Matrix-assisted laser desorption ionization (MALDI) tandem mass spectrometry (MSn) using a quadrupole ion trap

MISSING IMAGE

Material Information

Title:
Matrix-assisted laser desorption ionization (MALDI) tandem mass spectrometry (MSn) using a quadrupole ion trap
Physical Description:
xiv, 123 leaves : ill. ; 29 cm.
Language:
English
Creator:
Booth, Matthew M., 1960-
Publication Date:

Subjects

Subjects / Keywords:
Chemistry thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Chemistry -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1996.
Bibliography:
Includes bibliographical references (leaves 117-121).
Statement of Responsibility:
by Matthew M. Booth.
General Note:
Typescript.
General Note:
Vita.
General Note:
In MSn on the title page, n is a superscript.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 002205179
oclc - 36848316
notis - ALE5121
System ID:
AA00011775:00001

Full Text















MATRIX-ASSISTED LASER DESORPTION IONIZATION (MALDI) TANDEM
MASS SPECTROMETRY (MS ) USING A QUADRUPOLE ION TRAP












By


MATTHEW M. BOOTH


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


1996






























To my friends, my family, and my loving parents.












ACKNOWLEDGMENTS


I would first like to express my sincere gratitude to Dr. Rick Yost for allowing me

to be a part of his research group, and for the freedom he gave me to pursue my own

research ideas. He could always be counted on for critical insights into any project or for

solutions to almost any problem that came up. I especially appreciate the flexibility he

gave me in allowing me to work while completing my degree. I also would like to thank

my committee members, Dr. Jim Winefordner, Dr. Bob Kennedy, Dr. Nigel Richards,

and Dr. Sheldon Schuster, for the time they spent on my research.

Many Yost Group members were critical to the success of this project and in

contributing to the fun I had working towards my degree. I first worked with Dr. Nate

Yates when I came to UF, and this research eventually led to the work presented in this

dissertation. I most appreciated the friendship of both him and his wife, Jan, and will

never forget the camping trips in his Bus or the dinners at their apartment. Brody

Guckenberger spent his first year at UF working with me on the construction of the

MALDI QITMS instrument described in this dissertation. Most of the actual machine

shop work required to modify the original Vestec instrument to accommodate the

Finnigan QITMS was performed by him. Chris Reddick spent two years learning the

fundamentals of MALDI (using a Finnigan LaserMAT TOF MS instrument), and he was

of great assistance in the preparation of the samples that were analyzed for this

dissertation. Scott Quarmby provided help in both the construction of the instrument and

in discussions on both ion trap theory and data interpretation.

I would also like to credit several support personnel at UF who made it possible

for me to work full-time at Environmental Science & Engineering while performing

most of the research that went into this dissertation. Jeanne Karably and Donna Balkcom









were critical in helping me be at the right place at the right time with the right form in

order to graduate. Joe Shalosky was the machine shop supervisor (although mechanical

artist is probably a better title for him) who had a very large part in the construction of all

instruments I had a part in while at UF. Most of the instruments I designed and

constructed while at UF would not have even been attempted if Joe had not been there to

make a critical part or supply some much needed advice on how best to make something.

I must also thank Dr. David Powell for allowing me to work for him for a year in the

chemistry department mass spectrometry laboratory. He was also always there when I

needed someone to talk to (even if it was only ESE gossip).

I also need to thank my collaborator and good friend, Dr. Jim Stephenson. I never

would have even considered returning to school to earn a Ph.D. if Jim had not blazed the

trail for me. After three years with Finnigan, then four years at the University of Florida,

we have been through a lot together. It is no exaggeration to say that none of this would

have been possible without his help and encouragement. Thanks go to him and his wife,

Tracy, these were the best times of my life.

Finally, I wish to thank my wife, Pauline, and son, Christopher, for the love and

support they have given me. They truly are the most precious things that have every

come into my life. I could not have done it without them!















TABLE OF CONTENTS

page

ACKNOWLEDGEMENTS.......................................................... iii

LIST OF TABLES................................ ............ vii

LIST OF FIGURES....................... ... ........................................... viii

A B STRA CT ............................................ ................................................ xiii

CHAPTERS
1. IN TR O D U C TIO N ..................................... ..................... .............. 1

Matrix-Assisted Laser Desorption Ionization (MALDI)......................... 2
MALDI M atrices......................................... 4
M ALDI Sample Preparation............................. ..... ...... .......... 6
Time-of-Flight Mass Spectrometers.................................... 8
Mass Resolution in MALDI TOF MS................................. 11
The Quadrupole Ion Trap Mass Spectrometer..................................... 19
Fundamental Principles of Quadrupole Ion Trap Operation.... 19
Resonance Excitation and Resonance Ejection..................... 26
Mass Range Extension in the Quadrupole Ion Trap Mass
Spectrom eter.................................................... ................... 27
MALDI Using a Quadrupole Ion Trap Mass Spectrometer ................ 37
MALDI QITMS Using an External Ion Source................... 46
Overview of Dissertation................................................. .......... 49

2. INSTRUMENTAL DESIGN AND CHARACTERIZATION
OF A MALDI QUADRUPOLE ION TRAP MASS
SPECTROM ETER ........................................ .................... 55

Instrument Design.................. ......... ...... ..... .............. 55
Instrument Simulation.................................................. 64
Instrument Characterization.................................................. .. 64
MALDI QITMS of a Synthetic Tetrapeptide (MRFA).................... 66
M ALDI QITM S of Angiotensin II................................................... 72
Summary................... .................. ........... 87








3. THE ANALYSIS OF PHARMACEUTICAL COMPOUNDS
USING A MALDI QUADRUPOLE ION TRAP MASS
SPECTROMETER................. ... ............... .................. 88

Introduction................................................................... .. .............. 88
MALDI QITMS for the Structural Determination of Spiperone.......... 88
Taxol (paclitaxel) and Cancer ...................................... ................ 99
MALDI QITMS and MS2 of Taxol................................................. 103
MALDI QITMS for the Determination of Taxol in Rat Liver Tissue.. 107
Sum m ary............................................ ................................. 108

4. CONCLUSIONS AND FUTURE WORK........................................ 114

REFERENCE LIST ............................... ......................... 117

BIOGRAPHICAL SKETCH... ... .................................... 122













LIST OF TABLES


Table page
2-1. Scan table listing for the MALDI QITMS analysis of the synthetic
tetrapeptide MRFA. 71

2-2. Scan table listing for the MALDI QITMS analysis of angiotensin II. 79

3-1. Scan table listing for the MALDI QITMS analysis of spiperone. 93










LIST OF FIGURES


Fire pae
1-1. Representation of the MALDI process. 3

1-2. Typical MALDI matrices used with a nitrogen laser. 5

1-3. SEM micrograph of cytochrome C in a dihydroxybenzoic acid (DHB)
matrix. 7

1-4. MALDI TOF MS for the direct analysis of 14 salivary proteins. 9

1-5. Linear Time-of-Flight Mass Spectrometer (TOF MS) used for MALDI
analysis. 10

1-6. MALDI TOF MS spectrum of Glucose Isomerase. 12

1-7. MALDI TOF MS spectrum for a mixture of human and insulin mutants. 13

1-8. Factors affecting the mass resolution in a TOF mass analyzer. 15

1-9. Initial translational kinetic energy distributions for polypeptide ions. 17

1-10. Reflectron Time-of-Flight Mass Spectrometer (TOF MS) used for
MALDI analysis. 18

1-11. Quadrupole Ion Trap Mass Spectrometer. 20

1-12. Mathieu stability diagram for the quadrupole ion trap. Shaded region
represents stable region of a,q values. Black circles indicate ions of
successively higher mass lined up along the a=0 line. 23

1-13. Diagram of the mass-selected instability scan (scan function) showing
the RF amplitude as a function of time. 25

1-14. Photograph of 20 pum aluminum particles in a quadrupole ion trap. 29

1-15. Mathieu stability diagram, showing the point on the diagram where
instability occurs when no axial modulation is used. 32








1-16. Scan function and ion signal for the mass-instability mode of operation
where no axial modulation is used. Ejection corresponds to a qeject of
0.908. 33

1-17. Mathieu stability diagram, showing the point on the diagram where
instability occurs when using three different axial modulation
frequencies to lower the qeject. 34

1-18. Scan function and ion signal for the mass-instability mode of
operation when using axial modulation to lower the qeject to
0.906. This value corresponds to point A on figure 1-16. 35

1-19. Scan function and ion signal for the mass-instability mode of
operation when using axial modulation to lower the qeject to
0.454. This value corresponds to point B on figure 1-16. 36

1-20. Scan function and ion signal for the mass-instability mode of
operation when using axial modulation to lower the qeject to
0.0908. This value corresponds to point C on figure 1-16. 38

1-21. Relationship between the axial modulation frequency used during
the mass analysis scan and the maximum detectable mass. 39

1-22. Instrumental configuration for performing MALDI inside of an ion trap. 41

1-23. MALDI QITMS spectrum of bovine insulin performed inside of the
ion trap. 42

1-24. a). Laser desorption ionization (LDI) and b). MALDI QITMS spectra
of spiperone performed inside of the ion trap. 44

1-25. MS/MS daughter spectrum of protonated spiperone 45

1-26. MALDI QITMS analysis of spiperone in a Matrigel matrix. 47

1-27. External ion source MALDI QITMS instrument using a modified GC
ion source. 48

1-28. MALDI QITMS spectrum of bovine insulin using an external ion source. 50

1-29. MALDI QITMS spectrum of egg albumin using an external ion source. 51

1-30. MALDI QITMS instrument constructed at The Rockefeller University. 52









1-31. MALDI QITMS spectrum of bradykinin using the instrument constructed
at The Rockefeller University. 53

2-1. Simulation of MALDI-produced egg albumin ions (m/z 43300) for the
instrument used by Bier et al. Initial energy for this simulation was 126 eV,
initial angle was 0 degrees, grid spacing was 0.05 mm/grid 57

2-2. Simulation of MALDI-produced egg albumin ions (m/z 43300) for the
instrument used by Bier et al. Initial energy for this simulation was 126 eV,
initial angle was 1 degree, grid spacing was 0.05 mm/grid 58

2-3. Simulation of MALDI-produced bovine insulin ions (m/z 5730) for the
instrument used by Bier et al. Initial energy for this simulation was 17 eV,
initial angle was 1 degree, grid spacing was 0.05 mm/grid 59

2-4. Vestec ResearcHI MALDI TOF MS instrument used for this research. 61

2-5. MALDI QITMS instrument built for this research 62

2-6. SIMION simulation for bovine insulin using the design for the MALDI
QITMS instrument built for this research. 65

2-7. MALDI QITMS spectrum for 20 pmol of the synthetic peptide MRFA. 68

2-8. Comparison of the MALDI analysis of MRFA on a). a Finnigan LaserMAT
TOF MS instrument and b). on the MALDI QITMS instrument constructed
for this research. 69

2-9. Scan function used for the MALDI QITMS analysis of the synthetic
tetrapeptide MRFA. 70

2-10. MALDI QITMS spectrum for 20 pmol of angiotensin II. 74

2-11. Mechanism for the formation of the y7 ion for angiotensin II. 75

2-12. Mechanism for the formation of the b8 ion for angiotensin II. 76

2-13. Scan function used for the MALDI QIT MS/MS analysis of angiotensin II. 78

2-14. Sequential steps in the forward-and-reverse scan isolation procedure. 81








+
2-15. a). Mass-isolated MALDI QITMS specum ([M + H] ion) for 20 pmol
of angiotensin II, and b). MS/MS spectrum for this ion. 82

2-16. a). Mass-isolated MALDI QITMS specum ([M + H]+ ion) for 20 pmol
of angiotensin II, b). MS/MS spectrum for this ion, and c). MS/MS for
the mass-isolated b8+ ion (MS3 spectrum for the [M + H]+ ion of
angiotensin II). 84

2-17. a). Mass-isolated MALDI QITMS specum ([M + H]+ ion) for 20 pmol
of angiotensin II, b). MS/MS spectrum for this ion, and c). MS/MS for
the mass-isolated y7+ ion (MS3 spectrum for the [M + H]+ ion of
angiotensin I). 86

3-1. Spiperone, M.W. 395. 89

3-2. MALDI QITMS spectrum for 20 pmol of spiperone. 90

3-3. Scan function used for the MALDI QIT MS/MS analysis of spiperone 85

+
3-4. Stability diagram showing the path (---) taken by the [M + H] ion of
spiperone during the two-step isolation sequence. 94

3-5. a). Mass-isolated MALDI QITMS spectrum ([M + H]+ ion) for 20 pmol
of spiperone, and b). MS/MS spectrum for this ion. 95

3-6. Possible fragments that produce the daughter ions seen in the spectrum
for spiperone. 96

n
3-7. Series of mass-isolated spectra from the MS analysis of spiperone. 97

n
3-8. Possible fragment ions produced from the MS analysis of spiperone 100

3-9. Taxol (paclitaxel), M.W. 853. 101

3-10. MALDI QITMS spectrum for 20 pmol of Taxol. 104

+
3-11. a). Mass-isolated MALDI QITMS specum ([M + H] ion) for 20 pmol
of Taxol, and b). MS/MS spectrum for this ion. 106

3-12. MALDI QITMS spectrum for a thin slice of rat liver tissue incubated with
Taxol. 109








3-13. Blow-up of four regions for the MALDI QITMS spectrum for a). a thin
slice of rat liver tissue incubated with Taxol and b). 20 pmol Taxol
standard. 110

3-14. a). The MALDI QITMS spectrum and b). MS/MS spectrum of Taxol
incubated in a thin slice of rat liver tissue. 111

3-15. MALDI QIT MS/MS spectrum of Taxol in a). a standard and b). a thin
slice of rat liver tissue incubated with Taxol. 112













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

MATRIX-ASSISTED LASER DESORPTION IONIZATION (MALDI) TANDEM
MASS SPECTROMETRY (MSn) USING A QUADRUPOLE ION TRAP

By

Matthew M. Booth

December 1996

Chairperson: Richard A. Yost
Major Department: Chemistry

The specific aim of this study was to develop a quadrupole ion trap mass

spectrometer using a matrix-assisted laser desorption ionization (MALDI) source.

Currently, mass analysis of MALDI-generated ions is most often performed using a time-

of-flight (TOF) mass spectrometer due to its ability to record an entire spectrum for

every laser shot, as well as its high mass range and simplicity. The major drawbacks for

this type of mass analyzer are its poor mass resolution and lack of practical tandem mass

spectrometric capabilities. By using a quadrupole ion trap, unit mass resolution for low-

mass ions can be achieved, and tandem mass spectrometry can be performed for

structural determination.
TM
For this research, a Vestec ResearcH MALDI TOF instrument was modified by

placing a quadrupole ion trap immediately after the MALDI source. An exit tube lens

and dynode/electron multiplier were placed behind the exit endcap of the ion trap for ion

detection. Finnigan ITS40 electronics and Gatorware software (developed in our

laboratory at the University of Florida) were used to control the ion trap for this









instrument. The commerically available Vestec system was chosen because it was

designed specifically for MALDI TOF analyses.

The MALDI QITMS instrument designed and constructed for this research was used

for the analysis of biological and pharmaceutical compounds. Two peptides (MRFA, a

synthetic tetrapeptide, and Angiotensin II, an octapeptide) were analyzed to characterize

the performance of this instrument and to demonstrate its utility for the analysis of

peptides. The capability of this instrument for providing structural information using

tandem mass spectrometry was demonstrated by performing MS6 on Spiperone, an

antipsychotic drug developed by Bristol-Meyers Squibb. Finally, the anticancer drug

Taxol was analyzed in an incubated rat liver tissue sample to demonstrate the ability of
n
MS for the detection of target compounds in the presence of a large number of

interfering species.














CHAPTER 1
INTRODUCTION


It has traditionally been very difficult to generate intact gas-phase molecular ions

from compounds that are large, polar, and/or thermally labile since they are usually

solids and degrade or decompose when being exposed to thermal heating. The

vaporization and ionization that are required to perform a mass spectrometric analysis of

these compounds demand a method of energy transfer to the material which is fast

enough to avoid thermal degradation. Until 1988, the most successful methods of

desorbing these sensitive compounds were secondary ion mass spectrometry (SIMS),

which uses keV ions for desorption and ionization; fast atom bombardment (FAB),

which uses keV neutral atoms; plasma desorption (PDMS) which uses MeV particles

from the decay of radioactive compounds; and laser desorption (LDMS) which uses

nanosecond-pulsed high-power laser beams.

Although lasers had been used for the desorption of organic ions, all early

experiments revealed an upper limit to the size of molecules that could be desorbed as

intact ions. This limit, which depends on molecular structure and laser parameters, was
2
approximately 1000 Da for biopolymers and up to 9000 Da for synthetic polymers. The

laser desorption of molecules can occur either resonantly or nonresonantly. In the case

of resonant desorption, the direct resonant excitation of the analyte molecules also puts

energy into photodissociation channels. For nonresonant desorption, the high irradiances

required for desorption occur very close to the point of plasma generation, which also

destroys large organic molecules.









Matrix-Assisted Laser Desorption Ionization (MALDI)


In 1985, the Hillenkamp research group at the University of Frankfurt tried to use a

matrix to circumvent this problem, and first used the term matrix assisted laser

desorption ionization (MALDI). In MALDI, the analyte is mixed with a large excess of

either a solid or liquid matrix consisting of a small, highly absorbing species. Resonant

desorption of the matrix occurs, with the entrained analyte also being desorbed. Since it

is the matrix that interacts directly with the laser energy, the analyte molecules are spared
3
from the excessive energy that would lead to their decomposition. This also produces a

dilution of the analyte molecules which prevents associations that would otherwise lead

to complexes of masses too large to be desorbed and analyzed.

Using various matrices (most notably sinapinic acid), they showed the MALDI

time-of-flight mass spectrum of a dipeptide, Trp-Trp (M.W. 390) in 1985, Gramicidin

(M.W. 1900) in 1987 (at the University of Munster), bovine albumin (M.W. 67,000) in

1988, P-D-galactosidase (M.W. 116,900) in 1989, and cluster ions of a monoclonal

antibody (M.W. 300,000) in 1990.45

In MALDI, the matrix molecules absorb the energy from the laser pulse and

transfer it into excitation energy of the solid system (Figure 1-1).6 A supersonic plume

of matrix with entrained analyte is formed with a velocity of approximately 750 m/s (at

the threshold energy). Rapid adiabatic cooling of the expanding dense gas plume occurs.

Radical molecular ion intermediates of the matrix are formed by photoionization in the

condensed phase. Collisions of highly excited matrix molecules and ground-state analyte

molecules occur within the first tens of nanoseconds in the solid state or in the high-

density intermediate state of the desorbing material cloud just above the surface, leading

to even-electron quasi-molecular ions produced by proton transfer.






0 0000 c0 0ee
@o@@
o -o ee 6
0 0000 '000

oc eeoooe .
@o oso

0 @ @0000 -
0 0 0 0000 O00


S 0 @o @000 O0 *0



.

oeee see

0* 00060 0
00 0
00oO 00 000







0 a



s *@00 40 g
00000 e00 C




O0 809 oee00

0000* 00000 0 ^4



Q*o@@ 9 3 &3
S009090 00
e 0@@@ 0 ISO
0 0@ 0000000 C d
00000, 006e



00000049000009









MALDI Matrices


In order for a matrix to be usable for MALDI, it must have several characteristics :

1) The compound must strongly absorb the wavelength of the laser light. Using a

nitrogen laser (X = 337 nm), the molar absorptivity (e) should be greater than 1000

L cm mol Most of the compounds that are currently used as MALDI matrices

when using a nitrogen laser contain a phenolic hydroxy group and a carbonyl

function on an aromatic ring (Figure 1-2).

2) The matrix must contain -OH or -NH bonds for protonation. The excited state acidity

of the matrix is involved in the MALDI analyte formation process.

3) The matrix and analyte must be soluble in the same solvent and should crystallize

together (a matrix-isolated crystal containing the analyte must form).

4) The matrix must be stable under the vacuum conditions in the ion source of the mass

spectrometer.

5) The matrix preferably does not form a large number of cluster ions or adduct ions

with the analyte.


Although any compound that can perform a proton transfer under UV-irradiation is

actually usable as a matrix for UV-MALDI MS, only a select few are used on a routine

basis. The only differences that many matrices exhibit are in sensitivity, resolution,

tendency to form adduct ions, and the solvent system. Furthermore, although it is

possible to predict whether a candidate matrix material will absorb efficiently at the laser

wavelength, it is not possible to predict if a compound will form a homogeneous solid (or

liquid) solution with a given analyte molecule, particularly in the presence of ionic

contaminants.















HO O



CN



OH


Sinapinic Acid


S337= 14600


a-Cyano-4-hydroxy-
cinnamic acid

S337= 21875


2,5-Dihydroxy-
benzoic acid
E 337= 4250


Figure 1-2. Typical MALDI matrices used with a nitrogen laser.7









MALDI Sample Preparation


Typical sample preparation for MALDI analysis is a very simple procedure:

1) Matrix solution: Make a saturated solution (5 to 10 g/L) of the matrix material in

either pure water or a water/organic (ethanol, acetonitrile) solvent. Add 0.1%

trifluoroacetic acid (TFA) to the solution for proteins with low solubility in water.

2) Sample solution: Make a 105 to 10 M aqueous solution of the analyte.

3) Add 5 uL of the matrix solution and 5 uL of the sample solution to a microcentrifuge

tube and vortex. Add 1 uL of this mixture to the probe tip. Let this sample mixture

dry at room temperature.


Using this preparation scheme, the typical sample amount loaded on the probe tip is

in the range of 1 pmol or less, with the actual amount desorbed being in the amol range.

In the case of sinapinic acid, care has to be taken to protect the matrix solution from light

and to avoid heating in the evaporation step. Dried sinapinic acid preparations can

furthermore be cleaned from precipitated water-soluble contaminants such as inorganic

salts by immersing the dried sample in cold distilled water.

Depending on the matrix and the solvent used, the microscopic structure of the

preparations may vary between rather homogeneous amorphous layers, a dense packing

of tiny crystallites (e.g. sinapinic acid), or relatively large crystal needles grown from the

rim of the droplet (e.g. Figure 1-3, showing an SEM micrograph of cytochrome C in a

dihydroxybenzoic acid, DHB, matrix). The most intense ion signals are usually

associated with regions of the sample that appear crystalline, although this result depends

on the matrix substance and the presence of buffers or detergents in the analyte sample.

A key feature of MALDI first noted by Beavis and Chait is that commonly used

biological buffering agents, salts, and denaturants (for example phosphate, sodium, and













































































9ii ~~ 9i~`









urea) at concentrations up to several hundred millimolar do not have to be removed from
9-10
the sample prior to analysis. It is thought that crystallization of the matrix and analyte

molecules occurs but not of water-soluble contaminants, which may explain the

capability of some matrices to tolerate even a high excess of contaminants. This

tolerance to contaminants is in direct contrast to all other MS techniques presently in use

for the analysis of biological molecules. The ability to avoid purification prior to MS

analysis is highly advantageous because a great deal of time and material is often lost at

this step. As an example of the ability to analyze samples using MALDI without prior

purification, Nelson and Vestal of Vestec Corp detected 14 proteins present in the direct

analysis of a salivary sample (Figure 1-4). Although MALDI is highly tolerant of

many contaminants, sodium dodecylsulfate (SDS) contained in the sample can lead to a

total quenching of the protein ion production and must be removed prior to the sample

preparation.


Time-of-Flight Mass Spectrometers


MALDI has been used almost exclusively with time-of-flight (TOF) mass

analyzers, which can record all of the ions from each ionization event and thus offer the

possibility for higher sensitivity than scanning instruments. TOF mass spectrometers are

relatively simple, inexpensive instruments with high sensitivity and virtually unlimited

mass range. In a MALDI TOF mass spectrometer using a linear flight tube (Figure 1-

5),12 ions are formed in a short source region (s) by a pulse of laser light absorbed by the

sample. 3 A positive voltage (V) placed on the backing plate imposes an electric field (E

V/s) across the source region, which accelerates all of the ions to approximately the same

kinetic energy. As the ions pass through the extraction grid they have velocities which

depend inversely upon the square root of their mass. The ions then pass through a much


























ooo
v T



co



a


E
2


r ^r


0

S- 0












V.

















I*'


ClS
^r rt r)






10


0I







e
9H r.











.L.
I .1


g-o




I









0
d] v0
cC.









longer drift region (D, typically 0.5 to 2 m) and arrive at the front plane of the detector.

Because they spend most of their time in the drift region, their flight time measured at
V2
the detector is approximately t = (m/2zeEs) D. This time spectrum is then converted to

a mass spectrum using a mass calibration routine.

Results for the analysis of glucose isomerase (Figure 1-6) show the typically simple
+
spectra obtained when using MALDI, consisting predominately of [M + H] and [M +
2+ 14
2H] ions. Fragmentation of molecular ions due to the cleavage of covalent bonds in

the protein backbone is not apparent from the spectra. Structural information from

systematic fragmentations cannot be obtained directly (by fragment ion desorption), but

can be obtained by taking advantage of metastable decay reactions. The ionization step

seems to be extremely soft and no fragments are typically observed. The absence of

fragment ions and the dominance of singly and doubly protonated intact protein ions

make the interpretation of such spectra particularly straightforward.


Mass Resolution in MALDI TOF MS



In any mass spectrometer, mass resolution is defined as the separation of peaks of

different masses (Am) relative to the mass of interest (or more commonly, m/Am). The

rather poor mass resolution obtained in MALDI TOF MS of large compounds has been

realized as a drawback of the method from the very beginning. This leads to an

inability to resolve neighboring molecular ion signals in the high mass range and can lead

to poor mass accuracy. The precision attainable by mass determination is limited, since

peak broadening cannot be assumed to be absolutely reproducible, independent of

substance and homogeneous. The importance of mass accuracy and resolution is

demonstrated in Figure 1-7, where a mixture of human insulin and insulin mutants was






















0: .
V3 :


+



+


+ -






+


0>
.0 -





































C.
A -i





i i

o i









analyzed (Vestec Corporation, Personal Communication, 1993). Typical mass

resolutions of m/Am = 600-1000 for linear TOF systems are observed, but the resolution

degrades with mass (e.g. m/Am = 50 at m/z 75,000).

In a TOF mass spectrometer, peak widths (At) are determined by uncertainties in

the temporal, spatial, and initial kinetic energy distributions. Each of these will now be
17
discussed

Temporal distribution: A temporal distribution occurs when two ions of the same

mass and initial kinetic energy are formed at different times in the ion source (Figure 1-

8a). Because the ions have the same velocities, the time interval remains constant as they

exit the drift tube and are recorded by the detector. For MALDI, ions desorb off of the

probe surface within 20 gis.1

Spatial distribution: A spatial distribution occurs when ions of the same mass are

formed at the same time with the same initial kinetic energy, but are formed in different

locations in the extraction field (Figure 1-8b). Ions that are formed near the back of the

source will be accelerated to higher kinetic energy than ions formed closer to the

extraction grid. The ions formed at the rear of the source enter the drift region later, but

have larger velocities so that they eventually pass the ions formed closer to the extraction

grid and arrive at the detector sooner. For desorption techniques like MALDI, in which

ions are formed from a smooth conducting surface parallel to the extraction grid rather

than in the gas phase, the spatial distribution is greatly minimized.

Initial kinetic energy distribution: Ions formed with different initial kinetic energies

will have different final velocities after acceleration and arrive at the detector at different

times (Figure 1-8c). This is the most difficult initial condition to correct in a linear

instrument, since increases in drift length increase peak widths along with the separation

of peaks of different mass. Beavis and Chait9 determined the velocity and translational

kinetic energy distributions for polypeptide ions produced by MALDI. Using three












a). Temporal Distribution


Ato1 0o







b). Spatial Distribution

Aso 0


>- (> 0



c). Kinetic Energy Distribution
AU O 0
e+U G-

L~ (


Figure 1-8. Factors affecting the mass resolution in a TOF mass analyzer. 17









different polypeptides (angiotensin, M.W. 1030; insulin, M.W. 5730; and superoxide

dismutase, M.W. 15590), they determined that these molecules all had very similar initial

velocity distributions, independent of the molecule's mass, centered at approximately 750

m/s. Using this value as the most probable initial velocity, the average initial

translational kinetic energies for polypeptide ions are directly proportional to the

molecular mass



KEavg (eV) = 0.003 mass (u)


The initial translational kinetic energy distributions for the three polypeptide ions

studied by Beavis and Chait are shown in Figure 1-9. For insulin, the average initial

translational kinetic energy is 17 eV, with a distribution ranging from 5 eV to 25 eV.

This large spread in initial translational kinetic energy is the limiting factor in the

resolving power of a linear TOF mass spectrometer.

In order to deal with the initial translational kinetic energy distributions that lead to

a reduction in mass resolution in linear time-of-flight mass spectrometers, a different

geometry TOF MS was introduced by Mamyrin and coworkers in the early 1970s.20 This

instrument, called a Reflectron (Figure 1-10), incorporates a region containing both

decelerating and reflecting electric fields. When ions having the same m/z ratio but

different velocities enter the electric field region, those with a higher velocity (because of

their higher initial kinetic energy) will penetrate further into the decelerating field before

reflection than will those with lower velocities. By choosing appropriate reflection

voltages, the fast and slow ions will reach the detector together, which will optimize the

energy resolution.
























0


II?
E)













Cc
-JO


- w
Z C


~~-0c


I -- g

- (4%r ---
IL



^ '
<- -- ^


L//////-
s"^\\\\\'


-------- > >
O>


0


z
0

C)
w
-J
.LL


I I


















I,;










The Ouadrupole Ion Trap Mass Spectrometer



In 1953 Wolfgang Paul and his coworkers at the University of Bonn in Germany

first described both the quadrupole mass filter and its three-dimensional analog, the

quadrupole ion trap. While the quadrupole mass filter has become the most common

analyzer employed for mass spectrometry, until the 1980s the quadrupole ion trap was

only used by chemists and physicists to study trapped ions. The first commercial

quadrupole ion trap was introduced by Finnigan Corp. in 1984 as an inexpensive and

sensitive mass spectrometric detector for gas chromatography. This instrument was

based on a new method developed by George Stafford and his coworkers at Finnigan for

selectively ejecting ions of increasing mass-to-charge ratio from the ion trap, known as

mass-selective instability. The following discussion of the fundamental principles of the

quadrupole ion trap is adapted from March and Hughes, 1989 (unless otherwise noted).22



Fundamental Principles of Quadrupole Ion Trap Operation



The quadrupole ion trap (Figure 1-11) consists of three electrodes of hyperbolic

cross section: a central ring electrode and two endcap electrodes. The general

mathematical equations used for the construction of the ring and endcap electrodes for a

quadrupole ion trap are given by


1
-- (r- 2z2) = 1 (1-1)
2
r
0

1
(r2 22)= -1 (1-2)
2
2z
0










Q) Q)
0 -0
o 0

a) cz







1- E
w Q




V -

0 -

L E
o











,~I
T- 0 I
-- U )I
-7l- 4--^*
-^^a.ii^i | \\0
------ \\ 0
/ /^\ \\ '

^ r^ 2
T I T-i
i -- i 0c










where ro is the radius of the ring electrode and z. is the center-to-endcap distance. In the

case of a pure quadrupolar field, the relationship between the ring radius and the endcap

distance is given by


2 2
r =2Z (1-3)



Application of a radio-frequency (rf) potential to the ring electrode produces a

quadrupolar electric field within the ion trap, with the force on an ion within the potential

energy well formed by this field varying linearly with the ion's position from the center

of the device. The potential at the center of an ion trap is given by


U VcosOt
O -- (1-4)
2

where U is the DC potential and Vcosft is the rf potential (with V being the zero-to-peak

amplitude and M./2t being the frequency in Hertz, Hz). For an ion trap employing an

ideal quadrupolar trapping field, the potential at any point is given by




2 = i + (r- 2z2) (1-5)
2
r
o
Using equations 1-4 and 1-5 above, and setting the field strength at the exact center

of the ion trap to zero (in order to satisfy Laplace's equation), the resulting equations for

the motion of an ion in an ion trap can be given in terms of a radial (r) and axial (z)

displacement

d2r 2e
+ (U Vcosft)r = 0 (1-6)
2 2
dt 2mr
0










d2z 4e
+ -- (U Vcosft)z = 0 (1-7)
2 2
dt 2mr
0


where m is the mass of interest and t is the time variable.

These two equations of ion motion are in the form of linear second-order

differential equations, and the Mathieu equation can be used to obtain a solution which is

given in terms of stability parameters a and q


-8eU
a = az = -2ar = (1-8)
m(r2 + 2z 2)2


-4eV
qu = qz = -2q, = (1-9)
m(r 2 + 2z 2)2



These two equations are typically given as a and qu, since analogous equations can

be derived for ar and qr for ion motion in the radial direction. It should be noted that both

Mathieu stability parameters are inversely proportional to the m/z for the ion of interest,

and that a is proportional to the DC voltage and qu is proportional to the rf voltage

applied to the ring electrode. The range of au and q values that give rise to stable ion

trajectories in both the radial and axial directions can be expressed in terms of a Mathieu

stability diagram (Figure 1-12). It is evident from this diagram that there is a region

where ions are stable with respect to radial motion in the ion trap and a region where ions
are stable with respect to axial motion. When an ion has values of q and au that fall

within the area of the diagram where these two regions overlap (the shaded portion of the

stability diagram shown in Figure 1-12), it is stable within the body of the ion trap and is















03
ei


ed
-4
0





0






o .


a








.0
. .

















'-








o
... 1 0















o o
co












0








NN


CD C
Si
0 O




S--









stored. Because of the inverse relationship between m/z and qu, more massive ions will

have a lower q value than less massive ions and will appear towards the left side of the

stability diagram.

Mass analysis using the mass-selective instability mode of operation in the ion trap

is given as

4V
P-P
(m/z) = (1-10)
2 2 2
(r2 +2z2) f2 q.j
eject


where m/z is the mass-to-charge ratio of a given ion, V is the amplitude of the rf voltage
applied to the ring electrode, ro is the internal radius of the ring electrode, zo is the

distance from the center to the end-cap electrode, C is the angular component of the rf
drive frequency, and qeject is the point on the Mathieu stability diagram where ions

become unstable under the influence of the rf-only field and are ejected into the detector.

After a period of time where ions are created and stored in the ion trap, the amplitude of
the rf voltage applied to the ring electrode is ramped from approximately 100 V0 to

7500 V0-p in 110 ms (using a nominal scan rate of 5550 Da/s, scanning from 40 Da to

650 Da). As the amplitude of the rf voltage is increased, the qz values for all ions stored

in the ion trap will increase. Once the qz value for any given ion exceeds 0.908 (which

occurs from low m/z to high m/z in increasing fashion), the motion in the z-direction will

be unstable, with one-half of the ions hitting the top endcap and one-half exiting through

holes in the bottom (or exit) endcap, where they are detected by an electron multiplier.

This is shown in Figure 1-13 as a scan function, which is a plot of the rf voltage versus

time.




















cd
4" E
Ser

CE

0 o









C L

o
c u




mrd
a o
ed a
3 U






-a5
o




sO




.- I

V Cd
: E
03
i,















cg u,












IO
L4


(O 0
:0
I- B-
1^>|








Resonance Excitation and Resonance Ejection


When Finnigan Corp. introduced the Ion Trap Detector (ITD) in 1983 as an

inexpensive and sensitive mass spectrometric detector for gas chromatography, a 1.1
MHz (i/2r), 0 Vop to 7500 Vo0P rf potential was placed on the ring electrode with no DC

potential (au = 0), and the endcap electrodes were grounded. A significant advancement

was made in the operation of the ion trap with the introduction of the Finnigan Ion Trap

Mass Spectrometer (ITMS) in 1985. In this system, a DC potential could be applied to

the ring electrode along with the rf potential, which allows for mass-selective isolation.

In addition to this, a dipolar field could be created by the application of a low-voltage
supplemental rf signal (530 kHz, typically 3 Vp to 6 V ) placed 180 out of phase

across the endcap electrodes. For a fixed set of operating conditions, ions of each m/z

have characteristic frequencies of motion in the axial and radial directions. When the

amplitude of the rf placed on the ring electrode is ramped, ions of increasing m/z are

brought into resonance with the supplemental rf signal applied to the endcaps. As the

ions come into resonance with the supplemental signal, they absorb power (termed
"resonance excitation"), resulting in an increase in the amplitude of their axial

trajectories. If a very low voltage is used (typically 50 to 500 mV), the increase in ion

trajectory may result in fragmentation due to energetic collisions with background gases

(collision-induced dissociation, CID). When a much higher voltage (typically 1000 to

6000 mV) is used, the amplitude of an ion's axial trajectory will become so great that it

will exit the ion trap and be detected by the electron multiplier. This is termed
"resonance ejection" or "axial modulation".










Mass Range Extension in the Quadrupole Ion Trap Mass Spectrometer



For a quadrupole ion trap operated in the mass-selective instability mode, the

maximum mass that can be detected is given by



4V
max
(m/z)a = (1-11)
(ro + 2zo ) eject



where Vmax is the maximum radio frequency (rf) voltage that can be applied to the ring

electrode. Adding conversion factors to present this equation in terms of amu, volts, cm,

and MHz gives the following expression for the maximum mass:

2 219 26 1
(4 kg m2 s-2 C')* (1.6022 x 1019 C) (6.022 x 1026 mol) V
(m/z) =
(r. (cm)2 + 2(zo(cm2)) 104 m2/cm ((n (MHz))2 472 x 1012)(Hz2/MHz2 ) q


(0.09776) Vm
mmax
(m/Z)max =- (1-12)
(ro (cm))2 + 2(zo (cm) 2)) *( ( (MHz))2 jt



Using the operating conditions for the commercially available Finnigan ITS40 mass
spectrometer (Vmax = 15000 V p, ro = 1 cm, z = 0.792 cm, D/27n = 1.0485 MHz, qeje =

0.908), the resulting maximum mass that can be detected is 651.6 amu:


(0.09776) 15000 V
(m/z)max = = 651.6 (1-13)

(1 cm)2 + 2(0.792 (cm) 2)) (1.0485 MHz)2 0.908









It must be noted that this is not the maximum mass that can be stored within the ion

trap, but rather the maximum mass that can be ejected during a mass-selective instability

scan and detected. As an example of this, 20 gm diameter aluminum particles were

trapped by Wuerker, Shelton, and Langmuir (Figure 1-14 and detected using a camera.2

It was calculated that the relative molecular mass of these particles was 6 x 1015 and that

there were approximately 3 x 105 charges on these particles, giving an m/z value of 2

xl010

Using equation 1-12, the mass range of the ion trap can be extended, therefore, by

any one of four ways

(i) by increasing the maximum rf voltage, Vmax,
(ii) by adjusting the dimensions of the trap (reducing ro),

(iii) by decreasing the rf frequency, Q, and
(iv) by selecting a different point on the stability diagram, qej,,, to cross from

stability to instability rather than the normal qeject value of 0.908.



Each of these methods for extending the mass range of the ion trap shall be now be

discussed.


1) Operating at higher values of rf voltages (Vmax).

The current maximum rf voltage used for the ion trap is 15,000 V P. Any attempt to

increase the mass range accessible to the ion trap by increasing this voltage would result

in severe arcing between the ring and endcap electrode. This method has never been

attempted.


2) Reducing the size of the ion trap (r, zo).

Kaiser et al. evaluated the use of smaller ion traps to increase the accessible mass range.24

The electrodes used for these studies conformed to the theoretical calculations for the








29
















d.
en


0



2
a







C4
cl











E


0q
CO

















0
Cd



0







*-
cm

ipl









2 2 2 2 2
hyperbolic surfaces and spacing (r 2z = ro for the end-cap electrodes and r 2z = -
2 2
r2 for the ring electrode, and the spacing between the end caps and the ring electrode r2

= 2Zo2), with the dimensions scaled to produce an ion trap at half-, third- and quarter-

sizes. Data produced using desorption ionization and ion injection using the reduced size

ion traps showed that although the mass range was increased as expected, artifact signals

of low intensity were evident. These artifact peaks were believed to be due to ions

ejected prematurely because of imperfections in the rf field caused by the holes in the ion

trap electrodes (which were not reduced in size).


3) Reduction of the rf drive frequency (n).

In addition to changing the standard crystal used to generate the waveform for the rf

drive frequency, the air core coil (which is the tuned rf transformer on this system) has to

be modified to reduce the rf drive frequency. This transformer is essentially a tuned LC

circuit with a resonant frequency given as



f= 1/27r (LC)" (1-12)


where L is the inductance of the air core coil and C is the tuning capacitance of the

system (which includes the capacitance between the ion trap ring electrode and the

endcaps, the capacitance of any capacitors placed between the transformer coil's

secondary tap and ground, and the stray capacitance of the coil assembly). Since it is

extremely difficult to reduce the tuning capacitance of the system (C), the inductance of

the transformer coil (L) needs to be substantially higher to reduce the resonant frequency

to match the desired output frequency. In the work by Kaiser et al., a special coil

consisting of a smaller diameter coil mounted inside and connected in series with a

standard coil was used. This coil, used in conjunction with a lower frequency crystal,









allowed for an rf drive frequency 1.2 x lower (0.921 MHz), which led to an increase in

mass range by a factor of 1.7 times. Since a reduction in the rf drive frequency (n)

affects both the actual frequency placed on the ring electrode and the rf control DAC, the
increase in the mass range is proportional to (n / n ) where n is the drive

frequency on the commercial ion trap system (1.0485 MHz) and .new is the modified

drive frequency.


4) Reduction in qejet.

The last method to increase the mass range of the ion trap is to lower the q at which ions

are ejected from the ion trap. This is accomplished by the use of axial modulation. As

described above, a dipolar field is placed 1800 out of phase across the endcap electrodes.

As the ions come into resonance with this supplemental signal, they absorb power and

their trajectories become axially unstable, at which point they exit the ion trap and are

detected by the electron multiplier. At the standard axial modulation frequency used on

the Finnigan ITS40 ion trap system (485 kHz), the mass spectral peak shapes are

dramatically improved (particularly under space charge conditions).25 Another

consequence of this field is that the observed mass for a given ion decreases (e.g. the m/z

502 ion of perfluorotributylamine appears at m/z 498 in an uncalibrated system). Instead
of the ion being ejected when it reaches the stability boundary (qz = 0.908, shown in

Figure 1-15 and the corresponding scan function with detected ion signal, Figure 1-16, it
is ejected early (qz = 0.906 for the Finnigan ITS40, corresponding to point A in Figure 1-

17 and as a scan function and detected ion signal in Figure 1-18. When the q point is

further lowered (e.g., qet = 0.454, corresponding to point B shown in Figure 1-17, the

ions are ejected much earlier than at the original qj, of 0.908, and are hence ejected at a

much lower rf voltage (Figure 1-19. Since the rf voltage can still be ramped to its
maximum value of 7500 V0p a much higher maximum mass can be ejected. This effect



























0

C)
d

C


oe


.o
6-












--






o
-t


0
0
ed
















o








-t
O



. 0
cr












-0



















1.
50








*e
4-*











Ca





0



C
0
E







co


S .0


co
() o







S0o
a, o










0A0
C \ -






CCd
o o \__ ______





e- A eylO o plew Budde== =11 j


0LL
~~BO
(doA e~~ioAP~eI 6Udd~ ~i



























o 0 a 0
I E E E


N
cr


(L
I-


o






















E
(u


























c0
o












0
o
rA

















mo
ECr
ce C
-o


. b
cI




Ca
o ,






c"a













I-~

























C)

03











C,






0
0


C 0T
-- O









rr
o) I
OC ;




h CO 10


0A) e6illOA pl9 BuI !ddBj _-Iu


0
I-


0
0

0

o
0
r-
0
O


E


.0








Un
o0
4-


" 3
X:







-o
E2<


0 3
-.a
O- O

s 3

0.



c. 0


o






C.)


co
1gr


I -


1-


IAE
1 s
Co
(0










_o

CO




E E





N N
o o m
II-
S.oa
0













O' E
I-


O m




CJ -, <3
0) I


ac 1
O Q i i i i i i i i j I .j i




S'oA) e-eilOA ple- Buudd.Jl -IE -
o s


I
.2
0 -P
0









is even more exaggerated at a lower qej, (e.g., qeject = 0.0908, corresponding to point C

in Figure 1-16, where the mass range is extended by a factor of 10 since the q,,,t has

been decreased by a factor of 10, shown as a scan function and detected ion signal in
Figure 1-20 The mass range is effectively increased by this amount, and the new mass
limit is given as


(m/Z)new max = (m/Z)old max [(ld eject)/ qnew eject)] (1-13)


The axial modulation frequency (o) required to achieve a desired qeject cannot be

calculated directly, but must be obtained through P, which is related to the frequency of

an ion

a = 3 2/2 (1-14)


P is a complex function of a and q which can be approximated using the relationship

2 2 2 2 2 2 2
S= a + [q2/(2+p) -a-[q/(4+p) -a-[q/(6+p) -a-[...etc.]]]] +
[q2/(2-3) 2-a-[q /(4-3) -a-[q /(6-P) -a-[...etc.]]]] (1-15)


The relationship between the axial modulation frequency used during the mass analysis
scan and the maximum mass detectable is shown in Figure 1-21.



MALDI Using a Ouadrupole Ion Trap Mass Spectrometer


Mass analysis of MALDI-derived ions is most often carried out with time-of-flight
(TOF) mass spectrometers due to their ability to record an entire spectrum for every
single laser shot, as well as their high mass range and simplicity. As discussed earlier,






















































I I I I I I I


00
0C
0)
0


II
11
05
CD
-o*






Cu
0


C,


0)

V)


* *


I I I I I I I I
0 0O CM



-oA) BeIOA pl|i9 BU!ddleji L -j


0




0

4s
0


E








C4...
0
o 0

o
OU
in












0
,. 0


0o
0
E


,0,
cn



o












oU
o0U
Coi





0 60

Ci
03


















I4


E
o
s g
CO,


=t -5
-----


I
















6500

6000

5500

5000

4500

4000

3500

3000

2500

2000

1500

1000

500


Slo5x





5x


200 400
Axial Modulation Frequency (kHz)


600


Figure 1-21. Relationship 1,itween the axial modulation frequency used during the mass
analysis scan and the maximum detectable mass.









although the packet of ions produced by MALDI is well-suited for TOF mass analysis,

the initial kinetic energy spread of the ions leads to poor mass resolution. Typically,

mass resolutions of between 400 and 700 (m/Am, full-width at half-maximum, fwhm) are

observed for bovine insulin (M.W. 5733.5) using a TOF mass spectrometer with a linear

flight tube and between 1000 and 3000 using a reflectron.2 Another factor which

degrades the mass resolution in MALDI TOF MS is metastable decay along the flight

tube. This is particularly important when using a reflectron instrument, where it can
27
seriously degrade peak widths and sensitivity.

In order to overcome some of the disadvantages of using a TOF instrument with

MALDI, a quadrupole ion trap mass spectrometer (QITMS) has been used with MALDI.

Whereas the detection and ionization processes are coupled when using a TOF mass

spectrometer with MALDI, when using a QITMS the MALDI-generated ions are stored

inside of the ion trap prior to mass analysis.

The first attempt at using a QITMS for the mass analysis of MALDI-derived ions

was shown by Heller et al. in 1989.28 A Finnigan Corp. (San Jose, CA) ion trap detector

(ITD) was modified by drilling a hole through the center of the ring electrode. An

insertion probe lock was fitted to one side of the vacuum manifold and a flange with a

ZnSe lens was fitted to the opposite side (Figure 1-22). MALDI was performed within

the ion trap by placing a matrix/sample mixture on the tip of a stainless steel probe,

inserting the sample probe through the probe lock and into the hole in the ring electrode

so that it was flush with the surface of the ring electrode, and bringing in a laser beam

through the other side of the ring electrode.

Using a design similar to this, Chambers et al.29 showed the analysis of 163 pmol of

bovine insulin (M.W. 5733.6) (Figure 1-23). Although the molecular ion for this

compound can be seen, the sensitivity and resolution are very poor. This can be

attributed to two factors: (i) neutrals produced by the laser desorption process are present










41
















I-
N



-o


eS













4I.
0










/2
3.










o











42













0


.0 0




o


0















0











Sc-












00<
00

c4.
m am


mS
mc




-o

'
c~N









within the ion trap during the storage of the ions and (ii) the trapping efficiency of the

MALDI-produced ions decreases with increasing mass due to the initial kinetic energy of

the MALDI-produced ions. If an ion is produced with an initial kinetic energy greater

than the potential well created by the radio-frequency trapping field inside of the ion

trap, the ion will travel across the ring electrode without being trapped. Chambers et al.

calculated a high-mass limit of m/z 9830 when performing MALDI inside the body of

the QITMS.30 This leads not only to an upper mass limit but also to a loss in sensitivity

as the initial kinetic energy of an analyte increases (which increases linearly with mass).

This same instrumental configuration was used by Vargas and Yost31 at the

University of Florida for the analysis of spiperone, an antipsychotic drug developed by

Bristol-Myers Squibb. Both laser desorption ionization (LDI) and MALDI were

performed on this compound to demonstrate the differences between these two

techniques. LDI was performed by applying 2 IpL of a 0.1 p.g/tpL spiperone solution onto

a probe tip. MALDI was performed by depositing 2 iLL of a 0.1 mM spiperone solution

onto a probe tip followed by an equal volume of a 100 mM nicotinic acid MALDI matrix

solution. After air drying, the samples were analyzed by LDI (Figure 1-24a) and

MALDI QITMS using a nitrogen laser (Figure 1-24b). The major ions seen in the LDI

spectrum for this compound are low molecular weight fragment ions. As discussed

earlier, the high irradiance required for nonresonant desorption (approximately 107 W

cm2) is very close to the point of plasma generation, which leads to fragmentation of the
+
molecule. MALDI, however, produces mainly [M + H] ions, along with a small

amount of adduct ions and fragment ions.

In addition to performing MALDI of spiperone, Vargas and Yost also performed

tandem mass spectrometry on the MALDI-produced ions of this compound. Figure 1-25

shows the results for the MALDI MS/MS analysis for the m/z 396 ion of spiperone. The

results for the tandem mass spectrometric analysis of this compound using MALDI were







A 1000-



800-


O


165


98


1


123
!l


M'
395


>\ 600-
-t--



C
400-
c


200 -



0-
5

4000




3000-
X-



S2000
L-

C


1000




0


'00 150


ii, II I a.


200 250 300 350


S40 I 4 T ;5
400 450


(M+H)T
396


124


165


100 150 200 250 300 350


400 450


Figure 1-24. a). Laser desorption ionization (LDI) and b). MALDI QITMS spectra of
spiperone performed inside of the ion trap.31


ftt'ruik


.-v l l A j Ji.l..__ t klll^ ^. Al .i j, ,l.. ^^ 1 1 A. 1- ..j I.,,11


i I I I


L


I. ,I L.1 ..






















600





400-
>,

U)








0 --
150


65


292


232


I ,l JI,. I


262


P IFI I I I ll-


175 200 225 250 275


m/z


380


39


7


300 325 350 375 400
300 325 350 375 400


Figure 1-25. MS/MS daughter spectrum of protonated spiperone.31


1 1 .


r


I I









similar to those obtained using other methods of sample introduction and mass analysis

(e.g., solids-probe CI).32
Vargas and Yost also demonstrated the MALDI QITMS analysis of spiperone in a

Matrigel matrix. Matrigel (Collaborative Research, Inc.) is mouse extracellular tissue

used as a model biological matrix. 0.5 to 1 mg of spiperone was mixed in 1 mL of

Matrigel, and 4 upL of this solution was deposited onto a probe tip followed by an equal

volume of a 100 mM sinapinic acid MALDI matrix solution. After air drying, the

samples were analyzed by MALDI QITMS (Figure 1-26). Unlike the MALDI QITMS

analysis for spiperone analyzed with the matrix alone, addition of the Matrigel produced

intense sodium and potassium adduct peaks, due to the presence of these cations in the

Matrigel.


MALDI OITMS Using an External Ion Source


MALDI QITMS has also been performed using an external ion source. The

advantages of using an external ion source instead of performing MALDI directly inside

the ion trap are increases in sensitivity and resolution by not having the neutrals produced

by the laser desorption process present in the ion trap during the storage of the ions. In

addition, there is no upper mass limit due to the initial kinetic energy of the MALDI-

produced ions.

The first MALDI QITMS instrument designed with an external ion source was
33
constructed by Bier et al. of Finnigan. They modified a GC/MS ion source so that a

nitrogen laser beam at 337.1 nm was brought through a fiber optic and focused onto a

sample probe (Figure 1-27). Desorbed ions were extracted from the source and injected

into the ion trap for mass analysis. Using this design, they showed the analysis of 50

















250-
165

200


>150-
150- (M+Na)
1n :418

100-
C -
98
so- 123 (M(M+K)
(M+H) 434
396

., .,. 1


50 100


250
m/z


300 350


Figure 1-26. MALDI QITMS analysis of spiperone in a Matrigel matrix.31







48












0



0



*O
o
E

el







ca
E-



C










r.
C





v,



_0


Wc



Ib









pmol of bovine insulin (M.W. 5374) (Figure 1-28) and 600 pmol of egg albumin (M.W.

43300) (Figure 1-29). Although this instrument showed good results for the MALDI

QITMS analysis of bovine insulin, the spectrum for the higher mass egg albumin showed

a very poor signal-to-noise ratio and poor resolution.

At the same time that the MALDI QITMS instrument described in this dissertation

was being designed and built at the University of Florida, the Chait group at the
34
Rockefeller University was constructing a similar instrument.34 This instrument consists

of a sample probe plate, source ion extraction plate, three tube lenses, and the quadrupole

ion trap with a channeltron detector fitted with a conversion dynode (Figure 1-30).

Finnigan ITMS electronics were used to control the QITMS, with the addition of a scan

rate reduction board (Ionstor, Oak Ridge, TN) for increased mass resolution experiments

and an rf amplifier (Ionstor) used to increase the amplitude of the supplemental ac field

for mass range extension. Results from this system included the MALDI QITMS spectra

for 100 fmol of bradykinin (M.W. 1060.2) (Figure 1-31). As is typical with other ion
traps used for MALDI analysis, fragment ions (typically seen as a loss of H20 and NI3,

and loss of the C-terminal residue, in this case the amino acid arginine (R)) created upon

injection are evident.


Overview of Dissertation


This chapter described the development, mechanism, and application of MALDI

using both time-of-flight and ion trap mass spectrometers. Chapter 2 describes the

design considerations for the MALDI QITMS instrument built for this dissertation,

including simulations of both the modified-GC ion source MALDI QITMS instrument

used by Bier et al. of Finnigan and the Vestec MALDI ion source MALDI QITMS

instrument designed and constructed for this dissertation. The data for the two peptides












(M+ H)+


B Chain










(M + 2H)2z


5x


Figure 1-28. MALDI QITMS spectrum of bovine insulin using an external ion source.33









I LdL.I


(M + H)+


























10000 20000 30000 40000
m/z


Figure 1-29. MALDI QITMS spectrum of egg albumin using an external ion source.33












































6.

U) i
-J


0
L.


(E
to

U)
3


__


i


I I














4-.













CT





-0 -
o.


.0,






U
0 '
O 2











0
o U








o a
o -8






I-i
C.-











-0
U:
^___________= ==== = _0









(MRFA and angiotensin II) analyzed for the characterization of this instrument are also

included in this chapter. The MALDI QITMS instrument described in chapter 2 of this

dissertation was used for the analysis of two pharmaceutical compounds: spiperone, an

antipsychotic drug developed by Bristol-Myers Squibb, and the anticancer drug Taxol.

Chapter 3 contains the results for the MALDI QITMS analysis of spiperone using MS

and MSn. A fragmentation map for this compound was generated using multiple stages

of mass spectrometry, up to MS Also contained in this chapter is the MALDI QITMS

analysis of Taxol in an incubated rat liver tissue sample. This last analysis was

performed to demonstrate the ability of MSn for the detection of target compounds in the

presence of a large number of interfering species. Chapter 4 concludes by summarizing

the research described in this dissertation, and discusses the future work needed to use a

MALDI QITMS instrument for the analysis of tissue samples.














CHAPTER 2
INSTRUMENTAL DESIGN AND CHARACTERIZATION OF A MALDI
QUADRUPOLE ION TRAP MASS SPECTROMETER


Instrument Design


As discussed in the background section of this dissertation, the mass analysis of

MALDI-generated ions has been demonstrated using a quadrupole ion trap. Successful

instrument designs have been developed around the use of MALDI-generated ions

produced inside of the ring electrode of the ion trap and using an external ion source

along with ion injection into the ion trap. Since one of the ultimate goals of the research

undertaken as part of this dissertation is to produce a microprobe instrument, imaging of

the sample is critically important. Although an instrument which produces the MALDI-

generated ions directly inside of the body of the ion trap can be used, the use of an

external ion source allows for easier sample imaging and the potential to use a

translatable sample stage (using an x-y positioner coupled to the sample probe). Because

of these important features, a MALDI ion source externally coupled to an ion trap was

the instrument configuration chosen for this project.

Although the limitations of using internally generated MALDI ions has been

discussed by Glish et al., the use of a modified GC ion source to produce MALDI ions

also leads to instrumental compromises. These limitations have a direct impact on the

choice of the MALDI ion source that will be used for the research covered in this

dissertation. In order to understand the limitations in using a modified GC ion source to

produce MALDI ions, a study of the ion optics of the system used by Bier et al. was









performed using version 5.0 of the ion simulation program SIMION.3 Accurate

dimensions were taken of a Finnigan 4500 (TSQ70) ion source and a Finnigan ITS40

quadrupole ion trap, and these were used to generate a SIMION grid file. The voltages

used to produce the MALDI QITMS data shown in Figure 1-27 and Figure 1-28 were

supplied by Mark Bier (with no rf voltage placed on the ion trap and no collisions used

for these simulations). Most probable velocities of 750 m/s (as described by Chait20)

were used for these two different mass ions, which gave initial kinetic energies of 17 eV

for bovine insulin and 126 eV for egg albumin. The results from the SIMION analysis of

the modified GC ion source instrument used to produce the mass spectra of bovine

insulin and egg albumin are shown in Figures 2-1, 2-2, and 2-3.

It is evident from Figures 1-27 and 1-28 that the resolution and signal-to-noise ratio

(S/N) for egg albumin are very poor compared to that of bovine insulin. The poor

performance of this system for the higher m/z compound is a direct result of the use of a

modified GC source for performing MALDI. Even though this source had been

modified (the lens spacers were better anodized to allow for higher voltages before

arcing), only a limited range of voltages could be used to focus ions from the region

where the nitrogen laser strikes the sample to where they enter the ion trap. The

consequences of this are seen on ions with a significant radial velocity (that is, ions

leaving the probe tip at an angle with respect to the ion trap. When the ion beam leaves

the probe tip at an angle of 0 (directly towards the QITMS), the ions enter the ion trap

regardless of mass (Figure 2-1). In MALDI, however, only a small percentage of the

ions generated leave the probe tip at this angle.36 Since this source is designed to be used

with relatively low voltages, ions produced with radial velocities can not be well focused.

Figure 2-2 shows the SIMION results for egg albumin when the ion beam leaves the

probe tip at a 10 angle (corresponding to an ion leaving the probe tip with only a slight

radial velocity). When compared to the 00 angle case (where the ion leaves the probe tip









57




F(U
E-
(e







. :

COs













.o


o oo


t5
o



4)
-o
C3

























o r

.5 c













.-





> M






oa



mo























Cl
u-.














tij
8




Q E




3 ..
'<3 C
a 5
3 S












.<



a-





0 06
.c





"0





















OC
03







> ?
















c o









c0


a *,











a
L.0
>


r-









with no radial velocity) as shown in Figure 2-1, there is a large loss in sensitivity due to a

lack of focusing of the ion beam caused by the radial velocities present in the 1 angle

case. In the case of a lower molecular weight molecule such as bovine insulin (Figure 2-

3), some focusing does occur, leading to a much more intense mass spectrum. This is

reflected in the data presented in Figure 1-28.

The SIMION results for the use of a modified GC ion source as a MALDI ion

source shows the limitations in this type of instrument configuration (that is, limited ion

source voltages due to the lens placement in the source). This type of source also is very

difficult to operate, since the laser beam intensity is affected by the distance the fiber

optic is inserted in the ion source and the distance from the probe tip to the end of the

fiber optic. Sample imaging is also very difficult because of the compact size of the ion

source.

For the MALDI quadrupole ion trap that was to be constructed for the research to

be performed in this dissertation, a MALDI ion source taken from a commercially-

available time-of-flight mass spectrometer (a Vestec ResearcH" linear time-of-flight

mass spectrometer) was chosen. This ion source was specifically designed for the

MALDI analysis of proteins and polymers, and should perform this function much better

than an ion source originally designed for another purpose (e.g. gas chromatography). A

schematic of the original Vestec time-of-flight system is shown in Figure 2-4. The final

design of the MALDI QITMS instrument is shown in Figure 2-5.

The vacuum chamber from the Vestec ResearcH' was used for this instrument.

The original Vestec MALDI sample probe and air-actuated probe inlet mechanism were

removed from this chamber, and a standard Finnigan probe and bellows valve assembly

were added to the flange that holds the MALDI ion source and ion trap. The much larger

probe tip on the Finnigan probe allowed for easier placement of tissue samples on the



















































d

0) 0
+4-
L


CU

U 0
O-o
U



0)
U

:3
0
C/)
c
O-



.-J

2:





62











I







2
or`\
E-
Oa

0



/~ -









surface of the probe, which allowed for a greater choice of the site where MALDI will be

performed on the sample (since the site that the nitrogen laser will strike the sample can

be controlled by the laser optics). Four stainless steel rods were added to the flange as an

optical rail to facilitate alignment of the MALDI source and the ion trap. The flange

opposite the laser window flange was modified to allow for the addition of an rf high

voltage feedthrough for the ion trap ring electrode, a gas feedthrough for the helium

buffer gas needed for cooling of the ions in the quadrupole ion trap, and four single-

ended MHV feedthroughs rated to 1 kV (Insulator Seal Incorporated, Hayward, CA).

Two of the high voltage feedthroughs were used for the axial modulation signal placed

on the endcap electrodes of the ion trap, and two were used for the entrance and exit tube

lenses of the ion trap. The high voltage feedthroughs that are standard on the Vestec

system (rated at 35 kV) were used for the sample lens and focusing lens of the MALDI

ion source. The flight tube and MCP detector used for the MALDI TOF instrument were

removed, and the vacuum chamber that attaches to the back of the MALDI source

vacuum chamber was modified to allow for the mounting of a 20 kV off-axis

conversion dynode detector. A Finnigan ITS40 quadrupole ion trap with quartz spacers

was used for this instrument, with the electronics from the ITS40 used to control the ion

trap and detector.

The software used to control the quadrupole ion trap was Gatorware software,

created by Tim Griffin of the Richard Yost research group at the University of Florida.3

This software allows for the creation of the scan functions used for all of the MALDI

QITMS analyses shown in this dissertation. Major features of this software are the

ability to control axial modulation frequencies and voltages during each step of a scan

table (allowing for software control of the mass range extension and CID parameters),

the ability to use a TTL pulse to control the firing of the nitrogen laser, and the ability to









acquire profile data (as uncorrrected mass, acquisition DAC, and intensity) to disk for

later processing (to allow for signal averaging and mass assignment).


Instrument Simulation


Before this instrument was constructed, the proposed design was simulated using

SIMION. For this simulation, bovine insulin (M.W. 5734) was chosen as the model

compound, and only the transmission of MALDI-produced ions from the ion source into

the quadrupole ion trap was considered (that is, no rf was placed on the ring electrode to

study the trapping of the injected ions). The results for this simulation are shown in

Figure 2-6, with the voltages used displayed above the lenses. The grid spacing for the

simulation was set to 0.05 mm/grid (which allowed for the maximum resolution

available), and an initial kinetic energy of 17 eV and an initial angle of 10 degrees were

used. From Figure 2-6 it is clear that this source, used with much higher lens voltages,

should be much better at focusing ions produced with a greater radial velocity (which is

typically the case for MALDI-generated ions). This simulation demonstrated the

feasibility of this instrument design.


Instrument Characterization


In order to characterize the performance of the MALDI QITMS instrument

constructed for this research, a series of peptide standards were analyzed using a

dihydroxybenzoic acid (DHB) matrix. A variety of peptides were chosen because of the

mass range they covered and their availability: the synthetic tetrapeptide MRFA (M.W.

523.65); Angiotensin II (M.W. 1046.2); Bradykinin (M.W. 1060.2); Angiotensin I

(M.W. 1296.5); Substance P (M.W. 1347.6); Dynorphin B (M.W. 1570.9); and Melittin





I I









(M.W. 2847.5). Two of the peptides analyzed during this characterization are presented

in this dissertation as being representative of the ability of the MALDI QITMS

instrument to analyze these types of compounds.

Primary standards for these peptides were prepared at lxi 0 M in deionized water.

Working standards were prepared at 2x105 M in deionized water. The DHB matrix

solution was prepared by adding 0.125 g of DHB to 1 mL of 50:50 (v/v) deionized water

and acetonitrile. Equal amounts of the working peptide standards and the DHB solution

(10 4L each) were mixed in a 100 pL microcentrifuge vial, and 1 pL of this mixture was

placed on the mass spectrometer probe tip and allowed to air dry.


MALDI OITMS of a synthetic tetrapeptide (MRFA)


MALDI quadrupole ion trap mass spectrometry was first performed on the

synthetic tetrapeptide MRFA (Met-Arg-Phe-Ala) because it has a molecular weight

(523.65 amu) accessible to the standard quadrupole ion trap instrument. During this

stage of instrument characterization, proper lens voltage settings were obtained by

varying the voltages and observing the effect they had on the MALDI QITMS spectrum

for this compound. The settings determined for MRFA worked equally well for all of the

compounds analyzed for this dissertation. The setting used are


L 1 (sample plate lens) = +20 V

L2 (focusing lens) = -700 V

L3 (ion trap entrance lens) = -63 V

L4 (ion trap exit lens) = -500 V









For all of the MALDI QITMS spectra acquired for this dissertation, no offset

voltage was applied to the ion trap (that is, the ion trap was not floated), and an electron

multiplier voltage of -1600 V and a dynode voltage of -3 kV was used. The helium

buffer gas pressure was set to 2.5 x 10-5 torr (uncorrected for helium) as read on a

Bayard-Alpert type ion gauge (Granville Phillips) mounted above the diffusion pumps

used to achieve the high vacuum needed for this system. This corresponds to

approximately 1.8 x 10- torr within the ion trap when conductance limits and ion gauge

correction factors are accounted for.

The MALDI quadrupole ion trap spectrum for 20 pmol of MRFA is shown in
+
Figure 2-7. This spectrum shows the [M + H] ion at m/z 524 as the base peak along
+ +
with an intense [M + Na] ion typical of biopolymers. Also present is the [M + CO] ion

which is routinely seen in MALDI mass spectrometry and is produced from attachment

of photofragments from the matrix This spectrum shows the unit mass resolution

attainable by using a quadrupole ion trap as the mass analyzer for performing MALDI

analyses. The presence of adduct peaks in the mass spectrum seen here is typically

observed in MALDI mass spectra, and requires excellent mass resolution in order to

determine correct mass assignments. The mass resolution required to resolve these mass

peaks, especially in the low mass region, is very difficult for a time-of-flight mass

spectrometer to achieve. A comparison of the analysis of MRFA on the MALDI QITMS

instrument built for this dissertation and on a Finnigan LaserMAT linear time-of-flight

mass spectrometer (which has a 0.5 m flight tube) is shown in Figure 2-8. Whereas the

TOF instrument gives a broad peak at m/z 524 with a high-mass tail, the QITMS

instrument resolves these peaks into the [M + H]+ peak, the 13C isotopes associated with

the [M + H]+ peak, the [M + Na] and [M + K]+ peaks, and the [M + CO]+ adduct peak.

The scan function used to produce this spectrum is shown in Figure 2-9, and is

listed in Table 2-1. At the start of the scan function, the rf level is set so that the [M +






























+
5 l


A!sueu I


r, NNTo8-8






55

50-

45


20

15-


500
700-


600-


500-


400-


300-


200-


100-


0-


500
500


524


I
520


540


560 Mass (m/z)


S[M + Na]+
546

[M+CO]+
551


540
540


560
560


m/z


Figure 2-8. Comparison of the MALDI analysis of MRFA on a). a Finnigan LaserMAT
TOF MS instrum-nt and b). on the MAT DI QITMS in "-mcnt constructed
for this research.


! .


1












7500-

Mass Analysis Scan
6000-


^ 4500


3000
Eject Low Mass Ions

1500


0 I I I I I I I I I I I I I I I II



200 -
100 -









Figure 2-9. Scan function used for the MALDI QITMS analysis of the synthetic




tetrapeptide MRFA.
e 6^

0 50 100 150 200
Time (ms)



Figure 2-9. Scan function used for the MALDI QITMS analysis of the synthetic
tetrapeptide MRFA.








Table2-1. Scan table listing for the MALDI QITMS analysis of the synthetic
peptide MRFA.


Scan Table
First Table
Ionize
Cool
Ionize
Cool
Eject Low Mass Ions
Prescan
Acquire
Empty Trap


rf Voltage
Start End
1328 1328
1328 1328
1328 1328
1328 1328
1328 1328
3985 3985
1328 1328
1328 15000
0 0


Axial Modulation
Frea. (kHz) Amn (V)


Time
(ms)
0.5
0.001


50
0.001
50
5
5


--








+
H] ion for this compound (m/z 524) is injected at a q of 0.1 (scan table 1). This

corresponds to an rf voltage of 664 VO.p. The next table (scan table 2) is used to trigger

the nitrogen laser, which has a pulse width of 3 ns (FWHM) in the triggered mode. Next,

a cool time of 50 ms (scan table 3) is used to allow the ions that are injected into the ion

trap to have sufficient time to be buffered by the helium gas present. The laser trigger

and cool tables are then repeated (scan tables 4 and 5) to give a total of two laser shots

per scan function, which increases the number of ions present in the ion trap. The ability

of this system to store ions from multiple ionization events to increase the sensitivity of

the analysis is not available on a time-of-flight mass spectrometer, where each ionization

event (defined by the laser pulse) defines the start of the timing cycle to determine the

flight time of the ions (and thereby the mass of the ion). The rf is then increased (scan

table 6) to eject low mass ions (primarily produced from the DHB matrix) from the ion

trap. The mass analysis scan is then performed, in which the electron multiplier is turned

on and the rf is ramped to scan out the ions present in the ion trap to produce a mass

spectrum (scan table 8). This scan function was repeated 20 times (referred to as 20

microscans), and the mass assignment and sum of the ion signals were acquired as an

ASCII text file (using Gatorware software). Mass assignments for this compound were

made using the standard calibration software available with the ITS40 software.


MALDI QITMS of Angiotensin II


The next peptide analyzed was Angiotensin II (Asp-Arg-Val-Tyr-Ile-His-Pro-Phe,

molecular weight 1046.2 Da). This octapeptide is one of the major components of the

renin-angiotensin system of renal hypertension and acts on the adrenal gland to stimulate
39
the release of aldosterone. Since the mass of this compound is greater than the

maximum m/z achievable with the standard quadrupole ion trap instrument, the axial









modulation frequency was decreased during the mass analysis scan to increase the mass

range of the ion trap. For this analysis, an axial modulation frequency of 210.75 kHz
was used, giving a qejea of 0.533 and a maximum attainable mass of 1100 Da (giving a

mass range extension of 1.7x). Since the mass resolution is inversely related to the mass

range extension factor,40 the maximum attainable mass is usually set to just above the

maximum mass for the ion of interest (in this case, m/z 1046).

The MALDI quadrupole ion trap spectrum for 20 pmol of this compound is shown

in Figure 2-10. The scan function used to produce this spectrum is the same as the one
used for MRFA (Figure 2-9), except that the rf voltage which corresponds to a qije of

0.1 is 2655 V. This spectrum shows the [M + H] ion at m/z 1047 as the base peak along

with lower mass ions resulting from cleavage of the amide bonds along the peptide

backbone. These fragment ions, typically seen when performing MALDI using ion trap

instruments (e.g. FTMS and quadrupole ion traps)4142 but not with time-of-flight

instrument, are produced by collision-induced dissociation (CID) with helium and other

background gases present in the ion source and quadrupole ion trap, by surface-induced

dissociation, SID, with ion optic surfaces or the ion trap electrodes, or by metastable

decay, and can be very useful in sequencing the peptide.43 The mechanism for the

formation for the y- and b-ions (following the product ion nomenclature of Roepstorff

and Fohlman4) seen in this spectrum is shown in Figures 2-11 and 2-12, respectively.

Formation of y-type ions (e.g. the Y7+ ion, Figure 2-11) involves cleavage of the amide

bond along the peptide backbone, with the charge being retained on the amide nitrogen.

The fragment shown in Figure 2-11 is termed y7 because it contains seven amino acid

residues from the original peptide structure (counting from the carboxy terminus).
Formation of b-type ions (e.g. the b6+ ion, Figure 2-12) occurs by P-cleavage and charge

migration to the carbonyl end of the amino terminus fragment. The fragment shown in




















+ c


+ I









I


I I I I I I I '
8 O 8 8 8 8 o
m W 1


0
_0


0
0
0













-o






0
-o
N-


0
0
CD


A!suglul


+
T-
+3,-


--










ANGIOTENSIN II

R8 = Asp
R7 = Arg
R6 =Val
R5= Tyr
R4 = Ile
R3 = His
R2 = Pro
RI = Phe





R H 0 R O R4 O R 0
ANH NH NH NH
H-N +H NH NH NNH OH
H O R7 0 Rs 0 R3 0 R1




CID


H O R6 O R4 0 R2 0
H-NH NH NH NH
+ NH / NH) NH) OH
R7 0 R5 0 R3 0 Ri

Y7 ion


Figure 2-11. Mechanism for the formation of the y7+ ion for angiotensin II.58











ANGIOTENSIN II

R1 =Asp
R2 = Arg
R3 = Val
R4 =Tyr
Rs = Ile
R6 = His
R7 = Pro
Rs = Phe





R1 0 R3 0 R O R7 0
NH NH NH NH
H-N NH NH NH OH
H O R2 O R O R H Rg




CID


Ri O R 0 R5 0

H-N NH NH NH NH NH +
H O R2 O R4 O R

b6 ion


Figure 2-12. Mechanism for the formation of the b6+ ion for angiotensin II.58









Figure 2-12 is termed b6 because it contains six amino acid residues from the original

peptide structure (counting from the amino terminus).

Although fragment ions produced during the injection process can be used to help

sequence a peptide, this is only the case when the analyte is present as the sole

constituent in the sample (since it is impossible to tell which fragment ion came from
+
which [M + H] ion in a mixture of unknowns). When a mixture is present, MS/MS

must be performed. When performing MS/MS using a quadrupole ion trap, the ion of

interest must be isolated from the remainder of the ions present in the ion trap prior to

resonance excitation. A number of techniques have been successfully used to isolate ions

using a quadrupole ion trap, including apex,46 two-step,47'48'49 SWIFT (stored-

waveform inverse Fourier transform),5,'5' random noise,52 FNF (filtered noise field),53

and forward-and-reverse scanning.54,55 For the MSn results for Angiotensin II shown in

this dissertation, the forward-and-reverse isolation technique as shown by Schwartz et

al.545 was used. Although apex and two-step isolation are effective, they can only

isolate ions up to approximately m/z 600 on standard ion trap instruments because of the

high dc voltages necessary to be applied to the ring electrode. The forward-and-reverse

scanning technique, however, only requires differing frequencies and voltages to be

applied to the endcap electrodes.

The scan program to implement this method of ion isolation is shown in Figure 2-
13 and is listed in Table 2-2. After the ionization and cool steps (labeled as "Injection

Period"), the axial modulation is turn on (210.75 kHz, 8 V) and the rf is ramped from a q

of 0.1 to a q of 0.562 (which corresponds to an ion frequency just below that for the ion

of interest). Since the axial modulation frequency and voltage used for isolation is the

same as that used for mass analysis, the final q that will be ramped to is calculated based

on the observed (rather than the actual) m/z for the ion of interest (that is, the final q that

will be ramped to is varied until the corresponding mass, as displayed by the Gatorware









7500-
7000-
6500
6000-
5500-
5000-
4500
4000-
3500
3000 -
2500-
2000-
1500-
1000-
500
0--

0
400


a , I


50 100I 15
50 100 15C


S 200 250


I I II I I I I 1 1 1I I I I I I 1 1
50 100 150 200 250
Time (ms)


Figure 2-13. Scan function used for the MALDI QIT MS/MS analysis of Angiotensin II.


50 100 150 200 250


200 -


8-
6-
4-
2-
0 -








Table 2-2. Scan table listing for the MALDI QITMS analysis of angiotensin II.


# Scan Table
1 First Table
3 Ionize
4 Cool
5 Ionize
6 Cool
7 Forward Scan
8 Cool
8 Reverse Scan
9 Cool
10 MS/MS
11 Cool
12 Prescan
13 Acquire
14 Empty Trap


rf Voltage
Start End
2655 2655
2655 2655
2655 2655
2655 2655
2655 2655
2655 12274
10090 10090
15000 12444
7965 7965
7965 7965
7965 7965
2655 2655
2655 15000
0 0


Axial Modulation
Frea. (kHz) Amp (V)


210.75

210.75

115


210.75


Time
(ms)
0.5
0.001
50
0.001
50
70
5
19
5
10
5
5
108
2


Frea. (kHz) Amn M









software, is set to just below the displayed mass for the ion of interest). As ions of

increasing mass come into resonance with the supplementary field, they become excited

and are ejected from the ion trap. Just before the ion of interest gets to this frequency,

the axial modulation is turned off and the rf is set to cool the remaining ions in the ion

trap. This is termed a "Forward Scan" because the rf is scanned in an increasing manner

and ions of increasing m/z are ejected from the ion trap. The rf is then set to the

maximum available (7500 V0.p, q = 0.89) and the axial modulation is turned back on.

The rf is then ramped down to just above that for the ion of interest (the rf is ramped

from a q of 0.89 to a q of 0.570 in this case). This is termed a "Reverse Scan" since the

rf is scanned in a decreasing manner. The axial modulation is then turned off and the rf

is set to cool the remaining ions in the ion trap. The sequential steps involved in the

forward-and-reverse scan isolation procedure is shown in Figure 2-14. At this point a

narrow m/z range is isolated in the ion trap. In the case of the mass spectrum of

Angiotensin I shown at the top of Figure 2-15, the mass analysis scan immediately

follows the forward-and reverse scan. This gives the spectrum of the mass-isolated [M +
+ +
H] ion for Angiotensin II. To perform MS/MS on this ion, after the [M + H] ion for

Angiotensin II is isolated the rf is set to correspond to a q of 0.3 for the ion of interest

and the axial modulation frequency is set to 115 kHz. The axial modulation voltage is

varied until the mass-isolated ion (in this case the [M + H] ion) begins to decrease and

daughter ions begin to appear. If too much voltage is applied, resonance ejection takes

place instead of resonance excitation and the signal is lost. If not enough voltage is

applied, the spectrum remains unchanged. The daughter ion spectrum produced from the
+
fragmentation of the [M + H] ion of Angiotensin II using an axial modulation frequency

of 115 kHz and voltage of 2.8 V is shown in the bottom spectrum of Figure 2-15. Along
+ + +
with residual [M + H] ion, ions corresponding to the cleavage of amide bonds (Y7, b ,
+
and b6 ions) are observed. Although these ions correspond well with those seen for the







Mass of


Interest

0 0._


Resonance
Point
I
AN^- ^


Stability
Limit
I


0


e


G.e. 0


ee.


Ge.


n 0.908


Figure 2-14. Sequential steps in the forward-and-reverse scan isolation procedure.55


H


v


----















+ .0

4 2





-8 C
I.0
0
0

+0 +






o C



o










8 8 o
.u0
0 00
to o
-+

+ +



0o






0 C 0 O0 w Cu










fragment ions produced during the injection process, the isolation and MS/MS process

assures that these fragment ions are only produced for the ion of interest (in this case, the
+
[M + H] ion of Angiotensin II). It is also interesting to note the greater increase in

intensity for these peptide backbone fragment ions. Although it is difficult to correlate

separate MALDI spectra (due to a decrease in the signal intensity over time at a single

spot due to ablation of the sample), the intensities of the MS and MS/MS spectra taken in

Figure 2-15 can be roughly compared to estimate the MS/MS efficiency (defined as the

fraction of ions collected that are product ions, and is given as the ratio of the sum of the

intensities of the product ions divided by the intensity of the parent ion remaining

following excitation and the sum of the intensities of the product ions5) since they were

taken from the same laser spot and from back-to-back acquisitions. After 20 microscans

were acquired for the MS spectrum (with 2 laser shots per microscan, giving a total of 40

laser shots summed for this spectrum), the resonance excitation frequency was turned on,

and 20 microscans were acquired for the MS/MS spectrum (again with 2 laser shots per

microscan). Another MS spectrum taken immediately after the MS/MS spectrum

confirmed that the intensity had not decreased by more than 10% of its original value.

Comparing these two spectra gives an MS/MS efficiency of approximately 50%.

MS/MS was then performed on the daughter ions produced by the fragmentation of

the [M + H] ion of Angiotensin II. This is termed MS since three stages of MS are

performed. As is the case with MS/MS, MALDI ions from Angiotensin II were injected
+
and cooled, then the [M + H] ion for Angiotensin II was isolated using the forward-and-

reverse scanning technique. After fragmentation using resonance excitation, one of the
+
daughter ion, the y7 ion, was mass isolated (again using the forward-and-reverse
+
scanning technique). The rf level was set to a q of 0.3 for the y7 ion, and resonance

excitation was performed (115 kHz, 2.5 V). The mass-isolated spectrum for Angiotensin
is shown in Figure 2-16a, the daughter ion spectrum for the [M + H ion is shown in
II is shown in Figure 2-16a, the daughter ion spectrum for the [M + H] ion is shown in










1500 a). MS m

* 1000-

S500-

0-
600 700 800 900 1000
800 b). MS/MS of [M + H]+
600o- +
S400 b8+
200
200 bs
0 1 ---.Ji'
600 700 800 900 1000

7- H2 +
5 c). MS/MS of y7
100 [b6Y7- H2o

50 .b6y7-co]T /WY Y6+ b6+
50 F


600 700 800 900 1000







Figure 2-16. a). Mass isolated MALDI QITMS spectrum ([M + H]+ ion) for 20 pmol of
Angiotensin II, b). MS/MS spectrum for this ion, and c). MS/MS for the
mass-isolated y7+ ion (MS3 spectrum for the [M + H]+ ion of Angiotensin II).









+
Figure 2-16b, and the daughter ion spectrum produced from the y7 ion is shown in
+
Figure 2-16c. This is also termed a granddaughter ion spectrum for the [M + H] ion of
+ +
Angiotensin II, since the y7 ion is the daughter of the [M + H] ion and the resulting
+ +
ions for the MS/MS of the y7 ion are daughter ions of that ion. Besides the [y7 H20]

ion, which is typically seen in MS analysis of peptides, y- and b-ions due to further

peptide backbone cleavages are observed along with the internal sequence fragment ion
+
b6y, (which corresponds to the amino acid sequence Arg-Val-Tyr-Ile-His produced

from the center of the intact peptide) and ions due to the loss of water and carbon

monoxide from the internal fragment ion. The internal sequence fragment ion is most

likely driven by the presence of the highly basic proline residue on the seventh amino

acid of this peptide.8 The presence of the proline residue at this position also leads to a
+
lack of a b7 ion in this spectrum.
2 + 2 +
In addition to performing MS2 of the Y7 ion, MS of the b, ion was also

performed. The mass-isolated spectrum for Angiotensin II is shown in Figure 2-17a, the
+
daughter ion spectrum for the [M + H] ion is shown in Figure 2-17b, and the daughter
+
ion spectrum produced from the b8 ion is shown in Figure 2-17c. The scan function for
2 + 2 +
the MS of the b8 ion was identical to the one used for the MS of the y7 ion, except

that the rf levels used for the mass isolation and resonance excitation steps are set to

correspond to the different ion of interest. In contrast to the previous MS3 data, the only

easily identifiable peak in this MS3 mass spectrum is the neutral loss of ammonia from
+
the b8 ion. The lack of backbone fragments in this spectrum is due to the presence of
+
the proline residue as the first amino acid of the b8 ion. The very high basicity of this

residue leads to cyclization rearrangements, making identification of these mass peaks

very difficult.













1500 -a). MS

1000-

500-

0-
600 700 E

8004 b). MS/MS of [M + H]


600
400
200
0



200
150


100-
50-
0-


600 700 800


c). MS/MS of b8+


[M+H]


(00 900 1000

V7


900 1000


[g- NH\ 1
\


I--cLI-YYYL ~ L- llIIL~ I MCCI --(r -- -(C


600


700


800


900


1000
1000


Figure 2-17. a). Mass isolated MALDI QITMS spectrum ([M + H]+ ion) for 20 pmol of
Angiotensin II, b). MS/MS spectrum for this ion, and c). MS/MS for the
mass-isolated b8+ ion (MS3 spectrum for the [M + H]+ ion of Angiotensin II).


III 1 I I J I 1. .1 1 .1.1 i A L.


~WLYL -_L*I-LI ULllrY LYLYI L~L-