Citation
Separation of Tryptic Peptide Ions by High-Field Asymmetric-Waveform Ion Mobility Spectrometry

Material Information

Title:
Separation of Tryptic Peptide Ions by High-Field Asymmetric-Waveform Ion Mobility Spectrometry
Creator:
Bryant, Jennifer Alisia Garrett
Place of Publication:
[Gainesville, Fla.]
Publisher:
University of Florida
Publication Date:
Language:
English
Physical Description:
1 online resource

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Chemistry
Committee Chair:
Yost, Richard A.
Committee Members:
Smith, Benjamin W.
Stewart, Jon D.
Powell, David H.
Denslow, Nancy D.
Graduation Date:
8/9/2008

Subjects

Subjects / Keywords:
Amino acids ( jstor )
Electric fields ( jstor )
Electric potential ( jstor )
Electrodes ( jstor )
Gas composition ( jstor )
Geometry ( jstor )
Ions ( jstor )
Mass spectrometers ( jstor )
Mass spectroscopy ( jstor )
Signals ( jstor )
Chemistry -- Dissertations, Academic -- UF
faims, phosphopeptides
Genre:
Electronic Thesis or Dissertation
bibliography ( marcgt )
theses ( marcgt )
Chemistry thesis, Ph.D.

Notes

Abstract:
Proteins play a key role in cell function regulation. Understanding how a protein modulates cell activity may give new insight into disease mechanisms and elucidate biomarkers. A protein's structure can affect its biological activity; therefore, recent research has focused on determining a protein?s structure, especially where modifications occur, such as phosphorylation. Identifying alterations in a protein's phosphorylated state can be used as therapeutic and preventative biomarkers. Mass spectrometry (MS) has been a growing technique for identifying and sequencing proteins. Proteins are digested into their respective peptides for analysis by MS. In order to identify and sequence numerous proteins, data-dependent analysis is often employed. The computer selects a predefined number of the highest abundance peaks for further analysis by tandem MS, regardless of their relevance. Often, analytes of interest are in low abundance in a complicated biological matrix, making them difficult to analyze. Sample clean-up methods have been utilized, including immobilized metal affinity chromatography (IMAC) and high performance liquid chromatography (HPLC), but they are time-consuming. Here is presented high-field asymmetric-waveform ion mobility spectrometry (FAIMS) for the rapid separation of tryptic peptide ions. FAIMS separates ions in milliseconds based on their difference in mobility at high and low fields. The FAIMS device can be used as a stand-alone device or can be coupled to a mass spectrometer in between the ionization source and the mass spectrometer entrance. Here FAIMS is explored as a separation technique for tryptic peptides ions, particularly phosphopeptides. Variation in FAIMS parameters, such as gas composition and flow rate, dispersion voltages, electrode temperature and geometries are explored for their effect on tryptic peptide ion separation. The addition of FAIMS to a data-dependent HPLC-MS analysis is addressed.
Statement of Responsibility:
by Jeniffer Alisia Garrett Bryant

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Jennifer Alisia Garrett Bryant. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
8/31/2010

Downloads

This item has the following downloads:


Full Text

PAGE 1

1 SEPARATION OF TRYPTIC PEPTIDE I ONS BY HIGH-FIELD ASYMMETRICWAVEFORM ION MOBILI TY SPECTROMETRY By JENNIFER ALISIA GARRETT BRYANT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Jennifer Alisia Garrett Bryant

PAGE 3

3 To TJ, Mom, Dad, and Greg fo r all their love and support

PAGE 4

4 ACKNOWLEDGMENTS I would like to thank m y advisor, Richard Yo st, for supporting and advising me during my time in his lab. Thank you to all the past and pr esent group members who have made my stay at UF so enjoyable. A special thanks go to Rachelle Landgraf, Frank Kero, Mike Napolitano, Alisha Roberts and Dave Richards on for their enduring friendship over these past four years and for all the fun times both in and out of the lab. Thanks go to Marilyn Prieto, Todd Prox, and Joe Shalosky for all their support a nd excitement in developing hemispherical FAIMS. I have to give a special thanks to Tim Garrett for all his in sightful conversations and input. I would like to thank Dan Magparangalan for always sharing a laugh and putting things in perspective. I am especially thankful to Dodge Ba luya for all his help, friendship, and the fun bus/bike rides home. I would like to thank my husband, TJ; my parent s, Gary and Alison; and my brother, Greg, for putting up with me and loving me during my four years at UF, especially during the period when I was writing and preparing to defend.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4LIST OF TABLES................................................................................................................. ..........8LIST OF FIGURES.........................................................................................................................9ABSTRACT...................................................................................................................................12 CHAP TER 1 INTRODUCTION TO THE ANALYSIS OF TRYPTIC PEPTIDES BY ION MOBILITY SPECTROMETRY ............................................................................................14Introduction................................................................................................................... ..........14Peptide and Protein Identific ation by Mass Spectrometry..................................................... 14Data-dependent Analysis........................................................................................................ 16Introduction to Ion Mobility................................................................................................... 17Traditional IMS............................................................................................................... 17Peptide Analysis by IMS................................................................................................. 19Introduction to FAIMS...........................................................................................................21Fundamentals of FAIMS.................................................................................................21High-Field Asymmetric-Waveform Ion Mobility Spectrometry Modes.........................24 High-Field Asymmetric-Waveform Ion M obility Spectrometry Modes Parameters Affecting Ion Transmission................................................................................................ 25Carrier Gas.......................................................................................................................25Dispersion Voltage.......................................................................................................... 27Electrode Temperature.................................................................................................... 27Cell Geometry................................................................................................................. 28Selection of the Ionization Source..........................................................................................29Instrument Description......................................................................................................... ..30Overview....................................................................................................................... ..........302 CHARGE-STATE SEPARATION OF TRYP TI C PEPTIDE IONS BY FAIMS................. 41Overview....................................................................................................................... ..........41Advantages and Disadvantages of Electrospray for the Analysis of Tryptic Peptide Ions.... 41Data-dependent Analysis of Tryptic Peptides........................................................................42Simplifying Tryptic Peptide Ion Product Spectra................................................................... 43Reducing Background Noise........................................................................................... 43Current Charge-State Separation Technology.................................................................43Application of FAIMS to Charge-State Separation of Tryptic Peptide Ions.......................... 44Experimental................................................................................................................... ........45Synthetic Tryptic Peptides...............................................................................................45Cytochrome C Tryptic Digest......................................................................................... 46

PAGE 6

6 Instrument Parameters..................................................................................................... 46 High-Field Asymmetric-Waveform Ion M obility Spectrometry Modes Parameters..... 46Results and Discussion......................................................................................................... ..47Optimizing FAIMS Parameters for Charge-State Separation................................................ 47Dispersion Voltage Optimizati on for Tryptic Peptide Ions............................................. 47Optimizing Carrier Gas Flow Rate and Composition for Tryptic Peptide Ions.............. 48Experimentally Determined Optimized FAIMS Parameters for Charge-State Separation....................................................................................................................49Evaluating Charge-state Separation of DLLXR..................................................................... 49Charge-State Separation of a Cytochrome C Digest.............................................................. 51The Effect of Amino Acid Sequence on Tryptic Peptide Behavior in FAIMS...................... 52The Effect of Glycine......................................................................................................52The Effect of the C-Terminal Amino Acid..................................................................... 53Conclusion..............................................................................................................................533 SEPARATION OF PHOSPHOPEPTIDE S BY HIGH-FIEL D ASYMMETRIC WAVEFORM ION MOBILI TY SPECTROMETRY............................................................80Overview....................................................................................................................... ..........80Importance of Phosphopeptides.............................................................................................. 80Analytical Methods to Ch aracterize Phosphopeptides........................................................... 81Chromatography..............................................................................................................81Mass Spectrometry.......................................................................................................... 82Limitations in Phosphopeptide Analysis by Mass Spectrometry.................................... 83Application of FAIMS to Phosphopeptide Analysis.............................................................. 84Experimental Design............................................................................................................ ..85Sample Preparation..........................................................................................................85Instrumentation................................................................................................................ 85Instrument Modification.................................................................................................. 85Optimization of FAIMS Parameters fo r the Detection of Phosphopeptides..........................86Results.....................................................................................................................................86The Effect of Carrier Gas Com position on Phosphopeptide Separation................................ 86Variation of CO2/He Carrier Gas Compositions............................................................. 87Variation of CO2/He/N2 Carrier Gas Compositions........................................................ 88Variation of CO2/N2 Carrier Gas Compositions.............................................................. 89Variation of He/SF6 Carrier Gas Compositions.............................................................. 89The Effect of Heated Electrodes on Phosphopeptide Resolution...........................................91Conclusion..............................................................................................................................944 DESIGN OF A NOVEL HEMISPHE RICAL FAIMS GE OMTERY FOR PHOSPHOPEPTIDE SEPARATION..................................................................................115Overview....................................................................................................................... ........115Current FAIMS Electrode Geometries.................................................................................115Planar Geometry Electrodes..........................................................................................115Cylindrical Geometry.................................................................................................... 116Dome Cell Geomet ry.....................................................................................................118

PAGE 7

7 Spherical FAIMS Electrodes................................................................................................ 118Designing Spherical FAIMS Electrodes...............................................................................119Design of Novel Hemispherical FAIMS Electrodes............................................................ 120Experimental................................................................................................................... ......121Samples..........................................................................................................................121Instrumentation.............................................................................................................. 121Results...................................................................................................................................122Data Collection from Sphe rical FAIMS Electrodes......................................................122Data Acquired on Hemispherical FAIMS.....................................................................122Conclusion............................................................................................................................1235 CONCLUSION..................................................................................................................... 142APPENDIX AMINO ACIDS........................................................................................... 145LIST OF REFERENCES.............................................................................................................146BIOGRAPHICAL SKETCH.......................................................................................................152

PAGE 8

8 LIST OF TABLES Table page 2-1 The m/z of DLLXR, where X can be any of the listed am ino acids. The variation in expected m/z versus observed m/z could be due to the loss of an oxygen or hydroxyl group..................................................................................................................................57 2-2 The m/z of DLLXK, where X is a ny of the listed am ino acids.......................................... 58 2-3 The m/z of AWXVAR, where X is any of the listed am ino acids..................................... 59 2-4 The m/z of VAXL R, where X is any of the listed amino acids.........................................60 2-5 Instrumental parameters utilized fo r charge-state separation experim ents........................ 62 2-6 Optimized FAIMS conditions for tryptic peptide ions...................................................... 68 2-7 Charge-state separation of DLLXR is achieved by FAIMS. ............................................. 70 2-8 Effect of varying amino acid sequence on CV (V). AW XVAR was taken at 50% He, while DLLXR, DLLXK and VAXLR were ta ken at 40% He due to instrument malfunctions with the vacuum...........................................................................................78 3-1 Electrospray and mass spectrometer para m eters used for these experiments.................... 98 3-2 Variations of ca rrier gas com positions.............................................................................. 99 4-1 Comparison of resolution and resolvi ng power between the hem ispherical FAIMS and the commercial Thermo FAIMS............................................................................... 141

PAGE 9

9 LIST OF FIGURES Figure page 1-1 Nomenclature of tryptic peptide ion fragm ents................................................................. 33 1-2 Structures of three comm on phosphorylated am ino acids................................................. 34 1-3 Typical ion mobility spectrometer..................................................................................... 35 1-4 Cut away of the side-by-side FAIMS cell. ....................................................................... 36 1-5 Side-by-side FAIMS cell...................................................................................................36 1-6 Altered ion movement between electrodes as the field varies. .......................................... 37 1-7 Actual asymmetric waveform applied to the inner electrode. ........................................... 37 1-8 Classification of i on behavior in FAIMS. .......................................................................... 38 1-9 Net displacement of an ion over ti m e due to the asymmetric waveform........................... 38 1-10 Potential path of the ion in the FAIMS cell....................................................................... 39 1-11 The top electrodes show a mixtur e of ions entering FAIMS. ..........................................39 1-12 Thermo LCQ ion trap used in these experim ents.............................................................. 40 2-1 Data from Guevremont et al. demonstr ating the effect of varying carrier gas com position on the resolution of peptid es from an enolase 1 digest................................. 56 2-2 Plot of m/z versus CV for 254 tryptic pe ptides from data adapted from Guevremont et al. ....................................................................................................................... ...........61 2-3 Ion intensities of DLLPR+ at varying DVs. Optimal DV voltage for tryptic peptide ions is -5000 V................................................................................................................ ...63 2-4 Optimal charge-state separation of EFSGDK at varying DV. ......................................... 64 2-5 The effect of carrier gas fl ow rate on signal intensity. .................................................... 65 2-6 Variations in the carrier gas flow rate................................................................................ 66 2-7 The optimal carrier gas composition for the ch arge-state separation of tryptic peptides was 50/50 v/v He/N2............................................................................................67 2-8 Charge-state separation of DLLXR. ................................................................................69 2-9 Charge-state separation of DLLPR+ and DLLPR2+...........................................................71

PAGE 10

10 2-10 Charge-state separation is achieved for DLLXR. ............................................................ 72 2-12 CV required by select ch arge-states of DLLXR2+.............................................................74 2-13 Cytochrome C digest ESI-MS spectrum............................................................................ 75 2-14 Charge-state separation of cytochrome C by FAIMS. ....................................................76 2-15 Improvement in charge-state sepa ration of Cytochrom e C by FAIMS............................. 77 2-16 Tryptic peptides with 5 a nd 6 am ino acid (aa) residues..................................................... 79 3-1 Neutral loss scan. Q1 and Q3 are scanned at a sp ecified offset in order to detect neutral losses......................................................................................................................95 3-2 Parent ion scan. Q1 scans the m/z, ions are fragm ented in q2, and a selected ion is monitored in Q3.................................................................................................................95 3-3 Neutrals and ions used in neutral loss and parent ion scanning modes to selectively detect phosphopeptides. ...................................................................................................95 3-4 Structures of EFXGDK, where X = S, T, Y, pS, pT, or pY.............................................. 96 3-5 Interface between the FAIMS device a nd the LCQ m ass spectrometer. Length 83.9 mm, outer diameter 15.0 mm, inner diameter 1.3 mm...................................................... 97 3-6 CV scans of a mixture of EFXGDK ( 20 ppm ) with two common FAIMS carrier gas compositions................................................................................................................... .100 3-7 Comparison of MS spectra from FA IMS-MS using diffe rent carrier g as compositions................................................................................................................... .101 3-8 CV scans with variation in percent He in CO2 carrier gas. ............................................ 102 3-9 Comparison of MS spectra from FAIMS-MS using diffe rent carrier g as compositions................................................................................................................... .103 3-10 Variation of % CO2 in the He/N2 carrier gas. ................................................................104 3-12 Variation of % CO2 in the N2 carrier gas. ...................................................................... 106 3-13 CV peaks for varying SF6/He carrier gas combinations.................................................. 107 3-14 Mass spectra of vary percentages of He in SF6 carrier gas.............................................. 108 3-15 Plot of ion intensity versus CV in FAIMS with carrier gas He/SF6. ............................109 3-16 Comparison of He/N2 spectra and He/SF6 spectrum. ....................................................110

PAGE 11

11 3-18. The effect of electrode temperature on resolution between E FYGDK and EFpSGDK..112 3-19 The +2 charge-state ion inte nsities at different CVs. ..................................................... 113 3-20 The +1 charge-state ion inte nsities at different CVs. ..................................................... 114 4-1 Planar geometry FAIMS.................................................................................................. 124 4-2 Cylindrical FAIMS electrodes......................................................................................... 124 4-3 Measurements for a, b, and r used to calcu late the elec tric fields within a curved surface..............................................................................................................................125 4-4 Comparison of the electric fields in cylindrical and planar geom etries...........................126 4-5 Non-uniform electric field creates a focusing effect........................................................127 4-6 Focusing effect caused by the curvature of the electrodes. ........................................... 128 4-7 Dome cell geometry FAIMS............................................................................................ 129 4-8 Different paths between the electrode s for spherical and cylindrical FAIMS electrodes. .................................................................................................................. ....130 4-9 Spherical FAIMS electrodes with rods............................................................................ 131 4-10 Spherical FAIMS electrodes............................................................................................ 132 4-11 Hemispherical FAIMS electrodes. .................................................................................133 4-12 Inner and outer electrode s of he mispherical FAIMS.......................................................134 4-13 Hemispherical FAIMS disassembled............................................................................... 134 4-14 Assembled hemispherical FAIMS................................................................................... 135 4-15 Gold-plated outer electrode was destroyed due to arcing. ............................................... 136 4-16 CV scans for EFXGDK using hemispherical FAIMS at a DV of -2700 V (13,500 V/cm ).......................................................................................................................... .....137 4-17 The mass spectrum of tryptic peptide ions are shown using the hem ispherical FAIMS at low DV (-2700 V). .................................................................................................. 138 4-18 The separation of 3,4-DNT and TNT using hemispherical FAIM S at a DV of -2700 V. The resolution is 1.0. ............................................................................................... 139 4-19 Separation of TNT and 3.4-DNT. ...................................................................................140

PAGE 12

12 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SEPARATION OF TRYPTIC PEPTIDE I ONS BY HIGH-FIELD ASYMMETRICWAVEFORM ION MOBILI TY SPECTROMETRY By Jennifer Alisia Garrett Bryant August 2008 Chair: Richard A. Yost Major: Chemistry Proteins play a key role in cell function regulation. Unders tanding how a protein modulates cell activity may give new insight into disease mechanisms and elucidate biomarkers. A proteins structure can affect its biological activity; therefore, recent research has focused on determining a proteins structure, especially where modifications occur, such as phosphorylation. Identifying alterations in a proteins phosphoryl ated state can be used as therapeutic and preventative biomarkers. Mass spectrometry (MS) has been a growing technique for identifying and sequencing proteins. Proteins are digested into their respectiv e peptides for analysis by MS. In order to identify and sequence numerous protei ns, data-dependent analysis is often employed. The computer selects a predefined number of th e highest abundance peaks for further analysis by tandem MS, regardless of their relevance. Often, analytes of interest are in low abundance in a complicated biological matrix, maki ng them difficult to analyze. Sample clean-up methods have been utilized, including i mmobilized metal affinity ch romatography (IMAC) and high performance liquid chromatography (HPLC), but th ey are time-consuming. Here is presented high-field asymmetric-waveform ion mobility spectrometry (FAIMS) for the rapid separation of tryptic peptide ions.

PAGE 13

13 FAIMS separates ions in milliseconds based on their difference in mobility at high and low fields. The FAIMS device can be used as a st and-alone device or can be coupled to a mass spectrometer in between the ionization source an d the mass spectrometer entrance. Here FAIMS is explored as a separation t echnique for tryptic peptides i ons, particularly phosphopeptides. Variation in FAIMS parameters, such as gas composition and flow rate, dispersion voltages, electrode temperature and geometries are expl ored for their effect on tryptic peptide ion separation. The addition of FAIMS to a data -dependent HPLC-MS analysis is addressed.

PAGE 14

14 CHAPTER 1 INTRODUCTION TO THE ANALYSIS OF TRYPTIC PEPTIDES BY ION MOBILITY SPECT ROMETRY Introduction Proteins are responsible for m any functi ons in the body, including gene expression, cell function regulation and cell structure.1,2 Understanding a protei ns structure-function relationship is paramount in unders tanding how proteins affect biol ogical activity. However, the complexity of biological matrices challenges researchers to deve lop new analytical techniques to identify, sequence and understand peptides a nd proteins. Conventional sample clean-up techniques have been employed, but are time -consuming and frequen tly lack selectivity.3 Here a novel technique utilizing high-field asymmetric -waveform ion mobility spectrometry (FAIMS) for the rapid, selective separation of tryptic digests is presented. Peptide and Protein Identification by Mass Spectrometry Elucidating the biological function of a prot ein involves understandi ng its structure. Proteins have up to four degrees of substructure; the prim ary structure cons ists of the amino acid sequence.1,2 The different functional groups on ami no acids affect how a protein can form subsequent substructures. In order to identify the primary struct ure of a protein by mass spectrometry (MS), typically the protein is digested by an enzyme in to its respective peptid es. Several enzymes are commercially available to digest proteins (including pepsin, trypsin, and chymotrypsin) for analysis by mass spectrometry, but the most co mmonly used is trypsin. Trypsin cleaves a peptide on the carboxyl side of am ino acids lysine and arginine, except when a proline follows.4 Based on characteristic mass-to-charge ratios ( m/z ), an amino acid sequence of a protein can be identified by MS.

PAGE 15

15 Once digestion has been performed, the pept ides are subsequently identified according to their m/z measured by mass spectrometry. This process of digesting the protein and subsequently analyzing by MS is known as bottom up sequencing.5 The amino acid sequence of a tryptic peptide can further be determ ined by tandem mass spectrometry (MS/MS). Fragmentation of peptides yields product ions that can facilitate determining the amino acid sequence. Depending on where the amino acids fr agment and where the charge is retained, characteristic fragments are produced, as shown in Figure 1-1. Peptide fragment nomenclature distinguishes ions that are cleav ed from the peptide backbone ( a, b c x, y, z ) from those that involve side chain cleavage ( d v, w ).5,6 The ions most commonly observed from collision-induced dissociation (CID) are b and y ions, which conveniently cleave at the amide bond, preserving the intact amino acid. In CID, parent ions collide with a gas and energy is redi stributed in the ion. causing fragmentation to occur. Doubly charged ions typically produce bn/ ym complementary ion pairs (where n + m = total number of amino acid residues), facilitating protein identification.5 Though less common, other ion types may be observed as well from CID, such as a type ions. The m/z difference of 28 between two peaks in a tryptic digest MS spectrum may be indicative of a ab ion pair, as a ions correspond to the loss of CO from a b ion.5 As an alternative to collis ion-induced dissociation, other activation methods for MS/MS have been explored in order to gain addi tional fragmentation data. Infrared multiphoton dissociation, surface-induced dissociation, blackbody infrared dissociation, photodissociation, electron capture dissociation (E CD) and electron transfer dissociation (ETD) have been examined.7-14 ETD and ECD produce similar ions (predominantly c and z versus c and z. ions, respectively), which fragment fr om the tryptic peptide backbone without losing modifications,

PAGE 16

16 such as phosphorylation or glycos ylation, which are usually lost during CID. By analyzing fragmentation data, the sequence of the amino acids can be determined. Data-Dependent Analysis To increase throughput of protei n analysis, data-dependent analysis is often em ployed to collect large quantities of data. In data-depe ndent analysis, the peaks of highest abundance are selected for further tandem analysis regardless of their relevance to the study. This method could be improved if peptides of particular inte rest, such as doubly charged peptides or posttranslationally modified peptides, were enriched or exclusively separated from the digest to ensure they are the peak s of highest abundance. Doubly charged tryptic peptide ions are of in terest because they yield the most useful tandem MS spectra for identification by data-dep endent analysis. More fragments retain a charge and most product ions are singly charged, making computer identification of the peptide less complicated and more informative. As previously mentioned, complementary bnym ion pairs are often produced from doubly charged ions yielding more structur al information. By selectively separating doubly charged ions from the digest, data-dependent scanning is more likely to select relevant ions for further anal ysis and more accurately identify the peptide. The separation of phosphopeptides is of inte rest as well because phosphorylation is the most common protein post-translational modifica tion and plays a key role in many biological functions including signal transduc tion, gene expression, and cell cycle.15-17 Phosphorylation is a post-translational modification where peptides contain a phosphate covalently bound to the serine, threonine, or tyrosine amino acid, as s hown in Figure 1-2. Tran slation, the process of translating mRNA into protei ns, occurs in the ribosome.2 After transla tion, in order to biologically activate proteins, th ey must undergo modifications th at include folding, amino acid removal, functional group addition (i.e. acetyl, phosphoryl, methyl, or carboxyl) or the addition

PAGE 17

17 of an oligosaccharide.2 Many proteins are post-translatio nally modified, and understanding the effect of phosphorylation on a pr oteins biological function can clarify biological mechanisms, including disease dysfunctions. However, el ucidating phosphopeptides is challenging. Phosphopeptides are inherently difficult to anal yze due to their low abundance in a complex mixture. Enrichment methods, such as TiO2 columns and immobilized metal affinity chromatography, have facilitated phosphopeptide discovery, but are time-consuming.3 Immunoassays have difficulty accurately de tecting phosphopeptides du e to interferences.3 Therefore, a high-throughput separation method is needed that can separate post-translational modifications from a complicated matrix. FAIMS, a type of ion mobility, was utilized in these studies to accomplish the separation of phosphopeptides and charge-states. Introduction to Ion Mobility W ith the growing need for faster analytical methods for data-dependent analysis, the benefit of ion mobility spectrometry (IMS) has been explored in recent years. IMS is a rapid gas-phase separation technique that offers orthogonal separati on to liquid chromatography. A variant of IMS, FAIMS, was introdu ced in 1993 by I. A. Buryakov, et al.18 Since its inception, it has become a valuable analytical device capable of more selective separa tion of gas-phase ions than traditional ion mobility spectrometry. Ions are separated in approximately 100 ms, making FAIMS an excellent resource for high-throughput an alysis. Both IMS and FAIMS provide rapid and selective separation more rapidly than conve ntional separation techno logy, such as capillary electrophoresis or high performance liquid chromatography. Traditional IMS IMS has bee n used in recent years for a wide variety of applications, including military, biological, and synthetic polymer analysis.11,19-23 Ion mobility separates ions in 500 s to 2 ms based on their mass, charge, and collisional cross section.19,24 IMS can be used in tandem with a

PAGE 18

18 mass spectrometer or stand-alone using detection by a Faraday cup.20,25 Ions are pulsed into a drift region where they travel th rough a buffer gas at atmospheric pressure while being pulled by a uniform electric field (1 500 V/cm) in the dire ction of the detector, as shown in Figure 1-3. Collisions slow down an ion as it travels through the buffer gas. Thus, the greater the collisional cross section, the more time an ion requires to re ach the detector. Therefore, smaller ions will have higher mobilities, as shown in Equations 11 to 1-4. Equation 1-1 demonstrates that an ions velocity (vd, cm s-1) is equal to the mobility (K, cm2 s-1 V-1) times the electric field. (E, V/cm) An ions mobility can be calculated using the length of the drift cell (d, cm) and the drift time (td, s), as shown in Equation 1-2. vd = KE (1-1) K = d/(tdE) (1-2) To correct for the standard temperature (T, in Kelvin) and pressure (P, in torr) conditions, an ions reduced mobility is calculated, as shown in Equation 1-3. K0 = K T P 760 273 (1-3) Equation 1-4 describes the relationship between analyte characteristics and experimental conditions at low electric fields, where q is the ions charge, N is the number density of the drift gas, is the ions reduced mass, k is the Boltz mann constant, T is the drift gas temperature in Kelvin, and D is the collisional cross section.19,26

PAGE 19

19 (1-4) The sensitivity of IMS can be enhanced by in creasing the field strength or by altering the drift gas composition.27 Different drift gases will have differe nt ion-neutral inte ractions and alter the drift time. Peptide Analysis by IMS Conventional IMS has been used for a vari ety of proteomic applications, including determination of gas-phase conformation11,28-32, separation of isomeric peptides33 and the separation of proteins and pep tides from complex mixtures.22,23,25,28,34 Factors affecting ion mobility in IMS are well understood (see Equation 1-4), and general trends of peptide ion behavior have been observed including the effect of charge state, molecular weight and cross section on mobility. For example, charge-state affects ion drift time, as previously shown in Equation 1-4, because doubly charged ions experi ence twice the drift force compared to singly charged ions. For peptides with similar m/z the [M + 2H]2+ have higher mobilities than the [M + H]+ ions.28 Like many other compounds, peptides also have a correlation be tween drift time and mass, as indicated by Equations 1-2 and 1-4.35 Compounds with higher molecular weight require increased drift time. In addition, the cross section of a peptide or prot ein affects ion residence time. Larger cross sections will slow down the compound due to collisions, thus increasing the drift time. Based on the drift time, the cross sec tion of a peptide can be estimated using Equation 1-5, which is important in gas-pha se ion conformation determination.24 In Equation 1-5, kb is the Boltzmann constant, T is the temperature in Kelvin, mI is the mass of the ion, mB is the mass of the buffer gas, tD is the drift time, L is the drift tube length, P is the pressure in Torr, N is the

PAGE 20

20 number density and ze is the charge on the ion.24 However, determining the cross-section from ion mobility loses accuracy with peptides exceeding 15 amino acids.28 Furthermore, conformational differences, like helices, or phos phorylated or metal-cationized peptide ions demonstrate abnormal behavior in IMS.11,36-38 These drift time variations that are due to structure facilitate peptide separation. (1-5) The exclusive separation of charge-states and phosphopeptides by IMS has been studied without much success. Ruotolo et al. was not ab le to exclusively separate phosphopeptides from nonphosphorylated peptides via IMS.36 However, Ruotolo was able to classify phosphopeptides according to their behavior. One class of phos phopeptides had a higher mobility than their nonphosphorylated analogues, while the other class of phosphopeptides did not deviate in mobility. It was theori zed that the increased mobility of phosphopeptides was due to their comparatively decreased collision cross section.36 Despite the fact that phosphopeptides cannot be separated in time from nonphosphorylated peptides, phosphopeptides can still be distinguished using IMS because most phosphop eptides fall on a different trend line than nonphosphorylated peptides when drift time versus m/z is plotted. Separation of peptide charge-states by IMS ha s also been studied. Barnes et al. and Valentine et al. demonstrated that +2 charge-states had shorter drift times than +1 charge-states of similar molecular weight, but could not ex clusively separate them based on drift times.39,40

PAGE 21

21 Because IMS could not separate charge-states and phosphopeptides, FAIMS was explored as an alternative separation devi ce due to its enhanced selectivity over IMS. Introduction to FAIMS High-field asymmetric-waveform ion mobility spectrometry (FAIMS) is a technique used to separate ions in the gas phase. It is comm only used in tandem with a mass spectrometer, but other detectors have been reported.41,42 FAIMS-MS is a growing technique and has been used in a variety of fields including bioanalytical forensic, geological, and homeland defense applications.42-50 Because FAIMS separation is ort hogonal to liquid chromatography, LCFAIMS-MS can be employed for samples with complicated matrices. 51 Previous literature has reported the ability of FAIMS to lower the limit of detection for perchlorate in drinking water, while also eliminating isoba ric interferences from bisu lfate and dihydrogenphosphate.45 FAIMS can also separate isotopes, isomers, and charge-states.45,52,53 Versatility, selectivity and speed of analysis are all excellent benefits FAIMS offers. Fundamentals of FAIMS FAIMS separates gas-phase ions at ambient pressure due to their difference in mobility at high fields and low fields.18 FAIMS can be used alone with a Faraday cup as a detector or, as more commonly utilized, FAIMS can be used in tandem with a mass spectrometer. The FAIMS device is located between the ionization source and the entrance to the mass spectrometer, as shown in Figure 1-4. A variety of ionization sources can be used with FAIMS, including atmospheric pressure chemical ionization(APCI), electrospray ioniza tion (ESI), heated electrospray ionization (HESI), and distributed plasma ionization source (DPIS).43,54,55 Once generated by the ionization source, ions enter the FAIMS electrodes through the entrance plate. The entrance plate is typically held at +1000 V for positive ions and -1000 V for negative ions. The entrance plate

PAGE 22

22 helps focus ions from the source into the FAIMS device and reduces the am ount of neutrals that enter. Once inside the FAIMS device, ions trav el between two electrodes until reaching the exit, as shown in Figure 1-5. A carri er gas flows in the direction of ion travel facilitating the desolvation and pneumatic movement of the io ns through the FAIMS electrodes. Carrier gas flow rates depend on the cell geometry but vary from 1 to 4 L/min.56 The composition of the carrier gas can be varied for optimal FAIMS perf ormance. The most common carrier gas used is nitrogen, but helium, oxygen, SF6 and CO2 have also been used in a variety of combinations.57 The two electrodes the ions tr avel between in the FAIMS device are typically made of stainless steel. The outer electr ode is held at ground or a small bias. An asymmetric waveform is applied to the inner electrode. The wavefo rm consists of a high voltage (> 10,000 V/cm) portion applied for a short time period and a low voltage portion of opposite polarity (<500 V/cm) applied for a longer period. The time-vo ltage integral for the two fields is equal. The amount of voltage between zero and the peak of the waveform is known as the dispersion voltage (DV).42 Commercial FAIMS devices currently have a DV range between -5000 V and +5000 V. Figure 1-6 shows the ideal waveform for FAIMS. However, the power consumption of the square waveform makes using it unrealistic at high voltages.58 Instead, the asymmetric waveform is produced from a 750-kHz sinusoidal wave and its first harmonic in a 2:1 ratio, as shown in Figure 1-7 and Equation 1-6, where Vmax is equal to the DV, h is 2, is the phase shift and is equal to /2, and f = 2 (most often), 3, or 4.58 V(t) = ([f sin wt + sin (hwt-)] Vmax) / (f+1) (1-6)

PAGE 23

23 Given that the time-voltage integr al for both fields is equal, if an ions mobility were equal at both high and low fields it would not encounter any net disp lacement between the electrodes over time, as shown in Equations 1-7 to 1-9, where vhigh is the velocity at high field, vlow is the velocity at low field, Kh is the mobility at high field, Kl is the mobility at low field, thigh is the time at high field, tlow is the time at low field, dhigh is the distance at high field and dlow is the distance at low field. vhigh = KhEhigh (1-7) dhigh = vhigh thigh = KhEhighthigh (1-8) dlow = vlow tlow = KlElowtlow (1-9) However, an ions mobility is altered at high fi elds. The mobility of some ions increases at high fields (Kl
PAGE 24

24 times. The net displacement of the ion will cause it to drift toward one of the two electrodes where it will be extinguished, as shown in Figure 1-10. To counteract this drift and allow the i on to pass through the FAIMS device without extinguishing on an electrode, a DC voltage calle d the compensation voltage (CV) is applied to the inner electrode in addition to the asymmetric waveform. The CV required for an ion to traverse the FAIMS device is compound-dependent. Commercial FAIMS instruments can scan CV over a range from -200 V to +200 V. Most ions have CVs between -30 and 30, although CVs of 120 V have been reported.57 The CV can be scanned ove r a specified range for the passage of different ions over time, or the CV can be set at one voltage allowing the passage of only the ions that require that part icular CV. If the voltage is varied with time, certain ions are allowed to traverse the FAIMS cell to the detector while others with the incorrect CV extinguish on the electrodes, as shown in Figure 1-11. When the CV value changes, ions that previously traversed the device may extinguis h on the electrodes, while other ions are now allowed to pass through the cell. CVs can be scanne d as quickly as 20 V/s, but are typically scanned at a rate of 10 V/s, in order to allow sufficient mass spectr al data collection (at least 10 data points per peak). In contrast to IMS, where an ions behavior is typically well-characterized, an ions required CV cannot be predicted at this time. Ho wever, four different modes have been defined to describe the typical parameters required for an ion to traverse the FAIMS electrodes based on its size and charge. FAIMS Modes The two-dimensional atmospheric pressure i on focusing that occurs within a cylindrical FAIMS device facilitates ion transmission, but only when the correct DV and CV polarity are utilized. Otherwise, de-focusing occurs and ions ar e lost to the electrodes. The focusing effects

PAGE 25

25 dependence on polarity classifies ions in to four modes: P1, P2, N1, and N2.56 The polarity of the ion is distinguished by N or P (negative or positive, respectively). The Kh/K behavior of the ion is characterized as either or 2 (increasing mobility or decreasing mobility, respectively). Therefore, type A ions are classified as P1 and N1 because their mobility increases with electric field. Type C ions are focused in N2 or P2 modes because their mobility decreases at high fields. FAIMS Parameters Affecting Ion Transmission A variety of instrumental parameters affect the CV an ion needs to traverse the FAIMS device, including the shape of the waveform, gas pressure, cell geometry, gas composition, dispersion voltage, and temperature.42 In order to gain a be tter understanding of FAIMS separation, the latter four parameters will be investigated in this work. Carrier Gas The carrier gas used in FAIMS can dramatical ly affect separation. The composition and flow rate of carrier gas can be optimized for enhanced resolution. Typical carrier gas flow rates for commercial FAIMS devices range from 2.5 L/min to 4 L/min and help to desolvate and pneumatically move ions to the MS entrance. The excess carrier gas that does not enter the MS exits out of the entrance plate to further deter neutrals from enteri ng and facilitate desolvation. Common carrier gas compositions are N2 and N2/He combinations, but combinations of CO2, SF6, O2, and N2O have been used.57 The mobility of ions varies with differing gas compositions. Blancs law describes the mobility of an ion at low field in a gas mixture, shown in Equation 1-10, where Kmix is the mobility of an ion in the mixture of gas, xi is the abundance of a gas and Ki is the mobility in the individual gas. For instance, a CV needed for a particular ion in N2/O2 can be calculated using Blancs law.59 However, other gas mixtures, such as CO2/He, SF6/He and He/N2, deviate from

PAGE 26

26 Blancs law at high fiel ds and have much higher CVs than anticipated.57 Higher CVs create more focusing due to higher fields and th erefore are desired. It was dete rmined that mixtures of gases that greatly varied in molecular weight and cross section, such as He and SF6, deviated the most from Blancs law.57 (1-10) While N2 is the most common carrier gas, the add ition of He to the carrier gas has shown improvements in separation and sensitivity. For example, the use of a N2/He carrier gas was shown to lower the detection limits for codeine.60 The addition of He to the carrier gas is also known to shift CVs to greater absolute values.27 Higher CVs mean the ion will see higher fields and is more focused and able to reach the det ector, thus increasing the limit of detection. However, even though increasing the percentage of He in the carrier gas improves the resolution of analytes, there is some limitation to He conten t in the FAIMS carrier gas. At concentrations higher than 50% He in nitrogen, electrical break down within the FAIMS device occurs. Also, He is a much lighter gas than N2 and can challenge the pum ping capacity of the mass sepctrometer. SF6 is an inert gas commonly used in the semi conductor industry because of its ability to suppress electrical discharge. Therefore, it has been investigated as a FAIMS carrier gas complimentary to He. The addition of SF6 to He in the carrier gas increases the breakdown voltage to approximately 4 to 6 kV across a 2 mm gap.57 Traditional carrier gas compositions are limited to only 50% He, but the addition of SF6 allows He concentrations of up to 95%.57

PAGE 27

27 Combinations of SF6 and He have shown non-Blanc beha vior and benefit from enhanced resolution and peak capacity.57 This work demonstrates the application of SF6 to the separation of tryptic peptides. CO2 has also been evaluated as a FAIMS carrier gas. It has shown re markable abilities to separate isomers of phthalic aci ds, as well as isobaric compou nds bisulfate and perchlorate anions.59,61 Though not much is known about why CO2 facilitates ion separation, it is theorized that it is caused by ion-gas complexes that form at low fields, slowing down the mobility of the analyte ions. At high fields, these complexes dissociate, decreasing the collisional cross section and allowing the ion to travel at a higher mobility. The formation of CO2/analyte ion complexes at low fields and their subseque nt dissociation at high fields improves separa tion of analyte ions. This work evaluates the effect of different gas compositions, includi ng varying percentages of He, N2, CO2 and SF6 on the separation of tryptic peptides. To our knowledge, this is the first report of the application of FAIMS carrier gas compositions on phosphopeptide separation. Dispersion Voltage The dispersion voltage (DV) is the maximu m peak of the voltage applied and will be varied from -2500 V to -5000 V for these experi ments. A higher DV will provide greater ion focusing. It has been shown that increas ing the DV improves separation of analytes.62 The effect of dispersion voltage on tryptic peptide ion separation and transmission will be evaluated here. Electrode Temperature Temperature is also known to have an effect on the separation of analytes by FAIMS. Improved separation for a solution of taurocholic acid and methotrxate has been demonstrated using FAIMS with elevated electrode temperature.63 The effect of temperature on ion separation can be described by Equations 1-11 to 1-13. The mobility of an ion at high fields is described by

PAGE 28

28 Equation 1-11, where K is the mobility constant, E is the electric field, and N is the gas number density. The number density (N) can be found using Equation 1-12, where n/V is determined from the ideal gas law and NA is Avogadros number. From the equations, it is determined that as temperature increases, n/V decreases causing th e number density to decrease and therefore the effective E/N to increase, as shown in Equation 113. Due to the increase in the value of E/N, stronger fields will be required to maintain balanced conditions for an ion of interest within the FAIMS cell. The higher fields decrease ion focusing between the electrodes, thus improving resolution. Temperature of the electrodes will be independently varied in the range of 3090C to evaluate the effect on tryptic peptide resolution and transmission. Kh(E/N) = K[1 + f(E/N)] (1-11) N = (n/V)NA (1-12) n/V = P/(RT) (1-13) Cell Geometry The geometry of the FAIMS electrodes affects ion transmission and resolution. The first FAIMS geometry patented was a planar geometry by Buryakov et al. in 1993; since then, other geometries including the dome, cylindrical, and spherical geometries have been patented.18,64-66 The planar geometry has superior resolution to the cylindrical geometry, though a cylindrical geometry benefits from improved ion transmission when compared to the planar geometry due to the non-uniform electric field.65,67 A novel hemispherical geometry will be presented that improves the resolution when compared to th e commercial Thermo cylindrical FAIMS cell.

PAGE 29

29 Selection of the Ionization Source The transfer of large biomolecules from the condensed phase to th e gas phase has been reported by several methods.68,69 Matrix-assisted laser desorption ionization (MALDI) is one method that can ionize peptides and proteins. In MALDI, the analyte is mixed with a matrix, such as 2,5-dihydroxybenoic acid.68 Laser radiation penetrates the matrix, ionizing the matrix and subsequently the analyte. MALDI does not usually produce multiply charged ions and would not interface well with FAIMS because of coupling a pulsed ionization source with a scanning separation device. ESI is the ionization technique of choice beca use it is able to generate multiply charged gas-phase ions from ions that are preformed in solution. ESI produces a continuous flow of multiply charged ions and is therefore more compatible for coupling with FAIMS than MALDI. The sample will be directly injected into the ESI-FAIMS-MS via a syringe pump leading to the ESI source. The flow rate of the sample and ESI parameters will be op timized with the FAIMS device in place. In ESI, the sample is sprayed through an elec trically charged tip creat ing a mist of charged droplets containing the mixture of peptides. A high voltage (3-6 kV) is applied to the ESI needle causing multiple charges to distribute along the surface of the droplets. The voltage causes electrophoretic separation and pulls the solution into a Tayl or cone, which produces droplets 110 m in diameter with multiple charges.69 One theory of gas-phase ion formation suggests asymmetric droplet fission causes multiple charging during ion formation. In this theory, solvent evaporates from each drop until reaching the Rayleigh stability limit where electrostatic forces overcome surface tension and break the droplet apar t into smaller droplets with fewer charges.69 If the ions entering FAIMS are too solvated, resolution is reduced due to peak broadening.

PAGE 30

30 Instrument Description Once ions are generated by ESI, they will enter into the FAIMS device. The FAIMS devices used in this work have cylindrical geometry (Thermo FA IMS, Thermo Fisher Scientific, San Jose, CA and Ionalytics FAIMS, Ionalytics, Ottawa, Canada), as shown in Figure 1-4. After passing through the FAIMS device, ions enter the Thermo-Finnigan LCQ ion trap. A schematic of the LCQ ion trap is shown in Figure 1-12. Ions pass through the skimmer cone and through a series of ion guides and lenses. They are then pulsed into the ion trap mass analyzer, which is composed of two e ndcap electrodes and a ring electrode. The ions enter the ion trap mass analyzer thr ough a hole in the entran ce endcap electrode. They collide with helium and are slowed down and maintained in the center of the trap. Oscillating RF voltages are app lied to the ring electrode. The hyperbolic endcap electrodes are held at ground, biased at a constant DC or held at an oscillating AC potential.70 This creates a 3D quadrupole field in the mass analyz er, which traps ions in their stable trajectories. As the amplitude of the RF voltage is ramped, ions of increasing mass no longer have a stable trajectory and are ejected from the trap. This process is known as mass selective instability scanning.70 Ions of lower mass are ejected first. Upon exiting the trap, ions are converted to electrons using a conversion dynode. These electrons enter the el ectron multiplier and ar e amplified for signal detection. Overview of the Dissertation This dissertation presents the application of FAIMS to tryptic peptide analysis. In particular, the separation of ch arge-state and phosphopeptides is addressed. Development of a novel FAIMS electrode geometry to improve resolution and sensitivity is presented. Chapter 1 has presented an introduction to tr yptic peptide analysis by mass spectrometry, with an emphasis on the importance of doubly charged and phosphorylated tryptic peptide

PAGE 31

31 separation. Current techniques for enriching doubly charged and phosphorylated tryptic peptides have been discussed. The fundamentals of IMS and its application to th e analysis of tryptic peptides were included. FAIMS was discussed in detail, including f undamentals of operation and factors affecting resolution and transmission. The chapter concluded with a brief overview of electrospray ionization and ion trap mass spectrometry. Chapter 2 explores charge-state separati on by FAIMS for the improvement of protein identification in data-dependent scanning. A br ief history and the importance of data-dependent scanning is presented. The adva ntages and disadvantages of elec trospray ionization as it relates to charge-state separation are discussed. Altern ative techniques for charge-state separation are detailed. FAIMS parameters are optimized for charge-state separation, including dispersion voltage, carrier gas flow rate and carrier gas composition. Charge-state separation is demonstrated on a tryptic digest of cytochrome C. In Chapter 3, the application of FAIMS to the separation of phosphopeptides is presented. The biological importance of phosphopeptides is disc ussed, as well as the challenges in enriching and detecting phosphopeptides in complex mixtures. Utilizing non-traditiona l carrier gases such as SF6 and CO2 for the separation of phosphopeptides is ex plored. Chapter 3 also examines the effects of heated FAIMS electrodes on phosphopeptid e detection. The chapter concludes with a discussion of integrating FA IMS into a LC-MS method. In Chapter 4, a novel hemispherical FAIMS elec trode design is presen ted. A history and a discussion of electrode geometries in relation to resolution and transmission is presented. Data obtained on the hemispherical and spherical FAIMS electrodes are compared to data obtained on the commercial cylindrical FAIMS electrodes. The application of th e novel FAIMS electrodes to the separation of phosphopeptides is discussed.

PAGE 32

32 Chapter 5 discusses conclusion and future work. The benefits and disadvantages of utilizing FAIMS for the separati on of charge-states and phosphopeptid es is discussed, as well as future work with hemispherical FAIMS.

PAGE 33

33 a ion y ion y ion b ion b ion c ion a ion y ion y ion b ion b ion c ion d ion z ion x ion w ion d ion z ion x ion w ion v ion Figure 1-1. Nomenclature of tryptic peptide ion fragments. Adapted from Wysocki et al.5

PAGE 34

34 N O O O R P OH O OH R phosphoserine N O O O R P O OH OH R phosphothreonine N O O O R P O OH OH R phosphotyrosine N O O O R P OH O OH R phosphoserine N O O O R P O OH OH R phosphothreonine N O O O R P O OH OH R phosphotyrosine Figure 1-2. Three common phosphorylated amino acids.

PAGE 35

35 Figure 1-3. Typical ion mobility spectrometer.19

PAGE 36

36 Figure 1-4. Side-by-side FAIM S cell expanded. Illustration by Thermo-Fisher Scientific, 2005 Ions from Source Ions to Mass Spectrometer Inner Electrode Outer Electrode Outer Electrode Entrance plate Figure 1-5. Side-by-side FAIMS cell. Adapted from an illustration by Richard Yost, 2008.

PAGE 37

37 Figure 1-6. Representation of a square waveform and altered ion movement between electrodes as the field varies.62 Figure 1-7. Actual asymmetric wavefo rm applied to the inner electrode. DV (-5000 to 5000) CV (-200 to 200)

PAGE 38

38 Figure 1-8. Classification of ion behavior in FAIMS.62 The mobility of type A ions increases at high fields, the mobility of type C ions decreases at high fields and the mobility of type B ions first increase before decreasing at even higher fields. Figure 1-9. Net displacement of an ion over time due to the asymmetric waveform. Illustration by Christ opher Hilton, 2004. Gas Flow ~ Inner Electrode Outer Electrode Net Displacement

PAGE 39

39 Figure 1-10. Potential path of the ion in the FAIMS cell.71 Figure 1-11. The top electrodes show a mixture of ions entering FAIMS. All ions exit because the CV is scanned. In the bottom electrodes, the CV is set to a voltage so only the blue ions enter the mass spectrometer, while the other ions are lo st to the electrodes. Adapted from an illustration by Alisha Roberts, 2008. Application of the compensation voltage allows the ion to traverse the FAIMS cell. Here the compensation voltage has not been applied and the ion drifts towards an electrode as its mobility increases at high fields. Here the compensation voltage has not been applied and the ion drifts towards an electrode as its mobility decreases at high fields. CV 1

PAGE 40

40 Figure 1-12. Thermo LCQ ion trap used in these experiments. Adapted from R. Cole.69

PAGE 41

41 CHAPTER 2 CHARGE-STATE SEPARATION OF TR YPTI C PEPTIDE IONS BY FAIMS Overview This chapter discusses charge-state separation by FAIMS as a method to increase the signal-to-noise (S/N) of releva nt tryptic peptide ions to im prove the accuracy of protein identification by data-dependent analysis. The advantages and disadvantages of electrospray ionization as it applies to charge-state separati on for data-dependent scanning is addressed. Alternative techniques for the si mplification of tryptic peptid e ion product spectra will be discussed. Three experiments exploring the opt imal parameters for charge-state separation by FAIMS were designed. The first experiment optimized the dispersion voltage, whereas the second and third experiments focused on optimizing the carrier gas flow rate and composition for tryptic peptide charge-state separation. Charge -state separation is demonstrated on a tryptic digest of cytochrome C. The effect of amino acid sequence on tryptic peptide ion behavior was explored. Finally, the advantag es and disadvantages of implementing FAIMS as a charge-state separation technique to benefit data-d ependent analysis are addressed. Advantages and Disadvantages of Electrospray for the Analysis of Tryptic Peptide Ions Protein digest analysis is an important technique for the id entification and sequencing of proteins. Analysis of protei n digests by mass spectrometry (MS) has been a growing field in recent years due to new ionization technol ogy, like electrospray ionization (ESI).37,72 One of the main benefits of ESI is that it produces multiply charged peptide ions, which give large peptide ions a mass-to-charge ratio (m/z ) that is in the detectable range for triple quadrupole or ion trap mass analyzers.73 However, a disadvantage of this ioni zation technique is th at each peptide may have several different charge states that ma ke the subsequent tandem mass spectra more complicated. The desired multiply charged tryptic peptide ion may be present at analytically

PAGE 42

42 relevant levels, but can be obscu red by interferences from other charge-states and other tryptic peptide ions. Therefore, the an alysis of protein digest spectra becomes very complicated with numerous charge-states and inte rferences, making it difficult to elucidate the protein sequence by data-dependent analysis. Data-dependent Analysis of Tryptic Peptides With the tremendous number of protein digest sp ectra that need to be analyzed, the concept of database searching and id entification was first introduced in 1977 by Cleveland et al.74 However, it was not until 1993 th at algorithms for database se arching were simultaneously developed by five independent laboratories.75-79 The most popular algorithm, SEQUEST, automatically analyzes MS data by picking a predefined number of peaks for tandem MS. Digested proteins spectra are then compared to predicted mass spectra from proteins in the database in order to identify them.80 A problem with SEQUEST or any other computer algorithm is that the peaks of the highest abundance are selected for further analysis regardless of their relevance. A solvent cluster ion could be erroneously selected for further anal ysis by an automated program if it happened to have the highest abundance, despite that it w ould provide no useful information. Although SEQUEST is capable of identifying some common contaminants, the computationally expensive process is not amenable to high-throughput analysis.6 Therefore, relevant data are sometimes overlooked. Automated data analysis would become more efficient if MS spectra contained predominantly relevant peaks in high abundance. As the +2 charge-state of tryptic peptide ions yields the most relevant tandem MS sequencing information for automated analysis, in an ideal situation, each MS spectrum for a tryptic digest would contain all tryptic peptide fragment ions solely in the +2 charge-state in high abundance.81 The subsequent product spectrum from a +2

PAGE 43

43 charge-state parent ion would c ontain more structural informati on because more fragments will have a charge and be detected. By exclusiv ely separating the +2 charge-state, and thus simplifying the MS spectra and targeting more re levant ions, data-depen dent analysis would become more accurate. Simplifying Tryptic Peptide Ion Product Spectra Reducing Background Noise Current technology attempts to overcome the dilemma of complicated tryptic digest spectra in numerous ways. Methods such as high performance liquid chromatography (HPLC), capillary electrophoresis (CE), and tandem MS have been implemented to help eliminate interferences.81,82 However, these techniques are incapable of isolating a partic ular charge-state of a peptide exclusively. Both CE and HPLC are also time-consuming separation techniques (minutes to hour time frame) as neither is able to separate as quickly as the mass analyzer can scan in most circumstances. When CE and HPLC are used in conjunction with QITMS, they are often the time-limiting step in the analysis of tryptic digests.82 On the other hand, mass spectrometric analysis sometimes is the time-lim iting step when advanced scanning is necessary, such as MS/MS of many coeluting ions. Peak parking can be employed to give the mass analyzer time for complete scans.6 Current Charge-State Separation Technology Other technology has limited success in simplifying tryptic peptide product ion spectra by exclusively separating the +2 charge-state. Rudi mentary charge-state separation devices such as secondary electron resolved mass spectrometry (SERMS) or Fourier transform (FT) filters function more quickly than alte rnative separation techniques, but encounter difficulties with successfully separating the +2 ch arge-state from interferences.83,84 Software methods, such as FT filters, require highly repetitive chemical noise to succeed.84 SERMS is a low resolution

PAGE 44

44 technique and fails to distinguish accurately between +1 and +2 charge-states, leaving a majority of the interferences remaining in the spectrum.81,85 Perhaps the most promising of current technolo gy are charge-state separation devices that utilize ion mobility or multiply charged scan for separation of tryptic peptide ions. 86,87 81 They distinguish between certain charge -states, but cannot reproducibly separate lower charge-states. Therefore, there is a need for a selective, hi gh-throughput separation device for charge-states of tryptic peptide ions. Application of FAIMS to Charge-State Separation of Tryptic Peptide Ions The speed and selectivity of FAIMS makes it an excellent candidate for the application of charge-state sepa ration. Separation of tryptic peptide ion charge -states by FAIMS has shown promising data.41,54,88,89 Guevremont et al.54 demonstrated that variation in the carrier gas composition improved the separation of the tr yptic peptide ions by FAIMS, as shown in Figure 2-1. However, to date, previous work has not established FAIMS as a charge-state separation device for high-throughpu t tryptic peptide analysis. Here is presented a series of experiments to evaluate the optimal FAIMS parameters for charge-state separation. These parameters in clude variation of carrier gas composition and electrode temperatures. These parameters were first evaluated on the synthetic tryptic peptide DLLXR, where X is all of the natural amino acids except I, Q, and C, prior to their evaluation on a tryptic digest of cytochrome C. The expect ed molecular weights and experimentally observed m/z for ions of DLLXR are s hown in Table 2-1. All m/z for DLLXR ions differed from the expected value by 16 or 17, which is likel y the N-terminal under going cyclization to succinimide. (Alfred Chung, personal communication) The effect of varying amino acid sequence on CV was evaluated briefly using synthe tic tryptic peptides. These experiments were intended to evaluate the ability of FAIMS to sepa rate and improve the S/N of charge-states in a

PAGE 45

45 tryptic digest. An increase in S/N of +2 char ge-state tryptic peptide ions would improve the success of data-dependent analysis. Additionally, FAIMSs ability to filter ions on the millisecond time-frame can greatly reduce the time consumed by peak parking or LC gradients, thus decreasing the analysis time. Experimental Synthetic Tryptic Peptides The first analyte evaluated for charge-sta te separation by FAIMS was DLLXR, where X can be any of the natural amino acids with the exceptions of I, C, and Q. Both I and Q were not used because they are isobaric with L and K, respectively. The amino acid C was not used due to the possible formation of disulfide bridges. Appendix 1 shows all the amino acid structures and their abbreviations. Working solutions of each of the synthetic tryptic peptide standards were made at a concentration of 10 ppm in 50/ 50 methanol/ HPLC grade water with 1% acetic acid. Synthetic tryptic peptides were also used to evaluate the effect of amino acid sequence on CV. DLLXR, DLLXK, AWXVAR, and VAXLR, whose molecular weights are shown in Table 2-1 to 2-4, were the synthetic tryptic peptid es used in the sequence experiments. These sequences were selected because of their behavi or in a large data pool consisting of 254 tryptic peptides, as shown in Figure 2-2.54 AWSVAR was an outlier in the +2 charge-state with a CV of -34.1 V, while it required a more average CV (-8.7 V) in the +1 charge-state. VASLR2+ and VATLR2+ had somewhat lower than average CV valu es of -11.3 V and -15.6 V, respectively, but only differ by one amino acid from each other. DLLER and DLLFK had average CV values of -18.2 V and -16.5 V, respectively, for +2 tr yptic peptide ions. Ther efore, DLLER and DLLFK demonstrated average tryptic pe ptide behavior while also allo wing the study of the effects of the terminal amino acid. Working solutions of each synthetic tryptic pe ptide standard mixture

PAGE 46

46 were made at a concentration of 10 ppm in 50/ 50 methanol/ HPLC grade water with 1% acetic acid. Cytochrome C Tryptic Digest A tryptic digest was performed by adding 0.1 mg of protein (cytochrome C, SigmaAldrich) and 1 g Trypsin Gold (mass spectrometry grade, Promega) in a buffer of 50 mM NH4HCO3 (pH 8.4). The digest was placed in a 1 mL Eppendorf tube and allowed to sit overnight at room temperature and subsequently quenched with 5% acetic acid. De-salting was not necessary since the tryptic peptides were then diluted to 10 ppm in 50/50 HPLC grade water/methanol with 1% acetic acid. Instrument Parameters The instrumental parameters used in these experiments are shown in Table 2-5. 10 ppm samples were directly infused at 3 L/min and ionized by electrospray. For these experiments, two cylindrical FAIMS devices were used separa tely (Ionaltyics, Ottawa Canada or ThermoFisher Scientific, San Jose, CA) coupled to a Th ermo LCQ ion trap. A detailed explanation of coupling a Thermo FAIMS to the LCQ will be shown in Chapter 3. FAIMS Parameters The two FAIMS devices used were the Ionaly tics FAIMS with cylindrical electrodes with a 2 mm analytical gap between the electrodes and the Thermo FAIMS with cylindrical electrodes with a 2.5 mm analytical gap. The DV was varied from -5000 to +5000 V on the Thermo FAIMS and -4000 to +4000 V on the Ionalytics FAIMS in order to maintain the same field across the analytical gap. The carrier gas flow rate was varied between 2.5 L/min and 4.5 L/min to determine the optimum. The carrier gas co mposition was varied from 0 to 50% He in N2. Excess of 50% He in the N2 carrier gas could not be used due to arcing.

PAGE 47

47 Results and Discussion A series of experiments were designed to te st the charge-state separation capabilities of FAIMS. The FAIMS parameters (DV, carrier gas flow rate and composition) were optimized for charge-state separation of synt hetic tryptic peptides DLLXR. Subsequently, charge-state separation was demonstrated on a digest of cyto chrome C. Finally, the effect of amino acid sequence on charge-state separa tion was evaluated using synthe tic tryptic peptides DLLXR, DLLXK, AWXVAR, and VAXLR. Optimizing FAIMS Parameters for Charge-State Separation In order to achieve optimal charge-state se paration, it was necessary to optimize FAIMS parameters, including DV and carrier gas flow rate and composition. DV Optimization for Tryptic Peptide Ions The DV is the maximum peak of the applied voltage. In agreement with recently published literature, it was found that incr easing the DV increases the ion separation.62 The higher fields improve focusing of ions towards the detector. The current commercial Thermo FAIMS cell offers a DV of up to 5000 V. The DV was varied from +2500 to +5000 V and -2500 to -5000 V in 500 V increments. The optimal DV for tryptic peptides was found to be 5000 V, as shown in Figure 2-3. When positive DV was used, FAIMS peaks were very broad, nonGaussian and had low signal intensity. Optimal ch arge-state separation was also achieved at a DV of -5000 V, as shown in Figure 2-4. As th e absolute value of the DV decreases, the signal intensity and separation decrease. The red peak on the right in Figure 2-4 is the doubly charged dimer of EFSGDK and overlaps with the +2 charge-s tate of EFSGDK. This peak is suspected of being the doubly-charged dimer of EFSGDK becau se it increased as the concentration of EFSGDK increased, while the peak for EFSGDK+1 decreased. The zoom scan function of the

PAGE 48

48 LCQ was not available; therefor e, charge-state could not be confirmed for the doubly-charged dimer peak. Signal decreases substantially starting at a DV of -4000 V. A DV of -5000 V yielded improved signal because higher fields allow impr oved ion focusing. Better resolution may be achieved at higher fields because, as the voltage increases and more energy is imparted to the tryptic peptide ions, they can vary conformation between low and high fields. The different conformations may have different mobilities. The more the mobilities vary between high and low fields for different ions, the better the separation capabilities in FAIMS. The Ionalytics FAIMS has a 2 mm analytical gap, whereas the Thermo FAIMS has a 2.5 mm gap. In order to maintain the same electric fi eld within the analytical gap, all charge-state experiments were performed at a DV of -4000 V when the Ionalytics FAIMS was utilized. Therefore, a field of 20,000 V/cm was maintained in both FAIMS devices. Optimizing Carrier Gas Flow Rate and Composition for Tryptic Peptide Ions Carrier gas flow facilitates de terring neutrals, desolvation, and the pneumatic movement of ions to the detector. As di scussed previously, varying carr ier gas composition alters the resolution of FAIMS.57 The optimal carrier gas flow rate was found to be between 3.5 L/min and 4.0 L/min, as shown in Figure 2-5. Ion signal was maximum at 3.5 L/min, with 4.0 L/min having statistically similar ion intensities. Below 3.5 L/min, ions were more likely to cluster or solvate. These complexes will have different CV values than the bare ion, causing a decrease in the intensity of the desired m/z. Peak shape varied little with flow rate, as shown in Figure 2-6. On average, the peak base wi dth was approximately 4 V wide. The optimal carrier gas com position containing He and N2 was determined to be 50/50 v/v, as shown in Figure 2-7. It wa s determined that increasing the helium percentage in the carrier gas provides increased resolution because higher fields are needed to focus the ions. Higher

PAGE 49

49 fields may be necessary due to the differen ce in size and polarizability between He and N2. He is smaller and less polarizable than N2; therefore, the chance of collisions or interactions with the analyte ion is reduced. With minimal interactions with the carrier gas, the analyte ion may have more flexibility to alter conformations at high and low fields, which would affect its mobility. Tryptic peptide ions have improved intensity at increased He concentrations, as was shown in Figure 2-7. Fewer interactions with the carrier gas may allow improved focusing. There was a slight decrease in tryptic pep tide ion intensity at 50/50 He/N2 for the +2 charge-state DLLPR when compare to 40/60 He/N2, but this is abnormal behavior. Nearly all tryptic peptides ions experience increased ion intensities as the heliu m percentage of the carrier gas increases. Experimentally Determined Op timized FAIMS Parameters for Charge-State Separation These experiments determined the optimized FAIMS conditions for tryptic peptide ion charge-state separation, which are shown in Table 2-6. The best DV was found to be -5000 V, which both improved resolution and peak intensity. The best carrier gas flow rate was between 3.5 and 4.0 L/min. The best carrier gas composition of He and N2 was found to be 50/50 v/v, which improved peak intensity and separation. Evaluating Charge-state Separation of DLLXR The evaluation of the charge-state separati on of DLLXR, where X is any of the natural amino acids except I, C, and Q, was performed using the Ionalytics FAIMS. Instrumental parameters are shown in Table 2-5 and 2-6. Th e ESI-MS spectrum is shown in Figure 2-8a. The +1 charge-state and +2 charge -states of DLLXR ar e present in high a bundance. With the implementation of FAIMS, charge-state separatio n was achieved for the s ynthetic tryptic peptide DLLXR, as shown in Figure 2-8b, 2-8c and Table 2-7. In Figure 2-8b, the CV was scanned from -5.3 to -17.1 V, and only DLLXR ions in the +1 charge-state are observed in the MS spectrum. The +2 charge-state ions of DL LXR are exclusively separated by scanning the CV from -18.9 V

PAGE 50

50 to -23.4 V, as shown in Figure 2-8c. As shown in Figure 2-9, DLLPR+ (CV = -11.9 0.3 V) was separated from DLLPR2+ (CV = -20.1 0.1 V) with a resolu tion of 1.2. The resolution was calculated from equation 2-1, where the difference in CV is multiplied by 2 and divided by the base width of the CV peaks. 212bbWW CV R (2-1) The doubly charged DLLXR ions required a more negative CV than the singly charged DLLXR ions, as shown in Figure 2-10. The CV required for the singly ch arged peptides ranged from -10.9 to -16.4 V, while the doubly charged D LLXR ions required a CV that ranged from -19.4 to -20.9 V. The variation in CV between different doubly charged tryptic peptides ions may have been smaller than the variation in CV between different singly ch arged tryptic peptides ions because two charges on the doubly charge ions may have created less opportunity to form variable structures at high-field due to charge repulsion, making their high-field structures similar. Therefore, +2 charge-state tryptic pe ptide ions of DLLXR woul d have a similar change in mobility between high and low fields and thei r CVs would not vary widely. However, ions with only one charge may form different low en ergy structures from each other and would have very different changes in mobilitie s at high and low field. Therefor e, it would be expected that the +1 charge-state ions would have a greater variation in CV than the +2 charge-state ions. Figures 2-11 and 2-12 show the plot of the tryptic peptide io n intensities versus CV over four CV scans. Singly charged ions (Figure 211) have a higher mobility at high fields than doubly charged ions (Figure 2-12), as indicated by the more negative CV values required for doubly charged ions to traverse the FAIMS device. Repulsion of the two charges may give the

PAGE 51

51 doubly charged ions more flexibility to have di fferent conformations at high and low fields (though they all may have a similar conformation to each other), which would affect their mobility more. Despite the success of charge-state separation of DLLXR tryptic peptides, it was necessary to evaluate a more structurally di verse sample. A digest of Cytochrome C was subsequently evaluated. Charge-State Separation of a Cytochrome C Digest Charge-state separation on a cytochrome C digest was evaluate d using the Ionalytics FAIMS. Instrumental parameters used are s hown in Table 2-5 and 2-6. The ESI-MS spectrum of a cytochrome C digest take n on a Thermo TSQ Quantum (Therm o-Fisher Scientific) is shown in Figure 2-13. The cytochrome C +2 charge-state tryptic peptides are not the most intense peaks. However, with the implementation of FA IMS, charge-state separation was achieved, as shown in Figure 2-14. The doubly charged cy tochrome C ions required a more negative CV than the singly charged ions. Figure 2-15 show s the ESI-FAIMS-MS spectra of cytochrome C. The CV was scanned from -12.1 V to -17.5 V and shows the +1 charge-state in high abundance, as shown in Figure 2-15a. In Figure 2-15b, the CV was scanned fr om -24.9 V to -35.0 V. While the +2 charge-state cytochrome C tryptic pep tide ions are now of higher abundance in the mass spectrum, they are not the only peaks of high abunda nce. Even though it is not the optimum CV for the +1 charge-states, they are in high abundance. The +1 charge-state may have traversed FAIMS in the +2 charge-state and subsequent ly lost a charge prior to mass analysis. Alternatively, the +1 ch arge-state ions may be in such high abundance that even with a small amount passing at an unoptimized CV, they sti ll appear in high intens ity. Data-dependent scanning might overlook EDLIAYLK2+, but the chances of selecting TGPNLHGLFGR2+ are improved by the implementation of FAIMS.

PAGE 52

52 The Effect of Amino Acid Sequence on Tryptic Peptide Behavior in FAIMS The Effect of Glycine The effect of amino acid sequence on tryptic peptide behavior in FAIMS was evaluated using the four synthetic trypt ic peptides DLLXR, DLLXK, AWXVAR, and VAXLR, where X is any of the natural amino acids except I, C, and Q. Instrumental parameters are shown in Table 2-5 and 2-6. It was determined that when glycine (G) wa s present as the variable amino acid of the synthetic tryptic peptides, the CV consistently shifted to a more negative value, as shown in Table 2-8. This may be because the R group of glycine is a hydrogen, minimizing any steric hindrance an alternative amino acid would have provided. Without the steric hindrance of a bulky R group, the peptide would be able to ro tate more freely and may have demonstrated various conformations. With different conformations at low and high field, the mobility between low and high field may have differed, causing th e CV to shift to a more negative value. However, peptides with bulky R groups, such as W or P where the mo st significant steric hindrance is expected, did not demonstrate any noticeable difference in CV values from other peptides. While this is an interesting observation in the synthetic tryptic peptides, it does not conclude that every peptide containing a glycine amino acid will have a much more negative CV than peptides that do not. When observing CV data of peptides from a tryptic digest, those that contained glycine did not always have more ne gative values than thos e that did not contain glycine, as shown in Figure 2-16. Thus, addi tional structural effects are considered in determining the required CV va lue for a tryptic peptide.

PAGE 53

53 The Effect of the C-Terminal Amino Acid The effect of the terminal amino acid was evaluated using DLLXR and DLLXK. In general, the CV values for the peptides were si milar, as shown in Table 2-8, with the exception of when the variable amino acid was G, S, T, or H, which differed by 1.1, 2.6, 1.9, and 1.8 V, respectively, when the terminal amino acid was varied. In general when the CVs varied, DLLXR required a more negative CV than DL LXK. This may be caused by hydrogen bonding between the two nitrogens at the C-terminus of DLLXR and the amino acids S, T, and H, which contain an oxygen or nitrogen th at could form hydrogen bonding. The variation in CV for DLLGR and DLLGK may be due to the flexibi lity of the molecule to form hydrogen bonds between the two nitrogen s on the C-terminus of DLLGR more strongly than could DLLGK. DLLGK only has one nitrogen at the C-terminus and may not have as strong hydrogen bonds as DLLGR. The different conformations at high fields may result in th e differences in the ability to form hydrogen bonding and would explain the vari ation in CV between DLLXR and DLLXK. Conclusion The FAIMS parameters to best separate charge-state ions were determined to be a carrier gas composition of 50/50 v/v He/N2, a DV of -5000 V, and a carrier gas flow rate of approximately 3.54.0 L/min. FAIMS separate d the charge-states of DLLXR and increased the intensity of +2 charge-state ions relative to othe r ions in the spectrum in both a tryptic digest of cytochrome C and with the synthetic peptides D LLXR, allowing a greater chance of selection for tandem analysis by data-dependent scanning. Im proved charge-state separation due to heated FAIMS electrodes is discussed in Ch apter 3. Variation in the CV due to a change in the terminal amino acid was seen when amino acids G, S, T, and H were used in DLLXR and DLLXK. DLLXR required a more negative CV when those amino acids were used, which may have been

PAGE 54

54 due to increased hydrogen bonding. DLLXR and D LLXK did not show a signi ficant variation in CV when other amino acids were used. FAIMS shows promise for high-throughput analysis by filtering ions of interest from the background. For high-throughput char ge-state separation to be pr actical, FAIMS would have to be scanned slowly enough over a range of -10 to -20 V that the MS has enough time to complete MS/MS analysis. Scanning a specified range would overlook some peptides, but may elucidate others due to the increased S/N. LC is still necessary for complex mixtures because FAIMS is a post-ionization separation device and woul d not alleviate ion suppression. There are several factors that must be considered when integrating FAIMS into a LC-MS method. First, the time it takes for the mass spectrometer to complete scanning must be addressed. In data-dependent analysis, the parent scan can take 2 s and the product scans can take 1 s each. If one parent scan and thr ee product scans are completed, then the mass spectrometer needs steady conditions for five seconds The second factor to consider is the LC peak width. If the peak is 30 s wide, 6 parent/p roduct scan sets can be obtained across the peak if the MS needs 5 s to complete a set of data -dependent scans. Lastly, how quickly the FAIMS can scan the CV must be considered. The averag e tryptic peptide ion CV peak in a cylindrical FAIMS is five volts wide and it takes approxima tely 100 ms for an ion to traverse the FAIMS cell. If the CV is stepped 10 V/s, a 10 V range could be scanned in the time the MS is acquiring data. Six points could be acquired at any one CV if the LC peak wi dth is at least 30 s. However, scanning the CV over a large range so quickly can lead to poor quality signal and peak shape. The sample concentration required by FAIMS ma y limit the practicality of its application to charge-state separation as an order of ma gnitude of signal is lost typically.

PAGE 55

55 Due to the complications in scanning FAIM S consecutively with LC-MS, an excellent application of FAIMS would be protein biomarker analysis. The CV could be set for a known +2 charge-state peptide biomarker for rapid id entification in screeni ng. FAIMS could help reduce the time spent on LC gradients, as it can separate co-eluting compounds.

PAGE 56

56 Figure 2-1. Data from Guevremont et al. de monstrating the effect of varying carrier gas composition on the resolution of peptides from an enolase 1 digest.54 a.) The carrier gas was 100% N2 b.) The carrier gas was 50% He 50% N2.

PAGE 57

57 Table 2-1. m/z of DLLXR, where X can be any of th e listed amino acids. The variation in expected m/z versus observed m/z could be due to the loss of an oxygen or hydroxyl group. DLLXR Expected m/z Observed m/z Expected m/z Observed m/z +1 charge -state +1 charge-state +2 charge-state +2 charge-state G 573.7 556.3 287.9 279.2 A 587.7 570.3 294.9 286.2 S 603.7 586.3 302.9 294.2 P 613.7 596.3 307.9 299.2 V 615.7 598.3 308.9 300.2 T 617.7 600.3 309.9 301.2 L 629.8 612.4 315.9 307.2 N 630.7 613.3 316.4 307.7 D 631.7 614.3 316.9 308.2 K 644.8 628.4 323.4 315.2 E 645.7 629.3 323.9 316.7 M 647.8 630.4 324.9 316.2 H 653.8 636.4 327.9 319.2 F 663.8 646.4 332.9 323.2 R 672.8 655.4 337.4 328.7 Y 679.8 662.4 340.9 332.2 W 702.8 685.4 352.4 343.7

PAGE 58

58 Table 2-2. m/z of DLLXK, where X is any of the listed amino acids DLLXK m/z _+1 charge-state G 545.3 A 559.3 S 575.3 P 585.4 V 587.4 T 589.4 L 601.4 N 602.4 D 603.3 K 616.4 E 617.4 M 619.3 H 625.4 F 635.4 R 644.4 Y 651.4 W 674.4

PAGE 59

59 Table 2-3. m/z of AWXVAR, where X is any of the listed amino acids AWXVAR m/z of +1 charge-state G 695.4 A 673.4 S 689.4 P 699.4 V 701.4 T 703.4 L 715.4 N 716.4 D 717.4 K 730.4 E 731.4 M 733.4 H 739.4 F 749.4 R 758.4 Y 765.4 W 788.4

PAGE 60

60 Table 2-4. m/z of VAXLR, where X is any of the listed amino acids VAXLR m/z of +1 charge-states G 515.3 A 529.3 S 545.3 P 555.4 V 557.4 T 559.4 L 571.4 N 572.4 D 573.3 K 586.4 E 587.4 M 589.3 H 595.4 F 605.4 R 614.4 Y 621.4 W 644.5

PAGE 61

61 m/zvsCV for 254 tyrpticpeptides0 200 400 600 800 1000 1200 -40-35-30-25-20-15-10-50 CV (V)m/z VASLR AWSVAR DLLFK DLLER m/zvsCV for 254 tyrpticpeptides0 200 400 600 800 1000 1200 -40-35-30-25-20-15-10-50 CV (V)m/z m/zvsCV for 254 tyrpticpeptides0 200 400 600 800 1000 1200 -40-35-30-25-20-15-10-50 CV (V)m/z VASLR AWSVAR DLLFK DLLER Figure 2-2. Plot of m/z versus CV for 254 trypti c peptides from data adapted from Guevremont et al.54 The four synthetic peptides used to ev aluate the effect of amino acid sequence on CV were selected from this group. Tw o synthetic tryptic peptides were based on outliers (AWSVAR and VASLR) and two we re based on tryptic peptides with average CV (DLLER and DLLFK).

PAGE 62

62 Table 2-5. Instrumental parameters utili zed for charge-state separation experiments Instrumental Parameter Setting Sheath Gas (arbitrary units) 20 Auxillary Gas (arbitrary units) 0 Spray voltage 4.5 kV Heated Capillary Temperature 140C Capillary Voltage 22.5 V Tube Lens Offset (arbitrary units) 25

PAGE 63

63 Signal Intensity vs DV-5000 -4000 -3000 -2000 0123456 Signal Intensity (counts x10,000)DV (V) Figure 2-3. Ion inte nsities of DLLPR+ at varying DVs. Optimal DV voltage for tryptic peptide ions is -5000 V.

PAGE 64

64 -10 -20 -30 CV (V) 0 50 100Relative Abundance DV = -5000 VRelative Abundance -10 -20 -30 CV (V) 0 50 100 DV = -4500 V -10 -20 -30 CV (V) 0 50 100Relative Abundance DV = -4000 V -10 -20 -30 CV (V) 0 50 100Relative Abundance DV = -3500 V -10 -20 -30 CV (V) 0 50 100Relative Abundance DV = -5000 V -10 -20 -30 CV (V) 0 50 100Relative Abundance DV = -5000 VRelative Abundance -10 -20 -30 CV (V) 0 50 100 DV = -4500 V Relative Abundance -10 -20 -30 CV (V) 0 50 100 DV = -4500 V -10 -20 -30 CV (V) 0 50 100Relative Abundance -10 -20 -30 CV (V) 0 50 100Relative Abundance DV = -4000 V -10 -20 -30 CV (V) 0 50 100Relative Abundance DV = -3500 V -10 -20 -30 CV (V) 0 50 100Relative Abundance DV = -3500 V Figure 2-4. Optimal charge-state separation of EFSGDK at varying DV. All intensities are scaled to 7.45 x 104. The left red trace is EFSGDK+1 and the black trace is EFSGDK+2.

PAGE 65

65 Effect of Carrier Gas Flowrate on Signal Intensity4.0 5.0 6.0 7.0 8.0 22.533.544.5Carrier Gas Flowrate (L/min)Signal Intensity x 10,000 Figure 2-5. The effect of carrier gas flow rate on signal intensity. DLLPR+1 showed optimum signal at 3.5 L/min. 4.0 L/min was within the error bars as well. Error bars correspond to 1 standard deviati on of the mean for 3 replicates.

PAGE 66

66 Figure 2-6. Variations in the carrier gas flow rate. Intensity

PAGE 67

67 10 11 12 6.92 x 1055.06 x 1032.35 x 104 0 -10 -20 a b c 100% N21.31 x 1066.92 x 1032.40 x 104 4 6 0 -10 -20 a b c 10% He / 90% N2 4 5 6 0 -10 -20 1.35 x 1065.84 x 1032.33 x 104 a b c 20% He / 80% N2 10 11 12 6.92 x 1055.06 x 1032.35 x 104 0 -10 -20 a b c 100% N21.31 x 1066.92 x 1032.40 x 104 4 6 0 -10 -20 a b c 10% He / 90% N2 4 5 6 0 -10 -20 1.35 x 1065.84 x 1032.33 x 104 a b c a b c 20% He / 80% N2 1.15 x 1044.99 x 104 1.46 x 106 0 -10 -20a b c 30% He / 70% N2 1.50 x 1061.86 x 1044.38 x 104 0 -10 -20 -30 a b c 40% He / 60% N2 4 5 6 7 1.49 x 1062.55 x 1043.08 x 104 0 -10 -20 -30 a b c 50% He / 50% N21.15 x 1044.99 x 104 1.46 x 106 0 -10 -20a b c 30% He / 70% N21.15 x 1044.99 x 104 1.46 x 106 0 -10 -20a b c a b c 30% He / 70% N2 1.50 x 1061.86 x 1044.38 x 104 0 -10 -20 -30 1.50 x 1061.86 x 1044.38 x 104 0 -10 -20 -30 a b c a b c 40% He / 60% N2 4 5 6 7 1.49 x 1062.55 x 1043.08 x 104 0 -10 -20 -30 a b c a b c 50% He / 50% N2 CV (V) Figure 2-7. The optimal carrier gas composition for the charge-state separation of tryptic peptides was 50/50 v/v He/N2. a.) TIC b.) DLLPR+ c.) DLLPR2+ Intensity Intensity

PAGE 68

68 Table 2-6. Optimized FAIMS conditions for tryptic peptide ions FAIMS Parameter Optimized Setting DV -5000 V (-4000 V on th e Ionalytics FAIMS) Outer bias voltage 0 V (when using the capillary extender) Carrier gas flow rate 3.5-4.0 L/min Percent He in N2 carrier gas 50%

PAGE 69

69 Figure 2-8. Charge-state separa tion of DLLXR. a.) ESI-MS without FAIMS b.) CV scan from -5.3 to -17.1 V. Only +1 charge-states of DLLXR are visible. c.) CV scan from -18.9 to -23.4 V. Only +2 charge-states of DLLXR are present. +1 char g e-state +2 char g e-state TIC a b c

PAGE 70

70 Table 2-7. Charge-state separati on of DLLXR is achieved by FAIMS. CV (V) for DLLXR+ CV (V) for DLLXR2+ DLLGR -16.4 0.3 -20.7 0.0 DLLAR -13.0 0.0 -20.3 0.1 DLLSR -15.1 0.2 -20.3 0.1 DLLPR -11.9 0.3 -20.1 0.1 DLLVR -12.4 0.1 -20.4 0.1 DLLTR -14.1 0.1 -20.2 0.1 DLLLR -12.2 0.2 -20.4 0.1 DLLNR -12.5 0.1 -20.3 0.0 DLLDR -12.4 0.0 -20.2 0.1 DLLKR -12.1 0.1 -20.3 0.2 DLLER -12.1 0.1 -20.5 0.1 DLLMR -12.0 0.2 -20.6 0.1 DLLHR -11.6 0.5 -19.8 0.1 DLLFR -11.4 0.3 -20.8 0.1 DLLRR N/A -19.4 0.1 DLLYR -10.9 0.1 -20.7 0.5 DLLWR -11.2 0.2 -20.9 0.1

PAGE 71

71 Figure 2-9. Charge-state separation of DLLPR+ and DLLPR2+. Resolution of 1.2 was achieved. a 0.0 -10 -20 -30 -40 -10 -20 -30 -40 -10 -20 -30 -40 -10 -20 -30CV (V) 2 .41 x 10 74.22x 10 55.44 x 10 5TIC DLLPR+DLLPR2+ b c

PAGE 72

72 Charge State Separation of DLLXR0 200 400 600 800 -30 -20 -10 0 CV m/z Figure 2-10. Charge-state separation is achie ved for DLLXR. The +1 charge-states (blue diamonds) are around a CV of -10 V, whereas the +2 charge-states (pink squares) are around a CV of -20 V.

PAGE 73

73 0.0 -10 -20 -30 -40 -10 -20 -30 -40 -10 -20 -30 -40 -10 -20 -30 CV (V) 2.41x 1076.51x 1056.40x 1053.80x 1054.22x 1054.03x 1053.81x 1054.24x 105TIC DLLGR+DLLAR+DLLSR+DLLPR+DLLVR+DLLTR+DLLLR+ Figure 2-11. CV required by se lect charge-sta tes of DLLXR+.

PAGE 74

74 -20 -40 -20 -40 -20 -40 -20 CV (V) Intensity 2.41x 1074.40x 1056.58x 1056.93x 1056.46x 1058.82x 1057.16x 1051.18x 106 TIC DLLGR2+DLLAR2+DLLSR2+DLLPR2+DLLVR2+DLLTR2+DLLLR2+ -20 -40 -20 -40 -20 -40 -20 CV (V) Intensity 2.41x 1074.40x 1056.58x 1056.93x 1056.46x 1058.82x 1057.16x 1051.18x 106 -20 -40 -20 -40 -20 -40 -20 CV (V) Intensity 2.41x 1074.40x 1056.58x 1056.93x 1056.46x 1058.82x 1057.16x 1051.18x 106 TIC DLLGR2+DLLAR2+DLLSR2+DLLPR2+DLLVR2+DLLTR2+DLLLR2+ Figure 2-12. CV required by sele ct charge-states of DLLXR2+.

PAGE 75

75 400 600 800 1000 1200 1400 m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 344.60 389.19 545.57 748.94 483.18 736.54 678.95 964.91 840.33 1169.28 1073.90 1290.27 1496.64 Intensity 1.67 x 107 MIFAGIK+ EDLIAYLK2+ MIFAGIK2+ EDLIAYLK+ TGPNLHGLFGR2+ 400 600 800 1000 1200 1400 m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 344.60 389.19 545.57 748.94 483.18 736.54 678.95 964.91 840.33 1169.28 1073.90 1290.27 1496.64 400 600 800 1000 1200 1400 m/z 0 10 20 30 40 50 60 70 80 90 100 400 600 800 1000 1200 1400 m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 344.60 389.19 545.57 748.94 483.18 736.54 678.95 964.91 840.33 1169.28 1073.90 1290.27 1496.64 Intensity 1.67 x 107 MIFAGIK+ EDLIAYLK2+ MIFAGIK2+ EDLIAYLK+ TGPNLHGLFGR2+ Figure 2-13. Cytochrome C digest ESI-MS spec trum. No FAIMS was used for separation.

PAGE 76

76 Figure 2-14. Charge-state separation of cy tochrome C by FAIMS. The CV was scanned twice from 0 to -40 V over 4 minutes.

PAGE 77

77 Figure 2-15. Improvement in charge-state separation of Cytochrome C by FAIMS. a.) The CV was scanned from -2.1 to -7.5 V. b.) The CV was scanned from -14.9 to -17.8 V.

PAGE 78

78 Table 2-8. The effect of varying amino acid sequence on CV (V). AWXVAR was taken at 50% He, while DLLXR, DLLXK and VAXLR were taken at 40% He due to instrument malfunctions with the vacuum. DLLXR DLLXK AWXVAR VAXLR CV (V) CV (V) CV (V) CV (V) G -10.5 0.1 -9.4 0.2 -10.8 0.2 -8.6 0.1 A -8.7 0.1 -8.7 0.2 -9.6 0.3 -7.4 0.1 S -10.2 0.1 -7.6 0.3 -9.4 0.2 -7.4 0.1 P -7.7 0.1 -8.3 0.2 -9.0 0.1 -8.2 0.1 V -8.4 0.1 -8.5 0.1 -9.1 0.1 -7.7 0.1 T -9.7 0.0 -7.8 0.1 -9.1 0.1 -7.4 0.1 L -8.2 0.1 -8.5 0.0 -9.1 0.2 -7.5 0.1 N -7.9 0.1 -8.1 0.1 -9.0 0.1 -7.4 0.1 D -7.6 0.1 -8.2 0.2 -10.0 0.3 -7.4 0.1 K -7.6 0.1 -7.4 0.5 -8.8 0.2 -7.9 0.1 E -7.7 0.1 -7.3 0.5 -8.6 0.1 -7.9 0.1 M -7.6 0.1 N/A -9.1 0.2 -8.1 0.1 H -9.0 0.2 -7.2 0.0 -9.0 0.3 -8.1 0.6 F -7.5 0.1 -8.3 0.1 -9.5 0.1 -7.8 0.1 R N/A -7.5 0.5 -10.4 0.4 N/A Y -7.7 0.1 -7.9 0.2 -9.3 0.1 -7.8 .1 W -7.4 0.1 N/A -8.7 0.2 -7.8 .1

PAGE 79

79 CV for glycine contain peptides vs peptides with no glycine residues 300 400 500 600 700 800 -13-11-9-7-5 CV (V)m/z 5 aa; G 5aa; no G 6 aa; G 6 aa; no G Figure 2-16. Tryptic peptides with 5 and 6 amino acid (aa) residues.

PAGE 80

80 CHAPTER 3 SEPARATION OF PHOSPHOPEPTIDE S BY HIGH-FIEL D ASYMMETRIC WAVEFORM ION MOBILI TY SPECTROMETRY Overview This chapter addresses strategies for phosphopeptide separation by FAIMS. The benefits and limitations of current technology in phosphopep tide monitoring will be addressed. FAIMS parameters (carrier gas composition and elec trode temperature) will be optimized for phosphopeptide separation. Phosphopeptide separa tion by FAIMS is evaluated on a mixture of synthetic phosphopeptides and their nonphosphorylated analogues. Importance of Phosphopeptides The process of transla ting mRNA into proteins is called translation.2 After translation, proteins require further modifica tion, such as folding or amino acid removal, before becoming biologically active. Posttransl ational modifications (PTMs) ar e the addition or removal of chemical functional groups to the pr otein after transla tion has occurred.2 Protein phosphorylation, the most common PTM, plays a ke y role in the modulation of cellular activity such as signal transduction, cell cycle, apoptosis, metabolism, gene expression15,16, protein synthesis, and cellular morphology.17,90 In fact, most cellular pr ocesses are regulated by reversible phosphorylation of the amino acids serine, theronine and tyrosine.2,6 Phosphorylation has also been shown to rarely occur on histidine91, aspartate75,91, and glutamate75 amino acid residues. It is estimated that at least 30% of proteins undergo phosphorylation, and there are 3,652 known phosphorylated sites on only 1,240 human proteins alone.14 Phosphorylation is a highly dynamic pro cess, modulated by both phosphatases and kinases.92-94 Protein kinases, a type of phosphotransferase enzyme, catalyze the addition of phosphate groups to proteins, whereas protein phosphatases remove phosphate groups from

PAGE 81

81 proteins.95 The addition or removal of a phosphate gr oup to/from a protein can regulate its function by altering enzymatic ac tivity, affecting the formation of complexes, and modulating protein-protein interactions.6 Phosphorylation typically occurs on loops or turns of the protein near the surface where the amino acids are mo re accessible to kinases and phosphatases for modification.5 Alterations in a proteins phosphoryl ated state can lead to disease.6,16 For instance, heart disease, the leading cause of death in the world, can be caused by alterations in the phosphorylated states of cardiac proteins.96 Therefore, the ability to locate sites of phosphorylation and understand th eir biological regul atory mechanisms is critical in understanding diseases. However, the analysis of PTMs has challenged researchers to develop more sensitive, selective and efficient methods. Analytical Methods to Char acterize Phosphopeptides Numerous methods to enhance the analysis of phosphopeptides have been established, including Edman sequencing3,16,97, isotopic labeling with 32P or 33P 15, antibodies98, and fluorescent tagging99. However, these methods may lack selectivity, are time-consuming, and can be expensive. Chromatography Many methods for characterizing phosphope ptides focus on providing phosphopeptide enrichment prior to analysis. HPLC is often ut ilized for sample clean-up prior to MS analysis, but efficiently retaining hydrophilic phosphopeptides on re verse-phase columns is challenging.3,95 The most selective phosphopeptide enrichment columns include TiO2, ZrO2, and immobilized metal affinity chromatography (IMAC). In IMAC, phosphopeptides bind to chelated Fe3+ along the stationary phase; however, th is process is strongly pH-dependent.6 Furthermore, interferences can arise from salts buffers, and strongly aci dic peptides containing

PAGE 82

82 aspartate or glutamate residues that also ha ve an affinity for the stationary phase.6 Therefore, the sample composition prior to IMAC column lo ading is critical. New IMAC technology implements nitrilotriacetic chelat ing resin instead of the traditi onal iminodiacetic chelating resin to reduce interferences from nonphosphorylated peptides.11 However, IMAC still has the disadvantages of limited selectiv ity, long analysis times and a dependence on loading procedure.3 TiO2 and ZrO2 based solid phase columns are also used for phosphopeptide analysis and are more amenable to automation than IMAC.3 These new column materials consist of porous titanium oxide or zirconium oxide spherical pa rticles that have amphoteric ion-exchange properties, although the intera ction between phosphopeptides a nd the metal oxide spherical particles is not well understood.3 However, like IMAC, TiO2 and ZrO2 columns still suffer from interferences from acidic peptide affinity. Es terification of the peptide sample mixture can reduce the interferences from nonphosphorylated peptides; however, this increases sample preparation time.3 Mass Spectrometry Mass spectrometry has emerged as an alte rnative technique for the analysis of phosphopeptides, offering high sensitivity and the ability to discover, identify, and quantitate PTMs in complex mixtures. However, ev en MS has faced difficulties analyzing phosphopeptides. The common ionization source for an alyzing peptides and proteins, ESI, has difficulties ionizing phosphopeptides due to their hydrophilic nature.17,27 Negative ion mode works well, but the subsequent MS/MS analys is does not give much structural data.100 Triple quadrupole mass spectrometers can use varying s can functions to identify phosphopeptides. The two scan modes used commonly for phosphopeptide detection are parent and neutral loss scans.6 In a neutral loss scan, the quadrupol es are scanned at a constant offset. Thus, the only ions that are detected are those that lose a specific mass (neutral fragment) during collision-induced

PAGE 83

83 dissociation in q2, as shown in Figure 3-1. For example, in positive ion mode, a neutral loss scan of 80 (loss of HPO3) or 98 (loss of phosphoric acid, H3PO4) is can selectively identify phosphopeptides. Q1 and Q3 are scanned at an offset of m/z 80 such that only ions with a loss of 80 will be detected. Unfortunately, neutral loss scans do not yield enough signal in most cases to be useful.6 Parent ion scans are the most common scan f unction used to detect phosphopeptides using a triple quadrupole mass spectrometer. In a parent ion scan, Q1 completes a full scan, while Q3 is set to pass only ions with a specified m/z as shown in Figure 3-2. In a negative ion mode parent scan, Q3 can be set to pass only m/z 79 (PO3 -), m/z 31.67 (PO4 3-) or m/z 97 (H2PO4 -) for select phosphopeptide ion detection. In a positive ion mode parent scan, Q3 can be set to pass m/z 81 (H2PO3 +). However, in practice, many fragment ions are not produced in sufficient quantities for adequate detecti on. Mass-to-charge 97 corresponds to a b1 ion of Pro and is therefore not useful. The most us eful parent ion scan with adequa te signal and high selectivity is for m/z 79 (PO3 -) in negative ion mode.11,27 However, the solvents required to produce ions in negative ion mode electrospray are not very comp atible with reverse-phase chromatography and sample clean-up is usually done offline. In Figure 3-3, fragment ions for neutral loss and parent scanning for phosphopeptides are shown. Limitations in Phosphopeptide Analysis by Mass Spectrometry The advantages of mass spectrometry have been well reported in the characterization of low abundance biomolecules; however, this technique is not without limitation when applied to phosphopeptides. This is because th e fraction of a protein in its phosphorylated state is often less than 10% in vivo.6 The sensitivity of electrospray i onization is concentr ation-dependent and does not favor low-concentration, hydrophilic phosphopeptides. Therefore, phosphopeptides exhibit poor ionization efficiency in positive i on mode electrospray and are difficult to detect.

PAGE 84

84 In addition to the challenges associated with ion suppression, a sec ond challenge exists. The covalent bond between the phosphate and serine or threonine is labile as well, further challenging method design.95 During collision-induced dissociation in MS/MS, the phosphate group may be eliminated, making it difficult to elucid ate its original locati on. If the product ion does not retain the phosphate, it is difficult to determine where the phosphate was located in the amino acid sequence. Application of FAIMS to Phosphopeptide Analysis Given the lack of rapid and selective separation methodology for phosphopeptides, the separation of phosphopeptides by FAIMS for mass sp ectrometry is of interest. FAIMS has previously demonstrated the ability to separate isomers and improve signal-to-noise of analytes in complex mixtures.27,50,59 Here we test the hypothesis that FAIMS can selectively separate phosphopeptides from nonphosphorylated peptides. The transmission of an ion through FAIMS can be affected by factors that include the shape of the waveform, gas pressure, cell geom etry, gas composition, dispersion voltage, and temperature.19 In these experiments, both heated el ectrodes and carrier gas compositions (SF6, CO2, He, N2) were used to evaluate th eir effect on phosphopeptide separation. Combinations of SF6 and He have shown non-Blanc behavior and have improved re solution and peak capacity.57 CO2 was selected as a carrier gas for anal ysis of phosphopeptides because it has shown remarkable abilities to separate isomers of phthalic acids, as described in Chapter 1.59 The addition of CO2 also allowed for the separation of the isobaric bisulfate and perchlorate anions.45 Heated FAIMS electrodes have been shown to increase peak capacity.63

PAGE 85

85 Experimental Design Sample Preparation Phosphorylated peptides and their non-phosphor ylated analogues were obtained from the Interdisciplinary Center for Biot echnology Research at the Univer sity of Florida (Gainesville, FL). The sequences of the synthesized tryptic peptides were EFXGDK, wh ere X could be S, T, Y, pS, pT, or pY. Their structures and expected m/z values are shown in Figure 3-4. The tryptic peptides were made into a 500 ppm stock solu tion in 49.5/49.5/1.0 water/methanol/acetic acid and subsequently diluted to 10 ppm in 49.5/49.5/ 1.0 water/methanol/acetic acid for analysis. Although a variety of acids were evaluated (acetic acid, formic acid, phosphoric acid, TFA), acetic acid was chosen because it offered the best signal-to-noise for the tryptic peptides. Instrumentation The instrumentation used in these experiments was a Thermo LCQ ion trap mass spectrometer (Thermo-Scientific, San Jose, CA) equipped with a Thermo FAIMS. Mass spectrometer and Ionmax positive ion mode electrospray parameters for optimal signal are shown in Table 3-1. The carrier gases were dried using a carbon/mois ture trap (Activated carbon/molecular sieve, Tr igon Technologies). N2 and He flow rates were controlled by the FAIMS software. CO2 and SF6 flow rates were monitored using a rotameter. All CVs were scanned at a rate of 10 V/min. Instrument Modification The Thermo FAIMS was not designed to couple to the LCQ, so a brass capillary extender was designed as an interface, as shown in Figure 3-5. The capillary extender was placed behind the FAIMS outer electrode and connected to the heated capillary. Ion losses were near 70%; however, this design allowed for the coupling of FAIMS to the LCQ, which otherwise would not have been possible. Ion losse s were determined by comparing ESI-MS with and without the

PAGE 86

86 capillary extender in place. High concentrations (10 ppm) were used to offset ion losses from the use of the brass capillary extender. FAIMS is sold commercially coupled to a Thermo TSQ Quantum triple quadrupole mass spectrometer and the commercial interface does not suffer from huge losses. Optimization of FAIMS Parameters for the Detection of Phosphopeptides Parameters affecting transmission of an ion through FAIMS incl ude the shape of the waveform, gas pressure, cell geometry, dispersi on voltage, gas composition, and temperature.19 The latter two parameters were evaluated fo r their effect on phosphopeptide separation. A carrier gas composition of 50/50 He/N2 has been shown previously to provide improved separation for tryptic peptides.89 Alternative carrier gas compositions were explored for their possible improvement of phosphopeptide separation. The temperature of the inner and outer electrodes was varied independently from 30 to 90 C. The carrier gas composition was altered using varying amounts of N2, He, SF6, and CO2, as shown in Table 3-2. Results The following sections evaluate phosphopep tide separation by FAIMS as a result of varying the composition of the carrier gas (mixtures of N2, He, SF6, and CO2) and varying the temperature (30 C to 90 C) of the i nner and outer electrodes independently. The Effect of Carrier Gas Composition on Phosphopeptide Separation All gases were dried by a hydrocarbon/moisture trap. The He and N2 gas flows were controlled with flow controllers in the co mmercial Thermo FAIMS design, while the CO2 and SF6 gas flows were controlled using a calibrated rotameter. N2 and 50/50 N2/He are common carrier gases used in FAIMS.88,89 For proteins and peptides, 50/50 He/N2 is typically utilized and provides increased resolution an d signal. In Figure 3-6, CV scans for a mixture of EFXGDK with both N2 and 50/50 N2/He carrier gases are shown. The second peak in the right side of

PAGE 87

87 Figure 3-6b was determined to be the doubly ch arged dimer. In agreement with current literature, tryptic peptides show impr oved resolution and signal in 50/50 N2/He when compared to 100% N2 carrier gas.54 With the addition of helium, the magnitude of the required CV increases. As shown in Figure 3-7, the CV increases from around -3 to -5 V in 100% N2 carrier gas to -5 to -8 V in 50/50 He/N2 carrier gas. It is believed that peptide ions shift to a more negative CV when He is added to the carrier gas composition because He is both smaller and less polarizable, minimizing interactions with the analyte ions.57,101 Signal-to-noise also increas es, as shown in Figure 3-7. Alternative carrier gas compositions, shown in Table 3-2, were explored to analyze their effect on phosphopeptide separation. CO2/He Carrier Gas Compositions As described in Chapter 1, Shvartsburg suggest ed that the large vari ation in collisional cross section between CO2 and He would cause deviations in the predicted CV values calculated from Blancs law. Despite Shvartsburg s prediction of non-Blanc behavior, He/CO2 carrier gas compositions did not offer enhanced separation of phosphopeptides. Figure 3-8a shows CV scans with a carrier gas composition of 10% He and 90% CO2, where the CV scans for each individual tryptic peptide peak ar e wide and not separated from ot her CV peaks in the mixture. The baseline noise is large in the mass spectrum, shown in Figure 3-9a, a nd intensity is only 1 x 104, which is four times lower than the 50/50 He/N2 carrier gas composition. The second He/CO2 carrier gas mixture evaluated was 50% He and 50% CO2. This mixture offered minimally better signal than the 10/90 He/CO2 combination, but there was still significant overlap of peaks and resolution was p oor, as shown in Figure 3-8b. Peak shape was also wide and distorted for most of the tryptic peptides. The MS spectrum shown in Figure 3-9b has increased signal-to-noise and increased peak intensity when compared to Figure 3-9a.

PAGE 88

88 However, when compared to the MS spectrum obtained from 50/50 He/N2 carrier gas, shown in Figure 3-6, the phosphopeptide signal-to-noise and peak intensities were lower. The carrier gas mixture of He/CO2 could not resolve phosphopeptides from their nonphosphorylated analogues. No other combinations of He/CO2 were attempted due to the poor peak shape of the first two experiments. The 50/50 He/CO2 combination offered an increase in signal-to-noise versus the 10/90 He/CO2 mixture. However, the 50/50 N2/He combination still offers the best signal-to-noise and peak shape. Overall, the He/CO2 offered no advantage for the resolution of phosphopeptides and th eir nonphosphorylated analogues. CO2/He/N2 Carrier Gas Compositions Since He/N2 carrier gas mixtures offer increas ed resolution, the addition of CO2 to this combination was explored Previously, a carrier gas composition of CO2/He/N2 was shown to improve resolution of cisplatin.102 Three gas combinations (10/45/45 CO2/He/N2, 33.3/33.3/33.3 CO2/He/N2, 50/25/25 CO2/He/N2) were evaluated, as well as 100% CO2 as carrier gases for the separation of phosphopeptides. No improvement in resolution or signal-to-noise over a 50/50 He/N2 carrier gas was observed, as shown in Fi gures 3-10 and 3-11. As the amount of CO2 was increased in combination with He and N2, the CV peaks shifted to a more positive value. This is likely due to the fact that the percentage of He in the carrier gas is decreasing, replaced by a larger, more polarizable CO2, which may interact with the peptides more than would He. Interactions with CO2 may cause the peptides to be less fl exible and alter their mobility at highfield less than would He. The 10% and 33% CO2 in He/N2 carrier gas demonstrated improved resolving power when compared to 50% CO2 in He/N2 or a 100% CO2 carrier gas; however, the 10% and 33% CO2 addition to the He/N2 carrier gas did not improve resolution when compared to the traditional 50/50 He/N2 carrier gas. The best signal-to-noise was observed for the carrier gas 50/25/25

PAGE 89

89 CO2/He/N2, as shown in Figure 3-11. However, the CV peaks for 50/25/25 CO2/He/N2 carrier gas were broad and distorte d, as shown in Figure 3-10. The peaks in 100% CO2 carrier gas had the worst peak shape and intensity. Overall, the addition of CO2 to the He/N2 carrier gas did not improve phosphopeptide resolution. CO2/N2 Carrier Gas Compositions The addition of CO2 to a nitrogen carrier gas has been shown to improve the separation of phthlates, as discussed in Chapter 1.59 However, the addition of CO2 to the N2 carrier gas did not improve phosphopeptide separation. Overall, th e peak shape was distorted and broad and no separation was achieved, as s hown in Figure 3-12. 10% CO2 in the nitrogen carrier gas improved peak shape compared to 100% N2, but not when compared to 50/50 He/N2. Peak intensity was also less in 10% CO2 90% N2 than in 50% He 50% N2. These experiments determined that adding CO2 to a He or He/N2 carrier gas did not improve phosphopeptide separati on or signal-to-noise. CO2 may have improved phthlate separation because phthlates are small type A ions It has been suggested that at low fields, phthalic acids complex with CO2, but high fields break apart the complex, altering the mobility of the ions though the FAIMS electrodes.59 However, larger type C ions, such as peptides, may have experienced poor resolution with the addition of CO2 to the carrier gas because CO2 is a larger and more polarizable molecule, which may cau se distortions in the pe ptide structure or the formation of adducts that alter their mobility at high fields. He/SF6 Carrier Gas Compositions While others have found a carrier gas composition of 50/50 He/N2 to be the best for tryptic peptide separation, data published by Shva rtsburg et al. demonstrated the effect of a SF6/He carrier gas composition on ce sium ion behavior in FAIMS.27,57 Cs+ was shifted to a more negative CV value when SF6 was included in the carrier gas.57 The benefit of including SF6 in

PAGE 90

90 the FAIMS carrier gas for phosphopeptide separation was explored by varying SF6 from 50 to 100% in He. Helium percentages coul d not exceed 50% due to arcing. Both 10% and 25% He in SF6 gave poor CV spectra, as s hown in Figure 3-13. Peaks were broad and distorted. The MS spectra, shown in Figure 3-14, had poor signal-to-noise and not all of the peptides are dete ctable when using the 10% 90% SF6 carrier gas. However, the resolution of the +1 charge-sta te phosphopeptides was improved by using a carrier gas of 50/50 He/SF6 instead of 50/50 He/N2, as shown in Figure 3-15. A more negative CV was required when using a He/SF6 carrier gas compared to a He/N2 carrier gas. Despite the improvement in resolution, due to the decrease in ion signal it was determined that SF6 was not a suitable carrier gas for phosphopeptides. The +2 charge-state was not observed, while the +1 charge-state in SF6/He carrier gas had poor signal-to-noise when compared to 50/50 He/N2 carrier gas, as shown in Figure 3-16. The signal of the singl y charged peptides dropped from 4.96 x 104 in 50/50 He/N2 to 5.06 x 103 when using a carrier gas of 50/50 He/SF6 and the signal-to-noise increased. Signal reduction by the addition of SF6 to the FAIMS carrier gas could be detrimental to the elucidation of low-abundance phosphopeptides. SF6 is also significantly more expensive than N2 or He and may be cost prohibitive for routine samples analysis. These studies determined that 50/50 N2/He remains the best carrier gas composition for the separation and detection of phosphopeptides. CO2 offered no advantages, while SF6 improved resolution but significantly decreased ion intensities. Even though published data have shown interesting trends in FAIMS with the addition of CO2 and SF6 to the carrier gas, their analytes were small type A ions.57 Phosphopeptides may not have benefited from the addition of SF6 and CO2 because they are type C ions. Small type A ions can form complexes with the carrier gas that deteriorate at high fields and alter their mobility, but type C ions are much larger

PAGE 91

91 and complexes may not alter their mobility as mu ch. In the case where sample concentration was irrelevant, 50/50 SF6/He may be useful for improved separation. The addition of SF6 to He may have shifted CVs to a more negative value and improved resolution because SF6 is a large molecule and would collide more frequently with the peptide ions, decrea sing their mobility at high fields. However, these collisions may have also prevented the ions from being focused as effectively, thereby reducing the signal. The Effect of Heated Electrode s on Phosphopeptide Resolution The benefit of heated FAIMS elect rodes has been reported previously.63 As described by Barnett et al., the temperature of the electrodes affects the number density, and thus the effective field that influences the ion.63 Referring to Equation 3-1, Kh is dependent on E/N. N can be calculated from Equation 3-2, where n/V is from the ideal gas law, n is the number of moles of the carrier gas, V is the volume (cm3) and NA is Avogadros number. Th e ideal gas law is shown in Equation 3-3, where P is the pressure (atm), T is temperature (K), and R is a constant (0.08206 L atm mol-1 K-1). Thus, it can be shown that as te mperature increases, the number density decreases, and E/N increases. Kh(E/N) = K [1 + f(E/N)] (3-1) N = (n/V) NA (3-2) n/V = P/(RT) (3-3) Therefore, as temperature increases, the eff ective field (E/N) that influences an ion increases. Because the asymmetric waveform is applied to the inner electrode only, the field between the concentric cylindrical electrodes is non-uniform. Ions near the inner electrode are subjected to stronger fields than those near the outer electrod e. Improved ion focusing is

PAGE 92

92 achieved by heating the outer electr ode to higher temperatures than the inner electrode. A higher temperature at the outer electrode decreases th e number density near the outer electrode and focuses ions toward the center of the electrodes. If the inner electrode is hotter than the outer electrode, then the peaks become broad and unresolved. This is because ions are unfocused due to the temperature gradient. A hotter inner elec trode increases E/N and repels ions towards the outer electrode. If the electrode s are set to the same temperature, resolution decreases due to the lack of the temperature gradient. Figure 3-17 de monstrates the effect of temperature gradients on peak behavior. The 40C /70C (inner and outer electrode temperature, respectively) electrodes show the best peak resolution and resolving power for this compound. As the electrode temperatures are chan ged to 70C /70C, the resolution decreases. Once an electrode temperature of 90C/ 70C is achieved, resolutio n and resolving power decrease. Resolution was calculated from Equation 3-4, where the change in CV is multiplied by two and divided by the sum of the peak base width (Wb1 + Wb2). Resolving power is calculated from Equation 3-5, where the CV is divided by the peak width at full width half max. R = 212bbWW CV (3-4) Resolving Power = CV/ W1/2max (3-5) In these experiments, the inner and outer elec trode temperatures were varied independently from 30C to 90C to determine the best temper atures for phosphopeptide separation. The best resolution was achieved when the inner electrode was 50C and the outer electrode was 90C, as shown in Figure 3-18. As the electrode temper atures and gradient increase, the resolution

PAGE 93

93 between phosphopeptides and nonphosphorylated peptid es increases in the +2 charge-state, as shown in Figure 3-19. At higher temperature gradients, one explan ation has the +2 charge-state nonphosphorylated peptides requiring a more negative CV than th e +2 charge-state phosphorylated compounds due to their less rigid na ture. Phosphopeptides in the +2 charge-state are likely more compact and rigid due to interac tions between the charges (located on either end of the peptide) and the phosphate group, which limits their flexibility to change conformation and interact with the carrier gas.36 On the other hand, +2 nonphosphorylated tryptic peptides experience repulsion of the charges and are less rigid; thus, they interact with the carrier gas and change conformation more easily, lowering mobility at higher fields. The ion intensity of the +2 charge-state peptides decreased slightly at hi gher electrode temperatures ; however, the resolution between phosphorylated and nonphosphorylated peptides doubled from 0.8 to 1.6. Unlike the +2 charge-state, the +1 char ge-state showed mini mal improvement in resolution due to the increased temperature gradie nt, as shown in Figure 3-20. The +1 chargestate phosphopeptides required a more negative CV than +1 charge-state nonphosphorylated peptides. Phosphopeptides in the +1 charge-s tate may require a more negative CV than nonphosphorylated tryptic peptides because with only one charge, the phosphopeptides may not be as compact or rigid, but may be more flexible than a non-phosphor ylated tryptic peptide, causing more interactions with the carrier gas and a greater ability to change conformation. The second peak in the CV scans of Figure 320b-g at electrode temperatures of 30C was determined to be the doubly-charged dimer. As the concentration of the peptides was increased, the dimer peak increased while the singly charged peak decreased. Similar behavior to the +2 charge-state phosphopeptides is observed in the phosphorylated doubly-charged dimer. The phosphorylated dimers required a more negativ e CV than the nonphosphorylated dimers. The

PAGE 94

94 phosphate group may interact with the two char ges, making the phosphorylated doubly-charged dimer more rigid and less likely to change c onformations than the nonphosphorylated dimer. Therefore, the phosphorylated peptide ions would not interact with the carrier gas or change conformations as much as the nonphosphorylated pe ptide ions. When the temperature gradient was increased, as shown in Figure 3-20, the di mer signal decreased. The higher temperatures may have reduced the formation of dimers. Ion in tensity increased slightly at higher temperature gradients for the +1 charge-state. Increased +1 charge-state signal may have resulted from less dimer formation. At higher temperatures, more of the ions may have remained singly charged instead of dimerizing, which would e nhance the +1 charge-state signal. Conclusion The ability of FAIMS to separate phosphopeptides from their nonphosphorylated analogues in the +2 charge-sta te has been demonstrated. This is the first report of phosphopeptide separation by FAIMS. It was de termined that heated electrodes improved phosphopeptide separation by FAIMS without sign ificantly decreasing signal. The best temperature for resolution was 50 C for the inner electrode and 90 C for the outer electrode. Adding SF6 to the carrier gas was shown to improve resolution; however, due to its expense and the significant decrease in phosphope ptide signal, it was determined not to be a suitable carrier gas for phosphopeptide separation. Adding CO2 to the carrier gas offered no benefit in phosphopeptide separation.

PAGE 95

95 Q1 q2 Q3 Q1 q2 Q3 Figure 3-1. Neutral loss scan. Q1 and Q3 are scanned at a specifi ed offset in order to detect neutral losses. Q1 q2 Q3 Q1 q2 Q3 Figure 3-2. Parent ion scan. Q1 scans the m/z, ions are fragmented in q2, and a selected ion is monitored in Q3. Figure 3-3. Neutrals and ions used in neutral loss and pare nt ion scanning modes to selectively detect phosphopeptides. On the left, are ne utrals and ions that would be expected from an acidic solution. On the right, are neutrals and ions that would be expected from a basic solution.6

PAGE 96

96 EFYGDK +1 m/z 758.3 +2 m/z 379.7 EFpYGDK +1 m/z 838.3 +2 m/z 419.7 N O O O H N O N O O N O N O O O N O N+ OH H H H H H H H H H H P O O O H H H HEFpSGDK +1 m/z 762.3 +2 m/z 381.6 N O O O H N O N O O N O N O O O N O N+ OH H H H H H H H H H H H H N O O O N O N O O N O N O O O N O N+ OH H H H H H H H H H H H P OH OH OH H N O O O N O N O O N O N O O O N O N+ OH H H H H H H H H H H H H H EFSGDK +1 m/z 682.3 +2 m/z 341.7 N O O O N O N O O N O N O O O N O N+ OH H H H H H H H H H H H P OH OH OH H EFpTGDK +1 m/z 776.3 +2 m/z 388.6 EFTGDK +1 m/z 696.3 +2 m/z 348.7 N O O O H N O N O O N O N O O O N O N+ OH H H H H H H H H H H H H Figure 3-4. Structures of EFXGDK, wher e X = S, T, Y, pS, pT, or pY. The m/z values of the [M+H]+ and [M+H]2+ ions are listed.

PAGE 97

97 Brass Capillary Extender FAIMS Cell Inner Electrode Outer Electrode Heated Capillary (Entrance to the Mass Spectrometer) Curtain Plate Brass Capillary Extender FAIMS Cell Inner Electrode Outer Electrode Heated Capillary (Entrance to the Mass Spectrometer) Curtain Plate Figure 3-5. Interface between the FAIMS device and the LCQ mass spectrometer. Length 83.9 mm, outer diameter 15.0 mm, inner diamet er 1.3 mm. Illustration by Alex Wu, 2008.

PAGE 98

98 Table 3-1. Electrospray and mass spectromete r parameters used for these experiments Instrument Parameter Setting Sheath gas 20 arbitrary units Auxillary gas 0 arbitrary units Spray voltage +5 kV Sample flow rate(direct infusion) 3 L/min Heated capillary temperature 140C Tube lens offset 25 arbitrary units

PAGE 99

99 Table 3-2. Variations of carrier gas compositions %N2 % He % SF6 %CO2 Pure Gases 100 0 0 0 0 0 100 0 0 0 0 100 N2/He 50 50 0 0 N2/He/CO2 45 45 0 10 33.3 33.3 0 33.3 25 25 0 50 N2/CO2 90 0 0 10 70 0 0 30 50 0 0 50 30 0 0 70 10 0 0 90 He/CO2 0 10 0 90 0 50 0 50 He/SF6 0 10 90 0 0 25 75 0 0 50 50 0

PAGE 100

100 Figure 3-6. CV scans of a mixture of EFXGDK (20 ppm) with two common FAIMS carrier gas compositions. a.) total ion trace b.) EFSGDK+ c.) EFpSGDK+ d.) EFTGDK+ e.) EFpTGDK+ f.) EFYGDK+ g.) EFpYGDK+ 100% N2 50% He 50% N2

PAGE 101

101 Intensity 8.7 x 1050 696.3 682.3 758.2 838.2 776.1 200 300 400 500 600 700 800 900 1000 m/z 0 4.5 x 105 682.2 696.3 758.2 780.3 838.2Intensity 8.7 x 1050 696.3 682.3 758.2 838.2 776.1 8.7 x 1050 696.3 682.3 758.2 838.2 776.1 200 300 400 500 600 700 800 900 1000 m/z 0 4.5 x 105 682.2 696.3 758.2 780.3 838.2 200 300 400 500 600 700 800 900 1000 m/z 0 4.5 x 105 682.2 696.3 758.2 780.3 838.2 Figure 3-7. Comparison of MS spectra fr om FAIMS-MS using di fferent carrier gas compositions. a.) The carrier gas composition was 50%N2/50% He. The CV was scanned from CV -4.1 V to -8.0 V. b.) Carrier gas composition was 100% N2. The CV was scanned from -2.3 V to -6.9 V. a b

PAGE 102

102 Figure 3-8. CV scans with va riation in percent He in CO2 carrier gas. a) 10/90 He/CO2 b) 50/50 He/CO2 i.) total ion trace ii.) EFSGDK+ iii.) EFpSGDK+ iv.) EFTGDK+ v.) EFpTGDK+ vi.) EFYGDK+ vii.) EFpYGDK+ 50/50 He/CO2 has minimally improved signal compared to 10/90 He/CO2. i i i i ii i v v vi vii a b i i i i ii i v v vi vii

PAGE 103

103 Figure 3-9. Comparison of MS spectra from FAIMS-MS using di fferent carrier gas compositions. a.) Carrier gas composition of 10% He and 90% CO2. CV was scanned from -0.9 V to -4.9 V. b.) Carrier gas composition of 50% He and 50% CO2. CV was scanned from -2.1 V to -5.9 V. a b Intensit y Intensit y

PAGE 104

104 Figure 3-10. Variation of %CO2 in the He/N2 carrier gas. i.) total ion trace ii.) EFSGDK+ iii.) EFpSGDK+ iv.) EFTGDK+ v.) EFpTGDK+ vi.) EFYGDK+ vii.) EFpYGDK+ i ii iii iv v vi vii Intensity

PAGE 105

105 Figure. 3-11. 10% CO2 45% He 45% N2 b.) 33.3% each c. 50% CO2, 25% He, 25% N2 d.) 100% CO2 Intensity

PAGE 106

10% CO290% N2 10 0 -10 -20 CV (V) Intensity 2.19x 1061.36x 1055.76x 1041.20x 1056.26x 1045.16x 1043.89x 104 10 0 -10 -20 CV (V) Intensity 2.19x 1061.36x 1055.76x 1041.20x 1056.26x 1045.16x 1043.89x 104 10 0 -10 -20 CV (V) 3.65x 1061.56x 1057.14x 1041.73x 1052.25x 1056.69x 1047.88x 104 10 0 -10 -20 CV (V) 3.65x 1061.56x 1057.14x 1041.73x 1052.25x 1056.69x 1047.88x 10450% CO250% N2a b c d e f g a b c d e f g 10% CO290% N2 10 0 -10 -20 CV (V) Intensity 2.19x 1061.36x 1055.76x 1041.20x 1056.26x 1045.16x 1043.89x 104 10 0 -10 -20 CV (V) Intensity 2.19x 1061.36x 1055.76x 1041.20x 1056.26x 1045.16x 1043.89x 104 10 0 -10 -20 CV (V) 3.65x 1061.56x 1057.14x 1041.73x 1052.25x 1056.69x 1047.88x 104 10 0 -10 -20 CV (V) 3.65x 1061.56x 1057.14x 1041.73x 1052.25x 1056.69x 1047.88x 10450% CO250% N210% CO290% N2 10 0 -10 -20 CV (V) Intensity 2.19x 1061.36x 1055.76x 1041.20x 1056.26x 1045.16x 1043.89x 104 10 0 -10 -20 CV (V) Intensity 2.19x 1061.36x 1055.76x 1041.20x 1056.26x 1045.16x 1043.89x 104 10 0 -10 -20 CV (V) 3.65x 1061.56x 1057.14x 1041.73x 1052.25x 1056.69x 1047.88x 104 10 0 -10 -20 CV (V) 3.65x 1061.56x 1057.14x 1041.73x 1052.25x 1056.69x 1047.88x 10450% CO250% N2a b c d e f g a b c d e f g a b c d e f g a b c d e f g Figure 3-12. Variation of % CO2 in the N2 carrier gas. i.) total ion chromatogram ii.) EFSGDK iii.) EFpSGDK iv.) EFTGDK v.) EF pTGDK vi.) EFYGDK vii.) EFpYGDK

PAGE 107

107 Figure 3-13. CV peaks for varying SF6/He carrier gas combinations. a.) total ion trace b.) EFSGDK+ c.) EFpSGDK+ d.) EFTGDK+ e.) EFpTGDK+ f.) EFYGDK+ g.) EFpYGDK+ a b c d e f g a b c d e f g

PAGE 108

108 Figure 3-14. Mass spectra of vary percentages of He in SF6 carrier gas a.) 10% He 90% Spectrum collected over the range of CV +0.8 to -4.0 V. SF6 b.) 25% He 75% SF6; Spectrum collected over the range of CV -0.2 to -4.9 V. a b

PAGE 109

109 Figure 3-15. Plot of ion intensity versus CV in FAIMS with carrier gas He/SF6. a.) total ion trace b.) EFSGDK+ c.) EFpSGDK+ d.) EFTGDK+ e.) EFpTGDK+ f.) EFYGDK+ g.) EFpYGDK+ a b c d e f g

PAGE 110

110 Figure 3-16. Comparison of He/N2 spectra and He/SF6 spectrum. +2 charge -states are not seen in SF6/He spectra scanned from CV 50 to -100 V. a) He/N2 carrier gas. The spectrum is shown while the CV was scanned from -4.1 V to -8.0 V. +1 charge-state ions are seen b.) He/N2 carrier gas. The CV was scanned from -9.2 V to -14.4 V. +2 chargestate ions are seen. c.) He/SF6 carrier gas. The CV was scanned from -6.3 V to-11.3 V. Poor signal-to-noise is seen in He/SF6 carrier gas. Furthermore, not all phosphopeptides are seen. a b c

PAGE 111

111 Inner Electrode 40 C -30-24-18-12-60 Compensation Voltage Value (V) 0 20 40 60 80 100Relative Intensity Inner Electrode 40 C -30-24-18-12-60 Compensation Voltage Value (V) 0 20 40 60 80 100Relative Intensity -30-24-18-12-60 Compensation Voltage Value (V) 0 20 40 60 80 100Relative Intensity Inner Electrode 70 C -30-24-18-12-60 Compensation Voltage Value (V) 0 20 40 60 80 100Relative Intensity Inner Electrode 70 C -30-24-18-12-60 Compensation Voltage Value (V) 0 20 40 60 80 100Relative Intensity -30-24-18-12-60 Compensation Voltage Value (V) 0 20 40 60 80 100Relative Intensity Inner Electrode 90 C -30-24-18-12-60 Compensation Voltage Value (V) 0 20 40 60 80 100Relative Intensity Inner Electrode 90 C -30-24-18-12-60 Compensation Voltage Value (V) 0 20 40 60 80 100Relative Intensity -30-24-18-12-60 Compensation Voltage Value (V) 0 20 40 60 80 100Relative Intensity Figure 3-17. The effect of heated FAIMS electrodes on resolution and peak shape. Outer electrode is 70 C. The blue line is MIFAGIK+2 and the black line is MIFAGIK+.

PAGE 112

112 The Effect of Electrode Temperature on Resolution30/3040/50 30/70 50/70 50/9070/90 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 Temperature of the Inner/Outer Electrodes ( C)Resolution 30/30 40/50 30/70 50/70 50/90 70/90 Figure 3-18. The effect of electrode te mperature on resolution between EFYGDK and EFpSGDK ( the two closest peaks be tween phosphorylated and nonphosphorylated peptides). ( C)

PAGE 113

113 Figure 3-19. The +2 charge-state ion intens ities at different CVs. The electrode temp eratures were varied independently. a.) total ion chromatogram b.) EFSGDK+2 c.) EFTGDK+2 d.) EFYGDK+2 e.) EFpSGDK+2 e.) EFpTGDK+2 f.) EFpYGDK+2

PAGE 114

114 Figure 3-20. The +1 charge-state ion intens ities at different CVs. The electrode temp eratures were varied independently. a.) total ion chromatogram b.) EFSGDK+ c.) EFpSGDK+ d.) EFTGDK+ e.) EFpTGDK+ e.) EFYGDK+ f.) EFpYGDK+

PAGE 115

115 CHAPTER 4 DESIGN OF A NOVEL HEMISP HERICAL FAIMS GE OMTERY FOR PHOSPHOPEPTIDE SEPARATION Overview This chapter discusses the design of novel he mispherical FAIMS electrodes for improved phosphopeptide separation. Current cell geometries will be discussed, as well as the limitations in the design of spherical and hemispheri cal FAIMS electrodes. The application of hemispherical FAIMS electrodes to phosphope ptide detection will be demonstrated. Current FAIMS Electrode Geometries The geometry of the FAIMS electrodes affects ion transmission and resolution. The first electrode geometry for FAIMS implemented planar electrodes developed by Buryakov et al. in 1993.18 Sionex (Waltham, MA) currently manufactures FAIMS with a planar geometry as a stand alone detector.18,103 The dome cell geometry was developed in 1995 by Carnahan et al., but is no longer commercially available.64 In 2004, Guevremont et al. patented cylindrical FAIMS electrodes, which are currently commer cially produced by Thermo-Fisher Scientific (San Jose, CA) as a separation device to be used in conjunction with a mass spectrometer.65 Each electrode geometry has their advantages an d disadvantages in the optimization of resolution and ion transmission. Planar Geometry Electrodes Flat plate FAIMS electrodes, also known as planar electrodes, were first introduced in 1993 by Buryakov et al.18 A basic schematic of the flat plat e geometry is shown in Figure 4-1. The asymmetric waveform can be applied to either electrode, while the othe r is held at ground or a small bias. The main advantage of the plan ar geometry is improved resolution. Resolving powers of 40 have been reported, and resolu tion improves as ion residence time increases.104 Therefore, even ions with small variations in CV can be separated if the planar electrodes were

PAGE 116

116 long enough. However, planar geometry suffers from poor transmission. The electric field between the inner and outer electrode s is uniform, allowing ions to diffuse laterally. The lack of ion focusing results in poor ion transmi ssion into the narrow entrance of the mass spectrometer.65,104 Ion transmission has been improved by implementing an ion funnel at the FAIMS-MS interface.104 Cylindrical Geometry The cylindrical geometry, also known as th e side-by-side geometry, is currently commercialized by Thermo-Fisher Scientific and is sold as a separation device to be used in conjunction with a mass spectrometer. Schematic s of the cylindrical FAIMS electrodes are shown in Figure 4-2. This geometry has better transmission than the flat plate electrodes and only decrease signal by an order of magnitude Improved transmission is due to twodimensional atmospheric pressure ion focusing b ecause the non-uniform electric field created by the curved surface of the electrodes facilitates fo cusing ions into the center of the analytical gap.64,65 Decreased collisions with the electrode s due to the enhanced focusing results in improved sensitivity. The electric field at any point between the electrodes can be cal culated using Equation 4-1. Va is the applied voltage, r is the radius of the ion at any point between the electrodes, a is the radius of the center electrode, and b is the radius of the distance between the center of the inner electrode and inner surface of the outer electrode, as shown in Figur e 4-3. On the other hand, the electric field between planar electrodes can be calculated using Equation 4-2, which demonstrates the uniformity of the electric field created between planar electrodes. Figure 4-4 demonstrates the difference in the field between the planar and cylindrical geometries.

PAGE 117

117 (4-1) (4-2) Due to the non-uniformity of the electric field in cylindrical electrodes, ion focusing can be achieved, as shown in Figure 4-5. Ions near th e inner electrode are repe lled to the center of the analytical gap, whereas ions near the outer electrode are pulled to the center of the analytical gap. Barnett et al. described this effect, which can be experimentally observed in Figure 4-6.63 The solid line is the actual DV an ion requires to tr averse the electrodes. However, the electric field variations allow ions of many different ion mobility characteristics to be focused between the electrodes as well. For instance, the dotted line CVcorrect in Figure 4-6 represents the electric field gradient for an example DV, where the CV is the opposite polarity of the ion. Ions near the inner electrode, where the field is highest, w ould be repelled away towards the center of the electrodes where they will fall on the solid line an d are focused. Ions near the outer electrode, where the field is weak, can be pulled towards the higher fields in the cente r of the electrode and are focused. The dotted line CVlow represents ions that would have normally collided with the outer electrode. However, because the non-uniform field is now higher ne ar the inner electrode, the ion is pulled away from the outer and is focuse d. On the other hand, ions that fall on the dotted line CVhigh would have collided with the inner el ectrode if the field was nonuniform. The atmospheric pressure ion focusing causes them to be repelled from the higher fields near the inner electrode and be focused to the detector. The line CVincorrect represents a CV that, despite focusing, is not strong enough to pull the ions aw ay from the outer electrode and they are not focused. This demonstrates the effect of i on focusing due to curved geometry at atmospheric pressure.

PAGE 118

118 However, the use of curved surfaces to increase sensitivity compromises resolution.65 Guevremont et al. theorized that curved surfaces focus ions with a wide range of mobility behavior, which decreases resolu tion and increases peak width.65 As the radii of curvature increases, focusing improves and ion intensity increas es at the expense of resolution. Ions with very different mobility behavior can be focused together when utilizing curved electrodes. As the electrode curvature decreases, the focusing effect lessens and reso lution improves at the expense of sensitivity. Without the curvature, th ere is no focusing effect and ions with different mobility behavior can be distinguished. Therefore, as the electrode radii increases, resolution increases. Planar geometry electrodes have no curvature, thus they exhibit the maximum attainable resolution.65 Dome Cell Geometry The dome cell geometry was first proposed by Carnahan in 1995, but is not currently commercialized. The schematic of the dome cell geometry is shown in Figure 4-7. Modeling demonstrates that the ions are separated in the cylindrical portion and focused in the spherical portion.105 The gap between the inner a nd outer electrode in the sphe rical portion is adjustable. A gap between 1.7 and 2.5 mm has been found to be optimal for most compounds. The gap between the inner and outer electr ode on the cylindrical portion is 2 mm, so the electric field may differ between the cylindrical and spherical por tion of the dome electrodes allowing enhanced resolution and transmission. However, this design is not commercially available. Spherical FAIMS Electrodes Guevremont et al. patented spherical electr odes in 2004, but the patent was never reduced to practice.66 Spherical FAIMS electrodes were predic ted to improve ion focusing because all ion path lengths are equidistant, and therefore all ions experience the same field and travel the same distance. With spherical geometry, i ons can distribute uniformly around the inner

PAGE 119

119 electrode. On the other hand, in a cylindrical design, when ions spread out due to chargerepulsion and diffusion they may take different paths along the inner electrode, causing peak broadening. Figure 4-8 demonstr ates the different paths an i on can take through spherical and cylindrical electrodes. Furthermore, the radial field in the spheri cal geometry varies as a function of 1/r2, whereas the radial field in cylindrical cell only varies as a function of 1/r, allowing improved ion focusing capabilities between spherical electrodes.66 It has been hypothesized th at the resolution obtained by spherical FAIMS electrodes would be poor be cause of the focusing effects caused by the curvature.65,104 By varying the radius of the spherica l electrodes, it may be possible to alter resolution and sensitivity in this electrode design. A balance of resolution and sensitivity must be obtained when detecting phosphopeptides in tryptic digests. As many phosphopeptides are in low abundance, sens itivity is critical. Resolution is also important in the elucidati on of phosphopeptides because cell digests can be very complicated. A spherical cell was developed to test the ability of this electrode design to balance resolution and sensitivity. Designing Spherical FAIMS Electrodes An ideal spherical inner electrode would be perfectly centered with in a spherical outer electrode. However, suspending the inner electrode without disr upting the electric field is problematic. Several designs to suspend the inner electrode within the outer electrode were conceived. Figure 4-9 shows the schematics of s pherical FAIMS electrodes that were reduced to practice. A 0.7 mm rod was driv en through the center of an 8. 8 mm stainless steel ball to produce the spherical inner electrode. The stainle ss steel rod was necessary to both suspend the inner electrode, as well as to deliver the as ymmetric waveform and compensation voltage. MACOR sleeves (4 mm in diameter) were designed to insulate the stainless steel rod holding the

PAGE 120

120 inner electrode so that the asymmetric waveform would only emanate from the inner electrode. Cuffs were added onto the Maycor sleeves to minimize any horizontal movement of the inner electrode, to help ensure a uniform 2 mm gap betw een the inner and outer electrodes. The outer electrode was made of brass by joining two hemispheri cal cavities around the inner electrode, as shown in Figure 4-10. The outer electrode was then encased in Delrin ; however, subsequent designs were encased in Kel-F because Delrin is temperature sensitive. Kel-F has a working temperature range of -400 F to 400 F, so it is frequently used in gaskets, seals, valves and aerospace designs, making Kel-F more amenable for interfacing with the heated capillary of the mass spectrometer than Delrin.106 Design of Novel Hemispherical FAIMS Electrodes Due to the difficulties encountered using a spheri cal inner electrode, as will be described in the results section of this chapter, a hemisphe rical inner electrode was designed, as shown in Figure 4-11. The stainless steel inner electrode was made into a near hemisphere and screwed down to a Kel-F base plate, ensuring accurate cente ring and a uniform analytical gap. The inner electrode was extended 2 mm more than a true hemi sphere so that the ions entering the analytical gap would encounter a spherical electrode and any path taken ar ound the inner electrode to the exit would be identical. The asymmetric wa veform and compensation voltage were applied through the screw to the inner electrode. The stainless steel outer electrode was placed over the inner electrode such that there remained a 2 mm analytical gap. Photos of the hemispherical electrodes are shown in Fi gures 4-12 to 4-14. The original design of the out er electrode was PEEK machin ed into a hemisphere and coated with gold in order to apply an outer bias. However, imperfections in the gold coating led to arcing and destroyed the outer electrode, as shown in Figure 4-15. In the second design, the

PAGE 121

121 outer electrode was made from stainless steel, wh ich is much more difficult to machine, but is more resistant to damage from arcing. Improvements to the original spherical electr ode FAIMS curtain plate design were also implemented in the hemispherical electrode FAIMS design. The Delrin originally was designed such that it encompassed the stainl ess steel curtain plate. An excess of Delrin at the ion entrance could have led to charge build-up near the entr ance of the curtain plate; therefore, in the hemispherical design, the Kel-F was cut back, su ch that the ions only encountered the curtain plate. Furthermore, the space between the outer electrode and the curtain plate was minimized to 1 mm, allowing ions to encounter only a minimum amount of Kel-F before entering the analytical gap. These improvements in th e design yielded a working FAIMS device. Experimental Samples Explosives and phosphopeptides were tested on hemispherical FAIMS. A 10 ppm mixture of TNT and 3,4-DNT in 65/35 methanol/water was made. A solution of 20 ppm synthetic tryptic peptides EFXGDK, where X can equal S, T, Y, pS, pT, or pY in 50/50 water/methanol with 1% acetic acid was also used. Instrumentation An alpha waveform generator (Ionalytics, Ot tawa, Canada) was used to generate the 750 kHz asymmetric waveform. The spherical and hemispherical FAIMS el ectrodes were made by the Department of Chemistry Mach ine Shop at the Universi ty of Florida (Gainesville, FL). Data were compared to a commercial Thermo FAIMS (Thermo-Fisher Scientific, San Jose, CA) with cylindrical geometry electrodes.

PAGE 122

122 Results Data Collection from Spherical FAIMS Electrodes No data could be obtained from the spherical FAIMS electrodes. This is likely due to the inability to perfectly center the sp herical inner electrode to maintain a uniform 2 mm gap. If the gap is non-uniform then the variation in electric field within the cell may be too great for the ion to successfully traverse the electrodes. Th e diameter of the MACO R sleeves, which was significant in comparison to the size of the inner electrode, may have also affected the electric field. If the rods holding up the inner elec trode distorted the field, ion separation and transmission would be adversely affected. Fu rthermore, ions may have discharged on the MACOR sleeves and built up a charge. Scientists at Thermo-Fisher Scientific independently designed and tested a similar s pherical FAIMS geometry and concluded the same results. (Mike Belford, personal communication) To test the theory about a non-uniform gap not allowing the passage of ions, Thermo-Fisher Scientific scientists machined cylindrical geometry electrodes that purposely were off-center by a few percent a nd confirmed that they would not work either. Data Acquired on Hemispherical FAIMS The highest DV achieved on the hemispherical cell was -2700 V due to malfunctions in the alpha waveform generator. This DV was not sufficient for tryptic peptide ion separation in either the hemispherical or cyli ndrical electrodes. Figure 4-16 shows the CV spectrum acquired at DV -2700 V. The tryptic peptides ions have low intensities and the CV is around zero. The MS spectrum is shown in Figure 4-17. The signal-to-noise is low, and EFpYGDK and EFpSGDK are not distinguishable A higher DV is needed achie ve good S/N and resolution. TNT and 3,4-DNT were also tested on hemi spherical FAIMS and improved resolution and resolving power was achieved when compared to the commercial cylindrical FAIMS electrodes. 3,4-DNT and TNT were separated w ith a resolution of 1.0 at a DV of -2700, as shown in Figures

PAGE 123

123 4-18 and 4-19. The resolving power for TNT wa s 5.8 and the resolving power for 3,4-DNT was 6.7. At this same field, the commercial cylin drical electrodes had a re solving power of 1.9 and 2.5 for TNT and 3,4-DNT, respectivel y, and the resolution was 0.4, as shown in Table 4-1. The optimum separation, even at higher fields (18, 000 V/cm), was still poorer than the separation achieved by the hemispherical electrodes. At 18,000 V/cm, the commercial FAIMS produced a resolution of 1.0 and a resolving power of 4.2 and 4.9 for TNT and 3, 4-DNT, respectively. However, ion transmission is decreased in the spherical FAIMS when compared to the Thermo cylindrical FAIMS. Though the transmissions appear similar, as shown in Table 4-1, the Thermo FAIMS is attached to the LCQ via a capillary extender, which greatly reduces ion transmission. Therefore, if the Thermo FAIM S was attached to the TSQ, as commercially intended, the ion transmission w ould be higher. Increasing the radius of the inner electrode would improve ion transmission at the expense of resolution. The application of hemispherical FAIMS for the separation of explosives is currently being explored further. Conclusion The hemispherical electrodes show promise for improving the resolu tion of analytes over the commercial device without compromising too much sensitivity. Improved resolution and resolving power was observed when using the hemispherical electrodes compared to the commercially available electrodes. However, high er DVs are still needed in order to detect phosphopeptides. The hemispherical electrode diam eter could be varied to help improve the balance between resolution and signal intensity.

PAGE 124

124 ~ ((( Ion entrance Exit to detector ~ ((( Ion entrance Exit to detector Figure 4-1. Planar geometry FAIMS. Ions from Source Ions to Mass Spectrometer Inner Electrode Outer Electrode Outer Electrode Entrance plate Figure 4-2. Cylindrical FAIMS electrodes. Adapted from an illustration by Richard Yost, 2008.

PAGE 125

125 Figure 4-3. Measurements for a, b, and r used to calculate the electric fields within a curved surface.63

PAGE 126

126 Curved Geometry55 60 65 70 75 80 85 90 95 100 6.5 7 7.5 8 8.5 9 r (2.5 mm gap)Field (Td) Uniform 30 Uniform Flat 30 Planar vsCurved Geometry Field (Td) R (2.5 mm gap) 100 90 80 70 60 6.5 7.0 7.5 8.0 8.5 9.0 Curved Geometry55 60 65 70 75 80 85 90 95 100 6.5 7 7.5 8 8.5 9 r (2.5 mm gap)Field (Td) Uniform 30 Uniform Flat 30 Planar vsCurved Geometry Field (Td) R (2.5 mm gap) 100 90 80 70 60 6.5 7.0 7.5 8.0 8.5 9.0 Figure 4-4. Comparison of the elec tric fields in cylindrical and planar geometries. Adapted from an illustration by Leonard Rorrer, 2008.

PAGE 127

127 Inner Electrode Outer Electrode Outer Electrode Inner Electrode Outer Electrode Outer Electrode Figure 4-5. Non-uniform electric field creates a focusing effect.

PAGE 128

128 Compensation Field (Td)Ions repelled away from the inner electrode Ions pulled away from the outer electrode CVcorrect CVincorrect CVlow CVhigh Dispersion Field (DF) Compensation Field (Td)Ions repelled away from the inner electrode Ions pulled away from the outer electrode CVcorrect CVincorrect CVlow CVhigh Compensation Field (Td)Ions repelled away from the inner electrode Ions pulled away from the outer electrode CVcorrect CVincorrect CVlow CVhigh Dispersion Field (DF) Figure 4-6. Focusing effect caused by the curvature of the electrodes. Adapted from Barnett et al.63

PAGE 129

129 Figure 4-7. Dome cell geometry FAIMS.19

PAGE 130

130 Cylindicalelectrodes Spherical electrodes Lateral diffusion Cylindicalelectrodes Spherical electrodes Lateral diffusion Figure 4-8. Different paths between the elect rodes for spherical and cylindrical FAIMS electrodes. The spherical elect rode design, all ion paths are equal. In the cylindrical electrode design, lateral diffusi on can cause peak broadening.

PAGE 131

131 Stainless steel inner electrode Analytical gap Brass outer electrode Stainless steel rod MACOR sleeves Ion exit Ion entrance Stainless steel inner electrode Analytical gap Brass outer electrode Stainless steel rod MACOR sleeves Ion exit Ion entrance Figure 4-9. Spherical FAIMS elect rodes with rods. Ions entere d and traveled around the inner electrode and MACOR sleeves in th e analytical gap to the exit.

PAGE 132

132 Inner electrode MACOR sleeve over steel rod holding inner electrode Delrincasing Outer electrode Lead to outer electrode Inner electrode MACOR sleeve over steel rod holding inner electrode Delrincasing Outer electrode Lead to outer electrode Figure 4-10. Spherical FAIMS electrodes

PAGE 133

133 Figure 4-11. Hemispherical FAIMS electrode s. Illustrated by Todd A. Prox, 2008.

PAGE 134

134 Inner electrode Outer Electrode Kel-F Casing Inner electrode Outer Electrode Kel-F Casing Figure 4-12. Inner and outer elec trodes of hemispherical FAIMS. Inner electrode Kel-F casing over outer electrode Outer bias lead Inner electrode Kel-F casing over outer electrode Outer bias lead Figure 4-13. Hemispherical FAIMS disassembled.

PAGE 135

135 Outer bias lead Curtain plate Kel-F casing Outer bias lead Curtain plate Kel-F casing Figure 4-14. Assembled hemispherical FAIMS.

PAGE 136

136 PEEK housing Damage From arcing Gold-plated outer electrode PEEK housing Damage From arcing Gold-plated outer electrode Figure 4-15. Gold-plated outer electr ode was destroyed due to arcing.

PAGE 137

137 Figure 4-16. CV scans for EFXGDK using hemisphe rical FAIMS at a DV of -2700 V (13,500 V/cm). -10 -5 0 5 10 CV (V) Intensity 1.20x 106EFSGDK EFTGDK EFYGDK EFpSGDK EFpTGDK EFpYGDK TIC3.24x 1044.58x 1046.24x 1041.77x 1043.33x 1044.04x 104 -10 -5 0 5 10 CV (V) Intensity 1.20x 106EFSGDK EFTGDK EFYGDK EFpSGDK EFpTGDK EFpYGDK TIC3.24x 1044.58x 1046.24x 1041.77x 1043.33x 1044.04x 104

PAGE 138

138 620 640 660 680 700 720 740 760 780 m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 687.73 696.20 682.40 741.07 758.07 670.87 773.80 720.60 663.60 768.20 612.60 790.33 746.33 633.40 718.53 652.27 704.20 739.13 616.33 726.47 648.73 Intensity: 2.06x 103 EFTGDK EFSGDK EFYGDK EFpTGDK 620 640 660 680 700 720 740 760 780 m/z 0 10 20 30 40 50 60 70 80 90 100Relative Abundance 687.73 696.20 682.40 741.07 758.07 670.87 773.80 720.60 663.60 768.20 612.60 790.33 746.33 633.40 718.53 652.27 704.20 739.13 616.33 726.47 648.73 Intensity: 2.06x 103 EFTGDK EFSGDK EFYGDK EFpTGDK Figure 4-17. The mass spectrum of tryptic pept ide ions are shown us ing the hemispherical FAIMS at low DV (-2700 V). The CV was scanned from -2.5 to 0 V.

PAGE 139

139 4.0 8.0 0.0 CV (V) Intensity 1.75 x 1055.04 x 1049.01 x 103TIC TNT CV = 2.9 3,4-DNT CV = 4.1 4.0 8.0 0.0 CV (V) Intensity 1.75 x 1055.04 x 1049.01 x 103 4.0 8.0 0.0 CV (V) Intensity 1.75 x 1055.04 x 1049.01 x 103TIC TNT CV = 2.9 3,4-DNT CV = 4.1 Figure 4-18. The separation of 3,4-DNT and TN T using hemispherical FA IMS at a DV of -2700 V. The resolution is 1.0. The arrows on the peak represent where the spectra were obtained from in Figure 4-19.

PAGE 140

140 100 150 200 250 300 350 400 m/z 0 20 40 60 80 100 0 20 40 60 80 100Relative Abundance 197.0226.9 210.0 152.0 166.9 180.9 137.0 181.8 151.9Intensity = 6.72x 103Intensity = 1.41x 103 [3,4-DNT]-[3,4-DNT-NO][TNT]-[TNT-NO]100 150 200 250 300 350 400 m/z 0 20 40 60 80 100 0 20 40 60 80 100Relative Abundance 197.0226.9 210.0 152.0 166.9 180.9 137.0 181.8 151.9Intensity = 6.72x 103Intensity = 1.41x 103 [3,4-DNT]-[3,4-DNT-NO][TNT]-[TNT-NO]Figure 4-19. Separation of TNT and 3.4-DNT. CV was scanned from 2.7 to 3.2 V in the top spectrum and from 3.9 to 4.4 V in the bottom spectrum, as shown by arrows in Figure 4-18.

PAGE 141

141 Table 4-1. Comparison of resolution and re solving power between the hemispherical FAIMS and the commercial Thermo FAIMS. Hemispherical Thermo Commercial Electrodes Electrodes Resolution TNT and 3,4-DNT at 13,500 V/cm 1.0 0.4 at 18, 000 V/cm N/A 1.0 Resolving Power TNT at 13,500 V/cm 5.8 1.9 at 18,500 V/cm N/A 4.2 3,4-DNT at 13,500 V/cm 6.7 2.5 at 18,000 V/cm N/A 4.9 Intensity TNT at 13,500 V/cm 5.0 x 104 6.1 x 104 at 18,000 V/cm N/A 1.6 x 105 3,4-DNT at 13,500 V/cm 9.0 x 103 2.0 x 104 at 18,500 V/cm N/A 7.4 x 104

PAGE 142

142 CHAPTER 5 CONCLUSION FAIMS is an em erging separation technique that is fast and specific. Gaining a better understanding of how this technique separates tryptic peptide ions will lead to improved protein analysis. Previous tryptic peptide separation tec hnology is limited in its se lectivity and analysis time. The application of FAIMS to tryptic pept ides analysis can improve signal-to-noise, while reducing analysis time. The implementation of FAIMS could be expanded to other analyte ions as well for more efficient separations than curre nt technology allows. Pharmaceutical analysis, forensic studies and others coul d benefit from FAIMS technology. Data-dependent analysis has allowed the rapid screening and detection of tryptic peptides. Phosphopeptide ions and ions of +2 charge-states are two ions of interest targeted in datadependent analysis. The exclusive separation of +2 charge-state tryptic peptide ions by FAIMS for subsequent analysis by MS would yield produ ct spectra with more relevant data, improving the accuracy of protein identification. Data-depe ndent analysis frequently selects a predefined number of high abundance peaks regardless of thei r relevance. The exclusive separation of +2 charge-state tryptic peptide i ons would improve the eliminatio n of interferences from other compounds. Furthermore, the +2 charge-state product ion spectra acquired by tandem mass spectrometry would provide more structural information because the product ions are predominantly +1 charge-states and bm-yn ion pairs are more likely to form. FAIMS was shown to improve the separation of charge-states on bot h synthetic tryptic peptides and on a digest of Cytochrome C, improving the signal-to-noise and increasing the likelihood that the +2 ch arge-state would be selected by da ta-dependent analysis. However, FAIMS was not able to exclusively remove all +1 charge-states. The best conditions for the

PAGE 143

143 separation of charge-states in FAIMS were using a DV of -5000 V, a carrier gas composition of 50/50 v/v He/N2, and a carrier gas flow rate around 3.5 L/min. Identifying phosphopeptides is important in th e elucidation of phosphoprotein structure. Gaining and understanding of which proteins are phosphorylated, where they are phosphorylated and quantifying the phosphorylation can gain new insight on biomarkers and disease dysfunction. This dissertation presente d the separation of phosphopeptides from nonphosphorylated peptides in synthetic mixtures The use of heated electrodes improved phosphopeptide separation without compromising signal intensity and doubled the resolution between +2 charge-state phosphopeptides and +2 charge-state nonphosphorylated peptides. However, heated electrodes did not improve the resolution between the same tryptic peptide ions in the +1 charge-state. Carrier gas compositions were altered to include combinations of SF6, N2, He, and CO2. It was found that the use of CO2 in the carrier gas decreased resolution and signal intensity. The use of 50% SF6 50% He in the carrier improved resolution, but significantly decreased signal. Therefore, it was determined that the best conditions for the separation of phosphopeptides were 50/50 v/v N2/He in the carrier gas and heated electrodes set to 50C for the inner electrode and 90 C for outer electrode. Electrode geometry was also evaluated because it affects ion transmission and resolution. The use of planar electrode geometry is known to improve resolution, whereas the use of cylindrical electrodes is known to improve transmission. Spheri cal electrode and hemispherical electrode designs were tested for their effects on transmission and resolution. The spherical electrode design was never functiona l, likely due to difficulties centering the inner electrode to maintain a uniform analytical gap. The hemisp herical electrode design wa s only able to achieve a dispersion voltage of -2700 V due to problems with the Ionalytics waveform generator.

PAGE 144

144 However, even at low voltages, the hemispheri cal electrodes offered improved resolution and resolving power between TNT a nd 3,4-DNT over the commercial FAIMS device, even when the commercial FAIMS was at higher voltages. Ho wever, ion transmission was decreased when compared to the commercial FAIMS. There are many opportunities for future work in this project. The variation of the analytical gap, dispersion volta ge, and electrode diameters would affect the ion transmission and resolution in hemispherical FAIMS. By increasing or decreasing the electrode radii, a balance between ion transmission and reso lution can be optimized. Higher DVs also need to be achieved in order to improve transmission. Further stud ies can evaluate tryptic peptide ion transmission and separation at higher DVs on the hemispherica l FAIMS. Future work could also include studying FAIMS separations of diffe rent protein modifications, such as glycosylation, which is also important in biological regulatory mechanisms.

PAGE 145

145 APPENDIX AMINO ACIDS Figure A-1. Amino Acids.2

PAGE 146

146 LIST OF REFERENCES 1. Becker, W.; Kleinsmith, L; Hardin, J. The World of the Cell, 5th ed.; Electronic Publishing Services: New York, 2003. 2. Nelson, D.; Cox, M. Principles of Biochemistry, 3rd ed.; Worth Publishers: New York, 2000. 3. Pinkse, M. W. H.; Uitto, P. M.; H ilhorst, M. J.; Ooms, B.; Heck, A. J. R. Analytical Chemistry 2004, 76, 3935-43. 4. Biemann, K. Annual Review of Biochemistry 1992, 61, 977-1010. 5. Wysocki, V. H.; Resing, K. A.; Zhang, Q. F.; Cheng, G. L. Methods 2005, 35, 211-22. 6. Proteome Research : Mass Spectrometry, Springer: Berlin, NY, 2001. 7. Flora, J. W.; Muddiman, D. C. Journal of the American Chemical Society 2002, 124, 6546-47. 8. Ge, Y.; Lawhorn, B. G.; ElNaggar, M.; Strauss, E.; Park, J. H.; Begley, T. P.; McLafferty, F. W. Journal of the American Chemical Society 2002, 124, 672-78. 9. Price, W. D.; Schnier, P. D.; Williams, E. R. Analytical Chemistry 1996, 68 859-66. 10. Houston, C. T.; Reilly, J. P. Journal of Physical Chemistry A 2000, 104, 10383-91. 11. Ruotolo, B.;Verbeck, G.; Thomson, L.; Woods, A.; Gillig,K.; and Russell, D. Journal of Proteome Research 2008, 1, 303-306. 12. Zubarev, R. A. Mass Spectrometry Reviews 2003, 22, 57-77. 13. Giese, R. W. Journal of Chromatography A 2000, 892, 329-46. 14. Breuker, K.; McLafferty, F. Chemical International 2003, 42, 4900. 15. Tholey, A.; Toll, H.; Huber, C. G. Analytical Chemistry 2005, 77, 4618-25. 16. Flora, J. W.; Muddiman, D. C. Analytical Chemistry 2001, 73, 3305-11. 17. Molloy, M. P.; Andrews, P. C. Analytical Chemistry 2001, 73, 5387-94. 18. Buryakov, I. A.; Krylov, E. V.; Nazarov, E. G.; Rasulev, U. K. International Journal of Mass Spectrometry and Ion Processes 1993, 128, 143-48. 19. Creaser, C. S.; Griffiths, J. R.; Bramwell, C. J.; Noreen, S.; Hill, C. A.; Thomas, C. L. P. Analyst 2004, 129, 984-94. 20. Tam, M.; Hill, H. H. Analytical Chemistry 2004, 76 2741-47.

PAGE 147

147 21. Lawrence, A. H.; Neudorfl, P.; Stone, J. A. International Journal of Mass Spectrometry 2001, 209, 185-95. 22. Hoadlund-Hyzer, C. S.; Li, J. W.; Clemmer, D. E. Analytical Chemistry 2000, 72, 273740. 23. Taraszka, J. A.; Gao, X. F.; Valentine, S. J.; Sowell, R. A.; Koeniger, S. L.; Miller, D. F.; Kaufman, T. C.; Clemmer, D. E. Journal of Proteome Research 2005, 4, 1238-47. 24. Hilderbrand, A. E.; Myung, S.; Clemmer, D. E. Analytical Chemistry 2006, 78, 6792800. 25. Hilderbrand, A. E.; Myung, S.; Barnes, C. A. S.; Clemmer, D. E. Journal of the American Society for Mass Spectrometry 2003, 14, 1424-36. 26. Revercomb, H. E.; Mason, E. A. Analytical Chemistry 1975, 47 970-83. 27. Guevremont, R.; Barnett, D. A.; Purves, R. W.; Vandermey, J. Analytical Chemistry 2000, 72, 4577-84. 28. Counterman, A. E.; Clemmer, D. E. Journal of the American Chemical Society 1999, 121, 4031-39. 29. Hudgins, R. R.; Mao, Y.; Ratner, M. A.; Jarrold, M. F. Biophysical Journal 1999, 76, 1591-97. 30. Kinnear, B. S.; Kaleta, D. T.; Kohtani, M.; Hudgins, R. R.; Jarrold, M. F. Journal of the American Chemical Society 2000, 122 9243-56. 31. Hudgins, R. R.; Ratner, M. A.; Jarrold, M. F. Journal of the American Chemical Society 1998, 120, 12974-75. 32. Hudgins, R. R.; Jarrold, M. F. Journal of Physical Chemistry B 2000, 104, 2154-58. 33. Wu, C.; Siems, W. F.; Klasmeier, J.; Hill, H. H. Analytical Chemistry 2000, 72, 391-95. 34. Ruotolo, B. T.; Gillig, K. J.; Stone, E. G.; Russell, D. H.; Fuhrer, K.; Gonin, M.; Schultz, J. A. International Journal of Mass Spectrometry 2002, 219, 253-67. 35. Ruotolo, B. T.; Mclean, J. A.; Gillig, K. J.; Russell, D. H. Journal of the American Society for Mass Spectrometry 2005, 16, 158-65. 36. Ruotolo, B. T.; Verbeck, G. F.; Thomson, L. M.; Woods, A. S.; Gillig, K. J.; Russell, D. H. Journal of Proteome Research 2002, 1, 303-06. 37. Wyttenbach, T.; vonHelden, G.; Bowers, M. T. Journal of the American Chemical Society 1996, 118, 8355-64.

PAGE 148

148 38. Ruotolo, B. T.; Tate, C. C.; Russell, D. H. Journal of the Amer ican Society for Mass Spectrometry 2004, 15, 870-78. 39. Barnes, C. A. S.; Hilderbrand, A. E.; Valentine, S. J.; Clemmer, D. E. Analytical Chemistry 2002, 74, 26-36. 40. Valentine, S. J.; Counterman, A. E.; Ho aglund, C. S.; Reilly, J. P.; Clemmer, D. E. Journal of the American Soc iety for Mass Spectrometry 1998, 9, 1213-16. 41. Barnett, D. A.; Ding, L. Y.; Ells, B.; Purves, R. W.; Guevremont, R. Rapid Communications in Mass Spectrometry 2002, 16, 676-80. 42. Guevremont, R. Journal of Chromatography A 2004, 1058, 3-19. 43. Wu, A.; Prieto, M.; Rorrer, L.; Yost, R. 56th ASMS Conference on Mass Spectrometry and Allied Topics 2008. 44. Handy, R.; Barnett, D. A.; Purves, R. W.; Horlick, G.; Guevremont, R. Journal of Analytical Atomic Spectrometry 2000, 15, 907-11. 45. Ells, B.; Barnett, D. A.; Purves, R. W.; Guevremont, R. Journal of Environmental Monitoring 2000, 2, 393-97. 46. Smith, C. J.; Wilson, I. D.; Abou-Shakra, F.; Payne, R.; Parry, T. C.; Sinclair, P.; Roberts, D. W. Analytical Chemistry 2003, 75, 1463-69. 47. Gabryelski, W.; Wu, F. W.; Froese, K. L. Analytical Chemistry 2003, 75, 2478-86. 48. Barnett, D. A.; Ells, B.; Guevremont, R.; Purves, R. W. Journal of the American Society for Mass Spectrometry 1999, 10, 1279-84. 49. Ells, B.; Barnett, D. A.; Froese, K.; Purves, R. W.; Hrudey, S.; Guevremont, R. Analytical Chemistry 1999, 71, 4747-52. 50. Ells, B.; Barnett, D. A.; Purves, R. W.; Guevremont, R. Analytical Chemistry 2000, 72, 4555-59. 51. Guevremont, R.; Makowski, B. American Laboratory 2005, 37 11. 52. Barnett, D. A.; Ells, B.; Guevremont, R.; Purves, R. W. Journal of the American Society for Mass Spectrometry 1999, 10, 1279-84. 53. Barnett, D. A.; Purves, R. W.; Guevremont, R. Nuclear Instruments & Methods in Physics Research Section A-Accelerators Spectrometers Detectors and Associated Equipment 2000, 450, 179-85. 54. Barnett, D. A.; Ells, B.; Guevremont, R.; Purves, R. W. Journal of the American Society for Mass Spectrometry 2002, 13, 1282-91.

PAGE 149

149 55. Barnett, D. A.; Ells, B.; Guevremont, R.; Purves, R. W. Journal of the American Society for Mass Spectrometry 1999, 10, 1279-84. 56. Ionalytics, Ottawa Canada. Ionalytics FAIMS Manual. 2008. 57. Shvartsburg, A. A.; Tang, K. Q.; Smith, R. D. Analytical Chemistry 2004, 76 7366-74. 58. Guevremont, R.; Barnett, D. A.; Purves, R. W.; Viehland, L. A. Journal of Chemical Physics 2001, 114, 10270-77. 59. Barnett, D. A.; Purves, R. W.; Ells, B.; Guevremont, R. Journal of Mass Spectrometry 2000, 35, 976-80. 60. McCooeye, M.; Ells, B.; Barnett, D.; Purves, R.; Guevremont, R. Analytical Toxicology 2001, 25, 81. 61. Ells, B.; Barnett, D.; Purves, R.; Guevremont, R. Journal of Environmental Monitoring 2000, 2, 393. 62. Purves, R. W.; Guevremont, R.; Day, S.; Pipich, C. W.; Matyjaszczyk, M. S. Review of Scientific Instruments 1998, 69, 4094-105. 63. Barnett, D. A.; Belford, M.; Dunyach, J. J.; Purves, R. W. Journal of the American Society for Mass Spectrometry 2007, 18, 1653-63. 64. Carnahan; Tarassov, A. [5420424]. 1995. 65. Guevremont, R.; Purves, R. [6825461]. 2004. 66. Guevremont, R.; Purves, R.; Barnett, D. [6713758]. 2004. 67. Shvartsburg, A. A.; Li, F. M.; Tang, K. Q.; Smith, R. D. Analytical Chemistry 2006, 78 3706-14. 68. Dole, M.; Mack, L. L.; Hines, R. L. Journal of Chemical Physics 1968, 49, 2240-&. 69. R.Cole Electrospray Ionization Mass Spectrometry. Fundamental, Instrumentation, and Applications., John Wiley and Sons, Inc: New York, NY, 1997. 70. Watson, J. Introduction to Mass Spectrometry, 3rd ed.; Lippincott-Raven Publishers: Philadelphia, PA, 2004. 71. Guevremont, R. 2008. www.faims.com Accessed 1/1/08 72. Purves, R. W.; Guevremont, R. Analytical Chemistry 1999, 71, 2346-57.

PAGE 150

150 73. Jarrold, M. F. Annual Review of Physical Chemistry 2000, 51 179-207. 74. Cleveland, D. W.; Fischer, S. G.; Kirschner, M. W.; Laemmli, U. K. Journal of Biological Chemistry 1977, 252, 1102-06. 75. Henzel, W. J.; Billeci, T. M.; Stults, J. T.; Wong, S. C.; Grimley, C.; Watanabe, C. Proceedings of the National Academy of Sc iences of the United States of America 1993, 90, 5011-15. 76. James, P.; Quadroni, M.; Carafoli, E.; Gonnet, G. Biochemical and Biophysical Research Communications 1993, 195 58-64. 77. Mann, M.; Hojrup, P.; Roepstorff, P. Biological Mass Spectrometry 1993, 22, 338-45. 78. Pappin, D. J. C.; Hojrup, P.; Bleasby, A. J. Current Biology 1993, 3, 487. 79. Yates, J. R.; Speicher, S.; Griffin, P. R.; Hunkapiller, T. Analytical Biochemistry 1993, 214, 397-408. 80. Eng, J. K.; Mccormack, A. L.; Yates, J. R. Journal of the Americ an Society for Mass Spectrometry 1994, 5, 976-89. 81. Chernushevich, I. V.; Fell, L. M.; Bloo mfield, N.; Metalnikov, P. S.; Loboda, A. V. Rapid Communications in Mass Spectrometry 2003, 17, 1416-24. 82. Issaq, H. J.; Chan, K. C.; Liu, C. S.; Li, Q. B. Electrophoresis 2001, 22, 1133-35. 83. J.Axelsson; C.Reimann; B.Sundquist Journal of Mass Spectrometry and Ion Processes 1994, 133, 141-55. 84. Baranov, V.; Bonner, R.; Yang, M. 48th ASMS Conference on Mass Spectrometry and Allied Topics, 2000. 85. Thomson, B. A.; Chernushevich, I. V. Rapid Communications in Mass Spectrometry 1998, 12, 1323-29. 86. Giles, K.; Wildgoose, J.; Pringle, S. ; and Bateman, R. 52nd ASMS Conference on Mass Spectrometry and Allied Topics 2004. 87. Hoyer, J.; Wildgoose, J.; Giles,K.; and Pringle, S. 51st ASMS Conference on Mass Spectrometry and Allied Topics 2003. 88. Venne, K.; Bonneil, E.; Eng, K.; Thibault, P. Analytical Chemistry 2005, 77, 2176-86. 89. Guevremont, R.; Barnett, D. A.; Purves, R. W.; Vandermey, J. Analytical Chemistry 2000, 72, 4577-84.

PAGE 151

151 90. Hale, J. E.; Butler, J. P.; Gelfanova, V.; You, J. S.; Knierman, M. D. Analytical Biochemistry 2004, 333, 174-81. 91. WurglerMurphy, S. M.; Saito, H. Trends in Biochemical Sciences 1997, 22 172-76. 92. Ficarro, S. B.; McCleland, M. L.; Stukenberg, P. T.; Burke, D. J.; Ross, M. M.; Shabanowitz, J.; Hunt, D. F.; White, F. M. Nature Biotechnology 2002, 20, 301-05. 93. Hunter, T. Cell 1995, 80, 225-36. 94. Price, W. D.; Schnier, P. D.; Williams, E. R. Analytical Chemistry 1996, 68 859-66. 95. Huddleston, M. J.; Annan, R. S.; Bean, M. F.; Carr, S. A. Journal of the American Society for Mass Spectrometry 1993, 4, 710-17. 96. Stanley, B. A.; Gundry, R. L.; Cotter, R. J.; Van Eyk, J. E. Disease Markers 2004, 20, 167-78. 97. Larsen, M. R.; Trelle, M. B. ; Thingholm, T. E.; Jensen, O. N. Biotechniques 2006, 40, 790-98. 98. Kinoshita, E.; Yamada, A.; Takeda H.; Kinoshita-Kikuta, E.; Koike, T. Journal of Separation Science 2005, 28, 155-62. 99. Brittain, S. M.; Ficarro, S. B.; Brock, A.; Peters, E. C. Nature Biotechnology 2005, 23, 463-68. 100. Larsen, B.; McEwen, C. Mass Spectrometry of Biological Materials. 1998. CRC Press. 101. Barnett, D. A.; Ells, B.; Guevremont, R. Journal of the American Society for Mass Spectrometry 2000, 11, 1125-33. 102. Cui, M.; Ding, L. Y.; Mester, Z. Analytical Chemistry 2003, 75 5847-5853. 103. Eiceman, G. A.; Nazarov, E. G.; Mill er, R. A.; Krylov, E. V.; Zapata, A. M. Analyst 2002, 127, 466-71. 104. Shvartsburg, A. A.; Bryskiewicz, T.; Purves, R. W.; Tang, K. Q.; Guevremont, R.; Smith, R. D. Journal of Physical Chemistry B 2006, 110, 21966-21980. 105. Guevremont, R.; Purves, R. Journal of the American Society for Mass Spectrometry 2005, 16, 349-62. 106. San Diego Plastics, Inc. National City, CA, 2008. www.sdplastics.com/kelf.html Accessed 6/24/07

PAGE 152

152 BIOGRAPHICAL SKETCH Jennifer Alisia Garrett was born in 1982, in Ja cksonville, FL, to Gary and Alison Garrett. She attended Stanton College Preparatory Sc hool where she received an international baccalaureate diploma. She attended the University of North Florida on full scholarship from 2001-2004, where she received a bachelors degree in chemistry with a minor in biology. Her undergraduate research advisor was Dr. Stuart Chalk, who greatly influenced her interest in research. From 2003-2004, she interned at Mayo Clinic in Jacksonville, FL, under the direction of Dr. Rick Troendle and Dr. Chris Eckman, where she discovered her interest in mass spectrometry. In 2004, Jennifer joined the Yost Lab at the University of Florida to pursue a Ph.D. in analytical chemistry. In 2007, she married Gary T.J. Bryant on St. Augustine beach with her closest friends and family in attend ance. She received her PhD in August 2008 under the direction of Dr. Rick Yost.