|||||l|ll|HIlllllllllllllllllllllllllllllllllll\Illlllll / 3 1293 01563 This is to certify that the dissertation entitled Structural Characterization of Glycoconjugates and Proteins by Matrix—assisted Laser Desorption Ionization Mass Spectrometry presented by Naxing Xu has been accepted towards fulfillment of the requirements for Ph .D . degree in Chemistry; / m & Major professor Date l§fl1fi4 /§?9’ U / MSU is an Affirmative Action/Equal Opportunity Institution 0-12771 LIBRARY Michigan State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE MSU Is An Affirmative Action/Equal Opportunity institution c:\clrc\ddedm.pm3—p.1 STRUCTURAL CHARACTERIZATION OF GLYCOCONJUGATES AND PROTEINS BY MATRIX-AS SISTED LASER DESORPTION IONIZATION MASS SPECTROMETRY by Naxing Xu A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Chemistry 1997 ABSTRACT STRUCTURAL CHARCACTERIZATION OF GLYCOCONJUGATES AND PROTEINS BY MATRIX-ASSISTED LASER DESORPTION IONIZATION MASS SPECTROMETRY by Naxing Xu Matrix—assisted laser desorption ionization mass spectrometry (MALDI—MS) has quickly emerged as a powerfiil analytical tool for sensitive and accurate molecular weight determination of biomolecules. In this research, 2-mercaptobenzothiazole (MBT) and its analogs have been discovered as a class of new matrices for MALDI-MS. These compounds are structurally distinct from the conventional matrices. The information gained from the matrix studies has been applied to the structural characterization of peptidoglycans, a type of complex glycoconjugates consisting bacterial cell wall with important immunological significance. 5-Chloro—2-mercaptobenzothiazo1e (CMBT) was employed for the routine MALDI-MS analysis of muropeptides derived from peptidoglycans and structural information of muropeptide was obtained by further post- source decay (P SD) analysis. Glycoprotein characterization has been a challenging topic in bioanalytical field because their complexity and microheterogeneity. To minimize sample handling between the stages of lectin affinity binding and MALDI-MS, a micro—batch lectin binding process was also developed for direct MALDI-MS analysis for isolated glycopeptides. This approach is capable of identifying glycopeptides derived from ribonuclease B and avidin, proteins having a single glycosylation site. 0‘. d.» Md '1 A new glycosylamine derivative of oligosaccharides has been developed for facile sequence analysis in MALDI-MS. 3-Aminoquinoline was used as the derivatizing regent in making glycosylamine derivative at the reducing terminus. This type of glycosylamine derivatives produces primarily protonated molecules in MALDI-MS. Reaction cleanup is not necessary because excess reagent can be used as MALDI matrix without interfering MALDI-MS analysis. PSD analysis of the protonated molecule yields a spectrum dominated by reducing terminal fragments, allowing easy determination of the oligosaccharide sequence, which is required for the complete characterization of glycoconjugates and glycoproteins. ACKNOWLEDGMENT I would like to acknowledge Dr. Zhi-Heng Huang and Dr. Douglas A. Gage for their expert assistance in organic chemistry and biochemistry and enormous encouragement they have given me during my five years at Michigan State. A special thanks to the members of the staff of the MSU-NIH Mass Spectrometry Facility: Mel Micke for fixing the computer, Melinda Beming for fixing the paper work, and Bev Chamberlin for fixing JEOL 505, morning-coffee and delicious cakes. A thank-you to Mike Davenport and Tony Sills for their help with the ion-trap and Finnigan 4600 instruments, to Joe Hiller, Pao-chi Liao and Tun-1i Shen for sharing their experience on MALDI-MS. Thanks to Jian Wang, Huanqin Wu, and Kuruppu Dharmasiri for the useful discussion and making ASMS hotel reservations. I would also like to thank other Watson group members, Jiang Wu, Ken Roth, Jennifer Johnson, Ben Gardner, Ying Yang, and Qing Li, who made my journey through through graduate school a lot easier. And finally, I would like to thank Dr. Watson for his academic advice, research support, the freedom he allowed me in the lab, and for those great salmon steaks. TABLE OF CONTENTS List of Figures .......................................................... vii List of Tables .......................................................... xiii Key to Abbreviations .................................................... xiv CHAPTER 1. INTRODUCTION ............................................ 1 1. Recent Advances of Mass Spectrometry of Biological Molecules ................ l 2. Matrix-assisted Laser Desorption/Ionization Mass Spectrometry ................ 2 2.1 Operating Principles ............................................. 3 2.2 Sample Preparation .............................................. 4 2.3 Post-source Decay (PSD) ......................................... 8 2.4 Delayed Extraction (DE) ......................................... 11 3. Glycoconjugate and Protein Analysis by MALDI-MS ........................ 14 3.1 Status of Glycoconjugate Analysis ................................. 14 3.2 Status of Protein Analysis ........................................ 15 References ............................................................. 18 CHAPTER 2. Mercaptobenzothiazoles: A New Class of Matrices for Laser Desorption Ionization Mass Spectrometry ................................. 22 1. Introduction ......................................................... 22 2. Experimental ........................................................ 27 3. Results and Discussion ................................................ 29 3.1 Peptides and Proteins ........................................... 29 3.2 Carbohydrates ................................................. 37 3.3 Synthetic Polymers ............................................. 47 4. Conclusions ......................................................... 49 References ............................................................. 50 CHAPTER 3. Structural Characterization of Peptidoglycan Muropeptides by MALDI—MS and Post-Source Decay Analysis ............................... 53 1. Introduction ......................................................... 53 2' Experimental ........................................................ 56 3. Results and Discussion ................................................ 589 3.1 Molecular Weight Determination .................................. 25 3.2 Analysis by MALDI-PSD ................... . ..................... 70 3'3 LI’SOS’taphin Digestion and Mass Mapping Analy51s ................... 77 4- Conclusions ......................................................... References ............................................................. 78 CHAPTER 4. Structural Characterization of Glycoproteins by MALDI-MS .......... :1 1- Introduction ......................................................... 81 1'1 Glycoproteins .................................................. \‘ ”(a 1.2 Glycoproteins Analysis .......................................... 86 1.3 Fractionation and Structural Assessment of Glycoconjugates by Lectin Immobilized Column Chromatography .......................... 88 2. Experimental ........................................................ 93 3. Results and Discussion ................................................ 96 3.1 Structural Characterization of Porcine Oocyte Zona Pellucida Glycoprotein 30L (ZP3 or) ...................................... 96 3.2 A Micro-batch Lectin Affinity Binding Process for Direct MALDI-MS Analysis .................................................. 109 4. Conclusions ........................................................ 120 References ............................................................ 121 CHAPTER 5. A New Glycosylamine Derivative for Sequence Analysis of Oligosaccharides by MALDI-PSD-MS and ESl-MS ............ 125 1. Introduction ........................................................ 125 2. Experimental ....................................................... 129 3. Results and Discussion ............................................... 132 4. Conclusions ........................................................ 146 References ............................................................ 147 CHAPTER 6. Sinnmary ................................................. 150 References ............................................................ 1 54 vi Ppmnlnvmflm Diltnyulmim: List of Figures Figure 1.1 Voyage-Elite MALDI time-of—flight mass spectrometer .................... 5 Figure 1.2 Differentiation of PSD fragment ion, MS, from the intact ionized molecule MH+ by using the reflector. M2 is the neutral dissociated from MH+. With no potential applied to the reflector (top pannel), M: and M2 maintain the initial velocity of MH+ and reach the linear detector, L, simultaneously. With potential applied to the reflector (low pannel), MIJr is reflected to the reflectron detector, R, and recorded while M: and M2 still hit the linear detector ........ 10 Figure 1.2 Comparison of MALDI mass spectra of ACTH (18-39) obtained in linear mode without DE (Linear), linear mode with DE (Linear DE), and reflectron mode with DB (reflectron DE) showing the effect of DE on mass resolution of TOP—MS ............................. ' .......... 10 Figure 2.1 Light micrographs (magnification: x14) obtained from samples of bovine pancreatic insulin in different matrices (molar ratio of insulin to matrix: 1:1000). top left panel: MBT; top right panel: CMBT; (c) bottom left panel: DHB; bottom right panel: orCHCA ..................... 30 Figure 2.2 Positive MALDI mass spectra of three proteins with MBT as matrix. (a) human angiotensin I (MW 1,296.5 Da); (b) porcine pancreastatin (MW 5,103 Da); (c) bovine transferrin (MW 78000 Da). The data were aquired by summing 30 to 100 laser shots. The amount of proteins loaded on the sample plate was about 1 to 5 picomoles .................... 32 Figure 2.3 Negative ion MALDI spectrum of bovine serum albumin (MW 66,430 Da) with MBT as matrix; 200 laser shots summed. Total protein loaded: 5 pmol .................................................... 34 FiEllie 2.4 Positive MALDI spectrum of horse skeletal muscle myoglobin (10 pmol); 100 laser shots summed. The sample was dissolved in acetonitrile/0.l% (v/v) aqueous trifluoroacetic acid (1 :1) solution. (a) no sodium dodecylsulfate (SDS) added, matrix: MBT; (b)with the addition of 0.1% SDS, matrix: MBT; (c) with the addition of 10% SDS, matrix: MBT ..................................................... 35 Figure 2.5 Positive MALDI spectrum of bovine pancreatic insulin (MW 5,733.5 Da) using CMBT as matrix; 30 laser shots summed. The region of the spectrum shown corresponds to a mass range of5,500-6, 150. Total protein loaded: 2 pmol ............................................. 38 Figure 2.6 Positive MALDI spectrum of oligomannnose-type N-linked oligosaccharide [ManHGlcNAch (MW 1,235.1 Da); 40 laser shots summed. (a) sample loaded: 2 pmol, matrix: CMBT; (b) sample loaded: 2 pmol, matrix: DHB; (c) sample loaded: 100 fmol, matrix: CMBT .................................................... 40 Figure 2.7 Positive MALDI spectrum of dextran 5,000 with CMBT as matrix; 50 laser shots summed. Adjacent oligomer peaks in the spectrum are spaced by 162 mass units ........................................ 42 Figure 2.8 Negative MALDI spectrum of underivatized Gm, with CMBT as matrix; 40 laser shots summed. M1 and M2 are two molecular species in the sample differing by two methylene units in the long—chain base. [Ml + K -2H]' and [M2 + K -2H]’ of two molecular species are represented by dominant peaks in the spectrum .......................... 44 Figure 2.9 Structure of trimeric muropeptide derived from the peptidoglycan of S. aureus strain COL. The following abbreviations were used: G, N-acetylglucosamine; M, N-acetylmuramic acid .................. 45 Figure 2.10 Positive MALDI spectrum of trimeric muropeptide derived from peptidoglycan of S. aureus strain COL with (a) CMBT (b) ocCHCA as matrix. 40 laser shots summed. Total sample loaded was approximately 3 pmol .............................................. 46 Figure 2.11 Postive MALDI spectrum of PEG 3,350 with AMBT as matrix. 30 laser shots summed. No alkaline salts were added to the polymer solution ......................................................... 48 Figure 2.12 Positive MALDI spectrum of PEG 10,000 with AMBT as matrix. 100 laser shots summed. NaCl was added to the polymer solution ........... 48 Willie 3.1 (a) The Staphylococcus aureus bacterial cell wall peptidoglycan. (b) The repeating unit of peptidoglycan is a GlcNAc-MurNAc disaccharide whose lactyl side chain forms an amide bond with a peptide .......................................................... 54 Ifigure 3.2 Muropeptides derived from peptidoglycan of the highly methicillin-resistant Staphylococcus aureus strain COL. (a) unsubstituted monomer; (b) pentaglycine-substituted monomer (c) muropeptide oligomers (n = l, 2, 3, ...). The following abbreviations were used: GlcNAc, N-acetylglucosamine; MurNAc, N-acetylmuramic acid; Ala, alanine; iGln, isoglutamine; viii Lys, lysine; Gly, glycine ............................................ 60 Figure 3.3 Linear MALDI mass spectra of an unsubstituted monomer obtained in (a) positive and (b) negative ion mode with CMBT as matrix. The amount of analyte loaded on the sample plate was ca. 2 picomoles ........... 61 Figure 3.4 Linear MALDI mass spectra of pentaglycine-substituted monomer obtained in (a) positive and (b) negative ion mode with CMBT as matrix. The amount of analyte loaded on the sample plate was approximately 2 picomoles. Peaks labeled with an asterisk are separated by 57 mass units representing muropeptide impurities in the sample that are tetra-, tri-, di- glycine substituted ......................... 63 Figure 3.5 Composite positive ion linear MALDI mass spectra of muropeptides ranging from monomer to pentamer with CMBT as matrix. The amount of analyte loaded on the sample plate was between 1 to 5 picomoles ....................................................... 64 Figure 3.6 Positive ion MALDI—PSD mass spectrum of an unsubstituted monomer. Matrix: CMBT. Sample loading: 2 picomoles ........................... 66 Figure 3.7 Positive ion MALDI-PSD mass spectrum of a pentaglycine-substituted monomer. Matrix: CMBT; Sample loading: 2 picomoles .................. 68 Figure 3.8 Negative ion MALDI-PSD mass spectrum of a pentaglycine- substituted monomer. Matrix: CMBT; Sample loading: 2 picomoles ......... 69 Figure 3.9 (a) Negative ion MALDI mass spectrum of the lysostaphin digest of (b) muropeptide dimer 1. Matrix: CMBT. The digestion mixture was used directly for MALDI-MS analysis without further purification. The arrows in Fig. 3.9b indicate the favorable sites of hydrolysis, based on the intensities of peaks in the MALDI mass spectrum of the lysostaphin digest. Thicker arrows indicate more favorable hydrolysis sites ............................................................ 72 Figure 3.10 Negative ion MALDI mass spectrum of the lysostaphin digest of (b) muropeptide dimer II. Matrix: CMBT ............................ 75 Flgure 4.1 N-linked Oligosaccharides. (a) All N—glycosidic protein attachments occur through a [3-N-acetylglucosamino-Asn bond in which the Asn occurs in the sequence Asn-X-Ser/Thr. (b) N-linked Oligosaccharides usually have the branched (Man)3(GlcNAc)2 core structure ................. 82 Figure 4.2 Figure 4.3 Figure 4.4 Figure 4.5 Figure 4.6 Figure 4.7 Figure 4.8 Figure 4.9 Figure 4.10 Figure 4.1 1 Figure 4.12 Figure 4.13 Structures of major types of N-linked oligosacchardes. The box area encloses the core structure common to all N-linked Oligosaccharides ......... 84 Most common O—glycosidic attachments of Oligosaccharides to glycoproteins ..................................................... 85 Amino acid sequence of porcine oocyte zona pellucida Glycoprotein 3a (ZP3a) glycoproteins ........................................... 98 (a) HPLC of tryptic digest of ZP3a glycoprotein. (b) HPLC of the glycopeptide fractions from ZP3a glycoprotein after lectin (jacalin) affinity chromatography ........................................... 100 N-linked glycopeptides at Asn 220 of ZP3a glycoprotein. The MALDI-MS spectrum was obtained from a collected HPLC fraction ........ 101 O—linked glycopeptides of ZP3a glycoprotein after purification by lectin (jacalin) affinity chromatography and HPLC ...................... 102 N-linked Glycopeptides at Asn 220 of ZP3a Glycoprotein. The MALDI-MS spectrum was obtained on (a) after B-galactosidase treatment. (b) after B-galactosidase & N—acetylhcxoaminidase treatment ....................................................... 104 N-linked Glycopeptides at Asn 220 of ZP3a Glycoprotein. Observed W2 and calculated m/z were assigned to the [M+H]+ ions of the glycopeptides .......................................... 106 N-linked Glycopeptides at Asn 333 of ZP3a Glycoprotein. (a) MALDI-MS spectrum was obtained on original HPLC fraction. (b) MALDI-MS spectrum was obtained after B-galactosidase & N—acetylhexoaminidase treatment .................................. 107 N-linked glycopeptides at Asn 333 of ZP3a glycoprotein. Observed m/z and calculated m/z were assigned to the [M+Na]+ ions of the glycopeptides .......................................... 108 N-linked glycopeptides at Asn 203 of ZP3a glycoprotein. (a) MALDI-MS spectrum obtained on original HPLC fraction. (b) MALDI-MS spectrum obtained after B—galactosidase & N—acetylhexoaminidase treatment .................................. 1 10 N-linked glycopeptides (unfucosylated structures) at Asn 203 Figure 4.14 Figure 4.15 Figure 4.16 Figure 4.17 Figure 4.18 Figure 4.19 Figure 5.1 Figure 5.2 Figure 5.3 Figure 5.4 Figure 5.5 Figure 5.6 of ZP3a glycoprotein. Observed m/z and calculated m/z were assigned to the [M+Na]+ ions of the glycopeptides ...................... 111 N-linked glycopeptides (fucosylated structures) at Asn 203 of ZP3a glycoprotein. Observed m/z and calculated m/z were assigned to the [M+Na]+ ions of the glycopeptides ...................... 112 A micro-batch lectin affinity binding process for subsequently direct MALDI-MS analysis ........................................ 113 Direct MALDI-MS analysis of tryptic digest of ribonuclease B ............ 115 MALDI-MS Analysis of Glycopeptides from the Tryptic Digest of ribonuclease B after Micro—batch Lectin (ConA) Affinity Binding .......... 117 Direct MALDI-MS analysis of tryptic digest of avidin ................... 118 MALDI-MS Analysis of Glycopeptides from the Tryptic Digest of avidin after Micro-batch Lectin (ConA) Affinity Binding ................. 119 Structures of monosaccharides commonly found in eukaryotic glycoproteins and glycolipids ....................................... 126 Types of oligosaccharide fragmentation ............................... 127 Linear positive MALDI-MS spectrum of 3-aminoquinoline (3AQ) derivatized maltohexaose (G6). The reaction mixture was used for direct MALDI-MS analysis ...................................... 134 3-Aminoquinoline glycosylamine derivative formation using maltohexaose as the model compound ................................ 135 Positive ion MALDI-PSD mass spectrum of 3AQ-G6. The [M+H]+ ion at m/z 1 1 17.4 was selected as the precursor ion. The Oligosaccharide fragment ions in all figures were labeled according to the nomenclature of Domon and Costello ................... 137 (a) Linear positive MALDI-MS spectrum of 3AQ-derivatized LS-tetrasaccharide c (LSTc). The small peak at m/z 1021.3 represented the [M+Na]Jr ion of underivatized LSTc. (b) Positive ion MALDI—PSD mass spectrum of 3AQ-LSTc with [M+H]+ ion at m/z 1125.4 selected as the precursor ion. For symbols, see Figure 5.5 ...................................................... 138 xi Figure 5.7 Positive ion MALDI-PSD mass spectrum of 3AQ-derivatized (a) LNFP-l (b) LNFP-2. The [M+H]+ ion of either derivatized isomer had same m/z value, 980.4, and were selected as the precursor ions for both analyses. For symbols, see Figure 5.5 ............. 140 Figure 5.8 Positive ion MALDI-PSD mass spectrum of a 3AQ—derivatized triantennary oligosaccharide which was derived from Cyclamen seed xyloglucan. The [M+H]+ ion at m/z 1513.5 was selected as the precursor ion. For symbols, see Figure 5.5 .......................... 141 Figure 5.9 HPLC separation of 3-aminoquinoline (3AQ)-derivatized maltohexaose (G6). on a Phase Sep Spherisorb SS amino column. The system was run from 20% to 60% water-acetonitrile in 1 ml/min with a 30 minute linear gradient program ....................... 143 Figure 5.10 Positive ion ESI-MS spectrum of HPLC separated 3AQ-G6. Cone voltage (capillary-skimmer voltage): 100 V. The spectra shown represented an average over 1 min recording time. For symbols, see Figure 5.5 ............................................ 145 Table 2.1 Table 3.1 Table 3.2 List of Tables Physical properties of MBT and its analogs ............................. 26 MALDI Mass Mapping Analysis of Lysostaphin Digest of Muropeptide Dimer I .............................................. 74 MALDI Mass Mapping Analysis of Lysostaphin Digest of Muropeptide Dimer II .............................................. 77 xiii KEY TO ABBREVIATIONS MALDI matrix-assisted laser desorption/ionization ESI electrospray ionization MS mass spectrometry TOF time-of—flight PSD post-source decay MBT 2-mercaptobenzothiazole CMBT 5-chloro-2—mercaptobenzothiazole AMBT 6-amino-2-mercaptobenzothiazole DE delayed extraction FAB fast atom bombardment KB kinetic energy MS/MS tandem mass spectrometry HPLC high performance liquid chromatography UV ultraviolet 01CHCA oc-cyano-4-hydroxycinnamic acid DHB 2,5-dihydroxybenzoic acid HABA 2-[(4-hydroxyphenyl)azo]benzoic acid 3AQ 3-aminoquinoline THF tetrahydrofuran SDS sodium dodecylsulfate SDS—PAGE sodium dodecylsulfate-polyacrylamide gel electrophoresis DTT dithiothreitol CAD collisionally activated dissociation GlcNAc N—acetylglucosamine GalNAc N-acetylgalactosamine NeuAe N-acetylneuraminic acid MurNAc N-acetylmuramic acid Asn asparagine Ser serine Th? threonine Man mannose Gal galactose Glc glucose Fuc fucose Xyl xylose Con A concanavalin A XV in Chapter 1. Introduction 1. Recent Advances of Mass Spectrometry of Biological Molecules Mass spectrometry is based on producing, differentiating, and detecting ions in the gas phase. The transfer of small molecules into the gas phase has traditionally been accomplished by thermal vaporization; however, thermal vaporization for biopolymers and other nonvolatile or thermally unstable molecules has little use. Fast atom bombardment mass spectrometry (FAB-MS), discovered in the early 805 (1-2), opened the possibility for the analytical chemists to obtain mass spectral signals directly from larger, thermally labile molecules like peptides and Oligosaccharides. But FAB-MS analysis is limited to nanomole sensitivity range and usually not amenable to biomolecules with molecular mass higher than 5,000 daltons. It was the developments of two powerful ionization techniques, matrix-assisted laser desorption ionization (MALDI) (3-4) and electrospray ionization (ESI) (5-6), that dramatically expanded the field of biological mass spectrometry. Both techniques offer picomole to femtomole sensitivity in determination of molecular mass up to a few hundred thousand daltons. Since introduced at almost the same time in 1988, MALDI mass spectrometry (MALDI-MS), and E81 mass spectrometry (ESI-MS) have shown considerable promise in characterization of a wide range of biopolymers including proteins, peptides (7-10), carbohydrates (ll-12), and oligonucleotides (13-14). In combination with biochemical techniques, numerous applications in biomedical research have been achieved. The development of MALDI-MS and ESI—MS coincides with an increasing demand for more accurate and sensitive techniques in fields such as biochemistry, molecular biology, genetics and biotechnology (15). Their utility for mass iiIll/i0 measurement also facilitates routine characterization in small molecule synthesis, protein synthesis, and compounds obtained directly from biological matrices. In many respects, MALDI and E81 are complementary. MALDI-MS, used for most of this thesis work, is unique in its capacity to analyze a complex mixture. Compared to ESI-MS or F AB-MS, MALDI-MS is relatively tolerant to contaminants, such as buffers, salts, and denaturants which are common in biological samples. It has therefore the capacity to analyze unseparated, heterogeneous complex mixtures such as enzymatic or chemical digest mixtures of proteins. ESI-MS, on the other hand, is inherently compatible for analyzing mixtures separated on—line by HPLC; yet its ability to directly analyze very heterogeneous samples is limited. MALDI is typically used in conjunction with time-of-flight (TOF) mass analyzers because of the pulsed nature of laser desorption in MALDI. In addition, the time-of- flight mass analyzer has virtually no upper mass range and is compatible with MALDI, which can produce ions at very high mass-to—charge ratio (m/z). In the last few years substantial advancement also has been made in the two areas of the instrumentation of the time-of-flight mass analyzer: (a) enhanced capability of structural determination by means of so called post-source decay (PSD) analysis (16-17); (b) increased mass resolution by incorporating delayed extraction (DE) technique in the ion source (18-21). These latest developments, as described in the remainder of this chapter, have made MALDI TOF-MS an extremely useful technique not only for more accurate molecular weight determination (up to resolution in excess of 10,000 at mass up to 5 kDa), but also for structural elucidation of biomolecules. 2. Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry (MALDI-MS) The first attempt to generate ions of organic molecules by direct laser desorption/ionization dated back to early 19703 (22). However, the size of the analytes CT which can be desorbed and ionized was limited to about 1,000 daltons. In early applications (23-24) of laser desorption to volatilize large biopolymers, the neat analyte (no matrix) was irradiated directly with intense pulses of laser light for short durations. However, mass spectra for large molecular weight compounds were frequently dominated by fragment ions with few if any intact ionized molecules. This technique (direct laser desorption) had limited applications to biological samples and has been quickly supplanted by MALD-MS.. 2.1 Operating Principles The key to the MALDI process is the incorporation of the analyte into a matrix that absorbs radiation from the ultraviolet (or infrared) laser. The matrix compound is a small organic molecule and selected for its properties of laser energy absorption, solubility characteristics, and other criteria as described in detail in Chapter 2. The sample is typically mixed into a solution of the matrix which is in large excess with a matrix:analyte molar ratio in the order of 1000:1. Once the solvent is removed, the sample and the matrix cocrystallize after they have been deposited on a metal surface. Tile physicochemical events leading to the transfer of biopolymers to the gas phase and their ionization in MALDI have not yet been fully elucidated (25). The matrix is believed to serve to minimize sample fragmentation from the laser beam by absorbing the incident laser energy, resulting in the sample and matrix molecules being ejected into the gas phase. One model (15, 26) for the mechanism assumes that the uppermost layers of matrix are induced to undergo a phase transition from the solid to the gas phase. The Subsequent expansion of these matrix molecules into the vacuum drags the matrix- isolated biopolymer molecules into the gas phase. During the transfer from the solid Phase to the gas phase, the biopolymer undergoes ionization through proton transfer or cation attachment with the matrix by reaction processes that remain to be explained (27-30). Once ions are formed in the gas phase, they can be accelerated in an electric field into to a mass analyzer. The operating principle and the configuration of the MALDI instrument used for the thesis work are illustrated in Figure 1.1. This instrument, a PerSeptive Biosystems Voyager-Elite time-of—flight (TOF) mass spectrometer, is equipped with a nitrogen laser emitting at 337 nm, and a single-stage reflector (or reflectron). Dual microchannel plates were employed for ion detection. The sample plate holder used in Voyager Elite system can accommodate 100 sample wells and permits a high sample throughput for analysis. The laser radiation, controlled by an attenuator, is focused on the sample plate well through appropriate optics. A fiaction of light split from the laser pulse initiates the timing circuitry for measuring the time of flight for the ions produced by each laser pulse. In continuous extraction mode, a static electric field is imposed upon ions generated from the sample by application of a two-stage high voltage (20-30 kV) to the sample plate with regard to a closely spaced accelerating electrode, which corresponds to the grounded aperture in Figure 1.1. As all of the ions are given the same energy in the ion acceleration region, The lighter ions formed during each pulse travel faster in a field-free region, the flight tube as shown inFigure 1, and reach the detector earlier. The time required for ions to traverse the flight tube (t.o.f.) is dependent on their masses and described by the relationship: t.o.f. = L/v = L(In/22 eV)1/2, where L is the length of flight tube, v is the ion velocity, m is the mass of the ion, and V is the acceleration potential. Thus the ions are separated into a series of sPatially discrete individual ion packets, each traveling with a velocity characteristic of its m/z ratio. A detector positioned at the end of the field-free region produces a signal as each ion packet strikes it. The transient (duration of 300 microseconds typically) mass ng: Reflector (electrostatic mirror) Timed \ I — Linear detector Ion path In llnear or reflector mode Laser path ——..._‘ / \ _. ion selector lon source (and) Ion source (sample plate/stage) 11 gal l i\ Reflector detector __ Flight tube Sample loading chamber source chamber Figure 1.1: Voyager-Elite MALDI time-of—flight mass spectrometer. spectrum is recorded by a transient recorder at the detector. Commonly 10 to 250 of these transient mass spectra are summed up to produce a usable mass spectrum. To correct the initial kinetic energy spread of ions generated by MALDI, an electrostatic mirror (reflector) can be used to improve mass resolution for some applications. MALDI is typically considered as a soft ionization technique that yields little fragmentation. The most attractive feature of MALDI spectrum is its simplicity. Unlike ESl spectra which are dominated by multiply charged species, the predominant analyte signal in MALDI corresponds to the singly charged species, though the doubly-charged and triply-charged molecules are also observed to varying degrees. Also commonly observed in MALDI mass spectra are ions for gas phase aggregates such as dimers, trimers, etc. Other often observed characteristics of MALDI-MS spectra are adducts arising from analyte-matrix interactions or analyte-alkali metal ion interactions depending on the analyte, choice of matrix, and laser power. 2.2 Sample Preparation Proper sample preparation is critical for successfiil analysis by MALDI-MS (31- 32). Matrix solutions are typically prepared in water, water—acetonitrile or water-ethanol mixtures at a concentration of 5-100 mg/ml depending on the solubility properties of the matrix. The analyte is prepared at a concentration of about 0.01-0.1 mg/ml in a solvent that is miscible with the matrix solution (for peptides and proteins, aqueous 0.1% trifluoroacetic acid (TFA) is frequently used ). The matrix and analyte solutions are rrrixed to achieve a final matrix:analyte molar ratio in the range of 100:1 to 10,000: 1. An aliquot 0f 1 to 2 ul of the mixture is applied to a MALDI—MS sample plate well, and allowed to dry by either ambient evaporation, heating with a stream of warm air, or under vacuum During the drying process, the matrix codeposits from solution with the analytes. To date, the nature of the matrix-analyte interaction has not yet been defined, but it may involve analyte molecules that have become embedded into the matrix crystal lattice or on the surface of rapidly forming matrix crystals. Studies based on optical microscopy have shown that matrix crystal formation and size vary, depending on the matrix and solvent. In general, it has been observed that a decrease in the "quality" of matrix crystal morphology (where the matrix-analyte deposit appears as a film or non-crystalline glassy surface) is associated with a lower quality MALDI mass spectrum (33). These conditions can result from the presence of high levels of contaminating salts, buflers, or surfactants. Some means to desalt the sample may be required for successful analysis by MALDI (34). Alternatively, dilution of a contaminant which can interfere with analyte signal acquisition, may also lead to improved spectra. In general, sample preparation by different matrices, matrix additives, solvents, molar ratio between the analyte and the matrix affects the sensitivity and resolution of MALDI. Optimal results require parallel analyses under different conditions. Other than a metal surface, inert polymeric supports such as transfer membranes have also been used for picomole-level sample handling in protein mass spectrometry (35- 37). Proteins are often immobilized by adsorption onto synthetic supports at some stage as is performed routinely in automated Edman sequence analysis. There is also considerable interest now in the direct analysis by MALDI-MS of biological samples that have been adsorbed onto a transfer membrane because of implications in coupling gel electrophoresis with MALDI-MS. Nylon-66 and positive charge-modified nylon (ZETABIND) membranes were reported to be suitable for analyzing picomole levels of a protein that had been immobilized by adsorption onto the nylon membrane, washed to remove MALDI contaminants, and digested enzymatically and/or chemically prior to adding matrix (35) Another approach, aflinity mass spectrometry employs a modified target surface. In one exarnple, a target composed of porous agarose immobilized with single-stranded DNA has been developed as an affinity-capture medium in the assay of the 80-kDa glycoprotein lactoferrin from biological 5811113165 (3 8) Iii 2.3 Post-source Decay (PSD) MALDI TOF-MS has been initially recognized as “molecular weight machine” by the biochemists because of its superior capability in molecular mass determination of a wide range of biopolymers such as proteins, peptides, and carbohydrates. However, it lacked the capacity for structural analysis because relatively little or no fragmentation was observed in linear MALDI TOF-MS instrument. Metastable fragments formed in the flight tube could not be detected. This deficiency has been overcome since the introduction of the so called post-source decay (PSD) technique (16-17) which can separate and resolve metastable fragments on a reflector-equipped TOF instrument in 1993. In this technique, the excess energy deposited into the intact ionized molecules was found to result in the decomposition of a portion of these species in the flight tube after they leave the ion source. Ifthe ions dissociate within the source, the fragment ions and the intact ionized molecules would have obtained the same energy during acceleration, and because they have different masses, would travel with different velocities and have diflerent time of flight values. This type of fiagmentation, named as “prompt fragmentation” (39-40), does occur in MALDI TOP—MS, but was only observed for some very special analytes such as some carbohydrates and muropeptides derived from peptidoglycans as described in Chapter 2 and 3. However, for those ions that undergo PSD in the flight tube, the resulting fragment ions (or product ions) continue toward the detector with the same velocity as the intact ionized molecules (precursors), and thus are detected as a single peak in a conventional linear TOF- MS. But these product ions have kinetic energies (KE) that are different from each other (if they have different masses) and less than that of the precursor ion, and it is this characteristic that enables them to be separated by mass in the reflector. Consider a molecule M, which was 1“. it h | l is (m ionized in the MALDI source to form a protonated molecule, IVE-F, at m/z 1,000 and accelerated by a electric field of 20 kV. W dissociates into a fragment ion MC at m/z 500 and a neutral M2 during its flight from the source to the detector. W —> Mf + My, Figure 1.2 shows the projected flights of these ions within the reflector portion of the mass spectrometer. When MC ions were formed in flight, they retain the same velocity as MIT ions which remained intact, but the kinetic energy of MC is half that of MIT. In the top panel, when no potential is applied to the reflector, both ions MH+ and MK as well as neutral M2 arrive at the linear detector at approximately the same time, but with some variability, resulting in an increase of the peak width representing the intact MHl'. However, the ions MC can be distinguished very well fi'om MIT when a retarding potential, 12 kV, is applied to the reflector as shown in the low panel of Figure 1.2. The ions lVHrF entering the reflector with a KB of 20 keV will traverse the mirror with a final kinetic energy of 8 keV and hit the linear detector. The neutrals, M2, would be unaffected and hit the linear detector as before. However, the fragment ions, Mll, entering the reflector with a KB of 10 KeV, will be slowed down in the retarding potential first and then turned around by the reflector. Because of the dispersion process by the inherent radial velocity, M1+ ions will reach the reflectron detector and be recorded. By lowering the reflector potential in several intervals, the entire mass range of PSD ffigment ions for a particular precursor can be detected in the same manner. Like other tandem mass spectrometric (MS/MS) techniques, PSD can provide fiill or partial sequence information of molecules less than 2,500 daltons. Precursor ion selection for PSD was typically performed by an electrostatically switched ion gate, marked as timed ion selector Figure 1.1. Since its commercialization about three years MH+ ________ M 1:, M2 IIIIIIIIIIIII IIIIIIIIIIIII MH+ M]+ .flitlttEPe- M1+ “‘1‘: ______ M 5+ M2 M2 Figure 1.2: Differentiation of PSD fragment ion, M1+, from the intact ionized molecule MH+ by using the reflector. M2 is the neutral dissociated from MH+. With no potential applied to the reflector (top pannel), M1+ and M2 maintain the initial velocity of MH+ and reach the linear detector, L, simultaneously. With potential applied to the reflector (low pannel), M1+ is reflected to the reflectron detector, R, and recorded while M1+ and M2 still hit the linear detector. a," ago, MALDI-PSD has quickly evolved into a powerfiil technique for sequence detenrrination of biopolymers like peptides at the low picomole scale (41). 2.4 Delayed Extraction (DE) A major limitation of TOF mass analyzers has often been relatively poor mass resolution. The use of a reflector has overcome this limitation for some applications, but with MALDI in particular, resolution is often inadequate. Typically the resolving power of a linear MALDI TOF-MS is well below 1,000. With a reflector, the resolving power approaches 2,000 or sometimes even greater. Because of the poor resolving power and also the high mass of the ions analyzed by MALDI TOF-MS, it is not possible to discern individual isotopic peaks. The main problem is that the ions produced by MALDI exhibit a rather broad energy distribution. The initial velocity of desorbed analyte ions is nearly independent of mass and the initial kinetic energy is proportional to the mass of the analyte (18, 21). In addition, when desorption occurs in a strong electric field, energy is presumably lost by collisions within the neutral plume, resulting in further mass-dependent energy dispersion (21). The recently introduced technique of delayed extraction (DE) dramatically improves the resolution achievable in MALDI TOF-MS. In contrast to conventional MALDI instruments in which the ions generated by the laser beam near the surface of the sample plate are continuously extracted by a dc potential, in delayed extraction, a short time delay (typically less than 300 nanoseconds) is inserted between the laser desorption/ionization and ion extraction events. The basic idea is to allow those ions having the greater initial kinetic . ”fir: 12 energy to travel farther into the ion source than ions having a lesser kinetic energy, so that when the extraction field is applied, the ion having the lesser initial kinetic energy will traverse a greater distance through the electric field, thereby picking up more kinetic energy in a compensatory effort to give all ions of the same mass the same kinetic energy. Three important parameters need to be adjusted to optimize the performance. These include the field magnitude and direction during and immediately following ion production, the magnitude of the extraction field, and the time delay between the laser pulse and application of the extraction field. In principle, delayed extraction, causes ions of the same mass to arrive at the detector at nearly the same time giving rise to a sharp peak or greater resolution. Some improvement in resolution with delayed extraction may be due to dissipation of the very high local pressure (tens of atmospheres) in the laser plume following (for picoseconds) the laser pulse (18, 21). The benefit of delayed extraction in MALDI-TOF-MS is illustrated in Figure 1.3 which compares the MALDI mass spectrum of human adrenocorticotropic hormone fragment (ACTH) 18-39 obtained by conventional use of a linear TOF-MS (top panel) with those obtained by delayed extraction into a linear (middle panel) and a reflector (low panel) TOF- MS. Delayed extraction improves resolution, mass accuracy, and the quality of MALDI mass spectra by suppressing matrix background, reducing chemical noise, and minimizing the effect of laser intensity on performance. The potential of delayed extraction MALDI has been demonstrated for proteins and peptides as well as for oligonucleotides (42). The delayed extraction technique was added to our Voyager Elite MALDI-MS instrument in late 1996 and was used for the experiments described in Chapter 4 and 5. 13 Linear 2462 2464 2466 2468 2470 2472 Linear DE ' l fir Y I I 2462 2464 2466 2468 2470 2472 Reflectron DE 2462 2464 2466 2468 2470 2472 Figure 1.3: Comparison of MALDI mass spectra of ACTH (18—39) obtained in linear mode without DE (Linear), linear mode with DB (Linear DE),and reflectron mode with DE (reflectron DE) showing the effect of DE on mass resolution of TOF-MS. 3. Glycoconjugate and Protein Analysis by MALDI-MS Glycoconjugates and proteins are two major categories of molecules in biomedical research. Methods for the characterization of carbohydrates and proteins are both required for the analysis of glycoproteins. Like other applications of MALDI-MS, glycoconjugate and protein analysis is still a rapidly developing field. This section describes the latest developments in this field as well as major goals of the research projects carried out in this thesis work. 3.1 Status of Glycoconjugate Analysis Glycoconjugates include glycoproteins, glycopeptides, peptidoglycans, and glycolipids. Glycoconjugate analysis is usually complicated because of the heterogeneous nature of its carbohydrate moiety. The development of MALDI-MS has greatly facilitated structural analysis of glycoproteins. To date, the most reliable method of determining the molecular mass of a heavily glycosylated protein is MALDI TOF-MS (43-44) because the complexity derived from the multiple charge states present in ESI—MS generally renders it impractical for this purpose. New approaches incorporating endoglycosidase and exoglycosdase have been used to identify the site of glycosylation and obtain the sequence of the carbohydrate moiety (45-46). It was also demonstrated (12) that MALDI-PSD analysis could provide structural information for glycopeptides derived from glycoproteins after enzymatic digestion and separation by high-performance liquid chromatography (HPLC). MALDI TOF-MS has been shown to be a viable technique for analysis of lecolipids including gangliosides (47-48). Although satisfactory spectra have been obtained from underivatized molecules, permethylation has been shown to increase the ion at in yield considerably. In general, negative ion analysis has been preferred to work in the positive ion mode, an observation that also applies to the FAB-MS. Biemann et a1. (49) recently found that the molecular weight of highly sulfated glycoconjugates could be measured by MALDI-MS at the picomole level by mixing them with abasic peptide of known mass. These highly acidic compounds, derived from heparin and other polysaccharides, cannot be ionized in any other way with such sensitivity. Such methodology using MALDI-MS leads to a way of sequencing heparin, the detailed structure of which is not known, although the compound has been used clinically as an anticoagulant for many decades. The complete characterization of glycoconjugates usually requires structural analysis of their respective oligosaccharides. Analysis of oligosaccharide structure has been dominated by nuclear magnetic resonance (NMR) in the past. However, NMR lacks the sensitivity to address many biological problems. MALDI-MS not only has become a sensitive and accurate means in determining the molecular weight of oligosaccharides purified from various biological sources (50), but also provides the possibility of sequence and branching pattern determination of oligosaccharides at the picomole level (51). In addition, MALDI-MS has been used for the direct analysis of carbohydrate mixtures such as oligsaccharides from human milk (1 1). 3.1 Status of Protein Analysis Sodium dodecyl sulfate—polycrylamide gel electrophoresis (SDS-PAGE) has been used by the biochemists as a universal technique in protein molecular mass determination fora long period. Over the past few years, MALDI-MS has quickly gained its popularity 16 for protein molecular mass determination with much better accuracy relative to 'ther methods (e.g., SDS—PAGE). Because of the unlimited mass range of this technique, detection for very large proteins such as 500-kDa gramicidin S-synthase and 939-kDa human immunoglobulin IgM (52) was achieved. Incorporation with other biochemical tools like peptide mapping, MALDI-MS has been extensively used for the determination of the primary structure of proteins derived from both natural and recombinant sources. Peptide mapping is a technique whereby a protein sample is digested either enzymatically or chemically, and the resulting peptides are separated and analyzed. Overlapping sets of peptide fragments are generated using different enzymes or chemical agents and the peptides are separated and sequenced using standard Edman degradation reactions. Nowadays, peptide mapping based on MALDI- MS is used in the following applications: (a) confirmation of a protein sequence, especially of proteins produced through recombinant DNA techniques (53); (b) characterization of posttranslational modifications including covalently modified protein N- and C— termini (54), disulfide bond pairing (55-56), phosphorylation (57), glycosylation (58), and lipidation (59); (c) identification of variants in homologous proteins from different biological sources (60); (d) determination of the antigenic site or epitope, the specific region of the protein in intimate contact with the antibody (61). MALDI-MS provides routine means for direct analysis of a tryptic digest (62). A dramatic demonstration of the capability of MALDI in analyzing heterogeneous samples is shown with “protein ladder sequencing” which involves the simultaneous analysis of a mixture of peptide/proteins that have undergone a stepwise Edman degradation (63). Each of the fragments differs from the next by one amino acid residue, which can be identified from the mass difference between successive peaks. Another highly promising area is the unambiguous identification of protein spots from two dimensional gel electrophoresis by MALDI-MS (64-66). This method is Superior to Edman microsequencing for protein spot identification as the latter often suffers fiom problems such as insufficient protein quantities in 2D gel spots, widespread occurrence of N-temrinally blocked proteins, poor protein recoveries from gels, and other unknown factors. 3.3 Major Goals of this Research The three major goals of the research work described in this thesis are: (a) the investigation of a group of compounds, mercaptobenzothiazoles, as a new class of matrices for MALDI-MS and their application in structural characterization of peptidoglycan muropeptides; (b) the development of a new type of derivative for oligosaccharide sequence determination by mass spectrometry; (c) the study of lectin affinity binding in structural analysis of glycoproteins. References: —- . Barber, M., Bordoli, R. 8., Elliot, G. J ., Sedgwick, R. D., and Tyler, A. N., Anal. Chem, 54, 645A—657A, 1982. Rinehart, K. L., Science, 218, 254-260, 1982. Karas, M., and Hillenkamp, F., Anal. Chem, 60, 2299-2301, 1988. Hillenkamp, E, Karas, M., Beavis, R. C., and Chait, B. T., Anal. Chem., 63, 1193A- 1203A, 1991. Fenn, J. B., Mann, M., Meng, C. K., Wong, S. F., and Whitehouse, C. M., Mass Spectrom. Rev., 9, 37-70, 1990. . Fenn, J. B., Mann, M., Meng, C. K., Wong, S. F., and Whitehouse, C. M., Science, 246, 64-71, 1989. Smith, R. D., Loo, J. A., Ogorzalek Loo, R. R., Busman, M., and Udseth, H. R., Mass Spectrom. Rev., 10, 359-451, 1991. Beavis, R. C.; Chait, B. T. Anal. Chem, 62, 1836-1840, 1990. VI ON .\I 90 _\O Strupat, K.; Karas, M., Hillenkamp, F., Int. J. Mass Spectrom. Ion Processes, 111, 89- 102, 1991. 10. Beavis, R. C.; Chait, B. T., Proc. Natl. Acad. Sci. USA, 87, 6873-6878, 1990. ll. Stahl, B.; Steup, M., Karas, M., Hillenkamp, F., Anal. Chem, 63, 1463-1466, 1991. 12.Huberty, M. C.; Vath, J. B; Yu, W.; Martin, S. A, Anal. Chem, 65, 2791—2796, 1993. 13. Tang, K.; Allman, S. L.; Chen, C. H, Rapid Commun. Mass Spectrom, 6, 365-368, 1992. 14. Covey, T. R., Bonner, R. F ., Shushan, B. 1., and Henion, J., Rapid Commun. Mass Spectrom, 2, 249-256, 1988. 15. Siuzdak, G., Proc. Natl. Acad. Sci. USA, 91, l 1290-11297, 1994. h'. l‘, . 3a. tip 16. Kaufrnann, R.; Kirsch, D.; and Spengler, B., Int. J. Mass Spectrom. Ion Proc. 131, 355-385, 1994. 17. Kaufrnann, R.; Spengler, B.; and Lutzenkirchen., Rapid Commun. Mass Spectrom, 7, 902-910, 1993. 18. Brown, R. S., and Lennon, J. J ., Anal. Chem, 67, 1998-2003, 1995. 19. Colby, S. M., King, T. B., and Reilly, J. P., Rapid Commun. Mass Spectrom, 8, 865- 868, 1994. 20. Whittal, R. M., and Li, L., Anal. Chem, 67, 1950-1954, 1995. 21. Vestal, M. L., Juhasz, P., and Martin, S. A., Rapid Commun. Mass Spectrom, 9, 1044-1050, 1995. 22. Posthumus, M. A., Kistemaker, P. G., Meuzelaar, H. L. C., and Brauw, M. C., Anal. Chem, 50, 985-991, 1978. 23. Cotter, R., Anal. Chem. Acid, 195, 45-59, 1987. 24. Nuwaysir, L., and Wilkins, C., Anal. Chem, 60, 279-282, 1988. 25. Chait, B. T.; and Kent. S. B. H., Science, 257, 1885-1894, 1992. 26. Busch, K. L., J. Mass Spectrom, 30, 233-240, 1995. 27. Karas, M., Bahr, U., Ingendoh, A., Nordhoff, E, Stahl, B., and Hillenkamp, F., Anal. Chem. Acid, 241, 175-185, 1990. 28. Ehring, H, Karas, M., and Hillenkamp, F, Org. Mass Spectrom, 27, 472—480, 1992. 29. Gimon, M. 13., Preston, L. M., Solouki, T., White, M. A., and Russell, D. H., Org. Mass Spectrom, 27, 827-830, 1992. 30. Solovki, T. and Russell, D., Appl. Spectros, 47, 211-217, 1993. 31. Cohen, S. L. and Chait, B. T., Anal. Chem, 68, 31-37, 1996. 32' WITH, 0., Roepstorff, P., and Mann, M., Anal. Chem, 66, 3281-3287, 1994- 33- D0ktycz, s. J., Savickas, P. J., and Krueger, D. A., Rapid COmmun- Ma” Specm’m" 5,145-148, 1991. 20 34. Zhang, H., Andren, P. E, and Caprioli, R. M., J. Mass Spectrom, 30, 1768-1771, 1995. 35. Zaluzec, E. J., Gage, D. A., Allison, J ., and Watson, J. T., J. Am. Soc. Mass Spectrom, 5, 230-237, 1994. 36. Blackledge, J. A. and Alexander, A. J ., Anal. Chem, 67, 843-848, 1995. 37. Mock, K. K, Sutton, C. W., and Cottrell, J. S., Rapid Commun. Mass Spectrom, 6, 233—238, 1992. 38. Hutchens, T. W and Yip, T. T., Rapid Commun. Mass Spectrom, 7, 576-580, 1993. 39. Xu N., Huang Z. H., Watson J. T., and Gage D. A., J. Am. Soc. Mass Spectrom 8, 116-124,1997. 40. Scott, P. D. and Viswanatham, K, Anal. Chem. 66, 3727-3733, 1994. 41. Rouse, J. C., Yu, W., and Martin, S., J. Am. Soc. Mass Spectrom, 6, 822—835, 1995. 42. Juhasz, P., Roskey, M. T., Smirnov, l. P., Haff, L. A., Vestal, M. L., and Martin, S. A.,Anal. Chem, 68, 941-946, 1996. 43. Wada, Y., Gu, J ., Okamoto, N., and Inui, K, Biol. Mass Spectrom, 23, 108-110, 1994. 44- LaPolla, A., Baldo, L., Aronica, R., Fedele, D., and Traldi, B, Biol. Mass Spectrom, 23, 241-245, 1994. 45. Sutton, C. W., O’Neil, J. A., and Cottrell, J.S., Anal. Biochem, 218, 34-36, 1994. 46. Treuheit, M. J., Costello, C. E., and Kirley, T. L., J. Biol. Chem, 268, 13914-13918, 1993. 47.1uhasz, P. and Costello, C. E., J. Am. Soc. Mass Spectrom, 3, 785—796, 1992. 43- Harvey, D., J. Mass Spectrom, 30, 1311—1324, 1995. 49- JuhaSZ, P. and Biemann, K., Proc. Natl. Acad. Sci. USA, 91, 4333-4337, 1994- 50‘ Harvey, 13., Rudd, P. M., Bateman, R. H., Bordoli, R. S., Howes, K, HoyeS, I B, and Vickers, R. G., Org. Mass Spectrom, 29, 753-758, 1994. 51‘ Spengler, B., Kirsch, D., Kaufmann, R. J ., and Lemoine, J ., J. Mass Spectrom, 30, 782-787, 1995. 21 52. Nelson, R. W., Dogruel, D., and Williams, P., Rapid Commun. Mass Spectrom, 8, 627-631, 1994. 53. Burlingame, A. L., Carr, S. A., Eds, Mass Spectrometry in the biological sciences, Humana Press, Totowa, NJ, 1996. 54. Specht, B., Oudenampsen-Kruger, E., Ingendoh, A., Hillenkamp, F., Leaius, A. G, and Spener, F., J. Biotechnol, 33, 259-269, 1994. 55. Robertson, J. G., Adams, G. W., Medzihradszky, K. F., Burlingame, A. L., and Villafi'anca, J. J., Biochemisfly, 33, 11563-11575, 1994. 56. Chang, I. Y., Schindler, P., and Chatrenet, B., J. Biol. Chem, 270, 11992-11997, 1995. 57. Liao, P-C., Leykam, J ., Andrew, P. C., Gage, D. A., and Allison, J ., Anal. Biochem., 219, 9-20, 1994. 58.Bur1ingame, A. L., Curr. Opin. Biotechnol, 7, 4-10, 1996. 59. Wilcox, M. D., Schey, K. L., Busman, M., and Hildebrandt, J. D., Biochem. Biophys. Res. Commun., 212, 367-374, 1995. 60. Kerner, J., Zaluzec, E. J ., Gage, D. A., and Bieber, L. L., J. Biol. Chem, 269, 8209- 8219, 1993. 61. Zhao, Y., and Chait, B. T., Anal. Chem, 66, 3723-3726, 1994. 62. Billeci, T. M., and Stults, J. T., Anal. Chem, 65, 1709-1716, 1993. 63. Chait, B. T., Wang, R., Beavis, R. C., and Kent, S. B., Science, 262, 89-92, 1993. 64. Patterson, S. D., and Aebersold, R., Electrophoresis, 16, 1791-1814, 1995. 65. Erdjument,B. H., Lui, M., Sabatini, D. M., Snyder, D. H., and Tempst, P., Protein Sci, 3, 2435-2446, 1994. 66. Clauser, K. R., Hall, S. C., Smith, D. M., Webb, J. W., Andrews, L. E., Tran, H. M., Epstein, L. B., and Burlingame, A. L., Proc. Natl. Acad. Sci. USA, 92, 5072-5076, 1995. AI ,- 30 22 Chapter 2. Ma r‘ L " ' ' A New Class of Matrices for Laser Desorption Ionization Mass Spectrometry 1. Introduction An essential feature of the MALDI process is the use of a low molecular weight organic compound as a matrix to absorb the laser energy. The matrix probably plays multiple roles in the ion formation process (1-4). The step of absorbing the laser radiation leads to the breakup of a microvolume of the solid phase. Reactive precursors are formed and released from the sample surface as protonated and deprotonated matrix molecules, radical ions, and electronically and/or vibrationally excited neutral matrix molecules. Efficient matrices generate a high yield of ions from the analyte relative to the abundance of the matrix ions. To foster this highly complex process of ionization, we assume that a good matrix substance should have the following properties: a. A reasonably high molar absorptivity coefficient at the wavelength used (a > 104 L cm'1 mol'1 for ultraviolet (UV)-MALDI). A higher absorptivity coefficient at the given wavelength does not necessarily make a matrix work better. b. Miscibility with the analyte in the solid phase. The matrix should be soluble in the solvent of the analyte. For application to proteins, the usual solvent is water or a mixture of water and an alcohol or acetonitrile. 0. Good vacuum stability. (1. “Proper chemical composition”. Most matrices reported to date that promote Protonation of the analyte contain OH and/or NH groups. It is plausible that protons are to 23 efiectively transferred from these groups during ionization. However, many good “positive ion mode” matrices are also applicable in the negative ion mode. If the major ionization pathway corresponds to cation or anion attachment, the presence of OH or NH groups is not necessary. Most matrix development studies have been restricted to the positive ion mode; no generalization can yet be made for the negative ion mode because of the lack of sufficient experimental data. e. Other physical properties, such as the heat of sublimation (5) and the lattice structure (6), may affect the efficiency of a matrix. A number of additional features are desirable, but do not necessarily influence the efficiency of the ionization process. For instance, matrix-analyte photoadduct formation can adversely affect mass assignment accuracy when the adduct peaks are not resolved from sample peaks (7—9). The fialfrllment of some of these criteria can be assessed readily because physicochemical data on organic solids are available. Further, comparison of the relative abundance of analyte ions with the relative concentration of the analyte in the sample indicates that the ionization process is highly selective with efficient matrices. This selectivity requires that matrix-derived precursors ionize the analyte preferentially, otherwise a high matrix background and abundant metastable ions can be expected. Some matrices are selected for optimal performance with particular classes of analytes (1 1—14), For all of these reasons, new matrices are still usually found by trial and error. Since Karas and Hillenkamp first reported nicotinic acid (3-pyridine carboxylic acid) as a matrix to desorb large proteins with 266-nm laser radiation (15), a number of other matrices of substantial utility has been discovered. Most of these compounds 24 reported to date are substituted aromatic compounds containing a carboxylic acid group. Several cinnarnic acid derivatives, including ferulic, caffeic, and sinapinic acid, evaluated by Beavis and Chait (5), have made analysis by MALDI possible at laser wavelengths between 266 nm and 355 nm. These matrices, in particular sinapinic acid, form much less abundant photochemically—generated adducts with the analyte than does nicotinic acid. The matrix, 2,5-dihydroxybenzoic acid (DHB), introduced by Karas and co-workers (9), has been extremely useful for the analysis of proteins, carbohydrates, synthetic polymers and enzymatic digests at 337 and 355 nm, and it has a high tolerance to contaminants such as inorganic salts, buffers and detergents. Another matrix, or-cyano-4-hydroxycinnamic acid (otCHCA), is now widely accepted as the choice for the analysis of low-mass peptides and glycopeptides (2). The matrix 3-hydroxy—picolinic acid (16) was shown to desorb and ionize nucleotides effectively in the negative ion mode, and 2-[(4- hydroxyphenyl)azo]benzoic acid (HABA) was reported to be effective for the analysis of relatively large proteins (1 1). Some neutral and basic matrices also have been reported in the literature (17-19). The performance of coumarin laser dyes was investigated by Perera and co-workers (17). These laser dyes were shown to be useful in the analysis of a variety of macromolecules including peptides, proteins, deoxynucleotides, and polymers at 337 nm. The matrix 3- aminoquinoline was reported to be effective for the analysis of polysaccharides and Proteins (18). Fitzgerald et. al screened 37 substituted pyrimidines, anilines, and amino- pyridines as potential matrices (19). Although limited to the analysis of relatively small Proteins and oligonucleotides, these basic matrices extended the utility of MALDI to acid sensitive species. Other neutral matrices, such as 2,3,4-trihydroxyacetophenone and 25 2,4,6-trihydroxyacetophenone were also reported to be good matrices for DNA detection (20). A new liquid matrix which is a binary mixture of ocCHCA and 3-aminoquinoline has been reported by Orlando’s group for magnetic sector mass spectrometers with point detectors (21). This matrix was demonstrated to provide a very long-lasting and reasonably constant ion signal with each laser pulse, without the need to reposition the laser spot or adjust tuning of the mass spectrometer. The successful applications of this matrix included the analysis of oligosaccharides and polymers. A survey of the literature indicated that many compounds such as dithioacetate (I) and 5-ethylidene—rhodanine (H) which contain the common chromophoric functionality - C(=S)-S-, have strong UV-absorption in the region of 330-350 nm (22). This initiated our interest in these compounds as part of a search for new matrices which are structurally distinct from the conventional hydroxybenzoic acids and hydroxycinnamic acids. Considering other physico-chemical characteristics, including compatible solubility with the analytes, vacuum stability and solid state morphology, we chose to evaluate 2- mercaptobenzothiazole (MBT) and its analogs (III) as potential matrices in MALDI-MS. These analogs having the core structure III are: 5—chloro-2-mercaptobenzothiazole (CMBT), 6-amino-2-mercaptobenzothiazole (AMBT), 2—mercapto-5 methOXy- benzothiazole (MMBT), 6-ethoxy-Z—mercaptobenzothiazole (EMBT). Pertinent Properties of these mercaptobenzothiazoles are summarized in Table 2.1. The molar absorptivities at 337 mm (8337) were determined by measuring the absorbance of mercaptobenzothiazoles in a mixture of 1:111 EtOH/THF/water. \ 26 O \ CH //3 NH 3C\ _ / /i S S \S A Y 5 SH III Table 1. Physical properties of MBT and its analogs. Matrix Structure Madnm) 8337 MW (lmol'lcm'l) MBT X=H,Y=H 326.8 1.4x 103 167.25 CMBT X=C1,Y=H 331.7 1.4x 104 201.70 AMBT X=H,Y=NH2 339.0 1.0x 104 182.27 MMBT X=OMe, 37:11 3370 4.4 ><103 197.29 \ 131/Ii X=H,Y=OEt 330.0 1.1 x 104 211.32 27 2. Experimental Materials 2-Mercaptobenzothiazole and 5-chloro-2—mercaptobenzothiazole were purchased from Aldrich Chemical Co. (Milwaukee, WI, USA). 6-Amino—2-mercaptobenzothiazole was purchased from TCI America (Portland, OR, USA). 5-Methoxy-2- mercaptobenzothiazole and 6-ethoxy-2-mercaptobenzothiazole were purchased from Lancaster Synthesis Inc. (Windham, NH, USA). As purification by recrystallization, sublimation and ion exchange did not improve performance, all these matrix materials were used as received without further cleanup. Conventional matrices (including orCHCA, sinapinic acid, and DHB; Aldrich Chemical Co.) have been recrystallized and used according to literature protocols [5,6,10,18] to assure valid comparisons. Oligomannose- type N-linked oligosaccharide [Man]5[GlcNAc]2 (MW 1235.1 Da), Angiotensin I from human (MW 1,296.5 Da), porcine pancreastatin (MW 5,103 Da), bovine pancreatic insulin (MW 5,733.5 Da), horse skeletal muscle myoglobin (MW 16,951 Da), bovine serum albumin (MW 66,430 Da), bovine transferrin (MW 78,000 Da), PEG 3,350, and PEG 10,000 were purchased from Sigma Chemical Co. (St. Louis, MO, USA) and used without further purification. Ganglioside Gm, was purchased from Boehringer Manheim (lndianpolis, IN, USA). Dextran 5,000 was purchased from Fluka Chemical Corp. (Ronkonkama, New York, USA). 28 Instrument MALDI-MS experiments were carried out on the PerSeptive Biosystems Voyager- Elite time-of-flight mass spectrometer equipped with a Model VSL-337ND nitrogen laser (Laser Science, Newton, MA, USA) (337 nm; 3-nsec pulse length) and a dual microchannel plate detector (Galileo, Sturbridge, MA). The accelerating voltage in the ion source was 30 KV. All MALDI-MS results were obtained in the linear mode without DE. The most important operational parameter was the careful regulation of the laser intensity. The laser irradiance per shot was kept as low as possible, i.e., close to the threshold of analyte ion production. Sample Preparation 2-Mercaptobenzothiazole and analogs were dissolved in a mixture of ethanol (EtOPD/tetrahydrofiiran (TI-TF)/water (12121) to make a solution of lOg/l. This matrix solution concentration was found to be the most generally applicable for a variety of analytes. Peptide and protein samples were dissolved in a solution of 1:1 acetonitrile/O.1% (v/v) aqueous trifluoroacetic acid or 111:1 EtOH/THF/water. Carbohydrate samples were dissolved in a solution of 1:1 EtOH/water or 1:1:1 EtOH/THF/water. Polymer samples were dissolved in a solution of 1:1:1 EtOH/THF/water, The samples for mass spectrometric analysis were typically prepared by depositing about 1 it] of the matrix solution and an equal volume of the analyte solution On a sample plate well, which were then mixed by means of the pipette tip. The solvents were then removed by drying the samples in air at room temperature. 3. Results and Discussion Microscopic inspection of the matrix/analyte mixture on the sample plate surface after solvent evaporation revealed homogeneous crystalline phases for all five mercaptobenzothiazoles. Light micrographs obtained from samples of bovine pancreatic insulin in MBT and CMBT (molar ratio of analyte to matrix: 1,1000) are shown in Figure 2.1. For the samples with MBT, evenly spread large crystals were observed. The samples with CMBT, on the other hand, formed a uniform layer of very fine crystals distributed over the entire sample plate surface. Light micrographs obtained from the same sample in DHB and orCHCA are also included in Figure 2.1 for comparison. 3.1. Peptides and Proteins All five selected mercaptobenzothiazoles were capable of desorbing and ionizing angiotensin and bovine pancreatic insulin with very strong, stable signals for the protonated molecules, comparable to those obtained with orCHCA or sinapinic acid. Further experiments with higher molecular weight proteins showed that the quality of the mass spectra obtained and the mass range applicable varied with different mercaptobenzothiazoles. Despite a lower molar absorptivity at 337 nm relative to that of some of its an31083, 2-mercaptobenzothiazole (MBT) produced MALDI spectra with the best signal intensity and resolution for the peptides and proteins with molecular weights up to at least i ’ p rlg ) oCHCA, .5; 3 MI .. ‘m 'A N. 1 \‘AL 3. {l 100,000 daltons. Fig. 2 presents the positive MALDI spectra of human angiotensin I (Fig. 2a), porcine pancreastatin (Fig. 2b), and bovine transferrin (Fig. 2c) using MBT as the matrix. The amount of protein loaded on the sample plate cell was about 1 to 5 picomoles. In all these spectra, the major peak observed corresponds to the singly charged (protonated) analyte denoted by [M+H]+. Signals corresponding to the dimer of bovine transferrin (Fig. 2c) were also observed. For this, and other proteins examined (data not shown), the high-mass cluster ions are usually of significantly lower abundance. Multiply-charged analyte signals were also detected for all the analytes except angiotensin. The highest charge state observed was +4 in the spectrum of bovine transferrin (Fig. 2c). N0 obvious fragmentation was observed. The matrix adduct ion peaks were weak, but well resolved for small proteins such as porcine pancreastatin (MW 5,103 Da, Fig. 2b). Like orCHCA and sinapinic acid, MBT forms matrix-generated adducts of much lower abundance than that of the protonated molecules. In these spectra, the measured mass resolution (M/AM, FWHM) for the peak representing the protonated molecule, was 766 for angiotensin, 540 for porcine pancreastatin, and 63 for bovine transferrin, respectively. Additionally, MBT is a suitable matrix for the analysis of small peptides and proteins because of the absence of matrix ions above m/z 500. These and other results (data not shown) from examining over 30 different peptides and proteins demonstrate that MBT provides similar performance to that obtained with orCHCA or sinapinic acid in terms of sensitivity and mass resolution. 32 a 1M+H1 * .é‘ .5 [M+Na] T 2 ii & I I I 1200 1300 1400 b IM+HI * .E‘ m I: Q.) .E E E d.) 9‘ 1M+2Hi 2* I I” | I | 2000 3000 4000 5000 6000 MM 2* rM+Hi * E U) E .s e '5 E 32 [2M+H] " I I 50000 100000 150000 m/z Figure 2.2: Positive MALDI mass spectra of three proteins with MBT as matrix. (a) human angiotensin I (MW 1,296.5 Da); (b) porcine pancreastatin (MW 5,103 Da); (c) bovine transferrin (MW 78000 Da). The data were aquired by summing 30 to 100 laser shots. The amount of proteins loaded on the sample plate was about 1 to 5 picomoles. lion] iii-011 .‘n MBT also has been used successfully for protein analysis in the negative ion mode. Atypical negative ion mass spectrum of bovine serum albumin (BSA) with MBT as matrix is shown in Figure 2.3. In addition to the deprotonated molecule [M-H]', doubly- [M- 7.H]2' and triply- [M-3H]3' charged (deprotonated) molecular ion signals were observed in this spectrum. One of the strengths of MALDI-MS is its relative tolerance to contaminants (such as buffers, salts) in the sample. However, it is generally agreed that contaminants like detergents which are sometimes used for protein sample preparation and stability degrade the quality of spectra with respect to their signal—to-noise ratio and mass resolution, making the analysis of protein samples in the presence of surfactant agents by MALDI-MS very challenging (9-10, 25). The most deleterious spectral effects have been assigned to sodium dodecylsulfate (SDS), an ionic detergent (10, 25). With MBT as matrix, inorganic salts, and buffers in the analyte solution up to concentrations of at least 200 mM, as well as non-ionic detergents such as urea up at concentrations of at least 1M, do not lead to a Significant decrease in detection limits or quality of spectra. Furthermore, MBT was also found to have excellent tolerance to ionic detergents such as SDS. Figure 2.4 (a) presents the positive MALDI spectrum of myoglobin (10'5 M in acetonitrile/O.1% (v/v) aqueous trifluoroacetic acid (1 :1) solution) with MBT as matrix. The tolerance of MBT to SDS is illustrated in Figure 2.4(b) and Figure 2.4(c) where the myoglobin sample was applied in a 13/1 SDS (0.1%, w/v) and 100 g/l (10%, w/v) solution, respectively. SDS caused the background signal to be elevated in the low-mass region. The signal-to-background ratio for myoglobin has deteriorated compared to that in Fig. 4(a), but singly-charged and doubly-charged protein signals are clearly present in both Figure 2.4(b) and Figure 2.4(0). i‘ 34 IM-2I'112' [NI-Hl' .é‘ 2 8 .5 .9 3 3- f“. WI- ”1 82 I I 30000 50000 70000 m/z Figure 2.3: Negative ion MALDI spectrum of bovine serum albumin (MW 66,430 Da) with MBT as matrix; 200 laser shots summed. Total protein loaded: 5 pmol, a12.4; .’ 5th 35 2 2+ a M ”1 1M+Hi+ .9 a 8 £1 2 E o m L. I I I 10000 15000 20000 b IM+Hl+ i .é‘ ,5 [M+2H]2+ .2 5 <9 m 1 10000 rshoo 20000 C 1M+H1+ .é‘ a e 1M+2Hi2+ ‘5 E 4) or. 1 I I 10000 15000 20000 Figure 2.4: Positive MALDI spectrum of horse skeletal muscle myoglobin (10 pmol); 100 laser shots summed. The sample was dissolved in acetonitrile/0.1% (v/v) aqueous tnfluoroacetic acid (1:1) solution. (a) no sodium dodecylsulfate (SDS) added, matrix: MBT; (b)with the addition of O. 1% SDS, matrix: MBT; (c) with the addition of 10% SDS, matrix: MBT. 1:1 Ill. 36 It also was noted that the signal-to—noise ratio and mass resolution in Figure 2.4(c) are slightly better than that in Figure 2.4(b) in spite of more SDS added. Among the matrices reported in the literature, only DHB shows some tolerance to SDS (9,25). It has been reported before that even minimal contamination with sodium dodecylsulfate (SDS) completely suppresses the protein signals when matrices such as cinnarnic acid analogs including oc-cyano—4-hydroxycinnamic acid (aCHCA) are used in the sample preparation (10, 25). However, a recent article by Amado et a1. (26) argued that protein samples containing SDS can be analyzed by MALDI-MS with ocCHCA as the matrix. In their results, a decrease in the protein signal also was observed for increasing concentrations of SDS up to 0.1% (w/v), at which no signal was observed. But a recovery of the signals was observed at concentrations of SDS above 0.3%, with useful spectra at concentrations of SDS as high as 10%. They attribute the sample signal recovery to the formation of micelles above the critical micelle concentration (cmc, 0.23% for SDS). In their theory, at subcritical concentrations, the protein molecule may partially precipitate by the formation of ion pairs with the surfactant monomers, diminishing the quantity that cocrystallizes with the matrix, while the excess of surfactant may coprecipitate with the matrix, a part of the surfactant remains in the liquid phase. As the sample dries, the surfactant effectively coats the crystals, which may partially prevent desorption of the protein. Also in their theory, when SDS concentrations are above the mic, the protein is stabilized and solubilized by surfactant micelles, which may lead to a higher surface concentration of the protein resulting in an increasing ion formation Suflicient to compete with surfactant ion formation. Our results using MBT as the matrix Showed that the spectra quality with 10% SDS addition was better than that with 0.1% 37 SDS addition, which somewhat agreed with what Amado et al. have observed. However, contrary to what they observed, the protein signal was detectable with O. 1% SDS addition using MBT as the matrix as shown in Figure 2.4(b). Other analogs of 2-mercaptobenzothiazole were found to be effective matrices for the analysis of peptides and low-mass proteins in the range of 800 to 6000 daltons. Figure 2.5 shows a mass spectrum of bovine pancreatic insulin (sample loading: 3 pmol) using CMBT as the matrix. The mass resolution (FWHM) for the major peak at m/z 5734.5 is about 650. Peaks for the protonated molecule [M+H]+, sodium adduct [M+Na]+, potaSsium adduct [M+K]+, and matrix adduct [M+202]+ are well resolved. When these analogs are used for the analysis of higher-mass proteins, however, peak broadening becomes more obvious, resulting in lower signal intensity and mass resolution, which may partly be attributed to the formation of unresolved matrix adduct ion peaks. 3.2. Carbohydrates Although sodium and potassium adduct ions are almost always present in the MALDI spectra of peptides and proteins, the most abundant signals observed are from protonation of analyte molecules. However, when MALDI is applied to carbohydrates, cation attachment usually predominates, so that protonated molecules are rarely detected. The cinnamic acid-type matrices have been used successfully for the analysis of peptides and proteins, but these matrices are not useful for the analysis of carbohydrates. However, DHB, which is a good cationizing matrix, was found effective for carbohydrate analysis. 38 [WHY E‘ (I) a E: _“>_’, IM+Nr=Il+ E [M+2Na-H]* d) ‘3‘ [M+202]* I I I I 5600 5800 6000 6200 rn/z Figure 2.5: Positive MALDI spectrum of bovine pancreatic insulin (MW 5,733.5 Da) using CMBT as matrix; 30 laser shots summed. The region of the spectrum shown corresponds to a mass range of 5,500—6,150. Total protein loaded: 2 pmol. Effie] Z“-10 l inn SlilGlc] 315% is r lady 39 Among the five selected mercaptobenzothiazoles, only CMBT was found to be a successful cationizing matrix for the analysis of carbohydrates. The performance of CMBT as a cationizing matrix was compared with that of DHB in the analysis of an oligomannose-type N-linked oligosaccharide, [Man]5[GlcNAc]2 (MW 1,235.1 Da), as illustrated in Figure 2.6. The amount of carbohydrate loaded on the sample plate cell was 2pmol. With either CMBT or DHB as matrix, the major peak in the mass spectrum represents the sodium-cationized molecule, [M+Na]+. A potassium adduct ion, [M+K]+, is also observed with much less abundance. However, CMBT (Figure 2.6(a)) offers a much higher signal-to-background ratio for this oligosaccharide sample than does DHB (Figure 2.6(b)). When sample loading was decreased to 100 finol, no signal could be obtained with DHB as matrix. Even with binary matrices such as super—DHB (a mixture of DHB and 5-methoxysalicylic acid) (23) and DHB/HIC (a mixture of DHB and 1- hydroxyisoquinoline) (24), which have been reported to improve the performance of DHB, no signal could be obtained at the 100-fmol level (data not shown). However, with CMBT the carbohydrate signal is still detectable as shown in Figure 2.6(c), indicating the better sensitivity offered by this new matrix. Furthermore, cation doping experiments were performed to examine the effect of adventitious salts on performance. NaCl was added to the sample solution (at final concentrations of 50 leI) when CMBT or conventional matrices (DHB and and binary DHB mixtures) were used in the analysis of [Manls[GlcNAc]2. No significant difference in terms of the performance of the individual matrices was observed. As reported by other research groups (25), the solid state morphology of the matrix/analyte mixture plays an important role in MALDI ion yields. A practical 40 a [M+Na] * :5 Sample Loaded: 2 pmol 5‘? Matrix: CMBT .5 a 3:1 .1.“ 84’ | | | l r l 100 1200 1300 1400 1500 b [M+Na] * .é‘ a Sample Loaded: 2 pmol :3 Matrix: DHB ‘3 '33 IS 0 ad I I I I I 1100 1200 1300 1400 1500 C MNal + Sample Loaded: 100 fmol Matrix: CMBT Relative Intensity I I | I I I 100 1200 1300 1400 1500 m/z Figure 2.6: Positive MALDI spectrum of oligomannnose-type N-linked oligosaccharide [ManlsIGlcNAc]2 (MW 1,235.1 Da); 40 laser shots summed. (a) sample loaded: 2 pmol, matrix: CMBT; (b) sample loaded: 2 pmol, matrix: DHB; (c) sample loaded: 100 fmol, matrix: CMBT, .Ll 41 advantage of using CMBT as a matrix is the capacity to generate ions uniformly from sites across the target owing to its homogeneous crystallization with the analyte over the entire sample surface area. Almost any portion of the sample surface can be used to obtain the resulting spectra shown above. Good quality spectra could be obtained with DHB, but the results were not as reproducible as those obtained by CMBT because of the irregular crystallization process for DHB during sample preparation. DHB tends to form large crystals on the rim of the target (although DHB/HIC is more homogenous), leaving a microcrystalline region in the center, which yields the strongest signals for carbohydrates (9,12). In this study, different regions of the target were irradiated to generate the best signal possible. The potential of CMBT as a matrix for oligosaccharide analysis is demonstrated in Figure 2.7 in which dextran 5,000 was selected as a representative example of oligosaccharides with a high molecular mass range. Dextran 5000 consists of a mixture of polysaccharides containing a backbone of D-glucose units produced by bacteria growing on a sucrose substrate. The MALDI spectrum by CMBT shows oligomers up to 5,500 daltons with good resolution and signal-to-background ratio. Gangliosides are a class of glyCOSphingolipids that consist of a hydrophobic cerarnide and a hydrophilic carbohydrate chain containing one or more sialic acid residues at various linkage sites. These compounds can be analyzed by MALDI-MS with a number 0f matrices such as DHB, HABA, and 6-aza-2-thiothymine (26). CMBT also was found t0 be an efficient matrix for these glycolipids, in both the positive and negative ion mode. Negative ion mass spectra of the underivatized gangliosides were always of better quality Relative Intensity IIIIIIHIM... I I I 2000 4000 6000 rn/z Figure 2.7: Positive MALDI spectrum of dextran 5,000 with CMBT as matrix; 50 laser shots summed. Adjacent oligomer peaks in the spectrum are spaced by 162 mass units. 43 than the positive ion spectra, exhibiting better sensitivity, resolution, and less fragmentation fi’om loss of sialic acid residues. Figure 2.8 shows the mass spectrum of underivatized Gm, containing three sialic acid residues, obtained in the negative ion mode by using CMBT as matrix. No fragments were observed. Two prominent molecular species were detected in the sample, differing from each other by two methylene units. The most abundant peaks corresponded to [M-2H+K]' ions of each species, while [M-H]' peaks were less intense. CMBT also was found to have superior performance over conventional matrices in analyzing complex carbohydrates such as the muropeptides (27) derived from peptidoglycan, a structural component of the cell walls of bacteria. Figure 2.9 shows the structure of a trimeric muropeptide derived from peptidoglycan that was isolated from a highly methicillin-resistant Staphylococcus aureus Strain COL (28-29). Figure 2.10(a) displays the mass spectrum of this trimeric muropeptide using CMBT as matrix. Sodiated analyte signals are the dominant peaks in the spectra. A multiple sodium adduct ion, IM+2Na-H]+, and a prompt fragmentation peak [M+Na—GlcNAc]+ also are present in the mass spectra. Most of the other peaks present in the spectra are suspected to be impurities that coeluted in the HPLC fraction, or decomposition products. The mass spectrum presented in Figure 2.10(a) was recorded from a total sample load of 3 pmol indicating high sensitivity analysis can be achieved by MALDI-MS using CMBT as the matrix. In contrast, only weak signals were observed for this sample with conventional matrices such as otCHCA, DHB, or sinapinic acid as shown in Figure 2.10(b) where aCHCA was used as an example. 44 [Mz-ZHirK] ‘ Relative Intensity I I 2100 2200 2300 2400 m/z Figure 2.8: Negative MALDI spectrum of underivatized Gm, with CMBT as matrix, 40 laser shots summed. M1 and M2 are two molecular species in the sample differing by two melhl’lene units in the long—chain base. [Mi + K -2H]' and [M2 + K .21-1]~ of Mo molecular species are represented by dominant peaks in the spectrum. 45 G -r G—hl/I Alla G—M Ala G1" Ala Glx les— (61395 six Lsz—(Gly)5—Ala L'ys — (Gr/)5 — Ala Al. Al. Figure 2.9: Structure of trimeric muropeptide derived from the peptidoglycan of S aureus strain COL. The following abbreviations were used: G, N—acetylglucosamine; M, N-acetylmuramic acid. 46 a were]+ .4? g + [702Na-H] + E; UM+NafiyuNAq 0 M M I I I I I 3000 3200 3400 3600 3800 b IM+NaI * .Q‘ i ,N H, . 5 Na-GluNAc] * I)” a' s M / E D a: I I T T r 3000 3200 3400 3600 3800 m/z Figure 2.10: Positive MALDI spectrum of trimeric muropeptide derived from Peptidoglycan of S. aureus strain COL with (a) CMBT (b) OLCHCA as matrix. 40 laser shots summed. Total sample loaded was approximately 3 pmol. H n.- 47 3.3. Synthetic Polymers Synthetic polymers are another class of compounds that is typically more easily ionized by cation attachment (30-32). All five selected mercaptobenzothiazoles proved to be useful for analyzing synthetic polymers. The positive ion mass spectrum obtained from PEG 3,350 with AMBT as matrix is illustrated in Figure 2.11. No alkaline salts were added to this polymer sample prior to analysis. The spectrum clearly shows well resolved adjacent oligomer signals with masses ranging from 2,440 daltons to 4,600 daltons in Figure 2.11, reflecting the molecular weight distribution of PEG 3,350. Two series of peaks were observed originating from Na+ and K+ adducts. The peaks corresponding to the [M+Na]+ ions of the oligomers were centered at a value corresponding to a Mw of 3,362 daltons, close to the weight-averaged molecular weight provided by the manufacturer (Mw: 3,3 50 Da). The mass difference between adjacent [M+Na]+ ion signals was found to be approximately 44.2 daltons. The potential of using mercaptobenzothiazoles for the analysis of higher molecular weight synthetic polymers such as PEG 10,000 is demonstrated in Figure 2.12. Again, AMBT is used as the matrix, although similar results can be obtained with other mercaptobenzothiazoles, including MBT and CMBT. NaCl was doped into the PEG 10,000 sample to, assist cationization. The shape of the molecular weight distribution and mass of the repeat unit were well defined, although the oligomer ion peaks of PEG 10,000 could not be resolved completely. The oligomer distribution ranged from m/z 10,300 to ca. 14,600. The weight—averaged molecular weight obtained from tln's MALDI in thin 48 E U) 5 .5 .3 E 0 M MA». I I I ' I 2500 3000 3500 4000 45 0 m/z Figure 2.11: Postive MALDI spectrum of PEG 3,350 with AMBT as matrix. 30 laser shots summed. No alkaline salts were added to the polymer solution. Relative Intensity I I I I l 1000 12000 13000 14000 m/z Figure 2.12: Positive MALDI spectrum of PEG 10,000 with AMBT as matrix. 100 laser sllOts summed. NaCl was added to the polymer solution. 49 spectrum is ca. 12,600 which shows about 10% positive discrepancy from the GPC data provided by the manufacturer (Mw = 11,500 daltons). This observation agrees well with the results reported by two other research groups (33-34), where MALDI-MS showed about 10% higher weight-averaged molecular weight for PEG 8,000, relative to GPC. We also have applied mercaptobenzothiazoles in the analysis of nonpolar synthetic polymer molecules, such as poly(methyl methacrylate) 15,000. Although molecular weight distribution information could be determined from the spectrum (data not shown), the signal intensity is not as strong as that obtained from PEG samples. 4. Conclusions The performance of five selected mercaptobenzothiazoles has been evaluated as a matrix for MALDI-MS at 337 nm. It was demonstrated that 2-mecaptobenzothiazole is applicable to the analysis of peptides and proteins of molecular weight up to at least 100,000 daltons with comparable sensitivity and resolution to those achieved with conventional matrices such as orCHCA and sinapinic acid. 5-Chloro-2- mercaptobenzothiazole (CMBT) was found not only to be effective as a matrix for the analysis of peptides, low-mass proteins, and glycolipids, but also superior to conventional matrices for the analysis of oligosaccharides and muropeptides. Synthetic polymer analysis also is possible by MALDI-MS with these new matrices. In view of the fact that most matrices presently employed are essentially acidic in nature, introduction of a family of neutral materials (e.g., pKa for MBT is 7.5) will complement the limited choice of matrices available to deal with acid-sensitive biomolecules. 50 References: h—d . Ehring, H., Karas, M., and Hillenkamp, F ., 1992, 27, 472-480. 2. Beavis, R. C. and Chait, B. T., Org. Mass Spectrom., 1992, 27, 156-158. 3. Gimon, M. E., Preston, L. M., Solouki, T., White, M. A., and Russell, D. H., Org. Mass Spectrom., 1992, 27, 827-830. 4. Juhasz, P., Wang, B. H., and Biemann, K., Proceedings of the 40th Conference on Mass Spectrometry and Allied Topics, Washington, DC, May 31-June 5, 1992, p 372- 373. 5. Beavis, R. C. and Chait, B. T., Rapid Commun. Mass Spectrom., 1989, 3, 233-237. 6. Beavis, R. 0, Proceedings of the 40th Conference on Mass Spectrometry and Allied Topics, Washington, DC, May 31-June 5, 1992, p 5. 7. Hillenkamp, F.; Karas, M., Beavis, R. C.; and Chait, B. T., Anal. Chem, 1991, 63,1193A-1198A. 8. Beavis, R. C. and Chait, B. T., Anal. Chem, 1990, 62, 1836-1840. 9. Strupat, K., Karas, M., and Hillenkamp, F., Int. J. Mass Spectrom. Ion Processes, 1991, 111, 89-102. 10. Beavis, R. C. and Chait, B. T., Proc. Natl. Acad. Sci. 1990, 87, 6873-6878. 11. Juhasz, P., Costello, C. E., and Biemann, K, J. Am. Soc. Mass Spectrom., 1993, 4, 399-409. 12. Stahl, B., Steup, M., Karas, M., and Hillenkamp, F., Anal. Chem, 1991, 63, 1463- 1466. 51 l3.Huberty, M. C., Vath, J. 13., Yu, W., and Martin, S. A., Anal.Chem., 1993, 65, 2791- 2796. 14. Tang, K, Allman, S. L., Chen, C. H., Rapid Commun. Mass Spectrom., 1992, 6, 365- 368. 15. Karas, M. and Hillenkamp, F ., Anal Chem, 1988, 60, 2299-2303. 16. Wu, K. J., Steding, A., Becker, C. H., Rapid Commun. Mass Spectrom., 1993, 7, 142- 146. l7.Perera, I. K., Kantartzoglou, S., Dyer, P. E., Int. J. Mass Spectrom. Ion Processes, 1994, 137, 151-156. 18. Metzger, J. O., Woisch, R., Tuszynski, W., Angermann, R. Fresenius J. Anal. Chem, 1994, 349, 473-475. 19. Fitzgerald, M., Parr, G. R., Smith, L., M. Anal. Chem, 1993, 65, 3204-3211. 20. Zhu, Y. F., Chung, C. N., Taranenko, N. 1., Allman, S. L., Martin, S. A., Chen, C. H., Haff, L., Rapid Commun. Mass Spectrom., 1996, 10, 383-388. 21. K011i, V. S. K. and Orlando, R., Anal. Chem, 1997, 69, 327-332. 22. Organic Electronic Spectral Data, Vol. 1, 1946-52. 23. Rosinke, B., Strupat, K., Hillenkamp, F., Rosenbusch, J ., Dencher, N., Krilger, U.; Galla, H. J., Mass Spectrom., 1995, 30, 1462-1468. 24. Karas, M., Ehring, H., Nordhoff, E., Stahl, B., Strupat, K., Hillenkamp, F., Grehl, M., Krebs, B., Org. Mass Spectrom., 1993, 28, 1476-1481. 25. Mohr, M.D., Bornsen,'K.O., Widmer, H.M., Rapid Commun. Mass Spectrom., 1995, 9, 809-814. 52 26. Amado, F. M. L., Santana-Marques, M. G, Ferrer-Correia, J ., and Tomer, K. B., Anal. Chem, 1997, 69, 1102-1106. 27. Allmaier, G. and Schmid, E. R., in Bacterial Growth and Lysis; de Pedro M. A, Holtje J.- V., Loffelhardt, W., ED; Plenum Press, New York, 1993; FEMS Symposium No.65, p 23. 28. de Jonge, B. L. M., Chang, Y. -S., Gage D. A, Tomasz, A, J. Biol. Chem, 1992, 267, 11248-11253. 29. de Jonge, B. L. M., Chang, Y. —S., Gage D. A, Tomasz, A, J. Biol. Chem, 1992, 267, 11255-11259. 30. Bahr, V., Deppe, A., Karas, M., Hillenkamp, F., Anal. Chem, 1992, 64, 2866-2870. 31. Cotterl, J. S., Koemer, M., Gerhards, R., Rapid Commun. Mass Spectrom., 1995, 9, 1562-1564. 32. Larsen, B. S., Simonsick, W. J. Jr, McEwen, C. N., J. Am. Soc. Mass Spectrom., 1996, 7, 287-292. 33. Tang, X, Dreifilss, P. A., Vertes, A, Rapid Commun. Mass Spectrom., 1995, 9, 1141- 1147. 34. Dey. M., Castoro, J. A, Wilkins, c. L., Anal. Chem, 1995,67, 1575-1580. llUUI ”slant 53 Chapter 3. Structural Characterization of Peptidoglycan Muropeptides by MALDI- MS and Post-Source Decay Analysis 1. Introduction Peptidoglycan, a key component of bacterial cell walls, is a heteropolymer with unique chemical composition (1). It is composed of sugar chains consisting of alternating N-acetylglucosamine (GlcNAc) and N—acetyhnuramic acid (MurNAc) residues that are interlinked via the muramic acid residues through short peptides, which include D-amino acids as illustrated in Figure 3.1 with Staphylococcus aureus bacterial cell wall peptidoglycan used as an example. The network that this polymer forms surrounds the cytoplasmic membrane and protects the cell from osmotic rupture (2). Because of this property, its ubiquity among bacteria and absence in humans, antibiotics that interfere with the peptidoglycan metabolism have been successfully used to combat bacterial infections. Such antibiotics including penicillin and methicillin specially bind to and inactivate enzymes that function to cross-link the peptidoglycan strands of bacterial cell walls. Since cell wall expansion also requires the action of enzymes that degrade and resynthesize cell walls, exposure of growing bacteria to such antibiotics results in their lysis, that is, the normal balance between cell wall biosynthesis and degradation is disrupted. However, recently, growing resistance has emerged against this class of drugs (3- 51- Mounting numbers of bacterial species have been found to be multiply resistant to antibacterial drugs. In particular, infections caused by methicillin—resistant .0233 a £3» Eon 028m 8w meta Sago 02m 382 30:3 owtwnoommmw Ao4 linkages between N-acetylglucosamine and N-acetyhnuramic acid. These muropeptides can exist as monomers (the disaccharide unit plus short peptide chain) or can be cross-linked via their peptide chains to varying degrees to form dimers and higher oligomers. While FAB-MS and FAB-CAD-MS/MS have been successfully employed to Structurally characterize muropeptide monomers and dimers at the nanomole level, it is ‘33 thfl 1%? ”10‘ 56 difficult to obtain even molecular weight information from trimers, and FAB-MS usually fails entirely for higher oligomers. Within the past decade, the development of new ionization techniques such as electrospray ionization (ESI) MS (11) and matrix-assisted laser desorption ionization (MALDI) MS (12-15) have made significant contributions to the study of biological molecules. An early study by Allrnaier et al. (16) reported the molecular weight determination of muropeptides at the picomole level using these new ionization techniques. In this chapter, MALDI-MS, PSD, and specific enzymatic degradation with lysostaphin are demonstrated as an approach to characterize the muropeptide composition of a highly methicillin-resistant Staphylococcus aureus strain. With these techniques, structural information from muropeptides including peptide sequence and cross-linking patterns can be obtained with greatly improved sensitivity, relative to that of conventional mass spectrometric methods. 2. Experimental Preparation of Muropeptides from Peptidoglycan Muropeptides were obtained from a highly homogeneous methicillin-resistant Staphylococcus aureus strain COL as described by de Jonge et al. (9). Peptidoglycan was digested with muramidase (Sigma Chemical Co., St. Louis, MO) and the muropeptides were reduced with sodium borohydride. Separation of the reduced muropeptides by HPLC was performed essentially by the method of Glauner (22), with some modifications. Samples were applied to a 250 x 4.6-mm reversed-phase column ' 57 (ODS-Hypersil, 3 pm; Keystone Scientific, Bellefonte, PA), guarded by a (70 x 4.6-mm) perisorb RP-18 (EM Science, Elmsford, NY) precolumn. The column was eluted at a flow rate of 0.5 ml/min with a linear gradient starting immediately after injection of 5% ,(v/v) methanol in 100 mM NaH2P04 (pH 2.5), 0.00025% azide per liter to 30% (v/v) methanol in 100 mM NaHzPO4 (pH 2.8) in 150 min. Collected muropeptide fractions were desalted with acetonitrile according to Dougherty (23) and the amino acid composition was determined. Lysostaphin Digestion of Muropeptides for MALDI Mass Mapping Desalted muropeptide samples were digested with lysostaphin (Sug/ml, Applied Microbiology, NY) in 12.5 mM sodium phosphate buffer (pH 5.5) for 16 h and analyzed directly without firrther purification. Instrumentation MALDI-MS experiments were carried out on the PerSeptive Biosystems Voyager- Elite time-of-flight (TOF) mass spectrometer and continuous extraction with an accelerating voltage of 25 kV was used in all experiments. Linear mode and reflector mode mass spectra were calibrated with a model muropeptide GlcNAc-MurNAc-Ala-Gln (MW : 694.7, Sigma Chemical Co., St. Louis, MO) and human gastrin I (MW: 2098.2, Sigma Chemical Co., St. Louis, MO) as standards in both positive and negative modes. In PSD experiments, precursor ion selection was performed by means of an electrostatically switched ion gate, which allows a mass selectivity of about 90 WW. PSD fragment ion Spectra were sequentially recorded over 10-12 mass windows by incremental reduction of 58 the reflectron voltages. In each spectral window 30-100 single-shot spectra were averaged depending on the sample conditions. Accuracy and precision of PSD mass assignments reach i0.3-0.4 u in the mass range of m/z 300-1300. Compilation (stitching) of a set of PSD mass windows was performed by the data system. Average mass values are reported throughout this chapter. Sample Preparation 5-Chloro-2-Mercaptobenzothiazole (CMBT) was purchased from Aldrich Chemical Co. (Milwaukee, WI) and used as the MALDI matrix in muropeptide analysis without further purification (24). CMBT was dissolved in a mixture of ethanol/tetrahydrofirran/water (1 :2: 1) to make a solution of IOg/l. The samples for mass spectrometric analysis were typically prepared by depositing about 1 ul of the matrix solution and an equal volume of the analyte solution on a sample plate well and mixed. The solvents were removed by air-drying. 3. Results and Discussion Among many MALDI matrices employed to date, 5-chloro-2- mercaptobenzothiazole (CMBT) has been found to be most suitable for muropeptide mass determinations (24). For peptidoglycan analysis, CMBT offers sensitive detection of muropeptides at low picomole to femtomole range in both positive and negative modes. This matrix also exhibits excellent experimental reproducibility owing to the homogeneous ClYStallization of the analyte/matrix mixture over the entire sample surface area. In «Which “lyric-firmed 1% Weight bet: lhupeptides detivet rim in Figure 3.2. Th lilDl-TOF-MS with the i *WR and negative ic “I. Because of the Wm the positive My ion signals. Protc Mbmed with much 1: li’ld in the negative ion in the negative ion Wanted molecule [M- an The simplicity of floater weight determinati In the positive and "liner shown in Figure “W t. Some ion sign "him source) (25), donor [_____7 59 contrast to. the high sensitivity achieved by MALDI-MS, nanomole quantities of the sample are typically required in analysis by FAB-MS. 3.1. Molecular Weight Determination Muropeptides derived from peptidoglycan of Staphylococcus aureus strain COL are shown in Figure 3.2. The molecular masses of these compounds were determined by MALDI-TOF-MS with the instrument operated in the linear mode. Figure 3.3 presents the positive and negative ion MALDI mass spectra of the unsubstituted muropeptide monomer. Because of the high affinity of alkali “ions for the sugar moiety of the peptidoglycans, the positive ion MALDI mass spectrum (Figure 3.3(a)) is dominated by[M+Na]+ ion signals. Protonated molecules [M+H]+ and potassium adducts [M+K]+ are also observed with much less abundance. Alternatively, muropeptides can be easily analyzed in the negative ion mode without any special sample preparation. The major i peak in the negative ion MALDI mass spectrum spectrum (Figure 3.3(b)) is the deprotonated molecule [M-H]‘, accompanied by a [M+Na-2H]' ion peak of very low intensity. The simplicity of the negative ion MALDI mass spectra allows unambiguous molecular weight determinations of unknown samples. In the positive and negative MALDI spectra for the pentaglycine—substituted monomer shown in Figure 3.4, [M+Na]+ and [M-H]' ion signals, respectively, are Predominant. Some ion signals from prompt fragmentation (i.e., fragmentation occurring in the ion source) (25), denoted as [M-GlcNAc+N3]+ and [M'GICNAC‘HL resulting from L—_ (irNAc-MrrNAc (I) GlcNAc—l hi2: Muropeptides d WWW aureus St Muted monomer (c) Worn were used: Glc Marine; iGln, isoglutami 60 G1 cNAc -—MurNAc G1 cNAc —MurNAc Ala Ala l | iGln iGln Li's Li's —(G1y)5 I | Ala Ala I | Ala Ala (60 (b) _GlcNAc —MurNAc 7 GlcNAc -—MurNAc ALI Aia iGlln iGlln Li’s—(61305 Lylls —(G1Y)5 Alla Alla " ‘ n Aia (C) Figure 3.2: Muropeptides derived from peptidoglycan of the highly methicillin—resistant Staphylococcus aureus strain COL. (a) unsubstituted monomer; (b) pentaglycine- Substituted monomer (c) muropeptide oligomers (n = 1, 2, 3, ...). The following abbreviations were used: GlcNAc, N-acetylglucosamine; MurNAc, N—acetylmuramic acid; Ala, alanine; iGln, isoglutamine; Lys, lysine; Gly, glycine. [MrGlcl I 31323: Linear MALD W111 (a) positive (b) hunt of analyte load 61 i [M+Na]+ l a 991 r! .43 (I) c: ; o 1 E 1 H o .2 E. + g [M'GICTAC+Na]+ [M+I-Il] [M+K]+ r r r r 700 800 900 1000 l 100 m/z [M-Hl' b 967 .Q (A c: 0.) +4 c: F—1 4) .2. H 7‘: [M+Na-2H]‘ g 04 [M-GlcNAc-H]' 1 f A L“ A __J I, r r T —T —F ‘: 700 800 900 1000 1100 m/z Figure 3.3: Linear MALDI mass spectra of an unsubstituted monomer Obtained in (a) positive (b) negative ion mode with CMBT as matrix. The amount of analyte loaded on the sample plate was ca. 2 picomoles. — no ammonium“ well only small rm padlhermsubstimted item are likely to b figurednin rim coelul lhpniously been (181 W all elute withil idler-lee orrheee pee ens glycine substitut With much lower st “Mm (FAB-M mm and dimers. l Ihalide oligomers larg Wage in the detenninati lithe positive ion MAL NM. [M+Na]+ ion “peptide oligomer. Th lists with the size of fillers contain more te We separated by 11 62 the loss of N-acetylglucosamine (GlcNAc) residue, are also observed in the spectra. In comparison, only small fragment peaks for the loss of GlcNAc were observed in the spectra of the unsubstituted monomer (Figure 3.3). Other peaks labeled with an asterisk in Figure 3.4 are likely to be related muropeptide contaminants differing in the length of the glycine chain that coeluted in the HPLC fraction, rather than mass spectral fiagments. It has previously been demonstrated that penta—, tetra-, and tri- glycine substituted muropeptides all elute within one minute of each other under these HPLC conditions (10). The m/z values of these peaks in the spectra correspond to those expected for the tetra-, tri-, and di- glycine substituted muropeptides. With much lower sensitivity compared to MALDI-MS, fast atom bombardment mass spectrometry (F AB-MS) is able to determine the molecular masses of muropeptide monomers and dimers. However, it is difficult to obtain signals by FAB-MS for muropeptide oligomers larger than dimers. MALDI-TOF instruments have an inherent advantage in the determination of higher mass oligomers. This is illustrated in Figure 3.5 with the positive ion MALDI mass spectra of muropeptides ranging from a dimer to a pentamer. [M+Na]+ ion peaks are the dominant species in the spectrum of each muropeptide oligomer. The intensity of the fragment ion peak, [M-GlcNAc+Na]+, also increases with the size of the muropeptides. This is not unexpected given that higher oligomers contain more terminal GlcNAc groups. Between adjacent oligomers, [M+Na]+ peaks are separated by 1164 mass units, which characterizes the repeating units of the Relative intensity no.4: Linear MALDI lhiive and (b) negativ “onthe sample plate “tinned by 57 mass on ”hill; di- glycine substit 63 [M+Na]* a 1276 U; U) a 8 .5 Q) .2 3§ [nHaNer 3.4) [M-GlcNAc+Na]+ [M+H]+ lM+Kl+ >I< * * l r— ' I I 1000 1100 1200 1300 m/z [M-Hl' b 1252 Relative Intensity [M-GlcNAc-H1' * [Mama—2H]- W l l I r 1000 l 100 1200 1300 Figure 3.4: Linear MALDI mass spectra of pentaglycine-substituted monomer obtained in (a) positive and (b) negative ion mode with CMBT as matrix. The amount of analyte loaded on the sample plate was approximately 2 picomoles. Peaks labeled with an asterisk are separated by 57 mass units representing muropeptide impurities in the sample that are tetra-, tri-, di- glycine substituted. ‘ ‘ [Hr-‘1] fthis: Composite pos hummer to pentamer ”ill plate was between 1 64 Dimer [M+Na] + _ 2440 Trimer lM+Na1 + 3604 Tetramer [M+Na] + 4768 Pentamer [M+Na] + 5932 + +1 [M-GlcNAfiNa] [M-GlcNAc+Na] + [M-GlcNAc+Na] \ L [M-GlcNAc+Na] 2000 3600 4600 5000 6000 m/z Relative Intensity Figure 3.5: Composite positive ion linear MALDI mass spectra of muropeptides ranging from monomer to pentamer with CMBT as matrix. The amount of analyte loaded on the sample plate was between 1 to 5 picomoles. anion structure. :Ltlllysis by MALDI-l lltLDI-PSD fragn readies for structural miles of the analyte sar nnsn Analysis 0“” hpositit'e ion mod Tern for an unsubstit‘ 13310th protonated n iiizin Figure 36, all linens All peaks ll fined before in the anal 3 arise by the loss of thr 12920, 849, and 721, henclatural conventions iitnechanism of formant lion] group after the nntories (26, 2330) Ti :30 spectrum T . ~ hese lOn: :nd the sequence of the 65 peptidoglycan structure. 3.2. Analysis by MALDI-PSD MALDI-PSD fragment ion analysis has been performed on the monomeric muropeptides for structural elucidation. The sensitivity was such that typically only 1-10 picomoles of the analyte sample were loaded on the sample plate for the PSD experiment. MALDI-PSD Analysis of an Unsubstituted Monomer In positive ion mode analysis, the [M+Na]+ ion at m/z 991 was selected as the precursor for an unsubstituted muropeptide monomer because of its high abundance relative to the protonated molecules, [M+H]+. The post-source decay (PSD) spectrum is shown in Figure 3.6, along with its corresponding structure and fragment mass assignments. All peaks labeled in the spectrum are sodium-containing species. As reported before in the analysis of alkali metal-cationized peptides, some major fragment ions arise by the loss of the C—terminal amino acid residues (26). These ions, including m/z 920, 849, and 721, are designated as [bn.l+Na+OH]+ according to current nomenclatural conventions (26-27), but are labeled as “b+18” in the figure for simplicity. The mechanism of formation of ions of this type, involving retention of the C~terminal hydroxyl group after the rearrangement, has been described in detail by several laboratories (26, 28-30). Two [an+Na-H]+ ions at m/z 874 and m/z 546, also appear in the PSD spectrum. These ions, along with two [bn+Na-H]+ ions at m/z 331 and m/z 703, provide the sequence of the peptide portion of the molecule. Another prominent fragment n35. n:7_~ amHvOHnZ 75003 — 00 l0~uu0\ 0: ... — E Z .T (I: :0 i :0:- o Hasn‘t HOE 66 . .mBoSooE N nquNOM 638mm HEAD “5.52 38288 @8352st on mo 836on $9: QmmrHQASZ now 3.58% 66 among 0th 2% 2.6 fie NE .2: ewe : ,_ (22w n mg 33% 86 Keoeo Sm . 8m wee an 6H» H mm. 08 W. A 9 Wmumwvgm ”MSW—iv awn mmmmwvfi Nb m1 neg m sea a ma .me teem m mololeulszlruiEz . llmollmszlfimoolmulmzl ..Pmullmzlolmulow S a a ._._ _ a a a. _ w. _ a 1. .6 £28 6 ,6 ..e ..M _ as. am. 50 L535 it til 788, resulting ill: moiety in the moi groin also accompani mom/2920,1174, 84‘. 1:17.611, 646, 628, 51‘ Inuhande of this mc mimicry. lLDI-PSD Analysis of a 3 lie PSD spectrum m; attached via the 8 4 loomed monomer. T 111.1134, and 721 from Litedniigureftl. Se til-Hi ion at m/z 703 531th: fragment ion at j‘tleutral loss of GlcNAc 3111101 the pentaglycine in, 885, 517, and 4‘. W along the pentagl‘ StIuctural informat hid in the negative in "1 selected as the pm" f 67 ion at m/z 788, resulting from a loss of 203 mass units from [M+Na]+, confirms the GlcNAc moiety in the molecule. This fragmentation process, resulting in a loss of 203 mass units, also accompanied most of the peptide sequence fragments described earlier, including m/z 920, 874, 849, 831, 721, 703, and 546, producing a related series of ions at m/z 717, 671, 646, 628, 517, 499, and 343. The fragment peak at m/z 449 corresponds to the disaccharide of this monomer, with the loss of the three carbon unit at C-3 of the muramyl moiety. MALDI-PSD Analysis of a Pentaglycine-substituted Monomer The PSD spectrum of the [M+Na]+ ion of a monomer containing a pentaglycine moiety attached via the 8 amino group of the lysine residue is very similar to that of unsubstituted monomer. The peaks representing [bn.1+Na+OH]+ (“b+18”) ions at m/z 1205, 1134, and 721 from the stem peptide are also prominent in the PSD spectrum as illustrated in Figure 3.7. Several [a,,+Na-H]+ ions at m/z 1159, 1088 and 546, as well as a [brrtNa-H]+ ion at m/z 703 filrther characterize the peptide sequence of the molecule. As before, the fragment ion at m/z 1073, originates from [M+Na]+ after a loss of GlcNAc. The neutral loss of GlcNAc (-203) also occurred for most of the fragment ions described earlier for the pentaglycine-substituted monomer, resulting a series of ions at m/z 1002, 931, 956, 885, 517, and 499. It is interesting to note that no obvious fragmentation was observed along the pentaglycine chain in the positive ion PSD spectrum. Structural information concerning the pentaglycine moiety in this muropeptide can be found in the negative ion PSD spectrum, as shown in Figure 3.8. Here, the [M-H]' ion was selected as the precursor. Most of the fragments in the negative mode PSD spectrum a ...u a ...w a .5“ a .zzow IAII zl..:u|mllv9.9v I ova“ 5 +723): £003.: IO.=U\ .mofioaooa N ”magma: oanm MEG/HO “x532 .8882: wBBumnzmofiobwfiaom a mo 55.50on 3%: 05%ng no“ 338m Kan 9:.me N 000 H \E OO.N H F Omvw £0 . mo» mmv H MN H 3% H j £57, £2 03 3w \ n m QmHHL—VMHH fl Hmm HNB NOO H mm MBOH mo NH mmfiivmofi 2&9va— fizzfi :82 can: u 88 .32. $3 :0 IO x+MU ImHz W_+mnmlmz any ImHHU LIIZJ+@ InAumHUv ImHnW lHZleTm—UIIHZ I0 IHHUIIUn m .5 £on 0 .mo_ x _ game obNH _ n milk—5 lolAbOv Ilz: 58:2 +HNZ+H>HH r. .3"on K118191111 QAlmlafl .mofioaooa N ”@363 293m #920 ”gag .Hoaonoa “633333-030waqu a mo 888on gm: QmmADfiQ/H :8 03332 "QM 253 NE 82 8: 82 8m com com coo _ r r r k _ '7 4 e: Sm 3% En . mm: wok my: SSQSVNS com a? N m m ‘82 :8: am :am 2:», Ei « moluuqimulmZiolmulmz _|mulmz.Tu_x.,+£muvlmulmE+_u_lmu_lsz linoleum x .L r L. r mm o “EL owe lw_ _ ,3 ”may .... ex NmzlwfiloiTEIwU lot—Q harm—0+.bo .TF 5 zmounmu a x. r. x. ..a .. ,, 85% 8 mafi_ an: 5.: 3.: ea / m $2 mo o .54.): moumu . womb $2...- 51131191111 oAneroH fi probe formed wit] pm losses of: hauugfiagmem ion MWWM inredbyBowie er al. him peaks at ml: #lmkbone. Thelo: mwthedmvagr udsryr,MurNAc, w: imafZOSmassunit Winthenegativeit lupnimoethisobserva mus/us analysis ( “PM. We have employe “We characterize oli Worms aureus (33 Metro liberate N-termi We) from Sraphylococ f 70 appear to be formed with charge retention at the C-terrninal end with respect to the glycine chain. Losses of one to five glycine residues occurred from the [M-H]' ion, and the resulting fragment ion peaks were found at at m/z 1195, 1138, 1081, 1024, and 967, confirming the pentaglycine sequence. This type of fragmentation agrees well with that observed by Bowie et al. (31, 32). A similar fragmentation mechanism also produces fragment ion peaks at m/z 771, 700, and 572, providing the primary sequence of the peptide backbone. The loss of 203 Da from [M-H]', resulting in a fragment at m/z 1049 corresponds to the cleavage at the glycosidic bond and the loss of GlcNAc. The loss of the second sugar, MurNAc, was indicated by a fragment ion peak at m/z 844, representing a difference of 205 mass units from the peak at m/z 1049. The extent of fragmentation was diminished in the negative ion mode PSD analysis, relative to that in the positive mode. In our experience this observation is typical for other peptides and muropeptides in negative ion mode MS/MS analysis (including PSD). 3.3. Lysostaphin Digestion and Mass Mapping Analysis MALDI-PSD analysis of higher oligomers does not provide complete sequence and cross-linking information. These spectra can also be rather complex and difficult to interpret. We have employed a specific protease, lysostaphin, combined with MALDI-MS alyses to characterize oligomeric muropeptides. Lysostaphin is an antibiotic that lyses taphylococcus aureus (33). The active principal is enzymic in nature and has been eported to liberate N—terminal glycine and alanine (but not C-terminal alanine in the stem eIl’tide) from Staphylococcus aureus cell wall (34). It was also observed that — ”brine cross-linked iiiunque property, 1y ins, and was found t Two nmropeptidc r m for each was so thin The first rm Wear wens stn 1M! podfive and neg: Mmofthe lysostaphin lhspwtmn correspond Hinge at different Wedge, this is the first &510000in of lysostaphir tins, based on the int F11193.9(») are consisten “hand observed values ( ll lle most favorable hyr ’lllnsrepresented by prt Wiring both sides 0 W061 composition infor mglycine moiety with a f —»— 71 pentaglycine cross-linked muropeptides can be hydrolyzed by this enzyme (34). Because of this unique property, lysostaphin was used in the staphylococcal muropeptide oligomer analysis, and was found to be particularly useful for characterizing glycine cross-linked oligomeric isomers. Two muropeptide dimers were subjected to lysostaphin digestion and the reaction mixture for each was subjected directly to analysis by MALDI-MS without fiirther purification. The first muropeptide dimer, I, was from the highly methicillin-resistant Staphylococcus aureus strain COL. The molecular weight was determined to be 2417 Da by both positive and negative mode MALDI-MS. The negative ion mode MALDI spectrum of the lysostaphin digest of muropeptide I is shown in Figure 3.9(a). The peaks in the spectrum correspond to the product of hydrolysis, i.e., smaller muropeptides, after the cleavage at different sites between the Gly-Gly bonds of the dimer. To our knowledge, this is the first time that direct mass spectral evidence has been obtained for the specificity of lysostaphin cleavage of muropeptide oligomers. The favorable sites of ydrolysis, based on the intensities of peaks in the MALDI mass spectrum of the digest igure 3.9(b)) are consistent with those predicted from previous studies (34). Calculated alues and observed values of [M-H]' for resulting muropeptides are summarized in Table .l. The most favorable hydrolysis is at site A, as indicated in Figure 3.9(b), producing 0 ions represented by prominent peaks at m/z 1195 and m/z 1238 in the spectrum, haracterizing both sides of the pentaglycine bridge. Other peaks also provide the emical composition information of the cross-linking pentaglycine bridge and the entaglycine moiety with a free N-terrninal, attached to one of the monomer‘units of the Relative Intensity Lls—Gry—Gn 1 Ali. [aa' Figure 3.9: (a) Nega digest of (b) muropept Was used directly for Ileanows in Fig. 3.9 leirtensities of peak lllest Thicker arrows 72 a 1195 A? m 8 3 1238 d) ,> E 1081 11 8 34’ 1067J l 1124“ 1181 km 1 l l f fl 1000 1100 1200 1300 1400 m/z .G—|\I/| G-M Alla F E D Ala C B A Glln i l Gin l y Lys—Gly—Gly Gly—Gly Gly I Lys—Gly—Gly-GIy-Gly-Gly-Ala l .019- LAlal Figure 3.9: (a) Negative ion MALDI mass spectrum of the lysostaphin digest of 0)) muropeptide dimer I. Matrix: CMBT-The digestion mixture was used directly for MALDI-MS analysis without further purification. The arrows in Fig. 3.9b indicate the favorable sites of hydrolysis, based on the intensities of peaks in the MALDI mass spectrum of the lysostaphin digest. Thicker arrows indicate more favorable hydrolysis sites. — lupeptide. The absem intrusion that lysor mldydne residue of When this strain to in: the rmropeptide rind with muropeptid allAlDI-MS. Infor lweplide led to the (1 non of the lysostaph 121181 and m/z 1238 @518 obtained alter . 3100)). The masses of lltlrrrodon the left side or We peaks, as sumo lumen with PSD 105mg other types of s mIopeptide oligomers fro 73 muropeptide. The absence of peaks with a molecular weight less than 1024 Da supports the conclusion that lysostaphin is unable to hydrolyze the bond between the first and second glycine residue of the cross-linking bridge. When this strain was grown in the presence of high concentrations of exogenous glycine, the muropeptide composition of peptidoglycan changed (35). The dimer I was replaced with muropeptide dimer 11 having a molecular weight of 2403 Da as determined by MALDI-MS. Information obtained from the lysostaphin hydrolysate of this muropeptide led to the determination of its structure. The negative ion MALDI mass spectrum of the lysostaphin digest is shown in Figure 3.10(a). Two prominent peaks at m/z 1181 and m/z 1238 were detected in the MALDI-MS spectrum, representing the fragments obtained after cleavage at the most favorable hydrolysis site (site A, Figure 3.10(b)). The masses of these peaks indicate that amino acid substitution (Ala —> Gly) occurred on the left side of the muropeptide dimer as illustrated in Figure 3.10(b). Other hydrolysate peaks, as summarized in Table 3.2, also verified this anrino acid substitution. In conjunction with PSD analysis, lysostaphin digestion may be used in the firture for probing other types of structural modifications occurring in the glycine cross-linked muropeptide oligomers from Staphylococcus aureus. Uhmv thnuxvcnu—px owcaoaorSE wast—33M $93 :73: 833 -:._-_>: unncxxhv _ ..3 Z.th—S:< WE..QQQ~\/H WWNE NOiHé ..H .M. DHDfirfi — ..U:-A— 0—u.aA_Un_C...u—\/_ Q: ..Irnvnd..fl— :.:—A_: 74 $3 $3 m 98 < nooEom m was < vm: em: m was < noogom m was < S: S: Qw§fiom Qc§< $2 $3 an 0.8 fiwa $3 $3 86 0 mo flog U 82 32 an moo Ham mm: mm: 87. mac flog m $2 mm; so 50 Ema mm: mm: 8% «Co God < $38 -Eéé A038 -Hméé 023382: wafizmom BE owgmfio H 8E5 395832 mo “momma :Emfimoaq mo £93224 waqamz mmmz HQ§ ”Tm 2an l2 oUCQoQOSE chsmoM vim 093520 : ..U—plflr— 01.3.".020932 ..0 .Wnumwmflr— ..I—Q-SWOWXVHHQ w..w>.:w:< WEHQQNZ WWW: HQ1~§ ..N.m11 QHQNrH. 76 3.88% “on “dd .802 $3 $3 m 98 < nooEom m was < em: vim: m 98 < mooEom m can 4 3m: 3w: Qw§ 90% by weight. They occur in all forms of eukaryotes and have firnctions that span the entire spectrum of protein activities, including those of enzymes, transport proteins, receptors, hormones, and structural proteins. The polypeptide chains of glycoproteins, like those of all proteins, are synthesized under genetic control. Their carbohydrate chains, in Contrast, are enzymatically generated and covalently linked to the polypeptide without the rigid guidance of nucleic aid templates. The processing of proteins by glycosyltransferases is rfien incomplete leading to the synthesis of non-uniform glycoprotein products. ilycoproteins therefore have variable carbohydrate compositions, which is reflected by the rultiple carbohydrate variants at a single glycosylation site. This phenomenon, known as licroheterogeneity or heterogeneity, has complicated purification and characterization of ycoproteins. Carbohydrates form two types of linkage to proteins: (1) through an asparagine 1e chain (Asn- or N-linked), (2) through the side chain of serine (Ser) or threonine (Thr) linked). Sequence analyses of glycoproteins have led to the following generalizations tut the attachments of carbohydrates in glycoproteins. loll-glycosidic ( :rtiniinked to the 1 221m sequence Asn shoreline (Pro) or a 22s hinges usually 13110 either mannose mum, be linked 1c rnoigosaccharides is 111 1114.1: N—linked 01 111111 B-N‘acetylglu‘ 82 In N-glycosidic (N-linked) attachments, an N-acetylglucosamine (GlcNAc) is invariably B linked to the amide nitrogen of an asparagine (Asn) residue of the polypeptide chain in the sequence Asn—X-Ser or Asn-X-Thr, where X is any anrino acid residue except possibly proiine (Pro) or aspartic acid (Asp) as shown in Figure 4.1. The oligosaccharides in these linkages usually have a distinctive core whose peripheral mannose residues are linked to either mannose (Man) or N-acetylglucosamine residues. These latter residues may, in turn, be linked to yet other sugar residues so that an enormous diversity of N- linked oligosaccharides is possible. (3) OH O NH’C_CH2_ 9H Asn OH C = O HO ,1 NHCOCH3 Ser IO]. Thr GlcNAc (b) Manor (1 —‘> 6) > ManB -— GicNAc(I "" 4) GlcNAc—— Manor ( l —> 6) Figure 4.1: N-linked oligosaccharides. (a) All N-BIYCOSidiC protein attachments occur 1hr Ough a B-N-acetylglucosamino-Asn bond in which the Asn occurs in the sequence Asn- g-ltllll (bi N‘hnked :1: stature. Smures of N-ii 313$an comp 1.1:: Mandi-8012m- :: ranches (Figure 1 ;::;_'- u-mannose res gzachandes contain ' Titpe srructures ha ~431sccharides. and mos .l-‘. to the B—iinked manr The most commo 3371 l-galactosyi-(l -3)-( 11101 threonine as shr 11111101“, Q-O-glycos? 22111101116 S-hydroxyl 1111101131 generalizatior illl'ary in size from nitride units in pr. 83 X-Ser/T hr. (b) N-linked oligosaccharides usually have the branched (Man)3(GlcNAc)2 core structure. Structures of N-linked oligosaccharides fall into three main categories termed high-mannose—type, complex-type, and hybrid-type. They share the common core structure Manorl—>3(Man0t1-—>6)ManBl—>4GlcNAcBl—>4GlcNAc-Asn, but differ in their outer branches (Figure 4.2). High-mannose-type oligosaccharides have two to six additional a-mannose residues attached to the core structure. Typical complex-type oligosaccharides contain two to four outer branches with a sialyllactosamine sequence. Hybrid-type structures have the features of both high-mannose—type and complex-type oligosaccharides, and most of them contain bisecting N-acetylglucosamine, which is linked Bl-4 to the B-linked mannose residue of the core structure (Figure 4.2). The most common O-glycosidic (0—linked) attachment involves the disaccharide core B-galactosyl-(l-3)-01-N—acetylgalactosamine or linked to the OH group of either serine or threonine as shown in Figure 4.3. Less commonly, gaiactose (Gal), mannose, or xylose form or-O-glycosides with serine or threonine. Galactose also forms O-glycosidic bonds to the 5-hydroxyiysy1 residues of collagen. However, there seem to be few, if any, ' additional generalizations that can be made about O-glycosidicaily linked oligosaccharides. They vary in size from a single gaiactose residue in collagen to the chains of up to 1000 disaccharide units in proteoglycans. _ _ _ a _ _ __ coca 1113 m \1 EN . Ill haw a< l ”220.0: All a 03)-0t-N-acetylgalactosyl Ser/Thr Figure 4.3: Most common O-glycosidic attachments of oligosaccharides to glycoproteins. The carbohydrate chains of glycoproteins have begun to be widely acknowledged as having integral roles in the fiJnctional properties of glycoproteins. N—linked carbohydrates have been implicated in a wide variety of fimctions including modulation of biological activity, recognition between cells, interaction between host and pathogen, and distribution in tissues (1-4). The profound effects that carbohydrate moieties can have on other properties of glycoproteins such as solubility, antigenicity, immunogenicity, circulatory half-life, resistance to proteolysis, and thermal stability have come under increasing scrutiny with the desire of the biotechnology industry to produce these molecules for therapeutic use in humans (4-5). Although the biological and functional roles of O-Iinked carbohydrates are less well understood, it has been reported that O- glycosylation is necessary for efficient biosynthesis and secretion of certain glycoproteins such as erythropoietin and for prevention of antibody-based clearance of recombinant human granulocyte-macrophage colony-stimulating factor (1). For production of glycoproteins in recombinant systems, the structure of the protein, the specific cell type employed, and the bioprocess environment all influence BMW“ dire. (Hymproteir uniofspecific oli Imldmafrom muwwum inundation of thei liliympmtein ally: in meat years, u m of increasing in Hogital firnctions and Manual to this undi Mylation, determine ‘ hmbohydrate moieti lithe roles of the Eris partly caused by i mull quantifies Wbohydrates, such as require large amounts of haircare of carbohyd "l fluently being dlithography with P“ ill matrix-assisted 1 86 which glycosylation sites are utilized and the range of carbohydrate structures found at each site. Glycoprotein bioactivity or clearance can be affected by modification or removal of specific oligosaccharide structural classes from specific oligosaccharide structural classes from specific attachment sites. Thus the precise structures of the glycoforms, as well as their protein sequence context, must be understood in order to gain a fill] appreciation of their structure/function relationship. 1.2 Glycoprotein analysis Inrecent years, understanding and elucidating the role of glycosylation of proteins has been of increasing interest due to this posttranlational modification's influence over biological fimctions and implications on the therapeutic usage of recombinant proteins. Fundamental to this understanding is the necessity to recognize the presence of protein glycosylation, determine the site(s) of protein glycosylation, and elucidate the structure of the carbohydrate moieties. In many proteins, however, the carbohydrate heterogeneity and the roles of the carbohydrates are diverse, complex and not very well understood. This is partly caused by lack of information because many glycoproteins are available only in small quantities and because traditional methods for structural analysis of carbohydrates, such as NMR and fast atom bombardment (FAB) mass spectrometry (MS), require large amounts of the glycosylated protein and mainly show the major components in a mixture of carbohydrates (6). More sensitive methods for analysis of carbohydrates are currently being developed, including high-performance anion-exchange Chromatography with pulsed amperometric detection (7), capillary. zone electrophoresis (8), matrix-assisted laser desorption ionization (MALDI) MS (9-12), liquid maphy/electros mam analysis MALDI-MS has Manon). MA W which an 1 ma With know] we glycosylation machined from prc immune: Information about N from mass difi‘ere Ir-rridyoosidase). For Ellie MALDI-MS has in further developed iitlycoridnse mixtures in set of end product llithography to obtai lhflnilsttucture. This a “liming methods (ext 87 chromatography/electrospray ionization (ESI) MS (13-16) and enzymatic methods such as the reagent array analysis method (17). MALDI—MS has been gaining increasing acceptance for the detection, localization, composition determination, and sequence analysis of carbohydrates present in a glycoprotein (9—12). MALDI-MS experiments are typically performed on HPLC separated glycopeptides which are prepared from glycoproteins by proteolytic digestions or chemical methods. With knowledge of a protein’s primary structure available from DNA sequences, glycosylation modifications can often be inferred on the basis of anomalous masses obtained fi'om proteolytic peptides. Glycopeptide fractions also can be identified by their microheterogeneity as revealed in MALDI-MS spectra. Information about the type of modifications present and their compositions can be gained from mass differences after digestion with a specfic glycosidase (endoglycosidase or exoglycosidase). For example, a mass shift of 291 Da corresponds to a sialic acid residue. MALDI-MS has been exploited to monitor the degradation of glycopeptides by mixtures of specific exoglycosidases (9-10, 12, 18). The concept of using highly specific exoglycosidases sequentially to remove monosaccharide residues from oligosaccharides was first shown by Mizuochi et al. (19). A related non-mass spectrometric approach has been further developed into the reagent-array analysis method whereby a set of exoglycosidase mixtures are incubated with a chemically radiolabeled oligosaccharide to give a set of end products (17). These are pooled and analyzed by Bio-Gel P4 column chromatography to obtain a specfic fingerprint which can lead to identification of the original structure. This approach is dependent on isolation of oligosaccharides using time- consuming methods (extensive dialysis, hydrazinolysiS, radiolabeling, and a range of maphicandelei em moiety. Widesontheh limptrrun). Furt inmofisom mm In con with: and quick 11 Mon of exoglyc maldetails of olig lllmonhinant human imllopmteinases (18 ll Fractionation ant Meptides (20—22). White cells and/or l‘lmatography, the mi 88 chromatographic and electrophoretic purification steps) which also result in destruction of the protein moiety. In addition Bio-Gel P4 chromatography, which separates oligosaccharides on the basis of hydrodynamic volume, is a relatively slow technique (6 to 24 hours per run). Furthermore, as it is a low-resolution approach, it is possible work only with pure oligosaccharides and interpretation becomes difficult even when there are binary mixtures. In contrast, analogous MALDI-MS based procedures provide a more sensivitive and quick means for detection of the degradation products (18). The combination of exoglycosidase digestion and MALDI-MS has been used to obtain structural details of oligosaccharides derived from glycoproteins such as bovine fetuin (10), recombinant human macrophage colony stimulating factor (12), and tissue inhibitor of metalloproteinases (18). 1.3 Fractionation and structural assessment of glycoconjugates by lectin immobilized column chromatography The use of lectin immobilized column chromatography has been recently evaluated not only for the purification of glycoproteins with pathologically altered sugar chains but also for separating and determining structure of glycoprotein-derived oligosaccharides and glycopeptides (20-22). Lectins are proteins or glycoproteins of non-immune origin that agglutinate cells and/or precipitate specific complex carbohydrates. Different fiom fractionation techniques such as gel permeation, ion-exchange, partition, reversed-phase and normal-phase chromatography, lectin column chromatography is based on the carbohydrate-binding specificities of lectins. In typical lectin affinity column chromatography, the mixture of glycoconjugates to be resolved is applied to the wind lectin co hind oompOnents WWW lllll‘. specific hapten The first trial 0 Wed Asn-linked illulurose column myl residues m hrirwlumn This fin hell-linked sugar ch multitude, Manor in Since the core ha dun Because the f Iii still interact with hides the interactio lit sugar chain binds titer chains pass th Oligosaccharides, whici lindmore tightly to a liphrose column, 0 Membrane of polyo Wyndchod in sialic 89 immobilized lectin column. The glycoconjugates of interest are selectively bound to the lectin and components without affinity for the lectin are then washed away. The bound carbohydrate containing components are dissociated from the lectin by competitive elution with the specific hapten sugar for the lectin. The first trial of a lectin column for the structural study of the sugar chains of glycoproteins was reported by Ogata et al. (23). Study of the behavior of various radiolabeled Asn-linked oligosaccharides and oligomannosides in a concanavalin A (Con A)-Sepharose column revealed that at least two nonsubstituted or 2-0-substituted or- mannosyl residues must be present in a sugar chain to be retained on the immobilized lectin column. This finding was particularly useful for the structural analysis of complex- type N-linked sugar chains. For example, all complex-type oligosaccharides contain a pentasaccharide, Manoel—->6(I\/Ian0t1—->3)Man[31——>4GlcNAcB1——)4GlcNAc as a common core. Since the core has two or-mannosyl residues, this core binds to a Con A-Sepharose column. Because the first outer chain is linked at the C-2 position, the mannose residue can still interact with Con A. However, substitution with two outer chains totally abolishes the interaction of the oc-mannosyl residue. Therefore, a biantennary complex- type sugar chain binds to a Con A-Sepharose column, but those with more than three outer chains pass through the column without interaction. High mannose-type oligosaccharides, which have many Manorl—> residues and ~92Manocl—> residues, will bind more tightly to a Con A-Sepharose column. With use of this character of a Con A- Sepharose column, Ogata et al. found that the glycopeptides, obtained from the plasma membrane of polyomatransfered baby hamster kidney cells by pronase digestion, are not only enriched in sialic acid residues, but contain more branched structures than those from ”my cells (24 mm complex- ll'yrlso reported that rim with lOmM a. Ileduted with 250 Complex-type huyl reddue linked immosyl core. imsyhted at their cor merely be separat illobilized Aleuria 4 mireosylated core p: iii): with a fucosylatr linM fucose (Fuc) hatching nor by the pi tube used as a more hmcbilized lentil and liked to the core por hiding. The Fucotlt Gill-93(Fuwl-e4) FluteZGalBHBGl GlcNAc groups show With more than two 0 90 baby hamster cells (24). Later, Narasirnhan et al. added the evidence that bisected biantatennary complex-type sugar chains do not bind to a Con A-Sepharose column (25). They also reported that biantennary complex-type oligosaccharides can be eluted from the column with IOmM or-methylglucopyranoside, and high mannose-type oligosaccharides can be eluted with 250 mM or-methylmannopyranoside. Complex-type sugar chains can be classified by the presence or absence of an or- fucosyl residue linked at the C-6 position of the proximal N—acetylglucosamine residue of the trimannosyl core. It was found that some of the hybrid-type sugar chains are also fucosylated at their core (26). The fircosylated complex-type and hybrid-type sugar chains can easily be separated from their nonfucosylated counterparts by passing through an immobilized Aleurz'a aurantia lectin (AAL) column (27). Sugar chains with a nonfucosylated core pass through the column without interaction. In contrast, sugar chains with a fucosylated core bind to the column and are eluted with a bufl‘er containing 0.5 mM fucose (Fuc) (28). Because the binding is affected neither by the status of branching nor by the presence of bisecting N—acetylglucosamine residue, the fractionation can be used as a more reliable method to detemrine the fucosylated sugar chains than immobilized lentil and pea lectin columns. In these cases, not only the fircosyl residue linked to the core portion, but the structure of the outer chains are required for the binding. The Fucocl—>2Gall—>4GlcNAc, the GalB1-—>4(Fucoc1——>3)GlcNAc, and the GalBl-—>3(Fucocl—>4)GlcNAc groups weakly interact with an AAL column, whereas the Fucorl—>2GalBl—>3GlcNAc, and the GalBl——>4GlcNAcBl——>3GalB1—94(Fucocl+3) GlcNAc groups show almost no interaction with the matrix. However, oligosaccharide with more than two of these groups bind more strongly than those with only one of the we lherdore, manhforthe Beensemost more use! fine widely foun mm Many a tuning specificity . In hesively Sr mhlmmine resid Wdes with magnum (E-PP video: and indicated 1 Another variatic though mono- and inched complex-typ litmuharides in a it GalBl—>4GlcNAc “limit, as long as t 0l'louiccharidcs that c 91 groups. Therefore, care must be taken in interpreting the results of affinity chromatography for the structural determination of oligosaccharide. Because most of the N-linked sugar chains of glycoproteins are included in the complex-type sugar chain subclass, lectins that interact with the B-galactosyl residues, which are widely found in the outer chain moieties of complex-type sugar chains, are of particular use. Many galactose-binding lectins have been reported (29-30). Among them, the binding specificity of Ricinus communis agglutinin I (RCA 1) has been investigated most intensively. Several complex-type oligosaccharide contain a bisecting N- acetylglucosamine residue. Irimura et al. reported that bisected biantennary complex-type oligosaccharides with the GalBl—)4GlcNAc[31—> outer chains binds to an erythroagglutinin (E-PHA)-agarose column (31). Cummings and Komfeld confirmed this evidence and indicated that none of nonbisected sugar chains will bind to the column (3 2). Another variation in the complex-type sugar chains is in their branching pattern. Although mono- and biantennary sugar chains can be separated from other highly branched complex-type sugar chains by a Con A-Sepharose column, tri- and tetraantennary oligosaccharides cannot be separated by this column, An immobilized Datura stramonium agglutinin (DSA) column can be effectively used to fractionate these highly branched oligosaccharides (33-36). A study of the behavior of 34 complex—type oligosaccharides in a DSA-Sepharose column showed that oligosaccharides that contain the GalBl——>4GlcNAcB1—>4 (GalBl—>4GlcNAcBl—>2)Man group are retarded by the column, as long as the pentasaccharide group is not substituted by other sugars. Oligosaccharides that contain unsubstituted GalB1——)4GlcNAcB1—->6(GalBl—>4GlcNAcB 19le group bin: nipmaamine oligr tutu: of the inner ( rile Therefore, Wider into t Midas; the “r tthnrlr; and the “bc lldarrlteu‘aanteuna One lectin of tl Who), which has ill-lilN-acetylgalactt Mine exists, let lthique for isolating Willis sites of attacl ligmccharides by ja lIsland or inhibit 92 1—>2)Man group bind to the column and are eluted with buffer containing N- acetylglucosarnine oligomers. The binding characteristic was not affected by either the structure of the inner core portion or by the presence of bisecting N-acetylglucosarnine residue. Therefore, the column can fractionate a mixture of complex-type oligosaccharides into three fi'actions: the “pass-through fraction” contains biantennary oligosaccharides; the “retarded fraction” contains triantennary oligosaccharides, with a C- 2,4 branch; and the “bound fraction” contains triantennary oligosaccharides, with a C-2, 6 branch and tetraantennary oligosaccharides. One lectin of the greatest interest is jacalin, a lectin fiom jackfruit (Artocarpus integrifolia), which has high specificity and affinity for the core ‘ disaccharide, galactosyl [BU—>3) N-acetylgalactosamine, in O—linked oligosaccharides (37-3 8). As no universal 0- glycosidase exists, lectin aifinity chromatography using jacalin-agarose is a usefirl technique for isolating glycopeptides containing O-linked oligosaccharides and thereby localizing sites of attachment of these oligosaccharides. The specific binding of O-linked oligosaccharides by jacalin coupled to agarose has been exploited for affinity purifications of IgA and C1 inhibitor from human plasma. Recent characterization of the binding of glycoproteins to jacalin-agarose indicates that this immobilized lectin specifically binds other proteins that contain of O-Iinked oligosaccharides, including glycoproteins bearing only a single 0-linked oligosaccharide and those bearing sialic acid-containing oligosaccharides (3 7-3 8). Lectin affinity chromatography on jacalin-agarose also provides a usefirl technique for purifying glycopeptides from glycoprotein digests. This would greatly assist in localizing peptides with O-Iinked oligosaccharides. In many instances, this has been a difficult analytical problem even for extensively characterized proteins because, iethecaseofN-lin idglyoosyhfion be This chapter (1 out: zona pellucida murdMALDI‘ Hiatbohydnte all We: from tryp idmpoteins. Lect MD: protein with Ethics (01" and iv liltiprotein with bindi 1 Experimental hurls Ribonuclease lirepurchtsed from Write oocyte zona 93 unlike the case of N-linked oligosaccharides, it is not possible to reliably predict sites of O- linked glycosylation based on the amino acid sequence of a protein. This chapter describes our efforts in the structural characterization of porcine oocyte zona pellucida 30t-glycoprotein (ZP3a) using HPLC separation, exoglycosidase digestion, and MALDI-MS analysis. We also have investigated the utility of some specific lectin-carbohydrate affinity interactions in combination with MALDI-MS for identifying A glycopeptides from tryptic digest and determining site-specific carbohydrate heterogeneity of glycoproteins. Lectins studied in this chapter include: (1) Concanavalin A (Con A): an lO4-KDa protein with binding capability for terminal or-D-mannosyl and oc-D-glucosyl residues (Ca++ and Mn“ are required for its activity), and (2) J acalin: a SO-KDa glycoprotein with binding capability for only O-glycosidically linked oligosaccharides. 2. Experimental Materials Ribonuclease B(Type III-B, from bovine pancrease), avidin, oc-methylgalactoside were purchased from Sigma (St. Louis, MO). Endo-B-galactosidase digested and purified porcine oocyte zona pellucida glycoprotein 30c (ZP3a) was provided by Dr. E. C. Yurewicz (Department of Obstetrics and Gynecology, Wayne State University School of Medicine). Jacalin-agarose was obtained from Vector Laboratories (Watertown, MA). Sequencing grade trypsin was from Boehringer Mannheim (Indianapolis, IN). Neuraminidase (from A rthrobacter ureafaciens), B-galactosidase (from Diplococcus march), and Mar My!“ (Bedfo llllDl~MS MALDI-MS e W (Pasept l LS-Dihydroxybr iughou this chapt mm .Adelayt W of digested glyc hficrobore rew WA (MiChrom BioF "115-um particle, 31 45%45% acetonitrile it maintained at 5 Wind fractions w ligcstion. In this ihHCl buffer (pH ”him was mixed wi 94 pneumonia), and N-acetylhexosaminadase (fi'om Jack bean) were purchased from Oxford GlycoSystems (Bedford, MA). MALDI-MS MALDI-LMS experiments were carried out on a Voyager-Elite time-of-flight mass spectrometer (PerSeptive Biosystems Inc.) equipped with a nitrogen laser emitting at 337 nm. 2,5-Dihydroxybenzoic acid (DHB, recrystallized) was used as the MALDI matrix throughout this chapter. The ion source accelerating voltage was 20 kV in all T OF experiments. A delay time of 50 ns was used before the ion extraction. HPLC of digested glycopeptides Microbore reverse phase HPLC experiments were performed using a MiChrom UMA (MiChrom BioResources) instrument equipped with a PLRP-S column (1.0 x 150 mm, S-um particle, 300-A pore size). Elution was achieved with a linear gradient of 4.5%-45% acetonitrile in 0.1% TFA mobile phase over a period of 80 min. The flow rate was maintained at 50 ul/min and effluent was monitored continuously at 214 um. i . . . Indrvrdual fractions were collected and used for firrther MALDI-MS analysrs. l l Glycoprotein reduction and carboxymethylation The reduction and S-carboxymethylation for glycoproteins were performed prior to digestion. In this study, 300 pmol ZP3a glyc0protein was dissolved in 200 it] 0.5 M Tris-HCl buffer (pH 8.2) containing 2 mM EDTA and 6.0 M guanidine-HCI. This solution was mixed with 10 it] of 0.4 M dithiothreitol (DTT) and incubated at 37°C for 3 1m Avolume of 2 an Aridin and dbl LEM The reduce irisrubing (3,500 I Station with Irypsin For preparatior 33:. nbonuclease B, sensing grade try; :rabation was perform ligation with exoglyC Glycopeptides :uglycosidases were t itttiiiljacalin) affinity - Prior to affinity tdissolved in 50 mM 302% sodium azide ( llncentration of 2 mM hlsh activity. The illibrated with buffer ill pie has entered the < 95 hours. A volume of 20 ul of 0.8 M iodoacetic acid was then added and incubated for 1 hour. Avidin and ribonuclease B samples (100 pmol each) were treated in an identical manner. The reduced and carboxymethylated samples were desalted overnight using dialysis tubing (3,500 MW cut-oft) from Fisher and dried for further experiments. Digestion with Trypsin For preparation of tryptic digests, reduced and carboxymethylated preparations of ZP3a, ribonuclease B, avidin were dissolved in 100 mM ammonium bicarbonate (pH 8.0). Sequencing grade trypsin was added at an enzymezsubstrate ratio of 1:50 (w/w) and incubation was performed at 37°C for 24 hours. Digestion with exoglycosidases Glycopeptides were dissolved in 25 mM ammonium acetate buffer (pH 5.5). Exoglycosidases were then added and incubations were performed at 37°C for 24 hours. Lectin (jacalin) affinity chromatography Prior to affinity chromatography on jacalin-agarose, tryptic digests were dried and redissolved in 50 mM sodium phosphate buffer, pH 7.0, containing 500 mM NaCl and 0.02% sodium azide (buffer A). Phenylmethylsulfonyl fluoride was added to a final concentration of 2 mM and the sample incubated at 4°C for 30 min to inactivate residual trYpsin activity. The sample was then applied to a minicolumn of jacalin—agarose equilibrated with buffer A using a ratio of 1 mg digest per ml of settled resin. After the Sarnple has entered the column, flow was stopped and the column incubated for 30 min to we binding of 0’ shots Of Gold (40C: 2:76 sabse‘lumdy elu ,mpgajactoside- 3.Rrsults and mm” 3.1. Structural cha Die) The zona pellus gycoconjugates WhiC figmed manner l rotational processi ehelthreonine residu heterogeneous with res residues representing hacellular trafficking thirties of zona gl tithmably as ligands than (43). The pig represe tiillfuctural and func Z0 ' . “i'SCOmpnsed of ti 96 ensure binding of O-glycopeptides. The column was washed with 20 column—volume aliquots of cold (4°C) buffer A and each fraction collected separately. O-glycopeptides were subsequently eluted with ten column-volume aliquots of buffer A containing 50 mM oc-methylgalactoside. 3. Results and Discussion 3. 1. Structural characterization of porcine oocyte zona pellucida glycoprotein 3a (ZP3a) The zona pellucida of mammalian oocyte are comprised of sex and gamete-specific glycoconjugates which are expressed during oocyte grth in a developmentally programmed manner (3 9-41). Zona glycoproteins are characterized by extensive post- translational processing, including N- and O-glycosylation on asparagine and serine/threonine residues, respectively. Zona sugar chains examined to date are highly heterogeneous with respect to structure, size, and charge with polylatosamines and sulfate residues representing distinguishing characteristics (40, 42). In addition to a role in intracellular trafficking, zona carbohydrates strongly influence the physicochemical properties of zona glyCOproteins and participate in sperm—zona recognition events, presumably as ligands for complementary carbohydrate-binding proteins on the sperm surface (43). The pig represents an attractive alternative to rodent species for investigations of the structural and firnctional properties of zona pellucida glycoproteins. The pig oocyte zona is comprised of three structurally and immunologically distinct glycoproteins, which gill (82-KDa prote tong them, ZP3a is Erhsstudy. The ar isolation sites at A: '53 and Thr 304 ha‘ edition analysis. am the site-spa Reduced and car ”lies and glycopept 3LT hattions by MA tinnions. These for 3 HPLC Chromatogr :t-‘aned from one of t arrears, these pepiid “z: carbohydrates was dlttt Da corresponc Isolation and it il'toprotein can be a ti reverse phase H hotly to a reverse illicit program use iii htpeak was founi 97 are ZPl (82—KDa protein), ZP3a (SS-KDa protein), and ZP3B (SS-KDa protein) (44-47). Among them, ZP3a is a highly glycosylated protein and was used as a model compound for this study.. The amino acid sequence of ZP3a is shown in Figure 4.4. Three N- glycosylation sites at Asn 203, Asn 220, and Asn 333 and two O-glycosylation sites at Ser 293 and T hr 304 have been found before for this protein by HPLC and Edman degradation analysis. The main objective for our study on ZP3a glycoprotein is to characterize the site-specific microheterogeneity and carbohydrate structure. Reduced and carboxymethylated ZP3a glycoprotein was digested and the resulting peptides and glycopeptides were separated by reverse phase HPLC. Analysis of the HPLC fractions by MALDI-MS showed the presence of glycosylated peptides in four of the fractions. These four fractions are labeled as Asn 333, Asn 220, Asn 203, and 0GP in the HPLC chromatogram as shown in Figure 4.5(a). The MALDI mass spectrum obtained from one of the four fractions, Asn 220, was shown in Figure 4.6. Based on their mass, these peptides were identified as glycosylated peptides. Microheterogeneity of the carbohydrates was revealed by clusters of peaks separated by a mass difference of 162 and 203 Da corresponding to hexose and N-acetylhexosamine residues, respectively. Isolation and identification of O-Iinked glycopeptides from tryptic digest of ZP3a glycoprotein can be accomplished by sequential lectin (jacalin) affinity chromatography and reverse phase HPLC. The affinity purified 0-linked glycopeptides were applied directly to a reverse phase HPLC column for desalting and purification with the same gradient program used for the separation of the tryptic digest of ZP3a glycoprotein. A sing1e Peak was found in the HPLC run as shown in Figure 4.5(b). This peak, at the same GPAWCGGGTGGN mmmrcmcrv PLCLAL mascasc'r'r'rcarr P S S F E P WAEGCGCGSGCCKEC‘ DERGRL LET-C“ TGSGAGNG SSH G V E l.1'.".‘:3L"T7'GMGMG I G L E E A armaments :FLALQ sacmacccxrcn LBJ P I 'r TfiLfiM‘GSMACP C Y 1' G N '1 DSAECNACCTCCAC 'MSPPr Al‘GSAMC ACAC Am ”Tarn )K‘CAGEGGT AT XX E Q h V Y I coasacaocnm; R n “51?; cm: . a'r'fi:hCTCTc( " V Wmucccms mcMhGG‘l‘crcr i V E V 8 :MCAGACCMAW a K L i 5 T F S WWTCCAW D I H F GM’I'I‘CCGGGTGGMGTACCTGTTCTCCGCAGGCGCTATGTGG'I'I‘GCGGCCGTCCATCTGGCW 7 0 HWLRPSIHLCF 11 CCGC’I‘G‘I‘GTCI‘TGCTC'I‘GCCAGGCCAG'PCTCAGCCCMAGCAGCAGATGACC'I'PGGTGGCC'I‘CTAC‘I‘GTGGG l 4 2 PLCLALPGQBQPKAADDLGGLYCG 35 l CCMGCAGC'I‘I‘TCAT'I‘I‘CTCCATMATCT‘I‘CTCAGCCAGGACACAGCMC‘I‘CCNCTGCACTGGTGGTT’IGG 2 1 4 PSSFBPSINLLSQDTATPPALVVH 59 GMAGGCGCGGGCGGCTGCACMGCTGCAGMTGACTCTGGCTG‘I‘GGCACGTGGG‘I‘CCACPLAGGGCCCAGGC 2 8 6 DRRGRLBKLQNDSGCGTHVBKGPG 83 AGCTCCATGGGAG'I‘GGAAGCA'I‘CCTACAGAGGCTGCTATGTGACTGAG'I‘GGGACTCTCACTACCTCA'I‘GOCC 3 5 8 SSHGVEABYRGCYVTEWDSHYLHP 107 A'I'I‘GGAC'I‘I‘GAAGMGCAGATGCAGG‘I‘GGACACAGMCAGTCACAGAGACG WTG'I‘T‘I‘AAGTGCCCTGI‘G 4 3 0 IGLEEADAGGHRTVTETKLFKCPV131 GAI'ITCCTAGCTC'I‘PGA’I‘G‘I‘TCCMCCNI'PGGCC‘I'PTGTGATGCTGTCCCAGTGTGGGACCGATTGCCANT 502 DPLALDVPJIGLCDAVPVHDRJJC155 GC‘I‘CC'I‘CCACCCA'I‘CACTCMGGAGMTGCMGCAGC‘I'I‘GGCTGCNCTACMCTCGGAAGAGGTCCC‘I‘I‘C‘I‘ 574 AEEPITQGECKQLGCCYNBEEVPS179 TGTTACTATGGMACACAGTGACCTCACGCTGTACCCMGATGGCCACTTCTCCATCGCTGNTCTCGCMT 6 4 6 CYYGNTVTSRCTQDGBFSIAVSRN 203 GTGACCTCACCTCCACTGCTC‘I‘GGGXI'PC'I‘GTGCACCTGGCC'I‘I‘CAGAMTGACAGTGAATGTMACC‘I‘GTG 7 1 8 VTSPPLLWDSVBLAFRNDSECKPV 227 ANGMACACACACTTT‘I‘GTCCTCT'PCCGGI'PTCCATTPAGTPCCTGTGGGACTGCAMACGGGTMCTGGG 7 9 0 HETETFVLFRFPFSSCGTAKRVTG 251 AACCAGGCGGTATATGAAMTGAGC'I‘GGTAGCAGCTCGGGATGTGAGGACTTGGAGCCATGG’I'I‘CTATTACC 8 62 NQAVYENELVAARDVRTWSHGBIT275 CGAGACAGCATCT'PCAGGCTTCG AG'I‘CAG‘I'I‘GTA'I‘CTPCTCTG‘I‘AAGTAGCAGTGCTCTCCCAG'I'I'MCATC 9 3 4 RDSIFRLRVSCIYBVSSSJLPV’HI 299 CAGG'I'I'I’TCACTCTCCCACCACCGC‘ITCCGGAGACCCACCC'I‘GGACCTC‘I'I‘ACTCTGGAGC'I'I‘CAGA'I'I‘GCC 1006 QJFJLPPPLPETHJGPJJLELQIA323 MAGAI‘GMCGCTATGGC'I‘CCTACTACMTGCTAGTGACTACCCGGTGG'I‘GMA‘I'I‘GC’I‘TCGGGAGCCCATC 1078 KDERYGSYYNASDYPVVKLLREPI 347 TATGTGGAGGTCTCTA'I‘CCGTCACCGAACAGACCCCAGTCTCGGGCTGCACCTGCACCAGTGCTGGGCCACA 1 1 5 0 YVEVSIRHRTDPSLGLBLHQCWAT 371 CCCGGCATGAGCCCCCTGCTCCAGCCACAGTGGCCCATGCTAGTCMTGGATGCCCCTACACTGGAGACMC 1222 PGHBPLLQPQNPHLVNGCPYTGDR 395 TACCAGACCAMCTGATCCCTGTCCABMAGCCTCWCTGCTATI‘TCC’I'I‘CTCAC’I‘ACCAGCGTI‘TCAGT 1294 YQTKLIPVQKASNLLPPBHYQRFS 419 GTI'I‘CCACC’I'I‘CAG'I'I‘TTG'I‘GGACTCTGTGGCAMGCAGGCACTCMGGGACCGGTGTATCNCATI‘GTACT 1 3 6 6 VBTFSPVDBVAKQALKGPVYLHCT “3 GCATCGGTC'I‘GCAAGCCTGCAGGGGCACCGATC‘I‘GTGTGACMCC'I‘GTCCTGCTGCCAGACGAAGMGAAGT 14 3 8 ABVCKPAGAPICVTTCPAARRRRB 467 TCTGACATCCATT‘I'I‘CAGMTGGCACTGCTAGCA'I‘I‘TCTAGCMGGGTCCCATGATTCTACTCCAMCCACT 1 5 1 0 SDIHFQNGTABIBBKGPHILLQAT 491 CGGGACTC'I'PCAGMAGGCTCCATAMTACTCMGGCCTCCTGTAGAC‘I‘CCCNPGC'I‘CTGTGGGTGGCTGGC 1582 RDBBERLHKYBRPPVDBBALWVAG 515 CTCTPGGGAAGCTI‘MTTATTGG AGCCNGTTAGPGTCCTACCTGGTC'I'PCAGGAAATGGAGATGAG‘I'PACT l 6 5 4 LLGSLIIGALLVSYLVFRKHR‘ 536' camccmmewccurmmamccceum 1699 igure 4,4; Amino acid sequence of porcine oocyte zona pellucida Glycoprotein 3a ZP3a) notion time as th g‘renpeptides. Further MAl heterogeneity of the r ids represented the *lthaNAc) attac hidegh'cosylated p angHPLC and Edi rdllrr304, HOWCV f0martinis by gel . llredominam peak 0"[111de oligosaccha etention time as the 0GP fraction in Figure 4.5(a), positively identified the 0-linked glycopeptides. Further MALDI-MS analysis (Figure 4.7) of the collected fraction showed heterogeneity of the O—linked glycopeptides. The peak at the low-mass end at m/z 5146.9 hkely represented the [M+HT ion of a glycopeptide with two 0-linked disaccharide units :GalB-GalNAc) attached because the m/z difference between this ion and [M+H]+ ion of :he deglycosylated peptide (m/z 4416.1) is 730.1. Previous study by Yurewicz’s group rsing HPLC and Edman degradation has confirmed two O-glycosylation sites at Ser 293 and Thr 304. However, it was suspected that there may be one more O-glycosylation site iom analysis by gel electrophoresis on chemically cleaved 0-1inked oligosaccharides (47). A predominant peak at m/z 5716.5 in Figure 4.7 supported this hypothesis since most of 9-linked oligosaccharides are disaccharides or trisaccharides. .Z I -I I‘ll!) AITH()I‘I‘DSIIIUkJ ( A 195 1.6119341 figure 4'5: a Eillpepride fracrl "“mllOsraphy. I 100 OGP Asn 333 \ Asn 220 Ami” Ansoroance (.414 nm) V l l l 10 25 40 55 min b LA/r Ll . r r J 10 25 4O 55 min re 4:13: (a) HPLC of tryptic digest of ZP3a glycoprotein. (b) HPLC of the Opepti e fractions from ZP3a glyco rotein after 1 t' ' ‘ - matography. p CC 1n (Jacalrn) affimty hammflnofls 0>HHQnDMH 101 4224.2 4386.7 anuw... . - ”Wk.--“ 4548.9 m I I 3500 4060 4500 5000 nflz :ure 4.6: N-linked glycopeptides at Asn 220 of ZP3a glycoprotein. The MALDI-MS ctrum was obtained fi'om a collected HPLC fraction. Relative Intensity 102 co m' v» W. 3% 2 l\ :,. W 3 1 N ’ b' a :2 so, 3 09 "‘ £6 3 l\ "I” cog O\'—' m0 \o'W mm <1- ‘0‘“ W U I r I I 5000 5500 6000 6500 m/Z VSCIYSVSSgALPVNIQVFTakLPPPLPETI-IPGPLTLELQIAK *: O-glycosylation site we 4.7: O-linked glycopeptides of ZP3a glycoprotein after purification by lectin 11in) affinity chromatography and HPLC. fl Three other damned to be It Typiully, N-linked Ms: E (PM W) wbohydm rm to monito Waldo-bear ills glycopeptide. Mons, Asn 203 wide: an altemat glycopeptides, especi The MALDI ‘ town previously in i medial digestion ' has by MALDI- lwuled by a mass lolowed by N-acety Wed peaks at m/z [moved The mass 103 Three other glycopeptide fi'actions, Asn 220, Asn 203, and Asn 333 were mined to be N-linked glycopeptides as they were not retained ' on jacalin-agarose. cally, N-linked glycopeptides can be identified after the treatment with peptide-N- asidase F (PNGase 0F), an endoglycosidase that cleaves off asparagine-linked (N- :d) carbohydrates. MALDI-MS analysis can be incorporated with PNGase F merit to monitor the mass shift of the analyte. This enzyme does not fiinction if the )saccharide-bearing asparagine residue is located at either N- terminal or C- terminal ne glycopeptide. This happened to be the case for two of the N-linked glycopeptide tions, Asn 203 and Asn 220, from ZP3a glycoprotein. Jacalin affinity binding rides an alternative approach to distinguish N-linked glycopeptides and O-linked opeptides, especially when PNGase F treatment fails. The MALDI mass spectrum of one N-linked glycopeptide fraction at Asn 220, was vn previously in Figure 4.5. The carbohydrate structure was further characterized by rential digestion with specific exoglycosidases and by monitoring the resulting mass 33 by MALDI-MS. ' Since the microheterogeneity of this glycopeptide fraction was ed by a mass difference of 162 and 203 Da, it was digested with B-galactosidase twed by N-acetylhexosaminidase. After B-galactosidase digestion, a series of evenly ed peaks at m/z 3250.4, 3453.1, 3655.8, 3858.6 was observed in the MALDI mass trum (Figure 4.8(a)). After B-galactosidase and N-acetylhexosaminidase digestion, a 6 peak at m/z 32502 was produced as shown in Figure 4.8(b), which represented the HT ion of a glycopeptide with both terminal galactose and N-acetylhexosamine rved. The mass difference between this ion and the [M+H]+ ion of the deglycosylated N.OWNM _ NVflU m 3 v.0mfln Mum was obtain aillll-galactosldase .H 0 bmmnoafln 0>fi=wq0N~ bqunvn—EH U>.H..N~ Ma 104 8 £1 0 0°. .> 8 Es. "‘ s 32 g? ¢ T | I I 3500 4000 4500 5000 m/z 3?: ii 3 d) ,> .3 32 I l I I I 3500 4000 4500 5000 M igure 4.8: N-linked Glycopeptides at Asn 220 of ZP3a Glycoprotein. MALDI-MS >ectrum was obtained on the glycopeptide sample (a) after B-galactosidase treatment; (b) ter B-galactosidase & N-acetylhexoaminidase treatment. fl pptide(m/z 221 1.: to be (Man);(GlcN. mofigomccharidr and W wit Figure 4.10 pympeptide fractio rattled by mass di 3071):, which illust nl N—glywlylneura wanted the [M4 l'ne peaks for the l glyoopeptides were I wilrmueh lower inte floglycosidases: neu tingle peak at figure 4.10(b), whi 105 peptide (m/z 2211.5) was 1038.7 Da, which determined the core oligosaccharide structure :0 be (Man)3(G1cNAc)2(Fuc). Other glycopeptide structures were assigned based on this :ore oligosaccharide structure in Figure 4.9. Observed m/z values of [M+H]+ ions were in good agreement with calculated m/z values. Figure 4.10(a) showed the MALDI mass spectrum of the Asn 333 N—linked glycopeptide fraction. The microheterogeneity of this glycopeptide fraction was not only 'evealed by mass difference of 162 and 203 Da, but also by mass difference of 291 and :07 Da, which illustrate the existence of two types of sialic acid, N-acetylneuraminic acid md N—glycolylneuraminic acid. The peaks at m/z 3541.1 and 3704.2 in Figure 4.10(a) epresented the [M+I-I]+ ions of two N-glycolylneuraminic acid containing glycopeptides. The peaks for the [M+H]+ ions of two analogous N-acetylneuraminic acid- containing dycopeptides were also observed at m/z 3525.0 and 3688.4 (not labeled in the spectrum) vith much lower intensity. Glycopeptide fraction at Asn 333 was digested with a series of xoglycosidases: neurarninidase followed by B-galactosidase and N-acetylhexosaminidase. L single peak at m/z 2665.4 was produced after the sequential digestion as shown in igure 4.10(b), which represented the [M+H]” ion of a glycopeptide with terminal sialic cid, galactose and N-acetylhexose removed. Again, the mass difference between this ion nd the [M+H]+ ion of the deglycosylated peptide (m/z 1626.8) was 1038.6 Da, which Dnfirmed that the core oligosaccharide has a fucosylated structure, VIan)3(GlcNAc)2(Fuc). Other glycopeptide structures were assigned based on this core ligosaccharide structure in Figure 4.11. . 4 . ii s u Izc ( if rrtc. eon . \ZO \3~ “ml {/20 NTWW®M WOWWM ONszr.UZn~hv UUu.~—:U_:rv U.::U:.:T. UUZAuA—Avha— 106 .moanmoobm 2: mo maom +E+H>a 05 8 uoawmmma 6.83 Na: noun—36:8 v5 N\E 33030 5305830 wmmN mo ONN :2 3 3236335 v3??? "m6 «Sufi .20.. \zos . . \E to To 3w: mew? ommcérzOIZOLz _ 342 933 cérzorzorzhzZo To L szzofio CNN m szzo to ’20 ’20; w \Eioz \zoJ 384 384 92:2 :2 r2 _ \3720 To omm W 1295.10 Emma 03% ommaérzWIZOLz/Exzo _..o 20 123 orzorzhzrzu ”an; 98% 572 as... a saw New? ENS ommaéiWIZOIZ/EMW _10 Izo._ fiioimfihziofio 35m 92% 0mm M 12.20. \ZO . . \sfé 38¢ 38¢ ommaérzWiorzjzxzo rz oikzrzo ”mom 3%». a2 :2 ’20 02 w lzrzo N\S N\E ovumoaoobm 0:... mo N\E N\S ovumomoobm 06 .«o 602890 gafiofio chaosbm tomoaoi 33630 333030 2303.5. 88¢on l)’ l-.UWDO UUHQ~3U~SU mmm Cm< IZMU IZC 42 J4 I20 OfikuanQonvaTw 05 .NO 0h: u 03L}. .0 mUanOQOHHN 108 .movuaoqoobm 05 mo 32 JAE/2+3: of 9 coammmmw 0.63 QE 080328 98 N\E 003890 .Eopoaoobm ammN mo mmm cm< Hm moEEQQOQn—w we”??? ”:6 charm flocmm mémmm m.mowm wdowm Novom ©6va Ndobm Rheum flammm 0.3mm ”09% v.33“ N\E N\E 8:030 3:303 \wZL/TU J_IO 120 mmmeéuzmu I20 IE/E \20 _.I.00 120 _ x75 J \wZJ/HO To mmmcglm/MV 202/220 _ImV 20 _ \ZO J \2 20 _ IO mag 2% I20 IE/Exz0 _ I20 _ \EIZO l0 mmmcm\2 _0 mmm m /§ \20 _IO I20 \2 Izod mmm=m..92Aafiv Nxxzs 10:130—5. v jZA/ZLR JLO x ,OWNM Due—un—OnnavaA—Wm 0.: ._C N\-\ 0.: .03.. :1. «XJZSATVLL UU)LUWA~O V.~WNM N\-\ UOufi—301NHV \u). JAG J2 I20 UunQUQOUxQW 0:.- H0 USuUH-Luw VOWOQOLflN M0N:m< IZO I20 I2 111 0.3? mNNov 5.3me NE 32030 .300%082mgN 20.8 82 .3222; 20 8 020303 803 Q5 083828 000 N\E 030030 .530509»% ammN m0 mom 09¢ 8 $33033 0800302203 mo0wmomoobm 08200.? "2.0 ouswmm \ZUJ mflw; momcélzwiéwlwz/EAAC _% I20 0 L20 .. \2 20 T20 ~88 8.0.5. .20 .20 I212 \20 _I0 120 0 \2 ‘20 J_ . 20 o ommm mom=m<120 IZO $21220 _IpU 20 0 N\S 00000083» 20 .20 0030505 050:5 030005 oSmom NmSm 0&me wdmmm N\E 03.830 wfimom w.0$m 00me vgwmm es Baaoao \Zw \2 12 8qu I20 I20 ISM/22 2% I20 \2 I20 0 momcglzo I20 IVA/VA \ZO _I0 .20. E20 8&5. I20 I20 IE/E \ZO 120 \EIZO momcé I20 I20 .2 12 .20 0000000023 20 no 83032 0080022 NE N\5 CHESQOQOUxA—w 05 .00 UOEDMDO U0ud—30—60 OBSUSIJm UOWOQOLQ NE NE DUUQUQOQDW 05 k0 UOEOWDO UOuQ~JQ~QU OSHUDLum DOMOQOHQ 112 .mo0maomoobm 05 .«o meow 11? Z+0é 20‘ 8 02%33 803 N\E 0800808 030 N\E 002030 .Eouoaoohm GEN mo mom c2 3 $83033, 08038030 mo0umomoobm 0030.2 "3.0 ounwwm \EJZUJ LAO). 38m 38m magi IzoIE _Iw . . \220 +8 122$ Io m :2 v 3.3 mom:2.2%lzwI2 To 120‘. {\zo Io [2% J .z 140123.23. .75 0%? 22m mowed My \HZZO _IO {\Zwmv 510 0.8? gm; momaéiMuioIE/Eio _I0 I20 Izmi \EIZG \zotm 0 floom w Doom MONGQIZMVIZUIEJ/Hfimv \2 c 30% 08% momcénzwviwi/Exflw _IO 123 . . \223 w 3?. n mmmm momcéuzoIzqu Io _ ‘ 44.2% \220 m 120 \220 Sex 93% momaéifilzwIE/Exzo 30mm 0.3mm SNEIZGIZGQA 0 do 3 J38 N0: Q5 o0wmoQOobm 20 mo NS: N\S o0umomoobm 20 mo 002030 0303200 238.3 033on 09530 0803200 ogogm 033on Trypti glyco; & S-cz Figure 4'15: A mic 113 Lectin-immobilized beads put into a centn'fuge filter tube and equilibrated with binding buffer Tryptic digest of glycoprotein (reduced & S-carboxymethylated) ¢ Binding between glycopeptides and lectin 1 Wash with binding buffer 1 Binding buffer spun off & lectin- immobilized beads resuspended in matrix (DHB) solution 1 MALDI-MS Figure 4.15: A micro-batch lectin aflinity binding process for subsequent direct MALDI- MS analysis. imuse of its acid aidthelectin so th This appro hid avidin. Bovin iii and also ca sanded DNA. gy‘cosylation site a the and tissues shunts, each of w heat asparagine 1 ititheraiidin or F Concanavaii 53d asidin because Concanavalin A ha Binding buffer for ( 150 mM NaCi and MW or avidin Process whereas 0t bUffer. Direct MAL ——— 114 because of its acid nature, this acidity is able to disrupt the binding between glycopeptides and’the lectin so that glycopeptide signal can be obtained by MALDI-MS. This approach was used to identify glycopeptides derived from Ribonuclease B and avidin. Bovine pancreatic Ribonuclease B (RN ase B) is a glycoprotein that cleaves RNA and also can destabilize or unwind the DNA helix by- complexing with single- stranded DNA. RNase B comprises 150 amino acid residues and one N-linked glycosylation site at asparagine 60 (Asn 60). Avidin is a glycoprotein found in the egg- white and tissues of birds, reptiles and amphibians. Avidin contains four identical subunits, each of which comprises 128 amino acid residues and one N-linked glycosylation site at asparagine 17 (Asn 17). No 0-linked carbohydrate side chains have been observed in either avidin or Ribonuclease B. Concanavalin A-Separose was used to bind glycopeptides derived from RNase B and avidin because these glycopeptides consist high mannose-type oligosaccharides and Concanavalin A has high binding capability for terminal a—D-mannosyl residues. The Binding buffer for Concanavalin A was a sodium phosphate solution (pH 7.4) containing [50 mM NaCl and 1 mM MnClz and 1 mM CaClz. Glycopeptides in the tryptic digest of iNase B or avidin were retained by Concanavalin A-Separose beads during the binding )rocess whereas other peptides in the tryptic digest were washed away by the binding lufi‘er. Direct MALDI-MS analysis of tryptic digest of RNase B was shown in Figure '.16. The glycopeptide signals were seriously suppressed by signals of other peptides in Figure 4.1a . Irllh O wvw O.N_k. ii 9.0% hamZA-nva-u— fiv>muud~uvuy~ 115 m .vmmm wdomm mAwNNM 0.2M; Biz/I'D can: 59: M352 3&2 ii @002 MW 0.2a lImI 33w 3:. 0.2% IML m/z l 2500 2000 l 1500 l 1000 ‘.‘\‘“,-‘I ’ D ‘3‘"- all Figure 4.16: Direct MALDI-MS analysis of tryptic digest of ribonuclease B. 36 sample' MAI fier micro-batch l z-erepredominant temptic digest \ :1 2263.2 repres :«idues in the oli; 7363,1898], 2( gycopeptides. Dirt showed the presenc Jim-batch lectin NHLDI mass spec! 116 the sample. MALDI-MS analysis of glycopeptides from the tryptic digest of RNase B after micro-batch lectin (Con A) binding is shown in Figure 4.17. Glycopeptide signals were predominant in the MALDI mass spectrum while most other peptide components in the tryptic digest were not detected. The peaks at m/z 1714.8, 1876.9, 2039.1, 2201.6, and 2263.2 represented the [M+Na]+ ions of glycopeptides with 5, 6, 7, 8, 9 mannose residues in the oligosaccharide structure, respectively. Another series of peaks at m/z 1736.7, 1898.7, 2061.8, 2223.9, and 2385.4 represented the [M+2Na-I-I]+ ions of these glycopeptides. Direct MALDI-MS analysis of the tryptic digest of avidin (Figure 4.18) showed the presence of the glycopeptides in the region from m/z 3000 to m/z 3700. Afier micro-batch lectin (Con A) binding, these glycopeptides are more clearly visible in the MALDI mass spectrum (Figure 4.19). 1.-.-.Rw—MJV~ 118 0.32 odmvm webmm 0.2mm\\ 938 v ‘A M._ w v v 0.88 W Némfi 02w v.93 IIIW mamoo l 3000 l 1000 m/z 2000 ha unquDgu: D > 3.35/4 Figure 4.18: Direct MALDI-MS analysis of tryptic digest of Avidin. l Observed M I 3075.0 3237.4 ,1; 3319.2 3398.9 3! 3481.4 : i 3533.4 .. I 3621.3 3684.5 J ‘i [L a. l [(Glalliv ‘ Wm \;\ HMO FWthMAl WMMMMHe mmmqu+ 119 Observed Calculated Proposed structure : W2 W2 go 3075.0 3076.1 (Man)5(GlcNAc)2-Asn 17 to 3237.4 3238.3 (Man)6(GlcNAc)2-Asn 17 3319.2 3320.3 (Man)4(GlcNAc)4-Asn 17 3398.9 3400.4 (Man)7(GlcNAc)2-Asn 17 ox 3481.4 3482.5 (Man)5(GlcNAc)4-Asn 17 g: 3523.4 3523.5 (Man)4(GlcNAc)5-Asn 17 3 3621.3 3644.6 (Man)6(GlcNAc)4-Asn 17 3684.5 3685.7 (Man)5(GlcNAc)5-Asn 17 q E ‘9 <1: N 3 2' box <3- 00 (fir-1 \o \O 92 m j“ l I I m/ 1000 ' 2000 3000 Z We 4.19: MALDI-MS analysis of glycopeptides from the tryptic digest of avidin :r micro-batch lectin (ConA) affinity binding. Observed m/z and calculated m/z were lgned t0 the [M+Na]+ ions of the glycopeptides. 4' Conclusions he tryptic digest 30011831313 fract iiiiDi-Ms sped g-unguished by 16 mum of the gl '31 of the glyco soglycosidases v iiiiDI-Ms analy: aid unfiicosylated 111333 is compos oligosacchardes. ' binding and MAL] died MALDI-MS identifying glycope; h’cosylation site. 120 4. Conclusions The tryptic digest of ZP3a glycoprotein was separated by reverse phase HPLC and four glycopeptide fractions Were identified by their microheterogeneity as revealed in their MALDI-MS spectra. N-linked and O-linked glycopeptides of ZP3a glycoprotein can be distinguished by lectin affinity chromatography / HPLC / MALDI-MS. The MALDI mass spectrum of the glycopeptide at Asn 203 of ZP3a showed more microheterogeneity than that of the glycopeptide at An 220 or Asn 333. After the treatment with specific exoglycosidases which release most of the outer branches of these glycopeptides, MALDI-MS analysis revealed that the glycan at Asn 203 consists a mixture of fucosylated and unfucosylated bi- and tri-antennary structures, whereas the glycan at Asn 220 or Asn333 is composed of a mixture of fucosylated complex bi-, tri-, and tetra-antennary oligosaccharides. To minimize sample handling between the stages of lectin affinity binding and MALDI-MS, a micro-batch lectin binding process was also developed for direct MALDI-MS analysis for isolated glycopeptides. This approach is capable of identifying glycopeptides derived from ribonuclease B and avidin, proteins having a single glycosylation site. Rcferencfii i Cumming, D. . 2 Del A and R' 3’ $35311, M. W 1 Young N' M" Biop’D’s" 270’ 5 Green, E- D" ' Chem, 263, 18 5 Edge, C. J ., R Wing, D- R., ar ' Conboy, 1- J- a1 1 Hsi, K. L, Che 198,238-245, 1 '3. Mortz, E. Sari 1109-1118, 199 10, Yang, Y. and O lilieuheit, M. J., 12. Huberty, M. C., 1993. 121 References: 1. Cumming, D. A., Glycobiology, 1, 115-130, 1991. 2. Dell, A and Reason, A. J ., Curr. Opin. Biotechnol. 4, 52-56, 1993. 3. Sastry, M. V. K. and Surolia, A., Biosci. Repots, 6, 853-860, 1986. 4. Young, N. M., Johnston, R. A. Z., Szabo, A. G., and Watson, D. C., Arch. Biochem. Biophys., 270, 596-603, 1989. 5. Green, E. D., Adelt, G., Baenziger, J. U., Wilson, S., and van Halbeek, H., J. Biol. Chem, 263, 18253-18268, 1988. 6. Edge, C. J., Rademacher, T. W., Wonnald, M. R., Parekh, R. B., Butters, T. D., Wing, D. R., and Dwek, R. A., Proc. Natl. Acad. Sci. USA, 89, 6338-6342, 1992. 7. Conboy, J. J. and Henion, J. D., J. Am. Soc. Mass Spectrom. 3, 804-814, 1992. 8. Hsi, K. L., Chen, L., Hawke, D. H., Zieske, L. R., and Yuan, P. M, Anal. Biochem., 198, 238-245, 1991. 9. Mortz, E., Sareneva, T., Julkunen, 1., and Roepstorfi", P., J. Mass Spectrom., 31, 1109-1118, 1996. 10. Yang, Y. and Orlando, R., Rapid Commun. Mass Spectrom., 10, 932-936, 1996. ll. Treuheit, M. J ., Costello, C. E., and Halsall, H. B, Biochem, J., 283, 105-112, 1992. 12. Huberty, M. C., Vath, J. B, Yu, W., and Martin, S. A., Anal. Chem, 65, 2791-2799, 1993. 13. Carr, S. A., Huddleston, M. J ., and Bean, M. F., Protein Science, 2, 183-196, 1993. .4.Huddleston, M. J., and Bean, M. F., and Carr, S. A., Anal. Chem, 65, 877-384, 1993. 15. Schindler, P. Protein Sciem 16. Liu, J., Volk, Chromatogr., 17. in, D. and 1121' 13 Sutton, C. W., 19 Mizuochi, 1., 7404,7409, 19 2'1 Yamamoto, T. 21 West, I. And C 3 Endo, 1., J. C 3 Ogata, 3., Mur 2103333. 3., Mur. 35Narasimhan, S. immgwhi, T, Clement 17 21. Yamashjta, K, 26146884693 28‘ Komfeld, 8., R 1931_ 19. Baeflliger, J_ U. 122 15. Schindler, P. A., Settineri, C. A., Collet, X., Fielding, C. J ., and Burlingame, A. -L., Protein Science, 2, 791-803, 1995. 16. Liu, J., Volk, K. J ., Kerns, E. H., Klohr, S. E., Lee, M. S., and Rosenberg, I. E., J. Chromatogr., 632, 45-56, 1993. 17. Fu, D. and Halbeek, H. V., Anal. Biochem., 206, 53-63, 1992. ‘ 18. Sutton, C. W., O’Neil, J. A., and Cottrell, J. S., Anal. Biochem., 218, 34-46, 1994. 19. Mizuochi, T., Yonemasu, K., Yamashita, K., and Kotaba, A., J. Biol. Chem, 253, 7404-7409, 1978. 20. Yamamoto, T. Tsuji, and T. Osawa, M01. Biotech, 3, 25-36, 1995. 21. West, I. And Golding, 0., Mol. Biotech, 2, 147-155, 1994. 22. Endo, T., J. Chromatogr., 720, 251-261, 1996. 23. Ogata, S., Muramatsu, T ., and Kobata, A., J. Biochem. (Tokyo), 78, 687-696, 1975. 24. Ogata, S., Muramatsu, T., and Kobata, A., Nature, 259, 580-582, 1976.. 25. Narasimhan, S., Freed, J. C., and Schachter, H., Carbohydr. Res, 149, 65-83, 1986. 26. Taniguchi, T., Adler, A, J., Mizuochi, T., Kochibe, N., and Kobata, A., J. Biol. Chem, 261, 1730-1736, 1986. 27. Yamashita, K., Kochibe, N., Ohlcura, T., Ueda, I., and Kobata, A., J. Biol. Chem, 260, 4688-4693, 1985. 28.Kornfeld, S., Reitman, M. L., and Komfeld, R., J. Biol. Chem, 256, 6633-6640, 1981. 29. Baenziger, J. U. and Fiete, D., J. Biol. Chem, 254, 9795-9799, 1979. 30. Yamashita, K., Hitoi, A., and Kobata, A., J. Biol. Chem, 258, 14753-14756, 1983. 311mm, 7., Ts 5611566, 1981 32 Cummings, R. 33 Yamashita, K, Biol. Chem, 26 31Yoshima, H, 1 Chem, 256, 84 35 laniguchi, 1., Kobata, A., Bio 35 Parekn R. 3., ' 6111107., 6, 12 37Hortin G_ L, A 351101111 G. L. an lgiwassannan, P. 1 0. Mortillo, s. Maillmalian Egg 13119.8. s, v llGreve, J. M., S: 759. 1982, 428mm”) 3. Tsul ill/ion, E, Taka: Biochemistry, 30 123 31.1rimura, T., Tsuji, T., Tagami, S., Yamamoto, K., and Osawa, T., Biochemistry, 20, 560-566, 1981. 32. Cummings, R. D. and Komfeld, S., J. Biol. Chem, 257, 11230-11234, 1982. 33. Yamashita, K., Totani, K., Ohkura, T., Takasaki, S., Goldstein, I. J., Kobata, A., J. Biol. Chem, 262, 1602-1607, 1987. 34. Yoshima, H., Matsumoto, A., Mizuochi, T., Kawasaki, T., and Kobata, A., J. Biol. Chem, 256, 8476-8484, 1981. 35. Taniguchi, T., MizUochi, T., Beale, M., Dwek, R. A., Rademacher, T. W., and Kobata, A., Biochemistry, 24, 5551-5557, 1985. 36. Parekh, R. B., Tse, A. G. D., Dwek, R. A., Williams, A. F., and Rademacher, T. W., Ell/$0.1, 6, 1233-1244, 1987. 37. Hortin, G. L., Anal. Biochem., 191, 262-267, 1990. 38. Hortin, G. L. and T rimpe, B. L., Anal. Biochem., 188, 271-277, 1990. 39. Wassarman, P. M., Bleil, J ., Fimiani, C., Florman, H.,_ Greve, J ., Kinloch, R., Moller, C., Mortillo, S., Roller, R., Salzmann, G., Vazquez, M., In Diet], J. (ed): “The Mammalian Egg Coat”, Berlinz, Springer-Verlag, pp 18-37, 1989. 10. Dunbar, B. S., Wardrip, N. J ., Hedrick, J. L., Biochemistry, 19, 356-365, 1980. ll. Greve, J. M., Salzmann, G. 5., Roller, R. J., and Wassarman, P. M., Cell, 31, 749- 759, 1982. 12. Shimizu, S., Tsuji, M., Dean, J., J. Biol. Chem, 258, 5858-5863, 1983. 13-Mon', E. Takasaki, s., Hedrick, J. L., Wardrip, N. J. Mort T» Kobata, A» Biochemistry, 30, 2078-2087, 1991. 14.Yurewicz, E. 1987. 15. Yurewicz, E. molecular and 1989. 1,6. Yurewicz, E. ( llYurewicz, E. ( 124 44. Yurewicz, E. C., Sacco, A. G., Subramanian, M. G., J. Biol. Chem, 262, 564-571, 1987. 45. Yurewicz, E. C., Sacco, A. G., Subramanian, M. G., in Hedrick, J. L. (ed): “The molecular and cellular biology of fertilization”, New York, Plenum Press, pp 407-427, 1989. 46. Yurewicz, E. C., Pack, B. A., Sacco, A. G., M0]. Reprod Dev., 30, 126-134, 1991. 47. Yurewicz, E. C., Pack, B. A., Sacco, A. G., M0]. Reprod. Dev., 33, 1182-188, 1992. Chapter 1 1 Llntmduction The meI' 3513 requires 1 iigosacchanides C Jones between stood sugar unit). noun most of \n in), xylose (Xyl oi .1-acetylneurar commonly found in Sequence an that interest for ma tulisionally inducet last for the structui he Oligosaccharide “Wage ofthe glyci [351811 ring have 311111152, When thi ti a811mm are design 125 Chapter 5. A New Glycosylamine Derivative for Sequence Analysis of Oligosaccharides by MALDI-PSD-MS and ESI-MS 1. Introduction The complete characterization of glycoproteins, glycolipids, and polysaccharides usually requires the structural analysis of their respective oligosaccharides (1-2). Oligosaccharides consist of monosaccharide units jointed together by glycosidic bonds (linkages between C(1), the anomeric carbon, of one sugar unit and an OH group of a second sugar unit). About 80 different kinds of naturally occurring glycosidic linkages are known, most of which involve mannose (Man), glucose (Glc), galactose (Gal), fUCOSC (FUC), xylose (Xyl), N-acetylglucosamine (GlcNAc), N-acetylgalactosamine (GalNAc), and N—acetylneuraminic acid (NeuAc, sialic acid). The structures of monosaccharides commonly found in eukaryotic glycoproteins and glycolipids are illustrated in Figure 5.1. Sequence analysis of oligosaccharides by mass spectrometry has been a topic of great interest for many years. Fast atom bombardment mass spectrometry (FAB-MS) and collisionally induced dissociation (CID) MS/MS have been primarily used in the recent past for the structural analysis of oligosaccharides (3-4). The simplest fragmentation of the oligosaccharides occurring during FAB-MS and FAB-MS/MS results from the cleavage of the glycosidic bonds. More complex processes involving the fragmentation of the sugar ring have been observed, particularly in CID-MS/MS spectra. As illustrated in Figure 5.2, when the charge is retained on the non-reducing end of the oligosaccharides, fragments are designed as A1, Bi and C, where i represents the position of the glycosrdrc CH20H a-D-Gluco CHZOH O a-D~Manno B'D‘XYIOS. light 5.1: Stmctu- Endglycohpins. 126 HO CHZOH CHZOH . O , 0 HO H CGH12 06 HO C6H1206 0H 180.2 OH 180.2 OH OH a-D-Gtucose a-D-Galactose CHZOH 0H ' 0H 0 CH3 0 HO H HO OH Ce H1206 C(5141205 . OH 180.2 HO 164.2 a—D-Mannose a-L-Fucose ' HO CHZOH CHZOH O O HO H O . . H O NH NH / OH / CH C H O N _ C3 H15 05N /C=0 2281.215 6 /C—0 221.2 CH3 CH3 N-Acetyl-ot-D-glucosamine N-Acetyl-0t-D-galactosamine OH HOHZC\ CH/ OH O \CH HO / O CO H HO OH 2 HO HN H O / OH 05 H10 05 0:C\ C11H1909N 150.1 CH 309.3 3 B-D-Xylose a-N-Acetylneuraminic acid Figure 5.1: Structures of monosaccharides commonly found in eukaryotic glycoproteins and glycolipids. bond cleaved, cou: reducing sugar uni bond counted from H0 ) 0,2A 127 bond cleaved, counted from the non-reducing end. On the other hand, ions containing the reducing sugar unit are labeled as X3, Yj, and Z3, where j is the number of the glycosidic bond counted from the reducing end (5). _——.——.-—_—_—__ OH // / /I F/ / / V f r y/ 0721’31 B1C1 294A; 32 C2 275133 \'”""_"\ Figure 5.2: Types of oligosaccharide fragmentation. As an alternative to fast atom bombardment desorption ionization (FAB) (3-4), two powerful ionization techniques, matrix-assisted laser desorption ionization (MALDI) (6) and electrospray ionization (ESI) (7) provide means for more sensitive structural determination of oligosaccharides. MALDI, used typically with time-of-flight (TOF) mass analyzers, has been demonstrated to be a promising tool to determine the molecular mass of oligosaccharides from purified sample or complex mixtures with great sensitivity (6, 8- 9)- At an early stage of its development, MALDI TOF-MS lacked the capacity for Structural analysis because relatively little or no fragmentation was observed in the spectra, although the 00ml Nimble in de‘ mmsion of MA]- :olewlar weight ‘ noduced early in a portion of the :eiectron-equibpee ehnique for stru riddled ions of condin'ons, althoug ad intensive interr a complicate or e In addition indentation of r We hagmentatio 191 It was also re 1an unambiguous e 1111 may be inefl‘ic A variety of 11110131 for lmPro 1113310110 mass hinged conjug; Llltosylation (25.27) 128 although the combination of exoglycosidase digestion and MALDI-MS has been shown to be valuable in determining of the primary structure of oligosaccharides (10). A recent extension of MALDI TOF-MS allows structural information to be obtained in addition to molecular weight data, by means of so-called post-source-decay (PSD) analysis (1 1-13) as introduced early in Chapter 1. With this technique, fragments formed by decomposition of a portion of the intact ionized molecules in the field-flee region can be detected in reflectron-equipped instruments. Spengler et al. have demonstrated the feasibility of PSD technique for structural analysis of carbohydrates (14). They reported that sodium cationized ions of native oligosaccharides undergo extensive fragmentation under PSD conditions, although reducing and non-reducing terminal fragments, cross-ring fragments and intensive internal fragments were all produced from native oligosaccharides , which may complicate or even compromise sequence assignment (14-15). In addition to MALDI-MS, ESI-MS has been used successfully for the characterization of macromolecules including carbohydrates (16-19). ESI MS/MS and in- source fragmentation have been extended to the oligosaccharide structural analysis (17- 19). It was also reported that fragment ions from native carbohydrates do not always allow unambiguous detemiination of the sequence, and fitrther that underivatized samples in ESI may be inefficiently ionized (17-19). A variety of derivatization methods of oligosaccharides have been developed previously for improving sensitivity for HPLC separation and investigation of structural details prior to mass spectrometric experiments (20-27). Most of these methods employ reducing-end conjugation which involves either reductive amination (21-24) or direct glycosylation (25-27) depending on the conjugating reagents. Both types of derivatives provide a has possibly direc HPLC separa spectrometry. In this new glycosyla reducing end \ carbohydrate e useful for facil llS. Z-Exueriment Materials Lacto-h Sigma Chemjca NY)" 3‘Amino 309 Aldrich c will” further sample Was a git The catic MD) was PUrch 129 provide a basic moiety which serves as a good protonation site during ionization and may possibly direct fragmentation after protonation. All these derivatization products require HPLC separation, and sometimes, extra desalting procedures before analysis by mass spectrometry. In this chapter, we describe the preparation and mass spectral characteristics of a new glycosylamine derivative of oligosaccharides formed by direct glycosylation at the reducing end with 3-aminoquinoline (3AQ), a matrix material which has been reported for carbohydrate analysis in MALDI-MS (28). This new type of derivative was found to be usefirl for facile sequence analysis of oligosaccharides under MALDI-PSD-MS and ESI- MS. 2. Experimental Materials Lacto-N-tetraose (LNT), maltohexaose (G6), maltoheptaose were purchased from Sigma Chemical Co. (St. Louis, MO, USA). Lacto-N-fiicopentaoses (LNFP-l, LNFP-Z) and LS-tetrasaccharide c (LSTc) were purchase from Oxford Glycosystems (Rosedale, NY). 3-Aminoquinoline (3AQ) and 2,5-dihydroxybenzoic acid (DHB) were purchased from Aldrich Chemical Co. (Milwaukee, WI). These commercial materials were used without fiarther purification. Cyclamen seed xyloglucan triantennary oligosaccharide sample was a gift from Dr. C. Hervé du Penhoat (Department de Chimie, E.N.S. , France). The cation-exchange resin (i.e., analytical grade 200-400 mesh AG 50W-X8 in H+ form) was purchased from Bio-Rad Laboratories (Hercules, CA). Cation-exchange resin was COIlVCl Biemann (2! Derivatizati Oligi with 25 u] c of 200 mg/n ul acetic acir Sample Pre} One evaporated L n 100 mg/ Approximate delivered to and used dire beads lefi 0n r‘blown Ofl” ‘ HPLC 073A, HPLC PC Comilllte: chromatOgraF 130 was converted to the NH.” form according to the procedure outlined by Wang and Biemann (29). Derivatization of Oligosaccharides with 3-Aminoquinoline (3AQ) Oligosaccharide samples (500 pmol to 500 nmol ) were dissolved in reaction vials with 25 ul of a 8:2 mixture of methanol and water. To these solutions were added 25 pl of 200 mg/ml 3-aminoquinoline (3AQ) solution freshly prepared in the same solvent and 1 pl acetic acid. The reaction mixture was heated at 80 °C for 3-4 h. Sample Preparation for .MALDI-MS One ul of the reaction mixture was deposited on a sample plate well and evaporated under vacuum. The NH;+ form cation-exchange resin beads were suspended in 100 mg/ml DHB matrix solution which was dissolved in 8:2 methanol/water. Approximately 1 ul of such matrix solution which typically contains 10-20 resin beads was delivered to the target by a pipette tip. The sample plate was then dried under vacuum and used directly for the MALDI-MS analysis and MALDI-PSD-MS analysis with resin beads left on the target. Resin beads can also be loosened with a microspatula and then “blown ofl” with a pressurized air stream as described by Rouse et al. (30). HPLC of 3A Q-treated Oligosaccharides HPLC analyses were carried out using Waters model 6000 pumps controlled by a PC computer with the detector operated at 284 nm. 3AQ-treated samples were chromatographed on a Phase Sep Spherisorb S5 amino column (25 cm X 4.5 mm). The solvent sys linear gradi for ESI—ME Instrument MA Elite time-t MA, USA) delay time t PSD mode selection w. mass select were Seque reflectron \ dePending 1 reach 4102. ESI- Sllecllometr the Capillar Scanning or pump (Han 131 solvent system was run from 20% to 60% water-acetonitrile in 1 mein with a 30-minute linear gradient program. The derivatized oligosaccharide fraction was collected and used for ESI-MS or MALDI-MS. Instrumentation MALDI-MS experiments were carried out on the PerSeptive Biosystems Voyager- Elite time-of-flight (TOF) mass spectrometer (PerSeptive Biosystems Inc., Framingharn, MA, USA). The ion source accelerating voltage was 20 kV in all T OF experiments. A delay time of 50 us was used in the linear mode and a delay time'of 100 ms was used in the PSD mode before the ion extraction in the source. In PSD experiments, precursor ion selection was performed by means of an electrostatically switched ion gate, which allows a mass selectivity of about 90 M/dM (1 out of 90 mass units). PSD fragment ion spectra were sequentially recorded over 8-10 mass windows by incremental reduction of the reflectron voltages. In each spectral window 30-100 single-shot spectra were averaged depending on the sample conditions. Accuracy and precision of PSD mass assignments reach i0.2-0.3 u in the mass range of m/z 300-1600. ESI-MS experiments were performed on a F isons VG Platform electrospray mass spectrometer with a single-quadrupole analyzer. The ESI voltage was set to 2.5 KV and the capillary temperature was 100 oC. The pressure of the sheath gas was 30 psi. Scanning covered the range m/z 300-1400 with a scan rate of 2 s/scan. A Harvard syringe Pump (Hartford, CT, USA) was used in the infusion mode at a flowrate of 10 til/min. Res be labeled according t 3. Results Pre reagent we mixture, at section, wa spectrum (1 the derivati lMtNal in 1013.3 w, current fat than 90% Weight take Liki (25), the ( anllllatic a1 sappon the structure a] ll MALE 132 Resolution was sufficient so that all peaks in E81 and MALDI mass spectra could be labeled with their monoisotopic masses. Oligosaccharide fragment ions were labeled according to the nomenclature of Domon and Costello (31). 3. Results and Discussion Preliminary conditions for derivatization with 3-aminoquinoline (3AQ) as the reagent were initialized using maltohexaose (G6) as a model compounds. The reaction mixture, after on-plate desalting with ion-exchange beads as described in the experimental section, was directly subjected to linear MALDI-MS analysis. The resulting MALDI mass spectrum (Figure 5.3) revealed a major peak at m/z 1117.4 representing the [M+H]+ ion of the derivatized maltohexaose (3AQ-G6) and a minor peak at m/z 1139.4 representing the [M+Na]+ ion of the product. The [MJrNa]+ ion of underivatized maltohexaose at m/z 1013.3 was also observed in Figure 5.3 with much lower intensity. Based on the ion current ratio of derivatized and unreacted maltohexaose the reaction yield was greater than 90% after 3 hours at 80 °C. It has been noted that samples of larger molecular weight take longer for complete reaction. Like the N-(2-pyridinyl)-glycosylamines of carbohydrates studied by Her et al. (25), the derivative formed with 3-aminoquinoline is the dehydration product of an aromatic amine and the glycosyl hemiacetal (Figure 5.4). Observations in this work which SUpport the glycosylamine product instead of Schifi‘ base product or Amadori product structure are: 1.) 3AQ labeled glucose showed anomeric protons when studied by NMR; 2.) MALDI-MS molecular weight analysis of the peracetyl 3AQ-glucose derivative indicated acyclic S hydrolyn L MALDI- usually it molecule reported reductive liquid-liqi the requi; The prott advantagi The other o“gosacc desalting Na‘, in a resin lead With the sPecies in manuscri; seed 13,101 133 indicated the incorporation of four acetyl groups (five expected if the product was the acyclic Schifi‘ base); 3.) the amino group of the 3AQ derivative of maltohexaose can be hydrolyzed by acetic acid. An Amadori product, if formed, is stable to acid hydrolysis. Unlike peptides or proteins which are usually ionized as a protonated ion in MALDI-MS and ESI—MS, native neutral and acidic oligosaccharides lack basic groups and usually ionized by alkali metal cationization, especially in MALDI-MS. In fact, protonated molecules of native oligosaccharides are rarely detected in MALDI-MS. Lemoine et al. reported for the first time the protonated species of oligosaccharides in MALDI-MS after reductive amination with benzylamine (15). Nevertheless, such a protocol requires several liquid-liquid phase extractions and further reversed phase HPLC purification because of the requirement for excess reagents, especially sodium borohydride, used in the reaction. The protocol reported in this chapter allows direct analysis of glycosylamines by taking the advantage of the fact that excess reagent, 3AQ, can be used as part of the MALDI matrix. The other half of the matrix material, DHB, was used to improve the signal intensity of the oligosaccharide derivative. Cation-exchange resin was used in our protocol for on plate desalting of the derivatized products to control the variable amounts of metal ions, mainly, Na“, in different sample sources. In favorable cases, sample preparation without using the resin lead to the formation the protonated molecule of the 3AQ glycosylamine derivative. With the resin protocol, protonated molecules could always be generated as the dominant species in the MALDI mass spectrum of the other oligosaccharides investigated in the manuscript including LNT, maltoheptaose, LSTc, LNFP-l, LNFP-2, and the Cyclamen seed xyloglycan triantennary oligosaccharide. hams maltohex 134 [M+Hl+ 1117.4 .4?! .53 .s. .“2’ E O a: [M+Nart 1139.4 1013.3 . ._ ll-.. - L. 860 90b 10b0 11'00 1200 m/z Figure 5.3: Linear positive MALDI-MS spectrum of 3-aminoquinoline (3AQ) derivatized maltohexaose (G6). The reaction mixture was used for direct MALDI-MS analysis. Scheme the mod. 135 on OH OH “ZN / o o ( 3AQ) \ n H L 0 H 0“ > H0 80°C, 3-4 hr, HAc on on OH O O O / H o H L o H NH H0 \ OH OH 4 OH Scheme 5.4: 3-Aminoquinoline glycosylamine derivative formation using maltohexaose as the model compound. analys non-rt intern; sequer from I which ions 0: likely fragme suitabl MALI 1 125.4 m 11 1M+H; mass 5 gangljg The ab Hepr the fed and my. bonds) 136 As reported before by Lemoine et al. (15) and Spengler et al. (14), MALDI-PSD analysis on [M+Na]+ of native or derivatized oligosaccharide produced both reducing and non-reducing fragments. For native oligosaccharides, cross-ring fragments and intensive internal fiagments were also present in the MALDI-PSD spectrum, which can complicate sequence assignment. Figure 5.5 shows the positive PSD spectrum of the [M+H]+ ion from 3AQ-G6. A series of abundant, evenly spaced Y ions were observed in Figure 5.5 which allowed the sequence to be easily assigned. No significant fragments for B type ions or cross-ring cleavage were observed. In this case, the charge for [M+H]+ ion is most likely to reside on the quinoline ring due to its high proton affinity and observed the fragmentation pattern. The reaction conditions outlined in the experimental section were found to be suitable for acidic oligosaccharides like LSTc. Figure 5.6(a) presents the linear positive MALDI mass spectrum obtained from 3AQ derivatized LSTc. The peak of [M+H]Jr at m/z 1125.4 again dominated the spectrum accompanied by a relatively small [M+NaT peak at m/z 1147.4. A peak resulting from the loss of N-acetylneuraminic acid (NeuAc) from the [M+H]+ ion was observed at m/z 834.3. This type of neutral loss in the positive MALDI mass spectrum was also reported previously for NeuAc containing compounds like gangliosides (32-33). The PSD spectrum of the [M+H]+ ion is shown in Figure 5.6(b). The abundant Y ion series determined the sequence of the molecule to be NeuAc-Hex- HexNAc-Hex—Hex. The cleavage of the HexNAc glycosidic bond with charge retention on the reducing end led to the formation of the B3 ion at m/z 657.2. The ions at m/z 204.1 and m/z 366.1 corresponded to two internal fragments (i.e., double cleavage of glycosidic bonds) because of the presence of HexNAc moiety. Figui m/z ] carbo ‘, FL 3000“ 137 Relative Intensity .54 4L1“ A1 A , 20'0 460 660 860 loob m/z Figure 5.5: Positive ion MALDI-PSD mass spectrum of 3AQ-G6. The [M+H]+ ion at m/z 1117 .4 was selected as the precursor ion. The monosaccharide symbols in the carbohydrate structures in all figures are defined as follows: U, Glc; ; ', GlcNAc; A, Xyl; A, Fuc; 0, Gal; 0, NeuAc. Oligosaccharide fragment ions in all figures were labeled according to the nomenclature of Domon and Costello. Figure tetrasa Underit [Mm] 138 a lM+HI+ 1125.4 3.? 3 .93 ,5. .L'>f E [MNuA Ht 0 834.3 [WNW 1147.4 I 102‘13 805 90h 1000 1130 1:00 III/Z [MJrHTL 1125.4 Relative Intensity . .1, I I I I I 200 400 600 800 1000 III/Z Figure 5.6: (a) Linear positive MALDI-MS spectrum of 3AQ-derivatized LS- tetrasaccharide c (LSTc). The small peak at m/z 1021.3 represented the [M+Na]+ ion of underivatized LSTc. (b) Positive ion MALDI-PSD mass spectrum of 3AQ-LSTc with [M+H]+ ion at m/z 1125.4 selected as the precursor ion. For symbols, see Figure 5.5. analys derive ion se sh0w1 Were 1 0n the helped m/z 13 reSldm the p” 139 Differentiation of two isomeric fucosylated pentasaccharides, LNFP-l and LNFP- 2, also could be easily achieved by MALDI-PSD analysis after 3AQ glycosylation. The [M+I—I]+ ion of either derivatized products had the same m/z value, 980.4. LNFP-l, the isomer with linear structure, was characterized by Y ion series from Yo to Y4 in the MALDI-PSD spectrum as shown in Figure 5.7(a). The branching pattern of its isomer, LNFP—Z, was clearly indicated by two adjacent Y3 fiagments corresponding to the loss of either one or the other non-reducing terminal residue at the branching site as shown in Figure 5.7(b). A small peak at m/z 672.3 was observed in Figure 5.7(b) corresponding to the loss of both non-reducing terminal residues. Because of the presence of HexNAc, the ion of m/z 204.1, the ion of m/z 366.2, and a B type ion were produced in both Figure 5.7(a) and Figure 5.7(b), which agrees with the fragmentation pattern observed for 3AQ- LSTc. To test our protocol for larger and more complicated oligosaccharide structural analysis, a triantennary oligosaccharide derived from Cyclamen seed xyloglucan (27) was derivatized with 3-aminoquinoline and subjected to direct MALDI-PSD-MS analysis. Y ion series dominated the PSD spectrum of the [M+H]+ ion of the resulting derivative as shown in Figure 5.8. Some Y ions such as Ymp at m/z 469.2 and Yza/3p at m/z 601.2 were produced from double cleavages of glycosidic bonds with the charge retention still on the reducing-end AQ group. These two Y ions at m/z 469.2 and m/z 601.2 actually helped to determine the position of an internal Xyl residue. Other two adjacent Y ions at m/z 1351.4 and m/z 1381.4 illustrated that there were two types of non-reducing terminal residues, Xyl and Hex, at different antenna. Other fragments are in good agreement with the proposed structure. Figuri LNFP. were 3 140 Relative Intensity 2j00 400 6do “ sdo m/z b Y Y Y1 [M+H]+ Oi O—[D‘AQ 980.4 .2? ca) Y1 35" 307.1 Y2 ‘9 469.2 .2 E O Y m 204.1 Y3a 3‘} 366.1 512.2 Y3w313 818. 834.3 672.3 I“..-A--;‘.Lfl._I AJH-. ' I 200 400 600 800 m/z Figure 5.7: Positive ion MALDI-PSD mass spectrum of 3AQ-derivatized (a) LNFP-l (b) LNFP-2. The [M+H]+ ion of either derivatized isomer had same m/z value, 980.4, and were selected as the precursor ions for both analyses. For symbols, see Figure 5.5. Flgur Oligos W2 1 5 sh wives 0>CN~0NH am 141 [M+H]‘J Y3“. Yza [Hg 1513.5 29 Y 1 CF AQ g 40: Y3(\~_‘L Y213 3 A 0 5 O O Y3a' Y3p S Y3“. Y4a' 0 Yze 1351.4 94 Y1 307.1 Yza/zp Y3a'/4 a" 12194 . Y 401' 469 2 Y2a/3 [‘3 Y2a Y3 “73 0." 10574 1189.4 1381.4 ' 6012 763.3 925.3 1087.4 1 l l b I l m/z 250 500 750 10 O 1250 1500 Figure 5.8: Positive ion MALDI-PSD mass spectrum of a 3AQ-derivatized triantennary oligosaccharide which was derived from Cyclamen seed xyloglucan. The [M+H]+ ion at m/z 1513.5 was selected as the precursor ion. For symbols, see Figure 5.5. of t HPLt Chror ennt mohr mmgr found andtl produ nudh hund ngns addh SOluti< tempe alter tj analysj that 0] cone \ lM+H: about; 142 After the oligosaccharide has been labeled with the chromophore, 3AQ, separation of the resulting glycosylamine product from the reaction mixture can be achieved by HPLC as shown in Figure 5.9 in which derivatized maltohexaose was used as a model. Chromatographic separation revealed two isomeric forms of 3AQ-G6 because of the existence of two different positions (i.e., equatorial or axial position) of AQ group in the molecule. The yield of the total 3AQ-G6 obtained from this procedure was measured by integrating the peaks observed during the HPLC against an external standard and was found to be ~70%. The difference between the reaction yield from HPLC peak integration and that from MALDI-MS ion current integration may be due to the fact that derivatized product (3AQ-G6) had higher unit response than underivatized maltohexaose when studied by MALDI-MS. The AQ group of the HPLC-separated collected products was found to be resistant to hydrolysis under mildly basic or neutral conditions and showed no signs of hydrolysis after ten days in 50% aqueous acetonitrile solution. However, under acidic conditions the AQ group is gradually hydrolyzed. A 2% aqueous acetic acid solution brought about complete hydrolysis of the AQ group in one day at room temperature. This actually provides a way to regenerate the underivatized carbohydrate after the HPLC separation. The fractions of 3AQ-G6 from HPLC were collected and subjected to ESI-MS analysis. The intensity of the [M+H]+ ion was also found to be much more intense than that of [M+Na]+ ion or [M+K]+ ion. No significant fragmentation was observed when cone voltage (i.e., the capillary-skimmer voltage) was set at 20 V. Detection of the [M+H]+ ion was achievable at the level of 1 pmol of total sample infirsed, which provides about a 50-fold increase in sensitivity compared to that for underivatized maltohexaose on Figl Wate 143 - 410V - Y - - 10.0 V 1 16.0 Minutes Figure 5.9: HPLC separation of 3-aminoquinoline (3AQ) derivatized maltohexaose (G6). on a Phase Sep Spherisorb SS amino column. The system was run from 20% to 60% water-acetonitrile in 1 ml/min with a 30 minute linear gradient program. the 5 occur those When Yo iO 144 the same ESI instrument. It was further noted that intensive in-source fragmentation occurred at elevated cone voltage for infused 3AQ-G6. The fragment ions resembled those observed in Figure 5.2, which were generated during analysis by MALDI-PSD—MS. When the cone voltage was raised to 100 V, the complete series of Y ions ranging from Yo to Y5 were all produced as shown in Figure 5.10. F igl (cap mini 145 Relative Intensity 200 400 600 800 1000 m/z Figure 5.10: Positive ion ESI—MS spectrum of HPLC separated 3AQ-G6. Cone voltage (capillary-skimmer voltage): 100 V. The spectrum shown represented an average over 1- min recording time. For symbols, see Figure 5.5. deri prot forn ions dete react matr the E skirm analy 146 4. Conclusion In this study a new glycosylamine derivative was developed for facile sequence analysis of oligosaCcharides by matrix-assisted laser desorption ionization mass spectrometry (MALDI-PSD-MS) and electrospray ionization mass spectrometry (ESI- MS). 3-Aminoquinoline (3AQ) was used as the reagent for making glycosylamine derivatives at the reducing terminus. This type of derivative produces primarily protonated molecules for neutral and acidic oligosaccharides in MALDI-MS without the formation of significant [M+Na]+ ions. Post-source decay (PSD) analysis of such [M+HT ions yields a spectrum dominated by Y ions, allowing the sequence to be readily determined. Another attractive feature of this derivative in MALDI-MS analysis is that reaction cleanup is not necessary and excess reagent can be used as part of the MALDI matrix. Derivatized oligosaccharides also can be separated fi'om the reaction mixtures by HPLC and subject to ESI-MS study. Protonated molecules are also dominant species in the ESI spectrum. Intensive in—source fragmentation was observed at elevated capillary- skimmer voltage. The fragmentation pattern was very similar to that observed during analysis by MALDI-PSD-MS. Re 13. 147 References: 1. Dell, A. and Reason, A. J ., Curr. Opin. Biotechnol. 4, 52, 1993. 2. Cumming, D. A., Glycobiology, 1, 115-130, 1991. 3. Barber, M., Bordoli, R 8., Elliot, G. J ., Sedgwick, R. D., and Tyler, A. N., Anal. Chem, 54, 645A—657A, 1982. 4. Carr, S. A., Reinhold, V. N., Green, B. N., and Hass, J. R., Biomed Mass Spectrom., 12, 288, 1985. 5. Hillenkamp, F ., Karas, M., Beavis, R. C., and Chait, B. T., Anal. Chem, 63, 1193A- 1203A, 1991. 6. Fenn, J. B., Mann, M., Meng, C. K., Wong, S. F., and Whitehouse, C. M., Science, 246, 64-71, 1989. 7. Stahl, B., Steup, M., Karas, M., and Hillenkamp, F ., Anal. Chem, 63, 1463-1466, 1991. 8. Stahl, B., Thurl, S., Zeng, J ., Karas, M., Hillenkamp, F ., Steup, M., and Sawatzki, G., Anal. Biochem., 223, 218-226, 1994. 9. Harvey, D. J., Rudd, P. M., Bateman, R. H., Bordoli, R. S., Hayes, K., and Vickers, R. G., Org. Mass Spectrom., 29, 753, 1994. 10. Sutton, W. C., O’Neil, J. A., and Cottrell, J. 8., Anal. Biochem., 218, 34, 1994. 11. Huberty, M. C., Vath, J. B, Yu, W., and Martin, S. A., Anal. Chem. 65, 2791-2799, 1993. 12. Kaufinann, R., Spengler, B., and Ltitzenkirchen, F., Rapid Commun. Mass Spectrom. 7, 902-910, 1993. 13. Kaufinann, R, Kirsch, D., and Spengler, B., Int. J. Mass Spectrom. Ion Processes 131, 355-385, 1994. l4 l7 19. 20. 21. 22. 23. 24. 25. 26. 14. 148 Spengler, B., kirsch, D., Kaufmann, R., and Lemoine, J ., J. Mass Spectrom., 30, 782- 787, 1995. 15. Lemonie, J ., Chirat, F ., and Domon, B., J. Mass Spectrom., 31, 908-912, 1996. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. Smith, R. D., Loo, J. A., Edmonds, C. G., Barinaga, C. J., and Udseth, H. R, Anal. Chem, 62, 882, 1990. Yoshino, K., Takao, T., Murata, H., and Shimonishi, Y., Anal. Chem, 67, 4028-4031, 1995. Gu, J ., Hiraga, T., and Wada, Y., Biomed Mass Spectrom., 23, 212-217, 1994. Okamoto, M., Takahashi, K-I, and Doi, T, Rapid Commun. Mass Spectrom. 9, 641- 643,1995. ‘ Domon, B., Mueller, D. R., and Richter, W. J., Biomed. Mass Spectrom., 19, 390, 1990. Hase, T. Ibuki, and T. Ikennaka, J. Biochem., 95, 197, 1984. Dell, A., Carman, N. H., Tiller, P. H., and Thomas-Oates, J. B, Biomed. Mass Spectrom., 16, 19-24, 1988. Wang, W. T., LeDonne, N. C., Jr., Ackerrnan, B., and Sweeley, C. C., Anal. Biochem., 141, 366-381, 1984. Poulter, L. and Burlingame, A. L., in Methods Enzymol. (J. A. McCloskey, Eds), Academic Press, San Diego, 193, 661-689, 1990. Her, G. R., Santikarn, S., Reinhold, V. N., and William, J. C., J. Carbohydr. Chem, 6, 129-139, 1987. Caesar, J. P. Jr., Sheeley, D. M., and Reinhold, V. N., Anal. Biochem., 191, 247-252, 1990. —7— 149 27 . Braccini, 1., Hervé du Penhoat, C., Michon, V., Goldberg, R., Clochard, M., Jarvis, M. C., Huang, Z.-H., and Gage, D. A., Carbohydr. Res, 276, 167-181, 1995. 28. Metzger, J. O., Woisch, R., Tuszynski, W., Angermann, R., Fresenius J. Anal. Chem, 349, 473—475, 1994. 29. Wang, B. H. and Biemann, K., Anal. Chem, 66, 1918, 1994. 30. Rouse, J. C., and Vath, J. E., Anal. Biochem., 238, 82-92, 1996. 31. Domon, B., and Costello, C. E., Glycoconj. J., 5, 397-409, 1988. 32. Juhasz, P. and Costello, C. E., J. Am. Soc. Mass Spectrom., 3, 785-796, 1992. 33. Xu, N., Huang, ZZ-H., Watson, J. T., and Gage, D. A., J. Am. Soc. Mass Spectrom., 8, 116-124, 1997. 1 nm TC cha app disc dnt den indh exce men the con» 150 Chapter 6. Summary The deve10pment of matrix—assisted laser desorption ionization mass spectrometry (MALDI-MS) coincides with an increasing demand for more accurate and sensitive techniques in fields such as biochemistry, molecular biology, genetics and biotechnology (1). The MALDI time-of-flight (TOF) mass spectrometer has been quickly recognized as a “molecular weight machine” because of its superior capability in molecular mass determination of a wide range of biopolymers. The introduction of the post-source decay (PSD) technique allowed fiirther structural information to be obtained, in addition to molecular weight data, on reflectron-equipped TOF instrument. In this research, MALDI TOF-MS, as well as the PSD technique, has been used as the primary tool for structural characterization of glycoconjugates including glycoproteins. The success of analysis by MALDI-MS, in part, depends on the choice of apprOpriate matrix material. 2-Mercaptobenzothiazole (MBT) and its analogs have been discovered as a class of new matrices for MALDI-MS. These compounds are structurally distinct from the conventional matrices. MBT has been used successfirlly for the desorption of proteins up to 100,000 daltons. A comparison with conventional matrices indicates that MBT provides the same level of sensitivity and resolution and offers excellent tolerance to sample contaminants such as ionic detergents. 5-Chloro-2- mercaptobenzothiazole (CMBT), an analog of MBT, has been found not only effective for the analysis of peptides, low-mass proteins, and glycolipids, but also superior to conventional matrices for the analysis of glycopeptides and oligosaccharides. SE 151 The information gained fi'om the matrix studies has been applied to the structural analysis of peptidoglycans, a type of complex glycoconjugates consisting bacterial cell wall with important immunological significance. In analysis by MALDI-MS, peptidoglycan samples were subjected to enzymatic digestion with muramidase and the resulting muropeptides were purified by HPLC. CMBT was employed for routine analysis by MALDI-MS of muropeptides, which are problematic analytes in MALDI-MS by using conventional matrices. The results have demonstrated that sub-picomole to femtomole detection can be achieved in both positive mode and negative mode allowing unambiguous determination of molecular masses of muropeptides of peptidoglycans. Structural information of muropeptide monomers was obtained by further post-source decay (PSD) analysis. F ragrnentation patterns in positive mode and negative mode PSD are found to be complementary to each other for the elucidation of the composition of the peptide chains and oligosaccharide moieties. Enzymatic digestion was also incorporated with MALDI- MS in solving structural problems of muropeptide oligomers from pathogenic Staphylococcus aureus strains. Specific cleavage between pentaglycine bridge residues was performed by lysostaphin digestion for these oligomers and the resulting hydrolysates were subjected to direct MALDI-MS. The combination of an analysis by PSD and the lysostaphin digestion approach led to the determination of structural modification of peptidoglycans from Staphylococcus aureus strains. MALDI-MS also played an important role in the study of porcine oocyte zona pellucida 301 glycoprotein (ZP3a). Tryptic digest of ZP3a glycoprotein was separated by reverse phase HPLC and four heterogeneous glycopeptide fractions were identified by their MALDI-MS spectra. N-linked and 0—linked glycopeptides of ZP3a glycoprotein 1'61 re: CO 011 011 of the pit tea 152 were distinguished by lectin (jacalin) affinity chromatography / HPLC / MALDI-MS. The MALDI mass spectrum of the glycopeptide at Asn 203 of ZP3a showed more microheterogeneity than that of the glycopeptide at An 220 or Asn 333. After the treatment with specific exoglycosidases which released most of the outer branches of these glyc0peptides, MALDI-MS analysis revealed that the glycan at Asn 203 consists a mixture of fucosylated and unfucosylated bi- and tri-antennary structures, whereas the glycan at Asn 220 or Asn333 is composed of a mixture of fucosylated complex bi-, tri-, and tetra- antennary oligosaccharides. To minimize sample handling between the stages of lectin aflinity binding and MALDI-MS, a micro-batch lectin binding process was also developed for direct MALDI-MS analysis for isolated glycopeptides. This approach is capable of identifying glycopeptides derived from ribonuclease B and avidin, proteins having a single glycosylation site. The complete characterization of glycoproteins and other types of glycoconjugates requires the structural analysis of their respective oligosaccharides. However, the resulting spectra from MALDI-PSD analysis on native oligosaccharides are rather complex and difficult to interpret. Reducing and non-reducing terminal fragmentation, cross-ring fragmentation and internal fragmentation all occur to the native oligosaccharides, complicating sequence determination. A new glycosylamine derivative of oligosaccharides has been developed for facile sequence analysis in MALDI-MS. 3- Aminoquinoline was used as the derivatizing regent in making glycosylamine derivatives at the reducing terminus. ' This type of glycosylamine derivative produces primarily protonated molecules in MALDI-MS. Reaction cleanup is not necessary because excess reagent can be used as the MALDI matrix without interfering the analysis by MALDI-MS. 153 PSD analysis of the protonated molecule yields a spectrum dominated by reducing terminal fragments allowing easy determination of the oligosaccharide sequence. Reft 154 References: 1. Siuzdak, G., Proc. Natl. Acad. Sci. USA, 91, 11290-11297, 1994. HICHIGQN STRTE U l l Itlllttlilllt 312930156309