Citation
Exploring neuropeptide metabolism in the brain by mass spectrometry

Material Information

Title:
Exploring neuropeptide metabolism in the brain by mass spectrometry
Creator:
Blagojević, Tamara
Publication Date:
Language:
English
Physical Description:
x, 160 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Crystals ( jstor )
Ionization ( jstor )
Ions ( jstor )
Mass spectrometers ( jstor )
Mass spectroscopy ( jstor )
Metabolism ( jstor )
Neuropeptides ( jstor )
Quantification ( jstor )
Signals ( jstor )
Solvents ( jstor )
Chemistry thesis, Ph. D ( lcsh )
Dissertations, Academic -- Chemistry -- UF ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2004.
Bibliography:
Includes bibliographical references.
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by Tamara Blagojević.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
3100544 ( ALEPH )
AA00004689_00001 ( sobekcm )
770716518 ( OCLC )

Downloads

This item has the following downloads:


Full Text











EXPLORING NEUROPEPTIDE METABOLISM IN THE BRAIN
BY MASS SPECTROMETRY














By

TAMARA BLAGOJEVIC


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2004
































Copyright 2004

by

Tamara Blagojevid
































To my family, friends, and teachers.














ACKNOWLEDGMENTS

I wish to thank my whole family for all their love and support. I am especially

grateful to my parents and sister for understanding my adventurous spirit and supporting

some of my decisions that changed all our lives. I also wish to thank my husband for all

his love and help. He has always been there to comfort me, encourage me and help me

survive some of the most difficult times of my life.

I would like to acknowledge Dr. John Eyler, who helped me find the way to

accomplish my goals. He showed interest in my work, as well as in my progress towards

getting the degree. I am very thankful to Dr. Laszlo Prokai for being confident and

enthusiastic about my research, for offering me a great chance to apply my knowledge

and giving me the opportunity to be involved in the team work while opening new

perspectives of my research interests. I also express my gratitude to Dr. David Powell for

his guidance and patience as I was initially exploring the mass spectrometry world.

Special thanks go to Will Haskins for our endless talks, games and songs that made my

life in the lab bearable and for encouragement and understanding that helped me survive

the first few years of graduate school.

I am enormously grateful to Dr. James Deyrup- without his kindness and care I

would have never been given the opportunity to explore new avenues. He was involved

and helpful, always had a word of wisdom, and encouraged my spirits whenever I needed

it.








I learned a lot and was motivated to learn more by many of my teachers.

Therefore, I am grateful to all of them for inspiring my life-long search for enthusiastic

and interesting scientific ways.

Without the support and encouragement of my friends, my life would have

followed some other path. With her patience and persistence, Ivana was there to make

me eat and tolerate my antsy moods; we were broke and on our own in the new world,

but we never gave up on our journey to the stars. Irina was there to be our guide and a

parent as we came to the new continent, and Isa and Janina showed us that one can have

so much in common and find a best friend in someone who grew up in a different part of

the world. I also got enormous support from my best friends in Serbia; they were also

with me through all my ups and downs. If it weren't for all of my friends, I would have

not been as happy and fulfilled.















TABLE OF CONTENTS
Page

ACKNOWLEDGMENTS ........................................................................................... iv

A B ST R A C T ....................................................................................................................... ix

CHAPTER

1 INTRODUCTION..................................................................................................1

A Century of Mass Spectrometry ...............................................................................
Ionizing Large Molecules.............................................................................. 1
Matrix-assisted laser desorption/ ionization time-of-flight mass
spectrometry (MALDI-TOFMS) ...................................... ............. 2
Electrospray ionization quadrupole ion trap mass spectrometry (ESI-
Q ITM S).............................................................................................. 9
Quantification by Mass Spectrometry........................................... .............. 12
Coupling Liquid Chromatography (LC) to Mass Spectrometry...............................13
R reverse Phase LC .......................................................................................... 15
C apillary LC .................................................................................................. 15
N euroscience.........................................................................................................16
N europeptides................................................................................................ 18
Sampling Techniques ..................................................................................18
In vitro studies...................................................................................... 19
In vivo studies- microdialysis.................................................................... 20

2 INVESTIGATION OF SAMPLE PREPARATION METHODS TO IMPROVE
MALDI-TOFMS PERFORMANCE AND FACILITATE INTEGRATION OF
CAPILLARY LC SEPARATIONS............................. .......................................35

Introduction........................................................................................................... 35
MALDI Sample Preparation Considerations..................................................35
Matrix crystal size ................................................................................36
Sample spot size...................................................................................37
Capillary LC/MALDI Coupling ..................................................................... 38
On-line coupling...................................................................................38
Off-line coupling ..................................................................................38
Imaging Techniques ....................................................................................39
Scanning electron microscopy (SEM).......................................................40
Transmission electron microscopy (TEM)...................................................40









Fluorescence imaging...........................................................................40
Experimental Methods..........................................................................................41
Instrumentation..............................................................................................41
Chemicals and Reagents.................................................................................42
Electrospray Deposition Method (ESD)..........................................................42
ESD apparatus......................................................................................42
Electrospray tip production....................................................................... 43
Hydrophobic Sample Supports (AnchorChip Plates)......................................43
LC /M A LD I.................................................................................................... 43
Capillary column packing ..........................................................................43
Off-line method system............................................................................44
On-line collection, off-line analysis system.......................................... ..44
R esults................................................................................................................ 45
Minimizing Matrix Crystal Size.......................................................................45
Optimization of ESD parameters ............................................................45
Matrix crystal images........................................................................... 46
Minimizing Analyte Spot Size ........................................................................46
Optimization of ESD parameters ............................................................46
Analyte spot imaging ...........................................................................47
AnchorChipTM plates............................................................................ 48
Capillary LC/MALDI-TOFMS ......................................................................... 48
Preliminary investigations.........................................................................48
On-line collection, off-line analysis data .............................................49
Conclusions and Future Directions.....................................................................50

3 NEUROPEPTIDE FF METABOLISM STUDIES AND DEVELOPMENT OF
QUANTIFICATION METHODS BY MALDI-TOFMS AND ESI-QITMS ............74

Introduction........................................................................................................... 74
Quantification Using Mass Spectrometry ...................................................74
Relative quantification .............................................................................74
Absolute quantification ............................................................................77
Quantification Employing MALDI-MS ......................................................... 78
Irreproducibility influences.......................................................................79
Internal standard selection.................................................................... 79
Quantification Employing ESI-QITMS ......................................................80
Chromatographic separation requirements..............................................80
Mass spectroscopic requirements.............................................................81
Neuropeptide FF............................................................................................81
Experimental Methods..........................................................................................82
Instrumentation ........................................................................................82
MALDI-TOFMS .....................................................................................82
ESI-Q ITM S .......................................................................................8.... 83
Chemicals and Reagents..............................................................................83
Sampling Techniques ..................................................................................84
In vitro approach ..................................................................................84
In vivo approach ...................................................................................85









Sample Preparation Techniques........................................................................86
Sample desalting methods.........................................................................86
MALDI sample preparation....................................................................87
R esults................................................................................................................... 87
Metabolic Differences ................................................................................. 87
In vitro data ............................................................................................ 87
In vivo data .............................................................................................87
Quantification ofNeuropeptide FF Metabolism Using MALDI-TOFMS ..........88
Optimization of MALDI sample preparation...............................................88
Internal standard selection.................................................................... 89
Generating a calibration curve .................................................................89
Quantification ofNeuropeptide Metabolism Using ESI-QITMS .......................90
MS/MS optimization....................................... .....................................91
Generating the calibration curve ..............................................................92
Conclusions and Future Directions.....................................................................92

4 DEVELOPMENT OF DIFFERENTIAL QUANTIFICATION FOR MONITORING
METABOLISM OF DYNORPHIN 1-8 IN THE BRAIN ....................................... 112

Introduction......................................................................................................... 112
Opioid Peptides ...........................................................................................112
Discovery of dynorphins .............................................................. 113
Physiological role..................... ......................................................... 113
Neurochemical processing ......................................................................113
Differential Quantification Approach................................................................ 115
Chemical labeling of N-terminus .......................................................... 116
Chemical labeling of C-terminus ............................................................. 117
Experimental Methods..............................................................................................117
Sample Collection by Microdialysis ...............................................................117
Sam ple Preparation............................................................................................ 117
Derivatization methods ..........................................................................118
MALDI sample preparation .................................................................. 118
R esults................................................................................................................. 119
Selection of Derivatization Method..................................................................119
Method Examination ......................................................................................... 119
Influence of Inhibitors on Dynorphin 1-8 Metabolism in the Brain ................121
Time Dependent Metabolism of Dynorphin 1-8 in the Brain .........................121
Conclusions and Future Directions......................... .............................................122

5 CONCLUSIONS AND FUTURE DIRECTIONS ........................................ ...142

LIST OF REFERENCES .................................................................................................148

BIOGRAPHICAL SKETCH ..................................................................................... 160














Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

EXPLORING NEUROPEPTIDE METABOLISM IN THE BRAIN BY
MASS SPECTROMETRY


By

Tamara Blagojevid

August 2004

Chair: Dr. John R. Eyler
Major Department: Chemistry

The challenge of improving the reproducibility and sensitivity of matrix-assisted

laser desorption/ ionization time-of-flight mass spectrometry (MALDI-TOFMS) was

addressed by improving the sample preparation step. a-CHCA was used as the matrix

for all sample preparation methods. An electrospray deposition method and the use of

prestructured MALDI supports with dried droplet sample preparation were investigated

for improved sensitivity and reproducibility. AnchorChipTM plates focus the deposited

drop into a small, 200-Im diameter spot surrounded by a hydrophobic (Teflon-like)

coating. Detection limits below the attomole level were achieved when 1 jl of a 1 pM

peptide solution was analyzed. Electrospray deposition was explored for improvement in

crystal homogeneity and minimization of the sample spot size. Electrospray deposition

concentrated analyte into spots less than 300 jtm in diameter, while evenly distributing








and incorporating the analyte into the matrix layer. Optimized electrospray conditions

led to detection limits of< 10 amol when 15 nl of a 0.7 nM solution was electrosprayed.

Capabilities and performance of quantification approaches when using MALDI-

TOFMS and ESI-QIT were compared. Also, both methods were used for investigating in

vitro and in vivo processing ofneuropeptide FF infused into a mouse brain. It was shown

that MALDI-TOFMS provides better resolution and detection limits, but ESI-QIT

provides better reproducibility and dynamic range. Differences in NPFF metabolism

when sampled in vitro are present due to release of various enzymes upon cell lysis

during the in vitro experiment.

The approach of differential quantification was utilized for quantifying metabolized

dynorphin 1-8 fragments in the brain by MALDI-TOFMS. Dynorphin 1-8 introduction

and sample collection were performed in vivo via a microdialysis probe. Sample and

control were esterified with dO- and d3-methanol, respectively. The light and heavy

samples were mixed in a 1:1 v/v ratio, of which 1 tl was analyzed. Changes in fragment

abundance were then correlated to enzyme activity and the enzyme inhibitors involved in

processing of dynorphin 1-8. Enzyme activity was not changing with time, which was

confirmed by the ratio of unprocessed dynorphin 1-8 to the amount initially infused.













CHAPTER 1
INTRODUCTION

A Century of Mass Spectrometry

Sir J. J. Thompson won a Nobel Prize in 1912 for the pioneering research and

foundation of mass spectrometry. Almost a century later John Fenn and Koichi Tanaka

won a Nobel Prize for development of sophisticated ionization methods enabling

detection and characterization of proteins, peptides, nucleic acids, and polymers. Mass

spectrometry emerged from the need to define and explain basic physicochemical

properties of atoms, and yet became one of the most powerful methods for

macromolecular characterization of living systems. It has become an interdisciplinary

research methodology, impacting virtually every area of science from physics through

chemistry and biology, from geochronology and space research to studies of the physical

and chemical properties of materials, atoms, and particles, to proteomics and genomics

projects.

There are many different kinds of mass spectrometers, described generally by the

types of ionization sources, mass analyzers, or detectors used. The choice of ionization

method and the mass analyzer depends on the requirement of the particular application.

Mass spectrometers based on different approaches can, therefore, provide complementary

information.

Ionizing Large Molecules

Successful ionization is crucial for mass spectrometry. The method selection will

depend on the sample, sample origin, and data required. Ionizing volatile organic








molecules was routinely performed in mass spectrometric analysis by either electron

impact (El) or chemical ionization (CI). But, as the molecular mass of the analytes gets

bigger, they become less volatile. This challenge was faced and explored for years until

the successful development of matrix-assisted laser desorption/ionization (MALDI) and

electrospray ionization (ESI).

Matrix-assisted laser desorption/ ionization time-of-flight mass spectrometry
(MALDI-TOFMS)

Developed simultaneously in laboratories of Karas and Hillenkamp and Tanaka

during 1987,1,2 MALDI caused a revolution in mass spectrometry during the 1990s by

gaining the attention of biotechnological fields of research (like medicine, pharmacy,

bioorganic molecular synthesis). Results from both of these groups were published in

1988.3,4 Tanaka's work was based on use of a pulsed N2 laser (337 nm) and slurry of

glycerol with ultrafine metal powder as a matrix. Detection of the molecular ion of

carboxypeptidase-A (m/z 34,529) was demonstrated. Karas and Hillenkamp, on the other

hand, used a Nd:YAG laser (266 nm) to desorb proteins (like bovine serum albumin)

dissolved in a matrix solution containing nicotinic acid. The method was soon applied

for analysis of proteins with molecular weights in the range of 100 kD,s'8 with the idea of

using an organic UV absorbing matrices adopted (table 1-1).

Since presence of the matrix is what differentiates MALDI from other laser

desorption techniques, the importance of its role, selection and deposition is one of the

key issues to discuss. A high excess of matrix to sample is important, since the matrix

serves as the primary absorber of the UV radiation and breaks down rapidly, expanding

into the gas phase and carrying along undamaged analyte molecules (Figure 1-1).

Additionally, the high matrix/sample ratio reduces association between analyte molecules








and provides enough protonated and free-radical products to ionize the molecules of

interest.9 A number of detailed models for the MALDI ionization mechanisms have been

described,10-12 and will be discussed in more detail, but most of the experimental

parameters, like matrix selection, sample preparation method, and parameters used for the

ionization, are still predominantly empirically based.

Ion formation. Although having received noteworthy attention, the fundamental

processes of MALDI ion generation are not fully understood,13"16 and are still a matter of

active research. Good understanding of ionization pathways should help increase ion

yields, control analyte charge states and fragmentation, and enable studies of new classes

of analytes. A number of chemical and physical pathways for MALDI ion formation

have been proposed, but a general agreement is that there is no single mechanism that can

explain all the ions generated and observed within an experiment.3'17 A variety of

matrices, analytes, and experimental parameters, as well as different instrumental

characteristics, present major obstacles to ion formation mechanism studies. Some of the

parameters usually considered are laser wavelength and laser pulse width, but the choice

of matrix is known to be crucial for success in MALDI experiments.

Since the presence of matrix is necessary for MALDI, it is important to establish

the role of the matrix by studying its chemical nature (measuring proton affinity,

absorbance spectrum, ionization potential) and its crystal structure. Derivatives of

benzoic acid, cinnamic acid, and related aromatic compounds were recognized early on

as good MALDI matrices for proteins,7'18 2-(4-hydroxyphenylazo)-benzoic acid (HABA)

as a matrix for peptides and proteins,19 3-hydroxypicolinic acid (3-HPA) for

oligonucleotides,20'21 and the list went on as more matrices and new analytes were








introduced. Albeit, there are still no clear guidelines for matrix selection for a particular

analytical problem, since many of the uncertainties are related to matrix crystallization

and analyte incorporation. "Matrix assistance" and matrix selection should be better

understood through mechanistic studies of ionization process.

It is interesting to notice that only a few types of ions are commonly observed in

the positive mode MALDI experiments: radical cations, protonated molecular ions, and

cationized molecular ions. Several mechanisms have been proposed to explain ionization

processes during MALDI. The initial set of reactions is expected to provide "active"

matrix species that are ready to ionize the analyte molecules. These are followed by

processes that result in the final set of ions being detected. Some of the proposed

reactions for primary processes are summarized in figure 1-2.

One of the primary processes, multiphoton ionization of the matrix leading to a

formation of matrix radical cations,22 is shown in equation 1-1:

M n(hv) M+ +e- (1-1)

Another mechanistic possibility is that two or more separately excited matrix molecules

pool their energy to yield one matrix radical cation or a highly excited matrix molecule23-
27

MM ->M*M*-M+M" +e- (1-2)

M*M*+A -MM + A +e- (1-3)

This model is statistically favored, since it is more likely that a neighbor of an excited

matrix will be the next to be excited, rather than that the excited molecule will be hit by a

second photon. Besides, the matrix suppression effect also suggests a multicenter

ionization model,27 since it is observed that at sufficiently high analyte concentration








matrix ion signals can be completely suppressed in MALDI, lowering the overall

ionization yield. Excited-state proton transfer (ESPT) was a model proposed by Karas, et

al., in early MALDI studies. It is based on a presumption that the singly excited matrix

molecule is more acidic than when in its ground state, and the following is expected to

happen:

M+hv -> M* (1-4)

M*+A-(M H)- +AH (1-5)

M *+M (M H)- + MH (1-6)

However, none of the popular matrices is known to be particularly ESPT-active, resulting

in some doubts about this mechanism. The correlation of positive and negative ion

modes suggests that disproportionation reactions can be an active ionization mechanism:

2M "" >-(MM)* (M H)- + MH' (1-7)

(MM)* -> M + M (1-8)

Although this mechanism fulfils the energy requirements, the evidence for it is not yet

strong.29 An attractive idea is that the ions observed in the MALDI mass spectra are

already present in the solid sample and just ablated by the laser pulse,30 but it is difficult

to confirm that ions observed are truly preformed and not the result of secondary gas-

phase chemistry in the plume. Thermal ionization is also possible, and it can take place

directly at the particle surface or in the matrix bulk:

2M >M- + M (1-9)

It is found that the internal energy of molecules desorbed in an UV MALDI experiment

corresponds to effective temperature of around 500 K,31 a temperature at which a thermal

ionization/desorption reaction is not significant.








After the initial ion formation, sets of additional reactions occur in the MALDI

plume. This set of reactions provides intermediates that generate protonated and

cationized analyte ions, later observed in mass spectra. Some of the most important are

proton-transfer reactions, cationization reactions, and electron transfer reactions. Several

models include matrix-matrix proton transfer reactions.22,32'33

M 'M -- MH' + (M H)' (1-10)

(M H)' +e- -+ (M H)- (1-11)

M + e- -+ (M H)- + H' (1-12)

It is also found that the matrix-to-analyte proton transfer reaction:

MH + A -> M + AH' (1-13)

can be a thermodynamically favorable process.32 The presence of AH+ ions in the plume

was also confirmed in studies done by Bokelmann, et al.,34 and Kinsel, et al.,35

Furthermore, it was found that addition of alkali matrix salts to MALDI samples

promotes cationization of synthetic polymers.36 This encouraged the addition of other

ions, like silver and copper ions, to enhance polymer ionization by MALDI.37-41 In

general, cations do not even need to be added, especially when analyzing biological

samples, since the natural presence of Na and K+ is sufficient to give strong alkali

cationized signals. Several studies have provided strong confirmation of processes

involved in formation of cationized peptides.42'43 An electron transfer reaction is also

possible if the analyte has a lower ionization potential than the matrix:

M" +A-- M+A'" (1-14)

This process was observed when analyzing fullerenes, ferrocenes and metallocycles.44'45








A study done by Gluckmann and Karas on different matrices for various

preparation protocols and for different classes of analytes has shown that the magnitude

of the initial plume velocity can be a valuable and meaningful characteristic of the

MALDI process.46 It has been shown that the initial velocity is characteristic of a matrix

but may be affected by the preparation conditions and additives. Another characteristic

of the ablation process was demonstrated by Beavis and Chait, who showed that UV-laser

excitation results in the ablation of a shallow surface layer, up to a 1 Pm in thickness, due

to high absorption coefficients of matrix molecules.47

Experimental designs of MALDI experiments could be improved with further

exploration of the MALDI processes. Some of the expected benefits are manipulation of

fragmentation and charge states, improved ionization yields and better guidelines for

matrix selection, including possible new designs.

Time-of-flight ion separation. The first mass spectrographs designed by

Thompson48 and Aston49 used a magnetic field to separate ions according to their mass-

to-charge ratio. All the first mass analyzers were based on magnets until the 1940s when

advances in electronics and mechanics allowed development of ion separation based on

flight time. By the end of the 1940s Cameron and Eggers built an instrument they called

an ion velocitron, where the ion velocities were inversely proportional to the square root

of their mass-to-charge ratio.50 At the same time a patent for a similar instrument had

already been obtained by Stephens.51 In all the described instruments the ions of mass m

(kg) were accelerated to constant energy (eV) resulting in flight time t (s) across the drift

region L (m) given by:


t = 2 )2L (1-15)
2eV








The ions leaving the source and entering the time-of-flight analyzer have neither

exactly the same starting times nor exactly the same kinetic energy. Various TOF mass

spectrometer designs have been developed to compensate for these differences. During

the 1950s researchers suggested a delayed pulse for acceleration, enabling the ions to be

formed in a field-free region.52 This allowed an even start for all the ions and therefore

improved the resolution. In 1973 Mamyrin and co-workers introduced a first design for a

high resolution TOF MS, based on a dual-stage grid system.53 This innovation was

developed into the design of a reflectron, an ion optic device in which ions pass through a

"mirror" and their flight is reversed (Figure 1-3). The reflectron allows ions with greater

kinetic energies to penetrate deeper into the reflectron field than ions with smaller kinetic

energies. The ions that penetrate deeper will take longer to return to the detector. If a

packet of ions of a given mass-to-charge ratio contains ions with varying kinetic energies,

then the reflectron will decrease the spread in the ion flight times, and therefore improve

the resolution of the time-of-flight mass spectrometer. This invention started a new era

of development of high-resolution TOF instruments, based on delayed ion extraction and

use of a reflectron for longer drift tube length. A comparison of resolution that can be

achieved with these improvements is shown in Figure 1-4.

TOF is considered to be the fastest MS analyzer with the highest practical mass

range of all the MS analyzers. The pulsed desorption of ions favors the combination of

the laser desorption ion source with a time-of-flight mass analyzer54'55 or an FT-ICR mass

analyzer,5658 both capable of recording complete mass spectra for each laser shot. It can

provide MS/MS information from post-source decay, but is not commonly used for

MS/MS experiments due to the limited precursor-ion selectivity. The use and








development of TOF reemerged upon introduction of desorption ionization methods at

the end of 1980s. The introduction of MALDI and discovery of its many possibilities

have driven much of the current interest in TOF analyzers.

Electrospray ionization quadrupole ion trap mass spectrometry (ESI-QITMS)

Electrospray (ES) is a method by which ions present in a solution can be

transferred into the gas phase. However, electrospray existed long before its application

in mass spectrometry as a method of considerable importance for the electrostatic

dispersion of liquids and the creation of aerosols. Solvent evaporation from the charged

droplets leads to droplet shrinkage, increased coulombic repulsion, and release of the

electrostatic strain by droplet fission. Repeated evaporation and fission of parent and

offspring droplets leads to formation of a droplet population extending down to very

small multiply-charged droplets, with radii in the -10-100 nm range.

The first detailed studies of electrospray phenomena were carried out by Zeleny in

the early 1900s.59 In the late 1960s, Dole and co-workers were first to attempt using an

atmospheric pressure electrostatic sprayer as a means to produce gas-phase ions from

macromolecules in a liquid solution for mass spectrometric analysis.60-63 But the first

successful coupling of an ES ion source to a mass spectrometer is attributed to the Fenn64-

66 and Alexandrov67 research groups. This final confirmation of electrospray capabilities

caused an electrospray revolution and inspired rapid developments in the world of mass

spectrometry.

Ion formation. There are three major steps in the production of gas-phase ions

from electrolyte ions in solution by electrospray: 1) production of charged droplets at the

ES capillary tip; 2) shrinkage of the charged droplets by solvent evaporation and repeated

droplet disintegration, and 3) the actual mechanism by which gas-phase ions are








produced from the very small and charged droplets. There is still no consensus on the

mechanism by which analyte ions are formed from the charged droplets. Dole and co-

workers proposed the charge residue model (CRM), suggesting ions originate from small

droplets containing one molecule of the analyte.60 Evaporation of the solvent from the

initially formed droplet leads to a reduction in diameter, and an increase in surface field,

until the Rayleigh limit is reached. As the magnitude of the charge is sufficient to

overcome the surface tension holding the droplet together, a coulomb explosion occurs.

Continuation of this process results the formation of an ion containing a single analyte

molecule. The ion desorption model (IDM) is based on the work of Iribare and

Thomson.68,69 This model assumes that the surface electric field becomes sufficiently

large to lift an analyte ion at the surface of the droplet over the energy barrier enabling its

escape. Diagrams illustrating the CRM and IDM models of ion formation in electrospray

are shown in Figure 1-5.

A unique characteristic of ESI is extensive multiple charging of ionic species.

Number and placement of charge throughout a molecule depends on many experimental

factors. Some are dependent on the solution characteristics (like pH, solvent polarity,

analyte concentration and molecule conformation), while others depend on instrumental

operating conditions (like temperature, gas flow, and electrospray voltage). It has been

demonstrated that the number of basic sites on a peptide (or protein) directly relates to the

charge state of the specie observed in positive-ion mode ESI mass spectra.70-72

Quadrupole ion trap (QIT) operation and capabilities. Developments in this

field over the last 10 years have made the QITMS an excellent tool for biomolecular

analysis. The QIT is a three-electrode device (Figure 1-6). Two of the electrodes are








identical and are called end caps, while the third electrode is donut shaped and is called

ring electrode. The holes in the end caps are for ion injection into the device and ion

ejection out of the device to a detector. A three-dimensional quadrupole field is formed

when a suitable RF voltage is applied to the ring electrode, trapping the ions by

continuously forcing them toward the center of the device. Unlike triple-quadrupole

mass spectrometers, where each operation on the ion beam is separated in space, the QIT

operates on the ions over a period of time, but within the same analyzer. The ions are

dynamically trapped by the applied RF potentials, while the "bath gas" (He) helps contain

the ions in the trap. The RF and DC potentials can be scanned to eject successive mass-

to-charge ratios from the trap into the detector.

Wolfgang Paul first described a quadrupole ion trap mass spectrometer in 1953,73,74

a discovery for which he was awarded a Nobel Prize in 1989. The popularity of QIT has

been growing ever since, inspiring further developments and improvements in the design.

Coupling GC to QIT has shown that the resolution and sensitivity of the instrument were

greatly improved when approximately 1 mTorr of helium was present in the QIT

chamber.75 ESI was first coupled to a QIT in 1990,76 after which many upgrades led to

this system being a method of choice for various analytical applications.

Perhaps the biggest strength of the ion trap technique lies in its ability to perform

multiple stages of mass spectrometry, greatly increasing the amount of structural

information obtainable for a given molecule. The types of fragment ions observed in an

MS/MS spectrum depend on many factors, including primary sequence, the amount of

internal energy, or charge state. The accepted nomenclature for fragment ions is shown

in Figure 1-7.77,78 Fragments will only be detected if they carry at least one charge. If this








charge is retained on the N terminal fragment, the ion is classed as either a, b or c. If the

charge is retained on the C terminal, the ion type is either x, y or z. A subscript indicates

the number of residues in the fragment. In electrospray ionization, peptides generally

carry two or more charges, so that fragment ions may carry more than one proton.

Predominantly a, b and y ions are generated in ion trap CID experiments. In addition,

peaks are seen for ions that have lost ammonia (-17 Da) or water (-18 Da).

Quantification by Mass Spectrometry

Given its ability to provide specifics on molecular masses and fragmentation

information, mass spectrometry is predominantly used for qualitative analysis.

Quantitative methods have been underutilized and less explored, due to the variety of

other analytical techniques available and routinely performed in analytical, clinical,

environmental, and forensic labs. Mass spectrometry has been used to quantify drug

metabolites,79'80 sugars,81 and volatile organic compounds,8284 but MALDI and ESI

introduced new challenges to meet the need for quantifying proteins, peptides and their

metabolites. Given that this is an application driven method development, methods are

usually specific for a particular purpose. Further improvements in this field can be

expected upon automatization and additional instrumental developments.

Different quantification approaches have been developed using MALDI and ES

ionization techniques, some of which will be discussed in more detail later. While

MALDI MS analysis is tolerant to some impurities and reliable for mixture analysis, the

lack of reproducibility restricts its performance as a reliable quantification technique.

Large error is introduced due to heterogeneity of the sample distribution and matrix layer,

signal suppression, and non-linear detector response, causing spot-to-spot and shot-to-

shot irreproducibility. After many attempts, it has been shown that a highly homogenous








sample preparation is required for a reliable quantitative analysis.8588 Nevertheless,

MALDI-based quantification methods are still not routinely used and still are a challenge

for researchers. ESI MS, on the other hand, is much more sensitive to presence of salts

and other impurities, so most of the quantification attempts involve coupling with LC or

other separation technique.

A major problem of mixture quantification is the matrix effect of dominant

components causing signal suppression and loss of reproducibility. Also, mass resolution

and scanning performance of the mass spectrometer do not always allow for reliable

differentiation of two species similar in mass. Separation methods performed prior to

mass spectrometric analysis can significantly simplify and improve the quantification

analysis. By separating complex mixtures based on hydrophobicity, size, or

immunological properties of the components, each one of the components can enter the

mass spectrometer separately, minimizing the interference.

Coupling Liquid Chromatography (LC) to Mass Spectrometry

Chromatography includes a diverse and important group of methods that permit

separation of closely related compounds from complex mixtures. The applications of

chromatography have grown explosively in the last few decades, not only because of the

development of several new types of chromatographic techniques, but also due to a

growing need for better methods of characterizing complex mixtures. Besides becoming

a premiere method for separating closely related chemical species, LC can be employed

for qualitative identification and quantitative determination of separated species.

A chromatogram provides only a single piece of qualitative information about each

specie in a sample- its retention time. Even if it is widely used as a tool for analyte

detection and determination, the confirmation of identity requires a spectral or chemical








investigation of the separated compounds. Quantitative chromatographic analysis is

based on a comparison of either the height or the area of the analyte peak with that of a

standard. Peak areas are independent of broadening effects and therefore are more

satisfactory analytical parameters than the peak heights. The highest precision in

quantitative chromatography is obtained by use of internal standards. A measured

quantity of an internal standard is introduced into each standard and sample, and the ratio

of analyte to internal standard peak areas serves as the analytical parameter. It is required

for the internal standard peak to be well resolved from the peaks of other compounds, but

still to appear close to the analyte peak.

The origins of combining LC with MS can be traced to the 1960s.89 The main

challenge of such a coupling is in introducing a liquid sample into the vacuum of a mass

spectrometer, but utility of such a combined analysis still drives research towards further

advancements of the methods. One of the main advantages that ESI offers over MALDI

is a possibility of a direct liquid introduction. It is relatively straightforward to directly

introduce samples into the mass spectrometer while in liquid solution via electrospray. It

is more difficult to couple MALDI directly to liquid separations because samples have to

be dried on a solid surface prior insertion into the mass spectrometer. Therefore, MALDI

can serve as a useful off-line detector, which offers the advantage of an independent mass

analysis.9094 However, several attempts to develop on-line coupling of liquid separations

have been made, but they required instrumental changes and redesign.9597

As MS is a concentration dependent technique, using the LC prior the MS analysis

is beneficial in many ways; besides the fractionation of complex mixtures and salt

removal, LC also preconcentrates every analyte in the effluent. Chromatographic








separation of biomolecules can be based on different properties of the molecules, their

size, charge, hydrophobicity, or binding affinity.

Reverse Phase LC

Two types of partition chromatography are distinguishable based upon the relative

polarities of the mobile and stationary phase. Early work in liquid chromatography was

based on highly polar stationary phases such as water supported on silica or alumina

particles, while a relatively nonpolar solvent, such as hexane served as the mobile phase.

For historic reasons, this type of chromatography is now referred to as normal-phase

chromatography. In reversed-phase chromatography, on the other hand, the stationary

phase is nonpolar, often a hydrocarbon, and the mobile phase is relatively polar, such as

water, methanol or acetonitrile. Therefore, the most polar component appears first, and

increasing the mobile phase polarity increases the elution time. The most commonly

used packing for this type of chromatography contains a C8 (n-octyl) or a C18 (n-

octadecyl) chain attached to the siloxane coatings. The elution is usually carried out with

a highly polar mobile phase such as an aqueous solution containing various

concentrations of solvents such as methanol or acetonitrile.

Capillary LC

The ability to work with minimal sample sizes, at small flow rates, and the

enhanced detection performance with the use of concentration sensitive detectors due to

preconcentration are some of the main reasons for increased popularity and development

of capillary LC. Miniaturization of LC separations started in the 1960s with Horvath's

work,98'99 but became significantly more exploited as new mass spectrometric techniques,

especially ESI, were developed in 1980s, leading to low detection limits as well as

specificity in characterization.








The reduction of column i.d. to the capillary size (down to 20 gm i.d.) introduced

new requirements for the instrumentation applied, but offered the advantage of MS

compatible flow rates, so that the LC effluent could be directly introduced into a mass

spectrometer. It was also shown that inhomogeneities in mobile phase flow paths are

reduced due to the more uniform packing structures within smaller i.d. packed columns,

therefore reducing peak dispersion.100 Additionally, this results in more uniform

retention times and reduction in column band-broadening.

The UV detector is still commonly used for capillary LC.101-104 Fluorescence 105,106

and electrochemical detection107-109 are also employed. Nevertheless, mass spectrometry,

with all its advantages, is the most common for detection and characterization of LC

effluent. Mass spectrometric methods are commonly coupled on-line (ESI-MS) or off-

line (MALDI-MS), offering great sensitivity, minimal sample demand, and full

characterization of the effluent.

Neuroscience

All the sensations, feelings, thoughts, motor and emotional responses, learning and

memory, the actions of psychoactive drugs, the causes of mental disorders, and any other

function or dysfunction of a brain cannot be understood without the knowledge about the

processes occurring during communication between neurons. As far as we understand

the process now, the neurotransmitters are the messengers that are released in the

synaptic cleft and induce the signal to travel to the next nerve cell (Figure 1-8). Different

types of cells secrete different neurotransmitters. There are many classes of molecules

that can play that role, usually grouped into cholines, biogenic amines, amino acids, or

neuropeptides.








During the last decade, it has become evident that mass spectrometry is a key

analytical tool in neuropeptide research. For decades before that, less specific methods,

based on immunologic or structure specific assays, were used for neuropeptide

characterization and quantification. Some of the methods commonly used in

neuroscience are radioimmunoassay (RIA), radioreceptor assay (RRA), enzyme-linked

immunosorbent assay (ELISA), chromatographic, and electrophoretic separation with UV

detection.

The RIA method is based on the specific interaction between a given peptide (the

antigen) and its antibody (Figure 1-9). Although it offers good sensitivity and ease of

operation, cross-reaction with similar structures and fragments introduces a significant

error in the measurements. Some improvements in accuracy and precision due to better

selectivity are noticed when RIA is combined with a separation technique.110-112 hen

using RRA, the neuropeptide activity is detected and a suitable membrane preparation

containing the particular receptor is required. This technique was used to quantify the

opioid peptide activity in CSF from patients with chronic pain or neurologic

disorders.1133114 The error when using this method occurs since there are many different

opioid peptides that contribute to the detected activity, so it is unclear what particular

opioid system is responsible for the observed changes.

Among a variety of MS-based strategies available for peptide analysis, the most

commonly used techniques are MALDI-TOFMS and ESI-QITMS. These methods,

combined with separation techniques, offer high specificity and great sensitivity, even

when analyzing the biological samples directly."1517 However the search for new








endogenous peptides and their quantification is still a great challenge of mass

spectrometry in neuroscience.

Neuropeptides

Neuropeptides are a large family of bioactive compounds participating in the

transmission or modulation of signals between the nerve cells. They are involved in

many neurologic functions, like food intake, memory, mood, pain, reward, stress, and

many others. They are present and can be sampled from central nervous system (CNS)

and body fluids. Cerebrospinal fluid (CSF), because of its accessibility, offers a suitable

medium for neuropeptide analysis.

Synthesis of neuropeptide precursors occurs in ribosomes, at a considerable

distance from the secretary site of the axon terminal. Generally, the endogenous

neuropeptides are synthesized as large, biologically inert protein precursors

(prepropeptides). These precursors are subsequently cleaved throughout a series of

proteolytic steps, in a very sequence-specific and tissue-specific order, to generate the

active species. Other post-translational modifications, like amidation or phosphorylation,

have also been studied (Figure 1-10).

Determination of peptide concentration gives no information with regard to

turnover. Specific regional concentrations of peptides may be greater than those

described for the general area of central nervous system. This can significantly change

upon stimulation of certain brain regions. Alterations in concentrations of brain peptides

are being studied for their use as possible markers in a number of neurological diseases.

Sampling Techniques

One of the main limiting factors for reliable brain research is proper sampling

needed for valid data interpretation. Advances in various in vitro and in vivo sampling








techniques have considerably improved research into nervous system processes. Recent

improvements in chemical separations and mass spectrometry offer reliable chemical

analysis for brain research. Many assumptions and compromises are made when

processes in the living organisms are investigated, which can lead to misinterpretation of

the analytical results.

In vitro studies

Many in vitro models have been used in the biological sciences for preliminary

investigations, method development and various advanced methods in medicine,

including the most disputed methods of cloning living organisms. The benefits of such

approaches are enormous, since they minimize the use of living organisms for testing,

help eliminate some methods in the early stages of development, and give a broad

perspective for adventuresome and risky innovations. However, the results obtained by

in vitro and in vivo techniques can differ considerably, which hinders the extrapolation of

the in vivo results from the in vitro model. The major causes of these discrepancies occur

due to the properties of various in vitro and in vivo models used.

In vitro models can be designed according to experimental requirements. There are

many different concepts of in vitro brain analysis. Certain parts of the brain can be

studied separately, using the brain slices, primary cultures, or isolated brain microvessels.

Cell cultures can be generated and engineered according to experimental requirements.

One of the generally used approaches in brain analysis is to homogenize the brain tissue

in a buffer solution.118-120 This method offers fast and easy screening results, but can be

deceiving due to contents of the cell lysate that will be present, with many enzymes

which may disturb the environment. An alternative approach has been developed which

requires careful slicing of the brain into thin layers followed by washing with buffers to








remove any cell lysate contents.121 It is assumed that minimal damage to the cell is done,

and that no intracellular contents are present in the sample.

In vivo studies- microdialysis

Microdialysis is a sampling technique generally used for in vivo sampling, based on

a size-selective diffusion of the analyte through a semipermeable membrane. Molecules

smaller than the pore size may permeate into the perfusion medium following the

concentration gradient, while molecules larger than the membrane cut-off (like proteins,

enzymes), are excluded (Figure 1-11). It was initially developed by Delgado and co-

workers in 1972,122 but popularized by Ungerstedt and Pycock.123 The idea behind

microdialysis was to mimic the function of blood vessels and to achieve in situ sampling

and sample clean-up. It is performed by inserting a probe that contains a semipermeable

membrane into a medium of interest (Figure 1-12). The membrane, permeable to water

and small molecules, is continuously flushed by the physiological perfusion fluid, which

is collected as dialysate, and enriched with endogenous substances. Microdialysis may

also be used for local drug administration via the same probe at the same time as for

sampling. The versatility of microdialysis sampling has resulted in its application in

neurochemistry, pharmacokinetic studies, and biotechnology, mainly in vivo.

In vivo monitoring is ideally preferred over other sampling methods, since it offers

real life samples, as close as we can get with minimal disturbance of the regular

environment of an organism. Microdialysis provides direct measurements in

anaesthetized or freely moving animals, analysis of specific target tissues or metabolites

from enzymes in tissues, and continuous sampling with no liquid loss. It offers the

possibility of local drug administration and determination of drug influence on

metabolites. However, the dialysate is relatively dilute and contains high salt





21


concentrations, requiring further sample treatment prior the analysis. It is also an

invasive procedure, and it causes (although minimal) neuronal death. The method offers

limited spatial and temporal resolution, but is still the most commonly used due to the

ease of application and a routine experimental design.







Table 1-1. A table of the most used MALDI matrices. It can be noticed that all the
molecules are aromatic and have at least one acidic H. The aromatic or
conjugated structures are UV chromophores responsible for absorbing most of
the UV irradiance, while the H+ plays an important role in the analyte
ionization mechanism.

Matrix (abbreviation) Molecular Structure


Nicotinic Acid (NA) COOH
N
H1CO
Sinapinic Acid (SA) HO CH=CH-COOH
(3,5-Dimethoxy-4-hydroxycinnamic acid)
~~_______________H_______3CO
HO
DHBA COOH
(2,5-Dihydroxybenzoic acid)
OH
-N
HPA COOH
(3-Hydroxypicolinic acid)
OH

ca-CHCA HO
(a-cyano-4-hydroxycinnamic acid) COOH
NC









Toward the TOF Analyzer
t


Laser Beam


.a S


Electrostatic
Field


- I


Matrix with
the Analyte


Figure 1-1. A schematic presentation of MALDI process. The laser initiates an explosion
of the matrix/analyte plume, resulting in ionized analyte directed toward the
analyzer by the electrostatic field. (adapted from reference 124)


a


:-, t:















M**+ M

-e +e-

Mi' M-'
4M

+H' 4(M-H)' -H2


i i '(M+H) H '
M+ (M+H)+ (M-H) I (M-H)- (M-2H)-' M-'


(M+2H)+*
Figure 1-2. A summary of the initial ionization processes in MALDI. The primary
ionization results in activated matrix molecules (labeled M), which further
initiate analyte ionization. The process is usually observed only in positive
mode, but both mechanisms are expected to occur. (adapted from reference
13)








' 1 L, I .


II 1111111
I I
9.. 4


Detector


Ion Source


Figure 1-3. A schematic presentation of the reflectron TOF. The reflectron allows ions
with greater kinetic energies to penetrate deeper into the reflectron field than
ions with smaller kinetic energies. If a packet of ions of a certain m/z ratio
contains ions with varying kinetic energies, the reflectron will decrease the
spread in ion flight times, and therefore improve the resolution of the TOF
analyzer. (adapted from reference 53)


YJiIIi I






Static, Reflector
2600

K-1


, 1300 12


Delayed, Linear
R:4000

_JL_~h


R:4400

7iLl


Dela)
R:4600


R:5400


~-- ----R- -- I __ _I_ _


R:6600


S1300
nvz


R:8600


1290


ed, Reflector


IRPj


L~


1300


Figure 1-4. Comparison of signal in different types of TOF analyzers. Nine mass spectra
of angiotensin I are shown, using linear, reflectron and delayed extraction
modes of operation. Instrument geometries are labeled: RP, EL, and XL,
where flight distance increases in that order. (adapted from reference 125)


R:


R:320(


.a ~


R:5800


Ei


1290


90


JL.-


- __-I11-


U--MAY-


i t- ---- -








Solvent
Evaporation
hw


Coulomb
Explosion


Further Solvent Evaporation
and Coulomb Explosions Steps

r 1


Charge
Residue
Model

b)


Surface electric field
"lifts" analyte ion
from droplet


Solvent Evaporation
from a droplet
containing a single
analyte molecule

I


Figure 1-5. Schematic presentation of the ESI mechanism. Upon the initial droplet
formation, different models are suggested to explain the final ion formation: a)
ion desorption model (IDM), suggesting the ion "escape" due to the electric
fields present, and b) charge residue model (CRM), indicating the continuous
droplet evaporation as the ion is formed.


0


0


0


Ion Desorption
Model


a)








Ions


End-cap electrode


Ring


Resonance
AC voltage


RF


Detector


Figure 1-6. Schematic presentation of a quadrupole ion trap. The ions are selected and
ejected according to their m/z value by scanning the voltage on the electrodes.
(adapted from reference 126)










x3 y3 z3 x2 y2 z2 X, y, z, H
r- r- r- r- r r- r- r- r-
R, 0 ,1 R21 0 1 1 R, 0 1 1 R4

H2N-C C-N-C--C-N--C--C-N--C-COOH

HIH H HH HH

a, bI c, a2 b2 c2 a3 b, c,

Figure 1-7. Peptide ion fragmentation upon Collision Induced Dissociation (CID)
experiments. For low energy collisions, like those in a QIT fragmentation, the
peptide bond is cleaved and y and b type are predominantly formed ions.



































Figure 1-8. A presentation of a synaptic cleft. The neuropeptides are secreted as the
signal is given to the vesicle. Upon secretion in the synaptic cleft, they bind
to the corresponding receptors on the postsynaptic neuron, while some of the
material is reabsorbed by the presynaptic neuron. (adapted from reference
127)









REAGENTS. POSITIVE SAMPLE NEGATIVE SAMPLE
Ab specific for hormone high level of hormone low level of hormone
(coating the filter) I
Unknown sample with hormone g

Allow time to react
Wash away unbound substances
REAGENTS: %\i
I-labeled hormone

Allow tim to react
Wash away unbound
radiolbeled hormone


PROCEDURE: measure radioactivity
in a gamma counter

RESULT: amount of radioactivity
is inversely proportional to the
concentration of hormone in the
sample. _________
Figure 1-9. A diagram of the radioimmunoassay quantification method. The method is
based on antigen-antibody recognition of analyte. The amount present in the
solution and absorbed will be inversely proportional to the amount measured,
which results in great sensitivity of the method. (adapted from reference 128)








Biosynthesis
Pre-propeptide
Degrading Enzymes Prohrmone
Prohormone

Propeptide
Prohormone convertases
Active neuropeptide

Converting Enzymes
Active fragments
Degrading Enzymes
Inactive fragments or amino acids -
Figure 1-10. The pathways ofneuropeptide synthesis. After a protein is synthesized in
the ribosomes, the propeptides are stored in the neurons until the enzymes
start converting them into the active neuropeptides. The degrading enzymes
are also involved, converting the peptides into inactive fragments and amino
acids.












Fluid is pumped
V through inner cannula

Fluid Is collected


Brain


Substances from
extracellular fluid
diffuse through the
dialysis tubing


Figure 1-11. A schematic presentation of microdialysis. Microdialysis probe allows
diffusion of small molecules into and out of the probe. Therefore, this
sampling technique can be used to observe consequences of the perfused
compound, or to sample the environment for characterization of the
endogenous substances. (adapted from reference 129)


Dental
plastic'\













/ /\ Ingput Perfusion Luid


SPerfuson ........


SFWd Oulput
(to measuring equipment)




Collected C
Samples Outer Tube (400 mun)



Inner Tube (O10 pn)



\ Sen-permeable
S Membrane

TUssue
Figure 1-12. A presentation of a microdialysis experiment and the microdialysis probe.
The microdialysis probe is placed into the brain region of interest. The
sampling is performed on an alive and usually awake animal. (adapted from
references 130, 131)












CHAPTER 2
INVESTIGATION OF SAMPLE PREPARATION METHODS TO IMPROVE
MALDI-TOFMS PERFORMANCE AND FACILITATE INTEGRATION OF
CAPILLARY LC SEPARATIONS

Introduction

MALDI Sample Preparation Considerations

Sample preparation has a significant influence on the sensitivity and reproducibility

of MALDI-MS.132'133 Therefore, various sample preparation protocols have been

developed to improve the homogeneity of matrix crystals and the distribution of sample

within a deposited spot. The influence of different matrices, solvents, pH, counter ions,

and salts present or added were investigated using such deposition techniques as dried

droplet, thin layer, spin-dry, the sandwich method, and fast evaporation.'33 Selecting a

matrix for these techniques is considered to be somewhat of an empirical process;

however, guidelines for selecting matrices for various classes of molecules are available

(for instance, 4-hydroxy-a-cyanocinnamic acid (a-CHCA) is most commonly used for

peptide analysis).117,132-136

Solvent selection is dependent on both matrix solubility and the sample deposition

method used; variations in crystal formation are known to influence ionization,

distribution, and the incorporation of analyte into the matrix layer. Sadeghi, et al., have

demonstrated that ionization efficiency is directly dependent on crystal size due to the

predominant volatilization of smaller crystals by the laser beam.'37 Horneffer, et al,.138

have observed that incorporation (co-crystallization) of analyte molecules is not

necessary for all matrices; however, the efficiency of the MALDI process is improved








with a more homogeneous distribution of matrix crystals.'39 A detailed mapping of

conventionally deposited spots has revealed significant heterogeneity in analyte

distribution thought to be caused by uneven crystallization within the spot.140 Since spot

homogeneity and ionization efficiency are improved with the formation of smaller

crystals,'37 we investigated methods for reducing crystal size to further improve the

MALDI-MS technique as presented herein, which in summary, focused on two basic

goals:

minimizing crystal size for improved homogeneity of crystal distribution, and

minimizing sample spot size for improved detection limits.

Matrix crystal size

Vorm, et al.,141 proposed the fast evaporation method to reduce crystal size. Matrix

dissolved in a fast drying solvent forms a primary layer of small uniform crystals upon

which the analyte solution is deposited and incorporated into the matrix layer. An

additional improvement in sample distribution homogeneity was observed when analyte

was mixed with matrix and deposited as a second layer.142 Electrospray deposition

(ESD) of the matrix can be used to further reduce solvent evaporation time, producing

even and small crystals.143,144 A two layer approach for matrix and subsequent analyte

deposition can also be employed as it is hypothesized that small crystals formed in a first

layer seed the formation of small crystals in all successive layers.141'143

While electrospray is a well-established ionization technique for generating charge

on liquids and enabling their introduction into mass spectrometers, the same experimental

concept has been successfully employed for the deposition of fine, uniform layers of

particles as early as the 1950s.145'146 Small droplets are ejected from the Taylor cone

depositing particulates evenly onto the counter electrode due to the fast evaporation of








solvent (see Figure 2-1). Using electrospray deposition for MALDI-MS analysis has

shown improvement in spot-to-spot and shot-to-shot reproducibility due to the evenly

sized small crystals formed.143'144 In addition, the electrospray deposition method has

been successfully utilized for fabrication of protein films while preserving protein

reactivity when re-dissolved.147-150 This approach for storing biomolecule activity is

beneficial in conserving sample concentration, activity, or information.

Sample spot size

Minimizing sample spot size, thereby increasing analyte surface density, introduces

an improvement in detection limits. The laser conventionally irradiates less than 1% of

the sample spot size making a search for a "sweet spot" critical to successful MALDI.

Therefore, it is expected that concentrating the analyte into an area the size of the laser

spot should improve sensitivity.

Changing ESD parameters can effectively minimize spot size. Capillary i.d.,

electrospray voltage, emitter to plate distance, and flow rate can all be adjusted to achieve

smaller spots, and have been investigated in this work. ESD is also a non-contact

deposition method offering the possibility of depositing the matrix layer first

independently of the analyte or a matrix/analyte mixture second.

A competing approach to sample spot size reduction, in order to focus the analyte

into an area comparable to the laser spot size (-150 im in diameter), was the

development of miniaturized sample holders, termed AnchorChipTM plates by developers

at Bruker Daltonics.134'135 This can be accomplished with the use of a hydrophilic spot

area defined by a hydrophobic surface. When AnchorChipT plates are used,

crystallization of aqueous samples occurs in this hydrophilic spot, concentrating the








material into a 200 |pm diameter region.134'135 The "sweet spot" search is eliminated, with

all the material focused into the laser ablation area. Problems with shot-to-shot

reproducibility are diminished and sensitivity is increased given irradiation of the total

amount deposited.

Capillary LC/MALDI Coupling

Reverse-phase capillary liquid chromatography (LC) is commonly used for

separation of peptide mixtures and for cleanup of salts and other contaminants that

suppress the ionization process. Additionally, sensitivity and selectivity of mass

spectrometric analysis are expected to increase due to pre-concentration and separation of

the analytes when using capillary LC. Interfacing LC with MALDI-TOFMS requires

transporting analyte from liquid effluent into dry matrix crystals, which can be developed

for either an on-line or off-line mode of interfacing separations with mass spectrometry.

On-line coupling

Coupling LC to MALDI-TOFMS on-line is particularly challenging because it

requires delivery of the LC effluent directly into the vacuum of a mass spectrometer. On-

line coupling can be performed by continuous flow,95'96 or aerosol introduction into the

mass spectrometer.97 The LC effluent is directly introduced into the MALDI mass

spectrometer either on a moving belt or by spraying into the vacuum as aerosol. Both of

these approaches require major instrumental adjustments, which is probably the main

reason why they were never extensively used.

Off-line coupling

Off-line coupling involves the collection of fractions into vials, or directly onto a

MALDI plate-described by Caprioli's group as "on-line transfer, off-line analysis".90

This requires matrix introduction by either depositing the matrix prior to sample








deposition,9-94 or by introducing the matrix directly into the effluent either through a tee

union,151'152 a sheath flow,'53 or by adding it directly to the mobile phase.'54 Caprioli and

coworkers have automated dried droplet deposition, commonly used when analyzing

sample fractions, using a motorized XYZ stage that rosters sample spots across a MALDI

plate.9 For this technique, spots can be deposited by bringing the plate into contact with

a sample needle; however, to avoid such direct contact, nebulizing through a

microdispenser 91,92,94 or electrospraying the sample has been suggested.152'55 Detection

limits using these techniques range from low femtomoles to high attomoles.90'153'15

Electrospray deposition has also proven to be a successful way to deposit material for

further MALDI-TOFMS analysis, and since it can continuously deposit liquid onto a

solid support, it was expected to be a good intermediary for coupling liquid

chromatography with MALDI mass spectrometry. The flow rates and equipment used

for such methods are compatible with capillary LC; however, the change in solvent

composition during a gradient LC run can present a potential problem for the stability of

the Taylor cone of the electrospray. This can be overcome by ramping the electrospray

voltage during the run.156,157

Imaging Techniques

A few imaging techniques were employed in this research to investigate matrix

crystal and analyte spot size. Scanning electron microscopy (SEM) and transmission

electron microscopy (TEM) were used for measuring changes in crystal size and crystal

distribution as a function of solvent and sample deposition method. Fluorescence

imaging was used to observe distribution and incorporation of analyte into the matrix

layer. A fluorescent dye (rhodamine 6G) was added to the analyte solution to enable the

visualization of the spots.








Scanning electron microscopy (SEM)

The Scanning Electron Microscope (SEM) uses electrons rather than light to form

an image. The SEM is designed for the direct study of solid objects' surfaces. By

scanning with an electron beam generated and focused by electromagnetic lenses,

secondary electrons are generated from the sample's surface. The image is produced

from counting the number of electrons emitted from each spot on the sample. As a result,

SEMs have a greater depth of focus than optical microscopes and subsequently very good

resolution; however, samples have to be conductive. If not, they must be coated with a

thin layer of gold by a sputter coater prior to analysis.

Transmission electron microscopy (TEM)

In transmission electron microscopy (TEM), a thin solid specimen is bombarded in

vacuum with a highly focused, monoenergetic beam of electrons having sufficient energy

to propagate through the specimen, be collected and magnified to produce an electric

signal. Diffracted electrons are observed in the form of a diffraction pattern beneath the

specimen. Transmitted electrons form images from small regions of the sample that

contain contrast, due to several scattering mechanisms associated with electrons and

atomic constituents of the sample. High resolution images of the morphology or

topography of a specimen can be achieved by scanning an electron beam across a

specimen. Due to the shorter wavelength of electron beams, 0.2 nm resolution can be

achieved, compared with 200 nm when a light microscope is used.

Fluorescence imaging

In order to differentiate objects with similar optical properties that are irresolvable

by optical microscopy, fluorescence properties can be used. A fluorescence microscope

is basically a conventional light microscope supplemented by an excitation light source








and an array of filters. Much higher intensity light is required to illuminate the sample,

which in turn emits light of a longer wavelength. The image from the microscope is

viewed by eye or can be digitally collected via a CCD camera.

Experimental Methods

Instrumentation

Mass spectra were collected with a Voyager DE-PRO MALDI-TOFMS (Applied

Biosystems, Framingham, MA) and a Bruker Reflex II MALDI-TOFMS (Bruker,

Bremen Germany), each equipped with a nitrogen laser (\=337 nm) and operated at a

pulse rate of 3 Hz. Analysis was performed in reflectron mode using delayed extraction

(PE, Applied Biosystems) or pulsed ion extraction (PIE, Bruker). Positively charged ions

were accelerated to 20 kV, and 100 to 300 single shot spectra were averaged. All the

other parameters varied depending on the experiment and the instruments used, and are

discussed in detail for each experiment.

An Agilent 1100 series binary pump system (Agilent Technologies, Palo Alto, CA)

was used for preliminary off-line LC/MALDI experiments. ISCO model 100DM syringe

pumps (ISCO Inc., Lincoln, NE) were used for all advanced LC/MALDI experiments. A

variable wavelength UV detector, model TSP 1000 (Thermo Separations Products, San

Jose, CA), was used throughout for chromatography method development.

Electron microscopy was performed in the Major Analytical Instrumentation

Center (MAIC) of the University of Florida. Scanning electron microscopy was

performed with a JEOL SEM JSM-6400 (Figure 2-2), while transmission electron

microscopy was performed on a JEOL TEM 200CX (Figure 2-3), (JEOL USA, Inc.,

Peabody, MA). Fluorescence microscopy was performed in the laboratory of Dr.

Weihong Tan using an inverted microscope OLYMPUS 1X70-S1F1 (Olympus America








Inc. Melville, NY), an argon laser INNOVA 3000 (Coherent Inc., Santa Clara, CA), and

a CCD camera (Princeton Instruments, Trenton, NJ).

Chemicals and Reagents

a-cyano-4-hydroxycinnamic acid (a-CHCA) and all synthetic neuropeptides

(oxytocin, bradykinin, Arg -vasopressin, and neurotensin 1-11) were obtained from

Sigma (St. Louis, MO). Solutions were prepared with HPLC grade solvents (Burdick and

Jackson, Muskegon, MI). Mobile phases were filtered with 20-nm-pore size aluminum

oxide filters (Fisherbrand, Fisher Scientific, Pittsburgh, PA) to prevent capillary clogging.

Trifluoroacetic acid (TFA), acetic acid and hydrofluoric acid were purchased from Fisher

Scientific (Fair Lawn, NJ), while helium gas tanks were provided by PraxAir Inc.,

(Danbury, CT).

Electrospray Deposition Method (ESD)

ESD apparatus

Two different electrospray arrangements were used throughout this work. The first

used a pressure bomb for flowing liquid in all preliminary investigations with the

application of a high voltage to a gold wire placed at the capillary/emitter junction. A

MALDI plate was placed horizontally opposed to the ESD emitter, and was grounded to

complete the electrical circuit (as shown in Figure 2-4). An advanced electrospray device

was constructed based on preliminary results employing a syringe pump, fused silica

capillary, and a pulled capillary electrospray emitter (see Figure 2-5). The electrospray

voltage was applied directly to the syringe needle with the MALDI plate grounded. A

stepper motor was added to automate plate translation in one direction, and a camera was

placed behind the electrospray emitter to enable visualization of the electrospray process,

which helped maintain electrospray stability. Matrix and analyte solutions were








electrosprayed through a capillary emitter. The flow rate, applied voltage, and

electrospray tip to plate distance were optimized for achieving the smallest matrix crystal

size and an optimal analyte spot size.

Electrospray tip production

A flat-cut capillary tip was used as an electrospray emitter in preliminary

experiments. Improved electrospray emitters were introduced in the advanced setup,

made by pulling fused silica capillaries to a fine point using a CO2 laser puller (Sutter,

Narato, CA) followed by etching in 50% HF for 3-5 min. The capillaries were 50 or 25

p.m i.d. and 360 p.m o.d, where the former ones were used for electrospraying matrix

solution and the later once for peptide solution electrospray. The methodology of

capillary pulling is illustrated in Figure 2-6, with an SEM image of the capillary emitter

shown in Figure 2-7.

Hydrophobic Sample Supports (AnchorChip Plates)

Introduced by Gobom, et al.,132 AnchorChipTM plates were investigated to decrease

analyte spot size. These commercially available stainless steel plates were coated with

Teflon, leaving untreated spots of 200-600 ptm open as hydrophilic surfaces for sample

deposition. The plate with 200 pm diameter spots was used for all the experiments since

it was the most similar to the size of the laser spot and it concentrated the material the

most.

LC/MALDI

Capillary column packing

Capillary columns were prepared in the lab. Different lengths of 150 pm i.d./360

ipm o.d. capillary were packed with a C18 stationary phase. An acetone slurry (lOmg/ml)








of 5-jm Altima C18 reversed-phase particles (Alltech, Deerfield, IL) was pushed through

the column by 1000 psi pressure of helium from a pressure bomb. An in-line filter with a

sintered silica disk (Upchurch Scientific, Oak Harbor, WA) was used to prevent clogging

of the columns.

Off-line method system

For separation of a standard peptide mixture, a 2.5 cm section of fused-silica

capillary (150 pim i.d.) was packed with 5-jim C18 reversed-phase particles. Samples

were injected with a 5 jil sample loop (Valco, VICI Valco Instruments Co. Inc., Houston,

TX), as illustrated in the schematic shown in figure 2-8. The linear gradient was set from

95% A/5% B to 5% A/95% B in 8 minutes), at the flow rate of 8 pl/min (eluent A

consisted of 1% CH3COOH in H20, eluent B consisted of 1% CH3COOH in MeOH. The

UV absorbance chromatograms were collected using a LabView program (National

Instruments). Sixteen one microliter effluent fractions were manually collected over the

entire separation time from the capillary outlet into Eppendorf tubes already containing 2

jiL of matrix solution. Tubes were kept in an ice bath to prevent evaporation of the

solutions.

On-line collection, off-line analysis system

The schematic of this experiment is shown in figure 2-9. A 6 cm long capillary

column was prepared as described above. Identical solutions and mobile phases were

used as in the off-line separations, but effluent was directly deposited onto a MALDI

target plate that was coated with the matrix layer prior to analyte deposition.








Results

Minimizing Matrix Crystal Size

Optimization of ESD parameters

Solvent selection had to meet requirements for matrix solubilization and

electrospray stability. Mixtures of methanol and water with an acid modifier are usually

successful in most electrospray experiments, but a-CHCA was insoluble in these

mixtures. Instead, the matrix was soluble in acetone/0.1 %TFA, which was amenable to

electrospray requirements, but the high volatility of acetone was a problem since the

matrix solution would evaporate faster than it could be electrosprayed. This caused a

buildup of crystallized matrix on the capillary tip, subsequently clogging the electrospray

emitter. Therefore, instead of using a flat cut silica capillary (50 /m i.d., 360 pm o.d.), a

new pulled capillary emitter etched with 50% HF to open a small orifice was used to

alleviate this problem. The new emitter was beneficial in two ways: 1) less liquid was

sprayed into the air since the orifice was smaller than the capillary i.d. (usually -20 14m),

and 2) the excess solution (which was still present) was being pulled backwards, so it

wouldn't clog the emitter as fast. Even after this improvement, the emitter still had to be

washed with MeOH every 15 min to avoid clogging.

In preliminary experiments, a pressure bomb with helium gas was used to control

flow through the capillary. It was later shown that a syringe pump was easier to use and

provided more reliable control of the flow rate. An optimized flow rate was found as a

compromise between avoiding clogging and providing sufficient liquid for a stabile

electrospray. The most reliable electrospray operation was seen when the following

parameters were used: a saturated matrix (a-CHCA) solution in acetone/0.1% TFA (10

mg/ml) electrosprayed through a 50 gm i.d. capillary with a 20 lim i.d. emitter at +3 kV








and a flow rate of 500 nl/min. The distance from the emitter tip to the plate was set at 1

mm, producing matrix spots of- 5 mm in diameter. An optimized matrix layer thickness

was achieved by electrospraying each spot for 5 min.

Electrospray disturbances were observed by use of a camera and adjustments were

made to maintain a stable electrospray. A stepper motor facilitated plate translation in

one direction, enabling organized deposition of spots and the option to deposit tracks of

matrix and analyte. This principle would enable automated fraction collection of the LC

effluent in LC/MS experiments.

Matrix crystal images

Crystal structure and distribution were characterized by scanning electron

microscopy (SEM). The differences between electrosprayed matrix and matrix deposited

by conventional techniques are evident in Figure 2-10. A "network" of small matrix

crystals deposited by electrospray showed increased homogeneity when compared to the

same matrix deposited by dried droplet method. The size of the electrosprayed crystals

was confirmed by transmission electron microscopy (TEM), (see Figure 2-11) and

determined to be in 100 nm in diameter.

Minimizing Analyte Spot Size

Optimization of ESD parameters

It was found that 50% MeOH/0.1% TFA in H20 was an optimal solvent mixture

for dissolving and electrospraying peptides. Peptide solutions were electrosprayed at a

flow rate of 100 nl/min through a 25 pm i.d. capillary with a -5 pm i.d. pulled emitter

held at 1.2-1.3 kV. Deposition was made directly onto a matrix layer in 10, 30, and 60-

second intervals at a distance of 0.25 mm from the target. The spots were analyzed by








imaging techniques to determine their size and the distribution of the peptide solution

within the deposited spot.

Analyte spot imaging

The conditions used produced a spot of -200 1m in diameter (see Figure 2-12).

Even smaller spots could be made (Figure 2-13), but these were difficult to visualize

when acquiring data with the MALDI-TOF instrumentation (spots -100 nm in diameter

were not observable). On the other hand, results showed that under higher flow rates or

longer deposition times the electrosprayed solution would start redissolving/relocating

the matrix, which then migrated toward the edge of the formed "crater", thus

redistributing the analyte (methanol present in the electrospray solution redissolved the

matrix). This can be a limitation for analyzing larger volumes and on-the-spot

concentration of the sample. Matrix removal was observed in SEM imaging experiments

(Figure 2-14) and confirmed by fluorescence imaging after fluorescence dye (rhodamine

6G, soluble in MeOH) was added to the peptide solution in the pM concentration range to

enable visualization of the electrospray pattern by wide-field fluorescence microscopy

(see Figure 2-15). To overcome matrix removal by electrospraying, an analyte solution

on top of the matrix layer, these approaches were investigated: 1) the peptide mixture was

premixed with the matrix solution (1:1 v/v) and electrosprayed together; 2) the peptide

mixture was premixed with the matrix solution and electrosprayed on top of a previously

electrosprayed matrix layer; and 3) the peptide mixture alone was electrosprayed on top

of an electrosprayed matrix layer. It was shown that the least noise through interference

peaks and thus the highest S/N was present when electrospraying the peptide solution

alone on top of a previously electrosprayed matrix layer. Low attomole detection limits








were achieved by depositing 15-90 nl of the analyte at a concentration of 0.7 nM (see

Figure 2-16).

AnchorChipTM plates

Solution preparation was altered due to concentrating effects of all the material

deposited into a small spot. The optimal matrix solution (0.1 mg/ml in 50% MeCN/50%

H20/ 0.1% TFA) was diluted 1:100 compared to conventional sample preparation. The

peptide solution was prepared in 50% MeOH/50% H20/0.1% TFA and mixed with the

matrix solution in a 1:1 v/v ratio. A one microliter aliquot of this mixture was deposited

onto a 200-jm diameter AnchorChipTM spot using a gel-loader pipette tip. Spot size and

matrix crystal distribution were determined by SEM (Figure 2-17). In comparing

deposition of analyte alone with the deposition of an analyte/matrix mixture onto an

already deposited matrix spot, no significant difference was observed in both sensitivity

and reproducibility (Figure 2-18). Therefore, dried droplet deposition can be alternated

with the two-layer sample deposition, according to experimental demands. Detection of

five attomoles was achieved by depositing 0.5 iL of a 10 pM solution (see Figure 2-19).

Capillary LC/MALDI-TOFMS

Preliminary investigations

To evaluate the integration of MALDI-TOFMS with capillary LC, the influence of

different MeOH/H20 effluent ratios on the mass spectrometric response was studied.

The result showed some difference in the detection of peptides within the range of 25-

75% MeOH (see Figure 2-20). The variation in signal intensities can be attributed to

crystallization differences under different solvent conditions. Also, increased

irreproducibility when approaching detection limits can be expected. The important point

proven by this experiment is that all the peptides can be detected throughout a wide








difference of solvent composition, since we were not certain at what point in the gradient

each peptide was going to elute.

Preliminary LC/MALDI-TOFMS experiments were performed using the

experimental design shown in figure 2-8. Flow rate was adjusted to minimize band

broadening while maintaining the required peak width for fraction collection. The

optimal LC flow rate was determined to be -10 Al/min with an optimal gradient ramp

over 8 minutes. A UV detector was attached to the column outlet and used with

concentrated standards to optimize separations conditions on-line. The limits of detection

for the UV detector were in the /LM concentration range. The wavelengths of 214 nm and

280 nm were investigated for peptide detection. The former one was selected for its

chromophores being more universal (absorbs the peptide bond). After confirming

chromatographic separation of the standard mixture, experiments interfacing LC with

MALDI-MS off-line were performed. Two 11l aliquots of each fraction were deposited

onto the anchor chip plate (200 /m diameter spots). Mass spectrometric data collected

from each of the two spots were averaged for each fraction. MALDI-TOFMS data

confirmed separation of the neuropeptides (see Figure 2-21). The results indicate that

proper interfacing of capillary LC/MALDI-TOFMS has been achieved.

On-line collection, off-line analysis data

The interface of capillary LC and MALDI-TOFMS was made more efficient by

direct deposition of the LC effluent onto a MALDI plate that had been previously pre-

coated with a matrix layer (see Figures 2-9 and 2-22). Chromatographic separation was

improved by optimizing gradient settings, flow rate and using a 6 cm long packed

capillary column. This resulted in 30 s peak widths, detected when a 1 /M peptide

solution was tested (Figure 2-23). At the optimal flow rate of 6 Vl/min, 1 #il fractions








were deposited every 10 seconds from the capillary onto the matrix spots. The resolution

of a 100 nM solution of the four peptides by the capillary LC and analysis of each

fraction by MALDI-TOFMS is shown (Figure 2-24). While the UV detector couldn't

trace the peptides, the MALDI-TOFMS revealed the distribution of the

chromatographically resolved peptides.

Conclusions and Future Directions

Both electrospray deposition and use of AnchorChipTM plates resulted in increased

sensitivity of MALDI-TOFMS due to analyte material being focused into the laser

ablation area. Attomolar amount of peptides from picomolar solutions can be routinely

detected. Also, improved homogeneity of matrix crystal size and distribution was

achieved by use of a developed electrospray deposition method, due to fast solvent

(acetone) evaporation and even distribution of droplets over the targeted surface.

Optimal electrospray conditions were established for finest matrix crystal distribution,

required thickness of the electrosprayed matrix layer, and the most advantageous analyte

spot size. Minimizing sample spot size resulted in improved sensitivity with detection of

attomoles ofpeptide.

The possibility of coupling MALDI-TOFMS with capillary LC was demonstrated

by "on-line transfer and off-line analysis" of a standard peptide mixture that was

successfully separated and detected by MALDI-TOFMS. The UV detector used was not

sensitive enough to monitor the separation. It would therefore be useful to monitor the

separation with a fluorescence detector having greater sensitivity. Many possible

derivatization methods exist for attaching a fluorescencently active group to a peptide;

the method would simply require one of these as an extra step. One obstacle to this

approach might be analysis of the complex mass spectra of derivatized samples.





51


Also, since the method was being developed for analysis of neuropeptides sampled

from the brain, employment of the technique for that purpose would be the best criteria

for confirming its validity.













Reduction
G G)



90 00





G0
G


Electrons


Figure 2-1. Process of electrospray ionization of positively charged ions. The electric
field penetrates into the liquid, thus enriching the solution at the tip surface
with possible charge carriers forming a Taylor cone. Drops with excess
positive charge erupt from the cone and travel toward the counter-electrode.
Fast evaporation and fission of droplets produce free ion species. (adapted
from A.T. Blades, et. al, Anal. Chem, 1991, 63, 2110-6)


Oxidation































Figure 2-2. SEM-JEOL JSM 6400. This instrument was used for recording the scanning
electron microscopic images of matrix crystals and analyte spot size. The
instrument was available at the Major Analytical Instrumental Center (MIAC)
of UF.











































Figure 2-3. TEM-JEOL 200CX. This instrument was used for transmission electron
microscopic imaging of matrix crystals. The instrument was available at the
Major Analytical Instrumental Center (MIAC) of UF.










Pressure bomb

-


MALDI plate

Figure 2-4. Electrospray (ES) apparatus used for preliminary experiemnts. Solutions
were pumped through a capillary using a pressure bomb and helium pressure
tank. An ES emitter was placed above the MALDI plate that was grounded
and attached an XY stage.


Fused silica capillary
(50 prn i.d.)


MALDI p ate

X translation \


Syringe pump 3V
3 WV


Electrospray needle


YZ translation stage
Figure 2-5. The advanced electrospray deposition setup. A MALDI target plate was
mounted on an YZ translation stage while the electrospray emitter moved
back and forth orthogonal to the plate. High voltage was applied directly to a
syringe needle as the solution was pumped through a capillary to the
electrospray emitter.









Silica capillary


CO2 laser
Figure 2-6. Schematic presentation of a capillary puller. A CO2 laser was used to melt
fused silica, while the capillary was being pulled in opposite direction.
(adapted from reference 158)


Figure 2-7. SEM image of an electrospray emitter. After pulling and etching the capillary
tip in HF the emitter had an orifice smaller than the capillary i.d. and was
conical in shape. (taken from reference 159)








Sample


Capillary
LC column


Waste


plitter


LJ-Lf _J Ice Bath i-
Gradient
pump
Figure 2-8. Preliminary LC-MALDI set-up. Samples were injected through a 6-port
valve onto a 5 |t1 sample loop. A UV detector was placed post-column with
fractions collected into eppendorf tubes containing 2 gl of matrix solution.








Sample Loop


Capillary
LC column
Waste i--
Splitt6 Detector

AnchorChipTM
A Plate

Gradient
pump
Figure 2-9. The LC-MALDI set-up for on-line collection off-line analysis experiments.
After injecting the samples through a 6-port valve and detecting the analytes
with a UV detector, the fractions were directly deposited onto a MALDI plate,
having previously been coated with a matrix layer.

















































Figure 2-10. SEM images of matrix (a-CHCA) crystals. a) deposited by dried droplet
method, from a MeCN solution, allowing long crystallization time and crystal
aggregation, b) deposited by ES method from an acetone solution providing
for fast evaporation and small crystal formation.








SMatri,x (CHCA) crystal
-.0m


Figure 2-11. TEM image of electrosprayed matrix crystals. Carbon grid in the
background is a support for the electrosprayed crystals. Crystal size averaged
100 nm.


































e 2-12. SEM image of the analyte spot size. The analyte spot was observed as a
shaded region toward the image center and is approximately 200 pm in
diameter.

































Figure 2-13. SEM image of the analyte spot size. The analyte spot in the middle had a
diameter of less than 100 ptm.


















































Figure 2-14. SEM images of matrix removal by electrospraying analyte solution. a)
beginning of matrix removal after electrospraying the analyte solution for
30 s, b) total removal of the matrix layer after electrospraying the analyte
solution for 3 min.
























Figure 2-15. Fluorescence images of matrix removal by analyte solution deposition.
Fluorescent dye was added to the peptide solution, but some background
matrix fluorescence was also noticeable, a) after 10 s of analyte deposition,
the matrix layer is still intact, b) the relocation of the matrix layer by the
analyte solution is noticeable after 1 min of analyte deposition.











20,000 -




0
o



10,000


1060.5


1000


A L! I


M


1200


S14 6.76

L~I ji.jll",La'al jjJuLLL -I


1400


Mass (mlz)
Figure 2-16. MALDI-TOFMS mass spectrum of electrosprayed peptide solution.
Depositing 14.5 nl of 0.7 nM solution of bradykinin and neurotensin 1-11 by
electrospray led to detection of 10 amol of each peptide (S/N 70 and 15
respectively). This mass spectrum was taken using a DE-PRO MALDI-
TOFMS instrument. Ions were accelerated with 20 kV, with 92% grid voltage
and 0.005% guide wire voltage, and 100 ns delay time, while 100 spectra were
averaged.


IIYYLC---'--I IIIU-U


II


11~


I i' FTWH~~(I~na~c~~nl~pTO ~'T ..llNWIVi11nnill~llf'lH II




















































Figure 2-17. SEM images of matrix spot and matrix crystals deposited on an
AnchorChip plate. a) a thick matrix layer is deposited onto the 200 Am
diameter spot, b) small aggregates of matrix crystals, deposited from a diluted
MeCN solution.










Stwo-fyer method
40000 pmOrees

35000

30000

S25000




15000

10000

5000


Oxytocin Bradykinin Arg-8vasopressin Neurotensin 1-11
peptide
Figure 2-18. Relative signal comparison of two-layer deposition method and premixed
matrix/analyte solution. It was shown that no significant difference in signal
was observed intensities when 5 amol ofpeptides were deposited with either
of the investigated methods. Therefore, both methods can be used.







68









20000
1060.5




15000


C
1446.7
0

10000





5000 1084.4






1000 1100 1200 1300 1400 1500
Mass (mlz)
Figure 2-19. MALDI-TOF mass spectrum of peptide solution deposited on an
AnchorChip plate. 5 amol ofbradykinin, Args-vasopressin, and neurotensin
1-11 are deposited from in a 10 pM solution. This mass spectrum was
recorded using Bruker MALDI-TOFMS. It is a result of averaging 300 mass
spectra, recorded at 20 kV, and with 17.5 kV grid voltage applied.











25
*75% MeOH
*50% MeOH
20 25% MeOH

0
S15T
5---------4---------4--


5 --


0 4-


oxytocin


bradykinin


Arg-vasopressin
Arg-vasopressin


neurotensin 1-11


peptide

Figure 2-20. Comparison of signal intensity of MALDI-TOFMS mass spectra using
different solvent mixtures. The comparison was tested with a 10 pM solution
of the standard peptides oxytocin, bradykinin, Arg8-vasopressin, and
neurotensin 1-11.


I






70


35 oxytocincn
bradykinin

S1 I \" neurotensin 1-11

25

"I
0 20











6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 19 30
fraction number
Figure 2-21. Preliminary capillary LC/MALD separation. It is shown that the four








peptides are separated on the 6 cm long capillary LC column (150 pm i.d.
packed with 5 /im particles of C18 stationary phase). The MALDI-TOFMS
signal intensity of the four peptides in each faction were plotted.
5 10 f -1---- | ^--
I I




6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27"28 19 30
fraction number



Figure 2-21. Preliminary capillary LC/MALDI separation. It is shown that the four
peptides are separated on the 6 cm long capillary LC column (150 /m i.d.
packed with 5 pmi particles of C18 stationary phase). The MALDI-TOFMS
signal intensity of the four peptides in each fraction were plotted.



































igure 2-22. A picture of on-line collection of the capillary LC effluent onto the MALDI
AnchorChip plate. Matrix has been previously deposited. The final spot size
was 200 pm in diameter.









3
2.5
2-
8 1.5










time (s)

Figure 2-23. A UV chromatogram of the capillary LC separation. The four peaks are
resolved each peptides eluting for -30 s. This enabled collection of 2 to 3 12-
second fractions per elution peak for MALDI-TOFMS analysis.
> -0.5 A"- "''' A A'''
D IV VV .G 20

-1.5 -
-2 flLJ-I

time (s)

Figure 2-23. A UV chromatogram of the capillary LC separation. The four peaks are
resolved each peptides eluting for -30 s. This enabled collection of 2 to 3 12-
second fractions per elution peak for MALDI-TOFMS analysis.





























Oxytocin Arg'-vasopressin
Bradykinin Neurotensin 1-11
Figure 2-24. Capillary LC-MALDI-TOFMS data collected using on-line collection off-
line analysis. The mass spectra of all fractions were shown. Separation of the
four synthetic neuropeptides was confirmed.













CHAPTER 3
NEUROPEPTIDE FF METABOLISM STUDIES AND DEVELOPMENT OF
QUANTIFICATION METHODS BY MALDI-TOFMS AND ESI-QITMS

Introduction

Quantification Using Mass Spectrometry

Ion current intensity from mass spectrometric signals does not correlate precisely,

accurately, or directly with the amount of analyte present in a sample. Suppression of

ionization due to species of significantly higher abundance can skew quantification when

using either micro electrospray ionization or MALDI. Non-homogenous matrix

crystallization may introduce additional irreproducibility into quantitative analysis in

MALDI experiments. Choosing an approach that provides minimal bias with the highest

compatibility in terms of sample preparation and mass spectrometric analysis is crucial

for successful quantification.

Relative quantification

Relative quantification is based on comparing two states of a system (for instance,

control versus treated cell cultures). This does not quantify the exact amount of analyte

present, but does provide quantitative values related to the changes occurring as a

consequence of the observed process. This approach can be applied to observe the

influence of drugs, or to compare diseased to healthy cells.

Metabolic isotopic labeling. Metabolic labeling is based on the incorporation of

isotopic labels during cell metabolism (for example, during protein synthesis). Using

isotopically modified growth media enriched in '5N or 13C presents a simple, universal








method to metabolically incorporate an isotopic reagent.160-'62 Additionally, selected

stable isotope-incorporated amino acids can be used for metabolic labeling where instead

of isotopically labeling all amino acids, only selected ones are labeled.163-165 Several

isotopically-labeled amino acids have been used in this manner, such as (5, 5, 5-2H3)

leucine and (15N) methionine. In metabolic isotopic labeling, one sample set is grown in

a non-isotopic growth media, while the sample for comparison is grown in depleted

growth media that contains one or more isotopically labeled amino acid, and since the

isotopes are incorporated at an early stage experimental error is minimized. Despite the

advantages, this concept is limited to use with cell cultures, thereby excluding clinical

tissue samples.

Chemical labeling. Several chemical labeling approaches have been developed for

protein/peptide quantification when metabolic labeling cannot be performed. All of the

described methods are based on differential isotopic labeling of a standard and the

analytes. For accurate quantification it is essential that the labeling reaction be highly

specific and stoichiometric.

A light (60) or heavy (180) oxygen atom can be incorporated in the C-terminal

carboxylic group during a proteolytic digestion by way of the solvent used.166"168 It has

been shown, however, that proteolytic enzymes incorporate different amounts of isotope

during digestion, making accurate quantification difficult.169 In addition, the mass

difference with the incorporation of a single 180 is only 2 Da, which can be difficult to

resolve at higher m/z values. On the other hand, this small difference introduces minimal

isotopic effects on chromatographic retention times, which is beneficial to accurate








quantification by LC/MS. Moreover, oxygen incorporation can be applied universally to

all sample types, since it is done during sample processing, not protein synthesis.

Cysteine is a common amino acid of choice for label introduction. The free thiol

group is far more nucleophilic (meaning, reactive) than any other natural amino acid side-

chain group. Isotope-coded affinity tag (ICAT) reagents label the thiol group of cysteines

in a protein at the alkylation step of sample preparation.170 The labeled protein is then

enzymatically or chemically digested prior to MS analysis.171 ICAT reagents also

contain a biotin affinity tag to enrich cysteine-containing peptides via an avidin affinity

column. A mass difference of 8 Da is provided by incorporating eight deuterated

hydrogen (2H) atoms (Figure 3-1). To minimize chromatographic retention shifts, eight

13C can be used in place of deuterium atoms for labeling. Recently, an acid cleavable

linker has been incorporated to allow removal of the heavy biotin affinity tag before MS

analysis to minimize mass and interference during fragmentation.'72

Acid-labile isotope-coded extractants (ALICE) are a novel class of chemically

modified resins used as isotope-incorporating reagents that react with the thiol group on

cysteines. The reagents have an acid-labile linker (synthesized with either heavy or light

isotopes) that covalently attaches cysteine-containing peptides to a non-biological

polymer support. As Cys-containing peptides are covalently attached to the ALICE resin

nonspecifically bound species are removed by washing. Peptides are released from the

resin by mild acid-catalyzed elution, and can be directly analyzed by LC/MS.173

Other isotopic tag methods can be used to target different amino acids. Lysine

residues can be labeled with 2-methoxy-4,5-dihydro-lH-imidazole, which converts only

the e-nitrogen of lysine ('H4- or 2H4-forms of the reagent can be used).174 Such labeling








of basic residues can further complicate MS2 data interpretation.'75 Quantification of

phosphoproteins is another challenge; several methods have been recently reported.176'177

One is based on the p-elimination of phosphate from phosphoserine or phosphothreonine

to form dehydroamino-2-butyric acid or dehydroalanine, respectively. This is followed

by the addition of ethanethiol or 2Hs-ethanethiol to the a-P unsaturated amide. The

second method uses a phosphoprotein isotope-coded affinity tag (PhIAT) in which

phosphoserine and phosphothreonine residues are selected and quantified analogous to

cysteine labeling by ICAT.

There are several methods that label the N-terminus of a peptide or protein. The

key to successful derivatization is to provide selectivity in labeling only the N-terminal

amino group. Some of the reagents used are nicotinyl-N-hydroxysuccinimide,'78 acetic

anhydride,179 and S-methyl thioacetamidate.180 The C-terminus can also be labeled via

differential esterification.181 Both N- and C-terminal labeling will be discussed further in

Chapter 4.

Absolute quantification

The absolute measurement of protein/peptide concentration requires qualitative

knowledge of the species of interest, as well as their environment. This is critical for

method development due to crucial role played by internal standards, ionization methods,

and the instrumentation used for analysis. Quantification is based on the classic internal

standard addition and calibration curve. Several groups have developed reliable

quantitative methods by mass spectrometry,182,183 and all emphasize the choice of internal

standard as the key issue for successful quantification.








An appropriate internal standard will match the physical properties of size, charge,

hydrophobicity, and ionizability of the analyte(s), with the best internal standard being an

isotopically labeled version of the molecule of interest. Such an internal standard will

have identical extraction recovery, retention time, and ionization response in mass

spectrometry. Chemical analogs of analytes can also be used. Internal standards should

always be added at the same point to all samples, including controls, being added as early

as possible in the sample preparation process. The amount of internal standard should be

well above the limit of quantification, but not so high as to suppress ionization of the

analyte.

Quantification Employing MALDI-MS

The application of MALDI-MS in quantitative analysis of biomolecules was first

reported by Nelson, et al., in 1992,182'183 only a few years after introduction of the

ionization technique. They presented a standard curve for insulin quantification

exhibiting linearity over one order of magnitude with an error of 20%. It was noted that

the two main experimental parameters affecting the absolute intensity with MALDI were

laser irradiance and sample preparation. Matrix influences, signal suppression,

instrumental and data-acquisition variability all contribute to unreliable quantification by

MALDI-MS. Weinberger and Boernsen used different internal standard/analyte systems

and found that a second order polynomial best fit the data.'84 One major point in

MALDI-TOFMS is to avoid signal saturation during data acquisition.

Although several reports are found in the literature on quantitative analysis of

biopolymers by MALDI-MS, few reports address method development for quantification

of particular analytes in biological fluids.'83'185'188 Since biological matrices contain high

salt and often low analyte concentrations, a desalting/concentrating step is of great utility.








MALDI is considered more tolerant to salts and impurities than other ionization methods,

but the concentration of salt in biological fluids generally exceeds these tolerable limits.

Irreproducibility influences

Despite all the advantages MALDI offers in the ionization of biomolecules, one

major problem of the method is spot-to-spot and point-to-point irreproducibility

(discussed in Chapter 2). This especially inhibits reliable quantification, since significant

variability is introduced with every laser shot. It has been reported that improvements in

sample preparation can increase reproducibility and consequently quantification

reliability.183 Besides the methods for improving signal homogeneity discussed in

Chapter 2, there have been reports of a "hybrid" matrix made by mixing in a few unique

components for improving quantification.86 Also, it has been shown that reproducibility

is improved with smaller matrix crystals within a homogenous layer.87 Comparative

studies with different matrices have shown that the best reproducibility can be expected

when a-CHCA is used.86'88'189

Internal standard selection

The use of an internal standard has two major roles: 1) to account for losses

observed in biological sample pretreatment (this is why early introduction of the internal

standard is essential) and 2) to reduce the influence of MALDI ionization variability.

According to many researchers,86"89 an ideal internal standard: 1) must be completely

resolved from the analyte, 2) must be chemically stable during analysis, 3) should be

chemically similar to the analyte, 4) should be close to the analyte in mass and

concentration to avoid suppression errors, and 5) should not react with the analyte. Thus,

an ideal internal standard, with all these attributes, would be an isotopically labeled

analog.








Quantification Employing ESI-QITMS

The QITMS provides tandem MS functionalities; thus the combination of parent

and unique fragment ion is usually used to selectively monitor the analyte of interest.

This significantly improves analysis of complex mixtures, where the mass of the

compound is not selective enough.

When coupled to LC the mass spectrometer is typically set to scan a specific mass

range, which can be as narrow as a few m/z units in selected ion monitoring. LC/MS

data are represented by scanning the ion current in the individual mass scans and plotting

total current versus time. The most common modes of acquiring LC/MS data are: 1) full

scan acquisition providing a total ion current plot (TIC), which has minimal selectivity;

2) selective ion monitoring (SIM), where a single m/z value is monitored; and 3) selected

reaction monitoring (SRM), where a particular fragmentation reaction is followed.

Chromatographic separation requirements

Signal suppression in ESI due to presence of salts is a main reason for sample

purification prior to MS analysis. Electrospray ionization is, however, compatible with

common reversed-phase LC solvent mixtures. Flow rate requirements to combine the

two processes have recently been met with improvements in ESI devices and LC

miniaturization.116'90 LC separation of complex mixtures minimizes signal suppression

by separating compounds present in smaller concentrations from those of higher

concentrations. If separation of mixtures is not necessary, desalting of the samples can be

done by use of smaller reverse-phase columns, or capillary sample trap columns. These

columns are too short to separate mixtures, but will remove salts from samples.








Mass spectroscopic requirements

In selected reaction monitoring (SRM) a unique fragment ion is used to quantify an

analyte in the midst of a very complicated mixture. SRM plots are very simple, with only

a single MS2 peak, making this approach ideal for sensitive and specific quantification.

The SRM experiment is accomplished by specifying a parent mass for fragmentation and

monitoring a particular daughter ion. When optimizing MS conditions it is often better to

simulate the conditions that are appropriate to a particular assay. For example, if the

analysis is performed at a chromatographic flow rate of 400 Jl/min and the compound

elutes at 30%MeOH/69%H20/1%CH3COOH it is best to simulate these conditions for

optimization while the compound is being infused.

Neuropeptide FF

The cardioexcitatory peptide with the sequence FMRFamide was first isolated from

ganglia of the venus clam, Macrocallista nimbosa, by Price and Greenberg in 1977.19

This was the first of a group of neuropeptides with the characteristic RFamide C terminus

to be identified. Two mammalian RFamide peptides were identified from bovine brain

using anti-FMRFamide antisera, FLFQPQRFamide (NPFF) and

AGEGLSSPFWSLAAPQRFamide (NPAF).192 It was later confirmed that the same gene

and common precursor protein are responsible for synthesis of both peptides.193

Characteristic receptors of NPFF were also recently discovered,194,195 and it was

confirmed that RFamide C terminus was necessary for interaction with the receptor.196

Neuropeptide FF and its related peptides were found to be related to opioid

tolerance and dependence.'97 NPFF was therefore thought to be an endogenous

neuromodulator and a pain-enhancing substance that functionally antagonizes opioids in

the mammalian CNS. But recently, NPFF was found to have both anti- and pro-opioid








activity. It was demonstrated that NPFF functions as an anti-analgesic after

intracerebroventicular (i.c.v) injection, but as an analgesic when injected intrathecally

into the cerebrospinal fluid.198 The authors hypothesized that NPFF represents only one

of many RFamide containing peptides in the nervous system, and that it can exhibit the

biological activities of other structurally related RFamide peptides. NPFF may also

exhibit cardiovascular effects in mammals, since it has been found that it elevates arterial

blood pressure and heart rate in rats.199 It is also interesting to note that arterial blood

pressure increases upon i.c.v. administration of NPFF, but the effect is dose

dependent.200201

Metabolic pathways of NPFF in the brain have been recently investigated. A study

by Sol, et al., investigated extracellular metabolism of NPFF in mouse brain using

forebrain and midbrain slices only.121 Kinetics of metabolic degradation and the main

cleavage products were reported. It was also suggested that the peptide FLFQP (1-5

NPFF fragment) is a product of extracellular metabolism, though this finding is

questionable since it is known that an intracellular enzyme (prolyl oligopeptidase) is

responsible for post-proline bond cleavages.

Experimental Methods

Instrumentation

MALDI-TOFMS

Mass spectra were collected with a Voyager DE-PRO MALDI-TOFMS (Applied

Biosystems, Framingham, MA) equipped with a nitrogen laser (X=337 nm) and operated

at a pulse rate of 3 Hz. Data were acquired with the MALDI instrument operated in the

reflectron mode, using a delayed extraction time of 100 ns. Positively charged ions were








accelerated at 20 kV, and 100 single shot spectra were averaged. Baseline was corrected

using Data Explorer software, provided by the manufacturer.

ESI-QITMS

Mass spectra were acquired on a quadrupole ion trap instrument (LCQ,

ThermoFinnigan, San Jose, CA) operated with the Xcalibur software. Peptide solutions

in 50%MeOH/49%H20/l%CH3COOH solvent were introduced at the flow rate of 150

.l/min. Capillary voltage was set at 4.5 kV, temperature at 1500C, and sheath gas flow at

60 units for all the mass spectral acquisitions. Parameters were optimized for SRM for

all the investigated peptides and the internal standard. The parent ion isolation was in the

1 Da range, and the activation amplitude was optimized at 37% to obtain CID of the

analyzed peptides.

Surveyor LC pumps and the autosampler (ThermoFinnigan, San Jose, CA) were

used to inject 10 gtL of sample into the mass spectrometer. The pumps were operating at

a flow rate of 0.7 ml/min, but the flow was split before the injection valve, using a tee-

junction, producing the final flow rate of 150 pl/min. A schematic presentation of the

experimental set-up is shown in Figure 3-2.

Chemicals and Reagents

a-cyano-4-hydroxycinnamic acid (a-CHCA) and synthetic neuropeptide FF were

obtained from Sigma (St. Louis, MO). Internal standards were purchased from Bachem

Bioscience (King of Prussia, PA). Neuropeptide FF fragment 2-8 was synthesized in Dr.

Laszlo Prokai's laboratory at UF and the total peptide content was determined by the

elemental analysis, performed by Atlantic Microlab Inc. (Norcross, GA). It was

determined that 76% of the sample was the peptide fragment. All the standard peptide








solutions were prepared in artificial cerebrospinal fluid (Harvard Apparatus, Holliston,

MA) to resemble the biological samples. All the solutions used for calibration curves

were prepared from 1 mM stock solutions. Mobile phases were prepared with HPLC

grade solvents (Fisher Scientific, Fair Lawn, NJ) and filtered through 20-nm pore size

aluminum oxide filters (Fisherbrand, Fisher Scientific, Pittsburgh, PA) to prevent

clogging. Trifluoroacetic acid (TFA) and acetic acid were purchased from Fisher

Scientific (Fair Lawn, NJ), while helium gas tanks were provided by PraxAir Inc.,

(Danbury, CT).

Sampling Techniques

Both in vitro and in vivo strategies were employed to investigate the cleavage

patterns of neuropeptide FF. By repeating the reported protocol121 and comparing the in

vitro data with the in vivo processing, performed by using a microdialysis probe, we

hoped to reveal the origin of the FLFQP peptide fragment. Quantification methods were

explored using NPFF and its metabolites sampled in vitro from the brain homogenates

via a microdialysis probe.

In vitro approach

Brain slices were used to study neuropeptide FF processing in the brain in vitro. It

was expected that cell damage during the brain slicing process would be minimized when

compared to brain cell homogenates, since only the cells at the edges of the slices are

expected to be destroyed. The mouse brain was kept in a freezer prior to analysis. Only

the cerebrum forebrainn and midbrain) was used for the study. It was carefully sliced into

300 .im thick slices using a McIlwain tissue chopper (shown in Figure 3-3). All the

preparation was performed in the "cold room", at 40C.








Tissue preparation. Slices were washed with 15 ml of phosphate buffer (15 times

1 ml) to remove soluble peptidases, after which they were dispersed in 1 ml of the same

buffer. A 500-g1 aliquot was incubated for 20 min at 370C under gentle shaking.

Neuropeptide FF was added for 0.5 mM final concentration and incubation stopped by

sampling 45 1l and adding 5 pL of 1 M HC1 at 0, 15, 30, 45, 65, 85, 105, and 145

minutes and centrifuged for 5 minutes. The supernatants were stored in the freezer for

further analysis. All the samples were desalted prior to mass spec analysis using solid

phase extraction.

Determination of protein content. Protein content was determined using the

Bradford method.202 It is established on a BSA (bovine serum albumin) based standard

calibration curve, using photometric readings at 595 nm after dying the proteins with

Coomassie Blue dye (Figure 3-4). Total protein content was used to estimate the amount

of material sampled.

Microdialysis (in vitro). The microdialysate samples of NPFF and its metabolites

were analyzed in order to explore the possibilities and performance of quantification

methods using MALDI-TOF and ESI-QIT MS. Samples were collected in 30 min

intervals, as 100 iM NPFF solution was perfused through a solution of a mouse and rat

brain homogenates via a microdialysis probe (CMA/Microdialysis, Inc., Acton, MA)

(schematic diagram of the experiment is shown in Figure 3-5).

In vivo approach

Microdialysis experiments were performed by inserting a microdialysis probe

CMA 4 mm (CMA/Microdialysis, Inc., Acton, MA) into the striatum region of the mouse

brain. Synthetic neuropeptide FF, at the concentration of 100 pM, was perfused for 30








min before collection and the metabolites collected at the flow rate of 1 piL/min for 1

hour.

Sample Preparation Techniques

Sample desalting methods

Effective removal of salt contamination is particularly important in a quantification

method, since suppression of the signal should be minimized. Solid phase extraction and

"on the spot" clean up were performed for all the MALDI sample preparations. On-line

purification was used throughout all the ESI experiments.

Solid phase extraction (SPE). This method is based on separation of salts from

peptides on a reverse-phase packing. Cartridges used for these experiments contained C-

18 packing and were purchased from Supelco (Bellefonte, PA). The packing was first

activated with 2 ml of MeOH and than equilibrated with 2 ml of 3% aqueous acetic acid

solution. After applying the sample, the salts were eluted with 5 ml of 3% aqueous acetic

acid solution and the peptides with 400 pil of 70% MeOH, 27% H20, 3% CH3COOH.

The effluent was evaporated using the SpeedVac (Savant Instruments, Inc, Holbrook,

NY) and the analyte was redissolved in 0.1% TFA.

Washing the spot. When the developed sample preparation method was used,

crystallization of the matrix layer allowed for additional salt removal from the dry analyte

spot. It is postulated that the salts dissolve faster and easier in water than peptides.

Therefore, 2 p. of water were deposited on the analyte spot and picked up by a pipette tip

after 5 s. This process was repeated twice for optimal performance.

On-line approach. On-line salt removal was introduced by attaching a capillary

sample trap column (Upchurch, Oak Harbor, WA) in line prior to the mass spectrometer,








as shown in Figure 3-2. The sample trap column was filled with C-18 packing, allowing

for the salts to be removed when washed with an aqueous solution. The peptides were

eluted with a solution of 50%MeOH/49%H20/1%CH3COOH and introduced into the

mass spectrometer via the electrospray.

MALDI sample preparation

Matrix and analyte were deposited separately. Two different solutions of a-CHCA

dissolved in acetone/0.1%TFA were used: 1) diluted, 2 mg/ml and 2) concentrated, 10

mg/ml. A 0.3 pL aliquot of a diluted matrix solution was used as a seeding layer for

producing small matrix crystals. A 0.5 lI aliquot of a concentrated matrix solution was

layered on the top. Peptide solution (1 Il) was deposited on the top of the dried matrix

layer. One of the important parts of this sample preparation was the final removal of salts

by washing the spot with cold H20, as described above.



Results

Metabolic Differences

In vitro data

The presence of the FLFQP fragment was confirmed by MALDI-TOFMS and ESI-

QITMS after metabolism of 100 p[M NPFF within mouse brain slices (Figure 3-6).

Degradation of NPFF was noticed to be more rapid than reported by Sol, et al.,121

resulting in complete processing of introduced peptide within the first 30 min.

In vivo data

The FLFQP fragment was not detected by in vivo microdialysis (Figure 3-7). This

brought into question the validity of the conclusions presented by Sol, et al.121 We








postulate that the Ps-Q6 bond is not broken naturally in the brain, but only in damaged

tissue where enzymes like prolyl oligopeptidase are released as cells are ripped open. We

determined that the concentration of 100 [tM was too high for investigating the NPFF

metabolic pathway. This high concentration was only necessary in the previous work

because CE/UV analysis was employed, which is considerably less sensitive than mass

spectrometry detection. All samples were therefore diluted 100-fold prior to MALDI-

TOFMS analysis.

Quantification of Neuropeptide FF Metabolism Using MALDI-TOFMS

Optimization of MALDI sample preparation

After confirming that fast solvent evaporation led to the formation of small matrix

crystals and a homogenous matrix layer (chapter 2), new sample preparation methods

were developed to investigate neuropeptide metabolism quantitatively. A simple method

for depositing matrix with a fast evaporating solution was further improved by first

depositing a seeding matrix layer with 0.3 p1 of a diluted matrix solution (2 mg/ml in

acetone/0.1%TFA) using a gel-loader pipette tip. This layer evaporated rapidly,

producing a thin layer of small, even matrix crystals. This was followed by deposition of

0.5 pIL of a more concentrated matrix solution (10 mg/ml in acetone/0.1%TFA), forming

a thicker layer while maintaining the same homogeneity.

The analytes were deposited in 1 pl of a 0.1%TFA aqueous solution onto the dry

matrix layer, and allowed to air dry. By using this method, all of the analyte could be

focused into a -1 mm diameter spot due to the surface tension of the sample solution and

the hydrophobicity of the matrix layer. The spot was than washed with 2 pl\ of H20 that

was pipetted off after 5 seconds. The constant ion signal before and after this wash








confirmed that peptides were not lost, while most of the sodiated ion adducts were

eliminated, enabling more constant peak height determinations.

Internal standard selection

The main goal in searching for a suitable internal standard was to match structural

similarity and mass value with the investigated peptides. We first tested FMRFamide, a

naturally occurring analog of NPFF with the same characteristic RFamide sequence at the

C terminus. Even though the peptides were similar we were unable to use the

FMRFamide as an internal standard due to methionine oxidation resulting in the presence

of both unoxidized and oxidized forms in the sample (Figure 3-8). Therefore, as a

criterion in the search for a new internal standard, we stipulated that the peptide should

not contain methionine and cysteine.

A second internal standard with the sequence WXRFamide (X=Nle, norleucine),

was selected for further investigations. It also had a similar structural and mass to NPFF

with no potential for oxidized amino acids (Figure 3-9).

Generating a calibration curve

Calibration curves were generated for two neuropeptide FF fragments (1-8 and 2-8)

by ratioing peak heights with the internal standard. After generating the calibration

curves for each peptide fragment, the quantitative interaction of the two fragments was

evaluated in the system. Ion signals were recorded for both NPFF peptides to investigate

the influence of each on the signal intensity of the other and if signal suppression would

occur. The experiment was performed in two parts: 1) by preparing the same

concentrations of both peptide fragments present (for example, 0.5 JpM NPFF 1-8, 0.5

.M NPFF 2-8), and 2) by preparing different concentrations of the two investigated








peptides (like, 0.5 pM NPFF 1-8, 0.1 p.M NPFF 2-8). The peptide concentration ranged

from 0.1-1 pM. The calibration plots are shown in Figure 3-10.

Similarities between the two sets of conditions are noticed throughout the whole

range, except at the highest concentration tested, where the largest error is noticed. This

is due to signal suppression and competition during matrix-assisted ionization. No

significant signal suppression is noticed at lower concentrations, even when the other

peptide is present at 1 pM. It was also noticed that calibration curves spanning

throughout one order of magnitude in concentrations were not linear but polynomial.

This is caused by signal saturation and competition during ionization.

Calibration curves were generated daily for accurate quantification of analytes.

The samples were diluted in order to fit into the limited dynamic range of the instrument

and to not saturate the detector. Each sample solution was analyzed as three spots, with

error bars generated from the standard deviation of the peak height ratios measured.

Quantification of NPFF 1-8 and its metabolized fragment 2-8 was performed with

microdialysate samples collected after perfusion of a 100 gpM solution of NPFF into

mouse and rat brain homogenates. A curve showing metabolic equilibrium was

generated in both experiments with both peptides. It was demonstrated that 30-60

minutes were required to reach equilibrium (Figures 3-11 and 3-12). A discrepancy in

the trend was noticed for the 60 min sample point from the rat brain homogenate; this

was probably a result of experimental error during sample collection.

Quantification of Neuropeptide Metabolism Using ESI-QITMS

A mixture of the NPFF fragments 1-8 and 2-8 and the internal standard

(WXRFamide) was analyzed in both MS and MS/MS modes to investigate if there were




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID ENIXSRITU_PZWYB5 INGEST_TIME 2011-11-02T15:12:29Z PACKAGE AA00004689_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

EXPLORING NEUROPEPTIDE METABOLISM IN THE BRAIN BY MASS SPECTROMETRY By TAMARA BLAGOJEVIC A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 by Tamara Blagojevic

PAGE 3

To my family friends and teachers.

PAGE 4

ACKNOWLEDGMENTS I wish to thank my whole family for all their love and support. I am especially grateful to my parents and sister for understanding my adventurous spirit and supporting some of my decisions that changed all our lives. I also wish to thank my husband for all his love and help. He has always been there to comfort me, encourage me and help me survive some of the most difficult times of my life. I would like to acknowledge Dr. John Eyler, who helped me find the way to accomplish my goals He showed interest in my work, as well as in my progress towards getting the degree I am very thankful to Dr. Laszlo Prokai for being confident and enthusiastic about my research, for offering me a great chance to apply my knowledge and giving me the opportunity to be involved in the team work while opening new perspectives of my research interests I also express my gratitude to Dr. David Powell for his guidance and patience as I was initially exploring the mass spectrometry world Special thanks go to Will Haskins for our endless talks, games and songs that made my life in the lab bearable and for encouragement and understanding that helped me survive the first few years of graduate school. I am enormously grateful to Dr. James Deyrupwithout his kindness and care I would have never been given the opportunity to explore new avenues. He was inv olved and helpful always had a word of wisdom, and encouraged my spirits whenever I needed it. IV

PAGE 5

I learned a lot and was motivated to learn more by many of my teachers. Therefore, I am grateful to all of them for inspiring my life-long search for enthusiastic and interesting scientific ways. Without the support and encouragement of my friends, my life would have followed some other path With her patience and persistence, Ivana was there to make me eat and tolerate my antsy moods; we were broke and on our own in the new world, but we never gave up on our journey to the stars Irina was there to be our guide and a parent as we came to the new continent, and Isa and Janina showed us that one can have so much in common and find a best friend in someone who grew up in a different part of the world. I also got enormous support from my best friends in Serbia; they were also with me through all my ups and downs. If it weren t for all of my friends, I would have not been as happy and fulfilled. V

PAGE 6

TABLE OF CONTENTS ACKNOWLEDGMENTS ... .... . . . ... ... . .... ...... .... ..... ... . ........ ... ..... ............ ..... . ........... i v ABSTRACT ............. ... ... . ..................... ...... ......... .... ... . ... .......... . . ..... . . ... . ......... ........ i x CHAPTER 1 WTRODUCTION . ...... ................... . ..... ....... .............. .... . .............. ............. ... ... ... .... 1 A Century of Mass Spectrometry ......... ........... ... .... . ... ... ...... ...... ....... ... ...... . .... .... . 1 Ionizing Large Molecules ....... ....... ........... ... ..... .... .......... . ... ........ .... .... .... ........ .. 1 Matrix-assisted laser desorption/ ioni z ation time-of-flight mass spectrometry (MALDI-TOFMS) ............... .... ... ........ ............ ........ ........ ... 2 Electrospray ionization quadrupole ion trap mass spectrometry (ESIQITMS) ...... ..... ........... ....... ...... ... . .... ............. ............ . . . .... . ... .... . .... 9 Quantification by Mass Spectrometry . ... ... . ........ . .......... ... ... ... .... .... ....... . .... 12 Coupling Liquid Chromatography (LC) to Mass Spectrom e try . . ........... ... ... . .... ... 13 Reverse Phase LC ......... ...... ...... ..... ... ... .... ......... ...... ................ ... . . . ....... . .... . 15 Capillary LC . ............. . ... ....... .................. . .... ... .... ....... ............. ........ ...... .... .... 15 Neuroscience ..... ...... ....... ...... .... ... ....... . ... . .... . ... ... .............. ..... ........ ... ............ ... . 1 6 Neuropeptides .... . .... ........ ....... ..................... .... ... .... ... . ..... . ....... ....... ..... ... ...... 18 Sampling Techniques .. ..... ....... ..... . ..... .... ......... .... ............. ............ ................. . 18 In v i t ro studies ... ... .... ... .... ... . ... . ....... ...... .... ..... ... ... .......... ....... ......... .... 19 In viv o stud i esmicrodialysis .......... ... ...... .... .......... ..... ... . ... .......... ........ .2 0 2 INVESTIGATION OF SAMPLE PREPARATION METHODS TO IMPROVE MALDI-TOFMS PERFORMANCE AND F ACILI T A T E WTEGRATION OF CAPILLARY LC SEPARATIONS ..... ........... .... ..... .... . ... ... .......... ... ... . ........ .... ... 35 Introduction ... . ....... ...... .... .............................. .... . ...... ..... ......... ..... . ......... .... .... .... ..... 3 5 MALDI Sample Preparation Considerations .... ... ...... ...... ........ ....... ......... ..... . 35 Matrix crystal si ze ............ .... ........... ... ....... . ....... ........... ........ ................ . 36 Sample spot siz e . . ............ . .... . ..... ...... .................... ... .............. ..... ........ 37 Capillary LC/MALDI Coupling .... . ..... ... ........ ... . ............ ... . ............. .......... .... 3 8 On-line coupling .... . ........ ..... .... .............. ... . .... . ... . ..... .......... ........... .... 38 Offlin e coupling ... ... ..... ..... ....... .... ..... ........ ........ . .... .... .... . .... ..... ... ... .. 38 Im agi n g T e chniqu e s ... . . ... ... ... ... . . . . . ... ...... ........... ......... ..... ....... .............. . . 39 Scanning el e ctron microscopy (SEM) ...... ... ....... ..... .. ... ............ . ..... ...... .40 T ransm i ssion electron microscopy (T E M) . ..... ....... ......... ............ ...... .... . . .40 V I

PAGE 7

Fluorescence imaging .... ... . ... . ............ .. .... . ..... ... ........... .... ..... . . . ...... . 40 Experimental Methods ...... .... . ... . . .... ..... . . .. .... .......... ...... .... . .... . . . ... . ....... . ... 4 1 Instrurnentation ..... . . ... . ....... ... . ... .... ......... .................. ............. ... . . .. ... ... .... 4 1 Chemicals and Reagents ..... ... . ... ................................................................ ..... 42 Electrospray Deposition Method (ESD) . ... . .... . . . ... .. ..... .... ....... ......... ...... ...... .42 ESD apparatus .. .... ...... . ........ ......... ..... ..... .... .... .... ... . ............... . . .... . ... 42 Electrospray tip production ... . .................. ...... ... . .... .......... ...... ...... .. . . . .43 Hydrophobic Sample Supports (AnchorChip Plat es) ... ...... ... ..... .... ..... .... ... .43 LC/MALDI ..... .... . .... ... .... .... . . ..... ........ . ... ...... ......... ... ... ...... .. ... .... . . ...... .. 43 Capillary column packing ................. ... ................. ........ ...... ... ................... .43 Off-line method system ... ..... ... ......... . ...... .... .... ..... .................. .......... ... ... 44 On-line collection off-line anal ysis system ... .......... ... ............. ... ............. .44 Results .... .... ...... ..... ..... ... . . ... ..... .......... .... .... ... .... . .... . .......... .... .... . ... . ... ... ... .... 45 Minimiz ing Matrix Crystal Size ... . ... ....... .... ... .... .......... .... ...... .......... ... ..... ....... .45 Optimization of ESD parameters .... . .... ... ............ ............ ... ........................ 45 Matrix crystal images . ........ ... . .. ....... .......... . .... .. . .... ........ ....... . ....... ... ... 46 Minimizing Analyte Spot Size ....... ..................... ..... ................. ..... ............... .46 Optimi z ation of ESD parameters ..... ... ........ . . ............ ...... ... .... ... .. .... . ... .4 6 Analyte spot imaging ........ ..... . ........... ... ... .................. . ............ . .... ......... 4 7 AnchorChip T M plates ... ... .... ... .. . . .... .... .... . .... . ..... .... ....... ... . . . .... ... .... ... .. 48 Capillary LCIMALDI -TOFMS ........ ....... ......... ......... ...... .... ........ .... . ........... ... 48 Prel iminary in v estigations ... ........ ..... ... .... ... . . ... ...... .. . . . . .... ... ... ... ... .... ... .4 8 On-line collection off-line analysis data ... ......... ...... . .............. ....... .... ..... .49 Conclusions and Future Directions . . .... .... ..... . .... ....... .... ......... ..... ... ....... ...... ....... 50 3 NEUROPEPTIDE FF METABOLISM STUDIES AND DEVELOPMENT O F QUANTIFICATION METHODS BY MALDI-TOFMS AND ESI-QITMS ............ 74 Introduction ........ . . . ... . ... . .... ... ..... ....... . ..... ....... . ... ... . . ...... . ... . . .......... .... .... . .... 7 4 Quantification Using Mass Spectrom etry ......... ..... .... ... ...... .. ........ ... ............. . 7 4 Relati v e quantification .. .. ........ ........... ... . ....... ... ... ... ...... ..... . . ... .... ... .... ... 7 4 Absolute quantification .... ..................... ....... .. ...... ... ... ..... .... .... ......... ....... 7 7 Quantification Employing MALDI-MS . ....... . ... ... ... ...................................... 7 8 Irre producibility influences ......... ........... ................ .... ... ... ........ ............ ... 7 9 Internal standard s e lection . . ......... . . ....... . ..... .... ....... ..... ..................... ....... 79 Quantification Employing ESI-QITMS .................... .................. ....... ..... .......... 80 Chromatographic separation requirements ............. .............. ...................... 80 Mass spectroscopic r equire m e nts ........................ ... ............ . . ....... ............ 81 N e urop e ptid e FF . . .... .................. ............ ... ... ...... ................ .... ....................... ... 81 E x perimental Methods .... ..... .......... ... ....... . .... ....... ... .... .. ....... . .... . . ...... ............. .... 8 2 Instrumentation ...... ... . .... . . ........ .... .... .... ... .. . .......... . .. . .... ........ . . .... .... . .... . 82 MALDI-TOFMS ... ... . . . . .... . ... ... .... . . .... .... . .... ..... ...... . ... . .. ....... ...... ...... 8 2 ESI QITMS . .. .. .... . ... . ... ....... ......... ...... ..... .... . . ....... ..... ........... ..... .... ... ... 83 Chemicals and Reagents ..... ............... .......... .. ... .... . ....... .... ........ ................ ..... 8 3 Sampling Techniques ... . . ... .... ..... .... ... ....... .... .... .................. . . . . . ........ . .... ... 84 In v it ro approach ................... ..... ............ ................. ... ... ......... ... .................. 84 In vi v o approac h ... ..... ................... .......................... ..... ....... .... .... .... ...... ....... 8 5 v u

PAGE 8

Sample Preparation Techniques .. ...... .... .... .... .... ........ . . . ... ........... . .... ....... ... . 86 Sample desalting methods ... ..... ...... .... ......... ... ......... .... .... . ...... .... .... ....... 86 MALDI sample preparation ...... .......... . . . ... . ...... . . ...... . . . . . . ... ... .... . . . 8 7 Results . ....... ... ... ..... ............ . ............ .............. .... ..... .... ............. ........ ...... . . ... .... ... 8 7 Metabolic Differences ...... . .......... ............... .... ... ... .. ..... ......... . . ......... . ........... 8 7 In v itro data ...... ... . . .... . . ... . ....... ... .... ...... ...... ... .... .... ...... ... . .... ......... .... 87 In v i v o data ..... .... . ....... ..... ....... ..... .... ...... ..... ... . . . ......... . ....... ... ...... ... ..... 8 7 Quantification ofNeuropeptide FF Metabo l ism Using MALD ITOF M S .. ........ 88 Optimization ofMALDI sample preparation . ... ....... ... ... ... ... .............. ....... 88 Internal standard select i on .... .... . ... . .... ..... . ... ............ ... . . ......... .... . ... .... 89 Generating a calibration curve ... .......... ... . .... ... .......................... ... ......... .... 89 Quantification ofNeurope ptid e Metabolism Using ESI-QITMS ........ ... ........ ... 90 MS/MS optimi z ation .......... ...... . . . ... ............ .... ... ... . ... .... ....... .... . ..... ..... 9 1 Generating the calibration curve .... ... ... . ... .... .......... .... ... ......... ......... ... ...... 9 2 Conclusions and Future Directions . . . . ............................... ....... . ....... ...................... 9 2 4 DEVELOPMENT OF DIFFERENTIAL Q U ANTIFICATION FOR MONITORING METABOLISM OF DYNORPHIN 1-8 IN THE BRAIN ..... .... .... .... . . .... .... ....... 1 1 2 Introduction ..... . . ... .......... ... .... .............. . .... . . ............. ... ... .... ...... ... ...... ...... ...... ... 1 1 2 Opioid Peptides ................ ... .... ............... ...... ...... .... ... ........ . .............. . . . ... 1 1 2 Discovery of dynorphins . ....... . ......... ...... ........ .... ...... . . ... ....... . .......... . 113 Physiological role ... . . . .... ..... . . ........... ..... .... ... ................ ........ ...... .... . . 1 1 3 Neurochemical processing ............ ..... ... .... . ................. . ...... .... .............. 113 Diffe rential Quantification Approach ..... ... .... ......... .. ... ............... ...... .......... . ... 115 Chemical labeling of N-terminus . ...... ....... .......... ............ .... ... ... .... .... ...... 1 1 6 Chemical labeling of C-terminus ... ... .... .... .... ... ........ ... .... ............ ............. 11 7 Experimental Methods . .... . .... . ... ....... ..... .... .... ...... .... .... . . .... . ...... . . . . ...... ..... . 11 7 Sampl e Collection by Microdialysis ... ..... .... .... .......... ........ ..... ......................... 117 Sample Pre paration ... . ....... ..... .... .................. .... ...... . . ... ........ ............ ....... .... 1 1 7 Derivati z ation methods .... .... ............... ...... ....... ... ... ...... ... . .... ......... . ....... 118 MALD I sample preparation ... . ........ .... .................. ....... ... ......... .... .... ..... ... 118 Results ... ... .... . . . .... .... . ... ... . ... ... ..... .... .... ............. ... ....... .... . . . .......... .... ......... 119 Selection of D e rivati z ation M e thod ... ...... .. .. ... .......... .......... .... ........... ... ....... 1 1 9 Method Ex amination . ........... ........... .... ... ........... ... .... .......... .................... . ...... 1 1 9 Influence of Inhibitors on Dynorphin 1-8 M e tabolism in t h e Brain ...... ......... 121 T ime D e pendent Metabolism ofDynorphin 1-8 in t he Bra i n ...... ................ .... 121 Conclusions and Futur e Dir e ctions ............ ..... . ... .... .... ... . ... ....... ........ ...... ... .......... 122 5 CONCL USIONS AND FUTURE DIRECTIONS ....... .... .... ............. .... .... ............. 142 LIST O F RE FE RENCES ..... ........... .... ......... ..... . . .............. ... ... ...... ...... ..... ........ .... .... 148 BIOGRAPHICA L SKETCH ... ... ..... ...... .... . ... ........ ........ ... ... .... ...... ...... . . .... ....... ...... 160 Vlll

PAGE 9

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy EXPLORING NEUROPEPTIDE METABOLISM IN THE BRAIN BY MASS SPECTROMETRY By Tamara Blagojevic August 2004 Chair: Dr. John R. Eyler Major Department: Chemistry The challenge of improving the reproducibility and sensitivity of matrix-assisted laser desorption/ ionization time-of-flight mass spectrometry (MALDI-TOFMS) was addressed by improving the sample preparation step a-CHCA was used as the matrix for all sample preparation methods. An electrospray deposition method and the use of prestructured MALDI supports with dried droplet sample preparation were investigated for improved sensitivity and reproducibility. AnchorChip plates focus the deposited drop into a small 200-m diameter spot surrounded by a hydrophobic (Teflon-like) coating. Detection limits below the attomole level were achieved when 1 l of a 1 pM peptide solution was analyzed. Electrospray deposition was explored for improvement in crystal homogeneity and minimi z ation of the sample spot size Electrospray deposition concentrated analyte into spots less than 300 m in diameter while evenly distributing lX

PAGE 10

and incorporating the analyte into the matrix layer. Optimized electrospray conditions led to detection limits of< 10 amol when 15 nl of a 0. 7 nM solution was electrosprayed. Capabilities and performance of quantification approaches when using MALDI TOFMS and ESI-QIT were compared. Also, both methods were used for investigating in vitro and in vivo processing of neuropeptide FF infused into a mouse brain. It was shown that MALDI-TOFMS provides better resolution and detection limits, but ESI-QIT provides better reproducibility and dynamic range. Differences in NPFF metabolism when sampled in vitro are present due to release of various enzymes upon cell lysis during the in vitro experiment. The approach of differential quantification was utilized for quantifying metabolized dynorphin 1-8 fragments in the brain by MALDI-TOFMS. Dynorphin 1-8 introduction and sample collection were performed in vivo via a microdialysis probe. Sample and control were esterified with d0and d3-methanol, respectively. The light and heavy samples were mixed in a 1: 1 v/v ratio, of which 1 l was analyzed Changes in fragment abundance were then correlated to enzyme activity and the enzyme inhibitors involved in processing of dynorphin 1-8. Enzyme activity was not changing with time, which was confirmed by the ratio of unprocessed dynorphin 1-8 to the amount initially infused. X

PAGE 11

CHAPTER 1 INTRODUCTION A Century of Mass Spectrometry Sir J. J. Thompson won a Nobel Prize in 1912 for the pioneering research and foundation of mass spectrometry. Almost a century later John Fenn and Koichi Tanaka won a Nobel Prize for development of sophisticated ionization methods enabling detection and characterization of proteins, peptides, nucleic acids, and polymers. Mass spectrometry emerged from the need to define and explain basic physicochernical properties of atoms, and yet became one of the most powerful methods for macromolecular characterization of living systems. It has become an interdisciplinary research methodology, impacting virtually every area of science from physics through chemistry and biology, from geochronology and space research to studies of the physical and chemical properties of materials, atoms, and particles, to proteomics and genomics projects. There are many different kinds of mass spectrometers, described generally by the types of ionization sources, mass analyzers, or detectors used. The choice of ionization method and the mass analyzer depends on the requirement of the particular application. Mass spectrometers based on different approaches can, therefore, provide complementary information. Ionizing Large Molecules Successful ionization is crucial for mass spectrometry. The method selection will depend on the sample, sample origin, and data required. Ionizing volatile organic 1

PAGE 12

2 molecules was routinely performed in mass spectrometric analysis by either electron impact (El) or chemical ionization (Cl). But, as the molecular mass of the analytes gets bigger, they become less volatile. This challenge was faced and explored for years until the successful development of matrix-assisted laser desorption/ionization (MALDI) and electrospray ionization (ESI). Matrix-assisted laser desorption/ ionization time-of-flight mass spectrometry (MALDI-TOFMS) Developed simultaneously in laboratories of Karas and Hillenkamp and Tanaka during 1987,1 2 MALDI caused a revolution in mass spectrometry during the 1990s by gaining the attention of biotechnological fields of research (like medicine, pharmacy bioorganic molecular synthesis) Results from both of these groups were published in 1988.3 4 Tanaka's work was based on use of a pulsed N2 laser (337 nm) and slurry of glycerol with ultrafine metal powder as a matrix Detection of the molecular ion of carboxypeptidase-A (m/z 34,529) was demonstrated. Karas and Hillenkamp, on the other hand, used a Nd:YAG laser (266 nm) to desorb proteins (like bovine serum albumin) dissolved in a matrix solution containing nicotinic acid. The method was soon applied for analysis of proteins with molecular weights in the range of 100 kD ,5 8 with the idea of using an organic UV absorbing matrices adopted (table 1-1 ). Since presence of the matrix is what differentiates MALDI from other laser desorption techniques, the importance of its role, selection and deposition is one of the key issues to discuss A high excess of matrix to sample i s important, since the matrix serves as the primary absorber of the UV radiation and breaks down rapidly, expanding into the gas phase and carrying along undamaged analyte molecules (Figure 1-1 ). Additionally, the high matrix / sample ratio reduces association between analyte molecules

PAGE 13

3 and provides enough protonated and free-radical products to ionize the molecules of interest.9 A number of detailed models for the MALDI ionization mechanisms have been described, 1 0 -12 and will be discussed in more detail, but most of the experimen t al parameters, like matrix selection sample preparation method, and parameters used for the ionization, are still predominantly empirically based Ion formation. Although having received noteworthy attention, the fundamen t al processes of MALDI ion generation are not fully understood 13-16 and are st i ll a matter of active research. Good understanding of ionization pathways should help increase ion yields, control analyte charge states and fragmentation and enable studies of ne w classes of analytes. A number of chemical and physical pathways for MALDI ion format i on have been proposed but a general agreement is that there is no single mechanism that can explain all the ions generated and observed within an experiment.13 1 7 A variety of matrices, analytes and experimental parameters as well as different instrumen t al characteristics present major obstacles to ion formation mechanism studies Some o f the parameters usually considered are laser wav elength and laser pulse width but the choice of matrix is known to be crucial for success in MALDI experiments. Since the presence of matrix is necessary for MALDI it i s important to establish the role of the matrix by studying its chemical nature (measuring proton affin ity, absorbance spectrum ionization potential) and its crystal structure Derivatives of benzoic acid, cinnamic acid and related aromatic compounds were recogni z ed earl y on as good MALDI matrices for proteins ,7 1 8 2-(4-h y dro x yphenyla z o)-ben z oic acid ( HABA) as a matrix for peptides and proteins, 1 9 3-hydroxypicol ini c acid ( 3-HP A) for oligonucleotides ,20 2 1 and the list went on as mor e matrices and new a nal yte s we re

PAGE 14

4 introduced. Albeit, there are still no clear guidelines for matrix selection for a particular analytical problem, since many of the uncertainties are related to matrix crystallization and analyte incorporation "Matrix assistance" and matrix selection should be better understood through mechanistic studies of ionization process. It is interesting to notice that only a few types of ions are commonly observed in the positive mode MALDI experiments: radical cations, protonated molecular ions, and cationized molecular ions. Several mechanisms have been proposed to explain ionization processes during MALDI. The initial set of react ions is expected to provide "active matrix species that are ready to ionize the analyte molecules. These are followed by processes that result in the final set of ions being detected Some of the proposed reactions for primary processes are summarized in figure 1-2. One of the primary processes, multiphoton ionization of the matrix leading to a formation of matrix radical cations, 22 is shown in equation 1-1 : (1-1) Another mechanistic possibility is that two or more separately excited matrix molecules pool their energy to yield one matrix radical cation or a highly excited matrix molecule2 3 27 MM i h v >M*M*~M +M+ +e M M *+A MM + A + + e (1-2) (1-3) This model is statistically favored, since it is more likely that a neighbor of an excited matrix will be the next to be excited, rather than that the excited molecule will be hit by a second photon. Besides, the matrix suppression effect also suggests a multicenter ionization model ,27 since it is observed that at sufficiently hi g h analyt e concentration

PAGE 15

5 matrix 10n signals can be completely suppressed in MALDI, lowering the overall ionization yield. Excited-state proton transfer (ESPT) was a model proposed by Karas, et al., 28 in early MALDI studies. It is based on a presump tion that the singly excited matrix molecule is more acidic than when in its ground state, and the following is expected to happen: M +hv ~M* M*+A ~(M-Hf +AH+ M*+M~(M-Hf+MH+ (1-4) (1-5) (1-6) However, none of the popular matrices is known to be particularly ESPT-active, resulting in some doubts about this mechanism. The correlation of positive and negative ion modes suggests that disproportionation reactions can be an active ionization mechanism: 2M n(hv> >(MM)*~ (M -Hf +MH+ (MM)*~M-+M+ (1-7) (1-8) Although this mechanism fulfils the energy requirements, the evidence for it is not yet strong.29 An attractive idea is that the ions observed in the MALDI mass spectra are already present in the solid sample and just ablated by the laser pulse.3 but it is difficult to confirm that ions observed are truly preformed and not the result of secondary gas phase chemistry in the plume Thermal ionization is also possible, and it can take place directly at the particle surface or in the matrix bulk: (1-9) It is found that the internal energy of molecules desorbed in an UV MALDI experiment corresponds to effective temperature of around 500 K,3 1 a temperature at which a thermal ionization/desorption reaction is not significant.

PAGE 16

6 After the initial ion formation, sets of additional reactions occur in the MALDI plume. This set of reactions provides intermediates that generate protonated and cationized analyte ions, later observed in mass spectra. Some of the most important are proton-transfer reactions, cationization reactions and electron transfer reactions. Several d 1 1 d fi 22 3 2 3 3 mo es me u e matnx-matnx proton trans er react10ns ' (M-H)" +e-~(M-Hr M +e-(M-Hr +H It is also found that the matrix-to-analyte proton transfer reaction : (1-10) (1-11) (1-12) (1-13) can be a thermodynamically favorable process.32 The presence of AH+ ions in the plume was also confirmed in studies done by Bokelmann et al., 34 and Kinsel e t al.,35 Furthermore, it was found that addition of alkali matrix salts to MALDI samples promotes cationization of synthetic polymers.3 6 This encouraged the addition of other ions, like silver and copper ions, to enhance polymer ionization by MALDI. 3741 In general, cations do not even need to be added, especially when analy z ing biological samples, since the natural presence of Na+ and K + is sufficient to give strong alkali cationized signals. Several studies have provided strong confirmation of processes involved in formation of cationized peptides.42,43 An electron transfer reaction is also possible if the analyte has a lower ionization potential than the matrix: (1-14) This process was observed when analyzing fullerenes ferrocenes and metallocycles.44,45

PAGE 17

7 A study done by Gluckmann and Karas on different matrices for vanous preparation protocols and for different classes of analytes has shown that the magnitude of the initial plume velocity can be a valuable and meaningful characteristic of the MALDI process .46 It has been shown that the initial velocity is characteristic of a matrix but may be affected by the preparation conditions and additives Another characteristic of the ablation process was demonstrated by Beavis and Chait, who showed that UV-laser excitation results in the ablation of a shallow surface layer, up to a 1 m in thickness, due to high absorption coefficients of matrix molecules.4 7 Experimental designs of MALDI experiments could be improved with further exploration of the MALDI processes. Some of the expected benefits are manipulation of fragmentation and charge states, improved ionization yields and better guidelines for matrix selection, including possible new designs Time-of-flight ion separation. The first mass spectrographs designed by Thompson4 8 and Aston4 9 used a magnetic field to separate ions according to their mass to-charge ratio. All the first mass analyzers were based on magnets until the 1940s when advances in electronics and mechanics allowed development of ion separation based on flight time. By the end of the 1940s Cameron and Eggers built an instrument the y called an ion v e locitron where the ion velocities were inversely proportional to the square root of their mass-to-charge ratio s o At the same time a patent for a simi lar instrument had already been obtained by Stephens. s1 In all the described instruments the ions of mass m (kg) were accelerated to constant energy (eV) resulting in flight time t (s) across the drift region L (m) given by: ( m ) t= -L 2 e V ( 1-15 )

PAGE 18

8 The ions leaving the source and entering the time-of-flight analyzer have neither exactly the same starting times nor exactly the same kinetic energy. Various TOF mass spectrometer designs have been developed to compensate for these differences. During the 1950s researchers suggested a delayed pulse for acceleration, enabling the ions to be formed in a field-free region .52 This allowed an even start for all the ions and therefore improved the resolution. In 1973 Mamyrin and co-workers introduced a first design for a high resolution TOF MS, based on a dual-stage grid system .53 This innovation was developed into the design of a reflectron, an ion optic device in which ions pass through a "mirror" and their flight is reversed (Figure 1-3). The reflectron allows ions with greater kinetic energies to penetrate deeper into the reflectron field than ions with smaller kinetic energies. The ions that penetrate deeper will take longer to return to the detector. If a packet of ions of a given mass-to-charge ratio contains ions with varying kinetic energies then the reflectron will decrease the spread in the ion flight times, and therefo r e improve the resolution of the time-of-flight mass spectrometer. This invention started a new era of development of high-resolution TOF instruments, based on delayed ion extraction and use of a reflectron for longer drift tube length. A comparison of resoluti on that can be achieved with these improvements is shown in Figure 1-4. TOF is considered to be the fastest MS analy ze r with the highest practical mass range of all the MS analyzers The pulsed desorption of ions favors the combination of the laser desorption ion source with a time -offlight mass analyzer54 55 or an FT ICR mass analyzer, 5658 both capable of recording complete mass spectra for each laser shot. It can provide MS/MS information from post-source decay but is not commonly used for MS/MS experiments due to the limited precursor-ion select i vit y The use and

PAGE 19

9 development of TOF reemerged upon introduction of desorption ioniza t ion methods at the end of 1980s. The introduction of MALDI and discovery of its many poss i bilities have driven much of the current interest in TOF analyzers Electrospray ionization quadrupole ion trap mass spectrometry (ESI-QITMS) Electrospray (ES) is a method by which ions present in a solution can be transferred into the gas phase However, electrospray existed long before its application in mass spectrometry as a method of considerable importance for the electrostatic dispersion of liquids and the creation of aerosols. Solvent evaporation from the charged droplets leads to droplet shrinkage, increased coulombic repulsion and release of t he electrostatic strain by droplet fission Repeated evaporation and fission of parent and offspring droplets leads to formation of a droplet population extending down to v er y small multiply-charged droplets, with radii in the -10-100 nm range. The first detailed studies of electrospray phenomena were carried out by Zeleny in the early 1900s .5 9 In the late 1960s, Dole and co-workers were first to attempt using an atmospheric pressure electrostatic sprayer as a means to produce gas-phase ions from macromolecules in a liquid solution for mass spectrometric analysis.60 63 But the first successful coupling of an ES ion source to a mass spectrometer is attributed to the Fenn64 66 and Alexandrov67 research groups. This final confirmation of electrospray capabilities caused an electrospray revolution and inspired rapid developments in the world of mass spectrometry. Ion formation There are three major steps in the production of gas-phase ions from electrolyte ions in solution by electrospray : 1) production of charged droplets at t he ES capillary tip; 2) shrinkage of the charged droplets by solvent evaporation and repeated droplet disintegration and 3) the actual mechanism by which g as-phase ions ar e

PAGE 20

10 produced from the very small and charged droplets. There is still no consensus on the mechanism by which analyte ions are formed from the charged droplets. Dole and co workers proposed the charge residue model (CRM), suggesting ions originate from small droplets containing one molecule of the analyte.60 Evaporation of the solvent from t he initially formed droplet leads to a reduction in diameter, and an increase in surface field until the Rayleigh limit is reached As the magnitude of the charge is sufficient to overcome the surface tension holding the droplet together, a coulomb explosion occurs. Continuation of this process results the formation of an ion containing a single analyte molecule. The ion desorption model (IDM) is based on the work of Iribarne and Thomson.68 69 This model assumes that the surface e lectric field becomes sufficiently large to lift an analyte ion at the surface of the droplet over the energy barrier enabling its escape. Diagrams illustrating the CRM and IDM models of ion formation in electrospray are shown in Figure 1-5. A unique characteristic of ESI is extensive multiple charging of ionic species. Number and placement of charge throughout a molecule depends on many experimental factors. Some are dependent on the solution characteristics (like pH, solvent polarity analyte concentration and molecule conformation) while others depend on instrumental operating conditions (like temperature, gas flow and electrospray voltage). It has been demonstrated that the number of basic sites on a peptide (or protein) directly relates to the charge state of the specie observed in positive-ion mode ESI mass spectra .7 0 72 Quadrupole ion trap (QIT) operation and capabilities. Developments in this field over the last 10 years have made the QITMS an excellent tool for biomolecular analysis The QIT is a three-electrode device (Figure 1-6). Two of the electrodes are

PAGE 21

11 identical and are called end caps, while the third electrode is donut shaped and is called ring electrode. The holes in the end caps are for ion injection into the device and ion ejection out of the device to a detector. A three-dimensional quadrupole field is formed when a suitable RF voltage is applied to the ring electrode, trapping the ions by continuously forcing them toward the center of the device Unlike triple-quadrupole mass spectrometers, where each operation on the ion beam is separated in space the QIT operates on the ions over a period of time, but within the same analyzer. The ions are dynamically trapped by the applied RF potentials, while the "bath gas" (He) helps contain the ions in the trap. The RF and DC potentials can be scanned to eject successive mass to-charge ratios from the trap into the detector. Wolfgang Paul first described a quadrupole ion trap mass spectrometer in 1953 ,73 74 a discovery for which he was awarded a Nobel Prize in 1989 The popularity of QIT has been growing ever since, inspiring further developments and improvements in the design. Coupling GC to QIT has shown that the resolution and sensitivity of the instrument were greatly improved when approximately 1 mTorr of helium was present in the QIT chamber.75 ESI was first coupled to a QIT in 1990,7 6 after which many upgrades led to this system being a method of choice for various analytical applicat i ons Perhaps the biggest strength of the ion trap technique lies in its ability to perform multiple stages of mass spectrometry greatly increasing the amount of structural information obtainable for a given molecule. The types of fragment ions observed in an MS/MS spectrum depend on many factors including primary sequence, the amount of internal energy or charge state. The accepted nomenclature for fragment ions is shown in Figure 1-7. 77 78 Fragments will only be detected if they carr y at least one charge If this

PAGE 22

12 charge is retained on the N terminal fragment, the ion is classed as either a, b or c. If the charge is retained on the C terminal, the ion type is either x, y or z. A subscript indicates the number of residues in the fragment. In electrospray ionization, peptides generally carry two or more charges, so that fragment ions may carry more than one proton. Predominantly a, b and y ions are generated in ion trap CID experiments. In addition, peaks are seen for ions that have lost ammonia (-17 Da) or water (-18 Da). Quantification by Mass Spectrometry Given its ability to provide specifics on molecular masses and fragmentation information, mass spectrometry is predominantly used for qualitat iv e analysis. Quantitative methods have been underutilized and less explored, due to the variety of other analytical techniques available and routinely performed in analytical, clinical environmental, and forensic labs Mass spectrometry has been used to quantify drug metabolites,79 80 sugars,81 and volatile organic compounds,8 2 84 but MALDI and ESI introduced new challenges to meet the need for quantifying proteins peptides and their metabolites Given that this is an application driven method de vel opment methods are usually specific for a particular purpose. Further improvements in this field can be expected upon automatization and additional instrumental developments. Different quantification approaches have been developed using MALDI and ES ionization techniques, some of which will be discussed in more detail later. While MALDI MS analysis is tolerant to some impurities and reliable for mixture anal y sis the lack of reproducibility restricts its performance as a reliable quantification technique. Large error is introduced due to heterogeneity of the sample distribution and matrix la yer, signal suppression, and non-linear detector response causing spot-to-spot and shot-to shot irreproducibility. After many attempts, it has been shown that a highly homogenous

PAGE 23

13 sample preparation is required for a reliable quantitative analysis.85 88 Nevertheless MALDI-based quantification methods are still not routinely used and still are a challenge for researchers. ESI MS, on the other hand, is much more sensitive to presence of salts and other impurities, so most of the quantification attempts involve coupling with LC or other separation technique. A major problem of mixture quantification is the matri x effect of dominant components causing signal suppression and loss of reproducibility. Also mass resolution and scanning performance of the mass spectrometer do not always allow for reliable differentiation of two species similar in mass. Separation methods performed prior to mass spectrometric analysis can significantly simplify and improve the quantification analysis. By separating complex mixtures based on hydrophobicity size or immunological properties of the components, each one of the components can enter the mass spectrometer separately, minimizing the interferences. Coupling Liquid Chromatography (LC) to Mass Spectrometry Chromatography includes a diverse and important group of methods that permit separation of closely related compounds from complex mixtures The applications of chromatography have grown explosively in the last few decades not only because of the development of several new types of chromatographic techniques but also due to a growing need for better methods of characterizing complex mixtures Besides becoming a premiere method for separating closely related chemical species LC can be employed for qualitative identification and quantitative determination of separated species. A chromatogram provides only a single piece of qualitative information about each specie in a sampleits retention time. Even if it is widely used as a tool for analyte detection and determination, the confirmation of identity requires a spectral or ch e m i cal

PAGE 24

14 investigation of the separated compounds. Quantitative chromatographic analysis is based on a comparison of either the height or the area of the analyte peak with that of a standard. Peak areas are independent of broadening effects and therefore are more satisfactory analytical parameters than the peak heights The highest precision in quantitative chromatography is obtained by use of internal standards. A measured quantity of an internal standard is introduced into each standard and sample, and the ratio of analyte to internal standard peak areas serves as the analytical parameter. It is required for the internal standard peak to be well resolved from the peaks of other compounds but still to appear close to the analyte peak. The origins of combining LC with MS can be traced to the 1960s 8 9 The main challenge of such a coupling is in introducing a liquid sample into the vacuum of a mass spectrometer, but utility of such a combined analysis still drives research towards further advancements of the methods One of the main advantages that ESI offers over MALDI is a possibility of a direct liquid introduction It is relatively straightforward to directly introduce samples into the mass spectrometer while in liquid solution via electrospray It is more difficult to couple MALDI directly to liquid separations because samples have to be dried on a solid surface prior insertion into the mass spectrometer. Therefore, MALDI can serve as a useful off-line detector, which offers the advantage of an independent mass analysis.90 94 However, several attempts to develop on-line coupling of liquid separations have been made, but they required instrumental changes and redesign. 95 97 As MS is a concentration dependent technique, using the LC prior the MS anal y sis is beneficial in many ways; besides the fractionation of complex mixtures and salt removal, LC also preconcentrates every analyte in t he effluent. Chromatographic

PAGE 25

15 separation of biomolecules can be based on different properties of the molecules, their size, charge, hydrophobicity, or binding affinity. Reverse Phase LC Two types of partition chromatography are distinguishable based upon the relative polarities of the mobile and stationary phase. Early work in liquid chromatography was based on highly polar stationary phases such as water supported on silica or alumina particles, while a relatively nonpolar solvent, such as hexane served as the mobile phase For historic reasons this type of chromatography is now referred to as normal-phase chromatography In reversed-phase chromatography, on the other hand, the stationary phase is nonpolar often a hydrocarbon, and the mobile phase is relatively polar, such as water, methanol or acetonitrile. Therefore, the most polar component appears first, and increasing the mobile phase polarity increases the elution time. The most commonl y used packing for this type of chromatography contains a C8 (n-octyl) or a Cl8 (n octadecyl) chain attached to the siloxane coatings. The elution is usually carried out with a highly polar mobile phase such as an aqueous solution containing various concentrations of solvents such as methanol or acetonitrile. Capillary LC The ability to work with minimal sample si z es, at small flow rates, and the enhanced detection performance with the use of concentration sensitive detectors due to preconcentration are some of the main reasons for increased popularity and development of capillary LC. Miniaturization of LC separat i ons started in the 1960s with Horvath s work ,98 99 but became significantly more exploited as new mass spectrometric techniques especially ESI, were developed in 1980s leading to low detection limits as well as specificity in characteri z ation

PAGE 26

16 The reduction of column i .d. to the capillary size (down to 20 m i.d.) introduced new requirements for the instrumentation applied, but offered the advantage of MS compatible flow rates, so that the LC effluent could be directly introduced into a mass spectrometer. It was also shown that inhomogeneities in mobile phase flow paths are reduced due to the more uniform packing structures within smaller i.d. packed columns, therefore reducing peak dispersion.100 Additionally, this results in more uniform retention times and reduction in column band-broadening. Th UV d 1 d fi -11 LC 101-104 Fl 10s 106 e etector 1s sh common y use or cap1 ary uorescence and electrochemical detection107-109 are also employed Nevertheless, mass spectrometry, with all its advantages is the most common for detection and characterization of LC effluent. Mass spectrometric methods are commonly coupled on-line (ESI-MS) or off line (MALDI-MS), offering great sensitivity, minimal sample demand and full characterization of the effluent. Neuroscience All the sensations, feelings, thoughts, motor and emotional responses learning and memory, the actions of psychoactive drugs, the causes of mental disorders, and any other function or dysfunction of a brain cannot be understood without the knowledge about the processes occurring during communication between neurons. As far as we understand the process now, the neurotransmitters are the messengers that are released in the synaptic cleft and induce the signal to travel to the next nerve cell (Figure 1 8). Different types of cells secrete different neurotransmitters. There are many classes of molecules that can play that role, usually grouped into cholines, biogenic amines amino acids, or neuropeptides

PAGE 27

17 During the last decade, it has become evident that mass spectrometry is a k e y analytical tool in neuropeptide research For decades before that less specific methods based on immunologic or structure specific assays, were used for neuropeptide characterization and quantification Some of the methods commonly used in neuroscience are radioimmunoassay (RIA), radioreceptor assay ( RRA) enzyme-linked immunosorbent assay (ELISA), chromatographic, and electrophore t ic separation wi th UV detection. The RIA method is based on the specific interaction between a given peptide (the antigen) and its antibody (Figure 1-9). Although it offers good sensiti v i t y and ease of operation, cross-reaction with similar structures and fragmen t s introduces a significant error in the measurements. Some improvements in accuracy and precision due to better selectivity are noticed when RIA is combined with a separation technique.110 -112 When using RRA, the neuropeptide activity is detected and a suitable membrane preparat i on containing the particular receptor is required This technique was used t o quantify t h e opioid peptide activity in CSF from patients with chronic pain or neurologi c disorders.113 '114 The error when using this method occurs since there are many different opioid peptides that contribute to the detected ac t ivity so it is unclear wha t particular opioid system is responsible for the observed changes Among a variety of MS-based strategies available for peptide analysis the most commonly used techniques are MALDI TOFMS and E SI-QITMS. These methods combined with separation techniques offer high specificity and great sensitivity, e v en when analyzing the biological samples directly.1 1 5 -117 Ho w e ver the s e arch for n ew

PAGE 28

18 endogenous peptides and their quantification 1s still a great challenge of mass spectrometry in neuroscience. N europeptides Neuropeptides are a large family of bioactive compounds participat ing m the transmission or modulation of signals between the nerve cells. They are involved in many neurologic functions, like food intake, memory, mood pai n reward stress and many others. They are present and can be sampled from central nervous sys t em (C N S) and body fluids. Cerebrospinal fluid (CSF), because of its accessibil i ty, offers a suitable medium for neuropeptide analysis. Synthesis of neuropeptide precursors occurs m ribosomes at a considerable distance from the secretory site of the axon terminal. Generally, the endogenous neuropeptides are synthesized as large, biologically inert protein precursors (prepropeptides). These precursors are subsequently cleaved throughout a series of proteolytic steps, in a very sequence-specific and tissue-specific order to generate t he active species. Other post-translational modifications, like amidation or phosphorylation have also been studied (Figure 1-10). Determination of peptide concentration gives no information w ith regard to turnover. Specific regional concentrations of pept i des may be grea t er than those described for the general area of central nervous system. This can significantly change upon stimulation of certain brain regions. Alterations in concentrations of brain peptides are being studied for their use as possible markers in a number of neurological diseases. Sampling Techniques One of the main limiting factors for reliable brain research is proper sampling needed for valid data interpretation. Advances in various in v i tro and in v ivo samplin g

PAGE 29

19 techniques have considerably improved research into nervous system processes. Recent improvements in chemical separations and mass spectrometry offer reliable chemical analysis for brain research. Many assumptions and compromises are made when processes in the living organisms are investigated, which can lead to misinterpretation of the analytical results. In vitro studies Many in vitro models have been used in the biological sciences for preliminary investigations, method development and various advanced methods in medicine, including the most disputed methods of cloning living organisms. The benefits of such approaches are enormous, since they minimize the use of living organisms for testing, help eliminate some methods in the early stages of development, and give a broad perspective for adventuresome and risky innovations. However, the results obtained by in vitro and in vivo techniques can differ considerably, which hinders the extrapolation of the in vivo results from the in vitro model. The major causes of these discrepancies occur due to the properties of various in vitro and in vivo models used. In vitro models can be designed according to experimental requirements. There are many different concepts of in vitro brain analysis. Certain parts of the brain can be studied separately, using the brain slices, primary cultures, or isolated brain microvessels. Cell cultures can be generated and engineered according to experimental requirements. One of the generally used approaches in brain analysis is to homogenize the brain tissue in a buffer solution.118 120 This method offers fast and easy screening results but can be deceiving due to contents of the cell lysate that will be present, with many enzymes which may disturb the environment. An alternative approach has been developed which requires careful slicing of the brain into thin layers followed by washing with buffers to

PAGE 30

20 remove any cell lysate contents.121 It is assumed that minimal damage to the cell is done, and that no intracellular contents are present in the sample In vivo studiesmicrodialysis Microdialysis is a sampling technique generally used for in vivo sampling, based on a size-selective diffusion of the analyte through a semipermeable membrane. Molecules smaller than the pore size may permeate into the perfusion medium following the concentration gradient, while molecules larger than the membrane cut-off (like proteins enzymes), are excluded (Figure 1-11). It was initially developed by Delgado and co workers in 1972, 122 but popularized by Ungerstedt and Pycock.123 The idea behind microdialysis was to mimic the function of blood vesse ls and to achieve in situ sampling and sample clean-up. It is performed by inserting a probe that contains a semipermeable membrane into a medium of interest (Figure 1-12). The membrane, permeable to water and small molecules, is continuously flushed by the physiological perfusion fluid, which is collected as dialysate, and enriched with endogenous substances. Microdialysis may also be used for local drug administration via the same probe at the same time as for sampling. The versatility of microdialysis sampling has resulted in its application in neurochernistry, pharmacokinetic studies, and biotechnology, mainly in vivo. In vivo monitoring is ideally preferred over other sampling methods, since it offers real life samples, as close as we can get with minimal disturbance of the regular environment of an organism. Microdialysis provides direct measurements in anaesthetized or freely moving animals analysis of specific target tissues or metabolites from enzymes in tissues, and continuous sampling with no liquid loss. It offers the possibility of local drug administration and determination of drug influence on metabolites. However, the dialysate is relatively dilute and contains high salt

PAGE 31

21 concentrations, requmng further sample treatment pnor the analysis. It is also an invasive procedure, and it causes (although minimal) neuronal death. The method offers limited spatial and temporal resolution, but is still the most commonly used due to the ease of application and a routine experimental design.

PAGE 32

22 Ta b le 1 -1. A table of the most used MAL D I matrices It can be noticed tha t all the m o lecules are aromatic an d have at least o ne acidic H. The aromatic or c o njugate d structures are UV chromo p hores resp o nsible for absorbing most of the UV irradiance, whi l e t h e fr p l ays an imp o rtant role in the analyte i on izati o n me c hani s m Matrix (abbreviation) Molecular Structure Nicotinic Acid (NA) ~COOH H3CO Sinapinic Acid (SA) HOO CH=CH-COOH ( 3 5 Dlmethoxy-4h y d r o xycinnami c a c i d ) H3CO HO DHBA QcooH ( 2 5-Dihydroxybenzo l c acid ) OH HPA Q-cooH (3-Hydroxyp lcoll n i c a c i d ) OH a-CHCA HO-< h (O'.-cyano-4 -hydro x ycinnamic acid) COOH NC

PAGE 33

Laser Beam Matrix with the Anal:yte 23 Toward the TOF Analyzer Electrostatic Field Figure 1-1. A schematic presentation of MALDI process. The laser initiates an explosion of the matrix / analyte plume, resulting in ioni z ed analyte directed toward the analyzer by the electrostatic field. (adapted from reference 124)

PAGE 34

24 MM MM r M* M* M* 2hv M**+ M -e +e H H ( M +Ht/ (M-Hj+ l (M-HJ-(M+Ht (M-2H)-" (M+2Ht Figure 1-2. A summary of the initial ionization processes in MALDI. The primary ionization results in activated matrix molecules (labeled M), which further initiate analyte ionization. The process is usually observed onl y in pos itive mode, but both mechanisms are expected to occur. (adapted from reference 13)

PAGE 35

25 Detector t1 t -arl 111 1 1 1 1 1 11 Ion Source 1 1 1 111111 , Figure 1 -3. A schematic presentation of the reflectron T O F. The reflectron allows ions with greater kinetic energies to penetrate deeper into the reflectron fiel d than ions with s maller kinetic energies. If a packet of ions of a certain m/ z ratio contains ions with varying kinetic energies the reflectron will decrease the spread in ion flight times and therefore improve the resolution of the T OF analyzer. (adapted from reference 53)

PAGE 36

26 Static, Reflector Delayed, Linear Delayed Reflector R:2600 R:4000 R:4600 RP ~R:3200 R:4400 R:5400 E l r# R:5800 R:6600 R:8600 X 11 ,. ..... ....1290 mh. 1300 1290 m/z 1300 1290 mlt. 1300 Figure 1-4. Comparison of signal in different types of TOF analyzers. Nine mass spectra of angiotensin I are shown, using linear, reflectron and dela ye d extraction modes of operation. Instrument geometries are labeled: RP EL, and XL, where flight distance increases in that order. ( adapted from reference 125)

PAGE 37

Ion Desorption Mod_ l a) Su ce elecir c eld nurts" analyte ion from droplet 27 Coul' o b Explosion Further SOfwt Evaporation and C ouh1m1b Explosions steps Sotuent IEvapo a n fiion a droplet containinu a single ana1Yte I OI CUI b) Figure 1-5. Schematic presentation of the ESI mechanism. Upon the initial droplet formation, different models are suggested to explain the final ion formation: a) ion desorption model (IDM), suggesting the ion "escape" due to the electric fields present, and b) charge residue model (CRM), indicating the continuous droplet evaporation as the ion is formed.

PAGE 38

28 Ions End~ap electrode ----~ Resonance AC voltage Figure 1-6. Schematic presentation of a quadrupole ion trap. The ions are selected and ejected according to their m/z value by scanning the voltage on the electrodes ( adapted from reference 126)

PAGE 39

29 ---1 X3 Y3 Z3 X2 Y2 Z2 X1 Y1 21 : H+ r-r-rrr-r-rrr-1 I I I I I I I I R11 0 I I R21 0 I I ~I O I I I I II I I I I 11 I I I I II I I I H2N-C-1-c-1-N-1 c-1-c-1-N-1-c-1 C-1 N-1-C-COOH I I I I I I I I I I I I I I I I I I I I I I I I I H1 1H1H1 1H1H1 1H1H I I I I I I I I I __ I __ I __ I __ I __ I __ I __ I __ I __ I a1 b1 C1 a2 b2 C2 b3 C3 Figure 1-7. Peptide ion fragmentation upon Collision Induced Dissociation (CID) experiments. For low energy collisions, like those in a QIT fragmentation, the peptide bond is cleaved and y and b type are predominantly formed ions.

PAGE 40

30 Figure 1-8. A presentation of a synaptic cleft. The neuropeptides are secreted as the signal is given to the vesicle. Upon secretion in the synaptic cleft the y bind to the corresponding receptors on the postsynaptic neuron, while some of the material is reabsorbed by the presynaptic neuron. ( adapted from reference 127)

PAGE 41

REAGENTS: Ab specific for hormone (coat ing the filter) Unknown sample w ith hormone All"Ow time to react Wash away unbound substances R EAGENTS: 1 251-labele.d hormone A ll-0w t ime to react Wash away unbound rad iol<1~eled hormone PROCEDURE: measure rad i oact v ity I n
PAGE 42

Degradin 32 Bio synthesis Pre-propeptide Prohormone Propeptide Prohormone convertases Active neuropeptide I Converting Enzymes Active fragments Degrading Enzymes Inactive fragments or amino acids I ______ ___, Figure 1-10. The pathways of neuropeptide synthesis. After a protein is synthesi z ed in the ribosomes, the propeptides are stored in the neurons until the enzymes start converting them into the active neuropeptides. The degrading enzymes are also involved, converting the peptides into inactive fragments and amino acids

PAGE 43

33 Dental plastic Fluid Is pumped through inner cannula Skull Brain I Substances from extracellular fluid diffuse through the dialysis tubing Figure 1-11. A schematic presentation of microdialysis. Microdialysis probe allows diffusion of small molecules into and out of the probe. Therefore this sampling technique can be used to observe consequences of the perfused compound, or to sample the environment for characteri z ation of the endogenous substances. (adapted from reference 129)

PAGE 44

34 Input Perfusion L iquid 1 ,,. <>tier Tube (400 m) Inner Tube (150 m) Tissue Figure 1-12. A presentation of a microdialysis experiment and the microdialysis probe The microdialysis probe is placed into the brain region of in t erest. The sampling is performed on an alive and usually awake animal. (adapted from references 130, 131)

PAGE 45

CHAPTER2 INVESTIGATION OF SAMPLE PREPARATION METHODS TO IMPROVE MALDI-TOFMS PERFORMANCE AND FACILITATE INTEGRATION OF CAPILLARY LC SEPARATIO N S Introduction MALDI Sample Preparation Considerations Sample preparation has a significant influence on the sensitivity and reproducibility of MALDI-MS 132133 Therefore, various sample preparation protocols have been developed to improve the homogeneity of matrix crystals and the distribution o f sample within a deposited spot. The influence of different matrices, sol v ents, pH, counter ions, and salts present or added were investigated using such deposition techniques as dried droplet, thin layer, spin-dry, the sandwich method and fast evapora tion .1313 Selectin g a matrix for these techniques is considered to be somewhat of an empirical process; however, guidelines for selecting matrices for various classes of molecules are avai lable (for instance, 4-hydroxy-a-cyanocinnamic acid (a-CHCA) is most commonly used for peptide analysis).117 1321 36 Solvent selection is dependent on both matrix solubility and the sample deposition method used; variations in crystal formation are known to influence ionization distribution, and the incorporation of analyte into the matrix layer. Sadeghi, et al., ha v e demonstrated that ionization efficiency is directly dependent on crystal si z e due to the predominant volatilization of smaller crystals by the laser beam.137 Homeffer, et al,. 1 38 have observed that incorporation ( co-crystallization) of analyte molecules is not necessary for all matrices; however, the efficiency of the MALDI process is im pro ve d 35

PAGE 46

36 with a more homogeneous distribution of matrix crystals. 139 A detailed mapping of conventionally deposited spots has revealed significant heterogeneity in analyte distribution thought to be caused by uneven crystallization within the spot.140 Since spot homogeneity and ionization efficiency are improved with the formation of smaller crystals, 137 we investigated methods for reducing crystal size to further improve the MALDI-MS technique as presented herein, which in summary, focused on two basic goals: minimizing crystal size for improved homogeneity of crystal distribution, and minimizing sample spot size for improved detection limits. Matrix crystal size Vorm, et al. 141 proposed the fast evaporation method to reduce crystal size Matrix dissolved in a fast drying solvent forms a primary layer of small uniform crystals upon which the analyte solution is deposited and incorporated into the matrix layer. An additional improvement in sample distribution homogeneity was observed when analyte was mixed with matrix and deposited as a second layer.142 Electrospray deposition (ESD) of the matrix can be used to further reduce solvent evaporation time, producing even and small crystals.143 1 4 4 A two layer approach for matrix and subsequent analyte deposition can also be employed as it is hypothesized that small crystals formed in a first layer seed the formation of small crystals in all successive layers.141 143 While electrospray is a well-established ionization technique for generating charge on liquids and enabling their introduction into mass spectrometers, the same experimental concept has been successfully employed for the deposition of fine, uniform layers of particles as early as the 1950s.145146 Small droplets are ejected from the Taylor cone depositing particulates evenly onto the counter electrode due to the fast evaporation of

PAGE 47

37 solvent (see Figure 2-1). Using electrospray deposition for MALDI-MS analysis has shown improvement in spot-to-spot and shot-to-shot reproducibility due to the even ly sized small crystals formed.143 144 In addition, the electrospray deposition method has been successfully utilized for fabrication of protein films while preserving prote i n reactivity when re-dissolved.147 150 This approach for storing biomolecule activity is beneficial in conserving sample concentration, activity, or information. Sample spot size Minimizing sample spot size, thereby increasing analyte surface density, introduces an improvement in detection limits The laser conventionally irradiates less than 1 % of the sample spot size making a search for a "sweet spot" critical to successful MALDI. Therefore, it is expected that concentrating the analyte into an area the size of the laser spot should improve sensitivity. Changing ESD parameters can effectively minimize spot size. Capillary i d electrospray voltage, emitter to plate distance, and flow rate can all be adjusted to achieve smaller spots and have been investigated in this work. ESD is also a non-contact deposition method offering the possibility of depositing the matrix layer first independently of the analyte or a matrix/analyte mixture second. A competing approach to sample spot size reduction, in order to focus the analyte into an area comparable to the laser spot size (-150 m in diameter), was the development of miniaturized sample holders, termed AnchorChip TM plates by developers at Bruker Dalton ics. 134 135 This can be accomplished with the use of a hydrophilic spot area defined by a hydrophobic surface. When AnchorChip plates are used crystallization of aqueous samples occurs in this hydrophilic spot, concentrating the

PAGE 48

38 material into a 200 m diameter region.134135 The "sweet spot" search is eliminated, wi t h all the material focused into the laser ablation area. Problems with shot-to-shot reproducibility are diminished and sensitivity is increased given irradiation of the total amount deposited. Capillary LC/MALDI Coupling Reverse-phase capillary liquid chromatography (LC) is commonly used for separation of peptide mixtures and for cleanup of salts and other contaminants that suppress the ionization process. Additionally, sensitivity and selectivity of mass spectrometric analysis are expected to increase due to pre-concentration and separation of the analytes when using capillary LC. Interfacing LC with MALDI-TOFMS requires transporting analyte from liquid effluent into dry matrix crystals, which can be developed for either an on-line or off-line mode of interfacing separations with mass spectrometry. On-line coupling Coupling LC to MALDI-TOFMS on-line is particularly challenging because it requires delivery of the LC effluent directly into the vacuum of a mass spectrometer. On line coupling can be performed by continuous flow,95 9 6 or aerosol introduction into the mass spectrometer.97 The LC effluent is directly introduced into the MALDI mass spectrometer either on a moving belt or by spraying into the vacuum as aerosol. Both of these approaches require major instrumental adjustments, which is probably the main reason why they were never extensively used Off-line coupling Off-line coupling involves the collection of fractions into v ials, or directly onto a MALDI plate-described by Caprioli 's group as "on-line transfer off-line analysis".90 This requires matrix introduction by either depositing the matrix prior to sample

PAGE 49

39 deposition,9094 or by introducing the matrix directly into the effluent either through a tee union, 151 152 a sheath flow, 153 or by adding it directly to the mobile phase.154 Caprioli and coworkers have automated dried droplet deposition, commonly used when analyzing sample fractions, using a motorized XYZ stage that rosters sample spots across a MALDI plate.9 For this technique, spots can be deposited by bringing the plate into contact with a sample needle; however, to avoid such direct contact, nebulizing through a microdispenser 91 92 94 or electrospraying the sample has been suggested.152 155 Detection limits using these techniques range from low femtomoles to high attomoles.90 153 154 Electrospray deposition has also proven to be a successful way to deposit material for further MALDI-TOFMS analysis, and since it can continuously deposit liquid onto a solid support, it was expected to be a good intermediary for coupling liquid chromatography with MALDI mass spectrometry. The flow rates and equipment used for such methods are compatible with capillary LC; however, the change in solvent composition during a gradient LC run can present a potential problem for the stability of the Taylor cone of the electrospray. This can be overcome by ramping the electrospray voltage during the run.156 157 Imaging Techniques A few imaging techniques were employed in this research to investigate matrix crystal and analyte spot size. Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) were used for measuring changes in crystal size and crystal distribution as a function of solvent and sample deposition method. Fluorescence imaging was used to observe distribution and incorporation of analyte into the matrix layer. A fluorescent dye (rhodamine 6G) was added to the analyte solution to enable the visualization of the spots.

PAGE 50

40 Scanning electron microscopy (SEM) The Scanning Electron Microscope (SEM) uses electrons rather than light t o form an image. The SEM is designed for the direct study of solid objects' surfaces. By scanning with an electron beam generated and focused by electromagnetic lenses secondary electrons are generated from the sample s surface. The image is produced from counting the number of electrons emitted from each spot on t he sample. As a resu l t SEMs have a greater depth of focus than optical microscopes and subsequently very good resolution; however, samples have to be conductive. If not the y must be coated with a thin layer of gold by a sputter coa t er prior to analysis. Transmission electron microscopy (TEM) In transmission electron microscopy (TEM), a thin sol i d specimen is bombarded in vacuum with a highly focused, monoenergetic beam of electrons having sufficient energy to propagate through the specimen, be collected and magnified to produce an electric signal. Diffracted electrons are observed in the form of a diffraction pattern beneath t h e specimen. Transmitted electrons form images from small regions of the sample that contain contrast due to several scattering mechanisms associated w ith electrons and atomic constituents of the sample. High resolut i on images of the morpholo gy or topography of a specimen can be achieved by scanning an electron beam across a specimen Due to the shorter wavelength of electron beams, 0.2 nm resolution can be achieved, compared with 200 nm when a light microscope is used Fluorescence imaging In order to differentiate objects with similar optical propert i es t h at are irresolvable by optical microscopy, fluorescence properties can be used. A fluorescence microscope is basically a conventional light microscope supplem e nted by an e x ci t ation light sour c e

PAGE 51

41 and an array of filters. Much higher intensity light is required to illuminate the sample, which in turn emits light of a longer wavelength. The image from the microscope is viewed by eye or can be digitally collected via a CCD camera. Experimental Methods Instrumentation Mass spectra were collected with a Voyager DE-PRO MALDI-TOFMS (Applied Biosystems, Framingham, MA) and a Bruker Reflex II MALDI-TOFMS (Bruker, Bremen Germany), each equipped with a nitrogen laser (A=337 nm) and operated at a pulse rate of 3 Hz. Analysis was performed in reflectron mode using delayed extraction (PE, Applied Biosystems) or pulsed ion extraction (PIE, Bruker). Positively charged ions were accelerated to 20 kV, and 100 to 300 single shot spectra were averaged. All the other parameters varied depending on the experiment and the instruments used, and are discussed in detail for each experiment. An Agilent 1100 series binary pump system (Agilent Technologies, Palo Alto, CA) was used for preliminary off-line LC/MALDI experiments. ISCO model 1 00DM syringe pumps (ISCO Inc Lincoln, NE) were used for all advanced LC/MALDI experiments. A variable wavelength UV detector, model TSP 1000 (Thermo Separations Products, San Jose, CA), was used throughout for chromatography method development. Electron microscopy was performed in the Major Analytical Instrumentation Center (MAIC) of the University of Florida. Scanning electron microscopy was performed with a JEOL SEM JSM-6400 (Figure 2-2), while transmission electron microscopy was performed on a JEOL TEM 200CX (Figure 2-3), (JEOL USA, Inc., Peabody, MA). Fluorescence microscopy was performed in the laboratory of Dr. Weihong Tan using an inverted microscope OLYMPUS 1X70-S1Fl (Olympus America

PAGE 52

42 Inc. Melville, NY), an argon laser INNOV A 3000 (Coherent Inc., Santa Clara, CA), and a CCD camera (Princeton Instruments, Trenton, NJ). Chemicals and Reagents a-cyano-4-hydroxycinnamic acid ( a-CHCA) and all synthetic neuropeptides ( oxytocin, bradykinin, Arg8 -vasopressin, and neurotensin 1-11) were obtained from Sigma (St. Louis, MO). Solutions were prepared with HPLC grade solvents (Burdick and Jackson, Muskegon, MI). Mobile phases were filtered with 20-nm-pore size aluminum oxide filters (Fisherbrand, Fisher Scientific, Pittsburgh PA) t o prevent capillary clogging. Trifluoroacetic acid (TF A) acetic acid and hydrofluoric acid were purchased from Fisher Scientific (Fair Lawn, NJ), while helium gas tanks were provided by PraxAir Inc., (Danbury, CT). Electrospray Deposition Method (ESD) ESD apparatus Two different electrospray arrangements were used throughout this work. The first used a pressure bomb for flowing liquid in all preliminary investigations with the application of a high voltage to a gold wire placed at the capillary / emitter junction. A MALDI plate was placed horizontally opposed to the ESD emitter, and was grounded to complete the electrical circuit (as shown in Figure 2-4). An advanced electrospray device was constructed based on preliminary results employing a syringe pump, fused silica capillary, and a pulled capillary electrospray emitter (see Figure 2-5) The electrospray voltage was applied directly to the syringe needle with the MALDI plate grounded. A stepper motor was added to automate plate translation in one direction and a camera was placed behind the electrospray emitter to enable vi sualization of the electrospray process, which helped maintain electrospray stability. Matrix and analyte solutions were

PAGE 53

43 electrosprayed through a capillary emitter. The flow rate, applied v oltage, and electrospray tip to plate distance were optimized for achieving the smallest matrix crystal size and an optimal analyte spot size. Electrospray tip production A flat-cut capillary tip was used as an electrospray emitter m preliminary experiments. Improved electrospray emitters were introduced in the advanced setup, made by pulling fused silica capillaries to a fine point using a CO2 laser puller (Sutter, Narato, CA) followed by etching in 50% HF for 3-5 min. The capillaries were 50 or 25 m i.d. and 360 m o.d, where the former ones were used for electrospraying matrix solution and the later once for peptide solution electrospray. The methodology of capillary pulling is illustrated in Figure 2-6, with an SEM image of the capillary emitter shown in Figure 27. Hydrophobic Sample Supports (AnchorChip Plates) Introduced by Gobom et al., 132 AnchorChip TM plates were investigated to decrease analyte spot size. These commercially available stainless steel plates were coated with Teflon leaving untreated spots of 200-600 m open as hydrophilic surfaces for sample deposition. The plate with 200 m diameter spots was used for all the experiments since it was the most similar to the size of the laser spot and it concentrated the material the most. LC/MALDI Capillary column packing Capillary columns were prepared in the lab. Differen t lengths of 150 m i.d ./ 360 m o .d. capillary were packed with a C18 stationary phase. An acetone slurry (lOm g/ ml)

PAGE 54

44 of 5-m Altima Cl 8 reversed-phase particles (Alltech, Deerfield, IL) was pushed through the column by 1000 psi pressure of helium from a pressure bomb An in-line filter with a sintered silica disk (Upchurch Scientific, Oak Harbor, WA) was used to prevent clogg i ng of the columns. Off-line method system For separation of a standard peptide mixture, a 2 5 cm section of fused-silica capillary (150 ,m i.d.) was packed with 5-,m C18 reversed-phase particles Samples were injected with a 5 ,l sample loop (Valeo, VICI Valeo Instruments Co. Inc., Houston, TX), as illustrated in the schematic shown in figure 2-8. The linear gradient was set from 95% N5% B to 5% N95% B in 8 minutes), at the flow rate of 8 I/min (eluent A consisted of 1 % CH3COOH in H20, eluent B consisted of 1 % CH3COOH in MeOH. The UV absorbance chromatograms were collected using a LabView program (National Instruments). Sixteen one microliter effluent fractions were manually collected over the entire separation time from the capillary outlet into Eppendorf tubes already containing 2 L of matrix solution. Tubes were kept in an ice bath to prevent evaporation of t he solutions. On-line collection, off-line analysis system The schematic of this experiment is shown in figure 2-9. A 6 cm long capillary column was prepared as described above Identical solutions and mobile phases were used as in the offline separations, but effluent was directly deposited onto a MALDI target plate that was coated with the matrix layer prior to analyte deposition

PAGE 55

45 Results Minimizing Matrix Crystal Size Optimization of ESD parameters Solvent selection had to meet requirements for matrix solubilization and electrospray stability. Mixtures of methanol and water with an acid modifier are usually successful in most electrospray experiments, but a-CHCA was insoluble in these mixtures. Instead, the matrix was soluble in acetone/0.1 % TF A, which was amenable to electrospray requirements, but the high volatility of acetone was a problem since the matrix solution would evaporate faster than it could be electrosprayed. This caused a buildup of crystallized matrix on the capillary tip subsequently clogging the electrospray emitter. Therefore, instead of using a flat cut silica capillary (50 ,m i.d., 360 m o.d.), a new pulled capillary emitter etched with 50% HF to open a small orifice was used to alleviate this problem. The new emitter was beneficial in two ways: 1) less liquid was sprayed into the air since the orifice was smaller than the capillary i.d. (usually -20 m), and 2) the excess solution (which was still present) was being pulled backwards, so it wouldn't clog the emitter as fast. Even after this improvement, the emitter still had to be washed with MeOH every 15 min to avoid clogging. In preliminary experiments, a pressure bomb with helium gas was used to control flow through the capillary. It was later shown that a syringe pump was easier to use and provided more reliable control of the flow rate. An optimized flow rate was found as a compromise between avoiding clogging and providing sufficient liquid for a stabile electrospray. The most reliable electrospray operation was seen when the following parameters were used : a saturated matrix ( a-CH CA) solution in acetone/0.1 % TFA (10 mg/ml) electrosprayed through a 50 m i.d. capillary with a 2Q m i .d. emitter at + 3 kV

PAGE 56

46 and a flow rate of 500 nl/min The distance from the emitter tip to the plate was set at 1 mm, producing matrix spots of 5 mm in diameter. An optimized matrix layer thickness was achieved by electrospraying each spot for 5 min. Electrospray disturbances were observed by use of a camera and adjustments were made to maintain a stable electrospray A stepper motor facilitated plate translation in one direction, enabling organized deposition of spots and the option to deposit tracks of matrix and analyte. This principle would enable automated fraction collection of the LC effluent in LC/MS experiments. Matrix crystal images Crystal structure and distribution were characterized by scannmg electron microscopy (SEM). The differences between electrosprayed matrix and matrix deposited by conventional techniques are evident in Figure 2-10. A "network" of small matrix crystals deposited by electrospray showed increased homogeneity when compared to the same matrix deposited by dried droplet method The size of the electrosprayed crystals was confirmed by transmission electron microscopy (TEM) (see Figure 2-11) and determined to be in 100 nm in diameter. Minimizing Analyte Spot Size Optimization of ESD parameters It was found that 50% MeOH/0.1 % TF A in H20 was an optimal solvent mixture for dissolving and electrospraying peptides. Peptide solutions were electrosprayed at a flow rate of 100 nl/min through a 25 m i.d. capillary with a -5 m i d. pulled emitter held at 1.2-1.3 kV. Deposition was made directly onto a matrix layer in 10, 30, and 60second intervals at a distance of 0 25 mm from the target. The spots were analyzed by

PAGE 57

47 imaging techniques to determine their size and the distribution of the peptide solution within the deposited spot. Analyte spot imaging The conditions used produced a spot of -200 ,m in diameter (see Figure 2-12). Even smaller spots could be made (Figure 2-13), but these were difficult to visualize when acquiring data with the MALDI-TOF instrumentation (spots -100 nm in diameter were not observable). On the other hand, results showed that under higher flow rates or longer deposition times the electrosprayed solution would start redissolving/relocating the matrix, which then migrated toward the edge of the formed "crater thus redistributing the analyte (methanol present in the electrospray solution redissolved the matrix). This can be a limitation for analyzing larger volumes and on-the-spot concentration of the sample. Matrix removal was observed in SEM imaging experiments (Figure 2-14) and confirmed by fluorescence imaging after fluorescence dye (rhodamine 6G, soluble in MeOH) was added to the peptide solution in the ,M concentration range to enable visualization of the electrospray pattern by wide-field fluorescence microscopy (see Figure 2-15). To overcome matrix removal by electrospraying an analyte solution on top of the matrix layer, these approaches were investigated: 1) the peptide mixture was premixed with the matrix solution (1: 1 v/v) and electrosprayed together; 2) the peptide mixture was premixed with the matrix solution and electrosprayed on top of a previously electrosprayed matrix layer; and 3) the peptide mixture alone was electrosprayed on top of an electrosprayed matrix layer. It was shown that the least noise through interference peaks and thus the highest SIN was present when electrospraying the peptide solution alone on top of a previously electrosprayed matrix layer Low attomole detection limits

PAGE 58

48 were achieved by depositing 15-90 nl of the analyte at a concentration of 0.7 nM (see Figure 2-16) AnchorChip TM plates Solution preparation was altered due to concentrating effects of all the material deposited into a small spot. The optimal matrix solution (0.1 mg/ml in 50% MeCN/50 % H20 / 0.1 % TF A) was diluted 1: 100 compared to conventional sample preparation. The peptide solution was prepared in 50% MeOH/50% H20 /0. l % TF A and mixed with the matrix solution in a 1 : 1 v / v ratio A one microliter aliquot of this mixture was deposited onto a 200-m diameter AnchorChip T M spot using a gel-loader pipette tip. Spot size and matrix crystal distribution were determined by SEM (Figure 2-17). In comparing deposition of analyte alone with the deposition of an analyte / matrix mixture onto an already deposited matrix spot, no significant difference was observed in both sensitivity and reproducibility (Figure 2-18) Therefore dried droplet deposition can be alternated with the two-layer sample deposition, according to experimental demands. Detection of five attomoles was achieved by depositing 0 5 ,L of a 10 pM solution (see Figure 2-19). Capillary LC/MALDI-TOFMS Preliminary investigations To evaluate the integration of MALDI-TOFMS with capillary LC the influence of different MeOH/H20 effluent ratios on the mass spectrometric response was studied. The result showed some difference in the detection of peptides within the range of 2575% MeOH (see Figure 2-20). The variation in signal intensities can be attributed to crystallization differences under different solvent conditions. Also increased irreproducibility when approaching detection limits can be expected The important point proven by this ex periment is that all the peptides can be detect e d throughout a wide

PAGE 59

49 difference of solvent composition, since we were not certain at what point in the gradient each peptide was going to elute. Preliminary LC/MALDI-TOFMS experiments were performed usrng the experimental design shown in figure 2-8 Flow rate was adjusted to minimize band broadening while maintaining the required peak width for fraction collection. The optimal LC flow rate was determined to be -10 ,I/min with an optimal gradient ramp over 8 minutes. A UV detector was attached to the column outlet and used with concentrated standards to optimize separations conditions on-line. The limits of detection for the UV detector were in the ,M concentration range. The wavelengths of 214 nm and 280 nm were investigated for peptide detection. The former one was selected for its chromophores being more universal (absorbs the peptide bond). After confirming chromatographic separation of the standard mixture, experiments interfacing LC with MALDI-MS off-line were performed. Two l,l aliquots of each fraction were deposited onto the anchor chip plate (200 ,m diameter spots) Mass spectrometric data collected from each of the two spots were averaged for each fraction MALDI-TOFMS data confirmed separation of the neuropeptides (see Figure 2-21). The results indicate that proper interfacing of capillary LC/MALDI-TOFMS has been achieved. On-line collection, off-line analysis data The interface of capillary LC and MALDI-TOFMS was made more efficient by direct deposition of the LC effluent onto a MALDI plate that had been previously pre coated with a matrix layer (see Figures 2-9 and 2-22) Chromatographic separation was improved by optimizing gradient settings, flow rate and using a 6 cm long packed capillary column. This resulted in 30 s peak widths, detected when a 1 ,M peptide solution was tested (Figure 2 23). At the optimal flow rate of 6 ,I/min, 1 ,l fractions

PAGE 60

50 were deposited every 10 seconds from the capillary onto the matrix spots. The resolution of a 100 nM solution of the four peptides by the capillary LC and analysis of each fraction by MALDI-TOFMS is shown (Figure 2-24). While the UV detector couldn't trace the peptides, the MALDI-TOFMS revealed the distribution of the chromatographically resolved peptides. Conclusions and Future Directions Both electrospray deposition and use of AnchorChip T M plates resulted in increased sensitivity of MALDI-TOFMS due to analyte material being focused into the laser ablation area. Attomolar amount of peptides from picomolar solutions can be routinely detected. Also improved homogeneity of matrix crystal size and distribution was achieved by use of a developed electrospray deposition method due to fast solvent (acetone) evaporation and even distribution of droplets over the targeted surface. Optimal electrospray conditions were established for finest matrix crystal distribution, required thickness of the electrosprayed matrix layer and the most advantageous analyte spot size. Minimizing sample spot size resulted in improved sensitivity with detection of attomoles of peptide The possibility of coupling MALDI-TOFMS with capillary LC was demonstrated by "on-line transfer and off-line analysis" of a standard peptide mixture that was successfully separated and detected by MALDI-TOFMS. The UV detector used was not sensitive enough to monitor the separation. It would therefore be useful to monitor the separation with a fluorescence detector having greater sensitivity. Many possible derivati z ation methods e x ist for attaching a fluorescencently active group to a peptide ; the method would simply require one of these as an extra step. One obstacle to this approach might be analysis of the complex mass spectra of derivati z ed samples.

PAGE 61

51 Also, since the method was being developed for analysis of neuropeptides sampled from the brain, employment of the technique for that purpose would be the best criteria for confirming its validity.

PAGE 62

+ H i & h v ol t age power suppl y O x i dation T C 52 R educ t ion Electrons Figure 2-1. Process of electrospray ionization of positi v ely charged ions. The electric field penetrates into the liquid thus enriching the solution at the tip surface with possible charge carriers forming a Taylor cone. Drops with excess positive charge erupt from the cone and travel toward the counter electrode. Fast e v aporation and fission of droplets produce free ion species. (adapted from A.T. Blades et. al, A nal. Che m 1991, 63 2110-6)

PAGE 63

53 Figure 2-2. SEM-JEOL JSM 6400. This instrument was used for recording the scannin g electron microscopic images of matrix crystals and analyte spot size T h e instrument was available at the Major Analytical Instrumental Cen t er (MIAC) ofUF.

PAGE 64

54 Figure 2-3. TEM-JEOL 200CX. This instrument was used for transmission electron microscopic imaging of matrix crystals The instrument was available at the Major Analytical Instrumental Center (MIAC) ofUF.

PAGE 65

55 Pressure bomb HV supply -ES emitter MALDI plate~~~~ Figure 2-4. Electrospray (ES) apparatus used for preliminary experiemnts. Solutions were pumped through a capillary using a pressure bomb and helium pressure tank. An ES emitter was placed above the MALDI plate that was grounded and attached an XY stage. Fused silica capillary MALDI p~te (50 pm i.d.) \ X tra~lation stage~ Syringe pump j~V Electrospray needle YZ translation stage Figure 2-5 The advanced electrospray deposition setup. A MALDI target plate was mounted on an YZ translation stage while the electrospray emitter moved back and forth orthogonal to the plate. High voltage was applied directly to a syringe needle as the solution was pumped through a capillary to the electrospray emitter.

PAGE 66

56 Silica capillary CO2 laser Figure 2-6 Schematic presentation of a capillary puller. A CO2 laser was used to melt fused silica, while the capillary was being pulled in opposite direction (adapted from reference 158) Figure 2-7. SEM image of an electrospray emitter. After pulling and etching the capillary tip in HF the emitter had an orifice smaller than the capillary i.d. and was conical in shape. (taken from reference 159)

PAGE 67

Sample A = Gradient pump B 57 Capillary LC column UV detector Ice Bath Figure 2-8. Preliminary LC-MALDI set-up Samples were injected through a 6-port valve onto a 5 l sample loop. A UV detector was placed post-column with fractions collected into eppendorftubes containing 2 l of matrix solution.

PAGE 68

58 Sample Loop Gradient pump Capillary LC column Detector \ AnchorChip TM Plate Figure 2-9. The LC-MALDI set-up for on-line collection off-line analysis experiments After injecting the samples through a 6-port valve and detecting the analytes with a UV detector the fractions were directly deposited onto a MALDI plate having previously been coated with a matrix layer.

PAGE 69

59 Figure 2-10. SEM images of matrix (a-CHCA) crystals. a) deposited by dried droplet method, from a MeCN solution, allowing long crystallization time and crystal aggregation, b) deposited by ES method from an acetone solution providing for fast evaporation and small crystal formation.

PAGE 70

60 ( Matrix (CHCA) crystal Figure 2-11. TEM image of electrosprayed matrix crystals Carbon grid in the background is a support for the electrosprayed crystals. Crystal size averaged 100 nm.

PAGE 71

61 Figure 2-12 SEM image of the analyte spot size. The analyte spot was observed as a shaded region toward the image center and is approximately 200 m in diameter.

PAGE 72

62 Figure 2-13. SEM image of the analyte spot size. The analyte spot in the middle had a diameter of less than 100 m

PAGE 73

63 Figure 2-14. SEM images of matrix removal by electrospraying analyte solution a) beginning of matrix removal after electrospraying the analyte solution for 30 s b) total removal of the matrix layer after electrospraying the analyte solution for 3 min

PAGE 74

64 Figure 2-15. Fluorescence images of matrix removal by analyte solution deposition. Fluorescent dye was added to the peptide solution, but some background matrix fluorescence was also noticeable. a) after 10 s of analyte deposition the matrix layer is still intact, b) the relocation of the matrix layer by the analyte solution is noticeable after 1 min of analyte deposition

PAGE 75

20,000 l3 C: :, 0 u 10,000 1000 65 1060 5 I I 1200 1400 Mass (m/z) Figure 2-16 MALDI-TOFMS mass spectrum of electrosprayed peptide solution. Depositing 14. 5 nl of0.7 nM solution ofbradykinin and neurotensin 1-11 by electrospray led to detection of 10 amol of each peptide (SIN 70 and 15 respectively). This mass spectrum was taken using a DE-PRO MALDI TOFMS instrument. Ions were accelerated with 20 kV with 92% grid voltage and 0 005% guide wire voltage, and 100 ns delay time while 100 spectra were averaged.

PAGE 76

66 Figure 2-17. SEM images of matrix spot and matrix crystals deposited on an AnchorChip plate. a) a thick matrix layer is deposited onto the 200 m diameter spot, b) small aggregates of matrix crystals, deposited from a diluted MeCN solution.

PAGE 77

40000 35000 30000 l:i 25000 ;;; C QI .i 20000 C .2' "' 15000 10000 5000 0 67 Oxytocln Bradykin i n peptide Arg-8 vaso pressin a two-layer method l)( emlxed Neurotensln 1 -11 Figure 2-18. Relative signal comparison of two-layer deposition method and premixed matrix/analyte solution. It was shown that no significant difference in signal was observed intensities when 5 amol of peptides were deposited with either of the investigated methods. Therefore, both methods can be used.

PAGE 78

20000 15000 J!I C: :, 0 (.) 10000 5000 0 1000 68 1060.5 1446.7 1084 4 1100 1200 1300 1400 1500 Mass (m/z) Figure 2-19 MALDI-TOF mass spectrum of peptide solution deposited on an AnchorChip plate. 5 amol ofbradykinin, Arg8-vasopressin, and neurotensin 1-11 are deposited from in a 10 pM solution. This mass spectrum was recorded using Bruker MALDI-TOFMS. It is a result of averaging 300 mass spectra, recorded at 20 kV, and with 17 5 kV grid voltage applied

PAGE 79

69 25 -r--;:========;--------------------~ 075% MeOH so% MeOH 025% MeOH .20 --t--'----~--------------------.------, Ill -C :::, 0 u 0 g 15 -<----------f----1---+----------------t----t------i ..... )( ....... -'iii j 10 = "' C .21 Ill 5 oxytocin bradykinin Arg-vasopressin peptide neurotensin 1-11 Figure 2-20. Comparison of signal intensity ofMALDI-TOFMS mass spectra using different solvent mixtures. The comparison was tested with a 10 pM solution of the standard peptides oxytocin, bradykinin, Arg8 -vasopressin, and neurotensin 1-11.

PAGE 80

70 35 -.--------------------------1"'-oxytocin --bradykinin 30 +-----a--------..----,.-.----------1 I ---Arg[8]-vasopressin neurotensin 1-11 o 25 -+------------~ -~-----------------< 0 0~ .... I I ";: 20 -+---------~ ---..--------------~------------< ... 'iii I C: I c: 15 +-----------=--------------ilf-------l iii C: I I -~ 10 +---------------11-------------------------i------l I I 5 +------------'-----,.-------ii!----m----------1 I I 0 +-...-........... -,.... ... iJIIII ........................... ~""""' .............. .................. __, 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 19 30 fraction number Figure 2-21. Preliminary capillary LC/MALDI separation. It is shown that the four peptides are separated on the 6 cm long capillary LC column (150 m i.d packed with 5 m particles of C18 stationary phase) The MALDI-TOFMS signal intensity of the four peptides in each fraction were plotted.

PAGE 81

71 Figure 2-22 A picture of on-line collection of the capillary LC effluent onto the MALDI AnchorChip plate. Matrix has been previously deposited. The final spot size was 200 m in diameter.

PAGE 82

72 3 2.5 2 Cl) 1.5 (.) C: 1 C'CS .c ... 0 5 tn .c 0 C'CS > -0. 5 0 ::, -1 -1.5 -2 time (s) Figure 2 23 A UV chromatogram o f the capi ll ary LC separation The four peaks are resolved each peptides eluting for 30 s. This enabled collection of 2 to 3 12second fractions per elution peak for MALDI-TOFMS analysis.

PAGE 83

I I I A. I I I I I I I I ,, ... "'~" I I I l -'00 I I I I I I -Ul I l I I l I. . l O.ll) I I I SO 7 / ;~lOC UCO Oxytocin / A~g8-vasopressin Bradykinin 73 I l I I I 7 I 25 I I I I . I L l I I I 20 I I l Fract ion I Number 7 15 I I I --7 / 10 I 1 r :;oc "'/ Neurotensin 1-11 Figure 2-24 Capillary LC-MALDIT O FMS data collected using on-line collection off line analysis. The mass spectra o f all fractions were shown. Separation of the four synthetic neuropeptides was confirmed.

PAGE 84

CHAPTER3 NEUROPEPTIDE FF METABOLISM STUDIES AND DEVELOPMENT OF QUANTIFICATION METHODS BY MALDI-TOFMS AND ESI-QITMS Introduction Quantification Using Mass Spectrometry Ion current intensity from mass spectrometric signals does not correlate precisely, accurately or directly with the amount of analyte present in a sample. Suppression of ionization due to species of significantly higher abundance can skew quantification when using either micro electrospray ionization or MALDI. Non-homogenous matrix crystallization may introduce additional irreproducibility into quantitative analysis in MALDI experiments. Choosing an approach that provides minimal bias with the highest compatibility in terms of sample preparation and mass spectrometric analysis is crucial for successful quantification. Relative quantification Relative quantification is based on comparing two states of a system (for instance control versus treated cell cultures). This does not quantify the exact amount of analyte present, but does provide quantitative values related to the changes occurring as a consequence of the observed process. This approach can be applied to observe the influence of drugs or to compare diseased to healthy cells. Metabolic isotopic labeling. Metabolic labeling is based on the incorporation of isotopic labels during cell metabolism (for example during protein synthesis). Using isotopically modified growth media enriched in 1 5N or 13C presen t s a simple universal 74

PAGE 85

75 method to metabolically incorporate an isotopic reagent. 160-162 Additionally, selected stable isotope-incorporated amino acids can be used for metabolic labeling where instead of isotopically labeling all amino acids, only selected ones are labeled.163-165 Several isotopically-labeled amino acids have been used in this manner, such as (5, 5, 5-2H3 ) leucine and (15N) methionine. In metabolic isotopic labeling, one sample set is grown in a non-isotopic growth media, while the sample for comparison is grown in depleted growth media that contains one or more isotopically labeled amino acid, and since the isotopes are incorporated at an early stage experimental error is minimized. Despite the advantages, this concept is limited to use with cell cultures, thereby excluding clinical tissue samples. Chemical labeling. Several chemical labeling approaches have been developed for protein/peptide quantification when metabolic labeling cannot be performed. All of the described methods are based on differential isotopic labeling of a standard and the analytes. For accurate quantification it is essential that the labeling reaction be highly specific and stoichiometric. A light (160) or heavy (180) oxygen atom can be incorporated in the C-terminal carboxylic group during a proteolytic digestion by way of the solvent used.1 6 6 168 It has been shown, however, that proteolytic enzymes incorporate different amounts of isotope during digestion, making accurate quantification difficult.169 In addition, the mass difference with the incorporation of a single 180 is only 2 Da, which can be difficult to resolve at higher m/z values. On the other hand, this small difference introduces minimal isotopic effects on chromatographic retention times, which is beneficial to accurate

PAGE 86

76 quantification by LC/MS. Moreover, oxygen incorporation can be applied universally to all sample types, since it is done during sample processing, not protein synthesis Cysteine is a common amino acid of choice for label introduction. The free thiol group is far more nucleophilic (meaning reactive) than any other natural amino acid side chain group. Isotope-coded affinity tag (ICAT) reagents label the thiol group of cysteines in a protein at the alkylation step of sample preparation. 170 The labeled protein is then enzymatically or chemically digested prior to MS analysis.1 7 1 ICAT reagents also contain a biotin affinity tag to enrich cysteine-containing peptides via an avidin affinity column. A mass difference of 8 Da is provided by incorporating eight deuterated hydrogen (2H) atoms (Figure 3-1 ). To minimize chromatographic retention shifts, eight 13C can be used in place of deuterium atoms for labeling. Recently, an acid cleavable linker has been incorporated to allow removal of the heavy biotin affinity tag before MS analysis to minimize mass and interference during fragmentation. 172 Acid-labile isotope-coded extractants (ALICE) are a novel class of chemically modified resins used as isotope-incorporating reagents that react with the thiol group on cysteines The reagents have an acid-labile linker (synthesized with either heavy or light isotopes) that covalently attaches cysteine-containin g p e ptides to a non-biological polymer support. As Cys-containing peptides are covalently attached to the ALICE resin nonspecifically bound species are removed by washing. Peptides are released from the resin by mild acid-catalyzed elution, and can be directly analyzed by LC/MS.1 73 Other isotopic tag methods can be used to target different amino acids Lysine residues can be labeled with 2-methoxy-4 5-dihydro-IH-irnidazole which converts only the E-nitrogen of lysine (1H4 or 2H4forms of the reagent can be used) .1 74 Such labeling

PAGE 87

77 of basic residues can further complicate MS2 data interpretation.175 Quantification of phosphoproteins is another challenge; several methods have been recently reported.176 177 One is based on the P-elimination of phosphate from phosphoserine or phosphothreonine to form dehydroamino-2-butyric acid or dehydroalanine, respectively This is followed by the addition of ethanethiol or 2H5-ethanethiol to the a-P unsaturated amide. The second method uses a phosphoprotein isotope-coded affinity tag (PhIA T) in which phosphoserine and phosphothreonine residues are selected and quantified analogous to cysteine labeling by ICAT. There are several methods that label the N-terminus of a peptide or protein. The key to successful derivatization is to provide selectivity in labeling only the N-terminal amino group. Some of the reagents used are nicotinyl-N-hydroxysuccinimide, 178 acetic anhydride, 179 and S-methyl thioacetamidate. 180 The C-terminus can also be labeled via differential esterification .181 Both Nand C-terminal labeling will be discussed further in Chapter 4 Absolute quantification The absolute measurement of protein/peptide concentration reqmres qualitative knowledge of the species of interest, as well as their environment. This is critical for method development due to crucial role played by internal standards ionization methods, and the instrumentation used for analysis. Quantification is based on the classic internal standard addition and calibration curve. Several groups have developed reliable quantitative methods by mass spectrometry, 182 183 and all emphasize the choice of internal standard as the key issue for successful quantification.

PAGE 88

78 An appropriate internal standard will match the physical properties of size charge, hydrophobicity, and ionizability of the analyte(s), with the best internal standard being an isotopically labeled version of the molecule of interest. Such an internal standard will have identical extraction recovery, retention time, and ionization response in mass spectrometry Chemical analogs of analytes can also be used. Internal standards should always be added at the same point to all samples, including controls being added as early as possible in the sample preparation process. The amount of internal standard should be well above the limit of quantification but not so high as to suppress ionization of the analyte. Quantification Employing MALDI-MS The application of MALDI-MS in quantitative analysis of biomolecules was first reported by Nelson, et al in 1992, 1 8 2 1 8 3 only a few years after introduction of the ionization technique. They presented a standard curve for insulin quantification exhibiting linearity over one order of magnitude with an error of 20%. It was noted that the two main experimental parameters affecting the absolute intensity with MALDI were laser irradiance and sample preparation. Matrix influences, signal suppression instrumental and data-acquisition variability all contribute to unreliable quantification by MALDI-MS. Weinberger and Boernsen used different internal standard/analyte systems and found that a second order polynomial best fit the data.184 One major point in MALDI-TOFMS is to avoid signal saturation during data acquisition. Although several reports are found in the literature on quantitative analysis of biopolymers by MALDI-MS, few reports address method development for quantification of particular analytes in biological fluids .183 1851 8 8 Since biological matrices contain high salt and often low analyte concentrations a desalting/concentrating step is o f great utility.

PAGE 89

79 MALDI is considered more tolerant to salts and impurities than other ionization methods, but the concentration of salt in biological fluids generally exceeds these tolerable limits. Irreproducibility influences Despite all the advantages MALDI offers in the ionization of biomolecules, one major problem of the method is spot-to-spot and point-to-point irreproducibility (discussed in Chapter 2) This especially inhibits reliable quantification, since significant variability is introduced with every laser shot. It has been reported that improvements in sample preparation can increase reproducibility and consequently quantification reliability.183 Besides the methods for improving signal homogeneity discussed in Chapter 2, there have been reports of a "hybrid" matrix made by mixing in a few unique components for improving quantification.8 6 Also, it has been shown that reproducibility is improved with smaller matrix crystals within a homogenous layer.8 7 Comparative studies with different matrices have shown that the best reproducibility can be expected when a-CHCA is used .8 6 8 8 189 Internal standard selection The use of an internal standard has two major roles : 1) to account for losses observed in biological sample pretreatment (this is why early introduction of the internal standard is essential) and 2) to reduce the influence of MALDI ioni z ation variability. According to many researchers ,86 89 an ideal internal standard: 1) must be completely resolved from the analyte, 2) must be chemically stable during analysis 3) should be chemically similar to the analyte 4) should be close to the analyte in mass and concentration to avoid suppression errors and 5) should not react w i th the analyte. Thus an ideal internal standard with all these attributes would be an isotopically label e d analog

PAGE 90

80 Quantification Employing ESI-QITMS The QITMS provides tandem MS functionalities; thus the combination of parent and unique fragment ion is usually used to selectively monitor the analyte of interest. This significantly improves analysis of complex mixtures where the mass of the compound is not selective enough When coupled to LC the mass spectrometer is typically set to scan a specific mass range, which can be as narrow as a few m/z units in selected ion monitoring. LC/MS data are represented by scanning the ion current in the individual mass scans and plotting total current versus time. The most common modes of acquiring LC/MS data are: 1) full scan acquisition providing a total ion current plot (TIC) which has minimal selectivity; 2) selective ion monitoring (SIM), where a single m/z value is monitored; and 3) selected reaction monitoring (SRM) where a particular fragmentation reaction is followed. Chromatographic separation requirements Signal suppression in ESI due to presence of salts is a main reason for sample purification prior to MS analysis. Electrospray ionization is however compatible with common reversed-phase LC solvent mixtures. Flow rate requirements to combine the two processes have recently been met with improvements in ESI devices and LC miniaturization.116 190 LC separation of complex mi x tures minimizes signal suppression by separating compounds present in smaller concentrations from those of higher concentrations If separation of mixtures is not necessary desalting of the samples can be done by use of smaller reverse-phase columns or capillary sample trap columns These columns are too short to separate mixtures, but will remove salts from samples.

PAGE 91

81 Mass spectroscopic requirements In selected reaction monitoring (SRM) a unique fragment ion is used to quantify an analyte in the midst of a very complicated mixture SRM plots are very simple with only a single MS2 peak, making this approach ideal for sensitive and specific quantification. The SRM experiment is accomplished by specifying a parent mass for fragmentation and monitoring a particular daughter ion. When optimizing MS conditions it is often better to simulate the conditions that are appropriate to a particular assay. For example, if the analysis is performed at a chromatographic flow rate of 400 I/min and the compound elutes at 30%MeOH/69%H20 / 1 %CH3COOH it is best to simulate these conditions for optimi z ation while the compound is being infused Neuropeptide FF The cardioexcitatory peptide with the sequence FMRFamide was first isolated from ganglia of the venus clam Macrocallista nimbosa, by Price and Greenberg in 1977 .191 This was the first of a group of neuropeptides with the characteristic RF amide C terminus to be identified. Two mammalian RFamide peptides were identified from bovine brain usmg anti-FMRFamide antisera, FLFQPQRFarnide (NPFF) and AGEGLSSPFWSLAAPQRFamide (NP AF) .192 It was lat e r confirmed that the same g e ne and common precursor protein are responsible for synthesis of both peptides.193 Characteristic receptors of NPFF were also recently discovered, 194195 and it was confirmed that RFamide C terminus was necessary for interaction with the receptor .196 Neuropeptide FF and its related peptides were found to be related to opioid tolerance and dependenc e .197 NPFF was therefore thou g ht to be an endo g enous neuromodulator and a pain-enhancing substance that functionally antagoni z es opioids in the mammalian C N S But r e c e ntly NPFF was found to hav e both anti and pro opioid

PAGE 92

82 activity. It was demonstrated that NPFF functions as an anti-analgesic after intracerebroventicular (i.c.v) injection, but as an analgesic when injected intrathecally into the cerebrospinal fluid.198 The authors hypothesized that NPFF represents only one of many RF amide containing peptides in the nervous system, and that it can exhibit the biological activities of other structurally related RFamide peptides NPFF may also exhibit cardiovascular effects in mammals, since it has been found that it elevates arterial blood pressure and heart rate in rats.199 It is also interesting to note that arterial blood pressure mcreases upon i c.v. administration of NPFF, but the effect is dose dependent.200 201 Metabolic pathways ofNPFF in the brain have been recently investigated. A study by Sol, et al., investigated extracellular metabolism of NPFF in mouse brain using forebrain and midbrain slices only.121 Kinetics of metabolic degradation and the main cleavage products were reported. It was also suggested that the peptide FLFQP (1-5 NPFF fragment) is a product of extracellular metabolism, though this finding is questionable since it is known that an intracellular enzyme (prolyl oligopeptidase) is responsible for post-proline bond cleavages. Experimental Methods Instrumentation MALDI-TOFMS Mass spectra were collected with a Voyager DE-PRO MALDI-TOFMS (Applied Biosystems Framingham, MA) equipped with a nitrogen laser (A.= 337 nm) and operated at a pulse rate of 3 Hz. Data were acquired with the MALDI instrument operated in the reflectron mode using a delayed extraction time of 100 ns. Positively charged ions were

PAGE 93

83 accelerated at 20 kV, and 100 single shot spectra were averaged. Baseline was corrected using Data Explorer software, provided by the manufacturer. ESI-QITMS Mass spectra were acquired on a quadrupole ion trap instrument (LCQ, ThermoFinnigan, San Jose, CA) operated with the Xcalibur software. Peptide solutions in 50%MeOH/49%H2O/1 %CH3COOH solvent were introduced at the flow rate of 150 I/min. Capillary voltage was set at 4.5 kV, temperature at 150C, and sheath gas flow at 60 units for all the mass spectral acquisitions. Parameters were optimized for SRM for all the investigated peptides and the internal standard. The parent ion isolation was in the 1 Da range, and the activation amplitude was optimized at 37% to obtain CID of the analyzed peptides. Surveyor LC pumps and the autosampler (ThermoFinnigan, San Jose, CA) were used to inject 10 L of sample into the mass spectrometer. The pumps were operating at a flow rate of 0.7 ml/min, but the flow was split before the injection valve, using a tee junction, producing the final flow rate of 150 I/min. A schematic presentation of the experimental set-up is shown in Figure 3-2. Chemicals and Reagents a.-cyano-4-hydroxycinnamic acid ( a.-CHCA) and synthetic neuropeptide FF were obtained from Sigma (St. Louis, MO). Internal standards were purchased from Bachem Bioscience (King of Prussia, PA). Neuropeptide FF fragment 2-8 was synthesized in Dr. Laszlo Prokai's laboratory at UF and the total peptide content was determined by the elemental analysis, performed by Atlantic Microlab Inc. (Norcross, GA). It was determined that 76% of the sample was the peptide fragment. All the standard peptide

PAGE 94

84 solutions were prepared in artificial cerebrospinal fluid (Harvard Apparatus, Holliston, MA) to resemble the biological samples All the solutions used for calibration curves were prepared from 1 mM stock solutions Mobile phases were prepared with HPLC grade solvents (Fisher Scientific, Fair Lawn, NJ) and filtered through 20-nm pore size aluminum oxide filters (Fisherbrand, Fisher Scientific Pittsburgh, PA) to prevent clogging. Trifluoroacetic acid (TF A) and acetic acid were purchased from Fisher Scientific (Fair Lawn, NJ), while helium gas tanks were provided by PraxAir Inc., (Danbury, CT) Sampling Techniques Both in vitro and in vivo strategies were employed to investigate the cleavage patterns of neuropeptide FF. By repeating the reported protocol1 2 1 and comparing the in vitro data with the in vivo processing, performed by using a microdialysis probe, we hoped to reveal the origin of the FLFQP peptide fragment. Quantification methods were explored using NPFF and its metabolites sampled in vitro from the brain homogenates via a microdialysis probe In vitro approach Brain slices were used to study neuropeptide FF processing in the brain in v i tro. It was expected that cell damage during th e brain slicing process would b e minimi z ed when compared to brain cell homogenates since only the cells at the edges of the slices are expected to be destroyed. The mouse brain was kept in a freezer prior to analysis. Only the cerebrum (forebrain and midbrain) was used for the study. It was carefully sliced into 300 m thick slices using a Mcllwain tissue chopper (shown in Figure 3-3). All the preparation was performed in the "cold room" at 4 C.

PAGE 95

85 Tissue preparation. Slices were washed with 15 ml of phosphate buffer ( 15 times 1 ml) to remove soluble peptidases, after which they were dispersed in 1 ml of the same buffer. A 500-l aliquot was incubated for 20 min at 37 C under gentle shaking. Neuropeptide FF was added for 0 5 mM final concentration and incubation stopped by sampling 45 l and adding 5 L of 1 M HCl at 0 15, 30, 45 65, 85, 105 and 145 minutes and centrifuged for 5 minutes. The supematants were stored in the freezer for further analysis All the samples were desalted prior to mass spec analysis using solid phase extraction Determination of protein content. Protein content was determined using the Bradford method.202 It is established on a BSA (bovine serum albumin) based standard calibration curve using photometric readings at 595 nm after dying the proteins with Coomassie Blue dye (Figure 3-4) Total protein content was used to estimate the amount of material sampled. Microdialysis (in vitro). The microdialysate samples of NPFF and its metabolites were analyzed in order to explore the possibilities and performance of quantification methods using MALDI TOF and ESI QIT MS Samples were collected in 30 m in intervals, as 100 M NPFF solution was perfused through a solution of a mouse and rat brain homogenates v i a a microdialysis probe (CMNMicrodialysis Inc., Acton MA) (schematic diagram of the experiment is shown in Figure 3-5) In vivo approach Microdialysis experiments were performed by inserting a microdi a lysis prob e CMA 4 mm (CMA/Microdialysis Inc. Acton MA) into th e striatum re g ion o f th e mous e brain. Synth e tic n e uropeptid e FF at th e concentr a tion of 100 M, was p e rfused for 3 0

PAGE 96

86 min before collection and the metabolites collected at the flow rate of 1 L/min for 1 hour. Sample Preparation Techniques Sample desalting methods Effective removal of salt contamination is particularly important in a quantification method, since suppression of the signal should be minimized Solid phase extraction and "on the spot" clean up were performed for all the MALDI sample preparations On-line purification was used throughout all the ESI experiments. Solid phase extraction (SPE). This method is based on separation of salts from peptides on a reverse-phase packing. Cartridges used for these experiments contained C18 packing and were purchased from Supelco (Bellefonte, PA). The packing was first activated with 2 ml of MeOH and than equilibrated with 2 ml of 3% aqueous acetic acid solution. After applying the sample, the salts were eluted with 5 ml of 3% aqueous acetic acid solution and the peptides with 400 l of 70% MeOH 27% H20 3% CH3COOH The effluent was evaporated using the SpeedVac (Savant Instruments, Inc, Holbrook, NY) and the analyte was redissolved in 0 1 % TF A. Washing the spot. When the developed sample preparation method was used, crystalli z ation of the matrix layer allowed for additional salt removal from the dry analyte spot. It is postulated that the salts dissolve faster and easier in water than peptides. Therefore, 2 l of water were deposited on the analyte spot and picked up by a pipette tip after 5 s. This process was repeated twice for optimal performance On-line approach. Online salt removal was introduced by attaching a capillary sample trap column (Upchurch Oak Harbor, WA) in line prior to the mass spectrometer

PAGE 97

87 as shown in Figure 3-2. The sample trap column was filled with C-18 packing, allowing for the salts to be removed when washed with an aqueous solution. The peptides were eluted with a solution of 50%Me0H/49%H20/l %CH3COOH and introduced into the mass spectrometer via the electrospray. MALDI sample preparation Matrix and analyte were deposited separately. Two different solutions of a-CH CA dissolved in acetone/0.1 % TF A were used: 1) diluted, 2 mg/ml and 2) concentrated, 10 mg/ml. A 0.3 L aliquot of a diluted matrix solution was used as a seeding layer for producing small matrix crystals. A 0.5 l aliquot of a concentrated matrix solution was layered on the top. Peptide solution (1 l) was deposited on the top of the dried matrix layer. One of the important parts of this sample preparation was the final removal of salts by washing the spot with cold H 20, as described above. Results Metabolic Differences In vitro data The presence of the FLFQP fragment was confirmed by MALDI-TOFMS and ESIQITMS after metabolism of 100 M NPFF within mouse brain slices (Figure 3-6). Degradation of NPFF was noticed to be more rapid than reported by Sol, et a/.,121 resulting in complete processing of introduced peptide within the first 30 min In vivo data The FLFQP fragment was not detected by in vivo microdialysis (Figure 37). This brought into question the validity of the conclusions presented by Sol, et a/. 1 2 1 We

PAGE 98

88 postulate that the P5 -Q6 bond is not broken naturally in the brain, but only in damaged tissue where enzymes like prolyl oligopeptidase are released as cells are ripped open. We determined that the concentration of 100 M was too high for investigating the NPFF metabolic pathway. This high concentration was only necessary in the previous work because CE/UV analysis was employed, which is considerably less sensitive than mass spectrometry detection. All samples were therefore diluted 100-fold prior to MALDI TOFMS analysis. Quantification of Neuropeptide FF Metabolism Using MALDI-TOFMS Optimization of MALDI sample preparation After confirming that fast solvent evaporation led to the formation of small matrix crystals and a homogenous matrix layer ( chapter 2), new sample preparation methods were developed to investigate neuropeptide metabolism quantitatively A simple method for depositing matrix with a fast evaporating solution was further improved by first depositing a seeding matrix layer with 0.3 l of a diluted matrix solution (2 mg/ml in acetone/0.1 % TF A) using a gel-loader pipette tip. This layer evaporated rapidly producing a thin layer of small, even matrix crystals. This was followed by deposition of 0.5 L of a more concentrated matrix solution (10 mg/ml in acetone/0.1 % TF A), forming a thicker layer while maintaining the same homogeneity. The analytes were deposited in 1 l of a 0 .lTFA aqueous solution onto the dry matrix layer and allowed to air dry. By using this method, all of the analyte could be focused into a 1 mm diameter spot due to the surface tension of the sample solution and the hydrophobicity of the matrix layer. The spot was than washed with 2 l of H20 that was pipetted off after 5 seconds. The constant ion signal before and after this wash

PAGE 99

89 confirmed that peptides were not lost, while most of the sodiated ion adducts were eliminated, enabling more constant peak height determinations. Internal standard selection The main goal in searching for a suitable internal standard was to match structural similarity and mass value with the investigated peptides We first tested FMRFam i de a naturally occurring analog ofNPFF with the same characteristic RFamide sequence at the C terminus. Even though the peptides were similar we were unable to use the FMRFarnide as an internal standard due to methionine oxidation resulting in the presence of both unoxidized and oxidized forms in the sample (Figure 3-8) Therefore as a criterion in the search for a new internal standard, we stipulated that the peptide should not contain methionine and cysteine. A second internal standard with the sequence WXRFamide (X=Nle norleucine) was selected for further investigations. It also had a similar structural and mass to NPFF with no potential for oxidized amino acids (Figure 3-9). Generating a calibration curve Calibration curves were generated for two neuropeptide FF fragments ( 1-8 and 2-8) by ratioing peak heights with the internal standard. After generating the calibration curves for each peptide fragment the quantitative interaction of the two fragments was evaluated in the system. Ion signals were recorded for both NPFF peptides to investigate the influence of each on the signal intensity of th e other and if signal suppression w ould occur The experiment was performed in two parts: 1) by preparing the same concentrations o f both peptide fragments present (for example 0.5 M NPFF 1 8 0 5 M NPFF 2-8), and 2) by preparing different concentrations of the two investigated

PAGE 100

90 peptides (like, 0.5 M NPFF 1-8, 0 1 M NPFF 2-8). The peptide concentration ranged from 0.1-1 M. The calibration plots are shown in Figure 3-10. Similarities between the two sets of conditions are noticed throughout the whole range, except at the highest concentration tested, where the largest error is noticed. This is due to signal suppression and competition during matrix-assisted ionization. No significant signal suppression is noticed at lower concentrations even when the other peptide is present at 1 M. It was also noticed that calibration curves spanning throughout one order of magnitude in concentrations were not linear but polynomial. This is caused by signal saturation and competition during ionization. Calibration curves were generated daily for accurate quantification of analytes. The samples were diluted in order to fit into the limited dynamic range of the instrument and to not saturate the detector. Each sample solution was analyzed as three spots with error bars generated from the standard deviation of the peak height ratios measured. Quantification of NPFF 1-8 and its metabolized fragment 2-8 was performed with microdialysate samples collected after perfusion of a 100 M solution of NPFF into mouse and rat brain homogenates A curve show i ng metabolic equilibrium was generated in both experiments with both peptides. It was demonstrated that 30-60 minutes were required to reach equilibrium (Figures 3-11 and 3-12). A discrepancy in the trend was noticed for the 60 min sample point from the rat brain homogenate ; this was probably a result of experimental error during sample collection. Quantification of Neuropeptide Metabolism Using ESI-QITMS A mixture of the NPFF fragments 1 8 and 2-8 and the internal standard (WXRFamide) was analyzed in both MS and MS/MS modes to investigate ~ f there were

PAGE 101

91 any interferences or significant signal suppression. It was demonstrated that the two peptides did not suppress one another causing no difficulty in analyzing the mixtures. Therefore, we chose not to separate the peptides chromatographically prior to MS analysis. We did use a capillary sample trap column to remove salts. Different solvent mixtures of MeOH and H 2 O with 1 % CH3 COOH were tested for optimal capillary trap elution and ESI performance Peptide elution was too slow with 30% MeOH, while ESI performance was less efficient at 70% MeOH. A 50%MeOH/49% H 2 O with 1 % CH3COOH mixture was found to be optimal. The experiment was successfully automated with Xcalibur software controlling the autosampler, LC pumps and QITMS. Ten microliters of sample were injected with the autosampler onto the capillary sample trap with 1 % CH3 COOH/H 2 O and washed for 3 minutes to remove salts (during this period, the effluent was going to waste). Next the peptide elution mixture (50%MeOH/49%H 2 O/1 %CH3 COOH) was used, and sample was eluted into the QITMS after ionization via electrospray by switching the injection valve Elution occurred for 2 minutes after which the column was washed with 1 % CH3 COOH/MeOH and equilibrated for the next analysis with 1 % CH3COOH/H 2 O. MS/MS optimization Selected reaction monitoring (SRM) was used to quantify NPFF and its metabolites relative to the selected internal standard All peptides were tested individually and in mixtures for optimizing fragmentation parameters. It was shown that 37% collision energy was required to provide reproducible fragmentation patterns for all the peptides tested (Figure 3-13) The major daughter ions of the peptides and the internal standard were monitor e d for quantitative analysis.

PAGE 102

92 Generating the calibration curve A calibration curve was generated at optimized experimental conditions. The areas of the chromatographic peaks in the selected daughter ion current chromatograms were compared to the peak area of the internal standard (Figure 3-14). Calibration curves were generated in the presence of both peptides, as had been done in the MALDI experiments to investigate ionization suppression. Ratio values remained linear over an order of magnitude of concentrations (Figure 3-15). The mouse and rat brain homogenates that had been investigated with MALDI TOFMS were reanalyzed The same concentrations and trends during the sampling period were observed (Figures 3-16 and 3-17). This confirmed the validity of the previous data, and the applicable performance of both ionization methods as quantification techniques. Conclusions and Future Directions In vitro and in vivo experimental models have been employed to explore neuropeptide FF metabolism in the mouse brain. The peptide fragment FLFQP was detected by two mass spectroscopic methods, MALDI-TOFMS and ESI-QITMS, following in vitro sampling but not in vivo sampling. This confirmed the hypothesis that metabolic cleavage occurred as a result of extracellular enzymatic activity of prolyl oligopeptidase. We reproduced experimental conditions as described by Sol, et al., 1 2 1 using a 100 M concentration of neuropeptides. Better sensitivity of mass spectrometric methods allowed lower concentrations to be tested matching analogous neuropeptide levels. Therefore, these methods can be explored further at nanomolar to picomolar peptide concentrations metabolized

PAGE 103

93 Development of quantification methods using MALDI-TOFMS and ESI-QITMS was one of the main goals of this project. The results emphasize the main differences of the two ionization methods and the associated experimental methodologies. It was demonstrated that the same internal standard could be used for both of the approaches, though selection of an internal standard was challenging. The ESI-QITMS method offered the possibility of optimizing conditions for each peptide with great selectivity, since unique daughter ion spectra were used for quantification upon fragmentation in the ion trap. It was expected that the sensitivity of the method would increase in SRM mode. This was not always the case for different peptides due to fragmentation inefficiency. ESI-QITMS can be easily automated with a salt removal step performed on line, and with the possibility for chromatographic separation if necessary. When performing MALDI-TOFMS experiments, the focus was on improving sample preparation for increased reproducibility. Sample preparation was all performed manually, including a salt removal step, but analysis was still faster than with ESI-QITMS, since it took only -lmin to acquire a mass spectrum per sample. Another advantage of MALDI-TOFMS was that only 1 l sample volume was required for analysis, as compared with 10 l for ESI-QITMS. The differences in ionization concepts of the two methods were reflected in the calibration curve linearity and dynamic range. There was no signal suppression at higher concentrations in ESI experiments and the calibration curve remained linear over the investigated range Competition for ionization and signal suppression was noticed in MALDI, causing non-linearity in the calibration curve and miniming the dynamic range of the method to one order of magnitude

PAGE 104

94 In the end, it is important to state that successful quantification was achieved using both techniques with matching results. It can be expected that overcoming the irreproducibility problems of MALDI and miniaturizing ESI-QIT equipment would further strengthen these two techniques as reliable quantification tools.

PAGE 105

Biotin 95 0 X X 0 X X X X Linker, heavy or light Thiol-specifi c Reactive group Figure 3-1. ICAT reagent. The linker contains 8 hydrogen atoms (labeled X) that can be either 1H or 2H thereby introducing a different mass into the labeled peptides. This allows differential quantification of the same cysteine-containing protein from two different samples (for instance control versus treated).

PAGE 106

A Gradient pump Flow Rate Splitter 9 6 Capillary sample trap column Figure 3-2. Schematic of ESI-QITMS experimental set up. The samples were introduced using an autosampler. The flow rate was split before the injection valve. The samples were desalted using a capillary sample trap column packed with C-18 reverse phase packing. While washing the column, the effluent would go to waste, after which the valve wou l d switch and the eluted peptides would be electrosprayed into the mass spectrometer. The whole process was automated using Xcalibur software

PAGE 107

97 Figure 3-3. Mcllwain tissue chopper. The tissue chopper was used for in vitro sample preparation of mouse brain slices The brain sample was placed on the moving plate, while the blade was cutting slices of programmed thickness.

PAGE 108

98 0 6 0 5 a, 0 C: 0 4 ro .0 L0 fl) 0 3 .0 r., 0 2 0 1 0 I ' ' 0 1 00 2 0 0 300 400 5 00 600 700 8 0 0 900 1 0 00 concentration (mg/ml) Figure 3-4 Protein content determinati o n b y the Bradford meth o d This colorimetric method is used to determine the total p rotein content of brain samples. The ca l ibration curve is prepared using a BSA standar d so l utions The color intensities are rea d automatica ll y u sing a UV NIS microplate reader

PAGE 109

perfusion pump perfusion solution 99 m icrod ialysis / probe / magnetic stirer brain homogenate sample collectior Figure 3-5. A schematic presentation of microdialysis perfusion through a brain homogenate. The neuropeptide solution was pressurized using a perfusion pump, which introduced the peptide solution into the brain homogenate via a microdialys i s probe. Metabolism of the neuropeptide can be studied by analyzing the perfused material.

PAGE 110

100 a) FLFQPQRF -NH2 1081 59 6 5E+4 100 1103 .49 !I) FLFQP 9J >, 'vi 70 651.19 s :i: 60 BJ ., a: 861 .68 40 3) LFQPQRF-NH2 2) 934.91 10 717. 4 837. 8 958. 2 1078. 6 1199 0 Mass (mlt) b) FL FQP 11D 551.15 95 90 85 80 15 a> JC u <= 55 "' E 50 55 Q) 5 0 i!: -;;; ,5 ai a:: JS JO 25 20 15 10 5 '9] 9]0 59] iOO 69] ICC 19l !110 89] lllO 99] mlz Figure 3-6. Mass spectra of an in vitro meta b o l ism study ofneuropeptide FF: a) MALDI TOFMS, and b) ESIQ ITMS. Detection ofm/z 651 confirms the presence of the FLFQP fragment, proven to be a metabolite of an intracellular enzyme (prolyl oligopeptidase ).

PAGE 111

100 90 80 40 30 20 10 656 .19 101 821.48 934.42 956 .42 1081.48 1103. 43 6.2E+4 0 !--..J..l----~--~-...l......~.1...........--...L.J,l.,---=-,,=~~~~~~ 596. 0 714.6 8332 Mass (mlz) 951B 1070. 4 1189D Figure 3-7. MALDI-TOF mass spectrum of in vivo microdialysate The absence of FLFQP fragment is demonstrated, suggesting that the fragment was a product of intracellular enzymes from damaged cells and not a natural extracellular metabolic product in the brain

PAGE 112

100 90 80 7 0 >, ,,. 6 0 .. !i 50 .,_ 40 3 0 2 0 10 0 588.0 FMRFamide [M+H]+ 5 9 .61 6002 1 02 Oxidized FMRFamlde 615 60 6 1 2.4 Mass (m/z) 5. 0E+4 637 57 6 24 .6 636B 649.0 Figure 3 -8. MALD I-T O F mass s p ectrum of FMRFamide. O xidation o f the internal standard is shown. This was a pro bl em for reliable use of this molecule as an internal stan dard s ince th e reacti o n was n o t st oi chi o metric.

PAGE 113

103 100 620.66 4.3E+4 90 80 1081 .60 70 >60 .. C ... .!: 50 652.63 4 0 30 20 10 1103 .61 0 570.0 700.8 831. 6 962A 1093 2 1224D Mass (mlz) Figure 3-9. MALDI-T O F mass spectrum ofWXEFamide (X=Nle), (m/z 620) It is shown that the new internal standard was optimal since there were no potential oxidation sites, and the m/z range was within the range o f interest.

PAGE 114

6 00 5 00 4 00 0 en 3 00 ii: lL 2 00 1 00 0 00 104 ~------------------------1 FF 1-8 a -c --, .. ~---..,,,,-.6' -... --=--J"---:,.,----_ ... -'-I 0 0 2 0 .4 0 6 0.8 FF cone (uM) r .. -:-~l' .~ _,; FF 2-8 b FF 1-8 c FF 2-8 d 1.2 Figure 3-10. Calibration curves generated with the MALDI-TOFMS method. a) and b) were generated with the same peptide concentrations present; c) and d) generated with different concentrations of analyte present. Signal suppression and the largest errors were noticed at the 1 M point.

PAGE 115

105 a) 1-+-FF 1-8 Rat Brain, MALO 1 -TOFMS --FF2-8 6 .00 5.00 4 .00 t 2, 0 6 3 .00 t 0 0C) ,.!! u. u. 2 .00 1 .00 0 .00 - ,. 0 30 60 time (min) 90 120 150 b) Rat, FF 2-8 0.18 0 16 0 14 I i 12 ,2, 0 1 c.> 9). 08 f I LI. 11.Q 06 0.04 0.02 0 0 30 60 90 120 150 time(rrin) Figure 3-11. Neuropeptide FF metabolism in rat brain homogenate measured by MALDI-TOFMS. It was demonstrated that the equilibration of the metabolism and sampling processes was achieved after 30-60 minutes This is in agreement with previous findings and experimental concepts of microdialysis, where sampling usually starts 30-60 minutes after perfusion. b) The 2-8 fragment metab o lism changes are enlarged for a clearer view

PAGE 116

106 a) Mouse Brain MALDI-TOFMS l-+-FF1-8 ---FF 2-8 3 00 2 50 I I l 2 00 2, 0 1.50 C 8 u. u. 1.00 t 0 50 0 00 0 30 60 90 120 150 time(rrin) b) Mouse, FF 2-8 035 0.3 025 I I -2. 0 2 8 0.15 I I u.. u.. 0.1 ODS 0 0 30 60 90 120 150 time(mn) Figure 3-12. Neuropeptide FF metabolism in mouse brain homogenate measured by MALDI-T O FMS. The same trend in concentration changes was noticed as when the rat brain homogenate was analyzed. It is shown that l ess FF 1-but more metabolite 2-8 was present than in the rat brain homogenate suggesting faster conversion and meta b o l ic pathways b) The 2-8 fragment metabolism changes are enlarged for a clearer view

PAGE 117

107 a) Y B 821 .17 12 Y 5 15H .11 i 10 L F O P 2 8 459.04 [M-NH3)2 8 532, 4 5 <( bg-NH3 ~ 362.112 POR 51924 381512 a: b 6 D ... ml% b) HD 15415.25 Y 4 8 5 80 IS 10 0 5 .;'I ~H 150 ,. .. ;; a: 35 b 3 Y s J O Jl&O J t 5lt...29 25 O P O NH;, 2 0 331 5 1 b4-NH3 15 Y2 151"6 b6 1 0 FOP O R 321.17 bs-NH 3 Tl0.J2 Ys 5 65l J 4 82151 D 31D 3Sll IDD ... SID 5 Sll 61D .., 100 1 5 0 81D 8 Sll mlz c) 100 30 13 g5 b 2 g o 8 5 80 7 5 7 0 --0 C M ::, ::e 00 a, 55 ~ 50 "' -;;; 45 0:: 40 35 3 0 26 20 Y J 1 6 Y 2 43 2 0 10 32 1 3 6 0 150 200 250 300 3 50 400 m/z 450 500 550 600 Figure 3-13 ESI-QIT MS/MS spectra of all peptides investigated: a) NPFF 1 8 after CID ofm/z 541; b) metabolite 2-8, after CID ofm/z 468 ; c) internal standard after CID of m/z 310. The major daughter ions were used for quantification in SRM e x perim e nts.

PAGE 118

108 a) 1: .... ] _________ ___.,,,_tt_: .~~__OA_---"-"""""-....... ______ .,, J.l:J.81 c ,,. 1 M :&l!l9ll" I:~1 ____________ ~1_\_~-----------:5 J.1:)91 .:;:] ""' . .. """""'""' """'"~'1'3'2::~1, !'I P f.,,, """"" 821 m/z 546 m/z 302 m/z H H 10 lj. U lt lt lt t .t .tt ,_,_ U U TO U l t U ~ Tm(min) IJ 821Dg g s 541 ----:+ 821 m/z ~D~---------~,.-----------~ ]rno -516 .12 SJ .le b) 468 --+ 546 m/z ~o----------~------------: l~BJ D I I ( ( I 31 0 --+ 302 m/z I I I j I I I I j I I IC ( I I I j I I I j I I I I ( l I I I j I I I I I I I I I I JD ISJ SOD 56D 600 6SJ 70D ?SJ 80D rrlz Figure 3-14. Xcalibur presentation of: a) chromatogram of the major daughter ion current; b) SRM mass spectra. The peak areas were used to quantify neuropeptides

PAGE 119

8 7 6 o 5 :i, f!4 en u:: LL. 3 2 1 0 109 FF2-8a ~------------------------1 I 0 0 2 y = 6 .4772x R2 = 0 9933 0.4 0 6 0 8 FF cone (uM) y = 5.4863x ff = 0.9968 FF 1-8b F F 1-8 c FF 2-8d y = 1 .7365x R2 = 0 .9109 1 2 Figure 3-15 Calibration curves generated from ESI -Q ITMS experiments. No significant d ifferences were n oticed when the same ( a b) and different ( c d) concentrations were tested.

PAGE 120

110 Rat Brain, ESI-QrTMS 9 .00 I__._ FF 1-8 -FF 2-8 8 .00 -I 7 .00 6 .00 I I i' 2-5 .00 u C 8 4.00 LL. LL. 3 .00 2 .00 I 1 .00 1 1 0.00 I I I I 0 30 60 90 120 150 time (min) Figure 3-16. Neuropeptide FF meta b olism in rat brain homogenate by ESIQ ITMS. The same tren d in co ncentrati o n change s an d similar values were d etected as when the same samples were ana l yze d b y MALD I-T OFMS.

PAGE 121

111 Mouse Brain ESI-Q I TMS -+FF 1-8 4 00 FF2-8 3.50 I 3 00 I 0 2.50 :i:, f! 2.00 UJ -==:: l&.. 1.50 l&.. 1 00 0 50 f !! 0 00 I I I I I 0 30 60 90 120 150 time ( min ) Figure 3-17. Neuropeptide FF metabolism in mouse brain homogenate measured by ESI QITMS The same trend in concentration changes and similar values were detected as when the same samples were analyzed by MALDI-TOFMS

PAGE 122

CHAPTER4 DEVELOPMENT OF DIFFERENTIAL QUANTIFICATION FOR MONITORING METABOLISM OF DYNORPHIN 1-8 IN THE BRAIN Introduction Opioid Peptides The number of known neuropeptides is still increasing. The opioid peptides are among the best characterized They are coded by three distinct genes, whose products are preproopiomelanocortin (POMC), preproenkephalin, and preprodynorpin. The best known peptides released from these prepropeptides are P-endorphins, enkephalins, and dynorphins, respectively. Nerve cells or various glands throughout the body can naturally produce opioid peptides. When secreted by nerve cells, they act within the central nervous system (CNS) as neuromodulators but when produced by glands they can be delivered to distant target tissues in the body. They are known to induce the same effects such as alkaloid opiates (morphin, heroin). A broad variety of effects is known to depend on opioid peptides ranging from preventing diarrhea to inducing pain relief The active opioid peptides are formed from their precursors by enzymatic conversions producing smaller peptide fragments. Furthermore posttranslational processing can also include the addition of different chemical groups to the peptides (like, glycosylation acetylation, methylation) which can alter their biological activities. 112

PAGE 123

113 Discovery of dynorphins This whole group of peptides is named after the Greek word d y namis, meaning power, and (m)orphine, for the compound which gives similar effects in the brain. Dynorphin was first identified in the pituitary.204 205 Characterization and partial sequence of this highly potent endogenous opioid peptide were first reported in 1979 by Goldstein, et a/.,205 It was then known that this novel peptide contained the amino acid sequence of Leu-enkephalin at its amino-terminus, but the complete sequence of 17 amino acids was confirmed later.20 6 It is interesting to note that prodynorphin-derived peptides contain the Leu-enkephalin sequence. This limited early imrnunochernical studies due to cross reactivity of dynorphin antisera with proenkephalin It was later shown that the same gene is responsible for synthesis of both prepropeptides. These molecular, genetic and structural parallels have made it difficult to differentiate the dynorphin and enkephalin neurons in the nervous system. Physiological role Dynorphins are directly involved in many vital processes (in self-stimulation reward drug abuse, and so on) but their modulatory effects (in stress response response to brain injury and stroke feeding and many others) are significant as well. One of the most intriguing roles of dynorphins is their modulatory effect on morphine and influence over opiate dependence expression. Studies have shown that dynorphin can substitute for opiates in dependent situations which makes it able to attenuate the opiate withdrawal syndrome and opiate tolerance 207 Neurochemical processing As with many other biologically active substances dynorphins are synthesi z ed as an inactive pre hormone form (predynorphin) and then becom e biolo g ic a lly active aft e r

PAGE 124

114 posttranslational processing It is known today that this "disguise" protects the activity of many biologically active compounds .20 8 The active species Dyn A, Dyn B and aneoendorphin are created through a series of enzymatic cleavages after delivery to the active site ( see Figure 4-1 ). Further enzymatic conversion of these peptides produces shorter fragments, such as Dyn 1-8 (peptide sequence: YGGFLRRI) and Leu-enkephalin (sequence : YGGFL), which are also active. In addition, differential processing of predynorphin in different brain regions creates a variety of fragmentation patterns. 2 09 Natural processing of peptides depends on many factors. One of the most important is the activity of certain enzymes. Even though quantities of some compounds can be measured, determination of their "active concentrations" is what really matters in the signal processing within the brain. Knowing the influences of particular enzymes and their inhibitors would help define neuropeptide processing in the living organisms Even more, some of the natural inhibitors could be used as models for synthesis of possible drug candidates The most reliable results when analyzing the behavior and processmg of neuropeptides are expected from in vivo studies Processing of dynorphin fragments has been studied in tissues, body fluids (blood) and in brain tissue.118 1 20 2 1 0 These analyses revealed the main cleavage sites and the pathways of biotransformation. The critical sites for enzymatic cleavage of Dyn 1-8 are Tyr1-Gly2 Leu5-Arg6 and Arg6-Arg7 peptide bonds. It was also shown that the longer fragments are more resistant to further enzymatic degradation.120 2 1 0 The influence of selected peptidase inhibitors was also d d . b p k 211 stu 1e zn vztro y ro a1, et a

PAGE 125

115 Understanding the structure of dynorphins is important for investigating transport, binding and other peptide interactions, but the amount of active and metabolized specie is of the great importance for a more complete definition of a neuropeptide activity The f d hi . 1 bl d b 10-9 10-1 2 M 2 12 214 b t 1 1 concentrat10n o ynorp ns m c1rcu atmg oo 1s a out to u oca concentrations may be much higher at the sites of dynorphin release. Differential Quantification Approach Differential quantification enables an accurate comparison of a large number of peptide fragments in two samples It can be utilized to reveal dynamics of changes in proteins/peptides or response to stimuli. It is known that ion-current intensities of mass spectrometric signals alone may not always correlate directly or accurately with the amount of analyte present in a sample Instead of introducing internal standards for all the peptides in a sample to recognize changes in concentration between control and experimental samples all the components in the control could be used as internal standards for differentially labeled experimental samples. The analysis is then based on measuring the changes in isotope ratios of peptides in control and experimental samples based on the relative intensity of mass spectrometric signals of the heavy and light forms of the labeled species Peptide pairs will have the most similar physico-chemical properties so they are expected to behave identically during sample handling and ionization. Therefore it is expected that the species should provide an accurate measure of the relative abundance of the peptides. Peptide / protein labeling can be done by specific1 70 1 7 1 or non-specific 179 1 8 1 215 recognition. A stable isotope can be introduced into the peptide / protein molecule in many different ways It is important to note that the mass difference between the analyte and stable isotopomer should be 3 Da, to avoid overlap of naturally occurring isotope

PAGE 126

116 peaks. Also, the mass spectrometer used must be capable of resolving the two species. For larger molecules, like proteins, a larger difference in m/z should be introduced. The isotopic derivatives are made by attaching the groups to the original molecules where one class of samples would carry only light and the other only heavy isotopes After mixing the samples, and performing MALDI-TOFMS analysis, the ratio of signals for the "paired" species would correspond to the relative quantities present. Because all the species are analyzed from the same spot, some of the irreproducibility problems with MALDI sampling are minimized. Sample purification ( clean-up from salts and other impurities) is suggested for improved reproducibility This can be accomplished by a chromatographic separation or by washing the sample spot prior to the MS analysis (to be discussed later). Nand C-terminus labeling techniques do not require presence of any specific residue for labeling which makes them universally applicable. Also, the modifications of a peptide / protein are easily observed and analyzed. On the other hand no selectivity when COOH or NH2 are derivatized can introduce versatile products since these groups do not exclusively appear as C and N peptide terminus but also on the side chains of some amino acids Chemical labeling of N-terminus Munchbach, et al. described a technique in which the N-terminus is labeled with a succinic group without modification of other amino groups present in the side chains.1 78 Acylation of primary amines can also be employed for isotopic labeling Che and Fricker reported a reaction with 1H6 and 2H6 acetic anhydride that enabled quantification of neuropeptides in mice (reaction shown in figure 4-2).1 79 Another approach recently reported is based on mass-differential amidation of N-t e rminal and lysine residues with

PAGE 127

117 mass tags that differ by a methyl group.180 The methodology has been used successfully for protein quantification (QUEST) and is advantageous for the introduction of 14 Da in mass difference. Chemical labeling of C-terminus Godlett, et al employed esterification of carboxylic acid with methanol and 2H3methanol (originally developed by Hunt, et al.215 ) and demonstrated the use of this method for the protein quantification (reaction shown in figure 4-3).181 It was reported that all the free carboxylic groups (including the ones on aspartic and glutamic acid side chains) were also converted to their esters. It is important to state that no additional problem in terms of mass spectrometric analysis exists if side chains are derivatized but data interpretation can be more complex. Experimental Methods Sample Collection by Microdialysis A microdialysis probe CMMA, 4mm (CMA/Microdialysis, Inc., Acton, MA) was placed in the globus pallidus region of the Sprague-Dawley rat brain Dynorphin 1-8 was purchased from Sigma (St. Louis, MO). A 100 M peptide solution was perfused for 30 min before collection and then the microdialysate was collected for 1 h all at a flow rate of 1 I/min. The same timing intervals were used for perfusion of Dyn 1-8 with the enzyme inhibitor to be tested, bestatin (purchased from Sigma, St. Louis, MO). Sample Preparation The microdialysis samples were desalted using solid phase extraction cartridges (Supelco, Bellefonte PA), as described in the previous chapter. The final solutions of desalted samples were lyophilized in a SpeedVac (Savant Instruments, Inc Holbrook NY) since dry samples were required for the derivati z ation step.

PAGE 128

118 Derivatization methods Esterification.181215 The esterification mixture was prepared by drop-wise addition of 160 l of acetyl-chloride to 1 ml of d0or d3-methanol. Since this reaction is very exothermic, it was performed in an ice bath. After 5 min of mixing, 300 l aliquots of the prepared reagent were added to each sample: d0-methanol reagent to the control probes and d3-methanol reagent to the ones whose change was to be observed Esterification was stopped by lyophilization to dryness after 2 h. Samples were redissolved in a 5% aqueous solution of CH3COOH and used for further analysis. Controls and each treated sample were mixed in a 1: 1 v/v ratio and the mixture used for quantification Acetylation.179 Dried samples were dissolved in MeCN and 1 l of triethyl amine was added to give a resulting pH of 8 After adding 20 of acetic anhydride (Ac20), 20 l of triethyl amine was added, resulting in an optimal pH for the reaction (pH7) This quantity of base was sufficient to maintain an optimal pH throughout the reaction. The acetylation was terminated by lyophilization of the samples to dryness after 2 h The samples were redissolved in a 5% aqueous solution of CH 3 COOH for further analysis. MALDI sample preparation a-CHCA was a matrix of choice for all the MALDI-MS experiments The thin layer sample preparation method was used for all the MALDI experiments The matri x was dissolved in 0.1 % TFA in acetone (lOmg/ml) and 0. 7 l of this so lution deposited as a first layer Upon fast drying of a matrix layer 1. of a sample mixture was deposit ed. Dry sample spots were washed twice with ? l of cold H20 to remove any excess salts.

PAGE 129

119 Results Selection of Derivatization Method After generating a mass spectrum of dynorphin 1-8 and its fragments (Figure 4-4), different derivatization methods were tested for their reproducibility, selectivity, and ease of use. Mass spectra at all reaction steps were compared. Esterification proceeded to a high yield and produced a single set of methylated derivatives only on the C terminus of the peptides (see Figure 4-5) The signal-to-noise ratio improved upon derivatization due to higher hydrophobicity of the terminal ester compared to the acid. Figure 4-6 illustrates that more peptide fragments were detected upon esterification, and that the reaction goes to completion within two hours allowing addition of sufficient reagent. Acetylation yielded many different derivatization products since all amino groups reacted (see Figure 4-7). This resulted in a complex product mixture that complicated quantification Additionally, due to neutralization of the positive charge carried on the amino groups, the signal-to-noise ratio decreased significantly compared to the native peptides and many fragments were not detected. The esterification reaction was easier to perform with only a few steps of reactant addition, while it was difficult to maintain proper pH for the acetylation reaction due to the small sample volumes. Based on the data and observations presented here, the esterification reaction was determined to be more practical for quantification and was selected for further experimentation. Method Examination The experimental approach was evaluated to determine its reproducibility and reliability for differential quantification by MALDI-TOFMS analysis Specifically, two

PAGE 130

120 parts of the experiment were tested: 1) comparison of signal intensities within one spot versus a spot-to-spot comparison, and 2) reliability and reproducibility of the esterification reaction tested within the same spot. Though spot-to-spot reproducibility was significantly improved via the sample preparation protocol, comparing signal intensities across two different spots introduces significant error into the relative peptide quantification (a schematic of the experimental concept is shown in figure 4-8). We demonstrated that spot-to-spot signal comparison was not reliable, making it difficult to deduce relative quantities, due to the significantly different noise levels and absolute signal intensities (Figures 4-9 4-10) When comparing each fragment's derivatized pair within the same spot the change in concentration could be determined reliably (figure 4-11 ) The standard deviation of 3 averaged spots was <10% for all the peptide fragments, which was assigned to MALDI MS signal irreproducibility (figure 4-12) This observation supports the quantification concept that the internal standard must be treated and prepared with techniques as similar as possible to those used for the analyte To investigate reproducibility and reliability of the derivatization reaction, a sample was divided into 4 aliquots and each was derivatized as shown schematically i n figure 4 13. The differentially derivatized peptides were then pooled together and compared within the same MALDI spot. The results were quantitative with minimal deviation (<10%) (Figure 4-14). These results demonstrated that the developed approach could be used for reliable quantification of derivati z ed peptides with complete and reproducible esterification.

PAGE 131

121 Influence of Inhibitors on Dynorphin 1-8 Metabolism in the Brain The influence of certain peptidase inhibitors was investigated using the developed quantification method by MALDI-TOFMS. The objective of the project was to observe differences in fragmentation patterns due to inhibition of the enzymes responsible for the expected metabolic peptide cleavages We compared the fragmentation of dynorphin 1-8 with and without the bestatin inhibitor from in vivo sampled cerebrospinal fluid To compare the in vivo processing, dynorphin 1-8 was perfused prior to analysis of the inhibited fragmentation as diagramed in Figure 4-15 Bestatin is known to inhibit aminopeptidases and other exopeptidases. The results shown in figure 4-16 indicate a decrease of Dyn 2-8 in the presence of bestatin when compared to the control. A subsequent decrease in Dyn 4-8 was also noticed. Time Dependent Metabolism of Dynorphin 1-8 in the Brain The developed differential quantification approach was useful for monitoring dynamic processmg of dynorphin 1-8. We evaluated dynorphin 1-8 metabolism throughout a 4 day period by perfusing dynorphin 1-8 into the experimental animal (in vivo) and collecting metabolites via microdialysis one hour after repeated on the same animal for 4 day. The objective was to compare the differences in processing of this neuropeptide and discover any trends over time. The first day s sample was the control and was esterificated with d0-methanol, while all subsequent samples were derivatized with d3-methanol. Additionally the amount of peptide collected upon processing was compared to the amount perfused each day This was also accomplished by differentially labeling each pair It was demonstrated that metabolism of dynorphin 1-8 into Dyn 2-8 changed slightly with time (see Figure 4 17) This was confirmed by comparing the amount of

PAGE 132

122 Dyn 1-8 perfused to the amount collected (Figure 4-18). But, as the data was corrected for influence of microdialysis probe perfusion, it was confirmed that there was no change in dynorphin 1-8 metabolism with time (Figure 4-19). Conclusions and Future Directions A differential quantification concept was applied to analyze changes in in vivo dynorphin 1-8 metabolism. Development of the differential quantification method required that several crucial steps be investigated and optimized Choosing a specific and stoichiometric derivatization method was necessary for differential quantification. Since this issue is ambiguous for any given system, it has to be studied on the compounds of interest prior to analysis. It was demonstrated that esterification of the C-terminal carboxylic group with d0and d3methanol was optimal. Other derivatization approaches such as acetylation of amino groups or esterification of SH cysteine groups, could be optimal for peptides of other structures The sample preparation method for MALDI-TOFMS was optimized for improved reproducibility. This was achieved by using matrix solution prepared in fast evaporating acetone. Minimization of the matrix crystal size induced formation of a thin homogeneous layer significantly improving spot-to-spot reproducibility. Quantification was achieved within the same spot with less than 10% standard deviation w~en repeated The importance of both peptides being analyzed from the same spot was confirmed The developed method was successfully applied to the study of enzyme inhibition of dynorphin 1-8 metabolism. It was verified that bestatin inhibited aminopeptidases in vivo, since concentration of the metabolite Dyn 2-8 decreased when compared to basal levels.

PAGE 133

123 Overall, differential quantification has shown some great advantages over absolute quantification. One major simplification is that a calibration curve and search for an internal standard are not necessary Since the method does not provide absolute values of the analyte concentration, it is best exploited in comparative studies. This opens a broad perspective of studying dynamics of metabolic processes of peptides (for instance, before and after a drug is infused before and after an injury has occurred) or many other influences.

PAGE 134

124 Preprodynorphin (inactive precursor) + + Dyn A (1-17) YGGFLRRIRPKLKWDNQ i Dyn A (1-13) YGGFLRRIRPKLK Dyn A (1-8) YGGFLRRI i Dyn A (1-6) YGGFLR Leu-Enk YGGFL Figure 4-1. Schematic o f dynorphin meta bolism. The conversion process of dynorphin fragments is presented, starting fr o m the precursor and proceeding to the smallest active neuropeptide.

PAGE 135

125 0 a) H II H H~-c-t-N""" I R b) 0 H II H HtJ-C-tN'\/V\, I R 0 // HJC-C 0 H2 II H N-C-C-N"'-' I R 0 II DJC-C 0 H2 II H N-C-C-N"-'-' I R Figure 4-2. Acetylation reaction pathway. a) acetic anhydride reacts with amino groups resulting in acetylation of the N terminus. b) a deuterated analog of the anhydride reacts in the same fashion, but derivatizes the peptide with a heavy analog. a) b) ~ o "'"-"-C 1/"' OH + HO -CH 3 + HO -CO 3 ... 0 "'"'"' C / OCH 3 0 "-"-"C / "OCDJ Figure 4-3. Esterification reaction pathway. a) methanol reacts with the acetyl group of the peptide C terminus. This results in formation of a peptide ester b) a deuterated alcohol reacts the same way, while derivatizing the peptide with a heavy isotope form.

PAGE 136

126 981 .71 4.0E..4 100 90 BO 70 ,.. ""' 60 ] 50 .,_ 40 712 .63 30 704 .67 1003. 65 20 818 69 10 O 674 7ll 8 4 0 923 1006 1089 M a ss (m/ z ) Figure 4-4. MALDI-T O FMS mass spectrum of dynorphin 1-8 and its fragments. Shown are Dyn 4-8 (m/z 704), Dyn 1-6 (m/z 712) Dyn 2-8 (m/z 818) Dyn 1-7 (m/z 868), and Dyn 1-8 (m/z 981) fragments

PAGE 137

a) b) 100 00 726 34 718 .42 40 30 20 10 0 687. 0 1 00 90 80 40 30 775.39 -. J j I J 7 6 1 6 729 6 4 721.72 751.62 778 .71 127 995 25 13.283 832.35 882 .32 .w ;.J . 10i17-22 -l 8362 9108 !115. 4 100 0 0 Mass (m/z) 998 67 8794 9 102 0 67 835 69 Mass (m /z) Figure 4-5 MALDI TOFMS mass spectra of esterificated dynorphin fr agments. a) esterification with the d0-methanol resulted in + 14 Da derivat e s b) esterification with d3me t hanol resulted in + 1 7 Da deri v ates.

PAGE 138

a) 100 90 80 7 0 >, ;,:: 6 0 t,/1 C Qj ] 5 0 #4 0 712 .63 30 704 67 2 0 10 O 674 b) 100 90 8 0 70 726. 69 en C 60 Q) C 50 0 718. 75 40 29 72 30 721 74 20 10 0 683 128 818. 69 757 840 Mass fm/zl 868. 63 775.76 778. 78 776 .77 /.3,:rn ms m 775. 76 778.7 762 M a ss (m/ z ) 841 4.0E+4 923 4.3E+4 920 Figure 4-6. MALDI-TOFMS mass spectra of dynorphin fragments. a) before esterification; b) after esterification. Esterification improved ionization efficiency of the molecules, resulting in detection of fragments not noticeable before the derivatization (m/z 775/778 are esterificated Dyn 3-8 fragments)

PAGE 139

a) 129 100 I Dyn 1-8+ 2Ac 0-1 005 70 1 5E+ 4 9J Dyn 1-6 + 3 A c I ""''j' I 1!:' "lO ~ 60 0/n 4-a+ A c 838 63 796 6 5 j 1n 2-8 + A c I &) '# 758. 6 3 40 74 6 .71 860 69 ~--~ I Dyn 1-7 + 3Ac I 30 20 10 0 106. 0 800 4 t SOOS Mass (m /z) I Dyn 1-8 + Ac I t 1007 2 I Dyn 1-8 + 2Ac +N a Oyn 1-8+ 2Ac +2 N a 1087 6 5 1109 6 5 1149 67 1107.6 1208 0 b) m I (1/nr+Ac I I (n 1 8 + 2Ac 1005 .18 1666 0 so 80 4 0 30 20 1 0 74 6 36 0/n ~+ 2Ac I (1/ n 2 8 + A c I (n 1-6 2Ac l 860. 30 788 7 8 796.31 8 38 2 6 j (1/ n 1-8+ :'Ac I 0 ~'--"l--'...:ll-.!...L.!!.-ll--'---'--'-'----'---..,.....;...-'--'--...:.:._ __ c..,__,_ __ ...i_..:___:___:_....:..__:_--.1 61150 0 771 6 8882 Mass (mtz) 1 00 4 8 1121A 1Z38 0 Figure 4-7. MALDI-TOFMS mass spectra of acetylation products of dynorphin fragments Shown are many different derivatives of dynorphin fragments, since all the amino groups, not only the N terminal are derivatized with acetyl groups. A decrease in signal is also noticed Differences in a) and b) also demonstrate the irreproducibility of the method since the spectra are a result of derivatization of the same set of samples

PAGE 140

Sample 1 I I Sample 2 Spotting on the MALDI plate 130 I Sample 1 I Sample 2 iCH30H/AcCI iCD30H/AcCI I Sample 1 I I Sample 2 I ~Mixing (1:1)/ Sample 1 + Sample 2 I Spotting on 'Y the MALDI plate --Spot 1+ 2 Figure 4-8. Schematic presentation of spot-to-spot and single spot differential signal compression. It was important to emphasize the differences and errors introduced when the signals are compared between two spots.

PAGE 141

100 9'.) 00 ro 40 3'.) 20 10 726 34 729.64 718.42 721.72 775. 39 131 832 .35 778. 71 835 .69 995. 25 998 .67 102067 9 2E+J 8794 9 687. 0 761 6 83, 2 910. 8 l'v'lass (m/2) Figure 4-9. MALDI-TOFMS mass spectra from two different spots overlaid. It is noticed that the noise levels and the total signal intensity are different. This makes comparing the signal intensities unreliable and less accurate.

PAGE 142

132 Spot-to-spot o control sample 120 >, ;:! 100 VI C: ci, 8 0 C: "1a 60 C: .!2l 40 VI ci, > 20 ; 0 Oyn 4-8 Dyn 1-6 Dyn 2-8 Dyn 1-7 Dyn 1-8 fragment Figure 4-10 C o m p aris o n of s igna l intensities when peptides are analyzed fr o m different spot s demonstrating that such a c o mparison doe s not give meaningful data interpretati o n

PAGE 143

133 7 Ill 718. 7 5 7 .88 = 12 100 15. 0 738. 0 995 .78 4 .3E+4 90 75 7 e 80 778. 78 70 726 69 11e.11 Ill 998 76 i 60 50 0 40 g7g_43 30 81 20 10 0 683 762 841 M ass (m/ z ) 920 999 107 8 Figure 4-11 MALDI TOFMS mass spectrum of derivati v es taken from the same spot. Since analyzed from one spot the noise levels are of the same origin which makes signal to-noise levels of the two species comparabl e

PAGE 144

134 Within a spot 120 100 ~ Ill C: 80 QI .., ~ 60 C: .21 Ill QI 40 > .;::. "' '"ii ... 20 0 Dyn4-8 Dyn1-6 Dyn3-8 Dyn2-8 Dyn1-7 Dyn1-8 fragment Figure 4-12 Comparison of signal intensities when peptides are analyzed from the same spots, demonstrating that this provides meaningful comparison of data (since each fragment is compared to its own isotopically labeled pair).

PAGE 145

135 sample /+ [JJ [TI Figure 4-13 Schematic diagram of experimental concept analyzing the reproducibility of differential quantification. The original sample is divided into four aliquots, which are mixed after differential derivatization, and analyzed by MALDI TOFMS.

PAGE 146

120 100 iii C 4> 80 .... C "iii 60 C .!:!I 1/1 4> 40 ::,. +l 20 4> I.. 0 Dyn 4-8 136 Dyn 1-6 Dyn 3-8 Dyn 2-8 fragment Dyn 1-7 Dyn 1-8 O CH3 -CD3 Figure 4-14. Comparison of signal intensities of the differentially derivatized samples It is demonstrated that the errors are within the expected MALDI irreproducibility All the signals show the same amount of peptide present when tested from the same starting sample, confirming the method is reliable for quantification.

PAGE 147

137 1 ) Perfusion of Dy n 1-8 and 2) Observing its fragmentation i n the brain Control Sample I 1 Processing ~D-yn-1--8-a-nd-its-~-ra_g_m_e_n-ts~I .... in the Brain ....... DCH3OH/CH3COCI Chemistry in a Tube Esters of peptide fragments R-COCH 3 R-COCH3 R-COC D 3 MS: relative comparison I Added Inhibitor Dyn 1-8 and its fragments "l C D 3 OH/CH 3 COCI .,. Esters of peptide fragments R-COC D 3 Figure 4-15. Schematic diagram of experimental concept when introducing inhibitor. Control and samples are treated the same throughout the experiment, but are labeled with light and heavy derivatization reagents, respectively.

PAGE 148

138 0-CH3 120 ~-----------------------; > 80 Q) -C: ie 60 C: C) 'ii, 40 ?S 20 -CD3 .................. Dyn 4-8 Dyn 1-6 Dyn 3-8 Dyn 2-8 Dyn 1-7 Dyn 1-8 fragment Figure 4-16. Differential quantification of inhibitor influence on production of dynorphin fragments. It is demonstrated that the quantity of Dyn 2-8 fragments is decreased upon introduction of bestatin, which inhibits aminopeptidases.

PAGE 149

139 Dyn 2~ 1-8 I o,n 2-00yn 1-01 16~------------------, 14 -0 ; &! 12 10C 8 -6 N C 4-2I l f 0 ;------..---,-----,,-------.-,-----..---,------; cayO day1 cay 2 tirre day 3 d3y4 Figure 4-17 Time dependent Dynorphin 1-8 metabolism study. It is demonstrated that the Dyn 1-8/Dyn 2-8 ratio is increasing with time i ndicating a slight increase in Dyn 1-8 metabolism.

PAGE 150

140 Qln1-8 i~ec:ta:I cdle:ted 120 ~----------------------, ;; 100 C (1J ]l 80 (IS S, 60 U) 40 :.:; (IS m 20 0:: 0 -+-----"T"", -----.---,------,,.-------------.,---------t day0 day1 day2 time daj.3 d3y4 Figure 4-18. Quantification ofDynorphin 1-8 metabolism. The presence ofDyn 1-8 was monitored compared to the amount of the peptide injected. It is confirmed that the amount of metabolized Dyn 1-8 increased with time.

PAGE 151

10. 0 9 0 8 0 CD 7 0 I C 6 0 ., e 5 0 C. I 4 0 l!i 3 0 ., 2 0 1 0 0 0 day 0 141 Correcte d Dyn 1 processing I day 1 day 2 time I l day 3 da y 4 Figure 4-19. Time dependent Dyn 1-8 metabolism study, corrected for the influence of microdialysis perfusion. It was demonstrated that there was no significant difference in metabolism during a 4 day time period.

PAGE 152

CHAPTERS CONCLUSIONS AND FUTURE DIRECTIONS With recent developments in technology, new and advanced diagnostic tools have been applied to brain research studies Challenging the analysis of brain chemistry are the sample complexity and the limited amount of compounds present. Neuropeptides, an important class of small signaling molecules, are present in nM to pM concentrations (fmol-amol amount of material present per l sampled). The primary goal of this research was to develop mass spectrometry based analytical methods to characterize and quantify neuropeptides sampled from the brain The focal points of this goal were to improve sensitivity and reproducibility of MALDI-TOFMS and ESI-QITMS and to explore the strengths and weaknesses of the two approaches MALDI-MS offers a great advantage with limited sample amounts, since it onl y requires 1 l for analysis even at the pM concentrations Significant improvements in sensitivity were achieved by concentrating all the deposited material into a small spot (200 m in diameter). Attomoles of neuropeptides were detected by MALDI ionization from standard mixtures by ablating all deposited material. Different approaches were developed to minimize analyte spot size When using commercially available plat es a coating of hydrophobic material focused the deposited drop into a hydrophilic anchor that was clearly visible and ready for analysis. This technology does not require advanced sample preparation methods, making small area spottin g simple to p e rform An electrospray deposition method was also successfully used for deposition of small analyte spots. The major obstacl e to this method was irreproducibility from unstable parameter 142

PAGE 153

143 settings, which could be overcome by automating the deposition process. This method was complex, involved longer sample preparation times, and required larger sample volumes. Additionally, analyte deposition performed on top of the matrix layer redissolved the matrix crystals and dislocated material into the larger spot area, thereby further complicating analysis. Though picomolar solutions were successfully analyzed using the developed MS methods, it is important to emphasize that interferants present in biological samples can decrease sensitivity. Eliminating these interferences has been the driving force for developing LC methods coupled to MS analysis. Reverse phase chromatography is often employed to remove salts and other small molecules from biological samples, and also to separate components of the complex mixture and preconcentrate material. Coupling capillary-LC with MALDI-TOFMS was successfully performed by an on-line collection, off-line analysis approach. The amount of information and the specificity of characterization are also greatly improved with LC employed prior to MS analysis. A benefit of this method was that the samples (and information) could be stored and reanalyzed later with the same or other instrumental settings. For example, a rapid screening of the samples could be performed by MALDI-TOFMS, but the same sample could be reanalyzed using tandem mass spectrometry by MALDI-Q-TOFMS, which provides more specific information about the analyte. Speed and precision of the off-line sample collection could be improved by automation. Electrospray ionization has long been coupled to RPLC since both require liquid flow. Recently, refinements in ESI source technology have allowed for direct interfacing of small scale LCMS (micro and nanospray) that is better suited to analysis of sample

PAGE 154

144 limited biological material. This methodology was easily automated using available hardware and software ESI-QITMS compared to MALDI-TOFMS offered the option to perform collision induced dissociation (CID) of targeted peptides that provided unique fragmentation patterns. Post-source decay (PSD) performed with MALDI-TOF could also provide fragmentation patterns, but is not as effective with low concentrations of material. MALDI-Q-TOF instrumentation could provide a compromise offering the high mass accuracy and precision of MALDI-TOFMS with the MS/MS capabilities of a tandem mass spectrometer. Improvements in reproducibility have been explored for development of a reliable quantification method using MALDI-TOFMS. It was demonstrated that improvements in matrix layer homogeneity could minimize spot-to-spot irreproducibility. Matrix layer homogeneity was improved by minim i zing matrix crystal size, achieved by dissolving matrix in fast evaporating solvents such as acetone It was confirmed that faster evaporation produced smaller matrix crystals. Electrospray deposition of matrix in acetone solution, with rapid solvent nebulization, resulted in the most homogenous matrix layer. Improved matrix layer homogeneity was also achieved by using a diluted matrix solution as a seedin g layer. This approach was faster to perform and result e d in the homogenous matrix layer necessary for reproducibility improvements. The thin-layer based deposition method was used during the quantification investigations achievin g less than 10% RSD in spot-to-spot reproducibility. Two differ e nt MS based approaches have be e n e xplored for neurop e ptid e quantification While absolute quantification results in concentration determination

PAGE 155

145 relative quantification compares quantities of compounds from two samples (like, control to sample). The initial challenge for successful quantification was selecting an internal standard that worked with our samples. A peptide with similar physical properties was used for quantification with both MALDI-TOFMS and ESI-QITMS. The main differences between the two methods were linearity of response and dynamic range It was shown that signal suppression limits dynamic range when MALDI is employed. Other major differences were noticed in sample volume used and time needed for analysis, where MALDI-TOFMSbased methodology was ad v antageous. While 10 l of sample were needed for ESI-QIT analysis with the instrumentation used, only 1 l was needed for MALDI-TOFMS analysis The ESI-QIT method was fully automated with an on-line capillary sample trap for salt removal. It still required 14 minutes per analysis when using the optimized parameters. Differential quantification did not require an internal standard for each compound of interest, rather isotopic labeling of the investigated species performed by derivati z ation reactions The preliminary investigation was focused on choosing the derivatization method that would be specific and stoichiometric for the analyte of inter e st. This was th e most important step and required a detailed investigation. The goal was to establish a derivatization reaction and conditions for its optimal performance, based on chemical structures of the analytes and complexity of the sample Esterification was found to b e the most successful derivatization reaction for peptides with carboxylic groups only at the Cterminus. Acetylation reactions resulted in multiple derivati z ation products complicating further data analysis and use in quantification.

PAGE 156

146 Both absolute and differential quantification approaches were employed in neuropeptide metabolism studies. Absolute quantification offered more exact results, but required more complex method development. Differential labeling is most advantageous in MALDI-MS analysis, as the samples can be analyzed and quantified from the same spot, minimizing the error introduced with MALDI irreproducibility Additionally, multiple peptide fragments could be quantified at the same time with a differential quantification approach as compared to absolute quantification, which requires separate calibration curves for every analyte, and the synthesis of standards that match investigated species. Dynorphin 1-8 metabolism was investigated using differential quantification. The influence of enzyme inhibitors was quantified by comparison of differentially labeled peptide fragments We determined that Dyn 2-8 production decreased after the enzyme inhibitors were introduced, suggesting the mechanism of the peptidase inhibition by bestatin, an exopeptidase inhibitor. This method can further be used for further to study other influences on dynorphin 1-8 metabolism or other peptides of similar structure. The same approach of differential quantification can be applied for variety of peptides, but the derivatization method would have to be investigated for different structures. Neuropeptide metabolism was explored throughout this work using in vivo and in vitro sampling models. While in vitro models offer insight into metabolic pathways via simpler experimentation, the presence of intracellular components due to cell damage can introduce errors In vivo experiments offered greater reliability of results but were more complex to perform We have concluded that the FLFQP fragment was not a product o f the extracellular metabolism in the brain, but rather a product of enzymatic activity from

PAGE 157

147 an intracellular enzyme. Since our in vivo results of neuropeptide FF metabolism in mouse brain disputed previously reported in vitro results, 121 the next step would be to investigate the same metabolism in the rat brain to further substantiate our conclusions across multiple studies and species. Further improvements in brain research diagnostics can be expected from combining developments of analytical and sampling methods Miniaturization of instrumentation has already started dominating mass spectrometry, from small sample spots in MALDI-MS to nanoelectrospray coupled with capillary LC Miniaturization of brain sampling techniques, leading to less invasive analysis, should also be investigated to reduce damage to brain tissue. Early diagnosis of certain brain diseases is critical and the discoveries of powerful new drugs for treatment are expected to be the mainstream focus of future brain research Reliable and applicable analytical methods including those based on mass spectrometry will be promising tools in this pursuit.

PAGE 158

LIST OF REFERENCES 1. Tanaka, K.; ldo, T.; Akita, S. in Proceedings of the Second Japan-China Joint Symposium of Mass Spectrometry, Osaka, 1987 185-8 2. Karas, M.; Bachmann, D.; Bahr, U.; Hillenkamp, F. Int J Mass Spectrom Ion Processes, 1987, 78, 53-68 3. Tanaka, K .; Waki, H .; Ido, T.; Akita S.; Yoshida Y.; Yoshida, T. Rapid Commun. Mass Spectrom., 1988, 2, 151-3 4. Karas, M.; Hillenkamp, F. Anal. Chem., 1988, 60, 2299-301 5. Karas, M.; Ingendoh, A.; Bahr, U.; Hillenkamp, F. Biomed. Environ. Mass Spectrom., 1989, 18, 841 6. Karas, M.; Bahr, U.; Hillenkamp, F. Int. J Mass Sp ectr om Ion Process es, 1989, 92,231 7. Beavis, R C.; Chait, B.T. Rapid Commun. Mass Spectrom., 1989, 3 233-7 8. Beavis, R .C.; Chait, B.T. Rap id Commun. Mass Sp ectrom., 1989 3, 43 2-5 9. Hillenkamp, F.; Karas M .; Beavis, R.C .; Chait, B .T. Anal. Chem., 1991 63, 1193202A 10. Zare R.N.; Levine, R.D. Chem. Phys Letters 1987, 136 593 11. Vertes, A. ; Gijbels, R. ; Levine, R.D. Rapid Commun. Mass Spectrom., 1990, 4, 228-33 12. Vert es, A.; Irinyi, G.; Gijbels R. Ana l Chem., 1993 65 2389 13. Zenobi R. ; Knochenmuss R. Mass Spectrom. R ev 1998 17, 337-66 14 Knochenmuss, R; Zenobi, R. Chem. R ev 2003 1 0 3, 441-52 15. Karas, M ; Gluckmann M.; Schafer J. J Mass Spectrom., 2000 35, 1-12 16 Karas, M.; Kruger, R. Chem. R ev., 2003 103, 427-39 148

PAGE 159

149 17. Allwood, D.A.; Dyer, P.E.; Dreyfus, R.W. Rapid Commun. Mass Spectrom., 1997, 11, 499-505 18. Beavis, R.C. Rapid Commun. Mass Spectrom., 1992, 27, 864-9 19. Juhasz, P .; Costello, C.E.; Biemann, KJ Am. Soc Mass Spectrom. 1993, 4, 399405 20. Wu, K.J.; Odom, R.W. Anal. Chem, 1996, 68, 873-6 21. Wu, K.J.; Steding, A.; Becker, C.H. Rapid Commun. Mass Spectrom., 1993, 7, 1427 22. Ehring, H.; Karas, M. ; Hillenkamp, F Org. Mass Spectrom, 1992, 27, 427-32 23 Wang, B.H.; Dreisewerd K.; Bahr, U.; Karas, M.; Hillenkamp, F. J Am. Soc. Mass Spectrom, 1993, 4, 393-8 24 Johnson, R.E. Int. J Mass Spectrom. Ion Proc, 1994, 139, 25-9 25. Ehring, H.; Sundqvist, B .U. R J Mass Spectrom 1995, 30, 1303-8 26. Ehring, H.; Sundqvist, B.U.R Appl. Surf Sci, 1996, 96-98 577-83 27. Knochenmuss, R.; Dubois, F.; Dale, M.J.; Zenobi, R. Rapid Commun. Mass Sp ec trom., 1996 10, 871-7 28. Karas M.; Bachmann D.; Bahr U.; Hillenkamp F. Int. J Mass Sp ectrom Ion. Proc, 1987 78, 53-8 29 Breuker K.; Knochenmuss, R.; Zenobi, R. Int. J Mass Spe ctrom. Ion. Proc, 1987 78,53-8 30. Lehmann, E.; Knochenmuss R.; Zenobi, R. Rapid Commun Mass Spectrom ., 1997, 11, 1483 7 31. Mowry, C.D.; Johnston, M.V. JPhys.Che m 1994 98, 1904 8 32. Harrison A.G. Mass Spectrom R ev, 1997 16 201-15 33. Sunner, J Org. Mass Spectrom, 1993, 28, 805 9 34. Bokelmann V.; Spengler B ; Kaufmann R. Eur Mass Spectrom, 1995 1 81-6 35. Kinsel, G.R. ; Edmondson, R.D.; Russell D.H J Mass Spectrom, 1997 32, 714-9 36. Dogruel D. ; Nelson, R.W> ; Williams, P Rapid Commun. Mass Spectrom., 1996, JO, 801-6

PAGE 160

150 37. Llenes, C.F.; O'Malley, R.M. Rapid Commun. Mass Spectrom., 1992 6, 564-9 38. Belu, A.M.; DeSimone, J.M.; Linton, R.W.; Lange, G.W.; Friedman R.M J A m Soc Mass Spectrom, 1996, 7, 11-15 39. Chowdhary, A.K.; Crithley, G.; Diaf, A.; Beckmann, E.J.; Russell, A.J. Macromol 1996, 29, 2213-9 40. Deery M.J.; Jennings, K.R.,; Jasieczek, C.B. et al. Rap i d Commun. M a ss Spectrom 1997, 11, 57-62 41. Mowat I.A.; Donovan, R.J ; Maier, R.R.J. Rapid Commun. Mass Spect r om., 1997 11, 89-95 42. Wang B.H. ; Dreisewerd, K.; Bahr U ; Karas, M.; Hillenkamp F J Am. Soc Ma ss Spectrom, 1993 4, 393-8 43. Belov, M E ; Myatt C P.; Derrick P.J. Chem. Ph y s L ett, 1998 2 84, 412 7 44. Juhasz P. ; Costello C.E Rap i d Commun. Mass Spe c trom 1993, 7 343-7 45. Lidgard R.; Duncan, M.W. Rapid Commun. Mass Sp ec trom. 1995, 9 1283 2 46. Gluckmann M.; Karas M. J Mass Spectrom. 1999 34 467-7 7 47. Beavis R.C. ; Chait, B .T. Che m.Ph y s L e tt 1991 18 1, 479-83 48. Thompson J.J. Ra y s of Positi v e Electrici ty Longmans Greens 1913, London 49 Aston F W Phil. Ma g. 1919 38 707 50. Cameron, A.E ; Eggers, D.F., Jr. R e v Sci. Instr. 1948 19 605 51. Stephens W .E. US Patent 2 612 607 1952. 52. Wiley W.C., McLaren I.H R ev. Sci Instr. 1955 2 6 1150-7 53. Mamyrin B.A.; Karataev V.I.; Shmikk D.V.; Zagulin V .A. S o v Ph ys. JET P 1973, 37, 45 54. Hillenkamp, F., Unsold E., Kaufman R., Nitsche, R. Appl Ph ys, 1975 8, 341-8 55. Van Breem en, R. B., Snow M Cott e r R. J Int J Mass Sp ec t rom J o n Ph y s 1983 49 3550 56. Koster c Kahr M S., Castoro J. A., W i lkins C. L. M a ss S p ect ro m R e v iews, 1987 7 8 495 512 57. Buchanan M. V., Hetich R. L An a l C h em, 1993 65 245A 59A

PAGE 161

151 58. Castoro, J.A., Koster, C., Wilkins, C. L. Anal Chem, 1993, 65, 784-8 59. Zeleny, J. Phys. Rev., 1917, JO, 1-6 60. Dole M.; Mach, L.L.; Hines, R.L.; Mobley, R.C.; Ferguson, L.P.; Alice, M.B. J. Chem. Phys. 1968, 49, 2240-9 61. Mack, L.L., Kralik, P.; Rheude, A.; Dole, M J. Chem. Phys. 1970, 52, 4977-86 62. Clegg, G.A.; Dole, M. Biopolymers, 1971, JO, 821-6 63. Teer, D.; Dole, M. J. Polymer Sci., 1975, 13, 985-95 64. Yamashita, M.; Fenn, J.B. J. Phys. Chem., 1984 88, 4451-9 65. Yamashita, M.; Fenn, J.B J. Phys. Chem., 1984, 88, 4671-5 66. Whitehouse, C.M.; Dryer, R.N.; Yamashita, M.; Fenn, J.B. Anal. Chem., 1985, 57, 675-9 67. Alexandrov, M.L.; Gall, L.N.; Krasnov, N.V.; Nikolaev, V.I.; Shkurov, V.A. Zh. Anal. Khim., 1985, 40, 1227-36 68. Iribame, J V.; Thomson, B.A. J. Chem. Phys, 1976 64, 2287-94 69. Iribame, J.V.; Thomson, B.A. J. Chem. Phys, 1979, 71, 4451-63 70. Loo, J.A.; Udseth, H.R.; Smith, R.D. Biomed. Environ. Mass Spectrom., 1988, 17, 411-4 71. Loo, J.A ; Edmonds, C.G.; Udseth, H.R.; Smith, R.D. Anal. Chem., 1990, 62, 693-8 72. Covey, T.r.; Bonner, R.F.; Shushan, B.I.; Henion, J. Rapid Commun. Mass Spectrom., 1988, 2, 249-56 73. Paul, W.; Steinwedel, H. Z. Naturforsch, 1953, Ba, 448-50 74. Paul, W ; Steinwedel, H. German Patent 944900, 1956 75. Stafford, G.C.; Kelley, P.E.; Syka, J.E.P.; Reynolds, W.E.; Todd, J.F.J. Int J. Mass Spectrom Ion Phys, 1984 60, 85-98 76. Van Berke}, G.J.; Glish, G.L.; McLuckey, S.A. Anal .Chem, 1990 63, 1284-95 77. Roepstorff, P.; Pohlman, J. Biomed Mass Spectrom, 1984, 11, 601-7 78. Johnson, R.S.; Martin, S.A.; Biemann, K., Stults, J.T.; Watson, J.T. Anal. Chem 1987 59, 2621-5

PAGE 162

152 79 Jardine, I.; Fenselau, C. Anal. Chem, 1975, 47, 730-3 80. Martin, S B.; Knight, J.B. Biomed. Mass Spectrom, 1974, 1, 323-8 81. Bier, D.M.; Arnold, K.J.; Sherman, W.R.; Holland, W.B.; Holmes, W.F.; Kipnis, D.M. Diabetes, 1977, 26, 1005-15 82. Cailleux, A.; Turcant, A.; Premel-Cobic, A ; Allain, P. Chromatographia, 1993, 37, 57-9 83. Thomas, K.W.; Pellizzari, E.D.; Cooper, S.D. J Anal. Toxicol, 1991, 15, 54-9 84. Phillips, M.; Greenberg, J. J Chromatogr. Biomed. Appl, 1991, 564, 242-9 85. Ling, Y.; Lin, L,; Chen, Y. Rapid Commun. Mass Spectrom., 1998, 6, 317-27 86. Gusev, A.I., Wilkinson, W.R., Proctor, A., Hercules, D.M. Fresenius J Anal Chem 1996, 354, 455-63 87. Nicola, A.J., Gusev, I.A., Proctor, A., Jackson, E.K., Hercules, D.M. Rapid Commun Mass Spectrom, 1995, 9, 1164-71 88. Gusev, A.I., Wilkinson, W.R., Proctor, A., Hercules, D.M. Anal Chem, 1995, 67, 1034-41 89. Talroze, V L.; Skurat, V.E.; Karpov, G.V J Ph y s.Chem, 1969, 43, 241-5 90 Zhang, H.; Stoeckli, M.; Andren, E.; Caprioli, R. M. J Mass Spectrom.. 1999, 34, 377-83 91. Miliotis, T.; Kjellstrom, S.; Onnerjord, P.; Nilsson, J ; Laurell, T.; Edholm, L.; Marko-Varga, G. J Chromatogr. A 2000 886, 99-110 92. Miliotis, T ; Kjellstrom, S ; Nilsson, J ; Lauren, T.; Edholm L.; Marko-Varga, G. J Mass Spectrom. 2000 35, 369-77 93 Dai, Y.; Li, L.; Roser, D. C.; Long, S. R. Rapid Commun. Mass Sp e ctrom 1999, 13, 73-8 94. Esser, E.; Keil, C.; Braun, D ; Mondag, P.; Pasch, H. Polymer 2000, 41, 4039-46 95. Whittal, R. M. ; Russon, L. M.; Li, L. J Chromatogr. A 1998, 794 367-75 96. Preisler J.; Foret, F.; Karger, B. L. Anal. Chem. 1998, 7 0 5278-87 97. Murray K. K.; Russell D. H. J Am. Soc. Mass Spectrom. 1994 5, 1-9 98. Horvath, C.G .; Preiss, B.A.; Lipsky, S.R. Anal. Ch e m 1967, 39, 1422 8

PAGE 163

153 99. Horvath, C.G.; Lipsky, S.R. Anal. Chem, 1969, 41, 1227-33 100. Kennedy, R.T.; Jorgenson, J.W. Anal. Chem, 1989, 61, 1128-31 101. Reiger, D.N. ; Kaltenbach, P.; Siegert, H.G. Electrophoresis, 1994, 15, 1234-8 102. Verzele, M ; Steenbeke, G.; Vindevogel, J. J Chromatogr, 1989, 477, 87-95 103. Takeuchi, T.; Ishii, D. J Chromatogr, 1984, 288, 451-7 104. Ishii, D.; Goto, M.; Takeuchi, T J Chromatogr, 1984, 316, 441-8 105. Lin Ling, B.; Baeyens, W.R.G., Dewaele, C ; Del Castillo, B. JPharm. Biomed. Anal, 1992, 10, 985-92 106. Gluckman, J; Shelly, D.; Novotny, M J Chromatogr, 1984, 317, 443-8 107. Baur, J.E.; Kristensen, E.W.; Wightman, Anal. Chem, 1988, 60, 2334-9 108. Slais, K.; Oscik-Mendyk, B. J Chromatogr, 1988, 456, 370-6 109. Janecek, M.; Slais, K. J Chromatogr, 1989, 471, 303-9 110. Lewis, R.V.; Stein, A S.; Gerber, L.; Rubenstein, M.; Udenfiend, S. Proc. Natl. Acad. Sci (USA), 1978, 75, 4021-3 111. Lewis, R.V.; Stein, A.S.; Kimura, S.; Rossier, J.; Udenfiend, S. Science, 1980, 208 1459-61 112. Nyberg, F.; Nylander, L.; Terenius, L. Brain Res., 1986, 371, 278-86 113. Terenius, L. Adv. Pain Res. Ther., 1987, JO, 9-15 114. Terenius, L.; Nyberg, F. Progr Brain Res., 1985, 41, 805-8 115. Marquez, C.D.; Weintraub, S.T.; Smith, P.C. J Chromatogr. 1997, 694, 21-30 116. Emmet, M.; Caprioli, R.M. J Am. Soc. Mass Spectrom., 1994, 5, 605-13 117. Gobom, J.; Kraeuter, K.0.; Persson, R.; Steen, H ; Roepstorff, P.; Ekman, R. Anal Chem., 2000, 72, 3320-6 118. Prokai, L.; Zharikova, A.D. Rapid Commun Mass Spectrom. 1998, 12, 1796-1800 119. Chou, J.Z.; Chait, B.T.; Wang, R.; Kreek, M J. Peptides, 1996, 17(6), 983-90 120. Prokai, L.; Kirn, H.; Zharikova, A.; Roboz, J.; Ma, L.; Deng, L.; Simonsick Jr. W.J. J Chromatogr. A, 1998, 800, 59-68

PAGE 164

154 121. Sol, J -C.; Roussin, A.; Proto, S.; Mazarguil, H.; Zajac, J -M. Peptides, 1999, 20, 1219-27 122. Delgado, J.M.R.; Defendis, F.V.; Roth, R.H., et al. Arch. Int. Pharmacod yn, 1972, 198 9-21 123. Ungerstedt, U.; Pycock, C. Bull. Schweiz. Akad. Med. Wiss, 1974, 1278, 1-13 124. www.zoologie-skript.de/methoden/ progen/proseq2.ht 125. Vestal, M .L.; Juhasz, P.; Martin, S.A. Rapid. Commun. Mass Spectrom, 1995 9, 1044-9 126. www.rzuser.uni-heidelberg.de/ -bl5/ency/q.html 127. http://exoteric.roach.org/bg/extra/fractals /e fract2 html 128. http ://vtp b-www.cvm.tamu.edu/vtpb /vet_ micro / serology / rial default.html 129. http ://homepage. psy. utexas.edu/HomePage/class/Psy308 / salinas/Methods / Slide 10. GIF 130. http ://www.e mory edu/CHEMISTRY / justice / seminar/microdialysis.htm 131. www.mesaplus.utwente.nl/mutas / projects/fin_projects.html 132 Cohen S. L.; Chait B.T. Anal. Chem 1996 68, 31-7 133 Kaussmann, M.; Nordhoff, E.; Rahbek-Nielsen H.; Raebel, S.; Rossel-Larsen M.; Jakobsen L.; Gobom J.; Mirgorodskaya E.; Kroll-Kristensen, A.; Palm, L.; Roepstorff, P J Mass Spectrom. 1997, 32, 593-601 134. Gobom, J.; Schu e renberg, M .; Mueller, M ; Theiss, D.; Lehrach, H.; Nordhoff, E. Anal. Chem. 2001, 73, 434-8 135. Schuerenberg, M.; Leubbert, C.; Eickhoff, H.; Kalkum, M.; Lehrach, H.; Nordhoff, E. Anal. Chem. 2000 72, 3436-42 136 Ekstrom, S.; Ericsson, D .; Onnerfjord P.; Bengtsson M.; Nilsson, J.; Marko Varga, G.; Laurell, T Anal. Chem. 2001, 73, 214-9 137. Sadeghi, M.; Vertes, A. Appl. Surf Sci 1998 1 27-12 9 226-34 138. Horneffer, V.; Forsmann, A. ; Strupat K.; Hillenkamp, F.; Kubitscheck, U Anal. Chem. 2001, 73, 1016-22 139 Allwood, D. A.; Perera, I. K.; Perkins J. ; Oldershaw G. A. Appl. Surf Sci. 1996 103, 231 -44

PAGE 165

155 140. Garden, R. W.; Sweedler, J. V. Anal. Chem. 2000, 72, 30-6 141. Vorm, O.; Roepstorff, P.; Mann, M. Anal. Chem. 1994, 66, 3281-7 142. Dai, Y.; Whittal, R. M ; Li, L. Anal. Chem. 1999, 71, 1087-91 143. Axelsson, J.; Hoberg, A.; Waterson, C.; Myatt, P ; Shield, G. L.; Varney, J.; Haddlenton, D. M.; Derrick P. J. Rapid Commun. Mass Spectrom. 1997, 11, 209-13 144. Hensel, R.R.; King, R. C ; Owens, K. G. Rapid Commun. Mass Spectrom. 1997, 11, 1785-93 145. Robinson, P.S. Nuclear Instruments and Methods, 1966, 40, 136-40 146. Van der Eijk, W., Oldenhof, W., Zehner, W. Nuclear Instruments and Methods, 1973, 112, 343-51 147. Morozov, V.N., Morozova, T.Y. Anal Chem, 1999, 71, 1415-20 148. Morozov, V.N., Morozova, T.Y. Anal Chem, 1999, 71, 3110-7 149. Avseenko, N.V., Morozova, T.Y. Ataulakhanov, F.I., Morozov, V.N. Anal Chem 2002, 74, 927-33 150. Moerman, R., Frank, J., Marijnissen, J.C.M., Schalkhammer, T.G.M., van Dedem, G.W.K. Anal Chem, 2001, 73, 2183-9 151. Fitchett, J.; Brock, A.; Ericson, C.; Hom, D. ; Peters, E.; Phung, Q.; Salomon, A. Proc 49'-h ASMS Conj Mass Spectrom. And Allied Top. 2001, 144 152. Lou X.; van Dongen, J. L. J J. Mass Spectrom 2000, 35, 1308-12 153. Johnson, T ; Bergquist, J.; Ekman, R.; Nordhoff, E.; Schuerenberg, M.; Kloppel, K. D.; Muller, M.; Lehrach, H ; Gobom, J Anal. Chem 2001, 73, 1670-5 154. Gobom, J.; Nordhoff, E.; Mirgorodskaya, E.; Ekman, R.; Roepstorff, P. J. Mass Spectrom. 1999, 34, 105-16 155. Brock, A.; Ericson, C.; Fitchett, J.; Hom, D.; Peters, E.; Phung, Q.; Salomon Shaw, C M. Proc 49'-h ASMS Conj Mass Spectrom. And Allied Top. 2001 156. Lee, M.S.; Murphy, J.P.; Valaskovic, G. Proc. 5dh ASMS Conj Mass Spectrom. And Applied Top Orlando, FL, 2002 157. Valaskovic, G.; Murphy J.P ; Lee, M.S. Proc. 5dh ASMS Conf Mass Spectrom And Applied Top, Orlando, FL, 2002 158. www. chem.ufl.edu/tangroup

PAGE 166

156 159. www.newobjective.com 160. Oda, Y., Huang, K., Cross, F. R., Cowbum, D., Chait, B.T. PNAS, 1999, 96, 6591-6 161. Washburn, M.P.M., Ulasek, R., Deciu, C., Shieltz, D.M., Yates, J.R, III Anal Chem, 2002, 74, 1650-7 162. Stocklin, R., Arrighi, J.F., Hoang-van, K., Vu, L., Cerini, F ., Gilles, N., Genet R., Markussen, J., Offord, R.E., Rose, K Methods Mo! Biol, 2002, 146, 293-315 163. Meiring, H.D., Herberts, C.A., van der Heeft, E. ten Hove, G J van Els, C.A C.M., de Jong, A.P.J.M. Proceedings of the 49'h Annual Conference of the American Society of Mass Spectrometry, 2001 Chicago, IL 164. Ong Se.E Blagoev, B., Kratchmarova, I., Kristensen, D.B., Steen, H., Pandey, A. Mann, M. Mo! Cell Proteomics, 2002 1, 376-86 165. Zhu, H., Pan, S., Gu, S., Bradbury, E.M., Chen, X. Rapid Commun Mass Spectrom, 2002, 16, 2115-23 166. Yao, X., Freas, A., Remirez, J., Demirev P., Fenselau, C. Anal Chem, 2001, 73, 2836-42 167. Stewart, I.I., Thomson, T., Figeys, D. Rapid Commun Mass Spectrom, 2001, 15, 2456-65 168. Mirgorodskaya, O.A. Kozmin, Y.P., Titov, M.I., Komer, R., Sonken, C.P., Roepstroff, P. Rapid Commun Mass Spectrom, 2002, 14, 1226-32 169. Reynolds, K.J., Yao, X., Fenselau, C. Proceedings of the 49t h Annual Conference of the American Society of Mass Spectrometry, 2001, Chicago IL 170. Gygi, S.P., Rist, B., Gerber, S.A., Turecek, F Gelb, M.H. Aebersold, R. Nat Biotech, 1999, 17, 994-9 171. Griffin, T.J., Gygi, S.P., Rist, B., Aebersold, R. Anal Chem, 2001, 7 3, 978-86 172 Borisov, O.V., Goshe, M.B., Conrads, T.P., Penisko, E.A., Veenstra, T.D., Smith, R.D Proceedings of the 49'h Annual Conference of the American Society of Mass Spectrometry, 2001, Chicago, IL 173. Qui, Y Sousa, E.A., Hewick, R.M., Wang, J.H. Anal Chem, 2002, 7 4, 4969-79 174. Peters, E.C., Hom, D.M., Tully, D.C Brock, A Rapid Commun Mass Spectrom, 2001, 15, 2387-92 175. Ji, J., Chakraborty, A. Geng, M., Zhang, X., Amini, A., Bina, M., Regnier, F J Chromatogr B, 2000, 745, 197-210

PAGE 167

157 176. Weckwerth, W., Willmitzer, L., Fiehn, 0 Rapid Commun Mass Spectrom, 2000, 14, 1677-81 177. Goshe, M.B., Veenstra, T D ., Panisko, E.A., Conrads, T P. Angell N.H., Smith R.D Anal Chem, 2002, 74, 607-16 178. Munchbach M. Quadroni M., Miotto, G., James, P. Anal Ch e m, 2002 72 404757 179. Che, F.Y., Fricker, L.D. Anal Chem, 2002 74, 3190-8 180. Beardsley R.L. Reilly, J.P. J Proteome Res, 2003, 2 15-21 181. Goodlett D R. Keller, A. Watts J.D., Newitt, R., Yi, E.C., Purvine S. Eng J.K. von Heller P ., Aebersold R. Kolker, E. Rapid Commun M ass Spectro m, 2001 15, 1214-21 182. Nelson, R.W ., McLean, M.A. Vestal, M.L. Proceedings of th e 4dh Annual Confe r e nce of th e Ame rican Soci e ty of Mass Spectrom e t ry, 1992 Washington DC 183. Nelson R.W McLean, M.A. Hutchens, T.W Anal Chem 1994, 66, 1408-15 184. Weinberger S.R. Boemsen K.O. The Protein Socie t y 6th symposium 1992. 185. Muddiman, D.C., Gusev, A.I., Proctor, A., Hercules D M Venkataramanan R., Diven W. Anal Ch e m 1994 66, 2362-8 186. Muddiman D C. Gusev, A.I. Stoppek-Langer, K., Proctor A. Hercules D M Tata, P. Venkataramanan R., Diven W J Mass Spect r om 1995 30, 1463-75 187. Jesperson S., Niessen W M.A. Tjaden U R., v an der Greef, J J M ass Spectrom 1995, 30, 357-64 188. Nelson R.W. Krone J.R. Bieber A.L. Anal Ch e m 1995 67, 115 359 189 Craig A G Hoeger, C.A. Miller C.L. Goedken T., Ri v ier J.E. F isch e r W.H Biol Mass Spectrom, 1994 2 3 519 28 190 Wilm M .S.; Mann M. Int J. Ma s s Spectrom Jon Processe s 1994 136, 16 7 7 5 191. Pric e, D Z .; Gre e nb e r g, M.J. Sci e n ce, 1977 197, 6 70-1 192. Lake J.R. ; Hammond M V.; Shaddox RC; Hunsicker L.M. ; Yang H T ; Malin D H. N e u rosci L ett, 1991 13 2, 29 32 193. Vilim F S ; Aarnisalo A.A.; Nieminen, M.J .; L i ntunen M. ; Karls te dt K. ; Kontinen, V K .; Kalso E.; States B ; Panula P ; Ziff, E. Mo! Phar m a col, 1999 5 5 804 -11

PAGE 168

158 194. Liu, Q.; Guan, X.M.; Martin, W.J ; McDonald, T.P.; Clemens, M.K. et al. J Biol. Chem., 2001, 40, 36961-9 195. Kotani, M.; Mollereau, C.; Detheux, M.; Le Poul, E.; Brezillion, S. ; Vakili, J.; Mazarguil, H.; Vassart, G.; Zajac, J-M.; Parmentier, M J Pharmacol, 2001, 133, 138-44 196. Gicquel, S.; Mazarguil, H ; Desprat C.; Allard, M.; Devillers, J-P ; Simonnet G.; Zajac, J-M. J Med. Chem, 1994, 37, 3477-81 197. Rothman, R.B. Synapse, 1992 12, 129-38 198. Yang, H.T ; Iadarola, M.J Neuroscience, 2003, I 18, 179-87 199. Roth, B L.; Disimone, J .; Majane E.A.; Yang, H Y. Neurop e ptides, 1987 JO, 3742 200. Jhamandas, J.H.; MacTavish, D. J Neuroendocrinology, 2003, 15, 24-32 201. Jhamandas, J.H.; MacTavish, D. J Comparative Neurology, 2002, 447, 300-7 202. Bradford, M. M. Anal. Biochem 1976, 72, 248-54 203. Cox, B M.; Opheim K.E.; Teschemacher, H.; Goldstein A. Life Sci., 1975 16, 1777-82 204. Lowney, L.I.; Gentleman, S.B.; Goldstein, A. Life, Sci. 1979, 24, 2377-81 205. Goldstein, A. ; Tachibana, S.; Lowney, L.I.; Hunkapiller, M.; Hood, L. Proc. Natl. Acad. Sci. USA, 1979 76, 666670 206. Goldstein, A.; Fischli W .; Lowney, L.I.; Hunkapiller, M. ; Hood L. Proc. Natl. Acad. Sci. USA, 1981, 78, 7219-23 207. Nylander I.; Vlaskovska, M.; Terenius, L. Psychopharmacology 1995 118, 391400 208. Chretien, M.; Li, C.H Can. J Bioch em. 1967, 45, 1163 209. Sandin J ; Tan-No, K.; Kasakov, L.; Nylander, I.; Winter, A. ; Silb erri ng J.; Terenius, L. P eptides, 1997, 18 949-56 210 Nylander I.; Tan-No, K .; Winter, A.; Silberring J. Life Sci 1995 57, 123-29 211. Zharikova, A.; Zharikov S. ; Block E.R.; Prokai, L. Endothelium, 2002 9 37-44 212. Bernardi, P. ; Fontana, F.; Pich E.M.; Spampinato S.; Canossa, M. European Heart Journal 1992 13 1074-9

PAGE 169

159 213 Fontana, F ; Bernardi, P.; Pich, E.M.; Boschi, S.; DeTasio, R.; Spampipato S Peptides, 1998, 19, 21-6 214. Chou, J.Z.; Brian, T C.; Wang, R.; Kreek, M.J. Peptides, 1996, 17, 983-90 215 Hunt, D.F.; Yates, J.R.; Shabanowitz, J.; Winston, S ; Hauer, C.R. Proc. Natl. Acad. Sci. USA, 1986, 83, 6233-7

PAGE 170

BIOGRAPHICAL SKETCH Tamara Blagojevic was born on July 18th 1974 in Pancevo, Serbia (what used to be Yugoslavia). She gained interest in chemistry through enthusiastic teachers as she was attending the Gymnasium High School in Pancevo After getting a degree in chemistry at the University of Belgrade in 1997 she chose to stay in academia and start the masters program in the field of environmental chemistry under direction of Dr. Predrag Po lie Tamara joined the University of Florida, Chemistry Department in the summer of 1999 She entered the mass spectrometry world in Dr. David Powell's laboratory. She has continued her PhD program under the guidance of Dr. John Eyler working in Dr. Laszlo Prokai s lab at the College of Pharmacy. 160

PAGE 171

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope and quality, as a dissertation for the degree of Doctor of Philosophy . Eyler, Chair sor of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy. j_ Ci?L Laszlo Prokai Professor of Medicinal Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope and quality as a dissertation for the degree of Doctor of PhilosvN9~ David H. Powell Scientist of Chemistry I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy. I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate in scope and quality as a dissertation for the degree of Doctor of Philosophy. ~ ow cP~l.-Scientist of Biochemistry and Molecular Biology

PAGE 172

This dissertation was submitted to the Graduate Faculty of the Department of Chemistry in the College of Liberal Arts and Sciences and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August 2004 Dean Graduate School

PAGE 173

. 1 2o __ oq._ I Bl, ':)3 UNIVERSITY OF FLORIDA II IIIIIIIII IIIII I III I I I IIII I I IIIIII I I I I III I Ill 111111111111111111 3 1262 08554 2503