IL 7
1.3.2.! :2 vii
.(

v :5? .
“$4.14.:
1: ..

."II? 1.."

«3.. b

e..-
£va 5

. 4r.:.|...tk\.
. L I

I..-
[0413‘ 0’5
1| Icl

.I v.lnl.r.

. . 15!. .3 rt
. In. I)

II v‘rlniyoa
.L 4. , 1.3.1.! 1h .- : Au. . . , , ,

:2EZF ll?! 3.2.5....- :. :II 2. , ‘

A: 7| n. !.c3. .

‘ . . , fifimag
.. I. 5&9? ‘ . . 5.. . . .v-.. .. v . , .dJl-ifi.).(..1 ..... . .
Av.” jmfiiJ’un-vlzzu .I’I .I" .

 

.
. v

 

    

‘ snowmangipbw :
3|: 1‘» 31‘: ‘ .u. a onto!
.04 o { .

‘ w l. 'r
" l

 
 

     

 

      

5’ [II I
l '

 

FE

fi.
~

 

“3 STATE WNERSITY UBRARES

l2! Ill/lI/llll/I/l/Ill/ll!{Ill/IIHIIII/ll

 

ll
‘1!

 

\
E‘—
‘—
x
\—

1’ HI I

. 1 93 00794 4956

w

This is to certify that the

thesis entitled

INVESTIGATIONS OF TRIPHENYLPHOSPHONIUM
DERIVATIZATION OF PEPTIDES FOR IMPROVED ANALYSIS
BY FAST ATOM BOMBARDMENT

MASS SPECTROMETRY

presented by

Timothy Jay Nieuwenhuis

has been accepted towards fulfillment
of the requirements for

Chemistry

 

M.S ' degree in

 

 

Major rofessor _
P ofessor 0 Chemistry

and Biochemistry

Date October 2, 1992

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

LIBRARY
Michigan State
University

 

 

cMma-m

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
cWMS-e:

 

 

INVESTIGATIONS OF TRIPHENYLPHOSPHONIUM
DERIVATIZATION OF PEPTIDES FOR IMPROVED ANALYSIS
BY FAST ATOM BOMBARDMENT
MASS SPECTROMETRY

By

Timothy Jay Nieuwenhuis

A THESIS

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

MASTER OF SCIENCE

Department of Chemistry

1992

ABSTRACT

INVESTIGATIONS OF TRIPHENYLPHOSPHONIUM DERIVATIZATION
OF PEPTIDES FOR IMPROVED ANALYSIS
BY FAST ATOM BOMBARDMENT
MASS SPECTROMETRY

By

Timothy Jay Nieuwenhuis

Fast atom bombardment (FAB) is a popular ionization technique in
mass spectrometry (MS) because this method can desorb highly polar and
thermally labile analytes dissolved in a liquid matrix. Because the surface
activity of analyte in the hydrophilic matrices used for FAB affects their
ionization emciency, derivatization techniques to increase the hydrophobic
character of peptides are investigated. Reagents to selectively attach the
triphenylphosphonium (TPP) group to either terminus of a peptide are
synthesized or are commercially available. The hydrophobicity of the TPP
group increases the ionization efficiency of the derivatized peptides, and its
localized charge directs fragmentation from the terminus of derivatization.
The FAB-MS response of leucine enkephalin (YGGFL) modified with
various TPP-containing reagents is examined and the fragmentation
pattern of derivatized VGVAPG is analyzed by FAB-CAD-B/E.
Investigation of a series of chemical reactions proposed for the ethyl-TPP
derivatization of peptides in a solid phase reactor indicates that this
reaction scheme is feasible. Procedures for the isolation and future TPP-

derivatization of glycopeptides are described.

Dedicated to my parents

Henry and Arlene Nieuwenhuis

iii

ACKNOWLEDGMENTS

First, I wish to thank my co-advisors Dr. Rawle Hollingsworth and
Dr. J. Throck Watson for ideas, insights, support, and guidance that
were crucial to the completion my degree. I am grateful to Drs. John
Allison and J efl' Ledford for their time and advice as members of my
guidance committee. I acknowledge Dr. Z.-H. Huang for input on the
project simulating the solid phase derivatization of peptides and Dr. J.
Leykam for assistance and the use of equipment necessary for
glycopeptide isolation. The helpful advice of Drs. David Wagner and
Yoon-Seok Chang also is appreciated.

I would be remiss if I did not acknowledge the members of the
Watson group. I thank Kuruppu Dharmasiri, Jennifer Johnson, Jian
Wang, Ken Roth, and Ben Gardner for their moral support and
laboratory assistance. Thanks also are in order to the Mass Spectrometry
Facility gang; the assistance of Dr. Doug Gage, Bev Chamberlin, Mike
Davenport, and Mel Micke is appreciated heartily.

I am endebted and grateful to my parents for providing me the
opportunity to pursue a college education. The love and support and of ,
my dear friend, Carrie Blok, entitle you to share in my achievement. I
thank my siblings, Rob and Alyce, for the family activities and relaxation
that made my studies at MSU enjoyable.

Finally, but most importantly, "I thank Christ Jesus our Lord, who
hath helped me." (I Tim. 1:12 KJV).

iv

TABLE OF CONTENTS

List of Tables ............................................................................... xi
List of Figures ............................................................................ xii
Chapter]. InhoductionandOlfiectivee

A. Introduction ................................................................ .1

B. Fast atom bombardment mass spectrometry
1. Historical background ............................................... 2
2. FAB-MS of underivatized peptides and CAD ................ 3
3. Fragmentation of underivatized peptides by FAB-CAD-

M S / MS .................................................................... 5
4. Liabilities of FAB ionization and CAD-MS / MS ............ 11

C. Derivatization in conjunction with FAB-MS analysis

1. Derivatization schemes of other investigators ............. 14
2. Strategies of peptide derivatization ............................ 15
a) Enhancement of surface activity ........................... 16
b) Peptide degradation ............................................ 18
i) Hydrolysis .................................................... 18
ii) Methanolysis ................................................ 18
c) Modification of fragmentation .............................. Z) '
i) Fixed charge ................................................ Z)

ii) Adduct formation .......................................... 24

iii) Dialkylphosphorylation .................................. 25

(1) Determination of functional groups ...................... %

i) Acetylation ................................................... %

ii) Esterification ................................................ %

iii) Amino acid detection ..................................... 27

e) Miscellaneous .................................................... so

D. Research objectives ...................................................... 33
E. References .................................................................. 34

Chapter II. 'h'iphemdphosphonium Derivatization and the
SeparationofComponemaoftheReacfionMixhne

A. Introduction ................................................................ 39
B. Synthesis of reagents containing the TPP group
1. (2-Bromoethyl)TPP
a) Synthesis by Method I
i) Experimental ............................................... 41
ii) Results and discussion .................................. 41
b) Thin-layer chromatography
i) Experimental ............................................... Q
ii) Results and discussion .................................. 45
c) Flash column chromatography
i) Experimental ............................................... 47

ii) Results and discussion .................................. 47

vi

d) Synthesis by Method II
i) Experimental ............................................... 48
ii) Results and discussion .................................. 49
2. (2-Aminoethyl)TPP
a) Experimental ..................................................... 49
b) Results and discussion ........................................ 49
3. (4-Carboxybenzyl)TPP
a) Experimental ..................................................... 52
b) Structural investigations
1') Analysis by FAB-MS ...................................... 55
ii) Analysis by NMR .......................................... 55
. Peptide derivatization procedures
1. Bromoalkyl-TPP reagents and vinyl-TPP
a) Experimental ..................................................... (i)
b) Results and discussion
i) (2-Bromoethyl)TPP and vinyl-TPP ................... E)
ii) (3-Bromopropyl)TPP and (4-bromobutyl)TPP ..... 65

2. The “Kunz” reagent

a) Introduction ....................................................... 68
b) Experimental ..................................................... w
c) Results and discussion ........................................ w

3. (4-Carboxybenzyl)TPP
a) Experimental ..................................................... 71

b) Results and discussion ........................................ 72

vii

5.

(2-Aminoethyl)TPP

a) Experimental ..................................................... 74
b) Results and discussion ........................................ 74
Summary of derivatizing reagents ............................ 7'7

. Separation of the C-terminal derivatization mixture

1. Experimental .......................................................... 78
2. Results and discussion ............................................ 79

. Quantitation of the C-terminal reaction yield .................. 82
1. Determination of the absorption coefficient

a) Experimental ..................................................... 83
b) Results and discussion ........................................ 83
Direct quantitation of the reaction yield

a) Experimental ..................................................... 84

b) Results and discussion ........................................ 84

. FAB-CAD-B/E fragmentation of derivatized VGVAPG

1.
2.

Introduction ........................................................... 85

Mass spectrometry

a) Instrumentation ................................................. 87
b) Operating principles ........................................... 87
Analysis of underivatized VGVAPG

by FAB-CAD-B/E. .................................................... 89
(2-Bromoethyl)TPP and vinyl-TPP

derivatized VGVAPG .............................................. 92
(2-Aminoethyl)TPP derivatized VGVAPG .................. 94
(3-Bromopropyl)TPP derivatized VGVAPG ................. %

viii

7. (4-Bromobutyl)TPP derivatized VGVAPG ................... %

8. ‘Kunz’ reagent derivatized VGVAPG ........................ Q
9. (4-Carboxybenzyl)TPP derivatized VGVAPG ............. 101
G. References ................................................................. 103
Chapterlll. DevelopmentofOn-line Whiphmwlphosplmnium
A. Introduction .............................................................. 105
B. Advantages of an on-line derivatization reactor ............. 107
C. Proposed setup ........................................................... 109
D. Chemical methodology ................................................ 111
E. Separation of a peptide mixture by LC-MS
1. Experimental ........................................................ 113
2. Results and discussion ........................................... 114
F. Reaction in homogeneous solution ............................... 114
1. Formation of the ‘Kunz” reagent o-nitmphenolate
a) Experimental .................................................... 119
b) Results and discussion ...................................... 119
2. Peptide derivatization
a) Experimental .................................................... 123
b) Results and discussion ...................................... 124
G. Conclusions ............................................................... 132
H. Future work .............................................................. 132
I. References ................................................................. 133

ix

Chaptu'N. FutureDezivatizaflonofaGb'copepfide

A. Introduction .............................................................. 135
B. Procedures for glycopeptide isolation ............................ 137
C. Glycopeptide analysis ................................................. 139
D. The analysis of glycoproteins by FAB-MS ...................... 140
E. Derivatization of carbohydrates

and analysis by FAB-MS ............................................. 142
F. Advantages of ethyl-TPP derivatization

of a glycopeptide ......................................................... 144
G. References ................................................................. 145

Table2J.
Table2.2.

LIST OF TABLES

Investigated derivatizing containing the TPP group.

Ratio of signal intensity of derivatized leucine
enkephalin to that of underivatized, protonated peptide
contained in the reaction mixture during analysis by
FAB-MS.

xi

FigureLL

FigumLZ.

Scheme Ll.

FigureL3.

FigureL4.

Figure 2.].

LIST OF FIGURES

Representation of the processes occurring with FAB.
Protonated (M-I-H)+ and deprotonated (M-H)'
molecules of analyte are produced as well as many
neutrals, M, and ions of glycerol, G. If a species
exists as a “preformed ion”, it is desorbed as (M)+.

Structures of the common ions encountered in FAB-
CAD-MS/MS and their nomenclature as revised by
Biemann. Adapted from reference 1 2.

Fragmentation of the peptide backbone and
nomenclature proposed by Roepstorfl' and Fohlmann
[11].

Example of a FAB-CAD-B/E product ion spectrum
displaying a complete series of y. fragment ions and
an overlapping h, series, which permits complete

amino acid sequence determination of the
pentapeptide Y-G-G-F-M.

FAB-CAD-BIE product ion spectrum of the
heptapeptide R-V-Y-V-H-P-F that does not provide a
complete or overlapping series of peaks for complete
amino acid sequence elucidation.

The FAB-MS spectrum of the (2-bromoethyl)TPP
product synthesized by reaction of 1 ,2-dibromoethane
with triphenylphosphine. The molecular ion (M)+ of
(2-bromoethyl)TPP is represented by the doublet at m/z
369 and 371; the vinyl-TPP and (2-
diphenylphosphinoethyl)TPP impurities appear at
m/z 289 and m/z 475, respectively. Nitrobenzyl alcohol
(NBA) was used as the FAB matrix.

xii

13

Human

Figure 2.4.

HPLC chromatogram of the (2-bromoethyl)TPP
product mixture synthesized by reaction of TPP with
1 ,2-dibromoethane. For this separation,
approximately 100 nmol of product mixture was
injected. The parameters used in the separation are:
mobile phase, 45:55 (v/v) acetonitrile/water; flow rate,
1 mL/min; column, 4.6 x 250 mm; stationary phase, 5
pm diameter, C13 bonded phase; detector, 214 nm, 1.0
AUFS; chart speed, 20 cm/hr.

The FAB-MS spectra of the components of the (2-
bromo-ethyl)TPP reaction mixture prepared by the
method of Fredrich and Henning as separated by
flash column chromatography using a 14:100 (v/v)
methanol/chloroform mobile phase: a) (2-
methoxyethyl) TPP corresponding to the lower TLC
spot and last-eluting compound when separated by
flash column chromatography. b) vinyl-TPP and (2-
bromoethyl)TPP that were the first-eluting
compounds during separation by flash column
chromatography and was represented by the upper
spot after separation by TLC . NBA was the FAB
matrix used for both spectra.

HPLC separation of the (2-bromoethyl)TPP product
synthes-ized by reaction of (2-hydroxyethyl)TPP with
PBr3. Approximately 10 nmol of product was injected
for separation. The parameters used in the
separation are: mobile phase, 45:55 (v/v)
acetonitrile/water; flow rate, 1 mL/min; column, 4.6 x
250 mm; stationary phase, 5 pm diameter, C13 bonded
phase; detector, 214 nm, 0.5 AUFS; chart speed, 20
cm/hr.

Analysis of the product produced by reaction of (2-
hydroxy-ethyl)TPP with PBr3 by FAB-MS. The peaks
representing the molecular ion (M+) of (2-
bromoethyl)TPP dominate the FAB-MS spectrum (m/z
369, 371). Vinyl-TPP impurity appears at m/z 289.
The FAB matrix was NBA.

xiii

51

F‘isumZ-B.

Figure 2.7.

W23.

Figure 2.9.

Figure2.10.

Figure 2.11.

The chromatogram resulting from the separation of
the (2-aminoethyl)TPP product by HPLC. The first on-
scale peak represents the (2-aminoethyl)TPP product;
the second peak represents vinyl-TPP impurity in the
product. For this separation, approximately 200 nmol
of product was injected and the parameters used in
the separation are: mobile phase, 45:55 (v/v)
acetonitrile/water; flow rate, 2 mL/min; column, 10 x
250 mm; stationary phase, 5 pm diameter, 013 bonded
phase; detector, 214 nm, 0.5 AUFS; chart speed, 20
cm/ hr.

The spectrum resulting from the analysis of the (2-
aminoethyl) TPP product mixture by FAB-MS. The
molecular ion (M+) of (2—aminoethyl)TPP dominates
the spectrum (m/z 306), but some vinyl-TPP is
observed due to impurity and fragmentation (m/z 289).
The peaks at m/z 136 and 154 are characteristic of the
NBA matrix used.

Analysis of the (4-carboxybenzyl)TPP product by FAB-
MS that was produced by reaction of TPP with 4-
(chloromethyl)ibenzoic acid. The peak representing
the molecular ion (M+) of the (4-carboxybenzyl)TPP
product is at m/z 397. The base peak in this spectrum
represents protonated triphenylphosphine oxide and
the peak at m/z 201 represents diphenylphosphine
oxide. The FAB matrix was glycerol.

 

The spectrum resulting from analysis of the (4-
carboxybenzyl) TPP product by FAB-CAD-B/E. The
fragmentation observed is relawd to the structure of
the product in the inset. The FAB matrix was NBA.

Analysis of the (4-carboxybenzyl)TPP product by NMR
when dissolved in a) deuterated chloroform and in b)
d4-methanol.

The FAB-MS spectrum of 1 nmol of YGGFL N-
terminal derivatization mixture. Leucine enkephalin
was reacted with a 20-fold excess of a) vinyl-TPP and
b) (2-bromoethyl)TPP in an acetonitrile/pyridine 1:1
(v/v) reaction mixture. A higher response resulting
from the molecular ion (M+) of derivatized peptide
(m/z 844) was observed for reaction with vinyl-TPP
than (2-bromoethyl)TPP.

xiv

Figurem

Figure2.13.

Figure2.l4.

Figure2.l5.

Figure 2.16.

The FAB-MS spectrum corresponding to 1 nmol of
YGGFL reacted in a potassium phosphate/sodium
borate buffer solution (pH 9.0) with a 20-fold molar
excess of a) vinyl-TPP and b) (2-bromoethyl)TPP. The
spectrum represents 1 nmol of YGGFL reaction
mixture as isolated by the Sep-Pak® procedure. The
molecular ion of derivatized peptide (M+) appears at
m/z 844, and protonated underivatized peptide at m/z
556.

The FAB-MS spectra representing 1 nmol of YGGFL
reacted with a) (3-bromopropyl)TPP and b) (4-
bromobutyl)TPP. The reaction was performed with a
20-fold molar excess of reagent in an
acetonitrile/pyridine 1:1 (v/v) solvent mixture. The
molecular ion (M+) of propyl-TPP derivatized YGGFL
appears at m/z 858 and butyl-TPP at m/z 872.

The FAB-MS spectrum respresenting 1 nmol of
YGGFL reacted with a 20-fold molar excess of “Kunz’
reagent in an acetonitrile/pyridine solvent mixture.
The peak representing PEOC derivatized peptide (M+)
appears at m/z 888, decarboxylated, ethyl-TPP
derivatized peptide is at m/z 844, and underivatized
peptide appears at m/z 556.

The FAB-MS response of 1 nmol of YGGFL when
reacted with a 20-fold molar excess of (4-
carboxybenzyl)TPP and EDC in an aqueous solution
(pH 5.0) buffered with TFA. The peak representing
the molecular ion (M+) of derivatized YGGFL appears
at m/z 934, the protonated, underivatized peptide
appears at m/z 556.

The FAB-MS spectrum representing 1 nmol of
YGGFL reacted with a 20-fold molar excess of (2-
aminoethyl)TPP and EDC in an aqueous solution (pH
5.0). The peak representing derivatized peptide (M+)
appears at m/z 843 and underivatized peptide is at m/z
556.

XV

76

Figure 2.17.

Figure2.18.

Figure 2.19.

Figurem

Figure2.2l.

Flgure2.22.

Figure2.23.

Figure2.24.

Figume2.25.

HPLC separation of 20 nmol of the C-terminal
derivatization mixture of YGGFL. The reaction
mixture contained a twenty-fold excess of (2-
aminoethyl)TPP and EDC. The compounds that the
peaks represent are identified in the chromatogram.
The conditions of the separation are: flow rate
2mL/min; column, 10 x 250 mm; stationary phase, 10
um diameter, 013 bonded phase; detector, 214 nm, 2.0
AUFS; chart speed, 30 cm/hr.

Analysis of the peak representing separated, C-
terminally derivatized YGGFL by FAB-MS. The
molecular ion (M+) of derivatized YGGFL dominates
the spectrum (m/z 843), although signal representing
vinyl-TPP and (2-aminoethyl)TPP (m/z 306) is
observable.

The FAB-CAD-B/E spectrum of the molecular ion
(M+) of C-terminally derivatized YGGFL with (2-
aminoethyl)TPP appearing at m/z 843.

The FAB-CAD-B/E spectrum of underivatized
VGVAPG. The protonated molecule (M+H)+ is at m/z
499.

The FAB-CAD-B/E spectra of the molecular ion (M+)
of N -terminally, ethyl-TPP derivatized VGVAPG by a)
vinyl-TPP and b) (2-bromoethyl)TPP (m/z 787).

The FAB-CAD-B/E spectrum of the molecular ion
(M‘i') of C-terminally, ethyl-TPP derivatized VGVAPG
as modified by (2-aminoethleI‘PP (m/z 786).

The FAB-CAD-B/E spectrum of the molecular ion
(M+) of N-terminally, propyl-TPP derivatized
VGVAPG after derivatization with (3-
bromopropyl)TPP (m/z 801 ).

The FAB-CAD-B/E spectrum of the molecular ion
(M+) of VGVAPG modified at the N-terminus with a

butyl-TPP group as yielded from reaction with (4-
bromobutyl)TPP (m/z 815).

The FAB-CAD-B/E spectrum of the molecular ion
(M+) of N-terminally derivatized VGVAPG as
derivatized with the “Kunz” reagent to yield PEOC
derivatized peptide (m/z 831 ).

xvi

100

mm

Figure 3.1.

Scheme 3.1.

Figure3.2.

Figure 3.4.

Scheme3.2.

Figure 3.5.

The FAB-CAD-B/E spectrum of the molecular ion
(M+) of N-terminally modified VGVAPG as modified
by the (4-carboxybenzyl)TPP group (m/z 877).

Proposed on-line, ethyl-TPP derivatization setup. The
most effective design probably will perform solid
phase, precolumn derivatization prior to LC
separation and analysis by FAB-MS

Chemical reaction scheme for solid phase, ethyl-TPP.

derivatization. ‘Kunz" reagent (1) reacts with
immobilized o-nitrophenol to form the “Kunz”-o-
nitrophenolate (II). Peptide reacts with the covalently
bound reagent, derivatizing the peptide. The column
will be regenerated with the ‘Kunz’ reagent.

UV—Vis trace of the four-component peptide mixture
containing 100 pmol of VGVAPG, GLA, YGGFL, and
VQAADYING following separation with a microbore
column, prior to analysis by frit-FAB-MS.

Reconstructed total ion current (TIC) plot (a) obtained
when the peptide mixture entered the mass
spectrometer. Scanning was commenced when the
first-running peptide was detected by the UV-Vis
detector positioned prior to the mass spectrometer.
Figure 3.3 (b) contains the reconstructed mass
chromatograms of the individual components of the
peptide mixture.

The background-subtracted mass spectra of
individual peptides (a—d) eluted during LC-MS. The
scan rate was one scan every 10 seconds.

Homogenous-phase chemistry simulating the on-line
reaction sequence. Nonmobilized o-nitrophenol reacts
with the ‘Kunz’ reagent (I) to produce the “Kunz”-o-
nitrophenolate (II). This product (II), in turn,
derivatizes the peptide to yield the “Kunz” modified
peptide (III).

FAB-MS spectrum of reaction mixture resulting from
mixing a five-fold molar excess of o-nitrophenol and
0.5 uL 2,4,6-trimethylpyridine with 1 mg of “Kunz”
reagent for 10 sec. (b) Linked-scan spectrum at
constant B/E of the o-nitrophenolate of the “Kunz”
reagent.

xvii

102

110

112

115

116

117

118

12)

Fianna”.

Figure 3.7.

Scheme 3.3.

Figure 3.8.

Figure 3.10.

Investigation to determine the kinetics of the reaction
yielding the o-nitrophenolate of the “Km” reagent.
Reaction yield was defined as the ratio of peak
intensity at m/z 472 (11, Scheme 3.2) divided by peak
intensity at m/z 307 [(2-hydroxyethyl)TPP]. These
results show the product forms quickly, but is
unstable in a solution with water-containing
atmosphere above it.

Reaction of 1.25 nmol/uL AAA methyl ester with a
theoretical 25-fold molar exess of the “Kunz”-o-
nitrophenolate. The following spectra represent 1 1.1L
of the reaction medium containing 1.25 nmol of AAA
methyl ester after 32 minutes of reaction (a) at room
temperature, (b) at 37°C, and (c) in the presence of
TEA.

Suggested reaction pathway for decarboxylation of the
peptide derivatized with the “Kunz” reagent to form
an ethyl-TPP derivatized product.

FAB-CAD-B/E spectrum (a) of 1 nmol of AAA methyl
ester derivatized with vinyl-TPP and (b) FAB-CAD-B/E
spectrum of 1 nmol of AAA methyl ester derivatized
with the “Kunz”-o-nitrophenolate.

Spectra of the reaction mixtures theoretically
containing 1 nmol of AAA methyl ester after
derivatization: (a) ~8 hrs in a pyridine/acetonitrile (1 :1
v/v) solvent mixture at 37°C and a 20-fold molar excess
of vinyl-TPP, (b) 32 min in acetonitrile/TEA with a
approximate 25-fold molar excess of “Kunz”-o-
nitrophenolate.

 

Investigation of the speed of peptide derivatization
with the “Kunz”-o-nitrophenolate. The yield of
reaction was determined by the ratio of
decarboxylated, ethyl-TPP, derivatized AAA methyl
ester (m/z 534) to underivatized and protonated
AAAmethyl ester (m/z 246) peak intensity. This
reaction was performed at room temperature, 37°C,
and in the presence of TEA, as the legend indicates.
For these studies , 1.25 nmol of AAA methyl ester was
reacted with a 25-fold excess of the “Kunz”-o-
nitrophenolate.

xviii

122

125

1%

128

129

13)

Figure 3.11. (a) FAB-MS of reaction mixture containing 1.25 nmol
of VGVAPG after reaction 32 min with a 25-fold
molar excess of “Kunz”-o-nitrophenolate with 0.5 uL
of TEA present as a base catalyst. Decarboxylated,
ethyl-TPP, derivatized product appears at m/z 871,
while protonated, underivatized peptide appears at
m/z 499. (b) FAB-CAD-B/E spectrum of 1.25 nmol of
decarboxylated, derivatized VGVAPG.

xix

131

Chapter I
INTRODUCTION AND OBJECTIVES

A. Introduction

Although mass spectrometrists have endeavored to elucidate the
structure of biopolymers for over thirty years, the most significant
contributions of mass spectrometry (MS) to the analysis of important,
biological compounds have occurred in the past decade. Two major
developments in instrumentation revolutionized the analysis of biopolymers
by MS: i) high-field magnetic mass spectrometers and ii) fast atom
bombardment (FAB) ionization. These advances made it possible to obtain
mass spectra of intact polypeptides, oligosaccharides, oligonucleotides, and
other small biopolymers without the need for prior derivatization or
degradation [1]. However, sample derivatization also can be utilized to
augment the amount of information available from analysis by FAB-MS.

At this time, FAB-MS is the most popular of the desorption
ionization techniques and has proved to be a powerful and cost-effective
method for the structural determination of natural and recombinant
peptides [2]. The tandem FAB-MS technique has become the method of
choice to confirm the amino acid sequence of proteins predicted by DNA
sequencing [3]. Structural modifications of a biopolymer (posttranslational
modifications or alteration of amino acids) can be identified by FAB-MS, as
these cannot be deduced from DNA analysis. Even more importantly, FAB-
MS can analyze peptide mixtures without costly purification procedures.
This is an aspect foreign to Edman degradation.

1

2

There are many schemes for biopolymer modification that serve to
complement their characterization by FAB-MS. Many peptides, especially
those of higher molecular weight, show poor or incomplete sequence data,
even with the use of collisionally-activated dissociation (CAD) in
combination with tandem mass spectrometry (MS/MS). This necessitates
the use of chemical reagents to modify specific amino acid side chains, or
either terminus, to aid in identifying the amino acid sequence of a peptide.
Peptide modification is used to increase peptide ionization emciency and
direct fragmentation to simplify the elucidation of the amino acid sequence.

The objectives of this first chapter are to i) introduce the FAB
ionization and CAD techniques utilized for determining the amino acid
sequence of underivatized peptides, ii) discuss the strategies, procedures,
and efi‘ectiveness of chemical derivatization methods employed to enhance
the analysis of peptide biopolymers by FAB, and iii) detail the objectives
and criteria used for evaluating the derivatization schemes investigated in
this research project.
B. Fast atom bombardment mass spectrometry

1. Historical background

Before the introduction of FAB-MS by Barber et al., [4,5] chemical
derivatization was necessary to increase peptide volatility and thermal
stability for sequencing by gas chromatography-mass spectrometry (GC-
MS). In the late 19508, Biemann investigated polyamino alcohols derived
from peptides with chemical acetylation and reduction by LiAlH4 for GO-
MS analysis [6]. The derivatized peptides displayed improved volatility,
but also the specific fragmentation of specific bonds in the peptide
backbone. Later, this derivatization procedure was modified to yield N-
trifluoroethyl-O-trimethylsilyl polyamino alcohols. Peptides modified to N-

3

acetyl-N,O-permethylated derivatives were introduced into the mass
spectrometer by a direct probe for electron ionization [7]. Even with these
peptide modification schemes, mass spectrometry contributed little to
protein primary structure elucidation in the 1 9608 and 1 9708, however.
2. FAB-MS of underivatized peptides and CAD

During FAB-MS, a beam of xenon or argon atoms (or Xe'i' ions for
liquid secondary ionization mass spectrometry, LSIMS) is focused on a
viscous, liquid matrix (commonly glycerol) to desorb and ionize highly polar,
thermally labile analyte molecules dissolved in the matrix (Figure 1.1).
This "soft" ionization technique has expanded the scope of mass
spectrometric analysis to include a large number of biologically important
compounds. The secondary ion beam produced is long-lasting and stable,
which is another advantage of FAB ionization. Intact, high mass molecules
are desorbed into the gas phase. Little excess energy is imparted to the
molecule with FAB ionization when an ionic species is produced by the
addition or removal of a proton to form stable, even-electron ions, (M+H)+,
or (M-H)‘. These ions have a low tendency to fragment, which is useful for
the unambiguous determination of molecular weights, but is a liability if
fragmentation is desired for structural elucidation.

When FAB-MS was first implemented, researchers already realized
its potential for sequence determination of high-mass peptides [8].
However, further investigation showed FAB usually does not give enough
energy to the peptide to cause sufficient fragmentation for complete
structural determination. Tandem mass spectrometry (MS/MS) is the most
commonly used technique to increase the internal energy of a peptide and
induce fragmentation [9]. With MS/MS, the precursor ion fragments when
undergoing high energ CAD with a collision gas (e.g., helium, 10'3 torr).

Fast atom
beam

\5 §3 "Selvedge"

 

 

 

Figure 1.1. Representation of the processes occurring with FAB.
Protonated (M+H)+ and deprotonated (M-H)' molecules of analyte are
produced as well as many neutrals, M, and ions of glycerol, G. If a species
exists as a “preformed ion”, it is desorbed a8 (M)+.

S
Peptide fragmentation is thought to occur at bonds near and remote to the
site of protonation [10]. The degree of charge localization and the site of
protonation influences the relative intensity and structure of ions produced
with CAD fragmentation. With FAB ionization, the site of protonation
usually is at basic amino acid side chains; consequently, the positive charge
tends to localize on lysine or arginine residues rather than on the less basic
amide nitrogens and oxygens on the peptide backbone.
3. Fragmentation of underivatized peptides by FAB-CAD—MS / MS

Positive ion FAB-MS has sensitivity advantages for peptide analysis
because the basic nature of the polyamide backbone causes protonated
molecules (M+H)+ to form more readily than deprotonated molecules (M-
H)‘. The repeating units of a peptide biopolymer contain the twenty
common amino acids, which have a repeating structure —HN—CH(R)—CO—,
where R is the side chain characteristic of each common amino acid. When
the peptide bonds are cleaved with CAD after protonation, charge retention
on the carbonyl group generates the b" ion series (Figure 1.2). The mass
difference of the fragment ions reveals the identity of consecutive amino
acids with the exception of isometric leucine and isoleucine or isobaric
glutamine and lysine. Another set of peaks representing y" product ions is
due to a second fragmentation pathway involving the transfer of a hydrogen
atom from a nitrogen atom on the N-terminal side of the amide bond and
proton retention at the C-terminal portion of the peptide after amide bond
cleavage. Ifno basic amino acids are present in the peptide, mainly hp. and
some yn ions occur in the mass spectrum.

The first nomenclature for peptide fragmentation (Scheme 1.1) was
proposed by Roepstorfi' and Fohlmann [11]. Since then, the nomenclature

has undergone revision; Biemann labeled fragment ions by lower-case

N -terminal Ions

+
0.: H—(NH—CHR—CO),-,—NH=CHR,
or
Ho CknuRnb

H—iNH—CHR—Coi,-.—NH—CH

 

H. R,
a, + l: H—‘(NH——CHR-—c0)‘.-.——NH—CH-
1...: H—(NH—CHR—CO).-n—NH—CHR.—CEO‘

 

HQ
c,: H-‘(NH—CHR—CO)‘.—NH3

 

b
H. CR.

4,: H—INH—CHR—coi.-.—NH—cn

 

C-terminal Ion Types
H.
u,: HN=CH—co—iNH—CHR—c07._.—0H
CR,‘ H.
w,: CH—CO—iNH—CHR-COIrr—OH

x,: +OEC—NH—«CHR,,—CO---(NH—CHR—--CO),.. |—OH
or

H 0
o=C=N—'CHR.—CO—(NH—CHR—C(R.-.—0H
H?
y,: H-‘(NH—CHR—Coirou

 

 

 

 

 

H+
y, - 2: 'HN=CR.—co—(NH—CHR—coi,-.—0H
CK‘K. H4

 

z.: CH—co—iNH—CHR—coi.-.—ou

H4
2, + I: -CHR.—co-«'NH—CHR—cofi._.—0H

 

Figure 1.2. Structures of the common ions encountered in FAB-CAD-
MS/MS and their nomenclature as revised by Biemann. Adapted from
reference 12.

XaYszéYzzleYIZ.
F— F l— l— T T l—
T (Ii lie u 1‘3 (ll li‘ ii
HzN—(ll-"C"N"(|3- c- N--(l3- c- N--C—C—OH
H H H H H H H

 

 

 

 

 

 

 

 

 

_J .i__._i_iJ
B. A.B..A.B.

9

fl

Scheme 1.1. Fragmentation of the peptide backbone and nomenclature
proposed by Roepstorfi‘ and Fohlmann [11].

8
letters to differentiate them from immonium ions, which are identified with
upper-case letters. The structure of the common ions formed with CAD of
peptides are outlined in Figure 1 .2 [12].

Interpreting mass spectra usually requires more effort than
acquiring the actual data. Deducing the amino acid sequence from the
mass spectrum is made easier if other information about the peptide is
known; the source of the sample (tryptic fragments contain lysine or
arginine at the C-terminus), possible homology with other peptides, or
amino acid composition aid in interpreting the mass spectrum. Ideally, the
amino acid sequence of a peptide is determined by a complete series (e.g. ,
bu) or the overlap of two ion series, one derived from the N-terminal
sequence of amino acids (an, bu, or cn) and the other derived from the C-
terminal sequence of amino acids (x... y", or z").

The FAB-CAD-B/E mass spectrum outlined in Figure 1.3 from 1
nanomole of a peptide is an example of two overlapping ion series
unmistakably indicating the amino acid sequence of a peptide. Examining
the low-mass end of the spectrum, immonium ions are observed for
phenylalanine (m/z 120) and tyrosine (m/z 136), providing a partial amino
acid composition. Examination of the high-mass region reveals peaks
corresponding to the loss of water and the side chains of methionine,
phenylalanine, and tyrosine residues. To deduce the amino acid sequence,
the mass difference between (M+H)+ and a fragment ion that corresponds to
the loss of mass characteristic of an amino acid residue would represent the
yn.1 ion, n being the number of amino acids in the peptide. The peak at m/z
411 represents the loss of 163a from (M+H)+, the mass of a tyrosine
residue. Therefore, the tyrosine residue resides at the N-terminus. The

next reasonable loss in the y—series is to m/z 354 (57u), which indicates that

 

 

cum-

+8.36

.zacds
cosmonflaoa. .05 we 55835233 85:53 can Sada 33988 Egon
£033 .958 an $233196 do can meow unofimaum .3» me 858 32988

a mafia—9mm“. 83.8on :3 853.5 Em-Q<O-m<m a me 298an .mA 952%

 

 

 

 

 

 

 

 

 

 

 

as
com cowq com com 9:
«o «a
v0 «N «a no ah
b have ma
2.. «Q dd h
4m
as
n v m a H .é
‘8.“ as "so
«5 ER «Na in
«an a .c
w- a Ham 8a 82 a: .e
2- er a r.
I as.
9: Ba «Sc a. ”em
4
n H N a v m n:
*6 .

 

 

MmfimuM>m Hdawdmwuh

lO

glycine is the amino acid adjacent to tyrosine in the N -terminal series.
Another loss of 57u, represented by the peak at m/z 297, indicates that
another glycine is next in the series. A reasonable loss represented by the
y“ peak is 147u and this corresponds to a phenylalanine residue, bringing
us to the peak at m/z 150. Because two hydrogens are added to the y,
fragment ion, the final residue at the C-terminus i8 methionine. The y, ion
series indicates the peptide sequence is YGGFM.

This sequence can be confirmed by the ion series generated by
cleavage at the peptide bond and retention of charge at the N—terminus
(bu). The fragment ion represented by m/z 556 corresponds to the loss of
water from the (M+H)+. The peak at m/z 425 represents loss of 131u from
(M+H-H20)+, which corresponds to a methionine residue. Neutral losses of
147u (to m/z 278) and 57u (to m/z 221) indicate the sequence from the C-
terminus is GFM, confirming the N-terminal sequence. The an ion series
represented by peaks at m/z 193, 250, and 397 confirms the assignment of
the 1),. ion series. Informative fragmentation in this example reveals the
primary structure of the peptide as YGGFM, named methionine
enkephalin.

However, most peptides do not fragment completely because the
amide groups are not the most basic functionalities in the peptide. For
example, the amine group of lysine and and guanidine group of arginine are
more basic than amide groups and the CAD spectra of peptides containing
these residues are changed dramatically. Peptides obtained from tryptic
digestion, which contain a basic arginine or lysine residue at the C-
terminus, display a series of ions (v... w", x", y", 2,.) retaining the charged,
C-terminal residue, which effectively competes with amide groups for the
proton in (M+H)+ [13]. Wn ions are products of y" ions, resulting from

11

cleavage of the B,y-bond of the amino acid residue. The formation of w"
ions complicates the mass spectrum and decreases overall signal intensity,
but these unique ions difi‘erentiate leucine and isoleucine; the wn ion of
leucine is always 14 mass units lower than that of isoleucine [14]. Wn ions
are not produced from glycine, the aromatic amino acids, or alanine;
therefore, wn ions rarely represent a continuous series. Y" and w" ion
series predominate the FAB-CAD-MS/MS of tryptic peptide fragments with
a basic C-terminal amino acid, producing simple, informative spectra.

When basic residues are present near the N-terminus of a peptide,
the proton is retained at the N-terminus and series of fragment ions
retaining the charge at the N-terminus (an, bu, cu, (in) dominate the mass
spectrum. When the charge is retained at the protonated guanidino group
of arginine, an ions are produced by charge-remote fragmentation, and (1;.
ions result from cleavage of the 8,)‘bond of the amino acid side chain.
These ion series (an, (1,.) confirm the assignment of peaks caused by amide
bond cleavage. The relative abundance of an, bu, cu, x", y", and 2,, ions
indicates the position of basic residues along the peptide chain.

4. Liabilities of FAB ionization and CAD-MS / MS

A serious liability of FAB-MS is the dependence of peptide ionization
efficiency on its surface activity in the liquid matrix. Low surface activity
causes poor sensitivity for hydrophilic peptides and discrimination against
hydrophilic peptides in the analysis of peptide mixtures by FAB-MS. Often
the poor ionization efficiency of hydrophilic peptides prevents their
detection in the mass spectrum. This is a handicap for peptide mapping,
because it is desirable to obtain a peak of discernible intensity in the mass
spectrum with this technique, regardless of the degree of hydrophilicity of
the peptide. Although the lack of peaks representing key fragments with

1 2
analysis by FAB-CAD-MS/MS is not restricted to hydrophilic peptides, the
low signal abundance from such peptides can increase the difficulties of
deducing their complete amino acid sequence.

FAB-CAD-MS/MS relies on the capacity of the protonated peptide to
fragment eficiently and informatively during FAB ionization or with CAD,
prior to analysis by MS/MS. Unfortunately, not all underivatized peptides
fragment completely and the product ions required for structural
elucidation may be absent or in low abundance. Because the major factor
directing the fragmentation of underivatized peptides with FAB-CAD-
MS/MS is the site of proton attachment, the presence of a basic residue in
the center of a peptide chain complicates mass spectral interpretation. In
the absence of strongly basic residues, a mixture of N- and C-terminal ions
are observed. These ion series often are incomplete and result in
ambiguities when deducing the amino acid sequence. Immonium ions and
ions resulting from internal cleavage of the peptide further add to the
complexity of the mass spectrum, although they are beneficial for the
identification of amino acids near the N-terminus. Another liability of
MS/MS is that the efficiency of fragmentation caused by CAD decreases
with increasing molecular weight. Therefore, the complete amino acid
sequence is diflicult to obtain from the FAB-CAD-MS/MS of a peptide with
molecular weight above 1500-1700u.

For example, the mass spectrum in Figure 1.4 of 1 nmol of the
heptapeptide [val4]-angiotensin III (RVYVHPI) only allows partial
determination of the amino acid sequence. Peaks representing the
immonium ions of histidine (m/z 110) and tyrosine (m/z 136) are observed in
the low-mass end of the spectrum. The side chain losses of valine (m/z 874),
histidine (m/z 836), and tyrosine (m/z 810) residues from (M+H)+ are

13

.doflovmoflo 88:qu Soc 888 32988 you
$.on mo moron 98893.85 .8 32888 a 035.8 88 noon 85 m-m-H.H->.>
->.m 338888: on» no 838on com 888.8 Smddofism .v.H arm

       

 

 

 

 

 

 

 

 

 

 

38
com com OOH. com ocm OOHV com com 2:
_ , , _ o
3 m as V
\ m
r»
mm;
as
J
b o m u m N H at
86 \NE “:0
m8 BB ”en fl
5. The 8.. H8 8“ new
we a m , m by >H>| m
08. H
g g a "2%
8R :8
+8.85 H a m e a o s .2 .H

 

 

Mona.H>o HflaoflmoH-plh

14

observed at high mass. The peak at m/z 899 represents the loss of water
from the protonated peptide (m/z 917). N 0 peak in the mass spectrum
corresponds to the loss of an amino acid residue from (M+H-H20)+, which
would represent a b...1 ion. However, loss of 193u from the protonated
molecule produces an and ion (m/z 724), corresponding to the loss of a
phenylalanine residue plus CO. Subsequent losses of 97u, 137u, 99u, and
163u from the as ion indicate the C-terminal sequence is ?YVHPF.

A series of peaks at m/z 761, 662, 499, 400, and 263 representing a
complementary series of y" ions should confirm the C-terminal sequence.
However, only peaks at m/z 499, 400, and 263 are readily apparent from the
mass spectrum. Therefore, the incomplete C-terminal sequence cannot be
verified, and there is no direct information for the sequence commencing at
the N-terminus. Consequently, the complete amino acid sequence of this
peptide cannot be deduced from the mass spectrum, although the mass of
the protonated molecule and some peaks representing sequence ions are
observable in the FAB-CAD-B/E mass spectrum.

C. Derivatization in conjunction with FAB-MS analysis
1. Derivatization schemes of other investigators

Investigators [15-18] soon realized that chemical modification of an
analyte to produce the unconventional properties of nonvolatility and ionic
character enhanced its signal-to-background ratio during FAB analysis.
The “preformed” ions produced with analyte modification were thought to
be more surface active in the FAB liquid matrix than the unmodified
species. Protonation, cationization with alkali or transition metal ions, and
chemical modification to produce a “preformed” charge were three
derivatization techniques used to enhance the analysis of several large,
highly polar, and nonvolatile molecules.

15

A dramatic example of the effectiveness of analyte protonation for an
improved signal-to-background signal for desorption techniques was
demonstrated with the addition of p-toluene sulfonic acid to
glycylphenylalanine on a graphite surface [16]. Before treatment, the peak
of the protonated analyte was imperceptible from the background, but
addition of p-toluene sulfonic acid caused the (M+H)+ peak to dominate the
mass spectrum, with the intensity of (M+H)+ greatly increased relative to
the intensities of fragment ions. Chemical modification of specific
functional groups was also used for improved analysis by FAB-MS. The
carbonyl group of keto steroids was targeted with the Girard reagent to
form water-soluble hydrazones with a fixed positive charge [1 9]. Although
the protonated molecule could not be observed before derivatization,
physiological levels (1 ug/L) of androsterone, corticosterone, and
androsterone glucuronide were analyzed following derivatization.
Following derivatization, suppression of glycerol ion intensity was observed
relative to the signal of the modified species. Chemical modification with
reagents containing a “preformed” charge produced improved analysis by
FAB-MS for biomolecules, as was demonstrated with the modification of the
hydroxyl group of the steroid corticosterone with 2-fluoro-1-methyl-
pyridinium p-toluene sulfonate [1 8].

2. Strategies of peptide derivatization
Peptide modification is performed in conjunction with FAB-MS for
several reasons: i) enhancement of analyte surface activity; it) degradation
of the sample to smaller subunits; iii) enhancement of separability;
iv) modification of fragmentation: 8) enhanced formation of molecular
weight-related ions and b) enhanced formation of structurally informative

16

ions; and v) determination of functional groups. Useful derivatizations
should be specific, modifying only the intended portion of the target
molecule, and yield a single product. The modification procedure should be
fast, simple, and amenable to small sample sizes. Ideally, the
derivatization reagents should be safe and stable, but few derivatization
methods meet all of these criteria [20]. Many derivatization schemes for
peptides have been developed in the last decade. Reports of these
derivatization procedures often have appeared in the scientific literature for
known, commercially available peptides without a follow-up report on
application of the modification technique to unknown peptides isolated from
a biological matrix.

a) Enhancement of surface activity

When a peptide is placed in a drop of liquid matrix on a FAB probe
tip surface, a concentration gradient develops, resulting in the surface
enrichment of hydrophobic peptides. Inequalities in surface concentrations
are observed for different peptides with identical bulk concentrations. This
increase in peptide surface concentration or peptide adsorption at the
matrix/gas interface is termed “surface activity”. Physical and chemical
parameters such as pH, surface tension, micelle concentration, and the
hydrophilicity/hydrophobicity of the matrix, charge, size, stability, and
solubility of the peptide affect its surface activity [21]. At low bulk
concentrations of peptide, the surface peptide concentration is similar to its
bulk concentration. As the bulk concentration increases, the peptide forms
a monolayer and dislodges the glycerol molecules at the surface. Higher
concentrations of peptide cause the formation of micelles composed of
peptide. When the matrix surface is covered by a monolayer of peptide,

increasing concentrations of peptide do not affect the magnitude of the

17

signal due to the protonated peptide, (M+H)+. Consequently, there is a
direct correlation between the sputtering efficiency of protonated molecules
of hydrophobic peptides with their surface concentration, which is a
function of bulk peptide concentration.

One aim of peptide derivatization is to enhance the surface activity of
hydrophilic peptides. Many of the liquid matrices used in FAB-MS are
polar in character and adding hydrophobicity to a peptide causes it to be
drawn to the surface to desorbed preferentially. To increase the surface
activity of several dipeptides, Ligon derivatized them with an equimolar
amount of dodecanal [22]. Detectability of all the polar, derivatized
peptides increased, but the sensitivity for detection of a hydrophobic
derivatized peptide was actually less that of the underivatized peptide. Gas
chromatography and FAB-MS analysis indicated the reaction was complete
for all the investigated peptides. This suggests that the derivatization
procedure does not increase the surface activity of hydrophobic peptides.

By increasing the surface activity of all components of a peptide
mixture, chemical modification also can decrease the discrimination against
certain hydrophilic peptides that is observed with the screening of protein
digests by FAB-MS [23]. In the peptide mixtures investigated, the relative
ion abundance of some peptides was not proportional to their actual
concentration in the FAB liquid matrix. The surface activity of one peptide
could cause less surface active components in a mixture to be unobservable
[24]. Because derivatization inherently increases the homogeneity of
molecular mixtures, this technique can equalize differences in surface

activity for individual components of a peptide mixture.

1 8

b) Peptide degradation

With Edman sequencing, protein degradation is used to identify
amino acid subunits, but molecular size reductions used in conjunction with
MS are driven by instrument mass range. In the case of tandem mass
spectrometry, the need for degradation to smaller size fragments may be
motivated by molecular size limits of ion dissociation technology rather
than the analyzer mass range.

i) Hydrolysis

Kidwell et al. [25] first quaternized, then hydrolyzed, esterified, and
acylated peptides to achieve the subnanogram detection limits observed by
Cooks and coworkers in the static SIMS analysis of quaternary ammonium
salts [16]. This reaction sequence produced N-terminally charged products
with molecular weights differing by the mass of the amino acids attached to
the N -terminus after hydrolysis. Although uncharged side products result
from this procedure, they were not observed in the resulting static SIMS
spectra. This derivatization procedure produces detection limits in the low
nanogram range and compensates for the lack of fragmentation observed in
liquid FAB analysis, but deposition on solid silver is necessary and silver-
cationized amino acids were observed. Unfortunately, this derivatization
method is nonspecific because basic amino acid side chains would also be
derivatized. Another problem is that each peptide bond would be
hydrolyzed to an unequal extent in this procedure.

ii) Methanolysis

Treatment of a peptide with dry methanol containing a small amount
of hydrochloric acid (0.02-0.1 0N) converts all the free carboxyl groups into
methyl esters, but other functional groups remain intact. Reactions with
more concentrated HCl (5N, 37°C, 6 h) in dry methanol causes

19

methanolysis of peptide bonds in addition to esterification of the free
carboxylic groups. Consequently, methyl ester fragments of the original
peptide are produced by cleavage at the carboxylic end of amino acid
residues [26]. Treatment of 0.5 mg of peptide with 150 uL of 5N HCl in dry
methanol causes primary methanolysis, that is, cleavage of single peptide
bonds. The FAB spectra of the methanolysis products display an abundant
(M+H)+ ion corresponding to the completely esterified peptide. Information
about the original peptide can be obtained from the protonated peptide
methyl ester and the complementary fragment series that contain the N - or
C-terminal amino acid residues. The (M+H)+ signals for all the
methanolytic fragments are present in the mass spectra, but the intensity
of the (M+H)+ of fragments containing the esterified C-terminal residue are
much more intense than fragments containing the N-terminal residue. The
advantages of this procedure are that it is simple, fast, and does not require
separation or transfer of the sample. Only primary methanolysis is
observed under these reaction conditions and this derivatization technique
can be used in the analysis of N-blocked peptides. However, the position of
N- and C-terminal residues cannot be distinguished by this technique, and
leucine and isoleucine residues cannot be be distinguished. The
esterification of glutamic and aspartic acid side chains and methanolysis of
the asparagine amide group are also complicating factors. Peptides
containing cysteine require methylation of the sulflrydryl group by methyl
iodide to give a stable 5-methylcysteinyl residue prior to methanolysis.
Methionine residues undergo a partial methylation 'of sulfur during
methanolysis. These side reactions cause a complication of the mass

spectra of peptides containing cysteine and methionine [27].

20

c) Modification of fragmentation

i) Fixed charge

The advent of FAB ionization has led to the use of chemical
derivatization to introduce a fixed charge into a peptide, enhancing the
detectability of molecular weight related ions. The localized charge has a
stronger effect in directing fragmentation than any basic amino acids that
may be present in the peptide, drastically altering the appearance of the
CAD spectrum of peptides [28] and facilitating structure elucidation by
enhancing the formation of structurally informative ions. When a
preformed charge is introduced into a peptide, an and tip; ions are more
abundant for N -terminally derivatized peptides, while v", y", and w", ions
are formed when peptides are derivatized at the C-terminus. The altered
fragmentation pattern produced by derivatization with a fixed charge
simplifies mass spectral interpretation and usually allows differentiation of
the leucine and isoleucine residues.

Quaternization is a straightforward reaction that is used for the
chemical derivatization of peptides in FAB-MS. “Preformed” ions,
especially tetramethyl ammonium (TMA) ions, exist as preformed ions in
solution and are ionized preferentially above glycerol [29]. TMA suppresses
the ion signal of Bronsted bases in their protonated or deprotonated form.
Small suppression effects of TMA over alkali ions were also observed;
although when mixtures of alkali metals are made in glycerol, there is little
competition between alkali ions in the FAB matrix and cationized species
have similar sensitivities. The preference factors displayed by the TMA
ions over alkali ions results from the higher surface activity of the TMA

ions.

21

A derivatization procedure to convert the e-NHz group of lysine to a
2,4,6-trimethylpyridinium quaternized group was implemented by Johnson
et al. [30,31]. This derivative produced decreased abundance of immonium
ions, internal fragment ions, and b" ions when compared to the spectrum of
underivatized peptide. Under collisionally activated dissociation (CAD)
conditions, the derivatized peptide displayed fragmentation remote to the
fixed charge of the 2,4,6-trimethylpyridinium quaternized group. As
expected, an and (1,. ions predominated when the derivatized lysine was at
the N-terminus; v", w", and y" ions predominated when the derivatized
lysine was near the C-terminus. When the lysine was in the middle of the
peptide chain, the upper mass range displayed intense an and y 3 peaks,
but small peaks in the lower mass range were observed because the ions did
not contain the preformed charge of the derivatized lysine. ,

Stults also fixed a positive charge to the amino terminus by chemical
derivatization to increase the amount of information obtainable from CAD
tandem MS spectra [32]. The two-step procedure selectively quaternizes
the N -terminus, exploiting the pKa differences of the N -terminus (pKa~8)
and side chain amino groups of lysine (pKa~10.5) and arginine (pK..~12.0).
However, cysteine residues must be protected from derivatization by
reduction and alkylation. Other disadvantages of this derivatization
procedure include a decreased signal intensity of modified peptide and an
intense peak resulting from the loss of the -N(CH3)3 group. The use of
excessive reagent and a nonvolatile buffer also necessitates sample cleanup
(typically by HPLC). Because only well-separated an and d, ions
originating from the N-terminus were observed for CAD-MS/MS of the
derivatized peptide, enhanced signal due to the ion intensity spread over

fewer channels and the need for lower resolving power was noted.

22

Utilizing a microderivatization apparatus, Vath et al. [33,34]
modified peptides by covalently linking a trimethylammonium acetyl (TMA)
moiety to the N-terminus, converting the peptide to a quaternary
ammonium cation. This procedure reproduced the derivatization method
developed by Bennett and Day [35], but was altered to be performed in the
gas phase. Because no sample transfer was necessary in the
microderivatization procedure, the reaction was carried out at the
subnanomole level. Although analysis by HPLC and the FAB mass spectra
indicated the derivatization was complete, the ratio of signal to background
for the derivatized peptide was only one-half that of the underivatized
peptide. With FAB analysis, a peak caused by the replacement of the
-N(CH3)3 group by a hydrogen was also observed. Moreover, this peptide
derivatization procedure is a lengthy three-step process.

Despite these disadvantages, peptides derivatized by this method
produced simpler and altered CAD mass spectra, providing more sequence
information than the CAD spectra of unmodified peptides. The CAD mass
spectrum was dramatically changed in one example, giving a complete
series of an sequence ions where few sequence ions were observed before
derivatization. Fixing a positive charge at the N-terminus by derivatization
directs CAD fragmentation to produce an and (1,. ions. The presence of (1..
ions allows the differentiation of leucine and isoleucine. In this sense, the
CAD mass spectrum observed is similar to the CAD spectra of peptides
containing a basic arginine residue at the N-terminus. Moreover,
derivatized peptides with a basic residue at the C-terminus primarily
produce N-terminal fragment ions, demonstrating the strong directing
effect of a fixed charge on peptide fragmentation.

23

Recently in this laboratory, Wagner et al. [36] have developed (2-
bromoethyl)triphenylphosphonium and (2-aminoethyl)triphenylphosphon-
ium derivatives to covalently link the ethyl-triphenylphosphonium (ethyl-
TPP) group to the N- or C-terminus of a peptide, respectively. The ethyl-
TPP moiety incorporates a fixed positive charge and significant hydrophobic
character to the derivatized peptide, increasing its ionization efficiency.
With CAD, the fixed charge directs fragmentation from the terminus of
derivatization. Consequently, a predictable series of predominantly an and
(1,. ions is observed after N-terminal derivatization and v", y", and w" ions
are observed upon C-terminal derivatization. Significant ethyl-TPP
derivatization is accomplished in three hours with simple laboratory
procedures. These characteristics, especially the enhanced signal intensity
of low levels of peptides and simplified fragmentation observed with CAD of
the derivatized peptide, make this a sought-after derivatization approach.

The increased detectability and enhanced fragmentation caused by
incorporation of the ethyl-TPP group into the derivatized peptide are
features that cause this derivative to be ideal for deducing the amino acid
sequence of posttranslationally modified peptides. Often, the nature of
modification (glycosylation, phosphorylation, sulfation, etc.) increases the
hydrophilicity of peptides. The altered peptides often do not fragment
efficiently with CAD and a complete ion series for amino acid sequence
determination often is absent. Ethyl-TPP derivatization, however, can
minimize these problems [37]. Upon derivatization, N-blocked
(pyroglutamic acid), phosphorylated, and peptides containing a disulfide
bridge yielded a simplified spectrum of sequence information with ion
formation directed from the terminus of ethyl-TPP derivatization.

Incorporation of the ethyl-TPP group allowed the differentiation of leucine

24
and isoleucine and identification of the posttranslational modification by
increased abundance of d, and w" ions. The detectability of the ethyl-TPP
modified peptides was at least an order of magnitude greater than the
underivatized peptides. These features warrant an investigation of other
derivatizing reagents containing the triphenylphosphonium group.
ii) Adduct formation

Alkali, alkaline earth, and transition metal ions complex with
peptides to form cationized complexes. These metal-ligand interactions
have been investigated and have resulted in new approaches for mass
spectrometric structure determination of peptides. Addition of a drop of an
alkali iodide to a liquid matrix containing peptide often causes formation of
a peak representing the peptide-alkali cation (M+Cat)+, which many times
is of greater intensity than the (M+H)+ of the untreated peptide mixture.
Collisionally-activated dissociation studies of the gas phase complexes
(M+Cat)+ provide evidence that b, product ions result from the metal
cation binding to the N -terminus of a peptide. However, the alkali metal
ion can bind to several sites in the peptide molecule and these may include
amide oxygens and chelating side chains [38]. These investigators believe
that the gas-phase complexes (M+Cat)+ do not contain the metal ion bonded
to a deprotonated carboxylate anion, and consequently reflect binding of the

metal cation to basic sites in a neutral peptide.
Investigators also have studied H 3PO4 adducts that were produced

with the addition of concentrated H3PO4 to a glycerol matrix containing

leucine enkephalin [39]. Ion series of the general formula
(M+nH3PO4+H)+ due to adduct formation, and (M+H+62)+ and
(M+H3PO4+H+62)+ series resulting from the slow decomposition of
(M+nH3PO4+H)+ ions were observed under FAB conditions. These

25
investigators also used FAB to investigate the aggregation of polyamines,
leucine enkephalin, and phosphoric acid [40].
iii) Dialkylphosphorylation

Chai and Zhao have reported the analysis of positive FAB of amino
acids, dipeptides, and tripeptides after derivatization with
diisopropylphosphite [41 ,42]. Although the derivatization procedure is
lengthy and requires several steps of sample manipulation, improved
sensitivity and decreased background noise from the glycerol matrix were
observed after derivatization of the peptides. Most of the derivatized amino
acids produced a signal enhanced by a factor of ten in comparison to
underivatized amino acids. N-terminal fragment ions dominated the

fragmentation of the N -diisopropyloxy phosphoryl derivatized peptides,
giving evidence of directed fragmentation. An intense an series of ions was

observed, with an-42 and an-84 ion series representing the loss of one and

two propene molecules, respectively. Derivatized peptides also displayed
suppressed fragmentation of amino acid side chains. The dialkylphosphite
reagent used for derivatization also can be used for peptide synthesis [43]
and more studies are underway for this derivative [44].

d) Determination of functional groups

The presence of functional groups in a peptide can be identified by
measuring the mass shift after reacting the peptide with a site-specific
reagent. This procedure also can determine the number of targeted
functional groups.

i) Acetylation

To detect the presence (or absence) of a free amino terminus, a
peptide is reacted with acetic anhydride on a probe tip for several minutes.
An upward shift of (M+H)+ by 42u indicates the presence of a free N-

26
terminus. Another unique feature of this procedure is that any fragment
ions derived from the N-terminus of the acetylated peptide are shifted
upwards by 42u when compared to the CAD of the underivatized peptide.
Because the e-NH2 group of lysine residues will partially acetylate by this
procedure, care must be exercised. However, the 8-NH2 group can be
targeted for acetylation to distinguish between lysine and glutamine
because they have identical nominal mass. Hunt creatively applied an
acetylation procedure to deduce the amino acid sequence of individual
components of complex peptide mixtures, employing CAD-MS/MS with a
triple quadrupole mass spectrometer [45,46]. Acetylation is carried out
using an equimolar mixture of acetic anhydride and hexadeuterated acetic
anhydride. The product ions containing an acetyl group appear as a
visually conspicuous, equal-intensity doublet of peaks separated by 3 mass
units when (N-acetyl-M+H)+ and (N -acetyl-d3-M+H)+ are allowed to pass
through a 5u-window. Thus, N-terminal ions are easily distinguished from
C-terminal ions.
ii) Esterification

Although this technique is not widely used, esterification can greatly
reduce the detection limits of hydrophilic peptides. Falick and Maltby [47]
capitalized on this derivatization technique to decrease detection limits to
midpicomole levels, increasing the secondary ion yield produced with FAB
by a factor of twenty-five or more. Although methyl [48] and isopropyl [49]
esters were formerly used to determine the number of free carboxyl groups
in a peptide, these investigators developed a simple, rapid derivatization
procedure to improve sensitivity with FAB. Esterification was achieved by
reacting peptide with 0.2M acetyl chloride in 5 [L of the desired alcohol for
1 h at 45°C. The resulting solution was lyophilized to remove excess

27
reactants, dissolved in 0.1% aqueous TFA. and added to matrix on a probe
tip. The signal-to-background ratio increased for peptide esterification with
alcohols of increasing alkyl chain length, although 1-octyl and longer chain
alcohols were dificult to remove from the reaction mixture by evaporation.
The longer chain alcohols increased the surface activity of peptides, and
this characteristic was applied to a mixture of peptides of differing
hydrophilicity to eliminate discrimination against hydrophilic peptides.
However, peptides that contain more than one carboxyl group are not
completely esterified with this derivatization procedure, producing a
mixture of esterified products.
iii) Amino acid detection

Caprioli and Beckner [50] investigated reagents for the specific
derivatization of peptide side chain groups. The use of site-specific reagents
in combination with FAB-MS is a convenient method to confirm the
presence of specific residues in peptides. FAB-MS is an ideal technique to
determine the completeness of the derivatization reaction, assess the extent
of modification for the group of interest, and detect side reactions in the
derivatization procedure. The stability of the derivatized peptides also may
be established by FAB-MS.

Diethylpyrocarbonate (DEP) reacts with the imidazole nitrogen
atoms of histidine to give the corresponding diethoxyformyl derivative [50].
Dissolution of 5 x 10'6 mol of peptide in 1 mL of 0.05M phosphate bufi'er (pH
6.5) and the addition of 1 .25 x 10'5M diethylpyrocarbonate in 95% ethanol
causes both nitrogen atoms of the imidazole ring to be carboethoxylated
after irreversible cleavage of the imidazole ring. However, DEP also has
been used as a site-specific derivatizing reagent of free peptidyl amino

groups to produce the corresponding ethoxyl derivative by Foti et al. [51].

28

Consequently, peptides that contain a histidine residue and a free N-
terminus display a triderivatized product. FAB-MS indicated that peptides
with more than one amino group also produced multiply derivatized
products [50], representing the addition of one or more ethoxyformyl groups
to the peptide. Therefore, this derivatization procedure requires separation
of N-terminally monoderivatized peptide from reagents and side products
by HPLC. Because the reaction conditions required to produce the
diethoxyformyl derivative of histidine more effectively modify the N-
terminus (90% completion) [50], derivatization with DEP is commonly used
for the latter purpose.

BNPS-skatole oxidizes tryptophan residues to oxyindole and
dioxyindole (or dioxyindolelactone) analogues using a two-fold or ten-fold
excess of reagent, respectively [50]. A two-fold excess of 17.2mM reagent in
50% acetic acid reacted for 24 h at room temperature produces the
oxyindole analog in a 20-30% yield with no dioxy analog. However, a ten-
fold excess of BNPS-skatole causes formation of the dioxyderivatized
peptide (25-35%) and 30-50% of oxyindole analogue. Side reactions
observed in this derivatization procedure include cleavage of the peptide at
the peptide bond.

Several schemes to specifically modify tyrosine have been reported
[50,51]. Tetranitromethane (TNM) nitrates tyrosine residues to form an o-
nitrophenol analog. Ten micromoles of peptide dissolved in 1.4 mL of
0.001M Tris and treated with two equivalents of TNM dissolved in 95%
ethanol yield o-nitrophenolic tyrosine in a slow reaction. Only a 15% yield
of derivatized product was observed for an overnight reaction. However,
formation of the dinitrophenolic species was not observed. Tyrosine-

containing peptides treated with iodine rapidly react to form iodotyrosine

29

modified in the position ortho to the hydroxyl group. Five micromoles of
peptide and 20 micromoles of ICl dissolved in 0.01M glycine buffer nearly
quantitatively yielded the derivatized tyrosine residues in 18 hrs [50]. Non-
radioactive iodination of tyrosine is also accomplished [52] with Iodogen®
(Pierce Chem. 00.; Rockford, IL). However, diiodotyrosine analogues are
observed as impurities in both procedures. Reaction of 101 with the
imidazole group of histidine and the sulfhydryl of cysteine also are possible.
Another scheme identifying the presence of tyrosine with an upward mass
shift of 42u after acetylation of the phenolic group with N-acetylimidazole
also has been reported [52]. Good yields of monoacetylated tyrosine-
containing peptides were observed in one hour of reaction time. Treatment
with hydroxylamine confirmed acetylation, because this reagent only
reverses the acetylation of phenolic groups. However, amino and hydroxyl
groups react with N-acetylimidazole after longer exposure to the reagent
(24 h).

Oxidation of methionine and an increase in mass of the protonated
molecule by 16u confirms the presence of methionine in peptides.
Treatment of a peptide with 40 uL of 1 % hydrogen peroxide solution quickly
accomplishes oxidation and the resulting product can be directly analysed
by FAB-MS. The thioether can be regenerated by reduction with b-
mercaptoethanol or dithiothreitol.

A derivatization reaction specifically modifying arginine with 1,2-
cyclohexanedione to N7,N3-(1,2-dihydroxycyclohex-1,2-ylene)-L-arginine
has been reported [53]. In the derivatization reaction, 20 nmol of peptide is
reacted with 20 nmol of 1 ,2-cyclohexanedione in borate buffer (pH 8.5) for 2
h at 37°C. After evaporation of the solvent, approximately 1 nmol of
peptide is applied to the probe tip for FAB-MS analysis. In the mass

30

spectrum, the peak representing underivatized peptide has a greater
intensity than derivatized peptide in some cases, because the chemical
modification is not complete. The peak due to derivatized peptide is shifted
upward by 9411 for each arginine contained in the original peptide. The
intensity of the peak representing completely derivatized peptide decreases
with an increasing number of arginines, indicating the derivatization
reaction is not complete or the derivatized peptide has a poorer ionization
efficiency than the underivatized peptide.

8) Miscellaneous

Banner and Spiteller reported the acylation of N-terminally free
peptides to form dansyl derivatives [54,55]. Acylation converts the free N -
terminal amine to an amide nitrogen, decreasing its basicity and increasing
the tendency for protonation along the peptide backbone. N-terminal
fragment ions dominate the mass spectra of dansylated peptides; a
complete bu series was observed for a dansylated hexapeptide, while little
sequence information was apparent from the underivatized peptide. This
derivative greatly increases the sensitivity of ultraviolet detection for HPLC
and if 2-bromo-5-(dimethylamino)benzenesulfonyl chloride is used for
derivatization; the N-terminal fragments are easily detected because they
display the “goalpost” isotope pattern of peaks characteristic of bromine
separated by 2u.

Soon after the introduction of FAB, this ionization technique was
employed to identify and quanitate dansyl amino acids obtained in the N-
terminal analysis of proteins. The investigated protein was reacted with
dansyl chloride and then hydrolyzed to yield the dansylated amino acid for
FAB analysis [56]. The dansylation procedure is an involved, time-

consuming process. The protonated N-terminally derivatized amino acid is

31

the primary species observed, although fragments due to cleavage of the
dansyl group (m/z 170) and amino acid side chain are also observed. The
response of the dansyl amino acids is a linear function of concentration up
to 20 nmol/uL, with the limit of detection being approximately 100 pmol.
This procedure does not differentiate between isomeric leucine and
isoleucine, but isobaric lysine and glutamine can be distinguished because
glutamine is hydrolyzed to glutamic acid during peptide hydrolysis. The
accuracy of the procedure was substantiated with N-terminal analysis of a
series of known peptides.

To increase the amount of sequence information from FAB-MS
spectra, peptides converted to polyamino alcohols were analyzed by tandem
FAB-MS [57]. The peptide YAGFL was reduced using diborane. In this
procedure, the amide groups are converted to amines and the carboxylic
acid group is reduced to an alcohol. Analysis of the reduced peptide by
FABoCAD-MS/MS produced complete b... y... and 2,. ion series, which were
incomplete before reduction. Therefore, the simplicity of spectral
interpretation was increased markedly with reduction of peptides to
polyamino alcohols.

The t-butyloxycarbonyl (BOC) protecting group was also investigated
as a peptide derivatizing agent [58]. The loss of C4H3 and 002 ofien is
observed with FAB-MS analysis of BOC-protected peptides. Neutral loss of
the BOC group and water [59] is also observed in peptides containing a
BOC protecting group. Moreover, the (M+H)+ signal intensity of BOC-
protected peptides is less than the protonated molecule of underivatized
peptides. Due to the facile fragmentation of the t-butyloxycarbonyl group,
the CAD spectra of the derivatized peptides are more complex and yield less

sequence information than the underivatized peptides. Consequently, more

32

descriptive mass spectra result when the BOC promoting group is removed
after peptide synthesis.

N-benzyloxycarbonyl-protected tripeptide ethyl esters containing
proline have been investigated by negative ion FAB-MS. The results
implied that conformational differences due to the position and number of
proline residues may influence the fragmentation of tripeptides by FAB-
CAD-MS/MS in the negative ion mode. The fragmentation patterns of the
derivatized peptides varied significantly, particularly the intensities of the
fragment ions formed by cleavage of the benzyloxycarbonyl group,
depending on the number and position of proline residues in the derivatized
molecules. However, the fragmentation patterns of the derivatized peptides
were nearly identical in the positive ion mode. From their observations of
the fragmentation pattern of the N-benzyloxycarbonyl group in model
tripeptides, these investigators [60] have been able to empirically predict
the numbers and positions of proline residues in N-benzyloxycarbonyl-
protected peptide ethyl esters.

To develop a method to detect carcinogenic N-nitroso peptides in
gastric contents, Wait et al. [61] nitrosated a number of synthetic peptides
for examination by FAB-MS. The peptides were nitrosated with N02 in
dichloromethane and the N -nitroso peptides were readily desorbed under
positive FAB conditions. The structures of the fully nitrosated peptides (N-
terminus and peptide bonds) were unambiguous, but when a single NNO
group was introduced to the molecule, the product ion spectra produced by
FAB-CAD-B/E displayed a dominant loss of 30u (N 0 group). Consequently,
the site of introduction of a single NNO group could not be determined from
the spectra.

33

Backbone-modified peptides cannot be sequenced by normal
techniques, such as Edman degradation, maln'ng tandem FAB-MS a critical
technique for determining the amino acid sequence of these peptides.
Several types of backbone-modified peptides were synthesized and analyzed
by FAB-CAD-MS/MS [62]. The normal peptide bond in the modified
peptides was replaced with thiomethylene ether (Cst), thioamide (CSNH),
methyleneamine (CHzNH), and thiomethylene sulfoxide (CstO) groups.
With CAD, peptides containing thiomethylene and methyleneamine groups
predominantly produced C-terminal ions. The thiomethylene sulfoxide
modified peptides, on the other hand, produced both N- and C-terminal
fragment ions. However, the thioamide group causes only slight difi‘erences
in fragmentation in comparison to peptides containing the unmodified
amide bond. These investigations showed the presence and location of
peptide backbone modification is quickly determined by the fragmentation
pattern produced with tandem FAB-MS.
D. Research Objectives

The primary goal of this research is to develop techniques for
derivatization with the triphenylphosphonium group to enhance the
analysis of peptides by FAB and FAB-CAD-MS/MS. Because several
reagents are commercially available or have been synthesized in-house to
perform TPP derivatization, the effectiveness of these derivatizing reagents
will be described. Usually small (low picomole) amounts of peptides are
available for analysis, and to eliminate sample loss due to transfer, a second
goal of this research is to develop a method to derivatize peptides with the
ethyl-TPP group on-line. The feasibility of an on-line, ethyl-TPP
derivatization procedure will be investigated and described for a

homogeneous solution. Finally, firture projects, including the derivatization

34

of a glycopeptide, will be described. Covalently fixing a charge to the
peptide portion of a glycopeptide should cause selective fragmentation of
the peptide moiety for increased structural information and decreased
fragmentation of the carbohydrate moiety, which normally preferentially
fragments.

The criteria used to evaluate the effectiveness of the TPP
derivatizing reagents are the speed and completeness of the derivatizing
reaction, the improvement in ionization efficiency, and the efficiency and
simplification of fragmentation that results from derivatization. Other
criteria used to evaluate the derivatizing reagents include the ease of the
modification reaction and the number and complexity of the manipulations
necessary to accomplish the derivatization procedure.

E. References

1. Dell, A.; Rogers, M.E., Trends. Anal. Chem., 1989, 800), 375-378.

2. Greer, F., Lab. Pract., 1989, 38(10), 81 -2, 84-5.

3. Gibson, B.W.; Biemann, K., Proc. Natl. Acad. Sci. USA, 1984, 81,
1956-1960.

4. Barber, M.; Bordoli, R.S.; Sedgewick, R.D.; Tyler, A.N., J. Chem. Soc.
Chem. Commun., 1981, 7, 325.

5. Morris, H.R.; Panico, M.; Barber, M.; Bardoli, R.S.; Sedgwick, R.D.;
Tyler, A., Biochem. Biophys. Res. Comm., 1981, 101, 623-631 .

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

7. Biemann, K., Anal. Chem., 1986, 88, 1288A-1300A.

8. Barber, M.; Bordoli, R.S.; Elliott, G.J.; Sedgwick, D.; Tyler, A.N., Anal.
Chem., 1982, 54( 2), 645A-657A.

9. Gross, M.L.; Hayes, R.N., Methods Enzymol., 1990, 193, 237-263.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

2‘3

22.

25.

35
Carr, S.A.; Hemling, M.E.; Bean, M.F.; Roberts, G.D., Anal. Chem.,
1991, 63, 2802-2824.
Roepstorff, P.; Fohlmann, J ., Biomed. Mass Spectrom., 1984, 11, 601.
Biemann, K., Methods Enzymol., 1990, 193, Appendix 5, 886-887.
Biemann, K., Methods Enzymol., 1990, 193, 455-479.
Johnson, R.S.; Martin, S.A.; Biemann, K; Stults, J .T.; Watson, J .T.,
Anal. Chem., 1987, 59, 2621-2625.
Unger, S.E.; Ryan, T.M.; Cooks, R.G., Analytica Chimica Acta, 1980,
118, 169-174.
Busch, K.L.; Unger, S.E.; Vineze, A.; Cooks, R.G.; Keough, T., J. Am.
Chem. Soc., 1982, 104, 1507-1511.
Busch, KL.; Cooks, R.G., Science (Washington, D. C.), 1982, 218, 247-
254.
Flurer, R.A.; Busch, K.L., Pmc. Indiana Acad. Sci, 1985 (Pub. 1986).
95, 171 -5.
DiDonato, G.C.; Busch, KL., Biomed. Mass Spectrom., 1985, 12(8),
364-366.
Knapp, D.R., Methods Enzymol., 1990, 193, 314-329.
Allmaier, G.M., Rapid Commun. Mass Spectrom., 1988, 2(4), 74-76.
Ligon, W.V. J r., Anal. Chem., 1986, 58(2). 485-487.
Clench, M.R.; Garner, G.V.; Gordon, D.B.; Barber, M., Biomed. Mass
Spectrom., 1985, 12(7), 355-357.
Ligon, W.V.; Dorn, S.B.; Int. J. Mass Spectrom. Ion Proc. 1984, 57, 75-
90.
Kidwell, D.A.; Ross, M.M.; Colton, R.J., J. Am. Chem. Soc., 1984, 106,
2219-2220.

27.

29.

30.

31.

32.

33.

35.

36.

37.

38.

36
Foti, S.; Saletti, R., Biomed. Environ. Mass Spectrom. 1989, 18, 168-
173.
Foti, S.; Saletti, R.; Marletta, D., Org. Mass Spectrom., 1991, 26, 903-
907.
Johnson, R.S.; Martin, S.A.; Biemann, K., Int. J. Mass Spectrom. Ion
Proc., 1988, 86, 137-154.
Shiea, J .; Sunner, J ., Int. J. Mass Spectrom. Ion Proc., 1991, 109, 262-
278.
Johnson, R.S.; Biemann, K., Proc. 35th ASMS Conf. Mass Spectrom.
Allied Topics, 1987, Denver, CO., 556-557.
Johnson, R.S.; Martin, S.H.; Biemann, K., Int. J. Mass Spectrom. Ion
Proc., 1988, 86, 137-154.
Stults, J .T.; Halualani, R.; Wetzel, R., Proc. 37th ASMS Conf. Mass
Spectrom. Allied Topics, 1989, Miami Beach, FL., 856-857.
Vath, J .E.; Zollinger, M.; Biemann, K, Fresenius Z. Anal. Chem., 1988,
331, 248-252.
Vath, J .E.; Biemann, K., Int. J. Mass Spectrom. and Ian Proc., 1990,
100, 287-299.
Bennett, B.D.; Day, R.A., Proc. 35th ASMS Conf. Mass Spectrom.
Allied Topics, 1987, Denver, CO., 568-569.
Wagner, D.S.; Salari, A.; Gage, D.A.; Leykam, J .; Fetter, J .;
Hollingsworth, R.; Watson, J .T., Biol. Mass Spectrom., 1991, 20, 419-
425.
Watson, J.T.; Wagner, D.S.; Chang, Y.-S.; Strahler, J .R.; Hanash, S.M.;
Gage, D.A., Int. J. Mass Spectrom. Ion Proc., 1991, 111, 191 -209.
Teesch, L.M.; Orlando, R.C.; Adams, J ., J. Am. Chem. Soc., 1991, 113,
3668-3675.

39.

41 .

42.

$3:

47.
. Kausler, W.; Schneider, K.; Spiteller, G., Biomed. Environ. Mass

49.

51.

52.
53.

37
J ankowski, K.; Virelizier, H.; Lafontaine, P.; Pare, J .R.J ., Spectrosc.
Lett., 1991, 24(1), 35-41.

. Jankowski, K.; LeBlanc, J .-F.; Virelizier, H.; Tabet, J .-C.; Pare, J .R.J .,

Spectros. Int. J., 1989, 1, 49-56.

Chai, W.-G.; Yan, L.; Wang, G.-H.; Liang, X.-Y.; Zhao, Y.-F.; J ., Gai-
Jiao, Biomed. Environ. Mass Spectrom., 1987 , 14, 331 -333.

Zhao, Y.-F.; Zhang, D.-Q.; Xue, C.-B.; Zeng, J .-N .; J ., Gai-Jiao, Org.
Mass Spectrom., 1991, 26, 510-513.

Zhao, Y.-F.; Zhang, D.-Q.; Xue, C.-B., Int. J. Protein Res., 1991, 37, 457-
461.

Ma, X.-B.; Zhao, Y.-F., Biol. Mass Spectrom., 1991, 20, 498-502.

Hunt, D.F.; Bone, W.M.; Shabanowitz, J .; Rhodes, J .; Ballard, J .M.,
Anal. Chem., 1981, 53, 1706-1708.

Hunt, D.F.; Buko, A.M.; Ballard, J .M.; Shabanowitz, J .; Giordani, A.B.,
Biomed. Mass Spectrom., 1981, 8(9), 397-408.

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

Spectrom., 1988, 17, 15-19.

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

Caprioli, R.M.; Beckner, C.F., GC-MS News, 1984, 12(6), 152-159.
Foti, S.; Marletta, D.; Saletti, R.; Petrone, G., Biomed. Environ. Mass
Spectrom., 1991, 20, 336-339.

Silberring, J .; Nyberg, F., J. Chromatogr., 1991, 562, 459-467.
Silberring, J .; Nyberg, F., Biomed. Environ. Mass Spectrom., 1990, 19,
819-821.

Renner, D.; Spiteller, G., Angew. Chem., 1985, 97, 408-410.

55.

57.

59.

62.

38
Renner, D.; Spiteller, G., Biomed. Environ. Mass Spectrom., 1986, 13,
405-410.
Beckner, C.F.; Caprioli, R.M., Anal. Biochem., 1983, 130, 328-333.
Lippstreu-Fisher, D.L.; Gross, M.L., Anal. Chem., 1985, 57, 1174-1180.
Bathelt, E.R.; Heerma, W., Biomed. Environ. Mass Spectrom. , 1987 ,
14, 53-61.
Jankowsln', K.; Gaudin, D.; Pham Van Chuong, P.; Spectrosc. Lett. ,
1987, 20(3), 255-284.
Tsunematsu, H.; Nakashima, S.; Yoshida, S.; Yamamoto, M.; Isobe, R.,
Org. Mass Spectrom., 1991, 26, 147-150.

. Wait, R.; Leach, S.A.; Hill, M.J.; Thompson, M.H., Biochem. Soc.

Trans, 1988, 16, 740-744.
Deterding, L.J.; Tomer, KB; Spatola, A.F., J. Am. Soc. Mass
Spectrom., 1990, 1, 174-182.

39

Chapter II
'h'iphenylphosphonium Derivatizationand the

SeparationotComponentsofflreReactionMixtme

A. Introduction

Peptide derivatization at the N-terminus with (2-bromoethyl)TPP
and at the C-terminus with (2-aminoethyl)TPP has been performed in
this laboratory [1 -5]. The ethyl-TPP group significantly increases the
hydrophobic character of the derivatized peptides, enhancing their
ionization efficiency and increasing their signal intensity during FAB-
MS. The localized charge of the quaternized phosphorus atom in the
ethyl-TPP group directs fragmentation by remote site fragmentation from
the terminus of derivatization [6], simplifying the fragmentation pattern
produced by CAD-MS/MS, which otherwise produces fragments from
both termini of an underivatized peptide.

These benefits of derivatization with the triphenylphosphonium
group warranted the derivatization of peptides with other reagents
containing the TPP group. The signal enhancement produced with
derivatization depends on the extent of derivatization and the properties of
the modifying group. Consequently, the amount of signal enhancement
due to derivatization can be increased by employing derivatizing reagents
with optimal properties for the analytical technique employed and by
manipulating the reaction conditions to improve the reaction yield. A
number of reagents containing the triphenylphosphonium group are

commercially available or have been synthesized in this laboratory. The

4 0
compound (1 -bromomethyl)TPP contains one methylene group in contrast
to the two methylene groups of (2-bromoethyl)TPP. Commercially
available compounds include (3-bromopropyl)TPP and (4-bromobutyl)
TPP, which have an increasing alkyl chain length. The 2-(triphenyl-
phosphonio)ethoxycarbonyl chloride reagent (the so-called “Kunz”
reagent) is a reactive chloroformate. In addition, (4-carboxybenzyl)TPP
was synthesized in-house to modify the N-terminus of peptides with the
coupling action of a carbodiimide. The derivatization procedures used to
modify peptides with these magenta will be described in this chapter. The
derivatization of VGVAPG and the fragmentation patterns observed for
the modified peptides by FAB-CAD-MS/MS by linked scanning will be
explained in this chapter. The C-terminal derivatization of leucine
enkephalin (YGGFL) with (2-aminoethyl)TPP, separation of the
components of the reaction medium by HPLC, and ensuing quantitation
of the derivatized product by UV-Vis detection in combination with HPLC
chromatography will be related.

Table 2.1. Investigated derivatizing reagents containing the TPP group.

Beam Scum
(2-Aminoethyl)TPP In-house synthesis
(1 -Bromomethyl)TPP bromide Lancaster Synthesis Ltd.
(2-Bromoethyl)TPP bromide In-house synthesis
(2-Chloroethyl)TPP bromide Lancaster Synthesis Ltd.
Vinyl-TPP bromide Lancaster Synthesis Ltd.
“Kunz” reagent FlukaChemie AG
(3-Bromopropyl)TPP bromide Lancaster Synthesis
(4-Bromobutyl)TPP bromide Lancaster Synthesis

(4-Carboxybenzyl)TPP In-house synthesis

4 l
B. Synthesis of reagents containing the TPP group
1. (2-Bromoethyl)TPP
a) Synthesis by Method I
i) Experimental
The method of Frederich and Henning [7] was used to synthesize
(2-bromoethyl)TPP from 1 ,2-dibromoethane and triphenylphosphine in a
simple substitution reaction (Reaction 2-1). To 40 mL of xylene, 1.00 g of

(2-1) ©9/B}2—CH2—Br ——>- @PLmz—an—Br '4’ Br.

triphenylphosphine was added (3.81 x 10'3 mol) along with a hundred-fold
molar excess of 1,2-dibromoethane (33.0 mL). The mixture was heated to
50°C and stirred overnight. The desired product precipitated out of the
reaction solution and was collected on filter paper with suction filtration.
The white precipitate was washed with diethylether and dried for use as a
peptide derivatizing reagent.
ii) Results and discussion

One should note that peptides derivatized with a TPP group and the
derivatizing reagents containing a TPP group do not represent a
molecular ion according to a strict definition of the term, but a cation of
an ionic compound. Therefore, when a solution containing (2-
bromoethyl)TPP is analyzed by FAB-MS in the positive mode, the (2-
bromoethyl)TPP cation is desorbed (m/z 369, 371), but the compound also
contains a bromide anion, and the proper name of the compound is (2-

bromoethyl)TPP bromide.

4 2

Analysis of the reaction product by FAB-MS indicated that the
product contained (2-bromoethyl)TPP as represented by the bromine
isotope peaks of the molecular ion (M+) at m/z 369 and 371 (Figure 2.1 ).
The product contains a significant amount of vinyl-TPP impurity (Om/2
289) and some (2-diphenylphosphinoethyl)TPP (m/z 475) salt as a side
product. Direct analysis of the reaction product by HPLC confirmed the
presence of a significant amount of vinyl-TPP impurity. The
chromatogram of the product mixture (Figure 2.2) indicates two main
components are contained in the product. The fractions represented by
the peaks were collected and analyzed by FAB-MS. The first fraction
contained vinyl-TPP, while the second peak represented (2-
bromoethleI‘PP. Because this method of synthesizing (2-bromoethyl)TPP
also produced a vinyl-TPP impurity and the unpurified reaction mixture
was used for peptide derivatization, some necessary questions arose:
which compound was derivatizing the N-terminus of the peptides, and if
both reagents could react with a peptide, which reacted more efficiently?
To answer these questions, isolation of the individual compounds by flash
column chromatography was attempted [8].

b) Thin-layer chromatography

i) Experimental

The product synthesized by Method I was analyzed by thin-layer
chromatography (TLC) prior to separation of the mixture by flash column
chromatography. TLC separation was performed with Whatman‘”
(Whatman Paper Ltd.; Maidstone, Kent, England) premade TLC plates
that consist of a flexible aluminum backing coated with 250 pm of silica
gel. The plates were cut into 2.5 x 10 cm strips and spotted by micropipet
with solution containing the product to be analyzed. The sample spots

43

 

 

 

 

 

 

1:00: Vinyl-TPP" (2-Bromoethyl)TPP“'
e i

l 1

a 80‘

t .

i I

v 4

0 60‘

1 I

n .

t 40.

e .

n .

s .

i 20* TPP“

y i

«i. .. 18

10° 20° 300 460' . Y 500

Figure 2.1. The FAB-MS spectrum of the (2-bromoethyl)TPP product
synthesized by reaction of 1 ,2-dibromoethane with triphenylphosphine.
The molecular ion (M+) of (2-bromoethyl)TPP is represented by the doublet
at m/z 369 and 371; the vinyl-TPP and (2-diphenylphosphinoethyl)TPP
impurities appear at m/z 289 and m/z 475, respectively. Nitrobenzyl alcohol
(NBA)wasusedastheFABmatr-ix.

44

 

 

 

 

 

 

 

1.0-
a.
0.75- E:
‘8
A 3 '
O
i, a
E
0.5- «a
8
.2.
c3
4.
E“
"'8.
0.25- ,5 '
;>
i
0 I' F4 r l
Tweet 8 12 16 20 24 23
Time (min)

Figure 2.2. HPLC chromatogram of the (2-bromoethyl)TPP product
mixture synthesized by reaction of TPP with 1,2-dibromoethane. For this
separation, approximately 1 00 nmol of product mixture was injected. The
parameters used in the separation are: mobile phase, 45:55 (v/v)
acetonitrile/water; flow rate, 1 mL’min; column, 4.6 x 250 mm; stationary
phase, 5 um diameter, 013 bonded phase; detector, 214 nm, 1.0 AUFS;
chart speed, 20 cm/hr.

4 5
were 0.2 cm in diameter and made 1.0 cm above the bottom of the TLC
plate. For sample development, the spotted plate was placed in a solvent
chamber containing the mobile phase at a depth of 0.5 cm. The plate was
removed from the chamber when the mobile phase migrated to within
approximately 1 .0 cm of the top of the plate. The plate was allowed to air-
dry in a hood. When dry, the plate was observed in a viewing chamber by
fluorescence at 254 nm. The positions of the separated compounds
appeared as purple spots on the TLC plate. The positions of the spots
were marked and their Rf values were calculated. Several solvent
combinations were investigated for optimal separation, and when a
mobile phase produced a desired separation of the mixture, all the
fractions collected by flash column chromatrography were analyzed by
TLC using this solvent combination.
ii) Results and discussion

Optimal separation by flash column chromatography of a mixture
is accomplished in a reasonable length of time when the components
have Rfvalues between 0.35 and 0.50 and are separated by an Rfvalue of
0.1 . A methanol/chloroform 14:100 (v/v) mixture provided the desired
separation of the reaction product obtained by Method 1. The lower spot
resulting from TLC separation represents (2-methoxyethyl)TPP (m/z 321 ,
Figure 2.3 a)). The upper spot represents vinyl-TPP and (2-
bromoethyl)TPP (m/z 289 and 369, 371, respectively; Figure 2.3 b)). The (2-
methoxyethyl)TPP is a side product resulting from the Michael addition
reaction of methanol to vinyl-TPP [9-11]. Separation of the product
synthesized by Method I in a 1:1 (v/v) acetonitrile/chloroform solvent
combination produced a single spot with an Rf of 0.42. No side products

were observed with use of this mobile phase. However, no solvent

46

 

 

 

 

 

 

 

 

 

 

 

 

 

1100: (2-Methoxyetr.hyl)’I‘PP+

, . (a)

1 .

880‘

i 1

v ‘ 0

860- Vrnyl-TPP"

1 .

n .

t 40.

° i

n I

‘ ‘ r

:20-

y TPP"

100 200 ml 300 400 500

2

1%00‘

. . (b) mel-TPP“

1 .

1180‘

.t 4

r I (2-Bromoethyl)TPP"'

v 1

O 60‘ _

I l

n .

t 40.

e .

n .

s .

{so TPP”

’ l l... .

in. .8. .. . . .4.
100 200 300 400 500

m/z

Figure 2.3. The FAB-MS spectra of the components of the (2-bromo-
ethyl)TPP reaction mixture prepared by the method of Fredrich and
Henning as separated by flash column chromatography using a 14:100 (v/v)
methanol/chloroform mobile phase: a) (2-methoxyethyl) TPP corresponding
to the lower TLC spot and last-eluting compound when separated by flash
column chromatography. b) vinyl-TPP and (2-bromoethyl)TPP that were
the first-eluting compounds during separation by flash column
chromatography and was represented by the upper spot after separation by
TLC . NBA was the FAB matrix used for both spectra.

4 7

combination was discovered that could resolve vinyl-TPP and (2-
bromoethyl)TPP using TLC plates coated with silica gel.

c) Flash column chromatography

i) Experimental

Separation of the product produced by Method I was performed
with flash column chromatography. A 2-inch diameter column with a
stopcock and flow controller was filled with six inches of 200-300-um
diameter silica gel (Sigma Chem. Co.; St. Louis, MO). The column was
filled with the 14:100 (v/v) methanol/methylene chloride solvent mixture
that produced the TLC separation and pressure was applied to rapidly
push all air from the silica gel. One gram of the (2-bromoethyl)TPP
product mixture dissolved in this solvent mixture was applied to the
column and the flow controller was briefly placed on the column to push
the sample into the silica gel. The column was refilled with solvent and
eluted at a flow rate of 2 inches per minute (approximately 75 mL/min).
A rack containing forty 20 x 150 mm test tubes was used to collect the
eluted fractions. Small fractions were collected at the start of the run and
larger fractions were collected near the end of the run. Composition of
the collected fractions was determined by TLC and the fractions
containing the compounds corresponding to the lower and upper TLC
spots were concentrated by rotoevaporation.

ii) Results and discussion

Separation of the product produced by Method I by flash-column
chromatography provided two fractions. The first-eluting fraction
contained (2-bromoethyl)TPP and vinyl-TPP. The second fraction
contained the (2-methoxyethyl)TPP side product. The side product results

4 8
from the Michael addition of methanol to vinyl-TPP impurity in the
reaction product.

To avoid formation of the (2-methoxyethyl)TPP side product, flash
column chromatography was repeated in a 1:1 (v/v) acetonitrile/
chloroform mixture and a single fraction containing a mixture of vinyl-
TPP and (2-bromoethyl)TPP resulted, as investigation with FAB-MS and
separation by HPLC demonstrated. No Michael addition products were
observed when the reagents were isolated with these nonhydroxylic
solvents. Although separation of vinyl-TPP and (2-bromoethyl)TPP with
combinations of other solvents that do not contain hydroxyl groups was
attempted, satisfactory separation of the reaction mixture was not
achieved. Consequently, an alternative method to produce pure (2-
bromoethyl)TPP was pursued.

d) Synthesis by Method 11

i) Experimental

Nearly pure (2-bromoethyl)TPP was synthesized by reaction of (2-

hydroxyethyl)TPP with phosphonium tribromide (Reaction 2-2) [1 2,1 3]. A

as .(Q—g-m’flz—m) . Pan ——> s(©—§-Gz-mz—Br) + o=aoan

mixture of 5.00 g (0.0129 mol) of (2-hydroxyethyl)TPP and 5.0 mL (0.053
mol) of tribromophosphine were heated on a steam bath for 2 h with

occasional shaking. A viscous orange syrup resulted, which was
suspended in 100 mL of H20 and extracted with chloroform. The
chloroform fractions were washed with H20 and dried over N a2804. The

chloroform was removed by rotoevaporation, yielding yellow crystals.

4 9
These crystals were dissolved in acetonitrile with slight heating and
recrystallized to yield clear, white crystals. This procedure yielded 4.89 g
(0.0109 mol) of (2-bromoethyl)TPP; this is a 84.5% yield of this product.
ii) Results and discussion

Separation of the reaction product by HPLC and analysis by FAB-
MS indicated nearly pure (2-bromoethyl)TPP was produced by this
method. Because vinyl-TPP and (2-bromoethyl)TPP have identical
extinction coefficients, HPLC indicates that the product is 95% (2-
bromoethyl)TPP and 5% vinyl-TPP (Figure 2.4). FAB-MS also indicated a
highly pure product of (2-bromoethyl)TPP with slight vinyl-TPP
impurities was produced by this method (Figure 2.5).

2. (2-Aminoethyl)TPP

a) Experimental

The synthesis of high purity (2-aminoethyl)TPP from (2-
bromoethyl)TPP now was possible. To a stirred solution of acetonitrile
1.50 g (0.00333 mol) of (2-bromoethyl)TPP was added. When the (2-
bromoethyl)TPP dissolved, 12 mL of 27% ammonium hydroxide (~ 15-fold
molar excess) was added dropwise to the reaction mixture. The solution
was heated to 40°C and stirred for 1 hr. The acetonitrile and excess
ammonium hydroxide were removed by rotoevaporation. White crystals
of solid (2-aminoethyl)TPP remained in the round-bottomed flask. This
solid was used directly for C-terminal derivatization after being dissolved
to a known concentration in methanol.

b) Results and discussion

The reaction used to produce (2-aminoethyl)TPP is outlined in

Reaction 2-3. Separation of (2-aminoethyl)TPP synthesized by this method

50

 

 

 

 

 

 

 

1.0 -
+
D-I
0.75 .. E:
3’
A 8
8
u
m
93
0.5 -
0.25 -
4.: +
c: D-I
3 E?
H F!
5 .E‘
.2. >
a .
0 I I I *ILIJ I I I I
trmec't 6 12 18 24 30

Time (min)

Figure 2.4. HPLC separation of the (2-bromoethyl)TPP product synthes-
ized by reaction of (2-hydroxyethyl)TPP with PBr3. Approximately 1 0 nmol
of product was injected for separation. The parameters used in the
separation are: mobile phase, 45:55 (v/v) acetonitrile/water; flow rate, 2
mL/min; column, 10 x 250 mm; stationary phase, 10 um diameter, 013
bonded phase; detector, 214 nm, 0.5 AUFS; chart speed, 20 cm/hr.

51

 

 

 

 

 

 

100‘
3 i (2-Bromoethyl)’l‘PP+
1 .
a 80'
.t .
1 4
v .
O 60"
1 1
n .
t 40.
e .
g Vinyl-TPP"
; 20‘ ‘
y ‘ TPP"
T» u.+-;~L~TL“:‘LJ.L:J1 , t . in ,4- . .krl , A. . . .
100 200 300 400 500
m/z

Figure 2.5. Analysis of the product produced by reaction of (2-hydroxy-
ethyl)TPP with PBr3 by FAB-MS. The peaks representing the molecular
ion (M+) of (2-bromoethyl)TPP dominate the FAB-MS spectrum (m/z 369,
371). Vinyl-TPP impurity appears at m/z 289. The FAB matrix was NBA.

52

Q

+ +
(2-3) ©P—cr12—CH2_Br + NH. —> @P—CHg—CHg—NH. + HBr

by HPLC with subsequent UV-Vis detection, indicated that nearly pure
product was produced. The chromatogram of the reaction mixture
appears in Figure 2.6. Because the absorption coefficients of (2-amino-
ethyl)TPP and vinyl-TPP impurity are nearly identical, the product was
calculated to be 92% pure, based on the area of peaks in the
chromatogram. The analysis of product by FAB-MS also indicated the
product was of high purity (Figure 2.7).
3. (4-Carboxybenzyl)TPP

a) Experimental

To prepare the (4-carboxybenzyl)TPP derivatizing reagent, 8 1 :1
molar ratio of triphenylphosphine was reacted with 4-(chloromethyl)
benzoic acid. The reaction was performed in 300 mL of xylene; 1.54 g (1.18
x 10’2 mol) of triphenylphosphine and 1.00 g (1.18 x 10'2 mol) of 4-
(chloromethyl)benzoic acid were dissolved in this solvent with stirring.
The reaction was performed with gentle reflux at 144°C for 24 hours.
Analysis of the reaction mixture by TLC after reflux demonstrated that a
high yield of (4-carboxybenzyl)TPP was present and a small amount of 4-
(chloromethyl)benzoic acid and some TPP impurity remained in the
reaction mixture. The TLC separation was achieved in a 4:1 (v/v)
acetone/chloroform solvent mixture. After cooling, xylene was removed
by rotoevaporation until a light brown syrup remained. Further drying of

53

 

 

 

 

 

 

 

 

 

0.5- r
4.
E5
‘5.
fl
4)
0.375- g a
A £3 '55
w 8
8 .
N
a l E
0.25- 93
4.3
a
0.125— ‘E +
n. .
5 E:
,2, .
r3 '8
S
0 r HI I. r I
meet 8 12 16 2o 24
Time (min)

Figure 2.6. The chromatogram resulting from the separation of the (2-
aminoethyl)TPP product by HPLC. The first on-scale peak represents the
(2-aminoethyl)TPP product; the second peak represents vinyl-TPP impurity
in the product. For this separation, approximately 200 nmol of product was
injected and the parameters used in the separation are: mobile phase,
45:55 (v/v) acetonitrile/water; flow rate, 2 mL/min; column, 10 x 250 mm;
stationary phase, 5 um diameter, 013 bonded phase; detector, 214 nm, 0.5

AUFS; chart speed, 20 cm/hr.

54

 

-8

(2-Aminoethyl)TPP+

 

 

 

 

1

R

e 1

l ,

a 80‘

.t 4

1 1

V

e sol

I i

n J

t 40.

0 w

n 4

I .

:20. Vinyl-TPP

y 3 ‘
‘ A} A , . . . A. i, . .
q) 100 200 ml 300 400 600

z

Figure 2.7. The spectrum resulting from the analysis of the (2-aminoethyl)
TPP product mixture by FAB-MS. The molecular ion (M+) of (2-
aminoethyl)TPP dominates the spectrum (m/z 306), but some vinyl-TPP is
observed due to impurity and fragmentation (m/z 289). The peaks at m/z
136 and 154 are characteristic of the NBA matrix used.

5 5
the product was performed under vacuum until a solid product formed.
The yield of product was 3.09 g (60.6 %).
b) Structural investigations
i) Analysis by FAB-MS
The reaction used to produce (4-carboxybenzyl)TPP is described in
Reaction 2-4. Although analysis by FAB-MS revealed the expected

© . o
‘2'" Gig + a—Cflz‘Q’g"°" —_> ©6CH2-<>—CI=I—0H + cf

(<5

molecular ion (M+) for (4-carboxybenzyl)TPP at m/z 397 with ~ 10%
relative intensity, a base peak at m/z 279 was observed (Figure 2.8). The
peak at m/z 279 represents the protonated triphenylphosphine oxide
molecule. The ion observed at m/z 279 probably is formed in the FAB
selvedge region due to the high-energy conditions present in and above
the FAB-MS matrix. Analysis of the (4ecarboxybenzyl)TPP product by
FAB-CAD-B/E produced the fragmentation pattern appearing in Figure
2.9. The ion series at m/z 108, 183, 185, and 262 is characteristic of a
compound containing triphenylphosphine. The peaks at m/z 307, 320,
352, 369, and 381 support the identity of the product and fragment at the
positions described in the inset of Figure 2.9.
ii) Analysis by NMR

The identity of the (4-carboxybenzyl)TPP product was investigated
by NMR, which indicated the product was pure (4-carboxybenzyl)TPP.
Deuterated chloroform and d4-methanol were used for analysis by NMR
and the resulting spectra can be observed in Figures 2.10 a) and b),
respectively.

56

 

-3

('l‘l:'Poxlde+I-I)+

A l A A

A J A A A l A A

%fifl'flflOHuI-U Q<flofipma ”H
'° 8
O

 

 

 

 

I l I ‘ H (It-carbixybenzyll'EPP"
(i) 4“ A 1* *fl . , 4. . Tr .

100 200 300 . 400 500
m/z

Figure 2.8. Analysis of the (4-carboxybenzyl)TPP product by FAB-MS that
was produced by reaction of TPP with 4-(chloromethyl)benzoic acid. The
peak representing the molecular ion (M+) of the (4-carboxybenzy1)TPP
product is at m/z 397. The base peak in this spectrum represents
protonated triphenylphosphine oxide and the peak at m/z 201 represents
diphenylphosphine oxide. The FAB matrix was glycerol.

 

‘8

(30"

w

H
. 9."

 

I-I
°--.

9'---

$“DIBG95H O<""¢*fl"‘0

 

 

 

(9+ -

50 100 150 200 250 300 350 400
m/z

Figure 2.9. The spectrum resulting from analysis of the (4-carboxybenzyl)
TPP product by FAB-CAD-B/E. The fragmentation observed is related to
the structure of the product in the inset. The FAB matrix was NBA.

57

(a)

 

 

 

 

 

 

 

JJLU, .JJ L
V V I I T V I I V Y Y ' ' V ' ' ‘ ' Y V V I U ‘ ‘

 

4

d
J

d
‘

 

 

 

 

 

 

 

8 7 6 5 3 2 I 0 DO.
l 1“ JL M [J.n A 44L

l r 1 U j I ' ‘ 1 l ' ' ' ' I I I
I 7 6 5 4 3 2 I on.

Figure 2.10. Analysis of the (4-carboxybenzyl)TPP product by NMR when
dissolved in a) deuterated chloroform and in b) d4-methanol.

5 8

The NMR spectra for (4-carboxybenzyl)TPP dissolved in the two
unlike solvents differ markedly. The analysis by NMR of the sample
dissolved in deuterated chloroform appears in Figure 2.10 a). The three
signals furthest downfield (7.85, 7.80, and 7.70 ppm) represent the
hydrogens contained in the phenyl rings of the triphenylphosphonium
moiety. The peak at 7.24 ppm is caused by non-deuterated chloroform
impurities in the solvent. The phenyl hydrogens ortho to the phosphorus
atom are expected to cause the signal farthest downfield because they are
deshielded by resonance with the charged phosphorus atom (Scheme 2-1).
A resonance structure also can be drawn for the hydrogens para to the

phosphorus atom, (Scheme 22), but their signal is not shifted downfield

I \
Scheme 2-1 R-+P—R <-—> R _ P‘R
I l
R R
Scheme 2'2 R-+P—R <-——> R—P—R
l l
R R

as much as that of the hydrogens in the ortho position because the loss of
electron density by induction to the positively charged phosphorus atom is
not as strong. The peak at 7.70 ppm represents the hydrogens meta to the
phosphorus atom. These hydrogens have the least shift downfield
because no electron density is shifted to the charged phosphorous by
resonance. The quintet at 4.7 ppm and the twin triplets centered at 3.8

5 9

ppm are due to the geminal methylene hydrogens. The large coupling
constant observed for these signals is thought to be caused by the
dimerization of two (4-carboxybenzyl)TPP molecules to neutralize the
positive electrostatic charge of this molecule when it is dissolved in
relatively nonpolar methylene chloride. The dimer probably results from
the charged TPP group of (4-carboxybenzyl)TPP complexing with the
carboxy group of another (4-carboxybenzyl)TPP molecule. Both the signal
at 4.7 ppm and 3.8 ppm appear to be doublets of triplets. The splitting
observed is caused by interaction of the geminal methylene hydrogens
with one another and the triplet signals observed are caused by each
geminal hydrogen interacting with the hydrogens ortho to the methylene
group on the adjacent phenyl ring. Further splitting of the signal is
caused by hydrogens on the phenyl rings of the TPP moiety. The peak at
1.6 ppm is caused by water impurities in the sample.

The NMR spectrum of (4-carboxybenzyl)TPP dissolved in d4-
methanol (Figure 2.10 b)). on the other hand, displays absorbances at 7.0-
7.4, 7 .5-7.8, and 8.0 ppm due to hydrogens of the phenyl rings in the TPP
group. The peak at 4.9 ppm is due to water, and the peak at 3.3 ppm is
due to undeuterated methanol impurity contained in the solvent. The
peak (4.7 ppm) representing the methylene hydrogens a to the
quaternized phosphorus atom is a singlet because methanol is a more
polar solvent and can accommodate the electrostatic charge of (4-
carboxybenzyl)TPP more readily. Therefore, dimerization does not occur.
The environment of the methylene hydrogens is similar and a single
signal results. Conseqently, this spectrum indicates the product is (4-
carboxybenzyl)TPP.

6 0
C. Peptidedaivatimfionpncedmu
I. Bromoalkyl-TPP reagents and vinyl-TPP

a) Experimental

The peptide to be derivatized was reacted with a twenty-fold molar
excess of the bromoalkyl-TPP reagent or vinyl-TPP. First, the
derivatizing reagent was dissolved in 1 mL of the derivatizing medium
contained in Eppendorf tube. The derivatizing medium was either an
aqueous potassium phosphate/sodium borate buffer of pH 9.0 (Micro
Essential Laboratory; Brooklyn, NY) or a 1:1 (v/v) acetonitrile/pyridine
mixture (pr pyridine = 8.83 [14]). The desired amount of peptide to be
derivatized was placed in the reaction medium and the solution was
vortexed. The Eppendorf tube was placed in a water bath (37°C)
overnight. If the N-terminal derivatization was performed in the
acetonitrile/pyridine solvent mixture, the volatile solvents were
evaporated under a nitrogen gas stream from a N-EVAPQ
(Organomation Associates, Inc.; N orthborough, MA). The remaining
solid residue was redissolved in water and lyophilized to remove any
remaining pyridine. The solid residue was dissolved in 1-1.5 “L of 1:1
(v/v) acetonitrile/water mixture and added to 1 uL of glycerol positioned on
the FAB probe tip.

b) Results and discussion

i) (2-Bromoethyl)TPP and vinyl-TPP

All the bromoalkyl-TPP reagents derivatized the N-terminus of
peptides at an appreciable level, with the exception of (1-
bromomethyl)TPP. When the (1 -bromomethyl)TPP reaction mixture was
analyzed by FAB-MS, a signal representing the expected product was
barely perceptible. The (2-bromoethyl)TPP (Reaction 2-5) and vinyl-TPP

61

0
ll
(2-5) @PtCHZ—CHZ—Br + H—(lil—(IZ—C)fi-OH —->
H R
’I‘ (ll
HBr + @PLCHZ-CHz—(If—CIJ-C)n—OH
H R

reagents produced high yields of derivatized peptides modified with the
ethyl-TPP group at the N-terminus. Under non-aqueous conditions,
vinyl-TPP can under Michael addition (Reaction 2-6) with primary
amines and proteins, as reported by Swan and Wright [1 5].

W

(2-6) @p’f—CIFCHZ + H—(lil—C'Z—C)n-'OH ——>

H R
(C? 'r 0
+ II
©— —CH2— CHz—(rlI—ci—C);0H
H R

A volatile buffer system of 1 :1 (v/v) acetonitrile/pyridine reduced
sample loss when used for bromoalkyl-TPP derivatization. Initially, an
aqueous bufl'er system composed of potassium carbonate/sodium borate
salts was employed. Use of this buffer system necessitated separation of
the N-terminally derivatized peptide from the bufl'er salts by a Sep-Pak®
procedure. The Sep-Pak® procedure requires several sample transfer
steps contributing to sample loss. A study comparing the reaction yield
produced by vinyl-TPP and (2-bromoethyl)TPP when reacted with 1
nanomole of leucine enkephalin (YGGFL) in a 1:1 (v/v)
acetonitrile/pyridine solvent mixture overnight (~ 8 h) was made. A
higher FAB-MS response of N-terminally derivatized YGGFL was
observed for the reaction with vinyl-TPP than (2-bromoethyl)TPP (Figures
2.11 a) and b)). Also, a larger signal representing underivatized peptide

62

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

gm: (a) - . TPP+“" Vinyl-T133

e . X

1 . / -

a80‘

.t <

l

V

060

I .

n 4

1:40.

3 I +TPP-Ethyl-YGGFL

I 1 ' .

:20: t c -
. 1 . .o

y ‘ ll I II
C
' 100 200 300 400 500 600 700 800 900

- m/z

1"" ‘ TPP" 'Vin l-TPP“

R . y

1° , Cb) ‘ x /

£180 A

t 1

i .

v .

060‘

I I ‘

n .

1140-

o .

n .

.I .

l .

§ 2°. . . *"TPP-Ethyl-YGGFL
‘ lA $an “9* ' 11' . .A' l-
(I, 100 200 300 400 500 600 700 300 900

 

ml:

Figure 2.11. The FAB-MS spectrum of 1 nmol of YGGFL N-terminal
derivatization mixture. Leucine enkephalin was reacted with a 20-fold
excess of a) vinyl-TPP and b) (2-bromoethyl)TPP in an acetonitrile/pyridine
1 :1 (v/v) reaction mixture. A higher response resulting from the molecular
ion (M+) of derivatized peptide (m/z 844) was observed for reaction with
vinyl-TPP than (2—bromoethyl)TPP.

6 3

(m/z 556) is observed in the (2-bromoethyl)TPP reaction medium than in
the vinyl-TPP reaction medium, indicating that vinyl-TPP is a more
efficient derivatizing reagent. Because the reaction product for peptide
modification with these reagents is identical, the difference in the FAB-
MS response for ethyl-TPP derivatized peptide can be related to the
reaction yield. The results of this study indicated that vinyl-TPP is a more
effective derivatizing reagent than (2-bromoethyl)TPP when the
acetonitrile/pyridine reaction mixture is utilized.

When (2-bromoethyl)TPP and vinyl-TPP are reacted with YGGFL
under aqueous conditions (pH 9.0), vinyl-TPP produces a higher yield of
N-terminally derivatized peptide with these reaction conditions (Figure
2.12). Again, a larger amount unreacted YGGFL is present in the (2-
bromoethyl)TPP derivatizing medium than in the vinyl-TPP derivatizing
medium, indicating vinyl-TPP is a more effective derivatizing agent. The
FAB-MS spectra of 1 nmol of YGGFL reacted with vinyl-TPP and (2-
bromoethyl)TPP is presented in Figures 2.12 a) and b), respectively. The
majority of vinyl-TPP is converted to (2-hydroxyethyl)TPP (m/z 307) during
the course of the reaction, as only a small amount of vinyl-TPP (m/z 289)
is observable after 8 h of reaction time. However, the FAB-MS response
due to N-terminally, ethyl-TPP derivatized peptide produced in the
potassium carbonate/sodium borate aqueous bufl‘er (pH 9.0) is lower than
the FAB-MS response observed for the acetonitrile/pyridine buffer system.
The FARMS response results from the loss of sample incurred with the
transfer of sample during the Sep-Pak® procedure used to separate the
derivatized product from the buffer salts. Lower reaction yield also is
caused by the addition reaction of water and vinyl-TPP to form unreactive

(2-hydroxyethyl)TPP; the large excess of water present in the reaction

64

 

 

 

 

 

 

 

 

 

 

 

 

 

100- . a “M

3 1 (a) (zHydmxyet-hylfl‘ +TPP-Ethyl-YGGFL

1 4 .

9.80

t 4

i J

v 4

e 60- .

I 4

n 4 l

t 40. i

e 4

n . l

.8 4

l

t 2". i .

y ‘ ‘ . f ‘ q. i . .
4 llil I 'if ltl‘lllll ill lij|l!..1[“‘ . All-IAA_L-J.-L4.-.L. L
‘1) 100 200 300 400 500 600 700

 

8

-‘ (1))

AAA 1AA

+TPP-Ethyl-YGGFL

O

 

A E A A A

O

   
 

wen-unenu—t (semen-mg“

 

 

 

 

 

 

 

 

 

Whig 1-3.‘1 11., I,

100 200300400 500600 700'860'900

A A

 

Figure 2.12. The FAB-MS spectrum corresponding to 1 nmol of YGGFL
reacted in a potassium phosphate/sodium borate buffer solution (pH 9.0)
with a 20-fold molar excess of a) vinyl-TPP and b) (2-bromoethyl)TPP. The
spectrum represents 1 nmol of YGGFL reaction mixture as isolated by the
Sep-Pak® procedure. The molecular ion of derivatized peptide (M+) appears
at m/z 844, and protonated underivatized peptide at m/z 556.

6 5

medium makes this reaction likely. Although (2-bromoethyl)TPP may
react directly with the amine group of a peptide to form the ethyl-TPP
derivatized product through a substitution reaction, elimination of HBr
from (2-bromoethyl)TPP may occur, producing vinyl-TPP, which, in turn,
may react with the peptide (or with water). Consequently, the addition of
water to vinyl-TPP may lower the derivatization yield for both reagents in
the aqueous derivatizing medium.

Reaction of vinyl-TPP and (2-bromoethyl)TPP with peptides to form
the N-terminally, ethyl-TPP derivatized products could theoretically
derivatize lysine and arginine residues because they contain primary
amine groups on their side chains. Buffering the aqueous medium at pH
9.0, and using pyridine (pr = 8.83) as a component of the non-aqueous
reaction medium, maintains the amino groups on the side chains of
lysine and arginine predominantly in their protonated form (pKa’s 10.8
and 12.5, respectively). However, at this pH, the N -terminal amino group
predominantly is unprotonated. Because the derivatization with (2-
bromoethyl)TPP liberates HBr, this also requires a bufl'ered system to
keep the reaction medium basic. Because a buffered system is necessary
for N -terminal derivatization, the use of the vinyl-TPP derivatizing agent
in an acetonitrile/pyridine 1:1 (v/v) volatile, buffered system is an effective,
convenient, and time-saving procedure.

ii) (3-Bromopropyl)TPP and (4-bromobutyl)TPP

Derivatization of leucine enkephalin with (3-bromopropyl)TPP

(Reaction 2-7) and (4-bromobutyl)TPP to form propyl-TPP and butyl-TPP

66

H
I ll
(2-7) Q—PL(CH2)3 —Br + H—(lY—C—C);1—OH——->

(C? it

H R
HBr + ©PL(CH2)3—(hll -C)n—0H

© 31':

(Reaction 2-8) modified peptides did not produce as high a FAB-MS

O
l M
(2-8) ©PL(CH2)4—Br + H—(If—c—C)g-0H-—->
H R

0 .
HBr + ©PL(CH2)4—(li1—$—(ll)n—OH
H R
response (Figures 2.13 a) and b), respectively) as derivatization with vinyl-
TPP. A higher FAB-MS response caused by underivatized YGGFL (m/z
556) was observed with the (3-bromopropyl)TPP and (4-bromobutyl)TPP
reagents than with vinyl-TPP, indicating the TPP reagents with a longer
alkyl chain have a lower reaction efficiency when reacted under identical
reaction conditions. The reaction efficiency for derivatization with (3-
bromopropyl)TPP and (4-bromobutyl)TPP cannot be related directly to the
FAB-MS response of the derivatized peptides because the propyl-TPP and
butyl-TPP modified peptides presumably have a higher hydrophobicity
and ionization eficiency than the ethyl-TPP derivatized peptides. The
lower derivatization efficiency observed with (3-bromopropyl)TPP and (4-
bromobutyl)TPP reagents may be caused by their lower solubility in the
reaction medium or perhaps because elimination of HBr to form a

reactive alkene-TPP intermediate (analogous to vinyl-TPP) is not favored.

67

 

8

  

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

1
5, 4 (a) " " Allyl-TPP“
1 4
a 80--
t .
1
V
e 604
I I
n 1
t 40-
0 4
n 4
' ‘ .
i 20
; . *TPP-Propyl-YGGFL
4, «U
200 400 ml 600 300 1000
I
100- , ,
‘1: : (b) TPP+ Vinyl-TPP"
1 4
a; 80 H20=CH(CH2)2-TPP+
i 4 '
V
e 60-
1 i
n 4
f, 404
n j 4 ‘ 0140+
: . YGGFL
t 20: c
y W . +TPP-Butyl-YGGFL
‘ LA -AAA.-.A A A A .A1 A
‘1. 200 400 m" (700 800 1000

Figure 2.13. The FAB-MS spectra representing 1 nmol of YGGFL reacted
with a) (3-bromopropyl)TPP and b) (4-bromobutyl)TPP. The reaction was
performed with a 20-fold molar excess of reagent in an acetonitrile/pyridine
1 :1 (v/v) solvent mixture. The molecular ion (M+) of propyl-TPP derivatized
YGGFL appears at m/z 858 and butyl-TPP at m/z 872.

6 8
2. The “Kunz” reagent
a) Introduction

Horst Kunz has introduced a protective functionality for amino
acids, which is useful for peptide synthesis. Because amino acids
derivatized with the 2-(triphenylphosphonio)ethoxycarbonyl (PEOC) group
show enhanced solubility in aqueous solutions [1 6], they are efl‘ective in
the synthesis of peptides. An N-terminally PEOC-protected amino acid
was coupled with a tert-butylesterified amino acid by
dicyclohexylcarbodiimide (DOC) to form a dipeptide which was reacted
with TFA to deprotect the C-terminus. The free carboxy group can be
reacted with another esterified amino acid to lengthen the peptide as
desired. The protecting group is extremely stable under acidic
conditions, but its removal is possible under mildly basic conditions
because the methylene group located a to the phosphorus is acidic,
making B-elimination possible. During peptide synthesis, the PEOC
peptide esters precipitate as an oil from solution, making extraction with
dichloromethane possible [17]. The PEOC moiety has also been used to
aid in the synthesis of other naturally-occurring compounds, including
depsipeptides [18] and O-linked glycopeptides [19,20].

However, there are several disadvantages with using the “Kunz”
reagent: i) the PEOC group is a bulky residue that hinders the
condensation reaction in peptide synthesis, ii) the high lability of the
PEOC group in presence of base, which is the basis of its removal,
demands special care in its synthesis, and iii) the PEOC protective group
is a cationic substituent whose halogen counter ion leads to problems

with glycosylation with mercury or silver-containing reagents [21].

6 9

b) Experimental

N -terminal peptide derivatization with the commercially available
“Kunz” reagent (Fluka Chemie AG; Buchs, Switzerland) was performed
with a twenty-fold excess of this reagent. The derivatizing agent was
dissolved in 1 mL of an acetonitrile/pyridine 1 :1 (v/v) mixture, the desired
amount of peptide was added to the reaction mixture, the solution was
vortexed, and the reaction solution was heated at 37°C for 3 h in a water
bath. The volatile solvents then were evaporated under a N-EVAP®, the
remaining solid was dissolved in H20, and the water was lyophilized in a
Speed Vac‘D. The remaining solid residue was dissolved. in a
water/acetonitrile 1 :1 (v/v) solvent mixture and placed on a FAB-MS probe
tip.

Alternatively, 10 uL of 2,4,6 trimethylpyridine (TMP) were added to
1 mL of acetonitrile containing a 20-fold molar excess of the “Kunz”
reagent. The desired amount of peptide was added to this solution and
the solution was heated for 1 h at 37°C. The volatile components of this
solution were evaporated under a N -EVAP®, dissolved in a 1:1 (v/v)
acetonitrile/water solution and placed on a FAB—MS probe tip for analysis.

c) Results and discussion

As demonstrated in Figure 2.14, the N -terminal derivatization with
the “Kunz” reagent (Reaction 2-9) is an eficient reaction, when judged by

70

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1%001 0 . TPP+ o .

6 J ‘ ___———V'lnyl-TPP+

1 4 / - - . +

:80? / A (2 Hydroxyethylyrg

i 4

V t

e 60‘

I 1

n .

z 40 VGGFL '

n 4 (M +H)"' +TPP-OCOEthyl-YGG‘FL

i 204 “mama-Yea

y 4 , L \ .
W All ulm “14* l 1- ltni

(b 100 200

300 400 ml: 500 600 700 800 900

Figure 2.14. The FAB-MS spectrum respresenting 1 nmol of YGGFL
reacted with a 20-fold molar excess of “Kunz” reagent in an
acetonitrile/pyridine solvent mixture. The peak representing PEOC
derivatized peptide (M+) appears at m/z 888, decarboxylated, ethyl-TPP
'derivatiwd peptide is at m/z 844, and underivatized peptide appears at m/z
556.

7 1
the amount of underivatized peptide remaining in the reaction medium.
In all the cases investigated, only a small amount of unreacted peptide
remained in the reaction solution. This indicates derivatization with the
“Kunz” reagent is an efficient reaction that is probably over 75% complete.
However, as Figure 2.14 demonstrates, there is a large amount of signal
at m/z 787, which corresponds to the peptide derivatized with the ethyl-
TPP group, rather than the ethoxycarboxyl-TPP group expected with
acylation of the N-terminus of the peptide with the “Kunz” reagent. The
ethyl-TPP derivatized peptide arises from decarboxylation of the 2-
(triphenylphosphonio)ethoxycarbonyl (PEOC) group as described in
Section F.2.b) of Chapter III. This side reaction causes formation of a
mixed product, which decreases the signal produced if a single product
formed, making this process a disadvantage. Nonetheless, the FAB
response due to peptide derivatized with the triphenylphosphine(2-
ethoxycarbonyl) group is significant, which makes this a quick, effective
derivatization procedure. On-line derivatization with the “Kunz” reagent
as described in Chapter III could be adapted as an effective method to
produce ethyl-TPP derivatized peptides in a matter of minutes.
3. (4-Carboxybenzyl)TPP

a) Experimental

N -terminal derivatization with (4-carb0xybenzyl)TPP was per-
formed in 1 mL of a water/TFA buffer (pH 5.0). A twenty-fold molar
excess of (4-carboxybenzyl)TPP and N-ethyl-N’-(3-dimethylamino-
propyl)carbodiimide-HCI (EDC) (Aldrich Chem. Co.; Milwaukee, WI) as a
coupling agent were added to the buffer before the desired amount of
peptide. An Eppendorf tube containing the reaction medium was placed
in a water bath 3 h at 37°C. The water/TFA buffer was lyophilized in a

72

Speed Vac® and the solid residue remaining was dissolved in 1-1.5 11L of
an acetonitrile/water mixture 1 :1 (v/v) before being applied to the FAB
probe tip containing 1 uL of glycerol.

b) Results and discussion

Derivatization of the N-terminus with (4-carboxybenzyl)TPP was
performed by acylation of the N-terminus with benzoic acid (Reaction 2-
1 0). The reagent contains a methyl-TPP group rather than a vinyl-TPP

o H 0
+ I I I EDC
(210) ©p—m,—©—c—OH + a—(If—(lz—o—on ———>
a n

Q—E- CH2 ‘Q-E -(liT-(%—E) —OH + 1120

H B
group. The vinyl-TPP group is the most abundant fragment with CAD-
MS/MS presumably because formation of this fragment involves cleavage
of a bond near to localized charge on the TPP group and because of the
stability of the vinyl-TPP species. Replacing the ethyl-TPP group with a
methyl-TPP group decreases the likelihood of vinyl-TPP elimination.
However, the molecular weight of the derivatizing group is significant
(380u), increasing the molecular weight of the derivatized peptide. The
eficiency of - this derivatizing reaction was low, because underivatized
peptide was observed in the reaction mixture (mlz 499) and the signal due
to derivatized peptide (mlz 87 7) was not very intense when the reaction
medium for (4-carboxybenzyl)TPP derivatization was analyzed by FAB-
MS (Figure 2.15). Improvement of the reaction efi'iciency was attempted
by converting the peptide into a methyl ester in a 2% (v/v) HCl methanol
solution, but little improvement in the FAB-MS response due to

derivatized peptide was observed probably due to the additional sample

73

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

R004 "'71“ ‘ ' (2-Hydroxyethyl)TPP+
b0

10 j (4-Car xybenzylyrgllrtm
a 80‘ YGGFL
P t
l 1
X sol +TPP(4-benzylcarboxy)-YGGFL
I 4
n .
t 40. l
8 4
n .
i8 20.1 - ‘2 .0
t 1 4
y 4 4 4 . l II

4 ‘ 4] 'I' H 4 ‘ ‘ ‘

(1) 200 400 m/z 600 800 1000

Figure 2.15. The FAB-MS response of 1 nmol of YGGFL when reacted
with a 20-fold molar excess of (4-carboxybenzyl)TPP and EDC in an
aqueous solution (pH 5.0) buffered with TFA. The peak representing the
molecular ion (M+) of derivatized YGGFL appears at m/z 934, the
protonated, underivatized peptide appears at m/z 556.

7 4
handling required for this procedure. Derivatization under more acidic
conditions (pH 2.0 and 3.0) was also attempted, but little increase in the
derivatization efficiency was observed. In fact, no (4-carboxybenzyl)TPP
derivatized peptide product was observed when the reaction was
performed at pH 2.0.
4. (2-Aminoethyl)TPP

a) Experimental

The C-terminus of peptides were derivatized by (2-aminoethyl)TPP.
A twenty-fold molar excess of (2-aminoethyl)TPP and EDC were dissolved
in an aqueous solution (pH 5.0) buffered with trifluoroacetic acid. The
desired amount of peptide then was added to the solution, the solution
vortexed, and placed in a water bath (37°C) for 3 h. The water and TFA
bufl‘er were lyophilized before the reaction mixture was analyzed by FAB-
MS or HPLC. This procedure effectively coupled the aminoethyl-TPP
group to the peptide through an amide bond.

b) Results and discussion

The coupling reaction utilizing the water-soluble coupling reagent
EDC effectively linked the carboxy terminus of the peptide with (2-
aminoethyl)TPP [1]. Carbodiimides are routinely used for for peptide
synthesis and also convert carboxylic acids to active acylating agents.
This reagent also has also been used for high-sensitivity solid phase
sequence analysis [23]. The EDC is used to couple peptides to solid
supports with the favorable activation of the carboxyl group. The acidic
conditions used in the derivatization procedure with (2-aminoethyl)TPP
protonate one of the nitrogens of EDC and deprotonate the carboxyl group
of the peptide. Addition of the carboxylic acid to a C=N bond of the
carbodiimide generates an O-acylated derivative. The (2-aminoethyl)TPP

7 5
amine reacts with this agent because there is a strong driving force for
elimination of the urea unit, with formation of a stable amide carbonyl

group (Reaction 2-11). The (2-amin0ethyl)TPP reacted with the C-

'11]

(2-11) H-(N—C —C) —OH _ _ + EDC
I l n 4' HaN CHz—Cl-Ig P~© ——>-
H R
'i' u
H—(N—ci—C) firearm—Ag» . 44,0
H R H

terminal carboxy group in high yield, producing a high FAB response
(Figure 2.16). Although (2-aminoethyl)TPP is expected to derivatize the
side chains of aspartic acid and glutamic acid residues (pKa’s 3.9 and 4.3,
respectively), the double-derivatized or the side-chain derivatized peptides
were not observed. This reaction probably does not occur because the
formation of the O-acylated derivative after addition of the carboxylic acid
to the C=N bond of EDC is sterically hindered when the side chain is in
the middle of the peptide chain. If this complex is formed, attack by (2-
aminoethyl)TPP is very disfavored due to steric hindrance. Because a 20-
fold molar excess of (2-aminoethyl)TPP was used, the possible side
reaction involving the coupling of two peptides through their carboxy
groups was not observed. In these investigations, only decamers and
shorter peptides were derivatized by (2-aminoethyl)TPP. However,
derivatization efficiencies would be expected to decrease for larger
peptides because secondary structure would be expected to produce
further steric hindrance. For these smaller peptides, derivatization
efficiencies of 95% yield were observed, making this an effective

derivatization procedure.

76

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1i°°4 ‘ i / (2-Aminoethyl)TPP+ _
i 4 /p Vinyl-TPP
:80 l .L—w/ Cyclic YGGFL
x I
v
.. 1.
I
n .
t .
e 40‘ «4140+ YGGFL-NHEthyl-TPP"
n 4 YGGFL
: J °
t 2°]
3’ 4 ° °
200 400 m/z 600 800 1000

Figure 2.16. The FAB-MS spectrum representing 1 nmol of YGGFL
reacted with a 20-fold molar excess of (2-aminoethyl)TPP and EDC in an
aqueous solution (pH 5.0). The peak representing derivatized peptide (M+)
appears at m/z 843 and underivatized peptide is at m/z 556.

7 7
5. Summary of derivatizing reagents

All the reagents discussed yielded derivatized peptides containing a
triphenylphosphonium group. One measure of the effectiveness of a
derivatization method is to evaluate the signal intensity representing the
derivatized product when analyzed by an instrumental method. To
compare the yield of TPP-derivatized product produced by these
derivatization methods, 1 nanomole of leucine enkephalin was reacted
with a twenty-fold molar excess of the derivatizing reagents listed in
Table 2.1 for 8 h at 37°C in the appropriate reaction medium.
Derivatization with the bromoalkyl-TPP reagents, vinyl-TPP, and the
“Kunz” reagent were performed in an acetonitrile/pyridine 1:1 (v/v)
solvent mixture. Derivatization of leucine enkephalin with (4-
carboxybenzyl)TPP and (2-aminoethyl)TPP was performed with a twenty-
fold molar excess of EDC in an aqueous solution (pH 5.0) buffered with
TFA. The signal representing derivatized leucine enkephalin after
derivatization with the various reagents observed during analysis of the
reaction mixture by FAB-MS is largely dependent on the efiiciency of the
derivatization reaction. However, the signal produced during analysis by
FAB-MS is not a direct measure of the derivatization yield because the
derivatized leucine enkephalin produced with the various reagents has
diverse structures, causing the surface activity of modified leucine
enkephalin to vary with the reagent used. However, the surface activity of
TPP-derivatized leucine enkephalin probably differs minimally because
the hydrophobic effects of the TPP group dominate the surface activity of
the derivatized peptide. The FAB-MS response was calculated as the ratio
of signal intensity representing derivatized leucine enkephalin to the

signal produced by underivatized, protonated YGGFL (mlz 556) during

7 8
analysis of the reaction mixture. Table 2.1 (below) indicates that the
highest FAB-MS response for N-terminally derivatized peptide was
produced by modification with vinyl-TPP and that the C-terminus was
efficiently derivatized by (2-aminoethyl)TPP.

Table 2.1. Ratio of signal intensity of derivatized leucine enkephalin to
that of underivatized, protonated peptide contained in the
reaction mixture during analysis by FAB-MS.

 

D'I” B | Bl' [5' III 'I
Vinyl-TPP bromide 16
(2-Bromoethyl)TPP bromide 3.1
(3-Bromopropyl)TPP bromide 2.4
(4-Bromobutyl)TPP bromide 1 .2
2-(tri1phenylphosphonio)ethyl
chloroformate chloride 0-15
(the “Kunz” reagent) (0.25 for decarboxylated prod.)
(4—carboxybenzyl)TPP chloride 2.0
(2-aminoethyl)TPP bromide 15

When evaluated by the ratio of signal intensity produced by modified
peptide to the signal intensity due to underivatized peptide, vinyl-TPP is
the most effective TPP-containing derivatizing reagent. Only the (2-
aminoethyl)TPP reagent modifies the C-terminus of peptides, but
fortunately produces a high ratio of signal representing derivatized
peptide when compared to the signal of unmodified peptide.
D. SeparationoftheC-tenninalderivatizationmixtm'e
1. Experimental

The derivatization mixture was separated with a gradient of 30% to

70% acetonitrile in 25 mins and holding the acetonitrile concentration

constant for 15 mins. The separation was accomplished with a Beckman

7 9

342 Gradient Liquid Chromatograph system (1 1 2 Solvent Delivery Module,
420 Controller, and 340 Organizer; Beckman Instruments, Inc.;
Fullerton, CA). The HPLC column used was an Alltech EconosilO 013
column (Alltech Associates, Inc.; Deerfield, IL) with dimensions of 10
mm x 250 mm. The stationary phase was contained on 10-um spherical
particles. The flow rate during chromatographic analysis was 2 mL per
minute. The mobile phase exiting the column was monitored by a UV-Vis
detector at 214 nm (Kratos Spectroflow 757 Absorbance Detector, Kratos
Analytical; Ramsey, NJ). The flow cell had an 8-mm pathlength and 12-
pL volume. Alternatively, smaller amounts of the peptide derivatization
mixture were separated on an Alltech Econosphere® C13 column. This
column has dimensions 4.6 x 250 mm with a 5-um diameter particle size.
The flow rate used for this column was 1 mL/min.

2. Results and discussion

Separation of the (2-aminoethyl)TPP reaction mixture by HPLC
effectively isolated C-terminally derivatized leucine enkephalin from the
derivatizing reagent and resulted in a chromatogram which contains five
major peaks (Figure 2.17). Comparison of the FAB-MS spectrum of 1
nmol of the unseparated reaction mixture (Figure 2.16) with 1 nmol of the
product separated by HPLC (Figure 2.18) demonstrates that the signal
due to (2-aminoethyl)TPP and EDC are decreased with isolation of the
derivatized product. The solvent fractions represented by these peaks
were collected and analyzed by FAB-MS. The first peak is caused by the
difference in refractive indices of the solvent that the reaction mixture
was dissolved in and the mobile phase, which commonly is called the
solvent front. The coupling agent EDC is represented by the second peak,
and (2-aminoethyl)TPP was the primary compound in the fraction

80

 

 

 

 

 

 

 

 

 

 

 

 

2.0—
E.
1.5- *5 FE +
_ ,2 >1 0..
A ‘3 5 E
.2. a .:>~ g
:3 ,’ g
\ 2 r: -— E _
1.0- 8 5* El
1 z s
' S ' >4
o 4 J
D “I:
.44 ‘43,
0.5- f
0 l I I l I l l I f f
TIniect 4 '8 12 16 20
Time(min)

Figure 2.17. HPLC separation of 20 nmol of the C-terminal derivatization
mixture of YGGFL. The reaction mixture contained a twenty-fold excess of
(2-aminoethyl)TPP and EDC. The compounds that the peaks represent are
identified in the chromatogram. The conditions of the separation are: flow
rate 2mL/min; column, 10 x 250 mm; stationary phase, 10 um diameter,
018 bonded phase; detector, 214 nm, 2.0 AUFS; chart speed, 30 cm/hr.

81

 

 

 

 

 

 

 

 

 

1004—0” .
R
e 4
1 4
a 80-
P ‘ .
l I
Z 60 Vinyl-TPP“
1 ' ((2 Amino:
3 40 V ”ethyl)TPP“
f; 1 / / YGGFL-NHEthyl-TPF‘
8
t 204 / . .
Y 4 a

‘ 1 4 4 I i” 1 4 .

. wi 4 ' 1 44 ‘ . 4 I , - . 4 4 I , . 4

300 400 500 600 700 300 900
m/z

Figure 2.18. Analysis of the peak representing separated, C-terminally
derivatized YGGFL by FAB-MS. The molecular ion (M+) of derivatized
YGGFL dominates the spectrum (m/z 843), although signal representing
vinyl-TPP and (2-aminoethyl)TPP (mlz 306) is observable.

 

 

 

 

 

 

R10: n. 4 s 2 1 (M);

e ‘ "n3 474 am 1'

l ‘ x": m 46 '3

z 8: Y": 7 as: as an 275 an

i 13' an) a -Y

v

e 6? fl .

I v c c r 1. c 1Iv (cm, (on), r-@

n J I"; I" ©

1: 4.

0 4

n 4 4

9 Ya

I

t 21 06332 a 537 a

y 1 1131 507 l: ‘8
‘ 1*.1 l 4. . l 44 .4‘ l . l . . ' : '1 , ‘4‘
H 400 500 600 700 800

 

Figure 2.19. The FAB-CAD-B/E spectrum of the molecular ion (M+) of C-
terminally derivatized YGGFL with (2-aminoethyl)TPP appearing at m/z
843.

8 2

collected when the third peak eluted. Underivatized YGGFL is
represented by the fourth peak, and cyclized YGGFL appears in the main
component of the fifth peak. The final peak represents C-terminally
ethyl-TPP derivatized leucine enkephalin. In the FAB-CAD-B/E
spectrum of derivatized leucine enkephalin (Figure 2.19), directed
fragmentation from the localized charge of the ethyl-TPP group attached
to the C-terminus is observed. All the fragments observed in the
spectrum are derived from the C-terminus. No ion series dominates the
spectrum and the x, and y" ion series form a complete series, which are
complemented by y 1 — y3 fragment ions. Thus, the complete amino acid
sequence of the peptide can be elucidated with ease. The ion at m/z 680
represents the loss of a tyrosine residue from the N-terminus. Losses of
57u, 57u, and 147u indicate the amino acid sequence from the N-terminus
is Y—G—G—F—L. The ion at m/z 417 (y 1) is characteristic of a (2-
aminoethyl)TPP derivatized leucine residue.
E. Quantitationof the C-terminal mactionyield

Quantitation of the reaction yield was the ultimate goal after
separation of the component of the reaction mixture was achieved by
HPLC chromatography. Once the reaction mixture is separated into
individual components, UV-Vis absorption measurements with a
detector on-line can be performed separately for each of the individual
components, provided the compounds are entirely separated with
baseline resolution. Other requirements of this procedure include the
necessity that no other components of the reaction mixture can co-elute
with the compound of interest when the absorbance is being measured
and the identity of the peak of interest must be positively confirmed. The
detected amount of absorbance caused by each separated component then

8 3
may be used to quanitate the amount of that component, if the absorption
coefficient for the compound is known. Separation of the (2-
aminoethyl)TPP derivatization mixture of leucine enkephalin (Y GGFL)
by HPLC with on-line UV-Vis detection was used to quantitate the
reaction yield.
1. Determination of the absorption coefiicient

a) Experimental

A large quantity (1.50 mg, 2.70 x 10'6 mol) of YGGFL was reacted to
yield an adequate amount of C-terminally derivatized peptide to be
weighed with a reasonable amount of accuracy. The peptide was
dissolved in 100 mL of H20 (pH adjusted to 5.00 with TFA) with a twenty-
fold excess of EDC (0.01035 g, 5.40 x 10'5 mol) and (2-aminoethyl)TPP
(0.01035 g, 5.40 x 10'5 mol) added. The solution was stirred overnight at
room temperature. The solution was concentrated by rotoevaporation at
room temperature and subsequently lyophilized. Transparent crystals
were observed when all the water was lyophilized. The crystals were
dissolved in 250 uL of acetonitrile and 10 11L portions of the reaction
mixture were separated by HPLC under semipreparative conditions. The
fractions containing C-terminally derivatized YGGFL were collected,
pooled, concentrated by rotoevaporation and lyophilized.

b) Results and discussion

After large scale derivatization, semipreparative separation of the
derivatized peptide from the reaction mixture by HPLC allowed collection
of a fraction containing C-terminally derivatized leucine enkephalin.
Evaporation of the solvents contained in the collected fraction by
lyophilization and a mechanical pump yielded a yellow-brown, oily solid
that weighed twice the theoretical yield of C-terminally derivatized

8 4
leucine enkephalin expected. The FAB-MS spectrum of the collected
fraction displayed a peak at m/z 149 that is characteristic of phthalate
compounds. These observations indicated phthalate impurities were
contained in the collected fraction. Due to the presence of impurities with
C-terminally derivatized leucine enkephalin, an accurate weight of C-
terminally derivatized leucine enkephalin could not be obtained for
determination of its absorption coefficient. Thus, direct, on-line
quantitation of the reaction yield could not be performed based on the
absorbance of the peak corresponding to derivatized peptide in the HPLC
chromatogram. Rather, quantitation of the reaction yield was based on
the loss of peak area observed for the peak in the HPLC chromatogram
representing underivatized leucine enkephalin following derivatization
with (2-aminoethyl)TPP.
2. Direct quantitation of the reaction yield

a) Experimental

About 75 11L of the C-terminal reaction mixture (20 nmol YGGFL
and a twenty-fold excess of (2-aminoethyl)TPP and EDC dissolved in a 1:1
(v/v) HzO/ACN solvent mixture after lyophilization of the original H20 (pH
5.0) reaction medium was separated by HPLC under conditions identical
to those described in the separation of the C-terminal reaction mixture
described above except UV-Vis absorbance was monitored at 260 nm.

b) Results and discussion

As was already mentioned, phthalate impurities in the collected
fraction containing derivatized peptide were detected, consequently direct
quantitation of C-terminally derivatized YGGFL on-line was invalid
because an accurate extinction coefficient for this product could not be

obtained. Separation of the reaction mixture by HPLC produced a

8 5
chromatogram similar to that contained in Figure 2.17. The fourth
major peak in the chromatogram represents underivatized leucine
enkephalin. Consequently, quantitation of the C-terminal reaction yield
was based on the difference in absorbance observed for the peak
representing underivatized leucine enkephalin when 20 nanomoles of
this peptide were injected alone and the absorbance observed when 20
nanomoles of reaction mixture were injected. Absorbance was monitored
at 260 nm because the phenol group of the tyrosine residue and the phenyl
group of the phenylalanine residue in YGGFL absorb radiation at this
wavelength whereas the other portion of the peptide has little absorbance
at this wavelength. Although a side product of cyclized leucine
enkephalin is observed by the coupling of the N - and C-terminus by EDC,
the amount of side product produced by this reaction can be quantitated
accurately at 260 nm because the phenyl and phenol groups remain intact
during cyclization and the new structure of the peptide produced after
cyclization causes little difference in absorbance. The amount of side
product detected then can subtracted from the yield calculated by the loss
of signal intensity for the peak representing underivatized leucine
enkephalin after derivatization. Taking into account the quantity of side
product formed by cyclization of YGGFL, the yield of C-terminally
derivatized leucine enkephalin was 59 % after reaction with a twenty-fold
molar excess of (2-aminoethyl)TPP 3 h at 37°C.
F. FAB-CAD-BIE fragmentation of derivatized VGVAPG
1. Introduction

Investigations on the peptide VGVAPG derivatized with various
reagents containing the triphenylphosphonium group were performed by
FAB-CAD-B/E. These investigations will demonstrate the advantages

8 6

and the complications encountered when utilizing these reagents.
Although the TPP-containing derivatizing reagents were used to modify
other peptides, only the modification of VGVAPG with these reagents and
analysis by FAB-CAD-B/E will be reported here for the sake of brevity.
The fragments observed for the TPP-derivatized peptides containing a
localized charge likely are produced by remote-site fragmentation,
similar to the fiagmentation of fatty acids as described by Gross [24].

One should be aware that the m/z values of the peaks representing
fragments of the TPP-derivatized peptides (beyond the mass shift
associated with derivatization) differ from the m/z values observed for
underivatized peptides, although the same nomenclature is used in
symbolizing these fragments. This is because the peptides derivatized
with the TPP group contain a “preformed” charge. Consequently, the
derivatized peptides are not protonated and fragment by remote-site
processes. For N -terminally derivatized peptides, an-H, bn-H, and cn-3H
fragments are observed (A-H, B-H, and C-H according to the Roepstorff
nomenclature), insted of the a... b... and c" ions observed with
underivatized peptides. In the mass spectra of C-terminally, TPP
derivatized peptides, xn-H, yn-3H, and zn-H ions are observed (X-l, Y-l,
and Z-l according to Roepstorff nomenclature), rather than the x”, y...
and 2,. ions observed in the fragmentation of underivatized peptides.
Thus, the y" fragments observed are 3 mass units lower in mass than the
y, (Y” according to Roepstorff nomenclature) fragment expected for the

underivatized peptide.

8 7
2. Mass spectrometry

a) Instrumentation

FAB has greatly increased the number and variety of compounds
that can be analyzed by a mass spectrometer. Particularly because high-
mass molecules can be ionized by this technique, large-sized mass
spectrometers are necessary to analyze these molecules at high
resolution and sensitivity. The JEOL JMS-HX-llO is an ultra high-
performance mass spectrometer that meets these demands. A resolution
over 100,000 is possible because the JEOL HX-110 employs a perfect
second-order double-focusing system with an ion optics system whose
second-order aberration is less than one-tenth that of conventional
instruments. The velocity dispersion of isobaric ions that does exist in the
electric field is compensated for by direction-focusing of the magnetic
field. Ions are focused with such accuracy with instruments of the Nier-
Johnson design that the exact mass of ions can be measured to six
significant figures and ions of the same nominal mass, but different
elemental composition are separable. High sensitivity is possible because
ion beams of a wide angle distribution and energy width are converged by
a combination of quadrupole lenses and a large, uniform, and cylindrical
magnetic field. This configuration reduces ion loss in the vertical
direction by a factor of ten-fold. The JEOL HX-110 can detect ions with a
mass-to-charge ratio (mlz) over 15,000 with a 30-cm radius electromagnet
with a maximum strength of 1 8.8 kilogauss. Automatic evacuation of the
direct probe inlet and ion source make this instrument ideal for FAB-MS.

b) Operating principles

The fragment ions produced by collisionally-activated dissociation
in the first field-free region of the mass spectrometer can be separated

8 8
according to their mass in a double-focusing instrument with the B/E
linked scan. Linked scanning is performed at a constant accelerating
voltage. The electric and magnetic fields are scanned simultaneously so
that the ratio of their field strengths is constant [25]. These scans .
normally are carried out under computer control and prior mass
calibration is used to produce accurate mass-to-time correlations based
on the scan rate.
In the linked scan, the spectrum consists of product ions formed in
a collision cell placed close to the source-defining slit from a chosen
precursor ion. When the precursor ions (m1+) fragment to produce
product ions (m2+ ), these ions require different electric field strengths, E1
and E2, respectively, to be transmitted through the electric field sector,
and magnetic fluxes, B1 and B2, to be transmitted through the magnetic
sector. Because these scans are run at a constant accelerating voltage, V,
the energies of all the ions leaving the source are equal. Ions with mass,
m, and velocity, v, pass through the electric and magnetic sectors to be
detected when the electric field strength, E, meets the conditions
(Equation 2-3), where R, is the radius of the electric sector:
(2-3) eE = vaRe
For the magnetic sector field strength (Equation 2—4) where Rm is
the radius of the magnetic sector:
(24) eB = mv/Rm
Consequently, the conditions needed to pass m1+ and m2+ in the
electric sector are (Equations 2-5,6):
(2-5) eE1 = m1v12/Re
(2-6) eEz = m2v12/Re

 

89

The conditions required to pass these ions in the magnetic sector
are (Equations 27,8):

(27) eB1 = m1V1/Rm
(2-8) e32 = m2V1/Rm

Combining these equations gives Equation 2-9:
(2-9) m1/m2 = E1/Ez = 131/32

Consequently, B1/E1 = B9/E2 = constant. The value of the constant
or ratio of magnetic field strength to electric field strength (B/E) is
determined by the electric and magnetic field strengths needed to
transmit the precursor ion. The resolution needed to separate the
fragment ions is dependent on the translational energy released during
fragmentation. The resolution normally achieved in linked scanning
makes it possible to assign the product ion mass accurately to the nearest
integer. The resolution by which the precursor ion can be selected is only
from 300 to 400; consequently, m1+ ions containing one or more 13C
isotopes will be collected [26].

c) Experimental

FAB desorption of protonated and derivatized peptides was
performed with a 6-keV beam of xenon atoms. The desorbed ions were
analyzed in a JEOL HX-110 double-focusing mass spectrometer. The
resolution was set at 1000 and the accelerating voltage remained at 10 kV
for all scans. A JEOL DA-5000 data system generated linked scans at
constant B/E to produce CAD-MS/MS spectra. Helium was used as the
collision gas in the first field-free region to produce CAD.

3. Analysis of underivatized VGVAPG by FAB-CAD-B/ E

To compare the fragmentation of TPP-derivatized peptides with

underivatized VGVAPG, the FAB-CAD-B/E spectrum of underivatized

9 0

peptide is presented in figure 2.20. The I), and y, fragments resulting
from fragmentation of the amide bond [C(O)—N(H)] dominate the mass
spectrum. The amide functionality is the most basic functionality in
VGVAPG and is the site of predominant protonation. Consequently,
fragmentation of this functionality in the peptide backbone is probably
caused by charge-initiated chemistry. The 1),. fragments are fragments
resulting from cleavage of the peptide bond and retention of charge on the
N-terminal side of the molecule. The y, ions are formed by charge
retention on the C-terminal portion of the molecule and by hydrogen
transfer from the neutral fragment produced by fragmentation at the
amide bond. Specifically, protonation of the amine group of the amide
functionality occurs, and a hydrogen shift from a N (H)n group of the
neutral fragment causes formation of the y” fragments. In the FAB-
CAD-B/E spectrum of underivatized VGVAPG, the fragments
representing cleavage of the [CH(R)—C(O)] bond (an and x" ions) and the
[N(H)—CH(R)] bond (0,. and 2,. ions) are not nearly as abundant as
fragmentation of the amide bond.

Immediately, one observes that fragments (y, and bu) resulting
from cleavage of the peptide bond dominate the spectrum of underivatized
VGVAPG (Figure 2.20), because the amide bond represents the most
basic site in peptides that do not contain basic amino acid side chains.
Examining the low-mass end of the spectrum, a discernible immonium
ion is observed at m/z 72, corresponding to the mass of the valine
immonium ion, indicating that valine is present in the peptide being
analyzed. The high-mass region of the spectrum displays the loss of
water and also the side chain of a valine residue. Observing the direct

loss of residues (the y, series) from the protonated molecule (M+H)+, a

91

com 001"‘
1 v.»
.3».
>1 no on * a“
Cum-
+Am+u$

 

 

.mmv a): an 2 +Am+2v canoe—o8 33:38AM
25. finger» 833:85 as 5884 Emdééé 23. :8 u 28E

as

com

3 no

 

 

3

”a

an

 

 

 

0A.:
in
.H
m
4N
v a N a 3 fl
a use
"no 4
5 8a a: "an .m
a ”an .
< > Mpg—I .
”as .
a v a n: 4
r
m

 

mvfldum>0 Hflumfimwuh

9 2

small peak at m/z 400 (the y5 fragment) represents a reasonable loss,
indicating a valine residue resides at the N -terminus. Subsequent losses
of 57a, 99u, 71u, and 97u are represented by the y1 - y; fragments
appearing at m/z 343, 244, 173, and 76, respectively. The 25 + 1 fragment
confirms the assignment of the y5 fragment ion. Surprisingly, a y1 ion
can be observed at m/z 76, indicating that glycine resides at the C-
terminus. Consequently, the y" ion series identifies the amino acid
sequence from the N-terminus as V—G—V-A—P—G.

The b" ion series begins at m/z 424; this peak represents the loss of
a glycine residue (57u) and a water molecule from (M+H)+. The bu ion
series continues at m/z 327, 256, and 157, representing the b4 - I):
fragment ions. The bu ion series is confirmed by peaks representing the
a4, a5, c3 and c5 fragment ions, which appear at m/z 228, 299, 396, and
272, 441, respectively. The amino acid sequence deduced from the b,
series is ?-V-A-P-G. The amino acid sequence generated from the C-
terminus (b, ion series) verifies the amino acid sequence generated from
the N-terminus (the y" ion series), allowing the amino acid sequence for
this peptide to be deduced from these overlapping series of fragment ions.
The deduced sequence is VGVAPG. However, the spectrum of the
underivatized peptide is complicated by the generation of fragments from
both termini.

4. (2-Bromoethyl) TPP and vinyl-TPP derivatized VGVAPG

Analysis of the vinyl-TPP and (2-bromoethyl)TPP (Figures 2.21 a)
and b) respectively) derivatized VGVAPG revealed similar fragmentation
patterns, confirming that similar peptide modification is produced by
derivatization with these reagents. The fragmentation pattern for the N -
terminally, ethyl-TPP derivatized peptide changed dramatically when

93

 

‘4‘

H
-..“?

wowmnonnm o<wve—ow
QR.

 

? (a)

a1

 

as

 

.. c8

13 .1

man-om
a
d

GVAPG—OH

as -V

(“)4-

 

 

H
$.4.°

A l- L A A

.".‘. °’

t0
l A A

L

«er—managed o<~wn~ogu

 

A L

O

81

M

c1a,

400

”to:

500

83

 

m/z

b‘c‘

 

Figure 2.21. The FAB-CAD-B/E spectra of the molecular ion (M+) of N-
terminally, ethyl-TPP derivatized VGVAPG by a) vinyl-TPP and b) (2-
bromoethyl)TPP (mlz 787).

9 4
compared to that of the underivatized VGVAPG. With localization of
charge at the N-terminus, a, fragment ions dominated the mass
spectrum. All the ions observed in the spectrum correspond to N-
terminal ions (an, b», c,. and d"), however, the b, and en fragments have
much lower signal intensity than the an fragment ions. For this peptide,
a complete series of an, ha, and c, ions are observable. However, the
signal of the a: fragment ion is minimal. The formation of an fragment
ions is believed to occur through a 1,2-elimination reaction. The
fragmentation mechanism involves an H-shift from the b-carbon on the
amino acid side chain. Because the glycine residue does not contain a b-
carbon with attached hydrogens, formation of the a: fragment is
disfavored. However, the small abundance of as fragment ions that
appear probably are formed by another mechanism or perhaps involve a
hydrogen shift from one of the a-carbons. The spectrum of the ethyl-TPP
derivatized peptide provides informative cleavage of the amino acid side
chains ((1,. ions) and the complete ion series produced allows the complete
amino sequence of VGVAPG to be elucidated quickly.
5) (2-Aminoethyl) TPP derivatized VGVAPG

Derivatization of VGVAPG at the C-terminus by (2-aminoethyl)TPP
and subsequent analysis by FAB-CAD-B/E linked scanning also provides
a large amount of structural information (Figure 2.22). All the
fragments are formed by charge-remote fragmentation from the localized
charge at the C-terminus, due to the quaternized TPP group. The y,
fragments are the most intense in the spectrum, however, the x", y", z",
and v, ion series are easily recognizable in the FAB-MS spectrum.
Complete series of x" and y" ions were observed for this modified peptide,

but not a complete series of 2,, ions. However, the incomplete 2,. series

 

.83 Se dances?
-8 .3 Buses 8 em<>o> 883:8 $3.38 53553
.0 no ALE com 838—08 23 me 280% Sm.9<0.m<.m SE. dud earn

95

 

 

 

 

 

 

 

 

 

 

 

as
GE. 02. O8 cow 0mm com and.
...1 ; ...
an 4 an «N
J 4.» N 4.. a.»
v> an
a.»
u>
@ ._.
©I+m «58 «E8 2 m c 41% .> e.,->lz«m
0
© 53 QB v.5 PW ”EN
>. «8 Eh g Q» k L RM "ch
k a “an
o8 ”:3
Bu\ iv “A; «h
H a a v a ..c

 

 

Momaumbo HGH°QMHHH>S

96

was augmented by the formation of v" and w” ions, which result from
side chain cleavage of the 2,. + 1 ions. The abundance of the w" and v"
ions is enhanced by ethyl-TPP derivatization, from which significant
structural information about the amino acid side chains can be obtained.
From this mass spectrum, enough information is discernible to provide a
complete amino acid sequence and the spectrum is simplified because
only fragments for the C-terminus are produced.
6. (3-Bromopropyl)TPP derivatized VGVAPG

Although derivatization of VGVAPG by (3-bromopropyl)TPP was
not as efficient as derivatization with vinyl-TPP, a sufficient amount of
derivatized peptide was present to perform analysis by FAB-CAD-B/E
(Figure 2.23). Losses of the valine and alanine side chains from the
molecular ion of VGVAPG derivatized with (3-bromopropyl)TPP was
observed in the high-mass region of the spectrum. From the N-terminus
of derivatization, a complete series of an, b", and c, ions could be
observed. However, these ions were not as intense as those produced by
ethyl-TPP derivatization. The most intense peaks in the spectrum (m/z
339, 383, and 385) were produced by fragmentations that were not
accounted for. The peak at m/z 321 is due to cleavage of the [N(H)1 -
CH(R)1] bond and retention of charge on the modifying group. The peaks
at m/z 339, 383, 385, 444, 499, and 614 also could not be accounted for. The
c: fragment also seems unusually abundant. These results indicate that
an increase in the alkyl chain length of the TPP-derivatizing group
complicates the mass spectrum, perhaps causing fragmentation by

alternative mechanisms.

97

gm

 

 

+95

 

 

NOIOA<>U

>1? game
8»

as

No

A an use 8522888;
-8 fies 888388 8% cm<>c> 38.328 83.888 5:383
-z .8 Ea 8... 8:528. 2: .s 8888.. Eméééé 2:. dad 28a...—

 

a...

cow

man

 

wm

 

 

 

fir V T
H

v f

 

 

m0~w8-~>o Hduoflmmuh

9 8
7. (4-Bromobutyl)TPP derivatized VGVAPG

The FAB-CAD-B/E mass spectrum of VGVAPG derivatized with (4-
bromobutyl)TPP was similar to the spectrum of VGVAPG derivatized
with (3-bromopropyl)TPP, except the fragments were shifted upward in
mass by 14u (Figure 2.24). Again, fragments from the modified N-
terminus predominated in the spectrum because the butyl-TPP modifying
group contains a localized charge. A complete series of an ions was
observed in the mass spectrum, which is complemented by a nearly
complete series of bu and c" ions. At the high-mass end of the spectrum,
the losses of alanine and valine side chains provide a partial amino acid
composition. For the butyl-TPP derivatized peptide, the relative
abundance of the bu and cu fragment ions have a greater signal intensity
relative to that of the an ions. This may indicate that placing a fixed
charge more distal from the peptide (increasing the alkyl chain length)
causes the likelihoood of backbone fragmentation to be more equal for
each bond. Peaks also appear in this spectrum that can not be accounted
for. The ions at m/z 397, 399, 499, 560, and 628 can not be accounted for.
The peak appearing at m/z 335 is due to cleavage of the [N(H)1—CH(R)1]
bond, with charge retention at the derivatized portion of the molecule.
Perhaps these fragments are indicative of an alternative mechanism of
fragmentation. The high as fragment ion abundance also seems
unusual. Although these unknown fragments are observed for butyl-TPP
derivatized RVYVHPF, the unidentified peaks were not nearly as
intense.

8. “Kunz” reagent derivatized VGVAPG

The FAB-CAD-B/E spectrum of VGVAPG derivatized with the

“Kunz” reagent displays a simplified fragmentation pattern (Figure 2.25)

99

com

 

 

 

+35

 

 

:.
_
_

 

 

3

.a

MOI-G m <4 > 0 ?T game
. 5»
"no
1 A.

_ E

at

.38 c8 8:55.888 £3 8382 sec
822.1 8 88» 95-38 a £5 8:832 2: a Becca 24>?
co £6 :2 ceases 2: co a338,. Em-n<o-m<m 2a. .43 gr...

as

com

: _

4 .
4 am or

 

8’

no

255:"? .3“

98 as

me
We

 

«A.

 

 

3m

8m

«mm

 

 

3m.

 

 

mQu—cfl-JJu-t>0 Hfiuofimwuh

100

.33 say 039% 853:8 comm 20R
3 gauge ..Ear 2: £3 Euuguwu ma 84>? Bungtwu magmas
-z .3 £6 :3 3:538— 2: he again Emdééfi 2a. and 25E

com com com com cow

 

” u
a u 3 m5

>. H ‘b 0 Q3 Q
Iwuo

solo. a < > E >l

 

 

 

 

Z—fl:

 

 

 

gr

”8
izvmvmnfi

333.23

«a

 

 

 

Didi—.dum'ém Hduonmwuh

l 01
comparable to that of VGVAPG modified at the N-terminus by ethyl-TPP,
except that peaks representing fragment ions were shified upward in
mass by 44u. In this spectrum, only N-terminal ion series are observed,
with the an ion series dominating the mass spectrum, which is
complemented with the ha and c, ion series. From the fragmentation
pattern of the “Kunz” derivatized peptide, the amino acid sequence is
easily identified to be VGVAPG. However, the peaks at m/z 481 , 649, and
741 cannot be accounted for. The increased formation of d, ions also
facilitates elucidation of the amino acid sequence by informative
fragmentation of the amino acid side chains.
9. (4-Carboxybenzyl) TPP derivatized VGVAPG

The product of derivatization of VGVAPG with (4-
carboxybenzyl)TPP also was analyzed by FAB-CAD-B/E (Figure 2.26).
Fragments derived from the modified N -terminus dominated the mass
spectrum. The (4-carboxybenzyl)TPP modifying group causes
predominance of the an ion series, which represents a complete series.
In addition, complementary b, and en ion series are observable in the
spectrum at lower abundance. The high-mass region of the spectrum
displays side chain losses of alanine and valine from the molecular ion,
providing a partial amino acid composition. Peaks corresponding to the
descriptive (1,. ions are also observed. However, all the peptides
derivatized with this reagent displayed a peak at m/z 595, but the source of
this peak cannot be accounted for. These results indicate the (4-
carboxybenzyl)TPP derivatizing group effectively causes the formation of
a... fragment ions for facilitated peptide sequencing.

102

$5 £5 9.2»
mmhauamfiuoeswev 2: .3 guess 8 3.33 8528 £3353
.2 no is :3 3:338 2: me $88QO Emdééfi. 2:. .93 25E

 

 

 

 

 

 

e8
cow cor com can
:sw: ..i ... :.; n.2a . __ ,.
n m . . . . ., O «Q «a
0 fl «.0 _. a”
3
a“ v“
a
<. >. a .
A
._. © .
N .
mole m < > e.>#ZImIAUY;8.ElflV .
. 0 RM mg m
u G .
8n 3 a ".5 .
fi "co
3 Sic
we an ad .

 

 

3

 

Eu

mmmflam>fl Hnuonmwum

99°F.“

10.
11.

12.

103

References
. Wagner, D.S.; Salari, A.; Gage, D.A.; Leykam, J .; Fetter, J .;

Hollingsworth, R.; Watson, J .T., Biol. Mass Spectrom., 1991, 20, 419-
425.

Watson, J.T.; Wagner, D.S.; Chang, Y.-S.; Strahler, J .R.; Hanash,
S.M.; Gage, D.A., Int. J. Mass Spectrom. and Ian Process, 111,
1991, 191-209.

Wagner, D.S.; Salari, A.; Fetter, J .; Gage, D.A.; Leykam, J.;
Hollingsworth, R.; Watson, J .T., Proceedings of the 38th Annual
Conference on Mass Spectrometry and Allied Topics, Tucson, AZ,
pp. 307-308.

Wagner, D.S.; Gage, D.A.; Allison, J .; Watson, J .T., Proceedings of
the 39th Annual Conference on Mass Spectrometry and Allied
Topics, Nashville, TN, 1991, pp. 1330-1331.

Wagner, D.S.; Nieuwenhuis, T.J.; Chang, Y.-S.; Gage, D.A.;
Watson, J .T., Proceedings of the 40th Annual Conference on Mass
Spectrometry and Allied Topics, Washington, DC, 1992.

BS. Wagner, Ph.D. Dissertation, Michigan State University, 1992.
Friedrich, K.; Henning, H.-G., Chem. Ber., 1959, 92, 2756-2760.

Still, W.C.; Kahn, M.; Mitra, A., J. Org. Chem., 1978, 43, 2923-2925.
Keough, P.T., Grayson, M., J. Org. Chem., 1964, 29, 631 -635.
Schweizer, E.E.; Bach, R.D., J. Org. Chem., 1964, 29, 1746-1751.
Chistol, H.; Cristau, H.J.; Soleiman, Tetrahedron Lett., 1975, 16,
1381-1384.

Dubois, H.J.; Lin, C.-C.L.; Beisler, J .A., J. Med. Chem., 1978, 21(3),
303-306.

13.

14.

EEFSEE

.52

l 04
Sanyal, U.; Chatterjee, R.S.; Das, S.K., Chakraborti, Neoplasma,
1984, 31(2), 149-155.
Dean, J .A., Ed., Lange's Handbook of Chemistry, 13th Ed., 1985, Me-
Graw-Hill, Inc., NY, Table 5.8, 5-55.
Swan, J .M.; Wright, S.H.B., Aust. J. Chem., 1971, 24, 777-783.
Kunz, H., Chem. Ber., 1976, 109, 2670-2683.
Kunz, H., Angew. Chem. Int. Ed. Engl., 1978, 17, 67-68.
Kunz, H.; Lerchen, H.-G., Angew. Chem. Int. Ed. Engl., 1984,
23(10), 808-809.
Kunz, H.; Dauth, H., Leibigs Ann. Chem., 1983, 3, 337-359.
Kauth, H.; Kunz, H., Leibigs Ann. Chem., 1983, 3, 360-366.
Kunz, H., Angew. Chem. Int. Ed. Engl., 1987, 26, 294-308.
Sheehan, J .C.; Ledis, S.L., J. Am. Chem. Soc., 1973, 95, 875.
Aebersold, R.; Pipes, G.D., Wettenhall, R.E.H.; Nika, H., Hood, L.E.,
Anal. Biochem., 1990, 187, 56-65.
Jensen, N.J.; Tomer, K.B.; Gross, M.L., Anal. Chem., 1985, 57, 2018-
2021.
McLafl'erty, F.W., Ed. Tandem Mass Spectrometry, John Wiley and
Sons, New York, N.Y., 1983.
Jennings, K.B.; Dolnikowski, G.G., Methods in Enzymology, 1990,
193, 37-61.

Chapter III

DEVELOPMENT OF ON-LINE
ETHYL-TRIPHENYLPHOSPHONIUM DERIVATIZATION

A. Introduction

Analyte derivatization is a necessary evil for many chromatographic
separations to increase the detectability of targeted solutes. Advances in
the design of faster and more selective reactions, synthesis of new
derivatizing reagents, and development of novel derivatization methods for
almost every type of compound have occurred in the last several years.
Postcolumn derivatization offers inherent advantages over precolumn
methods in terms of precision and ease of automation; a single reaction
pathway is not necessary. However, these schemes require more care in
design to minimize extracolumn broadening, and they require reasonably
fast reaction kinetics [1].

Derivatization can be performed by homogeneous solution reactions
or heterogeneously with a solid phase reagent. Solid phase reactors
employing enzymes or chemical reagents are becoming increasingly
p0pular, especially in combination with flow injection analysis because the
disadvantages associated with homogeneous reactions can be overcome
with use of a solid phase reagent [2]. Krull’s group has been especially
active in the area of solid phase derivatizations. His group has reported the
use of polymeric agents for weak nucleophiles, including primary and
secondary alcohols [3], amines, amino alcohols, and amino acids [4]. They
also devised a mixed bed reactor that contains several different reagents to
simultaneously prepare derivatives of several functional groups on a single

105

106
analyte [5]. A mixture of silica gel and a polymeric anhydride containing

the o-acetylsalicyl group as the labeling moiety is useful for the
simultaneous collection and derivatization of airborne primary aliphatic
amines with a solid phase reactor [6]. Two different polymeric reagents to
derivatize fatty acids have also been described [7 ,8].

By placing a solid phase reactor on-line with a high performance
liquid chromatography (HPLC) column and a mass spectrometer, the
possibility of injecting picomole amounts of an enzymatic digest into this
apparatus and deducing a protein's entire amino acid sequence just hours
later is foreseeable. This is the ultimate goal for developing a system that
will combine a reactor for ethyl-TPP derivatization, HPLC separation, and
FAB-CAD-MS/MS. Further development would include an enzyme reactor
that could perform enzymatic digestion of proteins on-line [9]. Vestal et al.
reported the HPLC-thermospray MS detection of peptides resulting from
digestion in a reactor containing immobilized enzymes [1 0-1 3]. Voyksner et
al. [14] reported a 1 0 to 40-fold signal enhancement for peptides obtained
by hydrolysis in a enzyme reactor over unhydrolyzed peptide. The
increased sensitivity allowed detection and qunatitation of hydrolysis
products down to the 800 fmol level.

On-line coupling of HPLC and mass spectrometry (LC-MS) has
already been realized and has made a major contribution to analytical
methodology in protein chemistry [1 5]. One LC-MS configuration utilizes a
sample introduction probe that provides a continuous flow of solvent that
may contain sample into the mass spectrometer ion source and onto a frit
where FAB ionization takes place. The mobile phase flow rate is in the
range of 3-10 uL’min, generally. For peptide analysis with LC-MS, 1-5%
glycerol is added to the carrier solution to maintain stable operating

107
conditions. There are two advantages for using continuous flow CF-FAB-

MS in preference to FAB with a direct probe: i) lower limits of detection are
possible (the low picomole level for peptides) and ii) there is a decreased ion
suppression effect. Typically, packed, capillary-bore, fused-silica columns
are employed with LC-MS because they offer good separation and increased
sensitivity at low flow rates (3-10 [LL/min). although they are fragile and
dificult to work with.

Complex mixtures of tryptic peptides routinely are analyzed with
packed capillary columns coupled on-line with FAB-MS at picomole levels
[16]. Packed capillary columns allow all the injected peptide to be
introduced into the mass spectrometer and consequently are advantageous
for separating complex mixtures of peptides derived from the enzymatic
digestion of proteins. Enzymatic digestion in combination with LC-MS
quickly verifies the primary structure of recombinant proteins by a
technique called peptide mapping. When tandem MS is utilized in
combination with HPLC, identification of posttranslational modifications of
peptides and single amino acid substitutions is possible. Improved
sensitivity, minimized suppression of hydrophilic peptides, and efficient
sample introduction are just some of the advantages of coupling HPLC
online with FAB-MS.

B. Advantages of an on-line derivatization reactor

In general, the advantages already realized with the use of solid
phase reagents for sample derivatization in combination with
chromatography include: i) fast, mild, and efficient derivatization,
ii) improved reagent stability, iii) increased selectivity and production of
fewer side products, iv) less contamination caused by an excess of

derivatization reagent, v) higher reaction capacities due to high

108

concentration of derivatizing reagent, vi) ease of regenerating the solid
phase reagent, and vii) the option of performing on-line derivatization pre-
or post-column. In some cases, it is possible to filter and reuse the
derivatizing reagents.

Krull and co-investigators have developed several reactors for on-line
derivatization of amines employing a derivatizing agent covalently attached
to a polymeric support [1 7-1 9]. Covalent linkage of the derivatizing agent
to a polymer provides the best stability for the reagent and causes the
polymer-bound derivatizing reagent to be compatible with a larger range of
solvents. Derivatization yields near 90% for primary amines and 70% for
secondary amines were realized with these reactors. Detection limits in the
low parts-per-billion range were achieved in conjunction with improved
sample separation, achieving rapid, accurate, and precise quantitation of
amines in real-world sample matrices.

The proposed setup for on-line, ethyl-TPP derivatization in combina-
tion with LC-MS has several advantages beyond those common to solid
phase derivatization. Because only low-picomole amounts of peptides are
available for analysis in many cases, the most important advantage of
combining peptide derivatization, separation, and mass analysis on-line is
the reduced sample losses resulting from elimination of the sample
transfers necessary if these procedures were carried out manually. The
proposed setup should provide high derivatization yields, low detection
limits, increased sensitivity, fewer interferences, and faster quantitation
than is available by manual performance of these procedures. Further,
because these procedures will be automated, they can be performed quickly

and with decreased labor input.

109
The second important advantage of an apparatus to perform peptide

derivatization with the ethyl-TPP moiety on-line is the increased amount of
information generated with analysis by FAB-CAD-MS/MS of the
derivatized peptide beyond that obtained for the underivatized peptide.
Covalently linking the ethyl-TPP group to a peptide increases its
hydrophobicity and places a positive charge at a fixed position on the
peptide. Therefore, ethyl-TPP derivatization enhances the efficiency of
ionization for hydrophilic peptides and generates a series of ions from the
terminus at which the peptide has been modified. Fixing a positive charge
at the N-terminus or C-terminus facilitates the production of d" or w"
ions, respectively, which helps distinguish amino acid residues in the
peptide; these ions are especially helpful in difl'erentiating the isomeric
amino acids leucine and isoleucine.
C. Proposed setup

The proposed apparatus for on-line derivatization of peptides
consists of three major components: a reactor, a packed microbore LC
column, and a mass spectrometer capable of LC-MS (Figure 3.1 ). A
microsyringe pump will supply the solvent at low flow rates (3-10 uL/min).
The sample of peptide fragments obtained from an enzymatic digest will be
introduced to the derivatization reactor through an injector loop. Once
inside the reactor, the peptides will be tagged with the ethyl-TPP moiety.
The derivatized peptides then will pass through a reverse phase, packed
microcolumn for chromatographic separation. If the derivatization reactor
is pre-column, the derivatized peptides will be more homogeneous in
hydrophobicity and chromatographic properties, probably requiring a
shallower solvent gradient for separation. Following separation, the

presence of derivatized peptides will be indicated by an on-line UV-Vis

110

 

 

Reactor LC Column X0919."

 

 

 

 

 

 

 

 

 

Figure 3.1. Proposed on-line, ethyl-TPP derivatization setup. The most
effective design probably will perform solid phase, precolumn derivatization
prior to LC separation and analysis by FAB-MS.

1 l l
detector before they pass into the mass spectrometer. A mass spectrometer

adapted for sample analysis by LC-MS (a JEOL HX-llO modified with a
frit-FAB interface) will determine the mass of the derivatized peptides
when operated in the FAB ionization mode. Product ions resulting from
CAD-MS/MS will aid in determining the amino acid sequence of the
derivatized peptides.

A polymer-bound reagent containing the ethyl-TPP group will be the
active component of the reactor causing peptide derivatization.
Functionalized polymers have found widespread use in organic synthesis
and related fields [20]. They are employed as reagent carriers, ion
exchange resins, and matrices for immobilized enzymes and cells [21].
Styrene resins containing more than fifty percent vinyl benzene are
employed for HPLC applications because they can be safely subjected to
pressures above 1000 psi [22]. Polymer resins containing immobilized
reagent are beneficial for use in derivatization procedures because they can
be recycled repeatedly and conveniently, can be modified to contain a large
excess of reagent to drive reactions to completion, are compatible with a
wide range of solvents, and do not leach. Krull and coworkers have
reported the development of several reactors containing polymer-bound
magenta for on-line derivatizations; these immobilized reagents were stable
during more than fifty injection periods before regeneration was necessary
[17-19, 23].

D. Chemical methodology

The overall proposed procedure for preparing the immobilized
reagent for on-column derivatization of peptides is represented in Scheme
3.1 . This procedure utilizes a pre-made polymer support covalently linked
to o-nitrophenol. The o-nitrophenol group will be reacted with the “Kunz”

112

.uaowaou .uasM. on... 853
consensus». on a? 93:8 25. 633% ca... mfinflwzuov accuse." 333
330350 on... 4:3 338 333m .95 Safioqcnmoufid-e.a§umx on... Show
3 “oaongefidé panacea fit» Beach a accuse.“ gun—am. dofleuflwzuov
name—.33 .393 v38 .Sm c8058 .8533 33825 A.» sacs—om

@ AHHHV E Fm
“Sig + ©I+mlumolamoloI_o_l:¢_aln_.lolvom

_ __
Noz Q 0 mm 0

m.2...
a mlefizlwlolvom

___
no

A:
©Iflvalnmolamo I o IoLo fluu CC

+ mlamoiumoIOIolo

RM a ©I+ __
oawuonmmom Q o No

Q 6 a

©l+mlumolamoIOIol6 + no

:
e .

Q

113
reagent [2-(triphenylphosphonio)ethylchloroformate chloride] in a

nucleophilic acyl substitution reaction that eliminates HCl and immobilizes
the “Kunz” reagent. The immobilized o-nitrophenolate of the “Kunz”
reagent and its polymeric support then will be packed into a column. When
the polymeric reagent comes in contact with peptides dissolved in the
mobile phase containing peptides is pumped through the column, the N-
terminal amine group of peptides will act as a nucleophile in a nucleophilic
substitution reaction to produce peptide derivatized with the ethyl-TPP
group. The starting reagent, immobilized o-nitrophenol, will be the other
product of this reaction. The column reactor can be regenerated quickly by
simply passing a solution containing the “Kunz” reagent through the
reactor bed.
E. Separation of a peptide mixture by LC-MS
1. Experimean

The analysis of a mixture containing approximately 1 00 pmol of four
peptides (GLA, YGGFL, VGVAPG, and VQAADYING) was performed by
LC-MS with a JEOL HX-110 modified to house a frit-FAB interface. The
magnet was scanned from m/z 50 to m/z 1000 every 10 seconds. Two
buffers were used to introduce the sample into the mass spectrometer.
Bufi'er'A contained 98.4% H20, 1.5% glycerol, and 0.1% TFA; buffer B
contained 98.4% ACN, 1.5% glycerol, and 0.1% TFA. A gradient of 040% B
in 10 minutes, 40 to 100% B from 10 to 60 minutes, and 100% B from 60 to
70 minutes achieved chromatographic separation of the peptide mixture
with a fused-silica microbore column [0.32 x 300 mm, C13, 5 pm, capillary
outlet; LC Packings (U.S.A.), Inc.; San Francisco, CA]. The solvent flow
rate rate was 10 uL/min as delivered by a syringe pump (Brownlee Labs;
Santa Clara, CA). The mobile phase was monitored post-column with a

113
reagent [2-(tripheny1phosphonio)ethylchloroformate chloride] in a

nucleophilic acyl substitution reaction that eliminates HCl and immobilizes
the “Kunz” reagent. The immobilized o-nitrophenolate of the “Kunz”
reagent and its polymeric support then will be packed into a column. When
the polymeric reagent comes in contact with peptides dissolved in the
mobile phase containing peptides is pumped through the column, the N-
terminal amine group of peptides will act as a nucleophile in a nucleophilic
substitution reaction to produce peptide derivatized with the ethyl-TPP
group. The starting reagent, immobilized o-nitrophenol, will be the other
product of this reaction. The column reactor can be regenerated quickly by
simply passing a solution containing the “Kunz” reagent through the
reactor bed.
E. Separation of a peptide mixture by LC-MS
1. Experimental

The analysis of a mixture containing approximately 100 pmol of four
peptides (GLA, YGGFL, VGVAPG, and VQAADYING) was performed by
LC-MS with a JEOL HX-110 modified to house a frit-FAB interface. The
magnet was scanned from m/z 50 to m/z 1000 every 10 seconds. Two
buffers were used to introduce the sample into the mass spectrometer.
Bufl'erA contained 98.4% H20, 1.5% glycerol, and 0.1% TFA; buffer B
contained 98.4% ACN, 1.5% glycerol, and 0.1% TFA. A gradient of 0-40% B
in 10 minutes, 40 to 100% B from 10 to 60 minutes, and 100% B from 60 to
70 minutes achieved chromatographic separation of the peptide mixture
with a fused-silica microbore column [0.32 x 300 mm, C13, 5 pm, capillary
outlet; LC Packings (U.S.A.), Inc.; San Francisco, CA]. The solvent flow
rate rate was 10 uL/min as delivered by a syringe pump (Brownlee Labs;

Santa Clara, CA). The mobile phase was monitored post-column with a

1 14
Kratos Spectroflow 783 UV-Vis detector (Kratos; Foster City, CA) before

entering the mass spectrometer.
2. Results and discussion

The UV-Vis trace obtained with injection of the four-component
peptide mixture is shown in Figure 3.2. In figure 3.3 a) is a total ion
current (TIC) plot obtained when the peptides entered the mass
spectrometer. Data acquisition was begun when the fastest-running
peptide was detected by the UV-Vis detector. Figure 3.3 b) contains the
reconstructed mass chromatograms for the individual protonated peptides.
The order of elution was VGVAPG (mlz 499), GLA (m/z 260), VQAADYING
(mlz 950), and finally YGGFL (mlz 556). The background-subtracted mass
spectra of the peptides are contained in Figure 3.4. The peptides all display
good signal-to-noise ratios, with protonated GLA represented by the base
peak in Figure 3.4 b). Although the UV-Vis trace shows the
chromatographic separation of the reaction mixture was not baseline
resolved, Figure 3.4 demonstrates that individual peptides can be detected
by the mass spectrometer. The performance of the JEOL HX-110 modified
with a fiit-FAB interface demonstrates that it is adaptable for analysis of
peptide mixtures by LC-MS.
F. Reaction in homogeneous solution

To investigate the feasibility of the proposed reaction sequence
(Scheme 3.1), the chemical reactions were performed in homogeneous
solution with non-mobilized reagents according to Scheme 3.2. This allows
the feasibility of the reaction to be investigated quickly and conveniently
without actually fabricating the column to be used for on-column

derivatization.

115

 

;- -
e
g. ..
A 4 L
g- _

 

 

 

 

 

 

 

 

g fiM 4' __
atsgéégazéigfiuén
Time(min)

Figure 3.2. UV-Vis trace of the four-component peptide mixture
containing 100 pmol of VGVAPG, GLA, YGGFL, and VQAADYING
following separation with a microbore column, prior to analysis by fiit-FAB-
MS.

116

RetentionTime(min)
-9' r-§.- .19. -.12. . -15.

(a)

 

N
..s
8
O
49
M

V
gnu-ugoogr—q m<-¢+p~raw wnwnpanaH ocwnn-‘o
, A
1
1
l
1
l»
l
I
I
if
I
1

on
O

a
?

 

55

0A--

 

 

- - ~‘ ---- - - m/z950

h 132
m/z 556

A x2.7
m/z499

JL 11.8

f - j 1 - ' - - . j“ - f - . - ‘ - - M260
100 200 300 400
Scannumber

 

 

 

 

Figure 3.3. Reconstructed total ion current (TIC) plot (a) obtained when
the peptide mixture entered the mass spectrometer. Scanning was
commenced when the first-running peptide was detected by the UV-Vis
detector positioned prior to the mass spectrometer. Figure 3.3 (b) contains
the reconstructed mass chromatograms of the individual components of the
peptide mixture.

117

 

(a) 100 a‘ Sea 162
. n .
1 V-G-V-A-P-G "12%” I
80‘ t

601 p

. 134 ’
40‘ 1 3 _

T

 

 

 

%"fl'flflflflgt—i o<~.”’~°w
ffi T' Y

 

 

2“ 256
1 al.

7
100 ll kg“ Bindiil 14H. 1 1 .
(b) : Scan 185 (£33,. :

G-L-A
80‘ .

 

40‘ p-

92

20" 171 .

Killlfljn if '
J L .444?9_. A r a - a? J l

Scans182-194 i
V-Q-A-A-D-Y-I-N-G .

wow-"nuanur-n o<-on~ow

 

 

..__9__-—"
‘— H
P

 

100

fir “‘
10?—
w

(c)

on
-?--

a:
-?--
r

2):

.5
O

20

1 7 A” :°-: «139%»-

80 “8...

' U 1 V V j v u '

 

 

Vflfllflflf’flh‘ O<"'f’D—‘¢Dw
H

F" N
8---?---.--
I;

 

 

(d)

Scan 231 .
Y-G-G-F-L . _

 

 

 

 

 

 

wen-nonuv—n saw-«p—ow

 

 

 

 

ml: 600 80° . ......

Figure 3.4. The background-subtracted mass spectra of individual
peptides (a-d) eluted during LC-MS. The scan rate was one scan every 10
seconds.

118

a Q
@013 . m—c—o—cn,—cn,—p+@

(1) m/z 369

a <3
@o-c-o-cnz—cnz—(Crg + HCl

(II) m/z472
fl 3‘
Ho(-—c—c|:—N)n—H 1
En;
o H 0 Q No2
II I II +
Ho(—c—<':—1~II)n—c-o—CHZ—CH2—P-© + @011
“n“ (III)

Scheme 3.2. Homogenous-phase chemistry simulating the on-line reaction
sequence. N onmobilized o-nitrophenol reacts with the ‘Kunz’ reagent (1) to
produce the “Kunz”-o-nitrophenolate (II). This product (II), in turn,
derivatizes the peptide to yield the “Kunz” modified peptide (III).

1 1 9
1. Formation of the “Kunz" reagent o-nitrophenolate

a) Experimental

A five-fold molar excess of o-nitrophenol (Aldrich Chem. Co.;
Milwaukee, WI) and 0.1 uL of 2,4,6-trimethylpyridine (TMP) to act as a
weak base were dissolved in 20 pL acetonitrile. One milligram of “Kunz”
reagent (Fluka Chemika-BioChemika; Buchs, Switzerland) was added to
the solution. The reaction mixture was vortexed for l 0 sec and 1 51L of the
solution was placed on the FAB probe tip together with a 1:1 (v/v)
glycerol/thioglycerol matrix. The mass spectrum was recorded after 30 sec
of FAB bombardment with the conversion dynode voltage at -3.8 kV.

A study of the kinetics for the reaction of the “Kunz” reagent with o-
nitrophenol and stability of the product was performed by vortexing the
reaction mixture for 5, 10, 20, 45, and 90 sec and measuring the ratio of
peak intensity of the product (m/z 472) to that of (2-hydroxyethyl)TPP (m/z
307), which is a side product and also a product of degradation of the o-
nitrophenolate of the “Kunz” reagent with H20. The mass spectra for this
investigation were acquired 30 sec after FAB was commenced. All trials
were performed in triplicate.

b) Results and discussion

Because the “Kunz” reagent and the reaction products contain
positively charged triphenylphosphonium groups, the reaction was
conveniently monitored by FAB-MS. The mass spectrum of the reaction
mixture resulting from vortexing a solution containing a five-fold molar
excess of o-nitrophenol and 1 milligram of “Kunz” reagent in the presence of
trimethylpyridine (TMP) for 10 see is shown in Figure 3.5 a). The base
peak (m/z 472) represents the desired product of this reaction (Compound
11, Scheme 3.2), which forms quickly and in high yield. The peaks at m/z

120

 

H
O
P

(a)

on
O

14%; . l

O)
O

A
‘31..

122

M
0

& L A a l A a a

92
1 08

wnwnuonaH o<moe~ow

 

 

100

1 3 262

21

200

m/z

289 307

379

 

75 397

 

 

351

 

 

 

300

400'

 

‘?

D
.-3. . .23
Qua—Q
o
:z:
N
o
in
o

gowu=¢p¢+5H 0<-¢+
N A
O O

 

'160'

289

262 275

305 351

 

397

_ l.; 1
I V V r I

400

 

aeJ

Figure 3.5. FAB-MS spectrum of reaction mixture resulting from mixing a
five-fold molar excess of o-nitrophenol and 0.5 uL 2,4,6-trimethylpyridine
with 1 mg of “Kunz” reagent for 10 sec. (b) Linked-scan spectrum at
constant NE of the o-nitrophenolate of the “Kunz” reagent.

12 1
289 and 307 represent vinyl-TPP and (2-hydroxyethyl)TPP side products,

respectively. If this reaction were performed on-line with immobilized o-
nitrophenol, the vinyl-TPP and (2-hydroxyethyl)TPP side products could be
washed away by the mobile phase. In this mass spectrum, no peaks
representing the “Kunz” reagent (mlz 369) are observed, indicating that the
reaction has gone to completion. The FAB-CAD-B/E spectrum of the o-
nitrophenolate of the “Kunz” reagent (Figure 3.5 b)) confirms the identity of
this product. The predominant vinyl-TPP fragment (mlz 289) is
characteristic of compounds containing the ethyl-TPP group. The other
peaks observable at m/z 183, 185, 201, 262, and 275 represent fragment
ions characteristically formed upon CAD of the ethyl-TPP group.

The results of the kinetics study (Figure 3.6) indicate optimal yield of
the “Kunz” reagent o-nitrophenolate results with 10 sec of reaction time.
For this study, the yield of the o-nitrophenolate of the “Kunz” reagent was
determined by calculating the ratio of peak intensity at m/z 472 (“Kunz"-o-
nitrophenolate) to the peak intensity at m/z 307 [( 2-hydroxyethyl)TPP].
The ratio of the “Kunz”-o-nitrophenolate to (2-hydroxyethyl)TPP increases
rapidly when the reaction time is lengthened from 5 sec to 10 sec.
However, after 10 sec of reaction time, the ratio of peak intensity at m/z 472
to the ratio of peak intensity at m/z 307 decreases. This is probably
because the o-nitrophenolate of the “Kunz” reagent undergoes B-elimination
in the presence of excess trimethylpyridine base because o-nitrophenol is a
good leaving group. The products of this reaction probably are 0-
nitrophenol, vinyl-TPP, and carbon dioxide. The vinyl-TPP probably reacts
with water absorbed from the air above the reaction solution and dissolved
in the reaction medium to form (2-hydroxyethyl)TPP. Because increasing
amounts of (2-hydroxyethyl)TPP, which desorbs at m/z 307, and decreasing

122

Intensity (mlz 307)

 

 

 

Time (sec)

Figure 3.6. Investigation to determine the kinetics of the reaction yielding
the o-nitrophenolate of the “Kunz” reagent. Reaction yield was defined as
the ratio of peak intensity at m/z 472 (II, Scheme 3.2) divided by peak
intensity at m/z 307 [(2-hydroxyethyl)TPP]. These results show the product
forms quickly, but is unstable in a solution with water-containing

atmosphere above it.

1 2 3
amounts of the o-nitrophenolate of the “Kunz” reagent result, this accounts

for the decreasing ratio of peak intensity at m/z 472 to peak intensity at m/z
307. Nevertheless, this side reaction is not expected to be a problem in a
heterogeneous system, because the reaction solution should not be exposed
to atmospheric water when o-nitrophenol is bonded to a polymer in the
solid phase reactor enclosed in a HPLC column. Dry solvents also will be
used. Further, if the solution is acidified following formation of the “Kunz”-
o-nitrophenolate, base-assisted B-elimination is less likely, causing the
desired “Kunz”-o-nitrophenolate to be stable and available for reaction with
peptides. The speed of this reaction is appropriate for generating and
regenerating “Kunz” reagent covalently bound to a polymer.
2. Peptide derivatization

a) Experimental

To produce derivatized peptide, 2 uL of a 937.5 nmol/uL solution of o-
nitrophenol in acetonitrile (a 10-fold molar excess) was added to 2 “L of a
93.75 nmol/uL solution of "Kunz" reagent dissolved in acetonitrile and 0.1
uL of TMP. This solution was vortexed for 10 sec. Then 2 ”L of 3.75
nmol/uL of AAA methyl ester was added to this solution, giving final
concentrations of approximately 1 .25 nmol/uL AAA methyl ester, and
approximately 31 .5 nmol/uL (a 25-fold molar excess) of the o-nitrophenolate
of the “Kunz” reagent.

The formation of ethyl-TPP-derivatized peptide was investigated
with three different reaction conditions: vortexing the reaction medium,
heating the reaction medium, and vortexing the reaction medium in the
presence of 0.5 uL of triethylamine (TEA). To investigate the progress of
the reaction at different time intervals, the solution was vortexed or heated
(37°C) the desired amount of time (0.5, 1.0, 2.0, 4.0, 8.0, 16.0, or 32.0 min)

124
and 1 31L of the resulting solution was placed on a direct inlet probe

containing ~1 uL of a 1 :1 (v/v) glycerol/thioglycerol matrix. Acquisition of
the mass spectrum was commenced 30 sec after FAB bombardment was
begun. All trials were performed in triplicate with a conversion dynode
voltage of -20 kV.

b) Results and discussion

To investigate the feasibility of the o-nitrophenolate of the “Kunz”
reagent to derivatize a peptide, the methyl ester of the tripeptide AAA was
chosen as a model peptide, because it contains a deactivated carboxyl
group. The mass spectra of 1.25 nmol of the AAA methyl ester reacted with
the “Kunz”-o-nitrophenolate under normal reaction conditions (25°C), with
heating (37°C), or in the presence of TEA for 32 min (Figures 3.7 a), b), and
c), respectively) demonstrate the FAB-MS response for the derivatized
peptide (m/z 534) is highest in the presence of TEA, which acts as a base
catalyst. Increased reaction yield is also observed when the reaction is
performed at 37°C rather than 25°C. The expected product of the peptide
modified with the PEOC group should be observed at m/z 578, but little of
this product appears. Instead, the peak at m/z 534 represents the
decarboxylated product of the peptide methyl ester tagged with the "Kunz"
reagent (HI, Scheme 3.2). The decarboxylation process for PEOC [2-
(triphenylphosphonio)ethoxycarbonyl] derivatized peptides was proposed by
Kunz [24]. In weakly alkaline conditions (Scheme 3.3), B-elimination
causes formation of vinyl-TPP and a carbamate anion, which is protonated,
forming an amino acid carbamic acid. The carbamic acid decarboxylates
with a hydrogen shift, and the resulting nucleophilic amine group of the
amino acid then can attacks vinyl-TPP to form peptide derivatized at the N -
terminus with ethyl-TPP. The decarboxylated product has a simpler

125

 

H
O
O

A A A I

307 397

(a)

on
.11?

379

.$

93 185

a
.CB- .

+TPP-Ethyl-AAAOMe

to
$3..

(Mt-H)"

--JLL .1; unmih LLJAAL_11 L
93

185 307' 397

 

L A

 

 

 

 

n_

(b)

m
C

A l A A A j

262

 

a
o A A

re
°...

 

(M+_H)+
+TPP-Ethyl-AAAOMe

 

gau-uuon-gv-c °<""""”"'°W

 

 

 

 

 

 

L:

. link L
93 185 261 307 397

 

(c) +TPP-Ethyl-AAAOMe

on
O
A l #A A l A A A

A A l A A

a
.‘P.

(M-rH)+

 

A l A A

 

 

‘40“‘050950-4 (Dd-“rO-p—‘(bw

 

 

 

 

 

 

 

% 100 200 300 400 soo 600

m/z
Figure 3.7. Reaction of 1 .25 nmol/uL AAA methyl ester with a theoretical
25-fold molar exess of the “Kunz’-o-nitrophenolate. The following spectra
represent 1 uL of the reaction medium containing 1 .25 nmol of AAA methyl
ester after 32 minutes of reaction (a) at room temperature, (1)) at 37°C, and
(c) in the presence of TEA.

126

H O a ‘ ' o
I II | u
©‘P+- (3112- CH;- O-fi -N-peptido-C-OH %’ ©1’-ca,- ca,-N-pepude-c—on

O\I:*R ©/

H o
l u .. l
(fig-peptide-c—oa —> HaN-MfldO-C-OH + co,
o=c\ Kit
*8;

Scheme 3.3. Suggested reaction pathway for decarboxylation of the
peptide derivatized with the “Kunz” reagent to form an ethyl-TPP
derivatized product.

1 27
structure than the original product and is identical to peptides modified

with the ethyl-TPP group by vinyl-TPP. The fragmentation pattern of the
decarboxylated product of the AAA methyl ester produced by this reaction
sequence (Scheme 3.2) is identical to the vinyl-TPP derivatized AAA methyl
ester as the FAB-CAD-B/E spectra demonstrate (Figure 3.8 a) and b)). The
yields of derivatization with the o-nitmphenolate of the “Kunz” reagent in
the presence of TEA for 32 min and derivatization with vinyl-TPP in a 1 :1
pyridine/acetonitrile solvent mixture for ~8 hrs at 37°C are quite similar as
Figures 3.9 a) and b) demonstrate.

The reaction kinetics of the simulated on-line derivatization reaction
under different conditions are shown in Figure 3.10. The extent of the
reaction was determined by measuring the ratio of derivatized,
decarboxylated AAA methyl ester peak intensity (mlz 534) to underivatized
AAA methyl ester (mlz 246 represents the protonated molecule of AAA
methyl ester). As expected, faster reaction kinetics were observed at higher
temperatures (37°C vs. 25°C) and in the presence of TEA as a base catalyst.
This study indicates the optimal derivatization yield would be produced
with the use of a base catalyst (such as TEA) and heating of the reaction
mixture. The practicality of this reaction sequence for higher molecular
weight peptides was investigated by performing the reaction with the
hexapeptide VGVAPG to simulate on-line, ethyl-TPP derivatization.
Although this peptide contains an unprotected carboxyl group at the C-
terminus, the reaction proceeded in high yield as Figure 3.11 a)
demonstrates. The FAB-CAD-B/E spectrum of the “Kunz” derivatized,
decarboxylated product of VGVAPG (Figure 3.11 b)) is similar to that of the
peptide derivatized at the N-terminus with the ethyl-TPP group by vinyl-
TPP, indicating the products are identical (compare Figure 2.21).

128

 

 

 

 

 

 

 

 

 

 

 

 

 

 

20-
R « . (M)+
2!?
f (a)
a 1
t15‘ '
1 . 262
V
e
O
110 Q
:1 -©-P+ CH, CH, A A A (ll—OCH, 289
. : a
n .
8 5-
I .
t al
Y 1E5 l ”I (M-0020H3)+
100 200 m/z 300 400 500
1" : (b) 277
a 1
t15-
i . 262
V 1
e 1
110
n . 289
t .
O .
n t
9 5-
1 1 .,
y . 1E5 l 1' (M-COzCI-Isr'
'160‘ ' ' '260' " 360' ' ' '460' ' ' 560
m/z

Figure 3.8. FAB-CAD-B/E spectrum (a) of 1 nmol of AAA methyl ester
derivatized with vinyl-TPP and (b) FAB-CAD-B/E spectrum of 1 nmol of
AAA methyl ester derivatized with the “Kunz”-o-nitrophenolate.s

129

 

93

...:
O
O

(a)

.‘é. . 3. - 8..

wewusoeuH o<wnm~ow
N
O

 

 

Co-.

100

185

 

 

200

262

can»

 

289

 

300

349

 

494

 

+TPP-Ethyl-AAAOMe
534

 

 

400 500

 

100 93

(b)

on
O

b O)
?...‘?..r...
q

0‘

%Gwaaonat-c o<fl°fim~ow
to
o

 

 

 

 

100

Figure 3.9. Spectra of the reaction mixtures theoretically containing 1
nmol of AAA methyl ester after derivatization: (a) ~8 hrs in a
pyridine/acetonitrile (1 :1 v/v) solvent mixture at 37°C and a 20-fold molar
excess of vinyl-TPP, (b) 32 min in acetonitrile/TEA with a approximate 25-

185

 

 

262

tht

 

s67*’**
2s9

 

 

fold molar excess of “Kunz”-o-nitrophenolate.

39f

 

*TPP-Ethyl-AAAOMe

 

400 500

 

 

130

 

 

4'?
{a .
a 3,;
E - ' TEAaddcd
I . ° 37degreesC
- a? 2‘. ' 25degreesC
Cl
3 d
.5 1';
I /
O I I 1 I f I
0 10 2) 3)
Time (min)

Figure 3.10. Investigation of the speed of peptide derivatization with the
‘Kunz’o-nitrophenolate. The yield of reaction was determined by the ratio
of decarboxylated, ethyl-TPP, derivatized AAA methyl ester (mlz 534) to
underivatized and protonated AAA methyl ester (mlz 246) peak intensity.
This reaction was performed at room temperature, 37°C, and in the
presence of TEA, as the legend indicates. For these studies , 1 .25 nmol of
AAA methyl ester was reacted with a 25-fold excess of the “Kunz”-o-
nitrophenolate.

131

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

$00, 93 183“ 262 3
f l
:1 so- (a)
i * +TPP-Ethy1-VGVAPG
v . 362
e 60-
I .
n .
t 40.
° l
3 : (M+H)+
i .
t 2°.
y . .
ll -" ‘ JJAA'LLLIIl thAL LrAALfiA 1“ 1'
(1’ 100 200 300 400 500 600 700 800 900
rn/z

R : Ila mm
{3 I (b) d»; at:

‘ c...
a t
t I b..: 5
l 101 .
v j (C? a,,. .
e .

‘ + b _.
I 1 Q'P (crayfiv G v A p a on
n .
. i b)
e 54
n
9 l “ do
1 .
t 3 N .v
y l .

1 ‘11 a. b: b: °s 1" °4 be

tho 400 500 600 700

 

Figure 3.11 (a) FAB-MS of reaction mixture containing 1.25 nmol of
VGVAPG after reaction 32 min with a 25-fold molar excess of ‘Kunz'-o-
nitrophenolate with 0.5 1.1L of TEA present as a base catalyst.
Decarboxylated, ethyl-TPP, derivatized product appears at m/z 871 , while
protonated, underivatized peptide appears at m/z 499. (b) FAB-CAD-B/E
spectrum of 1 .25 nmol of decarboxylated, derivatized VGVAPG.

1 3 2
G. Conclusions

Efficient production of ethyl-TPP derivatized peptides demonstrates
the feasibility of using immobilized “Kunz” reagent as part of an on-line
reaction scheme. Although the reaction conditions may be modified, this
reaction sequence (Scheme 3.1) could be utilized for the on-line, ethyl-TPP
derivatization of peptide mixtures in combination with analysis by HPLC-
FAB-CAD-MS/MS. The analysis of a peptide mixture by the JEOL HX-110
in the LC-MS mode demonstrates that this instrument could function as a
mass detector in an apparatus designed to derivatize, separate, and mass
analyze peptide mixtures.

H. Future work

The next logical step in development of the on-line reactor is to
prepare the polymer-supported reagent through heterogeneous phase
chemistry. O-nitrophenol covalently linked to a polymer will be reacted
with the “Kunz” reagent to produce immobilized reagent (11, Scheme 3.1).
This reaction should be performed with stirring to achieve an even
distribution of the reagent on the polymer. Stirring should be performed
only a couple of hours to prevent damage to the polymeric material. Once
produced, the amount of “Kunz” reagent loaded onto the polymer support
should be characterized by displacing the “Kunz” reagent and quantitating
the amount displaced or by performing elemental analysis.

When the polymer-bound reagent is produced, it will be placed in a
plugged pasteur pipet. A solution containing generous amounts of peptides
then can be placed in the pipet. After a selected period of reaction time, the
solution will be drawn out of the pipet and if peptides derivatized with the
ethyl-TPP group are detected in the solution drawn off the heterogeneous
reaction mixture, smaller quantities of peptides will be derivatized by this

133

procedure to determine the efficiency of the reaction. The heterogeneous

chemistry can be repeated using variations in reaction temperature, solvent

systems, pH, etc., to produce optimal reaction emciency with trace levels of

peptides. Once the reaction system is optimized, the polymer support

attached to immobilized “Kunz” reagent will be packed in a column to

achieve ethyl-TPP derivatization on-line. The ultimate triumph of any

biopolymer modification procedure would be realized if it were part of a

fast, automated process that would separate and sequence a mixture of

peptides at low detection limits.

1. References

1.

9‘99”.”

10.

Dorsey, J .G.; Foley, J .P.; Cooper, W.T.; Barford, RA., Barth, H.G.,
Anal Chem., 1992, 64, 353R-389R.

Luque de Castro, M.D., Trends Anal. Chem., 1992, 11(4), 149-155.
Gao, C-X.; Krull, 1.8., J. Chromatrogr., 1990, 515, 337-356.

Bourque, A.J.; Krull, 1.8., J. Chromatogr., 1991, 537(1-2), 123-52.
Gao, C.X.; Schmalzing, D.; Krull, I.S., Biomed. Chromatogr., 1991, 5(1),
23-31.

Jedrzejcsak, K; Gaind, V.S., Analyst (London), 1990, 115 (10), 1359-
1362.

Yasaka, Y.; Tanaka, M.; Matsumoto, T.; Funazo, K.; Shono, T., Anal.
Sci., 1989, 5(5), 611-612.

Tanaka, M.; Yasaka, Y.; Funazo, K.; Shono, T., Fresenius’ Z. Anal.
Chem., 1989, 335(3), 311-312.

Porter, D.H.; Swaisgood, H.E.; Catignani, G.L., J. Agric. Food Chem.,
1984, 32(2), 334-339.

Pilosof, D.; Kim, H.-Y.; Vestal, M.L.; Dyckes, D.F., Biomed. Mass
Spectrom., 1984, 11(8), 403-407.

11.

12.

13.

14.

15.

16.

17.
18.

19.
20.
21 .
22.
23.

24.

134
Pilosof, D.; Kim, H.-Y.; Dyckes, D.F.; Vestal, M.L., Anal. Chem., 1984,

56, 1236-1240.

Stachowiak, K.; Wilder, C.; Vestal, M.L.; Dyckes, D.F., J. Am. Chem.
Soc., 1988, 110, 1758-1765.

Kim, H.Y.; Pilosof, D.; Dyckes, D.F.; Vestal, M.L., J. Am. Chem. Soc.,
1984, 106, 7304-7309.

Voyksner, B.D.; Chen, D.C.; Swaisgood, H.E., Anal. Biochem., 1990,
188, 72-81.

Henzel, W.J.; Bourell, J .H.; Stults, J .T., Anal. Biochem., 1990, 187,
228-233.

Kassel, D.B.; Williams, KP.; Musselmann, B.D.; Smith, J .A., Anal.
Chem., 1991, 63(18). 1978-1983.

Gao, C.-X.; Chou, T.-Y.; Krull, I.S., Anal. Chem., 1989, 61, 1538-1548.
Chou, T.-Y.; Gao, C.-X.; Grinberg, N.; Krull, I.S., Anal. Chem., 1989,
61 , 1548-1558.

Gao, C.-X.; Krull, I.S.; Trainer, Tom, J. Chrom. Sci., 1990, 28, 102-108.
Akelah, A.; Sherrington, D.C., Chem. Rev., 1981, 81, 337-587.

Klein, J .; Buchholz, K., Methods Enzymol., 1987, 135, 3-30.

Moore, J .C.; J. Polym. Sci., Part A, 1964, 2, 835.

Gao, C.-X.; Krull, I.S.; Trainor, T.M., J. Chromatogr., 1989, 463, 192-
200.

Kunz, H., Chem. Ber., 1976, 109, 2670-2683.

CHAPTER IV
FUTURE DERIVATIZATION OF A GLYCOPEPTIDE

A. Introduction

Because FAB-MS has increased the capability of investigators to
address new problems, studies on new classes of biopolymers are now
possible. With the i) introduction of FAB ionization, ii) the discovery of new
methods to increase the internal energy of species with high molecular
weight, iii) the extension of the mass range for existing ion optics of mass
spectrometers, and iv) the development of methods to increase sensitivity
while maintaining high resolution, high-mass, highly polar compounds can
be analyzed by MS. The introduction of high-performance tandem mass
spectrometry (MS/MS) with CAD has made it possible to obtain structural
information and determine the nature of the posttranslational modif-
ications of peptides. With CAD-MS/MS, the size and sequence of
carbohydrate structures attached to peptides, the site of attachment, and
more rarely, the peptide sequence itself can be determined. Sample
derivatization also has enhanced the analysis of glycopeptides by FAB-MS.
The previous chapters have demonstrated how derivatization of peptides
with reagents containing the TPP group decreased the detection limits of
the FAB-MS technique and also can increase the amount of information
available with analysis by CAD-MS/MS. The utility of attaching the TPP
moiety to the desired terminus of a peptide to direct fragmentation has
been demonstrated. A unique application of the ethyl-TPP derivatization
method would be to derivatize glycoproteins to enhance their ionization

135

136
eficiency and to cause preferential fragmentation of the peptide relative to

the carbohydrate group.

Glycosylation is one of the most common and firnctionally important
posttranslational modifications of proteins. A number of difi'erent cellular
processes, including adhesion, metastasis, trafficking, and immune
recognition, have been attributed to the presence of oligosaccharides on
proteins. The glycosylation of some proteins is necessary for their secretion
from the cell [1]. Glycosylation of a protein also may determine if
antibodies are produced for a particular glycoprotein [2]. Because of the
implications of glycosylation, it is advantageous to determine the nature
and site of glycosylation in proteins.

The types of glycosylation are defined by linkage of the
oligosaccharide to a protein. Although the site of glycosylation for the N -
linked sugars is the amide group of the asparagine residue found in the
Asn-Xxx-Ser/Thr sequon, no consensus sequence for O-glycosylation has
been found. O-linked units are attached to the hydroxyl groups of serine or
threonine residues. The site of glycosylation for O-linked carbohydrates
has been vaguely assigned to serine and threonine-rich B-turn regions of
proteins. Structural analysis of O-glycosylated proteins is also complicamd
by the lack of an enzyme to specifically cleave the carbohydrate group fi-om
the protein. Consequently, B-elimination with alkali conditions has been
the primary method used to liberate O-linked oligosaccharides. Reducing
agents such as NaBH4 have been included in the reaction mixture during
the elimination process, but under these conditions the peptide suffers
degradation and information on the carbohydrate attachment site is lost.
There may be much heterogeneity in carbohydrate composition at a single

137
attachment site. In some cases, heterogeneity may be extreme, causing the

identification the glycosyl units to be a challenging task.

Complete characterization of glycoproteins should include
elucidation of the oligosaccharide structure, primary peptide sequence,
assignment of the glycosylation sites, and positioning of the
oligosaccharides in the protein structure. These types of studies are
especially important to investigate the many glycoprotein recombinant
products produced by the biopharmaceutical industry. Many of the
manufactured glycopeptides have a significantly different carbohydrate
composition relative to the native form of the glycoprotein; consequently,
they may have other functional characteristics than the native
glycopeptide, because the extent and type of glycosylation play a central
role in glycoprotein activity. Analysis of glycoprotein products to confirm
their identity and purity by FAB-MS is practical because only small
amounts of sample are requimd.

B. Procedures for glycopeptide isolation

Isolating a glycopeptide for chemical derivatization and analysis by
FAB-MS can be a formidable task. The most common procedure utilized to
isolate a glycopeptide is to digest a glycoprotein with a specific enzyme
(trypsin cleaves at lysine and arginine residues, Staphylococcus aureus V8
cleaves at aspartic and glutamic acid residues). However, a work-up
procedure involving reduction and alkylation of disulfide bonds often is
required prior to digestion of the glycoprotein. In many instances, digestion
is accomplished in a cocktail of proteases for an extended period of time.
Following digestion, the mixture containing peptides and glycopeptides can
be purified by reverse phase chromatography. Fractions containing

138
glycopeptides and their accompanying oligosaccharide groups can be

recognized by colorimetric or fluorescent methods.

Fractions containing glycopeptides can also be identified by treating
a small portion of the enzymatic digest with PNGase F (N-glycosidase F).
This enzyme cleaves all N -linked carbohydrates from the peptide to which
they are covalently bound. Separation of the digest treated with PNGase F
by reverse phase chromatography and comparison of this chromatogram to
the chromatogram of the peptide mixture not digested with PNGase F
should indicate which peaks represent glycopeptides. The absence of peaks
in the chromatogram of sample treated with PNGase F, when contrasted
with the chromatogram not digested with PNGase F, would indicate that
the peaks that are absent represent glycopeptides. The appearance of
signals in the chromatogram of the sample treated with PNGase F, when
collated with the chromatogram of the undigested sample, likely represent
peptides with the carbohydrate group cleaved from them. Analysis of the
glycopeptide mixture by FAB-MS, followed by digestion with PNGase F and
subsequent analysis by FAB-MS also can indicate the presence of
glycoproteins, in a technique called carbohydrate mapping [3]. In this
procedure, the materialization of signals following treatment with PNGase
F indicates the presence of peptides that formerly were glycosylated.
Conversely, the disappearance of signals from the high-mass region of the
FAB-MS spectrum not treated with PNGase F (probably with low peak
intensity if a high percentage of the peptide is glycosylated), indicates these
signals represent glycopeptides.

Bovine fetuin is the glchprotein most commonly used as a source of
model glycopeptides. The complete amino acid sequence of this protein is
known and the sites of glycosylation have been identified [4]. The

1 39
oligosaccharides isolated from bovine fetuin have been characterized by

conventional methods [5,6], as well as by FAB-MS [7]. If a single
glycopeptide is desired for analysis, a technique designed to isolate
nanomole amounts of a glycopeptide from ribonuclease B is described by
Carr [8].
C. Glycopeptide analysis

Alternative and complementary methods of glycopeptide analysis
employ automated techniques. Microsequencing with the Edman technique
elucidates the amino acid sequence of a glycopeptide from the N-terminus,
except at the glycosylated residue, which causes a blank cycle, or more
rarely, an incorrect amino acid assignment. Carbohydrate analysis can be
performed by high-pH ion exchange chromatography with a pulsed
amperometric detector. The presence of galactose, fucose, or sialic acid
residues indicates that the carbohydrate portion of the glycopeptide has a
complex or hybrid structure. The classical methods utilized for
carbohydrate analysis include stripping the O-linked glycosyl units from
the protein with B-elimination in the presence of a base. The liberated
oligosaccharides are separated by repeated fractionation with gel filtration
and/or ion exchange chromatography. Structural analysis of the
oligosaccharides can also be performed by GC-MS following chemical
derivatization, possibly in combination with sequential glycosidase
digestion. However, large sample quantities are required for GC-MS
techniques and information about the site of glycosylation on the peptide
chain is lost with B-elimination.

When information about the peptide sequence and carbohydrate
content have been obtained, a combination of endo- and exoglycosidase

digestions, chemical derivatizations, and FAB-MS can be used to determine

140
oligosaccharide structures contained in glycopeptides. Interpretable

fragmentation spectra of the intact glycopeptide can be obtained without
further chemical modifications. FAB-CAD-MS/MS provides sequence
information to complement the molecular weight information obtained by
FAB-MS, through the fragmentation of glycosidic bonds and sugar rings of
oligosaccharides [9]. However, a large amount of fragmentation is observed
in the analysis of carbohydrates by FAB-MS in the negative ion mode,
which also may utilized for structural elucidation.
D. The analysis of glycoproteins by FAB-MS

Investigation of glycoproteins by MS is a complex problem [10],
although FAB-MS has greatly strengthened the strategies used to elucidate
the structure of glycoproteins. Commonly, only small amounts of
glycoproteins are available (microgram amounts or smaller). Even these
quantities of sample cause difiiculties in analysis by FAB-MS, because
carbohydrates have poor surface activity in the polar FAB matrices
employed. The large size of glycoproteins and the high number of covalent
modifications (glycoproteins usually contain many disulfide bonds in
addition to glycosylation) hinder the methods normally used to break
proteins into smaller units that are easier to manipulate. Chemical
cleavage and enzymatic digestion cleave glycoproteins into lower-mass
glycopeptides (MW=1000-5000; 30-hexose units) that fall into the mass
range of many mass spectrometers equipped with FAB. However, a
complex mixture of peptides, glycopeptides, and oligosaccharides normally
results from enzymatic digestion, especially if nonspecific enzymes such as
pronase are employed to produce truncated glycopeptides. The enzymatic
digestion of glycoproteins also is complicated because the carbohydrate
group may shield the peptide chain from proteolysis. Although FAB

141
ionization may successfully ionize glycoproteins, subsequent CAD-MS/MS

often does not prove productive because fi-agmentation of the carbohydrate
group often dominates the mass spectrum, obscuring peaks produced by
fi'agmentation of the peptide moiety.

In some cases, intact glycopeptides can be analyzed directly by FAB-
MS without extensive chemical derivatization. Several examples of the
analysis of intact glycopeptides by FAB-MS have been reported, confirming
their amino acid sequence and sugar composition [11]. The CAD-MS/MS
spectrum of a glycopeptide obtained from the S. aureus V8 digestion of
thDGF-B provided the complete amino acid sequence and carbohydrate
composition. Because an arginine residue is located near the N-terminus,
the fragmentation of this molecule with CAD-MS/MS was dominawd by N -
terminal fragmentation. The fragment ions a13-a15 were observed with a
162-u mass shift, indicating the presence of a single hexose unit on a
threonine residue. Side-chain fragmentation and the loss of a mannose
residue from the protonated molecule probably produced this series of
sequence ions. The CAD-MS/MS spectrum of a glycopeptide isolated from
bovine fetuin, on the other hand, displayed extensive fragmentation of the
carbohydrate side chains rather than the peptide backbone. The FAB-
CAD-MS/MS spectrum indicated that two identical hexosyl-N-acetyl-
hexosamine glycans are attached to the peptide backbone of the
glycopeptide. Because of the extensive fragmentation of the carbohydrate
groups, the peptide sequence for this molecule could only be determined by
selecting the ion of the deglycosylated peptide for analysis by CAD-MS/MS
[l 2].

Some investigators contend that FAB-CAD-MS/MS should not be

used solely to solve glycoprotein structure because critical structural

142
information may be missed [13]. Further, FAB-MS is not recommended for

quantitating the amount of carbohydrate present in a sample. The most
reliable approaches to solve glycopeptide structure combine several
analytical methods whenever possible. If FAB-MS is performed in
combination with analysis by N MR, oligosaccharide isolation and
purification procedures should provide the large sample amounts (micro to
millimolar) required for analysis by NMR. However, often glycoproteins
are only available in the low nanomole and picomole amounts of sample. In
these instances, mass spectrometry should be used in combination with
chemical and biochemical approaches, capitalizing on the unique strengths
of the individual techniques.
E. Derivatization of carbohydrates and analysis by FAB-MS

Many groups have utilized chemical derivatization in combination
with FAB-MS to decrease the limits of detection and increase the amount of
structural information available in carbohydrate analysis. In fact, some
glycoprotein problems are solved most easily by incorporating chemical
derivatization prior to FAB-MS experiments. Dell et al. [14] have shown
that peracetyl and permethyl derivatives of oligosaccharides are useful for
FAB-MS, and the sensitivity of analyzing. the derivatized species is
increased by at least two orders of magnitude over that for underivatized
carbohydrates. Typically, 0.5-5.0 ug of oligosaccharide are used for sample
analysis. Peracetyl and permethyl derivatized carbohydrates ionize to give
abundant molecular ions plus reproducible and predictable fragment ions
that allow sequencing and determination of branching patterns. Tsai et al.
[9] have acetylated high-mass oligosaccharides obtained from yeast
glycoproteins by endoglycosaminidase H (Endo H) digestion. Treatment of

a few micrograms of isolated saccharide in a trifluoroacetic

143
anhydride/glacial acetic acid solution two or three hours yields the

acetylated product. Permethylation and acetylation aids the assignment of
linkage positions because of an easily interpretable fragmentation pattern.
Acetolysis with acetic anhydride/glacial acetic acid/I12804, followed by
FAB-MS of the acetylated fragments recovered by solvent extraction,
simplifies the structural elucidation of the carbohydrates because the
molecules fragment at all the N-acetylhexosamine (HexNAc) residues,
yielding patterns of signals that reveal the types of glycans in the
oligosaccharide sequence. A trimethyl p-aminophenyl ammonium chloride
(TMAPA) derivative with a “preformed,” fixed charge was used to modify
high molecular weight carbohydrates to increase their molecular ion
abundance and to cause specific cleavage of the larger molecules after
subsequent methylation and acetylation.

The structures of six major acetylcholine receptor oligosaccharides
were determined by positive and negative ion mass spectrometry after
derivatization with the p-aminobenzoic acid ethyl ester (ABEE) reagent.
Oligosaccharides released from glycoproteins contain a free reducing end
that may be coupled to a lipid-like chromophore by reductive amination.
The ABEE derivative increases the ionization efficiency of released
oligosaccharides by increasing their hydrophobicity and sputtering
efficiency. This derivative is of low molecular weight and produces a single
product in contrast to permethylation or peracetylation [1 5]. In the positive
mode, these derivatized molecules show mainly protonated molecules with
little useful fragmentation. However, in the negative ion mode, the
deprotonated species yields a relatively abundant set of ions used to
establish the sequence and branching of oligosaccharides [1 6]. Product ions
reflecting the configuration, linkage position, acetyl position, sequence, and

144
branching structure of oligosaccharides were produced from

oligosaccharides derivatized with p-aminobenzoic acid [1 2]. Studies of a N -
linked core oligosaccharide utilizing ABEE derivatization of an
oligosaccharide from the Saccharomyces cerevisiae yeast mutant corrected
the proposed structure for this oligosaccharide [1 7].

A review describes the investigation of O-linked and N-linked
glycans by FAB-MS following B-elimination with NaBH4 and subsequent
methylation, exoglycosidase digestion, and Cr03 oxidation [18]. N,N-(2,4-
dinitrophenyl)octylamine derivatives also have been developed for the FAB-
MS analysis of oligosaccharides in the negative ion mode. The derivatized
oligosaccharides have an increawd ionization and fragmentation eficiency
in the negative ion mode [1 9].

F. Advantages of ethyl-TPP derivatization of a glycopeptide

Derivatization with the ethyl-TPP group should increase the
ionization efiiciency of glycopeptides as it has increased the ionization
efficiency of other posttranslationally modified peptides. Derivatization of
the N—terminus of the glycopeptide with (2-bromoethyl)TPP or vinyl-TPP is
more feasible than derivatization of the carboxy terminus with (2-
aminoethyl)TPP because the latter reagent could react with the
carbohydrate portion of a glycopeptide, which would be an undesirable side
reaction. The object in derivatizion would be to cause preferential
fragmentation of the peptide portion of the glycopeptide so the amino acid
sequence of the peptide portion of the glycopeptide can be elucidated more
readily. Presumably, attachment of a fixed charge to the peptide portion of
the molecule would increase the amount of fragmentation observed in the
peptide backbone relative to the amount of fragmentation observed in
carbohydrate portion of the glycopeptide.

145

G. References

1.
2.

10.

11.

12.

Dube, J .; Fisher, J .W.; Powell, J.S., J. Biol. Chem., 1988, 263, 17516.
Gribbon, J .G.; Devereux, S.; Thomas, N.S.; Keim, M.; Jones, H.M.;
Goldstone, A.H.; Lunch, D.C., Lancet, 1990, 334, 434.

Carr, S.A.; Hemling, M.E.; Folena-Wasserman, G.; Sweet, B.W.;
Anumula, K.; Barr, B.; Huddleston, M.J.; Taylor, R, J. Biol. Chem.,
1989, 264(35), 21286-21295.

Dziegielewska, KM.; Brown, W.M.; Casey, S.-J.; Christie, D.L.;
Foreman, R.C.; Hill, R.M.; Saunders, N.R.; J. Biol. Chem., 1990,
265(8), 4354-4357.

Green, E.D.; Adelt, G.; Baenziger, J .U.; Wilson, 8.; VanHalbeek, H., J.
Biol. Chem., 1988, 263, 18253.

Nilsson, B.; Norden, N .E.; Svensson, S., J. Biol. Chem., 1979, 254,
4545.

Angel, A.S.; Nilsson, B., Biomed. Environ. Mass Spectrom., 1990, 19,
721 -730.

Carr, S.A.; Roberts, G.D., Anal. Biochem., 1986, 157, 396-406.

Tsai, P.-K.; Dell, A.; Ballou, C.E., Proc. Natl. Acad. Sci. USA, 1986,
82, 4119-4123.

Poulter, L.; Earnest, J .P.; Stroud, R.M.; Burlingame, A.L., Biomed.
Environ. Mass Spectrom., 1988, 16, 25-30.

Medzihradszky, KF.; Gillece-Castro, B.L.; Settineri, C.A.; Townsend,
R.R.; Masiarz, F.R.; Burlingame, A.L., Biomed. Environ. Mass
Spectrom., 1990, 19, 777-781.

Gillece-Castro, B.L.; Burlingame, A.L., Biological Mass Spectrometry,
A.L. Burlingame, J .A. McClosky, Eds., Elsevier Sceince Publishers
B.V., Amsterdam, 1991, 411-436.

13.

14.

15.

16.

17.

18.
19.

146
Hefta, S.A.; Shively, J .E., Biological Mass Spectrometry, A.L.

Burlingame, J .A. McCloskey, Eds., Elsevier Science Publishers B.V.,
Amsterdam, 1991, 259-283.

Dell, A.; Carman, N.H.; Tiller, P.R.; Thomas-Oates, J .E., Biomed.
Environ. Mass Spectrom., 1988, 16, 19-24.

Poulter, L.; Earnst, J .P.; Stroud, R.M.; Burlingame, A.L., Proc. Natl.
Acad. Sci. USA, 1989, 86, 6645-6649.

Carr, S.A.; Reinhold, V.N.; Green, B.N.; Hass, J .R., Biomed. Mass
Spectrom., 1985, 12, 288.

Hernandez, L.M.; Ballou, L.; Alvarado, E.; Gillece-Castro, B.L.;
Burlingame, A.L.; Ballou, C.E., J. Biol. Chem. , 1989, 264(20), 11849-
11856.

Egge, H.; Peter-Katalinic, J ., Mass Spectrom. Rev., 1987 , 6, 331-393.
Zhang, Y.; Cedergren, R.A.; Nieuwenhuis, T.J.; Hollingsworth, R.I.
(submitted for publication).

 

 

,I I III ‘14 T. iual ‘N

 

lll‘l‘l‘ll‘l‘lllll‘l‘