.1 53:1)1
w v: e I:

5

 

3:.

m.

Vin

 

—

 

“a,

an:

..
. .62.:

55.

k?

at:
:1 A?

%

.. .2 . .

i.
2

e
.
x

133.12%, ,0
.erw.§
.maflh‘

mm
V ,
LI

. G1,.

hfiufi

n3.

.lyP.al:vnl,
I

9.1}
s

~
42...... L...

 

LIBRARY
Michigan State

:2 University
2w , -

WES

This is to certify that the
thesis entitled

ANALYSIS OF ALKALI METAL-CATIONIZED
PHARMACEUTICALS USING ELECTROSPRAY IONIZATION
TANDEM MASS SPECTROMETRY

presented by

SUSAN LYNN ACHBERGER

has been accepted towards fulfillment
of the requirements for the

Master of degree in Biochemistry and Molecular
Science Biology

 

 

Major Profess/or’ Signature
8/“? 202’ 37

Date

MSU is an aflinnative-action, equal-opportunity employer

. u.—nfi.—.-.----u—n-n-g-u-u-u-n-ga-

.u—v--—.-c--o-o------u-a--.—-.-.-._.—----¢!.-n—u-n

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K‘IProi/Acc8PresIClRC/DateDue indd

ANALYSIS OF ALKALI METAL-CATIONIZED PHARMACEUTICALS
USING ELECTROSPRAY IONIZATION TANDEM MASS SPECTROMETRY

By

Susan Lynn Achberger

A THESIS
Submitted to
Michigan State Univeristy

in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE
Department of Biochemistry and Molecular Biology

2008

ANALYSIS OF ALKALI METAL-CATIONIZED PHARMACEUTICALS
USING ELECTROSPRAY IONIZATION TANDEM MASS SPECTROMETRY

By

Susan Lynn Achberger

The effect of alkali metal cations (Li+, Na+, and K+) on the collision-
induced dissociation of the pharmaceuticals ciprofloxacin, levofloxacin, and
erythromycin was investigated using electrospray ionization tandem mass
spectrometry (ESI-MS/MS). For both levofloxacin and erythromycin, alkali
metal cationization yields molecular fragments that were not obtained using
protonation. In the case of erythromycin, lithium and sodium cationization yields
fragmentation of the macrolide ring, thus providing information about this
particular structural feature of the molecule, which was not obtained for
protonated samples. Data also indicate that the varying ion mobility of H+, Li+,
Na+, and K+ yields different fragmentation pathways. These experiments

demonstrate that alkali metal cationization can provide a method for the structural

determination of pharmaceuticals and their metabolites.

DEDICATION

To my parents Bill and Jan Achberger
and my furry friend Skip Achberger,
for your unconditional love and support

And to my undergraduate adviser,
Dr. Norman J. Wells
(May 5, 1948 — January 7, 2006),
for everything you did for me

iii

ACKNOWLEDGMENTS

To my advisor, Dr. A. Daniel Jones: Thank you for helping/putting up with a stubborn,
misguided graduate student. You taught me so much and I cannot thank you enough for
your patience and compassion.

To Dr. Ruth Waddell-Smith: Thank you for serving on my committee and for your
willingness to provide a listening ear. Most importantly, thank you for being a mentor
and a friend. You were with me from the very beginning and I could not have done any
of this without you.

To Dr. Robert Hausinger: I have greatly enjoyed working with you as a committee
member and professor. Your time and input are very much appreciated.

To the members of the Jones Laboratory, especially Siobhan Shay, Michael Stagliano,
and Ruth Udey: I am so fortunate to have such wonderful coworkers who are equally
wonderful friends. Thank you for all of your help and support. Please accept a delicious
bass as a sign of my gratitude.

To my fellow MSU graduate students, Julie Bordowitz, David and Heather Dotzauer,
Kate Higginbotham, Aaron McBride, Eric Moellering, Danielle Nevarez, and Aggie
Steiner: I am so thankful that I have wonderful friends who shared this crazy journey
with me. Thank you for your friendship, encouragement, and unfaltering support.

To my friends, Kimberly Moherrnan, Jessica Davila, Anita Dasu, Laura Johns-Fowler,
Stephanie Schuster-Maglis, and Alexis Witt: Even though we were miles apart
throughout my graduate school experience, I never felt far from you. I appreciate your
friendship, guidance, and “pep talks,” and I cannot express how gratefiil I am for all you
have done for me.

To the congregation of the Okemos Community Church, especially the members of the
handbell and chancel choirs: Thank you for all of your prayers and support. You were
my family during my time at MSU and I am very blessed to have you in my life. You
also served as a constant reminder that I am loved and I could not have done this without
your love and support. Words cannot express my gratitude.

iv

Table of Contents

DEDICATION ................................................................................................................... iii
ACKNOWLEDGMENTS ................................................................................................. iv
List of Figures .................................................................................................................... vi
List of Tables ...................................................................................................................... x
Key to Abbreviations ......................................................................................................... xi

CHAPTER 1: INTRODUCTION AND BACKGROUND

Background ..................................................................................................................... 1
Introduction to Mass Spectrometry ................................................................................. 5
Electrospray Ionization (ESI) ...................................................................................... 6
Tandem Mass Spectrometry (MS/MS) and Collision-Induced Dissociation (CID) 8
Cationization ............................................................................................................. 12

CHAPTER 2: COLLISION-INDUCED DISSOCIATION MASS SPECTROMETRY
OF METAL CATIONIZED XENOBIOTICS AND THEIR METABOLITES

Introduction ................................................................................................................... l7
Methodology ................................................................................................................. 22
FIA-ESI/MS/MS of Protonated and Cationized Pharmaceuticals ............................ 22
Results ........................................................................................................................... 24
F luoroquinolones: Ciprofloxacin and Levofloxacin ................................................ 24
Protonated Species ................................................................................................ 24
Alkali Metal-Cationized Species .......................................................................... 34
Discussion ............................................................................................................. 44
Erythromycin ............................................................................................................ 46
Protonated Species ................................................................................................ 46
Alkali Metal-Cationized Species .......................................................................... 49
Discussion ............................................................................................................. 58

CHAPTER 3: CONCLUSIONS AND FUTURE WORK

Summary and Conclusions ........................................................................................... 61

Future Work .................................................................................................................. 63
APPENDIX I

Pseudo-MS3 spectra for Chapter 2 ................................................................................ 64
APPENDIX H

Tables of Relative Abundances for Mass Spectra in Chapter 2 ................................... 7O
BIBLIOGRAPHY ............................................................................................................. 77

List of Figures

Figure 1. Structures of omeprazole and two isomers of its oxidative metabolites [1, 2]... 2

Figure 2. Schematic of the mass spectrometry process ..................................................... 5
Figure 3. Schematic of an ESI source [11, 12] .................................................................. 6
Figure 4. A schematic of the E81 mechanism, shown in positive ionization mode [11, 12;
Figure 5. Schematic of the tandem mass spectrometry process ......................................... 9
Figure 6. Schematic of the collision-induced dissociation (CID) process ....................... 10
Figure 7. The possible fragmentation pathways for molecule [A-B]+ ............................. 11
Figure 8. Structure of cylindrospermopsin ...................................................................... 14
Figure 9. The backbone structure of a ginsenoside with R groups corresponding to
hydroxyl or sugar groups .................................................................................................. 15
Chapter 2

Figure 10. Chemical structures of the fluoroquinolone antibiotics ciprofloxacin (left)
and levofloxacin (right) ..................................................................................................... 18
Figure 11. Chemical structure of the macrolide antibiotic erythromycin. ....................... 20

Figure 12. CID product ion spectra for protonated levofloxacin (m/z 362) at collision cell
potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V using argon as a collision gas 25

Figure 13. CID product ion spectra of m/z 364 (A+2) for levofloxacin in D20 at collision
cell potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V ......................................... 25

Figure 14. Proposed fragmentation pathway for protonated levofloxacin ...................... 26

Figure 15. CID product ion spectra for protonated ciprofloxacin (m/z 332) at collision
cell potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V ......................................... 29

Figure 16. CID product ion spectra of m/z 335 (A+3) in D20 at collision cell potentials
of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V ................................................................ 29

vi

Figure 17. Proposed fragmentation pathways for protonated ciprofloxacin .................... 30

Figure 18. Pr0posed fragmentation pathway of m/z 314([M+H-H20]+ generated by in-
source CID of ciprofloxacin ions. ..................................................................................... 32

Figure 19. CID product ion spectra of lithium-cationized ciprofloxacin (m/z 338) at
collision cell potentials of (A) 35 V, (B) 30 V, (C) 25 V, and (D) 10 V .......................... 35

Figure 20. CID product ion spectra of m/z 340 (A+Li+2) in D20 at collision cell
potentials of (A) 35 V, (B) 30 V, (C) 25 V, and (D) 10 V ................................................ 35

Figure 21. Proposed fragmentation pathways for lithium-cationized ciprofloxacin ....... 36

Figure 22. CID product ion spectra of lithium-cationized levofloxacin (m/z 368) at
collision cell potentials of (A) 35 V, (B) 30 V, (C) 25 V, and (D) 10 V .......................... 38

Figure 23. CID product ion spectra for m/z 369 (A+Li+1) for lithium-cationized
levofloxacin D20 at collision cell potentials of (A) 35 V, (B) 30 V, (C) 25 V, and (D) 10
V ........................................................................................................................................ 38

Figure 24. Proposed fragmentation pathways of lithium-cationized levofloxacin .......... 39

Figure 25. Proposed mechanism for the loss of hydrogen fluoride from levofloxacin,
which is not observed for protonated species ................................................................... 40

Figure 26. CID product ion spectra of sodium-cationized ciprofloxacin (m/z 354) at
collision cell potentials of (A) 35 V, (B) 30 V, (C) 25 V, and (D) 10 V ......................... 41

Figure 27. CID product ion spectra of sodium-cationized levofloxacin (m/z 384) at
collision cell potentials of (A) 35 V, (B) 30 V, (C) 25 V, and (D) 10 V .......................... 41

Figure 28. CID product ion spectra of m/z 385 (A+Na+1) for sodium-cationized
levofloxacin in D20 at collision cell potentials of (A) 35 V, (B) 30 V, (C) 25 V, and (D)
10 V ................................................................................................................................... 42

Figure 29. CID product ion spectra of potassium-cationized ciprofloxacin (m/z 370) at
collision cell potentials of (A) 25 V and (B) 10 V, and potassium-cationized levofloxacin
(m/z 400) at (C) 25 V and (D) 10 V .................................................................................. 43

Figure 30. CID product ion spectra of erythromycin (m/z 734) at collision cell potentials
of(A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V ................................................................ 47

Figure 31. CID product ion spectra of m/z 739 (A+5) of erythromycin in D20 at collision
cell potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V ......................................... 47

vii

Figure 32. Proposed fiagmentation pathway for protonated erythromycin ..................... 48

Figure 33. CID product ion spectra of sodium-cationized erythromycin (m/z 756) at
collision cell potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V .......................... 50

Figure 34. CID product ion spectra of m/z 761 (A+Na+5) for sodium-cationized
erythromycin in D20 at collision cell potentials of (A) 55 V, (B) 40 V, (C) 20 V, and (D)
10 V ................................................................................................................................... 50
Figure 35. Proposed fragmentation pathway for sodium-cationized erythromycin ........ 51
Figure 36. Mechanism for the loss of 114 u from the macrolide ring of erythromycin .. 52
Figure 37. Proposed fiagmentation pathway of m/z 309 ................................................. 53

Figure 38. CID product ion spectra of lithium-cationized erythromycin (m/z 740) at
collision cell potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V ......................... 54

Figure 39. CID product ion spectra of m/z 745 (A+Li+5) for lithium-cationized
erythromycin in D20 at collision cell potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D)

10 V ................................................................................................................................... 54
Figure 40. Proposed fragmentation pathways of m/z 582 ............................................... 55
Figure 41. A portion of the proposed fragmentation pathway for lithium-cationized
erythromycin ..................................................................................................................... 57
Figure 42. CID of potassium-cationized erythromycin (m/z 772) at collision cell
potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V ................................................ 58
Appendix I

Figure 43. CID product ion spectra of m/z 318 for protonated levofloxacin at (A) 55V,
(B) 40 V, (C) 25 V, and (D) 10 V ..................................................................................... 65

Figure 44. CID product ion spectra of m/z 261 for protonated levofloxacin at (A) 40 V,
(B) 25 V, and (C) 10 V ..................................................................................................... 65

Figure 45. CID product ion spectra of m/z 314 for protonated ciprofloxacin at (A) 55 V,
(B) 40 V, (C) 25 V, and (D) 10 V ..................................................................................... 66

Figure 46. CID product ion spectra of m/z 288 for protonated ciprofloxacin at (A) 55 V,
(B) 40 V, (C) 25 V, and (D) 10 V ..................................................................................... 66

viii

Figure 47. CID product ion spectra of m/z 231 for protonated ciprofloxacin at (A) 25 V
and (B) 10 V ...................................................................................................................... 67

Figure 48. CID product ion spectra of m/z 294 for lithium-cationized ciprofloxacin at
(A) 35de (B) 10V ...................................................................................................... 67

Figure 49. CID product ion spectra of m/z 324 for lithium-cationized levofloxacin at (A)
25Vand (B) 10V ............................................................................................................. 68

Figure 50. CD) product ion spectrum of m/z 309 for sodium-cationized erythromycin at
10 V ................................................................................................................................... 68

Figure 51. CID product ion spectra of m/z 582 for lithium-cationized erythromycin at
(A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V .................................................................... 69

ix

List of Tables

Appendix II

Table 1. Relative abundances for product ion spectra of alkali metal-cationized
erythromycin; average values obtained from three sample replicates (values obtained
using Waters MassLynx software) .................................................................................... 71

Table 2. Relative abundances for product ion spectra of alkali metal-cationized
erythromycin; average values obtained from three sample replicates (values obtained
using Waters MassLynx software) .................................................................................... 72

Table 3. Relative abundances for protonated levofloxacin; average values obtained fi'om
three sample replicates (values obtained using Waters MassLynx software) ................... 73

Table 4. Relative abundances for alkali metal-cationized levofloxacin; average values
obtained from three sample replicates (values obtained using Waters MassLynx software)
........................................................................................................................................... 73

Table 5. Relative abundances for protonated ciprofloxacin; average values obtained from
three sample replicates (values obtained using Waters MassLynx software) ................... 75

Table 6. Relative abundances for lithium-cationized ciprofloxacin; average values
obtained from three sample replicates (values obtained using Waters MassLynx software)
........................................................................................................................................... 76

CID
CYP
ESI
FIA

GC

LC
Met+
MS

MS/MS

Key to Abbreviations
Collision-induced dissociation
Cytochrome P450 (metabolic enzymes)
Electrospray ionization
Flow-injection analysis
Gas chromatography
Hydrogen-deuterium exchange
Liquid chromatography
Alkali metal cation
Mass spectrometry

Tandem mass spectrometry

xi

CHAPTER 1: INTRODUCTION AND BACKGROUND

Background

Metabolomics, the term given to assessing the entire array of metabolites,
provides a powerful tool for the potential treatment and diagnosis of diseases and the
discovery and development of new pharmaceuticals. In the pharmaceutical industry,
comparing the metabolome of an organism with and without the presence of a drug is
critical for the discovery of new pharmaceuticals. This will allow for the elucidation of
the structure of the drug’s metabolites, as well as the determination of the effect of the
drug on biological pathways. Because many enzymes are capable of metabolizing more
than one compound, this can lead to drug-drug interactions, during which one drug
interferes with the metabolism of another, leading to adverse effects. As a result, the
requirement for analytical techniques that can analyze metabolites in order to determine
their molecular structure has greatly increased. In this project, mass spectrometry, a
valuable technique for the characterization of metabolites, will be used to provide
information regarding the molecular structures of various pharmaceutical substrates, with
the aim of developing tools useful for elucidation metabolite structures.

One of the greatest challenges involved in metabolite analysis and metabolic
technology development is the distinction between isomeric and isobaric metabolites.
Oxidation is one of the most common metabolic transformations, and many compounds
can undergo metabolic oxidation at multiple sites. For example, S-hydroxyomeprazole
and omeprazole sulfone are the main metabolites in human in vitro metabolism of

omeprazole, catalyzed by the cytochrome (CYP) 450 enzymes CYP3A4 and CYP2C19

respectively. Both of these metabolic alterations cause a mass shift of +16 units, making
distinguishing these two structures challenging [1, 2]. Also, because various isomers can
have different toxicological properties, structural information about these compounds can
provide information about the potential toxicity of a compound. For example, the drug
thalidomide was administered as a racemic mixture to pregnant women in the 19605 to
prevent morning sickness. Thalidomide was later determined to be teratogenic and this

toxic effect was eventually attributed to one of thalidomide’s hydroxylated metabolites,

though the exact metabolite was not specified [3, 4].

 

Omeprazole sulfone S-Hydroxyomeprazole

Figure 1. Structures of omeprazole and two isomers of its oxidative metabolites [1, 2]

Drug metabolism can also lead to adverse drug-drug interactions. CYP 450
enzymes are considered the most important first step in metabolism of most drugs and
humans express 18 families and 43 subfamilies of CYPs. CYP families 1, 2, and 3 are
responsible for xenobiotic metabolism and are the determinants of elimination and

bioavailability of pharmaceutical compounds. Specifically, CYPs 1A2, 3A4, 2D6, 2C9,

and 2C19 metabolize over 90% of human drugs [5]. CYP-dependent metabolism can
increase the efficacy of a drug (bioactivation), promote elimination of the metabolite
from the body by forming more polar metabolites, or can generate electrophilic
metabolites that exhibit toxicity. Drugs and other environmental and endogenous
substances can lead to the inhibition or induction of CYP enzymes, which influences
drug-drug interactions in patients undergoing multi-drug therapy [5].

One example of this phenomenon is the adverse interactions between haloperidol,
an antipsychotic drug, and valerian, an herbal sleep aid. The findings of Dalla Corte et a1.
[6] indicate that adverse drug-drug interactions take place between valerian and
haloperidol, leading to oxidative stress in the liver. They propose that valerian and other
herbal supplements can change the metabolism of haloperidol through the inhibition or
induction of CYP450 enzymes. CYP3A4 and 2D6 have inductive effects on valerian in
vitro, and such alterations can affect the disposition of other drugs that are taken
concurrently [7].

Umathe et a1. [8] observed a similar phenomenon with the compounds quercetin
and pioglitazone. Quercetin, often contained in herbal add-on therapy for diabetes,
inhibits CYP3A4, which is responsible for metabolizing the antidiabetic drug
pioglitazone. Since quercetin inhibits CYP3A4’s ability to metabolize pioglitazone, the
bioavailability of pioglitazone increases, thus lowering the rate of hepatic clearance.

CYP3A4 is involved in the metabolism of 60% of all therapeutically used drugs
and its levels in the human liver can be increased or decreased because of a patient’s
exposure to numerous drugs [6, 10]. This can create a problem during the pharmaceutical

development process. If the same enzyme is responsible for the metabolism of many

drugs, this can lower the efficacy of the respective drugs because they are all competing
to interact with that particular enzyme. Determination of CYP-mediated metabolism of
new candidate drugs, including identification of specific sites of enzyme-catalyzed
metabolism, is necessary during the development of pharmaceutical compounds to ensure
the safety and efficacy of those compounds. As a result, improved methods for
metabolite structure elucidation are needed in order to determine the metabolic pathways
responsible for biotransformation of that compound, and to determine whether toxic

metabolites are generated.

Introduction to Mass Spectrometry

Mass spectrometry (MS) is the dominant analytical technique that is used to
determine the molecular weights of chemical compounds during the drug development
process. This technique can also be used to help elucidate the structures of drugs and
their metabolites. In mass spectrometry, ions are separated based upon their mass-to-
charge ratio (m/z). These ions are created through acquisition of a charge, such as
protonation or cationization, by neutral molecules [9]. Figure 2 presents a schematic of

the mass spectrometry process.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Sample introduced Ions enter
into ionization mass
source analyzer
Ions separated according _
' ’ to mass/charge ratio ' Detector
Ionization
Source Mass Analyzer
Computer 4

 

 

 

 

 

Figure 2. Schematic of the mass spectrometry process

Upon introduction into the instrument, analyte molecules undergo ionization in
the ion source. The resulting ions are propelled by electric fields into the mass analyzer
and then separated based upon their mass-to—charge ratios through various combinations
of electric and magnetic fields. The detector enables the mass spectrometer to amplify an

electrical signal from the incident ions, convert the ion current to voltage, and convert the

analog voltage to a digital measure of ion abundance. That digital measure of ion signal

is then transferred to a computer that can be used for data processing [9].

Electrospray Ionization (ESI)

Electrospray ionization (ESI), an ionization technique used to generate gas-phase
ionized molecules from a liquid solution, has emerged during the past few years as an
essential tool for the analysis of proteins, polymers, and small polar molecules. It allows
for detection of femtomole quantities and is easy to couple to liquid chromatography

(LC) [10]. Figure 3 presents a schematic of an ESI source.

 

 

Skimmer
To mass
'1]
&Ca 1 E81 droplets/spray spectrometer
———> Analyte flow > >
—
‘———

Cone Nitrogen flow

Figure 3. Schematic of an ESI source [11, 12]

To generate these ionized molecules, a fine spray of highly charged droplets is created in
the presence of a strong electric field. In ESI, which occurs at atmospheric pressure, a
strong electric field is applied to a liquid passing through a capillary tube. Figure 3

illustrates the mechanism of ESI, which is described below.

Positively

 

  

 

 

charged
Spray .needle mo ecule Analyte
tr molecules
Taylor Cone
.. 0 ' 0 o . Solvent Coulombic 5. °-,
0. . .0 evaporation .0. O . . . 0.. explosion '0’. . . ..o.
i. : —-—* -: a 3 ——> 0*
.0.....O. .000. .+
0":
charged Counter
molecule electrode/Cone
Electrons .
High-voltage Electrons
power supply >

 

 

 

 

 

 

 

Figure 4. A schematic of the E81 mechanism, shown in positive ionization mode [11, 12]

The electric field causes charge accumulation at the surface of the liquid that is emerging
from the end of the capillary tube. The excess charge in the liquid solution distorts the
liquid shape to form a Taylor cone, and electrostatic repulsion drives the liquid to break
apart to form highly charged droplets. Then, the droplets travel through a heated
capillary or a curtain of inert gas, typically nitrogen, where collisions cause evaporation
of solvent molecules. The droplets shrink and their charge per unit volume increases as
the solvent evaporates. Ions are then generated as the charged molecules are desorbed
from the droplet’s surface [12, 15].

ES] typically yields ions via deprotonation (negative ions), or via protonation, or
cationization to form positive ions. In the case of protonation, a proton will be added to a

molecule to yield a net positive charge of 1+ for each proton that is added. For example,

large analyte molecules, such as proteins, can have multiple charges if multiple ionizable
sites are present.

ESI is considered a “soft” ionization technique because intact molecular ions can
be produced with minimal fragmentation [13]. This stands in contrast to traditional
electron ionization which usually generates many ions consisting of fi'agments of the
original molecule. Because ESI occurs at atmospheric pressure, minimal energy is
transferred to the analyte during ionization, and little or no fragmentation occurs in the
ion source. As a result, mass spectra of most compounds show only protonated or
cationized ions in positive ion mode. While this is useful for molecular weight
determination, these mass spectra lack fragment ions that give information regarding

structure.

Tandem Mass Spectrometry (MS/AIS) and Collision-Induced Dissociation (CID)

More structural information can be obtained using tandem mass spectrometry
(MS/MS), which involves the activation of a particular precursor ion, generation of
fiagment ions, and the mass analysis of the fragmentation products. Tandem mass
spectrometry refers to any method that involves two stages of mass analysis that are
combined with either a dissociation process or a chemical reaction that can alter the
charge or mass of an ion. Figure 5 is a schematic of the tandem mass spectrometry

process.

Ion Source

Quadrupole 1: Select precursor
ion

Precursor Ion

Quadrupole 2 (collision cell):
collide precursor ion with
neutral gas

Product Ions
Quadrupole 3: Product ion
scan

Figure 5. Schematic of the tandem mass spectrometry process

In one kind of MS/MS experiment, referred to as tandem mass spectrometry in space, the
first mass analyzer isolates a precursor ion of a specific m/z that will subsequently
undergo fragmentation to give product ions and neutral fragments. A second mass
analyzer is used to analyze the masses of the product ions. The number of steps in
tandem mass spectrometry can also be increased. For example, a user can select ions of a
first particular mass, then select ions of a second particular mass from the fragments that
were obtained, and analyze the fragments of the last selected ions. The quantity of these
steps can be augmented to give MS" experiments, where n is the number of generations
of ions being analyzed [10].

One way to conduct MS/MS experiments is to couple two physically distinct
instruments. This technique is known as tandem mass spectrometry in space. Tandem in
space instruments include two mass analyzers. Frequently, a type of analyzer called a

quadrupole is used in these types of instruments. The triple quadrupole, or QqQ

configuration, is an instrument that consists of three quadrupoles. This configuration was
developed at Michigan State University by Chris Enke and Richard Yost in the 1970s
[14]. The first quadrupole is used to select the precursor ions, the second quadrupole, q,
is the collision cell that performs no mass filtering, and the third quadrupole analyzes the
product ions [10]. This is shown in Figure 5. In the collision cell, ions selected by
quadrupole 1 are activated by collisions with neutral gas molecules. This technique is

known as collision-induced dissociation (CID), which is illustrated in Figure 6.

 

 

Collision Gas Fragment Ions
V l
Precursor Ions . ../.v. . . . . Fragment Ions (Product Ions)
° 'V
—-—-—> : °, :
a ..’0 .°__.’ ..__—.—L’
O .\ O O O O O

 

Activated Ion \ \

Fragmenting Ion Collision Cell

Figure 6. Schematic of the collision-induced dissociation (CID) process

When the precursor ion collides with a neutral target gas, the kinetic energy of the
precursor ion is converted into internal energy, causing subsequent fragmentation [15].
The ion will fragment when the internal energy has accumulated in the appropriate
vibrational modes of the bonds in the molecule to allow fragmentation reactions to occur

on the time scale of the ion’s transit through the collision cell.

10

Fragmentation of an activated ion can occur through multiple reaction pathways.
Figure 7 shows an example of two possible fragmentation pathways that can occur for the

CID of the ionized molecule A-B.

[A-B]+

Pathway \atlgway

A++B A+B+

Figure 7. The possible fragmentation pathways for molecule [A-B]+

The fragmentation observed in a mass spectrum is dictated by the amount of internal
vibrational energy in the precursor ion and the activation energy barrier that must be
overcome in order to produce the observed fragments. For example, if the activation
energy barrier for Pathway A is lower than that of Pathway B, and if the internal energy
is too low to yield an appreciable rate of reaction for Pathway B, the fragmentation
reaction that generates the A+ ion will occur more rapidly than through Pathway B, and
only the A+ ion will be observed as a fragment. Portion B of the molecule, which is lost
as a neutral species, will not be observed in the mass spectra. Increasing the internal
energy may provide enough energy for reaction through Pathway B to occur, but the
lower activation barrier to Pathway A may result in its product dominating the spectrum
regardless of experimental conditions. In such situations, if molecule A-B was a

metabolite, any structural features present on portion B would not be observed in the

11

form of fi'agment ions. In order to observe the B+ ion in the mass spectra, the activation
energy needed to enable the molecule to dissociate via Pathway A needs to be raised to
allow Pathway B to become competitive. A method to control and enhance the
fragmentation pathways of a molecule is needed in order to identify any metabolic
structural modifications that have occurred, and to determine the position of those

modifications.

Cationization

Metal cationization, when used with ESI mass spectrometry (ESI-MS), is an
ionization mechanism that can provide enhanced molecular structure information that
often cannot be obtained utilizing traditional protonation. A charged complex is formed
during the cationization process, in which a positively charged ion such as an alkali metal
ion forms a complex with a neutral molecule [9]. When using ESI, metal ion adducts of
organic molecules are observed. Analytes that lack basic ftmctional groups are often
difficult to protonate, whereas they readily form complexes with alkali metal ions. Metal
cationization is easy to perform, requiring only the addition of a salt to the electrosprayed
solution. Only a small amount (pmoles or less) of the sample is needed. Metal
cationized ions may fragment through different pathways compared to protonated ions
[16]. Attachment of alkali metal cations can localize charge on different sites on the
molecule than would be the case for proton attachment, owing to different affinities of
functional groups for protons compared to metal cations. Some fragmentation reactions

are driven by proximity to charge whereas others occur remote fiom charge attachment.

12

Altering the balance between pathways and encouraging charge-remote fi'agmentation
has been used to determine the structure of lipids [17, 18].

Charge-remote fragmentation occurs when gas-phase decompositions, like those
observed in CID, take place at a location that is physically remote from the charge site on
the molecule. Protonated molecules ofien fragment largely via charge-directed
fiagmentation. The proximity of a positive charge can draw electron density from nearby
bonds, weakening them and encouraging fragmentation. Since protons are mobile owing
to their low mass and high velocities and vibrational frequencies, proton migration can
induce fi'agmentation to occur at multiple locations in an ion [10, 19]. Because alkali
metal cations have much greater masses than a proton, their velocities are lower at the
same vibrational energies and they are less able to migrate across molecules to facilitate
multiple fragmentation reactions. For example, Lopes and coworkers conducted
fragmentation studies on the antibiotic monensin A, and determined that protonated and
sodium-cationized samples have different fragmentation pathways [20-22]. This method
has also been used in CID experiments involving sugars [23] and fatty acids [24-27]. For
example, lithium cationization, coupled with CID, had been used to determine the type of
carbohydrate linkage, such as a [31 —) 4 linkage between two glucose molecules [28], as
well as the location of double bonds in fatty acids, esters, and alcohols [29, 30].

Ddrr et a1. [31] used ESI-MS" to investigate the effect of differential ion mobility
of IF, Li+, Na+, and K“ on the fragmentation pathways of cylindrospermopsin, a

zwitterionic toxin produced by cyanobacteria (Figure 8).

13

 

C5H5N203 .-
OH
H
o N O\s//
l O// \9
HN HN /N
o HN

Figure 8. Structure of cylindrospermopsin

Upon comparison of protonated and cationized spectra, they observed a difference in the
balance of terminal ring elimination between the two species (this pathway becomes
more prominent as the mass of the cation increases) when ion intensity ratios for
appropriate fragment ions were compared. Also, the lithiated spectra showed more
diverse fragmentation than sodium-cationized species, but no proposed structures or
mechanisms were shown for the lithiated cylindrospermopsin. Potassium cationization
yielded unique spectra as well, but again, no proposed structures or mechanisms were
discussed for potassium-cationized species. As Dorr et a1. discuss, their proposed
fragmentation pathway 1, which involves a loss of C5H6N203, becomes more prominent
as the atomic mass of the charged species increases. As a result, Dorr et al. determined
that the higher atomic mass of the migrating cations can alter the fragmentation pathways
observed in CID spectra, but their work failed to provide detailed explanations for these
differences.

Cui et al.[16] also used ESI-MS" to investigate the effect of metal (LiI, Na+, KI,

Ag+) cationization on the CID of ginsenosides, the compounds that give ginseng its

14

pharmaceutical properties. The general structure of a ginsenoside is shown in Figure 9,

where the respective R groups can be various sugars or hydroxyl groups.

R3
0H R4

\

 

R2
Figure 9. The backbone structure of a ginsenoside with R groups corresponding to

hydroxyl or sugar groups
They determined that the species of metal ion and the structure of the ginsenoside affect
the interaction between the ginsenoside and the metal ion. Specifically, the number of
oxygen atoms present in each ginsenoside affects the metal ion coordination. For
example, they observed that lithium-cationization yields a greater array of fragment ions
when compared to the other metal-cationized species. This indicates that, because if its
small mass, lithium can move around the surface of the molecule and cause bond
cleavage at different locations in the molecule, yielding a wider range of charge-directed
dissociation reactions than sodiated species. They also observed that sodium-cationized
species have lower fragmentation efficiencies than lithium-cationized species. Since
sodium has a greater atomic mass than lithium, its velocity is lower, and it is less likely to
migrate across the molecule to the same extent as lithium. As a result, fewer

fragmentation reactions are observed with sodium cationization. Fewer fragmentation

15

reactions were observed for potassium—cationized species for the same reason. Also, Cui
and coworkers determined that, as the size of the metal ion increases, the number of
oxygen atoms in the ginsenoside that are able to coordinate to the metal ion increases,
which fixes the charge in fewer locations and leads to more prominence for charge-
remote dissociation.

The work conducted by Cui et al. and Ddrr et al. demonstrates that cationization
can yield different fragmentation pathways than those observed with protonation.
Applying this concept to the example in Figure 7, cationization could be used to raise the
barrier to Pathway A and direct ion A-B+ to fragment via Pathway B. Because different
cations can guide fragmentation to occur through different pathways for the same
molecule, this technique can be applied to the MS analysis of pharmaceutical and
biological compounds to aid structure determination. Cationization will provide a
method to control the fragmentation of pharmaceutical compounds and thus determine

the location of structural modifications that have occurred during metabolism.

16

CHAPTER 2: COLLISION-INDUCED DISSOCIATION MASS SPECTROMETRY

0F METAL CATIONIZED XENOBIOTICS AND THEIR METABOLITES

Introduction

Mass spectrometry (MS) is the primary tool used for identification of xenobiotic
compounds and their metabolites owing to the low detection limits afforded by this
technique. The typical approach for MS-based structure characterization involves
conversions of molecular ion species into fragment ions and determination of the
fi'agment ion masses. The power of MS lies in its ability to provide critical information
about molecular masses, and the presence of molecular substructures can be established
based upon masses of fragment ions.

One of the primary limitations to the use of MS for structure elucidation involves
limited control of the fragmentation process. Some ions are resistant to fragmentation
owing to large activation barriers to formation of fi'agment ions. In contrast, other
molecular ions undergo a single facile fiagmentation that leads to formation of a
dominant fragment ion that only conveys information about a single structural feature.
Examples of this include losses of neutral water or carbohydrate groups, yielding
fragments that present minimal information about molecular structure. To address these
limitations, improved mass spectrometry methods are needed to enhance and control
fragmentation. The most common ions generated during LC/MS analyses are protonated,
or [M+H]+ ions. Protonation of groups such as alcohols provides a low activation barrier,
compared to other fragmentation pathways, to elimination of groups such as water.

Formation of ions other than via protonation offers the possibility of increasing activation

17

. barriers, and thus. allowing a greater fraction of the ionized molecules to fragment via
other reaction pathways.

In this Study, protonated and alkali metal-cationized antibiotics ciprofloxacin,
levofloxacin, and erythromycin were generated using electrospray ionization to determine
whether metal cationization could be used to control the ion fragmentation dynamics.
Ciprofloxacin and levofloxacin (Figure 10) belong to a class of antibiotic compounds
known as fluoroquinolones which inhibit topoisomerases involved in bacterial DNA
metabolism [32]. Some fluoroquinolones have caused serious toxic reactions, leading to
their withdrawal from the market or restrictions on their use. Over a wide range of doses,
ciprofloxacin and levofloxacin are two of the best-tolerated and safest fluoroquinolones
[32]. Even though ciprofloxacin is one of the safest fluoroquinolones, when administered
concurrently with the asthma medication theophylline, it can inhibit CYP 450 enzymes,

thus blocking the metabolism of theophylline in vitro [33].

HO HO i

Figure 10. Chemical structures of the fluoroquinolone antibiotics ciprofloxacin (left)
and levofloxacin (right)

Fragmentation of protonated and metal ion-cationized erythromycin (Figure 11), a
macrolide polyketide antibiotic, was also investigated in this study. While erythromycin

is an effective antimicrobial compound, it has also been shown to cause liver dysfunction

18

[34, 35]. For example, Karthek and Casson [35] reported a case of erythromycin-
associated liver dysfimction in a ten-year-old girl. They state that erythromycin transient
liver dysfunction is rarely reported in the pediatric literature, but it has been commonly
reported in adults. While the exact mechanism of erythromycin-associated liver injury is
unknown, one hypothesis is that a metabolite of erythromycin could cause an
immunological response, thus resulting in liver injury.

Erythromycin and its analogs are metabolized by CYP4SOS, which can lead to the
formation of stable metabolic intermediate complexes with CYP4505 [36, 37]. These
complexes are inactive and thus inhibit the metabolism of other drugs that are taken
concurrently [37]. Also, metabolism of erythromycin by various CYP450 enzymes
creates isomers because multiple cites of hydroxylation are possible. Different CYP450
isoenzymes may hydroxylate erythromycin at different positions. To establish which
CYP4SOs might be involved in erythromycin metabolism, there is a need for analytical

tools that can distinguish sites of hydroxylation.

19

   

OH

Figure 11. Chemical structure of the macrolide antibiotic erythromycin.

Development of a suitable technique for the structural characterization of
xenobiotic metabolites is important as this could provide useful information about
involvement of specific CYP450 isoenzymes with xenobiotic biotransformation, and
could serve as indicators of likelihood that a xenobiotic could interfere with the
metabolism of other endogenous or exogenous compounds

Collision-induced dissociation MS/MS spectra of the protonated and cationized
forms of levofloxacin, ciprofloxacin, and erythromycin were compared in this study.
Because alkali metal cations have a greater mass than a proton and will therefore migrate
between functional groups more slowly, metal cationization increases energy barriers to
migration of the attached cation throughout the analyte molecule and can direct

fragmentation to occur via alternate reaction pathways. It is anticipated that

20

fragmentation directed to occur through different reactions will generate structurally

useful fragment ions that would not form analogous ions for protonated molecules.

21

Methodology

FIA-ESI/MSMS of Protonated and Cationized Pharmaceuticals

Standards of erythromycin, levofloxacin, ciprofloxacin, and all alkali salts were
purchased from Sigma—Aldrich. Standards were prepared in MilliQ water/HPLC grade
methanol (50:50, vzv) at a concentration of 10 M. The water/methanol solution was also
used as a blank. Cationized samples were prepared as above, with 10 [LL of a 1 M
solution of lithium acetate, sodium acetate, or potassimn acetate added to each sample to
give a metal salt concentration of 10 mM.

Samples were analyzed by flow-injection analysis (FIA) ESI-MS/MS. Injection
volume was 20 uL ESI mass spectra and CID spectra were obtained using a Quattro
Premier XE triple quadrupole mass spectrometer (Waters, Milford, MA). Samples were
introduced into the electrospray source from an Acquity Ultra Performance LC Waters)
and analyzed in positive ion mode with a cone voltage of 30 V. Solvents used were
0.15% formic acid in water and acetonitrile (50:50) with a flow rate of 0.2 mLmin'l for
erythromycin, and a flow rate of 0.1 mLmin'1 for the fluoroquinolone samples. The
[M+H]+, [M+Li]+, [M+Na]+, or [M+K]+ ion was selected as the parent ion for CID. For
CID experiments, collision voltages of 10, 25, 40, and 55 V were used. Collision
voltages of 10, 25, 30 and 35 V were used for lithiated and sodiated fluoroquinolone
samples. Argon was used as the collision gas at a pressure of 2.73 X 10'3 mbar.

Pseudo-MS3 experiments were conducted for lithium- and sodium-cationized
samples of levofloxacin, ciprofloxacin, and erythromycin as above with a cone voltage of

95 V used for MS analysis. These experiments use in-source collisional activation to

22

generate fragment ions, and these fi'agment ions can be induced to fi'agment finther in the
instrument's collision cell. This information assists with assignments of fi'agmentation
pathways that proceed through specific ion intermediates. Parent ions selected for CID
were based upon proposed fragment structures illustrated in Figure 14, Figure 17, Figure
21, Figure 24, Figure 35, and Figure 41. These ions were selected to confirm the
proposed fragmentation pathways.

Hydrogen-deuterium (H/D) exchange experiments were conducted to verify the
proposed fragmentation mechanisms for protonated and cationized samples of each
respective compound. These experiments indicate the number of exchangeable
hydrogens present on the analyte molecule and its fragments, thus suggesting which
hydrogen atoms are involved in migrations that lead to fragmentation. Samples were
prepared as above with the use of 99-atom% deuterium oxide (Isotec, Inc.) instead of

water. Samples were analyzed as described previously.

23

Results

F laoroquinolones: Ciprofloxacin and Levofloxacin

Protonated Sjecies

CID spectra for levofloxacin are shown in Figure 12 and Figure 13, and the
proposed fragmentation pathways are depicted in Figure 14. Similarly, CID spectra for
ciprofloxacin are shown in Figure 15 and Figure 16 and the proposed fragmentation
pathways are depicted in Figure 17. CID spectra shown in the aforementioned figures
serve as the primary data for the proposed pathways, and the rationale underlying these

conclusions is described below.

24

va 1_42-1

 

 

 

110ct2007SLA_11 5 (0.108) Cm (1:46) A 4: Daughters of 362ES+
205 2.5865
100
194
.° 20
. \ 1' 3 219E21231
0.11 1.1.1.2134, . .1233, 215.252.2951-.-. 286.,
110c12007SLA_11 5 (0.102) Cm (1:46) B 3: Daughters of 362ES+
10 205 261 5.61e5
0] 219221
a\°«193 203
0.11.11,2.‘,§..,.2393?.L33551725? ,,.-.?§§ 3.7.3.218222391..- $119219 ,, ..... 24,4, , .,
110ct2007SLA_11 6 (0.118) Cm (1:47) c 2: Daughters of 362ES+
261 2.84e6
100
318
as
0 .. 205 219221 233 241247255 266270273 298.301 - 3151 . 314 362
110c12007SLA_11 6 (0.113) Cm (1 :47) D 1: Daughters of 362ES+
362 6.24e6
100
,,\°
318

 

950" 266“ 216" 22'6""236'246 "256" 26'0""2io "236'" zéb'w'é'bd'" 3'16 """ 3'20" 336""346 350360 "3‘13"z
Figure 12. CID product ion spectra for protonated levofloxacin (m/z 362) at collision cell
potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V using argon as a collision gas

Levofloxacln D, 1_88-7

 

 

23May20088LA_09 12 (0.258) Cm (1:45) A 4: Daughters of 364ES+
100 206 2.6585

I195195 20.,
a °\°' ‘ » I

204 l 221223

Girl, 131,319.11, '23? , , ,261 g ,
23May20088LA_09 13 (0.275) Cm (1 :46) B 3: Daughters of 364ES+
100 263 9.27e5

 
   

233235246248 262 268274 302

 

 

o W""TWTWWWW’TWWW
23May20088LA_09 12 (0.248) Cm (1:47) C 2: Daughters 01364ES+
263 4.02e6
100 320
..\° 262
01 222 235 24a 1, . 317 344 364
23May2008$LA_09 13 (0.264) Cm (1 :47) o 1: Daughters of 364ES+
1 364 1.01e7
°\°
320

A
vvvvvvvvv

 

Figure 13. CID product ion spectra of m/z 364 (A+2) for levofloxacin in D20 at collision
cell potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V

25

 

Y\O (\‘N/ WAC (\R/
N ”\J -co2 N Nd
.. I O = l O
F
o o

F
O
m/z 362 m/z 318
- C3H7N
YO
N I MA
- C3H4 F
O - C3H40
m/z 261
OH

F
m/z 221 O m/z 205

- CO

OH
H +
N I HNA

F
m/z 193

Figure 14. Proposed fi'agmentation pathway for protonated levofloxacin

CID of protonated levofloxacin (m/z 362) yields loss of water (18 units (u)) from

the carboxyl group to give m/z 344 (not shown) or loss of C02 (44 u) to give m/z 318.

26

This loss of CO2 occurs with simultaneous migration of the hydrogen from the carboxyl
group to the fluoroquinolone backbone. CID of deuterated levofloxacin (m/z 364,
[M+D]+ for levofloxacin—dj) also exhibits a loss of 44 u, demonstrating that the
deuterium migrates from the carboxyl group to the quinolone backbone, as
levofloxacin’s only exchangeable hydrogen is located on the hydroxyl group. The
fragment at m/z 261 can be explained by loss of C3H7N (57 u) fi'om m/z 318. Pseudo-
MS3 experiments generated product ion spectra for m/z 318 (Appendix I, Figure 43),
which showed a loss of 57 u from m/z 318. Loss of 57 11, which is attributed to loss of
C3H7N, is presumed to occur by fragmentation of the piperazine ring, as this is the
region of the molecule most likely to contain such hydrogen-rich functionality. CID of
deuterated levofloxacin also exhibits a loss of 57 u, suggesting fragmentation of the
piperazine ring as no exchangeable hydrogen atoms were involved in this fragmentation
(no exchangeable hydrogens are present on the lost portion of the piperazine ring
suggesting that an exchangeable hydrogen migrates to remain on the aziridine ring).

The levofloxacin fragment ion at m/z 261 undergoes fiuther fragmentation to give
m/z 205 or m/z 221 in pseudo-MS3 experiments (Appendix I, Figure 44). The latter is a
loss of 40 u, attributed to C3H4, which would require hydrogen migration fiom the
morpholine ring to the quinolone backbone. The former is the result of a loss of 56 u,
attributed to loss of C3H40, which would require migration of two hydrogen atoms fi'om
the morpholine ring to the rest of the molecule. The aliphatic portion of the morpholine
ring contains six hydrogen atoms. A fragment ion at m/z 205 was also obtained for
ciprofloxacin, which will be discussed later in this chapter (Figure 17). D’Agostino et

al. [38] observed a similar phenomenon in their CID experiments with levofloxacin

27

(referred to as ofloxacin) and ciprofloxacin, as well norfloxacin and enoxacin. They
observed that various fluoroquinolones studied exhibited similar fragmentation behavior
to one another during ESI-MS analysis, including neutral losses of water, C02, the
cyclic substituent from the pyridine ring, and fragmentation of the piperazine ring. All
fluoroquinolones analyzed fragment to yield m/z 205 (or m/z 206 in the case of
enoxacin, which contains a nitrogen atom in the fluoroquinolone backbone), the
structure of which is shown in Figure 14. This demonstrates that these fluoroquinolones

fragment to yield a common structure.

28

Ciprofloxocln 1_85-1

 

13ApIiI2008SLA_10 13 (0.280) Cm (1246) A 4: Daughters 01 32253"
203 6.2665

100
231

I \°

205 215 I
0.11.. hale/2.1.7.- .222 1. .1 .fi 41 we mm-

13Apr1120088LA_10 12 (0.253) Cm (1246) B 3: Daughters of 332ES+
231 2.7466

100 203

 

'314

WWW

13Apri|20088LA_10 13 (0.269) Cm (1:47) C 2: Daughters of 332ES+
2

100 45 4.10e6
314

e\° 204 231 268 288
oi (H 212217219227 | 233 240 , 1271 2351 j 234 332

 

13Apri|20088LA_10 14 (0.285) Cm (1 :47) D 1: Daughters 0133ZES+
100 332 1.87e7

. ., . . .. ...... .. . .. 1.- -22.??? .2 .. 2.311.... ., . we.-. .m/z

200 210 220 230 240 250 260 270 230 290 300 310 320 330 340

Figure 15. CID product ion spectra for protonated ciprofloxacin (m/z 332) at collision
cell potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V

vvvvvvvv

Ciprofloxacln D, 1_88-4

23May2008$LA_11 13 (0.280) Cm (1:46) A 4: Daughters 6133655+
205 4.92e4

100

. g 206 232

0.111II11, 21721917231 333 1

vvvvv

 

rrfirrfirvr r r'rvvr

23May2008$LA_11 13 (0.274) Cm (1 :46) B 3: Daughters of 335ES+
100 232 2.3465
205

 

W

 

 

23May20088LA_11 14 (0.291) Cm (1:47) C 2: Daughters 01335ES+
100 247 4.22e5
,248 291
\a 315
° 206208 271 316
229 233 246
0,-1lltl,2‘5.21.9., 111,243.13- . . 11 . ?8‘1,.291 . , . -1. -- . 315-.
23May20088LA_11 13 (0.264) Cm (1:47) D 1: Daughters 01335ES+
335 2.87e6
100
,,\°
231

v‘rvv

 

200 W210 ' 220 ' 230 240 I 260" 260' "’270 I 260 . 200 300 ' 310 ' 320 ’ 330 ' 340"”
Figure 16. CID product ion spectra of m/z 335 (A+3) in D20 at collision cell potentials
of(A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V

29

o o — 002
m/z 332
- H20

0”“ '

N N
I o
m/z 288
I)? F - C2H5N
O m/z 314 ' HF
- co, C3H5N Y Y
NH
(\NH (:1 N Hi):A
F
0

N
N\/I | I
+
+
O F O m/z 268 m/z 245

m/z 23]
- C3114

N HNA
||\N [NU
+ /F F
o

"1’2 203 m/z 205

 

 

 

 

Figure 17. Proposed fragmentation pathways for protonated ciprofloxacin

3O

For CID of protonated ciprofloxacin (m/z 332), loss of either water (18 u) fi'om
the carboxyl group to give m/z 314 or loss of CO2 (44 u) to give m/z 288 are observed.
The fi'agment ion at m/z 314 appears to undergo firrther fragmentation to give m/z 231,
which is attributed to a loss of CO (28 u), followed by a loss of C3H5N (55 11), which
corresponds to elimination of the cyclopropyl group and attached nitrogen. These losses
are also confirmed in pseudo-MS3 spectra (Appendix I, Figure 45 and Figure 47). As
shown in Figure 18, the electrons from the C-C bond between the quinolone backbone
and the C0 group migrate to the CO group to give an intermediate at m/z 286 (loss of 28
u). This is followed by the migration of electrons from the pyridine ring to the nitrogen,
which may insert into the cyclopropane ring to give a neutral fragment of formula
C3H5N. Since this fragment is neutral, it is not observed in the mass spectra. Further

elimination of C0 yields a fragment of m/z 203, also shown in Figure 18.

31

Y C .
Al ——> 1
6% F F
C)

m/z3l4

.111 K OH

r1 C
\ Q __ B. 1

 

 

 

 

+ / F

m/z 203
m/z 231

Figure 18. Proposed fragmentation pathway of m/z 314([M+H-H20]+ generated by in-
source CID of ciprofloxacin ions.

A competing path for fragmentation of protonated ciprofloxacin (m/z 332) results
in formation of m/z 288 via the loss of C02 (44 u). This must be accompanied by the
migration of the hydrogen from the carboxylic acid group to the fluoroquinolone
backbone in a manner analogous to levofloxacin. CID of deuterated ciprofloxacin (m/z
335) also exhibits a loss of 44 u, consistent with migration of deuterium fi'om the
carboxylic acid group to the fluoroquinolone backbone. Further fragmentation of m/z
288, generated by in-source CID of protonated ciprofloxacin, yields a fi'agment at m/z
268 corresponding to loss of hydrogen fluoride (20 u) to give m/z 268, or fragmentation
of the piperazine ring (43 u) to give m/z 245. In the former case, elimination of HF

occurs when a mobile proton combines with the fluorine. This is supported by the loss of

32

21 u (DF) in deuterated spectra, as opposed to 20 u in protonated spectra, which shows
that the deuterium fluoride is lost. This finding suggests that loss of HP in the unlabled
ion involves loss of the fluorine plus the proton attached during ionization. For m/z 288,
fiagmentation of the piperazine ring gives m/z 245 (loss of 43 u). A loss of 44 u
(C2H4DN) is seen for deuterated spectra (m/z 291 to m/z 247), showing that a deuterium
from the piperazine ring leaves as part of C2H5N and a deuteron migrates to remain on
the aziridine.

For protonated samples of both ciprofloxacin and levofloxacin, fragmentation of
the piperazine ring is observed in the forms of m/z 245 for ciprofloxacin and m/z 261 for
levofloxacin, shown in Figure 15 and Figure 12 respectively. The losses of water and
C02 are observed for both protonated molecules as well, and both compounds also lose
CO from the pyridinone ring as a loss of 28 u is observed in the protonated spectra of
both compounds. Fragmentation of both protonated molecules yields the structure
corresponding to m/z 205. The loss of hydrogen fluoride is observed for ciprofloxacin,
but not for levofloxacin. This varying fragmentation is a result of one of the structural
differences between the two compounds; ciprofloxacin has a cyclopropane substituent on
the pyridine ring, whereas levofloxacin has a morpholine ring that cyclizes with the
fluoroquinolone backbone. Fewer fragmentation ions are observed for levofloxacin
compared to ciprofloxacin, which was also observed by D’Agostino et al. [38],
demonstrating that this additional cyclic structure alters the fi'agmentation of levofloxacin

compared to other fluoroquinolones.

33

Alkali Metal-Cationized Species

The CID mass spectra demonstrate that lithium cationization, compared to
protonation, yields different and less diverse fragmentation pathways for both
levofloxacin and ciprofloxacin. This is because lithium has a greater mass than a proton,
so it migrates across the molecule more slowly than a proton. Many fi'agmentation
pathways with low activation energy involve proton migration. If the migrating proton is
the proton attached during ionization, replacing it with a slower-moving lithium ion
raises the activation barrier for analogous chemical reactions. The chemical structures
shown in this section were drawn to represent the chemical formula of the corresponding
m/z values obtained in MS/MS experiments. Since mass spectra do not give unequivocal
information about ion structure, details of ion structure should be considered as unproven

at present, but the masses suggest elemental compositions of ions.

34

CIprofloxacIn - LI ‘I_85-2, 338

 

 

 

 

 

 

13April2008$LA_17 13 (0.280) Cm (1:46) A 4: Daughters of 33868+
211 5.0983
100 294
' °\° 210 23 338
0"?071)*' :1 {2'373' r ' Tf 7,273,; 1"1' 'zfrs—Vrr' r "Y—f" r 'r 27r4 rrrrrrrr * 3971m'rfi'r 'r "I‘ r "rffi'T' r
13Apri|2008SLA_17 15 (0.317) Cm (1 :46) B 3: Daughters of 338ES+
294 1.12e4
100
.,\°
0203 211 214 232237 246 255 272 291 3??
13Apri|20088LA_17 15 (0.312) Cm (1:47) c 2: Daughters of 338ES+
294 3.33e4
100
336
°\°.
o 202.211 :2 -- 229- .- -- - ., H ..... - : ,. 9.113-- : - 339
13April20088LA_17 16 (0.371) Cm (1 :47) D 1: Daughters of 338ES+
_ 336 1.34e5
100
.,\°
0 2196 337.
200 I 213de7220 ‘ 230 240"" 250 ' iéd'Tvéib Y 260 - T290 " 300 ,, V310 "320 7' 330' '7 340

Figure 19. CID product ion spectra of lithium-cationized ciprofloxacin (m/z 33 8) at
collision cell potentials of (A) 35 V, (B) 30 V, (C) 25 V, and (D) 10 V

Clprofloxacln D, Ll, LBS-6, 340

21May20088LA_31 15 (0.323) Cm (1:46) A 4: Daughters of 340£s+

100 213 296 772
I 241

. g 213, 227 257

 

, 1216222 1.211

 

227 2561256 258271 261266 2‘94), 303309 7317326828 3363219345

 

 

21May20088LA_31 15 (0.318) Cm (1:46) 3 3: Daughters o1340ES+
296 5.01e3
100 .
.,\°
21May20088LA_31 17 (0.356) Cm (1:47) c 2: Daughters of 340£s+
296 1.05e4
100
340
,,\°
0 - g 217 H 247 2399 - -
21May20088LA_31 14 (0.286) Cm (1:47) D 1: Daughters o134oes+
340 3.35e4
100
,\°
206 256 299

300 210 220 230 240 250 260 270 280 290 300 310 320 330 340 3531/:
Figure 20. CID product ion spectra of m/z 340 (A+Li+2) in D20 at collision cell

potentials of (A) 35 V, (B) 30 V, (C) 25 V, and (D) 10 V

35

HC)

 

 

 

 

N N
' ——> I
F
F
C) C)
'C3H5N3 O
n02294
—l Li‘”
(\NH
\J I'HF
N
l | \
/
F
nflz211

 

0 m/z 274

Figure 21. Proposed fragmentation pathways for lithium-cationized ciprofloxacin

For CID of lithium-cationized samples of ciprofloxacin (m/z 338), fragmentation
pathways different from those of protonated samples were observed. Unlike protonated
samples, no water loss was seen, suggesting that no mobile proton is available to combine
with the hydroxyl group, thus inhibiting loss of water. Similar to protonated samples,
both C3H5N (55 u, fiom the cyclopropylamine group) and CO (28 u, fiom the pyridinone)
are lost when m/z 294 (the lithiated decarboxylated group) fragments to give m/z 211.
CID of deuterated, lithiated ciprofloxacin (m/z 340) exhibits a loss of 83 u (m/z 296 to
m/z 213). As loss of 83 u is observed for both deuterated and nondeuterated samples, this
suggests that migration of exchangeable hydrogens is not involved in this pathway as the

neutral mass loss does not change upon hydrogen-deuterium exchange.

36

Unlike protonated ciprofloxacin, no fragment ions corresponding to the losses of
the neutral species C0 and C3H5N are observed for lithiated samples. Also, no
fragmentation of the piperazine ring was observed (no loss of 43 11), indicating that
lithium cationization precludes the hydrogen migration necessary for that fragmentation
by interacting with a nitrogen on the piperazine ring. Surprisingly, lithitnn-cationized
ciprofloxacin also loses hydrogen fluoride, which is suggested by H/D exchange
experiments. For CID of deuterated, lithiated ciprofloxacin, m/z 296 loses 21 u to give
m/z 275, which corresponds to a loss of deuterium fluoride. This behavior might be
explained if some of the lithiated molecules consist of a lithium salt of the carboxylate,
retaining protonation of the basic piperidine group. The proton attached on the piperidine
retains the mobility needed to participate in the elimination of HF. Because lithium has a
greater mass than a proton, it migrates more slowly across the analyte molecule, leading
to fragmentation pathways that are less diverse then those obtained for protonated

samples.

37

LVX - LI 1_81-6 - 338

 

 

 

 

 

 

10Aprit2006SLA_16 20 (0.430) Cm (1:46) A 4: Daughters of 368ES+
100 270 275 324 2.2763
°\° 269
0' v 1411/271r7'fiza'zfirfirffi‘fiT 29'8rfi30r4 fir '''''' .Y§273fim, - 'r 1' v7 ' 'rr '1 r '
10April20088LA_18 15 (0.317) Cm (1 :46) a 3: Daughters o136ees+
324 6.96e3
100
270
I 271 296 304 325 368
0 1291315311 11 29112917111 111 11 3.2111 1111111111111 1 .11. 1113:6931 1
10Apn'l20088LA_18 13 (0.269) Cm (1 :47) c 2: Daughters of 368ES+
366 1.72e4
,\°
293319 fl 283fi 2933 3‘1” 32‘? 327 1 ff 1 1 1 '36? 1
10Apri|2008$LA_18 20 (0.414) Cm (1 :47) o 1: Daughters of 368ES+
366 1.03e5
100
,,\°
327 353

\
3r13firrv .,1.3+,# rvrvs--, 1. , .- , -,. . ,2- 1, ---,fi ...... r' -, -4 1 -.,v.--,

0"' V' ' I" 'I""I" 'I‘ r
260 270 260 290 300 310 320 330 340 350 360 370
Figure 22. CID product ion spectra of lithium-cationized levofloxacin (m/z 368) at

collision cell potentials of (A) 35 V, (B) 30 V, (C) 25 V, and (D) 10 V

 

Lovofloxacln D, LI, 1_38-3, 369

21May2008SLA_45 17 (0.367) Cm (1:46) A 4: Daughters of 3695s+
270 1.2163
100
369
I °\° 325

264265
0 1,! f ,. 211781 291331299294 2:99, 307' 313 919 {32]I‘ 330331339 7345 351' 357 364366 379'

 

 

2110May20088LA_ 45 16 (0.340) Cm (1:46) 3: Daughters of 369ES+
369 5.35e3
10%I 2595711 296 2919 3I 309310 _3I:28 I
21May20088LA_ 45 15 (0 313) Cm (147) 2. Daughters OI369ES+
369 264e4
100
.11
265 271 304 310 325 326
0' 37°11 11 267 1 299 1.305,. I 4' 1 341 if 351 1 .1 VVVVV
21May20088LA_45 17 (0.351) Cm (1 :47) D 1: Daughters 01369ES+
369 7.56e4
100
5°
265

 

vvrrv 1w rr r vv vv v 1v vvv 7 7V vv+f ‘7 v 77 vr VrVVv—fw

260" I 270" T260 " 290" " 300 310 320 I 330 340 ' 350 ' 360 ' 370
Figure 23. CID product ion spectra for m/z 369 (A+Li+l) for lithium-cationized
levofloxacin D20 at collision cell potentials of (A) 35 V, (B) 30 V, (C) 25 V, and (D) 10

V

38

 

Li

 

an304

Figure 24. Proposed fragmentation pathways of lithium-cationized levofloxacin

In CID of lithium-cationized levofloxacin (m/z 368), loss of hydrogen fluoride
was also observed; this was not observed for protonated samples and provides a fragment
unique to lithium cationization. The decarboxylated ion at m/z 324 fragments to give m/z
304, which corresponds to a loss of hydrogen fluoride (20 11). For CID spectra of
deuterated, lithiated levofloxacin, a peak at m/z 325 is observed, which loses 21 u to give
m/z 304. The loss of 21 it corresponds to a loss of deuterium fluoride. As is the case
with both protonated and lithiated ciprofloxacin, the exchangeable hydrogen from the
hydroxyl group remains on the quinolone backbone and is lost as hydrogen fluoride for
lithiated levofloxacin. Pseudo-MS3 for which m/z 324 was selected as the parent ion

(Appendix I, Figure 49) also suggests that m/z 324 fragments to give m/z 304. Lithium

39

cationization of levofloxacin is proposed to raise the activation energy needed for
competing fragmentation pathways, and allows loss of hydrogen fluoride to become
competitive whereas it was not observed for protonated samples. Similar to lithiated
ciprofloxacin, no fragmentation of the piperazine ring (no loss of 57 u) was observed,
indicating that lithium cationization increases the activation barriers for competing

fiagmentation reaction, thus allowing the loss of hydrogen fluoride to be detected.

C”

 

 

Figure 25. Proposed mechanism for the loss of hydrogen fluoride from levofloxacin,
which is not observed for protonated species

Sodium cationization of levofloxacin also results in CID fragmentation different
from protonation as it also yields loss of hydrogen fluoride. Since sodium (23 u) has a
larger mass than lithium (7 u), it cannot move throughout the molecular surface to the
same degree as lithium and therefore pathways that involve either proton or lithium
migration will be slower (and less likely to yield observable fiagment ions) when these
cations are replaced by sodium. Again, loss of HF may involve protonation of the
sodium carboxylate salt, retaining a mobile proton needed for HF elimination. Spectra

for sodium-cationized fluoroquinolones are shown on the following page.

40

Clprofloxacln - Na LBS-3
13Apn'l2008SLA_12 43 (0.925) Cm (1:46)
79

A 4: Daughters 01354ES+
100 ‘84

         

   

106119131

 

152 166197 212223 233

  

263 266 290 306309 350 355

    
  

13April2008$LA_12 18 (0.382) Cm (1 :46) B 3: Daughters 01354ES+

 

 

 

 

24 97 354 230
100 204
°\° 98
61 73 8132 124 207 231 247 298 311
, 11111141414317, 1. $163117-L1-JI .1111 1,44 2,1111 ,Jggssjjllééglfiwsjgfg 71:11:1ng :4 , 239.1%”, *1; 3%229 Cwutfwfs11¥ggfi 1.,
13April20088LA_12 20 (0.420) Cm (1 :47) C 2: Daughters of 354ES+
354 3.14e4
100
.,\°
0" '*I 'r r r r r r '7' r'fir
13April2008SLA_12 21 (0.436) Cm (1 :47) D 1: Daughters of 354ES+
354 5.4365
100
o\°
01 "r' ' r "r' ' r" ‘1'

 

2 4o 60 I"6'0"“'"1'00“ 120"”1'40""160""1'60'" 200' "220' ' 240 260'T'260'I' 30'0""320 " 340""'3'1‘53‘Iz
Figure 26. CID product ion spectra of sodium-cationized ciprofloxacin (m/z 354) at
collision cell potentials of (A) 35 V, (B) 30 V, (C) 25 V, and (D) 10 V

LVX - Na 1_31-4

  
 

10Apr1120088LA_10 17 (0.366) Cm (1:46) A 4: Daughters of 384ES+
320 163
100
. ,2 23 81 112 365
.30 4" 63 96 114 133'148 173177 187 303 222 238 256255264 299 313335339 352
0 i 1 , . .
10Apri|20088LA_10 20 (0.425) Cm (1 :46) a 3: Daughters of 384ES+
384375
100
364
°\° 23
0 123 141 176 196 220 235 267276 ,293

   

10April20088LA__10 22 (0.463) Cm (1:47) C 2: Daughters of 384ES+

 

100 4.60e4
°\°
32o
10April20088LA_10 14 (0.285) Cm (1 :47) D 1: Daughters of 3845$+

100 4.94e5

e\°

 

2'0 ' 4'0 ' 6'0 1 60 I 100' "120 I'1'40"'"160 ' 160'r'200 T220 ' 240 ' 260' I'260'I 300 "320' ' 340" 3130“"3611'7",z
Figure 27. CID product ion spectra of sodium-cationized levofloxacin (m/z 384) at
collision cell potentials of (A) 35 V, (B) 30 V, (C) 25 V, and (D) 10 V

41

Lovofloxacln D, Na, LBS-9

21 May2008SLA_50 20 (0.431) Cm (1:46) A 4: Daughters of 365£s+
100 105 1.0463
91
I °\° 167

 

365
0 92649316371 67 96 [1103141154 173 191 209 219329251 272267302 320 334 353,356 1392

V

 

21May2008SLA_50 16 (0.340) Cm (1:46) B 3: Daughters of 385ES+
100 23 385 796
°\o 91 105 161

 

' 145 345
01.23259?" 75291193., 1232. ,1. 176.111 19] 23? 211 2513661.?9339832193211 35.5 335392

‘v

 

21 May20088LA_50 17 (0.356) Cm (1 :47) C 2: Daughters of 385ES+
385 2.23e4
100
69
o 9; 105 2‘1? 321 341
21May20088LA_50 16 (0.329) Cm (1 :47) D 1: Daughters of 385ES+
100 385 2.5895

a\°

 

G2'0 ' 4'0 ' 6'0 ' 6'0 '1'00""120"'140"1'60" 160 "200 "220 '240 "260 "260' 300 ' 320'34'0' 360' 360 """
Figure 28. CID product ion spectra of m/z 385 (A+Na+1) for sodium-cationized
levofloxacin in D20 at collision cell potentials of (A) 35 V, (B) 30 V, (C) 25 V, and (D)
10 V

In the case of sodiated ciprofloxacin (m/z 354), the sodium atom (m/z 23) is lost
during CID and no fragmentation of the ciprofloxacin molecule itself is observed. Since
sodium has a greater ionic radius than lithium, it forms longer bonds with the neutral
molecule, therefore its interaction with ciprofloxacin is weaker than that of lithium,
causing it to dissociate from ciprofloxacin during CID. In the case of sodiated
levofloxacin (m/z 384), loss of 64 11 (C02 (44 u) and HF (20 u)) yields m/z 320.
Analogous to lithiated levofloxacin, the sodium cation may be associated with the
carboxylate group, thus precluding the hydrogen migrations necessary to cause a loss of
water. A loss 65 u (C02 (44 u) and DF (21 u)) to yield m/z 320 is observed for

deuterated, sodiated levofloxacin (m/z 385). As with lithiated levofloxacin, sodium

cationization of levofloxacin raises the activation energy needed for other competing

42

fragmentation reactions, allowing loss of hydrogen fluoride to become competive, and
gives fragment ions derived from pathways not observed for protonated samples.

For potassium-cationized samples, minimal fragmentation of the compounds is
observed. While levofloxacin and ciprofloxacin do forrn[M+K]+, upon CID of [M+K]+
(m/z 400 and m/z 370 respectively), the [M+K]+ ion is observed at 10 eV, but only K+
(m/z 39) is observed for the other collision voltages used. This is illustrated in Figure 29.
Since potassium has a larger ionic radius than the other cations used, the interaction
between it and the two fluoroquinolones is weak, so less energy is needed to cause it to

dissociate from the respective compounds.

Clprofloxacln - K LBS-4

 

 

 

 

 

13Apri|2008$LA_14 23 (0.464) Cm (1:47) A 2: Daughters of 370ES+
39 7.69e4
100
- ..\°
0.. 1
13April20088LA_14 24 (0.500) Cm (1:47) B 1: Daughters of 370ES+
370 8.2595
100-
as.
039 .....
19Feb20088LA_18 6 (0.116) Cm (1:47) c 2: Daughters of400ES+
39 4.11e3
100
.\°
0.
19Feb2008SLA_18 5 (0.091) Cm (1 :47) D 1: Daughters of 4ooss+
400 7.09e4
100
°\°l
c. 3.9

40 ' 6'0 .1 6'0 ' 100 ' 120 ' 140 ' 160 ' 160 ' 200m 220 ' 24o ' 260 ' 260 "300 ' 320 ‘ 340 ' 360 ' 360 ' 400 ' 420
Figure 29. CID product ion spectra of potassium-cationized ciprofloxacin (m/z 370) at
collision cell potentials of (A) 25 V and (B) 10 V, and potassium-cationized levofloxacin

(m/z 400) at (C) 25 V and (D) 10 V

43

Discussion

Collision induced dissociation of lithium and sodium ionized species yields the
loss of hydrogen fluoride from levofloxacin, a phenomenon that is not observed for
protonation, thus providing a method to enhance the fragmentation of levofloxacin. For
both ciprofloxacin and levofloxacin, differing fragmentation pathways were achieved
using protonation and alkali metal cationization, consistent with a variation in the
mobility of the proton, lithium, and sodium. Protonation yielded the most diverse
fragmentation because the ionizing proton was able to migrate across the molecule more
rapidly than the alkali metal cations. Because lithium and sodium have larger masses
than a proton, they cannot migrate across the analyte molecules to the same extent as a
proton. As a consequence, some kinds of fragmentation reactions have higher activation
barriers and rates of fragmentation too slow to be observed as fragment ions.

Both lithium and sodium cationization yielded loss of hydrogen fluoride, a
pathway for levofloxacin not observed with protonation. Lithium and sodium
cationization raise the activation energy of competing pathways, allowing loss of HF to
become competitive with other fragmentation reactions. For both lithium- and sodium—
cationized species, the differing substituents of the fluoroquinolones ciprofloxacin and
levofloxacin result in different fragmentation behavior. The additional cyclic component
in levofloxacin causes it to yield fewer fragment ions than ciprofloxacin, as was also
observed by D’Agnostino et al. [38] .

These experiments demonstrate that lithium and sodium cationization can be used
to alter the fragmentation pathways of levofloxacin and ciprofloxacin and reveal the

presence of fimctional groups not evident from the CID of protonated species. While

protonation provides more diverse fragmentation, alkali metal cationization enhances the
information in CID spectra of levofloxacin by enabling a loss of hydrogen fluoride. In
the case of both analyte molecules, alkali metal cationization increases the activation
barrier needed to cause fragmentation of the piperazine ring, thus allowing other
fiagmentation pathways, such as loss of hydrogen fluoride, to become competitive
reaction channels.

While alkali metal cationization of levofloxacin allows observation of the loss of
hydrogen fluoride, fragmentation pathways achieved with protonation are more diverse
and sometimes more structurally informative. For example, Hemeryck et al. [39] studied
the metabolism of levofloxacin in Rhesus monkeys and determined that one of the
metabolites produced was desmethyl levofloxacin, a result of the demethylation of the
piperazine ring. Analogous to levofloxacin, CID of protonated desmethyl levofloxacin
yielded a loss of 44 u (C02), which then lost 43 u to give m/z 261. This demonstrates
that the methyl group was lost from the piperazine ring rather than the morpholine ring.
Since the piperazine ring does not fiagment in the case of cationized samples,
determining the location of demethylation would be challenging. CID can be applied to
the analysis of protonated levofloxacin and ciprofloxacin metabolites to help determine
where any structural modifications have occurred during metabolism, thus allowing for
the possible determination of the pathways involved in the metabolism of these

compounds.

45

Erythromycin

Protonated Species

Figure 30 and Figure 31 show CID spectra of protonated and deuterated
erythromycin, respectively. Figure 32 illustrates the proposed fragmentation pathway for
protonated samples of erythromycin. CID of the parent ion m/z 734 yields loss of the
cladinose sugar (158 u) and three water losses (18 u per loss) from the macrolide ring
(which appear in the spectra at m/z 558, 540, and 522). Loss and subsequent
fragmentation of desosamine is observed, with desosamine (m/z 158) undergoing fitrther
fragmentation at collision voltages of 40 and 55 eV to give m/z 116 through the loss of
42 u. Protonated desosamine (m/z 158) appears as the dominant fragment owing to its
greater basicity than the macrolide ring, which was not observed as a fragment ion. The
CID spectra for protonated erythromycin therefore conveys virtually no information
about the structure of the macrolide ring, thus limiting the prospects for using CID of
protonated species to distinguish positions of any metabolic alterations in erythromycin
metabolites or analogs.

Product ion spectra of deuterated erythromycin with five deuterons (m/z 739)
exhibits a loss of 159 u (m/z 580), which is attributed to the cladinose sugar as it has one
exchangeable hydrogen. The loss of 2H1-desosamine (m/z 159) and three deuterated
water (HDO) losses (19 u per loss) are also confirmed for deuterated, protonated

erythromycin samples to give ions at m/z values of 561, 542, and 523.

4.6

Erythromycin 1_18-3

 

 

 

 

21May2008$LA_07 13 (0.281) Cm (1:46) A 4: Daughters of 734£S+
116 5.2986
100
a 158
o\
123
Or f'fé' Y" T'_' 1" fr 1 Y‘fi_"_l"_v l I r 7 fi“ f' Y' I x—r 7 W1' 1 'V TV‘V'T' 7 " I*' ‘l
21May2008SLA_07 13 (0.275) Cm (1:46) B 3: Daughters of 734ES+
158 1.13e7
100
01 116 '
o\.
127
0 *'WWW'TWY ,, V'V‘V'rrm')
21May20088LA_07 12 (0.248) Cm (1:47) C 2: Daughters of 734ES+
158 1.0997
100
$ 116127 576
0 Jéfi.-. 1. 1. 111.11.. .fi-..11..1.111 1-1. #25?” .1 -1 11113341
540
21May2008$LA_07 14 (0.286) Cm (1:47) D 1: Daughters of 734ES+
734 1.50e7
100
I o\°
576

 

(100 1597200 ”'250 "T300m'" 3'50 "461? ' 450 500 550 600 650fim 7'00 “"753":
Figure 30. CID product ion spectra of erythromycin (m/z 734) at collision cell potentials
of(A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V

/

Erythromycin D, 1_88-1

 

 

 

21 May20088LA_16 12 (0.259) Cm (1:45) A 4: Daughters of 739ES+
117 5.0365
100
$1 159
127
o .1 .2--. . .--... . . 1-1.2-1-...
21May20088LA_16 11 (0.232) Cm (1:46) B 3: Daughters of 739ES+
159 8.63e5
100
117
°\°
J 127 160
G v1£f 1 vfi 1 1 v 'r‘" r V'r *fi'mr" 'r --1- 1' f' r **'—r -
21May2008SLA_16 13 (0.270) Cm (1:47) C 2: Daughters of 739ES+
158 8.87e5
100
6227 150 524542 561 (581 739
0. . .-'. . . r. .. . .11 .. -. . . . -1. 1V}... w. 1 . . rfi . . 4.
21May20088LA_16 13 (0.264) Cm (1:47) D 1: Daughters 01739ES+
100 739 2.00e6

580
100 150 200 250 300 350 400 450" "'500 -1 550 " '600fim 650 700 11175317:
Figure 31. CID product ion spectra of m/z 739 (A+5) of erythromycin in D20 at collision
cell potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv

47

  
 
  
 

0“ \./

I NH
I”///
O
0
’o o/
O OH
m/z 734
O
ol."||'|
11111111 OH \+/
/ ,’ NH
’Ill//
0
O

m/z = 158
Figure 32. Proposed fragmentation pathway for protonated erythromycin

48

Alkali Metal-Cationized Species

While CID of protonated erythromycin does not yield ions derived fi'om
fragmentation of the macrolide ring, CID of sodium-cationized erythromycin (Figure 33
and Figure 34) causes fragmentation of the macrolide ring (m/z values 423 and 309), thus
allowing for the examination any structural alterations that have taken place on the
macrolide ring. Fragmentation of the macrolide ring is observed at a collision voltage of
55 eV as seen in Figure 33 on the following page. The chemical structures shown in this
section were drawn to represent the chemical formula of the corresponding m/z values

obtained in MS/MS experiments.

49

Erythromycin 138-2 Na
180902007_SLA14 5 (0.108) Cm (1 :46) A 4: Daughters of 756ES+
100 115 309 598 3.2794

 
   

 

 

- .\°
180191235247 291 327351 392405 23 485 519 536 580 738 756
/ -r~ 'W
18Dec2007_SLA14 6 (0.124) Cm (1 :46) B 3: Daughters of 756ES+
756 3.35e5
100
so 596
309 422 465 580 739 57
0 ‘flTWerwrrfirfi-rvrmwnrfifi -
18Dec2007_SLA14 6 (0.119) Cm (1 :47) c 2; Daughters of 756ES+
756 6.86e5
10
.\°
75
0'1 -, 1 1‘ 'rT' 'Tvr TW '1 '1' 'r - r‘ r V '1 -m- r r Wr'fifr"‘*r 'r 1' I - r V'WW r' r
18Dec2007_SLA14 7 (0.135) Cm (1:47) D 1: Daughters 01756ES+
756 7.00e5
100
°\°
75

 

50 ' 100 150 2200 I arms "'3‘50' ' 400'W5450V'W500'" '550" " 600 """ 650% 7130'" "7'50”“,2
Figure 33. CID product ion spectra of sodium—cationized erythromycin (m/z 756) at
collision cell potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V

Erythromycln O. Na, LOB-3

 
 

 

 

 

21May2008$LA_12 13 (0.260) Cm (1:46) A 4: Daughters 01761ES+
167 1.7294
100
. ,4» 602
525 ,545 602) 761
W
21May2008SLA_12 13 (0.275) Cm (1:46) a 3: Daughters 01761ES+
761 2.21e5
100
.,\°
603
0‘!" I Y ' 1"' l ' "' *7 " " ‘r‘ I' 'V ' "I ' 'I "_"_T' v'r" 'Y "I ' 1' 'V'V'TY""' ' 1“ "Y 'Y'fr'I " I W
21 May20088LA_12 12 (0.248) Cm (1 :46) c 2: Daughters of 761ES+
761 3.95e5
100
,,\°
0 '1 r r rfi r r r Y r r r r '*r' V
21 May2006SLA_12 14 (0.286) Cm (1 :47) o 1: Daughters of 761ES+
761 5.33e5
100
.\°
0

50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 m
Figure 34. CID product ion spectra of m/z 761 (A+Na+5) for sodium-cationized
erythromycin in D20 at collision cell potentials of (A) 55 V, (B) 40 V, (C) 20 V, and (D)
10 V

50

O , Na‘”
o-tIIIIIII

(DH \\¢//

    

   
 

‘lIlllra.. -IOIIIIIII

OH \ / -cladinose

IIIIIIIII
N /
0,”, ——-——>
o 0

(DH

nfl2598

m/z 756 - desosamine

 

 

nuz309 nuz423
Figure 35. Proposed fragmentation pathway for sodium-cationized erythromycin
CH) of sodiated erythromycin (m/z 756) exhibits a loss of 158 u to give m/z 598,
which corresponds to loss of cladinose. A subsequent loss of neutral desosamine (175 u)
is observed similar to protonated samples. The difference between CID behavior of
protonated and sodiated species may be explained by different sites of ionization. In
contrast to proton attachment at nitrogen on the desosamine, the sodium cation is
expected to form complexes with oxygen substituents on the macrolide ring, giving m/z
423 as a fragmentation product. The macrolide ring can undergo firrther fragmentation to

give m/z 309, attributed to loss of neutral C6H1002 (114 u). For CID of deuterated,

51

sodiated erythromycin, a loss of 115 u is observed, giving m/z 311. This demonstrates
that one exchangeable hydrogen is lost, while another migrates and remains with the
macrolide ring as there is a mass shift of one for the neutral loss (114 and 115 u) and a
mass shifi of two for the remaining fi'agment (m/z 309 and 311), which only has one
exchangeable hydrogen. A pr0posed mechanism for this phenomenon is shown in Figure
36. CID of sodium-cationized samples of erythromycin yields fragmentation of the

macrolide ring, which is not observed for CID of protonated erythromycin samples.

(I o —'l Met+ OH ——| Met+

OH

HO

     

   
  
 
 
 

I’li'lit'.

OH -114u

I
’0’],
/

o
'1
’1
I
I,
I
1’

   

”0H
Figure 36. Mechanism for the loss of 114 u from the macrolide ring of erythromycin
Proposed structures for sodium-cationized erythromycin were confirmed by
pseudo-MS3 analysis. For these experiments, m/z 309 was selected as the parent ion for
CID (Appendix I, Figure 50). The proposed fragmentation pathway for m/z 309 is

illustrated in Figure 37 on the following page.

52

    
 

   
  

0H

"null" - C02
OH

V

’I
I, ’l
”o, ”/

\\\\\\\\\

' a
‘2, '9’,
’CH 0“

m/z 309 m/z 265
Figure 37. Proposed fragmentation pathway of m/z 309

Lithium-cationization yields different fiagmentation pathways and further
fragmentation of the macrolide ring compared to protonated and sodium-cationized
samples of erythromycin (Figure 38 and Figure 39. Unlike sodium-cationized samples,
desosamine is not lost as a neutral sugar in the case of m/z 520 and m/z 468. This
behavior is interpreted as being similar to the phenomenon seen in protonated spectra,
with the lithium cation attached at the nitrogen on the desosamine in these cases, as
desosamine is not lost (as a neutral of 175 u from the lithiated species). Since lithium has
a lesser mass than sodium, it is more mobile, allowing it to interact with either the
nitrogen on the desosamine or a carbonyl oxygen on the macrolide ring. The ion
resulting from loss of cladinose, m/z 582 can lose either 62 u, attributed to a loss of C02
(44 u) and water (18 u), to give m/z 520, or 114 u to give m/z 468 through fragmentation

of the macrolide ring, as was discussed in the case of sodium cationization.

53

Erythromycin - Ll 1_77-1, 740

 

 

 

15Ma120088LA_21 5 (0.108) Cm (1:46) A 4: Daughters of 740ES+
293 1.3664
100
' "\° 350 407
311 389 4
0' Ari1 3?3‘ 14' .—r- L'L, r‘ " 3174*?8r r *ég9'4fx'r'ffi4x ‘ r '**—r firm 7 " Y'V'Vrf'f'1
15Mar20088LA_21 6 (0.124) Cm (1:46) B 3: Daughters of 740ES+
740 8.6794
100
582
s\°
407 56L4 J 722 743
o 11111. . 11.411. 1414.1 1111.11111111111 .1 11111111 1 111111
15Mar2008SLA_21 6 (0.119) Cm (1:47) C 2: Daughters of 740ES+
740 2.8265
1001
°\°
011 11 111 -1. .1 1. .1 fi11111fi111 14111.1.fi 11111111111
15Mar20088LA_21 7 (0.135) Cm (1:47) D 1: Daughters 0174OES+
100 740 3.6285

a\°

tvvr

Figure 38. C11) product ion spectra of lithium-cationized erythromycin (m/z 740) at
collision cell potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V

Erythromycin D, Ll, 1_88-2, 745

 

 

 

21 May2008$LA_09 13 (0.260) Cm (1:45) A 4: Daughters 01745ES+
100 .296 6.3563
306
, 01. I314 410 472 509
323 351 372 391 440 483 585
o 1.. 1).... 11. 1.112,. Liar; .,.1 323549557, 1 . 619 2 ...... 629711.- 7:60
21 May20088LA_09 14 (0.297) Cm (1 :46) B 3: Daughters of 74SES+
745 8.6664
100
°\° 566
0. 395 323 351 4025119 22? 452 471 592525.559 5558 L 612 - (€9,746
21May2008SLA_09 12 (0.248) Cm (1 :47) c 2: Daughters of 745ES+
745 3.09e5
100
‘,\°
0 . 3771.39.12”? 507523.- - 1. .586. 1. - .6992 31574.6
21 May20088LA_09 15 (0.307) Cm (1 :47) D 1: Daughters of 745ES+
745 4.09e5
100
.,\°
745746
rn/z

 

300' 325350 '3i5' '400' '425 '450' 475' 5002325 1550 sis 606125 9650' 67577007 '725' 750
Figure 39. CID product ion spectra of m/z 745 (A+Li+5) for lithium-cationized
erythromycin in D20 at collision cell potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D)
10 V

54

0 _“ Li+

. Cfii \Nq//

2%»
3 (D C)
(D

  
     
 

"”1"...

CD

a
21
(DH

nV2582

' C6H1002

 

anSZO

an468
Figure 40. Proposed fragmentation pathways of m/z 582

While a loss of 114 u (C6H1002) from the macrolide ring was also observed for
sodium-cationized erythromycin, a loss of 62 u is unique to lithium-cationized species.
This loss of 62 u is attributed to a loss of water (1 8 u) from the macrolide ring to give m/z
564 (structure not shown), followed by a loss of C02 (44 u) from the macrolide ring to
give m/z 520 (Figure 40). For CID of deuterated, lithiated erythromycin (m/z 745), a loss
of deuterated water (19 u, HDO) from m/z 586 to give m/z 567 is observed followed by a

loss of 44 u to give m/z 523. Because a loss of 44 u is also observed for lithiated

55

erythromycin, this demonstrated that no exchangeable hydrogens are involved in the loss
of 44 11. In pseudo-MS3 spectra for which m/z 582 was selected as the parent ion (Figure
51), peaks at m/z 293 and m/z 450 (468 — water) were observed, thus demonstrating that
these ions result from the fragmentation of m/z 582 and not via some alternative pathway.

For CID of lithiated erythromycin (m/z 740), loss of the cladinose sugar is
observed (15 8 u), similar to both protonated and sodium-cationized samples. A portion
of the proposed fragmentation pathway is illustrated in Figure 41. As with sodium-
cationized samples, a loss of 114 u, which is attributed to 051-11002, from the macrolide
ring is observed. From m/z 582 to m/z 468, a loss of 114 u is observed and a loss of 115
u is observed in CID spectra of deuterated, lithiated erythromycin (m/z 745). Similar to
sodium-cationized samples, this demonstrates that one exchangeable hydrogen is lost,
while another migrates and remains with the macrolide ring, according to the mechanism
in Figure 36. The same phenomenon is observed for m/z 407, which loses 114 u to give
m/z 293. A loss of 115 u is also observed for deuterated analogs. Like sodium-
cationized samples, the lithium is interacting with a carbonyl oxygen on the macrolide

ring.

56

O l Li+

  
 
     

 

- cladinose OH \ /
_——> a"! N
”’I/
O ' 0
’6H
0H m/z 582
0
m/z 740 - desosamme
o——l Li+
H0 H0
Illlmu‘. "mm"

 

OH

 

I””’
"ll

/

 

m/z 293 m/z 407
Figure 41. A portion of the proposed fragmentation pathway for lithium-cationized

erythromycin
Figure 42 shows the CID spectra for potassium-cationized erythromycin (m/z
772). Only the M+K+ ion (m/z 772), m/z 614 (loss of cladinose, 158 u) and the
potassium cation itself (m/z 39) are observed. Because potassium has a greater mass than
the other cations used, it cannot migrate across the molecule to the same extent as the
other cations. Futhermore, its large ionic radius gives weaker interaction with the
erythromycin molecule, thus yielding less fragmentation except for the formation of an

abundant potassium cation.

57

Erythromycin - K 1_71—5

 

 

 

 

15Mar2008SLA_34 4 (0.086) Cm (1:46) A 4: Daughters of 772ES+
39 4.21e3
100
I \°e
o 614
40 281 “1’6 772
0.1 . 11L .1. . 1 .1 1111.11 . . 1 .111. 1 1. L1
15Mar2008SLA_34 5 (0.102) Cm (1 :46) B 3: Daughters of 77zes+
100 772 9.25e4
$139
0"L'l""'7"" m'vV '7" 'T'W'r“ "7' fiT"' r'VWWfir' ' r ' 1* 'r*' 1' 1..--.rre 'fiT’“r' """""" flu6'1'41 """"""""""""""" '7r§'5"r"
15Mar20088LA_34 5 (0.097) Cm (1:47) c 2: Daughters 01772ES+
772 4.06e5
100
g.
15Mar2008SLA_34 5 (0.091) Cm (1:47) 0 1: Daughters 01772ES+
772 5.26e5
100
.11
01 r' """"""""" 1""1 """"""""""""""" r """" “TV """"""""""""""""" r """"" 1""- """"""""""""""""""""""""""" 1""1/2
50 100 150 200 250 300 350 400 450 500 550 600 650 700 750
Figure 42. CID of potassium-cationized erythromycm (m/z 772) at collision cell
potentials of (A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V
Discussion

Alkali metal cationization of erythromycin allows cross-ring fragmentation of the

macrolide ring that is not obtained from protonated species. This finding suggests that

CD) of cationized erythromycin can be a useful approach for the structural elucidation of

erythromycin metabolites. The varied fragmentation pathways achieved using

protonation and alkali metal cationization demonstrate that cationization alters hydrogen

migration across the molecule, and alters the fragmentation behavior of the ionized

molecule. Both lithium and sodium cationization increase the activation barrier for other

fragmentation pathways, allowing for the fragmentation of the macrolide ring to be

observed. In the case of potassium cationization, the potassium ion has a greater mass

58

than other ions, and does not migrate throughout the analyte molecule to the same extent
to allow of an array of fragment ions. Also, the size of the potassium ion makes the
interactions between it and erythromycin weak compared to the interactions seen with the
other cations studied. As a result, when enough energy is applied to potassium-
cationized erythromycin, the potassium ion dissociates from the molecule and minimal
fragmentation of the analyte molecule is observed. Unlike the fluoroquinolones,
fragmentation of the analyte molecule itself is observed in the case of erythromycin (m/z
614 corresponds to [M+K-cladinose]+).

In the case of protonation, the charge is able to migrate across the erythromycin
molecule more than the lithium- and sodium-cationized species because it has a smaller
mass than the alkali metal cations and can therefore migrate faster. For the lithium- and
sodium-cationized species, the altered charge localization allows for different hydrogen
migrations (based upon H/D exchange experiments), and thus different fragmentation
pathways. For example, hydrogen migrations for protonated erythromycin caused water
losses, while hydrogen migrations for cationized erythromycin caused fragmentation of
the macrolide ring. Lithium cationization yields different fragmentation of the macrolide
ring compared to sodium cationization because the sodium ion has a larger mass than the
lithium ion and therefore the charge cannot migrate as rapidly to participate in transition
states to fragmentation.

Also, because the desosamine group is present in some proposed structures for the
fragmentation of lithium-cationized species, this suggests that the lithium cation is
capable of attachment at with both the nitrogen on the desosamine and the oxygen-

containing groups on the macrolide ring. Sodium cationization yields loss of both

59

cladinose and desosamine, suggesting that the sodium cation is interacting with the
carbonyl oxygens on the macrolide ring. Because lithium and sodium have greater
masses than a proton, they migrate across the molecule more slowly than a proton, which
allows them to be more selective about the sites on the molecule with which they interact.

For erythromycin, lithium and sodium cationization can be used to fragment the
macrolide ring and thus help determine the location of any structural modifications that
have occurred on that portion of the molecule during the course of metabolism. Sodium
and lithium cationization alter the dissociation pathways of erythromycin, thus yielding
fragmentation of the macrolide ring, a phenomenon not observed for protonated
erythromycin. Because cationization causes fragmentation of the macrolide ring, rather
than having it be lost at a neutral in the case of protonation, the macrolide ring can be
observed in MS spectra, allowing for the structural elucidation of erythromycin
metabolites. As a result, both of these alkali metal cations can be used to enhance the
fragmentation of erythromycin and aid in the identification of the position of any
structural alterations that have occurred during the metabolism of erythromycin and its

analogs.

60

CHAPTER 3: CONCLUSIONS AND FUTURE WORK

Summary and Conclusions

When used with BSI—MS/MS analysis, alkali metal cationization provides a
method to control the fragmentation pathways of pharmaceutical compounds during CID.
This concept has been demonstrated in this thesis for the compounds erythromycin and
levofloxacin. For both of these compounds, alkali metal cationization yields
fragmentation pathways that are not observed for protonated samples. For example,
lithium and sodium cationization of levofloxacin allow loss of hydrogen fluoride to be
observed, a phenomenon not observed for protonated samples. This allows detection of
fluorine substitution to be established. While protonation of erythromycin does not yield
fragments indicative of substitution locations on the macrolide ring, both lithium and
sodium cationization enable cross-ring fragmentation of the macrolide ring to take place.
This has also been observed in previous work investigating the structural elucidation of
erythromycin analogs and protecting groups used during their synthesis [40-42]. In the
case of erythromycin, the fragmentation obtained using alkali metal cationization
provides a means of enhancing fragmentation and determining the location of structural
modifications that have been made to the molecule during the course of metabolism.

For erythromycin, levofloxacin, and ciprofloxacin, as the mass of the alkali metal
cation used increased, the amount of fragmentation decreased. Dorr et al. [31] and Cui et
al. [16] observed a similar phenomenon, as discussed in Chapter 1. As the mass of the
cation increases, the extent to which the cation is able to move throughout the molecule

decreases. For example, the only fragmentation observed for potassium-cationized

61

samples of levofloxacin and ciprofloxacin is formation of the potassium cation from
[M+K]+, yielding no useful structural information. Similarly, potassium cationization of
erythromycin only yields neutral loss of cladinose and formation of potassium ion. Also,
for all three analyte compounds, sodium cationization yields fewer fragment ions than
lithium cationization.

Molecular structure also plays a role in the fragmentation pathways that are
observed. For sodium-cationized ciprofloxacin samples, the only observed fi'agment ion
is Na”. This is also observed for sodium-cationized levofloxacin, but these [M+Na]+ ions
also fiagment to lose both CO2 and hydrogen fluoride. Levofloxacin has an additional
cyclic structure that ciprofloxacin lacks, but the reasons why this influences the
fragmentation behavior remains unclear. Also, both potassium-cationized
fluoroquinolones fragment to lose potassium. Potassium-cationized erythromycin
samples behave similarly, but potassium is lost at a higher collision voltage and
fragmentation of the analyte molecule itself is observed. Because erythromycin has a
greater mass than the fluoroquinolones, it has a stronger interaction with potassium.

Alkali metals interact with neutral analytes differently than a proton. A proton
has a smaller mass, so it can more readily migrate across the analyte molecule. Because
alkali metal cations have masses greater than protons, they are less mobile than protons.
Some fragmentation pathways that are unique to protonated analytes may involve
migration of the attached proton, and yield ions that cannot be formed form metal
cationized analytes. Through blocking of fragmentation pathways that require migration
of the proton, alkali metal cationization provides a method to direct fragmentation of

pharmaceutical compounds into pathways that provide structural information not

62

available from protonated species. This feature can be applied to MS/MS metabolite

analysis in order to aid determination of structures of pharmaceutical metabolites.

Future Work

To further investigate alkali metal cationization of pharmaceuticals, similar
experiments should be conducted on other fluoroquinolones and other macrolide
antibiotics and their metabolites. Examination of other classes of pharmaceutical
compounds will be pertinent. In addition, generation of metabolites through in vitro
incubations with liver microsomes, and subjecting the isolated metabolites to alkali metal
cationization CID should aid in the determination of the locations and identities of any
metabolic alterations.

Other compounds that are metabolically important should also be investigated.
For example, fatty acids and their metabolites can serve as biomarkers for various
diseases [43]. Many oxylipin metabolites are regioisomers, and alkali metal cationization
could provide a means for distinguishing these compounds. LC-MS can also be applied
to separate the various regioisomers [44], and post-column addition of alkali metal salts

could be used to generate metal cationized ions for subsequent CID analyses.

63

APPENDIX I

Pseudo-MS3 spectra for Chapter 2

64

va 1_42-1

 

 

 

 

110612007SLA_15 5 (0.108) Cm (1:46) A 4: Daughters of 318ES+
205 7.9463
100
' o\° ‘ 206 219221 254
0l_ll'[|{r1 '1 2'15'V|IL*39\‘21?11'1 ' 24%??57T '7 TV T261 Y'_' 'W' T" *Tfi"7 1 fi' fiT‘ '7 r "jl'alig'
110612007SLA_15 5 (0.102) Cm (1 :46) B 3: Daughters of 318ES+
205 4.4794
100
219221 261
°\°' 216]
206 231
O,1L-z,{1‘. , .: 5:214. - 1 11:1??? 11111 214.512.4317 111111 .rr141erm11fi1r. 1.11-.-.fi11, 11.1--.1,-r- 1 ....... 31.8, .....
110012007SLA_15 6 (0.119) Cm (1 :47) c 2: Daughters of 316ES+
261 2.21e5
100
318
o\°
0 2(15'219221 233 241 247 301 31
110ct2007SLA_15 6 (0.113) Cm (1 :47) D 1: Daughters of 318ES+
318 7.17e5
100
.\°
''''''''''''' 1 111,11 "' Wrrr'frfir r' 1 r""l *"r '1'" r' "' "" rim

 

200 210 ' 220 230 240 250 260 270 260' 290 ' 300 73105'73207
Figure 43. CID product ion spectra of m/z 318 for protonated levofloxacin at (A) 55V,
(B) 40 V, (C) 25 V, and (D) 10 V

 

 

 

LVX1_42-1
1100t2007SLA_16 7 (0.145) Cm (1 :46) A 3: Daughters of 261ES+
100 221 361
“ 205
192 208
' °\°“ 258
1751791180182 188 195 200203 11 214 220 225 229 232 233 242243 249 255 261266266
0‘ .
110ct2007SLA_16 7 (0.140) Cm (1 :47) B 2: Daughters of 261ES+
179 6.4483
1007
32.
. 219
176 202 221 261
172 184 197 205 22 253
o .1121,” 1126. 133194111 2). .1 2:11.215- .1, 1.1.5-- .293, -2143: 35319351,. 3-,123222'11270
110ct2007SLA_16 5 (0.091) Cm (1 :47) c 1: Daughters of 261ES+
100, 261 1.24e4
220
l
331
1 176 179
In I) 187 188‘ 194 197 203 205 212 215 221 225231 234 238 244 252255 258 265 268

 

 

0
170 175 180 165 190 195 200 205 210 215 220 225 230 235 240 245 250 255 250 265 270
Figure 44. CID product ron spectra of m/z 261 for protonated levofloxacin at (A) 40 V,

(B) 25 V, and (C) 10v

65

18Jun620088LA_06 13 (0.280) Cm (1 :45) A 4: Daughters d 314ES+

 

10 216 297
216
- 5°
202 302
209 222229230 31 237245 246 255 253 267 272 261282 235 29° 297 306 306 316 322 323
. .
18Jun620088LA_06 12 (0.253) Cm (1:46) B31 Daughters 01314ES+
231 5.1964
100
$1
2‘33 208 316 216 321 23‘(”234,235 242 249 257260 266 272(274287?(W280V 293: w300 305,306 315313323
18JuneZOO8SLA_06 11 (0.226) Cm (1 :47) c 2: Daughters or 314ES+
100 231 9.7384

0
o\

314
220922992221: -224 2.13.3.3??? 244- -254. 2122?“? 27227.3, 3" -.2,85.293.E?4399991311” 1 9115223324

18Jun62008SLA_06 13 (0.263) Cm (2:47) 01: Daughters of 314ES+
100 314 1.8885

0
o\

o .202 2.04-212 2.17. 99 . 229.231- 237 21.2. 2.48 2.51 .258, 264326272229 ZwZM?§LZ%3§J05 3‘3, 2692913212
200 210 220 230 240 250 260 270 260 290 300 310 320
Figure 45. CID product ion spectra of m/z 314 for protonated ciprofloxacin at (A) 55 V,

 

 

 

(B) 40 V, (C) 25 v, and (D) 10 v
Clprofloxacln, 1_85-1
18Jun620088LA_09 12 (0.258) Cm (1:46) A 4: Daughters 01288ES+
203 2.5803
100
I a\°
215
18Jun620085LA_09 12 (0.253) Cm (1:46) B 3: Daughters of 288ES+
203 4.47e4
100
o\° 205
o.-1. . 2- 2:9--. .. 22%??12 - .. -. . v-2- ., fiwmu
18June20088LA_09 12 (0.247) Cm (1:46) C 2: Daughters 01288ES+
245 1.1465
100
0\" 204205 231 288
[H , 217 g 2'1” 1 240, , -fi fi 2138 -
18June20088LA_09 12 (0.242) Cm (1:47) D 1: Daughters 01288ES+
100 288 6.9165

0
o\

 

(200' 205 2107215 2207225 250' 255 124072515 '250' 255' '260' '255' '270 '275' 260 '265 290' 1251;",2
Figure 46. CID product ion spectra of m/z 288 for protonated ciprofloxacin at (A) 55 V,
(B) 40 V, (C) 25 V, and (D) 10 V

66

Ctprofloxacln 1_65-1. 231

 

 

 

 

21 June20066LA_31 14 (0.291) Cm (1 :47) A 2: Daughters of 231Es+
203 636
1001
1
I °\°«
201 231
. 203 215
l 204 207 206 211 2‘2 216217216 219 222 223225 227 226 229 231 233 234
G I ‘r 1 TJ #flTLIA r‘ r 1t 71 r‘ r filr 71 r1 I{ r 1‘ 1' LirlrFLr )r 1T1 J ‘r—Lr r 17 [F] I
21 June2006$LA_31 14 (0.266) Cm (1 :47) 3 1: Daughters 01231ES+
1001 231 7.9264

‘Zo

 

 

 

o t r r r f l r r r 1 f r 1 r r T r r r r r r r r r v 1 r r r r r r I
200 202 204 206 206 210 212 214 216 216 220 222 224 226 226 230 232 234
Figure 47. CID product ion spectra of m/z 231 for protonated ciprofloxacin at (A) 25 V
and (B) 10 V

Clpmfloxacln Ll, LBS-2
18Jun620088LA_17 19 (0.409) Cm (1:46) A 4: Daughters of 294ES+
220 192

1001

 

 

 

302
204 224226230 232 241243 264 271 274 260 264 266 30°

OlvlllL' LY '2111'12'1162 lAlrll 1,.(fiff. L'L £4 vgfivagl$.2?vovwl' ll '11 ll: V L117 1 IL '1'2fi7_5'l fllr397
18Jun620088LA_17 18 (0.371) Cm (1 :47) B 1: Daughters of 294ES+
10% 294 3.6563

3.

242 275
vvvvv 7 L rvvvrrv-Y'1""r"" "" l

 

 

 

200 ' ' 210' .- 220 230 " 240 250 260‘ 1 270 ' T 260' ""290 ‘7300 ' 311)“
Figure 48. CID product ion spectra of m/z 294 for lithium-cationized ciprofloxacin at
(A) 35 V and (B) 10 V

67

LVX Ll 1_8‘| J, 324

 

 

 

 

 

 

21June20088LA_37 10 (0.205) Cm (1:47) A 2: Daughters of 324ES+
268 234
1001
324
I $1 272
246 257 259 304
271
4111.1ij 30232 240 Nil: 253 274 280 288 294 301 309 319 321 325330
02ll_L21'.iilii12.1..iifiliiiitl'itiliiiiiii.li1ii.l.rll/_._ir - LIJLILJILiLllLJLTlAllAVAlllllllJ—[JlAlllL5111'All‘l gilgg
21Jun620088LA_37 12 (0.242) Cm (1:45) B 1: Daughters of 324ES+
206 1.1664
1001
°\°4
205
324
207
304

 

 

 

0 m4 fir..- . ......... .1. ,.- -,. ,- ..,. .4 . . ---, w. e .md. . .1- .mnvmrfi nmlz
200 210 220 230 240 250 260 270 260 290 300 310 320 330
Figure 49. CID product ion spectra of m/z 324 for lithimn-cationized levofloxacin at (A)
25 V and (B) 10 V

Erythromycln - Na 1_26-1. 309

22Mar2008$LA_31 17 (0.350) Cm (1 :47) 1: Daughters of 309ES+
309 1.4063

1001

010

265
268

273
350 315

r‘ ' YVTL 1‘ 1."11 L- {$2 1'1" 1 VL Fl 21'. z m,‘ 111. 11 z

250 255 260 265 270 275 280 285 290 295 300 305 310 315 320

Figure 50. CID product ion spectrum of m/z 309 for sodium-cationized erythromycin at

10 V

 

 

ll .‘ALA+
w—v— vY—v V'V‘fi

.
v—vrv

 

68

Erythromycin Ll1_77-1,582

21Jun620088LA_16 12 (0.259) Cm (1:46) 4: Daughters of 582ES+2
10 3995 359 388
346 374 H497 505

. °\o 313 320 340 364 403 408 420 43044546 528531547

l ‘ . l 3 . 1 ' l . :iJ

. n. . , 1 , I .

21June20088LA_16 11 (0.232) Cm (1:46) 3: Daughters of 582ES+

307 291
100

450

o\°

 

309
95 f 328 350 .3133? 372 388394 407 418429 433 462 465 482488 503 511526 540551562573 587

o JAlLLAJ—‘IAA‘IIILLT Al.‘_.'_.l Liul .11:..l..(
r'fi‘Tf' l"

vvvvvvvvvvvv

 

vvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvvv

 

21Jun620088LA_16 10 (0.205) Cm (1 :47) C 2: Daughters of 582ES+
58 1.6233
100
‘2 346
o 295 31.9 5.25525 1. .559. 518391192. 425.111 -450355 435 321.495, 55.53.1534 553558581 386
21June20088LA_16 12 (0.243) Cm (1 :47) D 1: Daughters of 582ES+
582 1.1764
100
o\°
31.50 437 450 464 478 582

 

G@3150 " 320 " "340 ' 360 330 ’V 460 ’ 420 ' 440 ' 450 ' 4léo ' 560'” 520' “"540WY556” 536’“
Figure 51. CID product ion spectra of m/z 582 for lithium-cationized erythromycin at
(A) 55 V, (B) 40 V, (C) 25 V, and (D) 10 V

69

APPENDIX 11

Tables of Relative Abundances for Mass Spectra in Chapter 2

7O

56 SAN

COCO
COCO

mm; _ 2 .mm
86

mod

36

N6

COCO

A> m3 ~> o3

loggesfi .x. leases? .x.

vmfim
$6
no.—
on.—
wm; _

8.?
mod.

EN

mud

2.2

Ga
89653.. ..\o

N. _ .o
36
o
mmd
SN

34.0
h _ .o
3 .o
0.0
YN

fl 2
.o_o..§....§..ig< .x.

32
mmm
9%
Sn
own

mm _
mmn
own
wwm
2%

g

and

E 53 555255
3. 553555
3.5
:12 Elena. Edda

A0338 Eammmz €063 mama: 3:590 mos—«3 mausommou 29:3
025 89a wofifipo 83$, owfiozw mfioxfiofieco “3330:8483“ 28% mo 88on :3 835a 8m meanness“? com gamma A 2an

71

bmdm

m _ .—
:6
Fwd
2.0

mm;

Ned
wad
EH
5.“.
2.2

j a
{355$ .x.

3.3
xv.—

mod
nmd
mad
c
8:0.
cod

3d
Sam
3;
2.9
3.”
Know

No.0
wed
ow. m
wad
35
31%

m> ovv

oo§|v§|n|<

o\o

vmd

o
o
86

~> may

Illll8§§< .x.

C

CCOCCC

CCOCCC

OO

O
o

H>o3

33

am
So

3N
o;
:v
mNm
now
own

mam
Nov
wow
omm
vow
Nwm

_ _m
mmv
moo

aom
mmv
mam

g
:3 Enamfim

~55

mi.

o3.

Sh

own

33

Icemllucoam.

M - Eifiohfibm

A95 5 - 503855

5 - 505255

83v “2 - ciaoafim

«Z - onfiofiecm

EB. .Im.o E._.o. OI

8.858 gamma—2 £885 mama: 3:390 8:226 anaemia 038%
ooh: 80% 33390 83.3 owflozw mfioxaoufibo voNEouaoASoS 28:5 mo «.50on a2 8389 .8.“ 8885?. now 2630M .N 2an

72

 

 

 

 

 

Sam 3:0. mod e mi
34 find vmd o Sm
mod 36 3.6 o mmm
36 Ed chum nod mom
o mod 9.3 $5 0mm
o o _ o 3% we». 53
guano—ungua—
Nhg _.m 26 0 ma—
x: was 86 o 8m
36 own a: o SN
26 36 36m 36 SN
o mod Gem 2:. w _ m
o 36 SN _ _.o 3m mom anxoao>3
9 9 9 CI, 33% SE :2 “.55 9%
we 8.3%? .x. as 8539‘ .x. we 8.35% .x. o: 835$ .x.
Aoagom gammmz E033 mama:

3:330 mofigv 830:3“ 295m 8% 88m 35330 83? owfiozw ”Eowxougfl 38:88.“ 8m moogasnm com 9630M .m 033.

73

o VmN mod o mm

:5 Ed ME O 8m 3m 25¢
a2 - 5088—354

8.» amfi an; o mm
Nod 56 EN 9 omm cum «2 - Eomxougog
3.». _.n no...” em; KN
o $4 :.m o 3m
omfi Q .m w _ H mad mmm awn and
E .. Sofie—«god
36 3.» mod o own
an; 36 mg. 86 gm
mmd ah: vm.m_ o vmm mom 3 - 5088—354
AN m8 ~> on“ A> mm“ ~> c; an}: an?“
8:32;? . p< .x. 85653.. o\o 85c§9< o\o oo§§n< .x. =3 “seams:— =3 38am. nggEoU

A0338 £4332 £855 wfim: 3:330 mos—«é
83232 038mm 025 89¢ 35390 838, owns; ”ngocgo— BantmoASuE max? .8,“ moogasnw :2 guflom .v 2an

74

84

w _ d

NV.—
0

o
o
o
00:
36
aoN
o

o
c
o

A> wmu

185%? .x.

mm.—
waN
3d
Ed

cod

0.:
mega
mung
mad
o
o
3.0

Clr o3
33%:

who 0 vow
0mg” 0 SN
mm: 0 NR
05.2 mod SN
wad 86 EN
mv.m_ aim 3m
2mm $6 an
00.: o mom
”mm 0 non
ms 8.0 ZN
aodm Ed mg
0%». mad men
8.» mad «mm
0:: mm; Zn
01.3 910:: g
31% 3 a
thgom 503302 0:803 mfim:

Rm 8:: 5225206
Nmm 5088—9055
:3:
3 0:5... 0. .9: 0.0

005050 8303 8302:»: 29:8 00:5 80¢ 3:830 003? $802» ”floaxomoamo 080808: :09 80:00:50 :3 3:23“ .m 2an

75

36 E..— wmd 36 SN
v _.~ 33 o mum
wvd imam Noam mod com 9% and
S - 5035895
36 2mm and o :N
$3 $6 amd o 3m
36 3.3 2.3 Ed 3N wmm 5 - Euaxomoamo
9 m8 ~> on“ 9 mm“ ~> c: an?“ HEEV
ologluqsun. <..°\.m wlocmvgfix .x. clog. ! 5:5. <. o\o_ .o. |o§v||§m< ..o\.. a; 3 lEuawlfim I|=£ Elofilh v. as... odE...|oU.-

AonBUuOm Eammmz E083 mam: 3:330
8353 830:9: 29:8 85 80¢ 3:590 829, omega mfiomxocoamo uoNEowmo-E=Eu—= 8m 8058:? now 0333M .w 2an

76

BIBLIOGRAPHY

77

10.

ll.

BIBLIOGRAPHY

Tolonen, A., et al., A simple method for difi'erentiation of monoisotopic drug
metabolites with hydrogen-deuterium exchange liquid
chromatography/electrospray mass spectrometry. European Journal of
Pharmaceutical Sciences, 2005. 25(1): p. 155-162.

Kanazawa, H., et al., Determination of omeprazole and its metabolites in human
plasma by liquid chromatography-mass spectrometry. Journal of Chromatography
A, 2002. 949(1-2): p. 1-9.

Bauer, K.S., S.C. Dixon, and W.D. Figg, Inhibition of angiogenesis by
thalidomide requires metabolic activation, which is species-dependent.
Biochemical Pharmacology, 1998. 55(1 1): p. 1827-1834.

Hatfill, S.J., et al., Induction of Morphological-Diflaentiation in the Human
Leukemic—Cell Line K562 by Exposure to Thalidomide Metabolites. Leukemia
Research, 1991. 15(2-3): p. 129-136.

Pumapatre, K., S.K. Khattar, and KS. Sain, Cytochrome P450s in the
development of target-based anticancer drugs. Cancer Letters, 2008. 259(1): p. 1-
15.

Dalla Corte, C.L., Fachinetto, R., Colle, D., Pereira, R.P., Avila, D.S., Villarinho,
J .G., Wagner, C., Pereira, M.E., Nogueira, C.W., Soares, F.A.A., Rocha, J .B.T. ,

Potentially adverse interactions between haloperidol and valerian. Food and
Chemical Toxicology, 2008. 46: p. 2369-2375.

Hellum, B.H., Z.H. Hu, and O.G. Nilsen, The induction of CYP1A2, CYP2D6 and
CYP3A4 by six trade herbal products in cultured primary human hepatocytes.
Basic & Clinical Pharmacology & Toxicology, 2007. 100(1): p. 23-30.

Umathe, S.N., et al., Quercetin pretreatment increases the bioavailability of
pioglitazone in rats: Involvement of CYP3A inhibition. Biochemical
Pharmacology, 2008. 75(8): p. 1670-1676.

Siuzdak, G., Mass Spectrometry for Biotechnology. 1996, San Diego: Academic
Press.

de Hoffmann, E., Stroobant, V, Mass Spectrometry: Principles and Applications
3rd ed. 2007, West Sussex: John Wiley & Sons Ltd.

Kebarle, P. and L. Tang, From Ions in Solution to Ions in the Gas-Phase - the
Mechanism of Electrospray Mass-Spectrometry. Analytical Chemistry, 1993.
65(22): p. A972-A986.

78

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

Gates, P. Electrospray Ionisation (ESI). 2004 Jan. 20, 2004 [cited 2008 May 23,
2008]; Available from: http://www.chm.bris.ac.ulgms/theorv/esi-ionisationhtrnl.

Gabelica, V., De Pauw, E., Internal energy and fragmentation of ions produced in
electrospray sources. Mass Spectrom Rev, 2005. 24: p. 566-5 87.

Yost, RA. and CG. Enke, Triple Quadrupole Mass-Spectrometry for Direct
Mixture Analysis and Structure Elucidation. Analytical Chemistry, 1979. 51(12):
p. 1251—&.

Sleno, L., Volmer D.A., Ion activation methods for tandem mass spectrometry. J.
Mass Spectrom, 2004. 39: p. 1091-1112.

Cui, M., Song, F., Liu, Z., Liu, S., Metal ion adducts in the structural analytsis of

ginsenosides by electrospray ionization with multi-stage mass spectrometry.
Rapid Commun. Mass Spectrom, 2001. 15: p. 586-595.

Madhusudanan, K.P., Raj, K., Bhaduri, A.P. , Effect of metal cationization on
the low-energy collision-induced dissociation of loganin, epi—loganin and
ketoganin studied by electrospray ionization tandem mass spectrometry. J. Mass
Spectrom, 2000. 35: p. 901-911.

Adams, J. and M.J. Songer, Charge-Remote F ragmentations for Structural
Determination of Lipids. Trac-Trends in Analytical Chemistry, 1993. 12(1): p. 28-
36.

Cheng, CF. and ML. Gross, Applications and mechanisms of charge-remote
fragmentation. Mass Spectrometry Reviews, 2000. 19(6): p. 398-420.

Lopes, N.P., et al., Fragmentation studies on monensin A and B by accurate-mass
electrospray tandem mass spectrometry. Rapid Communications in Mass
Spectrometry, 2002. 16(5): p. 414-420.

Lopes, N.P., et al., Fragmentation studies on monensin A by sequential
electrospray mass spectrometry. Analyst, 2002. 127(4): p. 503-506.

Lopes, N.P., Almeida-Paz, RA, and Gates, P.J., Influence of the alkali metal
cation on the fi'agmentation of monensin in ESI-MS/MS. Brazilian Journal of
Pharmaceutical Sciences, 2006. 42(3): p. 363-367.

Smith, G. and J .A. Leary, Mechanistic studies of diastereomeric nickel(II) N-
glycoside complexes using tandem mass spectrometry. Journal of the American
Chemical Society, 1998. 120(50): p. 13046-13056.

Ann, Q.H. and J. Adams, Collision-Induced Decomposition of Sphingomyelins for
Structural Elucidation. Biological Mass Spectrometry, 1993. 22(5): p. 285-294.

79

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

Ann, Q.H. and J. Adams, Structure-Specific Collision-Induced F ragmentations of
Ceramides Cationized with Alkali-Metal Ions. Analytical Chemistry, 1993. 65(1):
p. 7-13.

Ann, Q. and J. Adams, Structure Determination of Ceramides and Neutral
Glycosphingolipids by Collisional Activation of [M + Lz]+ Ions. Journal of the
American Society for Mass Spectrometry, 1992. 3(3): p. 260-263.

Adams, J. and ML. Gross, Tandem Mass-Spectrometry for Collisional Activation
of Alkali-Metal—Cationized F atty—Acids - a Method for Determining Double-Bond
Location. Analytical Chemistry, 1987. 59(1 1): p. 1576-15 82.

Asarn, MR. and G.L. Glish, Tandem mass spectrometry of alkali cationized
polysaccharides in a quadrupole ion trap. Journal of the American Society for
Mass Spectrometry, 1997. 8(9): p. 987-995.

Madhusudanan, K.P., Raj, K., Bhaduri, A.P., Eflect of alkali metal cationization
and multiple alkali metal exchange on the collision-induced dissociation of

loganic acid studied by electrospray ionization tandem mass spectrometry. Rapid
Commun. Mass Spectrom, 2000. 14: p. 885-896.

Adams, J. and ML. Gross, Energy-Requirements for Remote Charge Site Ion
Decompositions and Structural Information from Collisional Activation of Alkali-
Metal Cationized Fatty Alcohols. Journal of the American Chemical Society,
1986. 108(22): p. 6915-6921.

Dorr, F.A., Tomaz, J .C., Lopes, NP, and Pinto, E., Comparitive analysis of the
gas-phase reactions of cylindrospermopsin and the difference in the alkali metal
cation mobility. Rapid Commun. Mass Spectrom, 2008. 22: p. 2015-2020.

Sprandel, K.A., Bodvold, K.A., Safety and T olerability of Fluoroquinolones.
Clinical Cornerstone, 2003. Supplement 3: p. 829-836.

Sarkar, M., et al., Invitro Efi'ect of F luoroquinolones on Theophylline Metabolism
in Human Liver-Microsomes. Antimicrobial Agents and Chemotherapy, 1990.
34(4): p. 594-599.

Derby, L.E., et al., Erythromycin-Associated Cholestatic Hepatitis. Medical
Journal of Australia, 1993. 158(9): p. 600-602.

Karthik, S.V. and D. Casson, Erythromycin associated Cholestatic hepatitis and
liver dysfunction in children: The British experience. Journal of Clinical
Gastroenterology, 2005. 39(8): p. 743-744.

80

36.

37.

38.

39.

40.

41.

42.

43.

44.

Lindstrom, T.D., B.R. Hanssen, and SA. Wrighton, Cytochrome-P—450 Complex-
Formation by Dirithromycin and Other Macrolides in Rat and Human Livers.
Antimicrobial Agents and Chemotherapy, 1993. 37(2): p. 265-269.

Zweers-Zeilmaker, W.M., et al., In vitro complex formation and inhibition of
hepatic cytochrome P450 activity by diflerent macrolides and tiamulin in goats
and cattle. Research in Veterinary Science, 1999. 66(1): p. 51-55.

D'Agostino, P.A., J .R. Hancock, and LR. Provost, Electrospray Mass-
Spectrometric Characterization of Fluoroquinolone Antibiotics - Norfloxacin,
Enoxacin, Ciprofloxacin and Ofloxacin. Rapid Communications in Mass
Spectrometry, 1995. 9(11): p. 1038-1043.

Hemeryck, A., et al., Pharmacokinetics, metabolism, excretion, and plasma
protein binding of I4C—levofloxacin after single oral administration in the Rhesus
monkey. Drug Metabolism Reviews, 2005. 37: p. 43-43.

Crowe, M.C., et al., Characterization of erythromycin analogs by collisional
activated dissociation and infrared multiphoton dissociation in a quadrupole ion
trap. Journal of the American Society for Mass Spectrometry, 2002. 13(6): p.
630-649.

Gates, P.J., et al., Structural elucidation studies of erythromycins by electrospray
tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 1999.
13(4): p. 242-246. '

Kearney, G.C., et al., Structural elucidation studies of erythromycins by
electrospray tandem mass spectrometry 11. Rapid Communications in Mass
Spectrometry, 1999. 13(16): p. 1650-1656.

Waugh, R.J., et al., Identification and relative quantitation of F-Z-isoprostane
regioisomers formed in vivo in the rat. Free Radical Biology and Medicine, 1997.
23(6): p. 943-954.

Orellana-Coca, C., et al., Analysis of fatty acid epoxidation by high performance
liquid chromatography coupled with evaporative light scattering detection and
mass spectrometry. Chemistry and Physics of Lipids, 2005. 135(2): p. 189-199.

81

uIgtitigigmgtjmiijtI