|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 FOURIER TRANSFORM ION CYCLOTRON RESONANCE MASS SPECTROMETRY OF BIOLOGICAL NONCOVALENT COMPLEXES By MICHELLE MARGARET SWEENEY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLM ENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2009
2 2009 Michelle Margaret Sweeney
3 To Mom, Dad, and Ashley
4 ACKNOWLEDGMENTS My success in graduate school is the result of the attention and aid I have received from a number of people throughout the course of my life. My family has been there for me since I was born and deserve s more than praise for all they have done for me. I wish to thank my parents, who encoura ged me to achieve all that I could, and without whose support I never could have gotten this far. My sister Ashley, my best friend, who always had a funny comment for me in my time of need, kept me from going crazy at times. These three people are the most important to me in the world. What scientific progress I have made has been the result of the attention of my advisor, Dr. John Eyler. He graciously accepted me into his group and gave me advice and support in my scientific endeavors. He also provided me with the opportunity to travel for conferences and resea rch, which has been invaluable to my development as a scientist. My time at the University of Florida was facilitated by the award of an Alumni Fel lowship, for which I thank the D epartment of Chemistr y. I was also financially supported by the National Science Foundation through a Partnership for International Research and Education (PIRE) grant administered by Dr. Mary Rodgers at Wayne State University. Thanks to t his funding I travelled to the Netherlands to work at the FO M Institute for Plasma Physics Rijnhuizen My work there was greatly aided by Drs. Jos Oomens, Jeff Steill, and Brita Redlich. I wish to thank them and Dr. Rodgers for the opportunity. Lastly, I wish to thank Dr. Steven Benner at the Foundation for Applied Molecular Evolution, who provided some library samples and early guidance for the projects described in this dissertation I also wish to recognize certain of my labmates who m I now consider friends first, and coworkers second. Fir st, I wish to acknowledge a former Eyler Lab student, Dr. Kellie Woodling, who got me started working in the lab and gave me advice on surviving graduate
5 school in general. She was always ready to talk college basketball or gossip with me. Next, I wish to thank Julia Rummel, who entered the University of Florida at the same time I did. Surviving classwork, cumes, and oral exams was easier for me because we could study together Lastly, I wish to thank Sarah Stefan, who is always around lab when I need an extra pair of hands or to vent. She was also the first student I trained, and in my doing so, I learned a lot more than I thought I could. Outside of the lab, I need also to thank my friend Rachel Giessert, who encouraged me to apply to the University of Florida graduate school and who helped me feel at home during my first year here. My friend Monique Williams was always willing to eat sushi and discuss the Harry Potter books with me. She took me to and from the airport more times than I can count.
6 TABLE O F CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES ...........................................................................................................................9 LIST OF FIGURES .......................................................................................................................10 ABSTRACT ...................................................................................................................................14 CHAPTER 1 NONCOVALENT CHEMISTRY AND TARGET ASSISTED COMBINATORIAL SYNTHESIS ...........................................................................................................................16 Noncovalent Chemistry ..........................................................................................................16 Electrostatic Binding .......................................................................................................18 Hydrogen Bonding ..........................................................................................................18 Pi Interactions ..................................................................................................................19 Hydrophobic Interactions ................................................................................................20 Significance of Noncovalent Interactions in Biological Systems ...................................20 Mass Spectrometry of Biological Noncovalent Interactions ..................................................21 Challenges .......................................................................................................................21 Literature Survey of Mass Spectrometric Analysis of Biological Noncovalent Interactions ...................................................................................................................22 Review articles .........................................................................................................22 Ionization te chniques ................................................................................................23 Noncovalent interactions in protein structure ..........................................................24 Noncovalent complexes of proteins .........................................................................25 Noncovalent interactions with libraries ....................................................................26 Binding strength determination ................................................................................28 Target Assisted C ombinatorial Synthesis ...............................................................................31 TACS Systems Analyzed by NonMass Spectrometric Methods ...................................34 Conventional Combinatorial Systems Analyzed by Mass Spectrometry ........................34 Mass Spectrometric Analysis of TACS Systems ............................................................35 2 INSTRUMENTATION ..........................................................................................................43 Electrospray Ionization ...........................................................................................................43 History .............................................................................................................................43 Process and Mechanism ..................................................................................................44 Related Ionization Methods .............................................................................................45 Advantages, Disadvantages, and Figures of Merit ..........................................................46 Fourier Transform I on Cyclotron Resonance Mass Spectrometry .........................................48 History .............................................................................................................................48 Principles .........................................................................................................................49
7 Cyclotron motion ......................................................................................................49 Trapping motion .......................................................................................................50 Magnetron motion ....................................................................................................51 Excitation, isolation, and detection ..........................................................................52 Instrumentation ................................................................................................................54 Magnets ....................................................................................................................54 ICR cells ...................................................................................................................55 Vacuum pumps .........................................................................................................55 Ion optics ..................................................................................................................56 Electronics and computers .......................................................................................56 Advantages, Disadvantages, and Figures of Merit ..........................................................56 Mass resolution and resolving power .......................................................................56 Mass accuracy ..........................................................................................................58 Mass range ................................................................................................................58 Linear operating range .............................................................................................58 Improving FTICR MS performance ........................................................................59 Advantages and disadvantages .................................................................................59 Dissociation Techniques .........................................................................................................60 Collision Induced Dissociation (CID) .............................................................................60 Infrared Multiple Photon Dissociation (IRMPD) ............................................................61 Experimental Instrumentation ................................................................................................63 3 STUDIES OF RIBONUCLEASE A AND URIDINE 5 MONOPHOSPHATE ..................77 Nucleic Acid Nomenclature ...................................................................................................77 Systems of Interest ..................................................................................................................78 Optimizing Conditions for RNase A Solutions ......................................................................79 Optimizing Conditions for RNase A UMP Complexes .........................................................83 In Source CID of Complexes .................................................................................................85 Fixed Wavelength IRMPD of Complexes ..............................................................................87 IRMPD of UMP Clusters ................................................................................................87 IRMPD of RNase AUMP Clusters ................................................................................89 4 MULTIPLE WAVELENGTH IRMPD FOR COMPARISON OF SPECIFIC AND NONSPECIFIC NONCOVALENT INTERACTIONS .......................................................109 Systems of Interest ................................................................................................................109 Calculations and Measured Parameters ................................................................................110 Comparisons of Spectra ........................................................................................................110 5 STUDIES OF RIBONUCLEASE A AND TWO SMALL DYNAMIC COMBINATORIAL LIBRARIES .......................................................................................124 Systems of Interest ................................................................................................................124 Microdialysis Method for Reduced Signal Suppression ......................................................124 Studies with Library 1 ..........................................................................................................128 Studies with Library 2 ..........................................................................................................129
8 6 CONCLUSIONS AND FUTURE DIRECTIONS ...............................................................157 LIST OF REFERENCES .............................................................................................................160 BIOGRAPHICAL SKETCH .......................................................................................................167
9 LIST OF TABLES Table page 41 Comparison of wavelengths for complex dissociation ....................................................123 51 ........................................137 52 ...............141 53 Fragments from IRMPD of RNase A .........................147
10 LIST OF FIGURES Figure page 11 Examples of noncovalent complexes .................................................................................38 12 Hydrogen bonding between nucleotide base pairs .............................................................39 13 Target assisted combinatorial chemistry (TACS) scheme ................................................40 14 Comparison of free energy profiles for thermodynamic and kinetic control .....................41 15 Redundancy of TACS libraries ..........................................................................................42 21 Schematic representation of the ESI process. ....................................................................65 22 Motion of a charged particle in a magnetic field ...............................................................66 23 Ion path resulting from cyclotron, magnetron, and trapping motions ...............................66 24 Ion motion following the application of an excitation voltage ..........................................67 25 Relationship between transient and mass spectrum. ..........................................................67 26 Typical experimental sequence for F ourier transform ion cyclotron resonance mass spectrometry (F TICR MS ) ................................................................................................68 27 Variations of experimental sequences for FTICR MS .......................................................68 28 Simple ICR cell ..................................................................................................................69 29 Variations of the ICR cell ..................................................................................................70 210 FTICR MS performance improves with increasing magnetic field strength. ....................71 211 In source collision induced dissociation (CID) scheme ....................................................72 212 Infrared multiple photon dissociation scheme ...................................................................73 213 Schematic of FTICR mass spectrometer ............................................................................74 214 Layout of the CO2 laser and FTICR mass spectrometer used for experiments in Chapter Three .....................................................................................................................75 215 Layout of the CO2 laser and FTICR mass spectrometer used for experiments in Chapters Four and Five ......................................................................................................76 31 Numbering system for nucleic acid rings ..........................................................................91
11 32 Nucleic acid nomenclature. ................................................................................................92 33 Sequence of bovine pancreatic Ribonuclease A (RNase A ) ..............................................92 34 Mechanism for RNase A catalyzed hydr olysis of ribonucleic acid ...................................93 35 Structure of uridine 5 monophosphate (UMP). ................................................................94 36 Effect of organic phase on average charge state for RNase A. ..........................................94 37 Effect of increasing acid modifier on charge state .............................................................95 38 Effect of RNase A concentration on signal abundance .....................................................95 39 Effect of RNase A concentration on average charge state .................................................96 310 Effect of increasing HMC temperature on RNase A abundance .......................................96 311 Effect of HMC temperature on RNase A average charge state .........................................97 312 Effect of HMC temperature on RNase A charge states .....................................................97 313 Mass spectrum obtained when 50 [UMP] = 1 [RNase A] .................................................98 314 Mass spectrum obtained when 10 [UMP] = [RNase A] ....................................................99 315 Mass spectrum obtained when 5 [UMP] = [RNase A] ....................................................100 316 Mass spectra for in source CID of UMP clusters ............................................................101 317 Abundance of UMP clusters as a function of voltage difference em ployed for insource CID .......................................................................................................................102 318 Relationship between cluster size and stability (CID) .....................................................102 319 In source CID of RNase A UMP complexes ...................................................................103 320 Mass spectra of UMP clusters at increasing irradiation times .........................................104 321 Abundance of UMP clusters as a function of IRMPD irradiation time ...........................105 322 Relationship between cluster size and stability (IRMPD) ...............................................105 323 Stability of most abundant IRMPD fragments .................................................................106 324 Proposed fragmentation pathway for m/z 325 .................................................................106 325 Proposed fragmentation pathway of m/z 347 ...................................................................107 326 Mass spectra for IR MPD of RNase A UMP complexes .................................................108
12 41 Structures of Library 1 components .................................................................................114 42 Sequence of equine heart myoglobin ...............................................................................115 43 Calculating percent complexation for a RNase A Library 1 mass spectrum ...................115 44 Mass spectrum showing complexation between myoglobin a nd Library 1 .....................116 45 Mass spectrum showing complexation between RNase A and Library 1 .......................117 46 Percent complexation versus irradi ........................118 47 ........................119 48 Percent co .......................120 49 ..........................121 410 ..........................122 51 Structures of Library 2 components .................................................................................132 52 Mass spectrum obtained when RNase A and Library 1 concentrations are roughly equal .................................................................................................................................133 53 Mass spectrum obtained when Library 1 concentration is 100 times greater than that of RNase A .......................................................................................................................134 54 Setup for microdialysis procedure. ..................................................................................134 55 Comparison of free library and RNase A Library 1 complex abundances for microdi alysis procedure. ..................................................................................................135 56 Mass spectrum of Library 1 in the absence of RNase A ..................................................136 57 Putative structures for fragment assignments made in Table 51. ...................................138 58 Ribonuclease A Library 1 complexation in the +8 charge state ......................................139 59 IRMPD fragments from RNase A ..............................140 510 Putative structure assignments for peaks listed in Table 5 2 (Structures 1 11b). ............142 511 Putative structure assignments for peaks listed in Table 5 2 (Structures 11c 14d). ........143 512 Putative structure assignments for peaks observed in Table 5 2 (Structures 14e 18a). ..144 513 Putative structure assignments for peaks observed in Table 5 2 (Structures 18b18d). ..145
13 514 Complexation between RNase A and Li br ary 2 following two rounds of microdialysis. ...................................................................................................................145 515 Incomplete dissociation of RNase A Library 2 complexes with IRMPD .......................146 516 Putative structure assignments for peaks observed in Table 5 3 (Structures 1 10c). ......148 517 Putative structure assignments for peaks observed in Table 5 3 (Structures 10d13c). ..149 518 Putative structure assignments for peaks observed in Table 5 3 (Structures 13d15h). ..150 519 Putative structure assignments for peak s observed in Table 5 3 (Structures 15i 18a). ...151 520 Putative structure assignments for peaks observed in Table 5 3 (Structures 18b19f). ...152 521 Putative structure assignments for peaks observed in Table 5 3 (Structures 20a 21f). ...153 522 Putative structure assignments for peaks observe d in Table 5 3 (Structures 21g22b). ..154 523 Abundances of selected species as a function of irradiation time for IRMPD of RNase A Library 2 complexes. ........................................................................................155 524 Effect of increasing IRMPD energy on average charge state of RNase A Library 2 complexes. .......................................................................................................................155 525 Effect of increasing IRMPD energy on various charge states for RNase A Library 2 complexes. .......................................................................................................................156
14 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy FOURIER TRANSFORM ION CYCLOT RON RESONANCE MASS SPECTOMETRY OF BIOLOGICAL NONCOVALENT COMPLEXES By Michelle Margaret Sweeney May 2009 Chair: John R. Eyler Major: Chemistry Noncovalent chemistry is the basis of many important biological interactions, including enzyme ligand complex formation. As these interactions are typically weaker than covalent bonds, special analysis methods are needed for studying noncovalent complexes. Using gentle ionization methods like electrospray ionization (ESI), noncovalent complexes can be preserved a nd analyzed by mass spectrometry. Fourier transform ion cyclotron resonance mass spectrometry (ESI FTICR MS) provides superior mass accuracy, resolving power, and tandem in time capabilities. Among the dissociation techniques available are insource collis ioninduced dissociation and infrared multiple photon dissociation (IRMPD). Dynamic combinatorial chemistry uses a special type of library wherein reversible binding between library members and between library members and a target molecule expands the potential number of strong complex interactions Here several 2 3 cyclic monophosphate nucleotides were incubated with Ribonuclease A to generate enzyme ligand complexes. Enzymatic activity may drive the library members to form RNA chains via phosphodies ter bond generation. This dissertation shows the method development for screening a cyclic nucleotide based dynamic combinatorial library for tight binding ligands of Ribonuclease A using
15 ESI FTICR MS. A model system comprised of Ribonuclease A and a known inhibitor ( uridine 5 monophosphate ) was used to optimize mass spectrometry parameters. The resultant noncovalent enzyme inhibitor and inhibitorinhibitor complexes were studied by IRMPD. Tandem MS was demonstrated for complex dissociation and inhi bitor fragmentation in a single experiment. Next, the use of variable wavelength dissociation of noncovalent complexes was employed to optimize dissociation efficiency. S pecifically bound Ribonuclease A Library complexes were compared to nonspecific contro l complexes. Differences were observed at two of the wavelengths employed, with specific complexes requiring greater energy for complete dissociation th an nonspecific complexes. Lastly two small dynamic combinatorial libraries were screened for binding a ffinity to Ribonuclease A and certain of the binding species identified. A microdialysis method was developed to reduce signal suppression due to the high library concentrations required for proper library functioning. The screening method efficacy was li mited by exact mass redundancy between the species in the libraries. Additional dissociation steps were unsuccessful at generating sufficient signal to discriminate between the structures with identical masses. Future work should be devoted to expanding li brary sizes and developing methods to use higher library concentrations.
16 CHAPTER 1 NONCOVALENT CHEMISTRY AND TARGET ASSISTED COMBINATORI AL SYNTHESIS Noncovalent Chemistry Noncovalent bonding is described in terms of what it is not a noncovalent bond is not a covalent bond. This vague terminology means that noncovalent bonding encompasses a myriad of different interactions. For example, the hydrogen bonding network between water molecules is noncovalent, as are the much weaker interactions between benzene molecul es. Compare also a crown ether selectively binding a potassium cation, a cyclodextrin binding a cholesterol molecule, and an enzyme binding its substrate. Noncovalent bonds hold together the double helical structure of deoxyribonucleic acid ( DNA) a nd the four subunits of hemoglobin. Figure 11 shows some examples of complexes held together by noncovalent interactions. Noncovalent complex nomenclature varies. The term noncovalent complex refers to two or more species exclusively bound together by noncovalent bonds. One nomenclature system describes the complex as a host guest interaction. Here, the host possess es Lewis base donor sites and the guest possesses the Lewis base acceptor sites.1 Host guest nomenclature is typically used for small mol ecule and synthetic systems and t his dissertation will consequently not use th e host guest nomenclature and instead employ several alternate nomenclatures. When referring to noncovalent complexes generally, receptors and ligands will comprise the complex. In enzymatic systems, binding occurs between an enzyme and a ligand, between an enzyme and a substrate, or between an enzyme and an inhibitor. A substrate is a species on which the enzyme carries out its catalytic function. An inhibitor binds to the enzym e and retards catalysis. The term ligand encompasses both inhibitors and substrates. In combinatorial systems, the terms target and ligand will be used in place of receptor ligand and enzyme ligand nomenclature. Candidates or
17 candidate ligands describe mol ecules which may bind the target, but have yet been shown to do so. Jean Marie Lehn coined the term supramolecular chemistry to describe the chemistry of noncovalent interactions, and this term has since been used to refer to the man made equivalent of noncovalent interactions in nature.2 4 Much early work in supramolecular chemistry focused on the interactions of macropolycycles.4 The relative simplicity of these systems (in compar ison to most biomolecular systems) allows easier exploration of fundamental interactions. Among the principles studied are catalysis, molecular recognition, and molecular self assembly and self replication, all of which have their counterparts in natural biological systems .5 Catalysis is the process whereby the rate of chemical reaction is increased by the presence of a net unconsumed species (catalyst). There are many distinct modes of catalysis, with some catalysts actively participating in the reaction while others stabilize transition states or bring reagents into close proximi ty. Large catalysts, including most enzymes, may operate by several of these modes. For example, the enzyme ribonuclease A catalyzes the hydrolytic cleavage of single stranded RNA.6 Certain of the enzymes residues act as acids or bases to drive the reacti on; other residues recruit the strand into the active site and stabilize the transition state. Molecular recognition describes a specific pattern of binding governed by high complementarity between the molecules involved. Here, a good example is the highly specific recognition between antigen and antibody. In molecular self assembly, the molecules involved arrange themselves in a particular configuration without external regulation. Lipid bilayer f ormation in cells is an example of self assembly. There are a few more terms encountered in noncovalent chemistry which should be clearly defined. Binding strength is often related to the kinetics of the receptor li gand binding. It may be mathematical ly descri bed as an association constant or dissociation consta nt. When the binding
18 is strong, the complex is slow to break apart. Complementarity between receptor and ligand occurs when the properties of a site on the ligand favor binding at a particular site on the receptor. These properties may relate to size, shap e, or local chemical environment. For example, a positively charged amino acid residue on a protein will be attracted to a negatively charged ligand group nearby. Binding s pecificity describes the degree of complementarity between species; the more complementary sites for a system, the more specific the binding. Electrostatic Binding Electrostatic interactions, based on the familiar principle of opposite charges attracting, number among the strongest noncovalent interactions. Specie charge can be in the form of an ion, dipole, or induced dipole. One also observe s repulsive forces between like charges. Where the charge is static, the interaction will be stronger. D istance between charges also plays a role in the association. One other feature of the electros tatic interaction is that the binding is directionally dependent. Ionion bonds are the strongest subset of electrostatic noncovalent interactions, with typical in bo nd energies of ~100350 kJ mol1.7 Iondipole interactions, also quite strong, have bond energies ~50 200 kJ mol1.8 Lastly, dipole dipole interactions may have strengths of ~550 kJ mol1.8 It should be noted that there is a wide range of bond strengths within each of these categories, dependent on the species and distances involved. Hydroge n Bonding The significance and strength of hydrogen bonding has long been recognized due to its contribution to the properties of water. Indeed the term hydrogen bond was first used to describe the interactions between water molecules. Now it describes a b roader group of interactions as a special class of electrostatic interactions.
19 Perhaps the best working definition of a hydrogen bond was provided by Pimentel and McClellan.9 They describe binding between a functional group, A H, and a separate atom or gr oup of atoms, B. Both A and B are electronegative, and a new noncovalent link is formed between B and H, while A H is maintained. In this case, A can be referred to as the donor and B as the acceptor. C ommon donors are C H, N H, O H. Other groups which may also act as donors are S H, P H, F H, Cl H, Br H, and I H Acceptors include N,O, P, S, F, Cl, Br, H ydrogen bonding strength varies greatly, with a strong dependence on the electronegativity of A and B. These interactions are among the strongest found in noncovalent chemistry. For neutral species, the energy of hydrogen bonding is on the order of 5 to 60 kJ mol1.8 However, many hydrogen bonds may be formed between molecules with the streng th of binding being additive. Water is an excellent example, forming an intricate network of hydrogen bonds, acting as both donor and acceptor (Figure 11). A nother good illustration of hydrogen bonding can be seen in WatsonCrick base pairing of DNA and RNA nucleobases, shown in Figure 12. Bases on the sense strand hydrogen bond with bases on the antisense strand to form the classical double helix.10 Thymine (T) and adenine (A) pairs form two hydrogen bonds. Cytidine (C) and guanine (G) form three. In RNA gua nine is replaced by uridine (U). Pi Interactions Aromatic ring interactions can represent another category of electrostatic binding. Around the ring, there is a partialabove and below the plane of the ring are the L ocalized regions of charge can interact with other electrostatic charges or undergo stacking interactions with other aromatic rings. These forces can have binding energies ~0100 kJ mol 1.1 1 One example of this is the stabilizing stacking interactions
20 in DNA. Double strands with favorable stacking interactions have higher melting points than double strands with weaker stacking interactions.12 This demonstrates that noncovalent interactions play a role in the physical properties of many macromolecules. Hydrophobic Interactions The most relevant hydrophobic interactions occur for large biomolecules in the presence of polar media. By bringing hydrophobic r egions of two or more species into close proximity to the exclusion of polar solvent, solvation energy effects can be minimized. The biological importance of hydrophobic noncovalent binding can be seen in two key examples. First, the removal of hydrophobic amino acid residues from the surface to the interior is one of the driving forces in protein folding; this has sometimes been called the hydrophobic effect.1 3 Second, lipids and fatty acids congregate and exclude charged molecules and solvent in the forma tion of cellular membranes and micelles. Significance of Noncovalent Interactions in Biological Systems The biochemical world is rife with examples of important noncovalent interactions. Of especial importance is the principle of molecular recognition. In molecular recognition, two or more molecules have a high degree of complementarity with regard to local steric and chemical environment.1,5 When more than two molecules are involved, cooperative binding may occur. The more points of agreement with rega rd to sites of interaction, the stronger and more specific the binding will be. Proteins use this chemistry in several ways. Folding and stabilization are accomplished via iontworks.12 Furthermore, amino acid residues are capable of association with solvents, salts, and other small molecules via nonspecific noncovalent binding. Enzyme recognition of substrates and inhibitors is governed by those same sorts of binding. The activ e site of an enzyme provides a specific
21 chemical and steric environment for substrate or inhibitor. Metal ions and other small molecules may be recruited as cofactors. Mass Spectrometry of Biological Noncovalent Interactions Mass spectrometry (MS) is an an alytical method whereby the massto charge ratio ( m/z ) of a gas phase ion is determined. Analysis can be broken down into two steps; first, gas phase ions of the analyte are generated and second, these ions are detected. The details of these processes are further described in Chapter Two. Speed, selectivity, and sensitivity are advantages of mass spectrometry.1 4 Challenges The analysis of noncovalent complexes poses several challenges to mass spectrometry T he choice of ionization methods can affect the preservation of weak interactions. There exists the possibility of false positives and false negatives with regards to the complexes observed in the mass spectrum. Gentle ionization techniques are required to maintain weaklyassociated complexes. If too much energy is imparted via ionization, t he complex may be dissociated before mass analysis. However, these gentle techniques can maintain some interactions that are not of interest. Observing salt or solvent adduct peaks is a common example of this Another issue of concern is how reflective gas phase ions are of solution phase structure. Consider that in solution, there are many interactions between solvent and analyte. In the gas phase, analyte ions are naked, as solvent is intentionally lost prior to mass analysis. Complexes relying heavily on solvent interactions may be distorted when analyzed by MS. In short, interactions between charges, dipoles, and polarizability become more favorable when the interacting species are transferred to the gas phase, whil e hydrophobic interactions are reduced in importance.1 5 Despite the fact that solution phase and gas phase structures are nonidentical, there
22 is abundant evidence to suggest these structures are related. Some of this evidence will be covered in the literat ure section below. In practice, another issue encountered is that of sensitivity. Analytes may be present at low concentrations and samples or sample matrices may be complex. In the mass spectrum, the presence of a particular specie may be amplified or reduced in comparison to solution phase concentrations, depending on the other species present.1 6 In biological samples, the pr esence of salts and buffers may exacerbate the sensitivity problem. Some of these issues may be alleviated by purification prior to mass spectrometric analysis; however, these purification steps must be compatible with the preservation of the desired noncovalent interactions. Literature Survey of Mass Spectrometric Analysis of Biological Noncovalent Interactions Review articles While not focused on biomolecular complexes, a review article by Schalley discusses many of the issues concerned with mass spectrometric analysis of noncovalent species.2 This is a good source to review fundamentals of interaction, as well as demonstrating the scope of supramolecular chemistry. Crown ether interactions, macrocycles, etc are thoroughly covered. A number of good review articles have been published in the area of mass spectrometric analysis of biomolecular noncovalent interactions. To begin, the review by Smith et al from 1997 covers some theoretical as well as practical considerations of ESI MS in this field.1 7 A later article by Loo provides an historical perspective of the complexes studied.1 8 Di Tullio et al ., covered the more specific topic of molecular recognition studied by mass spectrometry.1 9 For those noncovalent interactions specific to protein structure, Kaltashov and Eyles have published a thorough review.20 Lastly, larger protein complexes were covered in a minireview by Hernandez an d Robinson.21
23 Ionization t echniques Studying noncovalent interactions using mass spectrometry is a relatively recent development and has to this point been limited by the available ionization techniques. The earliest ionization methods used in MS, such as electron i onization ( EI ) and chemical ionization ( CI ) are generally too energetic for analysis of noncovalent compounds. Weak interactions are not preserved. Certain stronger noncovalent interactions, such as those of some metal adducts with small molecul es, are strong enough that they could be studied by EI and CI, and thus, these are the earliest examples of noncovalent interactions in mass spectrometry.2 2 As a further limitation, the upper mass range of analytes for EI and CI is sufficiently low to excl ude most biomolecules. In the early 1990s, two significant new ionization techniques were introduced electrospray ionization (ESI) and matrix assisted laser desorption ionization (MALDI). Ionization by ESI and MALDI can be described as gentle, meaning t hat weak interactions are preserved from solution to gas phase. Noncovalent interactions are among those weak interactions that can be observed. At present, MALDI is considered harsher than ESI, and its use for analysis of noncovalent interactions has been limited to those stronger noncovalent interactions. John Fenn was the first to publish on the preservati on of noncovalent interactions in ESI.2 3 E arly work showed that protein structure, which involves several types of noncovalent interactions, can be lar gely maintained during the ESI process ; this will be discussed later MALDI analyses were demonstrated by Karas and Hillenkamp 24 and by Tanaka et al .25 A number of ESI based techniques have been shown to be even gentler than conventional ESI. The first demonstrated among these was nanoelectrospray or nanospray ionization (nESI).2 6 Operating under lower flow conditions, nESI produces smaller initial droplets than ESI. This is the basis of many of its advantages such as improved sensitivity and higher salt tolerance.2 7 The
24 next ionization breakthroughs were made in the 2000s, beginning with the Cooks group. These techniques are electrosonic spray ionization (ESSI) and desorption electrospray ionization (DESI). Their work, and that of others, has shown that proteins analyzed by ESSI and DESI maintain more of their solution phase character than those analyzed by ESI. 2 8 30 Noncovalent interactions in protein structure The Fenn group was the first to successfully ionize proteins by ESI.2 3 As mentioned previousl y, it is presently believed that the mass spectrum of a protein, particularly the charge state distribution, largely reflects the solution phase conformations. Protein structure is important to maintain when studying specific interactions. Unfolded proteins may lose active sites present in the natively folded structure. Additional regions of the protein may be revealed as potential binding sites. Chowdhur y et al ., working with cytochrome c, were the first to demonstrate that changes in the solution conditions in this case, changes in pH resulted in shifts of the charge state distribution of the mass spectrum.3 1 Loo et al showed a similar effect by changing the organic solvents used. Furthermore, they demonstrated that disulfide bond reduction altered the pa ttern of charging in the mass spectrum; i.e., that changes in charges shown in the mass spectrum were indicative of protonation sites available in solution.3 2 Solution phase structures correspond to gas phase structures. It has also been shown that the pr esence of multiple, intact disulfide bonds increases the preservation of natively folded protein conformation structure into the gas phase. Recall also that hydrophobic interactions are reduced in importance when transferred into the gas phase; those struc tures highly reliant on hydrophobic interactions are more likely to lose their native conformations.20 There exist s an energy barrier to unfolding. However, when this barrier is low, such as for hydrophobic interaction reliant, non disulfide containing spe cies, structures are likely to have sufficient energy imparted during the ionization process to unfold.
25 At present, mass spectrometry can discriminate between native state proteins and those denatured by pH 3 1 solvent conditions 3 2 and temperature. 3 3 3 4 Guevremont et al showed that the solution phase conformations a re the most important factor in determining the protein charging patterns in ESI MS.3 5 The quaternary structure of a protein can be described as the interaction between its various tertiary subunits. Frequently these interactions are noncovalent in nature. In 1994, the Smith group showed that these noncovalent interactions can be maintained through ESI MS.3 6 They studied three proteins concanavalin A, avidin, and hemoglobinand found that gas phase stability constants agreed well with the corresponding solution phase stability constants. Furthermore, they found that the ESI MS interface conditions affected the structures present, with gentle conditions producing species observed at physiologic al conditions and harsher conditions giving rise to species not observed at physiological conditions Noncovalent c omplexes of p roteins One of the first noncovalent complexes studied by ESI MS was the enzymesubstrate complex comprised of hen egg white l ysozyme (HEWL) and N acetylglucosamine (NAG). I n 1991, Ganem et al generated and observed both enzyme substrate and enzyme product complexes.37 Another key early work was performed by Huang et al ., wh o analyzed the proteinprotein intera ctions of interfe ron by ESI MS.38 T he relatively low energy needed to dissociate this complex was attributed to the electrostatic repulsion between identically charged subunits. They also looked at the protein ligand interaction between H ras and guanosine diphosphate, and showed that more energy was required to dissociate this interaction in comparison to the proteinprotein interaction.3 8 P resent ly ESI MS has been demonstrated for protein protein, protein carbohydrate, proteinnucleic acid, and various other prot ein small molecule noncovalent complexes.
26 Analysis of noncovalent complexes by ESI MS has been extended to greater complex sizes. One recent innovation has been the coupling of ion mobility spectrometry to MS to obtain information on both size and shape of these large species.3 9 The Robinson group has done much work in this field, analyzing complexes with sizes in excess of 800,000 Da. For example, they successfully generated a mass spectrum of the intact GroEL complex (m/z ~803,000 Da) using nESI MS.40 Th e interactions of this complex were sufficiently strong that they could not be dissociated via in source collision induced dissociation. Considerable early work in this field was devoted to supporting the idea that gas phase structures correlate to solution phase structures. One piece of evidence, shown initially by Veenstra, is that the specificity and binding stoichiometry of these gas phase complexes agrees well with the results obtained by solution phase spectroscopy.4 1 A recent comparison of ESI, elec trosonic spray ionization ( ESSI ) and nESI by the Zenobi group found that gas phase dissociation constants for lysozyme and its inhibitor agreed well with solution phase dissociation constants.42 Of the three methods, they found ESSI to have the best agree ment. They argued that this agreement was due to similarity between gas and solution phase structures. Noncovalent i nteractions with l ibraries Libraries are collections of molecules sharing similar structures, chemistries, and/or origins. In drug discovery, libraries are of particular importance. Small changes to similar structures (scaffolding) can provide many candidates for binding to a target. Analyzing a library to find molecules with particular properties is called screening. As a common example, one might screen a library for binding affinity to a particular enzyme. Often, a library has many molecules with similar structure where the shared features have designed complementarity to a target molecule. Those members of the library which have the highes t complementarity to the target generate the complexes with the highest binding affinities.
27 Library screening methods seek to isolate and identify those complexes. Libraries may be derived from natural products 4 3 4 4 or may be synthetic in nature.4 5 4 7 Ad ditionally, combinatorial techniques can be used to expand the number of library members ; this will be described in a later section. In 1996, the Smith group published an example of an MS based library analysis.4 5 4 6 W ork was done using bovine carbonic anhydrase II (BCAII), an enzyme frequently studied by ESI MS. Two sets of inhibitors were used, with the solution phase binding constants known for each inhibitor. Abundance of enzyme inhibitor complexes in the mass spectra directly correlated to the strengt h of the solution binding constants. Furthermore, they were able to observe competitive and noncompetitive binding modes. They asserted that their method worked best for libraries where the binding constants varied widely a nd where each library member had a unique exact mass. The Siuzdak group was able to demonstrate a quantitative ESI MS screening method in 1997.47 A library of nucleoside mono, di and tri phosphates were screened against galactosyltransferase. This method differed slightly from the pre vious ones in that the inhibitors were not pooled in the screen during quantitation. However, proof of concept with pooled inhibitors was demonstrated. Species other than proteins and peptides may also be used in noncovalent interaction library screens. One such example was shown by Griffey et al ., who screened a library of aminoglycoside antibiotics against multiple RNA targets.4 8 This multiplexing, which they termed multitarget affinity specificity screening (MASS), greatly increased the number of potential ligand target matches. They were able to show selective binding for certain complexes, as well
28 as to determine the binding constants. This same technique was also used to screen a natural products library against the same targets.4 4 Binding s trength d et ermination For a thorough treatment of the subject, the Zenobi group has published an excellent review article.49 Studies of noncovalent complexes by ESI MS have expanded to include determination of the binding strength between ligand and target. The descr iption of these associations may be qualitative or quantitative, depending on the methods used. In qualitative methods, binding strengths of ligands are ranked against each other. These methods have the advantage of simplicity. It was recognized that for solutions comprised of many ligands and a target, there was a correlation between the abundance of a particular complex and the binding strength of the specific ligand to the target. The more tightly bound the complex, the more abundant it will appear in the mass spectrum. Using this, it is possible to generate a list of relative binding strengths for a particular set of compounds. This method was demonstrated by Wu et al .50 For quantitative methods, there are two general classes of experiments. Solution ba sed methods with mass spectrometric detection comprise the first. Titration methods vary the concentration of the ligand(s) and monitor the changes in free and bound target. The results of these experiments are frequently reported as association constants, Kas, or as dissociation constants Kds. For a single ligand, single target system, the relationship between Ka and Kd is shown in Equation 1 1. In this description, [L] is the concentration of ligand, [T] is the concentration of target, and [LT] is the con centration of the ligandtarget complex. 1 a dK LT T L K (1 1)
29 Competition experiments are similar to titration experiments. The procedure whereby ligand concentration is varied is the same between the types of experiments; the difference is that multiple ligand, single target systems are used in the competition experiments. These calculations are more involved, but tend to give more accurate values. In a few solution based experiments, the concentration of the ligands is held constant while the target concentration is varied ; 5 1 however, this is not the preferred method. It is important to note that these measures of the binding strength are reflective of the solution phase binding. Mass spectrometry is one way to monitor the amount of fr ee and bound target; however, other analytical techniques like isothermal titration calorimetry (ITC), 5 2 5 3 circular dichroism (CD), 5 3 and spectroscopy5 3 can also be used to determine binding. Since all these methods reflect solution phase binding, compa risons between different methods is possible. In the second type of quantitative experiments, gas phase complexes are dissociated by the input of energy, and the abundances of free target and free ligand are compared to the abundance of the complexes. The energy for dissociation may come from collisions with background gas molecules (CID) 5 4 5 5 from background infrared photons (BIRD) 5 6 5 7 or from infrared or ultraviolet radiation (photodissociation) 5 7 among the many methods available. The information obtained from these experiments is reported by parameters like CV50%, a voltage at which 50% of the complex has dissociated. The values reflect gas phase binding strength, and do not necessarily correspond to solution phase constants. An example of a tit ration based method was reported by Heck and coworkers, who determined solution phase Ka values for glycopeptide antibiotics and several peptide ligands using mass spectrometry as a detection method.5 8 Their method for Ka determination employed competition experiments. In this setup, three peptide ligands at equimolar concentrations are
30 exposed to the target antibiotic. From the mass spectrum, the peak areas of the complexes and of the free target are measured. The process is repeated for different ligand c oncentrations. Making the assumption that response factors for the free and bound target are the same, solution phase Ka values can be calculated. These values were compared to those obtained by non mass spectrometric methods, and there was good agreement. Kempen and Brodbelt reported a novel method for determining Ka.5 9 Their method was a solution based competition method with ESI MS detection. What distinguished this method was the use of a reference complex with a known association constant. This reference complex has the same target or ligand as the complex of interest, and thus will compete with the complex of interest for either target or ligand. Before the competing complex is added, a calibration curve of reference complex intensity versus reference complex intensity can be generated. This calibration curve is then used to determine the reference complex concentration during the competition experiments. This method reduces the need to make assumptions about the relative ESI efficiencies. Tjemburg et al have demonstrated an automated method for calculating Kd values of complexation between protein and small molecules.60 Using ESI MS, their method calculated dissociation constants directly, for one ligand one target systems, and by competition methods with multiple ligands and one target. These values were also compared to values obtained from CD. To account for the variations and loss of weaker interactions, a response factor was introduced into the calculations to adjust the observed complex abundanc e. This response factor is dependent on the species comprising the complex, instrumental conditions, and on the number of scans used for integration.
31 The Smith group quantified complex binding affinity between BCAII targets and para substituted benzenesul fonamide inhibitors using sustained off resonance irradiation collision induced dissociation (SORI CID) as a method of complex dissociation.50 Normalized relative intensities of the complexes and of BCAII were plotted against the irradiation amplitude; the crossing point of the two curves, E50, was reported as a measure of binding strength. The Marshall group was able to determine binding energies, reported as activation energies Ea laser, using infrared radiation to dissociate complexes comprised of an ani onic tetrapeptide and an artificial cationic receptor.6 1 Against the laser irradiation period, they plotted the natural logarithm of the ratio of protonated complex abundance to fragment abundance. This was repeated for five laser powers. The first order r ate constant, kdiss, was obtained by application of a linear fit to the line for each laser power. The natural logarithm of the kdiss values was plotted against the natural logarithm of laser power density; the slope of this line yields the unimolecular di ssociation activation energy. Target Assisted Combinatorial Synthesis The term combinatorial synthesis encompasses a variety of processes.62 Consider first a conventional combinatorial library. While there are numerous processes to generate these librari es, the y share several features which distinguish them from dynamic combinatorial chemistry. First, the number of candidate ligands in conventional combinatorial chemistry is fixed by the synthesis. Once the chemist completes the reaction, no additional ca ndidate ligands are generated. Also, these candidate ligands are synthesized in the absence of the target. There are also a number of unconventional combinatorial experiments.6364 This has sometimes been referred to as in vitro combinatorial chemistry and it functions as a blanket description for a number of synthetic processes designed to mimic the immune systems ability to synthesize, select, and amplify certain compounds. Included in this umbrella term are
32 systematic evolution of ligands by expone ntial enrichment (SELEX), click chemistry, combinatorial antibody synthesis, chemetics, molecular imprinted polymers (MIPs), and targetassisted combinatorial synthesis.64 Target Assisted Combinatorial Synthesis (TACS) is a synthesis method which takes advantage of molecular recognition principles. Several European groups working along the same lines, including the Lehn group, call this dynamic combinatorial chemistry (DCC).63 The libraries may be referred to as TACS libraries, as dynamic combinatorial li braries (DCLs), or as virtual combinatorial libraries. The complex scheme for TACS is illustrated by a simple schematic in Figure 13. A TACS library is composed of a number of fragments which can reversibly bind each other to form new compounds. These new compounds, which we will call candidates here, make up a virtual library, as not all possible candidates may be present in solution at a given time. The chemistry of the virtual library is designed to interact with another molecule of interest, called the target. When the target is added to solution, candidates will interact reversibly with the target. The interaction between target and candidates is noncovalent in character. Weak binding candidates will bind and quickly dissociate from the target. The tightest binding candidates will be slow to dissociate, and the equilibrium between fragments, candidates, and complex will favor production of the tightest binding complexes. Target assisted combinatorial synthesis differs from traditio nal combinatorial synthesis methods in that TACS pro ducts result from thermodynamic control, rather than kinetic control.63 A comparison of the free energy profiles can be seen in Figure 14. Thermodynamic control of TACS libraries is achieved by composition; these libraries are composed of reversibly binding fragments. Furthermore, since the relative stabilities of the products determine the proportions in
33 which they are formed, the conditions can be modified to shift equilibrium to form a particular product. This shift in equilibrium is achieved by addition of the target molecule. In comparison to traditional combinatorial methods, the advantage of TACS is that much larger virtual libraries of candidate molecules can be generated with less synthesis work on the chemists part. Only the fragment molecules are synthesized; the candidates are transiently formed in solution at low abundances. Those candidates which best bind the target are amplified through the library chemistry. Weaker binders are suppressed. Thus, a purifica tion and/or detection scheme is only needed to isolate and identify the large complexes, rather than all species in solution. In spite of these key advantages, TACS has not yet found widespread use. First, the number of reversible chemical reactions suitable for TACS libraries is limited at present and not necessarily available for all targets. A brief review of some of the libraries used will be shown below. Second, even comparatively large libraries (104 members) have a low probability of forming tightbinding ligands without significant library design. The work necessary to design the ligands approaches the work required to conventionally design a lead compound, and consequently the advantage of speed is lost Most importantly, analysis of these librar ies can be difficult, with high redundancy of similar species; here, redundancy means that similar species cannot be distinguished by a particular detection method. An illustration of this redundancy in mass spectrometric analysis can be seen in Figure 1 5. This figure also shows the major advantage of TACS, i.e ., production of many ligands with minimal synthesis. The choice of detection method is critical for successful use of the TACS method. It needs to be capable of resolving and identifying similar str uctures, maintain the solution equilibrium, and be fast. Several methods are reviewed below.
34 To begin, the review article by Weber provides a good background on unconventional combinatorial methods.64 Along these same lines, Cousins, Poulsen, and Sanders c over dynamic combinatorial chemistry in their article on molecular evolution.65 Next, Rowan et al published an excellent review article on dynamic combinatorial chemistry covering the fundamentals and reviewing previously studied systems.6 3 Huc and Lehn have also reviewed virtual combinatorial libraries.6 6 Articles by Triolo6 7 and Sussmuth6 2 describe the relationship between combinatorial chemistry and mass spectrometry. TACS Systems Analyzed by Non Mass Spectrometric Methods Non mass spectrometric analysi s is best suited for cases where the DCLs are simple and all ligands are known. In these cases, the ligands may be identifie d using nuclear magnetic resonance (NMR) or UVVis spectroscopy where the properties of all ligands are known or predicted previous ly. The Ramstrm group demonstrated that a dynamic combinatorial library comprised of designed thioesters and thiols undergoes transthioesterification in the presence of the target to generate candidate ligands to bind acetylcholinesterase.6 8 Based on the library composition, a total of ten ligands were possible. Proton NMR was used to monitor the reactions between fragments in the absence and presence of the target. Particularly, the formation of acetic and propionic acids and hydrolysis of a butyrate group were able to be observed in the NMR spectrum, reflecting the catalytic process. Conventional Combinatorial Systems Analyzed by Mass Spectrometry The ligands of conventional combinatorial libraries are present in solution in the absence of the target, a nd consequently are not reliant on noncovalent interactions to form. This section, discussing analysis of conventional libraries in the absence of target, has been included because
35 it illustrates some of the challenges and methods used in library analysis. Many of these aspects are important in the analysis of dynamic combinatorial libraries as well. A collaboration between the Eyler and Benner groups produced a paper in 1996, demonstrating analysis of a peptide combinatorial library via ESI FTICR MS.6 9 Th is illustrated the efficacy of high resolution mass spectrometry for analysis of highly degenerate libraries, although even this was insufficient for complete member resolution. They were also aided by computer simulations which generated simulated mass sp ectra of the library complexes. Mass Spectrometric Analysis of TACS Systems Analysis of small libraries is a straightforward matter. All ligands can be predicted in advance, and the detection parameters for each established. Those ligands detected can the n be matched with the predicted properties to identify what is present. Another feature of analysis is the use of tandem methods, such as highperformance liquid chromatography mass spectrometry (HPLC MS). By increasing the amount of information available for each detected species, tandem methods improve the accuracy of identification. One common configuration is a chromatographic separation coupled to mass spectrometric detection. Mass spectrometry is a widely used method to identify the favored enzyme ligand complexes from solution. It has the advantage of speedcomplexes can be rapidly isolated and identified without significant perturbation of the equilibria. Mass spectrometric methods are also sensitive and selective. As an added advantage free enzyme, free ligand, and enzyme ligand complex may all be simultaneously detected Furthermore, judicious choice of mass analyzers can provide high mass accuracy and high resolving power which assist in unambiguous identification of ligands from a highly redundant library. Tandem MS methods can give additional information about identity and structure.
36 Ramstrm and Lehn have published on screening carbohydrate libraries for affinity to concanavalin A.70 Reversibility in the library was achieved via thiol disulfide interconversion chemistry between fragments. In this work, the target enzyme was immobilized on beads; this permitted separation of free target and targetlibrary complexes from the free library; the ligands bound to the immobilized target could be eluted by changing pH. Reversedphase HPLC MS was used to separate the eluted ligands. The library functioned successfully, and production of several ligands was amplified in the presence of the target. When the library size was increased, there was difficulty in resolving all of the ligands using HPLC. This problem of resolving and identifying ligands in a highly degenerate library is a theme in dynamic combinatorial chemistry. Amine and aldehyde based fragments, in the presence of the target neuraminidase, wil l generate ligands as shown by Hochgrtel et al .7 1 Amplification of the highest affinity ligand was calculated around 120 percent compared to the library in the absence of the target. Analysis was performed using HPLC MS. Instead of using a single molecul e target, Li et al used an entire cell to amplify aldimine reduction to the corresponding amine.72 This procedure acts as a screen for imine reductase activity in the cell. Those enzymes present can act on the library much as a single molecule target does The products resulting from the reaction between library and cells were monitored by gas chromatography (GC) MS with an electron ionization (EI) ionization source. The screen identified two putative candidates as resulting from imine reductase activity. Additional evidence suggested that the cells may contain two distinct imine reductases. Another example of a TACS system analyzed by mass spectrometry was published by Poulsen.7 3 Here, BCAII was the target, and the reversible fragment binding was based on hydrazone exchange. Analysis was performed using negative mode ESI FTICR MS with a 4.7T
37 magnet. It was not possible to identify ligands directly from the mass spectra of the complexes due to incomplete resolution. This issue was successfully resolved by using tandem MS to confirm ligand identity. This set of experiments demonstrated the utility of highresolution FTICR MS for TACS library analysis due to speed of analysis and unequivocal identification of the ligands. This thesis will show the development of a method for screening a nucleotide based dynamic combinatorial library for tight binding ligands of Ribonuclease A using Fourier transform ion cyclotron resonance mass spectrometry. In Chapter Three, a model system comprised of Ribonuclease A and uri dine 5 monophosphate was used to optimize mass spectrometry parameters. The resultant noncovalent complexes were studied by infrared multiple photon dissociation tandem mass spectrometry. Next, in Chapter Four, the effects of different wavelengths for dis sociation of Ribonuclease A library complexes were studied. A comparison of these specifically bound complexes was compared to a control nonspecific complex. Lastly, two dynamic combinatorial libraries were screened for binding affinity to Ribonuclease A a nd the binding species identified.
38 K+ Hydrophilic Heads Hydrophilic Heads Hydrophobic Tails Receptor Ligand AB C D K+ Hydrophilic Heads Hydrophilic Heads Hydrophobic Tails Receptor Ligand K+ Hydrophilic Heads Hydrophilic Heads Hydrophobic Tails Receptor Ligand AB C D Figure 1-1. Examples of noncovalent complexes. A) 16-crown-6 selectively binds a potassium cation. B) Water interacts in a hydrogen bonding network. C) Bi nding between ligand and receptor. D) Hydrophobic interactions drive formation of the lipid bilayer.
39 Thymine Adenine Uracil Adenine Guanine CytosineDNA Only RNA Only DNA and RNA Thymine Adenine Thymine Adenine Uracil Adenine Uracil Adenine Guanine Cytosine Guanine CytosineDNA Only RNA Only DNA and RNA Figure 1-2. Hydrogen bonding between nucleotide base pairs. The specific bi nding shown here is called Watson-Crick base pairing, and it elucidates the specific affinity of T-A, U-A, and C-G. Hydrogen bonding between other pair s is possible, but less energetic. Here, hydrogen-bond donors groups are shown in red, and hydrogen-bond acceptors are shown in blue.
40 Fragments Candidates Target Complexes Fragments Candidates Target Complexes Figure 13. T arget assisted combinatorial chemistry (TACS) scheme. At the top, the free library is in equilibrium, with fragments reversibly binding to form candidates. As the target is added at bottom, certain candidates tightly bind to the target and the equilibrium favors production of those complexes.
41 Free EnergyReaction Coordinate Thermodynamic Path Kinetic Path GT GK Transition State GT GK Free EnergyReaction Coordinate Thermodynamic Path Kinetic Path GT GK Transition State GT GK Figure 14. Comparison of free energy profi les for thermodynamic and kinetic control. Dynamic combinatorial chemistry reactions are thermodynamicall y controlled while conventional syntheses are kinetically controlled. Adapted from Rowan et al .63
42 U A UU UA AA AUUUU UUA UAU UAA AUU AUA AAU AAAUUUU UUUA UUAU UUAA UAUU UAUA UAAU UAAA AUUU AUUA AUAU AUAA AAUU AAUA AAAU AAAA306.0253 329.0525 594.0400 617.0673 617.0673 640.0945 882.0548 905.0820 905.0820 905.0820 928.1092 928.1092 928.1092 951.1364 1170.0695 1193.0967 1193.0967 1193.0967 1193.0967 1216.1239 1216.1239 1216.1239 1216.1239 1216.1239 1216.1239 1239.1512 1239.1512 1239.1512 1239.1512 1262.1784 U A UU UA AA AUUUU UUA UAU UAA AUU AUA AAU AAAUUUU UUUA UUAU UUAA UAUU UAUA UAAU UAAA AUUU AUUA AUAU AUAA AAUU AAUA AAAU AAAA306.0253 329.0525 594.0400 617.0673 617.0673 640.0945 882.0548 905.0820 905.0820 905.0820 928.1092 928.1092 928.1092 951.1364 1170.0695 1193.0967 1193.0967 1193.0967 1193.0967 1216.1239 1216.1239 1216.1239 1216.1239 1216.1239 1216.1239 1239.1512 1239.1512 1239.1512 1239.1512 1262.1784 Figure 15. Redundancy of TACS libraries. The exam ple shown here is a two nucleotide library, where RNA chains between fragments form. For chains composed of a single nucleotide, there are two possibilities. For chains of two members, there are four possibilities. When three fragments combine, there are e ight possible candidates, and when four combine, there are sixteen. Chain length may in practice be longer, but if we assume a maximum of four units per chain, there are 30 distinct candidates present from a library of two fragments. This figure also shows in red the exact m/z of these species. Note that although there are 30 candidates, there are only fourteen m/z values, demonstrating the redundancy of the library.
43 CHAPTER 2 INSTRUMENTATION This chapter is meant to introduce the principles, design, and f igures of merit for the instrumentation used in the research reported in this thesis. Ionization, particularly electrospray ionization, will be covered first. Next, Fourier transform ion cyclotron resonance mass spectrometry (FTICR MS) will be presented as a means of ion detection. Dissociation techniques will be briefly reviewed, especially infrared multiple photon dissociation with r egards to their role in tandem mass spectrometry. Lastly, the specific instrumentation used in this research will be descri bed Electrospray Ionization History Electrospray ionization (ESI) was first observed by Malcolm Dole, who was able to produce gas phase macroions of polystyre ne for detection in a Faraday cup.74 However, the possibility of using this phenomenon to generat e ions for mass spectrometric detection was not recognized until the early 1990s. Until recently, the field of mass spectrometry was limited to analysis of small molecules; traditional ionization methods like electron ionization (EI) and chemical ionizatio n (CI) are simply incapable of producing gas phase ions from larger species. Introduction of fast atom bombardment (FAB) ,7 5 secondary ion mass spectrometry (SIMS) ,7 6 and field desorption (FD)7 7 extended the usable ma ssto charge range, yet even these impro vement s w ere insufficient for analysis of peptides, proteins, and other large biomolecules. The technology to mass detect those ions existed, but there w ere simply no means of introducing these species as gas phase ions. In 1989, John Fenn altered the fi eld of mass spectrometry by demonstrating the utility of electrosprays to ionize large molecules for mass spectrometric detection .2 3 Along with the
44 introduction of matrix assisted laser desorption ionization (MALDI) in that same year ,2 4 2 5 ESI was recognized as expanding the utility of mass spectrometry. Consequently, John Fenn was awarded a shared Nobel Prize in 2002 for his work. Electrospray ionization is now among the most commonlyused and routine methods of ionization. Process and Mechanism In ESI, s olutionphase molecules are transformed into gas phase ions through application of a high voltage.2 3 Solutions generally consi st of water, organic solvent, and analyte.1 6 When operating in positive mode, a small amount of acid can be added to increase the available numbers of protons and to improve current. T he relative concentrations of the water and organic phases are governed by two factors. First, the solution must be capable of dissolving the analyte. Secondly, the surface tension of the final solution must fall within a desired range. The reasons for this will be discussed later. An acid modifier, often formic acid or acetic acid, is used to carry charge, to regulate pH, and as a source of protons. The final solution component is the dissolved analyte typically in the concentration range of 105 M or lower.1 6 While not necessary for the electrospray process, solutions of biomolecules often contain buffers for analyte stability. A schematic representation of the ESI process is seen in Figure 21. In e ssence, ESI can be described as an electrical circuit where charges move between the high voltage applied to a needle and the entrance to the mass spectrometer. Application of high voltage (typically 15kV) to the solution results in charge separation. At the ESI tip, there is a balance between Coulombic repulsion of charges and the surface tension of the solution, resulting in the outward bulge called a Taylor cone. The size, shape, and stability of the Taylor cone are influenced by the components of the f lowing solution, the magnitude of the applied voltage, the outer diameter of the capillary, and the presence of any nebulizing gas. As the Rayleigh limit is reached, charge repulsion overcomes the solution surface tension and droplets are emitted from the Taylor cone.
45 The mechanisms and processes involved in droplet formation are critical to results obtained via ESI. Droplets produced directly from the Taylor cone (primary droplets) contain solvent molecules, analyte molecules, and multiple charges. These d roplets undergo one of two proposed m odes of fission. In the ion evaporation model proposed by Iribane and Thompson, charged analyte is desorbed directly from the droplets due to the strong electric field between the needle and mass spectrometer entrance.7 8 For t he second mechanism, Dole theorized additional solvent evaporati on occurs and additional asymmetric droplet fission occurs to produce progressively smaller secondary droplets.7 9 This process continues until the remaining analyte molecules have inhe rited all the charge present in the secondary and tertiary droplets. It is thought that the charged residue mechanism is true for larger species while the ion evaporation model is true for smaller molecules. Both models use heat or nebulizing gas to aid in solvent evaporation. Additionally, both proposed mechanisms are capable of explaining the mul tiple charging phenomenon of ESI. Related Ionization Methods Recently, several new ionization methods have been demonstrated as variants of the ESI desig n. Most of these methods will not be discussed in detail, as they were not used in the project. The earliest was nanoelectrospray or nanospray ionization (nESI).2 7 It is quite similar to ESI, differing only through the use of smaller needles and lower solu tion flow rates. Compared to ESI, nESI has improved sensitivity, lower sample consumption, and higher salt tolerance. These advantages arise because the primary droplets in nESI are smaller than those in ESI, and consequently, fewer secondary fission event s are needed to produce desolvated analyte ions. In the early part of the 2000s, the Cooks group at Purdue introduced several new ionization methods, coinciding with a rise of interest in ambient ionization techniques. First introduced
46 were electrosonic spray ionization (ESSI)80 and desorption electrospray ionization (DESI).8 1 As a whole, these techniques are capable of ionizing large numbers of molecules and are gentler even than ESI.8 2 The Cooks group also demonstrated that ESSI and DESI of large biomole cules resulted in narrower charge state envelopes at higher massto charge ratios. The implications of this will be discussed in a later chapter. Advantages, Disadvantages, and Figures of Merit As mentioned previously, ESI was one of the first techniques c apable of ionizing large molecules. The significance of this cannot be understated. It should be noted that, unlike MALDI and other nonESbased techniques, ESI produces multiplycharged species as an artifact of the ionization process. Multiple charging h as the key advantage of reducing the m/z of large molecules, making detection with a range of mass detectors possible. Electrospray ionization of species with molecular weights greater than 800 kDa has been reported.3 9 It was also recognized early on that ESI is a gentle ionization technique. By this, it is meant that ESI has been demonstrated to maintain solution phase noncovalent interactions, protein conformations, and macromolecular structure.3 5 Along with MALDI, and to some extent fast atom bombardm ent (FAB), ESI was one of the first techniques that ionized without significant fragmentation. This gentleness is especially important for study of biological systems. Compared to earlier ionization methods, the sample demands for ESI are low. Flow rate s range from 10 L/hr to 300 L/hr with a nalyte concentrations in the micromolar range or lower.1 6 Limits of detection as low as sub attomolar have been reported .1 6 Signal response is linear over roughly four orders of magnitude.1 6 With simple changes to nESI, etc., the se nsitivity of ESI can be improved by another order of magnitude or more.2 7 However, the limiting factor in sensitivity
47 seems to be ion transmission limits in the mass analyzer rather than the efficiency of ion generation.1 6 Another advantage of ESI is th at it can readily be interfaced with chromatographic and capillary electrophoretic techniques. This reduces analysis time, as no separate steps are needed to prepare a sample for analysis after purification. It also makes use of automation easier. Unlike M ALDI or FAB, ESI has no matrix requirements. This means there are no matrix interferences at low massto charge ratios to complicate analyses. Unfortunately, there are disadvantages to ESI. First, ESI is best suited for analysis of polar compounds.22 Gas phase proton affinity of the analyte must be higher than that of the solvents used when generating protonated species.8 3 Second, ESI is prone to signal suppression ; 1 6 8 4 this phenomenon arises when there are more species than there are charges available. It is further exacerbated by the fact that certain species are more prone to charging tha n others and consequently, the electrospray ions produced may not accurately reflect the species in solution. In complex mixtures or solutions containing significant a mounts of salts or buffers, certain species can be overrepresented at the expense of other species. Thus, samples containing many components may require additional separation prior to analysis. In general, additional purification and desalting may be requi red to remove salts. Another disadvantage of ESI is the requirement of organic solvents and acid modifiers. These improve ionization but affect the nature of the solution, making it unlike physiological conditions. This does not negatively impact all anal yses; however, studies of complexation, stability, and structure may be affected.
48 Fourier Transform Ion Cyclotron Resonance Mass Spectrometry History Se veral excellent review articles on FTICR mass spectrometry are available.858 7 The current FTICR mass spectrometer owes much to the fundamental work done by scientific giants. James Maxwell, in his famous equations, shed light on the behavior of electric and magnetic fields. Building upon the work of Maxwell and Faraday, Lorentz thoroughly described the mat hematics of a charged particle in a magnetic field. Early instrumentation was developed for the field of nuclear physics. In the 1930s, E.O. Lawrence built the cyclotron, which he used to accelerate charged particles to a large radius. 8 8 Using the cyclotr on, Lawrence and others were able to prob e fundamental atomic properties; 8 9 indeed, current particle accelerators can be envisioned as the offspring of Lawrences earlier cyclotron design. By the 1950s the utility of cyclotron resonance in mass spectromet ry was recognized in the development of the omegatron. Sommer et al used this early ion cyclotron resonance (ICR) mass spectrometer design to obtain the mass to charge ratio of a proton. 909 1 The next major development was using ICR to trap ions for exte nded periods. McIver demonstrated in the 1970s that ions could be held and detected for milliseconds and greater, permitting study of ion evolution.9 2 In 1965, Cooley and Tukey developed the fast Fourier transform, which enabled simultaneous detection of m ultiple frequencies with relatively short computation times and therefore concurrent ion detection in mass spectrometry.93 Fourier transform ion cyclotron resonance mass spectrometry debuted in 1974 with publication by Marshall and Comisarow.9 4 9 5 Since t hen, interest and application of FTICR mass spectrometry has grown, with commercial instrumentation offered now by several companies. The year 2004 saw the introduction of a new FTMS design in the form of the Orbitrap.9 6 Like FTICR, the Orbitrap detects io n frequencies which are then Fourier transformed. However, the
49 principles involved are not those of cyclotron resonance, and the capabilities of the Orbitrap for specialized experiments are of yet limited. Principles Cyclotron m otion The fundamental princi ple of FTICR mass spectrometry is the behavior of charged particles (ions) in magnetic field. Empirically, this translates to the motion of ions in the ICR trap. The overall motion of the ion packet is comprised of several components, each of which is described mathematically through a series of equations. To begin, cyclotron motion arises in an electric field E and a magnetic field B where an ion with charge q will be subject to a force perpendicular to B and to the direction of ion motion. This force F t he Lorentz force is described in Equation 21 where m is the ion mass, q is the ion charge, v is the ion velocity, E is the electric field, and B is the magnetic field. B v q E q dt v d m ma F (2 1) When ion velocity is constant and no collisions o ccur, the ions travel along a circular path. The charge on the ion determines its path, with positive and negative charges orbiting in opposite directions. Let the plane perpendicular to B be defined as the x y plane and ion velocity in this plane as vxy. To achieve circular motion, the Lorentz force is equal to the centripetal force. The angular acceleration can be described by Equation 2 2, where r is the radius of cyclotron motion. r v dt dvxy 2 (2 2) Substituting into Equation 21, Equation 23 is obtained which can be rearranged to obtain the radius of cyclotron motion in Equation 24.
50 B qv r mvxy xy2 (2 3) qB mv rxy ( 24) Using additional equations, the cyclotron frequency can be det ermined Substituting the angular velocity xy/r into Equation 23 results in Equation 25, r qB r m 2 (2 5) which can be rearranged to give the cyclotron frequency c ( Equation 2 6) m qB c (2 6) Rememb ering that one can determine an alternate definition of the cyclotron frequency, fc ( Equation 27) m qB fc 2 ( 27) Thus, the cyclotron frequency is de termined by the magnetic field strength, which is constant in F TICR MS, and by the mass to charge ratio (m/q ) of the ion. The mass to charge ratio is frequently expressed as m/z where z is the number of fundamental charges. There is no dependence on ion velocity, and therefore, no dependence on kinetic energy. Of the three types of motion observed, the cyclotron motion dominates. Trapping m otion When an ion moves parallel to the magnetic field, its equation of motion has no contribution due to magnetic field force, and thus motion along this axis ( z ) is unconstrained. The escape of ions from the cell along this direction can be prevented by applying a voltage to the end caps or plates of the cell. Applying this electric potential adds a second type of ion motiontrapping motion. Typically, a three dimensional axial quadrupolar trapping potential is
51 used. It can be described by Equation 28 where Vtrap is the trapping potential, r is the radial position of the ion in the xy plane, a is a description of the trap size, and and are descriptors of the trap geometry. 2 2 22 2,r z Vtrapz y (2 8) Ion motion in the z direction is then described by Equation 29, z y x q dt z md Faxial, ,2 2 2 ( 29) from which one can determine the timedependent z axis position with the oscillating frequency of 22 2 1 ma qV vtrap z (2 10) Magnetron m otion A third type of motionmagnetron motionarises from the uncoupled electric and magnetic trapping potentials. To begin, the tra pping potential from Equat ion 28 gives rise to a radial force Fradial, described by r qV qE Ftrap r radial 2 (2 11) Assuming a static magnetic field and a threedimensional axial quadrupolar trapping force, Equations 2 5 and 210 can be combined to gi ve the following equation r a qV r qB r m Ftrap 2 2 ( 212) Setting the equation equal to zero, we obtain the quadratic equation 02 2 ma qV m qBtrap (2 13)
52 Solving for we find the two natural rotational frequencies 2 2 22 2 2 ) ( z Cw (2 14) a nd 2 2 22 2 2 ) ( z c ( 215) The first of these frequencies, Equation 214, describes the cyclotron frequency in the absence of a d.c. trapping potential. The second, Equation 215, is t he magnetron frequency. Combining the three types of ion motioncyclotron, trapping, and magnetron ions in the cell behave as shown in Figure 2 3. Again, the cyclotron motion contributes more than the trapping or magnetron motion, and consequently, the c yclotron motion is usually the only motion detected. Excitation, i solation, and d etection Ions trapped in the cell tend to have low kinetic energies and will consequently have very small radii of motion in the cell independent of m/ z. It is necessary to ap ply a potential to the excitation plates of the cell to increase the radius of motion and to generate a coherently moving packet of ions The energy imparted to the ions can be used for several different purposes, which are shown in Figure 24 and discusse d below. First, application of a uniform oscillating electric field of the proper frequency will excite all ions in the cel l to larger radii and coherent motion The radius of motion to which the ions are excited must be smaller than the cell size; otherwi se ions are lost to collisions with the walls. During the process, the ions will continue their cyclotron motion as a coherent ion packet. This is called broadband or chirp excitation. Ions can also be selectively excited by applying oscillating voltages w ith frequencies that will excite only ions of a particular m/z
53 Second, the energy imparted by excitation fields can be used to dissociate ions. One such instance of this is sustained off resonance irradiation collision induced dissociation (SORI CID). Lastly, ions in the cell can be excited to radii greater than the cell size as a means of removal by collisions with the wall. Application of broadband oscillating electric fields will remove all ions. In some cases, such as tandem MS, it may be desirable to retain certain ions while removing others. This can be achieved through application of a stored waveform inverse Fourier transform (SWIFT) potential to the excitation plates. Once ions have been excited to the larger radii, the detection plates record an image current which simultaneously reflects the frequencies of all ions C onsequently, FTICR MS has the benefit of the broadband detection or the multiplex advantage. Note that the signal intensity is independent of magnetic field strength. The resulti ng transient is processed using a fast Fourier transform algorithm and calibrated to obtain a mass spectrum. A sample transient and its corresponding mass spectrum can be seen in Figure 25. Experimental s equence Now that we have covered the principles of FTICR, lets look at how these events fit together to provide a mass spectrum. Figure 26 shows a typical experimental sequence for an FTICR experiment. It is comprised of several events: quench, ionization, trap, excite, and detect. Quenching removes ions remaining in the cell from previous experiments. For ionization, ions are directed from an external ion source into the cell, or when using an internal ionization source, ions are generated inside the cell. Once these ions are in the cell, they are trap ped by application of the trapping potential. These ions are excited to a larger radius and then detected. The above describes a basic experimental sequence, but there are many variations that can be used for specialized experiments. Figure 2 7 shows how the general experimental
54 sequence can be modified for ion isolation and/or tandem mass spectrometry This capability makes FTICR one of the more versatile mass spectrometric techniques available. Instrumentation In the next section, the instrumental components of the FTICR MS will be introduced, with explanations of principles and reasons for use. Lastly, some advantages and limitations of each component will be presented. Magnets Magnetic fields are required to generate the cyclotron motion of the ions i n the cell. The ICR cell resides inside the bore of the magnet where the ICR cell resides. These fields can be generated in several ways. Permanent magnets, having low magnetic field strength (<2 T), have been used in several instruments but are not widespread due to better alternatives. Electromagnets were frequently used in the early history of FTICR MS because they provide stronger fields than permanent magnets (~1T 3 T). Present instrument design incorporates superconducting magnets, which generate the strongest magnetic fields to date (~3T 25 T). As will be shown later, increasing magnetic field strength corresponds to significant improvement in FTICR MS performance. Successful experiments with field strengths up to 25 T have been demonstrated, although this is not typical. Field strengths of 9 T are now routinely used and available commercially. However, there are two disadvantages to superconducting magnets. First, they require the use of costly cryogenic helium and nitrogen to maintain a core temperat ure of 4.2 K. The introduction of refrigerated magnets has fortunately reduced the cryogen demand. Second, the large size of these magnets means that portability and miniaturizati on are not possible in FTICR MS unless lower field (and lower performance) pe rmanent magnets are used.8 5 8 6
55 ICR c ells The ICR cell is critical, as it is the place where ions are trapped, excited, and detected; it is also where tandem in time experiments are performed. These cells are also called Penning traps. A basic schematic of an I CR cell is shown in Figure 28. Simply, it consists of three pairs of plates arranged to form a cube. Ions are trapped in the interior. The plates perpendicular to the z axis are responsible for trapping the ions; it is to these plates that the trap ping potential is applied. Another set of plates is used to excite the ions. The last set of plates is used to detect the image current. Cell design has evolved beyond this simple scheme. Figure 2 9 shows a variety of cell designs which have been implement ed to smooth potentials and/or increase trapping efficiency.8 6 In particular, the use of cylindrical cells has been favored due to their ability to hold larger numbers of ions. This shape in particular is better suited to fitting inside the cylindrical bor e of superconducting magnet s Yet all these cells share the feature of trapping electrodes orthogonal to the direction of ion motion and pairs of excitation and detection electrodes.8 5 8 6 Vacuum p umps Another feature shared by all FTICR mass spectrometers (indeed, shared by all mass spectrometers) is vacuum pumps. High to ultra high vacuum (108 to 1010 torr) is needed in the ICR cell to minimize collisions and extend trapping periods. Slightly higher pressures can be used for events other than detection. The most common example of this is when a gas is added for collision induced dissociation. However, before detection can occur, the cell must pump back down to the lower pressure. Ultralow pressures are achieved by differential pumping, with several st ages of pumping present between the atmosphere at the source and the vacuum of the cell. Most often, turbomolecular or cryogenic pumps are needed to achieve the necessary pressure. Mechanical
56 pumps can be used for backing. Diffusion pumps also have seen li mited use. These demands increase the cost of an FTICR mass spectrometer, as turbomolecular or cryogenic pumps are costly and several are needed.8 5 8 6 Ion optics Ion optics play a critical role when ionization occurs e xternal to the cell. These lenses, f unnels, plates, and rings are responsible for transmitting ions into the ICR cell with minimal loss. Additionally, they can be used to cluster ions into a packet, to spread ions out, and to reduce kinetic energy. Electronics and c omputers Without modern electronics and computing power, FTICR mass spectrometry wo uld not be practical Advances in electronics have im prove d the speed of data collection and analysis as well as improved the size of datasets generated. Among the key components are a f requency s ynthesizer a d elay pulse generator a broadband r.f. amplifier and preamplifier and a fast transient digitizer. Lastly, a computer is used to control each of these components. The computer is also used for data analysis, most importantly in performing the fast Fourier transform.8 5 8 6 Specific software applications have been developed for specialized data analysis. Advantages, Disadvantages, and Figures of Merit Mass resolution and r esolving p ower FTICR MS provides superior mass resolution compared to al l other mass spectrometric techniques. 8 5 8 6 In practice, mass resolution, 50%, is the full width of a peak at half height. On a mass spectrum, this can be represented by as the point at which two peaks separated by 50% will have a valley between them. Another term related to resolution is resolving power, R or 50%. It is desirable to have high resolving power and low resolution.
57 For this technique, resolution is dependent on the length of the transient, or time domain signal. 8 5 8 6 As we will sho w, longer transients provide better resolution. But there are practical limitations to long detection times. While FTICR MS is performed at ultrahigh vacuum, there are still extraneous species present in the cell that can collide with analyte ions. Within the ion packet, ions may interact repulsively with each other. When these events occur, the coherence of the ion packet is reduced and at long enough times, will go to zero. Thus, t he transient signal will decay at long times. Within each transient, reso lution is affected by the number of data points collected and processed, with more points being desirable. However, collecting and processing more points requires greater computing power and results in longer analysis times. Presently there is an upper lim it of ~106 points. 8 5 8 6 The lower limit is defined by the Nyquist theorem, since a minimum number of points m ust be collected in order to accurately reflect the data. 8 5 8 6 The sampling frequency must be at least twice the frequency of the highest frequency (lowest mass) ion detected. The relationship between the sampling frequency S, the number of data points collected N and the length of the transient Tacq can be described by Equation 216. S N Tacq ( 216) For a particular experiment, the highest resolving power achievable is described by Equation 217, where R is the resolving power, Tacq is the length of the transient, and fc is the cyclotron frequency. m TqB Tf Rc 2 2 (2 17) The resolving powe r is directly proportional to the cyclotron frequency. Since cyclotron frequency decreases with increasing m/q the resolving power varies for ions of different m/q
58 The lowest mass, highest frequency ions can be better resolved than higher mass, lower fre quency ions. 8 5 8 6 Mass accuracy In addition to superior mass resolution, FTICR MS is known for providing excellent mass a ccuracy. As shown in Equation 27, the detected cyclotron frequency is dependent only on magnetic field strength and the mass to ch arge ratio. Calibratio n with known standards is used to give an accurate value for the magnetic filed strength and thus to determine the actual masses. The calibration can be performed in a separate run, or simultaneously as part of the sample run. Mass accuracy in ESI FTICR MS is typically on the order of ppm ppb. 8 5 8 6 Mass r ange Theoretically, there are no upper or lower limits to the size of the ions that can be analyzed by FTICR MS. Solving Equation 2 7 for ions in typical mass ranges 10 100,000 shows that these ions will have frequencies in the kHz to MHz rangeseasily detectable by modern electronics. 8 5 8 6 However, there are practical limitations. The lower limit boundary is defined by the Nyquist theorem. As ions get smaller, their cyclotron freque ncies and the required sampling rates get larger. The upper mass range is typically limited by the ionization techniques available. Furthermore, simultaneous detection of ions over a large mass range is difficult. Linear operating r ange A linear dynamic range of 105 has been reported elsewhere.8 5 8 6 The limit of quantitation is around 100200 charges in the trapthe amount needed to generate a detectable broadband image current. 8 5 8 6 In practice, this obstacle is overcome by the use of multipole storage units, which can concentrate and pulse ion packets into the cell. With this instrumental modification, performance is typically l imited by the ionization source, with regards to generating and transmitting ions.
59 Improving FTICR MS performance The field of FTICR mass spectrometry has undergone significant developments since first introduced in the 1970s. It is relevant to discuss some of the reasons for this. Developments in magnet technology have been one of the keys to improved FTICR MS performance. As shown in Figure 2 11, many key measures of performance improve linearly or quadratically with increasing magnetic field strength. 85 Among these is resolving power, which was shown in Equation 217. More powerful computers have improved the speed and effi ciency of data analysis. Additiona l l y, they have increased the size of datasets which can be handled, which in turn has improved resolving power. Lastly, a number of new ionization sources have been introduced, greatly expanding the classes and sizes of molecules that can be analyzed by FTICR MS. 8 5 8 6 Advantages and d isadvantages The main disadvantage of FTICR mass spectrometry is the high cost of purchasing and maintaining the instrument. However, the costs associated with high field magnets and ultral ow pressure vacuum systems have seen some reduction in recent years. Another limitation is the size of an instrument due to its superconducting magnet. Portability is not possible. Indeed, FTICR mass spectrometers are not even suited for a benchtop instru ment. Consideration must be given to where the instrument is located; the size and weight of some magnets can be limiting. Furthermore, unshielded magnets must be located away from electronics which might be damaged. Costly cryogenic systems are required to maintain superconducting magnets, and the handling of such carries the risk of personal injury. Despite these disadvantages, FTICR remains the premier type of mass spectrometry It provides superior resolution and mass accuracy. Virtually the entire mass range is detectable.
6 0 Nondestructive detection allows a variety of tandem in time experiments to be performed. Some of these techniques are only available on the FTICR at present. Dissociation Techniques There are numerous ways to dissociate and fragment the ions trapped in the ICR cell. Generally, these techniques can be divided into slow heating or nonergodic. 8 5 8 6 9 7 Slow heating describes the ability of an ion to internally redistribute whatever energy is imparted and include such techniques as collision induced dissociation and infrared multiple photon dissociation. In nonergodic processes, dissociation is faster than the internal randomization of energy. Some highenergy CID is nonergodic, along with electron capture dissociation (ECD). A good review of large ion dissociation was written by Laskin et al 9 8 Collision I nduced D issociation (CID) Collision induced dissociation (also called collisionally activated dissociation, or CAD) is the most commonly used technique for ion dissociation and it is available on any instrument capable of tandem mass spectrometry. There are many types of CID experiments which are variants of the same principles. Energetic collisions between analyt e ions and gas molecules result in transfer of energy sufficient to break bonds.545 5 8 5 These gas molecules may be background neutrals in the mass spectrometer or they may be speci fic ally added at specific times. Controlling the energy and amount of gas added can allow the available energy for dissociation to be tuned. These collisions can take place in any part of the mass spectrometer. The most common method is to leak gas into the reaction cell where collisions occur. In an FTICR mass spectrometer, the next step is to pump the cell back down to operating pressures. Th is type of experiment can be called in cell CID. The advantage of this particular scheme is that specific ions can be massselected for dissociation. However, the additional pumping requirement
61 reduces the effectiveness of this technique. The time to run a n experiment increases and some of the analyte ions or fragment ions may be lost. Also, introduction of the background gas results in a poorer signal to noise ratio. An alternative to in cell CID is in source CID. The nomenclature varies, depending on the particular components of the mass spectrometer. In source CID works by allowing highenergy collisions to occur in the source region where background pressures are much higher than the cell.9910 3 These types of collisions occur in all experiments, but the key to dissociation is the difference in applied voltages between certain optics. When the voltage difference is high, more dissociation is observed. Figure 211 shows an example of insource CID between a capillary and skimmer. Here, a relatively high voltage is applied to the end of the capillary and a relatively low voltage is applied to the skimmer. The advantage of this mode of dissociation is its relative ease of implementation ; without recognition of the technique, many mass spectrometrists use it t o reduce the appearance of solvent clusters and salt adducts in a mass spectrum. No additional gas is needed, and therefore no additional pumping step is required. There are two critical disadvantages to in source CID. First, ion isolation is not possible.9 9 All ions are subjected to collisions and the subsequent fragmentation may be obfuscated. Second, the magnitude of voltages needed to induce dissociation may be beyond the capabilities of a particular ion optic. Alternatively, the voltages needed may fal l outside the range of those needed for optimal tuning. Infrared M ultiple P hoton D issociation (IRMPD) In ion trap mass spectrometry, IRMPD has been used as an alternative to CID.9 6, 10 4 105 Here the trapped ions are i rradiated with infrared photons with i r radiation times that vary from milliseconds to hundreds of seconds. Each ion absorbs multiple photons and the energy is internally reorganized into a quasicontinuum of vibrational states. Once sufficient photons have been absorbed to equal the dissociation energy of the weakest bond, the ion fragments. The
62 absorption process is dependent on the energy of the photons and the energy states available to the ion. As in spectroscopy, different types of bonds will absorb at different wavelengths. Thus, IRMPD can be used to glean specific structural information about the ions. This has been demonstrated for amino acids, carbohydrates, and small peptides. Infrared photons may be generated from a number of laser sources. Carbon dioxide (CO2) lasers are the cheapest and therefore the most common method of generating infrared radiation.104 In t he cheapest of the CO2 lasers, lasing occurs at a single wavelength. This is sufficient for most IRMPD applications. Typically used is the P20 band, corresponding to This particular wavelength has a strong, stable intensity. For more specialized applications, tunable CO2 lasers may be used.10 4 Here, the las ing wavelength may be varied by the user exists some variation in this range, as well as the precision of wavelengths, depending on the laser model. It is important to note that part of this range is inaccessible to the CO2 laser, due to the gaps between emission lines Altern atively, optical parametric oscillator (OPO) lasers may be used for IRMPD There are several advantages to OPO lasers.104 First, they operate in a different wavelength region than CO2 lasers, so different types of information may be gleaned. Second, lasing wavelengths may be more precisely con trolled. These features make the OPO especially attractive for ion spectroscopy applications. However, outside of that, OPO lasers are rarely used for IRMPD because of the cost. Free electron lasers (FELs) are anothe r source for IR photons.104 They are prohibitively expensive for individual laboratories to purchase, and consequently, they are often located at
63 national facilities. Of these three sources, FELs are the least used for IRMPD, and like the OPO lasers, tend to be used for ion spectroscopy applications. IRMPD has several advantages over CID.98, 105 It is not necessary to leak in a gas reagent, and thus no additional pumping stage is required. T his may reduce the amount of time needed to perform an experiment. Furthermore, it may result in better signal quality, as background is smaller and fewer analyte ions are lost. Next, a wider range of internal temperatures may be accessible using IRMPD. These temperatures may be more precisely controlled as well. Most im portantly, IRMPD can be used to give specific structural information, especially when used in conjunction with theoretical calculations. However, there are several drawbacks to IRMPD in comparison to CID.9 8 10 4 105 First, IRMPD use has been limited to wo rk in traps, primarily in the FTICR. Second, the cost of IRMPD is higher, as it requires purchasing a laser as well as modifying the mass spectrometer to introduce the radiation into the cell. The information gleaned from IRMPD and CID may be complementary, and where available, using two dissociation methods proves advantageous. Experimental Instrumentation Figure 213 shows a schematic of the instrumentation used for ESI FTICR in the Eyler lab at the University of Florida. It is a Bruker 47e (Bruker Daltonics; Billerica, MA, U.S.A.) with 4.7 Tesla superconducting magnet (Magnex Scientific Ltd.; Abdington, U.K.) The Analytica electrospray source (Analytica of Branford; Branford, CT, U.S.A.) is a low flow rate (10 20 or curtain gas flows. A high voltage (~1.52.0 kV) is applied to the ground fused silica spray needle. Leading into the mass spectrometer is a user modified heated metal capillary (HMC) which is held at 0 300 V.69 Typically, it is heated to 120 C. Furthe rmore, the HMC has been modified such that the inlet on the ESI end is funnel shaped. Bruce et al. have reported this type of design to give improved sensitivity.106
64 After exiting the mass spectrometer end of the HMC, the ions encounter a skimmer. Th e region between the HMC and the skimmer can be used to carry out in source CID. Past the skimmer is a hexapole ion guide, which is used for accumulating then pulsing the ions to the cell as a packet. Along the way, the ions are guided by a series of ion optics, which can be tuned for optimal transmittance. The Infinity cell is located in the bore of the magnet.107 Opposite the ion entrance to the cell is a zinc selenide window for transmitting the IR beam in IRMPD experiments. Timing and events are controlled by computer IRMPD experiments discussed in this thesis used CO2 laser s as a photon source For the experiments in Chapter Three, an Apollo 570 Tunable CO2 laser (Apollo Lasers Inc.; Chatsworth, CA, USA) with accessible frequencies of 930950 cm1 and 10301060 cm1 and power intensities of 0 13 W was u sed The setup for these experiments is shown in Figure 2 14. Power readings were obtained by manually moving the power meter into the path of the laser beam. Timing and sequences for the laser were controlled by the computer system (Predator). For Chapt ers Four and Five, a different setup was used; this can be seen in Figure 215. A Lasy 20G AT adjustable wavelength CO2 laser (Access Laser; Marysville, WA, U.S.A) with accessible frequencies of 918 980 cm1 and 10111089 cm1 and powers of 07 W was used A gate d was used to limit irradiation to certain events, and also to obtain intermittent power readings during scan collection The gated mirror and FTICR were controlled by the Predator data system.
65 Figure 21. Schematic representation of the ESI process. A solution comprised of water, organic solvent, acid modifier, and analyte is pushed through a small line and tip. A high voltage is applied, and the solution outwardly bulges from the tip in a balancing act between charge repulsion and surface ten sion. Emitted d roplets progress from left to right and become desolvated The resulting gas phase ions enter the mass spectrometer at right. Reproduced with permission of John Wiley and Sons, Inc from Cech, N.B.; Enke, C.G. Mass Spectrom. Rev. 2001, 20, 364.
66 Figure 22. Motion of a charged particle in a magnetic field. The magnetic field is directed into the plane of the page. The motion of the charge is orthogonal to the field, with positive and negative charges orbiting in opposite directions. Reproduced with permission of John Wiley and Sons, Inc from Marshall, A.G.; Hendrickson, C.L.; Jackson, G.S. Mass Spectrom. Rev. 1998, 17, 3. Figure 23. Ion path resulting from cyclotron, m agnetron, and trapping motions. Reproduced with permission of John Wil ey and Sons, Inc from Marshall, A.G.; Hendrickson, C.L.; Jackson, G.S. Mass Spectrom. Rev. 1998, 17, 14.
67 Figure 24. Ion motion following the application of an excitation voltage. The cell at left represents normal excitation, whereby the ion packet is excited to a larger, detectable radius. The central cell shows excitation to produce ion dissociation. The cell at right shows excitation of ions to a radius larger than the cell, for ion ejection. Reproduced with permission of John Wiley and Sons, Inc. fr om Marshall, A.G.; Hendrickson, C.L.; Jackson, G.S. Mass Spectrom. Rev. 1998, 17, 5. m/z 1,600 1,500 1,400 1,300 1,200 1,100 1,000 900 800 time 960,000 640,000 320,000 0 Fourier Transform Transient Mass Spectrum m/z 1,600 1,500 1,400 1,300 1,200 1,100 1,000 900 800 time 960,000 640,000 320,000 0 Fourier Transform Transient Mass Spectrum Figure 25. Relationship between transient and mass spectrum. On the left side is a transient obtained for myoglobin, a small protein. This signal undergoes a fast Four ier transform to generate the spectrum on the right. Each m/z in the mass spectrum has a cyclotron frequency which appears in the transient. Each peak in the spectrum represents myoglobin, albeit with each peak having a different number of associated charg es. This is characteristic of an ESI mass spectrum.
68 Quench Ionization Excitation Detection QuenchTime Quench Ionization Excitation Detection QuenchTime Figure 26. T ypical experimental sequence for FTICR MS. Quench Ionization Excitation Detection Quench Isolation DissociationTime Quench Ionization Excitation Detection Quench Isolation Time A B Quench Ionization Excitation Detection Quench Isolation DissociationTime Quench Ionization Excitation Detection Quench Isolation Time Quench Ionization Excitation Detection Quench Isolation DissociationTime Quench Ionization Excitation Detection Quench Isolation DissociationTime Quench Ionization Excitation Detection Quench Isolation Time Quench Ionization Excitation Detection Quench Isolation Time A B Figure 27. Variations o f experimental sequences for FTICR MS. A) I solation and analysis of an ion. B) T andem in time isolation, dissociation, and analysis for an analyte
69 Ion Motion Excitation or Detection Excitation or Detection Trapping Ion Motion Excitation or Detection Excitation or Detection Trapping Figure 28. Simple ICR cell. The cell is comprised of three pairs of plates. Each pair occupies positions on opposite faces of the cell. The trapping plates are located orthogonally to the motion of ions into the cell. The location of excitation and trapping plates is not important.
70 Figure 29. Variations o f the ICR cell. Reproduced with permission of John Wiley and Sons, Inc. from Marshall, A.G.; Hendrickson, C.L.; Jackson, G.S. Mass Spectrom. Rev. 1998, 17, 12.
71 Figu re 210. FTICR MS performance improves with increasing magnetic field strength. Four performance parameters improve linearly with magnetic field strength as shown on the left. Five other parameters improve quadratically with field strength as shown on the right. This underscores the importance of improved magnet technology. Reproduced with permission of John Wiley and Sons, Inc. from Marshall, A.G.; Hendrickson, C.L.; Jackson, G.S. Mass Spectrom. Rev. 1998, 17, 22.
72 200V 10V + + ++Capillary Skimmer Gas and Solvent Molecules ( ), Analyte Ions ( ) Fragment Ions ( ) 200V 10V + + ++Capillary Skimmer Gas and Solvent Molecules ( ), Analyte Ions ( ) Fragment Ions ( ) Figure 211. I nso urce collision indu ced dissociation ( CID) scheme. Species are moving from right to left. Gas molecules are shown in light gray, analyte ions in dark gray, and fragment ions in red. Electrosprayed species (ions, gases, and small droplets) pass through the capillary and then t hrough the skimmer. Voltages are applied to these lenses to help guide the ions into the mass spectrometer. Pressures in this region are relatively high compared to those of the mass analyzer, and significant amounts of background gas are present. The pote ntial difference between capillary and skimmer accelerates ions. Energetic collisions between the ions and background gas molecules occur. This can cause energy transfer and fragment ions.
73 Dissociation threshold v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR Dissociation threshold v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR Dissociation Threshold Intramolecular Vibrational Energy Redistribution (IVR) IVR IVRSequential Photon Absorption Dissociation threshold v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR Dissociation threshold v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR v =0 v =1 v =2 v =3 v =4 v =5 j =1 j =0IVR Dissociation Threshold Intramolecular Vibrational Energy Redistribution (IVR) IVR IVRSequential Photon Absorption Figure 212. Infrared multiple photon dissociation scheme. When a ground state molecule absorbs a photon, it becomes excited. Subsequent absorption events lead to a quasicontinuum of excited states until the threshold for dissociation is reached. Figure modified with permission of Dr. Sarah Stefan.
74 4.7T Superconducting Magnet ZnSe Window Infinity Cell FOCL2 PL9 FOCL1 Gate Valve HVO YDFL XDFL PL4 PL2 PL1 Extract/Trap Plate Hexapole Skimmer Modified Heated Metal Capillary Electrospray Tip Ion Lenses Turbopump Region 1 Turbopump Region 2 Turbopump Region 3 Atmosphere 4.7T Superconducting Magnet ZnSe Window Infinity Cell FOCL2 PL9 FOCL1 Gate Valve HVO YDFL XDFL PL4 PL2 PL1 Extract/Trap Plate Hexapole Skimmer Modified Heated Metal Capillary Electrospray Tip Ion Lenses Turbopump Region 1 Turbopump Region 2 Turbopump Region 3 Atmosphere Figure 213. Sch ema tic of FTICR mass spectrometer in the Eyler Lab, University of Florida. Figure is not to scale.
75 Wavemeter Power Meter Readout External Ionization Source Ion Optics 4.7T Superconducting Magnet InfinityTMCell with ZnSe WindowCO2LaserM1 M2 S.B. FTICR P.M. Laser Table Predator Wavemeter Power Meter Readout External Ionization Source Ion Optics 4.7T Superconducting Magnet InfinityTMCell with ZnSe WindowCO2LaserM1 M2 S.B. FTICR P.M. P.M. Laser Table Predator Figure 214. Layout of the CO2 laser and FTICR mass spectrometer used for experiments in Chapter Three. Red dashed lines indicate the path of the laser beam. The beam path is determined by the placement of several mirrors (M). The power meter (P.M.) was manually moved in and out of the beam to obtain power readings. A computer system (Predator) controlled the event sequence for the FTICR and laser.
76 Water Chiller Wavemeter Power Meter Readout External Ionization Source Ion Optics 4.7T Superconducting Magnet InfinityTMCell with ZnSe Window Motor CO2LaserM1 M2 M3 M4 S.B. Mgate FTICR P.M. Laser Table Predator Controller Water Chiller Wavemeter Power Meter Readout External Ionization Source Ion Optics 4.7T Superconducting Magnet InfinityTMCell with ZnSe Window Motor CO2LaserM1 M2 M3 M4 S.B. Mgate FTICR P.M. Laser Table Predator Controller Figure 215. Layout of the CO2 laser and FTICR mass spectrometer used for experiments in Chapters Four and Five Red dashed lines indicate the path of the laser beam. The beam path is determined by the placement of several mirrors (M). One of these mirrors (Mgat e) had two possible positions. When the mirror was in the up position, the laser beam was directed into the power meter (P.M.) to measure the laser output power. When the mirror was down, the laser beam was directed into the ICR cell. A computer system (Pr edator) controlled the event sequence for the FTICR and gated mirror.
77 CHAPTER 3 STUDIES OF RIBONUCLE ASE A AND URIDINE 5 MONOPHOSPHATE The experiments described in this chapter were performed on a known enzyme inhibitor system to optimize conditions for mass spectrometric analysis before proceeding to the more complicated enzyme library system. First, free enzyme was employed to optimize electrospray conditions. Since the enzyme is much larger than the inhibitor or any component of the library, the ioniza tion of the enzyme inhibitor or enzyme library complex should be very similar to that of the free enzyme. Once these conditions were explored, a known inhibitor was incubated with the enzyme to study the co mplexation and mass spectrometric parameters. Tand em MS of the enzyme inhibitor was performed. Nucleic Acid Nomenclature To reduce confusion, a brief review of nucleic acid nomenclature will be presented here. There exists a good textbook chapter reviewing this basic information.108 Nucleic acids are the basic units of ribonucleic acid (RNA) and deoxyribonucleic acid (DNA). Starting with the smaller units of nucleic acids, there are five major bases (or nucleobases) represented in nature, divided into two cl asses. Pyrimidine bases are six membered rings w hile purine bases are comprised of connected five and six membered rings. The numbering system for these rings is shown in Figure 31. Purine bases (adenine and guanine) are found in both DNA and RNA. Of the pyrimidine bases, cytosine is found in both DNA and RNA, while uracil occurs only in RNA and thymine is limited to DNA. The next component to consider is ribose. While free ribose interconverts between a linear structure and a ring, it adopts a furanose ring structure in nucleic acids. In RNA, the 2 and 3 position s are occupied by hydroxyl groups For DNA, the 2 site is a hydrogen. Binding between nitrogen 1 of a nucleobase and the 1 position of the ribose ring results in a nucleoside. Addition
78 of one, two, or three phosphate groups at the 5 posit ion of the ribose ring gives nucleotide mono, di and triphosphate structures, respectively. The general term nucleotide refers to any phosphorylated nucleoside. These structures are shown in Figure 32. C hains of RNA or DNA have phosphodiester linkage s between a 5 phosphate group and the 3 hydroxyl group. Phosphodiester bond formation results in the loss of water; conversely, water addition cleaves the phosphodiester bond. In nature, RNA or DNA chain formation is driven by the loss of phosphate group. Indeed biologists may recognize the nucleotide adenosine triphosphate as the energy currency ATP. Systems of Interest Ribonuclease A (RNase A) is a digestive enzyme secreted by the pancreas. It is composed of 124 amino acids (sequence shown in F igure 33) including eight cysteine residues.109 The cysteine residues link to form four disulfide bridges in the active enzyme.110 The natively folded, nonreduced protein has a monoisotopic molecular weight of 13,696 Da.105 In the body, RNase A catalyzes the hyd rolysis of the RNA phosphodiester bond at the 3 side of pyrimidine bases (cytidine and uridine) by the mechanism shown in Figure 34.1111 12 Histidine 12 acts as a general base, abstracting a proton from the 2 hydroxyl group of the RNA nucleobase. Acting as a general acid, His 119 protonates the leaving group (the RNA unit on the 3 end of the chain) and the phosphodiester bond between RNA nucleobases is cleaved. At this stage the nucleotide at the 5 position of cleavage adopts a transition state struct ure described as a nucleotide 2 3 cyclic monophosphate. To continue the mechanism, His 119 removes a hydrogen from a free water molecule present in the active site, with the hydroxyl group binding to the 2 3 cyclic phosphate structure. Lastly, the oxygen bound at the 2 site of the ribose ring removes the proton from His 12 to regenerate the original active site. The RNA chain now terminates in a 3monophosphate structure.
79 The specificity of this reaction is governed by several features. First, ther e are a number of amino acid residues along the binding pocket which are responsible for binding the RNA chain. Among these residues are Thr 45, Asp 83, Phe 120, Asn 71, Glu 111, Lys 7, Arg 10, and Gln 11.1 13 These are responsible for creating a steric and chemical environment specific for RNA chains. Discrimination between RNA chain sequences is weak; however, there is some specificity to place a C or U in the active site where its 3 side of the phosphodiester bond can be cleaved.1 131 1 4 Another important residue is Lys 41, responsible for transition state stabilization.1 13 Lastly, the cleavage mechanism is specific for RNA and will not cleave DNA. The reason for this is apparent from the mechanism. Since a 2 hydroxyl group is required for the reaction, a nd DNA has a 2 hydrogen, the reaction cannot proceed to cleave DNA. This system is particularly well suited for study via ESI FTICR mass spectrometry. R ibonuclease A is one of the most studied proteins, with a well characterized structure .6, 112127 Havi ng this information is advantageous, in the sense that potential protein fragments can be identified. Furthermore, the sequence contains several basic sites, making RNase A suitable for protonation in positive mode electrospray ionization. Lastly, the abundant literature on RNase A suggests that it will be stable in most electrospray solutions and quite possibly maintain its structure into the gas phase. Here, the disulfide bridging will be helpful.6, 1 1 5 1 21 Uridine 5 monophosphate (UMP) is a known weakly binding inhibitor of RNase A, with a 6 Its chemical structure is shown in Figure 35. Binding between UMP and RNase A can occur specifically with the UMP inside the RNA binding pocket, or via numerous nonspecific binding interactions along the enzyme surface. O ptimizing Conditions for RNase A Solutions The following electrospray conditions were optimized for RNase A: RNase A concentration, organic phase composition, and acid modifier composition. Since RNase A is
80 significantly larger than UMP, the optimal conditions for the RNase A UMP complex should be relatively close to those of the free RNase A. There are several spectral features to be considered when optimizing. First, the spectrum should contain RNase A at high abundance. Second, in order to maintain a na tively folded protein with intact active site, the conditions should be as mild as possible. In solution, this refers to low organic phase and acid concentrations. It is also important to keep temperatures and voltages as low as possible. The spectral sign ature of a folded protein is a narrow charge state distribution with relatively few charges and no degradation products. Third, it is beneficial to have as few non RNase A peaks in the spectrum as possible. This includ es reducing the number solvent and sal t adduct peaks, as well as contaminants. Ideally, the RNase A peaks will show no adducts, but this is not always feasible. Shifts in the charge state were quantitatively determined by calculating the average charge state. From the spectrum, the peak area for e ach charge state was measured using the MIDAS data analysis software. The peak area for each charge state was divided by the value of the charge to correct for the effect of ion charge on the image current; e.g., the +7 charge state peak area was divi ded by seven. This was termed the corrected peak area. Corrected peak areas for all charge states were summed. The corrected peak area for each charge state was in turn divided by the sum for all charge states; this gave the normalized contribution for eac h charge state. Normalized contributions for each charge state were multiplied by the charge and summed to give the corrected average charge state. Figures 36 through 3 9 are plots for the optimization of certain parameters. Each experiment comprised fifty averaged scans; the experiments were repeated three times and averaged. The error bars represent two standard deviations of the mean in each direction. When
81 these error bars do not overlap, the difference between points is statistically significant. Fig ures 310 through 312 are also optimization plots. Each point represents a single experiment of fifty averaged scans. The two most commonly used organic phases in electrospray ionization are acetonitrile and methanol. The ratio of organic phase to water i s governed by several features. Most importantly the solvent mixture must be capable of dissolving the analyte. In addition, the stability of the Taylor cone and the operating range for the current voltage diagram of electrospray will be governed by the am ount of organic phase. Highly stable and easily controlled Taylor cones occur most frequently when the organic phase and water are present in roughly equal parts. Beginning with equal parts water and organic solvent mass spectra were obtained for RNase A with acetonitrile and methanol. All other experimental parameters were the same. The spectrum for acetonitrile shows a slightly broader charge state envelope at slightly lower m/z than the spectrum for methanol. Based on this, methanol was chosen as the organic phase for the remainder of experiments. Next, the relative concentration of methanol was studied. The results of varying this concentration as well as that of 50% acetonitrile, are shown in Figure 36. There is a statistical ly significant differen ce in the average charge states for acetonitrile and methanol. However, over the range of 20% to 80% methanol, there is no significant change in the average charge state. More extreme organic phase compositions were tested, but were judged impractical due to instability of the electrospray current. For practicality, 50% organic phase was used for experiments. Next, the effect of acid modifier concentration was studied. In the literature, values of 0.5% 2.0% acid are typical for electrospray. R ibonuclease A is stable at a wide range of pHs as
82 indicated by enzyme activity and structure.6, 115 1 21 Figure 3 7 shows the effects of increasing acid modifier on the average charge state. A denatured or degraded protein would have a higher average charge state than an intact protein. There seems to be no significant change over the range 0.05% to 5% formic acid. However, signal stability and abundance are poorer and less stable at extremely low acid contents. Based on this, a value of 0.5% formic acid was chosen for t he remainder of the experiments. Lastly, the effect of RNase A concentration was studied. Figure 3 8 shows the effect of increasing RNase A concentration on RNase A signal abundance. Signal quality is poor for e signal abundance does not change much. At concentration on average charge state is also plotted in Figure 39. Over this concentration range, there is no difference in t he observed charge state. Aside from solution conditions, a few instrumental conditions were optimized. The effect of capillary and skimmer voltages is discussed in a later section. HMC temperature was explored with regards to RNase A signal intensity and character. Based on the nature of ESI processes, different conditions may give rise to ion populations having different internal energies. A toolow HMC temperature could result in incomplete desolvation, with analyte signal not being observed or observed in a partially solvated state. When the HMC temperature is too high, desolvation is complete but excess energy imparted to the ions can result in complex dissociation, protein unfolding, or fragmentation. The effects of HMC temperatu re on RNase A signal abundance are shown in Figure 310. At low temperatures, signal abundance is poor, and temperatures greater than 80C are required to observe any signal. Over the range of 100 to 120C, signal abundance plateaus. Greater
83 temperatures were not explored for several reasons. First, the mass spectra indicate that solvent adducts are not present which suggests that desolvation is complete at these temperatures. Second, an examination of Figure 311 shows that the average charge state is relatively stable over this temperature range, although a slight decrease might occur. Since the excess energy imparted would be expected to unfold the protein and increase the average charge state, the increasing temperature is not thought to be excessive. Considering the abundance of each charge state versus HMC temperature ( Figure 312), increasing the temperature does not appear to change that of the +6, +8, or +9 states. However, the most abundant charge state, the +7, does show a dramatic increase at higher temperatures, wh ich would change the values for average charge state. Optimizing Conditions for RNase A UMP Complexes Discussion of the optimization of m ost solution conditions was included in the previous section. T he concentration of UMP was next optimized to provide t he maximum amount of observable complex. Since each RNase A molecule is capable of binding several UMP molecules, the concentration of UMP was always at least as large, and usually several times greater than the RNase A concentration. Solutions containing UMP at concentrations between were prepared and mass spectra were obtained. Figure 3UMP. At this concentration ratio, no RNase A UMP complex is observed. However, there are other peaks present, identified as protonated UMP clusters. For example, the peak at m/z 973 corresponds to three UMPs with a single proton. At m/z 1459, is the doubly charged nine UMP cluster with two protons. At relatively high UMP concentrations (compared to RNase A), these clusters are observed. In addition to the protonated species, UMP clusters are also observed as sodiated or
84 potassiated cations ( m/z 22 and 39 higher than the protonated specie, respectively) These adducts will not be labeled in the mass spectra shown. Additionally, clusters may associate with three or more protons. However, the abundance of these more highlycharged species is low so they will not be discussed in this work. The extent to which UMP clustering is obs erved can be explained by hydrogen bonding opportunities. Significant reductions in UMP concentrations must be made before an RNase A UMP complex is observed. Figure 314 shows a mass spectrum for a solution containing A and 100 In this spectrum, and throughout spectra in the dissertation, the le tter R will be used to indicate RNase A. While UMP clusters are observed, RNase A UMP complex is also present. Looking at the most abundant RNase A charge sta te (+7), complexes comprised of RNase A with one, two, three, and four UMP molecules are present. Furthermore, the most abundant species is not the free enzyme, but RNase A with a single UMP bound. However, signal to noise is relatively poor and the absol ute abundance of any single form of complex is relatively weak. Competition between RNase A UMP complexes and UMP clusters for available protons in solution may reduce the amount of RNase A UMP complex transferred to the gas phase. This may be supported fu rther by the lack of higher RNase A charge states typically observed at these solution conditions. If the UMP concentration is reduced further, the stoichiometry of the RNase A UMP complexes is reduced further. As seen in Figure 315, only one or two UMP molecules bind t o significantly reduced compared to the spectrum in Figure 314. The enzyme is also observed in its +8 and +9 charge state, which was not the case for the solutions with greater UMP concentrations, supporting the idea of competing protonation pathways.
85 In Sou r ce CID of Complexes The fundamental processes of in source CID were discussed in Chapter Two. In the relatively high pressures of the source region, collisions betwe en analyte ions and gas molecules occur. By adjusting the voltages applied to source lenses the capillary and skimmer for the purposes of this dissertationthe energetics of the collisions can be controlled as a consequence of the electric field by which t he ions are accelerated. To some degree, these processes are always us ed both to improve desolvation and to guide the ions efficiently. S ome particular limitations apply to perform ing i nsource CID on the instrument as described in Figure 213. The experi ment can best be performed as ions exit the HMC and enter the skimmer. In positive mode, the capillary voltage can theoretically be set from 0 to 300V, while the skimmer can be set 20 to +20V. However, not every combination of these voltages permits effective analyte ion generation or transmittance. Introduction of gas beyond what exists from the ESI process is not possible. Furthermore, all species which pass through the HMC from the electrospray are subject to the same conditions and no selectivity of re actions at this stage can be achieved. For the in source CID experiments described here the term voltage difference will be reported as a measure of the energy involved. This term can be obtained by subtracting the value of the voltage applied to the skimmer ( in volts ) from the value of the voltage applied to the HMC ( in volts ) In source CID was initially explored as an easy method to remove the highlyabundant UMP clusters from the solutions such as th ose shown in Figure s 313 and 314. If successful, this would simplify the mass spectrum and make it easier to monitor RNase A UMP complexation. To test the method, solutions o Figure 316 shows representative mass spectra of this solution at varying voltage differences. Singlyprotonated clusters containing between one and six UMPs can be observed, while
86 doublyprotonated c lusters can contain up to eleven. Clusters with three or more bound protons are possible, but cannot be confirmed due to low abundance. The HMC and skimmer voltages were varied, and the presence of UMP clusters mon itored and plotted in Figure 317 Conside r the example of [3UMP+H+]+. As the voltage difference increases, its abundance increases and reaches a maximum at a difference of ~150V. Beyond 150V, its abundance decreases. Each of the cluster species appears to follow a similar pattern, with the maxima occurring at different values for the voltage differences. However, the abundances for higher order clusters are relatively low and it is not possible to say that any change in abundance is significant. Yet the effects for clusters containing between one and six UMP are dramatic. Figure 318 shows how the voltage difference at maximum abundance decreases as the number of UMPs in the cluster increases. This agrees with the predicted cluster behavior. As more energy is imparted, noncovalent interactions bet ween molecules in the cluster are disrupted, resulting in the sequential loss of UMPs. Surprising though, are the bond strength s of the lower order clusters. Those containing between one and three UMPs are present at the most extreme voltage differences achievable for this instrumentation. This leads to the original idea of using in source CID to remove UMP clustering and promote RNase A UMP complexation and demonstrates that it is not feasible. While clustering can be greatly reduced by this method, it ca nnot be fully achieved and the voltage differences necessary to achieve substantial reductions in clustering are quite severe. Furthermore, using large voltages to reduce UMP clustering did not result in signal for RNase A UMP complexes. Using insource C ID to dissociate RNase A UMP complexes was also explored. The abundance of the complex was monitored as a function of increasing voltage differences.
87 Initially, th ese data looked promising ( Figure 3 19) As the voltage difference increases, the abundance o f the complex decreases. This pattern would be observed if the complex dissociated to its discrete components. However, these data alone are insufficient to support the proposed behavior. If the same experiment is performed on the free RNase (also shown in Figure 319), one can see that the free enzyme abundance decreases in much the same way The reason for the free enzyme abundance decrease cannot be ascertained from these data; it may be the result of poor ion steering outside this range or incompatible degrees of desolvation. The behavior of the complex, however, can be attributed to the behavior of the enzyme. Since the enzyme is much larger than the inhibitor, the behavior of the enzyme is the primary guide of the complex behavior. Fixed Wavelength IR MPD of Complexes As the previous section showed, insource CID is not a feasible method for dissociation of the complexes of interest. Infrared multiple photon dissociation (introduced in Chapter Two) was next explored as a means to achieve complex dissoci ation and fragmentation. A fixed2 laser was available for these experiments. IRMPD of UMP Clusters In source CID was insufficient to completely dissociate the observed UMP clusters. However, using the setups available, IRMPD could potentially provide greater energy to the system and therefore result in complete fragmentation. Although this experiment is not directly related to the main project, there is valu e in its execution. First, since UMP has the same nucleic acid structure as t he library components, fragmentation similar to that desired for the library components might arise and be identified. Second, it was hoped that this study would demonstrate the effectiveness of IRMPD for fragmentation of nucleic acid structures.
88 The IRMP D experiments were performed using a constant power and variable irradiation time. All species in the cell were irradiated over this period. Sample mass spectra at various irradiation times are shown in Figure 320. As in the CID experiments, the cluster a bundance is plotted as a function of imparted energy. In comparing IRMPD ( Figure 321 and 322) with CID ( Figure 3 17 and 318), the dissociation patterns show qualitative agreement. Higher order clusters are disfavored as more energy is applied. However, IRMPD is able to successfully remove all UMP clusters. The irradiation times necessary are relatively small (~ms), but this is related to the laser power employed (5 W). Lower powers would require longer irradiation times. As the UMP clusters dissociate into singl y protonated UMP molecules ( m/z 325), additional energy imparted by IRMPD results in covalent bond cleavage. Various fragments arise as a function of the irradiation time, with two of these shown in Figure 323. In this figure, the abundance of protonated UMP is initially low, but rises rapidly as the ions are irradiated. This increase is due to larger clusters producing this species as a result of dissociation. As the irradiation times lengthen and more energy is imparted, the abundance of the pr otonated species declines and eventually goes to zero. This disappearance can be attributed to fragmentation. Indeed, the abundance of fragments increases over this period of UMP decline. Of the two fragments shown, m/z 213 appears at lower energies. It re aches a maximum and begins to decline as the abundance of m/z 97 increases. The abundance of m/z 97 plateaus at longer irradiation times, suggesting that it is a stable fragment. While it cannot be definitively stated, these curves are suggestive of a path way in which m/z 325 fragments to give m/z 213 which fragments to give m/z 97. Figure 324 shows the proposed pathway for fragmentation of protonated UMP ( m/z 325) Neutral base loss results in m/z 213. Additional energy results in the loss of the phosphat e group
89 and generation of m/z 97. This structure, the oxocarbenium cation, has been reported previously as a product of nucleotide fragmentation by CID.128 Support for the proposed pathway can be gleaned from the fragmentation of the sodiated UMP (top sche me, Figure 325). The m/z 347 has two possible structures with the same exact masses. Consider the first case, where UMP is associated with a sodium cation. If this structure undergoes the same fragmentation pathway as in Figure 3 24, m/z 213 and 97 are obtained. However, the other possible structure for m/z 347 is protonated UMP where one H on the phosphate group has been substituted with a sodium atom. If this species undergoes neutral base loss followed by phosphate group loss (bottom scheme, Figure 325 ), the resulting structures are m/z 235 and 97. A peak at m/z 235 is in fact observed in the spectrum (Figure 3 20, bottom left). Since all three pathways terminate in the oxocarbenium cation, this explains why m/z 97 has a greater abundance than the other species. T he proposed fragmentation pathways are consistent with the results in Figure 3 23. IRMPD of RNase AUMP Clusters For this experiment, the goal was to demonstrate the utility of fixed wavelength IRMPD for dissociating the RNase A UMP complex to g enerate unbound UMP related ions. The exper imental design generated here was the basis for later experiments involving combinatorial library complexes. As shown in the previous section, irradiating UMP ( m/z radiation will generate abundant fragments at m/z 213 and 97. Dissociation of the RNase A UMP complex to produce m/z 325 was taken as indicative of a successful experiment. Observing fragments m/z 213 or 97 in the absence of m /z 325 indicated that IRMPD was taking place for this experiment but cause d concern for later experiments. When only one nucleic acid species is present in the solution, the putative source of these ions is known. However, if several similar
90 structures are present, they may produce the same fragments, and thus the origin of the fragments cannot be known. A solution containing 10 used to generate the mass spectrum shown on the left in Figure 326. Ribonuclease A UMP complexation is observed in the region m/z 19502500; this range was isolated using a SWIFT waveform to remove any free UMP present. The m ass spectrum indicates that SWIFT isolation was successful, and thus any fragments appearing in the MS MS spectrum must be the result of complex dissociation. IRMPD was performed at a fixed wavelength of 10.6 i rradiation time of 1.0 s gave the best results; this MS MS spectrum is shown on the right in Figure 326. Irradiation gives rise to several subtle changes in the mass spectrum of the complex. The most abundant species becomes the free enzyme. Additionally, a shift to less highly charged complex seems to be the case, although the statistical significance of this has not been tested. Lastly, the complexes, on average, contain fewer bound UMP molecules. Photodissociation of the complex does generate protonated UMP at m/z 325; however, the absolute abundance of this ion is relatively weak and the spectrum has a poor signal to noise ratio. The complex is persistent and is present even as the UMP daughter ion appears. Imparting additional energy (increasing laser power or irradiation time) further decreases the complex abundance but surprisingly does not increase the abundance of protonated UMP beyond what is shown. One other point of interest is that significantly more energetic IRMPD conditions are necessary to dissociate the RNase A UMP complex compared to the UMP clusters. This indicates that the interactions between enzyme and inhibitor are significantly stronger than those interactions within clusters.
91 Infrared multiple photon dissociation of the RNase A-UMP complex was nominally successful with regards to the original goals. Yet the issues of poor daughter ion abundance and poor MS-MS signal-to-noise are cause for conc ern about possible success when the method was applied to the more complex librar y project. Another possible concern is that, even if significant amounts of daughter nucleotide fragments can be produced, the ultimate products of the pathways shown in Figures 3-24 and 3-25 do not include the base group. Thus, it might not be possible to identify the original binding nucleotides. These issues were addressed in two ways. First, the method will be repeated with the RNas e A-library system to confirm the expectations. Second, IRMPD of those complexes was performed at several different wavelengths in the hope of improved fragmentation efficiency and an alternate fragmentation pathway. 1 2 3 4 5 6Purine 1 2 3 4 5 6 7 8 9Pyrimidine Ribose 1 23 4 5 1 2 3 4 5 6 1 2 3 4 5 6Purine 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9Pyrimidine Ribose 1 23 4 5 1 23 4 5 Figure 3-1. Numbering system for nucleic acid rings. Pyrimidine and purine form the basis of the RNA and DNA bases. Five-membered ribos e rings can bind the bases to form nucleosides or nucleotides.
92 O OH OH O BaseP O OH HO O OH OH O BaseP O OH O P O OH O P O OH HO Base Base Nucleoside Nucleotide Monophosphate Nucleotide Diphosphate Nucleotide Triphosphate O OH OH O BaseP O OH HO O OH OH O BaseP O OH O P O OH O P O OH HO Base Base Nucleoside Nucleotide Monophosphate Nucleotide Diphosphate Nucleotide Triphosphate Figure 3-2. Nucleic ac id nomenclature. VSADFHVPVY PNGE CAVIIH KQANTTKY CA NPYKSGTSER CDTISMTSYS QY CQNTGNK C AVNKQSC VAQ VDALSEHVFT NVPK CRDKTK NRSKMMQN C Y NSSSAASMHN SSTEREFKAA ATEK 30 60 90 120VSADFHVPVY PNGE CAVIIH KQANTTKY CA NPYKSGTSER CDTISMTSYS QY CQNTGNK C AVNKQSC VAQ VDALSEHVFT NVPK CRDKTK NRSKMMQN C Y NSSSAASMHN SSTEREFKAA ATEK 30 60 90 120 Figure 3-3. Sequence of bovine pancr eatic Ribonuclease A (RNase A).109 There are 124 amino acids, among which are eight cysteine residu es (in red) that fo rm four disulfide bridges.
93O O O H H O P O OH O H H O HO O H N NH N N H H Base Base His12 His119 O O O H H O O OH O H H O N NH N N H Base Base His12 His119H P O OH H O O O H H O N NH N N H Base His12 His119H P O OH O H H O OH O H H O Base H O OH O H H O N NH N N H Base His12 His119 H P HO O OH O OH O H H O Base H 5' 3' + H2OO O O H H O P O OH O H H O HO O H N NH N N H H Base Base His12 His119 O O O H H O O OH O H H O N NH N N H Base Base His12 His119H P O OH H O O O H H O N NH N N H Base His12 His119H P O OH O H H O OH O H H O Base H O OH O H H O N NH N N H Base His12 His119 H P HO O OH O OH O H H O Base H 5' 3' + H2O Figure 3-4. Mechanism for RNase A-catalyzed hydr olysis of RNA. Structures shown in purple are histidine residues in the RNase A active site. Figure adapted from Voet, D.; Voet, J.G.; Pratt, C.W. Fundamentals of Biochemistry, Upgrade Edition John Wiley and Sons: New York, 2002. 292-293.
94 Figure 3-5. Structure of urid ine-5-monophosphate (UMP). 6.5 7 7.5 8 8.5 9 0102030405060708090 MeOH ACN Corrected Average Charge StateOrganic Phase Concentration (%) 6.5 7 7.5 8 8.5 9 0102030405060708090 MeOH ACN Corrected Average Charge StateOrganic Phase Concentration (%) Figure 3-6. Effect of organic phase on average charge state for RNase A. The remainder of each solution was comprised of 10 M RNase in Milli-Q water with 0.5% formic acid. Details on statistical analysis can be found on pages 80-81. Using methanol instead of acetonitrile resulted in a lower average char ge state, which may correspond to a more folded protein. Changing methanol concen trations in the range of 20-80% did not change the average charge state.
95 6.5 7 7.5 8 8.5 9 0 1 2 3 4 5 6 Corrected Average Charge StateFormic Acid Concentration (%) 6.5 7 7.5 8 8.5 9 0 1 2 3 4 5 6 Corrected Average Charge StateFormic Acid Concentration (%) Figure 37. in 1:1 methanol: water. Details on statistical analysis can be found on pa ges 8081. There was no statistically significant difference in corrected average charge state for formic acid concentrations of 0.1 to 5.0%. 0 1000 2000 3000 4000 5000 6000 7000 0 10 20 30 40 50 60 RNase A Concentration ( M)Absolute Abundance 0 1000 2000 3000 4000 5000 6000 7000 0 10 20 30 40 50 60 RNase A Concentration ( M)Absolute Abundance Figure 38. Effect of RNase A concentration on signal abundance. Details on statistical analysis can be found on pages 8081
96 6.5 7 7.5 8 8.5 9 0 10 20 30 40 50 60 Corrected Average Charge StateRNase A Concentration ( M ) 6.5 7 7.5 8 8.5 9 0 10 20 30 40 50 60 Corrected Average Charge StateRNase A Concentration ( M ) Figure 39. Effect of RNase A concentration on average charge state. All samples were constituted in 1:1 methanol: water with 0.5% formic acid. Details on statistical analysis can be found on pages 8081. No significant differences in corrected average charge state was observed for this concentration range. 0 10 20 30 40 50 60 70 80 90 75 80 85 90 95 100 105 110 115 120 125 Absolute AbundanceHeated Metal Capillary Temperature ( C) 0 10 20 30 40 50 60 70 80 90 75 80 85 90 95 100 105 110 115 120 125 Absolute AbundanceHeated Metal Capillary Temperature ( C) Figure 310. Effect of increasing HMC temperature on RNase A abundance. Details on statistical analysis can be found on pages 8081. No signal was observed for temperatures less than 80C. Abundance generally increases over the range 80100C. Between 100 and 120C, the abundance plateaus.
97 6.5 7 7.5 8 8.5 9 75 80 85 90 95 100 105 110 115 120 125 Corrected Average Charge StateHeated Metal Capillary Temperature ( C) 6.5 7 7.5 8 8.5 9 75 80 85 90 95 100 105 110 115 120 125 Corrected Average Charge StateHeated Metal Capillary Temperature ( C) Figure 311. Effect of HMC temperature on RNase A average charge state. Details on statistical a na lysis can be found on pages 8081. As no increase in corrected average charge state is observed, it is unlikely that protein unfolding occurs over this temperature range. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 75 80 85 90 95 100 105 110 115 120 125 9+ 8+ 7+ 6+ Heated Metal Capillary Temperature ( C)Corrected Contribution to Average Charge State 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 75 80 85 90 95 100 105 110 115 120 125 9+ 8+ 7+ 6+ Heated Metal Capillary Temperature ( C)Corrected Contribution to Average Charge State Figure 312. Effect of HMC temperature on RNase A charge states. Details on sta tistical anal ysis can be found on pages 8081.
98 m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 Abundance [2UMP+H+]+[3UMP+H+]+[4UMP+H+]+[5UMP+H+]+[6UMP+H+]+[UMP+H+]+[5UMP+2H+]2+[9UMP+2H+]2+[11UMP+2H+]2+[3UMP+2H+]2+ 2,000 m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 Abundance [2UMP+H+]+[3UMP+H+]+[4UMP+H+]+[5UMP+H+]+[6UMP+H+]+[UMP+H+]+[5UMP+2H+]2+[9UMP+2H+]2+[11UMP+2H+]2+[3UMP+2H+]2+ m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 Abundance [2UMP+H+]+[3UMP+H+]+[4UMP+H+]+[5UMP+H+]+[6UMP+H+]+[UMP+H+]+[5UMP+2H+]2+[9UMP+2H+]2+[11UMP+2H+]2+[3UMP+2H+]2+ 2,000 Figure 313. Mass spectrum obtained whe n 50 [UMP ] = 1 [RNase A]. At relatively high UMP concentrations, UMP clustering occurs and RNase A UMP complexation is not observed. Major protonated peaks are labele d.
99 m/z 2,000 1,500 1,000 Abundance [2UMP+H+]+[3UMP+H+]+[4UMP+H+]+[R+7H+]7+[R+UMP+7H+]7+[R+2UMP+7H+]7+ [R+3UMP+7H+]7+ [R+4UMP+7H+]7+ [R+6H+]6+[R+UMP+6H+]6+ m/z 2,000 1,500 1,000 Abundance [2UMP+H+]+[3UMP+H+]+[4UMP+H+]+[R+7H+]7+[R+UMP+7H+]7+[R+2UMP+7H+]7+ [R+3UMP+7H+]7+ [R+4UMP+7H+]7+ [R+6H+]6+[R+UMP+6H+]6+ Figure 314. Mass spectrum obtained when 10 [UMP ] = [RNase A] ; R indicates RNase A. As the UMP concentration is reduced, RNase A UMP complexation is observed in the spectrum. Major peaks are labeled.
100 m/z 2,500 2,000 1,500 1,000 Abundance [3UMP+H+]+[2UMP+H+]+[R+8H+]8+[R+UMP+8H+]8+ [R+9H+]9+[R+7H+]7+[R+UMP+7H+]7+[R+2UMP+7H+]7+ [R+6H+]6+[R+UMP+6H+]6+ m/z 2,500 2,000 1,500 1,000 Abundance [3UMP+H+]+[2UMP+H+]+[R+8H+]8+[R+UMP+8H+]8+ [R+9H+]9+[R+7H+]7+[R+UMP+7H+]7+[R+2UMP+7H+]7+ [R+6H+]6+[R+UMP+6H+]6+ Figure 315. Mass spectrum obtained when 5 [UMP ] = [RNase A]. The presence of UMP clusters is further reduced, and the relative abundance of RNase A UMP complexation is increased relative to Figures 3 13 and 314. Major peaks are labeled.
101 m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 Abundance 5.8 V m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 Abundance 47.0 V m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 Abundance 139.3 V m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 Abundance 262.3 V3UMP 4UMP 5UMP 6UMP 2UMP UMP (11 UMP)+2(9 UMP)+2(5 UMP)+2UMP 2UMP 3UMP 4UMP 5UMP (3 UMP)+2(3 UMP)+23UMP 2UMP UMP (3 UMP)+24UMP UMP (3 UMP)+22UMP m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 Abundance 5.8 V m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 Abundance 47.0 V m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 Abundance 139.3 V m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 Abundance 262.3 V3UMP 4UMP 5UMP 6UMP 2UMP UMP (11 UMP)+2(9 UMP)+2(5 UMP)+2UMP 2UMP 3UMP 4UMP 5UMP (3 UMP)+2(3 UMP)+23UMP 2UMP UMP (3 UMP)+24UMP UMP (3 UMP)+22UMP Figure 316. Mass spectra for in source CID of UMP clusters. A s the voltage difference is increased, smaller m/z become more prominent in the spectrum and UMP clustering is reduced. Major peaks are labeled.
102 0 200 400 600 800 1000 1200 1400 1600 0 50 100 150 200 250 300 (UMP+H)+ (2UMP+H)+ (3UMP+H)+ (4UMP+H)+ (5UMP+H)+ (6UMP+H)+ Absolute AbundanceVoltage Difference (V)Absolute Abundance 0 200 400 600 800 1000 1200 1400 1600 0 50 100 150 200 250 300 (UMP+H)+ (2UMP+H)+ (3UMP+H)+ (4UMP+H)+ (5UMP+H)+ (6UMP+H)+ Absolute AbundanceVoltage Difference (V)Absolute Abundance Figure 317. Abundance of UMP clusters as a function of voltage difference employed for in source CID Increasing the voltage difference reduces the presence of higher order UMP clusters and favors the presence of protonated UMP. 0 50 100 150 200 250 0 1 2 3 4 5 6 7 Voltage Difference at Max (V)Number of UMPs in Cluster 0 50 100 150 200 250 0 1 2 3 4 5 6 7 Voltage Difference at Max (V)Number of UMPs in Cluster Figure 318. Relationship between cluster size and stability (CID). Larger UMP clusters are more stable at smaller voltage di fferences.
103 0 20 40 60 80 100 120 140 160 180 -12 -10 -8 -6 -4 -2 0 2 4 6 Rnase Rnase+UMP Rnase+2UMP Voltage Difference (V)Absolute Abundance 0 20 40 60 80 100 120 140 160 180 -12 -10 -8 -6 -4 -2 0 2 4 6 Rnase Rnase+UMP Rnase+2UMP Voltage Difference (V)Absolute Abundance Figure 319. In source CID of RNase A UMP complexes. While changing the voltage difference decreases the abundance of observed complex, a similar effect is observed for the free enzyme. This suggests that the disappearance of RNase A UMP comp lex is due to disfavored ionization conditions, rather than complex dissociation.
104 m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance 1 UMP 2 UMP 3 UMP 4 UMP 5 UMP 6 UMP 1 UMP 2 UMP 3 UMP 4 UMP Laser Off 0.07 ms 0.75 ms2 UMP 97.08 0.15 ms m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance 1 UMP 2 UMP 3 UMP 4 UMP 5 UMP 6 UMP 1 UMP 2 UMP 3 UMP 4 UMP Laser Off 0.07 ms 0.75 ms2 UMP 97.08 0.15 ms Figure 320. Mass spectra of UMP clusters at increasing irradiation times. UMP clusters progressively dissociate to UMP and then undergo fragmentation. The most favored fra gment at high energies is m/z 97. Major peaks are labeled.
105 0 10 20 30 40 50 60 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 (5UMP+H+) (4UMP+H+)+ (3UMP+H+)+ (2UMP+H+)+ (UMP+H+)+ Absolute AbundanceIrradiation Time (ms)Absolute Abundance 0 10 20 30 40 50 60 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 (5UMP+H+) (4UMP+H+)+ (3UMP+H+)+ (2UMP+H+)+ (UMP+H+)+ Absolute AbundanceIrradiation Time (ms)Absolute Abundance Figure 321. Abundance of UMP clusters as a function of IRMPD irradiation time Increasing irradiation time results in large clusters dissociating to produce smaller clusters. These clusters can be completely dissociated to produce protonated UMP. 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 1 2 3 4 5 6 7 Irradiation Time at Max ( ms)Number of UMPs in Cluster 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1 0 1 2 3 4 5 6 7 Irradiation Time at Max ( ms)Number of UMPs in Cluster Figure 322. Relationship between cluster size and stability (IRMPD). Larger UMP clusters are more stable at shorter irradiation times.
106 0 20 40 60 80 100 120 00.20.40.60.818.104.22.168 325 213 97 AbundanceIrradiation Time (ms) 0 20 40 60 80 100 120 00.20.40.60.822.214.171.124 325 213 97 AbundanceIrradiation Time (ms) Figure 3-23. Stability of most abundant IRMPD fragments. Here, m/z 325 is protonated UMP and m/z 213 and 97 are fragments. The absolute abundance of m/z 97 is plotted here, while the absolute abundances of m/z 325 and 213 are multiplied by a factor of 1.5 and 3, respectively, for better visualization. m/z 325 m/z 213 m/z 97 m/z 325 m/z 213 m/z 97 Figure 3-24. Proposed fragmentation pathway for m/z 325. The abundances of selected IRMPD fragments as a function of irradia tion time are shown in Figure 3-23. The m/z 325 can fragment to m/z 213. Loss of water from m/z 213 is present in the mass spectrum at m/z 195. Neutral loss of the base gives m/z 213, observed in the Figure 3-20 spectrum (0.15 ms). Further fragmentation leads to phosphate group lo ss, generating the m/z 97 oxocarbenium cation.
107NH O O N O OH OH H H H H O P HO OH O Na m/z 347O OH OH H H O P OH O HO m/z 213O OH H H m/z 97 m/z 347O OH OH H H O P OH O O Na m/z 235 m/z 97 A BNH O O N O OH OH H H H H O P HO OH O Na m/z 347O OH OH H H O P OH O HO m/z 213O OH H H m/z 97 m/z 347O OH OH H H O P OH O O Na m/z 235 m/z 97 NH O O N O OH OH H H H H O P HO OH O Na m/z 347O OH OH H H O P OH O HO m/z 213O OH H H m/z 97 NH O O N O OH OH H H H H O P HO OH O Na m/z 347O OH OH H H O P OH O HO m/z 213O OH H H m/z 97 m/z 347O OH OH H H O P OH O O Na m/z 235 m/z 97 m/z 347O OH OH H H O P OH O O Na m/z 235 m/z 97 m/z 347O OH OH H H O P OH O O Na m/z 235 m/z 97 A B Figure 3-25. Proposed fragmentation pathway of m/z 347. There are two possible structures (with the same exact mass) for m/z 347. A) Instead of the proton shown in Figure 3-24, UMP is charged by a sodium cation. Aside from the initial sodiated species, fragmentation follows an identical pathwa y to Figure 3-24. B) As in Figure 3-24, UMP is protonated but one of the OH groups of the phosphate has been substituted to O-Na. This substitution is maintained, giving rise to m/z 235 (observed in Figure 320) instead of m/z 213. Since the phosphate group is lost, the m/z 97 oxocarbenium cation is the final product just as in the previous two pathways.
108 A B A B Figure 3-26. Mass spectra for IRMPD of RNase A-UMP complexes. A) Significant complexation in the range m/z 1950-2500 before SWIFT. This range is isolated using SWIFT, and irradiated to produce B).
109 CHAPTER 4 MULTIPLE WAVELENGTH IRMPD FOR COMPARISON OF SPECIFIC AND NONSPECIFIC NONCOVAL ENT INTERACTIONS This chapter describes the multiple wavelength IRMPD of two complexes. One is the specifically bound complex between RNase A and dynamic combinatorial L ibrary 1 (described below) Myoglobin nonspecifically binding the same library is the second complex. This set of experiments was performed for two distinct reasons. First, in Chapter Three, dissociation of the RNase A Then, in the work reported in this chapter, IRMPD at other wavelengths was explored to determine if better dissociation could be achieved. Second, the dissociation of specific and nonspecific complexes was compared. Systems of Interest As in Chapter Three, complexes of bovine pancreatic RNase A wer e studied. Instead of using a known inhibitor, a dynamic combinatorial library was used. This library (Library 1, shown in Figure 41) is comprised of uridine and adenosine 2 3 cyclic monophosphates. The behavior of the library will be discussed in depth in Chapter Five. In this chapter, library binding will be described solely in terms of nucleotides without regards to the ir identities. The nucleotides specifically bind to RNase A, with the possibility of multiple library members binding to a single RN ase A molecule. For comparison, the library was also s tudied with an enzyme to which it wa s not expected to interact specifically. For such a comparison, t he enzyme selected should be as close as possibl e in size and gas phase basicit y to RNase A and will ideally be well studied in the gas phase. Equine heart myoglobin, sequence shown in Figure 42, was selected as a control enzyme for these experiments.129 1 30 It is slightly larger than RNase A with a molecular weight of 16,971 Da. Like RNase A, it is wel l studied and suitable for positive mode ESI. The complexes formed
110 between myoglobin and the library serve as examples of nonspecific complexation. For these experiments, solutions contai ibrary 1 in 50:50 water: methanol with 0.5% formic acid were used. Calculations and Measured Parameters The extent of IRMPD is described in terms of percent complexation. Consider the spectrum in Figure 43. Several charge states are present, each with peaks representing the free enzyme and RNase A nucleotide complexes. In addition, there are adducts to sodium, potassium, and/or phosphate. Peak areas were calculated for each. The values for all peaks representative of RNase A nucleotide binding, even those with other adducts, were summed. When multiple charge states were present, areas were obtained for all present. This sum is termed the total complexation. To this value is added the peak area of the free enz yme and its addu cts. The sum is termed the total abundance. Dividing total complexation by total abundance and multiplying by 100 gives the percent complexation. Comparisons of Spectra To reduce complex degradation, solutions for the myoglobinand RNase A L ibrary 1 comp lexes had to be fresh; thus, solutions more than two weeks old were excluded from these studies. When not in use, solutions were stored in a freezer. Comparisons of freshlymade solutions and solutions two weeks old did not show significant differences. To control variability in the laser power and positioning, experiments for both the nonspecific and specific complexes were performed on the same day where possible. Before performin g IRMPD, the mass spectra for myoglobinLibrary 1 and RNase A Library 1 complexes were compared. That of myoglobin can been seen in Figure 44. Charge states between +12 and +21 were observed (bottom). A more detailed view of each charge state is pictured at the top of the figure. Unbound myoglobin is the most abundant peak
111 for each charge state. Additional peaks for myoglobin with one and two library nucleotides are present. Even under gentle mass spectrometry conditions, a small amount of free library is observed. For the IM RPD experiments, these peaks were ejecte d using SWIFT. Increasing the concentration of library did not result in a higher degree of complexation. Reducing the concentration decreased the amount of observed complex but did not significantly reduce the amount of free library in the spectrum. For contrast, the spectrum of RNase A Library 1 can be seen in Figure 4 5. The most abundant peaks in the spectrum are not free enzyme, but enzyme bound to library. As with myoglobin, peaks representing the binding of one and two nucleotides are present. Unlike with myoglobin, complexes with RNase A binding three and four nucleotides are observed. While free enzyme is present, its abundance is relatively low compared to RNase A Library 1 peaks. All this suggests that RNase A binds th e library more readily and t o a greater degree than myoglobin does, agreeing with the behavior expected for specific and nonspecific complexes. Several wavelengths were selected at which to perform IRMPD of the specific and nonspecific complexes. The choices were made based on sever al reasons relating to the laser properties. Wavelengths in the range of 9.174Covering the entire range was desirable. However, there are many wavelengths accessible within this span. To maximize IRMPD capabilities, the wavelengths for which maximum lasing power could be obtained were u sed. Plots of percent complexation versus irradiation time fo r various laser powers were generated These will be referred to as the IRMPD plots. Of the s the lowest laser output energy. The IRMPD plots at this wavelength can be seen in Figure 4 6. No significant difference is observed between the specific and nonspecific complexes. The highest stable power accessible at this wavelength
112 was ~1.0W. Although not shown on the plots, irradiation times in excess of ten seconds at maximum power did not produce measurable complex dissociation. Figure 47. Substantially greater laser powers are available at this wavelength, compared to W, compared to 1.0 W). Looking at similar powers betwe en the two wavelengths, the behaviors of the complexes are similar with little dissociation. For myoglobinL ibrary 1 complexes, little dissociation is seen at 2.0W. As the power is increased beyond 2.0 W, complexation drops off rapidly as a funct ion of irr adiation time. Even 1.0 s irradiation time by the laser produces complete dissociation when a power equal to or greater than 3.0 W is used. The complexation of RNase A with the library is more persistent, though it too declines with increasing laser power. At 4.0W and 5.0W, the RNase A Library 1 complex can be completely dissociated on a short time scale. At this wavelength, the specific complexes are more persistent and require greater energy to completely dissociate. I nfrared multiple photon dissociation of the complexes was next perfo ( Figure 4 8). At this wavelength the maximum laser output power is 3.0 W. The behavior of the specific and nonspecific complexes is very similar. In both cases, some dissociation can be achieved with maximum power at relatively long irra diation times. Figure 4 9. For myoglobin, low power (0.5W) at short irradiation times dramatically reduced the amount of complex present. The RNase A Library 1 complex proved more resilient. Laser power s greater than 1.0W were necessary to reduce the degree of complexation. However, at a laser power of 2.0W, the percent complexation declined smoothly as a function of irradiation time. Under the same conditions, the myoglobin complex was not observa ble
113 The last wavelength studied was 9.192 10 shows that neither the RNase A nor the myoglobin complexes readily dissociated at powers less than 2.0W. At a power of 2.0W, both types of complexes readily declined in abundance. Complete dissociation was achievable at this power. One of the reasons for these experiments was to improve the dissociation efficiency over that reported in Chapter Three with the eventual goal of using the determined IRMPD conditions to study the combinatorial libraries. For this purpose, there are several desir ed characteristics. First, the chosen wavelength must have sufficient energy to completely dissociate the complex. Second, because the energy imparted by the laser is a function of both laser power and irradiation time, greater available power can be used to reduce the necessary irradiation times. This reduces the time required to complete an experiment. Third, varying the irradiation times and laser powers should allow a degree of control with regard to complex dissociation. If a complex does not dissociat e ever or if it dissociates too rapidly, the dissociation and fragmentation behavior may be skewed. Controlling the degree of dissociation allows more information to be gleaned from the complex. For RNase A Library 1 complexes, these experiments show that dissociation efficiency is strongly dependen t on the wavelength employed for IRMPD. Of the five wavelengths used for the IRMPD experiments whose results are shown in Figures 4 6 through 4 10, the laser produced sufficient energy for complete complex disso ciation for three of them This eliminate d two wavelengths for practical use in the experiments reported in Chapter Five. Of the three remaining wavelengths, 9.588 we re similar with respect to the accessible laser power and degree of dissociation. At t was required to completely dissociate the complex, and it was thus also eliminated as a possible irradiati on
114 wavelength. Since it was most convenient to use a single w experiments discussed in Chapter Five. These considerations are summarized in Table 4 1. The other reason for these experiments was to compare the behavior of specific (RNase A L ibrary 1) and nonspecific (myoglobinL ibrary 1) complexes. It was expected that specific complexes would require greater energy to dissociate and that the nonspecific complexes would disappear more readily. At the five st udied wavelengths, the RNase A L ibrary 1 complexes required the same amount or greater power to completely dissociate. At 9.588 very little power was required to break apart the nonspecific interactions while the specific binding persisted. In none of the cases did the specific complex prove to be easier to break. It must also be pointed out that the quantitative reproducibility of this experiment is poor. However, the general patterns of behavior were consistent on a day to day basis. Using this method to determine the degree of specificity of a complex is not recommended, although it may be used as evidence to support a partic ular mode of binding. Adenosine -2 ,3 -cyclic monophosphate Exact mass: 329.053 Uridine 2 ,3 -cyclic monophosphate Exact mass: 306.025 Adenosine -2 ,3 -cyclic monophosphate Exact mass: 329.053 Uridine 2 ,3 -cyclic monophosphate Exact mass: 306.025 Figure 41. Structures of L ibrary 1 components. U ridin e 2 3 cyclic monophosphate and a denosine 2, 3 cyclic monophosphate comprised Library 1.
115 GLSDGEWQQV LNVWGKVEAD IAGHGQEVLI RLFTGHPETL EKFDKFKHLK TEAEMKASED LKKHGTVVLT ALGGILKKKG QSHATKHKIP IKYLEFISDA IIHVLHSKHP GDFGADAQGA MTKALELFRN DIAAKYKELG FQG1 31 61 91 121GLSDGEWQQV LNVWGKVEAD IAGHGQEVLI RLFTGHPETL EKFDKFKHLK TEAEMKASED LKKHGTVVLT ALGGILKKKG QSHATKHKIP IKYLEFISDA IIHVLHSKHP GDFGADAQGA MTKALELFRN DIAAKYKELG FQG1 31 61 91 121 Figure 42. Sequence of equine heart myoglobin.1 26 m/z 1,950 1,900 1,850 1,800 1,750 1,700 1,650 1,600 1,550 1,500 R R+1N R+2N R+3N R+4N R R+1N R+2N R+3N R+4N +9 CS +8 CS m/z 1,950 1,900 1,850 1,800 1,750 1,700 1,650 1,600 1,550 1,500 R R+1N R+2N R+3N R+4N R R+1N R+2N R+3N R+4N +9 CS +8 CS Figure 43. Calculating p ercent complexation for a RNase A L ibrary 1 mass spectrum. Here, R is RNase A and N is a nucleotide. The R+nN peak areas for the +9 and +8 charge states are summed. This value is divided by the sum of it and the areas for the +8 and +9 R peaks. Multiplying by 100 gives the percent complexation.
116 m/z 1,050 1,000 950 900 900 950 1000 1050 m/z M M M M+NMP M+NMP M+NMP M+2NMP M+2NMP M+2NMP M+2NMP +19 +18 +17 m/z 1,050 1,000 950 900 900 950 1000 1050 m/z m/z 1,050 1,000 950 900 900 950 1000 1050 m/z M M M M+NMP M+NMP M+NMP M+2NMP M+2NMP M+2NMP M+2NMP +19 +18 +17 m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance +19 +18 +20 +21 +17 +16 +15 +14 +13 +12 m/z 297.84 m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance +19 +18 +20 +21 +17 +16 +15 +14 +13 +12 m/z 297.84 Figure 44. Mass spectrum showing c omplexation between myoglobin and L ibrary 1. Myoglobin is labeled as M, and NMP is a nucleotide monophosphate. Unbound myoglobin is the most abundant specie s with decreasing ab undance for myoglobinL ibrary 1 and myoglobintwo L ibrary 1 species. The +19, +18, and +17 charge states are shown in the top figure; the pattern is repeated for other charge states in the full mass spectrum at bottom.
117 m/z 1,950 1,900 1,850 1,800 1,750 1,700 1,650 1,600 1,550 1,500 1,500 1,550 1,600 1,650 1,700 1,750 1,800 1,850 1,900 1,950R+2NMP R R+NMP R+3NMP R+4NMP R+2NMP R R+NMP R+3NMP R+4NMP +9 +8m/z m/z 1,950 1,900 1,850 1,800 1,750 1,700 1,650 1,600 1,550 1,500 1,500 1,550 1,600 1,650 1,700 1,750 1,800 1,850 1,900 1,950R+2NMP R R+NMP R+3NMP R+4NMP R+2NMP R R+NMP R+3NMP R+4NMP +9 +8m/z Figure 45. Mass spectrum showing c omplexation between RNase A and Library 1 Two charge statesthe +9 and +8 are present. In each case, the most abundant peak is that of RNase A bound to a single nucleotide. Each complex peak also has a corresponding phosphate adduct peak to its right (unla beled on spectrum). Progressively decreasing amounts of RNase A bound to two, three, and four library nucleotides are present. A small amount of free RNase A is also present.
118 0 5 10 15 20 25 30 35 40 0 2 4 6 8 10 12 0.5W 1.0W Irradiation Time (s)% Complexation 0 5 10 15 20 25 30 35 40 45 0 2 4 6 8 10 12 0.5W 1.0W Irradiation Time (s)% ComplexationMyo RNase 0 5 10 15 20 25 30 35 40 0 2 4 6 8 10 12 0.5W 1.0W Irradiation Time (s)% Complexation 0 5 10 15 20 25 30 35 40 0 2 4 6 8 10 12 0.5W 1.0W Irradiation Time (s)% Complexation 0 5 10 15 20 25 30 35 40 45 0 2 4 6 8 10 12 0.5W 1.0W Irradiation Time (s)% Complexation 0 5 10 15 20 25 30 35 40 45 0 2 4 6 8 10 12 0.5W 1.0W Irradiation Time (s)% ComplexationMyo RNase Figure 46. Percent complexation versus irradiation time for m. Neither complex shows appreciable dissociation at the accessible laser powers.
119 0 10 20 30 40 50 60 70 0 2 4 6 8 10 12 0.5W 1.0W 2.0W 3.0W 4.0W % ComplexationIrradiation Time (s) 0 10 20 30 40 50 60 70 80 90 100 0 2 4 6 8 10 12 0.5W 1.0W 2.0W 3.0W 4.0W 5.0W % ComplexationIrradiation Time (s)RNase Myo 0 10 20 30 40 50 60 70 0 2 4 6 8 10 12 0.5W 1.0W 2.0W 3.0W 4.0W % ComplexationIrradiation Time (s) 0 10 20 30 40 50 60 70 0 2 4 6 8 10 12 0.5W 1.0W 2.0W 3.0W 4.0W % ComplexationIrradiation Time (s) 0 10 20 30 40 50 60 70 80 90 100 0 2 4 6 8 10 12 0.5W 1.0W 2.0W 3.0W 4.0W 5.0W % ComplexationIrradiation Time (s) 0 10 20 30 40 50 60 70 80 90 100 0 2 4 6 8 10 12 0.5W 1.0W 2.0W 3.0W 4.0W 5.0W % ComplexationIrradiation Time (s)RNase Myo Figure 47. Percent complexation versus irradiation time for Both complexes can be completely dissociated. However, myoglobin complexes require less po wer for complete dissociation to occur. At maximum laser output power, dissociation is fast, with 1 s irradiation time sufficient for complete dissociation.
120 0 5 10 15 20 25 30 35 0 2 4 6 8 10 12 0.50W 1.0W 1.5W 2.0W % ComplexationIrradiation Time (s) 0 10 20 30 40 50 60 70 80 0 2 4 6 8 10 12 0.5W 1.0W 2.0W 3.0W % ComplexationIrradiation Time (s)Myo RNase 0 5 10 15 20 25 30 35 0 2 4 6 8 10 12 0.50W 1.0W 1.5W 2.0W % ComplexationIrradiation Time (s) 0 10 20 30 40 50 60 70 80 0 2 4 6 8 10 12 0.5W 1.0W 2.0W 3.0W % ComplexationIrradiation Time (s) 0 10 20 30 40 50 60 70 80 0 2 4 6 8 10 12 0.5W 1.0W 2.0W 3.0W % ComplexationIrradiation Time (s)Myo RNase Figure 48. Percent complexation versus irradiation time for Little dissociation is observed for either complex even at the maximum laser output power
121 0 10 20 30 40 50 60 0 2 4 6 8 10 12 0.5W 1.0W 1.5W % ComplexationIrradiation Time (s) 0 5 10 15 20 25 30 35 40 45 0 2 4 6 8 10 12 0.5W 1.0W 2.0W % ComplexationIrradiation Time (s)RNase Myo 0 10 20 30 40 50 60 0 2 4 6 8 10 12 0.5W 1.0W 1.5W % ComplexationIrradiation Time (s) 0 5 10 15 20 25 30 35 40 45 0 2 4 6 8 10 12 0.5W 1.0W 2.0W % ComplexationIrradiation Time (s) 0 5 10 15 20 25 30 35 40 45 0 2 4 6 8 10 12 0.5W 1.0W 2.0W % ComplexationIrradiation Time (s)RNase Myo Figure 49. Percent complexation versus irradiation time for IRMPD at 9.588 For RNase A, laser powers of 0.5 and 1.0 W do not dissociate the complex. However, irradiation at 2.0 W gradually dissociated the RNase A Library 1 complex to free enzyme and free library. The myoglobinLibrary 1 complex was completely dissociated at 0.5, 1.0, and 1.5 W laser power s Additionally, shorter irradiation times were required for complete m yoglobinLibrary 1 dissociation compared to those for RNase A Library 1.
122 0 5 10 15 20 25 30 35 40 0 2 4 6 8 10 12 0.50W 1.0W 2.0W % ComplexationIrradiation Time (s) 0 5 10 15 20 25 30 35 40 0 2 4 6 8 10 12 0.50W 1.0W 2.0W % ComplexationIrradiation Time (s)RNase Myo 0 5 10 15 20 25 30 35 40 0 2 4 6 8 10 12 0.50W 1.0W 2.0W % ComplexationIrradiation Time (s) 0 5 10 15 20 25 30 35 40 0 2 4 6 8 10 12 0.50W 1.0W 2.0W % ComplexationIrradiation Time (s) 0 5 10 15 20 25 30 35 40 0 2 4 6 8 10 12 0.50W 1.0W 2.0W % ComplexationIrradiation Time (s) 0 5 10 15 20 25 30 35 40 0 2 4 6 8 10 12 0.50W 1.0W 2.0W % ComplexationIrradiation Time (s)RNase Myo Figure 410. Percent complexation versus irradiation time for IRMPD at 9.192 Neither complex is dissociated at laser powers of 0.5 or 1.0 W. Both complexes can be completely dissociated using 2.0 W.
123 Table 4 1. Comparison of wavelengths for complex dissociation. Wavelength refers to the laser wavelength used to perform IRMPD of the complexes. The second and third columns describe if complete complex dissociation was achieved at that wavelength. The column maximum power gives the greatest stable power output used in the experiment for that wavelength. The last column shows if significant differences between nonspecific and specific complex dissociation were observed at that wavelength. Wavelength Complete RNase A d issociation ? Complete m yoglobin d issociation ? Maximum p ower Observed d ifference s between s pecific and n onspecific complexes ? No No 1.0 W No Yes Yes 5.0 W Yes No No 3.0 W No Yes Yes 2.0 W Yes Yes Yes 2.0 W No
124 CHAPTER 5 S T UDIES OF RIBONUCLEAS E A AND TWO SMALL DYNAMIC COMBINATORIAL LIBRARIES Systems of Interest Although library complexes were introduced in the previous chapter, the chemistry behind them was not much discussed. The libraries in this project share a similar scaffolding of a nucleotide 2 3 cyclic monophosphate. Recall from Figure 33 that the RNA chain adopts a 2 3 cyclic monophosphate configuration as a transitory structure in its RNase A catalyzed cleavage. This is not coincidental, and the library members were designed with this in mind. Since RNase A catalyzes forward and reverse reactions, when two nucleotide 2, 3 cyclic monophosphate structures are present at the active site, RNase A may join the two together via phosphodiester bond in the reverse reaction of RNA cleavage. This process may repeat to generate an RNA chain. The components of L ibrary 1 were introduced in Chapter Four ( Figure 41 ). The components of L ibrary 2, shown in Figure 51, share the same 2, 3 cyclic monophosphate scaffolding. Ribonuclease A was incubated with each library in an attempt to generate novel, tight bindi ng inhibitors. The abbreviation NMP will be used throughout this chapter for a nucleotide monophosphate of unknown structure. Microdialysis Method for Reduced Signal Suppression For proper library functioning, the concentration of library components should be well in excess of the enzyme concentration. When the two concentrations are roughly equal, a relatively small degree of complexation is observed, as seen in the mass spectrum in Figure 52. Complexation between a single library fragment and a single RNase A occurs, but binding between multiple library fragments and a single RNase is disfavored due to probability. The most abundant species in the spectrum are free enzyme peaks.
125 At much higher library concentrations, complexation between the library and enzyme is not observed in the mass spectrum as shown in Figure 53. However, this spectrum greatly resembles the mass spectrum of the free library, suggesting that signal abundance for the free library is much greater than for that of the complex. Applica tion of a SWIF T waveform to remove library did not result in a spectrum containing the complex. Given the greater numbers of library fragments available for protonation and the need for each complex to undergo several protonation events, it was thought tha t the absence of the complex in the mass spectrum was due to signal suppression during ESI. As shown in the literature, ESI can be prone to distortion of species distribution in complex samples.15 This is especially exacerbated when species have substantia lly different proton affinities. Theoretically, there are several ways in which signal suppr ession can be resolved. First, solvents with higher gas phase basicities could be used, increasing the efficiency of proton transfer in the gas phase. Second, the number of available solution phase protons could be increased by adding more acid modifier (in this case, formic acid) to the solution. However, these alternate solvent systems are more lik ely denatur e the protein. Furthermore, if the concentration of the free library is sufficiently high, continued protonation of the library may occur, masking the presence of the complex. The most common method of reducing ESI signal suppression is through sample purification. By removing large numbers of gas phase proton acceptors, there are more protons available for the species of interest. Various chromatograph ic and gel methods have been reporte d in the literature. However, the viability of these methods for a DCL is somewhat questionable. No doubt methods could be de veloped to remove the free library, but such an approach might run the risk of disrupting the equilibrium and dissociating the complex.
126 Microdialysis membranes are used for desalting proteins and peptides. They are small circular discs (around 1 cm diamete r) made of nitrocellulose and contain small pores of a known size. Species smaller than the pore size can diffuse across the membrane while larger species are retained on one side. The procedure is illustrated in Figure 5 4. A membrane is floated on a pure solution against which you wish to exchange (exchange solution). A small volume of the solution to be dialyzed is pipetted on the top of the disc (sample solution). Then the entire setup is covered and allowed to dialyze. After a set time, the sample solu tion is pipetted from the surface for further analysis. For desalting, this method has the advantage of ease and speed. Recovery of the desalted species is high. Most importantly, the method is suitable for desalting small volumes (10200 Instead of exchanging salts, it seems possible to exchange small molecules like those in the library. Free enzyme and enzyme library complex would not be removed from solution, as they are too large to pass through the membrane pores. Both the sample and exchange solution can be water, making the procedure very gentle. Additional rounds of microdialysis with fresh exchange solution can be performed to remove more small molecules. Solutions containing 20 Library 1 were prepared i aqueous aliquots. Higher concentrations of library were desirable, but were limited by the availability of the library as a pooled solution. A plastic Petri dish was filled with 10mL Milli Q water and a microdialysis membrane was floated on the surface. The 100 carefully onto the membrane. The dish was covered with parafilm to reduce evaporation and then the plastic lid placed over top. Exchange proceeded for 30 minutes, after which the sample solution was pipetted from the m embrane surface. This process could be repeated as desired with a fresh exchange solution. When microdialysis was assumed to be complete, 100
127 with 1.0% formic acid was mixed with the sample solution. Assuming no loss of RNase A (free or complex was unknown. The effectiveness of the microdialysis procedure can be seen in Figure 5 5. The mass spectrum of original solution of RNase A and L ibrary 1 is shown at top left. Free library peaks dominate the spectrum. Varying the capillary and skimmer voltages greatly changed the appearance of the mass spectrum, but the peaks observed were always due to free library and RNase A L ibrary 1 complex was not observed. To support the hypothesis of ESI signal suppression, the free library ions in the cell were selectively ejected but this did not give rise to complex signal. Performing one round of microdialysis on the same solution greatly changed the appearance of the mass spectrum, shown at top right of Figure 55. Free library peaks are seen to be significantly reduced in abundance yet still h ave a significant contribution to the spectrum. Furthermore, significant amounts of RNase A Library 1 complex are observed. A small amount of free enzyme is observed, but it is dwarfed by the signal of the complex. A second round of microdialysis (bottom left) results in a similar spectrum. Abundance of free library is further reduced and complexation remains high. In an attempt to remove the remainder of the free library, a third round of microdialysis was performed (bottom right). This mass spectrum shows a subtle difference from those of one and two rounds of microdialysis. Free library signal is minimal and at first glance, R Nase A Library 1 complexation appears good. However, free enzyme has a greater contribution to the signal in this region. The degree of complexation between library and enzyme also decreased in this third round. Recall from Chapter One that proper function ing of the dynamic
128 combinatorial library is highly dependent on concentration. This suggests the undesirable outcome of equilibrium shifting due to decreased library concentration. To prevent this, prepared solutions in subsequent experiments were limited to two rounds of microdialysis. Peaks due to the remaining free library were removed through selective ion ejection from the ICR cell. Studies with Library 1 Complexes formed between RNase A and L ibrary 1 members were introduced in Chapter Four. Library 1 was obtained as a pooled solution, so analysis of the individual components was not possible. However, a mass spectrum was obtained for the library in the absence of the enzyme. This is shown in Figure 56. By varying the capillary and skimmer voltages, different peaks were observed in the mass spectrum. IRMPD at 9. 588 of the major peaks observed under all conditions is shown in Table 51. Peaks representative of both the adenosine and uridine moieties were observed. Those structural assignments which could be made from Table 51 are show n in Figure 57. It is important to note that some structures cannot be assigned absolutely, as they have the same exact masses. The microdialysis procedure described above was used to generate solutions to study the RNase A L ibrary 1 complexes. The full spectrum of such was shown at the bottom right of Figure 55. Free library peaks were removed by a chirp isolation event. A portion of the resulting mass spectrum for the +8 charge state is shown in Figure 58. Complexes of RNase A with one, two, three, four, and five nucleotides are observed. The identity of the bound nucleotides cannot be ascertained from this mass spectrum. Signal abundance of the complex when viewed at maximum resolution is poor. The spectrum in Figure 58 is demonstrative of much lower resolution. Furthermore, as seen from the free library results, there are a number of structures with similar, if not identical molecular weights. In addition to RNase A L ibrary 1 peaks, phosphate adducts wer e also observed.
129 Isolation of the complex foll owed by IRMPD at 9.588 identifying the bound nucleotides. One such spectrum is shown in Figure 59. Nucleotide fragments resulting from the dissociation of the complex are observed. Peaks at m/z 136, 294, and 351 were observed. Of these, m/z 294 and 351 were observed in the free library system. While the structure of m/z 294 is unknown, it seems to be a common fragment for this library. It is even observed following in source CID of the free library. Protonated adenine occurs at m/ z 136. The structure of m/z 351 is given in Figure 56 (structure 5) as a uridine based 2 3 cyclic monophosphate. Thus, both nucleobases are represented in RNase A Library 1 complexes. Fragments with m/z indicating a di nucleotide chain were not obse rved. There are two possible explanations for this. In the first, the nucleotides bind to the enzyme, but do not form covalent bonds with each other. In this case, as IRMPD breaks the noncovalent enzyme ligand interactions, each nucleotide is lost in tur n. An alternate explanation is that the nucleotides noncovalently bind the enzyme and form covalent bonds between fragments. Then IRMPD preferentially breaks the newly formed phosphodiester bonds before the noncovalent interactions are disrupted to dissoci ate the complex. Since the employed IR wavelengths are in the O P O stretching region, this explanation is not unlikely. However, the energy required to break a covalent bond is greater than to break a noncovalent bond, so this explanation does have its dr awbacks. No definitive conclusions can be drawn. Studies with Library 2 The library components of L ibrary 2 were purchased separately, and thus, it was possible to obtain mass spectra for the individual components. These were then subjected to IRMPD in or der to develop a list of the resulting fragments. This enabled fragments from the IRMPD of the RNase A Library 2 complexes to be assigned as originating from specific components.
130 Also, a mass spectrum of the pooled library in the absence of the enzyme was obtained. IRMPD (at 9. 588 the most abundant of which are shown in T able 5 2. Among these fragments, the identity of a number of structures can be assigned, shown in Figure s 5 10 to 5 13. Because of m/ z redundancy, it is not po ssible to absolutely assign certain of the structures. In these cases, all possible structures are shown. It is possible that structures 10 18 incorporate a phosphodiester bond between nucleotides in addition to existing as a loosely bound dimer. Complexes incorporating three or more library members are not observed in the absence of RNase A. This suggest s that chains of greater than two nucleotides do not spontaneously form, although phosphodiester bonds between two library members may (or may not) occur. Using the method developed above, Library 2 was incubated with RNase A and electrospray solutions were prepared. A mass spectrum is shown in Figure 514. Binding between RNase A and one or two nucleotides is observed. These complexes were isolated and fra gmented using IRMPD at 9.588 shown in Figure 515. Table 53 lists the major fragments observed from the IRMPD of the RNase A Library 2 complexes. The proposed structures corresponding to the fragments are shown in Figures 516 through 522. Where RNA chains are possible structures, the sequence is listed in red on these figures. One interesting feature of this list is the appearance of larger m/z species than appear in the free library. While these are low abundance species, this co uld be indicative of covalent bond formation between library fragments in the presence of RNase A suggesting proper library functioning. Recall that complexes comprised of more than two library members were not observed in the absence of the enzyme.
131 T hese larger species are not observed as abundant products of the IRMPD complex dissociation. The possible reasons for this are the same as those explained for the RNase A L ibrary 1 complexes. However, the following RNA sequences may have been observed: C, A, CC, CA or AC, AA, CCC, and AAA. Additional dissociation events were employed in an attempt to gain further structural information, but MS3 signal quality was too poor to make any additional conclusions. Figure 523 shows how the abundance of several species in the RNase A Library 2 solution varies as a function of the irradiation time. Enzyme ligand c omplexation shows an initial decrease, but then plateaus The abundance of m/z 330 increases rapidly, but then begins a decline. This pattern is suggestive of an ion undergoing a second fragmentation event. Peak abundance at m/z 136 increases more slowly and reaches a plateau at longer irradiation times. This ion may be the result of a secondary dissociation event. One possible pathway is the diss ociation of the RNase A adenosine complex to give charged adenosine ( m/z 330) and RNase A with one fewer charge. Additional energy fragments m/z 330 to m/z 136, the protonated base. Another interesting phenomenon can be noted for the IRMPD of the RNase A Library 2 complexes. First is the average charge state at increasing IRMPD energies. One would expect that imparting energy to the system could disrupt noncovalent networks and disulfide bridges to give an unfolded or partially unfolded protein. As discuss ed previously, unfolded proteins are expected to have relatively large charge state envelopes at relatively low m/z (more highly charged) in comparison to their folded counterparts. The average charge state is plotted as a function of IRMPD energy (irradia tion time multiplied by laser power) in Figure 524 Surprisingly, the average charge state decreases. While the protein may be unfolding, this effect is likely not a manifestation of such. Rather, this may be the result of where protons end up after
132 dissociation. If in fact nucleotides have higher gas phase ba sicity than RNase A (and this would be supported by the idea of ESI-based signa l suppression), dissociation of the RNase A-nucleotide complex would likely result in a protonated nucleotide. The complex, having lost a proton, can still be observed since it still has charges asso ciated with it. Although the structure may not have changed, the charge state has decreased by one charge. This may explain the effect in Figure 5-24. Instead of looking at the average charge state, the behaviors of individual charge states can be plotted. In Figure 5-25 the +9, +8, and +7 charge states behave similarly. However, at the earliest time points, the +6 charge state increases more dramatically than the other charge states, suggesting that the +6 charge state may be th e product of deprotonating dissociation events for the more highly charged states. Adenosine-2,3-cyclic monophosphate Exact mass: 351.034 Cytidine-2,3-cyclic monophosphate Exact mass: 327.023 Adenosine-2,3-cyclic monophosphate Exact mass: 351.034 Cytidine-2,3-cyclic monophosphate Exact mass: 327.023 Figure 5-1. Structures of Li brary 2 components. Cytidine2, 3-cyclic monophosphate and adenosine-2, 3-cyclic monophos phate comprised Library 2.
133 m/z 2,000 1,500 +9 +7 +8 R+NMP R+2NMP R R+NMP R R R R+NMP R+2NMP R+NMP +10m/z m/z 2,000 1,500 +9 +7 +8 R+NMP R+2NMP R R+NMP R R R R+NMP R+2NMP R+NMP +10 m/z 2,000 1,500 +9 +7 +8 R+NMP R+2NMP R R+NMP R R R R+NMP R+2NMP R+NMP +10m/z Figure 52. Mass spectrum obtained when RNase A and L ibrary 1 concentrations are roughly equal. R ibonuclease A complexation with nucleotide monophosphate (NMP) library members is observed.
134 m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance m/z 850 800 750 700 650 600 550 500 450 400 350 294.25 417.21 623.53 351.02 439.13 Noise 711.06 789.14 596.33 743.12 671.20 m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance m/z 850 800 750 700 650 600 550 500 450 400 350 m/z 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance m/z 850 800 750 700 650 600 550 500 450 400 350 294.25 417.21 623.53 351.02 439.13 Noise 711.06 789.14 596.33 743.12 671.20 Figure 53. Mass spectrum obtained when L ibrary 1 concentration is 100 times greater than that of RNase A. Selected peaks are labeled. No RNase ALibrary 1 complexation is observed. Petri Dish Parafilm Cover Microdialysis Membrane Sample Solution Exchange Solution Petri Dish Parafilm Cover Microdialysis Membrane Sample Solution Exchange Solution Figure 54. Setup for microdialysis procedure. An aqueous sample solution comprise d of RNase A and concentrated library is pipetted onto a microdialysis membrane. Free library is small enough to exchange through the membrane, while complexes between library and enzyme are too large. The solution with reduced free library can be pipetted from the membrane surface and used to prepare ESI solutions.
135 m/z 2,400 2,200 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance 294.25 417.21 623.53 m/z 2,400 2,200 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance Free Library +9 +8 +7 m/z 2,400 2,200 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance Free Library +9 +8 +7 m/z 2,000 1,500 1,000 500 Abundance +9 +7 +8 No Microdialysis Microdialysis Once Microdialysis Twice Microdialysis Thrice m/z 2,400 2,200 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance 294.25 417.21 623.53 m/z 2,400 2,200 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance 294.25 417.21 623.53 m/z 2,400 2,200 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance Free Library +9 +8 +7 m/z 2,400 2,200 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance Free Library +9 +8 +7 m/z 2,400 2,200 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance Free Library +9 +8 +7 m/z 2,400 2,200 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 Abundance Free Library +9 +8 +7 m/z 2,000 1,500 1,000 500 Abundance +9 +7 +8 m/z 2,000 1,500 1,000 500 Abundance +9 +7 +8 No Microdialysis Microdialysis Once Microdialysis Twice Microdialysis Thrice Figure 55. Comparison of free library and RNase A L ibrary 1 complex abundances for microdialysis procedure. At top left, no microdialysis has been performed and only free library peaks are p resent. One round of microdialysis has been performed for the spectrum at top right. Although library abundance has been reduced, it is still abundant. Significant RNase A library complexation can be observed. A second round of microdialysis (bottom left) further decreases the presence of free library and significant amounts of complex remain. However, performing further microdialysis (bottom right) begins to reduce the amount of RNase A L ibrary 1 complex observed.
136 m/z 800 700 600 500 400 300 200 100 Abundance 294.25 348.03 362.05 370.01 384.02 392.00 328.98 630.93 m/z 800 700 600 500 400 300 200 100 Abundance 294.25 348.03 362.05 370.01 384.02 392.00 328.98 m/z 800 700 600 500 400 300 200 100 Abundance 294.25 348.03 362.05 370.01 384.02 392.00 328.98 630.93 Figure 56. Mass spectrum of L ibrary 1 i n the absence of RNase A. Selected peaks are labeled. The peaks present were highly dependent on capillary and skimmer voltages. A summary of the major peaks present can be found in Table 5 1.
137 Table 5 1. Fragments from IRMPD of Library 1 complex es These IRMPD fragments were obtained for Library 1 in the absence of RNase A. The first column gives the m/z for each ion. In the second column, assignments as to structure are made, where possible. Those numbers correspond to structures shown in Figure 5 7. m/z Structure 147.15 Unknown 212.17 Unknown 294.23 Unknown 307.01 1 325.01 2a or 2b 328.98 3a or 3b 348.03 4a or 4b 350.95 5 362.05 Unknown 370.01 Unknown 372.93 Unknown 384.02 Unknown 391.99 6a or 6b 394.06 Unknown 406.00 Unknown 412.06 Unknown 413.96 Unknown 418.93 Unknown 452.88 Unknown 473.36 Unknown 612.95 7 630.93 Unknown 634.90 8 648.94 Unknown 652.91 Unknown 656.87 Unknown 670.91 Unknown 674.87 Unknown 678.86 Unknown 692.89 Unknown 696.84 Unknown
138 Structure 1 Structure 3a Structure 3b Structure 2a Structure 2b Structure 4a Structure 4b Structure 5 Structure 1 Structure 3a Structure 3b Structure 2a Structure 2b Structure 4a Structure 4b Structure 5 Structure 6aN N N N N H2 O O OH H H H H HO P O OH Na O Na Structure 6b Structure 7O O O H H H H HO P HO O Na NH N O O O O O H H H H HO P HO O NH N O O Structure 8 Structure 6aN N N N N H2 O O OH H H H H HO P O OH Na O Na Structure 6b Structure 7O O O H H H H HO P HO O Na NH N O O O O O H H H H HO P HO O NH N O O Structure 8 Figure 5-7. Putative structures for fragm ent assignments made in Table 5-1.
139 m/z 1,950 1,900 1,850 1,800 1,750 Abundance R+H2PO4R+NMP R+2NMP R+3NMP R+4NMP R+5NMP R+NMP+H2PO4 R+2NMP+H2PO4 R+3NMP+H2PO4 R+4NMP+H2PO4 m/z 1,950 1,900 1,850 1,800 1,750 Abundance R+H2PO4R+NMP R+2NMP R+3NMP R+4NMP R+5NMP R+NMP+H2PO4 R+2NMP+H2PO4 R+3NMP+H2PO4 R+4NMP+H2PO4 Figure 58. R ibonuclea se A L ibrary 1 complexation in the +8 charge state. R ibonuclease A is observed with a bound phosphate group. This adduct is observed even when nucleotides bind. The limitations of signal abundance and resolution coupled with the plethora of nucleotide structures present make it impossible to state the identity of the bound nucleotide(s).
140 m/z 450 400 350 300 250 200 150 100 Abundance 136.07 294.25 350.98 m/z 136 m/z 351 m/z 450 400 350 300 250 200 150 100 Abundance 136.07 294.25 350.98 m/z 450 400 350 300 250 200 150 100 Abundance 136.07 294.25 350.98 m/z 136 m/z 351 Figure 5-9. IRMPD fragments from RNase A-Library 1 complexes (9.588 m). The three labeled peaks are the result of complex dissociation. The structures for tw o of these peaks are shown.
141 Table 5 2. Fragments from IRMPD of Library 2 in the absence of RNase A This t able represents the IRMPD fragment for Library2 in the absence of RNase A. The first column gives the m/z of the fragment. The second column indicates if that ion was observed for the pure adenosine or cytidine 2, 3 cylic monophosphate solutions, or if its origin is unknown. The last column gives structural assignments where possible. The numbers correspond to structures in Figures 1013. m/z Origin Structure 134.03 C 1 136.06A2 226.07 C 3 306.02 C 4 314.91 A Unknown 327.99 C 5a-5c 330.02 A 6 349.97 Unknown 7 352.00 A 8a or 8b 360.01 C Unknown 373.98 A 9 389.95 C Unknown 428.90 A Unknown 454.96 C Unknown 470.92 C Unknown 494.94 C Unknown 499.22 C Unknown 623.26 Unknown Unknown 628.97 C 10 632.94 C 11a-11c 634.97 Unknown 12a-12c 642.98 Unknown 648.90 C 13a-13d 650.94 C 14a-14f 658.97 A 15a or 15b 664.96 Unknown Unknown 670.88 C Unknown 672.90 C Unknown 680.95 A 16a-16e 686.92 C Unknown 694.90 C 17a or 17b 702.90 A 18a-18d 704.91 A Unknown 710.95 Unknown Unknown 713.86 A Unknown 719.87 Unknown Unknown
142N N H NH2 O Na Structure 1 Structure 2 Structure 3 Structure 4N N H2 O N O OH OH H H H H N NH2 O N O O O H H H H HO P O O H H Structure 5a Structure 5b Structure 5c Structure 6 N N H NH2 O Na Structure 1 Structure 2 Structure 3 Structure 4N N H2 O N O OH OH H H H H N NH2 O N O O O H H H H HO P O O H H Structure 5a Structure 5b Structure 5c Structure 6 Structure 7 Structure 8a Structure 8b Structure 9 Structure 10 Structure 11a Structure 11b Structure 7 Structure 8a Structure 8b Structure 9 Structure 10 Structure 11a Structure 11b Figure 5-10. Putative structure a ssignments for peaks listed in Table 5-2 (Structures 1-11b).
143 Structure 11c Structure 12a Structure 12b Structure 12c Structure 13a Structure 13b Structure 11c Structure 12a Structure 12b Structure 12c Structure 13a Structure 13b Structure 13c Structure 13d Structure 14a Structure 14b Structure 14c Structure 14d Structure 13c Structure 13d Structure 14a Structure 14b Structure 14c Structure 14d Figure 5-11. Putative structure assignments for peaks listed in Table 5-2 (Structures 11c-14d).
144 Structure 14e Structure 14f Structure 15aO OH O H H H H HO O OH O H H H H O P O HO P O HO H N N NH2 O N N NH2 O O Na N N N N NH2 O O O H H H H HO P HO O N N N N NH2 O O O H H H H HO P HO O H Structure 15b Structure 16a Structure 16bO OH O H H H H HO O O O H H H H O P O HO P O HO H N N N N N H 2 N N N N NH2 N N N N NH2 O O O H H H H HO P HO O N N N N NH2 O O O H H H H HO P O O H Na Structure 14e Structure 14f Structure 15aO OH O H H H H HO O OH O H H H H O P O HO P O HO H N N NH2 O N N NH2 O O Na N N N N NH2 O O O H H H H HO P HO O N N N N NH2 O O O H H H H HO P HO O H Structure 15b Structure 16a Structure 16bO OH O H H H H HO O O O H H H H O P O HO P O HO H N N N N N H 2 N N N N NH2 N N N N NH2 O O O H H H H HO P HO O N N N N NH2 O O O H H H H HO P O O H Na Structure 16c Structure 16d Structure 16e Structure 17a Structure 17b Structure 18a Structure 16c Structure 16d Structure 16e Structure 17a Structure 17b Structure 18a Figure 5-12. Putative structure assi gnments for peaks observed in Ta ble 5-2 (Structures 14e-18a).
145 Structure 18b Structure 18c Structure 18dN N N N NH2 O O O H H H H HO P HO O N N N N NH2 O O O H H H H HO P O O Na Na O OH O H H H H HO O O O H H H H O P O O P O O H N N N N NH2 N N N N NH2 Na Na Structure 18b Structure 18c Structure 18dN N N N NH2 O O O H H H H HO P HO O N N N N NH2 O O O H H H H HO P O O Na Na O OH O H H H H HO O O O H H H H O P O O P O O H N N N N NH2 N N N N NH2 Na Na Figure 5-13. Putative structure assignments for peaks observed in Table 5-2 (Structures 18b18d). m/z 2,50 0 2,000 1,500 1,000 500 Abundance R R+NMP R R+NMP R+2NMP R R+NMP R+2NMP 2,500 +9 +8 +7 m/z 2,50 0 2,000 1,500 1,000 500 Abundance R R+NMP R R+NMP R+2NMP R R+NMP R+2NMP 2,500 +9 +8 +7 Figure 5-14. Complexation between RNase A and Library 2 following two rounds of microdialysis.
146 m/z 2,500 2,000 1,500 1,000 500 Abundance 2,500R R+NMP R R+NMP R R+NMP R+2NMP +9 +8 +7 Noise 330.03 m/z 2,500 2,000 1,500 1,000 500 Abundance 2,500R R+NMP R R+NMP R R+NMP R+2NMP +9 +8 +7 Noise 330.03 Figure 515. Incomplete dissociation of RNase A L ibrary 2 complexes with IRMPD. Several peaks o ccurring in the range m/z 500600 are due to electronic noise. RNase A Library 2 complexation is observed. The peak at m/z 330 is the result of complex dissociation.
147 Table 5 3. Fragments from IRMPD of RNase A Library 2 complexes ( ) Fragments are listed by m/z The second column shows whether or not that ion was present in the free library spectrum. There is a trend for smaller ions to be present in both free library and enzyme library solutions, while larger ions only appear when the library is in cubated with RNase A. The third column refers to the presence of an ion in the mass spectra of the individual library component solutions. Lastly, those ions which can be identified are listed in the fourth column, with chemical structures shown in F igures 516 to 522. m/z Free library? Origin Assignment 134.03 Yes C 1 136.06 Yes A 2 165.03 No A Unknown 234.97 No Unknown Unknown 306.03 Yes C 3 328.00 Yes C 4a, 4b, or 4c 330.03 Yes A 5 348.04 Yes A 6 349.97 Yes C 7 362.08 Yes A Unknown 367.99 No Unknown 8 373.99 Yes A 9 469.92 No Unknown Unknown 607.91 No Unknown Unknown 650.96 Yes C 10a-10f 653.08 Yes Mixture 11a-11d 659.07 Yes A 12a or 12b 669.06 Yes Unknown Unknown 672.95 No Unknown Unknown 675.01 Yes A Unknown 681.01 Yes A 13a-13e 691.04 Yes Unknown Unknown 694.92 Yes C 14a or 14b 697.04 Yes Mixture 15a-15k 700.91 No Mixture 16a, 16b, or 16c 703.04 Yes A 17a-17e 712.91 No Unknown Unknown 718.91 No Mixture 18a-18d 976.09 No Unknown Unknown 982.09 No Unknown 19a-19f 992.06 No Unknown Water loss from 1010? 998.07 No Unknown Unknown 1004.07 No Unknown 20a-20c 1010.07 No Unknown 21a-21i 1014.04 No Unknown 22a or 22b 1020.04 No Unknown Unknown 1026.05 No Unknown Unknown 1033.00 No Unknown Unknown 1042.01 No Unknown Unknown 1048.02 No Unknown Unknown
148 Structure 1 Structure 2 Structure 3 Structure 4a Structure 4b Structure 4c Structure 5 Structure 6 N NH2 O N O O OH H H H P O Na O HO H Structure 1 Structure 2 Structure 3 Structure 4a Structure 4b Structure 4c Structure 5 Structure 6 N NH2 O N O O OH H H H P O Na O HO H Structure 7 Structure 8 Structure 9 Structure 10a Structure 10b Structure 10c Structure 7 Structure 8 Structure 9 Structure 10a Structure 10b Structure 10c Figure 5-16. Putative structure a ssignments for peaks observed in Table 5-3 (Structures 1-10c).
149 Structure 10d Structure 10e Structure 10f Structure 11a Structure 11b Structure 11c CC CCCC CA Structure 10d Structure 10e Structure 10f Structure 11a Structure 11b Structure 11c CC CCCC CA Structure 11d Structure 12a Structure 12b Structure 13a Structure 13b Structure 13c AC AA AA Structure 11d Structure 12a Structure 12b Structure 13a Structure 13b Structure 13c AC AA AA Figure 5-17. Putative structure assign ments for peaks observed in Table 5-3 (Structures 10d-13c).
150 Structure 14b Structure 15a Structure 15b Structure 13d Structure 13e Structure 14a AA AA CC Structure 14b Structure 15a Structure 15b Structure 13d Structure 13e Structure 14a AA AA CC Structure 15c Structure 15d Structure 15e Structure 15f Structure 15g Structure 15h CA CA CA Structure 15c Structure 15d Structure 15e Structure 15f Structure 15g Structure 15h CA CA CA Figure 5-18. Putative structure assi gnments for peaks observed in Tabl e 5-3 (Structures 13d-15h).
151 Structure 15i Structure 15j Structure 15k Structure 16a Structure 16b Structure 16c O OH O H H H H HO O OH O H H H H O P P O O OH Na O HO Na N N N N NH2 N N NH2 O O OH O H H H H HO O OH O H H H H O P P O O OH H O O Na N N N N N H2 N N NH2 O Na O OH O H H H H HO O OH O H H H H O P P O O OH Na OH O N N N N N H2 N N NH2 O Na AC AC ACO O O H H H H HO O O O H H H H HO P O O Na P OO N N NH2 O N N N N NH2 Na Na O OH O H H H H HO O O O H H H H O P P O O Na O O Na Na N N N H2 O N N N N NH2 AC CA Structure 15i Structure 15j Structure 15k Structure 16a Structure 16b Structure 16c O OH O H H H H HO O OH O H H H H O P P O O OH Na O HO Na N N N N NH2 N N NH2 O O OH O H H H H HO O OH O H H H H O P P O O OH H O O Na N N N N N H2 N N NH2 O Na O OH O H H H H HO O OH O H H H H O P P O O OH Na OH O N N N N N H2 N N NH2 O Na AC AC ACO O O H H H H HO O O O H H H H HO P O O Na P OO N N NH2 O N N N N NH2 Na Na O OH O H H H H HO O O O H H H H O P P O O Na O O Na Na N N N H2 O N N N N NH2 AC CA Structure 17a Structure 17b Structure 17c Structure 17d Structure 17e Structure 18a AA AAAA Structure 17a Structure 17b Structure 17c Structure 17d Structure 17e Structure 18a AA AAAA Figure 5-19. Putative structure a ssignments for peaks observed in Table 5-3 (Structures 15i-18a).
152 Structure 18b Structure 18c Structure 18d CA AC Structure 19a Structure 19b Structure 19c CC Structure 18b Structure 18c Structure 18d CA AC Structure 19a Structure 19b Structure 19c CC Structure 19d Structure 19e Structure 19f CCC CC CCC Structure 19d Structure 19e Structure 19f CCC CC CCC Figure 5-20. Putative structure a ssignments for peaks observed in Table 5-3 (Structures 18b-19f).
153 Structure 20a Structure 20b Structure 20c CC CCCO O O H H H H HO P O O Na O O O H H H H HO P O O Na O O O H H H H HO P O O Na N N NH2 O N N NH2 O N N NH2 O Na O OH O H H H H HO P O Na O O O H H H H O P O Na O O O H H H H HO P O O Na N N NH2 O N N NH2 O N N NH2 O Na O O O OH O H H H H HO P O Na O OH O H H H H O P O Na O O O H H H H O P O O Na N N NH2 O N N NH2 O N N NH2 O Na O O Structure 20a Structure 20b Structure 20c CC CCCO O O H H H H HO P O O Na O O O H H H H HO P O O Na O O O H H H H HO P O O Na N N NH2 O N N NH2 O N N NH2 O Na O OH O H H H H HO P O Na O O O H H H H O P O Na O O O H H H H HO P O O Na N N NH2 O N N NH2 O N N NH2 O Na O O O OH O H H H H HO P O Na O OH O H H H H O P O Na O O O H H H H O P O O Na N N NH2 O N N NH2 O N N NH2 O Na O O Structure 21a Structure 21b Structure 21c AA Structure 21a Structure 21b Structure 21c AA Structure 21d Structure 21e Structure 21f AA AA AA Structure 21d Structure 21e Structure 21f AA AA AA Figure 5-21. Putative structure a ssignments for peaks observed in Table 5-3 (Structures 20a-21f).
154 Structure 21g Structure 21h Structure 21i AAA AAA AAAO OH O H H H H HO P HO O OH O H H H H O P HO O O O H H H H O P O O H N N N N NH2 N N N N NH2 N N N N N H2 O O N a O OH O H H H H HO P HO O OH O H H H H O P HO O O O H H H H O P HO O Na N N N N NH2 N N N N NH2 N N N N NH2 O O O OH O H H H H HO P HO O OH O H H H H O P O O O O H H H H O P HO O H N N N N NH2 N N N N NH2 N N N N NH2 O O Na Structure 21g Structure 21h Structure 21i AAA AAA AAAO OH O H H H H HO P HO O OH O H H H H O P HO O O O H H H H O P O O H N N N N NH2 N N N N NH2 N N N N N H2 O O N a O OH O H H H H HO P HO O OH O H H H H O P HO O O O H H H H O P HO O Na N N N N NH2 N N N N NH2 N N N N NH2 O O O OH O H H H H HO P HO O OH O H H H H O P O O O O H H H H O P HO O H N N N N NH2 N N N N NH2 N N N N NH2 O O Na Structure 22a Structure 22b Structure 22a Structure 22b Figure 5-22. Putative structure assi gnments for peaks observed in Tabl e 5-3 (Structures 21g-22b).
155 0 5 10 15 20 25 0 2 4 6 8 10 12 136 330 Complexation Irradiation Time (s)Abundance 0 5 10 15 20 25 0 2 4 6 8 10 12 136 330 Complexation Irradiation Time (s)Abundance Figure 523. Abundances of selected species as a function of irradiation time for IRMPD of RNase A Library 2 complexes 6.8 7 7.2 7.4 7.6 7.8 8 8.2 0 5 10 15 20 25 30 35 40 Irradiation Time (s) Laser Power (W)Average Charge State 6.8 7 7.2 7.4 7.6 7.8 8 8.2 0 5 10 15 20 25 30 35 40 Irradiation Time (s) Laser Power (W)Average Charge State Figure 524. Effect of increasing IRMPD energy on a verage charge state of RNase A L ibrary 2 complexes.
156 0 10 20 30 40 50 60 70 0 5 10 15 20 25 30 35 40 9+ 8+ 7+ 6+ Irradiation Time (s) Laser Power (W)Complex Abundance 0 10 20 30 40 50 60 70 0 5 10 15 20 25 30 35 40 9+ 8+ 7+ 6+ Irradiation Time (s) Laser Power (W)Complex Abundance Figure 525. Effect of increasing IRMPD energy on various charge states for RNase A L ibrary 2 complexes.
157 CHAPTER 6 CONCLUSIONS AND FUTU RE DIRECTIONS Studying noncovalent chemistry provides insight into a numbe r of important biological systems. Dynamic combinatoria l chemistry uses noncovalent interactions between a target molecule and a library to select and amplify the production of the tightest binding species. One method for detecting and identifying novel binders is ESI FTICR MS, which offers tandem in time capa bilities, high mass accuracy, and superior resolving power. Ribonuclease A is a digestive enzyme which cleaves RNA. In the course of this reaction, a nucleotide 2 3 cyclic monophosphates structure is adopted Two libraries with this structure as scaffo lding and variable nucleobases were screened in the course of this dissertation. In the presence of RNase A, these nucleotides can be formed into chains of RNA. The first noncovalent complex presented here was that between RNase A and its weak inhibitor U MP. This complex was used as a simplified model of the RNase A library systems to perform optimization experiments. Optimal electrospray conditions were determined and two types of dissociation were explored. Insource CID was shown to be ineffective for d issociating these complexes, while proof of concept was shown for IRMPD. While studying RNase A UMP complexes, it was found that UMP is capable of forming noncovalent clusters with other UMP molecules. These UMP clusters are prevalent at high UMP concentr ations. Insource CID can break up some of these clusters but IRMPD can cause complete dissociation. I nfrared multiple photon dissociation of RNase A Library 1 complexes was performed at multiple wavelengths to improve dissociation efficiency. Ideally, th ere exists a wavelength at which the laser has sufficient power to completely dissociate the RNase A Library 1 complex at moderate irradiation times. Two of the five wavelengths used were incapable of inducing
158 significant dissociation; the remainder were able to dissociate the complex by varying the irradiation times. A wavelength of 9.588 dissociation characteristics and this wavelength was used for subsequent studies. The photodissociation behavior of specific and nonspecific complexes was also studied at multiple wavelengths. R ibonuclease A Library 1 complexes were used as specific complexes, while myoglobinLibrary 1 complexes were used as nonspecific complexes. At t wo of the five studied wavelengths, significant dissociation of either complex was not observed. This can be attributed to insufficient laser output power or poor absorption by the complexes. complete dissociation was achieved, but the behavior of specific and nonspecific complexes was indistinguishable. For the two remaining wavelengths, greater energy input was required to break apart the specific complexes than the nonspecific complexes. The energy input was increased by operating the laser at higher powers or by irradiating the ions for longer periods. As the specific complexes are expected to bind more tightly, the need for greater energy to achieve complete dissociation fits the expected behavior. Using the previously determined conditions, a single experiment was designed to incubate the enzyme and library, ionize it, dissociate it, and fragment the freed library member. This was performed for Library 1 and Library 2. Proof of concept for this experiment was demonstrated, and several binding nucleotide species were putatively identified. Absolute identification was hampered by the presence of multiple species with the same exact mass. Most importantly, dimers and/or trimers co uld not be discriminated from RNA chains without MS3. Nor could RNA chain sequence be determined. Low ion abundances limited this as a method. Furthermore, relatively large m/z nucleotide species were not abundant products of IRMPD, possibly as a result of phosphodiester bond cleavage prior to enzyme ligand complex dissociation. The
159 presence of larger complexes in the presence of the enzyme suggests that chain formation may be occurring, although this cannot be absolutely stated. Several avenues for future work could be pursued. First, the size and scope of the nucleotide libraries could be expanded. Recall from Chapter 1 that even large libraries have low probabilities of containing tight binding species. While the dynamic combinatorial nature of the libra ry increases the number of potential ligands, the libraries used in this dissertation were still extremely small. Furthermore, the chemistry of libraries used is limited. Expansion to include unnatural nucleobases as well as variations on the scaffolding m ay yield better results. Future developments might also be made to increase the incubation concentrations of the libraries. Extremely high concentrations of the library are necessary for proper functioning, but this need must be balanced against the resul ting signal suppression. A microdialysis method was presented here to overcome this challenge, but perhaps more effective methods exist. The possibility of nucleotide phosphodiester bond cleavage prior to dissociation from the complex was mentioned. If t his occurs, the binding chains of RNA cannot be identified. Using IR wavelengths outside the O P O stretching region may be successful at dissociating enzymelibrary complexes without breaking covalent bonds between library members.
160 LIST OF REFERENCES 1. Li ndoy, L.F.; Atkinson, I.M. Stoddart, J.F ., Ed. Self Assembly in Supramolecular Systems; 7; Monographs in Supramolecular Chemistry; Royal Society of Chemistry: Cambridge U.K., 2000. 2. Schalley, C A. Int. J. Mass Spectrom 2000, 194, 139. 3. Lehn, J.M. Supramolecular Chemistry. Concepts and Perspectives., WileyVCH: Weinheim, 1995. 4. Lehn, J.M. Science. 1985, 227, 849856. 5. Lehn, J.M. Angew. Chem. Int. Ed. 1990, 29(11), 13041319. 6. Anfinsen, C.B.; White, F.H. In The Enzymes, 2nd Edition; Ed. Boyer P.D.; Lardy, H.; Myrbck, K. Academic Press: New York, 1961, 5, 95122. 7. Schalley, C.A. Analytical Methods in Supramolecular Chemistry Schalley, C.A., Ed.; Wiley VCH Verlag: Weinheim, 2007; Chapter 1. 8. Steed, J.W.; Atwood, J.L. Supramolecular Chemistry ; John Wiley & Sons, Ltd.: Chichester, 2000. 9. Pimental, G.C.; McClellan A.L. The Hydrogen Bond. Freeman: San Francisco, 1960. 10. Watson, J.D.; Crick, F.H.C. Nature 1953, 171 (4361), 964967. 11. Hunter, C.A.; Sanders, J.K.M. J. Am. Chem. Soc. 1990, 112, 55255534. 12. Yakovchuk, P.; Pr otozanova, E.; FrankKamenetskii, M.D. Nucl. Acids. Res. 2006, 34(2), 564574. 13. Kaltashov, I.A.; Eyles, S.J. Mass Spectrom. Rev. 2002, 21, 3771. 14. Daniel, J.M.; Friess, S.D.; Rajagopalan, S.; Wendt, S.; Zenobi, R. Int. J. Mass Spectrom.2002, 216, 127. 15. Wu, Q .; Gao, J .; Joseph McCarthy, D .; Sigal, G .B.; Bruce, J E.; Whitesides, G .M.; Smith, R .D. J. Am. Chem. Soc. 1997, 119, 1157158. 16. Cech, N.B.; Enke, C.G. Mass Spectrom. Rev. 2001, 20, 362387. 17. Smith, R.D.; Bruce, J.E.; Wu, Q.; Lei, P. ChemSoc. Rev. 1997 26, 191202. 18. Loo, J.A. Int. J. Mass Spectrom. 2000, 200, 175186. 19. Di Tullio, A.; Reale, S.; De Angelis, F. J. Mass Spectrom. 2005, 40, 845865.
161 20. Kaltashov, I.A.; Eyles, S.J. Mass Spectrom. Rev. 2002, 21, 3771. 21. Hernandez, H.; Robinson, C.V. J. Biol Chem. 2001, 276 (50), 4668546688. 22. Fales, H.M.; Milne, G.W.; Pisano, J.J.; Brewer, H.B.; Blum, M.S.; MacConnell, J.G.; Brand, J.; Law, N. R ecent Prog. Horm. Res. 1972, 28, 591626. 23. Fenn, J.B.; Mann, M.; Meng, C.K.; Wong, S.F.; Whitehouse, C.M. Science. 1989, 246 ( 4926), 6471. 24. Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 22992301. 25. Tanaka, K.; Waki, H.; Ido, Y.; Akita, S.; Yoshida, Y.; Yoshida, T. Rapid Commun. Mass Spectrom. 1988, 2, 151153. 26. Wilm, M.; Mann, M. Anal. Chem. 1996, 68, 18. 27. Juraschek, R.; Dulcks, T.; Karas, M. J. Am. Soc. Mass Spectrom.1999, 10, 300308. 28. Takats, Z.; Wiseman, J.M.; Gologan, B.; Cooks, R.G. Anal. Chem. 2004, 76, 40504058. 29. Wiseman, J.M.; Takats, Z.; Gologan, B.; Davisson, V.J.; Cooks, R.G. Angew. Chem. Int. Ed. 2005, 44, 913916. 30. Yan g, P.; Cooks, R.G.; Ouyang, Z.; Hawkridge, A.M.; Muddiman, D.C. Anal. Chem. 2005, 77 (19), 61746183. 31. Chowdhury, S.K.; Katta, V.; Chait, B.T. J. Am. Chem. Soc. 1990, 112, 90129013. 32. Loo, J.A.; Ogorzalek Loo, R.R.; Udseth, H.R.; Edmonds, G.C.; Smith, R.D. R apid Commun. Mass Spectrom. 1991, 5, 101. 33. Wang, G.; Cole, R.B. Anal. Chem. 1995, 67, 28922900. 34. Craig, T.A.; Veenstra, T.D.; Naylor, S. Biochem. 1997, 36, 1048210491. 35. Guevremont, R.; Siu, K.W.M.; LeBlanc, J.C.Y.; Berman, S.S. J. Am. Soc. Mass Spectrom. 19 92, 3, 216224. 36. Light Wahl, K.J.; Schwartz, B.L.; Smith, R.D. J. Am. Chem. Soc. 1994, 116, 52715278. 37. Ganem, B.; Li, Y.T.; Henion, J.D. J. Am. Chem. Soc. 1991, 113, 78187819. 38. Huang, E.C.; Pramanik, B.N.; Tsarbopoulos, A.; Reichert, P.; Ganguly, A.K.; Trotta, P.P.; Nagabhushan, T.L.; Covey, T.R. J. Am. Soc. Mass Spectrom. 1993, 4, 624630. 39. Rostom, A.A.; Robinson, C.V. J. Am. Chem. Soc. 1999, 121 (19), 47184719.
162 40. Benjamin, D.R.; Robinson, C.V.; Hendrick, F.; Ulrich, H.; Dobson, C.M. Proc. Natl. Acad. Sci. 1 998, 95, 73917395. 41. Veenstra, T.D. Biophys. Chem. 1999, 79, 63. 42. Jecklin, M.C.; Touboul, D.; Bovet, C.; Wortmann, A.; Zenobi, R. J. Am. Soc. Mass Spectrom. 2008, 19 (3), 332343. 43. Poulsen, S.A.; Davis, R.A.; Keys, T.G. Bioorganic and Medicinal Chem. 2006, 14, 510515. 44. Cummins, L.L; Chen, S.; Blyn, L.B.; Sannes Lowery, K.A.; Drader, J.J.; Griffey, R.H.; Hofstadler, S.A. J. Nat. Prod. 2003, 66, 11861190. 45. Gao, J.; Cheng, X.; Chen, R.; Sigal, G.B.; Bruce, J.E.; Schwartz, B.L.; Hofstadler, S.A.; Anderson, G.A.; Smith, R.D.; Whitesides, G.M. J. Medicinal Chem. 1996 39, 19491955. 46. Cheng, X.; Chen, R.; Bruce, J.E.; Schwartz, B.L.; Anderson, G.A.; Hofstadler, S.A.; Gale, D.C.; Smith, R.D.; Gao, J.; Sigal, G.B.; Mammen, M.; Whitesides, G.M. J. Am. Chem. Soc. 1995, 117, 88598860. 47. Wu, J.; Takayama, S.; Wong, C.H.; Siuzdak, G. Chem. Biol. 1997, 4, 653657. 48. Griffey, R.H.; Hofstadler, S.A.; Sannes Lowry, K.A.; Ecker, D.J.; Crooke, S.T. Proc. Natl. Acad. Sci. USA 1999, 96, 1012910133. 49. Daniel, J.M.; Friess, S.D.; Rajago palan, S.; Wendt, S.; Zenobi, R. Int. J. Mass Spectrom. 2002, 216, 127. 50. Wu, Q.; Gao, J.; Joseph McCarthy, D.; Sigal, G.B.; Bruce, J.E.; Whitesides, G.M.; Smith, R.D. J. Am. Chem. Soc. 1997, 119, 11571158. 51. Sannes Lowry, K.A.; Griffey, R.H.; Hofstadler, S .A. Anal. Biochem. 2000, 280, 264271. 52. Doyle, M.L. Current Opinions Biotech. 1997, 8, 3135. 53. Loo, J.A.; Mass Spectrom. Rev. 1997, 16, 123. 54. Senko, M.W.; Speir, J.P. Anal. Chem. 1994, 66, 28012808. 55. Carlin, T.J.; Freiser, B.S. Anal. Chem. 1983, 55, 571574. 56. Wang, W.; Kitova, E.N.; Sun, J.; Klassen, J.S. J. Am. Soc. Mass Spectrom. 2005, 16, 15831594. 57. Dunbar, R .C. J. Chem. Phys 1991, 95 (4), 25372548. 58. Jorgensen, T.J.D.; Roepstorff, P.; Heck, A.J.R. Anal. Chem. 1998, 70 (20), 44274432.
163 59. Kempen, E.C.; Brodbe lt, J.S. Anal. Chem. 2000, 72, 54115416. 60. Tjemberg, A.; Carno, S.; Oliv, F.; Benkestock, K.; Edlund, P.O.; Griffiths, W.J.; Hallen, D. Anal. Chem. 2004, 76(15), 43254331. 61. Schfer, M.; Schmuck, C.; Heil, M.; Cooper, H.J.; Hendrickson, C.L.; Chalmers, M.J.; Marshall, A.G. J. Am. Soc. Mass Spectrom. 2003, 14, 12821289. 62. Weber, L. Drug Discovery Today 2004, 1 (3), 261267. 63. Rowan, S.J.; Cantrill, S.J.; Cousins, G.R.L.; Sanders, J.K.M.; Stoddart, J.F. Angew. Chem. Int. Ed. 2002, 41, 898952. 64. Cousins, G.R.L.; P oulsen, S.A.; Sanders, J.K.M. Curr. Opin. Chem. Bio. 2000, 4, 270279. 65. Huc, I.; Lehn, J.M. Proc. Natl. Acad. Sci. USA. 1997, 94, 21062110. 66. Triolo, A.; Altamura, M.; Cardinali, F.; Sisto, A.; Maggi, A. J. Mass Spectrom. 2001, 36, 12491259. 67. Sussmuth, R.D .; Jung, G. J. Chrom. B. 1999, 725, 4965. 68. Larsson, R.; Pei, Z.; Ramstrm, O. Angew. Chem. Int. Ed. 2004, 43, 37163718. 69. Nawrocki, J.P.; Wigger, M.; Watson, C.H.; Hayes, T.W.; Senko, M.W.; Benner, S.A.; Eyler, J.R. Rapid Commun. Mass Spectrom. 1996, 10, 18601864. 70. Ramstrm, O.; Lehn, J.M. ChemBioChem. 2000, 1, 4148. 71. Hchgrtel, M.; Kroth, H.; Piecha, D.; Hofmann, M.W.; Nicolau, C.; Krause, S.; Schaaf, O.; Sonnenmoser, G.; Eliseev, A.V. Proc. Natl. Acad. Sci. 2002, 99 (6), 33823387. 72. Li, H.; Williams, P.; Micklefield, J.; Gardiner, J.M.; Stephens, G. Tetrahedron 2004, 60, 753758. 73. Poulsen, S.A. J. Am. Soc. Mass Spectrom. 2006, 17, 10741080. 74. Dole, M.; Mack, L.L.; Hines, R.L. J. Chem. Phys. 1968, 49(5), 22402249. 75. Barber, M. Bordoli, R.S.; Sedgwick, R.D.; T yler, A ...N. J. Chem. Soc. Chem. Commun. 1981, 325327. 76. Benninghoven, A. Angew. Chem. Int. Ed. 1994, 33, 10231043. 77. Beckey, H.D. Principles of Field Ionization and Field Desorption Mass Spectrometry Pergamon Press: New York, 1977. 78. Iribane, J.V.; Thomson, B.A. J. Chem. Phys. 1976, 64, 22872294.
164 79. De La Mora, J.F. Analytica Chim. Acta. 2000, 406, 93104. 80. Takats, Z.; Wiseman, J.M.; Gologan, B.; Cooks, R.G. Anal. Chem. 2004, 76, 40504058. 81. Takats, Z.; Wiseman, J.M.; Gologan, B.; Cooks, R.G. Science. 2004, 306, 471473. 82. Jecklin, M.C.; Touboul, D.; Bovet, C.; Wortmann, A.; Zenobi, R. J. Am. Soc. Mass Spectrom. 2008, 19(3), 332343. 83. Schnier, P.D.; Gross, D.S.; Williams, E.R. J. Am. Soc. Mass Spectrom. 1995, 6, 10861097. 84. King, R.; Bonfiglio, R.; Fernandez Metzler, C.; Miller Stein, C.; Olah, T. J. Am. Soc. Mass Spectrom. 2000, 11, 942950. 85. Amster, J. J. Mass Spectrom. 1996, 31 (12), 1325 1337. 86. Marshall, A.G.; Hendrickson, C.L.; Jackson, G.S. Mass Spectrom. Rev. 1998, 17, 135. 87. Marshall, A.G. Int. J. Mass Spectrom. 2000, 200, 331356. 88. Lawrence, E.O.; Edlefsen, N.E. Science. 1930, 72, 376. 89. Lawrence, E.O.; Livingston, M.S. Phys. Rev. 1932, 40, 1937. 90. Hipple, J.A.; Sommer, H.; Thomas, H.A. Phys. Rev. 1949, 76, 18771878. 91. Sommer, H.; Thomas, H.A.; Hipple, J.A. Phys. Rev. 1951, 82, 697702. 92. McIver, R.T. Rev. Sci. Instrum. 1970, 41, 555558. 93. Cooley, J.W.; Tukey, J.W. Math. Comput. 1965, 19, 297301. 94. Marshall, A.G.; Comisarow, M.B.; Parisod, G. J. Chem. Phys. 1979, 71, 44344444. 95. Comisarow, M.B.; Marshall, A.G. Chem. Phys. Let t. 1974, 25 282283. 96. Hardman, M.; Makarov, A.A. Anal. Chem.2003, 75 (7), 1699 1705. 97. Cooper, H.J.; Hakansson, K.; Marshall, A.G. Mass Spectrom. Rev. 2005, 24, 201222. 98. Laskin, J.; Futrell, J.H. Mass Spectrom. Rev.2005, 24, 135167. 99. Gabelica, V.; DePauw, E. Mass Spectrom. Rev. 2005, 24, 566587. 100. Zhai, H.; Han, X.; Breuker, K.; McLafferty, F.W. Anal. Chem. 2005, 77(18), 57775784. 101. Yamada, N.; Suzuki, E.I.; Hirayama, K. Anal. Biochem. 2006, 348, 139147.
165 102. Griffey, R.H.; Sannes Lowry, K.A.; Drader, J.J.; Mohan, V.L Swayze, E.E.; Hofstadler, S.A. J. Am. Chem. Soc. 2000, 122, 99339938. 103. Ogorzalek Loo, R.R.; Smith, R.D. J. Am. Soc. Mass Spectrom. 1994, 5, 207220. 104. Eyler, J.R. Mass Spectrom. Rev. Published Online Feb 24 2009. 105. Polfer, N.C.; Oomens, J. Mass Spectrom. R ev. Published Online Feb 24 2009. 106. Wu, S.; Zhang, K.; Kaiser, N.K.; Bruce, J.E.; Prior, D.C.; Anderson, G.A. J. Am. Soc. Mass Spectrom. 2006, 17, 772779. 107. Caravatti, P.; Alleman, M. Org. Mass Spectrom 1991, 26, 514518. 108. Voet, D.; Voet, J.G.; Pratt, C.W. Nucleotides and Nucleic Acids, 4, 4176.Fundamentals of Biochemistry, Upgrade Editition. John Wiley & Sons: New York, 2002. 109. Smith, D.G.; Stein, W.H.; Moore, S. J. Bio. Chem. 1963, 238, 227234. 110. Graziano, G.; Catanzano, F.; Notomista, E. Thermochimica Acta. 2000, 364, 165172. 111. Voet, D.; Voet, J.G.; Pratt, C.W. Fundamentals of Biochemistry, Upgrade Editition. John Wiley & Sons: New York, 2002, 292293. 112. Usher, D.A. Proc. Natl. Acad. Sci. USA 1969, 62, 661667. 113. Chatani, E.; Hayashi, R. J. Biosci. Bioeng. 2001, 92(2), 98107. 114. Borkakoti, N. FEBS 1983, 162, 367373. 115. Raines, R.T. Chem. Rev. 1998, 98, 10451065. 116. Neira, J.L.; Sevilla, P.; Menendez, M.; Bruix, M.; Rico, M. J. Mol. Biol. 1999, 285, 627643. 117. Volynskaya, A.V.; Kasumov, E.A. Goldanskii, V.I. Int. J. Biological Macromolecules. 2006, 39 (4 5), 256264. 118. Ahmad, F. J. Biol. Chem. 1983, 258, 1114311146. 119. Pan, X.M.; Sheng, X.R.; Yang, S.M.; Zhou, J.M. FEBS Lett. 1997, 402, 2537. 120. Neira, J.L.; Rico, M. Folding and Design. 1997, 2(1), R1R11. 121. Pace, C.N.; Laurents, D .V.; Thomson, J.A. Biochem. 1990, 29, 25642572, 122. Mendez, T.J.; Johnson, J.V.; Nichols, L.S.; Lang, G.H.L.; Eyler, J.R.; Powell, D.H.; Richardson, D.E. Bioconjugate Chem. 2000, 11, 182194.
166 123. Nogues, M.V.; Vilanova, M.; Cuchillo, C.M. Biochimica Biophysica Ac ta 1995, 1253, 1624. 124. Jenkins, C.L.; Thiyagarajan, N.; Swee ney, R.Y.; Guy, M.P.; Kelemen, B.R.; Acharya, K.R.; Raines, R.T. FEBS 2005, 272, 744755. 125. Sundqvist, G.; Benkestock, K.; Roeraade, J. Rapid Commun. Mass Spectrom. 2005, 19, 10111016. 126. Park, C.; R aines, R.T. FEBS Lett. 2000, 468, 199202. 127. Beintema, J.J. Cell Mol. Life Sci. 1998, 54, 763765. 128. Wu, J.; McLuckey, S.A. Int. J. Mass Spectrom. 2004, 237, 197241. 129. Antonini, E., Brunori, M., Neuberger, A., Tatum, E.L., Eds. Hemoglobin and Myoglobin in Thei r Reactions with Ligands ; North Holland Pub. Co.: Amsterdam, 1971. 130. Dautrevaux, M.; Boulanger, Y.; Han, K.; Biserte, G. Eur. J. Biochem. 1969, 11, 267277.
167 BIOGRAPHICAL SKETCH Michelle Sweeney was born in 1982 in Austin, Texas. During her childhood, she a lso lived in Mobile, Alabama, and Huntersville, North Carolina. At all stages of her life, she was encouraged by family and friends to read, think, and experiment. Her early scientific thought was shaped by her father, himself a Ph.D. chemist. After a rigorous and broa d high school education, Michelle graduated from North Mecklenburg High School in 2000 with an International Baccalaureate diploma. From 2000 to 2004, she attended the University of North Carolina Chapel Hill, double majoring in c hemistry and biology and e arning a Bachelor of Science degree. Michelle conducted undergraduate research under the supervision of Dr. Marcey Waters in the area of bioorganic chemistry. In 2004, Michelle moved to Gainesville and joined the University of Florida Department of Chemis try. Under the supervision of Dr. John Eyler and the Analytical Chemistry D ivision, she performed research in the field of noncovalent complexes and mass spectrometry, receiving a Ph.D. in Spring 2009.