2.3%??? awry” ‘L..‘ E il?fla5. afiaf‘g ' " .5" P p '1 :V. t’ 3.3. “— ;‘ v.1”,- 1.31..“ £.& F33»?! . ’v . u. .2 z." P‘:_I'.f, ,.4 , . mu .1 c u ‘4 up If we"; 4}“, PLACE IN RETURN BOX to remove We checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE F—Wl ll MSU Is An Affirmative ActIon/Equel OpportunIIy Institution cmm3o1 CHARACTERIZATION OF THE TRIPHENYLPHOSPHONIUM DERIVATIVE OF PEPTIDES BY FAST ATOM BOMBARDMENT-TANDEM MASS SPECTROMETRY, AND INVESTIGATIONS OF THE MECHANISMS OF FRAGMENTATION OF PEPTIDES By David Scott Wagner A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1 992 tfi gyg-o70 ABSTRACT CHARACTERIZATION OF THE TRIPHENYLPHOSPHONIUM DERIVATIVE OF PEPTIDES BY FAST ATOM BOMBARDMENT-TANDEM MASS SPECTROMETRY, AND INVESTIGATIONS OF THE MECHANISMS OF FRAGMENTATION OF PEPTIDES By David Scott Wagner Fast atom bombardment collisionally activated dissociation tandem mass spectrometry is a powerful technique for the determination of the primary structure of peptides. However, there are factors that frequently prevent successful sequence analysis by mass spectrometry. Two such factors are the poor ionization efficiency of some hydrophilic peptides and, for many peptides, ambiguities in interpretation of the spectra when key sequence ions are weak or absent A Novel and simple procedures for preparing ethyl-triphenylphosphonium derivatives of peptides are described. These procedures allow an ethyl- triphenylphosphonium moiety to be selectively attached to either the N- or C- terminus. Modification of peptides by these chemical methods significantly enhances the efficiency of fast atom bombardment ionization. Moreover, upon collisionally activated dissociation, the derivatized peptides generate a predictable series of sequence ions from either the C-terminus or the N- terminus, depending on the location of the ethyl-triphenylphosphonium moiety. The potential utility of the ethyl-triphenylphosphonium derivative in structure elucidation is illustrated by a comparison of the mass spectra of underivatized and derivatized peptides containing up to 20 amino acid residues, or contain an N-terminal blocking group, or contain a phosphate group, or contain a disulfide bond, or contain a backbone modification. When protonated peptide molecules and cationized peptide molecules are subjected to high-energy collisionally activated dissociation, skeletal bonds cleave generating sequence-specific fragment ions. These bond cleavages usually involve H-shifts. The utility of selective deuterium labeling was applied here to elucidate fragmentation mechanisms. Skeletal bond cleavages in the ionized peptide H-VGVAPG-OH were investigated, in which the molecule was analyzed in the protonated form, cationized form, or as the charge-localized ethyl-triphenylphosphonium derivative. These data provided insights into both charge-induced and remote-site cleavages. Mechanisms have been proposed for each generic type of ions resulting from cleavages along the peptide backbone. to my grandparents Wesley and Verda Thomas iv . ACKNOWLEDGMENTS I would like to thank my research advisor Dr. J. Throck Watson for his support and guidance throughout my graduate school career. I would also like to thank Dr. John Allison, who helped me realize that mass spectrometry is not just a numbers game but involves physical chemistry. I am also grateful to Doug who continually pushed and prodded me to be the best I could be. I am also grateful to Joe for giving me the ability to work with peptides. I need to thank the staff of the Mass Spectrometry Facility. Mel for keeping the computer allve when it's plug should have been pulled, Melinda for all the help you gave, Bev for your continuous hard work that allowed me to focus on research. I want to especially thank Mike, who made the facility more enjoyable with his colorful personality. I also appreciate all the help he gave me in maintaining the instrument and the free electronic lessons which accompanied maintenance. I am grateful for the people which allowed me to keep my sanity. Pete who made my first year here most memorable, especially the weekend parties (Lick-Drink—Suck)ll. Your friendship got me through some hard times. I want to thank Gary for all that he thinks he has done for me, especially the round of golf that you now owe! Mike "Bush Man” Waldo, Big Al ”ate the worm” Flory, and the rest of the slime-balls you made MSU unforgettable. Remember the moped incident, 0.28, Boy's Night Out, the tricycle and the oven doorl What's up with that. I also need to thank the many fine establishments, B'Zars, Dooleys, Rick's and the Dollar, whose happy hours allowed my frustrations to be expelled in a non violent manner. I am also grateful to the fine cuisine of El Azteco, Peanut Barrel and Crunchies, which lined my stomach with best grease money can buy. I am very grateful for the love, encouragement and support I received from my family, especially my parents, while pursuing my degree. Also the turbo twins made life more exciting, especially on birthdays. And last but most importantly, I thank my wife, Laura. I am very blessed to be married to someone as awesome as you. Your love, support and encouragement let me get through the most stressful times of the past few years. We will go Diamondll My final words belong to 0106 "SEEEE YAAAAAIIII LIST OF TABLES... ................ . .................................... . ..................... TABLE OF CONTENTS CHAPTER 1: INTRODUCTION AND OBJECTIVES ..................... I. II. III. IV. Introduction.. .......... ..... . .......................... . ...... . ..... 1 Fast Atom Bombardment .............................................. 6 A. Operating Principles............... ............ .. ................ 6 B. The Matnx .................................... 8 C. The Desorption Process ......................................... 10 D. Surface Activity ..................................... . .............. l3 Collisionally Induced Dissociation ....... ........ 18 A. Introduction ........... .............. . ................ . .......... l 8 B. High Energy Activation ......................................... 20 C. Low Energy Activation .......................................... 21 D. High-Energy vs. Low-Energy ................................. 23 Instrumentation ................... .................................... 2 4 A. Introduction ....... . ............................... . .......... . ....... 2 4 B. Electric Sector ........................................................ 26 C. Magnetic Sector ................. .. .......... .. ............... 27 D. Linked Scans ................................................... . ...... 28 B. Other Instrumentation ......... . ........... ..... . ............ 31 Peptide Sequencing ........................................................ 3 2 vi 1 VI. References ..................................................................... 40 CHAPTER 2: DERIVATIZATION TO ENHANCE THE IONIZATION EFFICIENCY OF FAB AND TO CONTROL THE FRAGMENTATION OF PEPTIDES. I. Introduction. ......... ......... . ............................ 4 8 11. Experimental Section ...... ............................. . ............. 51 III. Peptide Derivatization................... ................................ 54 A. Derivatization of the N -Terminus ........................... 54 B. Derivatization of the C-Terminus ............................ 57 C. Reaction Efficiency .................................................. 59 IV. Signal Enhancement...........................‘. ......................... 60 V. Directing Fragmentation By Derivatization .................... 66 A. Peptides Containing Basic Residues ........................ 69 B. Post-translationally and Chemically Modified Peptides......... ............... ..... ....... 7 3 i) N-terminally Blocked Peptides... .................... 73 ii) Peptides Containing a Disulfide Bond ............. 78 iii) Phosphorylated Peptides .............................. 90 iv) Backbone Modified Peptides ......................... 93 v) Cyclic Peptides..... ......................................... 97 vi) Lipoidal Peptides ......................................... 100 VI. Conclusion. ............. . ........ ................... . ........ . ............. 102 VII. References ................. .. ............ . ............... ........... 106 CHAPTER 3: FRAGMENTATION MECHANISMS OF PEPTIDE BONDS BY FAB-CAD-MS/MS I. Introduction ................................................................... 110 II. Experimental Section ..................................................... 120 vii A. Peptide Synthesis .......................................... ‘ ........ 120 B. Peptide Derivatization ............................... . ............ 121 C. Mass Spectrometry ................................................ 123 III. Protonated Peptides ..................................................... 123 IV. Fragmentation of Charge-Localized Derivatives ............. 145 V. Comparison of Charge—Directed and Remote-Site Fragmentation.... ........................................................ 164 VI. References....................................................... ............. 166 CHAPTER 4: FRAGMENTATION OF CATIONIZED PEPTIDES I. Introduction .................................................................. 178 11. Experimental Section .......................................... . ......... 180 III. CAD of Cationized Peptides..... .............. . ....................... 183 A. [an-H]Cat+ Fragment Ion Series ........................... 195 B..[xn-H]Cat+ Fragment Ion Series ............................. 197 C. [bn-H]Cat+ Fragment Ion Series ........................... 197 D. [yn+H]Cat+ Fragment Ion Series ........................... 199 E. [yn-H]Cat+ Fragment Ion Series..... ...................... 199 F. [cn+H]Cat+ Fragment Ion Series ............................ 201 G. [znlCat+ Fragment Ion Series ........................... 204 H. [dn]Cat+, [wn]Cat+, and [vn]Cat+ Fragment Ions ...... ....... ............... . ........ 204 IV. Fragmentation Mechanisms ...... . ...... ..... .......... 207 V. References .................................................................... 214 viii Table Table Table Table Table Table Table Table Table Table Table Table Table Table Table 3.1a 3.1b 3.2 3.3a 3.3b 4.1a 4.1b 4.2a 4.2b 4.3a 4.3b 4.4 4.5 4.6 4.7 LIST OF TABLES N-terminal CAD Fragments From The Protonated Peptide ................................................................... 124 C-terminal CAD Fragments From The Protonated Peptide ................................................................... 125 Intensities of the Fragment Ions ............................. 126 Fragments from the N-terminal Derivative ............. 146 Fragments from the N-terminal Derivative ............. 147 N-terminal ions from CAD of [peptide]Li+ ................ 186 C-terminal ions from CAD of '[peptidelLi‘I' ................ 187 N-terminal ions from CAD of [peptide]Na+ ............... 18 8 C-terminal ions from CAD of [peptide]Na+ ............... 189 N-terminal ions from CAD of [peptide]K+ ........ . ......... 190 C-terminal ions from CAD of [peptide]K+ .................. 191 Intensities of Fragment Ions Produced from CAD on [VGVAPG]Li+......................................... .............. l 9 2 Intensities of Fragment Ions Produced from CAD on [VGVAPG]Na+ .......... . ............................... ....... ..193 Intensities of Fragment Ions Produced from CAD on [VGVAPG]K+ ........................................................ 194 Proton Affinities of the Fragment Ions .................... 209 ix Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 1.1 1.2 1.3 1.4 1.5 1.6a 1.6b 1.7 1.8 1.9 1.10 2.1 LIST OF FIGURES Tandem Mass Spectrometry ................................. 4 Fast Atom Bombardment ...................................... 7 View of the sample surface during fast atom bombardment .......................................... .. ........... 9 Representation of the desorption process by a collisional cascade and a thermal spike regime ..... 12 Comparison of hydrophobic and hydrophilic peptides in the FAB matrix ................................... 14 Representation of hydrophilic peptides interacting with the matrix molecules ..................................... l6 Representation of hydrophobic peptides interacting with the matrix molecules ..................................... 16 Double Focusing Mass SpectrOmeter ...................... 25 Generic peptide displaying the commonly accepted nomenclature for the fragmentation along the backbone .............. .......................... ........ 3 3 Assumed structures for the commonly detected fragment ions ........................................................ 34 FAB-CAD-B/E mass spectrum of Met-Enkephalin.. 34 Structures of the two reagents developed to derivatize peptides ................................................ 52 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 2.2a 2.2b 2.3 2.4 2.5a-c 2.6a-b 2.7a-b 2.8a-c 2.9a-b 2.10a-b 2.11a-b Synthesis of bromoethyl-triphenylphosphonium reagent ............................................................... 53 Synthesis of aminoethyl-triphenylphosphonium reagent........ ........... . .......... . ................................ 5 3 N-terminal Ethyl-TPP derivatization of peptides. 55 C-terminal Ethyl-TPP derivatization of peptides. 5 8 The FAB mass spectra of a) 50 pmol underivatized VQAADYING, b) 5 pmol of N-terminally derivatized VQAADYING, and c) 5 pmol of C-terminally derivatized VQAADYING ............................ ...... 61 FAB mass spectra of a) 5 pmol of the protonated peptide YGGFL, b) 1 pmol of the N-terminal derivative and C) 1 pmol of the C-terminal derivative ........................................................... 6 3 FAB mass spectra of a) 100 pmol of a mixture containing ALG, VGVAPG, and YGGFL, and b) N- terminal. derivatization on 25 pmol of the peptide mixture .............................................................. 6 5 FAB-CAD-B/E mass spectra Of a) 200 pmol of VGVAPG, b) 100 pmol of the N -terminal derivative and c) 100 pmol of the C-terminal derivative ..... 68 FAB-CAD-B/E mass spectra of a) the protonated peptide YQEAFRRFFGPV, and b) 250 pmol of the N-terminal derivatiVCOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO000...... 7 l FAB-CAD-B/E mass spectra of a) the protonated peptide RVYHPF, and b) the C-terminal derivative .......................................................... 72 FAB-CAD-B/E mass spectra of a) the protonated peptide pQGLPPGPPIPP, and b) the C-terminal derivative .................................................. . ....... 75 xi Figure Figure Figure Figure Figure Figure Figure Figure Figure 2.12a-b 2.13 2.14a-b 2.15a-b 2.16a-c 2.17a-b 2.18a-b 2.19a-b 2.20a-b FAB—CAD-B/E mass spectra of a) the protonated peptide Ac-RRPYIL, and b) the C-terminal derivative ......... . ................................................. 7 7 The normal process to analyze a protein or peptide containing a disulfide bond .................... 80 FAB-CAD-B/E mass spectra of a) the protonated peptide PFCNAFTGC—NH2 containing a 8-8 bond between the cysteine residues, and b) the N- terminally derivatized peptide ........................... 81 FAB-CAD-B/E mass spectra after 5 minutes of bombardment on a) the protonated peptide PFCNAFTGC-NHz containing a 8-8 bond between the cysteine residues, and b) the N-terminally derivatized peptide .............. , .............................. 83 FAB mass spectra of a) +TPP-PFCNAFTGC-NH2 , containing a S-S bond between the cysteine residues, at 0 min. of bombardment, b) after 10 min. of bombardment, and c) the reduced form of the N-terminal derivative ............................... 84 FAB-CAD-B/E mass spectra on a) the protonated peptide PFCNAFTGC-NHz, with a reduced S-S bond and b) the N-terminal derivative ........................ 8 6 FAB-CAD-B/E mass spectra on a) the protonated peptide AGAADCFWKYCV, containing a disulfide bond between the cysteine residue and b) the reduced form of the protonated peptide ............. 8 7 FAB-CAD-B/E mass spectra on a) the N-terminal derivative of AGAADCFWKYCV, containing a disulfide bond between the cysteine residue and b) the reduced form of the N ‘tcrminal derivative ........................................................... 87 FAB-CAD-B/E mass spectra of a) protonated phospho-kemptide, and b) C-terminal derivative of phospho-kemptide .......................................... 92 xii Figure Figure Figure Figure 2.24a-b Figure Figure Figure Figure Figure Figure Figure 2.21a-b 2.22a-b 2.23 2.25 2.26 2.27 2.28 3.1 3.2a 3.2b FAB-CAD-B/E mass spectra of a) protonated peptide HPFHLvVY (containing a methylene- amine group on L), and b) the N-terminal derivative ........................................................... 95 FAB-CAD-B/E mass spectra of a) protonated YAGFwLRRI (containing a thiomethylene ether linkage), and b) the N-terminal derivative.......... 96 Structure of a cyclic peptide containing a free amino-terminus .............. . ................................... 9 8 FAB-CAD-B/E mass spectra of a) the protonated cyclic peptide and b) the N-terminal derivative.. 99 Spectrum of the N-terminally derivatized lipoidal protein ................................................................ 101 FAB mass spectrum of a mixture of three peptides after N-terminal derivatization. The mixture was obtained from a tryptic digest of B-lactoglobulin 103 A CAD- B/E mass spectrum of the N-terminally derivatized peptide (GLDIQK) from the tryptic fraction in figure 2. 25 ......................................... 105 A CAD-B/E mass spectrum of the N-tcrminally derivatized peptide (IIAEK) from the tryptic fraction in figure 2.25 ................................. . ....... 105 Possible structures of the fragment ions occurring along the backbone of the peptide .......... , ........... 114 Fragmentation of a protonated peptide to form a and x via a common intermediate involving a hydride ion ......................................................... 116 Fragmentation of a protonated peptide to form a and x via a common intermediate involving the expulsion of an H2 molecule ............................... 117 xiii Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 3.3 3.4a 3.4b 3.5 3.6 3.7 3.8 3.9a 3.9b 3.10 3.11 3.12 3.13 Structures of the deuterium labeled peptide analogues of H-VGVAPG-OH ........... . ....... .. ........... 119 N-terminal ethyl-TPP derivatization of a peptide 122 C-terminal ethyl-TPP derivatization of a peptide 122 Formation of the an ion form the bn ion ............ 130 Formation of the xn ion via a 1,2-elimination reaction, remote from the charged site........ ....... 131 Formation of the bn ion from protonation of the amide nitrogen.... ................... ................ 133 Two most probable mechanisms for the formation of the yn ion ...................................... 135 Formation of the yz ion from the a-labeled alanine containing peptide. Figure demonstrates that the deuterium on the a-carbon is not transferred ......................................................... 1 3 6 Formation of the y; in from the d7-labeled peptide. Figure displays that a deuterium on the nitrogen is transferred in the fragmentation process ........... . .......... ....... ........................ 13 8 Formation of the y4 ion. Figure shows the formation of a neutral 6-membered ring and H transfer through an 8—membered ring .......... . ..... 139 Two possible mechanisms for the formation of the en fragment ion ........................................ 141 A possible mechanism for the formation of the C3 fragment ion ............... ......... . ..... 142 Formation of the bn ion form protonation of the amide oxygen ....... . ............................................. 144 xiv Figure Figure Figure Figure Figure Figure Figure Figure 3.21a-b Figure Figure Figure Figure 3.14 3.15 3.16 3.17 3.18 3.19 3.20 3.22 3.23 3.24 4.1 Formation of and ion by charge-remote fragmentation involving a 1,2-elimination reaction ..................... ........ ...... . ...... ......... 15 1 Formation of xn-l ins by charge-remote fragmentation involving a 1,2-elimination reaction ....... . ....... .. ......... . ..... ......... . ............. 153 Formation of the a and x ions via a common remote-site mechanism, where an H2 molecule is CXPelledOOOOOOOOOOOOOO ......... 0...... OOOOOOOOOOOO 000...... ....... l 5 4 A possible mechanism for the formation of the X2 fragment ion... ............ ........... ..... 155 A remote-site fragmentation mechanism for the formation of b and y ions involving the elimination of H2. The detected ion contains the ethyl-TPP derivative .................................... 159 Formation of the bn-I-l and yn-l via a common remote-site mechanism, involving a 1,2- elimination reaction .......... .............................. 160 A possible mechanism for the formation of the C3 fragment ion............ .............................. . ....... 162 The FAB-CAD-B/E mass spectra of a) protonated H-VGVAB3dPG-OH displaying the a4 ion and b) the protonate peptide H-RVGVAB3dPG—OH displaying the a3 ion ........ ..... ....... 168 Structures of the various an ions that form ........ 169 Structures of the fragment ions produced from protonated peptides ............................................ 173 Structures of the fragment ions produced from derivatized peptide..... .............................. . ......... 174 Structures of the deuterium labeled peptide analogues of H-VGVAPG-OH ................... ....... 181 XV Figure 4.2 Figure 4.3 Figure 4.4 Figure 4.5 Figure 4.6 Figure 4.7 Figure 4.2 Figure 4.9 Figure 4.10 Structure Structure Structure Structure Structure Structure Structure of the [an-H]Cat+ fragment ion ............. 196 of the [xn-H]Cat+ fragment ion ............. 198 of the [yn+H]Cat+ fragment ion ............ 200 of the [yn-H]Cat+ fragment ion ............ 202 of the [cn+H]Cat+ fragment ion ............. 203 of the [zn]Cat+ fragment ion ................. 205 of the [dn]Cat+, [wn]Cat+, and [vn]Cat+ fragment ions ................................ .. ................... 206 ' The formation of the a-type and x-type ions through a common remote-site mechanism involving the expulsion of an H2 molecule .......... 211 Formation of the b-type and y-type ions through a common remote-site mechanism involving the expulsion of an H2 molecule........... ........... .. ........ 213 xvi CHAPTER 1 INTRODUCTION AND OBJECTIVES I. Introduction The applicability of mass spectrometry in elucidating the primary structure of peptides was established in the late 1950s (1). However, its application was restricted through the 1970s to a determination of the molecular weight and the amino acid sequence of only small peptides after extensive chemical derivatization, to make the peptide volatile (2,3). Converting peptides to N-trifluoroethyl-O—trimethylsilyl polyamino alcohols made them amenable to analysis by GC-MS(4). An enzymatic digest or partial hydrolysis of large peptides was used to produce di- to hexapeptides. These small peptides could be derivatized as a mixture, then analyzed by GC-MS. Alternatively, N-acetyl- N,O-permethylated peptide derivatives (3) were analyzed by direct probe introduction. These early mass spectrometric procedures were mainly used to determine the modification of N-terminally blocked peptides, those not having a free amino on the N-terminus. These blocked peptides could not be analyzed by the traditional Edman degradation (6), which is the most frequently used sequencing technique. Pehr Edman devised a chemical method for labeling the amino terminal residue and cleaving it from the peptide without disturbing the peptide bonds between the other amino acids (6). Phenyl isothiocyanate reacts with the uncharged terminal amine of the peptide to form a phenylthiocarbamoyl derivative. Under mildly acidic conditions, a cyclic derivative (phenylthiohydantoin, PTH, amino acid) is liberated, and the amino acid is 1 2 identified by HPLC retention time. The remaining peptide is put through subsequent cycles of degradation. By this procedure, the sequence of 30 amino acid residues can be determined. Since the introduction of fast atom bombardment (FAB) in 1981 (7), mass spectrometry has become an important method for analyzing peptides, because the requirements for sample volatility have been eliminated. The most abundant ions found in FAB are protonated molecules (M+H)+ when analyzing in the positive mode, and deprotonated molecules (M-H)‘ in the negative mode, thus allowing molecular weight determination. One of the most important contributions of FAB mass spectrometry to protein structure determination has been the capability to obtain precise, accurate molecular weights of peptides on 10—50 pmol of sample, which is comparable to the amount required for Edman degradation (8). In the past, mass spectrometry was used in conjunction with the Edman technique for analysis of unknown peptides. The molecular weight information was valuable in confirming the results of amino acid analysis, determining the number of cycles to program the automated mlcrosequenator, improving the confidence of residue assignments, showing whether the peptide was fully sequenced, and identifying the presence of post-translational modifications (9- 11). Although the molecular weights of peptides can be easily obtained from the mass of (M+H)+, the stability of the protonated peptide prevents the formation of fragment ions to any relatively significant degree, which are needed to determine the amino acid sequence from the mass spectrum. However, there have been a number of reports of using the fragment ions obtained by FAB mass spectrometry to determine the amino acid sequence of both known and unknown peptides (12-16). The principal problem with this 3 method is that large amounts peptide are necessary to obtain a mass spectrum with a sufficient number of fragment ions. Collisionally activated dissociation tandem mass spectrometry (17-20) is a technique which overcomes this problem and frequently produces sequence specific fragment lens. This method requires two mass analyzers and a collision cell, Figure 1.1. The selection of a precursor or parent ion (M+H)+ is achieved in the first mass spectrometer, followed by collisions with an inert gas in a collision cell, a process which supplies energy for fragmentation; the second mass spectrometer analyzes the resulting fragment (daughter) ion spectrum. Although the Edman degradation technique is sensitive and can be automated, the process is relatively slow, taking approximately one hour per residue. Ambiguities in residue assignments occur when the technique is applied on small quantities or mixtures because more than one signal results. As the sequencing proceeds toward the C-terminus of a peptide, the residue signals constantly decrease because of imperfect reaction yields in previous operational cycles. As the peptides become smaller, they tend to become more hydrophobic in character and become more soluble in the extraction solvents, thus tend to be lost, consequently leading to an incomplete sequence determination due to loss of material. Furthermore, some post-translational modifications, important for protein activity, are lost by the Edman process because of the harsh chemical conditions employed. Modified amino acids can be incorrectly assigned as another amino acid with identical retention time or simply labeled as an unknown residue. Finally, one of the major problems is that the Edman procedure cannot be used with N-terminally blocked peptides, because of the necessity for a free amine at the N-terminus. 1W ' I f . I”? ' I 3.0: A . p . ugh 3 1 a»; I II II | a too zoo coo coo «W766 coo Mass Selection ,4' "i. Collisionally Activated ,0: m . Dissociation CAD / \ AA ‘ Collision Product Mass Analysis - “Lu- Qua-‘38.) oc-uu-ea Figure 1.1: Tandem Mass Spectrometry. 5 Tandem mass spectrometry has some advantages over the traditional Edman degradation for the sequencing of peptides. Sequencing peptides by tandem mass spectrometry is considerably faster, on the order of minutes, to obtain the mass spectrum and a couple of hours to resolve the amino acid sequence. This time can be reduced considerably by the use of a computer program (21-22) designed to interpret the amino acid sequence from a mass spectrum. Contrary to requirements for Edman degradation, peptide mixtures are not usually a problem in FAB-MS and CAD-MS/MS can be performed on the protonated peptides One at a time. Also tandem mass spectrometry is especially useful for identifying N-terminally blocked and post-translationally modified peptides. In addition to the capacity to increase the fragmentation by collisionally activated dissociation, tandem mass spectrometry has other related advantages. Matrix peaks and impurities are excluded (4.18.2326) because all ions in the daughter spectrum are directly related to the protonated peptide, which is selected as the precursor ion. This technique has proven to be useful for differentiating isomeric compounds (27). This is not to suggest that the conventional methods should not be utilized, rather that tandem mass spectrometry should be a complementary method of analysis. ' The objective of this chapter is to (i) introduce FAB ionization, (ii) describe the different modes of collisionally activation dissociation (CAD) used during MS/MS, (iii) introduce the instrumentation used in this research, that being a double focusing mass spectrometer, (iv) introduce some other types of mass spectrometers and ionization techniques that have been used to sequence peptides, and (v) introduce the methodology of sequencing peptides by mass spectrometry. 1]. Fast Atom Bombardment A. Operating Principles The ionization technique fast atom bombardment (7) made it possible to ionize large polar molecules directly without chemical modification of the peptide. FAB employs a neutral Ar or Xe atom beam, having several keV of translational energy, to sputter ions (secondary ions) from samples dissolved in a liquid matrix. This process is depicted by Figure 1.2. The liquid matrix plays a crucial role in reducing sample damage under high flux primary particle bombardment (7) by constant diffusion of fresh sample to the surface from the bulk of the solution. In addition to neutral atom sources, cesium ion guns (28) also can be used for bombardment. It is generally accepted that mass spectral data obtained during bombardment with ions or atoms are similar (28). This implies that the use of a liquid matrix is more important than the charge on the primary particle, thus leading to the term liquid secondary ion mass spectrometry (29). Cesium ion sources can produce more focused primary beams giving rise to more efficient ionization than neutral atom sources, and result in lower detection limits (30). Also, the increase in ion yield is related to the absence of a supporting gas and the larger mass of the primary particle (29-30) Mass Analyzer l gEl Secondary ___> Ion Beam A m l “g I}; 3 33—3—3 [mple Fast Particle Gun Figure 1.2: Fast Atom Bombardment B. The Matrix The presence of the liquid matrix provides some significant advantages including a decrease in fragmentation of the adduct ion in comparison to the case involving desorption from a solid surface (31 ). The matrix also promotes a stable, reproducible ion current lasting for longer periods of time (32), due to the diffusion process. Figure 1.3 represents a closer view of the sample surface. This figure illustrates the surface destruction of the sample by the high flux particle beam and the desorbed species produced. The viscous liquid matrix replenishes the destroyed area and provides additional analyte to the sample surface. Without this renovating effect, the analyte signal would not last nearly as long nor be as stable. Although the FAB matrix provides benefits for desorption ionization, the presence of a viscous matrix material in a high concentration presents several notable problems. These include poor sensitivity and limit of detection, high background ion counts at every mass, intense cluster ions from the matrix, and ion suppression effects, in which the formation of certain ions from the sample is inhibited by the presence of other compounds in the sample. Some of the requirements for the liquid matrix are (33-34): 1. the sample must be soluble in the matrix 2. low vapor pressure 3. chemically inert, must not react with the sample 4. the viscosity should be low enough to ensure the diffusion of solutes to the surface 5. ions from the matrix must be unobtrusive as possible V) FAST ATOM \3 BEAM \3 @ \8 © @ ® @ \3 ® ® @ @ GLYCEROL SOLUTION \3 OF ANALYTE Figure 1 .3: View of the sample surface during fast atom bombardment. IO It should also be noted that matrices can be chosen to help promote the formation of secondary ions for a particular analyte. Glycerol has been the most extensively used and reported FAB matrix. Thiols have been shown to improve the sensitivity at high mass for several analytes (35-36). Some commonly used matrices in the negative ion mode include diethanolamine (DEA) and triethanolamine. Less polar matrices aid in the analysis of less polar molecules (37-39). C. The Desorption Process Although the desorption process is not completely understood, it is based on a combination of factors (40-41). One mechanism that explains the desorption process is the direct desorption of prefdrmed ions formed in the matrix (42), including both protonated molecules and adducts with metal cations (Na+, K+, LP) (43). Some explanations leading to this conclusion include the fact that addition of acids to the matrix enhances the relative (M+H)+ ion abundance (44-46) and the concentration of the protonated molecule in the matrix is proportional to the relative abundance of the (M+H)+ ion observed in the mass spectrum (42,47). The second main mechanism describing the desorption process is that involving desorption of neutral molecules into the gas phase, followed by gas phase ionization (48-49) in the high-pressure region just above the sample surface, also called the selvedge region (50). The formation of (M+H)+ in the gas phase is substantiated by the similarity of FAB and chemical ionization (Cl) mass spectra, calCulations indicating similar source pressure requirements (51), and experiments demonstrating that gas phase basicity of a compound is more important for the production of an ion than the solution basicity (52). l I In all probability, both mechanisms are concurrently active. The mechanism that predominates depends on the matrix pH, the presence of metal cations or other contaminants, and the relative solution and gas phase pKa's of the matrix and analyte species (8). The ion production mechanisms for FAB are considered identical to those of solid SIMS (34) except for an important difference in the extent of fragmentation, which is much greater in solids. Sputtering of solids is described as either a linear cascade of collisions or as a thermal spike regime (34). The capability of the analyte's surface (target) is responsible for the gradual loss of kinetic energy of the primary particle (projectile). Momentum is transferred from the projectile to the target particles. Many target atoms are displaced or recalled by a direct collision with the primary particle. These atoms, in turn , collide with additional target atoms in the sample which emerge in all directions from the point of impact. It the induced motion of the target molecules is directed toward the vacuum with enough kinetic energy, the particle will be sputtered (Figure 1.4). The sputtering yield is related to the size and energy of the primary particles and to the surface binding energy of the target.(53-56) In the liquid matrix the molecules and ions are bound less tightly than in solids. During the collision cascade, the excess energy is absorbed by the molecules and the energy dissipation occurs through translational excitation of the matrix (57). This may lead to production of larger solvated clusters in the gas phase, in which the energy dissipation is'readily attained by sequential losses of solvent molecules without breaking intramolecular bonds (58). 12 ’69 0 00}.— OOO’V- GO 0/. co co O Q @\Q 3 9 g Collsional Cascade Regime \ / ../."\.... wowed/Tee O O O 9 C O O 9 Thermal Spike Regime Figure 1.4: Representation of the desorption process by a collisional cascade and a thermal spike regime. D. Surface Activity The penetration depth of the incident particle into the liquid is believed to lie within 50 to 100 angstroms of the gas liquid interface (55). When the sample volume is approximated as a portion of a sphere, the maximum depth of the sampled volume is equivalent to 5 molecular layers (34). Therefore all the processes associated with the energy deposition and collision cascade are centralized near the surface of the matrix. Since desorption depends on sputtering from the matrix surface, the surface concentration of the peptide is an important factor. Studies correlating the detection limits for a given peptide with its surface activity have been reported (56-58). The capability of an analyte to occupy the surface area is dependent upon interactions (hydrophobic, hydrophilic, hydrogen bonding...) between the matrix molecules and the analyte. Most of the matrices tend to be hydrophilic in nature. The more surface active compounds can interact favorably with the matrix molecules and distribute themselves throughout the entire sample. Surface active compounds do not favorably interact with the matrix, thus the matrix molecules interact with themselves and exclude the surface active compounds. A segregation of the surface active compounds occurs, moving them towards the surface and away from the interactions between the matrix molecules. Thus the surface active compounds will have a higher concentration of molecules near the surface than surface inactive compounds (Figure 1.5). . A=SURFACE ACTIVE (HYDROPHOBIC) ' I-SURFACE INACTIVE (HYDROPHILIC) Figure 1.5: Comparison of hydrophobic and hydrophilic peptides in the FAB matrix. 15 Hydrophilic and hydrophobic interactions of peptides with water molecules have been explained in great detail (59-60). By using the same rationale except replacing the water molecules with polar matrix molecules, one can imagine the interactions occurring in the FAB matrix. Hydrophilic effects, bindings, or interactions comprise the attractions of peptides for matrix molecules, (the polar sites and hydrogen bonding forces), and result in matrix- peptide associations (Figure 1.6a). This association increases the dipole strength of the attached matrix molecules. Attached matrix molecules may be forced out of their associations with hydrophilic sites into the bulk of the surrounding matrix or into association with other, more hydrophilic sites. Thus the degree of interaction between a given peptide and the matrix depends on the capability of a given peptide to associate with the matrix over other peptides or molecules in the sample. . Hydrophobic groups do not have attractions for matrix molecules; when hydrophobic groups are in contact with the matrix molecules, they are subject to a _“thrust" by the surrounding matrix molecules, which are trying to interact with one another. Because matrix molecules tend to reduce their associations with hydrophobic groups, the interactions between matrix molecules. prompted by Van der Waals’ forces, is considerable enhanced. Interactions of matrix molecules results in extrusion of hydrophobic groups initially between them (Figure 1.6b). Thus. the association of hydrophobic groups can be considered in terms of loss of the matrix/peptide surface free energy as well as entropically by an increase in hydrogen bonding between matrix molecules that where surrounding the hydrophobic peptide. 16 @ mom @ Figure 1.6a: Representation of hydrophilic peptides interacting with the matrix molecules. @ mom Figure 1 .6b: Representation of hydrophobic peptides being concentrated in order to increase the entropy of the matrix mo ecu es. I7 For the purposes of this discussion, the suppression effect is the result of a diminished tendency of compounds to occupy the surface layers of the sample solution. Suppression effects in desorption ionization of less surface active components by more surface active components (when both are present in equimolar mixtures) have been well documented (56-57,61-63). In addition, the FAB signal for hydrophilic peptides is often weak or cannot be detected because of the suppression effects by the matrix and/or when other more hydrophobic peptides are present in the sample (40.60-61.73-75). This idea was further developed when correlations were made between the hydrophobicity scale of Bull and Breese (64) and the response of a peptide mixture to FAB analysis (65-72). It was concluded that the hydrophobicity/hydrophilicity index of peptides is a major determinant as to whether a signal for the molecular ion species will be recorded using FAB MS (73). The more hydrophilic the peptide the more likely its (M+H)+ signal will be suppressed. A study by Capriloli and Moore (62) presented evidence that a number of physical and chemical properties of peptides, including their hydrophobic or hydrophilic nature, play important roles in their capacity to form ions by FAB MS. They indicated that factors such as the charge state of the peptide in the matrix and its tendency to form secondary structures in the water/matrix and solution can also play important roles. However, these factors combine to determine the suppression effect of an ion, it is clear that competition of the various species for the surface layers of the liquid sample is the dominant factor for the formation of ions from a particular peptide. Continuous-flow FAB (CF-FAB) was developed to minimize difficulties associated with the matrix by making use of a water matrix with only small concentrations of glycerol or another viscous organic liquid (74). This 18 technique has been shown to provide a higher sensitivity and lower limit of detection by considerably reducing the background chemical noise and the abundance of the matrix ions (75). Also, CF-FAB was reported to diminish the ion suppression effect in cases where mixtures of peptides were analyzed (76). Other successful approaches used to increase the FAB signal response of peptides have involved the formation of hydrophobic derivatives of hydrophilic peptides (66-68,73,77-79). III. Collisionally Induced Dissociation A. Introduction FAB is considered a soft ionization technique because very little fragmentation occurs in the desorption process. Methods of supplying an ion with enough energy to dissociate include collision with a surface(80-82), absorption of energy from a laser beam (83), use of an electron beam (84), and the most commonly used method of collisions with an inert gas, collisional activation (85-88). Although collisionally activated dissociation (CAD) was observed in 1913, it was not used by mass spectrometrists until the late 19603 (89-90) and only recently applied to peptide ions. (91-92) The CAD process can be broken down into two consecutive steps (93). The first is a fast (1045- 10'14 sec) collisional activation step in which some fraction of the initial-translational energy of the accelerated ion is transformed into internal energy of both ion and target. The second step in the CAD mechanism is the dissociation of the energized (and usually isolated) ion. The 19 yield of fragment ions after CAD depends on the possibility of unimolecular decomposition of the precursor (parent) ion after excitation. l The quasi-equilibrium theory (QET) is used to explain the rates of such dissociation reactions (94). GET theorizes that unimolecular decomposition reactions depend upon the random distribution of internal energy of the ion among all vibrational modes of that ion. The internal energy randomizes in the excited ion and eventual localization of the excess energy in a bond produces cleavage. In addition to dissociation, other processes occur from the interactions of the ion with a neutral target gas. The precursor ion can be scattered at an angle greater than the acceptance angle of the mass spectrometer, and thus not observed. The amount of scattering is proportional to (ac/v)2 where ac is the ion target interaction distance (impact parameter) and v is the ion velocity. Scattering can be reduced by increasing the ion velocity or decreasing the target gas size. Another important process is "charge exchange”, where the target gas (9) is ionized at the expense of the precursor ion (mt). 9 " mi "">. Q+ * m Other less important reactions are charge stripping: m“ ---—> m2+ + e‘ and charge inversion: m- . g ____, m- . 92+ All of these alternative processes are in competition with CAD and reduce the tendency to produce fragmentation (84,86-87). There are 3N-6 vibrational modes in a nonlinear ion consisting of N atoms. It can be seen that the number of vibrational modes is directly proportional to the molecular mass. QET requires the internal energy to be randomly distributed between the vibrational modes in an ion, thus the average 20 energy in any given mode must decrease with increasing size of the ion. Since the decrease is inversely related to the mass of the ion, the fragmentation of the ion starts to decrease at a threshold (95-97). This section will focus on two methods of ion activation using collisions of accelerated ions with a stationary gas phase target in the (i) high energy (keV) and (ii) low energy (eV) ranges of laboratory ion kinetic energy. B. High Energy Collisional Activation Efficient conversion of ion translational energy into internal energy occurs when the collision interaction time (tc) and the period of the internal mode that is being excited (I) are similar (93,98). Collisions at KeV energies are anticipated to result in excitation of internal electronic modes because the test 0'14 see is comparable to the Bohr period of an electron in valence orbit of a polyatomic molecule (99). Redistribution of the excitation energy to vibrational modes, an assumption of QET, ultimately results in bond cleavage. A collision between a fast moving ion and a stationary target gas atom has a limited amount of kinetic energy obtainable for conversion into internal energy and is given by: Ecom "-' mt Elab/(mt‘fml) where m and mi are the masses of the target atom and the precursor ion, and Elab is the kinetic energy of the ion given by the acceleration potential (93). The energy available for fragmentation is increased by either increasing the kinetic energy (raising accelerating potential) of the ion or the molecular weight of. target gas. One can also see that, increasing the mass of the parent ion will decrease available energy by 1/mi. 21 Studies involving internal energy depo'sition (100-102) determined that although the average energy deposited into an ion in a high energy collision is 1-39V, the distribution contains a high energy tail extending beyond 15eV of internal energy. This is important for the formation of certain ions that require high internal energy. As discussed above, the major pathways in competition with CAD are charge exchange and scattering. Reduction of the ion target size and increasing the ion velocity reduces these competitive processes, thus He is the most commonly used collision gas for high-energy CAD (103-104). C. Low Energy Collisional Activation Low-energy collisions are usually carried out in a quadrupole reaction chamber. The quadrupole collision chamber uses a radio frequency (RF) field to contain and transmit ions. A number of parameters can be optimized including the collision energy (accelerating potential) and the number of activating collisions (a function of pressure) (85,105). Precursor ions are selected by either another quadrupole (triple quadrupole instrument) or a double focusing analyzer (hybrid configuration), and product ions are analyzed by a quadrupole mass filter. The transmission efficiency of a quadrupole is high because of the strong focusing properties in the RF-only mode; ion trajectories are actually stabilized after collision and dissociation. Operating at high pressures induces multiple collisions for the parent ions, thus increasing the fragmentation efficiency. Low energy collisions (<100 eV) do not transfer translational energy to electronic internal modes very efficiently. The collision interaction time is on the 22 order of 10'13 seconds, considerably longer than the period of the electronic internal mode, thus the possibility of excitation of such a mode is reduced with respect to excitation by high-energy CA (93). The interaction time of 10‘13 seconds is comparable to the reciprocal of most vibrational frequencies. These low-energy collisions are nonadiabatic and interaction is thought to have "impulsive character" that efficiently induce energy transfer (93). The transfer of the parent ion's translational kinetic energy to internal vibrational energy can occur when an ion collides with a neutral gas with kinetic energy in the 9V range (106). The maximum energy (center of mass kinetic energy Ecom) available for conversion from the kinetic to internal energy distributed within the incident ion is described by the expression: Ecorn = Elab mt / Imp 4' mt (mp / mpl)I where mp is projectile mass, mp; is the impact portion of the projectile, and rat is the target mass. Elab, the accelerating potential, determines the kinetic energy of the parent ion. Following collisional activation, the excited ion can be collisionally stabilized by third body collisions or if the ion has sufficient internal energy, it can undergo unimolecular decomposition to form fragment ions. The mathematical equation suggests a massive target gas is preferred as a collision gas for CAD experiments, but the size of the target gas is limited so not to produce unwanted reactions (107-109). The most commonly used gases for low-energy CAD of peptides is argon and an argon/xenon mixture. Also, the pressure of the target gas influences the average internal energy by multiple collisions, increasing the energy deposited in a stepwise fashion. The internal energy deposited by low-energy collisions is similar to high-energy collisions except the distribution does not contain the high energy tail. Thus ions requiring high energy cannot form through low-energy activation. 23 D. H igh-Energy CAD vs. Low-Energy CAD Several studies have been reported on the comparison of high-energy CAD and low-energy CAD (110-111). An extensive study that compared the high-energy CAD from a four-sector instrument and the low-energy CAD obtained on a hybrid (sector-quadrupole) instrument for the structural characterization of peptides was recently reported (111). It was determined that extensive structural information could be obtained from both instruments for peptides up to 1000 Daltons. High-energy CAD produced good data from the majority of the peptides, while low-energy CAD fragmentation was sensitive to composition of the residues (presence of basic amino acids) and mass (the best data being produced for smaller peptides). Above mass 1000, sequence data obtained by using low-energy CAD was of poorer quality than that from the high-energy CAD spectra. Also less material was usually required for analysis by the four sector . Differences of fragmentation paths were observed under the two CA regimes (111). Internal fragments, which help determine the amino acids present but give no information about the order of the amino acids, were predominant in the low-energy CAD spectra. Cleavages involving the side chain of a residue, essential for distinguishing the isobaric residues Leucine and lsoleucine, require high internal energies, thus, these ions are not detectable with low-energy CA and the ability to distinguish between the two residues is not possible. The high-energy CAD spectra were found to be highly reproducible, with qualitatively similar spectra obtained over a wide range of operating conditions. In contrast, it was necessary to carefully control collision gas pressures and 24 collision energies in order to obtain good reproducible data for low-energy CAD (111). The use of CA to induce dissociation is now routine for investigating molecules with masses up to 2000. The CA fragmentation of biomolecules is relatively simple and serves to break the protonated peptide into sequence- determining ions. However, there is a limitation on the amount of laboratory translational energy that can be converted into internal energy to induce decomposition, thus creating a limit on the degree of fragmentation. Given the shortcomings of CAD, the development of other methods of ion activation is currently of high interest. Three approaches that may be promising are photoactivation (112-113), electron impact activation (84), and collisions with surfaces (80). These methods may overcome the limitation of small energy transfer to projectile ions in collisional activation with small atomic gases. However, difficulty, expense of implementation, and lack of generality prevent these techniques from being routine methods. IV. Instrumentation A Introduction The majority of the work that will be presented was performed on a JEOL HX110 double focusing mass spectrometer with forward geometry (often denoted as an EB instrument because the electric sector, E, precedes the magnetic sector, B). This is shown in Figure 1.7. Double focusing mass spectrometers were developed primarily as instruments to make very accurate mass measurements (114). 25 Figure 1.7: Double Focusing Mass Spectrometer 26 All mass spectrometers consist of three basic components: the ion source, the mass analyzer, and the detector. Ions are produced from the sample in the ion source, in this work FAB ionization was utilized. The mass analyzer separates ions according to their mass-to-charge ratio, m/z. Each ion strikes the detector and produces a signal proportional to its relative abundance. Singly charged ions of mass m subjected to an accelerating voltage V1 acquire a translation energy 6V1 =_‘I/2 m1v12 where e is the electronic charge and v1 is the velocity of the ion m1+. Ions of different masses have the same kinetic energy, eV1, however, the velocity of an ion of a given mass is proportional to the reciprocal of the square root of its mass, where the momentum increases with the square root of its mass: m1v1 = (2V1e m1)1/2 at fixed accelerating voltage. Therefore, separation of ions can be based on their momentum or their velocity, and results in a separation based on m/z. B. Electric Sector Ions leave the source with a small spread of kinetic energy due to the ionization process and other factors. An electric sector consists of two parallel cylindrical plates across which an electric field E is applied, which deflects the ions in a circular path. The deflection radius, R9, is independent of mass, but is proportional to the energy. The electric sector acts as a focusing device for diverging ion beams and as a device for dispersing ions according to their kinetic energies by as = m1v12/R9 .. zewrte = 2(ion energy)Re 27 Double focusing instruments are arranged such that these paths are combined into a single image after passing through the magnet. C. Magnetic Sector A magnetic analyzer separates ions according to their relative mass-to- charge ratios. When accelerated m1+ ions enter a magnetic field of strength B1, the ions follow a circular path of radius Rm, perpendicular to the direction of the field, where Rm is given by: Rm: m1v1/B1e or m1v1= Rm B16 The radius taken by an ion in a magnetic field is proportional to its momentum. Thus, either by altering the accelerating voltage or by the common method of adjusting the field strength of the magnet, an ion of a chosen 'm/z can be ' directed to the detector. If B is scanned at a fixed value of Rm, ions of progressively different momenta and hence different m/z, can be made to pass through a detector slit to give a mass spectrum. Alternatively, ions that follow paths of different radii at a fixed magnetic field strength may be collected using an array detector (115). Combining the above equations gives the standard expression for the separation of ions by a magnetic sector: ' m1/e=Bsz2/2V Because ions are deflected according to their momentum, ions of the same mass, but different velocities (a function of energy), will follow different paths. Without the compensating effect of the electric sector, the image width produced by the magnet would be greater, and resolution greatly reduced. As double focusing instruments were used, it became obvious that they could be scanned in special ways so that metastable ion decompositions of a precursor ion could be observed without interference of source-produced 28 precursor and product ions (1.9., MS/MS data could be collected). To obtain these daughter ions, a B/E linked scan in which both B and E were scanned together, such that the ratio of the magnetic field strength and the electric field strength were held constant (116-117). D. Linked Scans When ions m1+ are generated in the source, those with a small internal energy are detected as m1+, because they are stable and do not undergo fragmentation. However, those m1+ ions containing large internal energy are detected as m2+ (fragment ions) which are produced within the ion source: m1+ ----> m2+ + M. where Mn represents the neutral fragment from the dissociation process. Ions having an intermediate internal energy, metastable ions, undergo fragmentation between the ion source and the detector during passage, produce mz“ ions. Linked scan methods are used to detect the metastable m2" ions, which are generated in the 1st field free region of the magnetic field-type double focusing mass spectrometer. Fragmentation is induced by collisional activation to produce daughter ions (m2+, m3+, m4+,...) from a precursor or parent ion m1+z "12+ 4* Mn He m1+ ------- > m3+ + Mn' m4+ + Mn“ During fragmentation, m1+ discharges internal energy and adds kinetic energy. However, this energy is known to be approximately 19V or less, and is very 29 small compared with the initial kinetic energy of m1+ (10 KeV when accelerated by 10 KV). The velocity of m2+ is approximately the same as that of m1+, which is essential. Thus, the kinetic energy of the ions can be given by: KE = 9V =1/2 m1v12 =1/2 m2v12 +1/2 an12 where V is the accelerating voltage and v1 is the velocity of m1+ ions. 11 was shown earlier that ions with mass m and velocity v pass through the two sectors and are detected when: Electric sector eE = mv2/ Re Magnetic sector 93 = mv / Rm Thus, when the ion m1+ fragments to m2* the conditions of the electric sector to pass m1+ and m2+ are: for m1+ 9E1 = m1v12/ R9 for m2+ 9E2 a m2v12/ Re thus it can be easily followed that m2/m1 = E2/E1. Similarly, the conditions to pass m1+ and m2+ through the magnetic sector are: for m1+ 981 = m1v1 / Rm for m2+ 932 = m2v1 / Rm giving rise to the conclusion: m2 / m1 =B2 / B1 Finally, combining the two equations m2 / m1= E2 / E1 = 82/ B1 thus, 31 / E1 = 32 / E2 = Constant. With this method, the magnetic field intensity (B) and electrostatic field voltage (E) are link-scanned so that a ratio of B/E remains constant. All daughter ions that are generated from the precursor of a certain mass number are detected (116-117). The resolving power for m2+ obtainable using this scan depends on the translational energy released during fragmentation, but is between 500 to 1000 with mass assignment of the product ion accurate to the nearest integer. The resolving power with which m1+ can be selected is only 30 300 to 400. This means that product ions formed by the fragmentation of precursor ions will contain one or more 13C isotopes (118-119). Two other scanning methods are available for scanning B and E simultaneously at a fixed relation. One of these is scanning with constant BZIE ratio (120). This scanning method detects all precursor ions which generate daughter ions with the same mass number. All daughter ions that will have the same mass number m and different velocity can be detected (121). m1+ -----> m2+ + Mn m1'+ ----> I112" + Mn' m1”+ ----> my + Mn" This results in a mass spectrum with a very broad peak width. The second method involves scanning with a constant B/E * (1-E)1/2. All daughter ions which generate neutral fragments of the same mass number are detected: ‘ m1+ -----> m2+ + Mn m1'+ ----> m2'+ + Mn m1"+ -----> m2'+ + Mn The spectrum of the daughter ions obtained here fulfills the double focusing conditions, therefore a spectrum having normal resolution is collected. In addition, all linked scans contain low intensity artifact peaks which arise from the collection of fragment ions formed during acceleration or during passage through the electric sector. The m2+ ions formed with the release of translation energy may give rise to such peaks. The linked scan certainly provides MS/MS information but, while linked B/E scans give reasonable product ion resolution, they have poor precursor ion selection and are sometimes subject to artifacts as described above. 31 E. Other Instrumentation The resolving powers for both m1+ and my“ are improved tremendously when the scan is used with a four-sector tandem mass spectrometer. The first two sectors are used to select the precursor ion (10 keV) of those ions containing only 120 isotopes. CAD occurs in a cell between the two mass spectrometers and the electric and magnetic sectors of the second instrument are scanned such that B/E is constant to provide a product ion spectrum of unit mass resolution and a mass assignment accuracy of about 0.3 u. Usually, the collision cell is floated at 2 to 5 W to improve the collection efficiency of the product ions (122). An additional incentive for the linked scanning technique is related to the advantages of high-energy CAD compared to low-energy CAD discussed earlier. Another type of instrument that is commonly used is the triple quadrupole , mass spectrometer. Following ionization in a source, the ions are focused and passed into the first mass analyzer 01. After the first stage of mass analysis, the ions continue their flight path into the second quadrupole 02, which does not function as a mass filter, but a reaction (or collision) chamber and an ion transmission device. Introduction of a collision gas in this region (02) provides the mass selected reactant ion (selected from 01) with a reagent molecule for interaction. The reaction products generated in 02 may then pass to the second mass filter in the instrument which is the third quadrupole 03. Currently triple quadrupole instruments provide the capability of analyzing gaseous ions up to 4000 mass units, although resolution of the ions is poor above m/z 2000. The major advantages of the triple quadrupole MS/MS include relatively low cost, instrumental simplicity, and ease of computer control. Disadvantages include limited mass range, inability to separate precursor ions with high 32 resolution or provide accurate mass measurements of ions and restriction to low energy CAD. Some hybrid instruments are now available (MSl=sector instrument, M82=quadrupole) and provide enhanced mass range and resolving power for MS1, but in practice do not provide any distinct advantage over triple quadrupole instruments for peptide sequencing (8). V. Peptide Sequencing by Mass Spectrometry The amino acid sequence can be deduced by interpreting the fragment ion peaks formed by the (M+H)+ of a peptide. There are three different bonds which can fragment along the backbone of a peptide. A formal method of aSsigning peptide fragments was first proposed by Roepstorff and Fohlman (figure 1.8)(123). A single letter is used to indicate the type of bond that is cleaved in the peptide and which fragment retained the charge. The letter A, B, and C denote amino-terminal fragment ions (ions that contain the N-terminus) while X, Y. and 2 represent carboxyl-terminal fragment ions (ions that contain the C-terminus). The assumed structures of the fragment ions commonly observed are shown in Figure 1.9 (124). _ Hydrogens lost or gained by the observed fragment ion during the fragmentation were denoted by primes either before or after the letter respectively. The most commonly observed fragment ions are A, B, C", Y", Z and 2" for the positive spectra. Thus, in Y" for example, the double prime indicates addition of two hydrogens to the cleaved C-terminal fragment, Y. These designations were changed for simplification, the V" was first replaced by Y+2 (125), then by Biemann (124) to yn because ions with addition of two hydrogens were the only ions observed for this fragment. 33 0 R2 R3 0 4 ii-NH-ciH-EJ NH- (12% 34 NH ciH-Lli-OH K\K \\ \\K X3 Y3 23 x2 Y2 22 x1 Y1 z. Figure 1 .:8 Generic peptide displaying the commonly accepted nomenclature for the fragmentation along the backbone. Rn H-(-NH-2H-CO-)n1-NH=(’JH an Rh (1 I H-(—NH- H-CO-)n-1-NH-CH—C 56 H4- RX H- --( -NH-CH-CO-)n-1-NH2 9n '14- CHRn H—(-NH-ClH-CO-)n_1-NH-(!H dn 34 oaC-NH- RH-g- --( -NH-gH-3-)n_1-OH .'-...-i.. 1-.....1 Yn axe/Rb H" H-E-fi-NH—EH-gflM-Ol zn “\afl H+ fiH-fi-kNH-CH-gfln- 1-Ol-ll Wn Figure 1 .:9 Assumed structures for the commonly detected fragment Ions. 35 The lower case letters distinguish the fragment ions from the single letter amino acid codes used to label immonium ions or the loss of residue side chains from the MH+ ion. Preference for fragmentation will be discussed in a latter chapter. Acquisition of the mass spectrum requires much less time than that for data interpretation to determine the amino acid sequence. Resolving the sequence is similar to putting together a puzzle. "The sequence ions, like puzzle pieces, must be connected together properly to show the sequence." (8) The sequencing process is made easier if supplementary data have been collected or known such as the amino acid composition from acid hydrolysis and the type of enzyme used to digest the peptide. The masses and the structure of the amino acid residues are given in Figure 1.10, along with their corresponding immonium ions and sidechain masses that can be lost from the MH+ ion For interpretation of the mass spectrum, one should first examine the low mass range to identify possible immonium ions, which indicate the presence of specific amino acids. It is important to note that the relative intensity of a peak for an immonium ion does not establish the number of residues of a particular amino acid present in the peptide. The high-mass region should be examined for possible side chain losses, which also indicates the presence of a particular amino acid. Next determine the presence of any peaks corresponding to loss of a fragment whose mass is exactly that of any of the amino acid residues, this is a candidate for the last member (highest mass) of a series of yn.1 ions. The (n- 1) subscriptdescribes the number of amino acid residues in that particular fragment ion. There probably will be more than one ym candidate. For each of the yn.1 candidates, one searches for the next ion in the series which 36 3.. .é‘l ‘0 1 .5 I C -2. 120136 397 278297 ( 425 g I 467 5" 221 556 G) c: j 150193205 26 354 411 499 l J L J 2511; 313 l 449 482 512526 J 1 IL A. A 11.1 [AJLJ -- AL _..— 160' ' ' '260' V ' '360' - 460 . - 560 [TI/Z Flgure 1.10: FAB-CAD-B/E Mass Spectrum of Met- Enkephalin. 37 corresponds to the exact loss of another residue, this corresponds to yn.2. This process must be done on all candidate yn.1 ions. Constructing a table or tree diagram makes it easy to keep track of the candidate sequences. The best candidate(s) are those for which the data produce the longest sequence. A verification of the y-ion assignment is the identification of accompanying 2, 2+1 or 2+2 ions which would appear 17,16, 15 u lower in mass, respectively, than the y ion. It is important to note that 2 ions do not always accompany y ions. An ideal spectrum would contain an entire set of y ions; most likely there will be a couple of incomplete sequences and, in the worst instance, many short sequences that give no sequence information. After locating several possible y-ion series, one can next look for the N- terminal fragments. The b series start 18 u below the MH+ ion mass, which represents the protonated peptide minus H20. Inspect the spectrum fer a peak that corresponds to (MH+-18+residue); this peak represents the bn.1 ion. Similar to the process for the y-ion series, find residue mass losses for the bn.1 candidate(s) and proceed as far as possible. One can also follow the same procedure for identifying the a-ion series, which starts 46 u (loss of HCOOH) below the MH+ ion. Also the c-ion series beginning with on which if it existed, would be 1 u below the MH+ mass. Similar to the x ions, a presumed b ion can be substantiated by the existence of an a or c ion, 28 u below it or 17 u above it, respectively. With a certain amount of skill and good fortune, a sequence that matches the data can be obtained. All the major peaks should be assigned; if not , one needs to find possible identities for them such as: internal fragments, rearrangement ions, matrix related peaks, contaminants, water losses and so forth. If major peaks cannot be assigned to a reasonable structure, then the assumed sequence is probably incorrect, even if most of the other peaks match. 38 A different strategy starts by looking at the middle of the mass spectrum. One picks a fairly intense peak and looks for higher or lower mass peaks that are separated by one amino acid residue. For all mates found, one attempts to continue the series up to the MH+ ion. This approach seems to be better for peptides that produce a number of fragment ions in the middle of the peptide, but few at either end. After one has located and labeled as many fragment ions as possible, the sequence or partial sequence can be determined. There are two different means one could determine the complete amino acid sequence of a peptide. The first is to obtain a complete series of a particular fragment ion. The complete series can be either an N-terminal fragment series, such as the bn ion, or C-terminal ions such as the y... ‘The complete series will give the consecutive losses of amino acid residues and hence the sequence. The second means of determining the complete sequence is to have overlapping N- terminal and C-terminal series. The partial C-terminal series (e.g., yn ions) identifies the first section of amino acids, while the N-terminal series determines the last section of amino acids in the peptide sequence. The spectrum in Figure 1.10 represents a simple ideal example; it displays peaks which represent two overlapping series of ions which can be used to determine the amino acid sequence. Because m/z 574 represents MH+, the fragment peak at m/z 425 indicates the loss of 149u; this corresponds to the loss of a methionine (M) residue mass plus H20. Thus the peak at 425 represents the highest mass b ion, (b4), indicating that methionine is the C- terminal amino acid of the peptide. Subsequent losses of 147u and 57u from 475 indicate that phenylalanine (F) and glycine (G) are in sequence form the C- terminus. Confirming the assignments of the peaks at m/z 425, 278, and 221 represent bn ions are peaks at m/z 392, 250, and 193 which represent the 39 corresponding an ions. This gives confidence to the sequence -G-F-M at the C- terminus of the peptide. The peak at m/z 411 represents a y ion because it corresponds to the loss of a tyrosine residue (Y) directly from the MH+; thus, tyrosine is the N-terminal amino acid. Subsequent losses of 57u, 57u, and 147u from m/z 411 indicate that the sequence from the N-terminus is Y-G-G-F-. Together the overlapping fragment series from both the N- and C-terminus suggests the complete sequence is Y-G-G-F-M. Without the side-chain cleavage fragments dn or wn ions present, it is impossible to distinguish leucine (L) from isoleucine (l) (125). The mass of this peptide when protonated, corresponds to 574u and agrees with the parent ions mass. Therefore the FAB- CAD-B/E spectrum in Figure 1.10 provides'the ideal situation in which different series of fragment ions confirm the amino acid sequence. Manual interpretation of spectra is time-consuming and may not produce all possible sequences that fit the data. Computer interpretation programs are designed to minimize both of these problems (28,126-127). 40 VI. References 10. 11. 12. 13. 14. 15 16. K. Biemann, F. Gapp, J. Seibi, J. Am. Chem. Soc, 81, 2274, (1959). K. Biemann, In Biochemical Applications of Mass Spectrometry, Waller G.R.l Ed.; John Wiley & Sons, New York, 1972, pp. 405. K. Biemann, Anal Chem, 58, 1288A, (1986). SA Carr, W.C. Herligy; K. Biemann, Biomed. Mass Spectrom, 8, 51, (1981). E. Vilkas, E. Lederer, Tetrahedron Lett. 3089, (1968). P. Edman, Acta Chem. Scand. 4, 238, (1950). M. Barber, R.S. Bordoli, R.D. Sedgwick, and AN. Tyler, J. Chem Soc. Comm, 7, 325, (1951). ' J.T. Stultz, Methods of Biochemical Analysis, 801. 34: Biomedical Applications of Mass Spectrometry, Wiley-lnterscience, New York 1990, p. 145. B.W. Gibson, K. Biemann, Proc. Natl. Acad. Sci. USA., 81, 1956, (1984). R. Freedman, B. Gibson, D. Donovan, K. Biemann, S. Eisenbeis, J. Parker, and P. Schimmel, J. Biol. Chem, 260, 10063, (1985). K. Fukuhara, T. Fuji, K. Toi, T. Tako, and Y. Shimonishi, J. Biol. Chem, 260, 10487, (1985). K. Biemann, S. Martin, Mass Spectrom. Rev., 6, 1, (1987). S. Seki, H. Kambara, H. Naokl, Org. Mass Spectrom, 20, 18, (1985). J.L. Witten, M.H. Schaffer, M. O'Shea, J.C. Cook, M.E. Hemling, K.L. Rinehart, Biochem. Biophysics Res. Comm, 124, 350, (1984). P. Pucci, R. Borgia, G. Marino, J. Chromatography, 331, 425, (1985). P. Roepstorff, P. Hojrup, J. Moller, Biomed. Mass Spectrom 12, 181, 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 41 (1985) SA. Martin, K. Biemann, Int. J. Mass Spectrom. Ian Processes, 78, 213, (1987) HA Scoble, J.C. Biller, K. Biemann, Fresenius 2. Anal. Chem. .327, 239, (1987). HA Scoble, K. Biemann, Science, 237, 992, (1987). DP. Hunt, J.R. Yates Ill, J. Shabanowitz, S. Winstom, C.R. Hauer, Proc. Natl. Acad. Sci. USA, 83, 6233, (1986). HA Scoble, J.A. Biller, K. Biemann, Fresenius 2. Anal. Chem, 327, 239, (1987). i K. lshikawa, Y. Niwa, Blamed. Environ. Mass Spectrom,, 13, 373, (1986). l.J. Amster, M.A. Baldwin, M.T. Cheng, C.J. Proctor, F.W. McLafferty, J. Am. Chem. Soc, 105, 1654, (1983). G.L. Glish, P.J. Todd, K.L. Busch, R.G. Cooks, Int. J. Mass Spectam. Ian Processes, 56, 177, (1984). ’ l.J. Amster, F.W. McLafferty, Anal. Chem, 57, 1208, (1985). PA Lyon, P.W. Crow, K.B. Tamer, M.L. Gross, Anal. Chem, 56, 2278, (1984). E.E. Kinston, J.H. Beyon, R.P. Newton, J.G. Liehr, Blamed Mass Spectrom, 12, 525, (1985). W. Aberth, K.M. Straub, A.L. Burlingame, Anal. Chem, 54, 2029, (1982). AL. Burlingame, T.A. Baillie, P.J. Derrick, Anal. Chem. 58, 165R, (1986). R.G. Stall, D.J. Harlan, J.R. Hass, Int J. Mass Spectrom. Ion Processes, 61, 71, (1984). L.L. Busch, B>H. Hsu, Y.X. Xie, R.G. Cooks, Anal. Chem. 55, 1157, (19831 M. Barber, R.S. Bordoli, R.D. Sedgwick, A.N. Tyler, Nature, 293, 270, 33. 35. 36. 37. 38. 39. 4o. 41. 42. 43. 45. 46. 47. 48. 49. 50. 42 (1981) J.L. Gower, Blamed. Mass Spectam.,12, 191, (1985). E. DePauw, Mass Spectrom. Rev, 5, 191, (1986). W.D. Lehmann, M. Kessler, W.A. Konig, Blamed. Mass Spectrom,, 11, 2173(19841 I AM. Buko, LR. Phillips, B.A. Fraser, Biomed. Mass Spectrom, 10, 324, (1983) RF. Lattimer, Int J. Mass Spectrom. Ian Processes, 55, 221, (1983). J. Melli, J. Seibi, Org Mass Spectrom, 19, 581, (1984). D.H. Williams, C. Bradley, G. Bojesen, S. Santikam, L.C.E. Taylor, J. Am. Chem. Soc, 103, 5700, (1981). C. Fenselau, R.J. Cotter, Chem. Rev., 87, 501, (1987). SJ. Pachuta, R.G. Cooks, Chem Rev. 87, 647, (1987). 8.0. Musselman, J.T. Watson, GK. Chang, Org. Mass Spectrom, 21, 215,(1986) ' T. Keough, Anal Chem, 57, 2027, (1985). SA Martin, C.E. Costello, K. Biemann, Anal. Chem, 54, 2362, (1982). D. Renner, G. Spitteller, Blamed. Environ. Mass Spectrom, 13, 401, (1986). A. Malarini, G> Marina, A. Milone, Blamed. Environ. Mass Spectrom, 13, 477,(1986) G.W. Huang, L Wu, H.T. Tang, Int J. Mass Spectam Ian Processes, 70, 145, (I 986). J.A. Sunner, R. Kulatunga, P. Kebarle, Anal. Chem, 58, 1312, (1986). J.A. Sunner, Fl. Kulatunga, P. Kebarle, Anal. Chem, 58, 2009, (1986). R.G. Cooks, K.L. Busch, Int J. Mass Spectam. lon Processes, 53, 111,(1983) 51. 52. 53. 55. _56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 43 E. Schroder, H. Munster, H. Budzikiewicz, Org. Mass Spectrom, 21, 707, (19861 J.A. Sunner, A. Morales, P. Kebarle, Anal. Chem, 59, 2009, (1987). R.A.W. Johnstone, l.A.S. Lewis, M.E. Rose, Tetrahedran, 9, 1597, (1983). G. Puzc, J.C. Prome, Org. Mass Spectrom, 19, 448, (1984). SS. Wong, F.W. Rallgen, Nucl. Instrum. Metrhads Phys. Res. Sect. 8,14, 436,(19861 M.R. Clench, G.V. Garner, D.B. Gordon, M. Barber, Blamed Mass Spectam., 12, 355, (1985). W.V. Ligon, S. Darn, Int J. Mass Spectam Ian Processes, 57, 75, (1984). RM. Capriali, W.T. Moore, Int. J. Mass Spectram Ian Processes, 86, 187 (19881 T.E. Creighton, Proteins Structures and Molecular Properties, W.H. Freeman and Co., New York, 1984. S. Lewin, Displacement of Water and It's Control of Biochemical Reactions, Academic Press, London, 1974. S. Naylor, W.J. Skelton, D.H. Vlfilliams, J. Chem. Soc. Chem. Cammun., 1 619, (1 986). 8.8. Wong, F.W. Rallgen, Int. J. Mass Spectram. Ian Processes, 70, 135, (19861 M.Y. Zhang, X.Y. Liang, Y.Y. Chen, X.G. Liang, Anal Chem, 58, 2288, (19841 H.B. Bull, K. Breese, Archives at Biochem. and Biaphys., 161, 665, (19741 -A.M. FaIICK, D.A. MaItby, Anal. Biochem., 82, 165, (1989). D. Renner, G. Spiteller, Agnew. Chem. Int. Ed. Eng/., 24, 408, (1985). W.V. Ligon, 83. Born, Int. J. Mass Spectram Ian Processes, 61, 113, 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 44 (19841 G.M. Allemaier, Rapid Cammun. Mass Spectrom, 2, 74, (1988). GM. Allmaier, Anal. Chim. Acta, 223, 349, (1989). S. Naylor, G. Moneti, S. Guyan, Blamed. Environ. Mass Spectrom, 17, 393, (1988). S. Naylor, A.F. Findeis, B.W. Gibson, D.H. WIlliams, J. Am. Chem. Soc, 108, 6369, (1986). CE. Heine, Dissertation, Michigan State University, 1990. W.G. Chai, L. Yan, G.H. Wang, X.Y. Liang, Y.F. Zhao, G.J. Ji, Blamed. Environ. Mass Spectram... 14, 331, (1985). RM. Caprioli, T.Fan, J.S. Cottrell, Anal. Chem, 58, 2949, (1986). RM. Caprioli, T.Fan, Biochem. Biaphys. Res. Cammun., 141, 1058, (19861 R.M. Caprioli, W.T. Moore, T. Fan, Rapid Commun. Mass Spectrom, 1, 15, (1987). W. Kausler, K. Schneider, G. Spiteller, Blamed. Environ. Mass Spectrom, 17, 15, (1988). W.V. Ligon, Anal. Chem, 58, 487, (1986). US. Wagner, A. Salari, D.A. Gage, J. Leykam, J. Fetter, R. Hollingsworth, J.T. Watson, Biol. Mass Spectrom, 20, 419, (1991 ). M.J. Dekrey, H.T. Kethamaa, V.H. Wysacki, R.G. Cooks, Org. Mass Spectrom, 21, 193, (1986). A. Mabud, M.J. Dekrey, R.G. Cooks, Int. J. Mass Spectram. Ian Processes, 67, 285, (1985). M.J. Dekrey, A. Mabud, R.G. Cooks, J.E.P. Syka, Int. J. Mass Spectram. I an Processes, 67, 295, (1985). 83. 84. 85. 86. 87. 88. 89. 90. 91 . 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 45 W.D. Bowers, S.J. Delbert, R.L hunter, R.T. Mclver, J. Am. Chem. Soc, 106.7288, (1984). RB. Cody, B.S. Freiser, Anal. Chem, 51, 547, (1979). F.W. McLafferty, (ed) "Tandem Mass Stectrometry", Wiley, New York, 1 983. R.G. Cooks (ed.) ”Collision Spectroscopy”, Phenum, New York, 1978. K. Levson, H. Schwarz, Mass Spectram. Rev., 2, 77, (1983). RA. Yost, C.G. Enke, Anal Chem, 51, 1251A, (1979). K.R. Jennings, Int. J. Mass Spectram. lan Processes, 1, 227, (1968). WP. Haddon, F.W. McLafferty, J. Am. Chem. Soc, 90, 4745, (1968). R. Weber, K. Levsen, Blamed. Mass Spectrom, 7, 314, (1980). T. Matsuo, H. Matsuda, l. Datakuse, Y. Shimonish, Y. Maruyama, T. Higuchi, l. Kubota, Anal. Chem, 53, 416, (1981). RM. Haynes, M.L. Gross, Methods of Enzymology: Mass Spectrometry, vol. 193, Academic Press, San Diego,1990, p. 237. H.M. Rosenstock, M.B. Wallenstein, A.L. Wahrhaftig, H. Eyring, Proc. Natl. Acad. Sci. USA, 38, 667, (1950). WA Chupka, J. Chem. Phys., 30, 191, (1959). B. Andlauer, C. Ohinger, J. Chem. Phys, 55, 1471, (1971). F.W. McLafferty, T. Wachs, C. Lifshitz G. lnnorta, P. lving, J. Am. Chem. sea, 92, 6867, (19701 S. Singh, F.M. Garris, R.K. Bowd, J.H. Beyon, Int. J. Mass Spectram. Ian Processes, 66, 131, (1985). W. Lichten, Phys. Rev., 131, 229, (1963). MS. Kim, F.W. Mclafferty, J. Am. Chem. Soc, 100, 3279, (1978). V.H. Wysacki, H.l. Kentlamaa, R.G. Cooks, Int. J. Mass Spectram. Ian Processes, 75, 181 , (1987). 46 102. Ht. Kentlamaa, R.G. Cooks, J. Am. Chem. Soc, 107, 1881, (1985). 103. PH. Hemberger, J.A. Laramee, A.R. Hubik, R.G. Cooks, J. Phys. Chem, 85, 2335, (1981). 104. J.A. Larmee, D. Carperon, R.G. Cooks, J. Am. Chem. Soc, 103, 12, (1981). 105. K. Busch, G.L Glish, S.A. Mcluckey, “Mass Spectrometry/Mass Spectrometry", VCH Publ., New York, 1988. 106. DJ. Douglas, J. Phys. Chem, 86, 185,(19821 107. A.J. Alexander, R.K. Boyd, Int. J. Mass Spectram. Ian Processes, 90, 211, (1989). 108. K.L Schey, K.l. Kentlamma, V.H. Wysacki, R.G. Coods, Int. J. Mass Spectram. Ian Processes, 90, 71, (1989). 109. A.J. Alexander, P. Thibault, Rapid Cammun. Mass Spectrom, 2, 224, (1988). 110. S.J. Gaskell, M.H. Reilly, S.A. Carr, M. Bean, R. Vickers, Proceedings of the 36th American Society for Mass Spectrometry Conference, San Francisco, 1988, p. 111. 111. M. Bean, S.A. Carr, G.C. Thome, M.H. Reilly, S.J. Gaskell, Anal. Chem. 63, 1473, (1991). 112. ML. Vestal, J. H. Futrell, Chem. Phys. Lett, 28, 559, (1974). 113. D.H. Russel, Mass Spec. Rev., 9, 405, (1990). 114. Biemann, K., Methods in Enzymology: Mass Spectrometry, vol. 193, Academic Press, San Diego, 1990, p. 295. 115. S. Evans, Methods in Enzymology: Mass Spectrometry, vol. 193, Academic Press, San Diego, 1990. p. 61. ' 116. 0.8. Millington, J.A. Smith, Org. Mass Spectrom, 12,264, (1977). 117. C.J. Porter, A.G. Brenton, J.H. Beynan, Int. J. Mass Spectram. Ian Phys, 47 86, 69, (1980). 118. RS. Stradling, K.R. Jennings, S. Evans, Org. Mass Spectrom, 13, 429, (1978). _ 119. K. Biemann, Methods in Enzymology: Mass Spectrometry, vol. 193, Academic Press, San Diego, 1990, p. 455. 120. E.A.I.M. Evers, A.J. Noest, O.S. Ackerman, Org. Mass'Spectram" 12,419, (19771 121. W.F. Hadden, Org. Mass Spectrom, 15, 539, (1980). 122. ML. Gross, Methods in Enzymology: Mass Spectrometry, vol. 193, Academic Press, San Diego, 1990, p. 131. 123. P. Roepstorff, J. Fohlman, Blamed. Mass Spectrom, 11, 601, (1984). 124. K. Biemann, Blamed. Environ. Mass Spectrom, 1-6, 99, (1988). 125. S. Johnson, S.A. Martin, K. Biemann, J. Stults, J.T. Watson, Anal. Chem, 59, 2621, (19871 126. T. Sakurai, T. Matsuo, H. Matsuda, l. Katakuse, Blamed. Mass Spectrom, 11,396, (19841 ‘ 127. MM. Siegel, N. Bauman, Blamed. Environ. Spectram, 15, 333 (1988). Chapter 2 Derivatization to Enhance the lonlzatlon Efficiency of FAB and to Control the Fragmentation of Peptides I. INTRODUCTION Fast atom bombardment mass spectrometry, FAB-MS, and fast atom bombardment-collisionally activated dissociation-tandem mass spectrometry, FAB-CAD-MS/MS, are powerful techniques for the determination of molecular weight and the primary structure of peptides, respectively (1-12). However, there are factors that frequently prevent successful sequence analysis of peptides by mass spectrometry. Two such factors are the poor desorption/ionization efficiency of some hydrophilic peptides and ambiguities in interpretation of the spectra when key sequence ions are weak or absent. The detection limits for a given peptide have been correlated with its surface activity (10-12), which, in turn, is related to its relative hydrophobicity. The unfavorable interactions between hydrophobic peptides and matrix molecules causes the hydrophobic peptides to concentrate at the surface of the matrix. Hydrophilic peptides interact favorably with matrix molecules, and thus are dispersed throughout the sample volume and do not concentrate at the surface layer of the matrix. Therefore, the FAB signal for hydrophilic peptides is usually weak or cannot be detected (1 0-14).. Various methodshave been employed to increase the FAB signal response of hydrophilic peptides. One successful approach used to increase the FAB signal response has involved formation of hydrophobic derivatives of hydrophilic peptides (15-20). These derivatization methods include a variety of peptide modifications, ranging from esterification to the attachment of a single hydrophobic moiety to a specific site on the peptide. 48 49 Although there have been a number of successful reports of using FAB- MS and FAB-CAD-MS/MS for the analysis of peptides, most peptides cannot be sequenced using these techniques. Fragment ion peaks that are observed in FAB-CAD-MS/MS spectra can provide significant structural information, but a complete sequence cannot always be obtained because every peptide will not necessarily yield a complete or overlapping series of daughter ions from cleavages between each residue in the chain. Thus, ambiguities in the primary structure of the peptide can remain even when the protonated molecule (MH+), amino acid composition (immonium ions), and most (but not a complete series) of the sequence ions are obtained from the FAB and FAB-CAD-MS/MS spectra. Localization of the positive charge on a peptide influences the fragmentation observed in the FAB-CAD-MS/MS spectrum (16,21-27). In a FAB experiment, if one amino acid residue in a peptide is considerably more basic ' than any other, then the positive charge due to protonation of the peptide during FAB ionization is most likely located on the basic functionality of that residue (17,28). Basic amino acids located at or near the N-terminus promote the formation of N-terminal ions (an and d", and to some extent bn, and on), while basic amino acids located at or near the C-terminus yield predominantly C-terminal ions (yn, vn, Wu and to a lesser extent xn, Zn) (1,17,24,28). Thus, peptides isolated from tryptic digests often produce abundant yn series ions because the C-terminal amino acid is either lysine or arginine (1). Because CAD fragmentation of these peptides can occur at a variety of peptide bonds, the cleavages are thought to occur remote from the charged site (21,29). The protonated residue is assumed not to be involved directly in the fragmentation process. When these charge-remote fragmentations occur, the detected fragments contain the charged residue. The absence of a basic residue in a 50 peptide will yield relatively uncomplicated spectra with few fragment ion series, principally bn and yn (1,21). To take advantage of the localized charge effect, derivatization has been used (1621-27) to fix a positive charge or attach a very basic group at specific sites on a peptide in order to gain some control over the fragmentation process during FAB-MS/MS. Collisionally activated dissociation (CAD) of these derivatized peptides generates a set of fragment ions that contain the charged or basic residue incorporated during peptide modification. The resulting spectrum is comprised of a set of daughter ions different from those of the underivatized, protonated peptide, which can be useful in resolving ambiguities in the sequencing process when key sequence ions are otherwise weak or absent. The localized charge also has been shown to simplify the mass spectrum by generating only those ions that contain the modified residue, thus facilitating interpretation. Protonated peptides that do not contain one very basic site will exist in more than one isomeric form. The protonated peptide is a collection of different species, differing in the location of protonation; Thus, the location of the positive charge varies. Because fewer fragment ion series are preseitt in the spectra of derivatized peptides, the ion current is distributed among fewer ions, thereby increasing the abundance of those fragment ions formed (4,30). Whereas chemical modification of peptides has been used by others either to improve the detection of hydrophilic peptides or control the fragmentation of the peptide under CAD conditions, our objective was to combine both of these aspects. Novel and simple procedures for preparing ethyl-triphenylphosphonium derivatives of peptides are described. These procedures allow an ethyl-triphenylphosphonium moiety to be selectively attached to either the N- or C-terminus thus, localizing the positive charge at 51 that site. The resulting derivatives contain a positive charge at a fixed position and have significant hydrophobic character. Modification of peptides by these chemical methods significantly enhances the efficiency of FAB ionization, especially of hydrophilic peptides. Moreover, upon collisionally activated dissociation, the derivatized peptides generate a predictable series of sequence ions from either the C-terminus or the N-terminus, depending on the location of the ethyl-triphenylphosphonium moiety. II. EXPERIIVIENTAL SECTION Two derivatizing reagents that are similar in structure (Figure 2.1) were synthesized. Both reagents contain three phenyl rings, which are responsible for the large increase in hydrophobic character, and thus, greater surface activity. The three phenyl groups are attached to a phosphorous atom that carries a positive charge, which directs the fragmentation obtained during a CAD experiment. The phosphorous atom is connected to an ethyl linkage that contains the functional group of the derivative. The ¢3-CH2CH2P+Br is used to attach to the N-terminus of a peptide, while the reagent containing a primary amine structure is utilized to attach to the C-terminus of a given peptide. The bromoethyltriphenylphosphonium (bromoethyl-TPP) reagent was prepared by dissolving 5g of triphenylphosphine in a 50-fold molar excess of 1,2-dibromoethane (31) (Figure 2.2a). The mixture was heated to 80°C and was stirred for a 24-hour period. The product (bromoethyl-TPP) formed a precipitate in the reaction mixture and was collected on a sintered glass filter and was washed with ether. The reaction produces a low yield; 1-20% of the 2- 52 BI" X—CHg—CHz—P X = Br attaches to amino terminus X = NH2 attaches to carboxyl terminus Figure 2.1: Structures of the two reagents developed to derivatize peptides. S3 8- 8+ Br-CHz-CHz-Br Q 00 x . €61.32. Flgure 28: Synthesis of bromoethyl- triphenylphosphonium reagent. i+ so. OP-CHg-CHg-Br + NH3 6 -HBr ' i-I- GP-CHg-CHz-NHQ Figure 2b: Synthesis of aminoethyl- triphenylphosphonium reagent. S4 diphenylphosphino-ethyl triphenylphosphonium salt and vinyl- triphenylphosphonium salt are formed as side-products. Without further purification, the bromoethyl-TPP-bromide product was dried under nitrogen and stored at 0°C. The aminoethyltriphenylphosphonium (aminoethyl-TPP) reagent was synthesized from the bromoethyl-TPP reagent. The bromoethyl-TPP reagent (29) was dissolved in acetonitrile. A 30-fold molar excess of ammonium hydroxide was added dropwise to the sOlution. The mixture was stirred for 3 hours at 50°C and the reaction followed as shown in Figure 2.2b. The aminoethyI-TPP was separated from the excess ammonium hydroxide and ethanol by evaporation in vacuo, leaving the solid product, which was stored under nitrogen at 0°C. III. Peptide Derivatization A. Derivatization of the N-terminus Derivatization of the N-terminal primary amine of a given peptide with the ethyl-TPP moiety is a one-step reaction, as illustrated by Figure 2.3. In order for the reaction to occur, the N-terminus of the peptide must be a primary amine, therefore this reaction is not appropriate for N-terminally blocked peptides. The two basic amino acids, arginine and lysine, could react with the bromo-ethyl-TPP, because they both contain primary amines on their side chains. The reaction for derivatizing the N-terminal primary amine was made selective for the N-terminus by controlling the pH of the reaction mixture. Only 55 O o O—ptCHz-Cflz-Br + H2N-?H-g-NH-?H-g-NH-9H-fi-OH "1 "2 Fis HBr o 0 II ll 0‘ P- C Hz-C H2-NH-qH-C-NHTqH-C-NH-QH-fi-OH n R R Figure 2.3: rN-Terminal Ethyl-TPP derivatization of a peptide. 56 unprotonated primary amines (i.e., those with a free lone pair of electrons) will react with bromoethyI-TPP. Therefore, the reaction is carried out at pH 9, at which the pH the side chain amino groups of lysine and arginine are 98.5% and 99.9% protonated, respectively (pKa's = 10.8 and 12.5, respectively), and cannot react with the bromo-ethyI-TPP, whereas the N-terminal amino group is predominantly in the unionized form. Maintaining basic conditions during the derivatization process requires a buffered solution because HBr is produced during the reaction. The chemical reaction was carried out by dissolving a peptide (0.1-2 nmol) in a potassium sodium borate buffer solution (1-1.5 ml) at pH 9. The bromoethyl-TPP reagent was added to the buffered peptide solution in a 10-fold molar excess. The sample was vortexed and allowed to react for approximately three hours in a 37°C water bath or eight hours at room temperature. The peptide derivative was isolated from the buffer salts by applying the reaction mixture to a 018 Sep- pak (Waters), washing with water, and then eluting with an acetonitrile/water mixture (1:1 v/v). The need to separate the derivatized peptide from the buffer solution reduced the overall efficiency of the procedure, because some of the peptide is lost during sample handling. Therefore, a volatile buffer system comprised of an acetonitrile/pyridine (1 :1,v/v) mixture was employed. The pH of this mixture is between -8-9, thus the N-terminal primary amine is mainly in the neutral form, while arginine and lysine are protonated. It has been shown by others that vinyl-TPP also reacts with the N- terminus, generating the same product. In order to determine which TPP reagent was most effective in derivatizing the N-terminus of peptides, the bromoethyl-TPP and vinyl ethyl-TPP were separated and collected using HPLC, then allowed to react separately under identical conditions with several 57 standard peptides. The results showed that both compounds attached to the primary amine with equal efficiency. Since the vinyl-ethyl-TPP is commercially available (Lanchaster), the vinyl ethyl-TPP was then used to derivatize the N- terminus using the acetonitrile/pyridine reaction mixture. B. Derivatization of the C-terminus Derivatization of the C-terminus is accomplished in a two-step reaction (Figure 2.4) utilizing 1-Ethyl-3-(3-dimethy aminopropyl) carbodiimide (EDC, Pierce Chem. 00., Rockford, IL) as a coupling reagent. A similar reaction is used in solid phase peptide synthesis (32) where an amino acid attaches to the C-terminus of the synthetic peptide. Here the amino-ethyl-TPP reagent is used to attach to the C-terminus of a peptide instead of an amino acid. The reaction was carried out by dissolving the peptide (0.1-2 nmol) in an aqueous TFA solution at pH 5. Added to the solution in a 20-fold excess are the E00 and the aminoethyl-TPP reagents. The mixture was allowed to react for approximately three hours in a water bath at 37°C or eight hours at room temperature. The solvent was then removed by lyophilization prior to mass spectral analysis. Modification of the C-terminus is accomplished in an acidic solution in order to protonate one of the nitrogens in EDC. Although the reaction might be expected to occur with the carboxylic group on side chains of aspartic acid and glutamic acid, we have found that the derivative only attaches to the carboxyl terminus of the peptide. The reagent may not react with the sidechain of these amino acids because of steric hindrance; the side chains may not be accessible for derivatization. Because the reaction contains a 20-molar excess of amino- ethyI-TPP, reaction with the amino terminus of the peptide to form a cyclic product or the coupling of two peptides is not observed. 58 0 O " II quchNHchuHchOH-I-O—n "_O F'1 R2 0 o "2N-?HmCNH?H--c-NH?HC.c\’o- "1 "a O .. Q l +u2u—cu2-cquQ o 0 II II HZN-cH-c-NH-cu- Ic HNH-CFH --c -HH-CH2-CH2-P—Q Figure 2. 4. C-Terminal Ethyl-TPP derivatization of a peptide 59 C. Reaction Efficiency Reaction efficiencies for three peptides (59 amino acids in length), based on analyses by HPLC separation of the remaining unreacted peptide from the reaction mixtures indicated that the derivatization procedure for either terminus had an approximate yield of 85%. However, as the size of the peptide increases, the reaction yield decreases. Derivatizing the N- or C-terminus of larger peptides (more than 12 amino acids) usually required longer reaction times (6-8 hours) and higher reaction temperatures (up to 75°C). This change in the reaction conditions produced some notable side products, which interfere with analysis by FAB-MS. It should also be mentioned that some modified peptides (phosphopeptides and sulfated peptides) did give good reaction yields using the described methods. Therefore, a need to improve the chemistry must be considered in order to derivatize these modified peptides. The present chemistry restricts the peptide size to an upper limit between 20 and 25 amino acid residues, above which peptides are not significantly derivatized. This size limit is probably due to larger peptides forming secondary structures, where the termini are either buried or sterically hindered. In order to overcome this problem, the peptide could be to denatured, making the terminal groups accessible. This can be accomplished by heating the peptide to higher temperatures, however, this has been noted to produce side reactions which interfere in mass spectral analysis. Another possibility is to use a chemical denaturant such as quanidinium chloride, but changing the chemical conditions of the reactions may affect the reaction rates of derivatization and the denaturant would interfere in mass spectral analyses by FAB-MS. Thus, the chemical method would also require the separation of the derivatized peptide and the chemical denaturant prior to FAB-MS. It is not possible to increase the 60 pH for denaturing the peptide, because this would definitely change the reaction conditions. ‘ It should also be mentioned that the lowest possible level at which a peptide can be derivatized by these methods has not been established. The lowest level derivatized is 20 pmol of bradykinin. The major problem with using lower levels may not be the reaction, but rather the loss of the peptide during sample handling. Derivatization procedures have been attempted on lower levels of peptides, however the sample was lost. IV. SIGNAL ENHANCENIENT Significant enhancement of the FAB signal for both N- and C-terminal ethyl-TPP-derivatized peptides has been observed. The more hydrophilic peptides demonstrated the greatest signal enhancement, while the more hydrophobic peptides showed a smaller increase (as little as a factor of two) in the FAB signal response. The improvement of the FAB signal due to the presence of the ethyl-TPP moiety is readily apparent from comparison of the FAB mass spectra of an underivatized peptide and those of its corresponding N- and C- derivatized species. Figures 2.5a-2.5c exhibit FAB mass spectra for 50 pmol underivatized, 5 pmol N-terminally derivatized, and 5 pmol C-terminally derivatized V-Q-A-A-D- Y-l-N-G. The FAB mass spectrum of 50 pmol of the underivatized peptide (Figure 2.5a MH+=950) indicates that, at this level, the peak for the protonated molecule is too weak to be detected. The FAB mass spectra from 5 pmol of the N-terminally (Figure 2.5b) and the C-terminally (Figure 2.5c) derivatized peptides demonstrate the magnitude of signal enhancement of the derivatives 6i I 'w t : V-Q-A-A-D—Y-l-N-G : t 50 pmol : 0 I I I 0 I 0 I I I 400 500 600 700 800"900, 1000 1100 1200 1300 ml: 100 I I E 1 TEP-V-Q-A-A-D-Y-LN-G 5 pmol : m,“ E ‘1‘ ‘ ' 1 . 1g; 2‘ \ 1 , °' ' 1" ‘HI 0 t I v 1‘ ‘1 '1 111' ,; °’ ‘ 111‘ “‘11 ‘3‘1 I‘l “ 1 ““ “ "H“ lll“‘HHH 1‘1‘1‘| 1‘1 ‘1‘l|‘1‘|1‘ 119511111 ‘11‘11 l‘111 1 400 500 600 700 800 900 1000 1100 1200 1300 mlz ll 0 ' + + . : V-Q-A-A-DsY-I-N-G-TPP M i 5 pmol 1237 O , . 8 "fi-OD.“." .; ““‘.“‘.'l1"‘]"11} _, ,1, .1 .1 . y , ,1.,',19,1141!)2.13;..5'1‘1‘1.‘ "“‘1“"“l‘1"1‘.‘ ‘11-‘91"‘111'13 1Ii ‘1'.) ..1-1 1 .1 m‘ 400 500 600 700 800 900 1000 1100 1200 1300 ml: Fi ure 2. 5a-c: The FAB mass spectra of a) 50 pmol underivatized V AADYING, b) 5 pmol of N-terminally derivatized peptide, and C) 5 pmol of C-terminally derivatized peptide. 1‘1 1-‘.: 11 ' 62 over that of the underivatized peptide. The N- and C-terminus derivatives both generate an abundant molecular ion (M+=m/z 1238 and m/z 1237, respectively) during FAB. The detectability of the N- and C- terminally derivatized peptide was estimated to be on the order of 100-fold greater then that for the underivatized peptide. The signal enhancement is related to an increase in the hydrophobic character of the peptide. This can be shown by the Bull and Breeze F values (13), which serve as an index for surface activity. The larger the F value (more positive) the more hydrophilic character and the lower the F value (more negative) the more hydrophobic character of the peptide, and thus the greater the surface activity. The underivatized peptide has an approximate F value of +870, while the derivatized peptide has an F value of -3690. This is a substantial difference and is attributed to the presence of the three phenyl rings on the TPP moiety. _ A second example of signal enhancement by TPP derivatization is shown in figures 2.6a-2.6c, by comparing the mass spectrum of underivatized peptide leu-enkephalin, YGGFL, with those of the N- and C-terminal derivatives. Again from the comparing the spectra, it is easily seen that the derivatized form of the peptide reduces the detection limits and allows molecular weight determination where the underivatized peptide is undetectable. Because of "suppression effects", it is often difficult to analyze mixtures of peptides by mass spectrometry. Separation of a large mixture of peptides, by their relative hydrophobic character, into smaller mixtures can be accomplished by HPLC. Separation prior to analysis by FAB-MS can. minimize the problem of hydrophilic peptides being suppressed by more hydrophobic peptides. However, in a mixture of hydrophilic peptides, one or more of the peptides can remain undetected. To demonstrate the capability of the derivatization technique to improve the analysis of mixtures by FAB, a mixture containing. 63 1 . a . GchoroIPonk . g: [M g. 3: £1 WI“? 2} n ///// a0 5115 0 9'0 585 “To 565 '5110 M 1”‘ b 3 WW %‘ [Ir-ml i‘ 3. 5: 3. 1 . 1 III/l 1W- 0 GlyceroIPocIt ‘5 [com] Mama-mum 5‘ [Hf-843] LE 4% da‘ 15* 'Ha oh us ch 650 on M Figure 2.6a-c: FAB mass spectra of a) 5 pmol of the protonated peptide YGGFL, b) 1 pmol of the N- terminal derivative, and c) 1 pmol of the C-terminal denvafive. 64 three peptides in equal molar concentrations was prepared. The mixture was comprised of A-L-G (MH+=260), Y-G-G-F-L (MH+=556), and V-G-V-A-P-G (MH+=499). Figure 2.7a is a FAB mass spectrum for 100 pmol of the three underivatized peptides. At this level, peaks corresponding to the three protonated peptides are not observed, thus molecular weight information was not obtained. The spectrum displays abundant glycerol matrix cluster ions, denoted by the asterisks. ‘ Figure 2.7b displays the mass spectrum of 25 pmol of the same mixture after employing the C-terminal derivatization reaction. After derivatization all three peptides are easily detected. This demonstrates the ability to derivatize a mixture of peptides and enhance the ionization efficiency of all components thus, lowering the detection limit of the peptides in the mixture. These examples demonstrate the capacity of the derivatives to increase the FAB ionization efficiency and lower the detection limits of a given peptide. The detectability of the derivatized peptide is usually at least an order of magnitude greater than that of the underivatized peptide. The more hydrophilic peptides exhibited a greater increase in the molecular ions' intensity than hydrophobic peptides after the derivatization. Also, derivatives of smaller peptides displayed a greater increase in detectability than those of larger peptides. 65 .2 O .. .. 12.221 \ I 23:7 339 451 Flgure 2.78 a: ’l‘ g‘zse ALG '0" 4 260 c . — . 553 2‘ .1. '1‘? VGVAPG YGGFL 75 .‘ 499 553 II 1 / / ‘ ‘ 300 400 500 m 500 100 :53 *6“ Flgure 2.7b :5" ‘ ALG 1: 547 .9 .E 737 _“2’ VGVAPG- E m U: 921 _______.___ 2 . “H ‘...““ 1111 Figure 2.75-b: FAB mass spectra of a) 100 pmol of a mixture containing ALG, VGVAPG, and YGGFL, and b) N-terminal derivatization on 25 pmol of the peptide mixture. ’ 66 V. Dinecting the Fragmentation by Derivatization An important feature of these derivatization techniques is the superior sequence information obtained from CAD experiments. As in peptides that have a basic amino acid at or near the N- or C-terminus, the Mr ions of ethyl- TPP derivatives generate a predictable set of daughter ions upon collisional activation. FAB-CAD-MS/MS spectra of ethyl-TPP peptides contain exclusively either amino- or carboxyl-terminus daughter ions, depending on the site of derivatization. The FAB-CAD-MS/MS spectra of the Mt ions of N-terminally derivatized peptides predominately generate an and dn ions, whereas bn and en ions also are observed at much lower abundances. Conversely, the C- terminal derivatives generates xn, yn, 2", V“, and wn ions having similar abundances. The fragment ions involving side chain cleavages, (dn and Wu), needed to distinguish the residues leucine and isoleucine are much more prominent in the spectra of the derivatized peptides. Thus, by attaching the ethyl-TPP moiety to either terminus, a selected series of daughter ions can be obtained by FAB-CAD-MS/MS. It is worthy of note that the m/z value for the peaks observed in the mass spectra of the TPP derivatives reported herein differ (beyond that associated with the mass shift due to derivatization) from those anticipated based on familiarity with normal peptide CAD fragments. This occurs because the TPP derivatives are not protonated, and because they fragment by remote site processes. For example, in the mass spectra of C-terminally TPP-derivatized peptides, the y ion peak, equivalent to the Y" (Y+2H) ion of the Roepstorff nomenclature (33), occurs at an m/z value three mass units lower than might be anticipated. That is to say, the equivalent y ion that we describe for the TPP derivative of a peptide has an m/z value that is one mass unit lower than the m/z 67 value for the theoretical fragment Y formed by homolytic cleavage of the bond between the carbonyl carbon and the nitrogen of the peptide linkage. Similarly, we observe a-H and b-H ions (A-H and B-H ions according to the Roepstorff nomenclature) for N-terminally derivatized TPP peptides, rather than the a and b (or A and B) ions commonly encountered in the spectra of underivatized peptides. The FAB-CAD-MS/MS spectra of derivatized peptides display very weak peaks corresponding to immonium ions. The abundant peaks observed below m/z 290 correspond to cleavages associated with the ethyl-TPP moiety (26,27). Figures 2.8a-2.8c illustrate the difference in fragmentation behavior between underivatized V-G-V-A-P-G and its corresponding N- and C-terminal ethyl-TPP derivatized peptides. The underivatized peptide (Figure 2.83) at 200 pmol, when analyzed under FAB-CAD-MS/MS conditions, does not yield a complete or overlapping series of fragment ions required to obtain the complete amino acid sequence. Although the spectrum displays peaks for both carboxyl- and amino-terminal fragment ions from the underivatized peptide, only a partial amino acid sequence can be deduced. In contrast, 100 pmol of the C-terminally derivatized peptide, yields a much more informative spectrum under FAB-CAD-MS/MS conditions. It is apparent that the spectrum (Figure 2.8b) only contains C-terminal fragment ions, because the positive charge is localized on the C-terminus of the peptide. The spectrum for the C-terminally derivatized peptide displays the complete series of X“, yn, and 2,. ions resulting from cleavages between every peptide bond with charge retention on the C-terminus (i.e., the ethyl-TPP moiety). Significant structural information about the side chains also is obtained from the presence of enhanced Wu and V" ions. This mass spectrum readily provides enough information for a complete amino acid sequence analysis. 68 '8' MH” 3’2 V-G-V-A-P-G ‘ 200 ‘pmol b4 p-v-c-v-A-P-e 1 100 pmol I d: 1'. '0‘ as _ :. 84 * 5 .',' . *‘ I TPP ; d.‘ C182 D264 b3 Q 04 c4 * b3 .1” I. 111 . . aooasoaoomsoossoeoo'osommaoo ' m Fi ure 2.8a-c: The FAB-CAD-B/E mass spectra of a) 200 pmol of V VAPG, b) 100 pmol of the N-terminal derivative and c) 100 pmol of the C-terminal derivative. 69 Similarly, the N-terrninally derivatized peptide, at 100 pmol, also yields a more informative FAB-CAD-MS/MS spectrum than that of the underivatized peptide. In this spectrum the observed ions are only N-terminal ions, because the charge is localized on the N-terminus. The spectrum (Flgure 2.8c) of the N- terminally derivatized peptide displays .a complete series of an, bn, and cn ions, and also provides the side chain information (dn ions). The data from the N-terminal derivative of the peptide again allows for a complete amino acid sequence analysis to be easily obtained. A. Peptides containing Basic Residues These N- and C- terminal derivatives direct the fragmentation of peptides containing basic amino acids (K or R) at different sites in the peptide. This can be demonstrated by comparing the underivatized and N-terminally derivatized forms of the peptide Y-Q—E-A-F-R-R-F-F-G-P-V. The FAB-CAD-B/E spectrum obtained from 2 nmol of the protonated peptide is displayed in Figure 2.9a. The spectrum clearly indicates that peaks corresponding to fragment ions below m/z 700 have low relative intensities. This is explained by the location of two basic arginines in the middle of the peptide. Because these arginine residues are more basic than any other site in the peptide, the majority of the protonated species will have a structure with the protonating hydrogen on one of these residues. Consequently, most of the fragment ions produced from CAD will contain one or both of the arginines and the N- and C-terminal fragment ion series begins in the middle of the peptide. Although the spectmm displays abundant fragment ion peaks, a complete ion series or two different overlapping ion series are not detected in this mass spectrum, therefore the complete amino acid sequence cannot be determined. 70 The FAB-CAD-B/E mass spectrum of 250 pmol of the N-terminal derivative (Figure 2.9b) gives rise to different results. The spectmm displays a complete series of an ions and some confirming cn and dn ions thus, the complete amino acid sequence of this peptide is readily obtained. Also in contrast to the CAD mass spectrum of the protonated peptide where the fragment ions that contained the arginine residues were much more abundant than the other fragment ions, the N-terminal derivatized peptide generates an ion fragment ions that have similar relative abundances. This shows that the basic site apparently does not have an effect on the fragmentation process. A second example of the capacity of the derivative to overcome the effects of basic amino acids is shown by comparing the FAB-CAD-B/E mass spectra of the protonated and C-terminal derivative of the peptide R-V-Y-H-P—F. The CAD mass spectrum of 500 pmol of the protonated peptide is displayed in Figure 2.10a. This spectrum shows that the majority of the observed ions are N- terminal fragments, with the an series being the most abundant. This is because the most basic site on the peptide is the side chain of the arginine residue, which is at the N-terminus of the peptide. Thus, protonation during FAB ionization predominately occurs at the N-terminus and subsequently leads to the formation of N-terrninal ions. The spectrum displays a complete series of an ions, thus the amino acid sequence is determined. The C-terminal derivative of 100 pmol of the same peptide, Figure 2.10b, leads to a much different product ion spectrum. All fragment ions are C-terminal ions, whereas the spectrum of the protonated peptide displayed few, low abundant C-terminal ion peaks. A complete series of x... y... and 2... fragment ions are present in the FAB-MS spectrum of the TPP derivative, thus the amino acid sequence is readily determined. 7i 100 I a g. ; YQEAFRRFFGPV g"? 2nmol at: to. 5: L5: " .3: ‘2 E: 6:). '11 J '1“ '10 0 200 400 600 v VNSOV v '1m20' ' V1200- ' '14‘00' V ml: 100. ‘ b A? 2‘ ethyl-TPP-YQEAFRRFFGPV 2. 250 pmol 5. g as a, 33. @1 a1 05 DZ. . . d2a2c2d383°3a49 d,a,c. a7 d‘ a. o ,. _j 400 soo 500 1000 1200 1400 1500 1500 ml: Flgure 2.9a-b: FAB-CAD-B/E mass spectra on a) 2 nmol of the protonated peptide YQEAFRRFFGPV, and b) 250 pmol of the N- terminal derivative. Relative Intensity. Flelatlve Intensity 72 3. 1er+ R'V‘Y'V'H'P'F a 5 91 7 21 a —H . a4 20 1 g :11 t .32 H be ‘ I 0 - 1 ‘ l E ‘ ‘ ’ H ‘ I h I ‘ ‘ ‘ 1 , 100 200 300 400 500 600 700 800 900 1000 1 MIZ 31 *7 '5‘ + . + '; R-V-Y-H-P-F-TPP M 1 1203 1 Y2 ‘ i 2‘ I + Ye TPP . X1 V2 0 | W, v‘ ' . ‘ 1 x y' w. “'5 Y5 w; 11 l ‘ 23 1 l l Figure 2.10a-b: FAB-CAD-B/E mass spectra of a) the protonated peptide RWH PF, and b) the C-terminal derivative. 73 The two examples above indicate that the basic amino acids have little effect on the fragmentation pattern of the derivatized peptides. The fact that derivatized peptides produce exclusively C-terminal or N-terminal fragment ions even when basic residues are present, demonstrates the capacity of these derivatives to simplify the spectra by only producing ions that contain the ethyl- TPP moiety. B. Post-translationally and Chemically Modified Peptides An important contribution that mass spectrometry can make in the analysis of peptides is the identification and location of post-translationally modified residues. In most cases, established biochemical approaches for determining modifications are not sensitive or well developed, since the more often a protein is chemically manipulated, the less the chance that various modifications will Survive. Although mass spectrometry has some strengths which make it efficient for analysis, it possesses two disadvantages of sequencing by MS/MS: 1) the poor quality and degree of fragmentation makes interpretation difficult and 2) large amounts of material are necessary for MS/MS (0.5-2nmol). In order to investigate the utility of the N- and C-terminal derivatizing methodologies to post-translationally modified peptides, several model compounds were investigated. i. N-terminally Blocked Peptides One of the goals was determining the influence of terminal blocking groups on the formation and fragmentation behavior of the ethyl-TPP derivatives of the opposite terminus. An example of the utility of the ethyl-TPP 74 derivatization procedure on N-terminally blocked peptides is given by bradykinin potentiator C (pQ-G-L-P-P-G-P-P-l-P-P, MW=1051 Da), which contains a pyroglutamic acid residue at the N-terminus. The traditional Edman degradation would not be useful on this peptide because the N-terminus is not a free amine. The FAB-CAD-B/E spectrum of underivatized bradykinin potentiator C (Figure 2.11a) does not yield a complete or overlapping series of fragment ions needed to deduce the primary structure. Most of the fragment ions observed are due to the basic proline residues controlling the fragmentation. The spectrum also contains abundant internal ions (PPG, PPI, PPGPPI, and GLPGPPIP). which only complicates interpretation. Furthermore, the side-chain cleavage ions (W3, W4, d3, and d9) necessary for distinguishing the amino acid residues leucine and isoleucine are not observed. Therefore, it can be seen that, from this spectrum, it is impossible to deduce the amino acid sequence. . In contrast, the C-terminal ethyl-TPP derivative of bradykinin potentiator C, when analyzed by FAB-CAD-B/E provides a more informative mass spectrum as shown in Figure 2.11b. The spectrum displays a complete series of yn ions resulting from cleavages between every peptide bond with charge retention on the C-terminus (i.e., the ethyl-TPP moiety). Also the presence of W3 and ws ions allow the relative positions of leucine and isoleucine to be recognized as residues 3 and 9, respectively. Furthermore, the abundant internal ions formed by the protonated peptide are absent in the derivatized peptides spectrum, again showing the capacity of the derivatized peptides to only produce sequence specific ions. Therefore, the FAB-CAD-B/E spectrum of the C- terminal ethyl-TPP derivative provides an easily interpretable spectrum for complete amino acid sequence analysis; 75 _L O O pE-G-L-P-P-G-P-P-I-P-P A ‘t . I A A A A Relative Intensity Yo GLPPGPPIP (midi 0 200 400 600 800 1000 III]! 100 I u“ 1 ' + 951‘; pE-G-P-P-G-P-P-I-P-P-othyI-TPP .53 5 . l O ’4 "El ' 1 aw in n I . Y2 V5 we . '3 I4 '9 0 . '1 '2 ,3 re 0 400 600 800 1000 1200 I'll/l Figure 2.11a-b: FAB-CAD-B/E mass spectra on a) the protonated peptide pQGLPPGPPIPP, and b) the C- terminal derivative. 76 Another example of an N-terminally blocked peptide is provided by neurotensin fragment 8-13 (Ac-R-R-P-Y-I-L, MW=858) which contains an acetylated N-terminus. Because the two basic arginine residues at the N- terminus provide the most likely site for protonation, this peptide is expected to give predominantly N-terminal fragment ions (24). It has been observed from the comparison of the N-terminally acetylated peptides and the identical, but unblocked peptide, produce very different CAD spectra. It was noted that CAD spectra produce similarly abundant N-terminal ions, while the abundance of the C-terminal ions is dramatically reduced when the N-terminus is acetylated (24). Thus, both the position of the basic amino acid and the character of the blocking group can have a pronounced effect on the fragmentation of a given peptide. . The FAB-CAD-B/E spectrum of the underivatized peptide (Figure 2.123) yields a complete series of an fragment ions as well as the ha and on ion series, with the exception of the absence of the 02 and b3 ions. Also, the presence of peaks for dsa, den. and d5 allows the assignment of isoleucine and leucine as amino acid residues 5 and 6, respectively. The presence of the basic amino acids and the acetyl group promote the formation of mainly N- terminal fragment ions. In this case, the underivatized peptide produces a spectrum that can be interpreted to give the complete amino acid sequence. The C-terminal ethyl-TPP derivative of the neurotensin peptide yields a completely different FAB-CAD-B/E spectrum (figure 2.12b); as expected, the spectrum contains only C—terminal ions. A complete series of yn ions is present, although y5 is of low abundance. Most of the xn and 2.. ion series are contained in the spectrum and provide confirmation of the yn peak assignments. Also, the presence of w5 and w; allows I and L to be assigned to the amino acid residues 5 and 6, respectively. In this case, the mass spectra of both the free and the derivatized peptide provide sufficient information to 77 587 MH+ ‘ Ac-R-R-P-Y-I-L’ 712 £2 a) C ‘ -320 Q) . . E . 700 . -. 171 661\ o . 1 9 673\ -Ac .3. 15 15 63 ' 545 7Q (U i '76 . 242 424 61x | “I ”\7 \ l l - 1 2 355 \ ‘ ' 46 05 \ ' 200 400 " 600 ' 800 C C M Z 3‘. 33% 63‘: 1455\ l b 199- 355\ 51.5 a 18711 $225 48% '53:? 7oso\ 4. Ac—R “ Pl Y r L AC—R R ‘F I L—AETPP ' Z ‘2455 \21 9235‘ ‘73. \ I s \I. Y'2 "2 ' 5'2 Viz Viz Viz \Yi'z 2 on 505 2‘2 rm no .u: 530 417 \x, s 3 \X, \ 974 721 55. Ms Relative Intensity Figure 2.12a-b: FAB-CAD-B/E mass spectra of a) the protonated peptide Ac-RRPYIL, and b) the C-terminal derivative. 78 determine the amino acid sequence of neurotensin 8-13. Thus, it appears that neither basic amino acids nor the acetyl group at the N-terminus has an effect on the fragmentation process when the peptide is derivatized, whereas they both affect the fragmentation of protonated peptides. ii. Peptides Containing a Disulfide Bond Disulfide bridge formation between two cysteine residues is a commonly encountered post-translational modification in a variety of proteins and peptides, and one of thier primary functions is to maintain secondary and tertiary structures (34). Disulfide bonds are the only type of covalent bond involved in stabilizing protein structures (along with hydrogen bonding, electrostatic, hydrophobic and dipole/dipole interactions), thus their assignments are important (35-36). When a protein undergoes enzymatic cleavage the disulfide bonds remain intact and these bridges can be intermolecular, between two different peptide chains, or intramolecular, between two residues in one peptide. The challenge of sequence elucidation of peptides containing a disulfide bridge by FAB-CAD-MS/MS is more difficult than that presented by linear peptides. A single peptide containing one or more intramolecular disulfide links forms a cyclic structure, thus the disulfide bond * interferes with formation of sequence-related fragment ions (37). The FAB- CAD-MS/MS spectra of disulfide bridge-containing peptides generally show less intense fragment ion peaks. Consequently, these compounds are often chemically reduced to open the disulfide linkage prior to analysis by tandem mass spectrometry (37). Methods have been described to completely reduce the disulfide bond on the probe tip by the matrix (e.g. DTT:DTE or thioglycerol). These reduced peptides often produce a greater number of abundant fragment 79 ions in the FAB mass spectra, giving rise to a more complete amino acid analysis (38-39). The normal process to analyze a protein and peptide containing a disulfide bond is illustrated by Figure 2.13. In order to examine the utility of the ethyl-TPP-derivatization approach to the structural analysis of intramolecular disulfide-bonded peptides, a number of peptides were investigated. All of the peptides studied contained the intramolecular disulfide linkage, thus peptides were cyclic in nature. The unreduced spectra were obtained by using NBA as a liquid matrix, and the reduction of the disulfide bond was accomplished by using thioglycerol as the liquid matrix in the FAB experiment. The FAB-CAD-B/E mass. spectrum of the Crustacean Cardioactive peptide, P-F-C-N-A-F-T-G-C-NHg, (MW=956, containing a S-S bond between the cysteine residues) is displayed in Figure 2.143. The peptide does not generate extensive fragmentation, the spectrum displays very few fragment ions other than side chain losses. This restriction of the fragmentation demonstrates . the interference of fragmentation by the disulfide bond. The FAB-CAD-B/E mass spectrum of the unreduced N-terminally derivatized peptide (M+=m/z1245) is displayed in Figure 2.14b. The mass spectrum displays an and a: fragment ions; ions that form without disrupting the disulfide bridge. Also present is the c3-S ion occurring from an unsymmetric cleavage of the disulfide bond. Thus, from the unreduced form of the peptide, only the two N-terminal residues could be obtained. However, after the sample had been bombarded by the primary atom beam for five minutes, the derivatized peptide produced a totally different daughter spectrum (Figure 2.15). The spectrum displayed intense peaks corresponding to a complete series of an ions, except for the as and as fragment. The formation of an ions that originally contained the disulfide bridge must be formed through symmetric 80 PROTEINS WITH DISULFIDE BRIDGES ENZYMATIC CLEAVAGE w '32 -——> MASS SPECTROMETER l REDUCING AGENT SH1—> MASS SPECTROMETER Figure 2 13. The normal process to analyze a protein or peptide containing a disulfide bond. 100- Relative Intensity Relative Intensity 8| % r 1 . ‘ P-F-C-N-A-F-T-G-C-NH2 ag' Unreduced (M+H)+=956 I 82 IlII II on. 200 400 600. 800 MIZ * ’ . I W _ ethyl-TPP-P-F-C-N-A-F-T-G-C-NH2 ‘ Unreduced M+=1245 A L A A A 03-8. H cs 84+ . 86+" Ir " "II III IIIIIIIII I 400 g 600 800 1000 1200 ' mfz A A A A A A A A A a1 a2 “3 ' Flgure 2.14a-b: FAB-CAD-B/E mass spectra of the protonated peptide PFCNAFTGC-NH2 containing a S-S bond between the cysteine residues, and b) the N-terminally derivatized peptide. 82 cleavage of the disulfide bond. In addition, the spectrum still contained the unexpected (:36 and c5 ions, and the appearance of a strong d3 fragment ion. Interestingly, the underivatized peptide did not generate a more informative CAD spectmm after five minutes of atom bombardment, Figure 2.15b. This is not to say that the disulfide bond is not being reduced, rather that the derivatized peptide controls the fragmentation by producing only N-terminal ions while the underivatized peptide produces both weak N- and C-terminal fragment ions and the total ion current is spread over more peaks, thus leading to low intensity peaks. In order to establish whether the NBA matrix was reducing the disulfide bond, the normal FAB mass spectra of the derivatized peptide was obtained after 0 minutes and after 10 minutes of bombardment with the primary atom beam. The parent ion region of these two scans along with the parent scan for the reduced species (obtained by using thioglycerol as the matrix) is displayed in Figure 2.16. These scans clearly display that the peptide remains in the unreduced form in the NBA matn'x. Although the abundance of the parent ion is greater after 10 minutes of bombardment (it is common for the FAB signal to increase with time when using NBA as the matrix), the relative abundance of the observed fragment ions are similar. Therefore, the appearance of the an fragment ions formed by symmetric cleavage of the disulfide bond is not a result of the increased parent ion peak intensity, but symmetric cleavage of the disulfide bond must be occurring in the matrix by one-electron reductions of the disulfide bridge. 83 1004 g. l I - I I a9. g ‘ P-F-C-N-A-F-T-G-C-NH2 ,, C . . 2 I Unreduced (M+H)+=956 > , . E . o 4 a: ‘ * . a2 _ ‘ a1 ,Y‘I b2 b3+H a3+g4+H b5+H a l IIIII " ”I ~ ' {I . 10 . .1 . I j t . ' ethyl-TPP-P-F-C-N-A-F-T-yG-C-NH2 g. . Unreduced M+=1245 ** E 1 . a1 . . * é, : CB'S. a4+H ' a6+H £9 ‘ . . o . m 1 .:a.' a?“ 87+” I I. w . "‘III‘ II I'lII II I IIIII 400 800 1000 1200 MIZ Figure 2.15a-b: FAB-CAD-B/E mass spectra after 5 minutes of bombardment on a) the protonated peptide PFCNAFTGC-NH2 containing a S-S bond between the cysteine residues and b) the N- terminally derivatized peptide. 84 100 ‘ Figure 2.16 a .b 551 5, 111m- 0 Mil. “1 .E, M" - 1245 g: £4 . “’I mI 1 . I o .......... ,v .................. 1 1235 1240 1245 1250 1255 1260 mfz 1 3 Figure 2.16 b .12" 2 11m - 10 nit. é M+-124s cl .>. 51 a4 1&50' ' ' 72235 v 712110' '7 '12'45' ' 7'12'50' ' ' '12'55 ' f '1260 ml: 1 , Figure 2.16 c Q: (I). C. g. -j I“ + 2H -1247 2: =1. 3!. o. “I 1230 7 71235 12110 1245 123507 1255 T 1260 m’z Figure 2.16 a-c: FAB-CAD-B/E mass spectra on a) the TPP+-PFCNAFTGC-NH2 containing a disulfide bond between the cystein residues, at 0 min. of bombardment, b) the TPP+- PFCNAFI’GC-NH2 after 10 min. of bombardment, and c) the reduced form of the N-termlnal derivative. 85 Using thioglycerol as the liquid matrix reduces the disulfide bond, and adds two hydrogens to the peptide to give a molecular ion of (M+2)+=1247, where M is the molecular weight of the unreduced derivatized peptide. The FAB-CAD-B/E mass spectrum of the reduced derivatized peptide, Figure 2.17a, exhibits a complete series of an ions, and many bn and cn ions to substantiate the an assignments. The appearance of the unusual c3-S peak is absent, however, the mass spectrum contains two other unexpected ions, c2-77 and c5-77, which apparently form by the loss of the phenyl group in the sidechain of the phenylalanine residue. Reduction of the disulfide bond in the derivatized peptide produces a mass spectrum which can easily be interpreted to give the complete amino acid sequence. In comparison, the FAB-CAD-B/E mass spectrum of the underivatized peptide with a reduced disulfide bond, does not give a complete amino acid sequence as shown in Figure 2.17b. Although the spectrum displays extensive fragmentation, there is not a complete or overlapping series to determine the ' complete amino acid sequence and only a partial sequence can be obtained. Another example displaying the advantage of the ethyl-TPP derivatization approach for characterizing disulfide-containing peptides, is demonstrated by comparing the B/E linked scan mass spectra of the underivatized mass spectra and the N-terminally derivatized peptide A-G-T—A—D-C-F—W-K-Y-C-V (containing a disulfide bond between the two cysteines, MW=1360). Figure 2.18a displays the CAD mass spectra on 1 nmol of the underivatized peptide using NBA as the matrix. The FAB CAD spectrum of the underivatized, unreduced peptide displays weak peaks representing only the as, 811, and b11 fragment ions, thus this spectrum does not produce useful sequence information. In comparison, the FAB-CAD-B/E spectrum of the reduced underivatized peptide (thioglycerol matrix) displays a greater number of, and more abundant fragment 86 100, 3 P-F-C-N-A-F-T-G-C-NH2 39* I; .j Reduced (M+H)++2H -.-. 958 .5. : ' E I o 3 '1; 1 a5 § 1 . a2 f I35 . 3a, V1 b2 ,2 b3 8454.15 0 ”6°63” 1,7,8 . . 5' 200 ' 400 800 MIZ 100‘ .* ethyl-TPP-P-F-C-N-A-F-T-G-C-NH2 % . a2 Reduced (M++2H)=1247 .1 g . E 36 o) 8‘ 3% d3 34 E . 33 ‘ - 02-77 ‘3 d4 5 If .7 - . d1. 121 a ‘ ”1 dz. ‘ b4 55.5 was 'III rI '1 II I‘I':|| I I 400 600 800 1000 1200 mlz Figure 2.17a-b: FAB-CAD-B/E mass spectra on a) the protonated peptide PFCNAFTGC-NH2 with a reduced S-S bond and b) the N- terminal derivative. 87 1“ . a? Figure 2.18 a g.3 Netlve Peptide In NBA (mum 301 3: £1 :2 3‘ a" I’11 Q. 1:. 1 as I 0. 22 200 400 000 000 1000 1200 l I m Ala-GIy-Thr-Ala-Asp-Cys-Phe-TrpjLys-Tyr—Cys al ‘e '11 I’11 100- 1 Figure 2.18 b ' h11 :1 Reduced Peptide (ll-5H)"’+2Hs1363 h mloglycerol £1 g 1 O 3 ‘ I’10 '11 5. $1 b: h a b-, Y7 '10 : I abshb. e710,...» 0‘ ........................ , , 200 400 000 000 1 1200 V7 "V2 Ye L 000 Ala-Gly-Thr AlaJAsp ys Phe ys-Ty ys al ‘4 '5 .e '10 '11 , b 5: b8b6b7bl I’11 Figure 2.18a-b: FAB-CAD-B/E mass spectra on a) the protonated peptide AGAADCFWKYCV, containing a disulfide bond between the cysteine residues and b) the reduced form of the protonated peptide. 88 ions. Although reduction of the disulfide bond increases the number of detected fragment ions, 3 complete or overlapping series of fragment ions is not observed; consequently, only 3 partial amino acid sequence can be determined. The spectrum also contains a number of low-mass fragment ions that do not correspond to fragmentations along the backbone of the peptide, which would complicate the sequence determination, if this were a true unknown. The FAB-CAD-B/E mass spectra of 400 pmol of the N-terminally derivatized peptide in the unreduced and reduced forms (Figures 2.193 and 2.19b respectively), are different from those of the protonated forms. The derivatized peptide in the unreduced form, generates 3 more informative spectrum than the underivatized peptide. A series of an ions is present, leading to partial sequence determination, but is interrupted at the cysteine residue linked by the disulfide bond. Although the complete amino acid sequence cannot be determined, the spectrum shows the relative positions of the two cysteine residues by the break in the formation of the fragment ions. The CAD mass spectrum of the reduced derivatized peptide displays a complete series of 3.. ions, from which the amino acid sequence is determined. Thus, the reduced form produces a more informative spectrum. In addition, the unreduced derivatized peptide does generate 3 more informative CAD spectrum after 5 minutes of bombardment, however, the complete sequence cannot be obtained. These two examples demonstrate that the derivatization of peptides can enhance the sequence information obtained in peptides containing a disulfide bond. When the disulfide bond is present, the derivatives increase the relative intensity of the observed peaks because only those ions containing the TPP moiety are formed. Furthermore, when the disulfide bond is reduced, the ethyl- TPP derivatives produce a more informative and less complicated CAD mass 89 100 . ; Figure 3 :2? 2 4 N-termlnal Derivatlve of Unreduced Peptide M+=1650 ln NBA 0 1 g . m . ..>. 1 E . £2 1 32 '33 a4 400 600 800 1000 1200 1400 - 1600» + V I m I TPRAh-GjTIjAIjAs . -Phe-Trp-Lys-Tyr-Cys Val . - I ‘ 31 100 . . ' . 4 Flgure b . I g j ' N-termlnus Derlvatlve of Reduced Peptide I 54 M++2H=1652 In Thloglycerol E‘ , o l ' .2 IE I 33 g: .- . . a‘l ‘2 1000 1200. 1400 1600 .... .4..4.4.4.4.4. a: 33 “J 355 a. a7 a, 30 810 311 Figure 2.193-b: FAB-CAD- B/E mass 1&8 ectra on a) the N-terminal derivative of the peptide AGTADCFW CV, containing a S- 8 bond between the cysteine residues, and b) the reduced form of the N- - terminal derivative. 90 spectrum then the reduced underivatized peptide. Although the derivatized peptides seem to fragment more better after five minutes of atom bombardment, the reduced derivatized form generates a much more informative spectrum, thus, the reduction of the disulfide bond is still necessary. Fortunately, reduction of the disulfide bond is simple and can be carried out on the probe tip by using the proper matrix. iii. Phosphorylated Peptides Phosphorlyation is another common post-translational modification of proteins and peptides, and is usually important in the regulation of activity of enzymes. However, recognizing the presence and determining the location of the phosphate group is often difficult by traditional methods (40-43). The phosphoester bonds of serine, tyrosine, and threonine are not stable in the chemical processes of Edman degradation; the phosphate group is hydrolyzed from these amino acid residues (41-42). Although it is possible to convert phosphoserine specifically to S-ethylcysteine that can be detected in the Edman process, phosphothreonine and phosphotyrosine are not amenable to this approach (44-45). Sequence determination of phosphopeptides can be accomplished by mass spectrometry, but usually large amounts of peptide are required to determine the location of the phosphate group by mass differences (46)- In order to investigate both the signal enhancement and the effectiveness of fragmentation associated with the ethyl-TPP derivatives, 3 comparison between the underivatized phosphorylated kemptide, which contains a phosphoserine (L-R-R-A-S-L-G, MW=852, where Sis a phosphorylated serine), with the C-terminal ethyl-TPP derivative of phosphorylated kemptide. The FAB- 9] CAD-B/E spectrum of 1 nmol of the underivatized phosphorylated kemptide (Figure 2.203) displays both HP03 and H3PO4 losses from the MH+ ion, indicating that a phosphate group is present. The spectrum contains mainly N- terminal ions because the phosphopeptide has two arginines near the N- terminus, which most likely are protonated during ionization and, thus, contain the positive charge. The relative position of the phosphoserine can be deduced from the expected mass difference (167 Da for the phosphorylated residue of serine) between the 34 ion (mlz 469) and a5 (m/z 636) and/or the loss of a phosphate group from 35 to form d5. The presence of other peaks in the spectrum, 32-35 and cz-c4, allow the partial sequence ~R-A-S- to be readily determined. However, the primary structure cannot be obtained easily because of the lack of a complete or overlapping series of fragment ions, including wn or dn ions. FAB-CAD-B/E analysis of the prominent (M-H)- ion in the negative ion mode produced fewer sequence-related fragments. The FAB-CAD-B/E spectrum of 100 pmol of the C-terminal ethyl-TPP derivative of phosphorylated kemptide (Figure 2.20b) also yields major ions resulting from losses of HP03 and H3PO4 from the parent ion. The spectrum contains only C-terminal ions (due to the ethyl-TPP moiety) instead of the N- terminal ions observed in the mass spectrum of the underivatized peptide. In this case, the spectrum of the derivative also does not contain a complete series of ions. However, the series of y1 to y4 allows the partial sequence -A-S-L-G to be obtained. The presence of xn and 2,. ions confirm the yn peak assignments. Further, the presence of ya allows the N-terminus amino acid to be assigned as either leucine or isoleucine. The position of the phosphate group can be recognized by the appropriate mass difference (167 03) in the following pairs of fragment ins peaks: y2 vs. ya, X2 vs. X3, and 22 vs. 23. Thus, only the amino acid sequence (L or l)-?-?-A-S-L-G can be obtained. Although a complete 92 "H3PO4 MH+ 14 A A A 724 A I Relative Intensity "'"260'”""460""""660"""'360' )4/UZ 114 270 428 . ah Ias Eh: I 135 I 242 393‘] 453W 636 2 2‘ '25 3:5 5:5 3?; 724 557 . Y1 “'2 Y2 2 k 70 1020 114 043 x \x 3 X5 x, :04 090 571 502 :91 -H3PO4J \JM ‘ J! 03 5“ (x3-H3PO4) * 2 I 511 H _ O I . ‘I I E ‘304 * * ‘ I' "' 4 8130 - 0 . .2 ‘ 417 0‘6 $250” 16 I '5 . 348 III I 332 I I I3E11 1100 £5()() III/2: Figure 2.203-b; FAB-CAD-B/E mass s ectra of a)protonated phospho-kemptrde, and b) C-terminal erivative. 93 amino acid sequence cannot be determined directly from these data, signal enhancement due to formation of the ethyl-TPP derivative improved the detectability of this phosphopeptide by at least an order of magnitude. iv. Backbone Modified Peptides Others have shown that altering the backbone of peptides is useful for determining the functional role of the peptide bond (CO-NH) (47-49). The modifications that have been used include: thiomethylene ether (CHz-S), thiomethylene sulfoxide (CHz-SO), thioamide (CS-NH), and methyleneamine (CHz-NH). These backbone modifications of peptides can enhance receptor selectivity, change the peptide solubility, and make enzymatic degradation by proteases more difficult (50). In addition, some of these modified peptides cannot be sequenced by conventional methods (3.9., Edman degradation), making mass spectrometry an important technique for their analysis. The peptide H-P-F-H-L‘P-L-V-Y (where ‘1' represents a methyleneamine group at the Leu5 position, MW=1011) demonstrates the usefulness of derivatizing backbone . modified peptides prior to analysis by mass spectrometry. Although others (50) state that (M+H)+ ions generate mainly C- terminal fragment ions that contain the methyleneamine group, the FAB-CAD- B/E spectrum of this peptide, Figure 2.21, displays mainly N-terminal fragment ions and few C-terminal fragment ions. This may be due to the basic histidine at the N-terminus being the main site for protonation, thereby directing the fragmentation. The appearance of the b1-b4 and b7 fragment ions allow for partial sequence determination. Although the spectrum displays extensive fragmentation, a complete amino acid sequence cannot be deduced from the observed ions. Furthermore, the spectrum does not contain sequence peaks 94 indicating the modification along the backbone of the peptide nor an intense peak corresponding to an internal fragment as a result of or-cleavage of the CHz-NH bond of the substituted methylene group earlier reported. In contrast, the FAB-CAD-B/E mass spectrum of the N-terminally derivatized peptide, Figure 2.21b, displays a more informative mass spectrum. A complete series of 3.. ions are observed and the appearance of d5 and d5, which determine that Ieu is in positions 5 and 6, allows the complete sequence to be determined. Intense peaks at mlz 921 and 905, representing unusually prominent cs and b5 fragment ions, respectively, clearly indicate the site of modification (-CH2-NH-). Another backbone modified peptide is Y-A-G-F‘P-L-R-R-l, which contains a thiomethylene ether linkage. Contrary to a previous (50) report that peptides containing (CHz-S) upon CA generate largely abundant C-terminal ions, this protonated peptide produced mainly N-terminal ions, Flgure 2.223. Most of the ions are of low abundance. The spectrum does not display a complete or overlapping series of ions, thus the amino acid sequence cannot be determined, however, two major ions yr; and d4 involve fragmentation at the modified residue. The CAD spectrum of the N-terminal derivative of this modified peptide is displayed in Figure 2.21b. As expected, all detected fragments are N-terminal ions. The complete series of 3.. ions, (except 33 occurring at a glycine), some confirming bn and en ions are detected, and the appearance of d4 and d3, which distinguish positions 5 and 8 are leucine, and isoleucine, respectively; thus, the complete amino acid sequence can be readily determined. Furthermore, the location and nature of the modified site is clearly indicated in the spectrum by intense peaks at mlz. 713 and 745 representing the unusually prominent b4 and c4 ions. 95 100- _ W' : Figure 2.21 a 1012 I 3‘: Underlvetlzed Backbonellod Ilodlfled Peptide % 3 Containing a methyleneamine group 3:” 1 o I '6 5 : ‘3? . 0‘ 3 Y7 1 . 4 '7 ‘ x OI___ ._. __.-_ __2..- _.l_42____.____._2_ _I_.__.__-._L -- ___l-1_ .-___.._._._ 200 400 600 000 1”!) Y1 Yo Yo W2 His ro he ls eu-[CHZ-NHJ-Lejwa Tyr '1 '3 '4 'e '1 I’1 I’2 be be I"I c, 100 . 3 Figure 2.21 b N-termlnel Derivatlve of Backbone Modlfled Peptide 5 a, '4 g T a c I a dc ‘ é .. I"I " a g ‘ 3 '2 '63 o 1 finnejvtjmjajeq refine 41.4w Tyr Flgure 2. 21 a-b: FAB-CAD-B/E Mass spectra of a) the protonated peptide HPFHLwW (containing a methylene- amine group on L), and b) the N-terrninal derivative. 96 ”"3 Figure 2.22 a «in ‘Underlvetlzed Fe tld Contel In §:ThlomethyieneE Etheriinkegenmag-S) ‘4 n, * E : s. % V4. d. 3‘ °' Ii ‘1: - Y3 Y2 b“ .7 \ 34' e. c. J | 0 - , - - - , r - . Y - - - 1 - - - zoo M1000 y JAla Gly Phe-[CH2 S] Leu Arg rg lle ‘1 b: 5: ‘°°7 Figure 2.22 b . N-T Imam derlvetlzed PeptideConteh E; e Titlbmeth'yylene Ether Unit-go (cue-5) g i E I £1 5 . O . I . v ‘ v v v I v TPP-TyJAI GlflPheTf J[CH2 S] La rg lie I: I: I: '4 to I7 c‘o Figure 2. 22a-b: FAB-CAD-B/E mass spectra of a) protonated YAGFVLRRI (containing a thiomethylene ether linkage), and b) the N-terminal derivative. 97 v. Cyclic Peptides Cyclic peptides are another type of backbone modification. The cyclic peptide Y-c[K-G‘I'-F-L], Figure 2.23, was used to analyze the effect of localizing the charge on a cyclic peptide. This peptide contains a thioamide group in the glycine residue. This modification has been shown by Tomer et al. (50), to not effect the CAD mass spectrum of a given peptide. The CAD spectrum of the underivatized peptide, Figure 2.24a, exhibits extensive fragmentation, mainly from internal cleavages of the ring structure, labeled by A's. Although it would be very difficult to determine the amino acid sequence, the main point is that the peptide displays good fragmentation. Since this cyclic peptide has a single amino acid tail which contains a free N-terminus, the peptide could be N-terminally derivatized. The CAD mass spectrum of the N-terminal derivative, Figure 2.24b, exhibits very few peaks. The ions detected are attributable to side chain losses and cleavage of the peptide outside the ring structure. Consequently, ring cleavage of cyclic peptides requires the charge to be localized on the ring. Unlike peptides containing disulfide bonds, one electron reductions to-open the ring structure do not occur after extensive bombardment of the sample. Because TPP derivatives contain a ”fixed" charge, they may not be useful in the sequencing of cyclic peptides. 98 o o s o 0. II II II II ll MHz-cl:I-I—c—NH—cH—c-NH—cH-c—NH—cH—c-NH—cH—c- CH2 cit; OH: on; CH2 cu l /\ cm ¢%°% +~= on NH Figure 23: Structure of a cyclic peptide containing a free amino -terminus. 99 100 ; Flgure 2.24a A . A a? . 8* I E; woo -co I Q A=lnternal Fragments x,. g‘ A .-Y CI .1 Al‘ I A .2 b1t:1 ’4 "F -L "160' ' f '260 '1 liefw ' 460 1 130-- 660 mlz 100‘ 3 Figure 2.24 b I g. g. :1 .5 El 0 m4 0360 Flgure 2.24a-b: FAB-CAD-B/E mass spectra of a) the protonated cyclic peptide and b) the N-terminal derivative. IOO vi. Lipoidal Peptides Dr. Pat Kanda has been developing synthetic peptide-based vaccines effective toward controlling virus infections such as Human Immunodeficiency Virus Type 1 (HIV-1) and Hepatitis B. Synthetic peptides can contain epitopes from antibodies which neutralize a given virus. These peptide antigens often require coupling to immunogenic carrier proteins, and attaching an adjuvant in order to induce sustained antibody responses. The disadvantages of this method include undesired immune responses and allergic reactions in the host. Recently, the lipoidal 'N-terminal portions of Braun's Iipoprotein, S-[2,3- Bis(acyloxy)-(2RS)-propyl]-N-acyI-(R)-cystein, when attached to the amino terminus of synthetic peptides, was observed to be a potent immounoadjuvant, enhancing both lgM an lgG antibody responses to the attached peptide. Synthetic analogues of the Braun's Lipoprotein increase the production of antibodies specific for the modified peptide (51). Analysis of these peptides is complicated because purification by HPLC can cause irreversible adsorption of the lipopeptides to the bonded phase. Further, self-association of these peptides contributes to ineffective separation and this aggregation in the FAB liquid matrix causes poor signal response. However, the derivatized lipoidal peptide, Flgure 2.25, is readily detected by FAB-MS, whereas the underivatized lipoidal peptide was not detected under identical conditions. Although it was not possible to obtain a useful CAD mass spectrum on the derivatized peptide, the capacity of the ethyl-TPP derivative to improve the analysis of this class of compounds perhaps by steric interference with the formation of the aggregates. IO] 100 < nco-uH-gomoncpomrvncx-ipp Q: {'2 0). (In-my §< on, me E.‘ HrO-CO-Fi o‘ I Wit .2 l ' E I” I I“ I - Ill , m I ‘ I‘ (I I , I I‘ . mac: Palmltoyl . ‘ ‘ M 8000' ' '2ioo' ' '24bofi"2ooo""zeooTvrfi Figure 2.25: spectrum of the N-terrninally Derivatized Lipoidal Protein. 102 VI. CONCLUSION The ethyl-TPP derivatization procedure provides the analyst with the option of generating exclusively C- or N-terminal fragment ions during analysis of the corresponding derivative by FAB-CAD-MS-MS. There are several advantages to derivatizing peptides by the ethyl-TPP moiety. The detectability of the derivatized peptide is usually at least an order of magnitude better than that of the underivatized peptide. Whether the peptide was C- or N-terminally derivatized, the FAB-CAD-MS/MS spectrum was simplified relative to that of the underivatized peptide by containing only N- or C-terminal ions. Thus, the total ion current is spread among fewer ions, which may explain the increase in some of the daughter ion abundances over those in the mass spectra of the native peptide. Nevertheless, a complete sequence could be obtained in most cases. The ultimate objective of this derivatization technique is the development of a methodology that can be utilized readily in the analysis of picomole quantities of peptides isolated from biological sources. A study involving the tryptic digestion of B—lactoglobulin has been performed. A tryptic digestion was carried out on one nmol of intact protein, than the tryptic fragments were separated and collected into smaller peptide mixtures using HPLC. The N-terminal derivatization technique was applied on a fraction that consisted of approximately 100 pmol of three peptides, and the resulting FAB- CAD-B/E mass spectrum of 50 pmols is displayed in Figure 2.26. The molecular ions of the three derivatized peptides were detected and are labeled by their corresponding single letter codes. The unlabeled peak in the spectrum, which is labeled by an asterisk, is a peptide resulting from trypsin self digest. After the FAB mass spectrum was obtained, two peptides were analyzed by CAD-B/E, the third peptide could not be analyzed because the sample had 103 ethyl-TPP-IIAEK ethyl-TPP-GLDIQK ethyl-TPP—TPEVDDEALEK 1000 1200 1400 1600 mlz Figure 2.26: FAB mass spectrum of a mixture of three peptides after N-termlnal derivatization. The mixture was obtained from a tryptic digest of p—lactoglobulin. 104 been depleted. The FAB-CAD-MS/MS spectra of two of the derivatized peptides, ethyl-TPP-IIAEK and ethyl-TPP-GLDIQK, are displayed in figure 2.27 and 2.28. The spectra exhibit extensive fragmentation leading to the complete amino acid sequence of the two peptides. Therefore, derivatization of the tryptic fraction and subsequent mass spectral analysis allowed the determination of the peptides in the mixture. This demonstrates the capability of using thederivatization techniques on peptides isolated from biological matrices and produce structural information from realistic quantities of sample. ACKNOWLEDGMENT This work was supported in part by the Research Excellence Funds from Michigan State University using instrumentation available in the MSU-NIH ~ Mass Spectrometry Facility which is sponsored largely by a grant (RR-00480- 25) from the Biomedical Research Technology Program of the National Center for Research Resources at NIH. 105 100« f Ethyl TPP-G-L-D-l-Q-K E, . U) I SI E . £1 04 a. as (E: 81 d2 82 a4 d5 as as * 0 500 700 800 Figure 2.27: A CAD-B/E mass spectrum of the N-terminally derivatized (GLDIQK), from the tryptic fraction in Figure 2.26. Ethyl TPP-l-l-A-E-K A A a l . d2 .2 Relative Intensity (:3 at b. d‘ .4 C4 500 600 700 800 Figure 2.28: A CAD-B/E mass spectrum of the N-terminally derivatized (IIAEK), from the tryptic fraction in Figure 2.26. m’z 106 VII References 10. 11. 12. 13. 14. K. Biemann, S. Martin, Mass Spectrom. Rev., 6, 1, (1987). Hunt, D.F., Yates,JR,lll, Shabanowitz, J., Winston, S., Hquer, C.R., Proc. Natl. Acad. Sci. USA. 83, 6233, (1986). Biemann, K., Scoble, H.A., Science, 231, 992, 1987. Stults, J.T., Methods of Biochemical Analysis:Vol. 34, Biomedical Application of Mass Spectrometry., Vlfiley-lnterscience, 145, 1990. W.R. Mathews, R.S. Johnson, K.L. Cornwell, T.C. Johnson, B.B. Buchanan, K. Biemann, J. Biol. Chem, 262, 7537, (1987). TC Johnson, B.C. Lee, D.E. Carlson, B.B. Buchanan, R.S. Johnson, W.R. Mathews, K. Biemann, Bacteriol., 170, 20406, (1988). RS. Johnson, W.R. Mathews, K. Biemann, S. Hopper, J. Biol. Chem., 264, 9589, (1988). S. Hopper, R.S. Johnson, J.E. Vath, K. Biemann, J. Biol. Chem, 264, 20438, (1989). RS. Johnson, K. Biemann, Biochemistry, 26, 1209, (1987). Clench, M.R., Garner, G.V., Gordon, 0.8., and Barber, M., Biomed. Mass Spectrom. 12, 355, (1985). Ligon, W.V., and Dam, 8., Int. J. Mass Spectrom. Ion Processes, 57, 75, (1984) Caprioli, RM, and Moore, W.T., Int. J. Mass Spectrom. Ion Proc. 86 187, (1 988). ‘ Bull, H.B., and Breese, K., Archives Biochem. Biophys. 161, 665, (1974). Heine, 0., Ph.D. Dissertation, Dept, of Chemistry, Michigan State University, 1990. 15. 16. 17. 18. 19. 20. 21 . 22. 23. 24. 25. 26. 27. 28. 29. 30. 107 Falick, A.M., and Maltby, D.A., Anal. Chem, 182, 165, (1989). Renner, D., and Spiteller, G., Angew. Chem. Int. Ed. Engl., 24, 408, (1 985). Chai, W.G., Yan, L, Wang, G.H., Liang, X.Y., Zhao, Y.F., and Ji, G.J., Blamed. Environ. Mass Spectram. 14, 331, (1985). Kausler, W., Schneider, K., and Spiteller,G., Blamed. Environ. Mass Spectrom, 17, 15, (1988). Naylor, S., Findeis, A.F., Gibson, B.W., and Williams, D.H., J. Amer. Chem. 800., 108, 6359, (1986). Ligon, W.V., Anal Chem, 58, 487, (1986). Johnson, R.S., Martin, S.A., and Biemann, K., Int. J. Mass Spectrom. and Ian Process, 86, 137, (1988). Kidwell, D.A,, Ross, MM, and Coltan, R.J., J. Am. Chem. Soc, 106, 2219, (1984). Renner, D., and Spiteller, G., Blamed. Environ. Mass Spectam. 13', 405, (1986) Vath, J.E., and Biemann, K., Int. J. Mass Spec. Ion Proc., 100, 287, (1990). Anderegg, R.J., Mass Spec Reviews 7, 395-424, (1988). 0.8. Wagner, A. Salari, D.A. Gage, J. Leykam, J. Fetter, R. Hollingsworth, J.T. Watson, Biol. Mass Spectrom, 20, 419, (1991). J.T. Watson, D.S. Wagner, Y.S. Chang, J.R. Strahler, S.M. Hanash, D.A. Gage, Int. J. Mass Spectram. Ian Process.,111, 191, (1991). Sunner J., Morales, A., and Kebarle, P., Anal. Chem, 59, 1378, (1987). NJ Jensen, K.B. Tamer, M.L. Gross, J. Am. Chem. Soc., 107, 1863, (1985). Wetzel, R., Halualani, R., Stults, J.T., Quan, C., Blocanjugate Chemistry, 2, 114, (1990).. 31 . 32. 33. 34. 35. 37. 38. 39. 40. 41 . 42. 43. 44. 45. . 46. 47. 108 Maercker, A., Org. React. 14,271, (1965). M. Badansky, Peptide Chemistry, Springer-Berolag, New York, 1988. Roepstorff, P., and Fohlman, J., Blamed. Mass Spectrom, 1 1, 601, (1984). T.E. Creighton, Proteins Structures and Molecular Properties, W.H. Freeman and Company, New York, 1984. S.A. Carr, M.E. Hemling, M.F. Bean, GD. Roberts, Anal Chem, 63, 2802, (1991) S.A. Carr, G.D. Roberts, M.E. Hemling, Mass Spectrometry of Biological Materials; C.N. McEwen, B.S. Larsen, Eds.; Marcel Dekker, Inc., New York, 1990,p.87. S. Akashi, K. Hirayama, T. Seina, S,l. Ozawa, K.|. Fukugara, N. Oauchi, A. Mural, M. Arai, K. Tanaka, l. Nojima, Blamed. Environ. Mass Spectrom,, 15, 541, (1988). AM. Buko, B.A. Fraser, Blamed. Mass Spectrom,, 12,577 (1985). R.G. Ginke, M.W. Draege, J.C. Cook, K.S. Suslick, J. Am. Chem. 800., 106.5750, (1984). SE. Mayer, E.G. Krebs, J. Biol. Chem, 245, 3153, (1970). DB. Rylatt, P. Cohen, FEBS Lett, 98, 71, (1979). CG. Proud, D.B. Rylatt, S.J. Yeaman, P. Cohen, FEBS left, 80, 435, (1977) T. Patschinsky, T. Hunter, F.S. Esch, J.A. Cooper, B.M. Sefton, Proc. Natl. Acad. Sci. USA, 79,973, (1962). HE. Meyer, E. Hoffmann-Parorske, H. Korte, L.M. G. Heilmeyer, FEBS Lett, 204, 61, (1966). C.F.B. Holmes, FEBS Lett, 215, 21 , (1987). B.W. Gibson, P. Cohen, Methods Enzymol, 193,480, (1990). AP. Spatola, In Chemistry and Biochemstry of Amino Acids, Peptides and 48. 49. 50. 51. 109 Proteins, Weinstein, B., ED.; Marcel Dekker: New York, 1983, Vol 7, p 267. J.V. Edwards, A.F. Spatola, C. Lemieux, PW. Schiller, Biochem. Biophys. Res. Cammun. 136, 730, (1986). K. Clausen, A.F. Spatola, C. Lemieux, P.W. Schiller, Biochem. Biophys. Res. Commun. 120, 305, (1984). L.J. Deterding, K.B. Tamer, A.F. Spatola, J. Am. Soc. Mass Spectram. 1, 174, (1990). K. Deres, H. Schild, K.H. WIesmuller, G. Jung, H. Rammensee, Nature, 342, 561, (1989). Chapter 3 Fragmentation Mechanisms of Peptide Bonds By FAB-CAD-MS/MS I. INTRODUCTION Fast atom bombardment (FAB) collisionally activated dissociation (CAD) of protonated peptide molecules yield a number of sequence ions that provide information about the primary structure (1-13). Specific mechanisms have been suggested and widely accepted for the commonly detected sequence ions (1 ,.2,5,14). While in-depth studies of fragmentations involving cleavage of the amino acid side chains yielding dn and wn ions have been reported (15), few stable isotope labeling studies have been performed to elucidate the structures of the sequence-determining fragment ions and the mechanisms through which they are formed. ' Although the identification of peaks representing an, b... on, x... y" and 2.. ions in a FAB mass spectrum has been demonstrated to be straight forward (4), the relationship between mlz values and the relative abundances has not been established. Determining fragmentation mechanisms requires knowledge of the structures of the fragment ion and the neutral fragment. In order to understand the fragmentation of peptides, some mechanistic questions of the chemistry proposed to date must be resolved. Peptides are protonated either in the liquid matrix, and desorbed as the protonated form, or become protonated following the desorption process. It has been shown that protonation of a peptide is external originating from the FAB matrix (16). There are several basic sites on a peptide and the protonation could occur on the nitrogen of an -NHn- group, on the oxygen of an amide 110 111 carbonyl group, on a basic side chain, or on either terminus. It has been suggested that the majority of the protonated molecules have a structure in which the most basic site in the peptide is protonated (14, 17-19). Early work utilizing chemical ionization (CI) of small peptides (20) suggested that protonation of carbonyl oxygens and basic side chains may give relatively stable forms that do not fragment before being detected as [M+H]+. More recently, it has been suggested that the N and O atoms in an amide group could be protonated with equal probability (18). Therefore, two extreme possibilities exist: 1) protonated peptides have one structure, where the most basic site is protonated, or 2) the protonated peptide has several forms, where one of several possible sites is protonated to some extent. The description of a protonated peptide becomes more complicated when one considers the secondary and tertiary structures. Hunt 91 al. (2) suggested that protonated peptides in the gas phase fold in order to delocalize the positive charge and stabilize the peptide. This folding process brings one or more amide linkages in the vicinity of the protonated site, which serves as a plausible means for proton- transfer reactions to relocate the charge, thus promoting fragmentation throughout the peptide (21). The commonly accepted mechanisms for fragmentation of protonated peptides assumes that the chemistry is charge-induced (2), that the charge from protonation is responsible for most of the fragmentation. The mass spectra of peptides contain many intense peaks representing different ions, suggesting that the protonated peptide molecule is comprised of a distribution of species, with protonation occurring at many different sites, thus leading to fragmentation throughout the peptide. However, peptides containing basic amino acids (arginine and lysine) at or near the ‘N-terminus promote the formation of N- terminal fragment ions, while those with basic amino acids near the C-terminus 112 generate mainly C-terminal fragment ions (22-24). This suggests that the most basic sites in a peptide are preferentially protonated, the positive charge is somewhat fixed at the basic site and the subsequent fragmentation is remote from the charge (14,25). During the FAB-CAD experiment, the peptide is protonated and excess energy is imparted into the ion through collisions with an inert gas. The additional energy is randomized and then used to induce bond cleavage at some site in the molecule, which need not be in close proximity to the charge. The site of the charge would only determine which of the two peptide fragments carries the charge and, thus, is detected. Although remote-site fragmentations have been demonstrated for long chain fatty acids (26), in which the C-C bond cleavage occurs at a location physically removed form the charged group, such fragmentations are more difficult to prove in molecules containing many functional groups, where the site of protonation is not localized. The capability of charge-localized derivatives of peptides (14,17, 19, 27-31) to generate extensive fragmentation by CAD, is evidence that remote-site chemistry can occur in peptides. Thus, the site of protonation is important to determine whether the fragmentation is charge- directed or remote-site. it is apparent that when some of the fragment ions are formed, the bond cleavage process involves a hydrogen shift. Identifying the origin of these hydrogens is important for determining the fragmentation mechanisms. For example, the yn ion contains two additional H atoms and a positive charge. One H atom is the protonating H (16), the other shifts from the neutral fragment. The origin of the transferred H atom can be proximal to the site of fragmentation, through an elimination reaction, or the shifting H can come from atoms distal from the site of bond cleavage, that are brought into close proximity by secondary structures. In the latter case, formation of ions such as the yn ions 113 could result in the formation of cyclic neutral fragments (20). Thus, the number of sites from which the H-shifts could originate increases with the peptides size of the peptide. In addition to the structure and mechanisms for the formation of the fragment ions, it is important to determine the possible relationships between the ion series. The nomenclature (32,33) for the two possible fragment ions produced when -CHR-C(O)- bond is cleaved, labels the N-terminal fragment and C-terminal fragment as an and xn ions, respectively. Similarly, (bn and ya) and (on and 2..) have the same skeletal bond broken for the formation of each pair of ions. Thus, are both members of these pairs formed through different processes, or through common intermediates? This depends on the ionic structures of the fragment ions of interest. Since the exact structures are not known, this question has not been answered. Figure 3.1 displays some possible structures for the fragment ions occurring along the backbone of a protonated dipeptide. Figure 1 is not meant to represent the actual structure of the ions, but rather focuses on the relationship of the fragment ions' to the protonated peptide molecule. For example, the structure a; would be more accurately written as [H2N+=CH(R)]. The structure 61 is intended to display the structural relationship of the fragment ion to the parent structure of the protonated peptide molecule. The important issues that need to be determined are whether the fragmentation chemistry is remote-site or charge-induced and the origin of the shifting H atoms. Thus, some aspects of the fragment ion structures are important. I a and x ions: Some of the fragments resulting from the FAB-CAD process are protonated even-electron molecules (13). Are all of the fragments protonated? For example, the an ions appear to result from a simple bond 114 6 0+ Ell-l z-CH 45.1111 -an 4:.) a o R1 131 “2 [mg—(in —3—NH -CH -Cl—OH] 11" Re [MHz-CH j-NH «EH-fl .41] 11" y R1bII Hz D Y [le-CH—fi—NH—EH 0 NH H—E-NH—it‘i—fi—alilr 81 2 3 a x c z . O 0 O O 7 + Hz—CH -8-NH —c + ‘ -NH -CH —3—NI~I —CH —3-aI-I 1” 1, II] 3 31 XI 0 o o o Hz—CH -8-NH -CH- 11" 3-NH -CH —8-NH -CH -8-OH . If P R1 F'iz-H] [ 11 Jill an 3 xII [VHz-CfH Iii—NH -EH iii—ma] 11" [EH j-NH -CH j-aHf R1 2 3 21 [CH j-NH -CH j—OHLH] ii" 1'13 1., 211 Figure 3.1: Possible structures of the fragment ions occurring along the backbone of the peptide. 115 cleavage in the form of a1. However, the an ions could form through a +1/-1 mechanism, where the peptide fragment loses a H atom, but appears in the protonated form, which would be more accurately represented by structure an. If this is the case, then one must consider the origin of the lost H atom. The same questions can be asked of the xn ions. The structure could be a simple ionic fragment of the peptide, x1, or a H atom could shift as depicted by X“. If the latter occurs, then the origin of the shifting H is important for mechanistic considerations. Since an and xn fragment ions cleave at the same skeletal band, it is possible that their formation involves a common intermediate. Figure 2 displays two possible mechanisms fro the formation of a and x ions via a common intermediate. Figure 3.2a shows the possibility of forming a; and X], where the hydride stays with the neutral fragment and leaves the detected ion. Figure 3.2b is similar to the mechanism proposed by Gross et al. for remote site fragmentation of fatty acids (26), where the C-C bond cleavage involves the elimination of a H2 molecule. Here, the structures an and xn are formed from a common intermediate, where the protonated fragment is detected. These considerations are not consistent with the experimental observation of fragment ions 1 Dalton higher (an-+1 and xn+1). b and y fans: It has been suggested that bond cleavage adjacent to a protonated amide nitrogen can generate a fragment with the structure b1 (Figure 3.1), with the loss of a neutral HzN-R molecule (14). It has also been suggested that the an ions are formed from the b" fragments by the loss of a neutral CO molecule (2,14). However, one cannot distinguish between the structures b1 and bn in Figure 3.1, where bu can be written as [(bI-H)+H]+. The stmcture bu 116 R a R 1 (Ia M II 12 II [HgN-CH+ ----- H- ----- C-NH-CH- C— CH] +CI- NH-CIH- C-OH 0 R2? R1 11 + HC-NH-CH-C-OH | H2N-CH2+ H2N—CH+ x1 a1 Figure 3.2a: Fragmentation of a protonated peptide to form a and x via a common intermediate involving a hydride ion. 117 "— <22» “‘ F11 ‘1 . 403 ”~.. [HzN-CH-C-NH- h /H H+ E N-CH-fi-OH __ Rs 0 _ -H2 _ __ 1111 122 ‘11 E fi‘za r’mHz-CH-c-NH-CH=c-RZD]H+ NHz-CH- -NH-CH=C-R2b all + H-I- OR O=C=N-CH- -CH 1'13 __ '53 _ _,[O=C=N-CH-C-CH]H+ xII Figure 3.2b Fragmentation of a protonated peptide to form a and x via a common intermediate involving the expulsion of an H2 molecule. 118 links bn and yn ions as possibly originating from a common intermediate as well as involving cleavage of the same band. The skeletal -C-NH— band would cleave with a H-shift from the N-terminal fragment to the C-terminal fragment. Here, the detected fragment ion contains the protonating hydrogen. a and 2 ions: These ions originate from cleavage of the same skeletal bond. The 6.. ions represent the N-terminal fragment plus two additional H atoms; presumably, the fragmentation involves the transfer of a H to the amide nitrogen and the resulting fragment appears in the protonated form as illustrated in Figure 3.1. Since a H shifts to form the on ions, then the formation of the 2.. fragment ions could involve a H shift giving rise to the structure 211 in Figure 3.1. However, 2 ions could be simple fragments from the protonated peptide, illustrated by structure 21 in Figure 3.1. The interesting observation about the formation of the clzn ions is that the mechanism cannot involve a 1,2- elimination because the C-terminal fragment does not have a H in the required position (carbonyl group). The nearest available H is located on the side chain group R2. * Consequently, there are many obvious qUestions about the structures of the ions observed, as well as the mechanisms through which they are formed. The possibility of remote-site mechanisms of- protonated peptides will be realized or disputed when a better understanding of the fragment ions . structures and their mechanisms of formation is available. In Order to investigate the cleavage of peptides, FAB-CAD-B/E was used to study the fragmentation patterns of a number of deuterium-labeled variants of the hexapeptide H-VGVAPG-OH. The structures of the labeled peptides are shown in Figure 3.3. The analogues contain deuterium-labels on the ot-carbon 119 o g-ND-CH—g-OE 1H O 2 _ w H\ I ._. m. H 2 0 TI“\ 0' .H m. %IAC M OIC 3 . H H H\c .62.... D C * OflC e." Jul-H D 0 fl 3 NEg-CH-C-N CH /\ Halfli C o-vdadvdAdPad-oo H—fi-ffli 41H H « ONO 2 _ H H\C’2 3.» H C 2 fl/W\ O"% H m. H16 4 OflC 3 _ H 4. 6.... _Ic/u H C u." _ H Jle H w 3 O"% H D H\C T4 H C N H-Vadeo‘dAPG-OH m OflH CIH _ H N on? 2 _ H “tel. H a .. .. /H\ on c. w .. H HIAC N _ OflC 3 _ H a In 3 . or." H C 4 OH% H n—VIH MHz-CH-§-¢H1 C I CH3 H-VGVAadPo-OH 5; . H—E-(Hi C 1 H \ CH / ‘11 -CH-1?-NH- 2 CH2 I ‘\ O N H—CH—C-1l CH3 CH2 CH-E-N C-1H+-?H-4>-NH- H CH3¢H3 MHz—CH-’ f2 H-deGdeAPG-OH C l H \ CH / O 11 -CH-1P-NH-FF4L—OH CH2 I ‘\ -wuHCH-oeu on. CH2 1 -?H-4>-HH-uou>é «+ «1 ~+ «+ M” «+ «o «+ a 3 3o: .. Z +:_:o.2oue>.:_ a. «+ «+ «+ NH 9. «+ u+ _.. I in: u 3 +=E92<3>03>i 9. 9+ 3. 9. ”H o o a u a a t. .. a +:.:o.o._3n<>o>.:_ _. 3 I 2 "H o o a _. o o _.. _. o +2.323<>o>.:_ gum :3. «a. a: 89 ”um. 2h .3 222.53. 3:25.852. £5.13 8 883... 93 no on as 3 on no no an «o «a «a we 5 pa. Eon-.538 Essa-c one vac can sun can z.“ emu ONN «hr hr... our 9:. auras \ \ \ \ \ \ \ \ \ \ \ \ \ \ 893.10 253.5... _ _ _ __ _ _ _ _ _ _ _ _ ._ _ _ _ _ _\J_ _ _ _ _ _ _ _ _ _ _ IOIMIIOIFIZH :01 12.._. FIAT—+—ZHWFZOPZZHwFIOFIZHWI—ttzz Q. __:._o_ _o_.:_._ _o_ ._._ _o_:"_. nzoxnyo azoxayo :0: o a > o > i... onion. 35:80:. :3: £559: nuou>.o. .. _.. o a .. _.. I I .+ 3 + +:.:o.o._<3>oue>.:. o .. o o a c 7 I I 7+ .. +3.:oé..<3>o3>.:. o .. .+.+ a+ 9+ _.. a+ _.. 9+ a+ _. +:.:o.o..u8<>o>.z. o _._ —+ I .+ u I a .3 .+ a +:.:o.o..3<>o>..... Egg 38.2.13: 2.23:3“ 3“ 22223232.: 2.18: 3.... +...: .o 98.8... 30 wu ; «x a» «x as a> ox cu ; vx an a» an .3:-:98: .5802". on I. «as p: a: «an a: on“ cu» we» a: no» no» one cog-0.0 0.5.2.3... / I / / / / / / l / / / / / _ .\J_ _ _ _ z?w|:o..zz.m.w.m.:olz _MHMHFzHW ._.:o._.:z-_rm.rzo._.:z._. Fold... \0/ ...o \»20 «106920 3.2.0.. 35:29:. Ea... $5892.... 32.5.3.0 "ed «in... a1 82 as 84 85 b2 b3 b4 c1 c2 ‘33 c4. 'c5 126 Table 3.2: Intensities of the Fragment Ions I II III Hammumtlismmmn,tuumumt 0 5.1 0 0.8 -1 '38 0 1 .2 0 0.3 '1 2.1 0 5.9 0 1.9 -1 100 0 3.8 0 3.8 -1 1 5 0 2.0 0 2.0 '1 24 0 0.9 +1 1 .5 0 5.5 o 1 +7 -1 21 0 20 0 10 +1 2.4 0 100 0 100 _1 3.2 0 17 0 0.5 +1 3.8 * . +1 1 .7 +2 0.1 +1 1 .8 +2 0.2 +1 3.9 X1 X2 X3 X4 X5 V1 Y2 Y3 Y4 Y5 21 22 23 24 25 I II 1‘7 -1 21 -1 5.0 0 0.1 -1 7.9 O 3.2 -1 12 -1 9.6 +1 0 '1 13 ~ +1 0 +1 14 +2 37 +2 0.9 _1 14 +1 6.8 +1 26 +2 2.4 _1 1 1 O 6.4 +1 5.9 +1 0.8 O 15 0 7.8 +1 0.4 0 6.4 127 easier to recognize. As an example, consider the data on the C-terminal fragments in Table 3.1b. If the C-N skeletal bond in alanine were broken homolytically, the C-terminal fragment, ya ion, would correspond to a mass of 242 Daltons. The FAB mass spectrum of the protonated peptide displays the peak corresponding to the ya ion at mlz 244. This can be designated as the protonated C-terminal ion, with an additional H, [242+H]H+. The CAD spectrum of the peptide in which the alanine has been monodeuterated at the a-carbon displays the ya ion at mlz 245 because this fragment contains the deuterium, thus the mass shift is +1. an Ions: All five possible an ions for VGVAPG are observed in the B/E linked-scan CAD mass spectrum of the protonated, unlabeled peptide. The an ion series corresponds to the mlz ratio of a simple bond cleavage between the a-carbon and the carboxyl group with charge retention on the amino terminus fragment. This can be accomplished by four possible mechanisms: A1). Simple cleavage of the C-C bond in the -C-CO- group. A2). A +1l-1 reaction involving protonation and the loss of a hydrogen from the a-carbon. A3). A +1/-1 reaction involving protonation and the loss a hydrogen from the B—carbon on the side chain. A4). A +1/-1 reaction involving protonation and the loss of a hydrogen from an amide. Some al.-+1 ions are also observed. The a: ion is the least abundant ion in the series, corresponding to cleavage of the peptide at a glycine residue, which does not contain a sidechain. Table 3.1a lists the data obtained for the partially deuterated forms of the peptide. For example, in the CAD data for the [H-VGVAGdPG-OH]H+, a1, a2 128 and a3 fragment ions appear at the same mlz values as, observed for the undeuterated form, with 34 and 35 being shifted by +1 m/z value because they both contain the deuterated label on the alanine residue. Thus, data for the d1- alanine indicate that the deuterium label on the a-carbon remains on the m; fragment ion (mass shift from m/z 299 to mlz 300) and the as (mass shift from m/z 396 to m/z 397). Also, the B/E spectrum of [H-VadGdeAPG-OH]H+ exhibits the a1 (mlz 73, an addition of one Dalton) and the 83 (mlz 230, an addition of 2 Daltons) indicating that the deuterium labels on the a-carbo‘ns are still present. Similar results are obtained when the peptides contain deuterium labels on the B-carbons. The CAD spectrum of the [H-VGVAflwPG-OH]H+ analogue displays an 84 ion with a mass shift of +3 (mlz 302), thus all three deuteriums are still present. Likewise, CAD on the [H-VBdGVBdAPG-OH]H+ analogue displays 81 having a mass shift of +1 (mlz 73) and as having a mass shift of +2 (mlz 230), showing that the deuterium labels are still on the respective ionic structures. From these data, the form of an ions could be a1, a simple fragment in ionic form, or an, where the H that shifts away from the N-terminal fragment does not originate form the a-carbon or the B-carbon of a residue, but would have‘to involve an exchangeable H, on one of the N atoms of the N-terminal fragment. The CAD data for the [D-VdeVdAdPGd-OD]D+ does not determine if a H-shift from a N site occurs, because if a D shifts, the final product would be deuterated, thus there would be on D lost and one D gained. If a ions involve the transfer of a H and the ions do have the stmcture of an, then there are two extreme possibilities. One is that the transferred H atom is always from the N- terminus; the other is that it shifts from a N which is closest to the skeletal bond being cleaved. Since the as ion is observed, and there is no H on the N of proline, the latter is not the case. If a H does shift, it is not from a "local” N atom. Most probably, a ions are formed as a simple fragment, following local ”M's- l29 protonation of amide groups, deriving from b ions, through loss of CO, illustrated by figure 3.5. If this is the case, we note that b1 completely dissociates to form 31, consistent with the very stable form of this particular a ion, [H2N=CHR]+. xn lons: In the CAD spectrum of [H-VGVAPG-OH]H+, only the x3 ion is observed, although for some of the deuterium labeled peptides, X4 was detected. CAD of some peptides does not produce simple cleavage product * ions, but generates x+1 and or x+2 fragment ions. The data in Table 1b lead to the similar conclusion made for the a ions. If the x ions are not simple fragments, but have the structure of XI], which loses a H atom and is protonated, then the H atoms are not lost from the a- or fl-carbons. A H could be lost through a 1.2-elimination process, involving the shift of a H atom on a nitrogen (Figure 3.6), however, this cannot be detected in these experiments. Clearly, deuteration experiments are not sufficient for determining the possible structures of a and x ions. bn Ions: The b" ions do not gain a hydrogen from the fragmentation process; peaks observed for these ions correspond to the mass of an ion formed by bond cleavage between the carbonyl and nitrogen with charge retention on the amino terminal fragment. As in the case of a ions, b ions could be simple fragments, formed following protonation at amide nitrogens, or they could have the form of bu. The bn ions can form by three possible mechanisms. I30 R1 0 R2 l g l {‘1 + HgN—CH— —NH—CH-—CEO U bn ion -CO ll 1 0 R2 R l )1 + l H2N—CH— —NH=CH anion Figure 3.5: Formation of the an ion from the bn ion. l3] - ” H+ - R1 0 | R2 0 R3 0 I HzN—JEHZ + £=N— H—g—NH—éH—g—OH N-terminal Xn ion neutral Flgure 3.6: Formation of the xn ion via a 1.2-elimination reaction, remote-from the charged site. I32 31). Protonation of an amide followed by simple bond cleavage. 82). A +1/-1 reaction involving protonation and the elimination of a hydrogen from the ct-carbon. 82). A +1 /-1 reaction involving protonation and the elimination of a hydrogen from the B-carbon. B4). A +1/-1 reaction involving protonation and the elimination of an amide hydrogen forming a cyclic product. In the data presented b1 is never observed, but all other b ions are observed. The b4 ion is the dominate fragment ion in all of the CAD mass spectra. The CAD spectrum of the [H-VGVAadPG-OH]H+ displays peaks at mlz 328 and mlz 425, which correspond to the b4 and b5 ions containing the deuterium label. In addition, the CAD spectrum of [H-VWGVWAPG-OH]H+ exhibits b3 (mlz 258), b4 (m/z 329), and b5 (mlz 426) indicating that the deuterium labels are still present in the (at-position on both of the valines. Thus, the formation of the bn ions does not involve the transfer of an tat-hydrogen. The mass shifts of the CAD data for [H-VGVAB3dPG-OH]H+ and [H-deGdeAPG-OHJH+ specifically show that the bn fragment ions are not formed by the loss of the B-hydrogen. The data from the isotopic variants indicate that b ions are either simple fragments with structure b], or involve H shifts from N atoms. We note that, in previous chemical ionization experiments from permethylated peptides, in which H atoms on nitrogens are replaced with methyl groups, the Cl mass spectra contain a and b ions (20); consequently, they are probably simple fragments, and do not involve H shifts, as depicted by Figure 3.7. H2N— I33 F|11 0 RL2 R3 CH— —£— —NH— CH— CED —NH —CH—i— —OH H l R1 0 R2 0+ R3 HgN—cH—g—NH—cH—g + NHz—CH—g— —0H bn ion C-terminal neutral Flgure 3.7: Formation of the bn ion from protonation of the amide nitrogen. . 134 y" ions: The formation of the yn fragment ion is thought to involve protonation of an amide nitrogen and the transfer of a second hydrogen to the amide nitrogen during bond cleavage. The hydrogen transferred to the amide nitrogen during bond cleavage can originate from three sites: the tat-carbon, an - NHn- group, or one of the side chains. The two most probable mechanisms for the formation of the yn ion are shown in Flgure 3.8. Evidence for mechanism 2 has been reported by Richter et. al (16). by using a peptide similar to the [D-vdevdAdPGd-OD]D+. Although the y1 ion is never observed, the my; ion series is observed for all isotopic variants, with y2 dominating. FAB-CAD-B/E analyses of the peptide VGVAPG and the deuterated analogues permitted recognition of the site of origin of the transferred hydrogen. If the hydrogen on the a-carbon were tranSferred in the cleavage process, then a one-mass unit shift should occur in the formation of the ya fragment ion in the peptide containing the deuterium label on the tat-carbon of the alanine, and in the formation of ya and y5 fragment ions from the peptide containing the deuterium labels on the tat-carbons of both valines. However, the mass spectrum of the deuterated analogues di_d_ng_t_show a shift in the peaks corresponding to the ya ion (figure 3.9a) for the [H-VGVAadPG-Ole or for the y3, y4, and y5 ions in the B/E spectrum of the [H-vadevadAPG-OHJH+ labeled peptide. The mass shifts which do occur in the yn series are due to the presence of the deuterium on either the alanine’s or the valine’s a-carbon which is retained in these fragment ions. Thus, the formation of the yn ions is not through a 1,2 elimination involving the transfer of a hydrogen from the a- carbon as previously reported. Furthermore, the CAD mass spectral data for [H-VBdGVBdAPG-OHlH+ and [H-VGVABMPG-OHlH’r analogues show that the H being shifted does not originate from the B—carbon. .135 2 ll HzN—C—C —h|l—i=c=0 H 2 1 Neutral 1 3|. IT 3 IT 3 + H—N—i— —N — -OH or J. 3 l 4 l:12 +2.. 74.—~73.“ ’2 ‘°" 1 I Neutral 2 Figure 3.8: Two most probable mechanisms for the formation of the yn ion. I36 Niki/EH—fl—NH-CH—fi-NH—EH-fl-NH—z- -fi*}l—CH—fi-NH-CH-g-OH l-l" 011.0":ch H 1 F on, CH, CH; 0". — H— -}1—CN—fi—NH—CH-fi-OH 11" Cch "1’0": H y2 ion Figure 3.9a: Formation of the y2 ion from the ct-labeled alanine containing peptide. Figure demonstrates that the deuterium on the ct-carbon is not transferred. 137 When the exchangeable H atoms are replaced by deuteriums, and CAD is performed on the ion [D-VdeVdAdPGdeODth it is apparent that the H shifts from a N atom, yielding cyclic neutral products. For example, the ya fragment ion (figure 3.9b) is observed at m/z 177, displaying a mass shift of +4, corresponding to the addition of four deuterons. Three of the deuterons can be explained by H/D exchange and FAB ionization from the dg-glycerol. The fourth deuterium must be transferred from one of the labeled amides on the peptide. This indicates that the hydrogen involved in the transfer is located on an -NHn- group. This has also recently been reported (16, 34, 35). Are the H shifts from local sites, or from more distant sites? While exchangeable Hs' clearly shift, these experiments do not indicate their source. There are data that would support both possibilities. The absence of the y1 ion supports the idea that the H atom is transferred from the local nitrogen atom, since the C-terminally adjacent residue is proline which does not have a hydrogen available to transfer. However, when the peptide is N-acetylated, a drastic reduction of abundances of all y ions .is observed, which suggests that, to some extent, H- shifts are from the N-terminus. If this were the case, then formation of ya ions would be interesting. For example, the formation of the y4 ion, Figure 3.10, has the H-shift via an 8-membered ring intermediate, while a 6-membered ring is being formed as a neutral cyclic dipeptide product. Likewise, the rest of the yn ion series could give rise to the elimination of a neutral cyclic compound involving a 3, 6, 9, or 12- membered ring. 138 ‘— _1 i i . 1' 1 + lug—CH-c—No—rlsH—c-No—CH— —ND—CH—C —CH- -ND—CH- —oo o [111 H [1111 CH, CH, [c112 H omen, 011.011. CH: o—yectl-fi—No—fH—g—oo 0 CH2 H2 H \cflafi y2 ion Figure 3.9b: Formation of the y‘2 ion from the d7-labeled peptide. Figure displays that a deuterium on a nitrogen is transferred in the fragmentation process. 139 l _ — o o I . r170— 11: ‘D’ NH-CH-llsl—APG-OH H+ ' " HE}; t ‘i H C— i NH /NH + [H-NH-CH-C-APG-OH]H+ \_ H V H 1 Y4 O N-terminal neutral Figure 3.10: Formation of the y4 ion. Figure shows the formation of a neutral 6-membered ring and H transfer through an 8-membered ring. 140 cu ions: Similar to the formation of ya series, the en fragment ions are derived from the protonated peptide by cleavage of a given C-N bond involving the transfer of a hydrogen to the amide nitrogen, thus gaining two hydrogens in the fragmentation process. Although on ion series is of very low abundance, all of the ions can be detected except for the c2 and the c4 ion which cannot form. The CAD spectra for the protonated peptide, [H-VGVAadPG-OH]H+, and [H- VGVAB3dPG-OH]H+ produced a peak at m/z 273 for the ca fragment ion. Thus, the transferred H does not originate from the a-carbon or the B-carbon via a 1,2- elimination. The CAD spectrum of [D-VdeVdAdPGd-OD]D+ exhibited a c3 peak at m/z 280, corresponding to the addition of seven mass units. Six of the mass units can be accounted for by the five exchangeable sites now bearing deuterons, not hydrogens, plus the deuteron responsible for deuteration instead of a protonation. The seventh additional mass unit must be from a deuterium shifting onto the N-terminal fragment ion. Thus, confirmation is obtained that it is an exchangeable H that shifts in the formation of the on ion, illustrated by Figure 3.11. Again, there is the question of whether the H is shifted from a local site or from a distant site such as the COOH at the terminus. The c3 ion is observed, and there is no local H on the N of proline, so the H may come from an atom many skeletal positions away. If a 6-membered ring were formed as a neutral product, as shown in Figure 3.12, then (:2 would not be formed, again due to the position of the proline residue. There is a small peak representing the 05 ion, indicating that the H-shift can originate from the -COOH terminus. If a D shifted from -COOD in the fragmentation of [D-VdeVdAdPGd-OD]D+, a mass shift of +8 mass units would be expected for C5, however, the ion is not observed. This is not surprising, since the neutral molecule eliminated would be a highly-strained 3-membered ring. Again, D-Iabeling alone does not provide the whole answer. 14] \ NH ,0 neutral 1 H \ Figure 3.1 1: Two possible mechanisms for the formation of the on fragment ion. 142 H- MGV-NHCCH- ; \ (EH2 COOHO ‘EH3 V H— ? \ [H-VGV—NH2]H+ / N\C_CH/N era ("3 \3 C3 COOH C-terminal neutral Figure 3.12: A possible mechanism for the formation of the C3 fragment ion. I43 2.. Ions: It is interesting that the fragmentation of a given bond can produce different forms of the same fragment ion. The 2 ions do not have one simple form. In some molecules, the 3 ions appear to be simple fragments, while in this peptide, two 2 ions, 23 and 25, are observed as zn+1 ions. The data were obtained from CAD of an even-electron ion, the protonated peptide. In order to form the 2+1 ion, a radical cation and a radical neutral fragment must be lost. The charged fragment gains one H, either the protonating H, or a transferred H from the neutral fragment. The a- and B—labeled peptides display masses that show that neither the a—carbons nor the B—carbons lose a hydrogen in the fragmentation process. The CAD on [D-VdeVdAdPGd-OD]D+ generates 2“ ions which gain a deuterium, and not a hydrogen, in the fragmentation. Thus, the 2.. ions are formed by the hydrogen shifting from an exchangeable site to the detected fragment. This does not provide information on the source of the additional H, except that it originates on a N atom or is the protonating H. As mentioned above the spectrum of N-acetylated VGVAPG is noticeably different. The relative intensities of the C-terminal ions, most notably yn ion peaks, are reduced dramatically from those of the underivatized peptide while the relative abundances of the N-terminal ions remained the same. The only difference between the acetylated and normal peptide is that the primary amine is changed to a less basic amine. This suggests that the N-terminal free amine can no longer compete for the proton during ionization and that the carboxyl oxygens could be theonly groups protonated. Thus, the formation of bn ions may indeed involve a 1,2-elimination reaction involving the protonation of carboxyl oxygens Figure 3.13. This would imply that the bn and yn ions are formed through different intermediates and their formation depends on the site 144 4. R1 0 R2 Oj—lj R3 0 H2N—CH—g—NH—CH—gIfiH—CH—g—OH l R1 0 R2 0+ , R3 0 H2N—éH—g —NH—CH—g + NHQ—éH—g —OH bn ion C-terminal neutral Figure 3.13: Formation of the bn ion from protonation of the amide oxygen. 145 of protonation. Another explanation may be that the primary amine is the major source for the hydrogen transferred during the formation of the yn fragment ions. However, the range of the primary amine must be limited by the peptide's ability to wrap around since the y1 fragment ion is never detected. N. Fragmentation of Charge-Localized Derivatives A number of charge-localized derivatives of peptides have been described in the literature (14, 17, 19, 27-31). In order to investigate the remote- site fragmentations, CAD data are presented for peptides with a positive charge on either the N- or C-terminus of a peptide by derivatizing the peptide with an ethyl-triphenylphosphonium moiety. The goal in this comparison is to find similarities and differences in the chemistry that accompanies skeletal bond cleavage for the protonated molecule, and from derivatives in which the charge is many skeletal atoms away from the bond undergoing cleavage. If the data suggest that fragment ions for charge-localized derivatives are due to remote- site chemistry, then similarities and differences may assist in sorting out the high-energy CAD fragmentation mechanisms of protonated molecules. Table 3.3 displays data for the N- and C-terminal derivatized peptide and the derivatized deuterated variants of the peptide. The relative intensities of the peaks corresponding to the fragment ions from the undeuterated derivatized forms are given in Table 3.2. When a charge is localized on the N-terminus, only an, b... and on backbone ions are formed. These are listed in Table 3a. The N-terminal ethyl-TPP derivative of the peptide, ¢3P+CZH4-VGVAPG-OH, will be referred to as +-VGVAPG-OH. When the positive charge is fixed at the I46 0+ 0+ 0+ 3 0+ 9. I n+ I I oQuoauuou>.$ «+ I «+ «+ «+ «+ .1 a. I I I I I 3 zoda03>.$ «+ «+ M“ «a. a. 3 I "H 3 3 0+ 9. :900<8>03>.$ 9+ 0... .2 a. 2 «+ 3 o. 3 3 9. 0+ 3 9. x925<>0>$ I I a” 3 a. 0+ 9. ”H 3 3 a. 30.0..8<>o>.$ .2. 8.3.» :0 Mn" 3» .3 3.2.8.; a: MN“ :0 3. a: 3n :0.0a<>o>.$ no on ml on v. no «3 a. «o «a «a '0 :- w. . coal—5.000 2.2.58“— .2 a: :0 03 0:. a: 2;. t... :3 a: a: .3 a: S» 3» 00260.0 0.3.2.8: \ \ \ \ \ \ \ \ \ \ \ \ \ \ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . : so: _._ _ :. _ :. @ :olwl:0_ :z 1:: z_0+: “+0.sz 21...: _o _ :H+ :z _w_:0_ ._.:0.__tzl«=0uzol Ono ”zoxwxo £035 7:... +|._+|l_tl._l>l_.@O 0352.8 35.2.2.2 05 E9: mEoEmEu and 030... 147 a. 0+ I 9. 3 9. I «+ 9+ «+ .1 .1 9+ @domouou>.: + 0+ 0+ 0+ 0+ 0+ 0+ 0+ 0+ M+ 0+ 9+ ”H 9+ 9+ 9+ 9+ $éa§>i 9+ 0+ 9+ 9+ 0+ 0+ 0+ 0+ 0+ 0+ 0+ 0+ 0+ 0+ 9+ 9+ 9+ 9+ 9+ 9+ $.2<3>§>.: . 0+ 0+ 0+ 0... 0+ 0+ 0+ 0+ 0+ 0+ 0+ 0+ 0+ 0+ 0+ 0+ 0+ 0+ 6.0003590??— 0+ 0+ 0+ 0+ 0+ 0+ 0+ 0+ 9+ 0+ 9+ 9+ 9+ 9+ 9+ 9+ 9+ $290<>0>i 9+ 9+ 9+ on» .«o 0.0 can «an 3» .3 a: .3 a: 3. 3... 3.. :0 on. :0 2.0 so. 3.. $ 9.55.. : 5:25.000 5 p» 5. a» «x as a» an ea c» 3. an a.» 0x .5502“. 00230.0 3n a: 3n «3 o: a: 3+ 0:. a: .3 :0 a: :0 I... a: +2. 02. 0.52.8: Ill/I //////////// _, __ _ __ __ _ __ _A__ _ ___ to :.:3..: :r:_.: ‘ .u«:0.«:0.:¥+m+:m_ :2 _MT :0Iz _:0+..:. _w_0 :w :z _o w__.0.9:z_ ._..._0._.«:z :0xwz0 0 a < > 0 > 02.02.00 32.5.3.0 Ea: 0.9.0802... 50.0 030... 148 C-terminus, only x... y... and z" backbone fragment ions are formed. These are listed in Table 3.3b. The amino ethyl-TPP derivative is used to fix the charge at the C-terminus; the ionic derivative H-VGVAPG-NHCZH2P+¢3 will be referred to as H-VGVAPG-+. The data, concerning H-shifts, for the fragmentation of the charge- localized derivatives are evaluated. Comparison of the ionic structures of the fragments from the derivatized peptides with those obtained from protonated peptides will confirm whether the protonated peptides fragment through charge- initiated or remote-site processes. a- and x-type lons: From a comparison of the mass of molecular fragments formed by simple homolytic cleavage of the skeletal bond C-C(O) bond, it is obvious that the N- terminal derivatives generate a ions which are one mass unit lower than that expected for a simple homolytic cleavage. The CAD spectrum of the +- VGVAPG-OH displays an fragment ions corresponding to all C-C(O) cleavages in the molecule with a hydrogen shift away from the charged fragment. The an fragment ions are the most abundant in the N-terminal derivatized peptides. Also, there is essentially no signal for the a2 fragment ion corresponding to fragmentation at the glycine residue. This has been noticed in other derivatized peptides containing the glycine residue (29). The expelled hydrogen atoms can originate from 3 possible sites. The first being on the a—carbon, in which case one would observe a mass shift in the spectra of the peptide derivative which contains the deuterium-label on the a-carbon. The second possible site would be on the B-carbon or the sidechain of the amino acid involved in the cleavage which would be indicated by the mass shifts in the mass spectra of the derivatized peptide analogues containing the deuterium labels on the B- I49 carbons. The third possible location is the amide nitrogen which would be indicated by the mass shifts in the mass spectra of the derivatives of the deuterated analogue which has deuterons replacing all exchangeable hydrogens. The mass spectra of the +-VGVAadPG peptide derivative displayed a peak for 34 ion at mlz 588, which is 1 Dalton higher than that observed for the unlabeled peptide. Likewise, the spectrum of the derivatized +-V°‘dGVadAPG analogue displays the 31 (mlz 361) and 83 (m/z 518) corresponding to retention of the deuterated labels on the detected ions. Thus, a-hydrogen is not lost in the formation of the an fragment ions. In contrast, the CAD mass spectra of +-VGVAB3dPG and the +-VBdGVBdAPG exhibit the loss of the deuterium label for the same an ions. These observations prove that formation of the an ion involves the expulsion of the hydrogen located on the B-carbon. For example, consider the 01 ion. The unlabeled derivative, +-VGVAPG-OH, produces an 31 which can be designated as [-i--NHCH(CSH7);u]. When the B-position of the valine is deuterated, a mass shift of 0 is observed, clearly indicating that the D shifts. Thus, 81 from +-VBdGVBdAPG indicates that the shifting H comes from the side chain, and 31 has the form +-NHCH=C(CH3)2] The same conclusion can be made for the as and a4 fragment ions from the data in Table 3a. The case for the an ion is different, since there is not a side chain in the glycine residue as a source for the transferred H atom. The en ions formed at glycine residues are always of very small abundance or absent when formed from N-terminal ethyl-TPP derivatized peptides (29), The data do not suggest the source of the H-shift for the 32 ion. The formation of the as ion from cleavage of the proline C-C(O) bond also displays that a H is lost, and one is available in the B-position in the proline ring. 150 Thus, it seems that a 1,2 elimination mechanism is occurring. This mechanism is represented in Figure 3.14, suggesting a 1,2-elimination reaction. Also, this proposed mechanism explains the absence or very low abundance of an ions for all cleavages occurring at a glycine residue, as it does not contain a hydrogen-bearing carbon sidechain. The C-terminal derivative H-VGVAPG-+ generates abundant Xn ions. The X" ions derive from cleavage of the same bond as do the an ions except here the charge is localized on the C-terminus. lf skeletal bonds are cleaved in these triphenyl phosphonium (T PP) derivatives by chemistry that is unaffected by the charge, then the N-terminal derivative should produce an Xn ion that gains a hydrogen in the fragmentation process since formation of the an ion shifts a H away from the N-terminal fragment, as illustrated in Figure 3.13. However, Xn ions represent skeletal bond cleavages minus one H atom, thus a hydrogen shifts away from the C-terminal fragment. The CAD mass spectrum of the H-VGVAGdPG-+ displays the a peak at m/z 558 representing the X3 ion which corresponds to a mass shift of +1. Similarly, the CAD spectrum of H-VadGVadAPG-+ displays a peak representing X4 at mlz 657, which corresponds to a mass shift of +1. Since both deuterated analogues of the derivatized peptide yield Xn ions containing the deuterium label, then the a-H is not shifted during the fragmentation process. The spectra of the H-VGVAfl3dPG-+ shows the X3 ion (m/z 560) with a mass shift of +3, thus all three deuterium labels are present. Furthermore, the CAD spectrum of H-VBdGVBdAPG-+ exhibits the X4 ion (m/z 657), corresponding to a mass shift of +1, thus containing the deuterium label. Therefore, the [3 carbon’s hydrogen is not shifted from the C-terminal fragment during bond 151 @ECHz-CH2V(NH- -CH-fi)fl—rfl-:’H—q —NH-CH-g—NH-:H-§—°H l , . ©—©—CH2—CH2-(NH-CH—fi);‘NH—CH + Hg-HH—EH—fi-NH-cH-g-ou 7 R3 (8'1) ,, lon neutral Figure 3.14: Formation of an-1 ion by charge-remote fragmentation involving a 1 ,2-elimination reaction. l 52 cleavage. The CAD mass spectrum of the D-VdeVdAdPGd-=‘i= displays a peak for the X3 ion at mlz 559. The mass shifts by +2, indicating that a D is lost from the C-terminal fragment upon cleavage of the C-C(O) bond. This suggests that the fragmentation could involve a 1,2-elimination (Figure 3.15). There are two possible explanations for generating these a and X ions from the charge-localized derivatives. The first is that the chemistry does not represent true remote site chemistry; secondary interactions bring the charged terminus into close proximity with the bond being cleaved. Thus, the direction of the H-shift (1 ,2-eliminations) is controlled by the end of the peptide that carries the charge. A more reasonable explanation is that the fragmentation chemistry is remote-site and the mechanism shown in Figure 3.16 is consistent with the remote-site fragmentation in the hydrocarbon chains of fatty acids as reported by Gross et al. (25-26). When the C-C(O) bond cleaves, a 6-membered ring is formed, then both fragments lose a H atom, thus yielding 3 products: a C- terminal fragment, an N-terminal fragment, and an H2 molecule. Therefore, the a and X ions from these derivatives are formed through a common remote-site mechanism and the detected fragment would be the one which contains the localized charge. The mechanism in Figure 3.16, involving the loss of a H atom from a side chain and an H(N), may not be the only mechanism for forming X ions. The labeling studies determined that an exchangeable H is lost, but not its exact location. The data contain a small peak representing the X2 ion, involving the proline residue which does not contain a H atom on the amide N. Regardless, the data indicate that an exchangeable H is lost in the fragmentation process. A possible pathway for the formation of the X2 ion is shown in Figure 3.17. 153 MHz-CH wfljéEE—ir—g—(HH—EL—fizjuH—cnz—curg—Q Hm—EH—g—NH—EZHZ + mH-ZL—g4NJH-EATH4HrcHrEfig—Q neutral (X-1) n lon Figure 3.15: Formation of xn-1 ions by charge-remote fragmentaion involving a 1 ,2-elimination reaction. 154 7b...cUH o 0 II II —(—NHI —CH-C O-)—NHH- JD gfi-cfH—c— —NH —CH—C_. Fl‘l'l fa fb acb\/ + II II lCl TPP —CH2—CH2-(-NH —(I3H —c —NH —CH —C—NH —CH Rn n-1 R1 (a-1)n ion Ol’ ll l I + O=C==N —(|)H "O -NH -?H —C —NH—CH2—C Hz—TPP R3 R4 (x-1)n ion Figure 3.16: Formation of the a and x ions vlaacommon remote- site mechanism, where an H2 molecule lS expelled. 155 f __ CH3—fi H-VGV-NH-H fi-TD “{C- CH CH2 | . co-rfip .H2 0 CH2 /g-N H-VGV-NH-EH ~ Al I > / C-CH cl... N-terminal neutral CO'TE’P x2 ion Figure 3.17: A possible mechanism for the formation of the x2 fragment Ion. 156 b- and y-type lons: Since the data for a and x ions generated by the charge-localized derivatized peptides suggested a common remote-site mechanisms, then b and y ions could also form through a similar mechanism. However, the analysis of b and y ions led to a different conclusion. The first observation is that, while bn ions give very strong signals from the protonated peptide, they are weak or absent in the CAD spectra of +-VGVAPG-OH [this is also observed in peptides containing basic residues at the N-terminus (14)]. The formation of the bn ions does not correspond to a simple cleavage as in the underivatized peptides. All of the observed bn ions are of very low intensity. This suggests that b ions, from the protonated peptide are formed through charge-initiated chemistry. A second observation is that, while b ions from the N-terminal charge-localized derivative are of low abundance or not observed at all, the y ions are always observed in the CAD spectra of C-terminally derivatized peptides. This is g contradictory to involvement of a common remote-site mechanism as for alX ions. The b ions observed in Table 3.3a represent N-terminal fragments minus one H and, to a lesser extent, plus one H. The fragmentation of the skeletal bond is always accompanied by a H-shift. . The data in Table 3.3a show that the b1 ion is never formed. The b2 and b4 fragment ions display peaks at mlz values which are one maSs unit lower than in the case where a simple bond cleavage were occurring as for the underivatized peptide, thus indicating the loss of a hydrogen in their formation. The +-VGVAadPG-OH generates the b4 ion at mlz 615 and the +- VadGVadAPG-OH peptide exhibits the b3 ion at mlz 546, which suggests that the deuterium label is lost in the fragmentation process. The CAD mass spectra of +-VGVA°‘dPG-OH and +-V°‘dGV°‘dAPG-OH display bn ions with all the deuterium labels present, thus eliminating the alkyl group as the source for the 157 expelled hydrogen. The CAD spectrum of +-VdeVdAdPGd-OD displays a peak for the b4 ion at mlz 619. The increase in four mass units corresponds to the four exchangeable sites on the fragment ion. Thus, the hydrogen lost in the formation of b4 does not originate an acidic site. The data suggest that the or- hydrogen is lost in the fragmentation process. The data in Table 3.3a do not indicate the origin of the additional hydrogen to the N-terminal fragment in the formation of bn+1. The yn ions have the same bond broken as in the formation of the bn ions, thus we would expect the yn ions to gain a H in the fragmentation process. However, all of the yn ions displayed peaks corresponding to the loss of a hydrogen in the fragmentation process. However, in the case of the yn ions, the data in Table 3.3b clearly indicate the source of the shifting H. The yn ions were slightly dominant ions in the mass spectrum. The CAD mass spectrum of the derivatized peptide yields a ya peak at m/z 529 and y4 at m/z 628. The CAD mass spectrum of H-VGVAadPG-+ displays a peak for ya at mlz 529 and the mass spectrum of H-VadeadAPG-+ displays a peak for w ions at mlz 628, both are identical to data from the unlabeled derivatized peptide. This indicates that the (II-hydrogen is eXpelled in the fragmentation process. The CAD mass spectra of H-VGVAB3dPG-+ and H-VBdGVBdAPG-+ display peaks at m/z 532 and 629, respectively, indicating that the deuterium labels on the B-carbons were still present, thus the formation of the yn ions do not involve the expulsion of the B-hydrogens. The CAD mass spectrum of D- VdeVdAdPGd-+ exhibited a peak for ya at mlz 532, a mass shift of +3, corresponding to the three exchangeable hydrogens present on the fragment ion. This is consistent with the loss of the a-hydrogen to form the y3 fragment ion. I58 The formation of bn-1 and yn-1 ions could be analogous to the mechanism proposed for the a/X ions, involving skeletal bond cleavage with loss of a H2 molecule as shown in Figure 3.18, or the mechanism could involve a 1,2-elimination mechanism to form the bn+1 and yn-1 fragment ions as illustrated by Figure 3.19. In addition to y-1 ions, the ethyl-TPP derivatized peptides also yield two y ions, ya and y5, that gain one H in the fragmentation process. Thus giving evidence for multiple pathways for the fragmentation process. The H atom was determined to originate from an NH group, similar to that in the fragmentation processes proposed for protonated peptides. c- and z-type Ions: The c and 2 ions in Table 3.3 aresimilar to b and y ions, in that the 0 ions from N-terminal TPP derivatives are formed in very low abundance, while the 2 ions from the C-terminal derivatized peptides are easily detected. While the c and 2 ions are formed by cleavage of the same skeletal bond, -NH-CH(R)- , they are clearly not formed through a common remote-site mechanism, suggesting that the charge-localized derivatives may not consistently provide- data on the cleavage of skeletal bonds of peptides unaffected by the charge- site. The data for theN-terminal ethyl-TPP derivatives in Table 3.3a indicate the formation of on ions that appear at an rer value one mass unit higher than if a simple bond cleavage were occurring, implying that the on ions gain one hydrogen in the fragmentation process. The spectrum of +-VGVAadPG-OH displays peaks for c2 and c3 fragment ions identical in mass to those from the unlabeled derivatized peptide, a mass shift of 0. Thus, the a-hydrogen is not 159 3 Bid} -N H -NH R R r H -RCI>1 "C égbgzcé’c-(WH -:|:H-C I: NH n l» TgP—CHz—CHz—NH -('3H j-NH —(|>-C-O or NH-(f j—(WH -<|3H jj —NH —CH2-CH2-T;P ' ' n-1 Rt '11 n. (B—1)n ion (Y-1)n ion Figure3.18': A remote-site fragmentation mechanism for the formation of b and y ions, involving the elimination of H2. The detected ion contains the ethyl-TPP derivative. 160 ”5.3.1. 4.4.3 %%_(MLJ)L.HWH;©@ © l NHz—éL-fi—an I-I-CI-I—CH + HH-Zn—g—(H H—ZH-B)~H H-cru—CHi—EQQ n1 bn+1 iOI'I yn_1 ion © Figure 3.19: Formation of the bn+1 and yn-1 via a common remote-site mechanism, involving a 1 ,2-elimination reaction. l6] transferred in the fragmentation process. The CAD mass spectra of +-VGVAB3dPG-OH and ='i=-VBdGVl3dAPG-OH display on ions which are identical in mass to those from the unlabeled derivatized peptide, thus the B-H is not transferred in the fragmentation process. However, the mass spectrum of +-VdeVdAdPGd-OD exhibits a c2 ion four Daltons higher and c3 ion five Daltons higher than those generated from the unlabeled derivatized peptide. The increase in mass is consistent with the number of exchangeable sites plus a deuterium shift instead of a hydrogen shift onto the on fragment ion, thus suggesting that the shifting hydrogen is from an acidic site on the peptide. The hydrogen is not required to originate from the adjacent amide nitrogen because c3 is formed next to a proline residue which does not have a hydrogen to transfer. This can be the same mechanism as proposed in the fragmentation of the underivatized peptide except here the derivatized moiety replaces the protonating hydrogen. Also notable is the presence of the (:5 ion, where the only exchangeable H, which does shift, is at the C-terminus. Thus, to form the c5 ion listed in Table 3.3a, the neutral lost is CzH202. To form the 03, if the source of the shifting H is from the N cf the terminal glycine, a cyclic peptide, containing a 6-membered ring, would be eliminated as the neutral product via a process analogous to that shown in Figure 3.20. Another similarity between the on ions in the mass spectra of the derivatized and underivatized peptides is that the relative abundance remains the same while in all other series the relative abundance changes drastically. ~ The 2" ions are analogous to the on ions in that during their formation, the same bond is broken. Thus, if the c ions and 2 ions are formed in a common remote-site mechanism, and a H shifts onto the N terminal fragment to l———- I62 lift TPP-VGV-NH- -CH- I-l/\I\I_C | :y/ CH | 2 0 CO-OH ?”3 v CH— / \/N N + [TPP-VGv-NHZ] / \C—CH ] (EH2 ll C3 COOH C-terminal neutral Figure 3 .:20 A possible mechanism for the formation of the 03 fragment Ion. 163 form 0 ions, then 2 ions must represent C-terminal fragments minus one H atom. However, the 2.. ions do not display any hydrogen shifts, instead the series appear at m/z values corresponding to homolytic bond cleavages. This is also observed for all of the deuterated derivatized analogues since all of the labels were shown to be present throughout the series. This was not expected for the TPP derivatized peptides, since all 2 ions in Table 3.3b must be odd-electron ions, formed from an even-electron precursor. The most probable form would be a distonic ion, such as -CH2C(O)NHCH2CH2TPP+ representing the 21 ion. It is unexpected that the remote-site mechanism would involve a high energy homolytic cleavage instead of a H-shift. The chemical environment for the formation of the 2.. ions are unique in that there cannot be a 1,2—elimination reaction; a H could not shift from a local site because the carbonyl group does not contain an available H atom. Therefore, the lack of a H atom that could shift, could determine that the C-N skeletal bond undergo a homolytic cleavage without an accompanying H shift. Another less plausible structure for the z-type ions from the TPP- derivatives is one that shifts a phenyl group from the derivative to the C involved in bond cleavage. The 21 ion would then be more accurately written as [¢-CH20(O)NHCH20H2P¢2]+-, again a radical cation, possibly containing the charge and the unpaired electron on the N atom. The formation of this sthcture requires secondary interactions, where the C-terminal TPP group interacts with an amide linkage bringing the phenyl group close to the bond being cleaved. The phenyl group could then shift off the P atom and onto the C. 164 V. Comparison of the Charge-Directed and Remote-Site Fragmentations: a and X Ions: The labeling studies of the protonated peptide did not determine whether a ions had the structure a1 or an, shown in Figure 3.1. If a H shift does occur in the fragmentation, then it is an exchangeable H atom. Most likely, the fragmentation is a simple cleavage, where the ion expels CO from the bn ions. The a-type ions formed from the N-terminal charge-localized derivatives involves the loss of a H from the B-carbon located on the side chain of the residue. The deuterated variants did not determine whether a H shift occurred in the formation of x ions generated form the protonated peptide, however, it was determined that the C-terminal derivatized peptides yielded xn ions which lost an exchangeable H atom. This suggests that CAD of the protonated molecule generates a and X ions by separate charge-directed pathways, while the charge-localized derivatives of the peptide undergo a common remote-site reaction, shown by Figure 3.16, where H2 is eliminated. b and y Ions: The labeling studies suggest that b ions are simple fragments from the protonated molecule, while the formation of the y ions involves a H transfer from an exchangeable site on the neutral N-terminal fragment. The difference between remote-site and charge-directed fragmentations can be used to explain the low abundance of the bn-type ions. Strong b ion abundances were observed for the underivatized peptide in contrast to the very weak abundances observed for fragment ions from the N-terminally derivatized peptides. Thus, in order to form the bn ion, the protonated charge must be located near or at the 165 amide nitrogen (most likely on the amide O), in order for the bond cleavage to take place. When the charge is localized away from the amide nitrogen, then the bn ions do not form. The formation of the y ion from a protonated peptide involves the addition of two H atoms, however, the C-terminal derivatized peptides generate y ions that lose one H atom in the fragmentation process. Thus, protonation of an amide group induces the H-shift onto the C-terminal fragment, and the formation of both y and b ions generated from protonated peptides involves charge-initiated chemistry, although the y+1 ions observed for derivatized peptides suggests that a remote-site mechanism could also be occurring. c and z Ions: The CAD spectra of both protonated peptides and N-terminally derivatized peptides produce low abundance c ions which gain an exchangeable H in the fragmentation process. The 2 ions are one of the most interesting series. While z+1 ions are observed from the protonated peptide, . the C-terminal derivatized peptides produce 2 ions corresponding to a homolytic cleavage, both of which are radical cations. Obviously, some additional information is required to determine the chemistry associated with the formation of this ion series. 166 VI. Thelnfluence oinghly Basic Reddueeoan-agmentation of ProtonatedPeptidee The mechanistic studies of the fragmentation chemistry of peptides can be elucidated from experiments such as those presented above, but only a small number of chemical environments have been examined, and similar studies with peptides containing other residues must follow. The mechanisms suggested here should be operative in a variety of peptides. One case where variations may occur is in peptides containing basic residues. It has been suggested that peptides containing basic residues are preferentially protonated at the basic site of the peptide, and that the subsequent fragmentation of such peptides are through remote-site mechanisms. If these observations actually do reflect a preferred site of protonation, one might expect the mechanisms discussed here for fragmentation of protonated peptides to be changed. For example, consider the an ion series. The CAD mass spectrum of the protonated peptide H-VGVAB3dPG-OH , Figure 3.21a,'exhibits two peaks at m/z 302 and 303, corresponding to the a4 and an fragment ions: [H-VGVAB3dPG-OH]H+ ----> [H-VGV—NH-CH(CDa)]+ a4 [H-VGVAB3dPG-OH]H+ ---> [H-VGV-NH-CH(CDa)+H]+ a4+1 These fragmentations are from the charge-directed chemistry. However, CAD on the N-terminal TPP derivatized peptide H-VGVAB3dPG-OH forms an alt-type ion which loses a H atom, (H2 is eliminated in the process): +-VGVA133dPG-OH ----> +-VGV-NH-CH=CDz 84-1 167 To test the influence of basic residues, the peptide H-RVGVANGPG-OH was synthesized. The CAD mass spectrum of the protonated molecule, Figure 2.21 b, exhibits three peaks corresponding to the a4 ion fragments: ‘ [H-HVGVABadPG-OH1H+ ---> [H-RVGV-NH-CH(CDa)]+ a4 [H-RVGVAB3dPG-OH]H+ ----> [H-RVGV-NH-CH(CDa)+H]+ a4+1 [H-RVGVABSdPG-OHlm ----> [H-RVGV—NH-CH=CDZ]H+ a4-1 Thus, the product a4-1 at m/z 457 is formed in a remote-site process, from the protonated peptides which contain the protonating H on the arginine residue. While both charge-induced and remote-site reactions occur here, clearly those which contribute to the CAD spectrum of other protonated peptides will depend on whether basic residues are present. In this case, the a ions provide clear evidence that the site of protonation can be localized to a substantial extent on basic residues, although they are not the exclusive site of protonation; thus more than one mechanism is operative for the formation cf the an fragment ions. The structures of the an ions formed are shown in figure 2.22. The use of deuterium labeling is evaluated here for investigating the formation of sequence-specific fragment ions from protonated peptide molecules. .A comparison of the processes leading to skeletal bond cleavages in protonated H-VGVAPG-OH, its N- and C-terminal TPPderivatives (charge- localized on that terminus) and deuterium-labeled variants of these species, provides a useful context for investigating the chemistry through which fragment ions are formed. Skeletal bond cleavages are frequently accompanied by H- shifts. The direction of the H-shift, and the source of the H-atom, as suggested by labeling studies, distinguishes charge-induced and remote-site processes. It 168 a4 ions from H—VGVA133dPG-OH 1co- ' a 1 1 :3 3% ‘4 §l (M+H)'D 84+“ “1 1 _ A-..“ A A A -- o1 ................. - . - - . - - ' 296 see 300 302 304 306 303 310 m 35 ion from H-RVGVAB3dPG-OH 1OOI b : a5+H i 1 ‘5 5w ° 1 H D 51 (35+ 1' N '51 K 0 vvvvvvvvvvvvvvvvvvvvvvvvvvvvv 440 452 454 456 453 460 462 454 mrz Flgure 3.21 a-b: The FAB-CAD-B/E mass spectra of a) protonated peptide H-VGVA133dPG-OH displaying the a4 ion, and b) protonated peptide H-HVGVABSdPG-OH displaying the a5 ion. 169 H+ l-R10Rx 0 _‘fn RIO Rx 0 + [in an: HZN- -CH- 3- -(NH -CH- -C)_ -NH=CH H+ l I}. o I}. I fiRz (an+H)-H: HzN-CH-E-(NH-CH-C)n_1-NH-CH Figure 3.22: Structures of the various an ions that form. 170 appears that most of the fragment ions formed by CAD of protonated molecule follow charge-induced processes, although remote-site processes occur when highly-basic residues are present. When H atoms shift from skeletal fragments of the protonated molecule, cyclic products result. In contrast, charge-localized derivatives of peptides appear to fragment largely through a mechanism which involves elimination of H2. Since the masses of the fragments observed in the spectra of ethyl-TPP- derivatized peptides differ in mass from the corresponding fragment ions observed in the spectra of underivatized peptides by one to two mass units (aside from the difference in mass due to the derivatizing moiety and the absence of protonation as the derivative contains a fixed charge), then it seems advisable to clarify nomenclature for the peptide fragmentation. For example, Roeptstorff and Fohlman designated the Yr. ion to represent the following theoretical structure resulting from cleavage between the amide nitrogen and carbonyl: _O 11 II _ N-CH-C— NH—CH—C —OH = Yn | I n-1 n, R . Further, Roepstorff and Fohlman used the symbol Yn" to represent the species actually detected; the two primes indicate the addition of two H atoms to the Y" species, one from protonation by the FAB ionization process and the other by a hydrogen transfer, which induces cleavage at the peptide bond as follows: 0 O u "1.1.2H—3—(HH-cH—e' —OI-I E Y" K II? I; n-1 1 ‘r‘ Anni." 171 Subsequently, Biemann simplified the nomenclature with the symbol yn to represent the species actually detected, thus, yn=Yn'i. The ethyl-TPP derivatives of peptides generate a fragment ion series represented by Yn-1 according to the Roepstorff nomenclature, or yn-3 to be consistent with the established nomenclature as modified by Biemann. Huge—E NH—(Izl-I-(I? n-I-IHH-cHz-CHzip-Q E yn-3 RI‘I Additionally the C-terminally derivatized peptides generate xn-1, zn-1, wn-1, and vn-1 fragment ions and the N-terminally derivatized peptides yield an-f, b"- 1, cn-1, and dn-1 fragment ions using the nomenclature suggested by Biemann. Since the application of charged derivatives is increasing, and the use of the current nomenclature for the fragment ions produced from charge localized derivatives is confusing, then a modified version of the existing nomenclature should be implemented to describe the fragment ions generated by peptides containing a charged-derivative. I propose that the fragment ions generated from charge localized derivatives have the Biemann's nomenclature accompanied with an asterisk. For example the yn-3 ion would be designated as yn‘. Thus, the asterisk easily distinguishes the fragment ions produced by derivatized peptides from the fragment ions generated by protonated peptides, while the rest of the ions designation is consistent with the commonly used nomenclature, to avoid additional confusion. The surmised generic structures of the fragment ions produced from protonated peptides involving charge-directed fragmentation are given in Figure 3.23. The structures of the fragment ions produced by remote-site fragmentation of the N- and C- terminally derivatized peptides are shown in 172 Figure 3.24, and introduce the proposed nomenclature for the fragment ions produced from peptides containing a charged derivative. Fix an H—(—NH-t‘!H-CO-)n_1-I¢H=CH an Rn A I H-(-NH- H—co.-),,.,-NH-CH-c =26 bn H-(-NH-(£H-CO-)n.1-NH2 Cn H+ I CHHn H- -(- -NH-ZH-CO-)M-NH- H dn 173 Rn (Sac-NH— H-fi-k NH—EH-Ifi-m—OH Xn . H+ x H-(-NH—CH-fi-)n-OI-li Yn R\ [Rb H+ fi-fi-I(-NH-6H-3-)M-OI-ll Ray iH—E—L- NH-EH-fi-m -OI-ll Figure 3.23: Structures of the fragment ions produced from protonated peptides. . R R R {c/b X TEF—(—NH-tj:H—co-)M-NH—uou>.e «+ x «+ «1 M” «+ .w 7. .w 7 ._. t. x * +..=:o.oe<3>o3>.z_ «+ x «+ .9... M“ «a. x «+ .w I .w I * ._n 1450853563.... 3 ._. n1 2 M” 9. x 3 a 9. ._. 9+ * * +3.:oess3<>o>.:_ p... p+ r+ * "H o... * o... * 0+ * 0+ * * +..==0.0aua<>0>.z~ e8 h: an? «0' «an no an no man Zuaosguaessusaa :3 :3: 8: :- to no as no «u «a «a —0 p0 Pa 0: Fat 00¢ nan can aha now an“ 05.. top cop «up hop as aov fat non sun can z.“ or.“ on“ «hp saw our \\\ ziwlxokzzhm+x _ __e\/___ __ :.+w......_...+w. \\\\\\\\ :0\ __ tar; ___ \/ +..=:o.os<>o>.z_ 8 season 98 nausea—non Essa-E +3 + 826.6 egress: 53,-0.0 0.52.3: 347.258.. ._e 96 as: 22 3525.2 a: 2.3 187 n+ ++ e+ + x a." x «+ *«+ n+ «+ n+ .1 0+ * w ++ 14306233536. .7 xxx u+ *3 s+a+.+.+.+ 2" 9+ 3 H P+ +:_:o.osoa>i_ * x x “H * o+ ”H o+ .+ ”H .+ 2 I I +_.__:o.oeous>.:_ O... x .w * o+ .w n+ M” n+ 9+ M” n+ n+ M“ 9+ +..=.._o.o%8<>e>.£ .w x x “H .w 3 ”H .+ .+ "H .+ I "H .+ +_.=:o.osus<>o>.a 3833.3: 2:33 +2 a..." 2a 8» e3 2.» 3. e3 «9 3.5 +35. F" F» «x «a «x as 9» ax en o» vx nu ma ax .3:-:23 .5802... so 3 2: 2.. no... .2 .3 E «an 3n o3 can use an. +3 + 8963 «areas: 3 1. «2 2.3252 «can: can :9 .8 8» 233.. 8.23.0 2:328. /// JJJJJ/JJJJJ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ >_._. _ _ _ _ _ _ _ _ _ _ :olmlzohzzhmkzol Iz._.m._. ._..._L+.z 2:1. ohzzh.._.xo._.zz._._.r:otu:z z. 1.0. :15... _o_ ._._ _o_.:_U azoxnyo ezoxayo 2:22.83 “_e 93 52. 2.2 35.528 3.3 22¢ 188 I11 . I" 91.94. L. 5+ 9.9 m... 9.9 m+ 0... 9* ¢+ * * 9.9 9+ 9* 9-9 +aznootuanuu0u>ta N... 9.9 N... N+ M” N... * 9+ 9.9 p... * 9+ * * +a2—I0.0Qth>.-.= 9+ .9 9+ 9+ NH 9+ .9 9+ .9 9+ .9 9+ .9 .9 +9z_:o.09<99>099>.:_ 9+ .9 9+ 9+ M” 9+ .9 9+ .9 9+ .9 9+ .9 .9 +9z_:o.09999<>0>.:_ 9+ 9+ 9+ .9 "H 9+ .9 9+ .9 9+ .9 9+ .9 .9 +9z_:o.0999<>0>.:_ 999 92.:o.09<>0>.:_ 9 999 999 9 9 .99: 9 + 999 9 9 999 .99 9.99.999. A 99:99. 39:99. A 9. .9 30 no no no to 9a an an an N0 Na Na 90 .5 pa 0030:2990 Eon—92". Nov hvv a: can NNN VON ohN me new 00.. Nap our «Npuo .32 + 00260.0 0:39:01 1 999 999 999 999 999 2.9 999 999 9: 99.9 999 9: 99.99 99.9.9.0 9.99.299: \\\ \\\\\\\\\\\ _ _ _ _ _ _ _ _ _ _ _ _ :olml:0._.:z._.w.“.:0 _0\l/_:0.—:0._.:z.". .T 2: _ _ w _ "_9 _ _ 0 9:0. 0 :\0/ o : o_:\0/ 9:0 9:0 9:0 9:0 +0z_0o_30& 9o 93 E0: 0:0. _0:_E.09-z "0N6 030... 189 l .. M. 9+ + 9+ + .9 .9 .9 9+ .9 9+ m+ 9+ 9+ 9+ 9+ .9 ”.9 9+ +9z.0o.9099<9>909>.9. + .9 .9 .9 u+ .9 9+ “H 9+ 9+ 3 9+ 9+ "H 9+ +9z.:o.09<99>099>.:. .9 .9 .9 ”H .9 9+ “H 9+ 9+ 9”” 9+ 9+ 9+ 9+ +9z.:o.09<99>099>.:. #9 .9 .9 .9 u+ .9 9+ MM 9+ 9+ 9+ 9+ 9+ 9+ 9+ +9z.:o.09999<>0>.:. .9 .9 .9 9+ .9 9+ 9+ 9+ 9+ 3 9+ 9+ "M 9+ +9z.:o.0999<>0>.:. 9+ 9+ .99. 9.99 999 999 999 3.... .25. 89.89.39: 999.999. 999 999999 999999 999 999 999 999 .9 + 30 99 9» 9x 9» 9x 99 9» 9x 99 9» 9x 99 9» 9.. 929990.90 9:259... No 9.9 mNF var NNN cmN 99.0N NNN awn con NNN 00* «NV New .62 .9 00250.0 0:20:90... 99 99999 999999999 999999 999 999999 999999999 99998.0 2.9.99.9: /// JJ/JJ/J/J/J _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ D. _ _ _ _ _ _ _ _ _ _ _ :olwl:0._.:z.rw._.:0 :Hwh:0._.:z._. _._.:0._.:z.. _._.:0._.::._. _._.:0l9:z ._._ _9_ _o_..._e_ {it _9_ ._._ {9... 9:0\9./.0 9:0x9m0 92.999999. .9 30 .99.. 999. 355.20 59.9 9.999 190 9w 9w m... * m... o... * 9+ 9* 9* n+ 9... 9w * +¥nootumvnuu0u>ta .9 9+ 9+ .9 MM 9+ .9 9+ .9 .9 9+ 9+ .9 .9 +9.:o.0e<99>099>.:. * N... N+ * M” N+ ._9 N... 9w * p... _.+ 9w * +¥~IO+GQ<38>0u8>tza .9 9+ 9+ .9 m.“ 9+ .9 9+ .9 .9 9+ 9+ .9 .9 +9.:909999<>0>.:. .9 9+ 9+ .9 "H 9+ .9 9+ .9 .9 9+ 9+ .9 .9 +xfodsus<>0>.:. .999. 999 999.999. ”MM 999.999.999.999..999.99. 992999.399. +9.:o.0..<>0>.:. .9 80999.9 30 no on an en ea 90 an as N0 Na Na 90 90 pa 5922.00 9:09:0-9m 999 999999 999 999999 999 999 999 999999999 999999 +9. + 99998.0 099.959: 999 999999 999 999 999 999999 999 999 999999 99999 99.9.9.0 999.959: \ \ \ \ \ \ \ \ x \ \ \ \ \ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ :olwl:._.._.:z._.w.“.:0lz+w._.: 9.9.9:: ._.w._.: 9.1.: 9.th: ._..P:z._.w._.:._9l9:z o :_ _0 ._._9:0_ _0_ :0.” _m_ :_ _9.._:\0/ 9:0\9:0 9:0 9:0 +£92.99... .9 2.0 .99.. 9:9. 955.992 99.92999 ; 191 n... + + .9 .9 .9 9+ .9 9+ m+ .9 9+ 9+ .9 w+ 9+ +x.90.9099<9>909>.9. o... o... 9... 9+ .9 .9 .9 9+ .9 9+ 9+ .9 9+ 9+ 9+ 9+ 9+ 9+ +9.:o.09<99>099>.:. .9 .9 .9 9+ .9 9+ 9+ .9 9+ 9+ 9+ 9+ 9+ 9+ +x.:0.09<99>099>.:. 0+ 0+ 0+ F... + . .9 .9 .9 n+ .9 9+ mu .9 9+ M“ 9+ 9+ mu 9+ +9.:o.09999<>0>.:. .9 .9 .9 “H .9 9+ ”H .9 9+ "M 9+ 9+ "M 9+ +9.:o.0999<>0>.:. .999. 99 999 999 999 .99... 9...... .99..999..999. 99w..999. 999 999.999.999 999 999 999 999 999 + .0 990000... 98 9N 9» «x N» «x as 9» 9x 99 9» 9.. mu 9» 9.. 00.89.9900 .00...qu 99 999 999 999 999 999 999999 999 999999 999 999999 +9. + 99998.0 099.959: 99 99999 99.999999 999999 999 999999 999 999999 99993.0 2.9.9.99: /// xxx/J/JJJ/J _ _ _ 3_ _ _ _ _ . _ _ _ _ _ :ol _:w+:z+w+:0lz+w+:w+:z+ _.“.:9_.+:.z.“. _F:w+:z+w._.:wl9:z o .. o 3... o 9., ._ .. ._....» 9:0 9:0 9:0 9:0 +£92.99... .9 o<0 59.. 999. 99.5.2-0 3.9.9 9.999 192 Table 4.4: Intensities of Fragment Ions Produced from CAD on [VGVAPG]Li+ 821211.12 W mLz. 111mm [Cl -2H]Li+ 121 0.45 [yz-H]Li+ 178 0.37 [y2+H]Li+ 180 0.10 [d3]Li+ 221 0.71 [A3-H]Li+ 235 4.01 [y3-H]Li+ 249 0.30 [y3+H]Li+ 251 0.40 [x3-H]Li+ 277 0.31 [c3+H]Li+ 280 _ 0.71 [a4-15]Li+ 292 0.58 [a4-H]Li+ 306 1.76 [a4]Li+ 307 0.82 [W4]Li+ 319 0.97 [z4]Li+ 334 0.71 [y4-H]Li+ 348 0.29 [y4+H]Li+ 350 0.41 [b4+OI-1]Li+ 352 0.26 [x4-H]Li+ 376 0.71 [25]Li+ 391 0.41 [as-H]Li+ 403 1.92 [ys-H]Li+ 405 0.46 [y5+H]Li+ 407 . 0.88 [bs-H]Li+ 431 0.75 [xs-H]Li+ 433 0.38 [b5+OH]Li+ 449 0.61 [as-H]Li+ 460 1.11 [w6]]Li+ 475 0.81 [2611.9 490 2.51 193 Table 4.5: Intensities of Fragment Ions Produced from CAD on [VGVAPG]N21+ 823.1122 W mLz ' 112222112 [01 -2H]Na+ 136 0.43 [Yg'HlNa-l- 193 0,34 [y2+H]Na+ 195 0.12 [d3]Na+ 236' 0.61 [A3-H]Na+ 250 ' 3.81 [y3-H]Na+ 264 0.24 [y3+H]Na+ 266 0.43 [x3-H]Na+ 292 0.26 [c3+H]Na+ 295 0.60 [a4-15]Na+ 307 ' 0.51 [a4-H]Na+ 321 1.67 [a4]Na+ _ 322 0.77 [W4]Na+ 334 0.97 [z4]Na+ 349 0.69 [y4-H]Na+ 363 0.32 [y4+H]Na+ 365 0.38 [b4+OH]Na+ 367 0.29 [x4-H]Na+ 391 , 0.64 [zslNa+ 406 0.40 [as-HJNa+ 418 1.86 [yS-HlNa+ 420 0.49 [y5+H]Na+ 422 0.90 [c6+H]Na+ 463 0.71 ‘ [bs-HlNzu- 446 0.39 [xs-H]Na+ 448 0.66 [b5+0H]Na+ 464 1.18 [as-H]Na+ 475 0.63 [W6]]Na+ 490 0.74 [Z6]Na+ 505 2.32 194 Table 4.6: Intensities of Fragment Ions Produced from CAD on [VGVAPG]K+ 821211112 F n 11 211.: 11112231): [cl-2H]K+ 152 0.31 [yz-H]K+ 209 0.29 [y2+H]K+ 211 0.05 [d3]K+ 252 0.61 [A3-H]K+ 266 6.04 [y3-H]K+ 280 0.30 [y3+H]K+ 282 0.32 [x3-H]K+ 308 - - - [c3+H]K+ 311 0.05 [214-15 ]K+ 323 0.66 [a4-H]K+ 337 1.42 [a4]K+ 338 1.68 [W4]K+ 350 - - - [z4]K+ 365 0.76 [y4-H]K+ 379 0.30 [y4+H]K+ 381 0.22 [b 4+OH]K+ 383 0.41 [x4-H]K+ 407 0.70 [25]K+ 422 0.60 [as-H]K+ 434 2.73 [ys-H]K+ 436 0.45 [Y5+H]K+ 438 0.88 [bs-H]K+ 462 0.38 [xs-H]K+ 464 0.43 [b5+OH]K+ 480 0.99 [as-H]K+ 491 1.07 [W6]]K+ 506' 0.83 [z6JKa+ 521 5.93 195 A. [an-HJCat+ Fragment Ion Series By comparing the mass of molecular fragments formed by simple homolytic cleavage of the skeletal C-C(O) bond, it is clear that the an ions in Table 2a represent N-terminal fragments plus the sodium ion, minus one H atom. The mass shifts for a3 and 8;; fragment ions clearly indicate the source of the shifting H. For example, consider the as ion. The unlabeled sodiated peptide produces ,an a 3 ion that can be designated as [H-VG-NHCH(C3H7)-H]Na+. When the a-position of the valines are deuterium- Iabeled, the CAD spectrum displays an :13 ion with the mass of 252, a mass shift of +2, thus the on-hydrogen is not shifted during the fragmentation process. When the B-position of valine is deuterated, a mass shift of +1 is observed, thus indicating that one of the D shifts. Therefore, the as ion from [VBdGVBdAPG-OH]Na+ indicates that the shifting H comes from the side chain, and a3 has the form [H-VG-NHCH=C(CH3)2]Na+. Furthermore, CAD on [D-VdeVdAdPGd-OD]Na+ generates a peak corresponding to the as ion at an mlz 254, a +4 mass shift. The a3 fragment ion has four exchangeable sites; since the data indicates a mass shift of four, the 33 ion is formed without a shift of an exchangeable hydrogen. The same results were obtained for the 34 fragment ion. Thus. the [an-HJNa+ ion has the structure as shown in figure 4.2. Note that the £12 ion is not observed. The case is different for the a: ion, since the glycine residue does not have a sidechain as a source of the H atom. It should be noted that [an-H]Cat+ ions have very small abundance or are absent when formed from skeletal bond cleavage in glycine residues. This has also been noted by others (5). 196 Cat“? '8 5b I fix 9 ix 9 I Vi MHz—CH—c NH—CH-c uH—cl-I n. Figure 4.2: Structure of the [an-H]Cat+ fragment ion. 197 B. [xn-HJCat+ Fragment Ion Series The formation of X3] ion involves cleavage of the same bond as that cleaved in the formation of the an fragment ion, only the metal cation, is retained on the C-terminal fragment. These ions are formed in low relative abundances, however their structures were determined by the selective deuterium-labeling experiments. It is obvious that the ion retains the sodium cation and loses a H atom during formation. The X3 and X4 ion mass shifts show that the a-H and B-H atoms do not shift. The CAD mass spectrum of [D-VdeVdAdPGd-OD]Na+ display a peak representing the X3 and X4 fragment ions with mass shifts of +2 and +3, respectively. The total number of exchangeable sites for the X3 and X4 ions are four and five, respectively, indicating that a D shifts from an (-NH-) group. Thus, the [Xn-H]Cat+ ions have the structure exhibited in Figure 4.3. C. [bn-HJCat’r Fragment Ion Series The [bn-H]Na+ ion series is always formed in very low abundance compared to the bn ions formed from protonated peptides. Although it was determined that cleavage occurs at the amide bond with the loss of a H atom and that the cation is maintained on the fragment, not all of the deuterated peptides produced distinguishable peaks corresponding to these ions, thus the origin of the shifting H was not determined. 198 Figure 4.3: Structure of the [xn-H]Cat+ fragment ion. 199 D. Lyn+HJCat+ Fragment Ion Series The formation of [yn+H]Cat+ ions involves amide bond cleavage with a H-atom transfer from the N-terminal neutral fragment. Two possible mechanisms for ion formation were considered: 1) the hydrogen shifts onto the detected fragment from the a-carbon (5), 2) the hydrogen shifts onto the detected fragment from an amide nitrogen (4). The deuterated peptides determined the location of the transferred H to be from an exchangeable site (amide nitrogen). For example, the ya displays +0 mass shift for the peptides containing the a- and B—Iabeled alanine, thus the H shift is not from these sites. Furthermore, the [D-VdeVdAdPGd-OD]Na+ peptide displays y2 at a +3 mass shift. The yz fragment contains two exchangeable sites, thus the transferred H that comes from an amide nitrogen. The rest of the [yn+H]Na+ fragment ions confirm that an amide nitrogen is the source of the transferred H atom, thus yielding cyclic neutral fragments. The structure of the [yn+H]Na+ ions are shown in Figure 4.4. The formation of these fragment ions are probably similar to the fragmentation process of protonated peptides, which were determined to yield cyclic neutral N-terminal fragments (13). E. Lyn-HJCat+ Fragment Ion Series The spectra of the cationized peptides also show that, when the amide bond is cleaved, the C-terminal fragment retains the metal cation, the positive charge, and loses one H atom. The ratio of (yn-H)Cat+to (yn+H)Cat+ increases as n decreases. The H atom can be lost from one of three sites: 1) the a-carbon, 2) the B-carbon, or 3) the amide nitrogen. It can easily be seen from the data for the ya and y4 fragment ions, where the lost H atom originates. 200 Cat+ l'ix 9 ‘ix 9 I MHz—CH—c NH—CH—c OH n-1 Figure 4.4: Structure of the [yn+H]Cat"’ fragment ion. 201 For example, the peptide H-VadGVadAPG-OH displays that the y:, ion has a mass shift of +0, indicating the loss of the deuteron from the detected fragment. Moreover, the peptide H-VBdGVBdAPG-OH yields a y3 ion with a mass shift of +1, thus verifying that the D label is present. Further, the y3 ion has three exchangeable sites and CAD on the peptide D-VdeVdAdPGd-OD generates a fragment ion with a mass shift of +3, thus the H is not transferred from an exchangeable site. The structure of the [yn-H1Na+ ions is shown in Figure 4.5. F. [en-i-HJCat+ Fragment Ion Series These ions yield weak signals in the CAD spectra. Similar to the y" ions, the on ions contain the metal cation along with an additional H atom. The origin of the transferred hydrogen was determined from the deuterium-labeled peptides. When the alanine and valines are partially deuterated, the fragment ions do not exhibit deuterium-shifts, however, deuterium shifts are observed in the CAD spectrum of the [D-VdeVdAdPGd-ODlNa+. Thus, on ions have the structure of as depicted in Figure 4.6. An exchangeable H shifts from the C- terminal (neutral) fragment to the N-terminal (charged) fragment as the N-C bond cleaves. From where does the H shift? A c3 ion is observed, and the nearest N is in the proline residue, in which the nitrogen does not contain a hydrogen atom. Thus, the source of the shifting H need not be local to the site of fragmentation. To form c3, if the source of the shifting H is from the N of terminal glycine, a cyclic peptide, containing a 6-membered ring would be the neutral fragment. Also noteworthy is the fact that there is a c5 ion. In this case, the only exchangeable H atom is from the C-terminus carboxyl group. Thus, to form the es ion, the neutral lost is CzH202. 202 Cat+ 9'9 9‘ 9 ‘ I NH=C-C NH-CH-C OH n-‘l Figure 4.5: Structure of the [yn-H]Cat+ fragment ion. 203 Cat'l' 1 a, a, l o o I 3 I L1 NH2—CH- NH-CH- 1NH2 "- Figure 4.6: Structure of the [cn+H]Cat+ fragment ion. 204 G. [z,.]Cat’r Fragment Ion Series The 2,. ions represent C-terminal fragments plus the cation without a H- shift, apparently being formed by homolytic cleavage of the N-C(R) skeletal bond. All 2.. ions reported in Table 2b must be odd-electron ions, formed from an even-electron precursor. The most probable form would be a distonic ion illustrated by Figure 4.7. The chemical environment for cleavage of the N-C bond, to form a 2" ion, is unique in that there cannot be reactions such as 1,2- eliminations; a H could not shift from a local site in this way because the carbonyl group does not contain a H atom. Could the 2.. ions be formed by simple homolytic cleavage of the N-C bond? Of the three types of skeletal bonds, the C(O)-N bonds are the strongest with C-C(O) and C-N bonds being comparable, depending on the specific chemical environments in which they are found. It may well be that due to the lack of a local H that could shift in a 1,2- elimination reaction, because of the relative positions of carbonyl groups on either side, the C-N skeletal bonds would be most likely to undergo homolytic cleavage without an accompanying H shift. H. [w,JCAT+, [anCAT'3 and [v,JCAT+ Fragment Ions These ions are in greater abundance than in the CAD-B/E spectra of the protonated peptide. The structures of these cationized fragment ions were found to be similar to the structures formed from the protonated peptide and are displayed in Figure 4.8. These ions are presumed to be generated through remote-site fragmentations, mechanisms similar to those proposed for protonated peptides. Localization of the charge on the peptide by the site of 205 _ Cat+ l‘ w I 3’" 9 '1'" -CH— NH-CH-l OH Figure 4.7: Structure of the [zn]Cat+ fragment ion. 206 Cat+ Lt... .. 9).“..-9: [dn]Cat+ Cat+ 73:14. ._.... 2).} [Wu] Cat" Cat+ 1.. NH=CH-8 NH —(|3H -8 10H n. [Vn] Cat“ Figure 4.8: Structures of [dn]Cat"', [wnJCat+, and [vn]Cat+ fragment ions. ' 207 cation attachment may enhance these fragmentation processes and thus, the relative abundances of these ions. 1V. Fragmentation Mechanisms ' The objective of this study was not necessarily to disprove the mechanisms for formation of the cationized fragments that have been proposed to date, but to show thatother mechanisms are possible. In all probability several mechanisms are concurrently operative. The relative intensities of the CAD products are difficult to explain, however, we present a few possibilities. It should be noted that the structures of the cationized fragment ions given above are identical to the fragment ions produced when a positive charge is localized on either the N- or C-terminus using the ethyl-triphenylphosphonium derivatization techniques, (14-15). A study of these derivatized peptides indicate that N- and C-terminal fragment ions occurring from the cleavage of the same bond proceed through a common remote-site mechanism, rather than separate charge-directed pathways of protonated peptides. Therefore, one possibility for the fragmentation of cationized peptides is that the [peptide]Cat+ ions represent a distribution of "isomers”, with the Cat+ residing at different locations on the peptide. If the fragmentation chemistry is remote-site, unaffected by the location of the Cat+, when a skeletal bond cleaves, an equal number of N-terminal and C-terminal fragment ions would be detected if evenly distributed. This is not observed. The peaks corresponding to the aa-type ions are more intense than the x-type ions, suggesting that Na+ usually resides near the N-terminus. However, the peaks representing the y- 208 type ions are more abundant than b-type ions, which would suggest that Cat+ favors C-tenninal sites. Thus, it seems that a selective fragmentation process is occurring. Another possibility for the fragmentation of cationized peptides is that the metal cation resides at an amide linkage, stimulating the fragmentation chemistry. Therefore, once the chemistry occurs, the terminal groups of the two fragments interact with and compete for the metal cation. It has been shown that the presence of metal cations changes the actual'bond lengths of amino acids (16), thus altering the chemical environment. It has been suggested that Cat+-molecule interaction energies parallel proton affinities (17-20). Table 4.7 lists the proton affinities for the terminal groups of the peptide fragments. Since the structures of the fragment ions from cationized peptides are the same as those formed by charge-localized derivatives, then cationized peptides could fragment by analogous mechanisms. For example, when the C-C(O) bond cleaves both fragments lose a H atom. The [an-H1Cat+ and [Xn-H1Cat+ ions can form through a common fragmentation mechanism, (Figure 4.9), involving the expulsion of an H2 molecule, where the a and X ions evolve from a common intermediate: ---NH-CH=C(R)2----CAT+----O=C=N-- Once the H2 molecule is expelled, the terminal groups of the two fragments compete for the metal cation. The proton affinities in Table 4.7 show that the terminal group of the a-type ion is much larger than that of the X-type ion, suggesting that Cat+ will bond more strongly to the N-terminal fragment. This explains why the a-type ions are much more abundant than the Hype ions. Since the fragment ions produced from [peptide]Cat+ are similar to the fragment ions produced by the “charged-derivatives, the H atom lost in the 209 Table 4.7: Proton affinities of the fragment ions. Representing PA Fragment Structure (Kcal/mol) an.” MHz-CH = CH2 . 219.1 Xn-H HN = C = 0 173.3 bn-H c112 = c = 0 198.0 Vn'H NH = C(CH)2 221.0 210 formation of [bn-H1Cat+ is assumed to originate from the a-carbon. Similar to the mechanism for the an/Xn ions above, the [bu-H]Cat+ and [yn-HJCafl' fragment ions can be shown to form by a common mechanism (Figure 4.10), involving the expulsion of an H2 molecule and the formation of a common intermediate: ---(R)C=C(O)----Cat+~---NH-C(R)-- where the detected fragment retains the alkali cation. Based on the proton affinities in Table 4.7, the Cat+ should bind to the C-terminal fragment. Thus, the majorityof the ions resulting from the cleavage of the amide bond will be. y- type ions. This justifies the appearance of intense peaks corresponding to [yn-H]Cat+ ions and weak or absent peaks representing the [bu-HICah sefies. In contrast, fragmentation mechanisms proposed by others do not indicate the location of the metal cation on the peptide. They suggest that other factors such as steric effects and peptide composition control the fragmentation and notthe location of the metal cation on the peptide. The use of deuterium labeling permitted the structural determination of the fragment ions produced from CAD of cationized peptides. These structures of the cationized fragment ions are similar to those formed by charge-localized derivatives, ethyl-TPP derivatives. CAD of cationized peptides produce a series of predictable fragment ions, thus, the binding cations to peptides can be considered as a means of derivatizing the peptide in order to control the fragmentation. This capability to produce predictable fragment ion series can be useful for sequencing unknown peptides. The cation appears to initiate the fragmentation of a given bond, where the location of the metal cation is the site of fragmentation. Cationized peptides 211 r r '\/b l n H u H-(—NH—CH-(':l —Nl-l—CHmiffl a °mN-t|: H-C-(-NH-(I>H-c— -OH I 4 Rx III-1 B‘ an n r. r., \ / ° N a o .' ll '0'“ 9'0 in) 11 H-(—NH-¢i3H—éI -N.H':CH Q]: o°=c=°N—(I:H—c—(—NH_?H_C_ -OH -1 Rx m-‘l R1 an n [an-H]Na+ ion [Xn-HlNa'l' Ion Figure 4.9: The formation of the a—type and x- pe ions through a common remote site mechanism invo ving the expulsion of H2 molecule. 212 appear to fragment largely through a mechanism which involves elimination of a H2 molecule. The relative abundances of the fragment ions suggest that a competition for Ca1+ follows the bond cleavage, where the fragment with the highest proton affinity retains the metal cation, thus the positive charge. This also infers that the metal ion is not specifically bound to any one location because of the appearance of both N-terminal and C-terminal ions that contain the metal ion. Therefore, the cationized peptide is made up of various forms with the metal ion located in various positions, thus leading to the different fragment ions observed. 213 [1009 R‘ o H—(—NH—CH—c-)- NH-c " I 1111 ,c— —-NH-CH-C n—OH a ,c c NI-I\R2 8. “2 n1 ' 0 II I II II H— -NI-I—CH—C —NH—Csc-Om1Nln°-«NHstf—c —NH—CH-c —OH I M" I ”.1 3‘ Hz 8‘ Matt ll 7'??? ' '. .'