Exploring neuropeptide metabolism in the brain by mass spectrometry

MISSING IMAGE

Material Information

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

Subjects

Subjects / Keywords:
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.
Statement of Responsibility:
by Tamara Blagojević.
General Note:
Printout.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 003100544
sobekcm - AA00004689_00001
System ID:
AA00004689:00001

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