.1!

$1.. .

II :54! 1
2

.(‘V C; Alt 6 L. .
. .'A ..3“:A>Jncn.fi.

x £32 2. t: 9 ..
é... . u a .x‘nhhwu‘.
.._..n...uux..u..w! u... i. 8..

., :35... F9. 52.. .

Lazihx.

- w

.4 52...... . v
1,. w

:33’ :‘
52...}..‘d5
\az.t.z

.5 :0). In .L .
.II..... 17.6; 5\ l:
‘91.: r 332:4 )3}:

‘ 4;. $941.51... ‘1‘
:‘1: IY§1IL5151.

 

 

 

 

was... lllllflllllllllll”I”MINIMUM!WW‘IlnlllHllll

‘3 3014112902 , LIBRARY
(R 03.5; ,, Mlehlgan State
University

 

 

This is to certify that the

dissertation entitled
Investigations of Modified Edman Reagents
and Modified Anabolic Steroids for the

Electrospray Mass Spectrometric Detection

prese1_ ted by

Kuruppu A. N. Dharmasiri
has been accepted towards fulfillment

of the requirements for

Ph . D . degree in Chemistry

swim

Major professor

Dated) NH ”7%

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

PLACE ll RETURN Boxmmnwofltbdnekommtnyourneow.
TO AVOID FINES Mum on or More date duo.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU leAn Afflnnltlvo mauve“ Oppommuy 1mm
WIN-9.1

INVESTIGATIONS OF MODIFIED EDMAN REAGENTS AND MODIFIED
ANABOLIC STEROIDS FOR THE ELECTROSPRAY MASS SPECTROMTRIC
DETECTION

By

Kuruppu A. N. Dharmasiri

A DESSERTATION

Submitted to
Michigan State University
in partial fulfilment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

1 994

ABSTRACT

INVESTIGATIONS OF MODIFIED EDMAN REAGENTS AND MODIFIED
AN ABOLIC STEROIDS FOR THE ELECTROSPRAY MASS SPECTROMETRIC
DETECTION

By

Kuruppu A. N. Dharmasiri

Electrospray mass spectrometry (ESI/MS) has become a widely used technique
owing to its excellent sensitivity and capacity to mass analyze high-molecular weight
compounds due to multiply charging. This technique has gained wide acceptance in
biological mass spectrometry due to the ease of coupling with liquid chromatographic
systems. In this study, an electrospray source was designed and fabricated for the HP
5985 single quadrupole mass spectrometer. An interface between the mass spectrometer
and a personal computer has been developed.

The molecules containing pre-formed charged groups or basic functionalities
greatly improve the ESI/MS response. The phenylthiohydantoin amino acids in the
Edman degradation methodology are identified solely on the retention time on the
chromatogram after the separation by high performance liquid chromatography followed
by UV detection. New Edman reagents containing a basic pyridyl group were introduced
to improve the detectability and the reliability of thiohydantoin derivatives using ESI/MS.

The low level detection of anabolic steroids is difficult due to the extensive
fragmentations under the conventional gas chromatography-mass spectrometric
techniques. Anabolic ketosteroids were modified to form hydrazone derivatives or oxime
derivatives with Girard’s T reagent or hydroxylamine hydrochloride, respectively. The
modifications improve the detectability into the femtogram level using electrospray mass

spectrometry. Tandem mass spectral studies showed that a loss of 59 Daltons was

common to all derivatives. A selective method to detect the hydrazone derivatives was
developed using the neutral loss scanning mode of a triple quadrupole mass spectrometer.
Feasibility studies based on analysis of stanozolol using electrospray mass spectrometry
combined with the neutral loss scanning mode or liquid chromatography showed that this

technology is suitable for detecting stanozolol in urine samples.

to my parents

iv

ACKNOWLEDGMENTS

I would like to thank my research advisor, Dr. J. Throck Watson, for his support
and encouragement throughout my graduate school career. I would also like to thank Dr.
Doug Gage for continual encouragement he has given, Dr. Z-H. Huang for sharing his
expertise on organic chemistry, Dr. Rawl Hollingsworth for his instruments, Dr. Jack
Holland for advice on instrumentation, and Drs. John Allison and John McKraken for
serving as committee members. I would also like to thank Mike Davenport for helping me
build my confidence with electronics. This list is incomplete without Joe Laykem who
gave me very helpful suggestions on HPLC and peptide chemistry.

I want to acknowledge Mel, Melinda, Bev, and Curt along with the rest of the
crew at the MSU-Mass Spectrometry Facility for their assistance and for making a
pleasant working environment. I would also thank to Dr. Eugene Zaluzec for his
automechanic lessons in addition to useful advice on chemistry.

I enjoyed the cooparation of my friends Tim Nieuwenhuis, Kate Noon, and Patrick
Lukulay. I thank them for their moral support in getting accustomed to MSU. I enjoyed
working with past and present Watson group members.

I am greatly indebted to my parents for creating in me a desire for the higher
education. Finally, I thank my family for being patient with me and for being

understanding.

TABLE OF CONTENTS

LIST OF FIGURES
LIST OF TABLES
CHAPTER 1 INTRODUCTION

1. Introduction and Objectives ........................................................................................... 1
I]. History and Development of Electrospray .................................................................... 2
111. Operating Principles and Mechanisms of Electrospray ................................................. 4
A. Production of Charged Droplets at the E81 Capillary tip .................................... 5
1. Efl‘ect of Surface Tension on E81 .................................. . .......................... 7

2. Dependence of Droplet Radius on Liquid Flow Rate ............................... 10

B. Shrinkage of Charged ES Droplets ..................................................................... 10

C. Formation of Gas-phase Ions ............................................................................. 12

1. Charged Residue Model .......................................................................... l3

2. Iribame and Thomson Ion Evaporation Theory ....................................... 13

3. Emission of Gas-Phase Ions from Taylor Cone Tip ................................. 16

D. Dependence of the Ion Intensity on the Concentration ....................................... 16

IV. Characteristics of ESI/MS Data .................................................................................. 20

A. Types of Analytes Required for ESI/MS ............................................................ 20

B. Interpretation of Multiple Charged Envelope ...................................................... 23

C. Why Derivatives ? .............................................................................................. 24

D. Oxidation of Analytes under the High Electric Field (ESI conditions) ................. 24

V. Mass Analyzers Coupled with ESI ................................................................................ 25

A Quadrupole Mass Analyzer ................................................................................ 25

B. Magnetic Sector Instrument ............................................................................... 25

vi

C. Quadrupole Ion Trap Mass Analer .................................................................. 26

D. Fourier Transform Mass Analyzer ...................................................................... 27

E. Time of Flight Mass Spectrometer ..................................................................... 27

VI. Sample Introduction Techniques ................................................................................. 28
A.Coupling with Liquid Chromatographic systems ................................................. 28

B. Capillary Electrophoresis (CE) ........................................................................... 29

C. Ion Chromatography (IC) .................................................................................. 30

VII. Applications of ESI-MS ............................................................................................ 30
A. ESI/MS of Small Molecules ............................................................................... 31

B. Multiple Charging and High Molecular Weight Determination ............................ 31

1. Large Polypeptides and Proteins ............................................................. 31
2. Other High Molecular Weight Compounds .............................................. 32

C. Studies of Higher Order Structures .................................................................... 33

VIII. References ............................................................................................................... 3 5

Chapter 2 DEVELOPMENT OF INSTRUMENTATION FOR ESI/MS

I. Introduction .................................................................................................................. 42
11. HP 5985 Single Quadrupole Mass Spectrometer .......................................................... 43
A Operating principles of Quadrupole Mass Spectrometer ..................................... 44
B. Vacuum System of HP 5985 MS ........................................................................ 45
111. Design Considerations for the ESI Interface ................................................................ 47
A. Vacuum System and Sensitivity ......................................................................... 47
B. Transport of Gas and Ions into Vacuum ............................................................. 48
VI. Fabrication of Source and Flange ................................................................................ 50
A Instrumentation for Therrnospray MS ............................................................... 52
B. SIMION Ion Simulation .................................................................................... 53
C. Source Modification ......................................................................................... 55
D. Vacuum Flange Modification ............................................................................. 58

VII

E. Design and Fabrication of Ion Transport Capillary .............................................. 58

F. Electrospray Needle Assembly ............................................................................ 63

G. Electropolishing of the Capillary Tubes .............................................................. 64

V. Results and Discussion ................................................................................................. 67
A Data Collection .................................................................................................. 69

VI. Design and Construction of Computer Interface .......................................................... 72
A. Introduction ....................................................................................................... 72

B. Computer Interface ............................................................................................ 76

C. Data Acquisition and Processing Program .......................................................... 77

VII. TQMS ....................................................................................................................... 82
A. MS/MS Scan Modes with TQMS ...................................................................... 85

l. Daughter Ion Scan .................................................................................. 85

2. Parent Ion Scan ...................................................................................... 86

3. Functional Relationship Scan .................................................................. 86

4. Selected Reaction Monitoring ................................................................. 87

VIII. References ............................................................................................................... 91

Chapter 3 PEPTIDE SEQUENCING

I. Introduction .................................................................................................................. 93
A. Edman Degradation ........................................................................................... 94

II. Thiohydantoin Amino Acid Analysis by Mass Spectrometry ......................................... 98
A Electron Impact (E1) ......................................................................................... 98
B. Chemical Ionization (CI) .................................................................................... 99
C. Two-Step Laser Desorption/Multiphoton Ionization .......................................... 100
D. Thermospray ................................................................................................... 101
E. Atmospheric Pressure Chemical Ionization ........................................................ 102
F. Electron Capture Negative Ionization ................................................................ 102
G. Electrospray ..................................................................................................... 102

viii

111. Other Derivatives ........................................................................................................ 103

A. Fluorescent Derivatives ..................................................................................... 103
B. Radioactive Derivatives .................................................................................... 105
IV. Automated peptide Sequenator ................................................................................... 107
A Liquid Phase ..................................................................................................... 107
B. Solid Phase ....................................................................................................... 108
C. Gas Phase ......................................................................................................... 109
V. Other Mass Spectrometric Techniques for Peptide Sequencing ..................................... 109
A. Fast Atom Bombardment .................................................................................. 110
B. Electrospray ..................................................................................................... 111
C. Matrix Assisted Laser Desorption ..................................................................... 115
VI. References .................................................................................................................. 116

Chapter 4. DEVELOPMENT OF EDMAN REAGENTS FOR ESI/MS DETECTION

I. Introduction .................................................................................................................. 122
H. Alternative Edman Reagents ......................................................................................... 123
A. t-BAMPITC ..................................................................................................... 123

1. Introduction ............................................................................................ 123

2. Experimental ........................................................................................... 125

3. Results and Discussion ............................................................................ 126

B. Designing New Reagents .................................................................................. 127

III. Preparation of Isothiocyanate ...................................................................................... 127
A Method I ........................................................................................................... 128

B. Method 11 .......................................................................................................... 128

C. Separation of Mixture ........................................................................................ 132

1. Preparative Thin Layer Chromatography ................................................. 132

2. Column Chromatography ........................................................................ 133

3. Results and Discussion ............................................................................ 134

ix

D. Characterization of Isothiocyanates .................................................................... 134

1. Colorimetric Detection ............................................................................ 134

2. Mass Spectrometry ................................................................................. 135

IV. Thiohydantoin amino acids .......................................................................................... 136
A. Synthesis ........................................................................................................... 136

B. Extraction .......................................................................................................... 136

C. Mass Spectral Analysis of Thiohydantoins .......................................................... 137

D. ESI/CID/MS/MS ............................................................................................... 144

V. Studies on Reaction Completeness ............................................................................... 150
A. Experimental ..................................................................................................... 150

B. Results and Discussion ....................................................................................... 150

l. Coupling Reaction with PMITC .............................................................. 150

2. Results and Discussion ............................................................................ 151

D. Cleavage Reaction of PyTC-peptides ................................................................. 151

E. Results and Discussion ....................................................................................... 153

VI. Peptide Sequencing ..................................................................................................... 155
VII. Post-translationally Modified Amino Acids ................................................................ 156
VIII. Conclusion ............................................................................................................... 160
IX. References .................................................................................................................. 161

Chapter V. MODIFICATION OF ANABOLIC STEROIDS FOR ESI/MS

I. Introduction .................................................................................................................. 163

11. Modification of Steroids for ESI/MS ............................................................................ 165

III. Hydrazone Derivative Formation ................................................................................. 166
A Mass Spectral Analysis ...................................................................................... 166
B. Detection Limits of Hydrazones ......................................................................... 170
C. ESI/MS/MS Studies of Hydrazones ................................................................... 174
D. Study of Fragmentations with Offset Energy and Pressure ................................. 175

X

E. FAB/MS Analysis of Hydrazones ....................................................................... 178

F. High Energy CID on F AB/MS ........................................................................... 180
G. Comparison of ESI/CID/MS/MS with F AB/CID/MS/MS .................................. 180
H. Neutral Loss Scanning Studies of Hydrazone Derivatives ................................... 182
I. LC/MS Analysis .................................................................................................. 182
IV. Feasibility Study of Steroid Detection in Urine ............................................................ 186
A. MS/MS Studies of the Derivatized Urine Extract ............................................... 188
V. Kinetic Studies of the Hydrazone Formation Reaction .................................................. 191
A. Dependence of Reaction Conversion (%) with Temperature ............................... 191
B. Dependence of Reaction Conversion (%) with Acid Concentration .................... 192
VI. Source CID during ESI/MS ........................................................................................ 192
A. MS/MS of the (M-59)+ Ion Derived from Methyltestosterone Hydrazone .......... 193
B. MS/MS of the (M-87)+ Ion Derived from Methyltestosterone Hydrazone .......... 194
VII. Oxime Derivative Formation ...................................................................................... 199
A. Results and Discussion ....................................................................................... 199
B. Tandem Mass Spectrometric Studies of Oximes ................................................. 205
C. Detection Limits of Steroid Oximes ............... ................................................... 205
VIII. Investigation of Stanozolol by ESI/MS ..................................................................... 207
A. Introduction ...................................................................................................... 207
B. Mass Spectral Analysis ...................................................................................... 208
C. Tandem Mass Spectrometry of Stanozolol ......................................................... 211
D. Comparison of ESI/MS/MS with FAB/MS/MS ................................................. 213
B. Effect of Solvent Composition ........................................................................... 215
F. ESI/MS/MS of CH3CN Adduct Ion of ST ......................................................... 215
G. HPLC of Stanozolol .......................................................................................... 218
H. LC/MS of Stanozolol ........................................................................................ 218
I. Urine Extract ...................................................................................................... 220

xi

IX. Conclusions ................................................................................................................ 221
X. References ................................................................................................................... 224

Figure. 1-1.
Figure 1-2.

Figure 1-3.

Figure 1-4.

Figure 1-5.

Figure 1.6.

Figure 2-1.

Figure 2-2.

Figure 2-3.

Figure 2-4.

Figure 2-5.

Figure 2-6.

List of Figures

Components of the electrospray mass spectrometer used in this study ....... 5
Schematic representation of life time of parent and offspring droplets ....... l4

Schematic representation of Initial state (a) and transition state (b),
proposed by Iribarne and Thomson. Evaporating ions leaves as a cluster
M+(SL)m, where SL are solvent molecules, r is the distance of ions from
surface ..................................................................................................... 18

ESI/MS mass spectrum of 1, 10 phenanthroline (2.7 x 10‘5M) and
(CH3)4N1 (6.5 x ro-8M) in methanol/water (SO/50% v/v). Adapted from
ref 12 ....................................................................................................... 21

The electrospray mass spectrum of myoglobin (M.W. = 16,951). The
data were collected on a TSQ 700 mass spectrometer. 10 pmol were
consumed to obtain the spectrum ............................................................. 22

Hypothetical mass spectrum containing multiply charged ions ................... 23
Stability diagram of (a, q) space showing regions (patchwork area) that
corresponds to mathematically stable ion trajectories in the quadrupole
mass spectrometer .................................................................................... 46
Expansion of gas into vacuum; (a). simplified scheme with shockwaves
and silence zone; (b). generation of a beam of gas and ions by sampling
from a silent zone, the skimmer penetrates the mach disk; (c). sampling of
gas and ions with a skimmer located beyond the mach disk ....................... 51
The Vestec thermospray interface for the HP 5985 mass spectrometer ..... 54
The plots from the SIMION ion simulation. (a) Heated capillary is right

angle to the axis of the mass analyzer. (b). Heated capillary is parallel to
the mass analyzer. In both cases, source block is at ground potential ...... 56

The ESI source designed for the HP 5985 mass spectrometer ................... 60
The ion transport capillary tube designed for the HP 5985 mass
spectrometer ....................... ............ .. .......................................... 62

xiii

Figure 2-7.

Figure 2-8.
Figure 2-9.

Figure 2-10

Figure 2-11.

Figure 2-12.

Figure 2-13.

Figure 2-14.

Figure 2-15.

Figure 2-16.

Figure 2-17.

Figure 2-18.

Figure 2-19.

Figure 2-20.

Figure 3-1.

Figure 3-2.

Electrospray Needle. The probe can be used with or without nebulizer
gas ........................................................................................................... 65

The complete electrospray interface on the HP 5985 mass spectrometer...66
The circuit diagram for measuring ESI current ......................................... 73
ESI/MS mass spectrum of Lys-Tyr-Lys (M.W. = 43 7) ............................. 74

ESI/MS mass spectrum of Angiotensin III (M. W. = 917). Flow rate =
lOuL/min. ESI needle at 3.9kV. Heated capillary at 100°C and 100V and

the Skimmer at 15V ................................................................................. 75
The block diagram of the digital interface for the HP 5985 mass

spectrometer ............................................................................................ 78
The schematic diagram of the CIO-DIO48 Interface board ....................... 79

The schametic diagram of the custom made adapter to go into the digital
board on the HP 5985 mass spectrometer ................................................. 80

The cables from CIO-DIO48 board to the HP 5985 mass spectrometer....8l

The control panel of the modified data system for the HP 5985 mass
spectrometer ............................................................................................ 83

The ESI/MS mass spectrum of stanozolol acquired using the new data
system ...................................................................................................... 84

The reconstructed total ion chromatogram of a LC/ESI/MS analysis of
(100pg) testosterone hydrazone. C-18 (15cm x 800 m, 3 urn particles)
column. Flow rate 40pL/min of CH3CN/HZO/TF A (40/59.9/0.1 %v/v/v).
Scan rate 350-450 in 1 sec ....................................................................... 84

The Finnigan Electrospray source on the TSQ 700 mass spectrometer. It
incorporates a tube lens and an octapole rods to improve the ion

transmission efficiency .............................................................................. 89
The TSQ 700 triple quadrupole mass spectrometer ................................... 90
Chemical Scheme for Edman Degradation of Peptide ............................... 97
Structures of proposed fluorescent reagents for Edman sequencing .......... 106

xiv

Figure 3-3.

Figure 3-4.

Figure 4-1.

Figure 4-2.

Figure 4-3.
Figure 4-4.
Figure 4-5.

Figure 4-6.

Figure 4-7.

Figure 4-8.

Figure 4-9.

Figure 4-10.

Figure 4-11.

Figure 4-12.

Fragmentation of the peptide backbone and nomenclature proposed by
Roepstorff and Fohlmann ......................................................................... 112

Structures of the common ions encountered in FAB-CAD-MS/MS and
their nomenclature as revised by Biemann ................................................. 113

The structures of conventional Edman reagent (a), phenylthiohydantoin
derivative (b), reagents proposed by Aebersold et al. (c), and Basic et al.
((1) ............................................................................................................ 124

Structures of t-BAMPITC and arninomethylphenylthiohydantoin amino
acid .......................................................................................................... 125

The modified Edman reagents .................................................................. 127
The reaction scheme of isothiocyanate formation by Makaiyama Method. 130
Reaction scheme for the synthesis of DPT ................................................ 131

Scheme of isothiocyanate formation using DPT isothiocyanate transfer
reagent ..................................................................................................... 131

FAB/CAD/MS/MS spectrum of val-PMTH .............................................. 138

Portion of the ESI/MS mass spectrum of PyTH-val (5pmol). Sample was
injected into a stream of CH3OH/HzO/HOAC (50/49.9/0.1 %v/v/v) at a
flow rate of IOOuUmin ............................................................................ 140

The mass spectral profile obtained by injecting increasing amounts of
purified PyTH-phe derivative into a stream of CH3 OH/HZO/HOAC
(50/49.9/0.1%v/v/v) at a flow rate of lOOuUmin ...................................... 141

Portion of the ESI/MS mass spectrum of PMTH-val (5pmol). Sample
was injected to a stream of CH3OH/HzO/HOAC (50/49.9/0.1 %v/v/v) at
a flow rate of lOOuI/min ......................................................................... 142

The profile obtained for the replicate injections of the PM-thiohydantoin
amino acid in selected ion (m/z 298) monitoring mode. The conditions

were maintained as in the previous experiment ......................................... 143
Proposed fi'agmentation patterns for the PyTH-lys under
ESI/CID/MS/MS ..................................................................................... 144

Figure 4-13.

Figure 4-14.

Figure 4-15.

Figure 4-16

Figure 4-17.

Figure 4-18.

Figure 4-19.

Figure 4-20.

Figure 4-21.

Figure 4-22.

Figure 5-1.

Figure 5-2.

ESI/MS/MS mass spectrum of PyTH-lys. Collision gas pressure was at
3mtorr and energy was at 30eV. Sample was infused at a flow rate of
3|.tL/min. Mass spectrometer was scanned 100 - 450 Daltons in 0.5 sec....145

ESI/CID/MS/MS spectrum of PMTH-asp. Collision energy at 30eV and
pressure at 3 mtorr and the proposed fi'agrnentations ................................ 146

ESI/CID/MS/MS mass spectrum of PyTH-leu and the proposed
fragmentation pattern. Collision energy 30eV and pressure 3 mtorr .......... 148

ESI/MS/MS mass spectrum of PyTH-ile amino acid derivative and the
proposed fragmentation pattern. Conditions were similar to those of
Figure 4-15 .............................................................................................. 149

Kinetic assessment of coupling reaction between pyridyl isothiocyanate
and bradykinin (RPPGFSPFR). The peptide was treated with 2.5%
PleC in a buffer solution and coupling reaction was assessed by
analyzing the reaction mixture by RP-HPLC ............................................. 152

Kinetic assessment of coupling reaction between pyridylrnethyl
isothiocyanate and bradykinin (RPPGFSPFR). The peptide was treated
with 2.5% PMITC in a buffer solution and coupling reaction was
assessed as in Figure 4-17 ........................................................................ 152

F AB/MS mass spectra showing the cleavage of thiocabamyl peptide
(VGVAPG) with anhydrous TFA at 50°C. (a) PyTC-peptide afier 10
min., (b). PMTC-peptide afier 2min., and (c) PTC-peptide after 5
min .......................................................................................................... 154

HPLC chromatogram for the PTH amino acid standard containing
hydroxy-proline. Hydroxy-proline eluted as two peaks ............................. 157

ESI/MS mass spectrum of hydroxy-proline (5pmol). Sample was injected
into a stream of stream of CH3OH/I120/I-IOAC (50/49.9/0.1 %v/v/v) at
a flow rate of lOOuL/min .......................................................................... 158

The ESI/MS mass spectrum of PMTH-O-phospho serine (5pmol).
Sample was injected to a stream of CH3OH/HzO/HOAC (50/49.9/0.1

%v/v) at a flow rate of 100uL/min ........................................................... 159
Numbering system for the steroid carbon skeleton .................................... 167
The Structures Anabolic Steroids banned By IOC .................................... 168

Figure 5-3.

Figure 5-4.

Figure. 5-5.

Figure. 5-6.

Figure 5-7.

Figure 5-8.

Figure 5-9.

Figure 5-10.

Figure 5-11.

Figure. 5-12.

Figure. 5-13.

The reaction scheme for the formation of steroid hydrazone. The Girard's
T reagent reacts with the keto group of the steroid to form hydrazone
derivative ................................................................................................. 171

Portion of the ESI/MS mass spectrum of fluoxymesterone hydrazone
chloride (molecular weight of cation is 450u) ........................................... 172

Detection limits for the methyltestosterone hydrazone. The samples were
injected into a stream of CH3OH/H20/HOAC (50/49.9/0. 1% v/v)
flowing at a rate of 100uI./rr1in in increasing concentration. The injection
volume was SuL. The amounts are indicated above the corresponding

peak ......................................................................................................... 173

ESI/MS/MS spectrum of the molecular cation of methyltestosterone
hydrazone. The spectrum was obtained by infusing the sample at a flow
rate of 3|.LL/min. and at a concentration of 100 pg/uL. The mass
spectrometer was scanned at a rate of 980u/sec and spectra were
averaged for one minute. The collision energy was set at 40eV (lab). Ar
was used as the collision gas and the collision cell pressure was 3 mtorr... 17 7

Variation of product ion abundances with collision offset energy at the
collision cell gas pressure of 3mtorr .......................................................... 179

The variation of product ion abundances with collision cell gas pressure
at the collision offset energy of 40eV ........................................................ 179

The FAB/MS spectrum of methyltestosterone hydrazone (1 ng). Glycerol
was used as the matrix and matrix peaks are indicated by an asterisk ........ 181

The FAB/CAD/MS/MS spectrum of methyltestosterone ........................... 180

The neutral loss scan mass spectrum of nandrolone hydrazone .................. 184

Total ion current (RTIC) chromatogram for LC/MS analysis of
hydrazones of model compounds (a) and mass chromatograms for (b)
fluoxymesterone hydrazone (m/z 450), (c) methyltestosterone hydrazone
(m/z 416), (d) testosterone hydrazone (m/z 402) and (e) nandrolone
hydrazone (m/z 388) ................................................................................ 185

Total ion current chromatogram (RTIC) for LC/MS analysis of
hydrazones of urine extract (a) and mass chromatograms for (b)
fluoxymesterone hydrazone (m/z 450), (c) methyltestosterone hydrazone
(m/z 416), and (d) nandrolone hydrazone (m/z 388) ................................. 187

xvii

Figure. 5-14.

Figure 5-15.

Figure 5-16.

Figure 5-17.

Figure 5-18.

Figure 5-19.

Figure 5-20.

Figure 5-21.

Figure 5-22.

Figure 5-23.

Figure 5-24.

Figure 5-25.

Figure 5-26.

Figure 5-27.

Portion of the electrospray mass spectrum of extracted urine sample after
spiking with methyltestosterone (10ng/mL) followed by derivatization
with Girard's T reagent. The sample was infused at a flow rate of
5 uL/min without separation. The spectrum is dominated by signals from
the impurities in the urine sample .............................................................. 189

The simplified mass spectrum of the sample used to obtain figure 5-14,
by neutral loss scanning (59u) of the quadrupoles (Q1 and Q3). The
collision energy was set to 40eV and collision gas pressure at 3mtorr ....... 190

Variation of percent reaction conversion with time and temperature for
hydrazone formation ................................................................................ 195

Efi‘ect of HOAC concentration on percent reaction conversion ................. 195

ESI/MS mass spectrum obtained by inducing CID at the octapole. The
octapole ofi‘set was set to 40eV ................................................................ 196

The MS/MS spectrum of the ion at m/z (M-59)+ derived fi'om
methyltestosterone (m/z 416) by source CID ............................................ 197

The MS/MS spectrum of the ion at m/z (M-87)+ derived fi'om
methyltestosterone (m/z 416) by source CID ............................................ 198

Formation of testosterone oxime fiom testosterone and hydroxylamine
hydrochloride ........................................................................................... 201

ESI mass spectrum of methyltestosterone oxime in CH3OH/H20/HOAC
(50/49.9/0.1 % v/v/v) ............................................................................... 202

ESI mass spectrum of methyltestosterone oxime in ACN/HZO/HOAC
(50/49.9/0.l % v/v/v) ............................................................................... 203

ESI/CID/MS/MS spectrum of (M+CH3CN+H)"” for testosterone oxime.
The offset energy was set at SeV and the collision energy at 3 mtorr (Ar).204

ESI/CID/MS/MS spectrum of testosterone oxime .................................... 206

Structure of anabolic stanozolol, and preferred metabolic hydroxylation
sites indicated by an arrow ....................................................................... 209

ESI/MS mass spectrum of stanozolol in CH3OH and injected to a stream

of CH3OH/H20/HOAC (50/49.9/0.1 % v/v/v)- The capillary at 200°C
and needle at 4.5kV ................................................................................. 210

xviii

Figure 5-28.

Figure 5-29.
Figure 5-30.

Figure 5-31.

Figure 5-32.

Figure 5-33.

Figure 5-34.

ESI/MS/MS mass spectrum of stanozolol. Collision energy was 55eV
and the gas pressure was 3mtorr ............................................................... 212

The possible fragmentation pathways of stanozolol molecule .................... 213
FAB/CAD/MS/MS mass spectrum of stanozolol

ESI/MS mass spectrum of stanozolol in CH3OH as injected into a stream
of CH3CN/H20/HOAC (50/49.9/0.1 % v/v/v)- The capillary was at
200°C and needle at 4.5kV ....................................................................... 216

The ESI/MS/MS mass spectrum of the CH3CN adduct ion of ST. The
offset energy was at SeV and the collision cell pressure was at 2 mtorr ..... 217

The reconstructed mass chromatograms for (Mi-H)+ (m/z 329) and
(M+CH3CN+I-I)+ (m/z 370) for stanozolol (1ng injected) during an
LC/MS run ............................................................................................... 219

Selected reaction monitoring profile of stanozolol in urine extract
(10uL). The solvent system was CH30H/H20/HOAC (50/49.9/0.1
%v/v) at a rate of 100uL/min ................................................................... 223

xix

LIST OF TABLES

Table 1-1. Onset voltages Von and surface tension 7 for ES with different solvents... 8
Table 1-2. Experimentally determined coefficients for the equation 1-14. All

coefiicients relative to kgs+ = 1. These coefiicients are valid only at
concentrations above 10' M .................................................................... 19

Table 2-1: Function of the quadrupoles in various TQMS scan modes ....................... 87
Table 5-2. The steroids used to prepare hydrazone derivatives .................................. 170

Table 5-3. A list of peaks observed in MS/MS studies of hydrazone molecular cation
M+. ......................................................................................................... 176

CHAPTER 1
INTRODUCTION AND OBJECTIVES

I. INTRODUCTION

Since its discovery in 1912, mass spectrometry (MS) has developed into one of the
most powerfirl and versatile techniques available for the study of biomolecules. Over the
years, MS has been applied not only to problems of structure determination, but to
assessment of varied processes ranging from drug metabolism to enzyme kinetics. Rapid
advances in the biological sciences, especially in molecular biology and biochemistry, have
led to an increase in the demand for chemical and structural information on biologically
active molecules. Recent innovations in ionization methods, improvements in ion
detection, increases in mass range of instruments, the use of tandem instruments, and
improved interfaces for separation techniques are providing mass spectrometrists with the
capabilities needed to analyze a wide variety of biologically active materials. Recent
advances in mass spectrometric techniques such as electrospray ionization mass
spectrometry (ESI/MS) and matrix assisted laser desorption ionization mass spectrometry
(MALDI) have the capacity to establish molecular weights of biomolecules, even those
larger than 100,000 Daltons. Analytical methods involving MS have been reported with
detection limits in the femtogram (10'15g) and attomole (10'1 3 mol) range (1, 2).

In trace analysis of biologically important compounds, a single peak representing
the analyte molecule is desired for sensitive detection. The soft ionization techniques such

as ESI/MS or MALDI have the capabilities to give a single peak representing the
1

2
molecule. ESI/MS has shown distinct advantages in biological mass spectrometry,
because of its capabilities to retrofit into almost all the types of mass analyzers and
amenable to various kinds, of separation techniques. In this manuscript, ESI/MS in the
area of bioanalytical applications will be discussed in greater detail.

The intention of this chapter is to introduce (i) the principles of electrospray
ionization (ESI), (ii) the operating principles of ESI/MS, (iii) coupling of ESI with various
mass analyzers, (iv) coupling of various separation methods with ESI/MS, and (v) a brief
summary of applications of ESI/MS.

A. Objectives of the Research Project

The primary goals of this research investigative project are (1) the modification of
an HP 5985 single quadrupole mass spectrometer for electrospray ionization, (2) to
improve the reliability of Edman degradation chemistry by using mass spectrometry as a
detector, and (3) to develop LC/MS techniques to detect anabolic steroids in biological
fluids.

II. History and Development of Electrospray

Electrospray ionization-mass spectrometry (ESI/MS) has its origins in research
that long preceded the current flurry of activity. The study of the electrospray
phenomenon extends back perhaps over two and one-half centuries to the work of Bose in
1745 (3) and certainly to that of Zeleny early in this century (1917) (4). The seminal
research into the use of electrospray as an ionization method for macromolecules was due
to Malcom Dole and co-workers (5, 6), who more than twenty years ago, performed
extensive studies into the electrospray process and defined many of the important

experimental parameters. The purpose of Dole's studies was to use ESI to produce gas-

3

phase macro-ions. Dole's approach was largely an adoption of the electrospray studies
performed in the 1930's by Chapman, in which ion mobility studies were conducted for
electrospray ionization of low-molecular weight salt solutions (7). Experimental evidence
was presented by Dole for ionization of Zein (M.W.=50,000 Da) (8) and hen egg
lysozyrne (M.W.=14306 Da) (9). Interpretation of their results were problematic,
however, because a mass spectrometer was not available, and only ion retardation (5, 6, 8)
and ion mobility (9) measurements were obtained. ~

In 1984, ESI combined with mass spectrometry was reported, essentially
simultaneously, by both Yamashita and Fenn (10) and Aleksandrov et al. (11). Fenn and
co-workers also demonstrated ESI/MS in the negative ion mode (12), building upon the
original negative ion work by Dole and co-workers (5, 6, 8). Using magnetic sector
instruments, the Russian researchers independently demonstrated the on-line combination
with liquid chromatography in 1984 (11), and over the next few years applied ESI-MS to
oligosaccharides (13), and intact polypeptides up to molecular weight of 1500 Daltons
( l4), and developed chemical digestion-based methods for their sequence determination
(15).

Of particular importance to ESI analysis of macromolecules is the phenomenon of
multiple charging. The multiple charging of lysozyrne by ESI was reported by Dole on the
basis of ion mobility measurements (9), but difficulties in the interpretation led them to
suggest a charge state of only 3+, substantially lower than that shown by subsequent MS
studies. In 1985, both Penn and co-workers (16) and Aleksandrov et al. (13, 14) reported
dominant contributions for doubly charged ions for polypeptides such as bradykinin
(M.W=1060) and gramicidin S (M.W=1141). Most significant in the work by Penn and
co-workers was the observation of multiple charging with as many as 23 charges with
polyethylene glycols due to the attaChment of sodium ions (17). Since the higher-
molecular-weight polyethylene glycol samples had relatively broad molecular weight

4
distributions, only unresolved ”humps" of ions were observed for these materials. In 1988,
F enn and co-workers (18) first reported ESI/MS spectra of intact multiply protonated
molecules of proteins up to 40,000 Daltons having as many as 45 positive charges.
Several research groups quickly duplicated and extended the applications of ESI/MS (19-
21). Since its introduction in 1988 as a tool for the ionization of large biomolecules, the
practice of ESI/MS has grown impressively. Although its inception was nearly
simultaneous with MALDI mass spectrometry for similar applications (22), use of ESI/MS
has been substantially greater. The most important reasons are the ability to retrofit an
ESI source to existing mass analyzers easily and the speed with which commercial
instrumentation became available. ESI/MS has clear advantages for analysis of liquid
samples and solutions, particularly the efiluents fiom rnicroscale or nanoscale capillary
separation methods. In addition, tandem mass spectrometry has been shown to be usefill
for dissociation studies of the multiply charged biomolecules produced by ESI, allowing
much greater dissociation efliciencies than for singly charged molecular ions (23). ESI
combined with high resolution instruments such as magnetic sector (24) and Fourier
transform mass spectrometers (25) gives a better accuracy for high molecular weight
compounds. By combining the above capabilities with the high sensitivity of ESI/MS
serves as a very attractive tool for the mass spectrometry community, especially those who

are working with biomolecules.
111. Operating Principles and Mechanisms of the Electrospray Ionization Process

Important components of the electrospray mass spectrometric system used in this
study are shown in Figure l. The solution is forced through a small capillary tube held at
high electric potential. A fine mist or spray is formed at the tip of the ESI needle and is
drawn toward a heated capillary tube held at a lower potential. The analyte and solvent

5
molecules emerging from the capillary tube are sampled through a skimmer and ions are

focused by a set of lenses into the mass analyzer. The ions are detected by a continuous

 

 

 

 

 

 

 

dynode electron multiplier.
Heated C '11
ESI Needle 8‘” my / . l ,
Mass
Analyzer
3-5 kV \ .
100V

Skimmer Lenses

Figure. 1-1. Components of the electrospray mass spectrometer used in this study

There are four major processes in ESI/MS: the production of charged droplets
fi'om electrolyte dissolved in a solvent, shrinkage of charged droplets by solvent
evaporation and repeated disintegration, the mechanism of gas phase ion formation, and
secondary processes, by which gas-phase ions are modified in the atmospheric and ion

sampling regions of the mass spectrometer.

A. Production of Charged Droplets at the ESI Capillary Tip

When a capillary tube is held at high potential and placed near a counter electrode
at ground potential, an electric field between the capillary tube and the electrode is
created. The magnitude of the electric field at the capillary tip, EC, for a given potential,
V0, is given by the equation 1-1. (26, 27)

6
Ec=2Vc/rc ln(4wrc) 1’1

where rc is the radius of the capillary tube and d is the distance between the capillary tube
and the counter electrode. Equation l-l provides the field at the capillary tip in the
absence of a spraying solution. The field EC is proportional to the applied potential Vc on
the capillary. EC is essentially inversely proportional to re; it decreases slowly, with a
logarithmic dependence, with the distance d. For example, in a typical system of an
operating voltage Vc of 2kV, capillary radius rc of 0.2mm and distance d of 2cm produces
an electric field EC of3.3 x 106 V/m.

The imposed field E will also partially penetrate the liquid at the capillary tip.
When a positive potential is applied to the capillary tip, some positive ions in the liquid
will drifl toward the liquid surface and some negative ions will drift away from it until the
imposed field inside the liquid is essentially removed by this charge redistribution.
However, the accumulated positive charge at the surface leads to destabilization of the
surface because the positive ions are drawn downfield but cannot escape fi'om the liquid.
The surface is drawn out downfield such that a liquid cone forms. This is called a Taylor
cone (28).

At a sufficiently high field E, the cone is not stable and a liquid filament with a
diameter of a few micrometers, whose surface is enriched with positive charges, is emitted
from the Taylor cone tip. At some distance downstream, the liquid filament becomes
unstable and forms separate droplets. The droplets' surfaces are enriched with positive
ions for which there are no negative counter ions in the droplet. The length of the
unbroken filament decreases if the field E is increased. At higher fields, a multispray
condition is reached in which the central cone disappears and droplet emission occurs
from a crown of four to six short liquid tips formed at the rim of the capillary (29). This

ion separation mechanism, which is called the electrophoretic mechanism, is also most

7

plausible on energetic grounds. Pfeifer and Hendricks (27) were probably the first authors
who explicitly proposed and discussed the electrophoretic mechanism. More recently,
Hayati, Bailey and Tadros (30) also endorsed the electrophoretic charging mechanism in a
comprehensive examination of features of the ESI process. The number of excess charges
Q on a droplet is large relative to the number of solute molecules in the droplets. For
example, flash microphotographs taken under typical operating conditions for large
molecules (3 uL/min of lumol analyte in 1:1 methanol/water) show initial droplet
diameters of approximately 2.8um (31). Simultaneous measurement of spray current
indicates that each of those initial droplets has approximately 44,000 charges,
corresponding to a Q/Ni value of approximately 6 (31).

1. Effect of Surface Tension on ESI
Smith who also assumed the electrophoretic charging mechanism provided a very

useful equation (1-2) for the voltage required for the onset of charged-droplet emission
(32).

2Tr cosO _ ”2
V0n= A [—c—eo—QJ ln(4h/rc) 1-2

where 90 is the cone half-angle, so is pennitivity of free space, and A is a constant
By combining equation 1-1 with 1-2, one obtains the onset voltage Von

Von e 2 x105(7rc)1/21n(4d/rd 1-3

8

where y is the surface tension of the solvent, rc is the capillary radius, and d is the distance
between capillary tip and the counter electrode. Experimental verification of equation 1-3
has been provided by Smith (32) and Kebarle et al. (33, 34). For stable ESI operation one
needs to apply a few hundred volts higher than Von- Table 1 shows the evaluated Von
values for difi‘erent solvents for d = 4cm and rc = 0.1mm (35). The surface of the solvent
with the highest surface tension is the most difficult to stretch out into a Taylor cone and a

liquid filament; this leads to the requirement for the highest Von-

Table 1-1. Onset voltages Von and surface tension 7 for ESI with different solvents

 

 

 

 

CH3 OH CH3CN (CH3)2$O H20
gm (kV) 2.2 2.5 3.0 4.0
y (N/m2) 0.0226 0.030 0.043 0.073

 

 

Henion et al. (36) and Chait et al. (37) heated the electrospray needle assembly to
promote droplet formation when aqueous solutions are electrosprayed. At high
temperatures, the surface tension becomes low and the spray process takes place readily.
Buchanan et al. (3 8) studied the effects of surface tension and surface activity on
electrospray response. They investigated the ESI/MS response with varying the
concentration of cationic, anionic, and non-ionic surfactants. An increase in ESI/MS
response was observed with increasing concentration of cationic surfactant and non-ionic
surfactants and at a certain limit (0.3mM for Triton X-100), the ESI/MS response drops
ofl' sharply. These surfactants decrease the surface tension with increasing concentration

and alter a certain limit, surfactants may compete for the excess charges in the droplets.

9
The capillary current I which results fiom charged droplets leaving the capillary, is
of interest because it provides a quantitative measure of the excess positive electrolyte

ions produced by the spray. A theoretical derivation giving the dependence of I on
experimental parameters has been proposed by Hendricks (27).

1=AHVfEeon n=0.2-0.3 14

where A H is a constant that can be evaluated and depends on the dielectric constant and
surface tension of the solvent, Vf is the flow rate (volume/time), E is the electric field at
the tip, a is the conductivity of the solution, and v, e, n are experimentally determined
Mcients. Kebarle et al. compared the capillary current with the mass spectrometrically
detected intensity of protonated molecules and they found no close relationship (35).
When strong electrolytes are used at concentrations not exceeding 10'2 M, the

conductivity follows the relationship
0' = Amo C 1-5
By combining equations 1-4 and 1-5, equation 1-6 can be obtained.
1=AHVfVE3(A,,,0 C)" 1-6
The following coefiicients were experimentally determined; v = 0.5, e as 0.5, n = 0.2-0.3

(27). Because of the small exponents, the changes of I with flow rate, electric field, and

concentration of the electrolyte is small.

1 O
2. Dependence of Droplet Radius on Liquid Flow Rate
Fernandez de la Mora (39) has proposed a relationship for the radius R of the ESI
produced droplets from sprays

R a (pl/127913 1-7

where p is the density, If, flow rate, and y is the surface tension of the solvent. This
predicts that the radius of the droplets will increase with increasing flow rate. But in
ESI/MS, smaller droplets are desirable. Therefore, smaller volume flow rates are suitable
for better charging of the droplets.

B. Shrinkage of Charged ESI Droplets

The work by Gomez and Tang (39) and Davis and co-workers (40) has provided
data for charged droplets at a flow rate of SuIJmin with the concentration not exceeding
10'3 M. The droplets are small and have a narrow distribution of sizes, so they can be
considered monodispersed. The size distribution peaks at radius R0 as 1.5 um, and the
droplets have a charge of Q0 as 10'14C, which corresponds to N = 50,000 singly charged
ions (39).

The Rayleigh equation which gives the condition in which the charge Q becomes

just sufiicient to overcome the surface tension that holds the droplet together is
QR2 = 641t280‘Y RR3 1-8

where so is the permittivity of vacuum, 7 is the surface tension, RR is Rayleigh limit radius,
and QR is the Rayleigh limit charge.

11
Q? = 1.25 x10"10 RR3 1-9

The equation 1-9 gives the condition for methanol, whose y = 0.0226 N/mz; the numerical
factor for Q in coulombs and R in meters.

The initial charge Q0 10' 14C observed by Gomez and (Tang corresponds to only
50% of the Rayleigh limit charge QR for R0 = 1.5 pm. Gomez and Tang, who also studied
the charge-to-volume ratio of larger ESI-produced droplets, found that Q0 comes closer
to the Rayleigh limit for larger droplets.

Droplets with the radius of micrometer or larger are known to maintain their
charge; they do not emit gas-phase ions (40). The droplets shrink by evaporation of
solvent molecules until they come closer to the Rayleigh limit where they become unstable
and undergo fission into smaller droplets as shown in Figure 1-2.

Recent work has shown that droplets with sizes in the l-llm range fission
somewhat before, (at ~ 80% of the calculated value) the Rayleigh limit (39, 40). The
droplet will shrink by solvent evaporation until the radius R meets that condition, and then
fission will occur. The droplet does not split evenly into two smaller droplets of
approximately equal mass and charge (40). Typically, the droplets are observed to vibrate
alternatively between oblate and prolate shapes. These elastic vibrations stimulate
disruptions in which the ”parent” droplet emits a tail of much smaller offspring droplets.
This disruption pattern is similar to the disruption at the tip of the Taylor cone. The
emitted stream of offspring droplets carries ofi‘ only about 2% of the mass of the parent
droplet but 15% of the parent's charge. The radius of the offspring droplets, which are
quite monodispersed, is roughly one-tenth of the radius of the parent (41). A simple
calculation (42) shows that a 2% loss of mass leads to ~20 offspring droplets of this
radius. The offspring droplets are not only much smaller than the parent but also have a

much higher charge-to-mass ratio.

12
The time required for the parent droplet to reach the size R, that leads to first
fission can be estimated (42) with use of expressions providing the rate of solvent
evaporation fiom small droplets (43). When relatively volatile solvents are involved and
the droplets are a few micrometers in diameter or smaller, the evaporation rate follows the
surface evaporation limit law (43), which leads to a simple dependence (5, 40) of the

droplet radius on time 1,

WM
' 4ngT

where v is the average molecular velocity of the solvent gas, p0 is the vapor pressure of

R=Ro

 

1 1-10

the solvent at the temperature of the droplet, M molar mass of the solvent molecules, p is
the density of the solvent, Rg is the gas constant, and T is the temperature of the droplet.
The condensation coefficient, and its value is ~ 0.04 for both water and ethanol (42, 43).
The following equation for methanol evaporation at 35°C was obtained assuming a

condensation coefiicient, a = 0.04 for methanol and substituting the corresponding values.

R=Ro-1.2xlo-3t 1-11

The time dependence process is given in Figure. 1-3. For a methanol droplet
having initial diameter of 1.5 um and charge of 8 x 10'15C, at 35°C. The time required
for the first fission is ~400 us. The change of R due to fission is very small. Subsequent
fissions occur when the parent droplet again reaches 80% of the Rayleigh limit. These

fissions require shorter times, ~ 60 us, which decrease fi'om fission to fission.

C. Formation of Gas-Phase Ions
Despite an explosive growth in their use over the past few years, especially in ESI,

there has been very little identification and elucidation of component mechanisms in the

13
process by which solute species in a charged droplet are transformed into free gas-phase

ions. Two difi‘erent mechanisms have been proposed to account for the formation of gas-
phase ions fiom the small charged droplets: (a). charged residue mechanism, and (b).

Iribame and Thomson ion evaporation theory.

1. The Charged Residue Mechanism

The charged residue mechanism depends on the formation of extremely small
droplets R z lnm, which contain only one ion. Solvent evaporation fiom such a droplet
will lead to conversion of the droplet to a gas-phase ion. Such a mechanism was assumed
by Dole (5), the first investigator of gas-phase ion production by ESI. A more detailed
consideration and support of this mechanism was given by Rollgen (44).

The charged-residue model for macromolecular gas-phase ion formation proposed
by Dole and co-workers (5) over two decades ago is based on the combination of solvent
evaporation and coulombic explosions as shown in Figure 1-2, resulting in a droplet size

sufficiently small such that it contains only one molecular ion.

2. Iribame and Thomson Ion Evaporation Theory

The Iribame and Thomson theory (45, 46) assumes ion evaporation fiom very
small and highly charged droplets. Typically, the droplets from which ion emission
becomes competitive with Rayleigh fission have R z 8nm and N as 70 elementary charges
(45, 46). Under these conditions, the droplets do not undergo fission but emit gas-phase
ions. As N decreases, emission can still be maintained as a result of a decrease of R by
solvent evaporation. Thus, the Iribame mechanism does not require the production of
very small droplets (R as lnm) that contain only one ion. Iribame emission can occur even
when the droplet contains other solutes such as charge-paired electrolytes.

Iribame and Thomson based their theory on a derived equation that provides

detailed predictions for the rate of ion emission fi'om the charged droplets. In particular, it

14

g: 152250 51250
’ 0.945

I 0
At = 462 ms
‘

43560
43560 At = 74 ms 0.343
—> 0
+ j
384
O O O 0 37026 37026
0.09
20 droplets 0.844 0.761
——.>
At = 70 ms
+
326 00 O 0 37026
0-08 00.756
+ 278
0.03
278 O O O O -—-> 0
007 Al ‘3 39 ms
236
0 0.003
+
2 o o o o
0.003

Figure 1-2. Schematic representation of lifetime of parent and offspring droplets

Droplet at top left is a typical parent droplet created near the ES capillary tip at low flow
rates. Evaporation of solvent at constant charge leads to uneven fission. The numbers
beside the droplets give the radius R (um) and number of elementary charges N on the
droplet; t corresponds to the time required for evaporative droplet shrinkage to size where
fission occurs. Only the first three successive fissions of a parent droplet are shown.

15

predicts the dependence of the rates on the chemical properties of the ions. Observed
differences in the gas-phase ion intensities I A and I B of ions A+ and B+ present at equal
concentrations in the sprayed solutions were compared with predictions of the theory by
Iribame and Thomson and subsequently by other workers (47-50) as well. In general,
qualitative agreement between experiment and theory was obtained.

The Iribame treatment is based on transition state theory. The rate constant It] for

emission of ions from droplets is

a:
k, = %T exp(-AG IR?) 1-12

where k is the Boltzrnan constant, T is thetemperature of the droplet, and h is the Planck
constant. The fi'ee energy of activation, AG: , was evaluated on the basis of the model
shown in the Figure 1-3.

The barrier in the Iribame transition state is due to the opposing electrostatic
forces: the repulsion of the escaping ion by the other charges of the droplet and the
attraction between the escaping ion and the droplet because of the polarizability of the
solvent medium of the droplet. The attraction is larger at short distances between the ion
and the droplet surface, but it falls off faster than the repulsion as the distance is increased.

The equation for AG: was found (45, 46) to depend on four parameters. The first
two are N, the number of charges on the droplet, and R, the radius of the droplet. The
rate constant It] increases with N and decreases with R. The other two parameters express
the specific properties of the ions involved; the solvation energy and the distance of the
ion charges from the surface of the droplet. The escaping ion in Figure 1-3 is not the
naked M+ but an ion-solvent molecule cluster M+(sol)m containing m solvent molecules.
AG"F for the naked M+ is not the lowest, but when it leaves with several solvent

. . iF .
molecules, it has a lower actlvatlon energy. (for example, AG required to transfer Na+g

16
to the gas phase is ~ 98kcal/mol, whereas, Na+(HzO)7 requires only ~ 56kcal/mol) (42,
51). The strongly solvated ions have large transfer energies, and for these ions the

Iribame equation predicts large activation barriers and hence low rate constants.

3. Emission of Gas-Phase Ions from the Taylor Cone Tip
Very recently Siu and co-workers (52) proposed that the gas-phase ions are
emitted from the Taylor cone tip. The authors present evidence by obtaining ESI/MS
spectra in two different acids. Also, they compared mass spectra obtained in water and in

a nonvolatile solvent, ethylene glycol.

D. Dependence of Ion Intensity in ESI/MS on the Concentration of the Analytes

in the Electrosprayed Solution

When two electrolytes such as A+X' and BTY‘ are present in the solution, both
A+ and B+ ions will be present among the excess positive ions that constitute the charges
of the droplets. However, because of the very weak dependence of I on the total
electrolyte concentration, addition of BY to AX will not materially increase the current,
i.e., the total excess charge. On the other hand, B+ will compete with A+ among the
excess charges on the droplets. This means that the amount of gas-phase ions A+
produced from the charged droplets will decrease as BY is added to the solution. Kebarle
et al. (42, 53) introduced the relationship between I and the concentration of components

present in the solution.

‘QMfl

= LB
u-figmfl+@mql

 

17

The authors investigated the ion intensity from several alkaloids, quaternary
ammonium ions, and alkali metal ions in the presence of N114+ ions. A decrease in analyte
(A+) ion current was observed as predicted by the equation, when B+ = NH4+
concentration is increased while the Al” concentration is kept constant. When analyte
concentration increases while [A+]=[B+], the predicted pattern was observed above
10—5M. Experimentally determined ratios of coefficients for some compounds are given
in Table 1-2.

1. Comparison of Coefiicients with Iribame Theory and Charged Residue
Mechanism

Comparisons of the experimentally determined coefficients with the Iribame rate
constants and expected surface activities should be made only in the high concentration
range, where depletion is absent. The range of coefficient values in Table 1-2 is not very
large. This is a desired result for ESI/MS standpoint, and we can expect to detect with a
fair sensitivity any analyte ion present in the solution (3 5). A general correlation between
low solvation energies and high surface activities are expected. Because of the absence of
suitable quantitative information of solvation energies and surface activities in the
literature, it is difficult to establish the extent to which each factor contributes to the value
of the experimentally determined coefficient.

The charged residue mechanism, as originally stated (5, 44), did not provide
criteria for selectivity on the basis of physicochemical properties of the ions. It is obvious
that the surface activity is the factor that should lead to selectivity for the charged residue
mechanism. Ions that are enriched on the surface will preferentially end up in a final
droplet that contains a single ion. At present, it is not possible to state with certainty
which theory, ion evaporation or charged residue mechanism, fits better with the available

evidence. Fortunately, from a practical standpoint, qualitative predictions for high

18

(a) Initial state

/a.%a

Ion solvent cluster

(b) Transition State

 

Figure 1-3. Schematic representation of Initial state (a) and transition state (b), proposed
by Iribame and Thomson. Evaporating ion leaves as a cluster M+(SL)m, where SL are
solvent molecules, and r is the distance of ions from surface.

l9
selectivity on the basis of low solvation energy or high surface activity will both be valid,

because the two parameters are positively correlated (3 5).

Table 1-2. Experimentally determined coefficients for the equation 1-14. All coeficients

relative to sz+ = 1. These coeficients are valid only at concentrations above 10'5 M.

 

 

 

 

 

Ion k
Cs” 1
Li+ 1.6
Na+ 1.6
K+ 1.0
NH4+ 1.3
MorphineH+ 3
CodeineH+ 5
HeroinH+ 6
CocaineH+ 10
Ni2+(trypyridyl)2 5
(Butyl)4N+ 2
(Ethyl)4N+ 5
(n-propyl)4N+ 8
(n-pentyl)4N+ 14
n-C7H15NH3+ 10
n-C11H21NH3T 10

 

20

IV. Characteristics of ESI/MS Data

In recent years, ESI mass spectrometry has captured the interest of several
researchers owing to its high sensitivity, the simplicity of the interface, the extreme
mildness of the process, and the possibility of high molecular weight determination using
multiply charged ions. Because of the mildness of the ESI process, fragrnentations were
not observed under normal ESI conditions for most of the compounds. In the ESI
spectrum of small molecules, only peaks representing the original molecule were observed
(1, 2, 10, 12). The molecules that can form ionic adducts or that have pro-formed charges
in solution are observed in the ESI mass spectra as singly charged ions or multiply charged
ions. Figure 1-4 shows the ESI mass spectra of 1,10 phenanthroline and (CH3)4NI (12).
The most important finding in recent years is the multiple charging in the electrospray
ionization of proteins and peptides. This gives the ability to analyze biomolecules directly
fiom aqueous phase. The ESI mass spectra appear as an envelope of peaks representing
the parent molecule with different charge states. The adduct ion is usually H‘l’, Na+ (10,
12), Ca2+, or Mg2+ (54) ions . Figure 1-5 shows the ESI mass spectrum of myoglobin
(M.W. = 16,951). The peaks in the ESI mass spectrum corresponds to the myoglobin

molecule, but with the addition of difierent number of H+ ions.

A. Types of Analytes Required for ESI/MS

On the basis of previous work, it has become clear that ESI is a method by which
ions present in solution are transferred to the gas phase (55-59). The types of analytes
that can be analyzed by ESI/MS are those can form ionic species in solution. ESI transfers
ions from solution to the gas phase for a variety of ions: the simple singly charged
electrolytes such as Na+ and Cl'; group 11 ions such as Ca2+, Sr2+, and Ba2+, as well as
doubly and triply charged transition metal and lanthanide ions and complexes thereof;

protonated basic and deprotonated acidic compounds, compounds having pre-formed

21

charges, bioorganic ions such as multiply protonated peptides and proteins; and

deprotonated negatively charged nucleic acids (3 5).

 

 

 

m1"
“' 9%

8‘

'63

5 “waist”

‘65

H

o

.E Ma'H+
.9.

o l I I

M W

m/z

Figure 1.4. ESI/MS mass spectrum of 1, 10 phenanthroline (2.7 x 10'5M) and (CH3)4NI
(6.5 x 10"8 M) in methanol/water (SO/50% v/v). Adapted from ref. 12.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

22

 

 

 

 

 

 

 

 

 

 

 

1131
100-
9r3
999
so~
893
60- 12,3 4-12
1414
8Q1
1395
,0- 7721
+-ll
1542
H 4-10
711 1696
20- W
707
‘1
TA LIMA
I ' l I ' l "—I ' 1
600 800 1000 1200 1400 1600 1800 2000

   

Figure 1-5. The electrospray mass spectrum of myoglobin (M.W. = 16,951). The data
were collected on a TSQ 700 mass spectrometer. 10 pmol were consumed to obtain the

spectrum.

23
B. Interpretation of Multiple Charged Envelope

(n+1)

(n+2) n+

Intensity

(n+3)

 

 

 

 

m1 m2 m3 m4

Figure 1-6. Hypothetical mass spectrum containing multiply charged ions.

where m1, m2, m3, and m4 are m/z values corresponds to peaks containing the number of
charges (n+3), (n+2), (n+1), and 11 respectively.

Assume the molecular weight of the compound is M, and the number of charges on the
other ions as labeled. The adduct ion is x+ and usually it can be 11+, Na+, 1014+, Ca2+

 

ong2+.
m3 = M+(n+1)X+ 144
n+1
+
m4 = fl 1-15
11

by solving the above equations, charge state 11 and molecular weight M can be obtained.

24

n = ...—{ELL}— 1'16
m4-m3

M =n (m4-1) 1-17

C. Why Derivatives ?

The ESI technique transfers ions fi'om solution to the gas phase. The molecules
without pro-formed charges, acidic or basic filnctional groups cannot be easily detected
under ESI/MS conditions because they do not exist as ionic species in solution. For a
molecule without a ESI active group, one of the above mentioned functional groups must
be introduced to obtain a reasonably high response (1, 54). In chapter 5, the enhancement
of response of neutral steroid molecules after modification with a quaternary ammonium

group will be discussed in detail.
D. Oxidation of Analytes under the High Electric Field

In the electrospray process, highly charged droplets are formed under a high
electric field. Normally, in the positive ion mode, a high voltage of positive value is
applied to the ESI needle tip carrying the sample solution. Proof for the occurrence of
electrochemical oxidation at the metal capillary was provided by Blades et al. (48). When
a Zn capillary tip was used, the release of Zn2+ to the solution could be detected with
ESI/MS. Similar results were observed with stainless steel capillaries that releases F «9+
to the solution. The electrospray-induced oxidation due to the high electric field has been
observed for peptides containing methionyl, tryptophanyl, or tyrosyl residues (57). The
oxidation process is dependent on the field strength and the flow rate of the solution

passing through the capillary tube.

25
V. Mass Analyzers for Electrospray Mass Spectrometry
The intention of this part is to introduce the various types of mass analyzers that
have been coupled with an ESI source. These mass analyzers have their own unique

capabilities and advantages over one another.

A. Quadrupole Mass Analyzers
The most commonly used mass analyzer for ESI/MS is the quadrupole mass filter.

The quadrupole mass analyzers can tolerate a little higher pressures than the other types of
mass analyzers. They have relatively higher scan rates. The mass range of new
quadrupole mass spectrometers have been extended to 3000 Daltons. The acceleration
voltage of a quadrupole mass spectrometer is about 5-15V. The resolving power of most
commercial quadrupole mass spectrometers is limited to unit mass resolution. The details

of quadrupole theory will be discussed in chapter 2.

B. Magnetic Sector Instruments

There are benefits in interfacing an electrospray ion source to a high-performance
magnetic sector mass spectrometer. The major benefits are better mass accuracy, higher
mass resolution, higher mass (mass/charge) range, and the possibility of high energy
collisional activation of multiply charged ions (58). Better mass accuracy, for example, is
useful in determining post-translational modifications as well as point mutations in
proteins. Higher resolution aids in mixture analysis and extends the mass limit of
biopolymers that can be mass analyzed by providing better separation of the multiple
charged ions. A higher mass-range instrument (4000 Daltons) fitted with an electrospray
ion source permits observation of high molecular weight materials that do not support

sufiicient charges to fall into the mass range of a quadrupole instrument (59).

26

Aleksandrov et al. (10) interfaced electrospray ionization to a magnetic sector
mass spectrometer and reported obtaining singly charged ions of small molecules using
10-100 nmol of material. Fragment ions were obtained by increasing the potential
difference between the aperture plates separating the atmospheric-pressure region fi'om
the vacuum region of the source-housing chamber. Allen and Lewis reported interfacing
an electrospray ion source to a single-focusing magnetic-sector mass spectrometer (60).
They were able to observe doubly and singly charged ions for grarnicidin (M.W.=1141
Da) and bradykinin (M.W.=1060 Da).

Early results with ESI on a magnetic instrument produced spectra of peptides but
the sensitivity was poor relative to that reported for quadrupole instruments, and in
addition, highly charged ions fi'om proteins were not observed (58). Meng et al. (61)
improved the sensitivity and resolution by reducing the pressure in the kilovolt
acceleration region. They accomplished this by incorporating two stages of rotary
pumvins-

ESI on the magnetic sector mass spectrometers has been used to determine the
accurate masses of proteins and peptides (62-64). Sequence information has been
obtained by in-source fragmentation of electrospray-generated peptide ions. The complete
sequence of several peptides has been determined fiom their fragment ions consuming

only low picomole amounts of material (64, 65).

C. Quadrupole Ion Trap Mass Spectrometry
Use of the quadrupole ion trap mass spectrometer has undergone remarkable

growth in the past decade due largely to new developments in ion trap technology and the
subsequent commercial introduction of analytical instruments based on the three-
dirnensional quadrupole (66). The development was catalyzed by the introduction of the

mass-selective instability mode of operation (67) and the use of about 1 mtorr of He as a

27
bath gas to improve sensitivity and resolution (68). The development includes, for
example, techniques for mass spectrometry/mass spectrometry (MS/MS) (69), multiple
stages of mass spectrometry (MSn, n>2) (70), mass extension (71). Dissociation
eficiency was reported to be greater than 50% (72).

Van Berkel et al. (73) reported the coupling of an ESI source with an ion trap
mass spectrometer. The authors described the ESI/MS analysis of various classes of
molecules including, direct red 81, bradykinin, melittin, cytochrome c, myoglobin, and
bovine albumin. The ability to obtain more information was demonstrated by employing
M83 and MS4 dissociation experiments for multiply charged peptide ions.

D. Fourier Transform Mass Spectrometry

The potential of FT ion cyclotron resonance mass spectrometry for precise mass
measurement has long been recognized (74). Fourier transform mass spectrometry
(FTMS) has a capability to achieve very high resolving power. The pioneering work by
Mclarfi'erty and co-workers (75) led to the development of ESI on FTMS. Recently they
reported the resolving power of 5 x 105 for a 29-kDa protein (for the ion having 21+
charges) with less than l-ppm mass measuring error (75). Winger et al. (76) reported the
resolving power of 700,000 for bovine insulin (4+ charge state).

E. Time of Flight Mass Spectrometry

The time of flight mass analyzer, is an inherently pulsed techniqUe requiring
spatially and temporally well-defined ion packets with which to conduct mass analyses.
Due to the high fi'equency of charged droplet formation and the droplet spatial
randomization during evaporation and ion production, the ion current entering the vacuum
through the nozzle or capillary from atmospheric pressure is essentially a continuous

current with no discernible periodic oscillation (77, 78). Attempts to couple these two

28
techniques necessarily impose a penalty in terms of sensitivity because of reduced duty
cycle. Boyle et al. (78) introduced an electrospray ion storage time-of-flight mass
spectrometer. Recently, Lubman and co-workers (79) reported on a time of flight mass
spectrometer coupled with an ion trap capable of storing and sampling ESI produced ions
continuously. This technique is capable of obtaining the typical sensitivity and speed of

time of flight mass spectrometry in coupling with a liquid chromatographic system.

VI. Sample Introduction Techniques

A. Coupling with Liquid Chromatographic Systems

The ESI/MS is a liquid ionization technique and it allows to introduce the sample
in solution. The most common way of introducing a sample is through a flow injection or
by infusion using a infilsion pump. High-resolution separation prior to mass spectral
analysis of biological materials is highly desirable since essentially all biochemical systems
of interest are comprised of mixtures of varying complexity. Even proteins in a highly
”purified” state are often mixtures due to naturally occurring microheterogeneity, or are
converted to mixtures prior to analysis by chemical or enzymatic degradation procedures.
Since sample sizes are generally limited, there are strong incentives to conduct
biochemical research on the smallest scale possible. It is not surprising therefore that
combined separations-ESI-MS analysis is of broad interest and that greater sensitivity is
almost always desired.

Liquid chromatography is used to separate a wide variety of biologically important
compounds, and has been interfaced to the ESI mass spectrometer. For unassisted ESI,
flow rates of l-IOuI/min are the best for stable operation, depending somewhat on the
solution composition (17, 80). Packed capillary LC has also been coupled to mass
spectrometry using both unassisted (81) and assisted ESI (82). Capillary diameters of
typically 250nm, and as low as 50 pm with the flow rates of less than a few llL/min have

29
been used. Now it is possible to introduce flow rates on the order of 300 uIJmin that are
compatible with 2.1 mm diameter columns into commercially available ESI sources (83).
Reversed phase chromatographic columns as well as normal phase columns have been
used prior to the ESI/MS analysis. Recently, perfirsion chromatography, a rapid
separation technique, has been combined with ESI/MS (84, 85). Separation times on the

order of minutes were reported.

B. Coupling with Capillary Electrophoresis

Capillary Electrophoresis (CE) is attracting attention as a method for rapid high-
resolution separation of very small sample volumes ‘ of complex mixtures. In combining
with the inherent sensitivity and selectivity of mass spectrometry, CE-MS becomes a
potentially powerful bioanalytical technique The combination of CE with a quadrupole
mass spectrometer based on ESI has been described. The coupling of CE with an ESI
source is not straightforward as that of HPLC. The main limitations are the nanoscale
flow rates and the terminating potential. To overcome these limitations, Smith et al. (86)
and Henion et al. (87) developed CE-ESI probes. These authors used sheath liquid flow
in order to introduce a flow of solvent to flush the end of the CE capillary continuously.
This removes the migrating solutes fiom the CE capillary and facilitates the electrospray
process.

The sheath liquid electrode allows the ESI interface to be operated for almost any
high ionic strength buffers that could not otherwise be electrosprayed. Using CE-ESI/MS,
various classes of molecules have been analyzed with good sensitivity and high resolution,
including quaternary phosphonium salts, amines (86), peptides and proteins (88). Small
peptides are easily amenable to CE-ESI/MS analysis with low femtomole or better
detection limits (89).

‘ 30

C. Coupling with Ion Chromatography
The ion chromatography (IC) system can be interfaced with ESI/MS via a
postsuppressor (90). 1C is widely used to separate ionic component in a mixture.
Coupling of IC with ESI/MS gives several advantages including high sensitivity. But the
ionic strength of bufi‘ers used in IC prohibits the formation of ions in ESI. Therefore, an
ion suppresser was introduced alter the column but before the ESI needle to remove
excess ions in the bufl‘er. The micromembrane suppresser selectively removes over 99.9%
of the ion-pair agents required for ion chromatography from the eluent. The resulting
solution consists of analyte, organic modifier, and water, a solution which is compatible
with ESI. This work describes the separation and detection of quaternary ammonium
drugs and tetraalkylammonium compounds of industrial importance. A limit of detection
of 40 pg injected on-column for tetraalkylammonium cations was reported. The
separation, detection, and identification of some alkyl sulfates and sulfonates were also

shown with operation in the negative ion mode (90).

VII. Applications of ESI-MS

Since the introduction of ESI/MS in 1988 as a tool for the ionization of large
biomolecules without significant decomposition, the practice of ESI/MS has grown
impressively. Several review articles appeared recently describing ESI/MS use in
molecular weight determination of large biomolecules (89, 91). ESI/MS provides
accurate molecular weights of biopolymers that cannot be obtained by conventional
methods such as fast atom bombardment (FAB) and thermospray mass spectrometry.
Multiple charging enables high molecular weight determinations. Molecular weights
greater than 197 kDa have been determined successfully by ESI-MS (92). No attempt will
be made here to exhaustively catalog this work. Instead, some major classes of

31
compounds which have been addressed by ESI/MS will be t0uched upon, with the

inclusion of some appropriate references.

A ESI/MS of Small Molecules

Fenn et al. (12, 16) described their early findings of ESI mass spectrometry using
small molecules. Fenn and co-workers tried a wide range of solute species including
alcohols, acids, esters, amines, catechol amines, cholines, nucleotides, amino acids, and
small peptides (91) to evaluate gas phase ion formation fiom liquid phase and detection by
ESI/MS. In every case, the spectra were comprised of peaks for the solute species itself
when it was an ion or for its adducts with anions or cations when it was not an ion. They
demonstrated the ability to obtain the mass spectra from nonvolatile and thermally labile
compounds without significant decomposition or fi'agmentation.

Bruins et al. (93) analyzed monosulfonated azo dyes, steroid glucuronides, and
steroid sulfates in the negative mode. A A detection limit of 10 pg for diethylstilbestrol was
achieved with nebulizer assisted ESI/MS. In recent years, ESI/MS has become a very
popular technique in small polar molecule analysis and still is growing its popularity.

B. Multiple Charging and High Molecular Weight Determination
1. Large Polypeptides and Proteins

A feature of ESI mass spectra of most proteins is that the average charge state
increases in an approximately linear fashion with molecular weight, extending the mass
range of the MS experiment by the addition of compensating charge. The net number of
charged sites in solution appears to be the principal factor affecting the maximum extent of
charging observed in ESI mass spectra. For most proteins (prepared in aqueous solution,
pH< 4) an approximate linear correlation has been observed between the maximum charge

state detected and the number of basic anrino acid residues (e.g., arginine, lysine, histidine)

3.2

plus the NH2 terminus (89). Under typical acidic conditions most, if not all, basic residues
will be protonated in solution. The acidic residues can be deprotonated in solution at pH
greater than ~ 5, and such conditions would be expected to lower the observed positive
ion ESI charge state distribution. Smith et al. (89) compiled a list of proteins successfully
analyzed by ESI/MS, along with the maximum number of charges observed and the
number of positive charges anticipated under acidic conditions. An example of multiple
charging resulting fi'om electrospray ionization is shown in Figure 1-5. An envelope of
multiple charged ions was observed for the myoglobin. g

Negative ion ESI has been demonstrated for a variety of small molecules with
acidic functionalities such as carboxylic, phosphoric, and sulfonic acid groups (89).
However, very few negative ion ESI mass spectra of polypeptides have been reported
(27). One early example was a negative ESI mass spectrum of aqueous solution of bovine
A-chain of insulin (M.W. = 2532) (89) in which 4 cysteine residues are oxidized. Multiple
charging to the 5- (deprotonated) charge was observed. Most polypeptides in neutral and
acidic solutions do not produce multiple charged ions in the negative mode, because acidic
residues such as glutarnic acid and aspartic acid typically have pKa's around 5 or higher in

proteins.

2. Other High Molecular Weight Compounds
Molecules having preformed charges, basic sites or acidic sites have been directly
analyzed by ESI mass spectrometry. Other classes of polymers that have no such groups
have been analyzed after oxidation, reduction or transformation to metal cation
complexes.
Fenn et al. (94) reported the formation of ions in an ESI source from solutions of

poly(ethylene glycol) (PEG) samples with average molecular weights ranging fiom 200 to

33
17,500. Mass analysis of these ions provided evidence that up to at least 23 sodium
cations could be deposited on the larger oligomers. ‘

The feasibility of ESI/MS in polymer analysis has been evaluated by Kallos'et al.
(95). using a polyarnidoamine Starburst polymer. The repeat unit of the polymer is a
polyamidoamine. The molecular weight information was obtained for a mixture without
prior separation.

The fullerenes C60 and C70 were investigated by using ESI/MS (96). C50',
C5002', C7o', and C7002“ ions were detected by reducing C60 and C70 using NaK
amalgam in benzene + dimethoxyethane solvent. Wilson et al. (54) derivatized C50 and
C70 with a reagent to obtain a signal on ESI/MS.

.C. Studies. on Higher Order Structures of Proteins

Information on higher-order structure (secondary, tertiary and quaternary) is of
fundamental biological importance. The native state of proteins is typically folded into
well-defined three-dimensional structures by relatively weak intramolecular forces (e. g.,
hydrogen bonding). Protein function in biological systems generally involves interactions
of specific structural conformations essential for activity (97).

In their native state, globular proteins are tightly folded, compact structures. They
can be denatured and caused to unfold by subjecting them to high temperatures, extremes
of pH, detergents, and solutions containing high concentrations of compounds such as
urea, guanidium chloride, and organic solvents (97). ESI/MS has been used to study the
higher order structural changes recently.

Chowdhury et al. (98) investigated the structural conformational changes of bovine
cytochrome c at various pH values (pH=2.6-5 .2). These authors found three different
charge state distribution patterns at difi‘erent pH values. These results agreed with the

earlier reported conformations.

34

L00 et al. found that the addition of methanol denaturant in excess of 40% v/v was
required to eliminate the lower charge state (globular or native state) of Ubiquitin peptide
(M.W.=8,564.8). The native globular form dominates (7+ and 8+) for Ubiquitin in a
completely aqueous solution with a small contribution from the denatured helix structure
(10+ to 13+), ascribed to mixing of the analyte solution with the acetonitrile sheath (99). .

Le Blanc et al. (100) studied the efi‘ects of heat on the electrospray spectra of a
number of globular proteins, observing a dramatic increase with increasing temperature of
the charge states of ions produced from equine cytochrome c and chicken egg lysozyrne as
well as a strong increase of the ion intensity fi'om cytochrome c.

Hydrogen exchange with deuterium followed by ESI/MS appears to be a sensitive
and useful method for probing secondary structure in peptides and proteins. Katta et al.
(101) demonstrated the possibility of using hydrogen exchange ESI/MS for probing the
conformational changes of proteins in solution. The ESI/MS measurements provide
information about the rate of exchange and the extent of exchange (102, 103).

Electrospray mass spectrometry is a very mild technique for producing ions
directly from solution at atmospheric pressure. ESI/MS can be used to observe non-
covalently bound adducts. Henion et al. (104, 105) described non-covalent receptor-
ligand complexes using ion-spray MS. In several reports, non-covalent enzyme-substrate

and enzyme-product complexes have been observed.

10.

ll.

12.

l3.

14.

References
Aebersold, R.; Bures, E. J .; Goghari, M. H.; Shusan, B.; Covey, T. C. Protien
Science, 1992, I, 494

Conboy, J. J .; Henion, J. D.; Martin, M. W.; Zweigenbaum, J. A. Anal. Chem,
1990, 62, 800

Bose, G. M. Researches sur 1e cause et sur la veritable theorie de I 'electricite,
Wittenberg, 1745

Zeleny, J. J. Phys. Rev. 1917, 10, 1-6

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

Mack, L. L.; Kralik, P.; Rheude, A; Dole, M. J. J. Chem. Phys. 1970, 52, 4977
Chapman, S. Phys. Rev. 1937, 52, 184
Clegg, G. A; Dole, M. Biopolmers 1971, 10, 821

Gieniec, J.; Mack, L. L.; Nakarnae, K.; Gupta, C.; Kumar, V.; Dole, M. Biomed.
Mass. Spectrom. 1984, 11, 259

Aleksandrov, M. L.; Gall, L. N.; Krasnov, V. N.; Nikolaev, V. I.; Pavlenko, V. A;
Shkurov, V. A. Dokl. Akad Nauk. SSSR 1984, 277, 379

Aleksandrov, M. L.; Gall, L. N.; Krasnov, V. N.; Nikolaev, V. 1.; Pavlenko, V. A;
Shkurov, V. A; Bararn, G. 1.; Gracher, M. A; Knorre, V. D.; Kusner, Y. S.
Bioorg. Kim. 1984, 10, 710

Yarnashita, M.; Fenn, J. B. J. Phys. Chem. 1984, 88, 4671

Aleksandrov, M. L.; Besuklandnikov, P. V.; Grachev, M. A; Elyakova, L. A;
Zyyaginsteva, T. N.; Kondratsev, V. M.; Kusner, Yu. 8.; Mirgorodskaya, O. A;
Fridlyansky, G. V. Bioorg. Kim. 1986, 12, 1689

Aleksandrov, M. L.; Barama, G. I.; Gall, L. N.; Krasnov, N. V.; Kusner, Yu. S.;

Mirgorodskaya, O. A; Nikolaev, V. 1.; Shkurov, V. A; Bioorg. Kim 1985, 11,
700

35

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

36
Mirgorodskaya, O. A; Nikolaev, V. 1.; Shkurov, V. A; Bioorg. Kim 1985, 11,
705

Whitehouse, C. M.; Dreyer, R. N.; Yamashita, M.; Fenn, J. B. Anal. Chem, 1985,
57, 675

Wong, S. F.; Meng, C. K.; Fenn. J. B. J. Phys. Chem. 1988, 92, 546
Meng. C. K.; Mann. M.; Fenn. J. B. Z. Phys. D. 1988, 10, 361

L00, J. A; Udseth, H. R.; Smith, R D.; Blamed Environ. Mass Spectrom. 1988,
I 7, 41 1

L00, J. A; Udseth, H. R.; Smith, R. D.; Rapid Commun. Mass Spectrom 1988, 2,
207

Covey, T. R.; Boner, R. F.; Shusan, B. I.; Henion, J. D. Rapid Commun. Mass
Spectrom 1988, 2, 249

Karas, M.; Killenkamp, F. Anal. Chem, 1988, 60, 2299

Smith, R. D.; Loo, J. A; Loo, R. R. 0.; Busman, M; Udseth, H. R Mass Spec.
Rev. 1991, 10, 359

Meng, C-K.; McEwen, C. N.; Larsen, B. S. Rapid Commun. Mass Spectrom.
1990, 4, 147

Beu, S. C.; Senko, M. W.; Quinn, J. P.; Wampler III, F. M.; Mclarfi‘erty, F. M. J.
Am. Soc. Mass Spetrom. 1993, 4, 557-565

Loeb, L. B.; Kip, A. F.; Hudson, G. G.; Bennet, W. H. Phys. Rev. 1941, 60, 714
Pfeifer, R. J.; Hendricks, C. D. AIAA J. 1968, 6, 496

Taylor, G. 1. Proc. R. Soc. London A 1964, A280, 383

Ikonomou, M. G.; Blades, A. T.; Kebarle, P. Anal. Chem. 1990, 62, 957

Hayati, 1.; Bailey, A. I.; Tadros, T. F. J. Colloid Interface Sci. 1987, 117, 205 and
222

Taflin, D. C.; Ward, T. L.; Davis, E. J. Langmuir 1989, 5, 376

Smith, D. P. H. IEEE Trans. Ind Appl. 1986, 1A-22, 527

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

47.

48.

49.

37

Ikonomou, M. G.; Blades, A. T.; Kebarle, P. J. Am. Soc. Mass Spectrom. 1991, 2,
497

Wamper, F. W.; Blades, A. T.; Kebarle, P. J. Am. Soc. Mass Spectrom. 1993, 4,
289 '

Kebarle, P.; Tang, L. Anal. Chem. 1993, 65, 972A
Lee, E. D.; Henion, J. D. Rapid Commun. Mass Spectrom. 1992, 6, 727
Mirza, U. A; Cohen, S. L.; Chait, B. T. Anal. Chem. 1993, 65, 1-6

Buchanan, M. V. Proceedings 41st ASMS conference on Mass Spectrometry and
Allied Topics, San Francisco, CA, May 31-June 4, 1993, 764

Gomez, A; Tang, K. Proc. 5th International Conference on Liquid Atomization
and Spray Systems I (CLASS-91); Smerjians, H. G., Ed: NIST: Gaithersburg,
MD; Special Publication 813

Davis, E. J. ISA Trans. 1987, 26, 1

Willoughby, R.; Sheehan, E.; Jarrell, A; Masecic, T.; Pedder, R.; Penn, S.
proceedings 41 st ASMS conference on Mass Spectrometry and Allied Topics, San
Francisco, CA, May 31-June 4, 1993

Tang, L.; Kebarle, P. Anal. Chem. 1993, 65, 3654-3668

Davies, C. N. In Fundamentals of Aerosol Science; Shaw, D. T., Ed; John Wiley
and Sons: New York, 1978, 154

Rollgen, F. W.; Bramer-Wegner, E.; Buttering, L. J. Phys. Colloq. 1984, 45,
supplement 12, C9-297

Iribame, J. V.; Thomson, B. A. J. Chem. Phys. 1976, 64, 2287
Thomson, B. A; Iribame, J. V. J. Chem. Phys. 1979, 71, 4451

Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science
1989, 246, 64

Ikonomou, M. 6.; Blades, A. T.; Kebarle, P. Anal. Chem. 1991, 63, 1989

Sakari, M.; Yergey, A. L.; Siu, K. W. M., LeBlanc, J. C. Y.; Guevremont, R.;
Berrnan, R. 8. Anal. Sci. 1991, 7, 199

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

65.

67.

38
Rafaelli, A; Bruins, A. P. Rapid Common. Mass Spectrom. 1991, 5, 269
Hiraoka, K. Rapid Common. Mass Spectrom. 1992, 6, 463

Siu, K. W. M.; Guevremont, R.; LeBlanc, J. C. Y.; O'Brien, R. T.; Berrnan, S. S.
Org. Mass Spectrom, 1993, 28, 579

Tang, L.; Kebarle, P. Anal. Chem. 1991, 63, 2709

Wilson, S. R.; Wu, Y. J. Am. Soc. Mass Spectrom. 1993, 4, 596-603

Pfeifer, R. J .; Hendricks, C. D. AIAA J. 1968, 6, 496

Blades, A. T.; Ikonomou, M. G.; Kebarle, P. Anal. Chem. 1991, 63, 2109
Morand, K.; Talbo, G.; Mann, M. Rapid Common. Mass Spectrom. 1993, 7, 738-
1411111, M. H.; Lewis, I. A. S. Rapid Common. Mass Spectrom. 1989, 3, 255

Meng, C.-K.; McEwen, C. N.; Larsen, B. S. Rapid Common. Mass Spectrom.
1990, 4, 369-372 .

Allen, M. H.; Lewis, I. A. S. Rapid Common. Mass Spectrom. 1989, 3, 255

Meng, C.-K.; McEwen, C. N.; Larsen, B. S. Rapid Common. Mass Spectrom.
1990, 4, 147-150

Wada, Y.; Tarnura, J.; Musselrnan, B. D.; Kassel, D. B.; Sakurai, T.; Matsuo, T.
Rapid Common. Mass Spectrom. 1992, 6, 9-13

McEwen, C. N.; Larsen, B. S. Rapid Common. Mass Spectrom. 1992, 6, 173-178

Chapman, J. R.; Gallagher, R. T.; Barton, E. C.; Curtis, J. M.; Derrick, P. J. Org.
Mass Spectrom. 1992, 27, 195-203

Meng, C. K.; McEwen, C. N.; Larsen, B. S. Rapid Common. Mass Spectrom.
1990, 4, 15-1-155

Nourse, B. D.; Coks, R. G. Anal. Chim. Acta 1990, 228, 1

Stafi‘ord, G. C.; Kelley, P. E.; Syka, J. E. P.; Reynolds, W. E. .1111. J. Mass
Spectrom. Ion Processes, 1984, 60, 85

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

39
Kelley, P. E.; Stafi‘ord, G. C.; Fies, W. J.; Syka, J. E. P.; McIver,Hunter, R L.;
Todd, J. F. J.; Bexon, J. J. ASMS San Antonio, TX, 1984, pp 505

Louris, J. N.; Cooks, R. G.; Syka, J. E. R; Kelly, P. E.; Stafi‘ord Jr, G. C. Todd, J.
F. J. Anal Chem, 1987, 59, 1677

Louris, J. N.; Brodbelt-Lustig, J. S.; Cooks, R G.; Glish, G. L.; VanBerkel, G. L.;
McLuckey, S. A. Int. J. Mass Spectrom. Ion Proc. 1990, 96, 117

Kraiser, R. E.; Louris, J. N.; Amy, J. W.; Cooks, R. G. Rapid Common. Mass
Spectrom. 1989, 3, 225

Cox, K. A; Williams, J. D.; Cooks, R. 6.; Kaiser Jr, R E. Biol. Mass Spectrom.
1992, 21, 226-241

Van Berkel, G. J .; Glish, G. L.; McLuckey, S. A. Anal. Chem, 1990, 62, 1284
Wilkins, C. L.; Gross, M. L. Anal. Chem. 1981, 53, 1661A

Ben, S. C.; Senko, M. W.; Quinn, J. P.; Wampler III, F. M.; Mclarfi‘erty, F. M. J.
Am. Soc. Mass Spetrom. 1993, 4, 557-565

Winger, B. E.; Hofstadler, S. A; Bruce, J. E.; Udseth, H. R.; Smith, R. D. J. Am.
Soc. Mass Spetrom. 1993, 4, 566-577

Charbonnier, F.; Rolando, C.; Saru, F. Rapid Common. Mass Spectrom. 1993, 7,
707-710

Boyle, J. G.; Whitehouse, C. M.; Fenn, J. B. Rapid Common. Mass Spectrom.
1991, 5, 400-405

Chien, B. M.; Micheal, S. M.; Lubman, D. M. Int. J. Mass Spectrom. Ion
Processes 1994, 131, 149

Smith, R D.; Loo, J. A; Edmonds, C. G.; Barinaga, C. J.; Udseth, H. R Anal.
Chem. 1990, 62, 882

Lane, 8. J.; Blinded, K. A; Taylor, N. L.; Watkins, P. J. F., Harrison, M. E.
Rapid Common. Mass Spectrom. 1993, 7, 953

Carr, S. A; Hemling, M. E.; Bean, M. F.; Roberts, G. D. Anal. Chem. 1991, 63,
2802

TSQ 7000 API manual, Finnigan Mat, San Jose, CA, 1993

84.

85.

86.

87.

88.

89.

91.

92.

93.

94.

95.

96.

97.

98.

100.

40
Kessal, D. B.; Shusan, B.; Sakuma, T.; Salzmann, J-P. Anal. Chem. 1994, 66, 236

Lin, H-U.; Voyksner, R. D. Rapid Common. Mass Spectrom. 1994, 8, 333
Smith, R. D.; Barinaga, C. J.; Udseth, H. R Anal. Chem. 1988, 60, 1948
Lee, E. D.; Muck, W.; Henion, J. D.; Covey, T. R. J. Chromtogr. 1988, 458, 313

Smith, R. D.;Udseth, H. R; Barinaga, C. J.; Edmonds, J. Chromatogr. 1991, 559,
197-208

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

Conboy, J. J.; Henion, J. D.; Martin, M. W.; Zweigenbaunr, J. A. Anal. Chem.
1990, 62, 800-807

Fenn, J. B. ,Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Mass
Spectrometry Reviews, 1990, 9, 37- 70

Feng, R.; Bouthiller, F.; Konishi, Y.; Cygler, M. Proceedings 39th ASMS
Conference of Mass Spectrometry and Allied Topics, Nashville, TN, May 29-June
3, 1991

Bruins, A. P.; Covey, '1". 11.; Henion, J. D. Anal. Chem. 1937, 59, 2642

Wong, S. F.; Meng, C. K.; Fenn, J. B. J. Phys. Chem. 1988, 92, 546-550

Kallos, G. J.; Tomalia, D. A; Hedstrand, D. M.; Lewis, S.; Zhou, J. Rapid
Common. Mass Spectrom. 1991, 5, 383-386

Hiraoka, K.; Kudaka, I.; Fujimaki, S.; Shinohara, H. Rapid Common. Mass
Spectrom. 1992, 6, 254-256

Creighton, T. E.; Proteins, structures and Molecular Principles, Freeman, W. H.
and Co., New York 1984

Katta, V.; Chait, B. T. J. Am. Chem. Soc. 1991, 113, 8534-8535

Loo, J. A; Loo, R. R. O.; Udseth, H. R.; Edmonds, C. G.; Smith, R. D. Rapid
Commun. Mass Spectrom. 1991, 5, 101

Le Blane, J. C. Y. ,Beuchemin, D.; Siu, K. W.; Guevremont, R. ,Berrnan, S. S.
Org.MassSpectrom. 1991,26, 831

101.

102.

103.

104.

105.

41
Katta, V.; Chait, B. T. Rapid Common. Mass Spectrom. 1991, 5, 214-217

Stevenson, C. L.; Andregg, R. J.; Borchardt, R. T. J. Am. Soc. Mass Spectrom.
l993,4,646

Smith, J. B.; Thevenon—Emeric, G.; Smith, D. L.; Green, B. Anal. Biochem, 1993,
193, 118

Ganem, 3.; Li, Y-T.; Henion, J. D. J. Am. Chem. Soc. 1991, 113, 6294-6296

Ganem, 3.; Li, Y-T.; Henion, J. D. J. Am. Chem. Soc. 1991, 113, 7818-7819

Chapter 2
DEVELOPMENT OF INSTRUMENTATION FOR ESI/MS

I. Introduction

Electrospray mass spectrometry (ESI/MS) has become a very powerful technique
in modern mass spectrometry in a short period of time. The analysis of thermally labile
and polar molecules using mass spectrometry became routine afier interfacing ESI/MS
with chromatographic systems. ESI/MS provides the ability to obtain accurate molecular
weight information of polypeptides and other polymer materials which are refractory to
conventional ionization techniques by forming a multiple charged envelope of ions with
high sensitivity. The sensitivity for polypeptides (1) and small molecules having pre-
formed charges (2) has been demonstrated at the femtomole level or better.

Fenn et al. (3, 4) described their first electrospray mass spectrometer with
impressive results for the analysis of non-volatile molecules in 1984. They attracted the
attention of mass spectrometrists for their re-invention of ESI/MS.

The essential features of the Yamashita and Penn source are a hypodermic needle,
a cylindrical electrode around the needle, a nozzle, and a skimmer. Liquid sample was
introduced through a stainless steel hypodermic needle and maintained at 3-10 kV relative
to ground. The distance from the end of the needle to an end plate containing the nozzle
was in the range of 20-30mm. Surrounding this region was a cylindrical electrode 30 mm
in diameter and maintained at 500-600V relative to ground. Dry nitrogen typically at 800
torr was passed through the nozzle region at a velocity of several cm/s in order to sweep
out solvent vapor from the evaporating droplets. The region downstream of the nozzle
was maintained at a few mtorr or less by an oil difliision pump with a speed of a 1000Us.

A portion of the resulting free jet was passed through a skimmer with a 4-mm aperture

42

43
situated after the nozzle into a vacuum chamber containing the quadrupole mass filter and
pumped by a pair of bafiled 6-in oil diflirsion pumps with a combined speed of the order of
3000113.

Afier the introduction of the original ESI interface, several research groups have
duplicated the interface and the results very quickly (5, 6, 7, 8). In all the ESI sources,
there are several common features. The ions are formed by forcing liquid through a small
stainless steel capillary tube held at high potential. At the tip of the capillary tube, a fine
mist of charged droplets is formed. The highly charged droplets are desolvated and ions
are transported to the mass analyzer through a set of ion focusing lenses or RF -only
quadrupoles. There are several different methods reported for declustering the solvated
ions during the ESI process. Mainly three techniques were employed to’ achieve
declustering by using a heated counter-current flow of gas through the nozzle-skimmer
region (3, 4, 5), a heated metal capillary to control desolvation and transport ions (7), or a
heated ion source block (8). The flow of ions and gas going into the mass spectrometer
was controlled by using a skimmer and a nozzle (4, 5, 6), or a small diameter capillary and
a skimmer (7). These systems use high pumping capacity pumps to achieve reasonable
low pressures (10'5 torr or low) inside the mass analyzers.

In order to take the advantage of the sensitivity of ESI/MS, an electrospray
interface was designed, built in-house, and coupled with a single quadrupole mass
spectrometer (HP 5985). An interface between the mass spectrometer and a 486, 33-
MHz personal computer has been developed. The interface was assessed for the
operational use with model compounds. The intention of this chapter is to introduce the

considerations for source design, fabrication, and performance of the newly built source.

44
11. HP 5985 Single Quadrupole Mass Spectrometer
This project was started with the HP 5985 single quadrupole mass spectrometer
and its components. A thermospray unit, another LC/MS interface fi'om Vestec
Corporation (Houston) was used with this mass spectrometer. In this section, the
operation principles of a quadrupole mass spectrometer and the vacuum system of the HP
5985 mass spectrometer will be briefly discussed.

A Operating Principles of Quadrupole Mass Spectrometer

The HP 5985 mass spectrometer is a single quadrupole instrument. The
quadrupole mass spectrometer employs a combination of direct current (DC) and radio-
fiequency (RF) potentials to separate ions passing through the quadrupole rods according
to the m/z values. Ideal performance is obtained with rods having a hyperbolic cross
section (9-11), however, cylindrical rods are most often used and provide satisfactory
results. Opposite rods are connected together electrically and to RF and DC voltage
generators. As the ions travel through the central longitudinal region between the rods,
they are influenced by the combined DC and oscillating RF fields. As the ions drift toward
the detector along the Z-axis, their motion in the X-Y plane is described by the Mathieu

equation (1 1), that defines ax and qx parameters as follows:

ax = -ay = 4ZeU/mro2ro2 eqn 2-1
qx = -qy = ZzeV/mcozro2 eqn 2-2

where r0 is half the distance between opposite quadrupole rods, 0) is the alternating
frequency (RF), U is the magnitude of the DC field, and V is the amplitude of the RF-field.
Figure 2-1 represents the a-q stability diagram; the tip of the diagram represents
coordinates of those ions which will have stable trajectories through the quadrupole field.

45
The slope of the m/z scan line is defined by the ratio 2U/V. Altering the slope Of the mass
scan line changes the mass-resolution Of the ions. In a typical quadrupole mass
spectrometric analysis, voltages are selected tO achieve unit mass resolution. If the DC/RF
ratio were adjusted to raise the Operating line even higher, the resolving power Of the
instrument would increase, but the sensitivity would decrease, because fewer ions would
follow a stable trajectory to reach the detector.

In the "RF-only" mode, there is no DC field, U = 0, the Operating line lies
horizontally along the abscissa, indicating that ions having Operating point between q = 0
and q = 0.908 could be transmitted through the RF-Only quadrupole filter. However, as
suggested by Miller and Denton (12), the transmission efficiency for ions Of all m/z values
is not 100% in an RF-Only quadrupole. Not only is the transmission Of ions through a
quadrupole influenced by the stability Of the ions as defined by the Mathieu stability
diagram, but by (i) the acceptance aperture Of the quadrupole and (ii) the spatial focusing
conditions that arise due to the trajectory Of the ions through the quadrupole (12).
Therefore mass discrimination efi‘ects are apparent, which implies that RF-0nly
quadrupoles are not strictly total ion transmission chambers, that is, ions Of different m/z
ratios are transmitted through RF-Only quadrupole rod assemblies with different degrees
Of efiiciency.

B. Vacuum System Of the HP 5985 Mass Spectrometer

The vacuum system of the HP 5985 mass spectrometer uses conventional
multistage Oil difl'usion pumps which are pumped by mechanical pumps. The difi‘usion
pumps and their associated baffles are water cooled. A diffirsion pump with 600L/sec
speed for the EI-CI ion source provides adequate pumping speed to handle the large gas

flow rates used in the CI mode. A smaller 150 Usec unit takes care of pumping the mass

0.2 '

0.1

46

Mass ...—v
Scan Line

 

 

Figure 2-1. Stability diagram Of (a, q) space showing regions (patchwork area) that

corresponds to mathematically stable ion trajectories in the quadrupole mass spectrometer.

47
filter and the detector portion where gas flow rates are much smaller. The pressure inside
the source housing and the analyzer assembly is monitored using the ion gauges at the
base Of the chambers. In electrospray operation, the vacuum system Of the mass

spectrometer was used without modifications.

III. Design Consideration for the ESI Interface

Electrospray ionization is an atmospheric pressure ionization (API) technique.
Two problems are intimately related in API mass spectrometry: the transport Of ions from
an atmospheric pressure region into the vacuum system Of the mass spectrometer, and the
strong cooling Of a mixture Of gas and ions when expanding into a vacuum. The resulting
condensation Of polar neutrals (notably water vapor) on analyte ions produces cluster ions
having a mass far beyond the range Of most mass analyzers (8). The problem Of
declustering was attacked by several methods such as heating the ion source region (8),
using heated counter gas flow (3,5,6), or passing through a heated capillary tube (7). In
this design, the ions will be transported into the mass spectrometer through a heated
capillary tube (100°C) to avoid declustering upon expansion into vacuum.

A Vacuum System and Sensitivity

The goal Of designing a vacuum system for atmospheric pressure ionization mass
spectrometry (API/MS) is to transport ions, produced in atmospheric pressure into the
mass analyzer and to introduce as many ions, but as little gas as possible. Ideally, ions
should be separated fiom gas. In a simple approximation, the sensitivities Of different API

mass spectrometers can be judged fiom the throughput Of gas in the relevant vacuum

stage (13).

48

API/MS instruments can be divided into single-stage vacuum system and multiple-
stage systems. In a single-stage vacuum system, the flow Of ions into the mass analyzer is
proportional to the gas flow through the sampling orifice. The throughput Of gas into the
analyzer is given by the pumping speed multiplied by the pressure in the vacuum system
(13).

The purpose Of the multi-stage vacuum system is to use a larger orifice, which is
not easily blocked and can take more gas and ions into the mass spectrometer. In a typical
differentially pumped system, the throughput of gas into the ion Optics stage is most
important. The bigger the pump and the higher the pressure tolerated by the ion Optics,
the higher the throughput Of gas plus ions. In the ion Optics region, ions are focused and
transmitted towards the mass analyzer, while gas is pumped away, thereby increasing the
ion-tO-gas ratio. The region between the nozzle and the skimmer is called the molecular-
bearn stage (Figure 2-2). The molecular-beam stage is supposed to transfer a portion Of
gas plus ion beam without changing the ion-tO-gas ratio. Therefore, the molecular-beam
stage does not help to increase the sensitivity by an increase Of ion-tO-gas ratio. However,
a better shaped beam is presented to the mass analyzer, which will increase the
transmission of ions through the analyzer. A significant advantage is that a bigger orifice
makes the instrument more rugged.

The prediction of sensitivity based on vacuum considerations alone is a first
approximation. The effects Of the quality Of the molecular beam components, ion Optics,

and mass analyzer cannot be included in a simple generalization.
B. Transport of Gas and lens into Vacuum

A simple scheme Of the expansion Of a gas into a low-pressure region is given in

Figure 2-2. The principles Of the generation Of molecular beams (14,15) are explained in

49

qualitative terms. Inside the nozzle Opening and behind the nozzle, the gas molecules
acquire a high velocity, and gas molecules with entrained ions follow straight streamlines
originating approximately in the nozzle, the highest intensity of the gas flow being on the
axis Of the nozzle. Far away from the nozzle, the gas is pumped away, and gas molecules
move at random. In the transition between directed motion and random motion, called
shock waves, ions and gas molecules undergo many collisions, with scattering Of the beam
of ions and molecules as the result. The region inside the barrel shock wave, and between
the nozzle and the mach disk, is called the silent zone, where ions and gas move at equal
speed in the same direction and undergo strong and rapid cooling.

The location Of the mach disk is important for the design Of an API mass

spectrometer. The distance fi'om the nozzle is given by eqn 2-3.

1M = 0.67 D0 (Po/P1)“2 eqn 2-3

where Do is the orifice or nozzle diameter
P0 is the upstream pressure (atmospheric pressure)

P1 is the downstream pressure in the vacuum chamber

xM is firlly determined by the pumping speed available in the volume near the nozzle
Opening (16). A small pump or narrow pumping line or other Obstruction will result in a
small x”.

Often the flow through the nozzle is tOO large for the vacuum pumps Of the mass
analyzer. The central portion Of the beam can be sampled into the next vacuum stage by
means Of a skimmer (14-17) as in Figure 2-2b and 2-2c.

In Figure 2-2b, the core of the beam is sampled from the silent zone, and the ions

plus gas continue their movement in straight lines. Collection Of ions fiom the gas flow

50
through the skimmer should be efiicient, because Of the directional effect Of the molecular
beam. However, since the gas flowing in the molecular beam has strongly cooled upon
expansion, ions will cluster with water molecules, if present.

In Figure 2-2c, a sample is taken from behind the mach disk. Ions and gas have
undergone extensive scattering, and the extraction and focusing Of ions is more difficult.
On the other hand, the gas temperature has risen through collisions in the mach disk, with
breaking Of hydrogen bonds in cluster ions.

The circumference Of the skimmer Opening should be very sharp and fi'ee fiom
burrs. The full angle Of the skimmer cone is usually about 60° in a molecular beam
apparatus (14,15); a full angle up to 110° appears to be usable without adverse effects. A
larger angle and/or imperfections at the skimmer orifice will result in disturbance Of the
molecular beam if sampling is attempted from the silent zone. A lower transmission of
ions has been reported in such a case (18). Even with nearly perfect Skimmers,
disturbance Of the molecular beam is unavoidable near the edges. As a result, the effective

skimmer Opening in molecular beam experiments is always smaller than the physical

Opening.
IV. Fabrication of Source and Flange

The HP 5985 mass spectrometer was modified to accomadate the ESI source. A
thermospray interface was earlier fitted in the mass spectrometer. The instrumentation for
thermospray mass spectrometry will be discussed before the details Of ESI source

fabrications and modifications.

5]

Vacuum
1 atm

 

# silent zone

Ions, N2 <—:—> (a)
<—— mach disk

Nozzle
barrel shock

-3
1 torr 1 X 10 torr
1 atm

Ions, N2 /' . . (b)

\ Ions, N2

NOZZIC skimmer
. -3
1 atm 1 X 10 ton-
/ J
Ions, N2 T» T. Ions, N2
(C)
Nozzle
skimmer
<———>
XM

Figure 2—2. Expansion Of gas into vacuum; (a). simplified scheme with shockwaves and
silence zone; (b). generation Of a beam Of gas and ions by sampling from a silent zone, the

skimmer penetrates the mach disk; (c). sampling Of gas and ions with a skimmer located
beyond the mach disk.

52
A. Instrumentation for Thermospray Mass Spectrometry

Thermospray (TS) ionization is a mild ionization technique that uses heat to
vaporize the liquid stream. A partial or complete vaporization Of the liquid stream is
carried out passing it through a heated capillary tube. At the tip Of the capillary tube a
supersonic jet of vapor is created, normally containing a mist Of fine particles and solution
droplets. The droplets are desolvated by the nearby hot surfaces and excess solvent vapor
is removed by an additional pumping stage. In the "filament Ofi" mode Of Operation,
normally the solutes are mixed with a volatile bufi‘er such as ammonium acetate. It is
assumed that a gas phase reaction occurs between analyte and N114+ molecules to transfer
protons tO analyte molecules or to form adduct ions with NH4+ (19,20). In the "filament
on" mode Of Operation, a discharge electrode is used tO ionize analyte molecules. The ions
are sampled through a skimmer and mass analyzed by a mass analyzer. The solvated ions
are desolvated while passing through the source block and a heated lens system.

The major components Of the thermospray source are: a heated source block, a
sampling cone, a vaporizer probe, a discharge electrode, a pumping line to remove excess
vapor, and thermocouples to monitor the temperature of the vaporizer probe, the source
block, and vapor. Figure 2-3 shows the Vestec thermospray source (for HP 5985 or
newer models). The TS probe produces a supersonic jet Of vapor containing a mist of fine
droplets or particles. As the droplets travel at high velocity through the heated ion source,
they continue to vaporize due to rapid heat input fi'om the surrounding hot vapor. The ion
entrance aperture to the mass analyzer is lOcated at the apex Of a cone perpendicular to the
vapor jet from the thermospray vaporizer and near the center. In a fi'eely expanding
supersonic jet, heavier solutes tend to remain close to the jet axis. The purpose Of the
sampling cone is to extract these heavy ions selectively in order to increase sensitivity.

The desolvation Of ions takes place at the heated TS source as well as at the heated lens

53

region. The solvent vapor was pumped down by the l/2-in pumping line connected to a
cold trap and a 300 um mechanical roughing pump. In this Vestec thermospray
interface, the vaporizer probe is placed right angles to the axis Of the mass analyzer. The
ions emerging from the skimmer are focused by a set of lenses and analyzed by the
quadrupole mass analyzer. The ions are detected by a continuous dynode electron
multiplier.

HP 5985 single quadrupole mass spectrometer with Vestec thermospray interface
having 1/2-in pumping line with the existing vacuum system were not able to maintain the
vacuum inside the mass spectrometer when the HPLC effluent was introduced into the
mass spectrometer. Therefore, this interface at the MSU Mass Spectrometry Facility was
abandoned without use for long time.

B. SIMION Ion Simulation

The SIMION simulation program (21) was used to study the trajectories Of ions
produced by electrospray. The SMON program is an electrostatic lens analysis and
design program that allow estimates Of the trajectories of ions in an electric field. When
the ESI probe and the ion transport capillary tube is perpendicular to the axis of the
skimmer, the ion trajectories did not divert towards the skimmer, when the ESI source
was at ground potential and the skimmer at 15V. The potential on the skimmer and the
pressure gradient was not enough tO draw ions towards the skimmer. A repeller at high
voltage (several kV) has to be introduced in front of the skimmer to divert ions towards
the skimmer.

When the probe was placed parallel to the axis Of the skimmer, the trajectories Of
the ion beam enter the mass analyzer through the skimmer. The simulation studies

indicated that the parallel geometry would be the best for the introduction Of ions into the

54

Ions to MS

Filament or

Discharge electrode
Sampling Cone

WWW/

   
    

 

J
\/

 

 

 

 

 

 

 

Solvent Vapor tO Trap and
Rotary pump

Figure 2-3. The Vestec thermospray interface for the HP 5985 mass spectrometer.

55
mass analyzer. Therefore, the parallel geometry was adapted. Figure 2-4 shows the plots
fiom the SIMION ion simulation.

C. Source Modification

As mentioned earlier, the most difiicult problem in API/MS technique is to achieve
high degree Of vacuum inside the mass spectrometer. By decreasing the pressure before
the skimmer (inside the E81 source), pressure inside the second vacuum stage can be
decreased. This can be accomplished by increasing the conductance Of the vacuum line
connected to the E81 source chamber.

Throughput is the quantity Of gas (the volume Of gas at a known pressure) that
passes a plane in a known time ant (PV) = Q (22). The flow Of gas in a channel is
dependent on the pressure drop across the tube as well as the geometry Of the channel.

For a long round pipe the throughput (in viscous flow) is given by the Hagen-

 

Poiseuille equation, eqn 2-4 (22).
Q: m4 (Pl+P2)(Pl-P2) eqn 24
1281]! 2

where dis the diameter Of the tube, t is the length Of the tube, a is the viscosity of the gas.

Division Of the throughput by the pressure drop across a channel held at constant
temperature yields a property known as the intrinsic conductance C of the channel as
given by equation 2-5.

Q

C = n 2-5
P2-P1 “1

 

56
\ Electric Field Lines
Mass Analyzer

(3)

Skimmer

ESI Needle (3000V)

Tr ectories
loo 3,] / Ion Tra/nsport Capillary (100W\\\

Skimmer ——>

C3

5: (h)

f
’ Ion Transport. Capillary (100V)

 

 

ll

 

 

 

we

ESI needle (sooev? @

Figure 2-4. The plots fiom the SIMION ion simulation. (a). Heated capillary is right
angle to the axis of the mass analyzer. (b). Heated capillary is parallel to the mass analyzer.
In both cases, source block is at ground potential.

Electric Field Lines

57
where P1 and P2 refer tO the pressures measured in large volumes connected tO each end
Of the channel or component. The units Of throughput Q is Pa-m3/s (or Pa-L/s) and that
Of conductance or pumping speed C is m3/s (or Us).

The conductance Of a long round pipe (for air at 0°C) is given by the equation 2.6.

+P
C (US) = 1.38 x106 9; (112—2) eqn 2-6

Therefore the conductance Of a pipe is proportional to the (diameter)4 in the
viscous flow regime (22). By increasing the diameter by a factor Of 2 causes tO increase
the conductance by a factor Of 16.

The conductance Of the existing pumping line can be increased by increasing the
diameter Of the vacuum line. The maximum diameter Of the vacuum line was determined
by the space available on the vacuum flange and that Of the source chamber. A l-in
pumping line was attached to the ESI source through the vacuum flange to increase the
conductance Of pumping line. The increased conductance of the pumping line can remove
excess solvent vapor entering the ESI source. A source was fabricated to accommodate
the l-in vacuum line. Figure 2-5 shows the new ESI source.

A source block was (Figure 2-5) designed to control the flux Of ions and gas
entering into the mass spectrometer. This design facilitates three stages Of differential
pumping: molecular beam, ion optics, and mass analyzer. A 0.5 mm-ID skimmer was used
to sample a fiaction of the flow Of ions and gas emerging from the ion transport capillary
before entering the ion Optics region. The skimmer is electrically isolated from the source
block using a rubber O-ring placed in between the skimmer and the source block. The
skimmer was connected to a voltage supply (0-50V). The place Of the skimmer was

designed so that it can be placed co-axially with the mass analyzer. The ion transport

58
capillary tube can be placed cO-axially with the skimmer and a fiaction of the emerging ion
beam can be sampled and focused into the quadrupole mass analyzer. A l-inch pumping
line was connected to the source tO remove excess solvent vapor. The pumping line was
electrically isolated fiom the source using a Vespel spacer. The ion transport capillary
was electrically isolated from the source using a Teflon spacer. The spacers provide
vacuum seals as well as electrical insulation for the individual components. The source

block was mounted on a platform inside the source housing.
D. Vacuum Flange Modification

The vacuum flange also had to be modified to accommodate the vacuum line and
the ion transport capillary. A l-in Swagelok® welding fitting was installed in the vacuum
line. Another Swagelok® welding fitting with the inner diameter Of l/4-in was installed to
insert the ion transport capillary tube cO-axially with the skimmer. The capillary tube was
electrically isolated from the flange by using a Vespcl® ferrule (ID l/l6-in enlarged to
3/16-in, OD l/4-in). The original electrical feedthroughs were used without

modifications.
E. Design and Fabrication Of the Ion Transport Capillary

A vital part Of the atmospheric pressure mass spectrometry is to transfer ions
produced at atmospheric pressure into the high vacuum region Of the mass spectrometer.
In ESI/MS, this is achieved by restricting the flow of ions and gas into the mass
spectrometer using small diameter skimmers, nozzles, or capillary tubes. The ESI
produced ions are generally solvated clusters. The solvated clusters must be desolvated to
be mass analyzed. Several researchers Observed no signal fiom an analyte without

59

external assistance for the desolvation. On the other hand, expansion Of gas and solvent
vapor molecules into a vacuum creates a strong cooling. The resulting condensation Of
polar neutrals (notably water vapor) on analyte ions produces cluster ions having a mass
far beyond the range Of most mass analyzers (13).

In this study, a stainless steel capillary tube (0.5 mm 1. D., 1/16" 0. D.) was used
to transport ions produced at atmospheric pressure or outside the mass spectrometer into
the high vacuum region Of the mass spectrometer. The declustering was achieved by
heating the capillary tube while ions and gas were passing through it.

A capillary tube with 0.5 mm inner diameter was used to transport ions produced
in atmospheric pressure into the ESI source region (several torr). The inner diameter Of
the capillary tube controls the flow Of gas and ions tO Obtain the adequate degree of
vacuum inside the mass analyzer using the present vacuum system. The ends Of a stainless
steel capillary tube (0.5 mm ID, 1/16" OD and 25 cm long) were electropolished tO
remove the sharp ends. A heating wire (1/16" OD) was wrapped around the capillary tube
tO control heating Of the capillary tube. The capillary tube and the heating wire were
placed inside another stainless steel tube (3/16” OD, 1/8” ID) as shown in Figure 2-6.

In this design, the ion transport capillary has to be placed inside another tube
because a 25-cm long capillary tube can be bent or deformed by heating and cooling. The
deformation Of the capillary tube can reduce the ion transmission emciency. One end Of
the capillary tube was silver soldered to one end Of the outer stainless steel tube to Obtain
a vacuum seal. The whole length Of capillary tube can be heated by this way tO achieve
complete desolvation Of ions. A controlled voltage was applied tO the heating wire using a
voltage regulator. A Cr—Ni thermocouple was attached tO the outer walls Of the capillary
tube and temperature Of the tube was monitored using a thermocouple gauge. The
capillary tube was connected to a voltage supply (SOV-ISOV) (Hewlett Packard 6207B
DC power supply) to apply voltage to the ion transport capillary.

60

skimmer (0.5mm ID.)

in A:

 

 

 

 

 

 

 

 

 

 

 

 

        

._ ._:::_:.:._1$?:-:-;12-':-:-: ; .:v:.:::-;;~‘-;

   

 

 

W -‘.'.'.'.‘.-.-'.-.-.-.‘.‘.'.~.‘;-.‘.'.'.'

Heated Ion Transport
Capillary Tube

1" pumping line

Insulator

O-ring

Figure 2-5. The ESI source designed for the HP 5985 mass spectrometer.

61
Alternatively, the ion transport capillary tube was resistively heated by applying DC

current through the Opposite ends of the capillary tube using a power supply (Sorensen
Power Supplies, Raytheon CO. SRL 20-25). The desolvation was achieved by both
methods separately.

It is important tO have a very good vacuum seal between the stainless steel
capillary tube and the vacuum flange during repeated heating and cooling processes. The
vacuum seal was made between the 3/16" tube and the vacuum flange using a stainless
steel Swagelok fitting and a Vespel fen'ule.

For this system, a capillary tube was used to transport ions into the vacuum
system. Therefore, it can be assumed that the nozzle diameter is that Of the capillary tube.
The pressure inside the ESI chamber is about 1.5 torr. The upstream pressure Of the
capillary tube is 760 torr. The silent zone exists up to 7.5 mm from the end Of the
capillary tube inside the ESI chamber (from eqn 2-3). For efiicient ion sampling, the
skimmer should be kept inside the silent zone (13). In this system the skimmer is kept at
about 3.4 mm fiom the end of the capillary tube. If the skimmer is kept very close to the
capillary tube, the pressure inside the source housing cannot be maintained properly. That
is mainly because Of most Of ions and gases emerging from the capillary tube through the

skimmer into the source region.

Pressure Inside the Mass Analyzer:
when heated capillary inlet closed ..................................... 5 X 10'7 to“
with capillary inlet Open to atmospheric air --------------------- ~5 x 10'6 torr

(at 100°C)

62

58888092 v.88 3% mm 05 com gammmov 33 ace—=98 team—:5 com 2F .orm 82mm

538m ems;
>

.2; , .6

an: own;
same b b m _>

k ' ’ ’ l 'l ’r ' ' ' ’ ’
1" 1" lllrl

a ‘v 1" 1" 1' ‘ ll: t‘v‘"" ..

a

 

 

 

 

 

 

0330688.;

63

F. Electrospray Needle Assembly (with nebulizer)

In an ESI interface, an analyte solution is electrosprayed into the atmosphere
through a metal capillary tube (150 pm) that is held at a potential of several kilovolts
relative to the counter electrode. The buildup of charges at the liquid surface at the tip of
the capillary tube creates such an instability in the liquid such that coulomb repulsion
forces are suflicient to overcome the surface tension so that small charged droplets
separate fiom the liquid emerging from the capillary tube. The technique works best with
flow rates in the range of 5-10 uL/min (2). At higher flow rates and/or higher percentages
of water, the electrosprayer produced a stream of larger droplets together with the desired
fog (3). An increased voltage applied to the spray capillary dispersed the liquid droplets,
but at the same time initiated a corona discharge (23, 24) that substantially lowered the
sensitivity for samples ionized in solution. The combination of pneumatic nebulization and
electric field tolerates higher eluent flow rates and a higher percentage of water in the
formation of a spray of charged droplets (6).

An electrospray needle assembly was constructed in house with a nebulizer as
shown in Figure 2-7. The construction of ESI needle assembly was canied out using
standard fittings and other components. The ESI needle is a lO-cm long stainless steel
capillary tube and it has an inner diameter of 150 um and outer diameter Of 250 pm. The
capillary tube was placed inside another stainless steel capillary tube with an inner
diameter of 275 um (outer diameter of 1/16”). The outer capillary tube was secured to a
stainless steel union Tee (Swagelok®) using a Vespel® ferrule (I. D. = 1/16”). The small
capillary tube protrudes about 0.2 mm from the outer capillary tube. The small capillary
tube was sent through the union Tee and tightened with another Vespel ferrule (I. D. =
250 pm) at the other end of the Tee. The other end of the stainless steel capillary tube
was connected to a fused silica capillary tube fi'om the liquid chromatographic system. A

64
butt connector was placed in between two pieces Of firsed silica capillaries to avoid a
leakage of high voltage into the HPLC pump. A Teflon tube (0. D. = 1/16") carrying N2
gas was connected to the remaining outlet of the union Tee. N2 at 35 psig was used for
pneumatic nebulization. The stainless steel union was connected to a high voltage (0 - 5
kV) supply (Bertan Associate, Inc, Model 603-50P). The whole needle assembly was
secured on a custom-made polycarbonate holder. The holder and the needle assembly
were placed inside a Plexiglas cylinder that is front-side Open, for safety purposes. The
Plexiglas cylinder was mounted on a movable rail to adjust the height and the position

relative to the ion transport capillary.

G. Electropolishing Of the Capillary Tubes

The sharp edges of the ion transport capillary and the ESI needle were removed by
electrOpolishing both capillary tubes: the ESI needle and the ion transport capillary. A
mixture of glyceroVH3PO4/HZO (1:1 :1) was used as the electrolyte solution. A stainless
steel electrode having a larger surface area was used as a cathode. The capillary tube was
used as the anode. A current of 50 mA was passed through the circuit for about 2 hr or
until the desired smoothness was Obtained. The capillary tubes were thoroughly cleaned

with water and methanol before use.

65

new bum—.502 505$» .5 55 33 on :8 38m 05. 6:802 xmuamoboflm firm 0.53m

cougfinoz e8 mew NZ

33 coach l

> o_ a - c + 3 .H a: om: 23
526 85 seamen

 

 

one rezone our 825% .o ._ :9:

 

 

 

 

 

 

 

wmqtvmrrw

”r: .'

 

95a 0A 88%

66

uoaoEobooam .932 3% mm 0.,: co coerce: moisten—m 829:8 2:. .w-m 05mm

9 so: menace - ..H

a mm H m
— 5:38
tonnage :8
«.882 Hmm

/ _ m

\ /////////////

E

a _.l

BASED 8588/

\’

 

 

 

67
V. RESULTS AND DISCUSSION

Several configurations were considered for the construction Of the ESI source for
the HP 5985 single quadrupole mass spectrometer. The Vestec corporation implemented
an ESI source for a single quadrupole mass spectrometer and in this design, a heated ESI
source was used to achieve declustering of solvated ions. Three skimmers in series were
used to sample the flow of ions and gas molecules into the mass spectrometer. The
introduction of two stages of differential pumping and two pumping lines before the mass
analyzer is not easily provided in our mass spectrometer, because the source housing and
the flange must be modified to introduce two additional pumping lines and the probe.

Heating gas has been employed to achieve declustering of solvated ions in some
designs. These configurations need high pumping capacity pumps to evacuate the source
housing and mass analyzer regions (Fenn et al. used 1000L/s and 3000Us pumps (3);
Henion et al. cryo surfaces at 20-30K in the source housing (5), Smith et al. 30,000Us
and 500113 (5a).

Chowdhury, Katta, and Chait (7) reported an ESI/MS interface that uses a heated
capillary tube to transfer ions produced at atmospheric pressure into the mass
spectrometer and to desolvate the ions passing through it. The advantages of this design
are: after the first adjustment of the position of the heated capillary tube no further
adjustments are needed, the spray plume can be optimized by visually Observing the spray,
no need for heating gas, and therefore no need for very high pumping capacity.

Chait's design was compatible with our modifications because of the simplicity of
the interface. Simulation studies show that the best ion transmission can be Obtained when
the ion transport capillary, the ESI needle, the skimmer, and the mass analyzer are placed

co-axially. Therefore, a new source, an ion transport capillary, and an ESI needle were

68
designed and built as shown in Figures 2-4, 2-6, 2-7. The vacuum flange was modified to
accommodate the ion transport capillary and the pumping line for the mechanical pump
(Figure 2-5). The ion transport capillary was installed and isolated from the source and
the flange by a Teflon spacer and a Vespel ferrule, respectively. The outside end of the
heated capillary tube was closed with a Teflon cap when the system is not in use. The
complete system is shown in Figure 2-8.

The pressure inside the mass analyzer was monitored using a vacuum gauge
located at the base of the analyzer chamber. The pressure inside the mass analyzer was
found to be 5 x 10'7 torr (capillary end closed) and 5 x 10'5 torr (capillary Open to the
atmosphere) when the capillary is at 100°C. Back streaming of difi‘usion pump Oil was
observed at the source region. Since one end of the heated capillary tube is Open to the
atmospheric air, it has no influence on the flow rate of the liquid inlet system provided a
stable spray. This feature is very important when combining liquid separation techniques
with the ESI/MS. The ESI needle is capable of producing a stable spray plume at high
flow rates (>300 uL/min) with the assistance Of nebulizing gas. The ESI needle was kept
a few mm Off-axis to the heated capillary tube to avoid sampling of larger droplets. The
heated capillary can be cleaned with a stainless steel capillary tube (350 um ID) when it is
plugged with residual solids. The distance between the capillary tube and the ESI needle
is about 1cm. The position of the needle relative to the capillary tube was optimized by
monitoring the current emitted by the ESI spray.

The analyte solution is electrosprayed fi'om the ESI needle that is maintained at
3kV relative to the ion transport capillary tube through which droplets, ions, and gases
enter the mass spectrometer. A flow rate of 10 uL/min was maintained by a Waters
HPLC pump with a splitter (split ratio 1:9). The distance between the electrospray needle
tip and the capillary tube is typically 1cm. The quality Of the mass spectrum is strongly
dependent on the quality of the spray emitting from the needle. In the present design, the

69

spray can rapidly be optimized by direct Observation from outside the vacuum housing and
by monitoring the current emitted fi'om the needle. The position Of the stainless steel
capillary tube was adjusted to give a fine symmetrical plume.

Electrospraying of the analyte solution produces fine, highly charged droplets.
These droplets attempt to follow the electric field lines and migrate towards the metal
capillary tube that projects into the first vacuum stage of the mass spectrometer. The first
vacuum region is evacuated by rotary pump (Edwards E8, ). A fi'action of migrating
droplets enters the long stainless steel capillary tube assisted by the strong flow of gas that
results fiom the large pressure difference between the extremities Of the tube. Droplets
entering the tube tend to be focused toward its axis by this strong gas flow (25) and are
thus transported through the tube (7). The droplets were continuously desolvated inside
the capillary tube by controlling the heat applied to the capillary tube. A fi'action Of the
material that emerges from the capillary tube passes into a second vacuum chamber
through a coaxial, 0.5 mm diameter orifice in a skimmer situated about 3.4 mm from the
end of the tube. The solvent vapor emerging from the heated capillary tube was removed
by the rotary pumping stage before the skimmer. The ions that pass through the skimmer
are focused by a set of ion focusing lenses into the mass analyzer. Further desolvation was
carried out in the skimmer region by collisional activation due to the voltage difference
between capillary and skimmer (26, 27).

A Data Collection

After the successful modification of the ESI source and the vacuum system, the
ESI interface was assessed for its ability to make electrosprayed ions. It is known that
molecules having pre-forrned charges show very good response (6) in electrospray mass
spectrometry.

Tetraphenylphosphonium iodide (molecular weight of tetraphenylphosphonium

cation is 339 u) was selected as the model compound for tuning the mass spectrometer.

70
Tetraphenylphosphonium bromide dissolved in 4% acetic acid in methanol/H20 (SO/46M
% v/v/v) was continuously infused into the ESI interface. A continuous flow rate of
lOuL/min was maintained by a Waters HPLC pump (model 410) with a split ratio of 1:9.
The Waters HPLC pump has the minimum flow rate of lOOuL/min and increments of
lOOuUmin. A voltage of +3.9kV was applied to the ESI needle. The temperature of the
ion transport capillary was elevated to about 85°C as measured by a Cr/Nr thermocouple
attached to the outside wall Of the ion transport capillary. The voltage on the capillary
tube was held at 100V. The skimmer was kept at 15V. The ESI current between the ESI
needle and the ion transport capillary was Optimized by looking at an oscilloscope
connected to a home-made picoammeter. Figure 2-9 shows the circuit diagram used for
the ESI current measurement. The distance between the capillary tube and the needle was
adjusted to obtain the optimum current. The ESI needle was kept at about 1cm away
fiom the capillary tube and a few mm ofl'-axis. A high frequency oscilloscope was
connected to the mass spectrometer output using an extension to monitor the detector
signal. The mass analyzer was continuously scanned between 330 and 345u. The voltages
on the lenses were varied while continuously scanning the mass analyzer. Once the signal
for the tetraphenylphosphonium cation (TPP+ m/z = 339) was observed on the
oscilloscope, it was easy to optimize the signal by changing lens voltages, skimmer
voltage, capillary voltage, capillary temperature etc.,. Further, the signal was Optimized by
moving the position of capillary tube relative to the skimmer. Afier optimizing the system
with the oscilloscope, the mass spectrometer was re-tuned with the data system. The
present PIP-5985 data system has no capability to calibrate the mass spectrometer without
the E1 source. Therefore, the mass spectrometer could not be calibrated with the ESI
interface. When the E81 needle was very close to the heated capillary tube, a discharge
could occur and a glow (a discharge between the needle and the heated capillary) could be
observed if the surrounding was dark enough. The discharge could be heard if the

71
surroundings were quiet. When a discharge occurs, the major ion in the spectrum was due
to H+ and no other major ion was observed.

The change in capillary voltage with the ion intensity of stanozolol (M. W. = 328)
in methanol/acetic acid (1%) containing very basic pyrazole ring in the molecule was
studied. The intensity of the signal did not change significantly when the capillary voltage
was varied between 75-125V. Outside this range a drop in signal intensity was observed.
As the skimmer voltage falls below 10V, the signal intensity started to drop. When the
skimmer voltage was above 20V, the intensity of the peak becomes weak. The
temperature of the capillary also can be varied without adversely affecting the signal.

Next, an arginine (50 pmol/uL) solution was introduced into the ESI interface at a
flow rate of 10 uL/min. The voltage on the ESI needle was at 3kV, heated capillary was
at 100V and 100°C, the skimmer was at 15V, and lenses were maintained at the same
values after tuning. Arginine (M. W. = 174) was detected as a protonated molecule under
the above conditions. The spectrum contains a single peak representing the protonated

molecule of arginine in solution. No fragmentation was Observed for this species.

The capability Of forming multiple charges was assessed by introducing a solution
of tri-peptide lys-tyr-lys (10 pmol/uL) (M. W. = 437). It contains two lysine residues
(two fi'ee side chain amino groups) and a terminal amino group. The spectrum of tri-
peptide contains two major peaks at m/z 438 and 220 corresponding to the protonated
molecule and the doubly charged molecule, respectively. The doubly charged molecule
has the m/z value of 219.5, but the data system presents it as 220. Figure 2-10 shows the
ESI/MS mass spectrum of the tri-peptide lys-tyr-lys.

Afier establishing spectra of low molecular weight compounds, higher molecular
weight peptides were tried on the ESI interface. A seven-residue peptide, Val4
Angiotensin III, (arg-val-tyr-val-his-pro-phe) (M. W. = 917) was successfirlly tried on the

72
ESI/MS interface. The spectrum (Figure 2-11) shows the protonated molecule at m/z =
918 and the doubly charged molecule at m/z 459. The Val4 Angiotensin III peptide has
arg, his, and a terminal amino group as possible charging sites; but the doubly charged ion
was observed as the highest charge state. The spectrum of bradykinin (M. W. = 1060)
having the sequence of arg-prO-pro-gly-phe-ser-prO-phe-arg was also acquired. This
molecule has two possible charging sites: a fi'ee amino terminal and the amine group on
the arginine side chain. Only doubly charged molecule was Observed in its spectrum, but
the singly charged molecule has an m/z value above the limit of the mass analyzer which is
1000 Daltons. These molecules did not show any fragmentation under these conditions.
The main goal of using the modified ESI interface is for the analysis of small molecules.

Therefore, further analyses Of large molecules were not attempted with the interface.
VI. Design and Construction of Computer Interface
A. Introduction

The HP 5985 mass spectrometer uses a HP 21MX-E series computer with the
Tektronix 4012 graphics display terminal to control the mass spectrometer. The data are
transferred to an Old fashioned cartridge or PDP-ll system for storage. The data system
on the HP 5985 mass spectrometer is no longer reliable for continuous operation. At the
same time it did not provide several desired capabilities that are required for modern mass
spectrometry such as fast data acquisition, profile data collection, etc.,. The calibration of
the mass spectrometer can only be done with the E1 source. For the calibration, the ESI
source had to be replaced with the E1 source. The scanning rate of the instrument cannot

be fully controlled by the user.

929:3 Hmm manage .8.“ 86me “Soho 05. .mrm Saw—m

 

 

 

 

2.823
3
7 3 ¢
>c_m-o my
«2 _

 

 

S596

tease :3

 

H‘

BE

 

 

 

“.10.. . :

 

 

 

"33 Ill”

 

 

.38wa use
>2+
. /18_|_|»
one. "5 Jlllwu a: .e
/ >8-

m

 

 

 

 

 

_\#

_/_

7
J\

 

 

 

 

 

 

>2 +

 

 

 

 

 

 

>2-

 

74‘

 

Relative Intensity

 

M210“

 

0 130 170 210

220

250

290

(Mr-H)“
438

330 370

Figure 2-10. ESI/MS mass spectrum of Lys-Tyr-Lys (M.W. = 437).

 

570

75

 

459

(M+2H)2+

 

 

 

 

90 130 170 210 250 290 330 370 410

 

Relative Intensity

 

 

 

810 850 890 930 970 1010 1050

570 610 650 690 730 770

Figure 2-11. ESI/MS mass spectrum of Angiotensin III (M. W. = 917). Flow rate,=
lOuUmin. ESI needle at 3.9kV. Heated capillary at 100°C and 100V and the Skimmer at

15V.

76
B. Computer Interface

An inexpensive interface was designed and constructed in house using readily
available components. The 21MX E-series computer on the HP 5985 mass spectrometer
was replaced with a 486 DX, 33MHz personal computer (Gateway 2000, S. Dakota).
Since the 486 computer is not directly compatible with the electronics of the HP 5985
mass spectrometer, an communication interface between the 486 computer and the mass
spectrometer has been developed. A block diagram of the digital interface for the HP
5985 mass spectrometer is shown in Figure 2-12.

The HP 5985 mass spectrometer has an asynchronous digital interface consisting
of separate l6-bit input and output ports with a write strobe to initiate commands to the
HP 5985 and a status flag from the HP 5985 indicating data validity for read and write
operations. The host computer uses an enhanced version of the industry standard D1024,
a CIO—DIO48 to drive this interface.

The CIO-DIO48 digital I/O board (Computer Products, Inc., Mansfield, MA) has
24 digital I/O lines and 24 high current (60 mA) output lines. Figure 2-13 shows a
schematic diagram of the CIO-DIO48 computer board (28). The board is based upon the
industry workhorse 82C55 Programmable Peripheral Interface chip. The 82C55 has 2-
byte wide bi-directional data ports (A and B), two 4-bit bi-directional data or control
ports bi (C high and C low), and a control/status register used to configure data direction
of ports and mode Of Operation. The first 82C55 has ports A and B configured as two 8-
bit strobed inputs and is used to receive 16-bit data fiom the 5895 via connector P1. The
second 82C55 on board has high current drivers bufl‘ering its cmos level signals. It is
configured as two 8-bit strobed output ports and is used to send 16-bit commands to the
HP 5985 through P2. The PC/XT/AT bus interface on the board was plugged into an

extra expansion slot in the system board of the computer. All signals are connected to the

77
HP 5985 through two 2-meter long ribbon cables and a custom adapter board as shown in

Figure 2-14.

The adapter board simplifies routing the 2 cables from the 37D connectors P1 and
P2 on the CIO-DIO48 to the single connector on the rear of the HP 5985. The digital
interface on the mass spectrometer uses a 48 contact, 0.156" on center, edge connector.
A standard 50-contact connector, IS, on the adapter has 50 pins and care is required to
insure proper alignment of contacts. The top two pins of the J 5 adapter (pin numbers 49
and 50) were not connected. Figure 2-15 shows the schematic diagram Of the adapter.
The P1 male connector of the CIO-DIO is connected to a female connector, 11, on the
ribbon cable. The connector, J3, at the other end Of the cable is connected to the P3
connector on the adapter. The P2 connector on the board was connected to the 12
connector on another 2 meter ribbon cable and J4 connector at the other end was
connected to the P4 connector on the adapter. The RF-fi'equency driver and the other
electronic components of the HP 5985 mass spectrometer were used without
modifications.

C. Data Acquisition and Processing Program

A window-based computer program to control the mass spectrometer and to
acquire data was developed in the laboratory with the assistance of a computer
programmer. This program allows a user to interact with the mass spectrometer using a
DX-486, 33MI-Iz computer. Figure 2-16 shows the control panel Of the program. The
voltages on difl'erent lenses can be changed with the sliding bars associated with the
variables. The resolution of mass spectrometer was adjusted by changing the Ofi‘set and
the Gain of the mass axis. The program displays the real-tirne data, the total ion
chromatogram, and a mass spectrum on the computer screen. Figure 2-17 shows the
ESI/MS mass spectrum of stanozolol (M. W.=328) that .was collected using new program.

78

5380.50on m8:— mwam mm 05 com 3a.t8£ 3&6 05 .«o :5qu 3003 oi. .NTN ocswfi

2 Al mlfil 2

3538 x83 3 AIS. AI E .8 3 4| El: 3

 

8282800

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

nu / mm

mm fl mm .8833

“M W ”....
mam 8338 u / .. B E a

13:41:43] : a m
a. u o no a
mm fl mm mew.
E 2 mm mm / m a
u. m m" fl "m m
1 .l. m fl "u ,
2 2 E

 

 

 

wwounromu NH 038 coca

79'

ounce coarse $05-06 cede cameo assess. are .m 3 seem

 

020
>+

2w m.
z. 5

ho

F0

um

pm

b<

—<

 

 

  

 

 

 

E

— mam 8.3332:an
mmmuuam
(.50
5
mo
0 0
mg
C Q.
as mam <20

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

80

 

38888on 82: 3% mm 2: :0 Been 3:va 2: SE ow 8 bacon ovaE 8830 05 .«o .53an 23828 2:. .Erm 23E

 

      

N a
v n
O m
0 5
0a 0
Na «a
Ca Ha
04 fin
.4 ha

ON 04

QFDH‘OFUH"

.h'n-fl
HHHHHNNNNN‘d4-9‘OFUH"

.soaueccou so... use. one... ..an on (0:. 37.219... ca Cuzco on as)... 0.30
.LOvuoccou out. ax unattacolcoc (0.: now a Orvooootcou ate... a co: 0..

O’N- ‘CN so «1.0 (20¢ v «.0& d 0.0t Du ‘OLOIESQ it.
It.” IVflOMO CO 0’“. «it‘s 50 n).- .huv «.0; d v.0; 0v DOvSOL ca 0483

"moeoz

 

 

81

 

coatoacm 88886on made 3% mm 05 2 Boos $05.05 Set $33 05.. «TN oeawfi

IQNOuO co
aha-0v « 0»

.nrt OOOFt! o

N
a
a
a
a
a
O
h
0
H
a

      
 

IVNO
abnoo

   

(0
0h

82
Figure 2-18 shows the reconstructed total ion current chromatogram obtained by injecting
100pg of testosterone hydrazone into a HPLC system on the ESI mass spectrometer
interface. The scan rate was 1 scan/sec (3 50-450u) and the flow rate of the liquid stream
was 40uL/min. A C13 column (15cm x 800nm) having 3-pm particles (LC packings) was
used. The mobile phase was CH3CN/1120I'I‘FA (40/59.9/0.1 %v/v).

The data system is still under development. The data system has no ability to
process data on the same program. Therefore, the data were transferred into itr-2,
another data system that was developed for a time-of-flight mass spectrometer after
transforming into a compatible format. The development of the data system is not
complete at this stage. The goal of the development of the data system is to collect and
process data on the same computer without transferring into another computer. Usually,
ESI/MS data is noisier than EI data and therefore, capabilities such as averaging and

smoothing are a very important part of the ESI/MS data processing and development.
VII. Triple Quadrupole Mass Spectrometer (TQMS)

In this research project, a triple quadrupole mass spectrometer (TQMS) fitted with
a Finnigan electrospray source was also used to study the fiagmentation patterns and its
other modes of Operations. The Finnigan electrospray source that replaced their earlier
source (from Analytica Branford) has some similar features to the ESI source developed
at this laboratory. The main different features are the incorporation of a tube lens and the
octapole rod assembly to focus ions emerging fi'om the skimmer. Figure 2-19 shows the

Finnigan ESI source.

83

38888on 88: 3% mm 05 new 8898 83 3538 05 me $53 35:8 2F OTN ouswfi

 

 

 

 

 

 

 

 

4.52

a 8.88. e “E E 8.8.
o 83> 2m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  
 
 

 

 

......an 8.8 fin M.,...
..ocw coboom
cod @ . . .n omd— .
:o_mm_Em
8.8- g g 8.8
mmmrx 8.33m

 

 

Home:

 

 

 

.8 96m 83 8:5 825

 

 

 

 

 

 

 

 

 

 

 

 

.9550

 

 

84

 

 

 

 

 

. mm)" 329

.arc50242

l

-r *
- . . . - , .l . .1

100 300

Figure 2-17. The ESI/MS mass spectrum of stanozolol acquired using the new data
system.

 

Testosterone hydrazone
‘ M+=402

 

 

 

at, -. . .I
0 200 400
scan#

 

Figure 2-18. The reconstructed total ion chromatogram of a LC/ESI/MS analysis of
(100pg) testosterone hydrazone. C-18 (15cm x 800 m, 3pm particles) column. Flow rate
40uL/min Of ACN/HZO/TFA (40/59.9/0.1 %v/v). Scan rate 350-450 in 1 sec.

85

The principles and modes of Operation of TQMS will be briefly discussed here.
The triple-quadrupole mass spectrometer (TQMS) is a tandem arrangement, in which the
first and third quadrupoles are mass-selective and the second quadrupole serves as
collision cell (29,30). A schematic of a triple quadrupole mass spectrometer is shown in
Figure 2-20. Following ionization in ion source, the ions are focused and passed into the
first mass analyzer, Q1. After the first stage of mass analysis, the ions continue their flight
path into the second quadrupole, Q2, which is a collision cell or a transmission device.
Introduction of a reagent gas into this region (Q2) facilitates the interaction of mass-
selected ions and the reagent gas. The reaction products generated in Q2 may then pass
to the third quadrupole Q3, which is the second mass analyzer. This technique is referred
to as tandem mass spectrometry, mass spectrometry/mass spectrometry, or simply
MS/MS. The technique provides a powerful tool to study the ion-molecule reactions and

collisionally activated dissociation processes of selected ions.
A. MS/MS Scan Modes with TQMS

There are four difl‘erent MS/MS scan modes available with a triple quadrupole
mass spectrometer, the daughter ion scan, parent ion scan, functional relationship scan and
the selected reaction monitoring (SRM) mode (31,32). The TQMS may also be Operated
as a single stage quadrupole mass spectrometer by scanning only one of the two mass
analyzers while the other two quadrupole rod assemblies in the RF -0nly mode, firnction as
ion transmission devices. Thefirnction of the quadrupoles in the different scan modes is
illustrated in Table 2-2.

1. Daughter Ion Scan

The most familiar and commonly used MS/MS scan mode is the daughter ion scan

also called product ion scan. In this mode, the first mass analyzer, Q1, is set to pass ion

86
current of a particular mass-to-charge (m/z) ratio which is referred to as the precursor ion,
parent ion, or reactant ion. Precursor ions selected by Q1 interact with the collision gas in
Q2 generating product ions or daughter ions. These fragment ions formed in the collision

cell enter the third quadrupole which is scanned to Obtain a product (or daughter) ion mass

spectrum.
2. Parent Ion Sean

A parent ion (also called precursor ion) mass spectrum is obtained if Q3 is set to
pass a selected product ion, while the first mass analyzer, Q1, is scanned sequentially
passing parent ions with different mass into Q2. The resulting spectrum will indicate all
those ions generated in the source which upon reaction with collision gas in Q2, yield a

specific m/z value as chosen by the second mass analyzer, Q3.
3. Functional Relationship Scan

In a functional relationship scan, both Q1 and Q3 are scanned with a mass offset.
The most familiar functional relationship scan is a neutral loss or neutral gain scan, where
the mass offset remains constant. Neutral loss scans are used when examining dissociation
reactions in the collision cell. A neutral gain scan may be utilized if a reactive collision gas
is introduced into the collision cell for associative ion/molecule reactions. In a neutral
loss/gain experiment, ions will be detected provided that the parent ion eliminates (or
gains) a neutral moiety prior to entering the second stage of mass analysis, the mass lost or
gained being equal to the constant mass offset of Q1 and Q3. V

87

Table 2-2: Function of the quadrupoles in various TQMS scan modes.

Q1 Q2 Q3 Mode

Scan RF-Only RF-Only MS

RF-0nly RF-Only Scan MS

Fixed RF-0nly RF-Only SIM (MS)

RF-0nly RF-Only Fixed SIM (MS)

Fixed RF-Only Scan Daughter Ion
Scan (MS/MS)

Scan RF-Only Fixed Parent Ion
Scan (MS/MS)

Scan RF-Only Scan Functional
Relationship
(MS/MS)

Fixed RF-Only Fixed SRM (MS/MS)

MS = single stage analysis SIM = selected ion monitoring

MS/MS = tandem analysis SRM selected reaction monitoring

4. Selected Reaction Monitoring (SRM)

Selected reaction monitoring refers to the MS/MS mode in which Q1 and Q3 are
set to pass ions Of only one m/z value. Neither of the mass analyzers is scanned, a feature
which provides greater sensitivity at the expense of flexibility. Selected reaction

monitoring (SRM) is analogous to a selected ion monitoring (SIM) experiment which is

88

employed with a single mass analyzer, and is more selective than SIM experiment. The
SRM and SIM techniques are Often used to detect low levels of a selected analyte.

89

0088003 «32 con 0mm. 05 :0 00%85 $538005 comm—BE 0.3. .34.. 08mg

 

 

 

 

 

 

 

 

 

855 09.5.“. 95.5 3082002
82 3 - 08 8.8a
€an 3 05805 0t0fimo§<
‘1 06:65
i. {
U 4
l— a 0:802 Hmm
Own. 83 23300 909 + r 8
See a -8 be as no

 

 

 

 

9O

08080Q

08088809." 0008 06020000 033 02. 0mg. 05. .oN-N 0.5mE

3980000 00.58 :8 H
00380000 0004 n

0003000 225.6020 vcooom n
3980000 0.80000 3:0-va 009805 comma—00 N
0003000 0:58.009: «00E H

 

 

 

 

91

VIII. References

10.

ll.

12.

13.

14.

15.

16.

Smith, R D.; Olivares, J. A.; Nguyen, N. T.; Udseth, H. R. Anal. Chem. 1988,
60, 436

Aebersold, R.; Bures, E. J.; Namchuk, M.; Goghari, M. H.; Shusan, B.; Covey, 'l‘.
C. Protein Science 1992, 1, 494-503

Yamashita, M.; Fenn, J. B. J. Phys. Chem. 1984, 88, 4451
Yamashita, M.; Fenn, J. B. J. Phys. Chem. 1984, 88, 4471

Olivares, J. A.; Nguyen, N. T.; Yonker, C. R.; Smith, R. D. Anal. Chem. 1987, 59,
1232

Bruins, A. P.; Covey, T. R.; Henion, J. D. Anal. Chem. 1987, 59, 2642.

Chowhury, S. K.; Katta, V.; Chait, B. T. Rcmid Commun. Mass Spectrom. 1990,
4, 81

Allen, M. A.; Vestal, M. L. J. Am. Soc. Mass Spectrom. 1992, 3, 18

Dawson, P. H. Quadrupole Mass Spetrometiy and Its Applications. Elsevier, New
York, 197 6

Dawson, P. H.; Yu, B. Int. J. Mass Spectrom. Ion Processes, 1984, 56, 25

Miller, P. E. Denton, M. B. J. Chem. Educat. 1986, 63, 617

Miiler, P. E. Denton, M. B. Int. J. Mass Spectrom. Ion Processes 1986, 72, 223
Bruins, A. P. Mass Spec. Rev. 1991, 10, 53-77

Anderson, I. B.; Andres, R. P.; Fenn, J. B. In "Advavced Atomic Molecular
Physics” Bates, D. R.; Estermann, I., Eds; Academic Press, New York, 1965,
Chapter 8

Camparque, R. J. Phys. Chem. 1984, 88, 4466-4474

Beijerinck, H. C. W.; van Gerwen, R. J. F.; Kerstel, E. R. T.; Martens, J. F. M.;

van Vlembergen, E. J. W.; Smits, M. R. Th.; Kaasheok, G. H. Chem. Phys. 1985,
96, 153

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.
29.

30.

31.

32.

92
Gray, A L. J. Anal. At. Spectrom. 1989, 4, 371

Seraukas, R. V.; Brown, G. R.; Pertel, R. Int. J. Mass. Spectrom. Ion. Phys. 1975,
16, 69-87

Alexander, A. J .; Kebarle, P. Anal. Chem. 1986, 58, 471

Arpino, P. Mass Spectrometry Rev. 1990, 9, 631

SIMION, Dahl, D. A.; Delmore, J. E. Idaho National Engineering Laboratory,
Idaho Falls.

O'Hanlon, J. F. A User Guide to Vacuum Technology, Wiley-Interscience, New
York, 1989, pp 27

Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. H. Mass
Spectrometry Rev. 1990, 9, 37

Smith, R D.; Loo, J. A.; Loo, R. R. 0.; Busman, M.; Udseth, H. R Mass
Spectrometry Rev. 1991, 10,359

Anderson, J. B.; Andres, R. P.; Fenn, J. B. Adv. Chem. Phys. 1966, 10, 275

L00, J. A.; Udseth, H. R.; Smith, R. D. Anal. Biochem. 1989, 179, 404 ‘

Loo, J. A.; Udseth, H. R.; Smith, R. D. Rapid Commun. Mass Spectrom. 1988, 2,
207

CIO-DIO user's manual, Rev. 3.0, Computer Boards, Inc., Mansfield, MA
Yost, R A.; Enke, C. G. Anal. Chem. 1979, 51, 1251A

Yost, R. A.; Enke, C. G. McGilvery, D. C.; Smith, D.; Morrison, J. D. Int. J. Mass
Spectrom. Ion Phys. 1979, 30, 127

Vincenti, M.; Schwartz, J. C.; Cooks, R. G.; Wade, A. P.; Enke, C. G. Org. Mass
Spectrom. 1988, 23, 579

Schwartz, J. C.; Wade, A. P.; Enke, C. G.; Cooks, R. G. Anal. Chem. 1990, 62,
1809

CHAPTER 3
PEPTIDE SEQUEN CIN G

I. INTRODUCTION

Many proteins of biOlogical interest are often only available in sub-picomole
quantities, and isolation and purification techniques reduce this amount further.
Consequently, there is a demand for an increase in sensitivity of current sequencing

methods or for developing new sequencing techniques.

Several chemical methods have been developed for determining the sequences of
proteins and peptides. All involve sequential degradation of the molecule fiom one end or
the other, removing and identifying the amino acids one at a time.

Sequencing peptides from the C-terminal is possible by a variety of chemical
methods (1). All suffer from poor yields, which limit the number of residues that can be
determined and most involve complex, multistep chemistry. Recent improvements in the
thiohydantoin method (2) show promise for making C-terminal sequencing more practical.
More often carboxypeptidases are used to remove amino acids one at a time fi'om the C-
terminus. The released amino acids are then identified. The enzymatic approach to C-
terminal sequencing sufi‘ers fiom varying rates of hydrolysis for different amino acid,
making its use less than ideal in practice.

The most accepted N-tenninal sequencing method is that first proposed by Edman
(3). This approach involves reaction of the amino-terminus of the peptide with
phenylisothiocyanate to form phenylthiocarbamayl derivative. The N-tenninal amino acid

93

94

residue is then cleaved, converted to the more stable phenylthiohydantoin, identified, and
quantitated. The remaining peptide is put through subsequent cycles of degradation.
Although amino terminal sequencing is sensitive and, as a result of automation, can be left
unattended for long periods of time, the process is relatively slow, normally taking about
lb per residue. The most critical limitation of the Edman sequencing of peptides and
proteins is the unavailability of a free amino terminus for coupling in N-terminal blocked
peptides and proteins. Microsequencing can yield uncertainty in certain residue
assignments, especially when the quantity is small or mixtures contribute overlapping
signals. As sequencing extends towards the C-terminus of a peptide, residue signals drop
off gradually owing to imperfect yields of the repetitive chemical and physical process.

Many post-translational modifications are not readily identified by Edman
chemistry methods because of the harsh conditions employed. The chromatographic
identification of the modified PTH-amino acids could be misassigned as another amino
acid with identical retention time, or simply taken as an unknown residue, especially when
low levels are sequenced. When a mass spectrometer is used for detection in a peptide
sequanator, reliable identification of the thiohydantoin (TH) amino acids can be

accomplished by their molecular weight information combined with HPLC retention times.
A Edman Degradation

The preferred procedure for determining the sequence of amino acids in a peptide has
involved the use of Edman degradation chemistry first published in 1950 (3). The Edman
procedure determines the sequence of amino acids in a peptide starting at the amino
terminus. The methodology employs phenylisothiocyanate (PITC) that attaches to the
amino terminus of the peptide as shown in Figure 3-1 to form the (phenylthiocarbamyl)
FTC-peptide under alkaline Conditions. Treatment of the FTC-peptide with anhydrous

trifluoroacetic acid (TFA) causes cleavage of the N-temiinal amino acid as a FTC-amino

95
acid in equilibrium with the anilinothiazolinone (ATZ) of the amino acid, which upon
exposure to aqueous TFA is converted to the phenylthiohydantoin (PTH) derivative of the
amino acid. The reaction solution is extracted under conditions to isolate the cyclized
phenylthiohydantoin derivative of the amino acid. Identifying the cyclized amino acid
PTH derivative by retention time on HPLC thus determines the amino acid at the amino
termimrs of the original peptide. The cycle is then repeated to determine the second amino

acid fi'om the amino temrinus, etc.

The validity of an analysis conducted with a conventional amino acid sequenator
rests on the stability of a set of parameters that affect the HPLC chromatogram. That is,
the conventional sequenator relies on a single channel of information, namely UV
absorbance versus time as a means of ”identifying” various amino acid derivatives that are
injected and eluted from an HPLC column automatically. Because the conventional
automated sequenator ”identifies” an amino acid solely by retention time, any operational
parameter that affects the elution profile or order of amino acids can seriously affect the
validity of the analysis. For example, in a typical HPLC protocol, histidine and arginine
are especially sensitive to ionic strength of the bufi‘er solution in the mobile phase and, to
some extent, are influenced by the condition of the end-capping of the column (4). Thus,
slight changes in the bufi'er solution from those used in analysis of the calibration sample
or the gradual deterioration of the column can cause either one of these two amino acid
derivatives to wander into time windows for other amino acids and give false-positive
responses. Another operational problem associated with HPLC and the use of chemical
reagents for derivatization prior to analysis by HPLC is the potential for interference from
the breakdown products of the reagents or compounds associated with the sample. For
example, one problem commonly encountered during low-level sequence analysis is the
potential interference of phenylthiourea (a breakdown product fi'om exposure of the

reagents to oxygen during sample application to the sequenator) which will co-elute with

96

tryptophan as the phenylthiohydantoin derivative (5). In this case, especially during the
first four to five cycles in a given assay, it is possible for phenylthiourea to cause a false
positive for tryptophan or to cause such a high chemical background that reliable detection
of tryptophan is not possible. A third problem in the use of the conventional automated
sequenator is evident in an analysis when any of the 20 common amino acids has been
modified. In these circumstances, the modified amino acid as the phenylthiohydantoin
derivative will elute at a non-standardized time window and, thus, will either be ”missed"
or it will elute in some other time window and will be misidentified (6). The latter
circumstance is frequently realized when hydroxyproline is present in the peptide; the PTH
derivative of hydroxyproline typically has the same retention time as the PTH derivative of
alanine or histidine, depending on the buffer . This is a common problem with sequenators
as indicated in a recent survey by the American Biomolecular Resource Service Facilities,
which showed that in a survey (7) of some 60 peptide facilities, 54% of them misidentified
hydroxyproline as being histidine.

Several factors limit the sensitivity of the Edman degradation procedure. The limit
of detection for the phenylthiohydantoin derivative of the amino acid (usually by UV
detection) during HPLC analysis establishes the minimum sample requirement for the
Edman degradation procedure for amino acid sequencing. The most common means of
detection is based on UV absorption of the phenylthiohydantoin derivative in a flow-
through detector at the exit of the HPLC instrument. Typically, five picomoles of the
phenylthiohydantoin derivative of the amino acid can be detected by UV absorption with
an acceptable signal-to-background ratio (8). Background in the HPLC eflluent either
fi'om biological sources or from chemical sources plays a significant role in the detection
limit issue. Under extraordinary conditions of using ultra-clean solvents and careful
cleaning of HPLC columns, it has been possible to achieve detection limits closer to 200
fmols (9) with UV detection ,of products from the Edman degradation procedure.

97

R1 0 R2 0 R....
ll

N=C =8 HzN—CH—C—NH—CH—C—NH—CH
Coupling

1p}! 8-10

H R] o a; o a...

l I II I ll

N—C—NH — CH —c—'—-NH—CH—c —'NH — CH

l

Anhydrous
J CF3COOH Cleavage
R2 0 R3 0 K...
NH
\c’N\ l l I II |
\ /CH-—R1 HzN—CH —C—NH—CH—C —NH — cu
S -—C
O
l CH3COOH/H20 Conversion
0
l
/
N \

CH—Rr PTH-AA
Ll.
/
s

Figure 3-1. Chemical Scheme for Edman Degradation of Peptide

. 98
A third factor limiting the sample size requirement in the Edman degradation procedure is
the cumulative efi‘ect of successive losses. (10-12) suffered by the residual peptide as one
amino acid after another is cleaved away and the peptide is repeatedly exposed to
extraction conditions for isolating the phenylthiohydantoin derivative of the newly cleaved

amino acid.

II. Analysis of Thiohydantoin Amino Acids by Mass Spectrometry

There have been several detection methods employed to increase the sensitivity
and the reliability of PTH detection. Mass spectrometry has been used to analyze PTH
amino acids reliably and accurately. The mass spectrometric techniques were able to
detect PTH derivatives successfully, but earlier investigators could not achieve the desired
sensitivity required for peptide sequencing. In this chapter, the use of mass spectrometry
in pm detection and peptide sequencing will be summarized . '

A. PTH Analysis by Electron Impact (E1) Mass Spectrometry

Pisano and Homing have separated the PTH and DNP derivatives of thirteen
amino acids by gas chromatography (13). Melvas separated five PTH amino acids by gas
chromatography and detected them by EI mass spectrometry (14). He used the PTH
amino acids of ala, pro, val, leu, and ile as model compounds. He was able to observe
very strong molecular ion peaks as well as extensive fragrnentations for the five PTH
amino acids.

Fairwell and Lovins (15) studied the suitability of ATZ amino acids for the
detection of the end product of Edman degradation chemistry. They compared the EI
mass spectra of several ATZ amino acids which are especially very fragile derivatives with
those of corresponding PTH amino acids. They observed that both ATZ and PTH amino
acids have very similar mass spectra. They demonstrated the ability to use ATZ amino

99
acids for mass spectral detection by sequencing the first five residues of Bovine
Ribonuclease. They concluded that during sample introduction to the mass spectrometer
through the direct inlet probe, ATZ amino acids spontaneously undergo thermal
rearrangement to the more stable corresponding PTH amino acid.

Sun and Lovins (16) studied the methyl thiohydantoin (MTH) and PTH derivatives
by EI mass spectrometry with low ionizing voltages. They obtained mass spectra of
eighteen MTH derivatives and thirteen er derivatives with ionizing voltages of 70eV
(usual. ionizing voltage), 20eV and 11eV. It could be readily observed that the spectra
using 11eV energy are much less complex than those obtained at higher electron voltages.
At the lower electron energies most if not all of the fi'agment ions begin to disappear and
the spectra contain in a number of cases only the molecular ion. With the exception of
arg, lys, and thr, the mass spectra of all derivatives investigated contained the molecular
ion as the base peak of the spectrum with only one or two and in some cases no fiagment
iom of significant intensity. The spectra of PTH derivatives of amino acids: arg, lys, and
thr gave an abundant fiagment ion and a small number of less abundant fragment ions in
their individual spectra at low ionizing voltages. They recorded the abundance of some
ions with the ionizing voltages and observed that the maximum sensitivity was observed at
about 20eV electron energy. The complexity of spectra greatly decreased at below 20eV,
but the sensitivity also decreased.

B. PTH Analysis by Chemical Ionization (CD-Mass Spectrometry

Fales et al. have analyzed PTH amino acids by CI/MS in order to achieve the high
sensitivity required by the Edman degradation (1?). The protonated molecule appeared as
the major peak in the CI spectrum in 16 PTH amino acids. The PTH derivatives of arg,
lys, cys, and asp gave fragment ions as the major peak in' the spectrum. This indicates the
stability of 16 PTH amino acids under the CI conditions. However under the CI

100
conditions, PTH-arg decomposes at high temperature (at 220°C), lys forms an unstable 8-
phenylthiocarbamyl PTH, PTH-cys undergoes extensive fragrnentations, and PTH-asp
loses CH202. The loss of CH202 can be observed with PTH-asp and PTH-glu that is
uncommon with the Other carboxyl amino acids under the CI conditions. Such an efi‘ect
had been observed previously with free amino acids under the CI conditions (18, 19).
These researchers compared the sensitivity of the PTH amino acids with E1 and CI mass
spectrometry and they reported that Cl is 100 times more sensitive than El because of
fewer fi'agmentations under CI conditions (20).

C. PTH-Amino Acid Analysis by TwooStep Laser Desorption/Multiphoton

Ionization

Engelke, Hahn, Henke and Zare reported their work on PTH-amino acid analysis
using two-step laser desorption/multiphoton ionization (21). In this experiment, the
sample was loaded onto a glass cup and spread out as a thin film over the inner surface.
The thickness of the sample varied fiom hundreds of monolayers to submonolayers. One
monolayer of the sample molecules on the glass substrate corresponds to about 10
picomoles in the desorbing C02 laser area of 0.01cm2.

After the sample is introduced into the vacuum system, as a first step, a C02 laser
beam is directed to the thin film of the sample. Neutral molecules are generated in a fast
desorption process and expand into the high vacuum between two electrodes that form the
acceleration region of a simple linear TOF mass spectrometer. In the second step, these
molecules are ionized by resonance-enhanced multiphoton ionization (REMPI) with an
UV laser (NszAG laser at 266 nm).

The main features of the spectra are: (l) a high yield of the molecular ion, (2) a
small degree of fiagmentation, and (3) considerable reduction of background. They

1 0 1
demonstrated the simplicity of the method by analyzing an equimolar mixture of five PTH-

arnino acids.

D. PTH Analysis by Thermospray Mass Spectrometry

Thermospray mass spectrometry (TS/MS) has been used to identify the PTH
aminoacidsbyPramaniketat. (22). Intheirearlywork, theyacquiredTSmassspectra
for 19 synthetic PTH-amino acids corresponding to those commonly observed during N-
terminal amino acid sequencing. All the PTH amino acids including PTH(PTC) lys, in
which the amino side chain additionally bears a PTC moiety, gave a prominent protonated
molecule. Ions corresponding to the addition of ammonium ions (M+NH4+) were not
observed. In most cases, a peak for the protonated molecule was the base peak in the
spectrum, although for PTH-thr and PTH(PTC)-lys, one or more fiagrnent ions were
always more abundant than the protonated molecule.

Preliminary investigations of the sensitivity in the full scanning mode (150 - 410 u,
entire mass range covering all the PTH amino acids), different quantities of standard PTH
amino acids ranging fi'om 1 - 150 pmol per component were subjected to LC/MS. All the
PTH amino acids could be identified down to the 50—pmol level. Six pm amino acids
could be identified at the l-pmol level with S/N ratio 10:1 or better (22).

TS/MS has been used to determine the composition of amino acids hydrolysates
after reaction with PITC to form PTC-amino acid (23). All the PTC-amino acids except
PTC-cys could be detected at 250-pmol level with S/N ratio 80:1. In the PTC-cys analysis
1 nmol was required, due to extensive fragmentation (23). These higher detection limits
of PTC-amino acids may be due to more extensive fiagmentation than those of PTH-

amino acids.

l 02
E. PTH Analysis by Atmospheric Pressure Chemical Ionization Mass Spectrometry

Sakari and Kambara (24) analyzed PTH amino acids using atmospheric pressure
chemical ionization (APCI) mass spectrometry. Samples and solvent molecules were
ionized by corona discharge and protonated molecules were observed with only a few
fi’agment ions in APCI mass spectra. They were able to detect PTH amino acids at the 10-

100 pg level using the selected ion monitoring mode.
F. Thiohydantoin Detection by ECNI/MS

Several reagents were introduced to increase the electron capturing properties of
the sequencing reagent in the Edman chemistry. The reagent
pentafluorophenylisothiocyanate (PFPITC) was proposed as a sequencing reagent and the
resulting PTH derivatives were analyzed by gas chromatography-electron capture
detection (25), or gas chromatography-negative ion chemical ionization mass
spectrometry (NICI-MS) (26, 27). The PFPTH-asn amino acid was observed at the 155-
frnol level with S/N ratio of 80 using NICI-MS. Andregg et al. (28) investigated the use
of 4-nitro phenylisothiocyanate (NPITC) that was first introduced by Tarr (29) to detect
the thiohydantoin amino acids by NICI-MS. They detected NPTH amino acids at the 50-
frnol level using selected ion monitoring (28).

G. Modified Thiohydantoin Detection by Pneumatically Assisted Electrospray
Mass Spectrometry

While we were pursuing this project to find modified Edman reagents for sensitive

detection of thiohydantoin amino acids by electrospray mass spectrometry, Aebersold et

103

al. introduced a new reagent for Edman degradation (30). The new reagent carries a
quaternary ammonium moiety separated by two methylene groups from the conventional
Edman reagent, PITC. They demonstrated detection limits for the thiohydantoin amino
acids at the attomol level. The reagent shows comparable coupling and cleavage rates to
that of PITC. Because of the quaternary group of the new reagent, it has a very polar
nature. This reagent is compatible with use only in a solid phase peptide sequenator. The
recently developed methods for solid phase sequencing has reduced the sample
requirement to picomole or subpicomole quantities (31-34). The advantages of very low
detection limits may not be obtained successfully with solid phase peptide sequenators.

Recently Basic et al. (3 5) introduced another reagent for Edman sequencing
followed by ESI/MS detection. This reagent features a N, Nedimethyl aminopropyl group

attached to the isothiocyanate firnctionality.
III. Other Modifications to Detect Thiohydantoin Amino Acids

A Fluorescent and Colored Derivatives

PITC is by far the most important reagent for N-terminal sequential degradation,
but the detection of PTH amino acids at low levels is difficult. Several fluorescent
reagents were developed to increase the detectability of TH amino acids. On the whole
the claimed advantages, mainly the ease of detection of the TH derivatives, are
outweighed by disadvantages such as low solubility in the coupling buffer, low reaction
yields, chemical stability and background contamination. Therefore, these reagents failed
to translate improved detectability into improved sequencing sensitivity. In this section,

some important reagents are briefly discussed.

104
l.'DABIT C (dimethylaminoazobenzene-isothiocyanate) Methods

The sequencing reagent DABITC has been widely applied for manual liquid and
solid phase rnicrosequencing (3 6). The reagent was described by Chang and Creaser in
1976 (3 7) and has several advantages over the usual degradation with PITC; firstly, it
releases the thiohydantoins as red-colored derivatives which are visible to the eye in
picomole quantities on polyarnide sheets; and secondly, the side product of the Edman
chemistry forms blue-colored compounds whose color easily difi’erentiates them fiom the
red color thiohydantoin amino acids. '

However, the coupling yield is lower than that of the PITC, due to the bulky side
group of the DABITC reagent. Therefore, a second coupling step with the PITC was
introduced to complete the reaction (38). Since the introduction of this DABITC/PITC
double coupling procedure this method has found wide application. The identification of
TH amino acids can most quickly and simply be performed on a stamp-sized polyamide
thin-layer sheet (3 9). HPLC systems for additional confirmation of TH amino acids and
quatitation have been developed (40-42).

2. Dansylamino-PITC Method

Jin et al. (43) introduced a new Edman reagent, Dansylamino-PITC, having the
dansylarnino functionality. The dansylarnino pm amino acids were detected at the 1-5
pmol level on TLC and at the 200-fmol level by HPLC (44). The dansylamino-PITC
carries a bulky naphthyl ring and gives a coupling yield about 75-95% (44). However, the
coupling with dansylamino-PITC is not complete even if the coupling reaction is
performed for 60 min at 52°C (44). The increase in coupling time causes an increase in
the by-products. Hirano and Wittmann-Liebold (45) improved the method by
incorporating a PITC coupling step after the coupling with dansylamino-PITC, but before

105
the cleavage. The resulting dansylamino-PTH amino acids were detected on TLC at 100-

pmol level.

B. Radioactive Derivatives

The use of a radioactively labeled coupling reagent in the Edman degradation was
first reported by Cherbuliez (47) and by Laver (48) who used [358] phenylisothiocyanate
in the end group determination of various proteins. Although a detection limit of 0.1nmol
was obtained, sequential analysis of protein was not attempted. Laursen (49), in 1969,
reported the use of 3H-labeled phenylisothiocyanate in a solid-phase sequenator. [14C]
PITC was introduced by Silver and Hood (50) in conjunction with an automated
sequenator. Jacobs and Niall (51), used 35S-labeled PITC in- an automated peptide
sequenator. The thiohydantoin amino acids which were generated at each cycle of the
degradation were identified by TLC and autoradiography at the lO—pmol level. Tsugita et
al. (52), described a method to form a radioactive derivative fi'om the intermediate
obtained from the Edman degradation instead of the usual conversion to the more stable
PTH amino acid. The ATZ amino acid was quantitatively reacted with
[12511iodohistatnine to form a sensitized phenylthiocarbamyl derivative. The derivative
was detectable at the finol level. Because the ATZ amino acid is used for the reaction, the
kinetics of the reaction were not affected by this derivatization. Although these reagents
improve the sensitivity compared with conventional PITC, radioactively labeled reagents
undergo autoradiodegradation, resulting in decreasing product yields and increasing
proportions of labeled by-products (52). Handling of highly radioactive materials is not

convenient for routine analysis.

106

>~H~=~~Cs

4-N, Ndimethylaminoazobenzene-4'-isothiocyanate (DABITC)

N

©©
son‘s.

4-(N-1-dimethylarrinonaphthalene-5—sdponylamino) phenylisothiocyanate

Figure 3-2. Structures of proposed fluorescent reagents

107

IV. Peptide and Protein Sequencing by Automated Sequenator

The manual peptide sequencing by Edman degradation is a time and labor
consuming process. In 1967, Edman and Begg (53) introduced the automated peptide
sequenator based on Edman chemistry to improve the peptide sequencing methodology.
Their sequenator was based on liquid-phase mobilization of peptide and called the
spinning cup sequenator. In the spinning cup sequenator, short peptides tend to be
washed out during the solvent extraction steps. Several research groups have improved
this technique in order to overcome this difficulty. Another two types of sequenators,
solid-phase sequenator (54,55), and gas-phase sequenator (56-58) were developed to
reduce the possibility of loses due to the extraction solvents. The intention of this part is
to provide a brief introduction of the development of automated peptide sequencing and
current instrumentation available for this technique.

Advantages of the liquid-phase sequenator includes the increased speed of
sequencing compared to that of manual methods, a barrier for atmospheric 02 that
decreases by-product formation, the effective extraction steps, and the improved

sensitivity by precise extraction procedures and miniaturizing the reaction vessel.
A. Liquid Phase Peptide (Spinning Cup) Sequenator

The liquid-phase automated peptide sequenators are widely used in determining
peptide sequence fiom the N-terminal of a peptide or protein by Edman degradation. A
short account is given here on liquid-phase peptide sequencing. The description of
automated methods were reviewed in detail by Nlall (51).

. 108
A Edman and Begg (53) introduced the first automated peptide sequenator to
perform peptide sequencing. This automated sequenator was able to carry out the
coupling and cleavage steps independently as instructed by a computer.

The spinning cup sequenator has encountered great difficulty making the required
discrimination with short peptides, particularly hydrophobic ones, because the extraction
solvents in combination with residual sequencing reagents cause the peptide to wash out
of the cup. In order to overcome this problem, a volatile bufi‘er (trimethylamine) was
introduce (59) to reduce the by-products and extensive extractions. Subsequently, several
protein carriers were introduced (60, 61, 62). The major breakthrough came with the
introduction of polybrene to spinning cup technology (61, 62). This substance efl‘ectiVely
anchors small quantities of both proteins and peptides in the cup and allows sequencing of
even short hydrophobic peptides to completion.

Hunkapillar et al. (63) combined two important developments in sequencing
chemistry, polybrene for non-covalent attachment and HPLC for PTH separation followed
by a UV-Vls detector for identification. The instrumentation was fidrther improved by

introducing a miniaturized glass disk as a sample support in place of a spinning cup (64).

B-. Solid Phase Peptide Sequenator

The solid phase approach to sequence analysis of proteins was introduced by
Laursen in 1971 (54). The essence of the procedure is that proteins and peptides are
covalently attached to an insoluble support. This is then placed in a glass column and the
reagents for Edman degradation are introduced sequentially by an arrangement of pumps
and valves as instructed by an electro-mechanical or electronic control device. A major
advantage of this approach is that because the protein is covalently attached, by-products
and excess reagents can be removed by solvent washes without concomitant risk of loss of

sample. Several. resins have been developed to attach proteins in solid phase sequencing

. 109
to improve extraction and sequencing emciency. They include porous glass beads,
polystyrene resins, glass fiber sheets with activated isothiocyanate surface (65 -67), and
polyvinyldene difluoride (PVDF) membranes (68).

C. Gas Phase Peptide Sequenator

In 1984, Hewick et al. (69) introduced a new sequenator which employed the
coupling and cleavage reagents in gaseous phase rather than liquid phases. Only sufiicient
PITC solution to completely wet the glass fiber disc is delivered (~20uL) to the sample
holder. The heptane (solvent used to dissolve PITC) is removed by briefly flushing the
cartridge with argon. The coupling stage is effected by slowly bleeding TMA/HZO vapor
through the cartridge to the waste. Likewise, cleavage is effected by slowly bleeding
trifluoroacetic acid vapor through the cartridge to the waste. The advantages of the new
miniaturized gas-liquid sequenator includes: high sensitivity sequencing, very low reagent
and solvent consumption, a much shorter degradative cycle, and low maintenance and

running costs (69).

V. Mass Spectrometric Methods to Determine the Sequence of Peptides and
Proteins

Mass spectrometric methods are gaining wide acceptance in peptide sequencing in
recent years. The Edman degradation chemistry cannot be used when the N-terminus of
the peptide is blocked. Mass spectrometry plays a significant roll in sequencing such
peptides. In this part, three major mass spectrometric techniques to get sequence
information will be discussed.

l 10
A. Fast Atom Bombardment Mass Spectrometry

The development of fast atom bombardment (FAB) ionization technique along
with mass spectrometry provides a method of ionizing non-volatile samples. This
development created growth in peptide sequencing using mass spectrometry (70).

FAB/MS was used to obtain molecular weight information (71-74) and tandem
mass spectrometry (MS/MS) were used to obtain fragrnentations of selected peptide ions.
The mass spectrometry is a very useful technique to get peptide sequence information of
N-terminal blocked peptides which cannot be sequenced by Edman degradation. During
FAB/MS, energy deposited on a peptide molecule is usually not enough to induce
complete fiagmentation pattern necessary to deduce the sequence. Collisionally induced
dissociation has been used to enhance the fi'agmentation and it provides several
advantages: all fiagment ion peaks are directly related to the peptide structure or the
selected ion, matrix peaks are excluded fi'om the spectrum, and mixtures can be analyzed
without purification.

The amino acid sequence can be deduced by interpreting the fi'agment ion peaks
formed by the (M-i-H)+ of a peptide. There are three different bonds that can fragment
along the backbone of a peptide. The first nomenclature for peptide fragmentation (Figure
3-3) was proposed by Roepstorfl‘ and Fohlmann (75). The letters A, B, and C denote the
N-terminal fiagment (ions that contain N-terminus) while X, Y, and Z represent carboxyl
terminal fragments (ions that contain C-terminus). The assumed structures are shown in
Figure 3-4. Since then, the nomenclature has undergone revisions (76). Biemann used
lower case letters to distinguish from the single letter code amino acids, irnmonium ions,
or the loss of side chain residues from (M+H)+ ions.

There are two types of strategies are used to induce fiagmentation, low-energy
and high-energy collisionally induced dissociation. This distinction is based on the kinetic

. 1 11

energies of the selected beam, being either in the order of 50V for quadrupole instruments,
or several kilo-volts in the sector instruments- The low-energy experiment deposited
lower internal energy (estimated to be 2-3 eV) deposited into the selected beam, thereby,
resulting in the induction of the fi'agmentation in the peptide backbone. In the high-energy
experiments (estimated to be lO-l 5eV deposited), fragmentation in amino acid side chains
and other moieties, and an array of irnmonium ions were observed in addition to the
peptide backbone fiagmentation (77).

Derivatization schemes have been developed to increase the fragmentation
efl'lciency and to control the fragmentation patterns during ionization. There are two
types of derivatives have been prepared for this purpose, charged or hydrophorbic
uncharged derivatives. A permanent charge has been deposited on the peptide molecule
to obtain complete sequence information. The derivatization reactions include
ethyltriphenylphosphonium reagents for N- and C-termini (7 8-80), and a quaternary
ammonium moieties for N terminus (81-83).

An on-line HPLC technique called "continuous flow FAB/MS" has been developed
for FAB mass spectrometric analysis (84, 85). Continuous flow FAB/MS has two
advantages with respect to the conventional FAB, suppression of matrix background ions
(86) and reduction of competitive ionization that is due to hydrophobicity difl‘erences in

peptide mixtures (87).

B. Electrospray Mass Spectrometry

Tandem mass spectrometry combined with ESI has proven to be a usefill technique
to obtain sequence data of proteins and peptides. Several research groups discussed
tandem mass spectral data obtained from low-energy CID studies (88). High-energy CID
studies were also reported (89).

112

 

 

 

 

 

 

 

 

 

X3 Y3 Z. X2 Y2 Z. X. Y. Z
[— F — '— — — T F— i—
i‘ ‘3. It“ u i3 ‘3. i‘ ii
HgN—C--C--N--C- C- III-'04 C--N—-C-C—OH
H H H H H H H
HERRERA—15.5...

Figure 3-3. Fragmentation of the peptide backbone and nomenclature proposed by
Roepstorfi‘ and Fohlmann.

113

N —terminal Ions

+
0,: H—(NH—CHR—C0).- t—NH=CHR.
or

H. CR;R:

H—ENH—CHR—coi‘.-.—NH—CH

 

H. R,
a, + l: H—lNH-CHR—COJPt—NH—CH-

b,: H—(NH—CHR—CO).-.—NH—CHR.—-C-=—O‘

 

H.
c,: H-TNH—CHR—CO)‘.—NH:

 

a
H' CR.

. r— 1 u
d,. H—(NH—CHR—CO).-t—NH-—CH

 

C-terminal Ion Types
l H.
0,: HN=CH—CO—(NH—CHR—CO).-.—OH
CR... H’
0,: CH—CO—lNH—CHR—COJM—OH
x,: ‘OEC—NH—CHR.—CO—(NH—CHR——CO)..-t—OH
or
H.
0=c=N—'CHR.—co—(NH—CHR—CO)‘.-.—0H
H.
y,: H-(‘FH—CHR—coil—on

 

 

 

 

 

"O
y. — 2: 'HN=CR,—co—(NH—CHR—c0)‘.-.—0H
can: H.

 

2,: CH—CO—lNH—CHR—COIH-OH

HO
2, + l: -CHR,—co-+'NH—CHR—c0i‘.-.—0H

 

Figure 3-4. Structures of the common ions encountered in FAB-CAD-MS/MS and their
nomenclature as revised by Biemann.

l 14

Fragmentation of the peptide backbone was induced by increasing the skimmer
voltage of an ESI source on a single quadrupole mass spectrometer (90). Another method
to induce the fi’agmentation on the peptide backbone involves increasing the temperature
of the metal capillary interface (91). The authors induced thermal fiagmentation of large
multiply charged ions (melittin and myoglobin). The spectra were comprised of multiple
charged fiagrnent ions.

Microseparation techniques are always very valuable in the biological sample
purification and analysis. On-line separation techniques with the mass spectrometry
provide the sensitivity and speed necessary to characterize peptides and proteins.
HPLC/ESI/MS/MS has been used to sequence peptide mixtures (tryptic digests).
Numerous examples were given to illustrate the use of mass spectrometry to sequence
blocked peptides, define N- and C- terminal heterogeneity, locate and correct errors in
DNA- and cDNA- deduced protein sequences, and identify sites of modifications such as
deamidation, isoaspartyl formation, phosphorylation, oxidation, disulfide bond formation,
and glycosylation (88, 92).

C. Matrix Assisted Laser Desorption Mass Spectrometry

Matrix assisted laser desorption ionization (MALDI) mass spectrometry is a soft
ionization technique that can be used to determine the accurate molecular weights of
peptides and proteins. Because of the softness of the technique, only singly charged or
doubly charged ions are frequently observed in the MALDI mass spectra of proteins.
Structural characterization of proteins cannot be done on MALDI due to the lack of
fragmentation. A tandem time-of-flight mass spectrometer is still under development (93-
95).

l 15
Spengler et al. (96, 97) studied the metastable decay of laser-desorbed ions in the
first field-flee drift region of a reflectron time-of-flight mass spectrometer. Fragment ions
fiom metastable decay were mass analyzed and a complete structural analysis has been
performed for several peptides and proteins. The postsource decay fi'agments of a protein
(M.W. = 12500) were compared with the ions generated by a tandem four sector mass
spectrometer (98).

1. Peptide Ladder by MALDI ‘

Chait and co-workers (99) have described a method to sequence a protein by using
multiple steps of wet chemistry with a final, single-step mass spectrometric read out of the
amino acid sequence. A sequence defining a concatenated set of peptide fragments, each
difi‘ering from the next by a single residue, is chemically generated in a controlled fashion.
This was achieved by carrying out rapid step-wise Edman degradation in the presence of a
small amount of terminating reagent, phenyl thiocyanate. A small proportion of peptide
chain blocked at the amino terminus is generated at each cycle. A predetermined number
of cycles is performed without intermediate separation or analysis of the released amino
acid derivatives. The resulting mixture is read out in a single operation by matrix-assisted
laser desorption mass spectrometry (MALDI). The mass spectrum contains molecule ions
corresponding to each terminated polypeptide species present. The mass differences
between consecutive peaks each correspond to an amino acid residue, and their order of
occurrence in the data set defines the sequence of amino acids in the original peptide
chain.

116

VI. List of References

10.

11.

12.

13.

14.

15.

16.

17.

Inglis, A. S. Anal. Biochem. 1991, 195, 183

Bailey, J. M.; Nikfarjam, F.; Shenoy, N. R; Shively, J. E. Protein Science, 1992,
1, 1622

Edman, P., Acta. Chem. Scand. 1956, 10: 761,

J6mvall, H.; HOOg, J .-O.; Gustavsson, A. -M. Methods in Protein Sequence
Analysis, Birkhuser, Basel, 1991

Burdon, R. H.; Kmippenberg, P. H. Van (Eds), Laboratory Techniques in
Biochemistry and Molecular Biology, 1989, Elsevier, New York, 295

Mische, S., et. al. (Eds), Techniques in Protoin Chonustry IV, 1993, 453
Calacay, J., Rusnak, M., and Shively, J.E., Anal. Biochem. 1991, 192: 23
Joe Leykam, Manager of MSU Peptide Facility, personal communication
Walker, J.E., Feamley, I.M., and Blows, R.A, J. Biochem. 1986, 237, 73

Hunkapiller, M., Kent, 8., Caruthers, M., Dreyer, W., Firca, J., Gimn, C .,
Horvath, S., Hunkapiller, T., Tempst, P., and Hood, L., Nature 1984, 310, 105

Vandekerckhove, J ., Bauw, G., Puype, M., Van Darnme, J ., and Van Montagu,
M., in onooooo Methodo in Protein Mimmoengo Analysio; B. Wittmann-
Liebold, J. Salnikov, and V. Erdmann, eds., Springer-Verlag, Berlin, 1986,179

Pisano, J. J.; VandenHeuvel, W. J. A.; Homing, E. C. Biochem. Biophys. Res.
Commun. 1962, 7, 82

Melvas, B. W. Acta. Chem. Scand 1969, 23 (No. 5) 1679

Fairwell, T.; Lovins, R. E. Biochem. Biophys. Res. Commun. 1971, 43 (No. 6),
1280

Sun, T.; Lovins, R. E. Anal. Biochem. 1972, 45, 176

Fales, H. M.; Nagai, Y.; Milne, W. A.; Brewer, H. B.; Bronzert, T. J. Jr.; Pisano,
J. J ;Anal. Biochem. 1971, 43, 288

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

Richards, F. F.; Barnes, W. T.; Lolviri7s, R E.; Salomone, R; Waterfeild, M. D.
Nature 1969, 221, 1241

Milane, G. W. A; Axenrod, T.; Fales, H. M.; J. Am. Chem. Soc. 1970, 97, 5170
Pisano J. J, Bronzert T. J J. Biol. Chem. 1969, 244, 5597

Engelke, F.; Hahn,J. H.; Henke, W.; Zare ,R NAnal. Chem. 1987, 59, 909

Pramanik, B. C.; Hinton, S. M.; Millington, D. S.; Dourdeville, T. A; Slaughter,
C. A. Anal. Biochem. 1988, 175, 305

Pramanik, B. C.; Moomaw, C. R; Evans, C. T.; Cohen, S. A; Slaughter, C. A.
Anal. Biochem. 1989, 176, 269

Sakari. M.; Kambara. H. Anal. Chem. 1988, 60, 774
Lequin, R. M.; Niall, H. D. Biochem. Biophys. Acta. 1972, 252, 76

Tomqvist, M.; Mowrer, J .; Jensen, S.; Ehrenberg, L Anal. Biochem. 1986, 154,
255

Gosh, A; Andregg, R. Proceedings of the 37th ASMS Conference on Mass
Spectrometry and Allied Topics, Miami Beach, FL, May 21 - 26, 1989, 917

Waidanatha, S.; Andregg, R. J. Proceedings of the 37th ASMS Conference on
Mass Spectrometry and Allied Topics, Miami Beach, FL, May 21- 26, 1989, 881

Tarr, G ”Methods of Protein Microcharacterization" Ed: Shiverly, J 1986
Humana Press, Clifton, N. J. pp 156

Aebersold, R; Bures, E. J.; Namchuk, M.; Goghari, M. H.; Shusan, B.; Covey, T.
C. Protein Science 1992, I, 494-503

Aebersold, R; Pipes, G.; Nika, H.; Hood, L. E.; Kent, S. B. H., Biochemistry

' 1988, 27, 6860

Pappin, D. J. C.; Coull, J. M.; Koester, H. In Current Research in Protein
Chemistry (Villafranca, J. ;Ed) pp 191-202 Academic Press Sandiego

Aebersold, R; Pipes, G.; Wettenhall, R. E. H.; Nika, H.; Hood, L. E. Anal.
Biochem. 1990 187, 56

Coull, J. M.; Papin, D. J. C.; Mark. J.; Aebersold, R;Koester, H. Anal. Biochem.
1991 I94, 110

35.

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

47.

48.

49.

50.

51.

52.

53.

1 18
Basic, C.; Bailey, J. M.; Shively, J. E.; Lee, T. D. Proceedings of the 41st ASMS
Conference on Mass Spectrometry and Allied Topics, San Francisco, CA, May 31-
June 4, 1993

“frttman-Liebold, B.; Kimura, M. 1984 in Methods in Molecular Biology, vol 1,
Walker, J. M. ed; Humana press, Clifton, NJ, pp 221

Chang, J. Y.; Creaser, E. H.; Bently, K. W. Biochem. J. 1976, 153, 607

Chang, J. Y.; Brauer, D,; Wittman-Liebold, B.; FEBS Lett. 1978, 93, 205
Chang, J. Y.; Creaser, E. H. J. Chromtogr. 1977, 132, 303

Lehmann, A; Wittman-Liebold, B. FEBS Lett. 1984, 176, 360

Chang, J. Y.; Lehmann, A; Wittman-Liebold, B. Anal. Biochem. 1980, 102, 380
Prieto, J. A,; Collar, C.; de Barber, C. B. J. Chromatogr. Sci. 1991, 29, 205
Tsuga, A.;Kamo, M.; Jone, C.; Shikarna, N. J. Biochem. 1989, 106 60

Jill, S-W.; Chen, G-X.; Palacz, Z.; Wittmann-Liebold, B. FEBS Lett. 1986, 198,
150'

Salnikov, J .;Palacz, Z.; Wittmann-Liebold, B. in Methods in Protein Sequence
Analysis, (Walsh ed;), Humana Press, Crifion, NJ, 247

Hirano, H.; Wittmann-Liebold, B.; Tsukuba, 1.; in Methods in Protein Sequence
Analysis, (VVlttmann-Liebold, B ed;), Springer-Verlag Berlin, 1989, 42

Cherbuliez, E.; Sussman, A. R; Rabinowitz, J. Helv. Chem. Acta. 1966, 44, 319
Laver, W. G.; Biochem. Biophys. Acta. 1961, 53, 469

Laursen, R. A. Eur. J. Biochem. 1971, 20, 89

Silver, J.; Hood, L. Anal. Biochem. 1972, 67, 392

Jacobs, J. W.; Niall, H. D. J. Biol. Chem. 1975, 250, 3629

Tsuga, A; Arai, I.; Kamo, M.; Jone, C. J. Biochem. 1988, 103, 399

Edman, P.; Begg, G. EuropeanJ. Biochem. 1967, 1, 80

54.

55.

56.

57.

58.

59.

61.

62.

63.

65.

68.

69.

70.

71.

72.

73.

119
Laursen, R A. Eur. J. Biochem. 1971, 20, 89—102

Laursen, R. A; Bonner, A. G.; Horn, M. J. 1975 in Instrumentation in Amino Acid
Sequence Detamination (Perharn, R. N. ed.) pp73-110, Academic Press, London.

Hewick, R M.; Hunkapillar, M. W.; Hood, L. E.; Dreyer, W. J. J. Biol. Chem.
1981, 256, 7990 '

Hunkpiller, M. W.; Lujan, E.; Ostrander,F.; Hood, L. E. Methods Enzymol. 1983,
91, 227

Hunkpiller, M. W.; Hewick, R. M.; Dreyer, W. J .; Hood, L. E. Methods Enzymol.
1983, 91, 399

Hewick, R. M.; Hunkpiller, M. W.; Hood, L. E.; Dreyer, W. J., J. Biol. Chem.
1981, 256, 7990 ~

Tarr, G. E.; Beecher, J. E; Bell, M.; McKean, D. J. Anal. Biochem. 1978, 84, 622
Klapper, D. G.; Wilde, C. E.; Capra, J. D. Anal. Biochem. 1978, 85, 126
Hunkapillar, M. W.; Hood, L. E. Biochemistry, 1978, 17, 2124

Shively, J. E.; Miller, P.; Ronk, M. Anal. Biochem. 1987, 163, 517

Aebersold, R; Pipes, G.; Nika, H.; Hood, L. E.; Kent, S. B. H. Biochemistry,
1988, 27, 6860

Pappin, D. J. C.; Coull, J.; Koester, H. Anal. Biochem. 1990, 187, 10

Hewick, R M.; Hunkapillar, M. W.; Hood, L. E.; Dreyer, W. J. J. Biol. Chem.
1981, 256, 7990

Cotrell, J. S.; Greathead, R. J. Mass Spectrom. Rev. 1986, 5, 215

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

Barber, M.; Bordoli, R. S.; Sedgewick, R. D.; Tyler, A. N. Nature, 1981, 293, 270

Barber, M.; Bordoli, R. S.; Sedgewick, R. D.; Tyler, A. N.; Whalley, E. T.
Biomed Mass Spectrom. 1981, 8, 337

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

85.

86.

87.

88.

89.

91.

Morris, H. R; Panico. M.; Barber, MioBordoli, R S.; Sedgewick, R D.; Tyler, A.
N. Biochem. Biophys. Res. Commun. 1981, 101, 623

Roepstofi‘, P.; Fohlman, J. Biomed Mass Spectrom. 1984, 11, 601

Biemann, K. Biomed Environ. Mass Spectrom. 1988, 16, 99

Burlingarne, A. L.; Baille, T. A; Russell, D. H. Anal. Chem. 1992, 64, 467R

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

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

Wagner, D. S.; Allison, J. Fragmentation of Peptides Cationed with Metal Ions;
Proceedings of the 40th ASMS Conference on Mass Spectrometry and Allied
Topics, Washington DC, May 31-June 5, 1992

Kidwell, D. A; Ross, M. M.; Colton, R. J. .I. Am. Chem. Soc. 1984, 106, 2219
Renner, D.; Spiteller, G. Biomed Environ. Mass Spectrom. 1986, 13, 405

Stults, J. T.; Lai, J .; McCune, S.; Wetzel, R. Anal. Chem. 1993, 65, 1703

Ito, Y.; Takeuchi, T.; Ishii, D.; Goto, M. J. Chromatogr. 1985, 346, 161

Caprioli, R. M.; Fan, T.; Cottrell, J. S.'Anal Chem. 1986, 58, 2949

Caprioli, R M.; Fan, T. Biochem. Biophys. Res. Commun. 1986, 141, 1058

Caprioli, R. M.; Moore, W. T.; Fan, T. Rapid Commun. Mass Spectrom. 1987, 1,
15 '

Carr, S. A; Hemling, M. E.; Bean, M. F.; Roberts, G. D. Anal Chem. 1991, 63,
2802

Murata, H.; Takao, T.; Shimonishi, Y.; Rapid Commun. Mass Spectrom. 1994, 8,
205

Katta, V.; Chodhury, S. K.; Chait, B. T. Anal. Chem. 1991, 63, 174

Rockwood, A L.; Busman, M.; Udseth, H. R; Smith, R. D. Rapid Commun.
Mass Spectrom. 1991, 5, 582

92.

93.

94.

95.

96.

97.

98.

121
Michael, H.; Dickinson, T. A; Shabanowitz, J.; Davis, T. A; Cox, A. L.; Hunt, D.
F. Proceedings of the 40th ASMS Conference on Mass Spectrometry and Allied
Topics, Washington DC, May 31- June 5, 1992

Seeterlin, M. A; Vlasak, P. R; Beussmarl, D. J.; McLane, R. D.; Enke, C. G. .1.
Am. Soc. Mass Spectrom. 1993, 4, 751

Cornish, T. J .; Cotter, R. J. Rapid Commun. Mass Spectrom. 1993, 7, 1037
Gardner, 13.; Enke, c. 6.; Watson, 1. T. Proceedings of the 44th ASMS
Conference on Mass Spectrometry and Allied Topics, Chicago, IL , May 29-June
3, 1994

Spengler, B.; Kirsch, D.; Kaufrnann, R; Rapid Commun. Mass Spectrom. 1992, 6,
105

Spengler, 13.; Kirsch, D.; Kaufmanrl, R; J. Phys. Chem. 1992, 96, 9678

Spengler, B.; Kirsch, D.; Kaufrnann, R; Rapid Commun. Mass Spectrom. 1993, 7,
902

Chait, B. T.; Wang, R; Beavis, R. C.; Kent, S. B. H. Science, 1993, 262, 89

CHAPTER 4
Development of Edman Reagents for the ESI/MS Detection

I. INTRODUCTION

Protein sequencing is a highly sensitive technique that has been invaluable for
providing critical primary structural data on isolated proteins. The isolation of low
picomolar amounts of purified proteins and peptides has become a common occurrence,
due largely to substantial advances in the techniques available for sample purification. As
a result, there is an increased need for routine protein sequence analysis at low picomolar
levels. The Edman sequencing chemistry remains the most popular method for the protein
and peptide sequencing (1). During the Edman sequencing, the peptide is cleaved one
amino at a time and the resulting thiohydantoin (TH) amino acids are detected by a UV
detector after the separation by HPLC (see Figure 3-1). The identification of TH amino
acids is done solely by using the retention time on the HPLC chromatogram. Recently,
the Association of Biomolecular Resource Facilities (ABRF) Sequencing Committee
evaluated the sequencing reliability among 75 facilities using an identical peptide
containing two modified amino acids. The report says that identification of the post-
translationally modified amino acids represents a challenge in sequence analysis (2).

Mass spectrometry can be used to identify the TH amino acids by molecular
weight that is more specific for the compound, rather than retention time. The previous
attempts to use El (3), CI (4) and thermospray (TS) (5) mass spectrometry for the
determination of thiohydantoin amino acid were not successful. These mass spectrometric
techniques required a fairly high level of samples. Under these mass spectrometric

conditions, the PTH amino acids undergo extensive fragrnentations resulting low

122

123

Aebersold et al. (6) described a modified reagent with very high sensitivity but due
to its high polarity, a solid phase sequanator has to be used. Basic et al. (7) characterized
a series of alkyl amine thiohydantoin derivatives by ESI/MS. Figures 4-lc and (1 show the
structures of the reagents proposed by Aebersold et al (Figure 4-1a) and Basic et al.
(Figure 4-1b) for ESI/MS.

In order to improve the response of TH amino acids towards ESI mass
spectrometry, a new class of modified TH amino acids was synthesized and characterized
by mass spectrometry. The detectability of the modified amino acids was assessed on
ESI/MS. '

11. Alternative Edman Reagents

The PTH amino acids were analyzed by ESI/MS. The detectability of PTH
derivatives was poor in ESI/MS except for the amino acids containing basic side groups
such as arginine and histidine. In order to detect all the amino acids at the same
sensitivity, another fimctional group that is readily amenable to ESI/MS, must be
introduced. In this context, three reagents were investigated. One reagent is t-BOC
aminomethylphenylisothiocyanate (BAMPITC) that was introduced to improve the
detectability of TH amino acids by. forming fluorescent derivatives after the HPLC
separation. Another two reagents were proposed and their properties were investigated.

The TH derivatives were characterized by mass spectrometry.

A. t-BOC aminomethylphenylisothiocyanate (BAMPITC)
1. Introduction
Reagent I, t-BOC aminomethylphenylisothiocyanate (BAMPITC) (Figure 4-2a)
was developed for post column derivatization of the modified PTH amino acids for

sensitive detection (8). This reagent carries a primary amino group blocked by a t-BOC

124
group and during the acid cleavage step of the sequencing process, a fi'ee primary anrino
group is generated (Figure 4-2b). In the original process, the free amino group was
subjected to fluorescent labeling at the time of detection to improve the detectability of the
modified thiohydantoin amino acids. It was reported that the coupling rate is comparable
to that of PITC. The authors claimed the repetitive yield is over 90% for the automated

solid phase sequencing.
i
/ C
N \CH —R
s/C—NH
(a). Phenylisothiocyanate (PITC) (b). Phenylthiohydantoin (TH) amino acid

/./NCS
(CH3)3N+—CH2CH2 :

(c). 3—(4'ethy1ene-N,N,N-trimethylamino)pheny1-2-isothiocyanate

H3
/N —CH2 —CH2 —CH2 -NCS
H3C

(d). Dirnethylaminopropylisothiocyanate

Figure 4-1. The structures of conventional Edman reagent (a), phenylthiohydantoin
derivative (b), reagents proposed by Aebersold et al. (c), and Basic et al. (d).

N—C — S
i
(CH3)3C— o— C-NH— c 2

(a). t-BOC aminomethylphenylisothiocyante

i
/C\
H—R
H2N- CH21\ f

/—NH

S
(b). Aminomethylphenylthiohydantoin amino acid

Figure 4-2. Structures of t-BAMPITC and aminomethylphenylthiohydantoin amino acid

2. Experimental
The TH amino acids were synthesized using the BAMPITC reagent in

pyridine/water (1 : 1) and the pH was adjusted to 9 with triethylamine. The reaction
mixture was incubated at 50°C for 30 minutes. The reaction mixture was evaporated to
dryness and excess reagent and the by-products were extracted with 2-mL portions of
toluene three times. The residue was dried and incubated at 50°C for 10 minutes with
anhydrous TFA This cleaves the t-BOC group from the amino terminus as well as forms
the unstable anilinothioazolinine (ATZ) amino acid. The ATZ-amino acid is converted to
the stable TH amino acid with aqueous TFA (25%).

126

Analysis by ESI/MS was undertaken on the TSQ 700 triple quadrupole mass
spectrometer fitted with a Finnigan ESI source. The sample was introduced at a flow rate
of 4uL/min. The AMTH derivatives were analyzed by ESI/MS using
CH3OH/HgO/HOAC (50/49.9/0.1 %v/v) as the solvent system. In all the experiments,
needle was held at 4.5kV and the heated capillary at 200°C.

3. Results and Discussion

Since the reagent generates a free amino group during the acid cleavage step
(Figure 4-2b), it will improve the sensitivity of ESI/MS. Therefore, the detectability of
modified TH amino acids were investigated using ESI/MS. The t-BOC group on the N
atom cleaves when treated with anhydrous TFA and the cleavage is essentially complete in
several minutes (9). Due to the free amino group on the TH derivative afier the acid
cleavage, the AMTH amino acids could be observed as the protonated molecules. The
TH derivatives of lys and arg were observed as doubly charged ions in the ESI/MS
spectrum. All the other TH derivatives were observed as singly charged species. All the
AMTH amino acids could be detected at the low femtomol level using ESI/MS. The TH
amino acids with acidic residues are expected to give a low response in positive ESI/MS
due to the -COOH group in the side chain because the -COOH group of the TH amino
acid can be deprotonated in solution to form an electrically neutral species. But the asp-
AMTH was observed at 200 finol level.

Wittmann-Liebold et al. (9) discussed the synthesis and use of BAMPITC in
microsequencing in detail. They compared the efficiency of manual sequencing and
concluded that the repetitive yield was similar to that of PITC, but it needed longer

reaction times. In the solid phase sequencer, the yield was about 48% for the second cycle

127
of the insulin B chain (oxidized). Since the reported repetitive yield was low for the

automated sequencer, two new reagents were synthesized and characterized by mass
spectrometry.

B. Designing New Reagents

In order to improve the ESI/MS response of TH amino acids, a basic or a pre-
formed charged group must be introduced into the reagent. Aebersold et al. introduced a
quartenary ammonium group to the PITC reagent. The reagent cannot be used in the
liquid phase sequanator due to its high polarity. The work described here uses a pyridyl
group as a basic residue to improve the ESI/MS response (Figure 4-3 a and b). Two new
reagents were synthesized and characterized. The isothiocyanate group was separated
hour the pyridyl ring in one reagent. Because of the pyridyl group, they are suitable for
the liquid phase sequanator and the aromatic properties desired for the Edman reagent are
retained.

@N=C=S gab—N=C=S

N N

(a). Pyridylisothiocyanate (b). 3-Pyridylmethylisothiocyanate
Figure 4-3. The modified Edman reagents.

IH. Preparation of Isothiocyanate
Several methods have been reported for the synthesis of isothiocynates. Among
them, most reactions gave low reaction yields (25-7 5%) (10-13). Two methods have been

selected to synthesize modified isothiocynates using their corresponding amines (14, 15).

128
These two methods gave reasonable reaction yields (over 90%) and relatively simple
purification procedures.

A Method 1.

The method proposed by Makaiyama et al. was adapted (14) for the synthesis of
isothiocyanates fiom the corresponding amine. Pyridylisothiocyanate was synthesized
using pyridylarnine, C82, and 2-chloro-l-methylpyridinium iodide (Figure 4-4). In a
round bottom flask, cooled by an ice-salt bath, 1.1 mmol of C82 and 2.1 mmol of
triethylammine were placed. The mixture was stirred with a magnetic stirrer and
aminopyridine ( 1 mmol) was slowly added to the cooled reaction mixture. The stirring
was continued for thirty minutes after all the amine had been added, and then the reaction
mixture was allowed to stand for another thirty minute. During this time a heavy
precipitate of triethylammonuim pyridyldithiocarbarnate (Reaction 4-1) was formed.

A solution of triethylanrine (111 mg, 1.1 mmol) in dichlromethane (4mL) was
slowly added at room temperature under an argon atmosphere to a mixture of
triethylaminium pyridylditiocarbamate (304 mg, 1 mmol) and 2-chloro- l -methylpyridinium
iodide (281 mg, 1.1 mmol), and then the reaction mixture was stirred at room temperature
for an additional two hours. After the removal of solvent, the residue was separated by
the preparative thin layer chromatography or the column chromatography. The reaction
yield was not optimized.

3-pyridylmethy1isothiocyanate was prepared using 3-aminomethyl pyridine by the
same method.

B. Method 11
In this method, isothiocyanates were synthesized by reacting an amine with a
thiocarbanyl transfer reagent: di-2-pyridyl thiocarbonate (DPT). This method seems to be

1 29
simple and is easy to clean up after the reaction. In the first step, DPT was synthesized .
and purified (Figure 4—5). At the second step the amine was reacted with DPT in CH2C12
at room temperature to get corresponding isothiocyanate (Figure 4-6) (15).

1. Preparation of DPT.
In a round bottom flask surrounded by an ice bath, 2 mol of 2-hydroxypyridine and
2 mol of triethylamine were dissolved in 50 mL of CHzClz. 1 mol of thiophosgene was
slowly added to the reaction mixture with stirring under a hood. The reaction mixture was
stirred for another 2 hours at room temperature. The reaction mixture was neutralized
with. an anion exchange resin and filtered through a Buchner flame]. The residue was
washed with distilled water to remove the unreacted starting materials and triethylamine.

The residue was recrystalized with CH2C12.

2. Preparation of Pyridylisothiocyanate using DPT

3-aminopylidine (0.1 mol) was dissolved in 25 mL of CH2C12, in a round bottom
flask using a magnetic stirrer. The thiocarbamyl transfer reagent, DPT (0.1 mol) was
added to the above mixture and stirred for about 2 hours. Figure 4-6 shows the reaction
scheme involved in synthesis of isothiocyanate using DPT. The 2-hydroxypyridine was
precipitated fi'om the reaction mixture by adding diethyl ether. About 80% of the 2-
hydroxypyridine could be recovered by this method. The reaction mixture was filtered and
dried over anhydrous Na2SO4 overnight. The filtrate was concentrated on a rotovap and

separated on a preparative thin layer chromatographic plate.

13o
R—NH2 + 052 + H31»: —» R—NHCSZ- l~:t3NH+

R—NHcsz' Et3NH+ + @\
c1

CH3
1 CH2CI2
©S — C — NHR + Et3NH+ Y.
x' II
I s
CH3
Et3NH+Y' + Q + R—N=C=S
S
xv=mml I

one

Figure 4-4. The reaction scheme of isothiocyanate formation by Makayarna Method

131
S
.. +2©1 —-» (,1in
/C\ "CH o/\

thiophogene hydroxypyfim di-Z-pyridyl thiocarbonate (DPT)

Figure 4-5. Reaction scheme for the synthesis of DPT

NH2
g @0/:\0/©

Pyridyl anine
DPT

lance
©n=c=s + 2 @on

Pyridyl lsotlnocyanate

Figure 4-6. Scheme of isothiocyanate formation using DPT isothiocyanate transfer

reagent.

132
C. Separation of Reaction Mixture
(i). Preparative thin layer Chromatography (PTLC)

The products synthesized by the methods 1 and H were analyzed by thin-layer
chromatography (TLC) prior to separation of the mixture by preparative thin-layer
chromatography (PTLC) or flash column chromatography (16). TLC separation was
performed with Whatman® (Whatman paper Ltd; Maidstone, Kent, England) pro-made
TLC plates that were coated with 250-um thickness of silica Gel and fluorescent indicator
on flexible aluminum plates. The plates were cut into 2.5 x lO-cm strips and spotted by a
micropipet with solution containing the product to be analyzed. The sample spot was 0.2
mm 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 about 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 under UV light. The
position of the migrated dr0plets appeared as purple spots on the plate. The positions of
the spots were marked and their Rf values were calculated. Several solvent combinations
were investigated for optimum separation.

The reaction mixture was applied onto a silica gel preparative thin layer
chromatographic (PTLC) plate with fluorescent indicator (2-mm thickness, S-um
particles, 20 x 20 cm) as a band at 1 cm from the bottom of the plate- It was kept under
hood for several minutes to evaporate the solvent. The PTLC plate was developed in a
closed tank with CHCl3/CH2C12 (50/50 v/v%) mixture. After the solvent front was 1 cm
away fi'om the top of the plate, removed fi'orn the tank, and air dried under the hood. The
plate was observed in a viewing chamber by fluorescence at 254 nm to identify the
separated bands. The bands were marked, cut and collected for extraction. The silica gel
fi'actions containing products were extracted by using CH2C12 as the solvent. The

133
presence of the reagent was confirmed by FAB-MS, EI-MS or the characteristic

colorimetric reaction with the sodium azide and the dimethylforrnamide (17).

2. Column Chromatography

Large scale preparative separations are traditionally carried out by tedious long
column chromatography. Although the results are somewhat satisfactory, the technique is
always time consuming and frequently gives poor recovery due to band tailing. In recent
years, several preparative systems have evolved which reduce separation times to 1-3 hr
and allow the resolution of components having 11sz 0.05 on analytical TLC (15, 17).
Recently Skill et al. developed a substantially faster technique for the routine purification
of reaction products and called the flash chromatography (16).

Flash chromatography is basically an air pressure driven hybrid of medium pressure
and short column chromatography that has been optimized for rapid separations. Optimal
separation by flash chromatography of a mixture is accomplished in a reasonable length of
time when the components have Rf value of about 0.35 and are separated by ARf value of
0.1.

Separation of the products was performed with flash chromatography. A 2-inch
diameter column with a stopcock and flow controller was chosen to carry out the
separation. A small plug of glass wool was placed in the tube connecting the stopcock to
the column body. A smooth 1/8-inch layer of 50-100 mesh sand is added to cover the
bottom of the column and dry 40-63 um diameter silica gel (Sigma Chem. Co.; St. Louis,
MO) was poured into the column in a single portion to give depth of about 6-inches. With
the stopcock open the column was gently tapped vertically on the bench top to pack the
silica gel. Next a 1/8-inch layer of sand was carefillly placed on the flat top of the dry gel
bed. The column was filled with the 50:50 (v/v) CH2C12 : CHC13 solvent mixture that

produced the TLC separation and pressure was applied to rapidly push all the air from the

134

silica gel. One gram of the reaction mixture dissolved in the 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 mixture 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 fi'actions were collected at the
start of the run and larger fractions were collected at the end of the run. Composition of
the collected fractions was determined by TLC and the fractions containing the same
compound were combined and concentrated by rotoevaporation. The presence of the
isothiocyanate was confirmed by EI-MS, FAB-MS, and a characteristic color change with
sodium azide and dimethylformarnide.

3. Results and Discussion

Pyridyl and pyridylmethyl isothiocyanates were prepared by both methods. In the
method 1, several reaction products were observed on the TLC plate. The Rfvalues of the
isothiocyanate and the nearest compound were 0.52 and 0.44 respectively for the PMITC
reaction mixture. On the PTLC plate, two bands were partially overlapped and bands
have to be selected carefillly.

About 80% of hydroxypyridine was precipitated using diethyl ether. This makes
the separation less complex owing to less reaction products. Three major spots on the

TLC plate were observed for the reaction mixture.

D. Characterization of Isothiocyanates
1. Colorimetric Detection
Mass spectrometric techniques can provide the molecular weight information and
fragrnentations necessary to identify a particular compound. Mass spectrometry is not
very good in distinguishing isomeric compounds. Therefore, a chemical method has to be

135
used to identify the correct firnctional group. The isothiocyanate (-NC 8) can exist in
another isomeric form as thiocyanate (-CN S). Therefore, a chemical method was used to
chemically characterize the isothiocyanate reagents. The isothiocyanate group of the
reagent reacts with sodium azide in the presence of dirnethylforrnamide to evolve C02 gas
and a blue (or green) colored solution (18). This is a characteristic chemical reaction for

the presence of -NCS filnctional group in the molecule.

2. Mass Spectrometric Detection
i. EI/MS

EI mass spectrometry was carried out on a PIP-5980 MSD mass spectrometer.
The samples were introduced through a gas-chromatographic inlet. The injector
temperature was held at 200°C and the transfer line at 220°C. The temperature of the
column was held at 100°C for 2 minutes and increased at a rate of 10°C/min to 250°C.
The EI mass spectrum of pyridylisothiocyanate (M.W. = 136 Daltons) shows the
molecular ion at m/z 136.

The EI mass spectrum of 3-pylidylmethylisothiocyanate (PMITC) shows a peak at
m/z 150 that corresponds to the molecular weight. An intense peak was observed at m/z
92 that represents a stable fragment C6H6N+ that has an aromatic character.

ii. FAB/MS
FAB mass spectrometric data were collected on a JOEL I-IX 110 double focusing
high resolution mass spectrometer having EB geometry. The samples were ionized by a
high energy Xe beam (6 keV). A portion of 2uL of sample was placed onto a FAB probe
tip containing a viscous matrix usually glycerol. The probe with sample was introduced
into the ion source chamber and evacuated for ionization. The ions were accelerated to 10

keV before going through the analyzers. The data were recorded on a DA 5000 data

136
system The reagent PyITC shows the protonated molecule at m/z 137 and 3-PMITC
shows an intense peak at m/z 151 in their FAB mass spectra representing (M+H)+ ion.

IV. Thiobydantoin amino acids

A. Thiohydantoin Synthesis

Representatives from each group of amino acids containing acidic, neutral, and
basic side chains were selected to synthesize the thiohydantoin derivatives of amino acids.
The following amino acids were selected as representative amino acids: asp, val, phe, leu,
ile, gly, ala, ser, pro, asn, and lys. An amino acid (2 mmol each) was dissolved in
H20/pyridine (1 : 1) mixture in two separate vials and the pH of the solutions was adjusted
to 9 with triethylamine. The reaction mixtures were flushed with N2 gas for about 10
seconds. The PyITC reagent (2%) in acetonitrile (25 uL) was added to one vial containing
the amino acid and coupling buffer. A portion of 25 uL of PMITC (2%) in acetonitrile
was added to the other vial. The two vials were incubated for 30 minutes at 50°C for
PyITC and at 60°C for PMITC. The solvent was evaporated in a Speedvac® and the
excess reagent and by products were extracted with toluene. The dried residue was
treated with anhydrous trifluoroacetic acid (TFA) to form anilinothioazolinone (ATZ)
derivatives. The ATZ amino acids were converted to the more stable thiohydantoin (TH)
amino acid by treating with 25% TFA at 50°C for 10 minutes. The resulting reaction
mixture was injected onto a pre-equilibrated reversed phase HPLC system for separation.
The TH derivatives were collected and dried under vacuum. The derivatives were stored

under nitrogen at -20°C until use.

B. Extraction of TH Derivatives
The extraction of TH amino acids fi'om the reaction mixture was evaluated using
FAB/MS. All the TH amino acids (PyTH and PMTH) were dissolved in ethyl acetate. In

l 37

the presence of a water layer, the derivatives were observed in both layers and F AB/MS
indicates that the solubility in water is greater than that in ethyl acetate.

Afier the cleavage of ACP-PyTC and the ACP-PMTC (lOOpmol) with anhydrous
TFA, dried residue was extracted with ethyl acetate and analyzed by ESI/MS. The ACP
peptide has the sequence of VQAADYING. The TH derivative of val was observed in the
extract. This indicates the possibility of extraction using ethyl acetate as the solvent. This
extraction solvent is compatible with the liquid phase peptide sequanator. Therefore, both

reagents are compatible with the liquid phase peptide sequenator.

C. Mass Spectral Analysis of Thiohydantoins

The PyTH and PMTH derivatives of amino acids were subjected to various mass
spectrometric analysis techniques. Amino acids with non-polar side chains were studied
by GC/MS using EI and CI mode of ionization. In the E1 mode, extensive fiagmentations
were observed for all the derivatives with a reasonably intense peak for the molecular ion.

In the CI mode, a major (M+H)+ was observed for all the non polar TH amino
acids. The fi'agmentation was minimal for the TH derivatives of amino acids containing
non-polar side chains. The derivatives with polar side chains were not analyzed by
GC/MS.

FAB/MS analysis was undertaken on the TH derivatives of both reagents. All the
TH amino acids showed intense peaks for (M+H)+. FAB/CAD/MS/MS studies showed
the fi'gmentations necessary to identify the TH amino acids in their mass spectra. Figure
4.7 shows the FAB/CAD/MS/MS mass spectrum of PMTH-val.

138

 

 

 

 

a b
. +§1
+H ‘1 I”
: ’1 IC
: /Sl’ :
@V “‘2' “-c a
r‘ ~-- -- ,r
O ’ /
I’l’ HC__CH3
5 1
Fl CH3 A0
9 .
I
l “
1
v C
e 31 a 207
Q as . b
u 2.
n . [M+H-Val]+
d 1 1
a .
n .
c 1
e .
0 100 150 200 “320

 

Figure 4-7. FAB/CAD/MS/MS spectrum of val-PMTH.

1. Detection by ESI/MS
The purified TH-amino acids were subjected to ESI/MS on a TSQ 700 TQMS
fitted with a Finnigan ESI source. The capillary was heated to 200°C and the ESI needle

was at 4kV. The nebulizing gas pressure was at 35 psi. The TH-amino acid sample was

139
introduced into the source through a flowing stream of flow at a rate of 100 ul/ min. The
liquid system used was a mixture of CH3OH/HzO/HOAC (50/49.9/0.1%v/v/v). The total

flow was directed to the ESI source without a splitter.

Under the above conditions, ESI/MS mass spectra of TH-amino acids gave single
peaks representing the (M+H)+- The basic nature of the pyridyl ring provides protonated
molecules in solution. The ions are converted into the gas phase during the ESI process.
The TH-amino acids did not show significant fi'agmentation under these conditions.
Figure 4-8 shows the ESI/MS mass spectrum of val-PyTH (5pmol). The amino acids
containing acidic, basic and neutral side chain groups show a similar response due to the
pyridyl group in the Edman reagent. The TH derivative of lys shows the attachment of an
additional PyNCS group to the side chain amino group, but it did not show a strong
doubly charged ion in ESI/MS. The lack of fiagmentation improves the sensitivity for
detecting the TH analogs by ESI/MS. This simplifies the detection technique for TH-
amino acids for peptide sequencing.

To evaluate the detection limits of the TH amino acids, increasing amounts of
purified phe-PyTH were injected to a stream of CH3OH/H20/HOAC (50/49.9/0. 1%
v/v/v) at a flow rate of lOOuL/min. Figure 4-9 shows the profile obtained for selected ion
monitoring of the ion at m/z 284 (M+H)+. This shows the detection limit of phe-PyTH at
the 5-finol level.

Since the PMTH derivatives contain a pyridyl ring external to the thiohydantoin
ring structure, it does not conjugate with the unsaturated functional groups in the TH ring.
The pyridyl ring is fi'eely available for protonation during the ESI/MS process. The
PMTH derivatives of amino acids gave only a single peak in the ESI/MS mass spectrum.
Figure 4-10 shows the ESI/MS mass spectrum of PMTH val (5pmol). The PMTH

derivatives of all the amino acids studied gave similar response during the ESI/MS. The

I40
detectability of phe-PMI'H was evaluated on the selected ion monitoring mode. Figure 4-
11 shows the profile obtained for the replicate injections of the PM-thiohydantoin amino
acid. The conditions were maintained as in the previous experiment. The phe-PMTH

derivatives can be detected at the low femtomol level using the selected ion monitoring
mode.

(M+H)

236.0
100-

80*

ON
0

Relative Intensity

b
O

204

 

 

 

100 150 200 250 300 350 450

m/z
Figure 4-8. Portion of the ESI/MS mass spectrum of PyTH-val (5pmol). Sample was
injected into a stream of CH3OH/HzO/HOAC (50/49.9/0.1 %v/v/v) at a flow rate of
lOOuL/min.

141

 

 

 

 

500 finol 5 Pm°1
F1

40:
"i
E
0
0
% 20~
a: 50 fmol

Sfmol

 

 

 

 

 

 

 

wr'wvvv

200 "'466" 6256"" 366771666 tenai'i‘oarsoo
scan #

Figure 4-9. The mass spectral profile obtained by injecting increasing amounts of purified
PyTH-phe derivative into a stream of CH3 OH/HzO/HOAC (50/49.9/0.1%v/v/v) at a flow
rate of IOOuL/min.

142

 

 

+
(M+H)
250
100
so
.5 60-
§
.5
_':‘>’
33
32 4o
20~

100 150 200 250 300 350 400
m/z

Figure 4-10. Portion of the ESI/MS mass spectrum of PMTH-val (5pmol). Sample was
injected into a stream of CH3OI-I/H20/HOAC (50/49.9/0.1 %v/v/v) at a flow rate of

lOOuL/min.

143

 

 

 

 

 

 

 

 

 

'1 T
50 pmol
40"
.29
E
.5
0
.2
L3
O
:4
20‘ 5 pmol
'5 fmol 50 frnol 500 fmol
I ' iolto' ' ' fl ‘ ' ' islto """" iooo """" isbo
Scan #

Figure 4-11. The profile obtained for the replicate injections of the PM-thiohydantoin
amino acid in selected ion (m/z 298) monitoring mode. The conditions were maintained as

in the previous experiment.

144

D. ESI/CID/MS/MS

In order to characterize the thiohydantoin amino acids, tandem mass-spectrometry
of TH amino acids was undertaken on a TQMS. Solutions of the TH-arnino acids were
infilsed into the ESI source at a rate of 3ul/min. The concentration of the samples were
1pmol/ul. The ESI needle was held at 4kV, capillary at 200°C. The collision chamber
(Q2) pressure was at 3 mtorr and energy was at 30eV (lab). The protonated molecule was
subjected to CID and MS/MS mass spectrum was collected at the Q3. Figures 4-13
shows the ESI/ClD/MS/MS mass spectra of lys-PyTH. The proposed fragmentation
patterns are indicated in Figure 4-12. Peaks representing a loss of the side chain group
and the reagent were observed in all the derivatives studied. The fragmentation pattern
helps to characterize the TH amino acid derivatives.

Figure 4-14 shows the MS/MS spectrum of PMTH-asp and the proposed

fragmentation pattern.
c ,c
+ H ~~~~~ ‘ O (1” |
ll 5 s
/ C\ i ll 1
©/ N\ /CH—(CH2)4— NH-l—C—t-NH O
l : N
N C+ N/H ’1' I}
S/ + H” _ H

Figure 4-12. Proposed fragmentation patterns for PyTH-lys under ESI/CID/MS/MS.

145

 

 

 

 

 

 

 

 

 

 

 

b
265
roe
C
137
80‘
£9 60*
E +
,5 (M+H)
3 a 401
"7' 307
3 4e 95
129
147
204
204 283
M “W.” U W W W .... W ... M ....
100 200 300 400
m/z

Figure 4-13. ESI/MS/MS mass spectrum of PyTH-lys. Collision gas pressure was at
3mtorr and energy was at 30eV. Sample was infused at a flow rate of 3 uL/min. Mass
spectrometer was scanned 100 - 450 Daltons in 0.5 sec.

 

 

 

 

b
+H a
cs ‘\ O -H (a)
s‘ \‘ ‘
1 i, ll :
I I
' l l
©‘CH2T1‘\/‘\ fH —§—C Hz—COOH
I l
| r
P‘ 4/gc+ ‘ hfli i
s \
(M+H)+
266
uxn
80y
601
40‘ b
151
C a
92 206
20‘ 99 225
169 m

 

 

fi’ ‘Tfi V v v v v v v v v v v v v v r V
T T I I

'50 100 150 200 250 3b0
m/z

Figure 4-14. ESI/CID/MS/MS spectrum of PMTH-asp. Collision energy at 30eV and
pressure at 3 mtorr and the proposed fragrnentations.

147

Using mass spectrometry, isomass compounds cannot be distinguished easily. This
is a drawback of the mass spectrometric techniques when used to detect the isomass
thiohydantoin amino acids without HPLC separation. Detection of Ian and ile is very
important in peptide sequencing to get the correct sequence information. The TH
derivatives of leu and ile were subjected to tandem mass spectrometry to evaluate the
possibility of distinguishing leu and ile based on their fi'agmentation pattern. Figures 4-15
and 16 show the MS/MS mass spectra of PyTI-I-leu and ile respectively. The spectra are
very similar in nature. Both PyTH derivatives of leu and ile amino acids show a loss of
side chain that has a similar mass. A strong peak for a (leu-PyTH - leu + H)+ was
observed. According to these results, leu and ile derivatives cannot be separated based on
the MS/MS data along. In order to identify ile and leu separately, a short HPLC column
has to be used.

 

 

 

b a
+H . 0 ‘~
\“ ‘
I
: $113
I
I
©://1‘\:\/;H -:r—CH2— on
I
I
b
137
100
80-
60-
40‘
+
(M+H)
250
20* 121
86
m M

 

 

Figure 4-15. ESI/CID/MS/MS mass spectrum of PyTH-leu and the proposed
fragmentation pattern. Collision energy 30eV and pressure 3 mtorr.

149

b
+
”0‘ as,
s“ W \‘
t‘ :
Ac 5
@/ 1‘\ fH —:r—CI H—CHZCH3
I
N /C—-‘~“— NH E ('3H3
I
s \
b
137
1007
sod
E?
.... 60..
3
.5
.2
3E
52 4o~
+
(14+ED
20~ so 250
121
163 205

 

 

 

so 100 150 206 'ffizso ...360
m/z

Figure 4-16 ESI/MS/MS mass spectrum of PyTH-ile amino acid derivative and the
proposed fiagmentation pattern. Conditions were similar to those of Figure 4-15.

150

V. Studies on Reaction Completion

A Experimental

The coupling reaction is very important for Edman degradation chemistry for
successful repetitive cycles. The coupling reaction efficiency was assessed with bradykinin
(RPPGFSPFR) peptide under conditions typically used in manual or automated
sequencing. Aliquots of lOpmol of bradykinin in water were placed in microfirge tubes,
supplemented with 50uL of coupling buffer (HzO/methanol/N-ethylmorpholine 25/70/5 %
v/v) and with 10 pL of reagent dissolved in acetonitrile. The atmospheric air was removed
by passing dry nitrogen through the solution for about 10 seconds. The samples were
mixed by vortexing and incubated at 45°C, 50°C, and 60°C respectively for times ranging
ham 1 minute to 10 minutes in a water bath. After incubation, the whole reaction mixture
was immediately injected into a pre-equilibrated reversed phase HPLC system. The native
peptide and the derivatized peptide was monitored using a UV-VIS detector. The
reaction products were analyzed on a C18-column (4.6mm x 25 cm) at a flow rate of
2mL/min. A chromatographic column was developed on a gradient program of 20%B to
100% B. The solvent A was H20/0.1% TFA and the solvent B was acetonitrile/0.1%
TFA The eluents fi'om the column were monitored at 214nm using a Kratos Spectrflow

747 variable absorbance detector.

B. Results and Discussion

1. Coupling reaction with PyITC
Figure 4-17 shows the variation of coupling reaction completion (%) with time at
various temperatures. The coupling reaction is clearly temperature and time dependent as
shown in Figure 4-17. The reaction yield with the reagent concentration (1% - 10%) was

also assessed for the completeness. The coupling reaction seems to be faster than that of

1 5 1
PITC under similar conditions. At 50°C, 2% PyITC seems to be reasonable for the
reaction. The use of low reagent concentrations helps to reduce amounts of by-products
formed during the coupling reaction. The formation of by-products is always a problem
with Edman sequencing and interferes with TH detection. In an automated sequenator, a
reducing agent such as dithiothreitol is added to the solvents to prevent formation of by-
products usually the oxidation products.

2. Coupling reaction with PMITC
Figure 4-18 shows the variation of reaction completeness with temperature and
time. It clearly demonstrates the dependence of reaction yield on temperature and time.
The percent reaction completeness of PMITC is lower than that of PyITC at 50°C, but at

60°C it reaches completion.

C. Cleavage Reaction of TC-peptides

VGVAPG, and ACP (V QAADYING) peptides (40 nmol) were derivatized with
the reagents PyITC, PMITC, and PITC, respectively. The products were purified by RP-
HPLC and the fi'actions were manually collected. The pooled solutions were equally
divided into four effendorf tubes and dried under vacuum to dryness. The dried products
were incubated with 20 11].. of anhydrous TFA in a water bath at 50°C time ranging fi'om 1
minute to 10 minute. The products were immediately analyzed by FAB/MS.

Alternatively, the incubated products were injected into a pre-equilibrated RP-
HPLC system and developed as before. The disappearance of TC-peptide and the
appearance of the shortened peptide was monitored on a UV-Vls detector. The peaks
were manually collected and confirmed by FAB/MS.

% Conversion

100
i
804
60$
405

20—

152

+45°C
-¢-5¢c
-* ”60°C

 

 

Time/min

Figure 4-17. Kinetic assessment of coupling reaction between pyridyl isothiocyanate and
bradykinin (RPPGFSPFR). The peptide was treated with 2.5% PyITC in a buffer solution
and the coupling reaction was assessed by analyzing the reaction mixture by RP-HPLC.

% Conversion

100.0—
1

80.0;

60.03
40.05

20.01

.1

 

+ 60°C
+ 50°C

 

0.0 ‘

0

d
-l
.1

Time/min

Figure 4-18. Kinetic assessment of coupling reaction between pyridylmethyl
isothiocyanate and bradykinin (RPPGFSPFR). The peptide was treated with 2.5% PMITC
in a bufl'er solution and the coupling reaction was assessed as in Figure 4-17.

153

D. Results and Discussion

The anhydrous TFA acid cleaves the thiocarbamyl peptide to give an ATZ amino
acid and a truncated peptide. The unstable ATZ amino acid is converted to the more
stable PTH amino acid before HPLC separation followed by UV detection. The cleavage
step is an important part of the peptide sequencing procedure. FAB/MS mass
spectrometry is a convenient technique to study the reaction products quickly by using the
direct insertion probe. Figure 4-19 shows that the TC-peptides undergo cleavage to give
unstable ATZ amino acids and truncated peptides in the presence of anhydrous TFA at
50°C. F AB/MS data shows that the PyTC-VGVAPG peptide underwent cleavage at 50°
C within 10 minutes (Figure 4-19a). This reaction is somewhat slower than the cleavage
of the other derivatives. It cleaved the PyTC-ACP peptide to nearly the same extent at 10
minutes.

Figure 4-19b shows the cleavage of PMTC-VGVAPG when incubated with
anhydrous TFA at 50°C for 2 minutes. The cleavage reaction of PMTC-VGVAPG is
complete within 2 minutes and faster than that of the PyTC-peptide.

Figure 4-19c shows the extent of cleavage of PTC-VGVAPG at 50°C after 5
minutes. After 5 minutes, the PTC-peptide was almost disappeared in the mass spectrum.
The cleavage reaction for PMTC-peptides is comparable to that of the PTC-peptide.
PyTC-peptide cleavage is slower than that of the PTC-peptide. But within 10 minutes,
both TC-peptides undergo complete cleavage.

The progress of the cleavage reaction was assessed by monitoring the appearance
of the earlier eluting truncated peptide and the disappearance of the original peptide.
HPLC analysis indicates that cleavage of PyTC- and PMTC- peptides with anhydrous

TFA is essentially complete within 10 minutes.

154

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1-
R 4
. l
1 . .
. 80‘ O
,t j 369
'V « (a)
e 60*
A
b
:40- (Gvapcetrn .
g . 4°° e.
l 0
“2g (VGVAPG—PyTC + H)+
c 553 .
’ l
owl—VA l 4% A'f A L Y L I A A JV it
350 400 450 500 550 600 650 100
M12
100
R J
c
l t
.801
t 4
i . “ 9 (cvapc + 10+ "9
v 1
0 so 335 400
1
A 1
b 1
“40—
II 1 .
d .
. 4 461
“20‘ ' (VGVAPG-PMTC + Hij+
c . 153
. t
1... AAA; '7 AA A1 _A. ‘1: ..L_._
0350 400 450 500 650 600 650 M12700
100
R 1
O l
l r
:30 (e)
i 1 O
" ‘ 3 9 (GVAPG + 10+
° so. if 400
A t
b
“40
ll ‘ g
d .
a 461
3” 1 '3 (vchc-erc + H)+
‘ 1 i at
M.;. Jul- ..L ... . i l; A ' AL .1. L'L‘ ._ .L‘ A
400 4E1 560 650 600 550 M12700

Figure 4-19. FAB/MS mass spectra showing the cleavage of thiocabamyl peptide
(VGVAPG) with anhydrous TFA at 50°C. (a) PyTC-peptide after 10 min, (b). PMTC-
peptide afier 2min., and (c) PTC-peptide alter 5 min.

1 5 5
VI. Peptide Sequencing

The ACP peptide (V QAADYING) was manually sequenced using both reagents.
The peptide (lOOpmol) was incubated in coupling bufi‘er (200uL) with the reagent for 30
minutes at 50°C for PyITC and 60°C for PMITC. The solvents were evaporated and
excess reagent and by-products were extracted with ethyl acetate twice. The dried residue
was incubated with anhydrous TFA (20uL) for 10 minutes at 50°C and dried. The ATZ
derivative was extracted with ethyl acetate and dried in vacuo. The residue was converted
to the more stable thiohydantoin derivative with 25% TFA The TH amino acids were
injected into a stream of CH3OH/HzO/HOAC (50/49.9/0.1 %v/v) at a flow rate Iota/min
and analyzed by ESI/MS.

The first five residues were manually sequenced and could be detected at lOOpmol
using ESI/MS. The overlap between residues were minimal. In this experiment long
coupling times (30 minutes) were used (typical times for manual sequencing). A fairly
high amount of material (lOOpmol) was used in this experiment; and the amount can be
decreased by using automated peptide sequanator. The detection limit of the sequencing
process was not assessed. In manual sequencing there can be sample loses due to
incomplete extractions, wash out, or adsorption to walls of the tubes. In the automated
sequanator, the sample loses are minimized by non-covalently immobilizing the sample on
a polybrene membrane to avoid wash outs, precise extraction steps, and using a micro
reaction chamber. Therefore, low-levels can be sequenced using the automated
sequanator. The thiohydantoin amino acids can be analyzed at the low femtomol level as
shown earlier. These reagents have the potential to prepare suitable derivatives of
peptides for low-level sequencing capabilities and detection by ESI/MS.

156

VH.. Post-translationally Modified Amino Acids

One of the great interests in protein sequencing today is post-translationally
modified amino acids. It is very important to know the correct sequence without any
ambiguity. In the peptide sequencer, the PTH amino acids are separated by HPLC and
detected by UV absorption of the eluent. The PTH amino acids are identified by their
retention time in the HPLC chromatogram. The retention times are standardized using the
common PTH amino acids. Post-translationally modified amino acids are missed or
unidentified by their retention times, if they are not included in the retention standard. The
results of the Association of Biomolecular Resource Facilities Sequencing Committee
survey shows that hydroxyproline and O-phosphoserine are diflicult to identify during
peptide sequencing for reliable confirmation (19). In this survey, 75 facilities participated
and only 46% facilities identified hydroxyproline correctly. The most difficult amino acid
to identify was phosphoserine. In this part, mass spectrometry will be used to detect
modified TH amino acids by their molecular weight information.

Hydroxyproline is a common modified amino acid present in the peptides. During
the HPLC separation and detection technique, there are two peaks present in the
chromatogram as shown in Figure 4-20. A peak lies close to the peak for PTH alanine
(20). Usually, this will be identified as alanine unless the accompanying peak is taken into
consideration.

Therefore, hydroxyproline was used as a modified amino acid to study the
suitability of mass spectrometry with the new reagents. The thiohydantoin derivatives of
hydroxyproline were prepared as described earlier using both reagents PyITC and PMITC.
The purified TH amino acids were analyzed by ESI/MS. The derivatives (5pmol) were
injected to a stream of CH3OH/HzO/TF A (50/49.9/O.1 %v/v) at a flow rate of

. 157
lOOuI/min. Figure 4-21 shows the ESI/MS mass spectrum obtained for the PMTH-
hydroxyproline. This can be identified without ambiguity using the molecular weight.
Another common modified amino acid present in proteins is O-phosphoserine. O-
phosphoserine was derivatized with the modified Edman reagents and analyzed by
ESI/MS. Figure 4-22 shows the mass spectrum obtained for PM'I'H-O-phosphoserine
(5pmol). The pyridyl group in the modified reagent makes it amenable to ESI/MS. The

O-phosphoserine derivatives can be identified using ESI/MS to determine its molecular

weight and HPLC its retention time.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

"RHYDROXYPROUN!
soon
D
N
«triob
D i Y M
14 r v D
E 1' P W
a U A P T
b O U
4000 s c P It
T H
r? r L
senor D
r
r
1 Lil 8' Al
one " AAiJJ_ A _- ..
.l A I 1 l 1 l l 1
so so 100 12.0 14.0 use no 220 240 200 20.0

Figure 4-20. HPLC chromatogram for the PTH amino acid standard containing
hydroxyproline. Hydroxyproline eluted as two peaks.

158

 

 

 

 

100-
O
/c\
Cflz—N CH—CHz
O c .../1 \
. _ CH H
so / ‘ \ / 2—0
8 C142
MW.-263
so-
>5
3::
ii
H
.5
0
.g
.9;
o 40‘
d.
20‘
150
Mini LA -- -11
T—‘ re ' r j l H H* r r
150 200 250 300 350 400 450
m/z

Figure 4-21. ESI/MS mass spectrum of hydroxyproline (5pmol). Sample was injected into
a stream of stream of CH30H/I-120/HOAC (50/49.9/O.1 %v/v/v) at a flow rate of

lOOuL/min

159

318.1

 

 

 

 

 

i
/C\
CH2—N CH—CHz—O—P03H
O \. .4
/
S
MW.-317
60"
a?
an
t:
3
.5 to-
0
..>..
E
O
a:
20‘
{112"
192.0 29?.9
150‘ 200 250 300 .350 400 450

Scan#

Figure 4-22. The ESI/MS mass spectrum of PMTH-O-phosphoserine (5pmol). Sample
was injected to a stream of CH3OH/HzO/HOAC (50l49.9/0.1 %v/v/v) at a flow rate of

100nL/min.

160
VIII. Conclusion

The responses of the modified amino acid derivatives were evaluated for ESI/MS.
They were detectable at the low femtomol level using selected ion monitoring. Percent
completeness of the coupling reaction at 50°C indicates that the PyITC reagent reacts
very fast with the fi'ee amino terminus of the peptide. The cleavage rate is little slower
compared to that of PITC reagent.

The PMITC reagent reacts slowly with the free amino terminus of the peptide at
50°C and the reaction goes to near completion at 60°C by 10 minutes. PMTC-peptide
cleaves very quickly with anhydrous TFA. The PMTH derivatives can be detected at the
low femtomol level using ESI/MS.

Both reagents were evaluated for their capacity to use as Edman reagents in
peptide sequencing. They both can be used in sequencing and the resulting thiohydantoins
can be detected by ESI/MS. In order to completely evaluate the reagents for low level

sequencing, the automated sequencer must be used.

16]

IX. References

10.

11.

12.

13.

14.

15.

16.

17.

Edman, P. Acta. Chem. Scand 1956, 10, 761

Mische, S.M.; Yfiksel, K. 0.; Mende-Mueller, L. M.; Matsudaira, P.; Crimmins, D.
L.; Andrews, P. C. Techniques in Protein Chemistry IV, 1993, 453

Melvas, B. W. Acta. Chem. Scand. 1969, 23, 1679

Fales, H. M.; Nagai, Y.; Milne, W. A.; Brewer, H. B.; Bronzert, T. J.; Jr.; Pisano,
J. J. Anal. Biochem. 1971, 43, 288

Pramanik, B. C.; Hinton, S. M.; Millington, D. S. Dourdeville, T. A.; Slaughter, C.
A Anal. Biochem. 1988, 175, 305

Aebersold, R.; Bures, E. J.; Namchuk, M.; Goghari, M. H.; Shushan, B.; Covey,
T. C. Protein Science, 1992, I, 494

Basic, C.; Bailey, J. M.; Shively, J. E.; Lee, T. D. Proceeding of 41st ASA/IS
Conference on Mass Spaectromerry and Allied Topics, San Francisco, CA, May
31-June 4, 1993

L'Italien, J. J.; Kent, S. B. H. J. Chromatogr. 1984, 283, 149

1111, S-W.; Chen, G-X.; Palacz, Z.; Wittmann-Liebold, B. FEBS, 1986, 198, 150
More, M. L.; Crossley, F. S. Org. Syn. 21 81

Dains, F. B.; Brewster, R. Q.; Olander, C. P. Org. Syn. Call 1, 447

Dyson, G. M. Org. Syn. Call! 165

L'abbe, G. Synthesis 1990 525

Shibanuma, T.; Shiono, M.; Mukaiyama, T , Chem. Lett., 1977, 573

Kim, 8.; Yr, K. Y., Tetrahedron Lett. 1985, 26, (13), 1661

Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43 (19), 2923

In ”The Chemistry of Cyanates and Their Thio Derivatives”, Patai, 8 (Ed) 1977
John Wiley & Sons, New York

18.

19.

20.

162
Hunt, B. J.; Rigby, W. Chem. Ind. (London) , 1967,1868
Hihara, J. Protein Analysis Renaissance, Applied Biosystems, 1993/94, 15

Sengoku, Y.; Shionoya, Y.; Hagiwara, K.; Kirsher, 8.; Yuan, P.-M Protein
Analysis Renaissance, Applied Biosystems, 1993/94, 25

CHAPTER 5
MODIFICATION OF ANABOLIC STEROIDS FOR ELECTROSPRAY MASS
SPECTROMETRY

I. INTRODUCTION

The illegal use of anabolic steroids to affect the outcome of both human and
equine sport events is of major concern and has been well documented (1,2). A concise
summary of the controversial literature on the effect of anabolic steroids on athletic
performance and their deleterious side effects has appeared (3). Most anabolic steroids
are synthetic derivatives of the male sex hormone testosterone, with high anabolic
potential and minimized androgenic activity. Due to extensive metabolism, confirmation
of the unlawfiil use of this group of doping agents requires an understanding of their
metabolic fate in the species and the development of more efficient analytical methods for
unambiguously detecting these steroids and their metabolites in biological fluids (4).

Several different analytical techniques have been employed to detect anabolic
steroids in biological fluids. The most common techniques are radioimuimnnoassay (5)
(RIA), high performance liquid chromatography (HPLC) (5), gas
chromatography/electron impact mass spectrometry (GC/EI-MS) (6), gas
chromatography/electron capture negative ionization mass spectrometry (GC/ECNI-
MS)(7), and more recently liquid chromatography/mass spectrometry (8-12). The major
advantage of RIA is its sensitivity, but it is not routinely used in screening because of
disadvantages such as cross-reactivity, difiiculty in measuring the testosterone-to-
epitestosterone ratio, and difficulty in detecting methenolone (13). HPLC has been used
mainly for the analysis of anabolics in the veterinary, pharmaceutical, and forensic fields,
rather than in human urine screening. The major problem with HPLC is absolute

confirmation of the analyte.
1 63

164

By far the most common analytical method for anabolic steroid analysis is GC/MS.
Steroids must be derivatized to protect fimctional groups and to enhance volatility. (6, 14,
15). Methoxime-trimethylsilyl derivatives were among the first to be used in the analysis
of anabolic steroids. The methoxime moiety protects the carbonyl functionality while the
trimethylsilyl reacts with the hydroxyl groups. Reports of detection limits for individual
steroids or metabolites range from 3 ng/mL to 20ng/mL of urine (16-18). Trimethylsilyl
(TMS) derivatives are now more widely accepted than methoxime-TMS derivatives (19-
22). Selective derivatization can be accomplished by selecting the silanizing reagent and
catalyst. I

In order to detect trace levels of anabolic steroids long after administration,
analytical techniques with better sensitivity and selectivity are needed. Electron capture
negative ionization (ECNI) mass spectrometry has the capacity to meet some of these
analytical needs. But ECNI has not been as widely used as had been anticipated, this is
partially due to problems associated with finding proper derivatives that are amenable to
the ionization technique. Chemical oxidation followed by GC/ECNI-MS has been
developed in this laboratory as a sensitive and selective detection technique for
dexarnethasone and prednisolone (23, 24). In this method, steroid molecules are
chemically oxidized to a highly electrophilic species while not significantly afl'ecting the
electrophilic character of the biological matrix, thereby pemritting selective and sensitive
detection of the analyte. Efl'ect of structure on the ECNI mass spectrometric response of
steroids has been reported (24). It was determined that the key to enhanced ECNI-MS
response of ketosteroids is extensive or, B-unsaturation.

The atmospheric pressure chemical ionization (APCI) (11) and thermospray (9,10)
approaches are the most widely used in LC/MS techniques to detect steroids and their
conjugates. The detection limits of low ng/uL to pg/uL has been reported for steroids
using thermospray mass spectrometry (8).

165

Usually, the forms of metabolites are conjugation with glucuronic acid or sulfate.
Since drug conjugates are not amenable to determination by GC/MS, owing to their low
volatility and thermal instability, they are hydrolyzed before analysis by GC/MS. Henion
et al. reported the detection of anabolic steroid sulfate conjugates by using negative
electrospray mass spectrometry (ESI/MS) with excellent detection limits (11).
Subsequently, several research groups have reported the use of negative ESI/MS to detect
steroid conjugates in biological-fluids (12).

Figure 5-1 shows the generic steroid structure and the numbering system used in
steroid nomenclature. The use of anabolic steroids to enhance the performance of athletes
are forbidden by the International Olympic Committee (10C). The structures of banned
anabolic steroids are given in Figure 5-2.

The objective of this part of the research project is to develop analytical
methodology to detect trace levels of anabolic steroids present in biological samples using
electrospray mass spectrometry with high specificity.

II. Modification of Steroids for ESI/MS

The anabolic steroids are a very important class of molecules with similar
structural characteristics. All the steroids have similar steroid skeleton except stanozolol.
The analysis of stanozolol is discussed later in this chapter. The unconjugated steroids
have no functional group to form ions readily in solution. The structures of common
anabolic steroids have a keto functionality in common as shown in Figure 5-2. The keto
group was modified using Girard's T reagent or hydroxylammine hydrochloride to
improve the sensitivity by ESI/MS. The Girard's T reagent forms a hydrazone derivative
with the keto group of the steroid and it carries a quartenaiy ammonium group which has

a pre-formed charge. The reagent hydroxyl amine hydrochloride forms an oxime

166

derivative with the keto functionality and can be protonated easily in acidic solution.

Formation of both derivatives with steroids will be discussed in detail later in this chapter.

111. Hydrazone Derivative Formation

The Girard's T reagent first introduced by Girard and Sandulesco (25, 26) reacts
with the keto functionality of steroid ketones to form a water soluble hydrazone. The
hydrazone derivative carries a quaternary ammonium moiety that is amenable to positive
electrospray mass spectrometry. The ESI/MS response of quartenary ammonium groups
is well documented to be in the low femtogram range (27). In this study, several
testosterone analogs (Table 5-1), were used as representative model compounds.

The reaction, shown in Figure 5-3, with testosterone, proceeded readily when 10
nmol of steroid in methanol was mixed with a 20-fold excess of Girard's T reagent
dissolved in methanol. The solution was brought to a final acid concentration of 1M
HOAC by adding glacial acetic acid. The reaction mixture was incubated in a water bath
at 50°C for 30 minutes. The reaction mixture was injected directly into a pre-equilibrated

RP-HPLC system for separation.

A Mass Spectral Analysis

In recent years, electrospray mass spectrometry has gained wide applicability in
biological mass spectrometry due to its high sensitivity and capacity to form multiply
charged ions in determining the molecular weight of polymers that cannot be determined
by conventional mass spectrometry. In this study, ESI/MS has been investigated for
analysis of anabolic steroids.

167

 

21 22 24 26
o
18
12 23 25 27
11
17
13
1 19 16
2 8 14 15
3
5 7

A = androstane, C19 skeleton
P = pregnane, C21 skeleton
C = cholestone, C27 skeleton
ene = double bond
01 = hydroxy group
one = carbonyl group
or = Stbstituents erdendingintothe plane ofthe pay

[3 = Stbstituents extendingoutofthe plane ofthe page

Figure 5-1. Nomenclature and numbering system for the steroid carbon skeleton

168

as“ as“ as:

Testosterone Nandrolone Methyltestosterone
OH OH on
M“ yogi-m3
1
H
Orandrolone Bolasterone Boldenone
0H OH
OH 0'13
' (113 I I I l '>
Cl
Fluoryrmsterone Mesterolone Costebol
OH OH 01-1
M. 013 X'GES (fi- 013
Mcthand'cnone Methenolone Dehydrochbromethyl
testosterone
OH OH OH
OHC
Stanozolol Norethandrolone Fomyklienolone
0H 0H 0H
- -CH3
«EL/95 CH3 11% “(MW
Oxyrmsterone Drannstanolone Oxyrmtholone

Figure 5-2. The Structures of Anabolic Steroids banned By IOC

169

It has previously been shown that quaternary ammonium moieties and strongly
acidic groups such as sulfates or sulfonates are detectable at low femtomole levels by
pneumaticatically assisted electrospray mass spectrometry (27, 28). Hydroxylated
compounds have been analyzed by positive electrospray mass spectrometry as a sodiated
molecule (M+Na)+ (29). We investigated a method to improve the detection limits of
anabolic ketosteroids by forming hydrazone derivatives with Girard's T reagent.

The mass spectral analysis was undertaken on the home-built electrospray interface
for the preliminary investigations. Other investigations were undertaken on a TSQ 700
triple quadrupole mass spectrometer fitted with a Finnigan ESI source. The ESI needle
was held at 4.5kV and the heated capillary tube was at 200°C. Samples were introduced
into the ESI source through an infusion pump or loop injection.

The ESI mass spectra were acquired by directly loop-injecting a SuL portion of the
reaction mixture containing 100ng 11L of each hydrazone into a stream of mobile phase
(CH3OH/HzO/CH3COOH 50:49.9:0.1 % v/v/v). Hydrazone derivatives were prepared
with several steroids (Table 5-1) and the ESI/MS mass spectra were acquired. All the
hydrazone chlorides show a strong peak for the molecular cation in the ESI/MS mass
spectra. The electrospray mass spectra of the hydrazone derivatives of model compounds
showed little or no fragmentation under normal ESI conditions as illustrated in Figure 5-4.
The lack of fragmentation coupled with the fixed charge on the quaternary ammonium

moiety facilitate efficient detection of these steroid derivatives by electrospray mass

spectrometry.

170

Table 5- 1. The steroids used to prepare hydrazone derivatives

 

 

 

 

 

Testosterone Boldenone
Methyltestosterone Bolasterone
Nandrolone Mesterolone
Fluoxymesterone Methandrostenolone
1,2 Dehydro 17a methyltestosterone

 

 

 

B. Detection Limits of Hydrazones

The strong response of quaternary ammonium functional groups using electrospray
mass spectrometry is well documented and detection limits in the femtogram range have
been reported (27, 28). To determine the detection limits for the hydrazone derivatives of
testosterone analogs, increasing amounts of methyltestosterone hydrazone chloride were
injected into the ion source of the triple quadrupole mass spectrometer. A volume of SuL
was flow injected into a stream of methanol/water/acetic acid (SO/49.9/0. 1%v/v) at a flow
rate of lOOuIJmin. Figure 5-5 shows the mass spectral response during selected reaction
(m/z 416 to m/z 357) monitoring in the triple quadrupole mass spectrometer. Detection

limits into the low femtogram level were achieved.

171

H
or 3

...
H3C— N— CH2CONH— NH2 +

 

CH3

Girard's T reagent Testosterone

Cl' H3

+ , + H20
H3C-- N— CH2CONH— N '

 

CH3

Testosterone Hydramne Chloride

Figure 5-3. The reaction scheme for the formation of steroid hydrazone. The Girard's T

reagent reacts with the keto group of the steroid to form hydrazone derivative.

172

 

 

M+
450.2
100 i
80‘
.5? 60
3
.5
3
3
0
M 40.
20‘
. it, JILL; Al... .1“ 11.111 1‘1“] hrtlerll 1...: ALL“... [.11

 

 

360 380 400 420 440 460 480 500

m/z

Figure 5-4. Portion of the ESI/MS mass spectrum of fluoxymesterone hydrazone chloride

(molecular weight of fluoxymesterone hydrazone cation = 450)

173

IM‘

hang

A

Rehdhn:hflen§ty

2119

 

 

 

2019 200., 2m 2099 20°” \J

L t a v

T—v—v Vrfi

 

 

unur#

Fig. 5-5. Detection limits for the methyltestosterone hydrazone molecular cation. The

samples were injected into a stream of CH3OH/HZO/HOAC (SO/49.9/O. 1% v/v) flowing

at a rate of lOOuL/min in increasing concentration. The injection volume was 5 1.1L. The

amounts are indicated above the corresponding peak.

l 74
C. MS/MS Studies of Hydrazones by ESI/MS

Tandem mass spectrometry (MS/MS), since its development in the 1970s, has
been gaining acceptance as a rapid, sensitive and selective analytical method for the
analysis of complex biological samples (3 0-33). The advantages gained by MS/MS often
complement those gained by extensive sample clean-up procedures and lengthy
chromatographic separations (34-3 6). Other capabilities of tandem mass spectrometry
such as selected reaction monitoring and constant neutral-loss scanning make this mode of
operation attractive in analyzing mixtures to obtain selectivity. It has been investigated the
capacity of MS/MS to reduce chemical background during an analysis by using constant
neutral-loss scanning with the triple quadrupole mass spectrometer.

The collision-induced fragmentation of hydrazones was studied with the aim of
discriminating between secondary ions generated by fiagmentation of the primary ion and
chemical noise. Such discrimination can be used as the basis for increasing the signal-to-
background ratio by selective electronic filtering. These experiments were of particular
importance as they addressed one of the major technical problems in highly sensitive
analyses of complex biological samples, chemical noise obscuring low-level signals from
the analyte. The experiment was undertaken on a triple quadrupole mass spectrometer
fitted with a F innigan ESI source. The salient fragmentation pathways comprised a
stepwise loss of neutral N(CH3 )3 (59u) followed by another loss of 28u. The
fi’agmentation of testosterone, methyltestosterone, nandrolone, fluoxymesterone, and
bolasterone hydrazones provided analogous spectra. The hydrazone derivatives of
boldenone, methandrostenolone, and mesterolone gave strong peaks at (059) , (C-8 7),
and additional peaks at (C-15), (C-102), and (C-117) in their MS/MS spectra where C
indicates the molecular cation of the derivative. Table 5-2 shows the list of peaks

observed in the tandem mass spectrometric studies. The collision energy was maintained

175

at 40eV relative to the laboratory frame. An increase in collision energy beyond 40eV did
not yield new fragments, but abundances of product ions increased with collision energy.
During the MS/MS study, the samples (100pg/uL) were infused at a flow rate of 3Wmin
with a syringe pump (Cambridge, MA ) while the mass spectrometer was scanned from
lOu to 500u in 0.5sec and the pressure inside the collision cell was maintained at 3.0
mtorr. Figure 5-6 shows the MS/MS spectrum of methyltestosterone hydrazone after
averaging data acquired over one minute. Therefore, all the hydrazone derivatives show a
common loss of 59u in their MS/MS spctra.

D. Study of Fragmentations with Offset Energy and Pressure

When MS/MS experiments are carried out on a triple quadrupole mass
spectrometer, two major variables affect the abundance of representative product ions.
One variable is the collision energy (offset energy) applied to the RF -only quadrupole.
Another variable is the pressure inside the collision cell. The effect of both variables were
studied while infirsing the sample at a constant flow rate of 3 uL/min. The ESI needle was
at 4.5kV and capillary tube was heated to 200°C. The nebulizing gas (N2) pressure was
at 35psi. All the experiments were conducted under similar conditions and data acquired
after averaging over one minute. The abundances of product ions were studied as a
function of collision energy fi'om 20eV to 40eV. Figure 5-7 shows the product ion
abundances for the testosterone hydrazone derivative with the ofi'set energy at the
collision pressure of 3mtror (Ar).

The absolute abundances of three characteristic ions of MS/MS spectra of
testosterone hydrazone (collision gas pressure at 3 mtorr) were plotted against the Vub.
Figure 5-6 shows that the m/z 343 (M-59)+ ion has the highest abundance at offset energy

of 30eV. When examining the respective spectra, the spectrum at 40eV has more

 

176

abundant fiagments than that of at 30eV and the ion current at M+ has distributed over
the large number of fragment ions at 40eV. At 30eV, the product spectrum has only a few
fiagment ions and the most abundant fragment ion is the (M-59)"' ion. More intense
fiagmentation can be induced at 40eV collision energy. The same pattern was observed at
the collision gas pressures l, 1.5, 2.0, 2.5, 3.0, and 3.5 mtorr.

At 40eV collision offset energy, as the gas pressure increases, the abundances of
m/z 343 and 163 increase and the abundance of the parent ion decreases as shown in
Figure 5-8. In order to do neutral loss scanning or selected reaction monitoring (SRM),

the energy of 30eV at 3 mtorr is appropriate for hydrazone derivatives.

Table 5-2. A list of peaks observed in MS/MS studies of hydrazone molecular cation M+.
(where C = M")

 

 

 

 

 

 

 

 

 

 

Steroid Derivative Major Peaks Observed in MS/MS spectrum

(C-15) (C-59) (C-87) (C-102) (C-1 17)
Methyltestosterone + +
Nandrolone + + +
Fluoxymesterone + +
Bolasterone + +
Boldenone + + + + +
Mesterolone + + +

 

 

 

 

 

 

177

 

 

 

 

 

 

 

(M-59)+
357.3
1001
80*
g 60.
a
0
,>
g (M-87)+
9‘ 40* 329.3
+
M
416.3
20, 163.1
134.9
.41 l--- , W., .1 , -
100 200 300 400

m/z

Fig. 5-6. ESI/MS/MS spectrum of the molecular cation of methyltestosterone hydrazone
(M"’). The spectrum was obtained by infirsing the sample at a flow rate of 3 uL/min. and
at a concentration of 100 pg/ 11L. The mass spectrometer was scanned at a rate of
980u/sec and spectra were averaged for one minute. The collision energy was set at 40eV

(lab). Ar was used as the collision gas and the collision cell pressure was 3 mtorr.

1 78
E. F AB/MS Analysis of Hydrazones

FAB/MS analysis was undertaken on a JOEL HX 110 high performance double
focusing mass spectrometer with forward geometry (often denoted as an EB instrument
because the electric sector, E, precedes the magnetic sector, B). The hydrazone samples
were dissolved in methanol and dispersed on a few microliters of matrix that was glycerol
on a probe tip. The probe was inserted to an ion source of the mass spectrometer and
evacuated. An analyte ion beam was generated by bombardment of high energy Xe atoms.
The ion beam was accelerated to lOkeV and mass analyzed by the FAB/MS. The data
were collected on a DASOOO data system. Figure 5-9 shows the FAB mass spectrum of
methyltestoterone hydrazone (1ng). The asterisks indicate the matrix peaks. The FAB
mass spectrum of the steroid hydrazone gave a strong peak representing the molecular
cation and several peaks for the glycerol adduct ions. The matrix peaks of the FAB mass
spectrum make the detection of low level of analytes difficult in most cases. Comparison
of Figures 5-4 and 5-9 clearly shows the simplicity of the ESI mass spectrometry of
hydrazone.

179

 

 

 

51¢—
41¢—
€31¢~
m
3
=21&4
H i
11$
010°“....q.Le-.~.-.','.':T.*?T .... ....
15 20 25 so 35 4o 45
Collision Energy (eV)

Figure 5-7. Variation of product ion abundances with collision ofi‘set energy at the

collision cell gas pressure of 3mtorr.

+402
+343

 

 

l l l I

I I
O .5 1 1 .5 2 2.5 3 3 .5 4
Pressure (mtorr)

Figure 5-8. The variation of product ion abundances with collision cell gas pressure at the

collision offset energy of 40eV.

180
F. High Energy CID on FAB/MS
The molecular cation was subjected to the high-energy FAB/CAD/MS/MS
analysis. The M+ was selected and fragmented with Ar gas. Figure 5-10 shows the
FAB/CAD/MS/MS spectrum of methyltestosterone chloride.

G. Comparison of ESI/CID/MS/MS with F AB/CID/MS/MS

The ESI/MS/MS spectra and the FAB/MS/MS spectra of the hydrazone
derivatives are significantly different. The energy deposited in the ions are different in
each case. The ESI/MS/MS spectrum on the TQMS shows more fragmentation at the
lower-mass end of the spectrum and the high-energy FAB/MS/MS spectra shows more
fragmentation at the higher-mass end of the spectra. Both spectra show the loss of 59u
and 87u fi'om the molecular cation. In the FAB/MS mass spectrum of the testosterone
hydrazone, no significant fragments were observed below 200u. But in the TQMS, there

were several strong peaks below 200u due to ring cleavages of the steroid skeleton.

181

 

 

 

 

 

 

M4-
185-0 416.0
100 - .
R .
. 4
l .
a 80 .
t .
i
V ‘ 0
a so .
. 277
A .
b ‘ 359.0
11 40 l
n .
d 4
‘ 4
n 20 «
C
o 1 302.1
1 245-0 477.9
10° 150 200 250 300 350 400 450 m 500

Figure 5-9. The F AB/MS spectrum of methyltestosterone hydrazone chloride (1 ng).

Glycerol was used as the matrix and matrix peaks are indicated by an asterisk.

M+

 

.<-o.-—.w
8

H
OI
A l A A A

 

 

 

 

 

 

 

A 4
b I
u 10 -
n 4
d 4
‘ I
n
5..
c ‘ 237.1
C
68.0
o l l 1 1 111111
c fi‘r I Y I V 7 Y Y I Y fi fr T r I' V I Y I V Y~r Tfi Y
50 100 150 200 250 300 350 m

Figure 5-10. The FAB/CAD/MS/MS spectrum of methyltestosterone hydrazone chloride.

l 82
H. Neutral Loss Scanning Studies of Hydrazone Derivatives

Because all hydrazone derivatives fragmented with a neutral loss of 59u,
corresponding to the elimination of trimethylarnine, a method of MS/MS scanning for this
class of compounds could be implemented by scanning both quadrupole mass analyzers in
tandem with a mass difference of 59 u. In this mode of operation, termed as MS/MS
constant neutral loss scanning, only those compounds that undergo a neutral loss of 5%
will appear in the final analysis, therefore permitting a means of detecting these derivatives
in the presence of significant levels of potentially interfering contaminants. A derivative of
model compound nandrolone hydrazone was subjected to neutral loss scanning of 59u and
the following spectrum (Figure 5-11) was obtained. The neutral loss spectrum of
nandrolone hydrazone contains only a single peak and therefore, hydrazones that lose 59u
in an MS/MS experiment, are responsive to this mode of scanning. The nandrolone
hydrazone (lOOpg/uL) was infirsed at a flow rate of 311L/min into the ESI source. The
mass spectrometer was scanned from 10-500u in 0.5 sec. The collision gas pressure was
at 3 mtror and energy was at 40eV. An application of the neutral loss scanning mode is

presented later in the analysis of a spiked urine sample.

1. LC/MS Analysis

Isolation and separation of trace quantities of biologically important materials is
always a primary concern of an analyst. Sample losses during purification and extraction
steps are unavoidable. By using on-line separation techniques, the losses due to
purification and extraction steps can be minimized. By carefully selecting the correct
separation technique, the signals from trace materials can be maximized. In order to

evaluate on-line HPLC with ESI/MS, a mixture of hydrazones of model compounds was

 

183
analyzed by using a C13 column with acetonitrile and water mixture containing 0.1% TFA
and gradient of 30% to 60% acetonitrile over 20 min. The total flow of 200pL/min was
introduced into the ESI interface and the individual derivatives were identified as shown in
Figure 5-12 by scanning Q1 between 350 and 500 u in 0.5 sec. The reconstructed total
ion current (RITC) and mass chromatograms show that all the four derivatives can be
identified after HPLC separation. Also, the chromatograms (Figure 5-12) show that the
derivatives form isomeric compounds that have slightly difl‘erent retention times. The
sensitivity and separation can be improved by using small diameter (800 pm or smaller)

columns.

184

 

+
at
100 388
80 ~
§§c114
is?
.8
33
$3 404
204

 

 

200" .-256.V.Sdo,. 5507' '400 4361 VYSBO
m/z

Figure 5-11. The neutral loss scan mass spectrum of nandrolone hydrazone derivative.

185

 

 

 

 

 

 

100«
(e).m/z 388
so .
100, m
501 (d). m/z 402
g 100« m
E (c).m/z416
o 50‘
,>
33.
0
9‘ 100‘
50, (b). m/z 450
100
50‘ (a). RTIC
VVVVVVV 500- 10001 1500

Fig. 5-12. Total ion current (RTIC) chromatogram for LC/MS analysis of hydrazone
derivatives of model compounds (8) and mass chromatograms for (b) fluoxymesterone
hydrazone (m/z 450), (c) methyltestosterone hydrazone (m/z 416), (d) testosterone
hydrazone (m/z 402) and (e) nandrolone hydrazone (m/z 388).

186

IV. Feasibility study of Steroid Detection in Urine

Sample preparation methods and extraction procedures have been well
documented in the literature (5, 20-22). Samples were prepared by spiking 5 mL of urine
(obtained from a male volunteer with no previous history of anabolic steroid use) with a
standard solution containing the model compounds, nandrolone, methyltestosterone, and
fluoxymesterone. A blank urine samples and fortified urine samples (10ng/mL) were
extracted using C13 Sep-pak® cartridges (Waters, Millipore Corp., Milford, MA)
previously washed with 5 mL of methanol and 5 mL of water. The urine sample was
passed through the cartridge followed by 5 mL of water to remove water-soluble urinary
components. Steroids were then eluted with 2 mL of methanol and evaporated to dryness
under a nitrogen stream. The residue was dissolved in methanol (0.5 mL) and derivatized
with the Girard's T reagent. The reaction mixture was subjected to analysis by
electrospray mass spectrometry afier HPLC separation. The reconstructed mass
chromatograms (Fig 5-13) for the model compounds (lOng/mL) show the presence of all
the steroids added to the spiked urine sample.

187

 

 

 

 

 

100 -
(d) m/z 388
50 .
3.. 100 1
“ m/ 416
'5‘, 50« (c) z
.>
3
a W
100 ~
(b) m/z 450
50‘
100 '
(a) RTIC
50‘
"”566 1666 {566 I
scan#

Fig. 5-13. Total ion current chromatogram (RTIC) for LC/MS analysis of hydrazone
derivatives of urine extract (a) and mass chromatograms for (b) fluoxymesterone

hydrazone (m/z 450), (c) methyltestosterone hydrazone (m/z 416), and (d) nandrolone
hydrazone (m/z 388).

1 88
A. MS/MS Studies of the Derivatized Urine Extract

The neutral-loss scanning mode of operation discriminated against the matrices
which do not lose a fragment corresponds to the offset mass between Q1 and Q3. This
method is very useful in analyzing hydrazone derivatives because all the hydrazones lose
59u in their MS/MS spectra. If the components in the urine matrix do not lose 59u in the
tandem mass spectra, they will not appear in the final spectrum.

In order to evaluate the neutral loss scanning mode of operation for urine analysis,
a spiked urine sample was prepared with methyltestosterone, extracted and derivatized
with the Girard's T reagent. A urine sample was spiked with methyltestosterone
(lOng/mL) and allowed 1hr to equilibrate at room temperature. A 5mL portion of the
urine sample was extracted and derivatized as described in earlier section. The derivatized
reaction mixture was evaporated to dryness and reconstituted in SOOuL of methanol. The
final concentration of the hydrazone in solution was lOOpg/uL. The sample was infirsed at
a flow rate of SuIJmin and data were collected in the full scanning mode over one minute.
The signal due to the steroid hydrazone was very small and signals fi'om other compounds
dominated the normal ESI mass spectrum that averaged over one minute (Figure 5-14).
The same sample was subjected to neutral loss scanning of 59u and the spectrum on
Figure 5-15 was obtained. This clearly demonstrates the possibility of avoiding
interference fi'om complex matrices present in the biological samples by selecting a
suitable scanning mode of operation of the triple quadrupole mass spectrometer. In both
cases the sample was infirsed into the ESI ion source at a rate of 511L/min. The collision
offset energy for the neutral loss scanning experiment was set at 40eV and the pressure

was at 3mtorr.

189

 

 

 

 

 

 

 

377.1
100“
80* 481.4
g 60-
E
“:3
3.3
o .
d 40
M+
416.4
20. l

 

' 3160* 380 460' 420 440 4160' 480' 500
m/z

Fig. 5-14 Portion of the electrospray mass spectrum of extracted urine sample after
spiking with methyltestosterone (1 Ong/mL) followed by derivatization with Girard's T
reagent. The sample was infirsed at a flow rate of 511L/min without separation. The

spectrum is dominated by signals from the impurities in the urine sample.

190

 

+
M
416.3
100 a
80-
.? 604
E
.'.>:
.13
£2 40«
20-

 

 

 

360 380 400 420 440 460 480 500
m/z

Figure 5-15. The simplified mass spectrum of the sample used to obtain figure 5-14, by
neutral loss scanning (59 u) of the quadrupoles (Q1 and Q3). The collision energy was set

to 40eV and collision gas pressure at 3mtorr.

191
V. Kinetic Studies of the Hydrazone Formation Reaction
The coupling reaction between the keto group of the steroid and the Girard's T
reagent is very important part of the reaction for successful analysis by the methodology
described in this chapter. The Girard's T reagent reacts with the ketone in the presence of
an acid; usually acetic acid. The completeness of the reaction was monitored at different

temperatures and acid concentrations.

A. Dependence of Conversion (%) with Temperature

The coupling reaction rate of hydrazone formation was studied using testosterone.
Aliquots of 10 nmol of steroid in methanol were placed in a microfuge tube, supplemented
with 20-fold excess of Girard's T reagent and acidified with glacial acetic acid to 1M. The
samples were mixed by vortexing and incubated at 30°C, 40°C, and 50°C, respectively,
for times ranging from 15 min to 1 hour in a water bath. After incubation, the whole
reaction mixture was immediately injected into a pre-equilibrated RP-HPLC system.
Progression of the reaction was quantitated by monitoring the disappearance of
testosterone and the appearance of a new peak corresponding to the expected product as
characterized by the UV spectrum and elution time. Testosterone hydrazone has a
maximum at 280nm in the UV absorption spectrum. Figure 5-16 shows the dependence
of percent reaction completeness with time and temperature. As the temperature
increases, the rate of the hydrazone formation increases. As expected, the rate of reaction

increases with time to a constant value.

192

B. Dependence of Reaction Conversion (%) with Acid Concentration

The dependence of acid concentration on the reaction rate also was investigated.
Aliquots of 10 nmol of testosterone were placed in microfirge tubes and a 20-fold excess
of Girard's T reagent was added to each. The reaction mixture was acidified with glacial
acetic acid to bring the solution to 0.1M, 0.2M, 0.5M, 1M, 2M, or SM, respectively. The
samples were incubated at 37°C or 50°C for 30 min. The reaction mixture was injected
into a pre-equilibrated RP-HPLC system. Progress of the reaction was monitored as
described earlier. The rate of reaction greatly increases in the presence of an acid. As
shown in the Fig. 5-17, the rate of reaction becomes almost constant at acid
concentrations 2 1M. Wheeler (20) investigated the rate of reaction of a series of steroid
ketones in 0.2M acetic acid/methanol at 25°C and established the order of 3 > 6 >> A4’3
~ 7 ~ 17 > 20 > 12 >> 11, where the position of the ketone is indicated by number. The
rates of hydrolysis are in the order of 114-3 ~ 3 < 6 < 17 ~ 20 < 7 ~ 12 ~ 11 (26).
Previous workers indicated the possibility of increasing the rate of hydrolysis at high acid
concentrations (> 1M), but the hydrolysis constants for A4'3 steroids (most of the
anabolic steroids are A4'3) are very low as shown earlier by Wheeler (26). By considering
the above factors, 1M HOAC in the reaction mixture for 30 minutes is suitable for the

formation of hydrazones with testosterone analogs.
VI. Source CID during ESI/MS
Collisionally induced dissociation in ESI/MS can be carried out at the source by

decreasing the electric field (in positive mode) between the skimmer and the nozzle in

single quadrupole mass spectrometers. This technique allows the researchers to do CID

193
on a single quadrupole mass spectrometer fitted with an ESI source. In a Finnigan
TQMS, source CID can be induced by applying an additional potential to the RF-only
octapole. At the octapole region, the pressure is in the mtorr region and is suitable for
CID. Source CID can be induced by using an additional offset potential on the octapole
to accelerate ions in this high pressure region to undergo CID. In order to mass analyze
the source-induced fi'agments, the potentials on the entire

Methyltestosterone hydrazone was infused into the ESI source of the TQMS
which was operated in the single quadrupole mode. The ESI needle was at 4.5kV and the
heated capillary was at 200°C. The sample (lOng/uL) was introduced at a rate of
SuIJmin. The other voltages on the lenses were kept at normal tuned values. The
potential on the RF-only octapole was decreased from -3V to -43V and mass spectra were
acquired. The mass spectrometer was scanned from 100 to 50011 in 0.5s and data
averaged for one minute. Figure 5-18 shows the mass spectrum resulting fi'om the source-
CID of the methyltestosterone at the octapole region of the triple quadrupole mass
spectrometer. It provides the characteristic peaks observed in MS/MS experiment; loss of
59u and 87u. Therefore, source-CID can be used to induce fi'agmentation in a single
quadrupole mass spectrometer in order to obtain more structural characterization
combined with ESI.

A. MS/MS of the (M-59)+ Ion Derived fi'om Methyltestosterone Hydrazone

Source CID induces strong peaks representing loss of 59u (C1159) and 87u (0"-
87) fi'om the molecular cation of methyltestosterone hydrazone in TQMS instrument
operated in the single quadrupole mode. This gave a valuable information necessary to
characterize the analyte using a single quadrupole mass spectrometer. The source-CID of

194
electrosprayed ions adds another dimension to the capability of the TQMS. Therefore the
ions induced by source-CID can be used to generate another set of ions at the Q2 which
canbemassanalyzed in the Q3.

In this experiment source CID was induced by decreasing the voltage on the RF-
only octapole and a selected ion (m/z 357) was allowed to pass through Q1. At Q2,
collisionally activated dissociation was done via collisions with Ar gas (3mtorr). The
ofl‘set energy of the Q2 was held at 40eV. The Q3 was scanned from 100-500 11 in 0.5s.
The MS/MS spectrum of the ion at m/z 357 is shown in Figure 5-19.

B. MS/MS of the (INA-87)+ Ion Derived from Methyltestosterone Hydrazone

In a second experiment, a peak representing the ion at m/z 329 was allowed to
pass through Q1 for collection of the MS/MS spectrum of the ion at m/z 329 (see Figure
5-20). The ESI interface and mass spectrometer parameters were kept as earlier
experiment. The two spectra from the ions at m/z 357 and 329 are remarkably different
from each other. The MS/MS spectrum of m/z 357 did not give a peak at m/z 329.
Therefore, ions represented by the peak at m/z 357 have no contribution toward ions
represented by the peak at m/z 329. This technique provides valuable information
regarding the structure of the molecule and the selected ions. From this study, it can be
concluded that on a TQMS instrument, MS/MS/MS studies can be done on ESI produced
ions effectively. The MS/MS/MS data of methyltestosterone hydrazone chloride
indicates that ions corresponding to (M-59)+ and (M-87)+ derive fiom two different
fiagmentation pathways. These two ions have no common fragments and they undergo
fi'agmentation giving different product ions. The ion at m/z 329 (M-87)+ may be due to a
rearrangement product. The MS/MS spectra of all the steroid hydrazones show a strong

peak at (M-87)+, therefore the rearrangement reaction may be common to all hydrazones.

195

100—

 

% Reaction Conversion

 

o IIIIIUI'IIUITi‘ITUIT1IIUIIIIITI]

1 O 2 0 3 0 4 O 5 O 6 0 7 0
Reaction time/min

Figure 5-16. Variation of percent reaction conversion with time and temperature for

hydrazone formation

 

 

100—
- .A —'—' — —_ — —A
8 3 A/ '
L3 80‘: -Jo-———I/-
Q,‘ . ..
8 605 l
G . —V-— 50°C
:8 40- / +3700
0 .1
cu .
. -/
m 20.4 ,
s l
o l'U'rl'U'II'V'Il'l'rliii'lllel
0 1 2 3 4 5 6

HoAc concentration! M

Figure 5-17. Effect of HOAC concentration on percent reaction conversion

196i

 

 

 

 

 

 

 

 

+
(Ni-59)
357
100 i
80‘
E 60‘
35:” 132
o + +
.5 M
g (M'37)
and) 40* 416
329
20 ~ 151
299
100m“"26'6”"W'soofi'wfidofiflr
m/z

Figure 5-18. ESI/MS mass spectrum of methyltestosterone hydrazone chloride obtained

by inducing CID at the octapole. The octapole ofl‘set was set to 40eV.

100-

804

60-

Relative Intensity

4o~

20~

 

93

.7

197

162.8

150.9

108.1

 

 

 

 

 

 

 

177.1

 

 

 

vvvvv

m/z 416 (molecular cation)

l-S9u

357.4

 

Figure 5-19. The MS/MS spectrum of the ion at m/z (M-59)"' derived fi'om

methyltestosterone hydrazone chloride (MW of cation is 416u) by source C1D.

198

 

 

 

 

 

135.0
100 ~
123.0
80 ~
3‘
g 60 «
L5
0
.2
g m/z 416 (molecular cation)
a:
40 d l -87u
93.9
148.9
329.1
20 - 174.8
79.9 1 186.8
160. ' 260 ' 360 ' 460
m/z

Figure 5-20. The MS/MS spectrum of the ion at m/z (M-87)+ derived from

methyltestosterone hydrazone cation (m/z 416) by source CID.

1 99
VII. Oxime Derivative Formation

The formation of oxime derivatives to protect the keto functionality of steroids is a
well known reaction (16-18). In GC-MS analysis, methoxime derivatives are widely used.
The oximes are Schiff bases that can be protonated easily in acidic solution. The ESI/MS
is well known technique that ions in solution transfers to the gas phase. Therefore, neutral
steroids are not easily detected during ESI/MS because they do not exist as ions. After
forming oxime derivatives with hydroxylammine hydrocloride, they can exist as
protonated molecules in acidic solution and thus give a good response to ESI/MS.

In this reaction, shown in Figure 5-21 with testosterone, 10 nmol of steroid in
methanol was mixed with 200 uL hydroxylamine hydrochloride in pyridine (1mg/mL).
The reaction mixture was incubated at 70°C for 30 minutes and then dried under vacuum
using a SpeedVac® centrifuge concentrator. The residue was dissolved in water and
extracted with ethyl acetate. The excess reagent is water soluble. The extract was

injected into a pre-equilibrated HPLC system for further purification.

A. Results and Discussion

Oxime derivatives have been widely studied and used in GC/MS methods to
protect the keto functionality of steroids. Theonot et al,. studied the reaction kinetics of
oxime formation with oxymethalonone (3 6). The oximes are Schilf bases and can easily
be protonated in acidic solution. Under normal ESI conditions, when the steroid oxime
was introduced in CH3OI-I, a peak representing the protonated molecule was observed
with very little or no fiagmentation. Figure 5-22 shows the electrospray mass spectrum of
testosterone oxime in methanol (100pg injected). When oxime derivatives of the model

compounds were dissolved in acetonitrile and injected into a stream of

200

acetonitrile/HzO/HOAC (50/49.9/0.1%), a peak representing (M+CH3CN+H)+ was
observed in addition to a peak for the protonated molecule. A significant peak for
methanol adduct was not observed when methanol was used as the ESI solvent. Figure 5-
23 shows the ESI mass spectrum of testosterone oxime in CH3CN/HzO/HOAC
(50/49.9/0.l % v/v). ESI/MS spectra of several anabolic steroid oximes were acquired in
acetonitrile/HZO/HOAC. It was found that all the oximes show (M+CH3 CN+I-I)+ in
addition to the (M+H)+. The intensities of peaks representing (M+CH3 CN+H)+ and
(M’rl-I)+ ions varies with variables such as flow rates, concentrations, etc. This indicates
that the oxime molecules can form an adducts with acetonitrile molecule. Therefore, this
process leads to loss of sensitivity when ESI/MS is carried out in CH3CN/I-120 as the
solvent system. At the same time, this process can be used to achieve additional
selectivity by selecting proper solvents. Therefore, methanol is a better choice for the
analysis of oximes by ESI/MS to obtain (M+H)"' as the major signal.

The MS/MS analysis of (M+CH3CN+H)"' at a collision cell pressure of 3mtorr
shows dissociation to the protonated molecule at a very low offset energy of 5eV. A
significant fraction of acetonitrile solvent adduct can be dissociated in the presence of only
collision gas without an offset energy. Figure 5-24 shows the ESI/CID mass spectrum of
(M+CH3 CN+H)+ at 5eV offset energy and 3mtorr collision gas pressure.

The CID studies show that the oxime derivatives form a weak interaction with
CH3CN, but one strong enough to go through the capillary tube heated to 200°C. The
adduct can be dissociated into the protonated molecule using a very low collision energy
applied to it. Therefore, source C11) can help to decompose adduct ions formed with
ACN. By selecting proper voltages for source CID, only the protonated molecule can be
generated in the source region of the ESI/MS instrument. By this method, the abundance

of the protonated molecule can be increased.

201

 

HO — NH3+ C1' +

Hydroxylam'ne Hydrochloride Testosterone

+ HCl + H20

 

HO—‘N

Testosterone Oxime

Figure 5-21. Reaction scheme for the formation of testosterone oxime from testosterone

and hydroxylamine hydrochloride.

202

 

 

 

+
(M+H)
304.3

100 -

80—

g 60.
§
5
o
.2.
E

32 40-

20-

50 100 150 200 250 300 350
m/z

Figure 5-22. ESI mass spectrum of methyltestosterone oxime in CH3 OH/I-120/I-IOAC
(50/49.9/0.l % v/v/v)

203

 

 

 

+
(M+CH3CN+H)
345.1
100 -
+
(M+H)
304.2
80 ~
60 -1
i=1
0
,>
.8
34’ 4o —
20 ~

 

 

 

 

Figure 5-23. Portion of the ESI mass spectrum of methyltestosterone oxime in

CH3CN/1120/I-IOAC (50/49.9/0.1 % v/v/v)

204

 

 

 

 

+
(M+H)
304.3
100 ~
80
g 60-
‘5’
,3
E
34’ 40
201 +
(M+CH 3CN+H)
345.1
“2 1‘, €4-1VL- T - f; . ,At‘?‘f‘;“: - .A‘- :Ma-L, L-kaf 9%,—
100 200 300 400
m/z

Figure 5-24. ESI/CID/MS/MS spectrum of (M+CH3 CN+I-I)+ for testosterone oxime.

The ofl‘set energy was set at 5eV and the collision energy at 3 mtorr (Ar)

205
B. Tandem Mass Spectrometric Studies of Oximes

The protonated molecule of the testosterone oxime in CH3OH was subjected to
collisionally induced dissociation to obtain the fragmentation pattern shown in Figure 5-
25. The MS/MS spectrum of testosterone oxime in CH3 OH shows several peaks related
to the steroid skeleton that are common to the MS/MS spectra of oxime derivatives of
testosterone, methyltestosterone, and nandrolone. The fragmentation pattern of the
oximes did not yield much structural information that could distinguish homologous

steroids.
C. Detection Limits of Steroid Oximes

The purified testosterone oxime derivative was dissolved in methanol and
increasing amounts of derivative were injected into a stream of CH3 OH/HzO/HOAC
(50l49.9/0.1 % v/v) at a flow rate of 10 uL/min. The detection limits in the low
femtogram range (100 fg injected) were achieved in the selected ion monitoring mode of
the TQMS.

. 206

 

 

 

 

 

123.9
100 -
80, 111.91
13 60.
t:
8
.s
O
.>
3
£2 404
(M+H)+
304.1
20‘
137.9
VAWVL‘AAA’AA- rfiféll

 

Vfiw v T

50 "166' ' '1‘566 ' 266' ' 250 306' ' 330
m/z

Figure 5-25. ESI/CID/MS/MS spectrum of testosterone oxime

207

VIII. Investigation of Stanozolol by ESI/MS

A Introduction

Although the synthetic anabolic steroid stanozolol (ST) has been available to the
medical profession since 1961 and has been widely used owing to its reputation for very
high anabolic to androgenic properties(3 7), the available literature covering analytical
methodologies for its determination is limited. Reported methods for the determination of
stanozolol in pharmaceutical formulations include gas chromatography (GC) (3 8),
colorimetry (3 9), liquid chromatography/ultraviolet (LC/UV) (40), gas
chromatography/mass spectrometry (GC/MS) (41-43 ), isotope dilution mass spectrometry
(44), and high performance liquid chromatography/tandem mass spectrometry using
atmospheric pressure chemical ionization (APCI) (45). The standard procedure for
doping control in urine samples, based on enzymatic hydrolysis before derivatization and
subsequent analysis by GC/MS, has proven difficult for ST. Since ST is structurally
difi‘erent fi'om most anabolic steroids, its GC behavior is poor and its detection in urine is
relatively difficult. Its pyrazole ring condensed into the androstane ring system induces
difi‘ering GC behavior from that of related compounds.

Another reason for the difficulty in its detection is due to its low excretion in urine.
When the excretion of stanozolol was investigated with radio-labeled drug, only 16% of
the radioactivity was excreted in the urine during the first day, while 40-60% was excreted
in the feces (46).

Derivatization techniques for GC/MS include preparation of N,O-bis, trimethylsilyl
(N, O-bis-TMS) ether derivatives (47), the trimethylsilylether (OTMS) (46), the
heptafluorobutyrarnide (NHFB) (46) and methylation (48).

Masse et al. (49) and Donike et al. (43) extensively studied stanozolol metabolism
and identified eleven metabolites in human urine samples. All the metabolites were

identified as mono- or di-hydroxylated stanozolol analogs, most of which are excreted in

208
urine in the form of conjugates (49). The structure of stanozolol and its metabolites are
shown in figure 5-26. Recently, Masse et al. reviewed the analysis of anabolic steroids
including stanozolol using GC/MS.
In this part of the dissertation, the analysis of stanozolol using positive electrospray
mass spectrometry is described. The HPLC behavior, the extraction of a stanozolol-
spiked urine sample, and its analysis by ESI/MS are discussed.

B. Mass Spectral Analysis

Stanozolol (ST) (Sigma chemical Co. St. Louis, MO) was dissolved in methanol
(lOOpg/uL) and introduced into the ESI source. The ESI needle was held at 4.5kV and
the heated capillary tube at 200°C. The sample (SuL) was flow injected into a stream of
CH3OH/HZO/HOAC (50/49.9/0.1 % v/v/v) at a flow rate of 100 uL/min. ST produces a
strong signal at m/z 329 that corresponds to the protonated molecule. The ST molecule
has a basic site in its structure that causes difficulties during analysis by GC/MS. Due to
the basic N atoms in the pyrazole ring, ST undergoes protonation easily under acidic
conditions. The protonated molecule in solution is transferred to gas phase by
electrospraying the solution. Figure 5-27 shows the electrospray mass spectrum of
stanozolol that is an average of 5 scans. The mass spectrometer was scanned 200-500u in
lsec. Under normal ESI conditions stanozolol undergoes little or no fragmentation. The
ion current due to the ST is concentrated at (M+H)+, hence creating a sensitive and
selective detection method for ST. Since the stanozolol molecules can be selectively
detected by ESI/MS without forming derivatives, it will be very advantageous to use

ESI/MS to detect stanozolol in complex biological samples.

209

 

 

OH
“\s CH3
. 17 ““\
11
\ C D 16 <———
31

HN B
6

—>A>.—

F igne 5-26. Structure of ambolic steroid stanozolol, and preferred metabolic
hydroxylation sites indicated by an arrow.

210

 

+
(Mm)
329
100~
80‘
i? (it
3
.5
.3
E.
,8 40~
20~
~.‘ A fifi e '40:.L- fa“. .A‘.+Ae+

 

 

 

A A

200' '230 300 350 400 450 500

m/z

Figure 5-27. ESI/MS mass spectrum of stanozolol in CH3OH and injected to a stream of
CH3OH/H20/HOAC (50/49.9/0.1 % v/v/v) The capillary at 200°C and needle at 4.5kV.

211

C. Tandem Mass Spectrometry of Stanozolol

Tandem mass spectrometry is a very useful technique for structural
characterization of various molecules. The stanozolol metabolites are structurally similar
compounds containing different numbers of hydroxyl groups at difl‘erent positions (49,
50). Henion et al. reported MS/MS properties of stanozolol metabolites by atmospheric
pressure chemical ionization (APCI) mass spectrometry on a TQMS (45). The authors
studied the fragmentation of stanozolol and three other metabolites and observed different
CID tandem mass spectra. The tandem mass spectrometry technique can be used to
identify difi‘erent metabolites in urine samples.

In this study, standard solutions of stanozolol were used to study the ESI/MS/MS
properties. The sample solution (lOpg/uL) was infirsed into the ESI source at a rate of
4p.L/min. The collision gas pressure was adjusted to 3 mtorr and the ofi‘set energy to
55eV (Lab). The mass spectrum obtained by ESI/MS/MS of stanozolol is given in Figure
5-28. The suggested fragmentation pathways are also given in Figure 5-29 (45). The
fragrnentations are similar to those obtained by APCI mass spectrometry on the TQMS.
The fi'agment ion at m/z 81, resulting from cleavage of the A-ring between carbon atoms
C1/C2 and C3/C4 (possibly C1/C2 and C4/C5), dominates the tandem mass spectrum of
stanozolol. Increasing the collision gas pressure or acceleration energy led only to
changes in the abundances of the fragment ion at m/z 81 and parent ion at m/z 329; an
increase of relative abundances for the other fi'agment ions was not accomplished with
these changes. Henion et al. (45) compared the fragmentation patterns for 17-
deuteromethyl stanozolol and unlabelled stanozolol to determine the ring cleavage
positions, but they were not able to unambiguously assign the position of cleavage to give

a specific fi'agment.

212

81
100 ‘

80‘
g 60
§
5 95
.3
... +
% 107 (M+H)
M 40.

121 329
20‘
135

 

 

 

 

 

 

 

 

 

 

 

 

Figure 5-28. ESI/MS/MS mass spectrum of stanozolol. Collision energy was 55eV and

the gas pressure was 3mtorr.

213
950 13.213).

109 (127-18)
. 1---

  

 

1
1
1
1
1
1
‘ 1
1
1
1
1
1
1
1
a

  

1-- L43 (58+2H-16)
71(72-11)

 

   

 

 

H+ : ' “1 163
: 1‘ 121 (163-42)
: 135 ~,
---J : 93 (135-42)
43 --J 95 n},
67 93 ('ZH) 149
‘59 (57"2H) 107 (149-42)

Figure 5-29. The possible fragmentation pathways of stanozolol molecule
D. Comparison of ESI/MS/MS with FAB/MS/MS

The high-energy F AB/CAD linked scan mass spectra were acquired on a JOEL
HX 110 double focusing mass spectrometer. The CID energy was lOkV and He was used
as the collision gas. The FAB/CAD/MS/MS mass spectrum (figure 5-30) shows a loss of
16 and 18 mass units from the protonated molecule, while ESI/MS/MS on the TQMS did
not give an intense peak for a loss of H20. The higher mass end of the spectrum is
significantly different from each other. ESI/MS/MS on the TQMS gave very weak peaks
at high mass whereas the FAB/CAD/MS/MS spectrum shows a series of intense peaks at
high mass. In the ESI/MS/MS spectrum of stanozolol from the TQMS, there is a series of

intense peaks at the lower mass end of the spectrum.

214

 

 

 

 

 

(M+H)+
10"
R r 32.3
e 4
1 4
a 8‘
t 4
1 ‘ (NI‘l'I'I-I‘12('))+
v 4
e 6"
A 4
b 4
U ‘d
n 4
d l '
I 4
n 2. 257.3
c 110.9
‘ J‘ 13 .9
9 J
1 .. 1 . ‘ l7 '0 ‘l' ‘
v r ' ' 1 fi' '
50 100 150 200 250 300 “/2350

Figure 5-30. FAB/CAD/MS/MS mass spectrum of stanozolol

215

The ESI/MS/MS spectrum of stanozolol is similar to that obtained by APCI mass
spectrometry on TQMS, but different from the FAB/CAD/MS/MS spectrum obtained on
the double focusing mass spectrometer. These differences are due to the different energies

employed in the collisionally induced dissociation process.
B. Effect of Solvent Composition

When stanozolol is dissolved in methanol and analyzed by ESI/MS, the major peak
in the mass spectrum is due to (M+H)+. A significant peak for (M+CH3 OH+H)+ was not
observed as shown in figure 5-27. When the mobile phase was changed to
CH3CN/HzO/HOAC (50/49.9/O.1 % v/v) system, there were two major peaks observed
in the ESI mass spectrum of ST, corresponding to (M+H)+ and (M+CH3CN+H)‘*' as
shown in figure 5-31. The association of acetonitrile (CH3CN) as an adduct ion is weak
but stable at the ESI/MS conditions used in this experiment.

F. ESI/MS/MS of CH3CN Adduct Ion of ST

The CH3CN adduct ion of ST was isolated at the Q1 and collected the CID mass
spectrum at the Q3. The solvent adduct ion dissociates into the protonated molecule in
the presence of 2 mtorr Ar gas in the collision cell (Q2), even with a small collision offset
energy (5eV) as shown in figure 5-32. Even without the collision energy, a significant
fiaction of (M+CH3 CN+H)+ ion current dissociates into the (M+H)+ ion. The adduct ion
also can be dissociated into the protonated molecule by increasing the ofi‘set voltage by
20V on the octapole of the TQMS. Therefore, it is possible to form only (M+I-I)"' as the

216
major ion in the ESI/MS mass spectrum by inducing source-CID in the region before the

skimmer in a single quadrupole mass spectrometer.

 

 

 

+
(M+ACN+H)
370
100‘
80‘
g 60-
r:
o
E
Q)
.2
E
0
a: 40‘
+
20* (101+H)
329
,1, fl ,,,,,,,,,,, 11L,” 1 L 1
150 200 250 300 350 400

Figure 5-31. Portion of the ESI/MS mass spectrum of stanozolol in CH3OH as injected
into a stream of CH3CN/I-IZO/HOAC (SO/49.9/0.1 % v/v/v)- The capillary was at 200°C
and needle at 4.5kV.

217

 

 

 

 

 

+
(M+H)
329
100
801
g 601
E
32’
E
g" 40
20+
+
(M+CH), CN+H)
L 370
'''''' 1'66" ”H.20OTVYHH3IO071H40O
m/z

Figure 5-32. The ESI/MS/MS mass spectrum of the CH3CN adduct ion of ST. The ofl'set

energy was at 5eV and the collision cell pressure was at 2 mtorr.

21 8
G. HPLC Analysis of Stanozolol

The screening of stanozolol was carried out with a 4.6mm id x 15cm column
packed with C13 particles (All tech, Deerfield, IL). The elution profile was monitored at
220nm. Two solvent systems were tested for reasonable peak shape. The
methanol/water/TF A (50/49.9/0.1 % v/v/v) solvent system at a rate of lmI/min gave a
broad peak for ST whereas CH3CN/HzO/TFA (50/49.9/0.l % v/v/v) at a rate of lmL/min
gave a good peak at retention time of 4min. Therefore the CH3CN solvent system is
better choice for a reasonable peak shape and resolution.

H. LC/MS of Stanozolol

Biological samples are mixtures of various compounds. Separation techniques
coupled with mass spectrometry provides the necessary selectivity, sensitivity, and
reliability required to identify mixtures. HPLC coupled with mass spectrometry provides
the capability to analyze thermally labile compounds without extensive derivatization.

A Vydac C13 column (2 mm x 25 cm.) was used in LC/MS analysis of stanozolol.
A portion of 10 11L of stanozolol (100ng 11L) was injected and an isocratic run of
CH3CN/I~120/'I'FA (40/59.9/0.1 % v/v) at a rate of 200 al./min was used to develop the
chromatogram. Figure 5-33 shows the mass chromatograms for m/z 329 (M+I-I)+ and
m/z 370 (M+CH3CN+H)+. The both peaks were present in the chromatogram and the
intensity of the peak at m/z 370 is higher than that at m/z 329 in CH3CN. A tailing of the
elution profile was observed due to the basic nature of the pyrazole ring of the stanozolol
molecule. The sensitivity of stanozolol on the ESI/LC/MS was reduced using acetonitrile

as the solvent due to the formation of adduct ion with ST. However, this ACN protocol

2 19
can be used to obtain additional selectivity by monitoring ions representing both (M+I-I)+

and (M+CH3CN+I-I)+ in the ESI/MS mass spectrum.

 

 

 

100
m/z 329
50 '
'3
t:
0
H
G LA MA; A A
1—1 ._- -.. A L.
2 100
.15
34’
m/z 370
50‘
fi* """" s """"" s """""" s """"" s """"" I"

 

 

 

Figure 5-33. The reconstructed mass chromatograms for (M-l-I-I)+ (m/z 329) and
(M+CH3CN+H)+ (m/z 370) for stanozolol (1ng injected) during an LC/MS run.

220
I. Urine Extract

A sample of urine from a male volunteer was collected and a 5mL of portion was
spiked with stanozolol standard (Sng/mL). The sample was allowed to equilibrate for
about one hour before extraction. A C13 cartridge (Waters, Milford, MA) was activated
using 5 mL of methanol; and excess methanol was washed with 2 mL of water. The
sample (5 mL) was passed through the cartridge at a rate of 20 mL/min and washed with 5
mL of water to remove water soluble urinary components. The cartridge was washed
with 0.1M NH40H to improve the extraction efficiency as described by Massé et al (49).
During elution of urine on the Sep-pak C13 cartridge, some of these compounds would be
partially retained or adsorbed owing to specific interactions with the silica core of the
stationary phase. Higher recoveries were obtained when treated with 0.1M NH40H after
passing urine through the column and before elution of stanozolol (49). The stanozolol
was eluted with 5 mL of methanol. The extract was evaporated to dryness under a stream
of N2 and reconstituted in 500 uL of methanol.

The urine extract was subjected to selected reaction monitoring (SRM) on the
TQMS. The SRM is a selective technique to identify compounds based on specific
fragment ions. The selected product ions can be monitored in the SRM mode of the
TQMS. In this study, the most abunth product ion, m/z 81 was chosen for SRM. The
following reactions were monitored in the SRM mode of the TQMS to detect stanozolol
in the spiked urine sample; m/z 329 —> 329, and 329 —-> 81. Figure 5-34 shows the
SRM profile of the above reactions for the urine extract. A portion of lOuL of urine
extract was injected into a stream of CH3OH/H20/HOAC (50/49.9/O.1 %v/v) at a rate of
100 uUmin. The SRM mass spectrum of urine extract indicates the presence of
stanozolol in the urine extract. The complete MS/MS mass spectrum of stanozolol was

collected using the urine extract. The appearance of the spectrum was similar to that of

221
the standard sample. Next, the detection limits of ST were determined using positive
electrospray mass spectrometry. It was observed that ST can be detected at the low

femtogram level using selected reaction monitoring mode.

IX. Conclusions

In this report, the feasibility of forming oxime or hydrazone derivatives of anabolic
ketosteroids for electrospray detection has been described. Under normal ESI conditions,
hydrazone derivatives of ketosteroids gave only a single peak in the mass spectrum
representing the intact molecule. The feasibility has been demonstrated for detecting these
derivatives at the low-femtogram level in biological fluids. The hydrazone formation
reaction was investigated in detail because it is a promising candidate for method
development in determining anabolic ketosteroids in biological fluids. Tandem mass
spectrometry (MS/MS) combined with ESI/HPLC/MS is a valuable tool for identifying the
steroids present in biological fluids by combining fragmentation patterns with retention
time and molecular weight information. Collisionally induced dissociation studies of
hydrazone derivatives demonstrated a major loss of 59u fiom the molecular ions of model
compounds. The characteristic loss of 59u can be used in the neutral loss scanning mode
to obtain additional selectivity for the analyte. The neutral loss scanning mode can be
used to discriminate against the biological matrix as shown above. This method has a high
potential for detecting low levels of anabolic ketosteroids present in biological samples.

The oxime derivatives of anabolic steroids show an enhanced response in the
ESI/MS analysis. The selection of CH3CN for the solvent system for oxime analysis is
very important since they form an adduct with acetonitrile. The source CID can be used

to decompose the adduct ions to obtain (M+I-I)+ as the major ion.

222

The stanozolol molecule has a very basic pyrazole ring which makes analysis by
GC/MS difficult. In ESI/MS, the pyrazole ring can be protonated in acidic solution and
transferred to the gas phase without significant decomposition. Stanozolol can be
detected with high sensitivity using ESI/MS. Detection limits at the low femtogram level
were achieved with the selected reaction monitoring mode. Since stanozolol forms an
adduct with acetonitrile, care must be taken to select the correct solvents for analysis.
Tandem mass spectrometry combined with liquid chromatography can be used to confirm

the presence of stanozolol in biological samples.

223

 

 

 

 

 

 

g SRM329 ——>31
0
.5 50.
O
.2.
1.5
0
ad
50 100 150 200 250 300
100 H
,5; SRM329 —> 329-
,3, 50«
E
0
..>.
.‘é’
&
so 100 150 200 250 300

scan #

Figure 5-34. Selected reaction monitoring profile of stanozolol in urine extract (lOuL).

The solvent system was CH3OH/HzO/HOAC (50/49.9/O.l %v/v/v) at a rate of

lOOuL/min.

224

X. References

10.

ll.

12.

l3.

14.

R H. Strauss, Drugs and Performance in Sports, W. B. Saunders, Philadelphia
(1987)

T. Tobin, Drugs and Performance in Horse, Charles C. Thomas, Springfield,
Illinois (1981)

H. A. Haupt and G. D. Rovere, Am. J. Sports Med 1984, 12, 469
W. M. Muck and J. D. Henion Biomed Environ. Mass Spectrom. 1990, 19, 37

Sample, R. H. B. and Baenziger, J. C. In Analytical Aspects of Drug Testing, D
Deutch, ed., Wiley, New York, pp 247-272 (1988)

Goudreault, D.; Masse, R. J. Steroid Biochem. Molec. Biol. 1990, 3 7, 137

Gaskell, S. J.; "Analysis of Steroids" In Methods of Biochemical Analysis, Vol. 29,
D. Glick, ed., Wiley (interscience), New York, NY, pp 385-434 (1983)

Curcuruto, 0.; Franchi, D.; Hamdan, M.; Favretto, D.; Traldi, P. Rapid Commun.
Mass Spectrom. 1993, 7, 673

Shacketon, C. H. J. Steroid Biochem. Mol. Biol. 1993, 45, 127

Watson, D.; Taylor, G. W.; Murray, S. Biomed Environ. Mass Spectrom. 1986,
13, 65

Weidolf, L. O. G.; Henion, J. D. Biomed Environ. Mass. Spectrom. 1988, 15,
283

Kletke, C.; Roitman, E.; Bitsch, P.; Shaldeton, C. Proceeding of 40th ASMS
Conference on Mass Spectrometry and Allied Topics, Washington DC, May 31-
June 5, 1992

Hatton, C. K.; Catlin, D. H. Clin. Lab. Med, 1987, 7, 655-668

D. Goudreault; R. Masse J. Steroid Biochem. Molec. Biol. 1990, 37, 137

15.

16.

17.

l8.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

225
B. Chung; H-Y. P. Choo; T. Kim; K. Born; 0. Kwon; J. Suh; J. Yang; J. Park J.
Anal. Toxicol. 1990, 14, 91

Honour, J. W.; Brooks, C. J. W.; Shackleton, C. H. L. Biomed Mass Spectrom.
1982, 9, 505

Bjorkhem, 1.; Ek, H. J. steroid Biochem. 1983, 18, 481
Bjorkhem, 1.; Lantoo, 0.; Lof, A. J. steroid Biochem. 1980, 13, 481

Moss, M. S.; Cowan, D. A , in Clarke's Isolation and Identification of Drugs,
2nd Ed, ed. by Mofi‘at, A. C, Pharmaceutical Press, London 1986

Masse, R.; Ayotte, Dugal, R. J. Chromatogr.1989, 489, 23
Shakleton, C. H. L.; Whitney, J. 0. Clin. Chim. Acta 1980, 107, 231

Kruger, T. L.; Litton, R. W.; Kondrat, R. W. Cooks, R. G. Anal. Chem. 1976, 48,
2113

K. Kayganich; J. T. Watson; C. Kilts; J. Richie Biomed Environ. Mass. Spectrom.,
1990, 19, 341

H. K. Mayer; W. Reusch; J. T. Watson Org. Mass Spec. 1992, 27, 560
Girard, A.; Sandulesco, G. Helv. Chim. Acta. 1936, 19, 1095
Wheeler, 0. H. J. Chem. Edu. 1968, 45, (6)435

Bruins, A. P.; Covey, T. R.; Henion, J. D. Anal. Chem. 1987, 59, 2642

Conboy, J. J.; Henion, J. D.; Martin, M. W.; Zweigenbaun, J. A. A. Anal. Chem.
1990, 62, 800

Fenn, J. B.; Mann, M.; Meng, C. K; Wong, S. F.; Whitehouse, C. M. Mass
Spectrometry Reviews, 1990, 9, 37

Kondradt, R. W.;Cooks, R. G. Anal. Chem. 1978, 50, 81A
McLarferty, F. W.; Bockhofl‘, F. W. Anal. Chem. 1976, 50, 69
Yost, R. A.; Enke, C. G. Anal. Chem. 1979, 5], 1251A

Glish, G. L.; Shaddock, V. M.; Harmon, K.; Cooks, R. G. Anal. Chem. 1980, 52,
165

34.

35.

36.

37.
38.

39.

40.

41.

42.

43.

45.

46.

47.

48.

226
Brotherton, H. 0.; Yost, R. A Anal. Chem. 1983, 55, 549
Johnson, J. V.; Yost, R. A Anal. Chem. 1985, 5 7, 758A
Thenot, J-P.; Homing, E. C. Anal. Lett. 1972, 5(1), 21

Bates. P. C.; Chew. L. P.; Milward. D. J. Endocrinol. 1987, 114, 373
Magin. D. F. J. Chromatogr. 1975, 115, 687

Shingbal. D. M.; Barad. U. G.; J. Assoc. Ofii Chem, 1985, 68, 98,

Cavrini. V.; Pietra. M.; Raggi. M. A; Mioli. M. G. Analyst, 1987, 112, 1671
Rossi, S.-A.; Johnson, J. V.; Yost, R. A Biol. Mass Spectrom. 1992, 21, 420
Schanzer, W.; Donike, M. Anal. Chim. Acta. 1993, 23, 275

Schnzer, W.; Opfermann, G.; Donike, M. .1. Steroid Biochem. 1990, 153, 36
Lantto, 0.; Bjorhem, 1.; Ek. H.; Johnson, D. J. Steroid Biochem. 1981, 14, 721
Muck. W. M.; Henion. J. D Biomed Environ. Mass Spectrum. 1990, 19, 37
Choo, H-Y. P.; Kwon, O-S.; Park, J. J. Anal. Toxicol. 1990, 14, 109

Masse. R.; Ayotte. C.; Bi. H.; Dugal. R J. Chromatogr. 1989, 497, 17

Schanzer, W.; Opfermann, G.; Donike, M. Proceedings of 36th ASMS Conference

on Mass Spectrometry and Allied Topics, San Francisco, CA, June 5- June 10,
1988, 397