 kcal/mol)}
Both processes make the bonds at and near the deprotonation sites the easiest to cleave.
Therefore, the charge site (protonation or deprotonation site) within a molecule is crucial to
the formation of fragment ions.

When protonated or deprotonated peptides are formed from fast atom
bombardment, a charge site-the protonation/deprotonation site, is not at a single position.
In positive mode, the charge site is determined by the most basic sites present in a peptide
molecule such as that at an arginine side chain. To be more precise, protonation occurs
with a statistical distribution such that the most basic site is protonated preferentially. In
negative ion mode, the most acidic site is most probably the site of deprotonation. Still the
deprotonated peptides (M-H)‘ exist as a collection of isomers with a proton removed from
various sites within the molecules. Therefore, the peak representing (M-H)‘ does NOT
represent a single chemical species. For (M-H)’ ions, different ions may have different
deprotonation sites. Ions with different charge sites may have different fragmentation
patterns.

Since charge site is important in the fragmentation of protonated or deprotonated
molecules, efforts have been made in previous research to put a charge at a fixed site in a
molecule, by making charged derivatives, mainly to control/direct the fragmentation of
MH+ of the analyte molecules [16-19]. Charge-localization in positive ion mode was also

proposed to provide a way of differentiating leucine and isoleucine in peptide sequencing

163

by Biemann and his coworkers [18-19]. It has been noticed that dn ions or wn ions in
positive ion mode, formed from specific side-chain cleavages necessary for the
differentiation, can only occur when charge is at or very close to either termini of a peptide
[19]. Recently, Wagner et al. [20] used triphenylphosphonium (TPP) derivatives to
achieve a different purpose. Charge localization was used to enhance the response of a
peptide in FAB by charge localization and surface activity enhancement, in addition to
improved fragmentation.

Charge-localization for positive ions are frequently achieved by attaching tertiary
ammonium or phosphonium groups at or near either terminus of a peptide. References for
the synthesis of these derivatives can be found in a recently work by Biemann and his
coworker [2 1 ].

Charge-localization for peptide negative ions has not been reported previously.
However, W. J. Griffiths and coworkers [22] used a series of aminosulphonic acid
derivatives to improve the fragmentation of bile acids via charge-remote fragmentation. In
this study, amino-sulphonic acids and aminophosphonic acids have been used as ’the

derivatization reagents for charge localization on peptides.

B. Fragmentation of Deprotonated Peptides with Charge-localization.

A list of amino-sulfonic and amino-phosphonic acids, as described in the
experimental section, have been used for derivatization reagents for charge-localization.
Three peptides, H-VGGFL-OH, H-YGGFL-OH and H-AG(Sar)FL-OH, have been
derivatized with the procedures described in the experimental section. The CID (including
contribution from metastable ions) spectra were obtained.

Based on CID/metastable spectra of these peptide derivatives, sulphonic acid
derivatives have shown better charge-localization effect than phosphonic acid derivatives.
Among sulphonic acid derivatives, aminobenzenesulphonic-acids gives the best results for

charge localization. For example, when the C-terminus of a peptide is derivatized by

 

164

AMSA and ABSA (Figure 6.3), the dominant ions are C-terminal ions, indicating a fairly
strong charge-localization efficiency. However, a few N-terminal ions such as C3, with
low abundances, are always observed, suggesting an incomplete charge-localization (e.g.
some peptide molecules are deprotonated at sites other than the C—terminus in this case,
which leads to the formation of C3 ions). When the C-terminus of a peptide is derivatized
with ABSA, all ions observed are C-terminal ions (Figure 6.3b), indicating a complete
charge-localization at the C—terminus.

Fragment. ions from these three sulphonic acid derivatives are not exactly the same.
AMSA derivatives produce mainly (Z-H)‘ ions and 2 ions (uncommon!), and little Y ions
and W ions are observed (Figure 6.3a). With AESA derivatives, major ions are almost
exactly the same as those for AMSA derivatives. With ABSA derivatives, the ions are
mainly Y ions, (Z-H)‘ ions, and W ions ( corresponding negative counterparts of Wu in
positive mode).

(It should be noted that CID spectra provide more fragment ions than metastable
spectra. For these peptide derivatives, W ions and (Y-2H)’ ions are enhanced with
collisional activation.)

Charge-localization at the N -terminus of a peptide was also studied. The CID
spectrum of SBA-VGGFL-OH clearly indicate a complete charge-localization ( Figure
6.5). The dominant ions are found to be (An-2H)‘ ions, (Bn-H+OH)' ion - all N-terminal
ions. Therefore, benzenesulfonic acids moiety can be used effectively to fix/localize a

negative charge on peptides.

l. Fragment Ions from C-terminally Charge-localized Peptides
The major ions from C-terminally charge-localized peptides are Y ions, (Zn-HY

ions, (Ya-2H)‘ ions, and sometimes W ions.

 

165

(Zn-HY ions: These ions appear in CID/metastable spectra of charge-localized
peptides, regular peptides, even in C-terminally blocked peptides (sometimes). (Zn-H)‘
ions must be in the form of radicals also (-CHRCO—---COO'). Their formation can be best
described as homolytic cleavages of NH-CHR bonds from C-terminally deprotonated
peptides (equation 6), or peptides with deprotonation sites (amides) near the C-terminus

(equation 7).

-~NH-CHRCO-~~COO' ----- >~-NH- + oCHRCO~-COO' (6)
~~NH—CHR~~CON'-OCH3 ----> -NH~ + °CHR~~CON'~~OCH3 (7)
(M-H)' (Zn-H)’

mm'jgns: It is the radical nature of (Z-H)’ ions that probably leads to the
formation of (Z-2H)’ ions by loss of H atom from the sidechain (equation 8, R =

-CHR'R").

°(CHR'R")CHCO~~~COO' ----- > R"R'C=CHCO~~~COO' + '1'] (8)
(Zn-HY (Zn-2H)"

Therefore, glycine may not form (Zn-2H)' ions due to its lack of side chain atoms This is
confirmed by the spectra of H-VGGFL—OH and H-VGGFL-ABSA. For (Z3-H)’ and (Z4-

H)‘, there is no corresponding (Z—2H)’ ions observed because of the presence of glycine

residue at the cleavage sites. (Zz-ZH)' IS observed because a phenyl alanine side chain is

at the cleavage site.

ms; The radical nature of (Zn-Hr ions also lead to the formation of specific
sidechain losses forming W ions (equation 9). Due to the stable structure of products -

alkene in conjugation with a carbonyl group, this reaction is favored.

166

°(R"R'HC-)CHCO~~~COO' ----- > R'HC=CHCO~~~COO' + 0R" (9)
(Zn-HY Wn ions

The significance is that this ion series can be used to differentiate isomeric leucine and

isoleucine. For leucine and isoleucine, the side-chain loss is -CH(CH3)2 and -Cl-I2CH3

respecdvely, which provides the basis for mass differentiation.

Zn Ions: Zn ions are observed with AMSA and AESA derivatives of peptides
(Figure 6.3). Since the charge of such a derivative is localized at the C-terminus, a
hydrogen atom must shift from the N-terminal side to the C-terminal side of the cleavage
site via reactions such as 1,3-elimination. However, the source of the shifting hydrogen

has not been studied.
was: For charge-localized peptides, a hydrogen shift is involved, and the
shifting hydrogen must be from either the amides or the N -terminus of a peptide, as is

shown for C-terminally deprotonated peptides (1,3-elimination mechanism).

(Y-Z) ions. A hydrogen shift away from the C-terminal side of the peptide must

be involved. Therefore, H shifts from CHR must be involved (equation 10).

H
{H ——> N H2~~~C—H + HN=CR-~COO’

(10)
(M-H)‘ (Yn-ZH)'

167

Another possibility is that this ion series is formed from (Y-H)‘ ions (equation 11),

followed by ejection of a hydrogen atom (equation 12).

NH2~~CO-NHCHR~~COO‘ ----> N H2~~COo + oNHCHR~~COO' (11)
(M-H)' (Y-H)‘

oNHCHR~~~COO' ----- > H- + NH=CR-~COO' (12)
(Y-H)‘ (Y-ZH)‘

According to Wagner [23], the following charge remote fragmentation can occur for
charge-localized peptides in positive mode because it is very similar to an reaction (equation

13) observed by Gross and coworkers [24]:
Peptide ------ > (B-H) + (Y-H) + H2 (13)

With charge-localized peptides, the ions in positive mode will be (B-2H)"’ and (Y-ZH)+
respectively. If this mechanism is true for the (Y-2H)’ formed in CID of negative charge-
localized peptides, then the corresponding (B-2H)’ ions should be observed when a N-
terminally charge-localized peptide is used. However, such (B-2H)‘ ions have not been
observed in the spectrum of SBA-VGGFL-OH. These results suggest that the above

reaction VIII is not the reaction that occured.

2. Fragment ions of N-terminal charge-localized peptides
The major ions from N -terminally charge-localized peptides are (An-2H)’, (A-H)‘
ions and (Bn-H+OH)' ions. The latter are also formed when proline is present at or near

the C-terminus of a peptide, such is the case in H-VGVAPG-OH.

 

168

(An-H)‘ Ions: (A-H)’ ions are produced by homolytic cleavages of CHR-CO
bonds of a N-terminal charge-localized peptide (equation 14). (A-2H)' ions are formed by

ejection of a hydrogen atom from an (A-H)’ ion (equation 15).

'O3S-NHCHR-CO~~COOH ----- >‘O3S~~NHCHR-+°CO-COOH (14)
(M-H)’ (A-H)'
'O3S~~NHCHR- ------ > ‘O3S~~N=CHR + -H (15)
(A-H)‘ (A-ZH)‘

Or this reaction can proceed via 1,2-elimi nation reaction (equation 16).

'03s~-}~I-—CHR ——>'O3S~~~N=CHR+ HCO-~COOH (16)
H )r—aooOH
0

Dn ions: D ions have been observed for N -temrinally deprotonated peptides such

as SBA-VGGFL-OH. The presence of this ion series indicates the presence of an
intermediate species (An-Hr ion - a radical anion - a product of homolytic cleavage of N-
terminally charge-localized peptide (equation 17). As is the case for W ions, D ions can

also be used to differentiate leucine and isoleucine in the same manner.

R
\‘i/Rb Ré/H
CH

'O3S~-NHCl' —> 'O3S~~~NHC| + Rb (17)
H H

(M-H)' Dn ion

 

169

(Bm-O)‘ ions: This ion series is actually (Bn-H+OH)’ ions from cyclization
process by shifting the C-terminal OH group to the cleavage site (equation 18), as is the
case in thermal degradation of peptides [15]. In the spectrum of SBA-VGGFL-OH, this is

the major ion series. Such ion series usually only occur at the high mass range of a

peptide spectrum.
0 M.
n H\
035...... C- NH\ N H II

V“ [CHR ——> + 'O3S-~C-OH (18)

HO—fi R ‘
O O

(M-H)‘ (Bn-H+OH)'

In the spectrum of H-VGVAPG-OH, (B4+O)‘ and (B5+O)' ions are observed.
This forms a cyclic peptide by transferring a 0' to the cleavage site (equation 19). This ion

is not C-ion related because the proline prevents the formation of C ions.

H2N~-C— NH 0
5K? CHR --> :)><: + H2N-~|C—O (19)

170

IX. The Applicability of CID of Deprotonated Peptides

for Peptide Sequencing

A. Peptide sequencing with CID-MS/MS of Negative Ions

Peptide sequencing by CID-MS/MS in positive mode has been demonstrated to be
one of the most successful applications of mass spectrometry [25-27]. In negative mode,
however, detailed studies on CID-MS/MS have been extended only to di- and tri—peptides
before this disseration work [1].

As has been demonstrated in the preceding sections of this Chapter, the dominant
ions from CID spectra of deprotonated peptides are usually Cn ions, Yn and (Zn-H)“ ions.
These ion series, derived from skeletal bond cleavages of a peptide, contain important
information of amino acid sequence of the peptide. With one complete ion series or more
partial ion series identified, the sequence of a peptide can be deduced.

One of the most important approaches for sequencing starts by calculating the mass
difference between the (M-H)‘ ion and any fragment ion that could correspond to loss of
one residue, for example between (M-H)’ and (M-H)'-7l. Those with mass differences
that exactly match a residue mass (e. g. 71 = alanine) are candidates for Yn-1 ion. There
will likely be several of these, especially for larger peptides. For each of these Yn-1
candidates, one search for the next residue mass that could correspond to Yn-2. For
example, if there were peaks corresponding to (M-H)‘, (M-H)'-71, (M-H)‘-71-97, and so
on, the mass differences could correspond to Ala, Pro, and so forth. These data then
suggest a sequence starting at the N-terminus of Ala-Pro- Constructing a tree diagram
makes it easy to keep track of the candidate sequences. The best candidate(s) are those for
which the data produce the longest sequence. A check of the validity of Y ion assignment
is the identification of (Z-H)’ and (Z-ZH)’ ions, which would appear 16 and 17 mass units

lower in mass, respectively, than the corresponding Y ion. Ideally, there will be a

 

171

complete Yn series that provide the entire sequence. More likely, there will be several
partial sequences.

When Y series is absent or weak, (Zn-H)‘ ion series should be checked starting at
(M-H)'-16. A peak at (M-H)'-16-71 may suggest a (Zn-1-H)‘ ion and the sequence
starting at the N-terrninus of Ala- m, and so forth.

After finding the possible Y ion series, one should look for N-terminal fragments.
The C ions start one mass unit below the (M-H)’ ion mass. Search for peaks
corresponding to (M-H)'- (1+residue); this peak correspond to Cn-1 ion. As with Yn-1
ion, search for residue mass losses from the Cn-1 candidate(s) until one or more of a long
C-ion series are found.

° For example, the spectrum of H-VGGFL-OH (Figure 6.1a) as an "unknown" can
be analyzed as follows: The identities of the peaks in the above spectrum should be
identified first. For Y ions, the following series can be found: 490 --> 391 (-99 = Val)-->
334 (-57 = Gly) —-> 277 (-57 = Gly) --> 130 (-147 = Phe), which suggests a partial
sequence starting at the N-terminus of VGGF- (the next one may be Leu/Ile, to be
confirmed). For C ion series, start at 490-1 = 489, 489 --> 376 (-113 = Len/He) --> 229
(-l47 = Phe) --> 172 (-57 = Gly), suggest a sequence starting at the N-terminus of -GFL/I.
Therefore, the Cu and Yn ion series can provide enough information to deduce the
"unknown" sequence of this peptide, which is found to be V-G-G-F-L/I. Note that the
leucine and isoleucine can not be differentiated in this specific case (They can be
distinguished by charge-localization, which will be shown below.)

In peptide sequencing, ions from sidechain cleavages of a deprotonated peptide
should always be checked first. These ions appear at the high mass region of CID spectra
of deprotonated peptides, and provide important information about the presence of amino-
acid residues. For example, peaks at m/z values of 91, 92 and 93 mass unit lower than the
deprotonated peptides indicate the presence of a phenylalanine(s). Peaks at m/z values of

57 and 58 lower than the deprotonated peptides indicate the presence of leucine/rsoleucine.

172

Dominant cleavages can occur for sidechains such as those of threonine and serine,
which can be identified easily by neutral losses of 44 and 30, respectively, as demonstrated
in the spectra of H-VHLTEVEK-OH (Figure 6.16) and H-RPPGFSPF-OH (Figure 6.17).
The presence of these sidechains is obvious because it frequently produces a neutral loss
current much higher than any other ions. The dominant loss of a sidechain can also occur
for fragment ions. As is the case for the spectrum of H-RPPGFSPF-OH (Figure 6.17), in
which the ion current from Y-3O (CHzO; 30u) is almost as large as that from Cn ions.
Such ion series may be used to determine the position of these amino acid residues.

Charge-localized peptides (negatively charged) can be easily prepared for the
purpose of peptide sequencing when needed. Fragment ions from a charge-localized
peptide give simpler spectrum because these ions are either N-terminal or C-terminal ions.
For example, H-VGGFL-ABSA produces almost exclusively Y ions and (Zn-H)‘ ions,
which simplifies spectral interpretation. In addition, sidechain specific ions such as W and
D ions are obtainable in CID of deprotonated peptides, especially for charge-localized
peptides (see Figure 6.3b). The presence of these ions can be used to differentiate isomeric
leucine and isoleucine residues. Charge localization with sulphonic derivatives such as
ABSA and SBA derivatives can be served for these purposes. For example, the D1 ions
from H-VGGFL-ABSA indicates that the C-terminal amino-acid residue is leucine instead
of isoleucine.

Although CID spectra of deprOtonated peptides can provide structural information
of peptides, there are several serious limitations for the purpose of peptide sequencing
when it is compared to the corresponding positive mode.

1. The sensitivity in negative mode is usually much less than that in positive mode.
Even when the production of precursor ion is more efficient than that in positive mode, the
quality of CID spectra is also not as good. The reasons for this poor sensitivity will be

discussed in section II of this Chapter.

 

 

 

 

 

 

 

 

 

 

 

 

173

zqfi 441

5 3 H-VHLTPVEK-OH C 44

5
L (M—H)' at m/z 920.4 (Bum ,RE

6-
f . 618 859\
v . I -R \
e < I 33 ‘

1. \
A . \
b i Y H o jI740 (25-313. 601'

u 4 2— 2 644

3 . 252 7.55 on I 695 74s
. 581

a t \ (A4-H)‘ A5 1 645

n 212 \ 422 520

C

e 1

.2, . ,- . . ' ' . I a Cwmefioés
1" 200 300 400 500 600 700

m/z

Figure 6.16 CID-BIB linked scan mass spectrum of a deprotonated peptide H-VHLTEVBK-OH.

174

 

  

 

    

2-1
a I H-RPPGFSPF-OH
e 4
i j (M-H)‘ at m/z 902.5
I; ; Y5-30
v r 522
° ‘ |
1-1
A . 1(25-11)‘
b . Y4-30 536
l
U 1 Y3 3 C5
465 1
n ‘ 348 I y l 571
2 1 Y2 (Z3'“)" C4 I 49: 1 1Y5
n . 251 332! C3 423 | I552
C ‘ 355
e ‘ 1
t ' ‘ '
2'0 300 400 500

Y6-30
619

IY6
'649

  

C5
557

Y Ar sidechain
74,6 eavags
V860
R I \
- F \
v7.30 811 1341
715 1 I
1 | 1
'RR 1 :30
802

  
 
  

H2C=O a -30 ——1

 

     
 
  
 

 

 

 

 

 

Figure 6.17 CID-B/E linked scan mass spectrum of a deprotonated peptide H-RPPGFSPF-OH.

 

175

2. In positive mode, immonium ions (CHR=NH+) in the low-mass region of a
spectrum provide very important information of the amino acid constituents - the presence
or absence of some amino acid residues. In negative mode, such information is not
available because no corresponding ions are formed.

3. In positive mode, the high-mass region of a CID spectrum provides important
information about the presence or absence of an amino acid residue due to specific side-
chain losses. The neutral losses are very specific and peaks corresponding to these are
unique for each amino-acid residue. In negative mode, side-chain losses still occur, but
with much more complication. When the side chain of a leucine residue cleaves, there are
many ways by which the cleavages may occur. For example, loss of the whole side chain
[-C4H9 = -57] is observed. Loss of 58 is also observed in about the same abundance, loss
of 56 is also observed. Loss of 16 (CH4), 44 (C3H8), 15 (CH3) and 43 (C3H7) can also
occur for this amino acid residue. These non-unique cleavages can be very confusing , as
can be seen from the spectrum of H-FLEEI-OH (Figure 6.18). For instance, if both valine
and leucine residues are present in a peptide (Figure 6.1a - the spectrum of H-VGGFL-
OH), the spectrum can not suggest the presence of valine because leucine also give some
sidechain losses which are common for valine.

4. Sidechain cleavages of some amino acid residues such as threonine and serine
can be so dominant that peptides containing one of these residues may produce spectra
dominated by the specific cleavages [ Figure 6.17 - the spectrum of H-RPPGFSPF-OH,
and Figure 6.16 - the spectrum of H-VHLIPVEK-H]. For serine and threonine, the
dominant losses are -30 (CH2=O; 30u) and -44 [CH3CH=O; 44u]. The dominant
cleavages lead to the decrease in the total ion current of other fragment ions, which reduces
the structural information that is available for peptides without such sidechains. In
addition, these specific losses extend to fragment ions also, making sequencing more

complicated.

 

176

 

 

 

 

 

 

 

 

 

 

 

4
a :21 men)
1 I H-FLEEI-OH ,3» I
a d 1
1:3, (M-H)‘ at m/z 648.3 ”III I
3, * -<RE+H>I I I
‘ . 551 5MI I I
b I 372 I

« <Y2-H)- ' 575 | '
g . 241 <Y3-H20)- I 431 II I
d1« 372 I C3 1 C4 |
a . 1- - 534
g 1’12?” Y2 w: \‘Y 405 'm (2449- I
a , 259 313 3838 442 485

1 I

. ‘ l _
0W 0
100 200 300 ml 400 500 500
Z

4‘ b
R 1 (23°2H'H20)- 660
§ ; H-RKDV Y -OH 359
a .1 -
c3. (M-H) at m/z 678.3 MSidechain
1 . Cleavage:
" ‘ 123-2H)- 6?“
O 4
A27 3" (Aa-ZH)' 59H 23‘?
b I 370 I
d ‘ 300 / (A4-2H)' “RY

1‘ “'3 459 571 I I
a Y / 618

y 2 333 c3 D4 . -

c . 1521110 258 415 455 497 . 1
c , I . 11

,______2____._.___---,-____2--____.. -___, 122......3WMWM... -__.._--_2.:;¢.2.2_.

10 200 300 400 500 500

m/z

Figure 6.18 CID-B/E linked scan mass spectra of deprotonated peptides:
a) H-FLEEI-OH; b) H-RKDVY-Ol-I.

177

The presence of such residues, however, does not always produce the dominant
cleavages. This can be seen from the CID speCtrum of H-WAGGDASGE-OH (Figure
6.19).

5. Ion series can be very sensitive to changes in deprotonation sites. For example,
esterification of a peptide may cause dramatic changes in total ion current . When sites
such as carboxyl groups are blocked, ion current will decrease.

6. As is the case in positive mode, sequence ions are not always complete,
especially for large peptides. It is also observed that small fragments are not observed for

relatively large peptides, as is the case for H-RVYIHPF-OH (Figure 6.20).

B. Why is negative mode not favorable?

l. The poor sensitivity

The poor sensitivity of negative mass spectrometry (not with NCI or EC-NCI) has
been recognized for many years. The sensitivity of regular FAB in negative mode is found
to be about 100 times lower than that in positive mode. The efficiency of fragmentation of
(M-H)‘ in CID-MS/MS is also much less than that of (M+H)+. With experiments on a
JOEL BX 110 instrument, with full scale of precursor ion beam, positive mode produces
signal at 50% level while negative mode produces signals usually around 2% levels. What

are the reasons?

The above observations can be answered partly by the low electron affinity values
which lead to low efficiency for the production of (G-I—l)‘ (g) and (M-H)‘ (g) ions due to
their poor stability.

The production of (G-H)’ ions: With gas-phase ionization mechanism,

deprotonation of peptides occurs due to the following reaction:

 

178

 

4

A

A A

A A A A

’To4wnnwoau

A A A

 

OODDQDCU’?

0

H
.

 

 

 

H-WAGGDASGE-OH
(M-H)- at m/z 847.4
(ls-ZR)-
459 (A7-H)‘
i 616
/ A5-H)- I
C4 441|Y5 599 05 [(23.11)—
387 76 573
(Yz-Hzor P C5 ' “5C,
185 361 lsoz | 660
2— ’ if “"”‘_' —'

m/z

 

 

 

C8
717
-R'w -RtD
/ 802
.R'E
787
'RE
774
700
__ ___ ‘t
700 8 00

Figure 6.19 CID-BIB linked scan spectmm of (M-H)- of H-WAGGDASGE-OH.

 

 

 

179

 

 

 

 

 

 

 

 

 

2~ -RY

‘ 822
g H-RVYIHPF-OH
l ‘ _ -Rn
, . (M-H) at m/z 929.4 848
I .
i 1
v * (ls-H)-
° ‘ 658

1* l
A 4
656 -H

3 C2 Y3 Y4 . Y5 (Z7657 )’
n 271 398 C3 511 I 674
d I 34 I (A6-ZH)'” | Y6
a Y2 I (Z3-l-l)- (743" 539 (A7-2H)-| 773
n 261 I 382 I 495 l 547 736

._ _ _..—___ ___ _ _ _- ._ _ .7 _ ___.___.._2ia, “___;N

“ 300 400 500 600 700 800 900

m/z Mil

Figure 6.20 CID-BIB linked scan spectrum of (M-H)‘ of H—RVYIHPF-OH.

 

 

 

180

M(g) + (G-H)' (g) ------ > (M—H)’ (g) + 0(8)

where G(g) and M(g) designate gas-phase glycerol and peptide molecules respectively.
Obviously, (G-H)‘ (g) has to be produced first for the above reaction to occur. The
reaction rate R (deprotonation rate) is therefore directly proportional to the gas phase
number density of deprotonated glycerol, assuming that the reaction is first order With
respect to b0th M and (G-H)‘ (g).
R = k [M(g)] [(G-H)’(g)l
where k is the rate constant for the reaction. The formation of deprotonated glycerol in the
gas phase can be assumed to be from the direct bombardment from high energy Xe atoms
(3 keV):
G(g) + Xe ------ > (G+H)+ (g) + H‘(g) + Xe
or G(g) + Xe ------ > (G-H)‘ (g) + H+(g) + Xe
or 2G(g) + Xe ------ > (G-H)‘ (g) + (G+H)+(g) + Xe

The above are simplified equations for protonation and deprotonation of glycerol
matrix molecules. The dominant ions such as (G+H)+ (g) and (G-H)’ (g) in a mass
spectrum are the major evidence for the above reactions.

During fast atom bombardment (protonation and deprotonation) processes, glycerol
molecules are excited first by the high energy Xe atoms, and a small portion of their
translational energy is converted into the internal energy of the neutral glycerol molecules,
leading to the excitation of the glycerol molecules. The excited glycerol molecules then
undergo ionization via reactions such as those shown above.

The ions formed during such processes may still be in excited states and they can
undergo further degradation reactions. This statement can be partly supported by the ions

formed via dehydration of protonated or deprotonated peptides.

 

181

(G-H)‘ (g) (m/z 91) ------ > (G-H-H20)' (g) (m/z 73) + H20
(G+H)+(g) (m/z 93) ------ > (G+H-H20)+(g) (m/z 75) + H20
(G+H-H20)+(g) (m/z 75) ...... > (0+H-2H20)+(g) (m/z 57) + H20

With protonated glycerol, loss of an electron will produce multiply charged ions,
which is not favored. Other pathways will produce positive ions which can act as
protonation reagents. For deprotonated glycerol molecules, there is another important
pathway which produces neutrals and electrons ( can not be detected by a regular mass

spectrometer) due to the very small eleCtron affinity value:

(HOCH2)2CH-O' (g) ----> (HOCH2)2CH-O- (g) + e' (g) AH = 1.86 eV

Due to the small activation energy needed for such processes {EA[(CH3)2CH-O‘)] = 1.86
eV = 40 kcal/mol}, this process is expected to be at least one of the major pathways for
deprotonated glycerol. Note that the electron detachment reaction is expected to be much
faster than other fragmentation reactions because there is no geometry requirement.
Because of this additional pathway for deprOtonated glycerol, the concentration (or number
density) of deprotonated glycerol in the gas phase would be much lower than that of
protonated glycerol.

If the rate constants for deprotonation and deprotonation of peptides are the same,
and considering that the concentration of (G-H)‘ (g) is much less than that of (G+H)+ (g)
in the gas phase, it is obvious that the rate of deprotonation of peptides is much less than
that of protonation of peptides. Therefore, the number density of (M+H)+(g) is much
higher than that of (M-H)‘ (g), and the sensitivity of positive mode is better than that of

negative mode.

 

182

The stability of deprotonated peptides: The poor Stability of deprotonated
peptides is anOther reason why the sensitivity of the negative mode in FAB is not as good
as that of the positive mode. Protonated peptides are unlikely to be deprotonated
(becoming neutral) due to the large pr0ton-affinity values. Since an ion is positively
charged, loss of an electron forming doubly charged ions, is also not favored. Other
pathways such as loss of sidechains and hydrogen atoms, will produce fragments which
can be detected.

With deprotonated peptides, the ions are not as stable as positive ions. One major
pathway responsible for the instability is the easy loss of an electron and become

electrically neutral species, which can not be detected by a mass spectrometer:
-~COO' ----- > ~~~COO° ( neutral radical) + e’ AH = 3.4 eV

For this process, the energy required is only 3.4 eV (79 kcal/mol, from the data of acetic
acid and alanine). With deprotonation at Other sites, this number can be much lower. To
cleave a peptide skeletal bond (homolytic cleavage), the lowest energy required is 79
kcal/mol (for CHR-CO bonds). Therefore, if the energy available can break a bond leading
to fragmentation, then the same amount of energy can also remove an electron from a
negatively charged species (the above equation). Removal of an electron could be favored
because loss of an electron requires no arrangement of any atom and the steric hindrance is
not a factor. Therefore, such processes directly lead to the loss of negative ion sensitivity
in a regular FAB mass spectrometry.

In CID experiments, the average internal energy deposition is determined to be
about 3 eV with a high energy tail above 15 eV [28]. Therefore, immediately following
collision activation, a large portion of the excited ions will undergo electron stripping and

form neutrals, which is directly related to the poor fragment ion currents for negative ions.

 

E‘i

183

2. Why are peak clusters formed?

The appearance of CID spectra of deprotonated peptides is generally not as good as
that of protonated peptides. In positive mode, a peak representing an ion is accompanied
by an isotopic peak. In negative mode, peak clusters for sequence ions are most frequently
observed. For example, a (Z-H)‘ ion is frequently accompanied by a (Z-ZH)’ ion.
Sidechain loss of leucine can lead to loss of 57 and 58, etc. In many cases, the formation
and the nature of ions in peak clusters cannot be explained. It should be pointed out that
the formation of clusters lowers the sensitivity of the technique because ion currents are
distributed among more species.

One explanation can be suggested based on the CID spectra of di- and tri- peptides
in Bowie’s work [1-9]. It has been shown that loss of a hydrogen atom from di- and tri-
peptides is very common, especially when the precursor ion current is high. This loss of
hydrogen radical lead to new series of new chemistry such as loss of H2 molecule, or loss
of another side chain radical to stabilize the radical formed. The easy loss of hydrogen
radical is because the new species formed, also a radical, can be stabilized by radical
delocalization.

When a hydrogen atom is removed from a negative charged peptide at the
deprotonation site, the radical ion can be stabilized due to resonance structures (both the
charge and the radical are stabilized).

-~(C°)R-COO‘ < ----- > ~~~(C-)R-C(O')=O and

-~(C-)R-COO' < ----- > ~~~CR=C(O°)O‘
Therefore, the product from loss of a hydrogen atom of CHR at a peptide chain is relatively
stable.

{In positive mode, there is no similar situation. For example, if the N-terminus of

a peptide is protonated.
H3N+(C-)RCO~~~ < ----- > H3N+-C(R)=C(o.)~-

184

Obviously, the charge cannot be delocalized by resonance structures. In such cases,

H3N+-C(R)=C(O-)-~ may be the dominant species because the positive charge will make

the double bond stay closer to the charge site. J
X. Conclusions

Preformation of (M-H)‘ (l) in the liquid phase does not enhance the formation of
(M-I-l)‘ (g). Therefore, deprotonation of peptides occurs in the gas phase. Deprotonation
of amides is favored over that of (rt-carbon positions because the former provide favorable
intermediate leading to deprotonation.

Ions from CID of larger peptides are found to be different from what have been
reported for smaller peptides such as di- and tri-peptides due to the smaller internal-energy
deposition during a collision and the long life time of (M-H)‘ ions. For di- and tri-
peptides, the major sequence ions have been reported to be Y ions. With larger peptides, C
ions are found to be another important ion series, which are as abundant as Y ions in a C11)
spectrum. V

In CID spectra of deprotonated peptides, the situation becomes more complicated
than that in the case for protonated peptides, especially for the high—mass ions. When a
high-mass peak appears in a spectrum, it usually is not a single peak with its isotopic peak,
instead a cluster of peaks is seen, which apparently cannot be explained simply by isotopic
contributions. Because of the complexity, our efforts can only be put in explaining the
formation of major ion series.

By using deuteration, derivatization and charge-localization, we have concluded that
there is significant deprotonation at amide sites, which has n0t been considered as
important sites of deprotonation previously by other groups. Also found is that
deprotonation at the C-terminus is not as prevalent as peOple thought. These conclusions

are important because they are the starting point for peptide fragmentation.

in
'1

 

185

For most of the C ions observed, their formation does not involve the shift of a-
hydrogen atoms. However, a small portion of C ions are formed via shifting of et-
hydrogen atoms.

For the formation of Y ions in dipeptides, it has been reported that an a—hydrogen
shift is involved. With larger peptides such as hexa-peptides, we find that a-hydrogen
shifts do not occur, indicating totally different mechanisms.

With peptides of various deprotonation sites (derivatized and charge-localized
peptides, we found that both Y ions and C ions are formed mainly from amide
deprotonated peptides and the fragmentation processes are mainly charge-induced. Based
on these conclusions, mechanisms for the formation of C ions and Y ions are proposed.

With charge-localized peptides, ions are formed via charge-remote fragmentation
mechanisms. With the formation of Y ions, a hydrogen has to be shifted to the C—terminal
side of the cleavage site. The formation of (Z-l—l)‘ ions is simply from homolytic cleavage
of NH—CHR bond.

CID-B/E linked scan spectra of deprotonated peptides can provide useful
information for peptide sequencing. The major limitation of peptide sequencing with this
technique is the‘relatively low sensitivity due to the facile loss of electron from
deprotonated molecules. In addition, the easy loss of hydrogen atoms leads to the

formation of clusters of peaks, which may complicate the interpretation of mass spectra.

 

References

l. Waugh, R. J. ; Bowie, J. H. Rapid Commun. Mass Spectrom. 1994, 8, 169-173.

2. Echersley, M.; Bowie, J. H.; Hayes, R. N. Org. Mass Spectrom. 1989, 24, 597-
602.

3. Waugh, R.J.; Bowie, J. H.; Hayes, R. N. Int. J. Mass Spectrom. Ion Processes
1991, 107, 333-347.

4. Waugh, R. J.; Bowie, J. H.; Hayes, R. N. Org. Mass Spectrom. 1990, 26, 250-
256.

5. Waugh, R. J.; Ekersley, M.; Bowie, J. H.; Hayes, R. N. Int. J. Mass Spectrom.
Ion Processes 1990, 98, 135-145.

6. Waugh, R. J.; Bowie, J. H.; Gross, M. L. Rapid Commun. Mass Spectrom. 1993,
7, 623-625.

7. Waugh, R. J.; Bowie, J. H.; Gross, M. L. Austral. J. Chem. 1993, 46, 693.

8. Waugh, R. J.; Bowie, J. H.; Gross, M. L. Int. J. Mass Spectrom. Ion Processes
1994, 133, 165-174.

9. Barber, M.; Bordoli, R. S.; Sedgwick, R. D.; Tetler, L. W. Org. Mass Spectrom.
1981, 16, 256.

10. Raftery, M. J.; Bowie, J. H. Int. J. Mass Spectrom. Ions Processes 1988, 85, 167.

11. Knapp, D. R. "Chemical Derivatization for Mass Spectrometry" in Methods in
Enzymology, vol. 93, McCloskey, J. A., ed., Academic Press: San Diego, 1990;
pp314—328.

12. Sehgal, D.; Vijay, l. K. Anal. Biochem. 1994, 218, 87-91.

13. Lias, S. G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Levin, R. D.; Marlard,
W. G. J. Phys. Chem. Ref. Data, 1988, 17, Suppl. No. 1.

14. Roepstorff , P.; Fohlman, J. Biomed. Mass Spectrom. 1984, 11, 601.

15. Huaiqin Wu and John Allison, J. Am. Soc. Mass Spectrom. 1994, 5, 564-475.

186

16.

17.

18.

19.

20.

21.

22.

23.
24.

25.

26.

27.

187

Renner, D.; Spiteller, G.; Biomed. Environ. Mass Spectrom. 1986, 13, 405.
Johnson, Martin, Biemann, Int J. Mass Spectrom. Ion Processes 1987, 86, 137.
Kidwell, D. A.; Ross M. M.; Colton, R. J. J. Am. Chem. Soc. 1984, 106, 2219-
Vath, J. E.; Biemann, K. Int. J. Mass Spectrom. Ion Proc. 1990, 100, 187.
Wagner, D.S.; Salari, A.; Gage, D. A.; Leykam, J.; Fetter, J.; l-lollingsworth, R.;
Watson, J. T. Biological Mass Spectrom. 1991, 20, 419-425.

Zaia, J.; Biemann, K. 41th Am. Soc. Mass Spectrom. Allied Topics 1993, San
Francisco, pp358.

Griffiths, W. J.; Zhang, J.; Sjovall, Rapid Commun. Mass Spectrom. 1993, 7, 235-
240.

Wagner, D. 8. PhD Dissertation, Michigan State University, 1992.

Jessen, N. J.; Tomer, K. B.; Gross, M. L. J. Am. Chem. Soc. 1985, 107, 1863-
1868.

Roepstorff, P.; Richter, W. J. Int. J. Mass Spectrom. Ion Processes 1992, 118/119,
789-809.

Stults, J. T. Peptide Sequencing by Mass Spectrometry in "Biomedical Application
of Mass Spectrometry, vol. 34", Suelter, C. H. & Watson, J. T., ed.; Wiley: New
York; 1990.

Biemann, K.; Martin, S. A. Mass Spectrom. Rev.l987, 6, 1-76.

APPENDICES

' 1

APPENDIX I

Local Pressure at the Sample Holder in the Ion Source
and

Collision Frequency In K+IDS Experiments

I. Introduction

As has been seen from the previous chapters, K“’IDS-MS can be viewed as a
chemical ionization technique with K+ ions as the reagent ions. The reagent ions are
generated by a thermionic emitter. They react with thermally desorbed intact analyte
molecules and their thermal degradation products forming K+-adducts. Because the adduct
formation may be a second order reaction, the collision frequency between K+ and
desorbed species will be a direct indication of sensitivity. Since collision frequency, a
function of source pressure and K+ flux, is a function of pressure, let's first consider the
ion-source pressure during the thermal desorption process.

The pressure in the ion source under K+IDS condition has been assumed to be high
during early development of this technique. It was believed that the high pressure to a great
extent stabilized the adduct ions after their formation.

However, there are some obvious observations which indicate low pressure in a
K+IDS ion source, especially the local pressure at the sample holder. 1. There are no
competitive reactions observed for K+ attachment to mixtures. For example, in studying
relative responses of various amino acids, ion chromatograms from all amino-acid are

complete and symmetric, indicating that there is no equilibrium - no high collision

188

189

frequency - existing in K+IDS ion source. 2. The sensitivity of K+IDS technique is
relatively low, even less sensitive than that of chemical ionization (CI). This suggests that
there might be a low pressure in the ion source which leads to a low collision frequency,
thus low sensitivity. 3. An assumed high pressure should lead to a high collision
frequency, thus high adduct formation rates, high stabilization effects and thus high
sensitivity of the technique. Unfortunately, this is not true, indicating that the high-
pressure assumption may not be true.

The questions are: What is the local pressure at the sample holder in a typical
K+IDS experiment? What is the encounter frequency between K+ ions and desorbed
neutral molecules? After adduct formation, what is the average collision number between
any adduct ion formed and desorbed neutral species? If the average collision number is
less than one, what is the percentage of K+ ions that undergo at least one collision?

Since the collision frequency is highly dependent on the pressure in the ion source
(especially the small volume around the sample holder), a rough estimation of the ion
source pressure or number densities for both K+ ions and the neutral molecules will be
necessary. Because the separation between the sample holder and K+ emitter is 1 mm, a

spherical volume with r = 1m is considered in this study.
11. Calculation of the local pressure at sample holder in the ion source

To find out the local pressure around the sample holder in the ion source, two
methods for estimation are used. The consistency from these two methods suggests that

the results are reliable.

A. Estimation Method #1
Considering a space 1 mm around the sample holder. If a gas is confined in this

volume (V = 1.3337tr3 = 4 mm3) and the volume has a imaginary surface A (A = 47tr2 =

190

10 m2), from which gas can diffuse to a lower pressure region (assuming 0 Pascal for
the ion source).

Typical ion source conditions in K+IDS are as follows:

Exit area: A = 4m2 = 4x3.l4xl mm2 = 0.1 cm2 = 10-5 m2

Volume considered (around bead): V = 1.3337tr3 = 4 mm3 = 4x10—9 m3)

Molar mass: 200 g/mol Sample load: 20 pg = 10'7 mol
Desorption time: 10 sec K+ emission current : 10 LIA
Average desorption rate: ZOug/ 108 = 10‘8 moles/s.
m=M/Nav=3.3x10‘25Kg T=3OOK

Source Chamber Pressure: 10‘4 Pa

If there is no diffusion from the imaginary surface from which gases diffuse, the
pressure in the confined volume will build up to the following after one second of

heating/desorption according to the ideal gas equation:
P = nRT/V = (10-8 mol x 8.3141 K'lmol'l x 300K )/4x10'9 m3 = 6000 Pa

however, the diffusion (the source pressure is only 10—4 Pa, which can be considered as 0;

thus diffusion) of gas with a pressure of 6000 Pa from the imaginary surface cannot keep

this high pressure at all, the gas flux Jn through the surface can be calculated as follows

[1]:

1;; = P/(thka)1/2
= 6000 Pa/ (2 x 3.14 x 3.3 x 10‘25 Kg x 1.38 x 10'23 JK‘1 x 300 K)”2

= 6 x 1025 s-1 m-2

191

The diffusion rate of analyte in the gas phase through the imaginary surface at a

pressure of 6000 Pa:

1,, x A / Nay = 6 x1025 5'1 m'2 x 10'5 m2 / 6.022x1023 mol‘1

= 1 x10‘3 mol s-1

It is easy to see that the diffusion rate (assuming that the pressure is maintained at
6000 Pa) is much faster than the desorption rate.

During desorption process, a steady-state desorption and diffusion can be assumed.
Thus the diffusion rate must be equal to that of the desorption rate 10'8 mol/sec ( steady-
state approximation). Since the diffusion rate in mol/s is directly proportional to ion source
pressure, the corresponding diffusion pressure with which steady-state condition is
maintained is:

6000 Pa x (10-8/10-3) = 0.06 Pa.

B. Estimation Method #2

The throughput of gases through the exit of the ion source can be calculated as
follows [2]:

_ 4132; l/2_2_1/(r-1)
Q—AP1C(-fn— r+l) ( )

r+l (valid for szP1 5 (2/(r+1)r/r-1)

where Q is the gas flow through an orifice, A is the area of the orifice (a imaginary surface
1 mm around the sample holder), P1 is the pressure inside the confined volume, P2 is the
pressure in the ion source (ion source chamber), C’ is a correction factor for the orifice area
(0.85 for thin orifice), r is the ratio of specific heat . The specific heat can be roughly
estimated as r = cp/cv = (2n+3)/(2n+l), where n is the number of atoms in a molecule. For
small peptide molecules with molecular weights of 200, n is about 30 (for peptides, 11 z

MW”). and r is calculated to be about 1.024.

192

For r = 1.024, [2/(r+1)]llr/(r-1)l = 0.608. It is apparent that the P2/P1 « 0.608
since the pressure outside the ion source is below 10'4 Pa. So the above equation is valid
for the calculation of the flow rate at the imaginary surface 1 mm around the sample holder.

Given P2 = 10-4 Pa, T = 600 K, r = 1.024, A = 0.1cm2 =10-5 m2

Q = 10:5 m2 x P1 Pa x {(8.314JK-1mol-1x600Kx2x1.024)/[.2Kg/molx2.204]}1/2

x [(2/(1+1.024)]1/(1.024-1) x 0.85
105 m2 x P1 Pa x 158.9m/s x 0.608 x 0.85
8.2 x10‘4 P1 (Pa m3 s-1 )

Given r= 1.0328, Q = 8.1 x 104 P1 (Pa m3 s-1 )

Given r = 1.01 (n=100), Q = 8.2 x 10-4 (Pa m3 8'1 )
this clearly shows that r hardly affects the flow rate when it is very small.

The desorption rate is about 10'8 mol/sec, which is corresponding to

APV = AnRT =10-3 mol s-1 x 8.314] K-1 mol-1 x 600 K = 5 x 10'5 J 8'1
= 5 x10'5 Pa m3 s-l.

Obviously, to maintain the equilibrium, the diffusion rate (flow rate through the exit)
should be equal to the desorption rate:

8.2 x10-4 P1 m2 s-1 = 5 x10‘5 Pa m2 s4.

P1 = 0.06 Pa
This means that the analyte pressure 1 mm around the emitter tip is kept at 0.06 Pa, which
is consistent from the value estimated by method #1. Obviously, these numbers are not as

high as many of us had imagined.

C. Collision diameter: The size of peptide molecules
To estimate the encounter frequency between K+ ions and neutral molecules, the
collision diameters of both peptide molecules and K+ ions have to be estimated because the

mean free path for analyte molecules can be calculated as follows:

A.,, = kT/(ZmrtdaZP)

193

The problem is the effective collision diameter has to be figured out.

To calculate the collision diameter, the molecules can be treated as a ball in the gas
phase. Suppose the density of peptide is 0.8 g/cm3, and the peptide is in the primitive
cubic crystal form, with MW = 200 g/mol.

Every one cm3 contains moles of peptides: 0.8 g/(200 g/mole) = 4x10'3 moles

# of molecules in one cm3 is 4x10-3 moles x 6.022x1023/mole

Each peptide molecule takes a volume of (1.3337tr3)/0.524

1 cm3 = 10-6 m3 = 4 x 10—3 moles x 6.022 x 1023/mole x (1.333m3) / 0.524

r = 6.6 x 10-10 m

dag: 2r = 1.32 x10'9 m

Am = kT/(Zl/zrtdaaZP)
1.38x10-23 J K-1 x 300 K / [1.414 x 3.14 x (1.32x10‘9 m)2 x P]

= 7.6X1O‘4 /P m (J m’2 Pa‘1 = m)
A.“ = 0.76/P mm with P in Pascal
rrr= 600k is used, 7.3, =1.5/P mm

D. The mean free path of K+ in the ion source
The K‘l' emission rate in the ion source is 10 uA = 10'5 A = 10'10 moles/sec
(note 1 mole charge = 96487 C as 105 C), which is only 1% of the desorption rate. The

mean free path for K+ ion in the analyte molecule gas can be calculated as‘follows:

xk= 1

v2 1:
[2"2nn1dfi+(1+—;)l’zn,z(dk+d,)2]
Vk
where nk is number density of K+ ion in the gas phase, na is the number density of gas-

phase analyte molecules; vk is the velocity of K+ ions, va is the velocity of analyte

molecules in the gas phase , d3 is the diameter of analyte molecules, dk is the diameter of

194

K+ ions. 7~k is the mean free path of K+ ions in the gas-phase analyte molecules. Since v2

is inversely proportional to the mass of a molecule, the above equation can be rewritten as

1

k: m
[21’2nn1dfi+(1+—m:)”2n,£(dk+d,)2]

 

since na » nk, d32 » dkz, the above equation can be simplified as following:

1
AR: m
[(1+fitfnn,%(dk+d,)2]

 

at this low pressure, the peptide in the gas phase may be treated as ideal gas, thus:

1“: ml: 1/2litT 2
[(113719 I(dk+d8) P]
Given mg =39, ma = 200, dk 2: 10%da, plug all the numbers in the equation, then

11: = 9.0 /P m
at the pressure of 0.06 Pa, 11: = 9.0/P mm = 150 mm.

For the ion source used, the source center (El source) to the exit is about 5 mm, which is
much less than the mean free path of K+ ions.

The low pressure of analyte and the low K+ emission rate is the key factors for the
low sensitivity of this K+IDS technique.

But why do we see adduct ions?

Do not misunderstand the mean free path, mean free path greater than the ion
source dimension does not mean no collision! In fact, kinetic theory describes the
distribution of free paths as follows:

N=Nem
where N' is the number of molecules in the volume and N is the number of molecules that

traverse the distance x before suffering a collision.

195

If the source dimension x = 10%1 , then N/N' = 90%, 10% of the molecules or ions
traversing the ion source will undergo a collision. If x = 1/5 1, N/N' = 82%, 18% of

molecules or ions will have a collision before getting out of the ion source.

 

XI! 3% 10% 1/5 2/5 1 5
%collision 3 10 18 33 63 99.7

 

so, even though the source dimension is only 3% (5 mm/ 150 mm) of the mean free path of
K+ ions, there are still 3% of K+ ions collide with analyte molecules before leaving the ion
source. Because of the low encounter frequency between K"' and neutral species, the
concentration of (M+K)+ is low; thus the possibility of three-body collisions - the
collisions between (M+K)+ ions and neutral species - is small. Therefore, three-body
collisions are unlikely to occur in K+IDS experiments.

Consider the above situation. 3% of the K+ ions under collisions with neutral
species is a reasonably high collision frequency. If each collision leads to the formation of
one adduct ion, then the sensitivity can be very high. Assume ionization efficiency of 100%
in FAB, the signal lasts 5 minutes, then it is only as sensitive as K+IDS. In fact, the FAB
responses of peptides are much higher than K+IDS responses. This may suggest that the

detachment of K+ ions is very fast, which leads to poor sensitivity.

E. Comparison with Chemical Ionization (C1)

In CI, there are several factors changed. First, the most significant change is that
the protonated molecules are much more stable than that of K+ adducts of analyte
molecules, Second, the collision diameter of CH5+ is increased by a factor of 2, Third,
loss of reagent ion CH5+ and analyte molecules is also reduced because the collisions with

the high-pressure reagent gas will keep them in the ion source for a much longer time.

196

kT

mCH‘ H2?! 2
[(1+——mCH:) 3(dcug+dc11) P]

 

XCHXCH‘) =

assume dc115+ = dCH4 = 5A, T = 300K, the result is 1CH5+(CH4) = 4/P mm.

The mean free path for collisions between CH5+ and analyte molecules are:

kT
m 0
[(1%)1/2%(dm3+ d,)2 P]

assume dCH5+ = 5A, d3 = 13 A, T = 200 K, MW = 200, 1CH5+(a) = 1.6/P mm.

 

A'CH;(a) =

In a CI source, the pressure is usually set at 1 torr (133 Pa), the mean free path of

the reagent ion with reagent gas is l = 4/(133 Pa) = 0.03 mm , which means that CH5+
will undergo hundreds of collisions with CH4 or analyte molecules in the ion source. This
high collision frequency ensures that the protonation process has a very high efficiency.
The frequent collisions also stabilizes the protonated molecules due to thermalization,
which results in little fragmentation of the protonated molecules.

The frequency of collisions between reagent ions and analyte molecules is
determined by the number densities of both. If the sample quantity is the same as given in
K+1DS, the pressure can be kept much higher since the exit is now much smaller, plus the
diffusion of the reagent gas is much faster than the analyte molecule. The diffusion rate
ratio is (16/200)"2 =- 1/3, so the reagent gas diffuses 3 times as fast as the analyte
molecules out of the ion source.

Note here the collisions between analyte molecules and reagent ion such as CH5+ is
much less frequent than collisions between neutral molecules because of the much lower
concentrations of both.

So what is the concentration of CH5+ like? Suppose filament current 300 11A and
electron energy 300 eV, each electron ionized two neutral species. 300 11A = 3x10'4 A =1
3x1079 moles/sec since 1 mole of charge = 105 C. If this concentration can build up, the

pressure can be as high as 10 Pa in the small ion source. If each collision between a C115+

197

and an analyte molecule forms a protonated analyte, the chemical ionization rate will be the

collision rate between CH5+ and analyte molecules.

There is a locally high concentration of CH5+ along the electron path, which will

result in high ionization yield.

References

1. Alberty, R. A.; Sibley, R. J. Physical Chemistry, lst ed.; Wiley & Sons: New
York, 1992; pp609.

2. O'Hanlon, J. F. A User's Guide to Vacuum Technology, Wiley & Sons: New York,
1989;pp29.

Appendix II
Enthalpies of Thermal Degradation of Peptides from Model Compounds

1. Model Compounds:

Direct calculation or estimation of reaction enthalpies for thermal degradation of
peptides are impossible at this time due to the lack of thermodynamic data of peptides. In
order to estimate these data for the cleavages of peptide skeletal bonds (AX, BY and CZ
bonds based on Roepstoff nomenclature) under various mechanisms, model compounds
are selected to simulate these processes.

The three skeletal bonds in peptides are simulated with the following models:

I. CH3CO-NHCH3 AX bond
11. CH3CH2-COOH; BY bond
NHzCHz-COOH;
CH3CH2-COCH3
III. CH3NH-CH2CH3 CZ bond

The reaction enthalpies of these model compounds, based on the following data

available, are used as the reaction enthalpies of peptides.

Table 1. Bond Dissociation Energies (BDE, kcal/mol)

BDE(~CO-NH~) 93.6 amide bond
BDE(H-NHR) 101 Primary N -H bond.

198

199

BDE(1-I-NHR1R2) 91.5 Secondary N -H bond.
BDE(H-i-C3H7) 95.9 secondary C-H bond.
BDE(H-t-C4H9) 93.3 tertiary C-H bond.

2. Estimation and Optimization of Bond Dissociation Energies

A. Modified Estimation of Amide N-H Bond Strength:

Consider the two N-H bonds in a peptide such as H-NHCHzCON(-H)CH2~~~,

which N-H bond is stronger ?
BDE(CH3NH-H) = 100
BDE((CH3)2N-H) --' 91.5

How does -CO- affect the bond strength?
BDE(CH30—H) = 104.4

BDE(CH3COO-H) = 105.8, increased by 1.4 kcal/mol.

If this is true for amide/amide, then
BDE((CH3)2N-H) = 91.5
BDE((CH3CON(CH3)-H) = 91.5+1.4 = 93

Note here that the trend for C-I-l bond is different:
BDE(CH3CH2-H) =100.3
BDE(CH3COCH2-H)=98.3 decreased by 2 kcal/mol.

B. Estimation of C-0 and C-N bond strengths

CH3COOCH3 (g) ------ > CH3COO- + ~CH3
-98.5 -49.6 34.8

AH” = 83.7kcal/mol

This bond is very similar to that formed in the (Z-H) fragment when cyclization occurs

from the C-terminus.

200

C2H50-C2H5(g) ------- > CszO- + ~C2H5 AH' = 84.2kcal/mol
60.3 -4.1 28
The small difference indicates that the carbonyl group near the O does not affect the bond

strength of O-C bond. In the same consideration, it is assumed that the carbonyl group

near N will not affect the bond strength of NC bond.

CH3NH-C2H5(g) -------- > CH3N1-1- + °C2H5 AH“ = 81.4 kcal/mol

-1 1 42.4 28
BDE (CH3CONH-C2H5)= BDE(CH3NH-C2H5) = 81.4 kcal/mol

The small affect of carbonyl group can also be seen from the following example.

CH3-CH2CHO ----- > CH3 + CH2CHO
-49 34.8 3.0 i 2.0 BDE = 86.8 kcal/mol.
CH3-CH2CH3 ----- > CH3 + CH2CH3
-24.8 34.8 28 BDE = 87.6 kcal/mol.

3. Enthalpies for Different Thermal-Degradation Pathways of Peptides

Model molecule: CH3CO-NHCH3

Model bond: -CHRCO-NHCHR'-, the BY bond.
CH3CO-NHCH3 ------ > CH3CO- + oNHCH3 AH' = 93.6 kcal/mol
-56 -6 43.6
CH3CO-NHCH3 ------ > CH3COH + NH=CH2 AH“ = 48.4 kcal/mol
-56 -39.6 32
CH3CO-NHCH3 ------ > CH2=C=O + NHzCH3 AH' = 39.1 kcal/mol

-56 -l 1.4 -5.5

201

Cyclization at BY bonds: for example, 1, 6 - Elimination:

P ------ > cyclo-(B-H) + (Y+H) AH“ = 93 -101 = -8 if cyclization from an amide.

One secondary NH bond is broken and one primary N-H bond is formed.

P ------ > cyclo-(B-H) + (Y+H) AH: 101-101 = 0 if cyclization from the N-

terminus. One primary N-H bond and an amide bond are broken while one primary N-H

bond and one amide bond are formed.

Model Molecules: NH2CH2-COOH, CH3-COCH3, CH3-COOH

Model bond: -NHCHR-CONH-, the AX bond.

NH2CH2COOH ------ > NH=CH2 + HCOOH
-93 32 -90.5
NH2CH2COOH ------ > NHCHz- + ~COOH
-93 38 -53.3
CH3CH2COCH3 ------ > CH3CH2-+ ~COCH3
-57 .5 28 -6
CH3CH2COOH ------ > CH3CH2- + ~COOH
-107 28 -53.3
CH3COOH ------ > CH3- + oCOOH

-103.3 34.8 53.3
CH3COCH3 ------ > CH3CO- + ~CH3

-52 -6 -34.8
CH3CONH2 ------ > CH4 + =C=N-CH3

-56 -l7.8 -31

AH' = 34.5 kcal/mol

AH“ = 78 kcal/mol

AH‘ = 79.5 kcal/mol

AH’ = 82 kcal/mol

AH' = 85 kcal/mol

AH' = 81 kcal/mol

AH' = 7 kcal/mol

Unusually low

202

Cyclization at AX bonds: P ------ > (A+H) + Cyclo-(X-H) AH: 93.3 - 96 = -3

because one secondary OH is converted into a tertiary C-H bond.

Model Molecule: CH3NH-CH2CH3
Model bond: -CHRNH-COCHR'-, the CZ bond.

CH3NH-CH2CH3(g) ------ > CH3NH2 + CH2=CH2 AH“ = 18 kcallmol

-11 —5.5 12.5

CH3NH—CH2CH3 ------ > CH2=NH + CH3CH3 AH“ = 23 kcallmol
-11 32 -20

CH3NH-CH2CH3 ------ > CH3NHo+ ~CH2CH3 AH“ = 81.4 kcallmol
-11 42.4 28

Cyclization: There are two possibilities.

P ------ > (C+H) + Cyclo-(Z-H) AH“ = 93-101== -8 kcallmol if cyclization is not
from the C—terminus. A primary N-H is formed while a secondary N -H bond is broken.

P ------ > (C+H) + Cyclo—(Z-H) AH“ = 81.4 (C-N) -84.3 (00) + 105.8 (O-H)
-101(N-H)== 2 kcallmol if cyclization is from the C-terminus. A O-H bond and N-C bond
are broken while a primary N-H bond and a C-O bond are formed.

4. Data Rearranged in Types of Reactions

Homolytic cleavage:

B-Y; CH3CO-NHCH3 ------ > CH3CO~ + ~NHCH3 AH“ = 93.6 kcallmol
A-X: NH2CH2COOH ------ > NH2CH2- + °COOH AH“ = 78 kcal/mol

 

203

C-Z: CH3NH-CH2CH3 ------ > CH3NH°+°CH2CH3 AH“=81Kcal/mol

1,2 - Elimination:

B-Y: CH3CO-NHCH3 ------ > CH3COH + NH=CH2 AH“ = 48.4 kcal/mol
CH3CO-NHCH3 ------ > CH2=C=O + NHzCH3 AH“ = 39.1 kcallmol

A-X: NH2CH2COOH ------ > N H=CH2 + HCOOH AH“ = 34.5 kcal/mol
CH3CONHCH3 ------ > CH4+ O=C=N-CH3 AH“ = 7 kcal/mol

C-Z: CH3NH-CH2CH3 ------ > CH3NH2 + CH2=CH2 AH“ = 18 kcal/mol
CH3NH-CH2CH3 ------ > CH2=NH + CH3CH3 AH“ = 23 kcal/mol

Cyclization: Cyclization involves bond changes from secondary amine to primary amine,

from secondary carbon to tertiary carbond, ect.

P ----> (A+H) + Cyclo-(X-H) AH“ = = -3 kcal/mol.
P ----> (C+H) + Cyclo-(Z-H) AH“ = -8 kcallmol if cyclization from an amide.
P ----> (C+H) + Cyclo-(Z-H) AH“ = 8 kcal/mol if cyclization from the C-terminus.

P ----> cyclo-(B-H) + (Y+H) AH“: ~8 kcallmol if cyclization not from the N -terminus.

P ----> cyclo-(B-H) + (Y+H) AH“ = 0 kcal/mol if cyclization from the N-terminus.

Reference

1. CRC Handbook of Chemistry and Physics, 71th ed.

"Illllllllllllllll'lllllll