UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Capillary electrophoresis mass spectrometry for the characterization of glycoproteins and N-glycans Jayo, Roxana Gabriela 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2015_february_jayo_roxana.pdf [ 5.01MB ]
Metadata
JSON: 24-1.0166079.json
JSON-LD: 24-1.0166079-ld.json
RDF/XML (Pretty): 24-1.0166079-rdf.xml
RDF/JSON: 24-1.0166079-rdf.json
Turtle: 24-1.0166079-turtle.txt
N-Triples: 24-1.0166079-rdf-ntriples.txt
Original Record: 24-1.0166079-source.json
Full Text
24-1.0166079-fulltext.txt
Citation
24-1.0166079.ris

Full Text

CAPILLARY ELECTROPHORESIS MASS SPECTROMETRY FOR THE CHARACTERIZATION OF GLYCOPROTEINS AND N-GLYCANS  by Roxana Gabriela Jayo  B.Sc., Pontifical Catholic University of Peru, 2005 Professional Degree in Chemistry, Pontifical Catholic University of Peru, 2006   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE   DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Chemistry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  December 2014  © Roxana Gabriela Jayo, 2014 ii  Abstract Glycosylation of proteins is ubiquitous and has the ability to significantly alter the biological and biophysical properties of proteins. The need to study structure-function relationships of glycans in a living organism requires the continuous development of rapid and sensitive technologies for the characterization of glycan components. In recent years, a broad range of technologies have evolved to provide new developments and emerging glycomics techniques. In this thesis we present the development of new methodologies based on capillary electrophoresis-electrospray ionization-mass spectrometry (CE-ESI-MS) for the study of protein N-glycosylation in complex biological systems and therapeutic recombinant drugs. Our approach involves the use of a flow-through microvial, a novel technology that provides a robust and easy-to use strategy for interfacing CE separation with MS detection.  In chapter 2, we report a simple and robust CE-ESI-MS methodology for comprehensive characterization of glycosylated proteins at the level of intact protein, enzymatically released glycopeptide and glycans. In chapter 3, we characterize a complex set of enzymatically released N-glycans from a recombinant therapeutic drug that revealed extensive glycan heterogeneity. The study demonstrated the potential of our approach to complement established techniques for glycan characterization of therapeutic glycoproteins in the pharmaceutical industry.  Chapters 4 and 5 of this thesis are devoted study protein glycosylation of relevant biological systems. In chapter 4, O-acetylated N-glycans from fish serum were characterized in their native state and the structural variations of their isomeric species were investigated by tandem MS approaches. The developed CE-MS methods may be useful not only for the characterization of acetylation of complex glycans but also to study other types of glycan modifications in different contexts.  iii  In chapter 5, we present CE-MS methodologies for characterizing protein N-glycosylation in human serum associated with prostate cancer and asthma. Comparison of glycan compositions and relative abundances revealed abnormal glycosylation in prostate cancer and asthma serum. The capability of our CE-ESI-MS method to perform global glycan profiling of human serum demonstrates its potential for comprehensive glycan profiling in the context of malignancies and for the discovery of glycan disease markers with high selectivity and specificity.   iv  Preface Aside from the exceptions listed below, all results presented in this thesis are my own work.  Analysis of the results was carried out in consultation with my supervisor, Professor David. D. Y. Chen.  Contributions from other researches: Chapter 2:   The enzymatic release and analysis of resulting glycopeptides derived from  RNase B was carried out in collaboration with Cai Tie. Chapter 3:   The CE-MS analysis of recombinant human Erythropoietin using an ultra-high resolution mass spectrometer was performed under the supervision of Dr. Rawi Ramautar, at Dr. Thomas Hankemeier’s research group, Leiden University, the Netherlands.   The LC-MS/MS analysis of N-glycans released from recombinant human Erythropoietin was performed in collaboration with Dr. Morten Thaysen-Andersen at the University of Macquarie, Sydney, Australia. Chapter 4:    The analysis of N-glycans samples from fish serum corresponding to different weeks of stress experiments was performed under the supervision of Dr. Jianjun Li at the National Research Council of Canada (Ottawa).  The APTS-derivatization of glucose ladder was performed in collaboration with Dr. Jane E. Maxwell.   Chapter 5:  The CE-MS analysis of glycosylation in human serum using an ultra-high resolution mass spectrometer was performed under the supervision of Dr. Peter Lindenburg, at Dr. Thomas Hankemeier’s research group, Leiden University, the Netherlands.     Publications arising from work presented in the dissertation: Simple Capillary electrophoresis-mass spectrometry method for complex glycan analysis using a flow-through microvial interface. Jayo, R. G.; Thaysen-Andersen, M.; Lindenburg, P. W.; Haselberg, R.; Hankemeier, T.; Ramautar, R.; Chen, D. D. Y. Anal. Chem. 2014, 86, 6479–6486.   Material from this manuscript is included in Chapter 3. Reused with permission. Copyright (2014) American Chemical Society. v  Capillary electrophoresis mass spectrometry for the characterization of O-acetylated N-glycans from fish serum. Jayo, R. G.; Li, J.; Chen, D. D. Y. Anal. Chem. 2012, 84, 8756-8762.  Material from this article is included in Chapter 4. Reused with permission. Copyright (2012) American Chemical Society.  A promising capillary electrophoresis-electrospray ionization mass spectrometry method for carbohydrate analysis. Maxwell, E. J.; Ratnayake, C.; Jayo, R. G.; Zhong, X.; Chen, D. D. Y. Electrophoresis 2011, 32, 2161-2166.   Material from this article is included in Chapter 4. Reused with permission. Copyright (2011) Wiley & Sons.                 vi  Table of Contents Abstract .......................................................................................................................................... ii Preface ........................................................................................................................................... iv Table of Contents ......................................................................................................................... vi List of Tables .............................................................................................................................. xiii List of Figures ............................................................................................................................. xiv List of Abbreviations ............................................................................................................... xviii Acknowledgements ......................................................................................................................xx Dedication .................................................................................................................................. xxii Chapter 1: Introduction to capillary electrophoresis-mass spectrometry for the characterization of glycosylated proteins and their glycan derivatives ....................................1 1.1 CAPILLARY ELECTROPHORESIS-SEPARATION OF CHARGED SPECIES IN SOLUTION............................................................................................................................... 1 1.2 CAPILLARY ELECTROPHORESIS- SEPARATION BASED ON DIFFERENTIAL MIGRATION IN ELECTRIC FIELD ...................................................................................... 3 1.2.1 Separative principles .................................................................................................. 3 1.2.2 Electroosmotic flow: a mode of mass transport ......................................................... 5 1.2.3 Capillary coatings ...................................................................................................... 7 1.3 ELECTROSPRAY IONIZATION ................................................................................... 10 1.4 CE-MS INTERFACING ................................................................................................... 13 1.4.1 Challenges for coupling CE with MS detection ...................................................... 13 1.4.2 Sheath-liquid CE-ESI-MS interfaces ....................................................................... 15 1.4.3 Sheathless CE-ESI-MS interfaces............................................................................ 17 vii  1.4.4 Flow-through microvial CE-ESI-MS interface ........................................................ 20 1.5 MASS SPECTROMETRY- SEPARATION OF CHARGED SPECIES IN GAS PHASE ……………………………………………………………………………………………21 1.5.1 Quadrupole mass analyzer77 .................................................................................... 22 1.5.2 3D ion trap mass analyzer81 ..................................................................................... 25 1.5.3 Time-of-flight mass analyzer83 ................................................................................. 29 1.6 ANALYSIS OF PROTEIN N-GLYCOSYLATION USING CE-MS .............................. 31 1.6.1 General considerations of protein glycosylation ...................................................... 31 1.6.2 Analytical techniques for N-glycosylation analysis ................................................. 33 1.6.2.1 Chromatographic separation .................................................................. 34 1.6.2.2 Electrophoretic separation ..................................................................... 35 1.6.3 Glycosylation analysis ............................................................................................. 37 1.6.3.1 Intact glycoprotein analysis by CE-MS ................................................. 37 1.6.3.2 Glycan analysis by CE-MS .................................................................... 38 1.6.3.3 Reductive amination .............................................................................. 39 1.6.3.4 Labels ..................................................................................................... 40 1.6.3.5 Strategies for glycan structural elucidation............................................ 41 1.7 RESEARCH OBJECTIVES ............................................................................................. 44 1.7.1 Application of a novel CE-MS interface for comprehensive glycoprotein analysis …………………………………………………………………………………………..44 1.7.2 CE-MS characterization of enzymatically released N-glycans from relevant biological sources.............................................................................................................. 45 viii  Chapter 2: Potential for comprehensive characterization of glycoproteins by capillary electrophoresis mass spectrometry .............................................................................................47 2.1 INTRODUCTION ............................................................................................................ 47 2.2 MATERIALS AND METHODS ...................................................................................... 49 2.2.1 Materials .................................................................................................................. 49 2.2.2 Digestion of RNase B with trypsin .......................................................................... 50 2.2.3 Peptide and glycopeptide purification using solid-phase extraction (SPE) ............. 50 2.2.4 Deglycosylation of RNase B with PNGase F .......................................................... 51 2.2.5 Glycan purification using solid-phase extraction (SPE) .......................................... 51 2.2.6 Glycan derivatization ............................................................................................... 52 2.2.7 CE-MS system ......................................................................................................... 52 2.2.8 CE procedure ........................................................................................................... 53 2.3 RESULTS AND DISCUSSION ....................................................................................... 53 2.3.1 Intact glycoprotein analysis ..................................................................................... 53 2.3.2 Analysis of enzymatically released N-glycans ........................................................ 59 2.3.3 Glyco-profiling of glycosylation site of RNase B ................................................... 62 2.4 CONCLUDING REMARKS ............................................................................................ 66 Chapter 3: A simple CE-MS method for complex N-glycan analysis using a flow-through microvial interface .......................................................................................................................68 3.1 INTRODUCTION ............................................................................................................ 68 3.2 MATERIALS AND METHODS ...................................................................................... 70 3.2.1 Materials .................................................................................................................. 70 3.2.2 rHuEPO purification ................................................................................................ 71 ix  3.2.3 Glycoprotein deglycosylation with PNGase F ......................................................... 72 3.2.4 Glycan purification using solid-phase extraction (SPE) .......................................... 72 3.2.5 Electrophoretic procedure ........................................................................................ 73 3.2.6 CE-ESI-MS system .................................................................................................. 73 3.2.7 LC-MS system ......................................................................................................... 74 3.3 RESULTS AND DISCUSSION ....................................................................................... 75 3.3.1 Analysis of N-glycans from human IgG .................................................................. 75 3.3.2 N-glycan analysis of rHuEPO using CE-MS ........................................................... 80 3.3.3 Detection of glycan modifications in rHuEPO ........................................................ 85 3.3.3.1 Acetylation of N-glycans ....................................................................... 85 3.3.3.2 Sialic acid variation................................................................................ 91 3.3.3.3 Elongation of the N-glycan chain .......................................................... 93 3.3.4 LC-MS/MS analysis of N-glycans from rHuEPO ................................................... 93 3.4 CONCLUDING REMARKS ............................................................................................ 94 Chapter 4: Capillary electrophoresis mass spectrometry characterization of O-acetylated N-glycans from fish serum ..........................................................................................................97 4.1 INTRODUCTION ............................................................................................................ 97 4.2 MATERIALS AND METHODS .................................................................................... 100 4.2.1 Materials ................................................................................................................ 100 4.2.2 Sample preparation ................................................................................................ 100 4.2.3 APTS labeling ........................................................................................................ 101 4.2.4 Capillary electrophoresis ....................................................................................... 101 4.2.5 Mass spectrometry ................................................................................................. 102 x  4.3 RESULTS AND DISCUSSION ..................................................................................... 103 4.3.1 CE-LIF ................................................................................................................... 103 4.3.2 CE-MS of native N-glycans of fish serum ............................................................. 104 4.3.3 Identification of isomeric O-acetylated species of native N-glycans ..................... 108 4.3.4 Analysis of potential N-glycans isomers................................................................ 110 4.4 CONCLUDING REMARKS .......................................................................................... 113 Chapter 5: Global human serum N-glycan profiling by capillary electrophoresis-mass spectrometry ...............................................................................................................................115 5.1 INTRODUCTION .......................................................................................................... 115 5.2 MATERIALS AND METHODS .................................................................................... 118 5.2.1 Materials ................................................................................................................ 118 5.2.2 Glycoprotein deglycosylation with PNGase F ....................................................... 118 5.2.3 Glycan purification using solid-phase extraction (SPE) ........................................ 119 5.2.4 Electrophoretic procedure ...................................................................................... 119 5.2.5 CE-ESI-MS system ................................................................................................ 120 5.3 RESULTS AND DISCUSSION ..................................................................................... 121 5.3.1 Optimization of enzymatic N-glycan release ......................................................... 121 5.3.2 CE-ESI-MS separation of N-glycans from human serum ...................................... 122 5.3.3 N-glycan detection and identification .................................................................... 124 5.3.4 CE-MS separation of isomeric glycans.................................................................. 130 5.3.5 Method reproducibility .......................................................................................... 131 5.3.6 Variations of glycosylation in different types of human serum samples ............... 132 5.4 CONCLUDING REMARKS .......................................................................................... 136 xi  Chapter 6: Concluding remarks and future work ..................................................................138 6.1 CONCLUDING REMARKS .......................................................................................... 138 6.2 FUTURE RESEARCH DIRECTIONS .......................................................................... 141 6.2.1 Exoglycosidase digestion ....................................................................................... 141 6.2.2 High-throughput glycan analysis for biomarker discovery ................................... 142 6.2.3 Use of microfluidic platforms with flow-through microvial interface for the analysis of glycoproteins and their derivatives ............................................................................. 143 Bibliography ...............................................................................................................................144 Appendices ..................................................................................................................................156 Appendix A ................................................................................................................................. 156 A.1 CE-MS spectra of the glycoforms of RNAse B ....................................................... 156 A.2 CE-MS spectra for T3-labeled glycans derived from RNase B ............................... 158 Appendix B ................................................................................................................................. 161 B.1 CE-MS analysis of N-glycans from rHuEPO using the maXisTM ultra-high resolution TOF MS .......................................................................................................................... 161 B.2 Analysis of glycosylation of rHuEPO by LC-MS/MS ............................................. 163 Appendix C ................................................................................................................................. 164 C.1 CE-LIF of Glucose ladders with MS-amenable buffers........................................... 164 C.2 CE-MS of APTS-labeled N-glycans of fish serum .................................................. 166 C.3 Optimization of CE-MS conditions for native N-glycan analysis............................ 169 C.4yPotential identification of isomeric N-glycans in O-acetylated bi-antennary oligosaccharides .............................................................................................................. 171 C.5 Potential N-glycan isomers....................................................................................... 174 xii  Appendix D ................................................................................................................................. 175 D.1XOptimization of denaturation temperature and time for enzymatic N-glycan  release…. ........................................................................................................................ 175 D.2 Representative CE-MS spectra of isomeric N-glycans from human serum............. 177  xiii  List of Tables Table 2-1 Identification and annotation of the intact glycoforms and glycan components of RNase B including their abundances. Glycoforms were detected as [M+8H]8+ and [M+9H]9+ ion species, with the lower charge state being the most intense ion. Data was acquired using the LCQ* Duo ion trap MS. Glycans were detected as [M+H]+ and [M+2H]2+ species using the API 4000 triple-quadrupole MS. Composition and structural schemes for glycans are given in terms of N-acetylglucosamine (blue squares) and mannose (green circles). The number indicated by an asterisk on the right hand side panel indicate the abundance of glycans determined with LC-MS,181 as discussed in the text. . ..............................................................................................58 Table 2-2 Glycopeptide identification including the observed m/z, experimental masses and relative abundances. Glycopeptides were detected as [M+H]+ and [M+2H]2+ ion species. Data was acquired using the LCQ* Duo ion trap MS. ...................................65 Table 3-1  Identification and annotation of native N-glycans of human IgG observed with CE-MS. Glycans were observed as [M- 2H]2-  and [M- H]- species. Data was acquired using the maXisTM Ultra High-Resolution TOF MS. ..................................................79 TableX3-2XIdentification and annotation of the seventeen most abundant N-glycan monosaccharide compositions of rHuEPO shown in Figure 3.2. Glycans were observed as [M-2H]2-, [M-3H]3- and/or [M-4H]4- ion species. Data was obtained using a triple quadrupole MS.................................................................................................84 Table 4-1  Identification and annotation of the bi- (column on the left) and tri-antennary (column on the right) N-glycans present in fish serum observed as [M-2H]2- and [M-3H]3-, respectively.Composition and structural schemes are given in terms of N-acetylglucos amine (blue square), mannose (green circle), galactose (yellow circle) and sialic acid (purple diamond). For ease of identification, the numbers of O-acetyl groups present on sialic acids are included as subscripts in the N-glycan composition. ...................111 Table 5-1 List of the distinct compositions of N-linked oligosaccharides observed in human serum (control sample) using the maXisTM Ultra High-Resolution TOF MS. ..........125     xiv  List of Figures FigureX1.1XSchematic diagram of a capillary electrophoresis instrument with online optical detection. ........................................................................................................................4 Figure 1.2 (A) Structure of the electrical double layer and resulting EOF due to the deprotonation  of the bare-fused silica capillary wall. (B) Potential profile with increasing distance from the capillary wall………………………………………………………………...6 FigureX1.3XSchematic representation of the electrospray ionization process. The formation of charged droplets via'Taylor' cone and the evolution of gas phase charged analytes traveling toward the MS inlet. .....................................................................................12 FigureX1.4XCommon arrangements for sheath-liquid interfaces. (A) Coaxial sheath-liquid interface with sheath gas, (B) Liquid junction interface, (C) Pressurized liquid junction interface.58 ......................................................................................................17 FigureX1.5XDifferent arrangements for creating electrical contact in sheathless interfaces. (A) Conductive coating applied to the emitter tip, (B) wire inserted at the tip, (C) wire inserted through hole, (D) split-flow interface with a metal sheath, (E) Etched capillary, (F) junction with metal sleeve, (G) microdialysis junction, and (H) junction with conductive emitter tip.58 .......................................................................................19 FigureX1.6XSchematic diagram of the flow-through microvial CE-ESI-MS interface used in this thesis. ...........................................................................................................................21 Figure 1.7 Schematic diagram of a quadrupole with hyperbolic rods and correspondingly applied potential. The space between the rods comprises the trajectory of the ion before it hits the detector. ..................................................................................................................23 Figure 1.8 First stability area of a quadrupole with an operating line. Adapted from78. ...............24 Figure 1.9 (A) Ion trap cut in half along the axis of cylindrical symmetry. (B) Schematic diagram of the cross section of and ion trap showing the asymptotes. ......................................26 Figure 1.10 First stability region of 3D ion trap.80 ........................................................................27 FigureX1.11 Schematic diagram of a time-of-flight mass spectrometer with reflectron. Adapted from88. ..........................................................................................................................30 Figure 1.12 Types of glycosylation. The common core structure of each type of glycosylation is indicated in the gray box. The terminal monosaccharide sequences in each type of glycans are representative structural variations. Adapted from98. ...............................33 FigureX1.13XStrategies for N-glycan release from isolated glycoproteins and glycopeptides. Adapted from146. ..........................................................................................................38 Figure 1.14 Reductive amination reaction for the labeling of N-glycans after their release from glycoproteins. ...............................................................................................................40 Figure 1.15 Types of carbohydrate fragmentation as proposed by Domon and Costello. Adapted from190. .........................................................................................................................43 FigureX2.1XWorkflow of the different glyco-profiling approaches used in this study for identification of glycoform distribution, glycosylation site and glycan composition of RNase B, as a model protein. CE-MS analysis were performed using an LCQ* Duo ion trap mass and an API 400 triple quadrupole MS, as indicated in the text. ............49 Figure 2.2  Contour plot of migration time v.s m/z of RNase B glycoforms showing charge states  corresponding to [M+8H]8+ and [M+9H]9+. Same conditions are used as in Figure 2.3C…………………………………………………………………….…………….55  xv  Figure 2.3 EIE displaying the migration of intact glycoform species of RNase B using a (A) PB-DS-PB coated capillary, (B) PEI-coated capillary and (C) polyacrylamide-based neutral capillary. The optimum BGE was 100 mM ammonium acetate. In (A) and (B), the pH of the BGE was 3.0, and the CE polarity was reversed. In (C) the pH of BGE was 3.1, and CE polarity was normal. Data was acquired using the LCQ* Duo ion trap MS.........................................................................................................................57 Figure 2.4(A) Structure of the T3-labeling reagent employed for derivatizing the N-linked glycans from RNase B. (B) Base peak electropherogram showing baseline separation for N-glycans of RNase B labeled with T3. Peaks are identified as follows: (1) GlcNAc2Man5, (2) GlcNAc2Man6, (3) GlcNAc2Man7, (4) GlcNAc2Man8, and (5) GlcNAc2Man9. The inserts in the left show the putative glycan structures. Glycans were detected in positive ESI mode. Data was acquired using the API 4000 triple-quadrupole MS………………………………………………………………………60 Figure 2.5 Extracted ion electropherograms for the CE-MS separation of glycopeptides of RNase B. BGE contains 1% formic acid in methanol: water (1:1). Glycopeptides were identified as: (1) GlcNAc2Man9 at m/z 1172.57, (2) GlcNAc2Man8 at m/z 1091.69, GlcNAc2Man7 at m/z 1010.67, (4) GlcNAc2Man6 at m/z 929.64 and (5) GlcNAc2Man5 at m/z 1695.59. Data was acquired using the LCQ*Duo ion trap MS. ........................64 Figure 3.1 Density map for the CE-ESI-MS separation of neutral and acidic N-glycans from human IgG. Conditions: neutral (HPC)-coated capillary 70 cm X 50 µm; BGE: 50 mM ammonium acetate /20% methanol (pH 3.1); separation voltage: -25 kV + 10 mbar; negative ion ESI-MS; modifier solution: 10 mM ammonium acetate (pH 3.1) containing 75% of 2-propanol: methanol (2:1). The intensity of the color reflects the glycan abundance with light green being less abundant that dark blue. ......................78 Figure 3.2  EIE for the CE-ESI-MS separation of N-glycans from rHuEPO. Only the seventeen most abundant N-glycans are displayed. The glycans were detected under the same CE-MS conditions as in Figure 3.1. The monosaccharide composition of the glycans is indicated in the graph. Nomenclature is as follows: H = mannose or galactose (Hex), N = N-acetylglucosamine (HexNAc), F = fucose, and S= sialic acid (Neu5Ac)………………………………………………………………………..........82   Figure 3.3 (A) EIE obtained for the CE-MS separation of the tetra-antennary tetra-sialic glycan H7N6F1S4 (peak 1) and its acetylated species containing from 1 to 7 O-acetyl groups (peaks 1a to 1g, respectively). (B) Average MS spectrum obtained for (A) where it can be observed the m/z values of the O-acetylated and Neu5Ac/Neu5Gc variations species of H7N6F1S4. (C) EIE obtained for the CE-MS separation of the tetra-antennary tri-sialic glycan H7N6F1S3 (peak 6) and its acetylated species containing from 1 to 4 O-acetyl groups (peaks 6a to 6d, respectively). (D) Average MS spectrum obtained for (C), where it can be observed the m/z values of the O-acetylated and Neu5Ac/Neu5Gc variant species of H7N6F1S3. (E) EIE obtained for the CE-MS separation the tetra-antennary tetra-sialic glycan containing one LacNAc unit H8N7F1S4 (peak 2) and its acetylated species containing from 1 to 6 O-acetyl groups (peaks 2a to 2e, respectively). (F) Average MS spectrum obtained for (E), where it can be observed the m/z values of the O-acetylated and species of H8N7F1S4. The glycans were detected under the same CE-MS conditions as in Figure 3.1. ...............88 xvi  Figure 4.1 Sialic acid structure and common natural modifications that occur in the “R” groups. At position R2, Sia stands for sialic acid. 297 ................................................................98 Figure 4.2 (A) CE-LIF separation of APTS-labeled N-glycans of control fish serum. Peaks (1a),   (2a), (3a), (4a), and (5a) are APTS-labeled glycan peaks. (B) Base peak electropherogram of underivatized O-acetylated N-glycans of control fish serum acquired using the LCQ* Duo ion trap MS. BGE was 40 mM ε-Aminocaproic acid with 20% methanol. Peaks (1b), (2b), (3b), (4b), and (5b) correspond to the peaks (1a), (2a), (3a), (4a) and (5a), respectively in (A). The migration time (above) and the observed m/z (below) are indicated for each peak……………………………….....103  Figure 4.3 Extracted ion electropherogram showing baseline separation for isomeric species at (A) m/z 1153.04 and (B) m/z 1132.02. CE-MS spectra was acquired using the LCQ* Duo ion trap MS. In all cases, [M-2H]2- was observed. The inserts show the CE-MS/MS spectra for the isomeric species acquired using the API 3000 triple quadrupole MS: Fragment ions from precursor ions at (a) m/z 1153.07 peak (2b), (b) m/z 1153.04 peak (5b), (c) m/z 1131.8 peak (1b) and (d) m/z 1132. 02 peak (4b) were observed…………………………………………………………………………….106 Figure 4.4 Base peak electropherograms showing separation of isomeric species for bi- and tri- antennary oligosaccharides present in samples taken from (A) week 1 and 4, (B) week 2 and (C) week 3 of the handling-stress experiment. Data was acquired using the API 3000 triple quadrupole MS…………………………………………………………113 Figure 5.1 Strategy for sample preparation and analysis of human serum glycans. Released N-glycans were analyzed using a flow-through microvial CE-MS interface connected to a Triple Quadrupole and a High-Resolution TOF MS. SDS in the second box refers to sodium dodecyl sulfate. .............................................................................................122 Figure 5.2 Base peak electropherogram for the CE-MS separation of N-glycans of a control sample of human serum using the maXisTM High-Resolution TOF MS. Only the putative structures of the most abundant glycans are shown. CE-MS analysis was performed. Monosaccharide legend for the putative structures as follows: blue squares: N-acetylglucosamine, green circles: mannose, yellow circles: galactose, purple diamonds: N-acetyl neuraminic acid and red triangles: fucose. The position of the N-acetylneuraminic acid correspond to the proposed structure and it has not been defined in the present study………………………………………………………...124 Figure 5.3 (A) Fragmentation spectra of glycan at m/z 1110.9 with composition H5N4S2 and (B) at m/z 959.0 with composition H6N5S3. The fragmentation spectrum of glycan shown in (A) corresponds to a bi-antennary sialylated glycan, while the spectrum shown in (B) corresponds to a tri-antennary sialylated glycan……………………..128 Figure 5.4 EIEs of acidic N-linked glycans released from human serum showing baseline separation for isomeric species at (A) m/z 1111.1, (B) m/z 959.4, (C) m/z 1008.3 and (D) m/z 1057.4. Glycan composition is indicated for each species below their corresponding m/z value. Glycans were observed as [M-2H]2- and [M-3H]3- ions species. Putative monosaccharide composition and the corresponding m/z are indicated for each glycan...........................................................................................131 Figure 5.5 (A) Comparison of the relative abundances of all the N-glycans found in common among the human sera samples. Pie charts illustrating the N-glycans types and relative abundances present in (B) normal human serum, (C) prostate carcinoma xvii  (stage III) serum and (D) asthma serum obtained using the optimized CE-ESI-MS conditions……………………………………………………………………….......134  xviii  List of Abbreviations amu atomic mass unit APCI atmospheric pressure chemical ionization  APPI atmospheric pressure photoionization  APTS 8-aminopyrene-1,3,6-trisulfonate BGE background electrolyte BHK baby hamster kidney BPE base peak electropherogram BRP biological reference preparation CAE capillary affinity electrophoresis CE capillary electrophoresis CGE  capillary gel electrophoresis CHO chinese hamster ovary CID collision-induced dissociation CUR curtain gas CZE capillary zone electrophoresis Da dalton DC direct current DS dextran sulfate   dielectric constant     electric field EIE extracted ion electropherogram EOF electroosmotic flow EPO erythropoietin ESI electrospray ionization  ESI/MS electrospray ionization-mass spectrometry      drag force     electrostatic force Fuc fucose Fuc-Sia index fucosylation- sialylation index FWHM Full width at half maximum G20 glucose ladder standard GlcNAc N-acetylglucosamine Gal galactose Hex hexose Hex-NAc hexose-N-Acetylglucosamine HPC hydroxypropyl cellulose HPLC high performance liquid chromatography ICP inductively coupled plasma ID inner diameter  IgG immunoglobulin G kV kilovolts KDa kilo daltons xix  LacNAc poly-N-Acetyllactosamine LC liquid chromatography  LIF laser induced fluorescence MALDI matrix-assisted laser desorption ionization Man mannose MEKC micellar electrokinetic chromatography  MS mass spectrometry  MS/MS tandem mass spectrometry MW molecular weight m/z mass-to-charge ratio η viscosity N-glycosylation nitrogen-linked glycosylation Neu5Ac N-acetylneuraminic acid NeuGc N-glycolylneuraminic acid O-glycosylation oxygen-linked glycosylation OD outer diameter PB polybrene PEI polyethyleneimine PGC porous graphitized carbon  PNGase F peptide-N-glycosidase F psi pounds per square inch PTM post-translational modification   net charge Q-TOF quadrupole time of flight mass spectrometer RF radio frequency rHuEPO recombinant human erythropoietin RNase A ribonuclease A RNase B ribonuclease B RSD relative standard deviation SPE solid phase extraction THF tetrahydrofurane TOF time of flight   electrical potential UV ultraviolet       apparent electrophoretic mobility      electroosmotic mobility      electrophoretic mobility      electrophoretic velocity   alternate voltage ζ zeta potential   xx  Acknowledgements I would like to express my eternal gratitude to my wonderful supervisor, Dr. David Chen, for giving me the opportunity to fulfill my lifelong dream to hold a Ph.D. degree in chemistry. David, you were a mentor, a friend and certainly, this thesis could not have been possible without your patient guidance, continuous encouragement and vast knowledge through every step of my Ph.D. studies. I am deeply thankful for the many opportunities you gave me, which helped me to grow both professionally and individually. Certainly, your optimistic view of life has positively influenced me.  Thank you to my fellow group mates: Jane, Xuefei, Koen, Joe, Chang, Sharon, Laiel, Alexis, Akram, Matthew, Cheng, Sherry, Jessica, Lingyu, Cai, Monsalud and Nikita for sharing their knowledge and being supportive during important academic steps. I would like to acknowledge Caitlyn and Xian Zhen for taking time to read part of this thesis and to provide constructive criticism. I owe particular thanks to Dr. Jianjun Li, from the NRC, and Dr. Rawi Ramautar, Dr. Peter Lindenburg and Dr. Thomas Hankemeier from Leiden University, for their extensive academic support during my research experience as a visiting scholar and their thoughtful comments during the manuscripts revisions. I am so grateful to have shared with Javier many years of my Ph.D. in Vancouver and to have formed a lovely family with him. Thank you very much for being a colleague, a friend, a husband, and for all your support and love during the tough times. It is you who made my everyday special in so many ways. Any place where life takes us will be more enjoyable with you by my side. Thanks to my little ‘happy boy’, Sebastian, for brighten my life and for giving me a reason to smile every day. I would particularly mention my gratitude to a few, but special xxi  friends, in Vancouver: Emmanuel, Fiona, Jennifer, Montse, Ainur, Jannu, and Gloria, for the good times during this long journey. Each one of you is special to me in so many ways. Last but not least, I would like to specially thank my parents and brother for their endless love and support. Thank you for believing in me and for encouraging me to pursue my dreams during all my life. Although the distance can only make me feel your strong and unconditional love, thank you for keeping the distance short and for visiting me in Vancouver almost every year. It is you who made me the person I am today, and I am blessed for having such a wonderful parents. Finally, as it is right, thanks to God in the highest for the exciting opportunity to pursue Ph.D. studies, and for what life brings ahead.            xxii  Dedication         To my parents…1  Chapter 1: Introduction to capillary electrophoresis-mass spectrometry for the characterization of glycosylated proteins and their glycan derivatives  1.1 CAPILLARY ELECTROPHORESIS-SEPARATION OF CHARGED SPECIES IN SOLUTION    Capillary electrophoresis (CE) is a technique that offers electric field mediated differential migration separation selectivity. The principle was investigated in 1967 by Hjertén who performed for the first time zone electrophoresis in linear tubes with a diameter of 1-3 mm.1 Compared with two dimensional gel electrophoresis, the use of narrow bore capillaries for CE offers the advantage to perform online detection with improved sensitivity and specificity, as well as increased separation speed because of the higher voltage used. In 1981, Jorgenson and Lukacs demonstrated the use of open tubular capillaries with an internal diameter of 75 μm.2 This launched a wide spread use of this technique, especially after the instrument was made commercially available. Currently, applications of CE can be found in both academia and industrial settings. In the fields of proteomics, metabolomics and pharmaceutical research, CE has become extremely popular due to its high separation efficiency, short analysis time, and minimum sample and buffer consumption.  The most common detection method for CE is the use of ultraviolet absorption (UV) detectors but its function is often limited by the lack of standards necessary for identification of unknown compounds in complicated matrices. Moreover, analytes with limited capacity to produce a recognizable spectroscopic signal for detection require pre-column derivatization. In 1987, Olivares et al.3 demonstrated the use of mass spectrometric (MS) detection in combination 2  with capillary electrophoresis. Mass spectrometer is a highly selective and a widely used detector for analytical purposes. It provides high resolution separation based on mass-to-charge ratio in the gas phase and the ability to identify analytes based on fragmentation spectra. The coupling of CE with MS not only has expanded the applicability of CE but becoming a turning point in the bioanalytical field. However, an aspect that has delayed the growing interest in capillary electrophoresis coupled to mass spectrometry (CE-MS) is the lack of robustness for the interface of this hyphenated technique. However, significant progress is being made in this area.4   Currently, the predominant way to couple CE with MS is through electrospray ionization (ESI) using the commercially available sheath-liquid interfaces, and to a lesser extent CE-MS interfaces constructed in research laboratories. In spite of the popularity of CE-MS in a variety of application areas, its full potential is still underexploited for the analysis of protein glycosylation.  Protein glycosylation is an important post translational modification that modulates the physical, chemical and biological properties of proteins.5 Glycans are implicated in a wide range of intracellular, cell-cell and cell-matrix signaling that has attracted major research interest.6 Changes in glycosylation have been correlated to various pathological states that make glycans an important target for analysis. Given the analytical complexity associated with glycan analysis, orthogonal techniques are often required to unravel their inherent structural heterogeneity. The interface of CE with MS offers two independent dimensions for separation and promises to provide sufficient analytical power for glycan identification and structural elucidation. However, CE-MS is still not used as commonly as liquid chromatography-mass spectrometry (LC-MS). More research is needed to improve and develop new methods to demonstrate the usefulness of CE-MS in the glycomics field.  3  This thesis presents work on the development of approaches toward the characterization of protein glycosylation with capillary electrophoresis coupled to mass spectrometric detection.  Sections 2-5 of this chapter are devoted to introducing the theory of electrophoresis and the technology for interfacing CE with electrospray ionization-mass spectrometry (ESI-MS) using a flow-through microvial, developed earlier in our research group. Chapters 2 to 5 demonstrate the utility of the interface for glycosylation analysis of glycoproteins from different biological sources. The theories presented are not meant to be exhaustive, but provide enough information for readers whose expertise lies elsewhere.  1.2 CAPILLARY ELECTROPHORESIS- SEPARATION BASED ON DIFFERENTIAL MIGRATION IN ELECTRIC FIELD 1.2.1 Separative principles Capillary electrophoresis is performed in narrow-bore fused silica capillaries with a typical inner diameter of 25-75 µm. The separation is based on the differential migration of analytes in the liquid phase within a uniform electric field applied across the length of the capillary. A typical capillary electrophoresis system using online optical detection is shown in Figure 1.1. A standard CE setup consists of a high voltage power supply, a polyimide-coated fused-silica capillary with inner diameter (ID) smaller than 100 μm, two chemically inert electrodes (e.g. platinum) placed in the two buffers and/or sample reservoirs.7 Typically, an online optical detection unit is placed near the outlet by removing a portion of the polyimide coating to serve as an optical detection window.   4   Figure 1.1 Schematic diagram of a capillary electrophoresis instrument with online optical detection.  Capillary zone electrophoresis (CZE), the most popular mode of CE, separates ionic analytes based on their charge-to-size ratio.7 Non ionic analytes can also be separated by CE using secondary equilibrium additives. To perform analyte separation, the capillary is filled with a background electrolyte (BGE), followed by pressure or electrokinetic injection of a small sample plug. After that, both ends of the capillary and the electrodes are immersed into the BGE vials, and voltage, usually up to a maximum of 30 kV, is applied across the capillary. Upon voltage application, ions in free solution, in the presence of the electric field experience an electrostatic force     (1-1) that is proportional to the net charge carried by the ion,  , and the magnitude of the electric field,   . In response to this force ions accelerate; however they also experience a drag force,    (1-2), from the solution, which is proportional to the radius of the solvated ion,  , the velocity of the ion,   , and the viscosity of the medium,  . The drag force increases quickly until it balances the electrostatic force and the ions migrate at a constant electrophoretic velocity,     .  5                                           (1-1)                                    (1-2)                                 (1-3)                                        (1-4)                                   (1-5)  Equation (1-4) indicates that in free solution CZE, the electrophoretic mobility of ionic species is proportional to their charge-to-size ratio. Therefore, the electrophoretic mobility of an analyte,     , depends on its charge-to-size ratio and the viscosity of the buffer solution, and it is independent of the applied electric field, as indicated in Eq. (1-5).  1.2.2 Electroosmotic flow: a mode of mass transport  In addition to the electrophoretic mobility of ionic species, electroosmotic flow (EOF) could simultaneously occurs upon the application of the electric field, and it can assist in transporting the analytes toward the capillary outlet.7 EOF refers to the bulk flow of solution that is generated when the applied electric field interacts with the excess of positive ions in the diffuse layer of an electrical double layer. Figure 1.2 shows the structure of the electrical double layer. When the capillary is rinsed with a BGE that has a pH greater than 3, the silanol groups deprotonate and the capillary wall surface is negatively charged. These negatively charged silanol groups attract counter ions from the solution to the wall, creating an electric double-layer at the solid-liquid interface. Because these counter ions experience a strong electrostatic force, they are bound to the surface and form a stagnant layer, also called the Stern layer.   6  The amount of counter ions in the electrical double layer are not enough to neutralize all the negative charges on the capillary surface, and a second layer is formed next to the stagnant layer. This second layer, known as the diffuse layer, contains both positive and negative ions, but has an excess of positive charges. The electrical potential decreases linearly within the Stern layer and exponentially within the diffuse layer until it approaches zero in the bulk solution. The potential at the interface of the Stern layer and the diffuse layer, is regarded as the zeta potential (ζ), and it plays a significant role in determining the magnitude of the EOF. Electroosmotic flow results when a voltage is applied across the length of the capillary because the excess of positive charges in the diffuse layer experience an electrostatic force and move toward the cathode. Because the ions in the diffuse layer are hydrated, they drag the bulk solution with them at the same velocity. The magnitude of the electroosmotic mobility (1-6) depends on  , the dielectric constant of the fluid,     the zeta potential of the capillary wall, and    the viscosity of the fluid.                                                                                             (1-6)  Figure 1.2 (A) Structure of the electrical double layer and resulting EOF due to the deprotonation of the bare-fused silica capillary wall. (B) Potential profile with increasing distance from the capillary wall. 7  The buffer pH has a strong influence on the zeta potential because the charge density on the capillary wall is detemined by the dissociation of the silanol groups. At pH above 10, the EOF mobility is the strongest because the silanol groups are fully dissociated, while at pH below 3, the deprotonation of the silanol groups is minimal. The ionic strength of the buffer also influences the zeta potential. At higher ionic strength, the magnitude of the zeta potential is reduced because the larger number of counter ions in the electrical double layer results in larger potential drop within the electrical double layer. For that reason, the magnitude of the EOF decreases with increased ionic strength.   The apparent electrophoretic mobility of an analyte is the vector sum of the electroosmotic mobility and its intrinsic electrophoretic mobilty, as shown in Eq. (1-7).                                                                                      (1-7) Because, the magnitude of the electroosmotic mobility is influenced by buffer pH, ionic strength and the nature of buffer counter ions, exploting these parameters, would result in differential migration of the analytes.  1.2.3 Capillary coatings  The use of conventional bare fused-silica capillaries in CE may compromise separation efficiencies due to analyte-capillary wall interactions, leading to significant changes in the EOF and the migration times of the analytes. Adsorption of analytes, such as proteins, can be partly avoided using buffers of extreme pH and high ionic strength. However, this approach may be limited due to analyte stability under such conditions. Therefore, coating the capillary wall is the more effective approach for minimizing analyte adsorption. Moreover, this procedure can modify the magnitude and direction of the EOF and improve CE performance.8-12  However, all of the coating strategies that have been proposed for CE are not compatible with MS detection. 8  Effective coatings for CE should be stable over time, easy to introduce onto the capillary wall and compatible with different BGEs and sample matrix components. Additionally, if the intended use of the coating is for CE-MS analysis, bleeding of the coating into the MS source should be minimized to avoid supression of analyte signals, contaminantion of the ion source and the presence of strong background noise.13 Deactivation of the silanol groups can be achieved by dynamic or covalent coatings. Dynamic coatings are present as additives in the BGE and compete with the analytes for the binding sites on the silanol so analyte adsorption is minimized.8-11 The common practice for coating the surface of the capillaries for CE-MS analysis is by static coatings that permanently modify the inner walls of the capillary by chemical reactions or by physical adsorption of the coating species.12 Covalent coatings reported to date, are either positively-charged coatings or neutral coatings that can be easily implemented in CE-MS analysis of biomolecules.14, 15 The major advantages of static-adsorbed coatings over static-covalent coatings are that (i) the coating procedure is achieved by simple rinsing steps, (ii) the coating regeneration is easily performed by using strong acidic or basic solutions, (iii) the coating is strongly adsorbed to the capillary wall, thus preventing bleeding on the mass spectrometer, and (iv) most of the coating materials are commercially available and only minute amounts are required. Excellent reviews discussing these coatings have been published.12, 13, 16 A variety of static-adsorbed positively-charged and neutral coatings have been employed in this thesis. Polyethyleneimine (PEI), Polybrene (PB), hydroxypropylcellulose (HPC), and commercially available polyacrylamide-based neutral capillaries, were employed for the analysis of glycosylated proteins and their glycan derivatives.   Positively-charged coatings offer a large number of cationogenic amine groups that interact with the negatively-charged silanol groups of the capillary through electrostatic 9  interactions. PEI17, 18 and PB19-21 are the most commonly used adsorbed coatings in CE-MS. The procedure for PEI coating is outlined in US Patent 6923895 B2 and it has been slightly modified in this thesis to shorten the coating time. A triple-layer coating of successive layers of PB, dextran sulfate (DS), and PB reported by Katayama et al.,19 has been used in this thesis. Both, PEI and PB are commercially available and give rise to a high-pH independent EOF that is reversed toward the anode. The cationic layer formed on the capillary wall reverses the EOF and the polarity required for the separation also needs to be reversed, to ensure that the cationic analytes migrate toward the detector.  The majority of neutral coatings are covalent coatings in which the EOF is strongly minimized or totally suppressed. HPC-coated capillaries were prepared following the method of Shen and Smith,22 in which the HPC is introduced into the capillary by performing rinsing steps and becomes permanently bound to the capillary upon heating.23, 24 This coating confers very high stability over a wide pH range. Hydrogen bonding interactions maintain the neutral coatings fixed to the silica capillary wall. Fused silica capillaries treated with a polyacrylamide-based neutral, hydrophilic wall coating, which are available from a number of manufacturers, have been employed for CE-MS analysis of N-glycans. While relatively fast EOF toward the CE outlet is preferred to enhance stability of the ionization process, CE-MS can also be performed at low or negligible EOF.24, 25 26  This approach has not being extensively reported in the literature and will be further explored in chapters 2, 3, 4 and 5 of this thesis, to enhance differences in the electrophoretic mobilities of structurally similar analytes.    10  1.3 ELECTROSPRAY IONIZATION Electrospray ionization is the most widely used ionization technique for online coupling of separation techniques such as capillary electrophoresis and liquid chromatography with mass spectrometric detection. The formation of charged species from the liquid to the gas phase is achieved by applying a high electric potential over the liquid phase at atmospheric pressure.27 The production of intact, multiple charged ions has revolutionized the analysis of large biomolecules28-30 and have found  a wide range of industrial applications.31  In 1968, Dole et al.32 were the first to attempt electrospray to produce gas phase ions of macromolecules from the liquid phase for mass spectrometric analysis. However, it was not until 1988 that ESI proved to be a successful ionization technique for MS, when John B. Fenn and Masamichi Yamashita reported the ESI mass spectra of multiply-charged proteins by coupling ESI to a quadrupole mass analyzer.33  This opened the possibility of using conventional mass spectrometers, capable of detecting ions in the low mass-to-charge ratio, for analysis of large biomolecules. The development of ESI as a method for identification and structure analyses of biological macromolecules won Fenn a Nobel Prize in 2002. Currently, ESI is considered the key enabling technology for biomolecule characterization due to its ability to form multiply-charged ions. Successful applications of ESI can be found in the fields of proteomics,30 metabolomics,28 and lipidomics.29      Figure 1.3 shows a schematic representation of the electrospray process. During typical ESI a dilute solution of analyte, dissolved in a polar volatile solvent, is pumped through a stainless steel or silica capillary (50 to 150 μm) at a flow-rate between 1μL/min and 20μL/min. A voltage of 2-5 kV, known as the ESI voltage, is applied on the ESI needle located a few centimeters away from the MS inlet.34 Because of the potential difference between the ESI 11  needle and the MS inlet, typically grounded, a strong electric field of about 106 V/m is generated at the capillary tip. This leads to the formation of charged droplets that evaporate and burst into smaller droplets as they move toward the entrance of the mass spectrometer.  The mechanism by which gas phase ions are produced from the charged droplets can be divided in three major steps: (1) production of charged droplets at the ESI needle; (2) solvent evaporation from the charged droplets; and (3) formation of gas-phase ions from high charged density droplets.35, 36 When the voltage is initiated, the high electric field penetrates partly past the surface of the liquid at the tip of the ESI needle, and acts on the ions in solution leading to charge separation between anions and cations. Under a positive electric field, positive ions migrate toward the meniscus of the liquid and accumulate at the tip of the ESI needle. The mutual repulsion among the positive ions at the liquid surface leads to the protrusion of the liquid toward the counter electrode, with the surface tension of the solvent holding the liquid still.37 When more charges gather on the surface, the Coulombic repulsion overcomes the surface tension and the liquid begins to expand downfield forming the ‘Taylor cone’. If the electric field at the capillary tip is high enough to exceed the surface tension, a thin surface layer at the cone tip moves toward the counter electrode like it is being pulled off from the cone, forming a liquid filament –liquid-jet– from which a fine spray of charged droplets of single polarity are emitted.38  Following dispersion of the charged droplets, they move in the air toward the counter electrode and shrink in size due to solvent evaporation, at constant charge. Shrinkage of the droplets occurs until Coulombic repulsion overcomes the surface tension, known as the Rayleigh limit. Beyond the Rayleigh limit, the droplets experience Coulombic fissions to produce smaller offspring droplets. Further fissions of the charged droplets eventually lead to the formation of gas phase ions, and two mechanisms have been proposed. The Charge Residue Model (CRM), 12  proposed by Dole,32 suggests that when the solution is dilute enough a series of Coulombic fissions produce droplets so small that each one bear one or more excess charges, but only a single analyte molecule. Iribarne and Thomson proposed the ion evaporation mechanism (IEM),39, 40 which considers that solvated ions are emitted directly from the charge droplet after the radii of the droplets decrease to a certain limit by solvent evaporation. The IEM suggests that the charge density on the droplet surface overcomes the Coulombic repulsion for droplets of radii smaller than 10 nm.40 Once the ions are transferred to the gas-phase either by the CRM or the IEM, they are ready to be analyzed for mass-to-charge ratio within the mass spectrometer. Normally a curtain gas, which flows counter to the ion flow, is used to completely dry droplets, and to prevent solvent and neutrals from entering into the vacuum system of the mass analyzer.     Figure 1.3 Schematic representation of the electrospray ionization process. The formation of charged droplets via'Taylor' cone and the evolution of gas phase charged analytes traveling toward the MS inlet.  13  1.4 CE-MS INTERFACING Liquid chromatography coupled with mass spectrometry through electrospray ionization has demonstrated the ability of this hyphenated technique to attain low detection limits and fast response times.41 Therefore, the coupling of capillary electrophoresis with mass spectrometry through an electrospray ionization interface is also expected to provide an extremely powerful two dimensional separation and detection tool for analytical purposes. In 1987, Smith and co-workers first attempted online electrospray ionization of CE effluents for mass spectrometric analysis.42, 43 Despite the limitations were not overcome in the following decade, ESI has now become the main ionization technique for combining CE with MS detection. Since then, other techniques with several types of ionization sources have been developed. Available ionization techniques include matrix assisted laser-desorption ionization (MALDI),44 inductively coupled plasma (ICP) ionization,45 atmospheric pressure chemical (APCI) ionization 46 and atmospheric pressure photoionization (APPI).47 The applicability of each ionization technique is primarily limited by the type of the analytes involved. MALDI is widely applied for the analysis of intact proteins and it is usually performed off-line given the difficulties of spotting the CE effluent and the UV absorbing matrix eluent onto a probe for co-crystallization. ICP is only useful for the analysis of metalloproteins; APPI excels in high efficiency ionization of low polarity compounds; APCI is applied almost exclusively for the ionization of low-mass pharmaceutical compounds and it is not suitable for the analysis of thermally labile compounds.  1.4.1 Challenges for coupling CE with MS detection In the last decade, improvements in terms of reproducibility and sensitivity have increased the popularity of CE-MS for analysis of biomolecules.48, 49 Although ESI have evolved as the preferred approach for CE-MS interfacing, its operation still encompasses a series of 14  particular challenges that need to be addressed. First of all, both the CE and ESI processes require stable electrical contact of the solution with an electrode at the capillary outlet without disrupting the EOF from the CE effluent. The use of a modified outlet CE electrode to act as an electrospray emitter that generates an intense electric field has to be considered. Secondly, the mismatch between the electrospray currents, usually confined below one microampere, and the CE currents, usually between tens to a hundred microamperes, can affect the stable electrical contact at the shared electrode. In addition, improper grounding of the CE electrical circuit can lead to arcing between the electrospray emitter and the counter electrode that affects the analyte ionization.  Finally, the flow-rate compatibility between CE, which operates between 1 and 100nL/min, and ESI-MS, which operates between 1-200 µL/min, also possess a difficulty in attaining a stable electrospray and must be considered. In addition, electrospray ionization place particular restrictions on the composition of the background electrolytes that can be used for CE separation. The non-volatile salts commonly used as buffers in CE could affect mass detection. The use of volatile acidic or basic electrolytes increases the compatibility of the CE effluent and improves the ionization efficiency of the analytes.  Over the past two decades, a variety of interfaces have been developed to maximize the potential of the coupling of CE and MS through electrospray ionization. Excellent reviews of CE-ESI-MS interfaces can be found elsewhere.50-53 Currently, available interfaces for CE-MS can be divided into two categories: sheath-liquid and sheathless interfaces. The flow-through microvial interface used throughout this thesis does not have as much dilution as commonly used sheath liquid interface, and is more versatile compared to the common sheathless interfaces.  15  1.4.2 Sheath-liquid CE-ESI-MS interfaces Currently, the sheath-flow or co-axial arrangement first developed by Smith et al.,42 is the most common type of CE-MS interface used. It requires an additional flow of liquid, known as sheath-liquid or make-up liquid, for maintaining the electrical contact necessary for CE separation and electrospray ionization. This added flow modify the composition of the BGE in order to attain good ionization yields and increases the liquid flow to levels comparable of those delivered by liquid chromatography.  The most common interface of this type is the coaxial sheath-flow arrangement where the CE capillary is surrounded by another larger capillary, which provides a continuous flow of sheath liquid, as shown in Figure 1.4A. The coaxial delivery of sheath-liquid aimed to solve the discrepancy between the amount of flow delivered by the CE effluent and the required flow for the ESI process. The currently available commercial sheath-liquid interface employs three concentric capillaries. The innermost capillary is used for the electrophoretic separation, while the second metal capillary continuously delivers the sheath-liquid. The third outer capillary delivers a flow of sheath gas, or nebulizing gas, to sustain the electrospray and to help with solvent evaporation. The major drawback of this interface is its low sensitivity because of the analyte dilution, and the chemical noise introduced by the large flow-rate required for the sheath-liquid to operate stably.54  Alternative set ups for sheath-liquid interfaces have been proposed to overcome the analyte dilution effect. Henion et al.,55 developed a liquid junction interface where a metallic tee union was used to provide a narrow gap (10-25 µm) between the ends of the CE capillary and the electrospray emitter, as illustrated in Figure 1.4B. Unfortunately, a highly precise alignment and spacing between the separation capillary and the spray needle is required to avoid band 16  broadening of the peaks caused by the significant post-column dead volume.56, 57 A  comparison study between the liquid junction and the coaxial interface, demonstrated that both approaches efficiently coupled CE to MS, but the sheath-liquid interface was more robust and reproducible.57  More recently, a pressurized liquid junction electrospray interface, that uses flow-rates in the nanoliter range, has also been described.23  In this arrangement, both the CE capillary and the electrospray emitter are located in a pressurized vessel of sheath-liquid, and the ESI potential is applied inside the liquid junction, as indicated in Figure 1.4C. Although the dilution factor is still an issue with this interface, the pressurized junction overcomes the peak broadening effect that often leads to reduced resolution.  Although sheath-liquid interfaces suffer from low sensitivity, they offer important advantages for CE-MS coupling. They allow the use of diverse types of CE electrolytes and buffer additives as the primary eluent exiting the interface is the sheath liquid, which usually contains organic solvents to facilitate electrospray ionization. Since the sheath-liquid mixes with the CE effluent outside the capillary tip, the products of electrolysis are kept away from the analyte path and do not interfere with the ionization process. Lastly, the robustness of sheath-liquid interfaces allows attain good reproducibility and are well-suited for commercialization.   17    Figure 1.4 Common arrangements for sheath-liquid interfaces. (A) Coaxial sheath-liquid interface with sheath gas, (B) Liquid junction interface, (C) Pressurized liquid junction interface.58  Reprinted from Analytica Chimica Acta 627, E. J. Maxwell and D. D. Y. Chen. Twenty years of interface development for capillary electrophoresis-electrospray ionization-mass spectrometry, 25, Copyright (2008), with permission from Elsevier.  1.4.3 Sheathless CE-ESI-MS interfaces Since mass spectrometry is a concentration dependent technique and analyte dilution is a major concern of sheath-liquid interfaces, sheathless configurations were proposed to improve detection sensitivity. The concept of sheathless interfaces, introduced by Olivares et al.,3 involves the use of the terminal portion of a CE capillary for electrical contact, to avoid the use of an additional sheath-liquid capillary. Typically, a fused silica capillary tip is pulled and etched with hydrofluoric acid to a thin tip, to avoid the need for sheath-liquid flow. Since the flow-rates produced by the electroosmotic flow in CE are usually in the nanoliter regime, the liquid droplet sizes are so small that gas phase analyte ions can be efficiently formed in the ion source. Therefore, the analyte ionization efficiency achieved with sheathless interfaces is much higher than that achieved with sheath-liquid interfaces. However, creating a stable electrical contact at 18  the capillary terminus is often a challenge for sheathless interfaces. Usually coating of the sprayer tip with a conductive layer such as gold,59 silver,60 copper,61 nickel,62 and graphite,63-65 and connecting this coating to a high voltage power supply is the preferred method of choice, as illustrated in Figure 1.5A. Unfortunately, rapid deterioration of the coating materials due to the high electrical fields applied, compromises the stability of the electrospray.  Other approaches for creating the electrical contact have been developed and include tapering the CE capillary outlet to form a microspray tip, inserting a thin wire into the capillary through an opening at the tip (Figure 1.5B)66 or, through a hole near the tip (Figure 1.5C);67 splitting the liquid flow from the capillary to make that a portion of the flow touches an outside electrode (Figure 1.5D);68 etching the outside surface of a fused silica capillary to obtain a porous tip that is inserted into a stainless steel ESI needle filled with a static conductive liquid (Figure 1.5E). 69-71 In cases where the separation capillary and the electrospray emitter are two independent pieces, electrical contact is created at the junction of the two portions. The separation capillary and the spray emitter are aligned in closed contact with no additional flow entering at the junction, and the surrounding electrolyte provides the electrical contact. This type of configuration has been constructed using silica emitters, metal sleeves (Figure 1.5F)72 or micro dialysis tubing to align the separation capillary and electrospray emitter (Figure 1.5G).73 As illustrated in Figure 1.5H, stainless steel emitters can also be used where the high potential can be applied directly onto it.74  Although optimal configurations for sheathless interfaces offer improved sensitivity, the limited robustness and cumbersome procedures for their manufacturing and operation have limited their wider acceptance. Moreover, this type of interface is restricted to CE systems 19  capable of generating a strong EOF toward the outlet, as the spray is only formed from the bulk flow in the separation column. In cases where the EOF is not large enough, the CE inlet vial has to be pressurized to sustain the bulk flow of the electrospray.           Figure 1.5 Different arrangements for creating electrical contact in sheathless interfaces. (A) Conductive coating applied to the emitter tip, (B) wire inserted at the tip, (C) wire inserted through hole, (D) split-flow interface with a metal sheath, (E) Etched capillary, (F) junction with metal sleeve, (G) microdialysis junction, and (H) junction with conductive emitter tip.58  Reprinted from Analytica Chimica Acta 627, E. J. Maxwell and D. D. Y. Chen. Twenty years of interface development for capillary electrophoresis-electrospray ionization-mass spectrometry, 25, Copyright (2008), with permission from Elsevier.  Recently, the porous tip sheathless CE-MS interface introduced by Moini69 has attracted much interest and has been made commercially available. In this design, the last portion of the fused-silica capillary is etched with hydrofluoric acid producing a porous capillary outlet that is conductive when in contact with an electrolyte. The ESI emitter, which is required to be filled with static conductive liquid, allows electrical contact and electrospray formation at the capillary tip. This interface has demonstrated improved detection limits in comparison to sheath liquid CE-MS interfaces for the analysis of model proteins.75 20  1.4.4 Flow-through microvial CE-ESI-MS interface The home-built interface constructed by our group is a type of sheath-liquid interface that operates in the nanospray regime. It uses a flow-through microvial strategy to effectively decouple the solution requirements of the capillary electrophoresis and electrospray processes for a more robust operation.76 The separation capillary is fully inserted into the electrospray emitter, which consists of a stainless steel beveled tip. A second capillary is inserted at the orthogonal entrance of a metal tee-union to deliver a chemical modifier solution, which improves the compatibility between the CE effluent and the electrospray ionization of the analyte. A schematic representation of the interface is shown in Figure 1.6. Typically, the modifier solution is composed of volatile organic solvents such as acetonitrile, methanol, isopropanol, and contains basic or acidic additives such as formic acid, acetic acid or ammonium acetate. The volume contained between the capillary terminus and the inner walls of the electrospray emitter creates a flow-through microvial that acts as the outlet vial and the terminal electrode for CE, providing the necessary electrical contact to allow the analytes to pass through the emitter. In order to provide a stable ESI spray, the flow-rate of the modifier solution is adjusted to a few µL/min, by using a syringe pump, to compensate the nanospray flow-rate provided by CE.     21      Reprinted from Electrophoresis 2010, 31, 1130-1137. E. J. Maxwell; X.  Zhong; H. Zhang; N. van Zeijl; D. D. Y. Chen. Decoupling CE and ESI for a more robust interface with MS. with permission from Elsevier.   1.5 MASS SPECTROMETRY- SEPARATION OF CHARGED SPECIES IN GAS PHASE Mass spectrometry is an extremely powerful analytical technique that separates and detects ionic species in the gas phase by measuring their mass-to-charge ratio (m/z). The high sensitivity, low detection limit, and high speed required for many applications have made this technology an irreplaceable tool in the analytical field.    Five basic components are, for the most part, common in all mass spectrometers and include: (i) a sample inlet to introduce the compounds to be analyzed, (ii) an ionization source to produce ions from the sample, (iii) a mass analyzer to separate the several ions, (iv) an ion detector to count the ions exiting the mass analyzer, and (v) a data processing system to generate the mass spectrum in a readable format. Despite some MS instruments combine the sample inlet and the ionization source, or the mass analyzer and the detector, all analyte molecules undergo   Figure 1.6 Schematic diagram of the flow-through microvial CE-ESI-MS interface used in this thesis.  22  the same processes regardless of the MS configuration. The most important component is the mass analyzer, and although it discriminates ions with respect to their m/z, not all mass analyzers operate in the same way. While some mass analyzers separate ions in space, others separate ions by time, and depending of the intended purpose, the choice of mass analyzers is fundamental. The following section discusses three different types of mass analyzers involved in applications described throughout this thesis. 1.5.1 Quadrupole mass analyzer77      The quadrupole analyzer separates ions according to their m/z based on the stability of their trajectories in oscillating electric fields. It is composed of four perfectly parallel cylindrical rods or, ideally, hyperbolic surfaces. The cross sections of the parallel electrodes and the electrical potential applied to them is shown in Figure 1.7. Ions travelling along the z axis are subjected to the applied potential,  , that contains both direct current (DC) and radio frequency (RF) components, and is expressed as in Eq. (1-8),                                                                     (1-8) Where   is the direct current (DC) potential applied pole to ground,     is the zero-to-peak amplitude of the radio frequency (RF) voltage applied pole to ground, and  is the angular frequency of the RF voltage. Since  is a function of   , the potential among the four rods can be expressed as in Eq. (1-9), in terms of   and   Cartesian co-ordinates,                                                                               (1-19) The origin of the Cartesian co-ordinates is located at the centre of the four rods, and    is the shortest distance from the center to the rod surface. 23   Figure 1.7 Schematic diagram of a quadrupole with hyperbolic rods and correspondingly applied potential. The space between the rods comprises the trajectory of the ion before it hits the detector.   Reproduced from Ref.78 with permission from The Royal Society of Chemistry.  The potential of the centre of the four rods is usually maintained at zero, regardless of the potential of the rods. When the ions are accelerated along the   axis, they maintain their velocity along this axis. However, the movement of ions in the x and y axis is subjected to forces induced by oscillating electric fields and is given by Eq. (1-10), where   is the mass of the ion,   is the charge of the ion,   is the fundamental electron charge, and   and   denote the position of an ion from the centre of the rods.                                ;                                  (1-10)  Obtaining the ion trajectories during time is analytically challenging; however, as long as the ion motion amplitude is smaller than     the ions will be able to pass the quadrupole without hitting the rods. A stability diagram is often used to determine whether an ion with a specific m/z will have a stable trajectory in a quadrupole. Figure 1.8 shows the first stability region represented in a  ,   diagram. The parameters   and   are given by Eq. (1-11), and represent a relationship between the position of an ion and time. 24                                                                       (1-11)  Figure 1.8 First stability area of a quadrupole with an operating line. Adapted from79.  For an ion of any mass with specific m/z, its trajectory inside the electrodes will be stable, if the values of   and   fall within the triangular stability region. Otherwise, the ion will discharge itself against the rods without reaching the detector. Typically, for commercially available quadrupoles,    is a constant and   is usually maintained constant so the stable trajectory of an ion will be determined by the values of   and    From Eq. (1-11), it can be deduced that   and   are directly proportional to m/z, so the successive detection of ions with different mass-to-charge ratios is achieved by maintaining the     ratio constant. Therefore, the quadrupole allows obtaining a mass spectrum by scanning the   and   parameters, i.e. the DC and RF voltage, respectively. Scanning along the operating line that passes through different stability areas, defines the resolution of the quadrupole; the higher the slope of the line, the better the resolution of the quadrupole. If the quadrupole has a value of    , the resolution becomes low and the quadrupole acts as a low molecular weight cut-off ion guide that allows ions with a 25  certain minimum mass to pass through and reach the detector. The minimum on stable masses is given by the value of  .    Quadrupole mass spectrometers can operate in tandem with other two quadrupoles to scan precursor ions or fragment ions in several ways. The scanning modes that can be performed on a quadrupole mass analyzer include the ‘fragment ion scan’, the ‘precursor scan’ and the ‘neutral loss’. The first scanning mode is useful for chemical structure elucidation and for identification of analytes. It operates by selecting an ion with a specific m/z, the so-called precursor ion. The precursor ion is then accelerated into the second quadrupole, operating with a DC voltage equal to zero, where it collides with an inert gas to produce fragment ions that are analyzed by the third quadrupole.  Quadrupole mass spectrometers are the instruments most often used for the selective detection of markers ions from glycoconjugates, given its ability to selectively detect glycan-specific ions at high sensitivity during their electrophoretic separation.80 The application of a triple quadrupole mass analyzer will be described in chapters 2, 3, 4 and 5 for the analysis of N-glycosylation of model glycoproteins, recombinant drugs, and animal and human serum.   1.5.2 3D ion trap mass analyzer81 The 3D ion trap operates based on the same principle as the quadrupole mass analyzer. It uses the radio frequency quadrupole field to separate ions according to their m/z. However, instead of letting the ions to pass through the quadrupole analyzer, the 3D analyzer stores the ions within the quadrupole field. The 3D ion trap is composed of a ring electrode with hyperbolic surface, as shown in Figure 1.9. By applying a potential that includes DC and RF voltage to the ring electrode while maintaining the cap electrodes grounded, a ‘3D quadrupole’ field is produced, in which ions of all masses are trapped. 26    Figure 1.9 (A) Ion trap cut in half along the axis of cylindrical symmetry. (B) Schematic diagram of the cross section of and ion trap showing the asymptotes.  Reprinted from Journal of Mass Spectrometry, 32, R. E. March, An Introduction to Quadrupole Ion Trap Mass Spectrometry, 351, Copyright (1997), with permission from John Wiley & Sons.  Similar to the expression of the potential applied in a quadrupole mass analyzer, the potential applied to the 3D ion trap can be expressed as in Eq. (1-12), in terms of cylindrical coordinates      The origin of the cylindrical co-ordinates is located at the intersection of the two asymptotes, and  the smallest distance between the origin and the ring electrode surface is given by   , while the smallest distance between the origin and the cap electrode surface is given by    81                                                                                   (1-12)   In order to find areas where ions of different masses have a stable trajectory, a stability diagram is used. Figure 1.10 shows the first stability region of a 3D ion trap. If the parameters    and    of the stability diagram are within the limits of the stability region, the ions will have a stable trajectory along both   and  , and will be efficiently stored in the ion trap. Stability of the (A) (B)27  ions within the trap is achieved if the trajectory of the ions never exceeds the dimensions of the trap,     and    .                                                                                     (1-13)  Figure 1.10 First stability region of 3D ion trap.81  Reprinted from Journal of Mass Spectrometry, 32, R. E. March, An Introduction to Quadrupole Ion Trap Mass Spectrometry, 351, Copyright (1997), with permission from John Wiley & Sons.  In order to operate the 3D ion trap as a mass analyzer, ions successfully stored in the trap are ejected by applying a variable RF voltage, with constant frequency and variable amplitude, to the end caps, while the DC voltage on the ring electrode is set to zero (     . When the RF voltage is increased, all the ions will have a higher    value, and will eventually reach the point          (0, 0.908), which indicates that the ions have reached their stability limit. Beyond this point a minimal increment of the RF value will cause each ion to have an unstable trajectory, and it will be expelled from the trap. The mass limit value for this ejection to occur is determined by Eq.(1-14). It can be deduced from Eq. (1-14) that the upper mass-to-charge ratio for ion ejection is determined by the maximum     that can be applied to the circular electrode.  28                                           (1-14) Although ion trap mass spectrometers work well in the low mass range, it is possible to increase the mass analysis range by resonant ejection at smaller   values.82 Within the ion trap, the ions oscillate in the   axis at a secular frequency (   , proportional to    . If a supplementary AC voltage at frequency    is applied to the ellipsoid caps, the ions will resonate along the   direction, and their oscillations will increase until they become unstable and eventually will be ejected through the ellipsoid cap. Under these conditions the mass analysis limit of the trap analyzer can be extended by applying a    of reduced    , such that the ions are ejected at a lower      82  Despite the sensitivity of the 3D ion trap analyzer can be improved by increasing the ion ejection time, the mass accuracy and resolution can be significantly affected.83 Because the electric field induced by the applied DC and RF voltage is modified by the electric field generated by the charged ions stored in the trap, the stability diagram will be shifted and will affect the accuracy and resolution of the mass analyzer. Thus a balance between sensitivity and space charge effects needs to be maintained.  The utility of 3D ion trap mass spectrometers for the characterization of oligosaccharides and glycan mixtures have been extensively reported in the literature.84-88 Collision-induced dissociation (CID) of glycans yields abundant fragment ions that provide key information for structural characterization of glycoconjugates. Moreover, tandem mass spectrometry with orders higher than two exemplifies the potential of 3D ion trap instruments for glycan analysis.88 Applications of this type of mass analyzer will be described in chapters 2 and 4 for the analysis 29  of glycosylated proteins and their derivatives, and preliminary studies of protein N-glycosylation in fish serum.  1.5.3 Time-of-flight mass analyzer83 The time-of-flight (TOF) mass analyzer (TOF-MS) separates ions of different mass-to-charge ratio based on the time they take to pass a field-free flight tube (usually 1-3 meters in length) after a preceding acceleration. In the TOF-MS, ions of mass   and charge  , are accelerated by a potential  , to specific velocities. As they leave the source, fully accelerated ions of the same charge acquire the same kinetic energy and they drift in a field-free region. The time of flight it takes for a ion to travel in the drift tube is related to its mass-to-charge ratio as determined by Eq. (1-15), where   represents the drift time and   is the length of the flight tube.                                                                                                 (1-15) A significant drawback of the linear TOF-MS is its poor mass resolution that is affected by factors that create a distribution in flight times among ions with the same m/z ratio. Increasing the length of the flight tube or increasing the flight time by lowering the acceleration potential, are some of the approaches that have been adopted to overcome this problem. However, lowering the acceleration potential decreases the sensitivity. Moreover, the spatial distribution of ions within the ion source also represents a complication because not all the ions receive the same initial kinetic energy and often result in broad peaks. Therefore, modern TOF mass spectrometers have several features that largely improve the mass resolution and also help to improve the mass accuracy. The reflectron, which consists of ring electrodes, acts as a mirror that deflects the ions and sends them back through the flight tube to a detector. In this way, improvements in mass resolution are achieved because the reflectron corrects the kinetic energy 30  dispersion of ions with the same m/z. Ions with more kinetic energy, and hence more velocity, will penetrate the reflectron more deeply and will spend more time in the reflectron. The longer time the faster ions spend in the reflectron will compensate for the shorter time they spend in the drift region, so ions of the same m/z but with different kinetic energy, will reach the detector at the same time.   Figure 1.11 Schematic diagram of a time-of-flight mass spectrometer with reflectron. Adapted from89.   According to Eq. (1-15), the bigger the ion is and the fewer charges it has, the longer it takes to reach the detector. Therefore, the upper mass range of the TOF has no limits in principle. However, the increased flight path for the ions improves the mass resolution, affects the sensitivity, and introduces a mass range limitation. TOF mass analyzers have the highest scanning speed among all mass analyzers and they are the cheapest high resolution mass analyzers. In comparison with scanning instruments, they offer a number of advantages in terms of acquisition speed, scanning range, sensitivity and resolving power. Hyphenation of CE with ultra high resolution TOF mass analyzers is increasingly being used because of a uniquely powerful combination of separation efficiency and accurate mass determination. In recent years, 31  the commercialization of ultrahigh resolution TOF (UHR-TOF) mass analyzers, a high-end MS technology, have made plausible the detection of less abundant biological compounds, in comparison  with standard TOF instruments in which the data quality is often insufficient for analyte identification. A combination of accurate mass, isotopic distribution and migration time makes them ideal for characterization of complicated glycan species. In this thesis, studies on N-glycosylation of recombinant human therapeutic proteins and disease-associated human serum (chapter 3 and 5, respectively) were conducted using an ultra high resolution time-of-flight mass analyzer (maXis Ultra-High resolution, Bruker Daltonik). The increased resolving power of the maXis (20,000-40,000 FWHM) and mass accuracy (600-800 ppb) allowed for an exact establishment of the monosaccharide composition of the glycans, confirming assignments previously made with the triple quadrupole. Excellent reviews on the use of high resolution TOF mass analyzers for biomolecule characterization can be found elsewhere.49, 90  1.6 ANALYSIS OF PROTEIN N-GLYCOSYLATION USING CE-MS  1.6.1 General considerations of protein glycosylation Compared with most natural biopolymers such as oligonucleotides and polypeptides, oligosaccharides cannot be considered simple string-like structure molecules because of the intricacies of their fundamental building blocks. Monosaccharides, the basic units of oligosaccharides, can be linked together by glycosidic linkages to form up to thirty two different disaccharides if all possible combinations of linkage position (at C-2, 3, 4, 6 or 8) and anomericity (α or β configuration) are considered.91  Common mammalian oligosaccharides are built from a limited number of monosaccharides that include: glucose (Glc), galactose (Gal), mannose (Man), fucose (Fuc), N-acetylglucosamine (GlcNAc), N-acetylgalactosamine (GalNAc) 32  and sialic acid. Additionally, multiple residues may be linked in several different ways to a single monosaccharide, which can lead to complex branched structures. Therefore, oligosaccharides are considered the most structurally diverse molecules known to date.  The enzymatic addition of oligosaccharides, also known as glycans, to proteins and lipids is a highly common post-translational modification. It has been widely reported that more than 50% of human proteins are glycosylated and almost all extracellular proteins are glycosylated to some extent.92 Glycosylation is essential for the survival of most organisms as glycans dramatically enhance the complexity of glycoproteins, and participate in a wide variety of biological process such as: protein folding, protein clearance, receptor binding and activation, and cell adhesion.6, 93 Moreover, as glycans can reflect changes associated with physiological states, they are considered potential disease biomarkers.94-97    Several types of glycosylation are known, such as N-glycans, O-glycans, glycosphingolipids and proteoglycans, as illustrated in Figure 1.12. Since all studies throughout this thesis are focused on N-glycosylation and N-glycan profiling of glycoproteins from human and animal serum, and synthetic and recombinant therapeutic glycoproteins, other types of glycosylation will not be discussed further in this chapter. The N-linked glycans are covalently attached to the asparagine residue with a conserved consensus sequence of Asn-X-Ser/Thr, where X cannot be Pro, via a GlcNAc2Man3 core.93 They are subdivided in three groups, based on the nature and the location of the sugar residues added to the core.6 Glycans that contain just mannose residues attached to the common trimannosyl chitobiose core are called high-mannose type N-glycans. Glycans that contain only N-acetyllactosamine (Gal-GlcNAc) in their antennal region are classified as complex type N-glycans. The third class of glycans, known as the hybrid type N-glycans, can be regarded as a 33  mixture of the two previous types: they contain both mannose residues and N-acetyllactosamine attached to the core. Additional monosaccharide residues such as Fuc or sialic acids commonly elongate the antenna region. The additional fucosylation site is at the core GlcNAc. Compared to the high-mannose glycans, the complex glycans exhibit great structural diversity.98   Figure 1.12 Types of glycosylation. The common core structure of each type of glycosylation is indicated in the gray box. The terminal monosaccharide sequences in each type of glycans are representative structural variations. Adapted from99.   1.6.2 Analytical techniques for N-glycosylation analysis Structural elucidation of protein glycosylation represents a considerable analytical challenge due to inherent structural complexity of glycans, augmented by the lack of chromophore or fluorophore groups, and the potential occurrence of several closely related structures (including positional and/or linkage isomers). The different glycosylation site occupancy (macroheterogeneity) and the variability of glycan moieties at the glycosylation site (microheterogeneity), further add to the challenges of glycosylation analysis.   34  Currently, protein glycosylation analysis is performed using mass spectrometric, fluorescence or pulsed amperometric detection. Both fluorescence and pulse amperometric detection require the use of chromatographic or electrophoretic techniques prior to detection, while mass spectrometric detection can be performed in a stand-alone manner, but it may also be hyphenated with an additional separation technique. Excellent reviews regarding the separation and detection of glycans have been published. 100-103 1.6.2.1 Chromatographic separation High-performance liquid chromatography (HPLC) is a commonly applied separation technique for glycosylation analysis. Several stationary phases, including hydrophilic interaction chromatography (HILIC), reverse phase (RP), and porous graphitized carbon (PGC) have been employed.  HILIC is a very common HPLC-mode technique for oligosaccharides separation that offers the advantage of predictive retention time with increasing glycan size and hydrophilicity.104-106 HILIC-separations are mainly achieved based on the number of polar groups on the oligosaccharides and their ability to interact with the hydrophilic stationary phase by application of a polar mobile phase rich in organic solvent. Although glycan derivatization with fluorescent tags may increase the hydrophobicity of glycans, it has demonstrated to perform well for the simultaneous analysis of fluorescently labeled neutral and acidic glycans.107 A review addressing in more details the recent advances of HILIC systems and their application for structural glycomics has been recently published.108 A major disadvantage of HILIC-based analysis is that run-times of 60 minutes or more are needed.  RP-HPLC is a viable option for glycan analysis that generally requires derivatization to achieve retention on the stationary phase.109 The retention properties of labeled glycans are 35  determined by the hydrophobic properties of the labeling agent and the monosaccharide residues.110  Commonly used labeling agents include: pyridylamine (PA),110  2-aminobenzoic acid ethyl ester (2-ABEE),110 and  2-aminobenzoic acid (2-AB).111 Recently, the separation power of RP-HPLC for  2-AB labeled oligosaccharides have been demonstrated by Chen and Flynn for the analysis of structural isomers.111 Another commonly reported derivatization procedure employed to increase glycan hydrophobicity, and thereby allowing RP-HPLC is permethylation.112 This derivatization procedure has become extensively used in combination with tandem MS due to the highly informative fragmentation of permethylated glycans.     Porous graphitized carbon (PGC) columns have been shown to be an attractive alternative to HILIC separation.113 PGC has demonstrated superior capability for performing separation of oligosaccharide alditols,114 fluorescently labeled glycans, and permethylated glycans,115 as recently reviewed.116, 117 The mechanism of glycan retention on PGC columns is based on analyte adsorption given the high hydrophobicity of the carbon columns, which require the use of a high content of organic solvent in the mobile phase for glycan elution. Due to the high selectivity of PGC columns for the separation of anomers, glycans need to be reduced before attempting separation. The main advantage of PGC is that retention times in combination with mass analysis are often sufficient for structural determination. Reviews addressing in more details the mechanistic of PGC systems and their application for glycomics can be found elsewhere.118, 119  1.6.2.2 Electrophoretic separation Capillary electrophoresis analysis of protein glycosylation can be regarded as a more  attractive technology, compared with HPLC approaches, due to its ability to work at a microscale level,120 its quantitative capabilities, and the reduced time required for the analysis. The 36  applicability of CE-based separation methods has been demonstrated in different formats, such as typical zone electrophoresis,72, 121-123 micellar electrokinectic capillary electrophoresis (MEKC),124 capillary gel electrophoresis (CGE),125-127 and capillary affinity electrophoresis  (CAE).128  CE analysis of glycans often requires derivatization procedures to introduce a charged group necessary for migration under the applied electric field, and to increase the sensitivity by laser-induced fluorescence (LIF) detection. As an exception, acidic glycans carrying sialic acids are negatively charged and can be analyzed without labeling or derivatization since they can be easily deprotonated even in slightly acidic conditions (pKa of sialic acids 2.0-2.8).129  While capillary electrophoresis with LIF detection enables high sensitivity and selectivity, it does not provide additional glycan structural information but it conjunction with the currently available databases (section 1.6.3.5) and exoglycosidase based sequencing it does. Therefore, the use of MS detection for CE glycan analysis is often a preferred choice over other detection techniques. CE-MS technologies offer a more efficient and versatile approach that combines the high sensitivity of MS techniques with the high separation efficiency of CE.130  Hyphenation of CE to MS also benefits from the high mass accuracy and the wealth of information obtained from specific fragment ions generated by MS/MS, in particular by low-energy collision-induced dissociation (CID).131-134 In spite of the mismatch between CE separation and MS detection conditions and the lack of robust interfaces, the coupling of CE to MS detection through electrospray ionization has proved to be an extremely useful technique for glycosylation analysis, especially for the analysis of positional and/or linkage isomers. Numerous applications of CE-MS technologies for glycan analysis have been demonstrated in 37  recent reviews.102, 135-138 Information about off-line detection for CE-MS using matrix-assisted laser desorption ionization (MALDI), can be found elsewhere.136, 139 140 1.6.3 Glycosylation analysis  Because of the enormous structural variation and the low abundance of glycoproteins and released glycans, glycosylation analysis is commonly performed by maintaining the focus either on the structural characterization of the glycans or on the identification of intact protein glycoforms. In this thesis these two approaches were applied. Because analysis of released glycans is more widely applicable, this approach was emphasized.  1.6.3.1 Intact glycoprotein analysis by CE-MS In order to perform a complete structural characterization of protein glycosylation by CE-MS, several steps are required. Determination of glycoforms, glycosylation sites, degree of site occupancy at each glycosylation site, and amount and structure of glycans at each glycosylation site are often required. Although a major advantage of the analysis of intact glycoproteins is the minimal need for sample preparation, MS-based characterization of glycoproteins is typically more difficult than that of proteins.141 The substantially heterogeneous nature of the glycan component and its reduced ionization efficiency, render the analysis more challenging compared with that of proteins. Additionally, the binding of salts to the glycan moiety can seriously compromise the MS signal quality.142 The high molecular weight of glycoproteins also poses a limitation for glycoform analysis in terms of the mass resolution of certain mass spectrometers such as quadrupole mass analyzers.80 In order to solve this problem, glycoproteins can be cleavage, either chemically or enzymatically, to reduce their molecular weight. Cleavage occurs at each glycosylation site such that each site is represented by an individual glycopeptide. Glycopeptides inform about the 38  location of the glycosylation site and its degree of glycan occupation.143 However, the MS signals of glycopeptides, usually obtained by tryptic digestions, are often suppressed in the presence of other peptides. The peptide and glycopeptide mixture can then be separated by CE-MS. The use of MS, as a second dimension for separation, offers more details about the complex enzymatic degradation and significantly enhances the high resolution electrophoretic separation by allowing mass resolution of co-migrating components.  1.6.3.2 Glycan analysis by CE-MS The first step toward the analysis of glycans involves the use of enzymatic or chemical means to liberate the oligosaccharides from their corresponding  glycoprotein, followed by the subsequent purification and sometimes, derivatization, prior to  CE-MS analysis.144-146 Several approaches for N-glycan release have been reported, which are shown in Figure 1.13.   Figure 1.13 Strategies for N-glycan release from isolated glycoproteins and glycopeptides. Adapted from147.   Typically, oligosaccharides are released using Peptide-N-Glycosidase F (PNGase F), which is the most effective enzyme for removing almost all N-linked glycans.148 PNGase F Isolated glycoprotein/glycopeptidePeptide-N-Glycosidase F (PNGase F): cleaves off all types of N-glycans exc pt wh n α(1-3)-linked fucose is r s nt.Peptide-N-Glycosidase A (PNGase A): cleav off any type of N-glycan. End g ycosidase H (Endo H):  cleaves off only high-mannose N-glycans, leaving one N-acetylglucosamine residue.β-elimination: highly alkaline conditions followed by reduction are required to avoid degradation of released glycans.Hydrazynolysis: glycans are released using anhydrous hydrazyne at 60˚C for 5-6 hours. Enzymatic Release Chemical Release39  cleaves between the innermost GlcNAc and asparagine residue of the N-glycans and deaminates the asparagine residue to aspartic acid, resulting in glycans with a free reducing end.149 Other enzymes such as PNGase A and Endoglycosidase H (Endo H) have also been employed. 123 Several chemical protocols for the use of β-elimination for the release of O-glycans have been reported.142, 150, 151 Since the conversion of glycans to their respective alditols precludes further derivatization,152, 153 attempts have been made to develop a non-reductive β-elimination procedure,150, 151 which demonstrated limited applicability for glycan analysis.154, 155 Despite hydrazynolysis has been employed for releasing N-, as well as, O-glycans, it constitutes a cumbersome procedure with many precautions to consider because of the use of anhydrous hydrazine that is a highly toxic and explosive chemical.156 1.6.3.3 Reductive amination  Upon release, the N-glycans may be derivatized with chromophoric or fluorogenic compounds to facilitate their analysis. Several derivatization procedures have been reported, among which reductive amination,157-159 permethylation,115, 160 Michael addition,161 and hydrazyde labeling162  have been the most commonly applied. In this thesis, only reductive amination has been employed.    Reductive amination takes place between the aldehyde group at the reducing end of the glycan and the amino group of the labeling reagent to produce a Schiff-base that is stable at neutral pH, but decomposes in the presence of acid. Sodium cyanoborohydride reduces the Schiff-base forming an acid-stable secondary amine linkage between the glycan and the label (Figure 1.14). The reaction is usually performed in dimethylsulfoxide (DMSO),163 but alternative methodologies using tetrahydrofuran164 and methanol165 have been reported.   40            Figure 1.14 Reductive amination reaction for the labeling of N-glycans after their release from glycoproteins.  1.6.3.4 Labels Since the first application of 3-(4-carboxybenzoyl)-2-quinolinecarboxaldehyde (CBQCA) as a the derivatizing agent for the analysis of N-glycans,166 several labeling agents have been reported for reductive amination of glycans.167, 168 The following labels have been widely applied for glycan analysis: 2-aminopyridine (2-AP),169 2-aminobenzamide (2-AB), 2-aminobenzoic acid (2-AA),170, 171 8-aminonaphthalene-1,3,6-trisulfonate (ANTS),172, 173 and 8-aminopyrene-1,3,6-trisulfonate (APTS).121, 127, 174, 175    APTS is currently one of the most popular labeling agents because the derivatization reaction generally allows preserving fucose and sialic acid residues. Although the three negative charges of APTS make it very suitable for capillary electrophoresis,176 the faster migrating rates of APTS-labeled glycan peaks may overlap with excess reagent-derived peaks, which introduces some difficulties for analyzing less abundant species.177 Successful characterization of APTS-labeled glycans derived from various glycoproteins such as ribonuclease B,121, 127, 178, 179 fetuin,121, 175, 179 recombinant human erythropoietin,121 kallikrein,121 and chimeric recombinant monoclonal antibody174 have been reported.     41  1.6.3.5 Strategies for glycan structural elucidation  Because capillary electrophoresis can be performed in different formats, it offers different selectivity with great potential for glycan structural identification. Under conditions of counter- EOF, increasing charged-based migration is favored with larger species migrating faster than smaller ones of the same net charge.180 Enhancement of selectivity for structurally related glycans can be achieved by interaction with ionic buffer additives that allows secondary equilibrium based on monosaccharide composition and different tri-dimensional structures. Also, separations performed in capillaries with suppressed EOF, enable high resolution of isomeric species because of the differences in charge-to-hydrodynamic volume ratio.181 After CE separation with LIF- or MS-detection, unambiguous identification of the glycans can be performed using several approaches including: glycan migration time, glycan fragmentation, glycan molecular mass analysis, and use of sequential enzymatic digests. In the studies presented in this thesis glycan-based approaches have been applied.  a) Glycan migration time Electrophoretic separations offer migration times that are considered a unique characteristic of the analytes. However, reliable migration times, as an approach for glycan characterization, are highly dependable on the reproducibility and precision of the electrophoretic method. Several research groups have published comprehensive data sets containing migration times of fluorescently labeled glycans expressed in glucose units.182-184  Glucose units consist on expressing the migration times of the analytes based on the chain length of a hypothetical dextran oligomer with the same migration time. Although the information obtained is valuable for glycan identification, it requires the use of oligosaccharide standards. Limited commercial availability of oligosaccharides and their production, often restricted by 42  limited separation efficiency, preclude the practical use of glycan standards for structural assignments. Therefore, different approaches are almost always necessary in addition to migration time comparison.     b) Molecular mass The major advantage of using mass spectrometry for glycan analysis is that it provides an information-rich complementary detection approach, and enables the measurement of glycan masses and charge properties, regardless of the use of oligosaccharide standards or exoglycosidase enzymes. Molecular mass is considered the most important strategy for glycan structural assignment as mass can be determined with great accuracy. Glycan composition can be deduced based upon the addition of monosaccharide constituent masses. Information regarding the identity of the monosaccharides can be deduced from the glycan type (N-, O-linked glycan) and the corresponding biosynthetic pathways of the glycans. Additionally, the availability of databases that offer detailed chemical information for thousands of glycans is a valuable tool for assigning glycan composition with high precision based on molecular mass. Major carbohydrate databases include: Bacterial carbohydrate structure DB (http://www.glyco.ac.ru/bcsdb3/), CCSD CarbBank (www.boc.chem.uu.nl/sugabase/carbbank.html),185 CFG Glycan database (http://www.functionalglycomics.org),186 Glycome DB (http://www.glycome-db.org),187 Glycosciences.de (http://www.glycosciences.de),188 GlycosuiteDB (http://www.glycosuitedb.expasy.org/glycosuite/glycodb),189 KEGG Glycan (http://www.genome.jp/ligand/kcam/),190 and the National Institute for Bioprocessing and Research and Training (http://www.nibrt.ie). Despite molecular mass information is considered fundamental for glycan composition, it has a limited potential for elucidating linkage and 43  positional isomers. Therefore, to maximize the level of information provided by MS detection, it can be performed in tandem following collision induced dissociation.      c) Glycan fragmentation Tandem MS experiments result in two major types of glycan cleavage, as illustrated in Figure 1.15. The glycosidic cleavage between two rings gives information on monosaccharide sequence and branching degree, whereas the cross-ring cleavage results from breaking two bonds on the same monosaccharide residue,191 and give information on linkages.192    Figure 1.15 Types of carbohydrate fragmentation as proposed by Domon and Costello. Adapted from191.   Fragmentation can be performed on both positive and negative ion modes that provide complementary information for structural assignment. Characteristic fragment ions are produced either by increasing the orifice voltage in the ESI source or by CID within the collision cell of a triple quadrupole mass spectrometer. The latter is preferred given the high selectivity offered by the triple quadrupole as it detects only the molecular ion that fragments to produce the characteristic ions.80 The occurrence of such ions, upon application of single or multistage fragmentation, depends on a combination of factors including the energy applied to the ion, the time allowed for fragmentation, the charge state of the ion, and the type of glycan.193-195 Despite  44  interpreting fragment spectra is a laborious task, the information obtained is valuable for defining all terminal sequences present in the oligosaccharides. Since the fragmentation of multiple glycan species often results in inaccurate spectral interpretation and erroneous glycan mass-based structural conclusions, glycans must first be CE-separated to reduce the MS spectral congestion. Moreover, CE can resolve isomeric glycans and can reduce the ionization suppression of less abundant species in the presence of predominant ones, given the high efficiency of CE-based separations.     Unlabeled glycans that carry charge, such as sialic acids, have demonstrated to be well CE-separated in counter-EOF and negligible EOF conditions based upon their charge and molecular size.26, 196, 197  Those acidic glycans are most sensitively detected in negative ESI ion mode, and their MS/MS fragmentation often leads to predominant B- and Y-ions due to the loss of terminal sialic acids, and lower amounts of C-type fragments.198 Fragmentation in the negative ion mode has been reported for acidic glycans in various forms, native and derivatized. Despite chemical derivatization via amidation, methylation and permethylation has shown to stabilize sialic acids,199 detrimental charge neutralization and increased hydrophobicity often limit their application in CE-MS, and its is generally recommended to avoid glycan degradation during the MALDI-TOF-MS analysis.200          1.7 RESEARCH OBJECTIVES 1.7.1 Application of a novel CE-MS interface for comprehensive glycoprotein analysis The potential of our flow-through microvial interface has been demonstrated for CE-MS analysis of a wide range of standard analytes. However it has not been fully explored for an analytical challenge such as glycoprotein analysis. In order to determine the applicability of our 45  microvial interface for CE-MS studies of glycoproteins, efficient and reproducible separations are needed. Typically, glycoconjugates require non-volatile buffers in order to achieve their electrophoretic separation. Those CE buffers are not ideal for ESI because they interfere with the ionization process and hinder the detection of analytes in MS. Additionally, the CE analysis of glycoproteins often need capillary surface modifications to avoid detrimental protein-wall interactions that reduce their reproducibility. Therefore the aims were to:  Develop a practical and robust CE-MS method for comprehensive glycoprotein analysis using MS-friendly buffer systems that allow efficient CE separations.   Investigate the potential of positively charged and neutral coatings for CE analysis of protein glycoforms; while paying attention to the compatibility between the CE flow-rate and the ESI-MS flow-rate requirements.  1.7.2 CE-MS characterization of enzymatically released N-glycans from relevant biological sources Because glycosylation largely determines the biological and biophysical properties of glycoproteins, its characterization is mandatory. Currently there is no single method that could provide thorough information on glycan profile due to the complex nature of glycans in terms of composition and structure. Additionally, glycan sample preparation is often laborious and it can involve chemical derivatization that is usually expensive and time-consuming. Glycosylation analysis of recombinant therapeutic glycoproteins in the pharmaceutical industry is often a lengthy process that requires anywhere from several minutes to several hours. Thus the goals were to: 46   Investigate the potential of the flow-through microvial interface for development of robust CE-MS methodologies for N-linked glycan analysis of a relevant recombinant human therapeutic protein.  Demonstrate the importance of separating glycan isomers at high efficiency and highlight the applicability of CE-MS/MS methods for structure elucidation of isomeric species in animal serum with biological significance.  Transfer established CE-MS and CE-MS/MS methods for N-glycan analysis of human serum glycosylation and highlight their inherent merits and shortcomings for identifying disease associated N-glycan alterations.  Design an improved protocol for enzymatically release of human serum N-glycans to ensure experimental precision and to maintain glycan stability that could potentially inform about disease related glycan changes.  Demonstrate the ability of established CE-MS methodologies for analyzing non-derivatized glycan species to simplify sample preparation and reduce total analysis time. Details of the work in achieving these goals are presented in this thesis. 47  Chapter 2:  Potential for comprehensive characterization of glycoproteins by capillary electrophoresis mass spectrometry  2.1 INTRODUCTION  Because protein glycosylation is the most common post-translational modification,93 the development of analytical techniques for glycoprotein characterization in the biomedical and clinical fields have dramatically increased in the last few years.197, 201 Combined methodologies,143, 202-204  which employ “bottom-up” and “top-down” strategies, have been used for characterization of intact protein glycoforms, determination of composition and structure of glycans, and identification of protein glycosylation sites. Excellent reviews about the different methods used for glycoprotein characterization can be found in the literature.100, 101, 136   In recent years, high-resolution separation techniques coupled to mass spectrometry detection have emerged as powerful tools for the analysis of protein glycosylation. 197, 205 CE, with its unmatched separation efficiency, directly coupled to high-resolution MS, could be an attractive tool for comprehensive characterization of protein glycosylation. However, due to the difficulties associated with CE-MS hyphenation, previously discussed in section 1.4.1, only a few reports for CE-MS characterization of intact glycoproteins and glyco-profiling have been published. 205-208  An indirect approach to obtain valuable information about the carbohydrate heterogeneity of a glycoprotein is the precise mass determination of every intact glycoform. Because different carbohydrate combinations can lead to the same mass,209 assignment of a single carbohydrate composition to a specific glycoform is difficult. Therefore, characterization of the released 48  glycan pool is mandatory and complementary to the analysis of intact glycoproteins.210 Characterization of glycoprotein microheterogeneity can be achieved by analyzing, either the enzymatically released N-glycans or the glycopeptides. Despite that glyco-profiling provides a map of all the glycan structures attached to a given glycosylation site, it fails to correlate specific glycan compositions with glycosylation sites. This information can be recovered by performing glycopeptide analysis. In contrast to the profiling of N-glycans, only a few CE-MS studies have been published on glycopeptide-based analysis of glycoproteins.211, 212 These analysis are mainly performed by direct matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS)143, 213 and nano ESI/MS214, 215 without any separation step prior to MS detection.  In this chapter we describe the development of a relatively simple procedure for comprehensive characterization of a model protein (Figure 2.1). RNase B, a small glycoprotein with a single N-glycosylation site at 60Asn, was employed to evaluate the potential of our flow-through microvial CE-MS interface for performing intact glycoprotein analysis. Our methodology compares the performance of several capillary coatings and ESI-compatible background electrolytes for analysis of intact glycoforms. Site-specific glycoprofiling is performed based on a combination of general proteolysis and solid-phase extraction methods, followed by CE-MS analysis of glycoprotein digests. For additional structural information, enzymatically released high-mannose N-glycans are characterized to unequivocally assign carbohydrate composition to the intact protein glycoforms. Since N-glycans of RNase B do not contain readily ionizable groups under the conditions of the analysis, they are derivatized with a novel labeling reagent, prior to CE-MS analysis. The labeling reagent, T3, has a  triazine based-structure that under acidic conditions introduces multiple positive charges in the glycans.216 This 49  labeling reagent significantly enhances the separation efficiencies and increase the sensitivity of glycans by up to 10-fold.216  The combination of data obtained from the glycopeptide and the glycan analyses is used to obtain the peptide mass and the amino acid sequence to confirm the glycosylation site of the glycoprotein. The method demonstrated in this work can potentially be applied to study other glycoproteins in order to increase the understanding of the role between glycan structure and function, and could provide a bridge for interfacing proteomics with functional glycomics.    Figure 2.1Workflow of the different glyco-profiling approaches used in this study for identification of glycoform distribution, glycosylation site and glycan composition of RNase B, as a model protein. CE-MS analysis were performed using an LCQ* Duo ion trap mass and an API 400 triple quadrupole MS, as indicated in the text.  2.2 MATERIALS AND METHODS 2.2.1 Materials Bovine ribonuclease B (RNase B), Polybrene ≥ 95%, dextran sulfate sodium salt M.W.> 500,000 and ε-aminocaproic acid were purchased from Sigma-Aldrich (St. Louis, MO, U.S.A.); GlycoproteinGlycopeptides Free glycans1. Trypsin 1. PNGase F2. SPET3-labeled glycansCE-ESI-MSDetermination of Glycoform Overall Composition, identification of Glycosylation Site and Glycans Quantification 3. Glycan labeling2. SPE50  RNase B was used without further purification. Trypsin was purchased from Promega Corp. (Madison, WI, U.S.A.). A Peptide-N-Glycosidase F (PNGase F) kit and GlycocleanTM H cartridges were obtained from Prozyme (Hayward, CA, U.S.A.). All chemicals and solvents were of analytical grade or better. Formic acid, acetic acid, trifluoroacetic acid, ammonium acetate, methanol, ethanol, acetone, and 2-propanol were purchased from Fisher Scientific (Nepean, ON, Canada). PEI coating reagent of trimethoxysilylpropyl-modified polyethyleneimine, 50% in isopropanol was purchased from Gelest Inc. (Morrisville, PA, U.S.A.). All aqueous solutions were prepared using purified water (18.2 MΩ.cm-1) with a Mili-Q purification system (Millipore, Bedford, MA, U.S.A.).   2.2.2 Digestion of RNase B with trypsin  50 µg of RNase B were adjusted to a final volume of 50 µL in 50 mM ammonium bicarbonate and denaturation was performed at 95  C for 10 min. 1.25 µL of 200 mM dithiothreitol was added and the mixture was incubated at 60  C for 60 min. After that, 5 µL of 200 mM iodoacetamide was added and the sample was incubated at room temperature for 45 min (in the dark), followed by the addition of 1.5 µL of 200 mM dithiothreitol and subsequent incubation for 30 min. The reduced and alkylated glycoprotein was digested with 1 µg of trypsin at 3   C overnight.   2.2.3 Peptide and glycopeptide purification using solid-phase extraction (SPE)   Tryptic-digested glycopeptides were purified by SPE using Oasis HLB columns (Waters Corporation, Milford, MA, U.S.A.). In brief, HLB columns were conditioned with 1 mL of methanol: water (1:1 ratio), followed by 1 mL of solvent A composed of water: methanol: trifluoroacetic acid (97.9:2:0.1 ratio). Glycopeptides were diluted to 500 µL with solvent A prior 51  to loading into the column. Glycopeptides were then washed with 0.5 mL of solvent B composed of water: methanol: formic acid (97.9:2:0.1 ratio) and eluted with a solution containing water: methanol: formic acid (19.9:80:0.1 ratio). Finally, the glycopeptides were dried down in a SpeedVac vacuum centrifuge (Eppendorf, Germany) and reconstituted in 100 µL of distilled water prior to CE-MS analysis.   2.2.4 Deglycosylation of RNase B with PNGase F  Denaturation of 50 µg of RNase B in 10 mM Tris-HCl, pH 8.0 was performed by heating at 100 C for 5 minutes with the subsequent addition of 2.5 µL of 0.1% sodium dodecyl sulfate and 50 mM β-mercaptoethanol, according to the manufacturer’s instruction. After that, 2.5 µL of 0.75% NP-40 detergent was added, followed by 2 µL of PNGase F (1000 mU). The mixture was then incubated at 37 C overnight. Finally, the digested sample was placed at -20 C for 20 minutes to terminate the reaction.   2.2.5 Glycan purification using solid-phase extraction (SPE)  Released N-glycans of RNase B were purified by SPE using GlycocleanTM H cartridges. The cartridges were initially washed with 3 mL of 1M NaOH, followed by 6 mL of distilled water, and 3 mL of 30 % acetic acid. Cartridges were then primed with 3 mL of 50% acetonitrile/ 0.1% trifluoroacetic acid (v/v) in water, followed by 6 mL of 5% acetonitrile/ 0.1% trifluoroacetic acid (v/v) in water. The glycan sample was diluted with water (1:1 ratio) and then loaded onto the cartridge. Non-glycan contaminants were removed by washing the cartridge with 3 mL of Mili-Q water followed by 5% acetonitrile/ 0.1% trifluoroacetic acid (v/v) in water. N-glycans were then eluted with 2 mL of 50% acetonitrile/ 0.1% trifluoroacetic acid (v/v) in water. 52  Finally, N-glycans were dried down in a SpeedVac vacuum centrifuge (Eppendorf, Germany) and reconstituted in 50 µL of distilled water prior to CE-MS analysis.   2.2.6 Glycan derivatization  Enzymatically-released N-glycans were derivatized with the recently reported labeling reagent T3 that improves separation efficiencies and MS sensitivity of glycans.216  The resulting T3-labeled N-glycans were diluted 10-fold with water prior to CE-MS analysis and were used without further purification.   2.2.7 CE-MS system   CE analysis of the intact glycoprotein and the glycopeptides were performed on a P/ACE MDQ capillary electrophoresis system (Beckman Coulter, Brea, CA) coupled with a Finningan LCQ* Duo ion trap mass spectrometer (Thermo Fisher Scientific, Waltham, MA, U.S.A.). A modified capillary cartridge that permits external detection was used for CE-MS. The CE-MS interface with a flow-through microvial was developed in our laboratory and has been described previously.26, 76, 217-219 Electrospray voltage was set at   3.5 k  and the temperature of the heated capillary at the MS inlet was set at 200  C. The detector scan range was 1500-2000 m/z for intact glycoproteins and 300-2000 m/z for glycopeptide analysis. The trap injection time was set at 50 ms. The MS scanning parameters were optimized using the “Autotune” function of the LCQ Xcaliber software (Thermo Fisher Scientific, Waltham, MA, U.S.A.) by continuous infusion of RNase A. The modifier solution was delivered by a syringe pump (Harvard apparatus, Holliston, MA, U.S.A.) at a flow rate of 0.3 µL/min and was composed of 75% methanol/0.1% acetic acid in water.  53   CE-MS analysis of N-glycans were carried out on a PA800 plus capillary electrophoresis system (Beckman Coulter, Brea, CA) connected to an API 4000 triple-quadrupole mass spectrometer (AB SCIEX, Concord, Canada) using the same interface setup described above. Electrospray voltage was set at +3.5 kV. Nitrogen (UHP) was used as curtain gas. Data acquisition and system control were performed using the Analyst® 1.4.2. software (AB SCIEX, Framingham, MA, U.S.A.). The conditions for modifier solution and its delivery rate are the same as the ones when the LCQ* Duo ion trap MS was used.  2.2.8 CE procedure Neutral coated capillaries with polyacrylamide-based hydrophilic wall coating (50 µm ID, 365 µm OD, 67 cm long) were obtained from Beckman Coulter Inc. (Brea, CA), and were employed for intact glycoprotein and N-glycan analyses. Glycopeptides were analyzed using a fused silica capillary (Polymicro Technologies, Phoenix, AZ, U.S.A.) coated with polyethyleneimine (PEI) (50 µm ID, 365 µm OD, 75 cm long). Separations were carried out in normal- and reverse-polarity mode, as indicated in the electropherograms. Background electrolytes were composed of 30-100 mM ammonium acetate (pH 3.0-4.0), 50-150 mM ε-aminocaproic acid (pH 3.0-4.0) or 1% formic acid in methanol: water (1:1).The samples were injected at 1.0 psi for 10 s, corresponding to a volume of 19 nL.       2.3 RESULTS AND DISCUSSION 2.3.1 Intact glycoprotein analysis  Electrophoretic separation of glycoforms of RNase B was performed using both a neutral coating, and positively-charged capillary coatings. A neutral coated capillary with polyacrylamide-based hydrophilic wall coating was tested. In this type of coating, the magnitude 54  of the EOF is negligible and separation occurs based exclusively on the intrinsic electrophoretic mobility of the analytes under normal-polarity mode. The use of these capillaries has been demonstrated with our flow-through microvial interface for protein analysis. 76 BGE’s consisting of ammonium acetate and ε-aminocaproic acids were tested in the concentration range of 30-100 mM (pH 3.0-4.0) and 50-150 mM (pH 3.0-4.0), respectively. Ammonium acetate BGEs provided better separation than ε-aminocaproic acid BGEs. An increase of ammonium acetate BGE concentration from 30 to 100 mM led to longer migration times with a concomitant improvement of glycoform separation. BGE concentrations higher than 100 mM led to increased migration times without improvement of the separation, and also led to unstable ESI operation due to high CE currents.219 A pH reduction of the BGE from 3.5 to 3.1 slightly improved the CE separation. At pH 3.1, it was observed five partially separated peaks. With an optimum BGE consisting of 100 mM ammonium acetate at pH 3.1, RNase B glycoforms were positively charged, and under normal-polarity mode, they migrated to the cathode, depending on their charge-to-size ratio with heavier glycoforms migrating later. Under the optimum conditions, the electrophoretic widths for the peaks at half-height were only 10 s, which demonstrates the good peak capacity of CE. Figure A.1 in the appendix section shows the mass spectra obtained for the glycoforms of RNase B. Glycoforms were observed in the 8+ and 9+ charge states, with the 8+ charge state generally displaying the highest intensity. The m/z window was restricted to 1500-2000 Da which increased the acquisition speed and the sensitivity of the analysis. Mass measurements using the LCQ* Duo ion trap MS allowed direct determination of the exact molecular mass of the glycoforms.  Figure 2.2 shows a contour plot of migration time vs. m/z for the separated glycoforms. As seen in Figure 2.2, only 8+ and 9+ charge state species are observed for the glycoforms 55  because their m/z falls within the selected range of the ion trap MS. Both charge states can be clearly observed with a trend moving to increasing migration times at increasing m/z, from RNase B-Man5 to RNase B-Man9. In addition to the expected molecular ions for RNase B glycoforms, there are two additional ions that appear at a lower m/z than the 8+ and 9+ charge states of the glycoforms, and correspond to RNase A (also detected as 8+ and 9+ charge states species). Usually RNase A is present as a contaminant in RNase B preparations and due to its non-glycosylated nature it displays a reduced molecular weight and it is detected as a single peak. Under the optimum conditions, RNase A shows a stronger mobility than RNase B glycoforms due to its higher charge-to-size ratio and it is baseline-separated from RNase B glycoforms.   Figure 2.2 Contour plot of migration time vs. m/z of RNase B glycoforms showing charge states corresponding to [M+8H]8+ and [M+9H]9+. Same conditions are used as in Figure 2.3C.  Positively-charged capillary coatings that were tested include: a triple-layer coating consisting of Polybrene (PB), Dextran Sulfate (DS) and PB, and a single layer coating of PEI. The combination of a positively-charged capillary wall, acidic BGE and reverse-polarity mode 02072012 rnase b cems 12 RT: 6.68 - 14.06 Mass: 1500.07 - 2000.00 NL: 2.26E5 F: + p NSI Full ms [1500.00-2000.00]7.0 7.5 8.0 8.5 9.0 9.5 10.0 10.5 11.0 11.5 12.0 12.5 13.0 13.5 14.0Time (min)15001550160016501700175018001850190019502000m/zRNase B RNase A[M+9H]9+[M+8H]8+Man 9Man 8Man 7Man 6Man 5Man 8Man 7Man 6Man 556  generated a strong anodal EOF. The magnitude of the electrophoretic mobilities of the glycoforms, migrating toward the cathode, was significantly smaller than the electroosmotic mobility, and they were detected at the anode in short times. Ammonium acetate BGE’s in the concentration range from 30 to 100 mM (pH 3.0- 4.0) were employed with PB-DS-PB and PEI-coated capillaries. Figure 2.3A shows the extracted ion electropherograms (EIEs) obtained for the analysis of glycoforms of RNase B in PB-DS-PB coated capillaries, using 100 mM ammonium acetate (pH 3.0) as the BGE. The EIE represents all ion species associated with a single glycoform. The migration time of the glycoforms depends on the number of hexose units of the attached glycans to the glycoprotein. Glycoforms containing the highest number of hexoses (RNase B-Man9) displayed a stronger apparent mobility towards the anode and thus, were detected earlier than those containing a lower number of hexoses (RNase B-Man5). Similar results for the migration order of the glycoforms were obtained when PEI-coated capillaries were employed. However, the magnitude of the EOF in PEI-coated capillaries was smaller than that of PB-DS-PB-coated capillaries (6.09 x 10-4 cm2 V-1 S-1 vs. 6.7 x 10-4 cm2 V-1 S-1 in 100 mM ammonium acetate pH 3.0) and this characteristic slightly improved the differences in the mobilities of the glycoforms, leading to better separation. As observed in Figure 2.3B, the migration time for the glycoforms in PEI-coated capillaries was longer than that in PB-DS-PB-coated capillaries (Figure 2.3A) due to the reduced EOF. However, this reduced EOF only increased the migration time window in 0.3 minutes with respect to the triple layer coating, without allowing an effective separation.  57   Figure 2.3 EIE displaying the migration of intact glycoform species of RNase B using a (A) PB-DS-PB coated capillary, (B) PEI-coated capillary and (C) polyacrylamide-based neutral capillary. The optimum BGE was 100 mM ammonium acetate. In (A) and (B), the pH of the BGE was 3.0, and the CE polarity was reversed. In (C) the pH of BGE was 3.1, and CE polarity was normal. Data was acquired using the LCQ* Duo ion trap MS.   Table 2-1 lists the observed m/z of the 5 glycoforms as well as their relative abundances (%). The observed m/z for the glycoforms concurs with those previously reported.220, 221 Monosaccharide composition, observed m/z and relative abundances (%) are also included for the enzymatically released N-glycans of RNase B, which will be discussed later.  The CE-MS method provides two dimensions of separation, and although some degree of overlapping is observed in time domain (i.e. CE separation) when using the neutral coated capillary, glycoforms can be fully resolved in the second dimension (i.e. MS detection). The separation power of CE-MS allows resolving structurally similar glycoforms with small mass differences (ΔM = 162 Da) on the level of a 15-KDa glycoprotein. The CE-MS method seems to be well suited for providing an overall picture of the number of protein glycoforms and their relative abundances. The approach is simple, fast and could be potentially useful for characterization of various intact glycoproteins without laborious sample preparation. However, detailed information about the glycan chains and glycosylation site is not possible to obtain with 2 4 6 8 10 12 14 16020406080100   RNaseBMan9 RNaseBMan8 RNaseBMan7 RNaseBMan6 RNaseBMan5Time (min)Relative Abundance (%)2 4 6 8 10 12 14020406080100  Time (min)2 6 8 10 2 14 16 18020406080100 Time ( in)(A) (B) (C )58  this approach. Therefore, bottom-up strategies on the level of enzymatically released glycans and glycopeptides are also examined in the following sections, to provide a comprehensive and detailed characterization of the protein glycoforms.  Table 2-1 Identification and annotation of the intact glycoforms and glycan components of RNase B including their abundances. Glycoforms were detected as [M+8H]8+ and [M+9H]9+ ion species, with the lower charge state being the most intense ion. Data was acquired using the LCQ* Duo ion trap MS. Glycans were detected as [M+H]+ and [M+2H]2+ species using the API 4000 triple-quadrupole MS. Composition and structural schemes for glycans are given in terms of N-acetylglucosamine (blue squares) and mannose (green circles). The numbers indicated by an asterisk on the right hand side panel indicate the abundance of glycans determined with LC-MS,181 as discussed in the text.      Intact Glycoprotein Glycan components Glycoform Obs.  m/z [M + 8H] 8+ Abundance  (%) Composition Glycan structure Obs.  m/z Abundance  (%) RNase B - Man5 1864.13 48 GlcNAc 2 Man 5 1471.1  [M+H] + 49.82 *(52±2) RNase B - Man6 1884.93 20 GlcNAc 2 Man 6 1633.3  [M+H] + 18.85 *(32±2) RNase B - Man7 1905.27 11 GlcNAc 2 Man 7 906.9  [M+2H] 2+ 10.57 *(6.3±0.6) RNase B - Man8 1925.47 17 GlcNAc 2 Man 8 987.9  [M+2H] 2+ 16.24 *(7.2±0.8) RNase B - Man9 1945.33 4 GlcNAc 2 Man 9 1069.0 [M+2H] 2+ 4.52 *(1.6±0.1) 59  2.3.2 Analysis of enzymatically released N-glycans Protein glycosylation was determined by enzymatic digestion of RNase B with PNGase F.  Direct infusion of the released glycans at m/z 900- 1900 Da showed a complex MS spectrum with multiple cation adducts signals. Therefore, prior electrophoretic separation before the MS analysis was necessary for obtaining better quality mass spectra that leads to an improved characterization of the oligosaccharides.  Following the one-step labeling procedure of N-glycans with T3,216 whose structure is  depicted in Figure 2.4A, the glycans were analyzed by CE-MS using a neutral coated capillary with polyacrylamide-based hydrophilic wall coating. Figure 2.4B shows the electrophoretic separation for the derivatized N-glycans using a BGE containing 50 mM ammonium acetate (pH 3.0). Under these conditions, the glycans migrate based on their intrinsic electrophoretic mobilities toward the cathode. In this sense, T3-labeled glycans with an increased amount of mannose residues in the antennae region have a lower mobility than smaller glycans and appear later in the electropherogram. For instance GlcNAc2Man9 has a longer migration time than GlcNAc2Man5.  As seen in Figure 2.4B, the glycan separation achieved is excellent, and glycans with the same charge but different numbers of mannose residues (distinguished by 162 Da per mannose residue) are baseline-separated. However, the peaks observed are particularly broad due to the presence of multiple isomers, corresponding to different types of branching within the oligosaccharide chains. These isomers can be slightly separated by CE based on the hydrodynamic volume of the branched glycans and led to the observed broad peaks.181, 222  According to Hua et al.,223 13 site-specific glycans for RNase B have been identified based on MS/MS characterization of the N-glycans by nano-LC-MS.223 The number of isomers for each 60  glycan composition has been reported as 2, 2, 3, 4 and 2, corresponding to GlcNAc2Man5, GlcNAc2Man6, GlcNAc2Man7, GlcNAc2Man8 and GlcNAc2Man9, respectively. However, Prien et al.222 have reported an increased number of isomers for GlcNAc2Man5 and GlcNAc2Man7, i.e., 4 and 5 isomers, respectively that were identified and characterized by ion trap MS.222  Table 2-1 shows the observed m/z for the T3-labeled glycans along with their relative abundances. Glycans identified as GlcNAc2Man5 and GlcNAc2Man6 are detected as [M+H]+ species while glycans corresponding to GlcNAc2Man7, GlcNAc2Man8 and GlcNAc2Man9 are detected as [M+2H]2+ species.  (A)  (B)  Figure 2.4 (A) Structure of the T3-labeling reagent employed for derivatizing the N-linked glycans from RNase B. (B) Base peak electropherogram showing baseline separation for N-glycans of RNase B labeled with T3. Peaks are identified as follows: (1) GlcNAc2Man5, (2) GlcNAc2Man6, (3) GlcNAc2Man7, (4) GlcNAc2Man8, Base Peak Chrom. of +Q1: from Sample 6 (02162012 RNase B PNGase F 06) of 02162012 RNase B PNGase F.wiff mass range 700.0... Max. 4.5e6 cps.28 29 30 31 32 33 34 35 36 37 38 39 40Time, min0.02.0e54.0e56.0e58.0e51.0e61.2e61.4e61.6e61.8e62.0e62.2e62.4e62.6e62.8e63.0e63.2e63.4e63.6e63.8e64.0e64.2e64.4e64.6e64.8e6Intensity, cps32.0833.0834.8534.0235.42(1)  GlcNAc2Man5GlcNAc2Man6GlcNAc2Man7GlcNAc2Man8GlcNAc2Man9(2) (3) (4) (5) 61  and (5) GlcNAc2Man9. The inserts in the left show the putative glycan structures. Glycans were detected in positive ESI mode. Data was acquired using the API 4000 triple-quadrupole MS.  Figure A.2 in the appendix section shows the MS spectra of individual T3-labeled glycans from GlcNAc2Man5 to GlcNAc2Man9 in the selected m/z range from 900-1900 Da. As seen in Figure A.2A and A.2B, the fragmentation performed on GlcNAc2Man5 and GlcNAc2Man6 by using a high sample cone voltage, produced CID of the selected glycans. The high voltage was experimentally determined by comparing the intensities of the fragment ions, and a cone voltage of 60 V was selected as the optimum. Nitrogen was used as the neutral collision gas.224 In Figure A.2A, the CE-MS/MS spectrum of precursor ion at m/z 1471.1 corresponds to T3-labeled GlcNAc2Man5 ([T3-HexNAc2Hex5+H]+) and shows fragments ions at m/z 1309.1, 1147.0 and 985.0. The fragment ions observed correspond to [T3-HexNAc2Hex4+H]+, [T3-HexNAc2Hex3+H]+ and [T3-HexNAc2Hex2+H]+, respectively,  due to the loss of 162 Da  from the precursor ion. The CE-MS/MS spectrum of T3-labeled GlcNAc2Man6 ([T3-HexNAc2Hex6+H]+), shown in Figure A.2B, also shows fragments at m/z 1471.1, 1309.0, 1147.0 and 985.1 that correspond to [T3-HexNAc2Hex5+H]+, [T3-HexNAc2Hex4+H]+, [T3-HexNAc2Hex3+H]+ and [T3-HexNAc2Hex2+H]+ respectively, due to the loss of multiple Hex residues from the precursor ion. In both cases, the loss of 162 Da provides conclusive evidence on the high-mannose nature of the released N-glycans of our model glycoprotein. Figures A.2C, A.2D and A.2E show the CE-MS spectra of T3-labeled GlcNAc2Man7, GlcNAc2Man8 and GlcNAc2Man9, respectively. For these glycans, doubly-charged species were detected in the m/z range from 900 to 1900 Da. The good resolution of the glycan peaks allowed performing full-scan analysis on each peak at 1000 amu/s.   62  According to Table 2-1, the relative proportions of the abundances of the intact glycoforms correlate well with those of the enzymatically released N-glycans. For a protein containing a single glycosylation site, this suggests that each oligomannose structure represents each of the five glycoforms of RNase B. Therefore, an indirect characterization of glycoform populations in terms of their oligosaccharide content is possible. Quantification of the relative glycan abundances in the glycoprotein is possible because the T3-labeled glycans have similar structures and their ionization efficiencies are similar as well. The relative abundances of RNase B glycans acquired using LC-MS of glycans labeled with negatively charged reagent and separated on a reversed-phase,181 are also included in Table 2-1. Although there are differences between the results, such differences might be attributed to the source of RNase B samples employed.   2.3.3 Glyco-profiling of glycosylation site of RNase B  Site-specific glyco-profiling of RNase B was performed by enzymatic digestion with trypsin to produce glycopeptides of adequate length ( 4 to 10 amino acid residues), which were purified from hydrophobic peptides using SPE methods. Because limited proteolysis is key to performing an adequate site-specific analysis of glycosylation sites, trypsin was chosen as the proteolytic enzyme.143 Furthermore, tryptic-digested glycopeptides can be detected as multiply charged ions that fall within the optimum m/z range of the LCQ* Ion trap MS, where it performs well in terms of resolution and sensitivity. Use of tryptic-digested glycopeptides allows determination of both the amino acid sequence and the glycosylation site.225 Mixtures of non-specific enzymes like Pronase have also been reported for performing site-specific glyco-profiling of glycoproteins.223 Although Pronase yields glycopeptides of adequate length, its low 63  activity makes it necessary to use relatively high concentrations, which requires lengthy and laborious purification prior to the analysis of the resulting glycopeptides. Moreover, while the utility of Pronase is advantageous for performing glycan profiling,213, 226 information regarding  the variability at the glycosylation site can be lost.    Considering the inherently low ionization efficiency of glycopeptides and potential signal suppression caused by nonglycosylated peptides, electrophoretic separation was employed before mass spectral screening was attempted. CE separation of glycopeptides was achieved using a PEI-coated capillary with 1% formic acid in methanol: water (1:1) as the BGE. The capillary coating minimizes adsorption of the glycopeptides to the wall of the fused-silica capillary and provides a strong, reversed EOF. EIEs of the glycosylated peptides isolated after trypsin digestion are shown in Figure 2.5. The EIEs were obtained after the glycopeptides were cleaned, desalted and concentrated by SPE. Glycopeptides containing larger glycans (i.e. GlcNAc2Man9) have a stronger apparent mobility toward the anode due to the reversed EOF and migrate faster than glycopeptides containing smaller glycans (i.e. GlcNAc2Man5). Although baseline separation in the time domain is not achieved for all the glycopeptides, the orthogonal baseline MS resolution separates them in the second dimension. Despite the MS signals corresponding to large, nonglycosylated peptides, while minimized, were not completely eliminated, did not preclude the observation of the corresponding glycopeptide patterns. The glycopeptides were mainly observed as doubly-charged species, [M+2H]2+, and  singly-charged species, [M+H]+, were only observed for the smallest glycopeptide. Glycopeptide peaks were observed at m/z 929.64, 1010.67, 1091.69, 1172.57 ([M+H]+ species) and m/z 1695.59 ([M+2H]2+ species) and corresponded to peaks that differ by 162 Da or one hexose residue.   64   Figure 2.5 Extracted ion electropherograms for the CE-MS separation of glycopeptides of RNase B. BGE contains 1% formic acid in methanol: water (1:1). Glycopeptides were identified as: (1) GlcNAc2Man9 at m/z 1172.57, (2) GlcNAc2Man8 at m/z 1091.69, GlcNAc2Man7 at m/z 1010.67, (4) GlcNAc2Man6 at m/z 929.64 and (5) GlcNAc2Man5 at m/z 1695.59. Data was acquired using the LCQ*Duo ion trap MS.  To identify the peptide moieties in Figure 2.5, the glycans were released from the glycopeptides with PNGase F. The peptide moieties were identified by subtracting the masses of the previously observed glycans from the masses of the glycopeptides. For example, the peptide mass (459.49 Da), obtained by subtracting the observed glycan mass 1235.09 Da ([M+H]+ minus T3-labeling and H20, and H+) from the observed glycopeptide mass 1695.59 Da ([M+H]+ minus H+), corresponded to the peptide sequence NLTK , with the glycosylation site at position 60Asn. Peaks at m/z 929.64, 1010.67, 1091.69 and 1172.57 correspond to protonated adduct, i.e.  glycosylated peptide NLTK with GlcNAc2Man6, glycosylated peptide NLTK with GlcNAc2Man7, glycosylated peptide NLTK with GlcNAc2Man8 and glycosylated peptide NLTK with GlcNAc2Man9, respectively. The peak at m/z 1695.59 corresponded to the glycosylated peptide NLTK with GlcNAc2Man5, and were therefore, obtained from the single glycosylation site of RNase B at 60Asn. The only peaks observed above m/z 900 unequivocally correspond to 0 2 4 6 8 10 12 140.02.0x1044.0x1046.0x1048.0x1041.0x1051.2x1051.4x105Intensity (cps)Time (min)(2)(1)(3)(4)(5)65  glycopeptides because the carbohydrate moieties have a mass of at least 800 mass units due to the presence of the chitobiose core (GlcNAc2Man3). Through the glycopeptide mapping, the glycosylation site of RNase B was confirmed and concurs with other reports in the literature.203, 213, 223  These findings also supported the results obtained for the analysis of enzymatically released N-glycans. Based on the regular 162 Da spacing of the glycopeptide masses, the high-mannose structure of the glycans was confirmed.  Table 2-2 shows the data from single-stage MS used for glycopeptide identification. According to Table 2-2, the most abundant glycopeptides were found at m/z = 1695.59, 929.64, 1010.67, 1091.69 and 1172.57 which corresponds to glycopeptides containing from GlcNAc2Man5 to GlcNAc2Man9, respectively. The relative abundance of glycopeptides accurately reflects the relative abundances of all five glycoforms presented in Table 2-1. Also, the relative abundances of N-glycans concurs with that of the glycopeptides, which indicates that it is possible to quantify the glycan contents at the glycosylation site for the case of proteins with little heterogeneity like RNase B.   Table 2-2 Glycopeptide identification including the observed m/z, experimental masses and relative abundances. Glycopeptides were detected as [M+H]+ and [M+2H]2+ ion species. Data was acquired using the LCQ* Duo ion trap MS.   Obs. m/z glycopeptideExperimental glycopeptide massglycopeptideabundance (%)Experimental glycan mass474.38 474.381695.59 1694.58 48.33 1235.09929.64 1857.26 20.96 1397.091010.67 2019.32 10.87 1558.781091.69 2181.36 13.96 1720.781172.57 2343.12 5.88 1882.9866  As the correlation between the two sets of data (Table 2-1 and 2-2) suggests, both enzymatic procedures released N-glycans and glycopeptides almost in molar proportions, and their relative abundances also agree with those of the intact glycoforms. For that reason, the CE-MS method also offers the possibility of indirect analysis of glycoforms at the intact protein level and can be applied when there is rarely enough material for a second dimension analysis.   2.4 CONCLUDING REMARKS  A combined approach for comprehensive analysis of intact glycoforms of RNase B and enzymatically released glycans as well as glycopeptides by CE-MS, with a flow-through microvial interface, is presented. The methodology offers a valuable tool for fast and reliable elucidation of protein N-glycosylation. Electrophoretic separations conducted under conditions of negligible EOF and acidic buffers allowed the resolution of both protein glycoforms and N-glycans with very similar electrophoretic mobilities. The glycan methodology successfully provides detailed information about the type and composition of the oligosaccharides which are in agreement with previous reports. The one-step derivatization of N-glycans with the T3 labeling agent is a simple, clean and efficient alternative that allows effective glycan ionization and detection by ESI-MS, and it can be suitable to characterize complex protein glycosylation from other sources. Glycopeptide mapping was used for elucidation of the glycosylation site, to circumvent the limitations of glycan profiling in which the loss of site-specificity prevents linkage of a glycan structure to its origin.  The experimental approach allowed quantitative comparison among the relative abundances of the intact glycoforms, glycopeptides and derivatized glycans which showed a high degree of correlation. Therefore, the presented CE-MS methodology may prove to be a useful 67  qualitative and semi-quantitative tool for obtaining the highest amount of information from glycoproteins. In this regard, the method succeeded in providing an overall composition of protein glycoforms, and its enzymatically released derivatives. This in-depth analysis could be used for the characterization of the glycosylation pattern of any glycoprotein, and could potentially aid in resolving biological aspects of recombinant therapeutic glycoproteins.   68  Chapter 3: A simple CE-MS method for complex N-glycan analysis using a flow-through microvial interface   3.1 INTRODUCTION Characterizing glycosylation of proteins is still one of the most challenging tasks for bioanalytical chemists. The non-template control of the biosynthetic pathway of glycosylation is the basis for the great complexity of glycans.135  Extensive microheterogeneity often occurs at the glycosylation site of proteins, leading to a large glycoform population. Because protein glycosylation plays a key role in a wide range of biological processes, aberrant protein glycosylation in mammalian systems is often associated with pathogenesis.93 Given the potential of recombinant glycoproteins as therapeutic drugs and diagnostic reagents, the analysis of protein glycosylation has attracted much interest in recent years.227-229 Glycotyping, a comprehensive and detailed characterization of the various glycoforms present in a glycoprotein, has been used to detect subtle yet biologically relevant differences in glycan composition of different batches of glycoprotein pharmaceuticals.230, 231 Such differences can affect their biological activities and influence drug safety, efficacy and stability.232 The common practice to determine the glycan composition of recombinant or plasma-purified protein therapeutics involves the enzymatic release of the carbohydrate moiety with endoglycosidases (peptide N-glycosidase F or A), followed by the analysis and detection of either labeled or underivatized glycans.229, 233 196 Various chromatographic methods have been used for the analysis of enzymatically released glycans. Hydrophilic-interaction chromatography (HILIC),228, 234 porous graphitized 69  carbon (PGC) LC-MS,227, 235 and reversed-phase liquid chromatography (RP-LC),236, 237 are routinely used for identifying glycan components in complex mixtures. However, since glycans are typically hydrophilic, they are not well retained by RP-LC, and common alternatives such as HILIC and PGC, that better retain hydrophilic glycans, often result in longer retention times versus RP-LC. Additionally, these processes often require time-consuming derivatization processes with fluorophores that can introduce bias in the detection sensitivity of the glycans because of the purity of the labeling reagents or the derivatization steps.238, 239 Over the past few years, on-line coupling of capillary electrophoresis (CE) to time-of-flight mass spectrometry (TOF-MS) detection has emerged as an attractive and complementary technique for characterization of biomolecules.48, 49, 90, 240 Because CE-TOF-MS combines high separation efficiencies with high mass resolution and mass accuracy, it may be a powerful method for glycan analysis. CE-MS of carbohydrates has been previously done with coaxial sheath-liquid241, 242 and sheathless interfaces.211 Although sheath-liquid CE-MS interfaces are relatively stable and robust, the required high volumetric flow of sheath-liquid contributes to considerable analyte dilution in comparison to sheathless CE-MS interfaces. The use of sheathless CE-MS interfaces often face difficulties in fabrication and operation. Unlabeled carbohydrate analysis using sheathless interfaces has been reported to be extremely slow  (  120 minutes) for routine use when CE in reverse polarity was performed.211  We have previously described a novel flow-through microvial interface in which a chemical modifier solution mixes with the CE effluent and delivers the analyte from the capillary terminus to the source.76 The major advantages of this interface are that (i) the modifier solution is delivered at low flow rates, typically 100-500 nL/min to minimize analyte dilution, and (ii) the stainless steel electrospray emitter holds the separation capillary inside, making the operation 70  very robust. The beveled shape of the electrospray emitter stabilizes the ESI spray in a wide range of flow rates.243  Successful applications of this low flow sheath-liquid interface have been reported for the analysis of fluorescently labeled carbohydrates,219 separation of isomeric O-acetylated N-glycans in fish serum,26 which will be further discussed in detail in Chapter 4, monitoring prostate cancer biomarkers,244 sensitive metabolic profiling studies245 and for the coupling of capillary isoelectric focusing (cIEF) to MS.218 In this chapter, we describe a relatively simple approach including enzymatic digestion, SPE purification, CE separation and MS detection of underivatized neutral and acidic N-glycans from immunoglobulin G (IgG) and recombinant human erythropoietin (rHuEPO). Unlabeled glycans were analyzed to preserve their native structural characteristics and to avoid time-consuming derivatization procedures. Glycosylation of rHuEPO was also studied using on-line liquid chromatography-mass spectrometry (LC-MS) to evaluate its ability, in comparison with CE-MS, to detect underivatized glycans along with their secondary modifications.  3.2 MATERIALS AND METHODS  3.2.1 Materials Standard rHuEPO produced in a Chinese hamster ovary (CHO) cell line was obtained as Biological Reference Preparation (BRP) from Pharmacopeia (EDQM, European Pharmacopeia, Council of Europe, Strasbourg, France) and stored according to the manufacturers’ specifications. IgG from human serum (containing all the IgG isotypes) was purchased from Sigma-Aldrich (St. Louis, MO, USA) and was used without any further purification. A Peptide-N-Glycosidase F (PNGase F) kit and GlycocleanTM H cartridges were obtained from Prozyme (Hayward, CA, USA). All chemicals and solvents were of analytical grade or better. Formic 71  acid, acetic acid, trifluoroacetic acid, ammonium acetate, methanol, ethanol, acetone, and 2-propanol were purchased from Fisher Scientific (Nepean, ON, Canada). All aqueous solutions were prepared using purified water (18.2 MΩ.cm-1) with a Mili-Q purification system (Millipore, Bedford, MA, USA).  3.2.2 rHuEPO purification    Each vial of rHuEPO BRP contained 0.25 mg of EPO, 0.1 mg of Tween 20, 30 mg of trehalose, 3 mg of arginine, 4.5 mg of NaCl, and 3.5 mg of Na2HPO4. rHuEPO vial contents were dissolved in 250 μL of Milli-Q water. The resulting solution was desalted and preconcentrated through ultracentrifugation using an Ultracel®-10K centrifugal filter unit (Millipore Corporation, MA, USA). In brief, the filter membrane was initially rinsed with 250 μL of Milli-Q water for 10 min at 13000 g; both retentate and eluent were discarded. Then, 50 μL of reconstituted rHuEPO solution was added to the membrane and centrifuged for 10 min at 13000 g. The eluent was discarded and the retentate was washed by adding 200 μL of Milli-Q water and centrifuged for 10 min at 13000 g. The retentate was washed 3 more times for 10 min under the same centrifugal force and the eluent containing the excipients of low molecular mass was discarded. The retentate which is the desalted rHuEPO was recovered from the cartridge by centrifuging the membrane upside down into a new microcentrifuge tube for 3 min at 1000 g. This step was repeated one time by adding 2 μL of Milli-Q water to the membrane in order to increase rHuEPO recovery. The recovered volume of rHuEPO solution was 30 μL. Sample was stored at 4 C.   72  3.2.3 Glycoprotein deglycosylation with PNGase F   N-glycans from purified rHuEPO and IgG from human serum were released by enzymatic digestion with PNGase F. 50 μg of rHuEPO or 100 µg of human IgG were adjusted to a final volume of 50 μL with 10 mM Tris-HCl at pH 8.0. Samples denaturation was achieved by adding 2.5 µL of denaturation solution, containing 0.1% sodium dodecyl sulfate and 50 mM β-mercaptoethanol, and heating at 100 C for 5 minutes. Samples were then treated with 2.5 µL of detergent solution containing 0.75% NP-40, followed by 2 µL of PNGase F (1000 mU). The solutions were incubated overnight at 37 C.   3.2.4 Glycan purification using solid-phase extraction (SPE)   Released N-glycans from rHuEPO and human IgG were purified by solid-phase extraction using GlycocleanTM H cartridges. The cartridges were initially washed with 3 mL of 1 M NaOH, followed by 6 mL of distilled water, and 3 mL of 30 % acetic acid. Cartridges were then primed with 3 mL of 50% acetonitrile/ 0.1% trifluoroacetic acid (v/v) in water, followed by 6 mL of 5% acetonitrile/ 0.1% trifluoroacetic acid (v/v) in water. The glycan samples were diluted with water (1:1 ratio) and then loaded onto the cartridges. Non-glycan contaminants were removed by washing the cartridges with 3 mL of Mili-Q water followed by 5% acetonitrile/ 0.1% trifluoroacetic acid (v/v) in water. Glycans were then eluted with 2 mL of 50% acetonitrile/ 0.1% trifluoroacetic acid (v/v) in water. Finally, N-glycans were dried down in a SpeedVac vacuum centrifuge (Eppendorf, Germany) and reconstituted in 20 µL of distilled water prior to CE-MS and/or LC-MS analysis.    73  3.2.5 Electrophoretic procedure   CE-MS analyses of SPE-purified N-glycans were performed using hydrophilic hydroxypropyl cellulose-coated capillaries (HPC-capillaries, 50 µm ID, 365 µm OD, 70 cm long) prepared in-house according to the method of Shen and Smith.22 Separations were carried out in reverse-polarity mode at -25 kV with an overimposed pressure of 10 mbar (0.145 psi) to reduce analysis time for neutral glycans. Background electrolytes (BGE) were composed of 30-100 mM ammonium acetate (pH 3.0-4.0) with 20% methanol. The samples were injected at 1.0 psi for 10 s, corresponding to a volume of 19 nL.   3.2.6 CE-ESI-MS system   CE-ESI-MS analyses of N-glycans were carried out with a PA800 plus capillary electrophoresis system (Beckman Coulter, Brea, CA) coupled to an API 4000 triple-quadrupole mass spectrometer (AB SCIEX, Concord, Canada). A modified capillary cartridge that permits external detection was used for the analysis. The CE-MS interface with a flow-through microvial was developed in our laboratory and has been described previously.26, 76, 217-219, 244, 245 The electrospray voltage was set at -3.5 kV. Data acquisition and system control were performed using the Analyst® 1.4.2. software (AB SCIEX, Framingham, MA, USA.). The modifier solution was delivered by a syringe pump (Harvard Apparatus, Holliston, MA, USA) at a flow rate of 0.3 µL/min and was composed of 10 mM ammonium acetate (pH 3.1) containing 75% of 2-propanol and methanol (2:1 ratio). The same interface setup and solutions were used for CE-ESI-MS with an HP 3DCE (Agilent Technologies, Waldbronn, Germany) coupled to a maXisTM Ultra High-Resolution TOF mass spectrometer (Bruker Daltonik, Bremen, Germany). The mass spectrometer was operated in negative ionization mode and acquired data in full scan mode with 74  a mass range from m/z 500-3000 at a spectra rate of 1 Hz. Optimization of the transfer parameters and MS calibration were performed with TuneMixTM (Agilent Technologies) in order to obtain the best sensitivity at satisfactory resolution (R≈ 20,000 at m/z 1334). Instrument control and data analysis were performed using ESI Compass 1.3 application software from Bruker (Bruker Daltonik, Bremen, Germany). The conditions for modifier solution and its delivery rate are the same as the ones when the API 4000 MS was used.   3.2.7 LC-MS system   Free non-reduced N-glycans were separated on Hypercarb porous graphitized carbon column (5 µm particle size, 320 µm (ID) x 10 cm, Thermo Scientific) on an Ultimate 3000 LC (Dionex) and analysed using an HCT 3-D ion trap (Bruker Daltonics) coupled directly to the LC. Separation was carried out at a constant flow rate of 2 µl/min using a linear gradient with 2–16% (v/v) ACN/10 mM NH4HCO3 for 60 min, followed by a gradient from 16–45% over 20 min before washing the column with 45% (v/v) ACN/10 mM NH4HCO3 for 6 min and re-equilibrating in 10 mM NH4HCO3. ESI-MS was performed in negative ion mode with the following scan events: MS full scan with mass range m/z 100–2,000 (scan speed: 8,100 m/z / sec) and data dependent MS/MS scans after CID of the top three most intense precursor ions (ion target was 200,000, max acc. time: 200 ms). Dynamic exclusion was disabled. The mass spectrometers were externally calibrated using tune-mix (Agilent) prior to data acquisition. The monosaccharide compositions and acetylation of the N-glycan structures were manually determined by their corresponding mass, fragmentation in CID-MS/MS and relative PGC LC retention time.246    75  3.3 RESULTS AND DISCUSSION  3.3.1 Analysis of N-glycans from human IgG Complex bi-antennary, core-fucosylated and partially galactosylated, and sialylated oligosaccharides are typically found at the single N-glycosylation site of human IgG, located in the constant region of the heavy chain.247  The variable presence of bisecting GlcNAc and α-2,3- or α-2,6-linked sialic acids on terminal galactose residues contribute to the glycan heterogeneity.248 Despite human IgGs are characterized by relatively simple glycosylation profiles, changes in their glycosylation are associated with immune responses and with the activity of recombinantly expressed monoclonal antibodies (mAbs) for therapeutic purposes.248, 249  Regarding the clinical role of mAbs, it is crucial an in-depth characterization of their glycosylation for activity, safety and quality control during development and production.250  Because of the utmost importance of glycosylation in human IgG for its biological and clinical functions, here we present a simple and efficient method for analyzing its N-glycosylation profile by CE-MS. The method will be further used to explore the N-glycosylation of rHuEPO. Based on our experience in analyzing native N-glycans from fish serum,26 we used neutral coated capillaries and acidic background electrolytes (BGEs) to optimize the CE-MS conditions for underivatized N-glycans from human IgG. To explore the possibility of performing CE-MS analysis of neutral and acidic (sialylated) glycans simultaneously, CE was operated in reversed polarity while ESI ionization was performed in the negative ion mode. Analysis of N-glycans have been reported to be exclusively performed in uncoated capillaries using highly alkaline background electrolytes (BGE’s) to assist in the separation and to promote the ionization.196, 207  However, separation in those conditions can compromise spray stability 76  and signal intensity. In the present study, neutral HPC-capillaries and acidic MS-compatible BGE’s based on ammonium acetate, formic acid and acetic acid were evaluated for glycoprofiling neutral and sialylated glycans of human IgG. The BGE’s showed good spray stability and signal intensity but those BGE’s in which the primary electrolyte component was acetic acid or formic acid did not provide successful separation between neutral and acidic glycans. Improved separations were obtained with ammonium acetate as the BGE at pH 3.1. Concentration of ammonium acetate, containing different amounts of methanol, was varied in the range 20-100 mM to find the optimum concentration that would give acceptable resolution without compromising spray stability. A gradual increase in BGE concentration from 20 mM to 50 mM resulted in improved separation efficiency and resolution for the glycan mixture. However, at BGE concentrations higher than 50 mM, MS signals decreased as the BGE concentrations increased without improving the separation. The reason for this could be the higher background redox currents obtained at higher BGE concentrations that would negatively impact the spray stability and ESI response. Methanol content in the BGE was evaluated in the range 10-50%. Addition of methanol to the BGE was necessary to reduce the high CE currents that affect spray stability and to make the BGE similar in composition to that of the modifier solution. The current limit for our flow-through microvial interface is 15 μA.219 The best separation between neutral and acidic glycans was obtained with a BGE containing 50 mM ammonium acetate with 20% methanol at pH 3.1. The optimum BGE represents a compromise among resolution, spray stability and efficient analyte ionization. The optimum BGE contains only 20% methanol in order to maintain the speed and efficiency of the separation. Moreover, higher methanol concentrations were not necessary as the surface tension of the CE effluent was further reduced by the use of the modifier solution to improve ESI-MS detection. Finally the 77  separation voltage was also optimized, and a value of -25 kV was selected. In order to reduce analysis time for neutral glycans, an overimposed pressure of 10 mbar (0.145 psi) was used. Separation at baseline between neutral and sialylated N-glycans was achieved (Figure 3.1) and glycans were detected as [M-H]- and [M-2H]2- ion species. Sialylated glycans migrated faster than neutral glycans due to their electrophoretic mobility enhanced by the acidic BGE in reversed polarity mode.211 Among sialylated species, separation occurred based on the number of sialic acid residues and the hydrodynamic volume of the glycans. The overimposed pressure of 10 mbar (0.145 psi) and the interaction of neutral glycans with acetate ions allowed the glycans to migrate toward the anode. Although baseline separation in the time dimension was not achieved for all the glycan species, the orthogonal baseline MS resolution allows direct analysis of underivatized neutral and acidic glycans in a single run, providing a detailed glycoprofile of the complex glycoprotein. The presented methodology also eliminates the need to label the glycans, as well as the need to remove sialic acid residues prior to their analysis as it has been commonly reported for acidic glycans of different sources.100, 251  As outlined in Table 3-1, 22 glycans of IgG were identified. Putative glycan monosaccharide compositions were assigned based on accurate precursor mass data and previous reports on N-glycosylation analysis of human IgG.252-254 The number of glycans observed in IgG contrasts with the large heterogeneity found in other glycoproteins of the immune system.255 For instance, a single GPI-anchored glycoprotein, CD59,  has been reported to contain over 120 glycoforms.255   78    Figure 3.1 Density map for the CE-ESI-MS separation of neutral and acidic N-glycans from human IgG. Conditions: neutral (HPC)-coated capillary 70 cm X 50 µm; BGE: 50 mM ammonium acetate /20% methanol (pH 3.1); separation voltage: -25 kV + 10 mbar; negative ion ESI-MS; modifier solution: 10 mM ammonium acetate (pH 3.1) containing 75% of 2-propanol: methanol (2:1). The intensity of the color reflects the glycan abundance with light green being less abundant that dark blue.                  Sialylated IgG glycans Neutral IgG glycans79   Table 3-1 Identification and annotation of native N-glycans of human IgG observed with CE-MS. Glycans were observed as [M- 2H]2-  and [M- H]- species. Data was acquired using the maXisTM Ultra High-Resolution TOF MS.                Nomenclature of the glycan monosaccharide composition as indicated in Figure 3.2.   GlycanComposition Theor.  Mass (Da)Obs. Mass (Da) Err. (ppm)Observed m/zH N F S [M-2H]2-[M-H]-H5N4S2 5 4 0 22224.4701 2224.46880.58 1111.2264H5N5S1 5 5 0 12136.9011 2136.89970.66 1067.4419H5N4F1S1 5 4 1 12079.9317 2079.9326-0.43 1038.9583H5N4F1S2 5 4 1 22370.0102 2370.0115-0.55 1183.9978H4N4F1S1 4 4 1 11916.9718 1916.9729-0.57 957.4785H5N4S1 5 4 0 11933.3989 1933.39720.88 965.6906H5N5F1S2 5 5 1 22573.8456 2573.8478-0.85 1285.9159H5N5S2 5 5 0 22427.3621 2427.3642-0.87 1212.6741H5N5F1S1 5 5 1 12282.0945 2282.09280.74 1140.0384H4N5S1 4 5 0 11974.8396 1974.8409-0.66 986.4125H3N4F1 3 4 1 01464.4508 1464.4513-0.34 1463.4433H4N4F1 4 4 1 01626.5117 1626.51040.80 1625.5024H4N5F1 4 5 1 01829.5348 1829.5361-0.71 1828.5281H3N5F1 3 5 1 01667.4445 1667.4456-0.66 1666.4376H3N4 3 4 0 01318.3101 1318.3112-0.83 1317.3032H5N3 5 3 0 01439.4867 1439.48560.76 1438.4776H4N4 4 4 0 01480.6557 1480.65460.74 1479.6466H4N5 4 5 0 01683.5676 1683.56680.48 1682.5588H5N4F1 5 4 1 01788.3823 1788.3832-0.50 1787.3752H5N5 5 5 0 01845.6863 1845.6875-0.65 1844.6795H3N5 3 5 0 01521.5435 1521.5443-0.53 1520.5363H3N3F1 3 3 1 01261.4101 1261.4109-0.63 1260.402980  3.3.2 N-glycan analysis of rHuEPO using CE-MS Endogenous and recombinant EPOs, engineered in different cell cultures, contain the same 165-amino acid polypeptide chain with one O-linked oligosaccharide (Ser126) and three N-linked oligosaccharides (Asn24, -38, and -83).256, 257 Recombinant human EPO exists as a range of isoforms whose glycoform population is mainly composed of complex-type N-glycans with variable number of terminal sialic acids.231, 232, 258 Previous CE-MS studies on N-glycosylation of rHuEPO were performed using bare-fused silica capillaries in basic background electrolyte (BGE) conditions.196, 207 To enhance mobility differences of glycoforms and to avoid compromising spray stability or signal intensity that may occur in basic BGE, our CE-MS method employs neutral hydrophilic (HPC)-coated capillaries under acidic BGE conditions. N-glycan population from rHuEPO is mainly composed of core-fucosylated tetra-antennary glycans containing variable numbers of sialic acids. Tri-antennary N-glycans containing up to 3 sialic acid residues and di-antennary N-glycans containing up to 2 sialic acid residues are also present but with lower abundance than tetra-antennary structures.231, 232, 258 Because N-glycans from rHuEPO are highly sialylated, they remain negatively charged over a broad pH range and can be easily CE separated and detected in negative ion ESI-MS without derivatization or glycan labeling. Analyzing underivatized glycans is desirable not only to simplify the sample preparation but also to preserve their native structural characteristics. The most common derivatization strategy for CE analysis of N-glycans is reductive amination,121 in which the combination of acidic conditions and high temperatures present a considerable risk for loss of sialic acids which could contribute to an increased heterogeneity.259  Removal of the excess of labeling reagent can also be problematic and labor-intensive because of the similar properties of the labeling and the derivatized glycans. Therefore, the analysis of underivatized glycans is 81  preferred to reduce the risk of bias because of sample handling and the potential loss of sialic acids.   Figure 3.2 shows the extracted ion electropherogram (EIE) of the seventeen most abundant N-glycans from rHuEPO detected with a triple quadrupole MS. Separation of N-glycans is achieved based on different electrophoretic mobilities influenced by the number of sialic acids residues, N-acetyl-lactosamine (LacNAc) repeat units and the hydrodynamic volume of the glycans. Glycans with an increased number of sialic acids display a stronger mobility toward the anode and appear earlier in the electropherogram. As expected, heavier glycans containing the same number of negative charges have a reduced mobility and are detected later in the electropherogram. Smaller species containing variable amounts of sialic acids migrate faster than larger species according to the LacNAc extension of the chain. Although the detected glycans are not fully time resolved, sialylated glycans displaying the same charge but containing variable number of hexoses (Hex) and/or LacNAc units (161.05 or 364.1 Da, respectively) can be profiled and distinguished in the MS, showing the excellent separation power of the technique to resolve structurally-similar N-glycans. Some abundant glycans, such as tetra-antennary glycans with 3-4 sialic acid residues are detected as broad peaks because of the co-migration of their acetylated and/or NeuAc/NeuGc variations and because of the diffusion process that can occur under negligible EOF conditions. Band broadening properties in CE-MS have been recently discussed in a publication from our research group.217   82    Figure 3.2 EIE for the CE-ESI-MS separation of N-glycans from rHuEPO. Only the seventeen most abundant N-glycans are displayed. The glycans were detected under the same CE-MS conditions as in Figure 3.1. The monosaccharide composition of the glycans is indicated in the graph. Nomenclature is as follows: H = mannose or galactose (Hex), N = N-acetylglucosamine (HexNAc), F = fucose, and S = sialic acid (Neu5Ac).   Putative monosaccharide compositions, observed masses and m/z values for glycans shown in Figure 3.2 are summarized in Table 3-2. The precursor mass data obtained from the triple quadrupole MS were used to search for matching compounds in the Consortium of Functional Glycomics (CFG) for N-linked glycans. In agreement with other reports, all released glycans of rHuEPO contain a fucosylated core with variable degrees of sialylation257, 260, 261  Using the developed method we detected the same population of glycans from rHuEPO expressed by CHO cells previously reported.196, 256, 257, 260, 262  We also identified additional species that were not reported before: two tetra-antennary glycans (mono- and di-sialylated) and one mono-sialylated tri-antennary glycan, with the monosaccharide compositions H8N7F1S1, H9N8F1S2 and H6N5F1S1, respectively. Glycans with acetylation and/or Neu5Ac/Neu5Gc variations with putative compositions H8N7F1S1 and H9N8F1S2 were also observed (Table 3-83  3) Assignment of glycan monosaccharide composition could be based on the observed precursor mass alone because of the restricted combinations of monosaccharide residues of N-glycans.263 For each glycan monosaccharide composition there are several possible isomeric structures and different types of glycosidic linkages that cannot be identified by single stage CE-MS analysis. Differences in electrophoretic mobilities, resulting in differential migration time for glycans, can be useful to predict the structure of certain glycans. For instance, the monosaccharide composition of glycan H8N7F1S3 (Table 3-2) matches both with the structure of a tri-antennary glycan with two LacNAc units and with the structure of a tetra-antennary glycan with one LacNAc unit. The longer migration time of glycan H8N7F1S3 in comparison to that of glycan H7N6F1S3, a tetra-antennary species,196 suggests that glycan H8N7F1S3 may be a tetra-antennary species with one LacNAc repeat. The increased volume of glycan H8N7F1S3 in comparison to glycan H7N6F1S3, reduces its electrophoretic mobility and appears later in the electropherogram.                   84  Table 3-2 Identification and annotation of the seventeen most abundant N-glycan monosaccharide compositions of rHuEPO shown in Figure 3.2. Glycans were observed as [M-2H]2-, [M-3H]3- and/or [M-4H]4- ion species. Data was obtained using a triple quadrupole MS.     Nomenclature of the glycan monosaccharide composition as indicated in Figure 3.2.  To accurately determine the mass of the glycans, our flow-through microvial interface was connected to a maXisTM ultra high-resolution TOF MS which offers a wider mass range, higher resolution and higher mass accuracy than the triple quadrupole MS. In addition to the glycans detected before, a mono-sialylated tetra-antennary glycan with three LacNAc units (monosaccharide composition H10N9F1S1) that has not been previously reported by CE-MS was detected with the TOF MS. In total, more than 70 N-glycans were observed of which several could be identified, as shown in Appendix B.1 (Table B.1). A complete list of the glycans observed for BRP rHuEPO is shown in Appendix B.1 (Table B.1), including monosaccharide GlycanComposition Theor. Mass (Da)Obs. Mass (Da)Observed m/zH N F S [M-2H]2-[M-3H]3-[M-4H]4-H7N6F1S4 7 6 1 4 3683.33 3683.21 1226.7 919.7H8N7F1S4 8 7 1 4 4048.67 4048.62 1348.5 1011.1H9N8F1S4 9 8 1 4 4415.07 4415.10 1470.6 1102.7H6N5F1S3 6 5 1 3 3026.74 3026.76 1512.3 1007.8H10N9F1S3 10 9 1 3 4490.10 4489.95 1495.6H7N6F1S3 7 6 1 3 3393.10 3392.97 1695.4 1129.9H6N5F1S1 6 5 1 1 2444.22 2444.23 1221.0H9N8F1S2 9 8 1 2 3833.51 3833.47 1276.8H9N8F1S3 9 8 1 3 4124.80 4124.69 1373.8H8N7F1S3 8 7 1 3 3758.20 3757.90 1251.6H8N7F1S2 8 7 1 2 3467.49 3467.60 1154.8H6N5F1S2 6 5 1 2 2736.50 2736.46 1367.1H5N4F1S2 5 4 1 2 2371.17 2371.30 1184.6H8N7F1S1 8 7 1 1 3171.85 3171.79 1056.2H9N8F1S1 9 8 1 1 3539.24 3539.40 1768.6 1178.7H10N9F1S4 10 9 1 4 4781.42 4781.37 1592.7H7N6F1S2 7 6 1 2 3101.84 3101.77 1549.8 1032.885  composition, theoretical mass, observed mass, and m/z values. Glycans were identified based on their true isotopic pattern provided by the ESI Compass 1.3 application software from the maXisTM TOF MS in combination with the information provided by the Consortium of Functional Glycomics (CFG) for N-linked glycans. The high mass accuracy, in the low ppm range (< 1 ppm) and the resolving power, of nearly 20 000 at m/z 1334 (Agilent ESI low concentration tuning mix) of the TOF MS was also useful to confirm glycan modifications previously observed with the triple quadrupole MS.   3.3.3 Detection of glycan modifications in rHuEPO 3.3.3.1 Acetylation of N-glycans In agreement with other reports, our CE-MS method revealed the presence of three types of secondary glycan modification in enzymatically released N-glycans from rHuEPO.196, 207, 228, 258, 264 These modifications were mainly observed in highly sialylated glycans and include acetylation of sialic acids, N-acetylneuraminic (Neu5Ac) / N-glycolylneuraminic acid (Neu5Gc) variation, and elongation of the glycan chains due to the presence of LacNAc repeats. Up to three modifications occurred per glycan. Acetylation of N-glycans was detected due to mass shifts of 42 Da from the naked glycan masses. The presence of a set of peaks differing in 10.5 m/z (Th) for glycans observed as [M-4H]4- ion species or 14 m/z (Th) for glycans observed as [M-3H]3- ion species, accounts for the acetylation of glycans. For instance, the presence of a set of seven peaks additional to the most abundant ion at m/z 919.53 for the tetra-sialylated glycan H7N6F1S4, indicates the occurrence of 7 acetylated forms that were resolved in the mass dimension (Figure 3.3A+ 3.3B). Acetylated species were predominantly detected as [M-4H]4- ion species, while [M-3H]3- ions species were 86  less abundant. Higher acetylated forms of glycan H7N6F1S4 have a reduced electrophoretic mobility compared to less acetylated species and are detected later in the electropherogram. This is illustrated in the extracted ion electropherogram (EIE) (Figure 3.3A), where the migration order of the glycans are in agreement with the number of acetyl groups. Our CE-MS method succeeded in detecting mobility differences due to only 42 Da, corresponding to acetylation, in high molecular mass analytes (2300-4800 Da, approximately) with an optimal MS resolution. Acetylation of other glycans such as the tri-sialylated H7N6F1S3 and the tetra-sialylated with one LacNAc repeat H8N7F1S4, containing 4 and 6 acetylated forms, respectively, were also observed. Figure 3.1C shows a set of five peaks corresponding to the separation achieved for the tetra-antennary tri-sialylated glycan H7N6F1S3, and four of its acetylated forms. The glycan and its acetylated forms were predominantly detected as [M-3H]3- ion species, where the most abundant ion at m/z 1129.35 corresponds to H7N6F1S3, as shown in the MS spectrum (Figure 3.1D). The same glycans were also detected as low intense peaks as [M-2H]2- ion species. Extensive acetylation of glycans was also observed in a tetra-antennary species containing one additional LacNAc unit H8N7F1S4 where six acetylated forms were observed, as shown in Figure 3.1E. The glycan and its acetylated species were detected as [M-4H]4- ions, where the at m/z 1011.06 corresponds to H8N7F1S4, as shown in the MS spectrum (Figure 3.1F). The same set of glycans was also detected as [M-3H]3- ion species, at lower intensity. The m/z values of the acetylated species of H8N7F1S4 differed in 10.5 m/z (Th) when the glycans were detected as [M-4H]4-ion species, and in 14 m/z (Th) when the glycans were detected as [M-3H]3- ion species. This demonstrates the occurrence of acetyl modifications corresponding to mass shifts of 42 Da from the non-acetylated glycan. All acetylated N-glycans observed in rHuEPO are included in Table B.1.  87  CE-MS analysis of sialylated glycans shows that a higher degree of acetylation is present in glycans containing a higher degree of sialylation. This indicates that the acetylation occurs at the sialic acid level due to the occurrence of Neu5,9Ac2, as it has been reported before in rHuEPO257 Based on the number of sialic acid residues and O-acetyl groups observed, it can be predicted that not only mono-O-acetylated sialic acid residues are present but also multiple O-acetylated forms of sialic acids occurs. For instance, the presence of 5, 6 and 7 O-acetyl substituents in tetra- and tri-sialylated glycans implies multiple O-acetylations occurring on a single sialic acid residue. This phenomenon has also been reported during the glycoprofile of rHuEPO,196, 207, 228 Dynepo,265 a novel recombinant human EPO, darbepoetin,261 an hyperglycosylated glycoform and other commercial EPOs produced in CHO cell lines.264 The degree of acetylation of sialic acid deserves special attention because it is correlated with the circulatory half-life of the glycoproteins in the human serum and its in vivo biological activities.265, 266 Also, variability in the glycosylation pattern can potentially affect the biological, physico-chemical and immunological properties of rHuEPO.267                   88  (A)     (B)      XIC of -Q1: 919.8 amu from Sample 8 (20120418 EPO Glycans 08) of 20120418 Glycans .wiff (Turbo Spray), Smoothed, Smoothed Max. 1.6e6 cps.20.7 20.8 20.9 21.0 21.1 21.2 21.3 21.4 21.5 21.6 21.7 21.8 21.9 22.0 22.1 22.2 22.3 22.4Time, min0.01.0e52.0e53.0e54.0e55.0e56.0e57.0e58.0e59.0e51.0e61.1e61.2e61.3e61.4e61.5e61.6e6Intensity, cps(1)(1a)(1b)(1c)(1d)(1e)(1f)(1g)Neu5Ac/Neu5GcNeu5Ac/Neu5GcNeu5Ac/Neu5GcNeu5Ac/Neu5Gc919.5379923.0365923.7875930.2911-MS, 17.7-18.7min #(1037-1095)0.00.51.01.52.0Intens.[%]918 920 922 924 926 928 930 932 m/z919.5379923.7875930.2911934.2784940.7919944.5465951.0456955.9399961.5470965.5600972.0496982.9639993.0448-MS, 17.7-18.7min #(1037-1095)0.00.511 52.0Intens.[%]910 920 930 940 950 960 970 980 990 m/zAcAcAcAcAcAcAcNeu5Ac/Neu5Gc89  (C)      (D)     XIC of -Q1: 1129.9 amu from Sample 1 (20120419 EPO Glycans 01) of 20120419 Glycans .wiff (Turbo Spray), Smoothed, Smoothed, S... Max. 6.5e6 cps.20.0 20.5 21.0 21.5 22.0 22.5 23.0 23.5 24.0 24.5 25.0Time, min0.05.0e51.0e61.5e62.0e62.5e63.0e63.5e64.0e64.5e65.0e65.5e66.0e66.5e6Intensity, cps21.88(6)(6a)(6b)(6c)(6d)1129.35611134.68721143.35901149.01361154.33321168.33511171.36291179.30971185.36241193.6424-MS, 20.1-22.0min #(1180-1290)0123456Inte s.[%]1130 1140 1150 1160 1170 1180 1190 m/zAcAcAcAcNeu5Ac/Neu5GcNeu5Ac/Neu5GcNeu5Ac/Neu5Gc1129.35611134.68721143.35901149.01361154.33321168.33511171.36291179.30971185.3624-MS, 20.1-22.0min #(1180-1290)0123456Intens.[%]1120 1130 1140 1150 1160 1170 1180 1190 m/z90  (E)    (F)    Figure 3.3 (A) EIE obtained for the CE-MS separation of the tetra-antennary tetra-sialic glycan H7N6F1S4 (peak 1) and its acetylated species containing from 1 to 7 O-acetyl groups (peaks 1a to 1g, respectively). (B) Average MS spectrum obtained for (A) where it can be observed the m/z values of the O-acetylated and Neu5Ac/Neu5Gc variations species of H7N6F1S4. (C) EIE obtained for the CE-MS separation of the tetra-XIC of -Q1: 1011.6 amu from Sample 1 (20120419 EPO Glycans 01) of 20120419 Glycans .wiff (Turbo Spray), Smoothed, Smoothed, S... Max. 4.4e6 cps.19.5 20.0 20.5 21.0 21.5 22.0 22.5 23.0 23.5 24.0 24.5 25.0 25.5 26.0Time, min0.02.0e54.0e56.0e58.0e51.0e61.2e61.4e61.6e61.8e62.0e62.2e62.4e62.6e62.8e63.0e63.2e63.4e63.6e63.8e64.0e64.2e64.4e6Intensity, cps21.49(2)(2a)(2b)(2c)(2d)(2e)(2f)1129.35611134.68721143.35901149.01361154.33321168.33511171.36291179.30971185.3624-MS, 20.1-22.0min #(1180-1290)0123456Intens.[%]1120 1130 1140 1150 1160 1170 1180 1190 m/z1007.64941011.06941014.81491021.32131026.06381031.82261036.05941042.57531046.82231053.07571063.3278-MS, 18.4-19.5min #(1080-1141)0.000.250.500.751.001.25Intens.[%]1010 1020 1030 1040 1050 1060 1070 m/zAcAcAcAcAcNeu5Ac/Neu5GcNeu5Ac/Neu5GcNeu5Ac/Neu5GcNeu5Ac/Neu5Gc91  antennary tri-sialic glycan H7N6F1S3 (peak 6) and its acetylated species containing from 1 to 4 O-acetyl groups (peaks 6a to 6d, respectively). (D) Average MS spectrum obtained for (C), where it can be observed the m/z values of the O-acetylated and Neu5Ac/Neu5Gc variant species of H7N6F1S3. (E) EIE obtained for the CE-MS separation the tetra-antennary tetra-sialic glycan containing one LacNAc unit H8N7F1S4 (peak 2) and its acetylated species containing from 1 to 6 O-acetyl groups (peaks 2a to 2e, respectively). (F) Average MS spectrum obtained for (E), where it can be observed the m/z values of the O-acetylated and species of H8N7F1S4. The glycans were detected under the same CE-MS conditions as in Figure 3.1.   3.3.3.2 Sialic acid variation rHuEPOs engineered in Chinese hamster ovary (CHO) cells have a distinct set of glycans relative to recombinant EPOs engineered in human fibrosarcoma cell lines and baby hamster kidney (BHK) cells, because of the slightly different glycosylation machineries including the sugar-transferring enzymes.268, 269  Contrary to human cells, CHO cells do not express sialyl-α-2-6-transferase, α-1-3/4-fucosyltransferase and bisecting N-acetylglucosamine transferase,260 but they can express alpha galactose residues that are highly immunogenic. However, these cells contain the enzyme responsible for N-glycolylneuraminic acid synthesis (CMP-N-acetylneuraminic acid hydroxylase [CMAH]) that is not present in humans due to an internal frame shift mutation in the CMAH human gene.270, 271 For that reason, rHuEPO expressed in CHO cell has been reported to contain between 1-1.5% of Neu5Gc residues relative to the total sialic acid content, while EPO products engineered by gene-activation in human cells contain no Neu5Gc.264, 269, 272  Our CE-MS method detected the presence of NeuGc, instead of Neu5Ac, in some N-glycans from rHuEPO. In some cases, this sialic acid variation occurred simultaneously with the acetylation of glycans previously discussed (Figure 3.3B). The presence of a set of peaks differing in 4 m/z (Th) for glycans observed as [M-4H]4- ion species, correspond to mass shift of 16 Da from the unmodified glycan masses, and accounts for the Neu5Ac/Neu5Gc variation. For instance, the tetra-sialylated glycan H7N6F1S4 and four of its acetylated forms, containing one 92  to four acetyl groups, displayed lower intense peaks at 4 m/z (Th) higher than the corresponding acetylated glycan. This type of modification was also observed for other glycans. For instance, a set of peaks with a constant difference of 4 m/z (Th) was observed for the tetra-antennary tetra-sialic glycan H8N7F1S4, as shown in the MS spectrum (Figure 3.3E). In Figure 3.3E the ion at m/z 1011.06 corresponds to H8N7F1S4 and the lower intense ions at 4 m/z (Th) higher, corresponds to the Neu5Ac/Neu5Gc variations. The same type of modification was observed for the acetylated forms of the tetra-antennary tri-sialic glycan H7N6F1S3, in which only glycans containing up to two acetyl groups showed the Neu5Gc/Neu5Ac variation. For these species, low intense peaks were observed at 5.3 m/z (Th) higher than the corresponding predominant ions at 1129.35 m/z and were detected as [M-3H]3- species (Figure 3.3D). Other glycans where the Neu5Ac/Neu5Gc variation also occurred are shown in Table B.1, along with their corresponding putative monosaccharide composition. This phenomenon has not been extensively reported and there are only a few studies in the literature dealing with the presence of Neu5Gc in rHuEPO.196, 258, 264, 265, 272 Even though Neu5Gc is a common sialic acid in most mammals, it has been demonstrated to be absent in healthy humans and only small amounts have been found in some tumors and meconium.273  The CE-MS method succeeded in detecting this heterogeneity of acidic monosaccharides without removing the sialic acid from the glycan chain by acid hydrolysis as it has been commonly reported.264, 265, 274 As reviewed lately, the LC-MS detection of sialic acids and its heterogeneous subtypes is challenging.117 Mass spectrometric results indicate that the variation occurred at the glycan level as shown in Figures 3. 3B, 3.3D and 3.3F.   93  3.3.3.3 Elongation of the N-glycan chain The third type of glycan modification corresponds to the elongation of the glycan chain due to the presence of LacNAc repeat units. For instance, tetra-antennary glycans, the most abundant oligosaccharides present in rHuEPO, were observed to contain up to three LacNAc extensions. Glycans with monosaccharide compositions H8N7F1S4, H8N7F1S2; H9N8F1S4, H9N8F1S3; H10N9F1S3 and H10N9F1S4 were identified to carry one, two and three units of LacNAc, respectively which are in agreement with other reports in the literature.196, 257, 260, 262 These extended oligosaccharides represent the most important difference between circulating human EPO and rHuEPO, and they have been reported as a critical structural feature of rHuEPO that should be considered for its biological significance.275 Recently, Bones et al.228 have reported an increased number of LacNAc extensions containing up to five units in BRP rHuEPO. In that study, successful detection of tetra-antennary glycans with extensive LacNAc repeats was based on a combination of weak anion exchange chromatography (AEC) and HILIC that allowed for the first time the detection of up to five LacNAc repeats.228   3.3.4 LC-MS/MS analysis of N-glycans from rHuEPO Underivatized, non-reduced N-glycans of rHuEPO were also analyzed using online porous graphitized carbon (PGC) LC-MS/MS to compare its ability to glycoprofile complicated glycan mixtures with that of CE-MS. Table B.2 of Appendix B.2 shows a list of the glycans detected with PGC LC-MS/MS. A total of 27 glycan monosaccharide compositions including a few acetylated species were detected. Glycan heterogeneity such as Neu5Ac/Neu5Gc variations and extensive acetylation of tri- and tetra-sialylated glycans H7N6F1S3, H7N6F1S4 and H8N7F1S4, observed with our CE-MS method, were not detected with PGC LC-MS. The tetra-94  sialylated glycan containing one LacNAc unit H8N7F1S4 was not detected by PGC LC-MS either.  Successful LC-MS/MS analysis of glycans using PGC often requires oligosaccharide reduction in the presence of mild reducing agents or derivatization with hydrophobic compounds prior to MS detection.227, 276, 277 However, those additional steps can eliminate subtle modifications including O-acetylation present in glycans (unpublished observation). Therefore, the analysis of native species is the preferred method of choice to preserve the structural characteristics of these complex analytes. Although both chromatographic and electrophoretic separation techniques are suitable for the profiling of glycans, CE is advantageous due to its demonstrated capability to analyze underivatized glycans and offers an alternative and complementary analytical tool to chromatographic based techniques. Our CE methodology has the advantage of keeping the detailed structural information about the glycans, which results in the detection of more species, not only because of the type of MS employed, but also because of the unique characteristics of the technique. Moreover, CE-MS requires short analysis time for the study of intact N-glycans, in comparison to LC-MS, which requires longer times for separation and column re-equilibration. Taken together, this makes CE a well-suited analytical approach for extensive glycan characterization.   3.4 CONCLUDING REMARKS  In this study, we demonstrated the excellent potential of CE-MS for glycan analysis using a flow-through microvial interface. With our CE-MS method, we were able to characterize heterogeneous mixtures of enzymatically released N-glycans from human IgG and rHuEPO. Glycans released from IgG were used to optimize the conditions for CE-MS analysis of complex 95  N-glycan mixtures. In total, 22 N-glycans, including neutral and sialylated species of IgG were identified. The use of hydrophilic HPC-coated capillaries that provide almost zero EOF was necessary to allow successful separation of the N-glycan mixtures. Underivatized glycans were used, because the labeling procedure often affects the detection of glycan modifications by inducing, for example, de-O-acetylation or complete destruction of chemically unstable substituted sialic acids. Also, it was not necessary to remove sialic acid residues. Sialylated and neutral N-glycans from IgG were simultaneously glycoprofiled to provide a general overview of the complex glycan mixture, although it was not possible to achieve high resolution for both neutral and sialylated species.   Application of the optimum CE-MS conditions for the analysis of rHuEPO glycosylation, revealed the presence of more than 70 N-glycans, including O-acetylation, Neu5Ac/Neu5Gc heterogeneity and extension of the glycan chains due to LacNAc repeats. The CE-MS method made it possible to unravel the complexity of the carbohydrates and showed detailed information about their minor glycan modifications. This revealed the ability of the method to glycoprofile glycans based on the degree of sialylation and the patterns of their substituents. The ultrahigh resolution TOF MS confirmed the presence of modified glycan structures, and putative glycan monosaccharide compositions were determined by using the high resolution and high mass accuracy of the TOF MS. Although, the method has several advantages for the glycosylation characterization of rHuEPO, in terms of the simple instrumentation, small sample consumption, relatively short analysis time and the feasibility of coupling CE with different types of MS instruments, its major limitation is probably the limited sensitivity. Developments in the field of online sample pre-concentration in CE are an active area of research, and have been recently reviewed.278 However, in the context of therapeutic glycoprotein characterization, highly 96  concentrated pharmaceuticals are commonly available. Therefore, the presented CE-MS method can be regarded as an orthogonal and complimentary approach for glycoprofiling protein pharmaceuticals.        In addition, PGC LC-MS/MS confirmed the presence of sialylated glycan structures containing variable degrees of acetylation. However, other glycan modifications were not detected with PGC LC-MS/MS which indicates the potential advantages of CE-MS over LC-MS for unraveling the structural complexity of glycans released from glycoproteins. Quality control of recombinant therapeutic proteins requires determining the glycosylation profile and the occurrence of secondary glycan modifications that may shed light on additional biological properties that affect both pharmacodynamics and pharmacokinetics. In combination with the high quality of MS data, an improved insight into the glycan heterogeneity of rHuEPO was possible. Our CE-MS analysis can be considered as a promising alternative to assess the glycosylation profile of therapeutic drugs and to evaluate the quality of biosimilar products. 97  Chapter 4: Capillary electrophoresis mass spectrometry characterization of O-acetylated N-glycans from fish serum   4.1 INTRODUCTION  The family of sialic acids includes over 40 naturally occurring, nine-carbon carboxylated sugars frequently found as the terminal units of glycoproteins and glycolipids.93, 279 The two most dominant species are N-acetylneuraminic acid (Neu5Ac) and N-glycolylneuraminic acid (Neu5Gc). Neu5Ac is the most diversely substituted sugar known in nature,280, 281 as shown in Figure 4.1. One of the better studied modifications of Neu5Ac is the addition of O-acetyl esters to hydroxyl groups at the 4, 7, 8, and/or 9 positions, giving rise to a great variety of possible compounds and isomers.281-283 O-acetylation of Neu5Ac is highly tissue- and species-specific, and is perhaps the most common modification found on sialic acids due to its involvement in a growing number of biological and pathophysiological phenomena.281, 284 Typical examples include the partial or complete block of neuramidase action,285, 286 bacterial antigenicity,287 binding of influenza C viruses to cells,288, 289 modulation of the alternative pathway of complement activation290 and cell-cell interactions.281, 285 The most abundant sources of O-acetylated Neu5Ac are bovine submandibular gland mucin and human colon mucosa.291  O-acetylation of Neu5Ac has also been identified on certain salivary mucins,283 neural gangliosides,292 rat liver membranes,293 cysteine proteinase inhibitors from Atlantic salmon skin,294 and kininogens from spotted wolffish and Atlantic cod.295  Recently, Liu et al.296 investigated the change of O-acetylated sialic acids in complex N-glycans in the sera of Atlantic 98  salmon (Salmo salar) as a response to varying periods of long-term handling stress. However, structural details were difficult to elucidate because of the possible presence of isomers.296    Figure 4.1 Sialic acid structure and common natural modifications that occur in the “R” groups. At position R2, Sia stands for sialic acid. 297  Reprinted from Glycobiology 19, J. Du, M. A. Meledeo, Z. Wang, H. S Khanna, V. D. P. Paruchuri, and K. J. Yarema. Metabolic glycoengineering: Sialic acid and beyond, 12, Copyright (2009), with permission from Oxford Journals.  Most commonly used analytical methods require releasing the sialic acids from glycoconjugates by enzymatic or chemical hydrolysis, followed by a combination of reduction and fluorescence labeling with 1,2 diamino-4,5-methylenedioxybenzene (DMB), and subsequent separation by reverse-phase HPLC (RP-HPLC) with fluorescence or mass spectrometric detection. 100, 251 However, serious limitations have been encountered when O-acetylated sialic acids are analyzed. Problems associated include: incomplete release of O-acetylated sialic acids from the glycosidic linkages, poor derivatization of some species, de-O-acetylation and migration of O-acetyl groups, and complete destruction of chemically unstable substituted sialic 99  acids during isolation and purification of the glycoconjugates. Furthermore, even if O-acetylated sialic acids could survive the isolation and purification steps, the linkage information to define the position of each substituent in the glycan chain is lost.289, 291, 298  Therefore, new methods that allow the detection of O-acetylated sialic acids without the liberation of glycans are of great interest. In recent years, profiling of intact glycans containing O-acetyl sialic acids has also become increasingly important for complete characterization of acidic glycoproteins. However, due to the structural variability and lability of O-acetylated N-glycans at the glycosylation sites of glycoconjugates, their analysis remains a challenging task.  Capillary electrophoresis (CE) has shown superior capability in separating positional and linkage isomers.299, 300 Typically, N-glycans enzymatically released from glycoproteins using Peptide-N-Glycosidase F (PNGase F) are labeled and characterized by CE with laser-induced fluorescence (LIF) detection.121, 301 A frequently used labeling reagent is 8-aminopyrene-1,3,6-trisulfonate (APTS) due to its high fluorescent yield and its ability to remain ionized over a wide pH range.24, 302-304 Although CE-LIF allows glyco-profiling, identification of new glycans is not straightforward. Combining the resolving power of CE with the high sensitivity and information-rich MS detection could be a powerful tool for glycan analysis. Isomer identification and structure determination for glycans can be achieved simultaneously by CE-ESI-MS/MS. Because O-acetylated N-glycans from fish serum are highly sialylated and remain negatively charged over a broad pH range, they can be CE separated and MS detected in negative ESI mode without the need for derivatization. There are only a few reports in the literature dealing with the CE-MS of intact glycans containing O-acetylated sialic acids.296, 305 However, comprehensive glycan analyses down to the level of isomeric differentiation for the characterization of native oligosaccharides have not been achieved.  In this study, a combined 100  approach for the analysis of APTS-derivatized and underivatized O-acetylated N-glycans enzymatically released from fish serum of Atlantic salmon (Salmo salar) is developed. CE-LIF was used for the separation of APTS-labeled O-acetylated N-glycans under acidic conditions. The results were compared with those obtained using the flow-through microvial interface for CE-ESI-MS of the derivatized and underivatized glycans for on-line mass and composition determination. Furthermore, CE-ESI-MS/MS was used to confirm the O-acetylation of sialic acids of isomeric glycans that were baseline separated with our methodology.  4.2 MATERIALS AND METHODS 4.2.1 Materials  A carbohydrate analysis kit, including 8-aminopyrene-1,3,6-trisulfonate (APTS) labeling dye and glucose ladder standard, was obtained from Beckman Coulter Inc. (Brea, CA).  Sodium cyanoborohydride (1 M in tetrahydrofuran) and ε-Aminocaproic acid were purchased from Sigma Aldrich (St. Louis MO, USA).  All chemicals and solvents were of analytical grade or better. Formic acid, acetic acid, ammonium acetate, methanol, ethanol, acetone and 2-propanol were purchased from Fischer Scientific (Nepean, ON, Canada).  All aqueous solutions were prepared using purified water (18.2 MΩcm-1) with a Mili-Q purification system (Millipore, Bedford, MA, USA).  4.2.2 Sample preparation  N-glycans present in serum of juvenile Atlantic salmon (Salmo salar) were obtained from two groups of fish: control and stress, according to the methodology described by Liu et al, 296 and were provided to us as such. According to Liu et al, 296 fish in the stress group were subjected to daily long-term handling stress (15 s out of water) for a period of 1 to  4 weeks. 101  Dried glycan samples were stored at -20°C. The dried samples were resuspended in 100 uL of purified water before labeling with APTS.   4.2.3 APTS labeling  N-glycans of Atlantic salmon from the control group of fish sera were derivatized with APTS by reductive amination via Schiff base formation according to the manufacturer’s instructions for the glucose ladder. Hereto, 50 uL of fish serum glycan solution was dried down in a SpeedVac vacuum centrifuge (Eppendorf, Germany). To the dried glycan, 2 µL of 0.2 M APTS diluted in 15% acetic acid and 2 µL of 1 M sodium cyanoborohydride/ THF were added. The resulting solution was vortexed and incubated overnight at room temperature in the dark.  Following incubation, the excess APTS reagent was removed by precipitating the labeled glycans with ice-cold acetone and centrifuging at 12000 rpm for 15 min.  After removing the supernatant, the APTS-labeled glycan pellet was washed with additional ice-cold acetone and reconstituted in 100 µL of distilled water.  This sample solution was used directly for all CE-LIF and CE-MS analysis.  4.2.4 Capillary electrophoresis  Capillary electrophoretic analyses of the APTS-labeled N-glycans were carried out with a Beckman Coulter P/ACE MDQ System (Beckman Coulter Inc., Fullerton, CA) equipped with a laser-induced fluorescence detection module. The Argon Ion laser was operating at 488 nm for excitation, and emission was monitored at 520 nm, which provides optimal sensitivity for APTS-labeled N-glycans. Separations were performed using neutral coated capillary with polyacrylamide-based hydrophilic wall coating (50 µm ID, 365 µm OD,  65 or 50.2 cm long) obtained from Beckman Coulter Inc. (Brea, CA). Separations were carried out in reverse-polarity 102  mode. Background electrolytes were composed of 0.2-2.0% formic acid or acetic acid, 10-50 mM ammonium acetate, ammonium formate or 25-50 mM ε-Aminocaproic acid. The APTS-labeled N-glycans were injected for 10 s at a pressure of 1 psi, which corresponds to a volume of 24 nL.  4.2.5 Mass spectrometry   CE-MS analysis were performed using a Finnigan LCQ* Duo ion trap mass spectrometer (Thermo Scientific, Waltham MA) operating in the negative ion mode. The CE-ESI-MS interface with a flow-through micro vial was developed in our laboratory and has been described previously.26, 76, 218, 219, 243-245 Electrospray voltage was set at -3.5 kV, and the temperature of the heated capillary at the MS inlet was set at 200°C. The detector scan range was set at m/z 900-1300 Da. The trap injection time was set at 50 ms. The MS scanning parameters were optimized using the ‘Autotune’ function of the LCQ Xcaliber software (Thermo Fisher Scientific, Waltham MA) by continuous infusion of APTS-labeled glucose ladder standard. The modifier solution was delivered by a syringe pump (Harvard Apparatus, Holliston MA) at a flow rate of 0.3 µL/min and was composed of 10 mM ammonium acetate (pH 3.17) containing 75% of 2-propanol and methanol (2:1 ratio). The samples were injected at 1.0 psi for 10 s, corresponding to a total volume of 19 nL. The same interface setup and solutions were used for CE-ESI-MS/MS with Prince CE system (Prince Technologies, The Netherlands) coupled to an API 3000 triple quadrupole mass spectrometer (Applied Biosystems/Sciex, Concord, Canada).  Electrospray voltage was set at -3.5 kV. The MS/MS spectra were acquired with a dwell time of 5.0 ms per step of 1 Th; the 103  precursor ion was selected at low resolution (about 5 Th window) and the Q3 peak width was set to approximately 2.0 Th.   4.3 RESULTS AND DISCUSSION 4.3.1 CE-LIF  CE-LIF separation with MS-amenable buffers gave good separation of APTS-labeled glucose ladder standard sample, as shown in Appendix C.1 (Figure C.1), the ladder can also be analyzed by MS.219 However, the optimum buffer conditions for CE-LIF analysis of fish serum glycans was found to be different from that of the standard glucose sample. Figure 4.2A shows the CE-LIF electropherogram for the separation of APTS-labeled N-glycans of fish serum using 25 mM ammonium acetate buffer (pH 3.1) which gave an acceptable current of 12 µA at -20 kV. According to according to Liu et al.,296  the glyco-profile of control fish serum contains only three N-glycans corresponding to O-acetylated bi-antennary oligosaccharides. Interestingly, the glyco-profile in Fig. 4.2A shows five species with distinctive electrophoretic mobilities that can be partially separated. In order to verify their identity, CE-ESI-MS was performed on the APTS-labeled O-acetylated N-glycans used, as described below.           56 58 60 62 64 66 68020406080100  (2b) 60.911153.07(3b) 61.621111.10(4b) 62.111132.02(5b) 62.631153.04Relative Abundance (%)Time (min)(B)(1b) 60.50  1131.812.5 13.0 13.5 14.0 14.5 15.0 15.5 16.00.02.0x1064.0x1066.0x1068.0x1061. x107Time (min)RFU    (1a)(2a)(3a)(4a)(5a)(A)104      4.3.2 CE-MS of native N-glycans of fish serum Attempts of using CE-MS for APTS-labeled N-glycans of control fish serum did not produce optimum results because the CE current under CE-LIF conditions was too high, as discussed in Appendix C.2. Therefore, we evaluated the CE-MS response of underivatized, native O-acetylated N-glycans present in 4 samples of control fish serum, relying on the dissociation of sialic acids that would provide the charges needed for electrophoretic migration and ESI ionization. Moreover, this approach was useful for preserving the structural characteristics of the natural low-abundance O-acetylated N-glycans.   At pH 3.03 in the negative ion mode, higher ionization efficiencies were observed in comparison to the positive ion mode (data not shown). The separation was carried out under reverse polarity, and worked for all the N-glycans released from fish serum which migrated under electrophoretic conditions according to the extension of chain length. For example, in Figure 4.2B, a set of ions detected as [M-2H]2- ion species were observed at m/z 1131.8 (1132.02), 1153.07 (1153.04) and 1111.10,  which are in agreement with samples of control fish sera according to Liu et al.296  Clearly, the presence of a set of doubly charged peaks differing by 42 Da, due to the presence of different numbers of acetyl groups in the sialic acids, can be observed. Therefore, the method allows the detection of mobility differences due to only 42 Da in high molecular mass analytes (2000-3200 Da).  As expected, heavier glycans bearing the same negative charge showed a lower mobility and are detected later than smaller glycans. This is       Figure 4.2 (A) CE-LIF separation of APTS-labeled N-glycans of control fish serum. Peaks (1a), (2a), (3a), (4a), and (5a) are APTS-labeled glycan peaks. (B) Base peak electropherogram of underivatized O-acetylated N-glycans of control fish serum acquired using the LCQ* Duo ion trap MS. BGE was 40 mM ε-Aminocaproic acid with 20% methanol. Peaks (1b), (2b), (3b), (4b), and (5b) correspond to the peaks (1a), (2a), (3a), (4a) and (5a), respectively in (A). The migration time (above) and the observed m/z (below) are indicated for each peak. 105  illustrated by the migration order of the fish serum glycans containing zero, one and two acetyl groups (at m/z 1111.10, 1132.02 and 1153.04, respectively) as can be observed for peaks (3b), (4b) and (5b) in Figure 4.2B. Earlier migration of N-glycans at m/z 1131.8 and 1153.07 is discussed in the following paragraph. Although some degree of overlapping is observed in the first dimension corresponding to CE separation, glycans can be resolved in the second dimension corresponding to MS detection, showing the excellent separation power of CE-MS.  Interestingly, two pairs of peaks corresponding to bi-antennary glycans containing one and two O-acetyl groups (at m/z 1132.0 and 1153.0 respectively) were baseline separated and provided evidence of structurally different isomeric species present in N-glycans of fish serum. The observed isomers: peaks (1b)/(4b) and (2b)/(5b) in Figure 4.2B, could be due to positional and/or linkage isomers that have different electrophoretic mobilities and therefore migrate with different velocities. The baseline separation achieved for the isomers, as can be observed in the base peak electropherograms in Figures 4.3A and 4.3B, demonstrates in more detail the separation power of the CE-MS method. The major reason for an improved selectivity of the native isomeric species could be the use of ε-Aminocaproic acid as buffer. Isomeric species may have distinctively selective interactions with the BGE, resulting in different hydrodynamic sizes and consequently dissimilar electrophoretic mobilities, yielding a baseline separation. Liu et al.296, has reported that the major ion at m/z 1153.0, corresponds to a bi-antennary complex glycan with two O-acetyl groups at its terminal sialic acid residues. However, they were not able to determine if two mono-O-acetylated sialic acids or a di-O-acetylated sialic acid were present as the terminal units in the complex glycan.  Moreover, only one glycan species was reported at m/z 1132.0.296 In order to identify the isomeric species already separated, CE-MS/MS analyses 106  were performed using the API 3000 triple quadrupole MS based on the adequate separation selectivity achieved.   (A)             -MS2 (1153.00) CE (-60): 37.407 to 38.210 min from Sample 1 (FSG 15 Week 0 MS/MS 1153.0) of DJLRJ110818B07.wiff (Ion Spray)... Max. 1001.0 cps.200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400m/z, amu120140160180200220240260280300320340360380400420440Intensity, cps373.5290.0% Intensity(a)Neu5Ac290.0Neu5,7(8), 9 Ac3373.5  -MS2 (1153.00) CE (-60): 34.499 to 35.803 min from Sample 1 (FSG 15 Week 0 MS/MS 1153.0) of DJLRJ110818B07.wiff (Ion Spray)... Max. 2677.7 cps.200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400m/z, amu20406080100120140160180200220240260280300320340360380400420440Intensity, cps331.9290.0% Intensitym/z(b)Neu5Ac290.0Neu5,7(8,9)Ac2        331.950 55 60 65020406080100    (A)Relative Abundance (%)Time (min)   (2b)  60.911153.07   (5b)  62.631153.04Peak (2b)Peak (5b)107  (B)  Figure 4.3 Extracted ion electropherogram showing baseline separation for isomeric species at (A) m/z 1153.04 and (B) m/z 1132.02. CE-MS spectra was acquired using the LCQ* Duo ion trap MS. In all cases, [M-2H]2- was observed. The inserts show the CE-MS/MS spectra for the isomeric species acquired using the API 3000 triple quadrupole MS: Fragment ions from precursor ions at (a) m/z 1153.07 peak (2b), (b) m/z 1153.04 peak (5b), (c) m/z 1131.8 peak (1b) and (d) m/z 1132. 02 peak (4b) were observed.    It is important to note that CE-ESI-MS detection of native N-glycans and CE-LIF detection of APTS-labeled N-glycans yielded the same glyco-profile. The five migrating species could be clearly seen in both the LIF and MS base peak electropherograms. In order to establish a correspondence between peaks detected with both methods, a quantitative assessment of the total peak area was made for the LIF trace (Figure 4.2A) and the MS base peak electropherogram (Figure 4.2B). In both cases, the percentage of a specific oligosaccharide present in the mixture was calculated from the area of the peak of that glycan relative to the total area of the 5 glycans, which is defined as 100%. The approximate percent peak area of the total glycans for peaks (1a-5a) in Figure 4.2A is as follows: peak (1a) = 2.87%, peak (2a) = 2.86%, peak (3a) = 13.79%, 50 55 60 65020406080100    (4b)  62.111132.02   (1b)  60.50 1131.8Time (min)Relative Abundance (%)   (B)  -MS2 (1132.00) CE (-70): 20.258 to 20.559 min from Sample 1 (FSG 15 week 0 MS MS 1132) of DJLRJ110829B13.wiff (Ion Spray), C... Max. 3055.1 cps.200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400m/z, amu20040060080010001200140016001800200022002400260028003000Intensity, cps331.8289.7271.51.0(c)% IntensityNeu5Ac289.7Neu5,7(8,9) Ac2331.8Neu5Ac–H2O271.5  -MS2 (1132.00) CE (-70): 19.423 to 20.793 min from Sample 1 (FSG 15 week 0 MS MS 1132) of DJLRJ110829B13.wiff (Ion Spray), C... Max. 4493.5 cps.200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400m/z, amu200460081021182022462830233840244Intensity, cps290.0331.9271.8408.0276.0243.7 262.0 390.0382.7228.4 286.0328.0 363.0(d)m/z% IntensityNeu5,7(8,9) Ac2331.9Neu5Ac290.0Neu5Ac –H2O271.8Peak (4b)Peak (1b)108  peak (4a) = 37.12%, peak (5a) = 43.33%, while for peaks (1b-5b) in Figure 4.2B is: peak (1b) = 1.93%, peak (2b) = 2.95%, peak (3b) = 14.26%, peak (4b) =36.03%, peak (5b) = 43.02%. The correlation shown between the relative peak areas gives us confidence to compare the peaks from the LIF trace and MS base peak electropherograms.  Reproducibility of the CE-ESI-MS method was also evaluated in terms of the migration times of the N-glycans present in 4 samples of control fish serum. The migration time % RSDs for the five components in three consecutive days with 5 injections on each day, were in the range of 0.15-0.31%. Of particular importance is that fact that under the ionization and detection conditions of the N-glycans, the in-source decay by desialylation is non-existent, which allows detection of native O-acetylated complex type oligosaccharides. These studies clearly demonstrate that the CE-ESI-MS approach is an excellent tool for the analysis of low-abundance acidic glycan species without the need of extra labeling steps.   4.3.3 Identification of isomeric O-acetylated species of native N-glycans   To further explore the identity of the isomeric species separated, native O-acetylated N-glycans were subjected to CE-ESI-MS/MS using the API 3000 triple quadrupole MS. Collision-induced dissociation was promoted by using a high sample cone voltage to produce CID of selected ions, using nitrogen as the neutral collision gas.224 The high voltage was experimentally determined by comparing the intensities of the fragment ions and a cone voltage of 75 V was selected as the optimum. The fragmentation was carried out using full MS in the range of m/z 200 to 400 to determine the extent of acetylation of sialic acids.  Insert (a) in Figure 4.3A shows the CE-ESI-MS/MS spectrum of precursor ion at m/z 1153.07. The fragment ions observed at m/z 290.0 and 373.5 correspond to terminal unmodified 109  Neu5Ac and di-O-acetylated Neu5Ac, respectively. Insert (b) in Figure 4.3A shows the CE-ESI-MS/MS spectrum of precursor ion at m/z 1153.04 with fragment ions at m/z 290.0 and 331.9, corresponding to terminal unmodified Neu5Ac and mono-O-acetylated Neu5Ac, respectively. In both cases, the fragments generated were produced by glycosidic cleavages of terminal Neu5Ac from the glycan antenna region. These fragmentation results confirmed that the bi-antennary complex type oligosaccharides, with two additional O-acetyl groups and nominal m/z 1153.0, were originated from two structural isomers containing two mono-O-acetylated Neu5Ac and another isomer with a Neu5Ac residue and a di-O-acetylated Neu5Ac. This observation is in agreement with those previously  proposed by  Liu et al.296, and, in addition provides conclusive evidence on the degree and distribution of O-acetyl groups on the sialic acids present as terminal units in mixtures of N-glycans from fish serum. CE-MS/MS analyses were also performed on precursor ions at m/z 1131.8 and 1132.0 corresponding to the mono-O-acetylated bi-antennary N-glycans of fish serum. In both cases the same fragments were obtained as observed in the inserts (c) and (d) of Figure 4.3B, which correspond to Neu5Ac and mono-O-acetylated Neu5Ac at m/z 290.0 and 331.8, respectively. A weak peak at m/z 271.8 was also observed and could be the result of a water loss adduct from the fragment ion at m/z 290.0.  According to previous studies in the literature, the major mono-O-acetylated sialic acid species identified was Neu5,9Ac2, while Neu5,7Ac2 and Neu5,8Ac2 were also present but in lower abundance leading to the occurrence of isomers for di-O-acetylated Neu5Ac.289, 293, 294, 298 The observed mono-O-acetylated Neu5Ac at nominal m/z 1132.0 could be comprised of a mixture of two possible sialic acid species, carrying O-acetyl groups in 9- and 7- or 8-OH of Neu5Ac leading to the occurrence of two isomeric species. However, at this stage we are not 110  able to identify the mono-O-acetylated isomers in CE-ESI-MS/MS electropherograms, shown in the inserts (c) and (d) of Figure 4.3B. We were not able to find studies in the literature where the location of O-acetyl groups in native, unreleased sialic acids were definitively reported.   4.3.4 Analysis of potential N-glycans isomers A further application of the CE-ESI-MS/MS method for the separation and identification of glycan isomers present in control fish serum was to study the potential occurrence of isomeric species in 10 samples of fish serum subjected to periods of stress from 1 to 4 weeks. Each sample was analyzed in three consecutive days with 3 injections per sample. In addition to the previously observed bi-antennary oligosaccharides with zero, one and two O-acetylated Neu5Ac, three and four O-acetyl groups in the terminal sialic acids were also detected. Moreover, tri-antennary structures ([M-3H]3-), containing from zero up to six O-acetylated Neu5Ac, were also observed. Table 4-1 shows the glycan composition and proposed structure for each of the N-glycans detected. A list of the bi- and tri-antennary glycans observed in fish serum samples, according to the week of the stress experiment, are shown in the Table C.5 (Appendix C.5).                     111  Table 4-1 Identification and annotation of the bi- (column on the left) and tri-antennary (column on the right) N-glycans present in fish serum observed as [M-2H]2- and [M-3H]3-, respectively. Composition and structural schemes are given in terms of N-acetylglucos amine (blue square), mannose (green circle), galactose (yellow circle) and sialic acid (purple diamond). For ease of identification, the numbers of O-acetyl groups present on sialic acids are included as subscripts in the N-glycan composition.      Figure 4.4 shows the base peak electropherograms for the separation of bi- and tri-antennary oligosaccharides present in samples taken from weeks 1 to 4 of the handling-stress experiment. Several peaks at the same m/z were identified and baseline separated, both for bi- and tri-antennary glycans. These peaks, which correspond to isomeric species, are possibly due to the distribution of O-acetyl groups in Neu5Ac, as discussed before. In general, three types of electropherograms were observed showing a distribution pattern, depending on the week of the stress experiment. The electropherograms of week 1 and week 4 samples show baseline Obs. m/z Glycan structure and composition  Obs. m/z Glycan structure and composition  1111.01132.01153.01175.01194.0959.73973.40987.131001.271015.501029.401043.40H5N4S2H5N4S2(1Ac)H5N4S2(2Ac)H5N4S2(3Ac)H5N4S2(4Ac)H6N5S3(1Ac)H6N5S3H6N5S3(2Ac)H6N5S3(3Ac)H6N5S3(4Ac)H6N5S3(5Ac)H6N5S3(6Ac)112  separation for isomers of bi- and tri-antennary oligosaccharides. The electropherogram of week 2 sample displays isomer separation for bi- and tri-antennary glycans, with an increased number of isomers for bi-antennary glycans. Week 3 sample only shows isomers corresponding to bi-antennary glycans, while isomers of tri-antennary species were not detected. Interestingly, the pattern for isomer distribution not only shows the occurrence of O-acetylated Neu5Ac in bi- and tri-antennary glycans but also follows the trend of the distribution of O-acetylated species reported by Liu et al.,296 i.e., week 2 shows a significant increase of O-acetylation in comparison to week 3, while week 1 and 4 have the same acetylation pattern. The distribution of O-acetylated Neu5Ac in bi-and tri-antennary glycans may lead to an increased number of isomeric species that shows a characteristic CE profile.   Of particular interest is the occurrence of three isomers for bi-antennary oligosaccharides in the week 2 samples, as seen in Figure 4.4B. To evaluate the possibility of obtaining structural information, CE-ESI-MS/MS was also performed under the optimum conditions. Comparison of the relative intensities of the fragment ions indicate that  the most intense peak of bi-antennary glycans at week 2 could correspond to Neu5,7,9Ac3 followed by Neu5,8,9Ac3 (Figure 4.4B), given the relative abundances discussed in the literature.283, 293 Also the 8-O-acetyl derivative, Neu5,7,8Ac3, regularly occurs, with lower abundance from the 7-O-acetyl migration.283, 293, 306 This potentially could correspond to the first peak of bi-antennary glycans at week 2, which explains the isomeric distribution of the acetylated species (see Figure 4.4B). Certainly, more effort needs to be put into the interpretation of MS/MS spectra; however, this is beyond the scope of the present study and requires the extensive characterization of individual standards of O-acetylated Neu5Ac (not commercially available) in order to compare fragmentation patterns and to assign a specific signature for each isomer. The potential of this approach to separate and 113  detect even less abundant components with high sensitivity—previously inaccessible due to overlapping of isobaric structures—can be considered as a significant contribution in the progress of detailed glycan analysis.  Figure 4.4 Base peak electropherograms showing separation of isomeric species for bi- and tri- antennary oligosaccharides present in samples taken from (A) week 1 and 4, (B) week 2 and (C) week 3 of the handling-stress experiment. Data was acquired using the API 3000 triple quadrupole MS.  4.4 CONCLUDING REMARKS A simple and effective separation method using CE-MS was developed for the analysis of enzymatically released O-acetylated N-glycans in Atlantic salmon (Salmo salar) fish serum. Characteristic CE glyco-profiles for both control and stress serum glycans were obtained. The CE method allowed unraveling the complexity of the glycans showing detailed information about the degree of O-acetylation present in the sialic acids. The use of neutral coated capillaries with polyacrylamide-based hydrophilic wall coating along with an acidic zwitterionic buffer, allows successful CE separations with ESI-MS and ESI-MS/MS detection of mixtures of complex N-glycans containing minor modifications. Although comparable CE-LIF and CE-ESI-MS glyco-profiles for APTS-derivatized and native glycans were obtained, it has been demonstrated that there is no need for derivatization or extra sample preparation when CE-ESI-MS is used for the analysis of glycans containing sialic acids. However, since derivatization with (A) Type I: week 1 and 4 50 55 60 65020406080100   Tri-antennary    glycansTime (min)    Bi-antennary glycans50 55 60 65020406080100   Time (min)Tri- tennary      glycanBi-antennary glycans50 5 60 65020406080100Time (min)    Bi-ante nary gl cans Tri-antennary     glycansRelative Abundance (%)(B) Type II: week 2 (C) Type III: week 3114  APTS is a relatively simple procedure, it can be used as a starting point for CE-profiling of glycan mixtures containing neutral and charged species. Clearly, the advantages of using mass spectrometry include valuable direct molecular weight and structural information for glycan identification with the potential of quantitative evaluation for glycan pools.   The CE-ESI-MS method developed was used to obtain baseline separation of isomeric species and can be a new tool for glycomic analysis. Moreover, CE-MS/MS provides an isomer-specific profile of O-acetylated Neu5Ac in fish serum glycans to unequivocally determine glycan structure for different degrees of O-acetylation. Isomer-specific analysis revealed up to three different isomers for di-sialylated bi- and tri-antennary oligosaccharides with highly reproducible migration times that have not been previously observed. In particular, it was shown that these structural isomers can be distinguished and assigned by CE-ESI-MS/MS. Further identification of linkage isomers will potentially require strategic glycosidase digestion but was not possible at this stage. However, we believe that the approach discussed in this study could be extended to other types of complex oligosaccharides and probably to all kinds of N-glycans.       115  Chapter 5: Global human serum N-glycan profiling by capillary electrophoresis-mass spectrometry  5.1 INTRODUCTION Over the last few years the advent of high-throughput analytical technologies for global glycan profiling of human sera has enabled researchers to pursue the role of aberrant glycosylation as potentially promising markers for several diseases.114, 158, 307  Glycans in serum are of particular interest because over 50% of human proteins are glycosylated.92 Since the biosynthesis of glycans involves competition among a set of glycosyltransferases, the glycans produced possess a vast structural heterogeneity. Therefore, monitoring the changes in serum glycosylation is a promising strategy to identify glycan biomarkers related to cell oncogenic transformation, acute phase condition and chronic diseases.94, 308 Because protein glycosylation is highly sensitive to the cell surroundings, alterations in N-glycan profiles have been reported  in ovarian,309  breast,310  lung,311  prostate,95, 96  pancreatic 312 and hepatocellular carcinoma.97 Also, the diagnostic and progression of diseases and disorders such as atherosclerosis,313 rheumatoid arthritis,123  Crohn’s syndrome314 and inflammation315 have been associated with changes in the glycosylation profile of serum glycoproteins. The ability to examine variations in glycosylation of serum glycoproteins could also serve for diagnostic and/or prognostic monitoring with sufficient specificity for clinical use.  The common practice to perform glycan profiling of human serum is limited to obtain a mass profile of the glycans which is not sufficient to analyze their vast structural diversity since many possible structures could correspond to a certain composition.316, 317 Matrix-assisted laser 116  desorption/ionization (MALDI)95, 96 and electrospray (ESI)310, 318 have been commonly reported for global glycan profiling of human serum. However, these approaches alone cannot identify structural and/or positional glycan isomers, and often require to be coupled with high-performance separation methods to identify robust and specific candidates for glycan markers.  Various combinations of separation techniques and detection methods have been used for the analysis of enzymatically released N-glycans from human serum. Hydrophilic interaction liquid chromatography (HILIC),319, 320 microfluidic chip-based nano LC with porous graphitized carbon (PGC),321, 322 capillary gel electrophoresis with laser-induced fluorescence detection (CGE-LIF),242, 323 and ion mobility mass spectrometry (IMS-MS)318 have been used to interrogate the human serum for glycan-based biomarker discovery. However, limited attempts to monitor changes in glycosylation with little regard for the specific protein to which the carbohydrate are bound, and to perform native glycan isomeric separation have been reported due to the lack of separation power necessary to identify closely related structures.     In the last few years, on-line coupling of capillary electrophoresis to time-of-flight mass spectrometry (TOF-MS) detection has demonstrated to be an attractive and complementary technique for characterization of biomolecules.48 The ability of CE to separate structural and/or linkage isomers makes it a valuable technique for glycan profiling, especially when reduced amount of samples are available.299, 300  Interfacing CE to TOF-MS combines high separation efficiencies with high mass resolution and accuracy, and provides a powerful tool for glycan analysis. To date, the standard approach for coupling CE to MS is the sheath-liquid interface, introduced by Smith et al.42 Although the use of this interface has been demonstrated for CE-MS analysis of low-abundant glycans,242  it often contributes to significant analyte dilution due to the mismatch in sheath-liquid and CE flow rate. The sheathless CE-MS interface, introduced by 117  Moini and coworker,69  has also been used for CE-MS analysis of oligosaccharides.211 In comparison to the sheath-liquid interface, the sheathless interface does not suffer from analyte dilution but often faces difficulties in fabrication and operation that affects its stability and widespread use.   In this work, we describe a simple CE-MS approach for analysis of human serum glycosylation using our earlier reported flow-through microvial interface. Our methodology involves optimization of the conditions for enzymatic digestion, SPE purification, CE separation and MS detection of underivatized N-glycans from human serum. Unlabeled glycans were preferred over derivatized analogues to preserve their natural structural characteristics and to avoid time-consuming derivatization reactions and post sample treatment. An advantage of the present methodology is that human serum is analyzed without any protein purification so the glycans released from tumor-related glycoproteins or from circulating abnormal cells-related glycoproteins are simultaneously being analyzed. Given the potential of structure-specific glycans as markers for diseases, our CE-ESI-MS method was also evaluated for the analysis of isomeric glycans. This simple approach was found to be applicable to N-glycans enzymatically released from glycoproteins present in prostate cancer (late-stage, III) and asthma serum. The capability of our CE-ESI-MS method to glycoprofile complex heterogeneous mixtures of N-linked glycans and to separate isomeric glycans species, could demonstrate its potential for monitoring glycan alterations in the context of malignancies.    118  5.2 MATERIALS AND METHODS 5.2.1 Materials   Commercially available human serum obtained from healthy donors and from patients suffering bronchial asthma and prostate carcinoma (late-stage, III) was purchased from Proteogenex (Culver City, CA, U. S. A.) and stored according to the vendors’ specifications. A Peptide-N-Glycosidase F (PNGase F) kit and GlycocleanTM H cartridges were obtained from Prozyme (Hayward, CA, U.S.A.). All chemicals and solvents were of analytical grade or better. Formic acid, acetic acid, trifluoroacetic acid, ammonium acetate, methanol, ethanol, acetone, and 2-propanol were purchased from Fisher Scientific (Nepean, ON, Canada). All aqueous solutions were prepared using purified water (18.2 MΩ.cm-1) with a Milli-Q purification system (Millipore, Bedford, MA, U.S.A.).  5.2.2 Glycoprotein deglycosylation with PNGase F   According to the protocol described by Morelle et al.,142 human serum glycoproteins were denatured using sodium dodecyl sulfate followed by enzymatic digestion with PNGase F. In brief, 20 μL of human serum solution was adjusted to a final volume of 50 μL with 10 mM Tris-HCl at pH 8.0. Sample denaturation was achieved by adding 2.5 µL of denaturation solution containing 0.1% sodium dodecyl sulfate and 50 mM β-mercaptoethanol and heating at 60 C for 10 minutes. The sample was then treated with 2.5 µL of detergent solution containing 0.75% NP-40, followed by 2 µL of PNGase F (1000 mU). Following overnight sample incubation at 37 C, digested samples were placed at -20 C for 20 minutes to terminate the reaction. Deglycosylated proteins were precipitated using 400 μL of ice-cold ethanol and the sample was then chilled at -80  C for 1 h. Upon centrifugation at 13 200 rpm for 30 minutes, the supernatant, containing the 119  glycans, were transferred to  new Eppendorf tubes, and stored at -20C before solid-phase extraction.   5.2.3 Glycan purification using solid-phase extraction (SPE)   Released N-glycans from human serum were purified by SPE using GlycocleanTM H cartridges. The cartridges were initially washed with 3 mL of 1M NaOH followed by 6 mL of distilled water and 3 mL of 30% acetic acid. Cartridges were then primed with 3 mL of 50% acetonitrile/ 0.1% trifluoroacetic acid (v/v) in water followed by 6 mL of 5% acetonitrile/ 0.1% trifluoroacetic acid (v/v) in water. The glycan samples were diluted with water (1:1 ratio) and then loaded onto the cartridges. Non-glycan contaminants were removed by washing the cartridges with 3 mL of Milli-Q water followed by 5% acetonitrile/ 0.1% trifluoroacetic acid (v/v) in water. Glycans were then eluted with 2 mL of 50% acetonitrile/ 0.1% trifluoroacetic acid (v/v) in water. Finally, N-glycans were dried down in a SpeedVac vacuum centrifuge (Eppendorf, Germany) and reconstituted in 20 µL of distilled water prior to CE-MS analysis.   5.2.4 Electrophoretic procedure   CE-ESI-MS analyses of SPE purified N-glycans were performed using hydrophilic hydroxypropyl cellulose-coated capillaries (HPC-capillaries, 50 µm ID, 365 µm OD, 70 cm long), prepared in-house according to the method of Shen and Smith.22  Separations were carried out in reverse-polarity mode at -25 kV with an overimposed pressure of 10 mbar (0.145 psi) to reduce analysis time for neutral glycans. Background electrolyte (BGE) was composed of 50 mM ammonium acetate (pH 3.1) containing 20% methanol. The samples were injected at 1.0 psi for 10 s, corresponding to a volume of 19 nL.  120  5.2.5 CE-ESI-MS system   CE-ESI-MS analyses of N-glycans were carried out with a PA800 plus capillary electrophoresis system (Beckman Coulter, Brea, CA) coupled to an API 4000 triple-quadrupole mass spectrometer (AB SCIEX, Concord, Canada). A modified capillary cartridge that permits MS detection was used for the analysis. The CE-ESI-MS interface with a flow-through microvial was developed in our laboratory and has been described previously.26, 76, 218, 219, 243-245 The electrospray voltage was set at -3.5 kV. Data acquisition and system control were performed using the Analyst® 1.4.2. software (AB SCIEX, Framingham, MA, U.S.A.). The modifier solution was delivered by a syringe pump (Harvard Apparatus, Holliston MA) at a flow rate of 0.3 µL/min and was composed of 10 mM ammonium acetate (pH 3.1) containing 75% of 2-propanol and methanol (2:1 ratio). The same interface setup and solutions were used for CE-ESI-MS with an HP 3DCE (Agilent Technologies, Waldbronn, Germany) coupled to a maXisTM Ultra High-Resolution TOF mass spectrometer (Bruker Daltonik, Bremen, Germany). The mass spectrometer was operated in negative ionization mode and acquired data in full scan mode with a mass range from m/z 700 to 2500 at a spectral acquisition rate of 1 Hz. Optimization of the transfer parameters and MS calibration were performed with a TuneMixTM (Agilent Technologies) in order to obtain the best sensitivity at satisfactory resolution (R≈ 20 000 at m/z 1334). Instrument control and data analysis were performed using ESI Compass 1.3 application software from Bruker (Bruker Daltonik, Bremen, Germany). The conditions for modifier solution and its delivery rate are the same as the ones when the API 4000 MS was used.   121  5.3 RESULTS AND DISCUSSION 5.3.1 Optimization of enzymatic N-glycan release Glycome screening of human serum requires a rapid and effective sample preparation method and purification for the analysis of the resulting glycans. Our sample preparation protocol is simple and effective and includes enzymatic release followed by protein precipitation using ethanol and graphitized carbon for SPE purification (Figure 5.1). Following purification, the released N-glycans are ready for CE-ESI-MS injection. In this study N-glycans released from human serum were analyzed using two different types of mass analyzers.   Morelle et al.142 have reported strategies for enzymatically release of glycans from purified proteins and other biological fluids. Using the protocol reported for releasing N-glycans as a starting point,142 we optimized the denaturation temperature and the time for the enzymatic digestion for the analysis of N-glycans from human serum. Experimental details for the optimization procedure are described in the Appendix D.1. The following method was chosen for the enzymatic release of N-glycans: 20 µL of human serum (adjusted to 50 µL with Tris-HCl at pH 8.0) was denatured by addition of 2.5 µL of 0.1% sodium dodecyl sulfate and 50 mM β-mercaptoethanol, followed by incubation at 60 C for 10 minutes. After samples were removed from the heat, 2.5 µL of detergent solution containing 0.75% NP-40 was added, followed by 2 µL of PNGase F (1000 mU). N-glycans were then cleaved off from their corresponding glycoproteins by incubation overnight at 37 C using the procedure described above. Following ethanol precipitation and SPE purification, glycans were dried down in a vacuum centrifuge and reconstituted in 20 µL of distilled water prior to CE-MS analysis.   122    Figure 5.1 Strategy for sample preparation and analysis of human serum glycans. Released N-glycans were analyzed using a flow-through microvial CE-MS interface connected to a Triple Quadrupole and a High-Resolution TOF MS. SDS in the second box refers to sodium dodecyl sulfate.  5.3.2 CE-ESI-MS separation of N-glycans from human serum  Based on our experience using neutral coated capillaries and acidic BGE’s for CE-ESI-MS analysis of N-glycans from rHuEPO and fish serum,26 described in chapters 3 and 4, respectively, we used similar conditions for the analysis of N-glycans from human serum. Underivatized N-glycans were chosen because the derivatization procedure may modify the original glycan profile giving rise to a reduced structural heterogeneity due to partial loss of sialic acid residues.  Since human serum N-glycans comprise a heterogeneous mixture—containing neutral and acidic species—with great structural diversity and dynamic range,324 prior electrophoretic separation is mandatory to minimize ion suppression effects that often occurs when direct infusion into the MS is performed. Baseline separation between neutral and acidic glycans was achieved (Figure 5.2) and both types of glycans were mainly detected as [M-H]- ion species. API 4000 Triple Quadrupole MS 20 uL Serum SampleDenaturation using SDSN-glycan release using PNGase F EtOH precipitationPorous Graphitized Carbon SPE CE-ESI/MS maXisTMUltra-High Resolution TOF MS 123  Sialylated glycans were also detected as [M-2H]2- ion species and less frequently as [M-3H]3- ion species. Under reversed polarity mode and acidic BGE conditions (pH 3.1), sialylated glycans migrated faster than neutral glycans due to their electrophoretic mobility and appeared earlier in the electropherogram (peaks around 20 and 25 min). Among sialylated glycans, differential electrophoretic mobilities were observed based on the number of sialic acids residues, hydrodynamic volume, size and elongation of the glycan chains. Of all those factors, the degree of sialylation seems to influence the most the electrophoretic mobility of the glycans. For example, tri- and tetra-sialylated complex-type glycans with presumably three or more antennas displayed a high electrophoretic mobility and migrated earlier (around 20 min) than mono- and di-sialylated complex bi-and tri-antennary glycans. Complex and hybrid-type mono-and di-sialylated glycans showed a reduced electrophoretic mobility and migrated later in the electropherogram (around 25 min). However, the influence of size and shape on migration time was less predictable for bi- and tri-antennary glycans since these two types of glycans can have for instance, similar size and composition but a significantly different three-dimensional structure, which accounts for the migration order observed in Figure 5.2. Interactions of neutral glycans with acetate ions and an over imposed pressure of 10 mbar (0.145 psi) allowed them to migrate towards the anode and appeared at longer migration times between 32 and 36 min.   Although neutral and acidic glycans, containing the same number of sialic acid residues are not fully time resolved, they can be profiled and distinguished in the mass dimension, given the orthogonal baseline MS resolution. The excellent separation power of CE-MS allows the analysis of neutral and acidic glycans under the same electrophoretic run. Also, this CE-MS approach eliminates the need to label the glycans, as well as the need to remove sialic acid 124  residues prior to their analysis, which is a common practice for analysis of acidic glycans from different sources.100     Figure 5.2 Base peak electropherogram for the CE-MS separation of N-glycans of a control sample of human serum using the maXisTM High-Resolution TOF MS. Only the putative structures of the most abundant glycans are shown. Monosaccharide legend for the putative structures as follows: blue squares: N-acetylglucosamine, green circles: mannose, yellow circles: galactose, purple diamonds: N-acetyl neuraminic acid and red triangles: fucose. The position of the N-acetyl neuraminic acid correspond to the proposed structure and it has not been defined in the present study.  5.3.3 N-glycan detection and identification  Glycan detection was initially performed using a triple quadrupole MS which revealed the presence of more than 25 neutral and sialylated glycans that are in agreement with other reports in the literature for the analysis of human serum glycosylation.114, 322, 325, 326  In order to accurately determine the mass of the observed glycans, our flow-through microvial CE-MS 0 10 20 30 40 50 Time [min]0123454x10Intens.23Aug2012_000005.d: BPC 700.0-2300.0 -All MS FullScan, Smoothed (1.02,1,GA), Smoothed (1.02,2,GA), Smoothed (1.02,2,GA)125  interface was connected to a maXisTM ultra high-resolution TOF MS which offers a wider mass range, higher resolution and higher mass accuracy than the triple quadrupole MS. In addition to the glycans detected before, a set of more than 30 glycans were observed with the TOF-MS, including some complex and hybrid fucosylated and sialylated glycans that have not been previously reported. In total, more than 60 glycans were detected of which several could be identified. Table 5-1 shows a list of the glycans observed, including putative monosaccharide compositions, observed masses and, theoretical masses for all the glycans detected. Glycans were identified based on their true isotopic pattern provided by the ESI Compass 1.3 application software from the maXisTM TOF MS in combination with the information provided by the Consortium of Functional Glycomics (CFG, http://www.functionalglycomics.org)186 for N-linked glycans. Since the serum sample is from a human source, two search criteria were used. First, the composition only included glycans containing hexose (H), N-acetylhexosamine (N), fucose (F) and N-acetylneuraminic acid (S) with a pentasaccharide core of minimum three hexoses and two N-acetylglucosamines. The second criterion was that glycan composition must belong to one of the three types of N-linked glycans: high-mannose, complex or hybrid type.      Table 5-1 List of the distinct compositions of N-linked oligosaccharides observed in human serum (control sample) using the maXisTM Ultra High-Resolution TOF MS.  Theo. glycan mass (Da) Obs. glycan mass (Da)  Oligosaccharide Composition a H N F S  High Mannose  1235.099  1235.092  5  2  0  0  1559.380  1559.401  7  2  0  0  1883.667  1883.711*  9  2  0  0  2045.809  2045.825*  10  2  0  0   126  Theo. glycan mass (Da) Obs. glycan mass (Da)  Oligosaccharide Composition a H N F S  Complex/Hybrid (non-fucosylated, non-sialylated)      1276.150  1276.139  4  3  0  0  1317.206  1317.188  3  4  0  0  1438.290  1438.272  5  3  0  0  1479.348  1479.335  4  4  0  0  1519.566  1519.598  3  5  0  0  1600.436  1600.425  6  3  0  0  1641.491  1641.473  5  4  0  0  1681.619  1681.639  4  5  0  0  1843.672  1843.633  5  5  0  0  2006.827  2006.804  6  5  0  0  2737.500  2737.467  8  7  0  0   Complex/Hybrid (fucosylated)   1056.957  1056.931   3  2  1  0  1260.151  1259.112  3  3  1  0  1422.293  1422.276  4  3  1  0  1463.347  1463.331  3  4  1  0  1584.442  1584.421  5  3  1  0  1625.490  1625.436  4  4  1  0  1666.543  1666.577  3  5  1  0  1787.631  1787.692  5  4  1  0  1828.671  1828.637  4  5  1  0  1949.770  1949.811*  6  4  1  0  1990.866  1990.902  5  5  1  0  2095.921  2095.891*  6  4  2  0  2111.912  2111.886  7  4  1  0  2152.975  2152.998  6  5  1  0  2194.022  2194.056  5  6  1  0  2283.119  2283.171  5  5  3  0  2632.448  2632.501  5  6  4  0  2792.542  2792.505  6  6  4  0  2810.593  2810.637  7  6  3  0   Complex/Hybrid (sialylated)   1567.411  1567.439   4  3  0  1  1729.552  1729.578  5  3  0  1  1770.611  1770.576  4  4  0  1  1932.750  1932.717  5  4  0  1  2224.006  2224.037*  5  4  0  2  2298.085  2298.123*  6  5  0  1  2589.343  2589.389*  6  5  0  2  2663.422  2663.478  7  6  0  1  2880.601  2880.559*  6  5  0  3  2954.682  2954.623*  7  6  0  2  3171.864  3171.808  6  5  0  4  3537.215  3537.149*  7  6  0  4     127  Theo. glycan mass (Da)  Obs. glycan mass (Da)    Oligosaccharide Composition a  H  N  F  S  Complex/Hybrid (fucosylated and sialylated)   1916.762  1916.793  4  4  1  1  2077.746  2077.705  5  4  1  1  2119.942  2119.908  4  5  1  1  2265.085  2265.130  4  5  2  1  2282.085  2282.126  5  5  1  1  2323.138  2323.099  4  6  1  1  2370.146  2370.088  5  4  1  2  2573.342  2573.289  5  5  1  2  2614.401  2614.356  4  6  1  2  2641.422  2641.454  5  6  2  1  2647.422  2647.401  6  6  1  1  2735.485  2735.516  6  5  1  2  2922.680  2922.705  5  6  2  2  2938.680  2938.717  6  6  1  2  3012.759  3012.801  7  7  1  1  3026.742  3026.687  6  5  1  3  3084.822  3084.856  6  6  2  2  3101.848  3101.801  7  6  1  2  3172.885  3172.843  6  5  2  3  3246.964  3246.917  7  6  2  2  3319.027  3319.056  6  5  3  3  3392.083  3392.112  7  6  1  3  3683.338  3683.297  7  6  1  4   aNomenclature identification: H= hexose, N= N-acetylhexosamine, F= fucose and S= sialic acid (NeuAc). Some glycans are marked with an asterisk (*) to indicate that isomer separation, at baseline, has been observed for that glycan.    In order to confirm the composition of the glycans previously identified, CE-ESI-MS/MS was performed. For instance, Figure 5.3 shows the fragmentation pattern of two acidic N-glycans (H5N4S2 at m/z 1111.0 and H6N5S3 at m/z 959.4). The CID spectrum in Figure 5.3A shows characteristic N-glycan fragment ions consistent with a composition corresponding to 5 hexoses, 4 N-acetylhexosamines and 2 N-acetylneuraminic acids. For example, the presence of a singly-charged fragment at m/z 655.5 corresponding to [Neu5Ac-Gal-GalNAc-H]- indicates the composition of the terminal sialylated antenna. The CID spectrum in Figure 5.3B shows clear characteristic N-glycan fragments confirming the glycan annotation corresponding to 6 hexoses, 128  5 N-acetylhexosamines and 3 N-acetylneuraminic acids. The fragment ion at m/z 655.5, corresponding to ion B3x according to the nomenclature of Domon and Costello,191 is also present indicating the composition of the antenna. In both cases an intense 2,5A7 ion due to internal fragmentation by cross-ring cleavage from the corresponding glycan at m/z 1060.4 and 925.3 was also observed.       (A)              -MS2 (1111.10) CE (-60): 19.211 to 19.712 min from Sample 4 (20120709  Bronchial Asthma Serum MSMS 04) of 20120709 Glycans .... Max. 5.7e5 cps.200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000 1050 1100 1150 1200m/z, amu5.0e41.0e51.5e52.0e52.5e53.0e53.5e54.0e54.5e55.0e55.5e55.7e5Intensity, cps289.81060.4958.61000.51030.51110.9835.5907.5655.5919.7671.51009.1949.91101.7424.0468.4306.0889.9817.9B1B32,4A4B52,5A72,5A6C4B6[2,4A4x– 290]-B3xC4xB4xB52,5A6B62,5A7[M– 2H]2-B1x129  (B)   Figure 5.3 (A) Fragmentation spectra of glycan at m/z 1110.9 with composition H5N4S2 and (B) at m/z 959.0 with composition H6N5S3. The fragmentation spectrum of glycan shown in (A) corresponds to a bi-antennary sialylated glycan, while the spectrum shown in (B) corresponds to a tri-antennary sialylated glycan.  Of particular importance is the fact that the CE-ESI-MS method alleviates the suppression effects often encountered when glycan mass profiling is performed with MALDI or ESI by direct infusion.95, 96, 310, 318 The most abundant glycans, fucosylated and sialylated, did not suppress the signals of less abundant species even in negative ionization mode. A further consequence of the minimized suppression effects is the possibility to detect glycan structures that cannot be observed when mass glycoprofiling, with no separation, is performed. Also, under  the detection conditions, the in-source decay by desialylation or the loss of fucose residues, previously reported327 is nonexistent and allows performing detection of underivatized glycans.     -MS2 (959.40) CE (-40): 16.915 to 17.283 min from Sample 6 (20120707 Prostate Cancer Serum MSMS 959  06) of 20120707 Glycans... Max. 2.4e5 cps.200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000m/z, amu1.0e42.0e43.0e44.0e45.0e46.0e47.0e48.0e49.0e41.0e51.1e51.2e51.3e51.4e51.5e51.6e51.7e51.8e51.9e52.0e52.1e52.2e52.3e52.4e5Intensity, cps925.3289.7959.0857.5905.2891.3953.0933.0823.3851.7655.0273.4 321.4234.7 746.0716.0B1xB3xB5B6[M– 3H]3-0,2A62,5A62,5A72,5A7B5B1B30,2A62,5A6B6130  5.3.4 CE-MS separation of isomeric glycans    Despite molecular mass information is considered fundamental for structural assignment of glycans, the occurrence of multiple isomeric forms preclude determining the exact sequence of monosaccharides and requires high-performance separation methods previous to MS detection. Moreover, separation of structure-specific glycans is of great importance because it may provide highly specific glycan markers for diseases. Baseline separation of structural and/or linkage isomers was achieved for N-glycans of human serum under the conditions of the CE-MS analysis. Representative extracted ion electropherograms (EIEs) for the separation of isomers of acidic N-glycans from human serum are shown in Figure 5.4 to depict the separation power of the method and the number of isomers that may comprise a specific composition. The mass spectra corresponding to the dominant peak for each EIE are shown in Appendix D.2. The presence of two peaks—separated at baseline— are probably due to positional or linkage isomers that have different electrophoretic mobilities and therefore, migrate with different velocities. Isomeric glycans, with unique three dimensional structures, may have distinctive interactions with the BGE components that affect their shapes and hydrodynamic sizes and account for their dissimilar electrophoretic mobilities leading to baseline separation.  Full structural elucidation of each isomeric peak will probably require the use of a cocktail of enzymes to differentiate, the occurrence of α2,3- and α2,6-linked Neu5Ac commonly found in acidic oligosaccharides, and the potential epimerization of N-acetylglucosamine to N-acetylmannosamine. The reproducibility in terms of migration times observed for each peak in Figure 5.4 (%RSD < 1%), confirms that each one represent a specific glycan structure although the exact glycan structure has not been elucidated yet.  The ability of our CE-ESI-MS method to separate—at baseline— underivatized isomers of human serum is of note given the limited 131  number of studies reported in the literature.114, 322 Derivatization procedures were not employed in this study because they increase the sample manipulation and processing times and alter the original glycan profile.     Figure 5.4 EIEs of acidic N-linked glycans released from human serum showing baseline separation for isomeric species at (A) m/z 1111.1, (B) m/z 959.4, (C) m/z 1008.3 and (D) m/z 1057.4. Glycan composition is indicated for each species below their corresponding m/z value. Glycans were observed as [M-2H]2- and [M-3H]3- ions species. Putative monosaccharide composition and the corresponding m/z are indicated for each glycan.     5.3.5 Method reproducibility   Variability of the CE-ESI-MS method was simultaneously evaluated in terms of the relative abundance and the migration time of the 10 major peaks observed for N-glycans released Base Peak Chrom. of -MS2 (1111.10) CE (-60): from Sample 4 (20120709  Bronchial Asthma Serum MSMS 04) of 20120709 Glycans .... Max. 3.2e5 cps.2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38Time, min0.02.0e44.0e46.0e48.0e41.0e51.2e51.4e51.6e51.8e52.0e52.2e52.4e52.6e52.8e53.0e53.2e5Intensity, cps19.5526.44Base Peak Chrom. of -MS2 (1008.30) CE(-40): from Sample 11 (20120707 Prostate Cancer Serum MSMS 1008  11) of 20120707 Glyc... Max. 2.1e5 cps.2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34Time, min0.01.0e42.0e43.0e44.0e45.0e46.0e47.0e48.0e49.0e41.0e51.1e51.2e51.3e51.4e51.5e51.6e51.7e51.8e51.9e52.0e52.1e5Intensity, cps17.1018.64Bas  Peak Chrom. of -MS2 (1057.10) CE (-30): from Sample 5 (20120709  Bronchial Asthma Serum MSMS 05) of 20120709 Glycans .... Max. 5.3e5 cps.2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44Time, min0.02.0e44.0e46.0e481.0e51.2e51.4e51.6e518 52. 52.2e52.4e52.6e52.8 533.2e53.4e53.6e53.8e5404.2e54.4e54.6e54.8e55.05.2Intensity, cps16.876.85XIC of -Q1: 959.4 amu from Sample 1 (20120707 Prostate Cancer Serum 01) of 20120707 Glycans .wiff (Turbo Spray), Smoothed, Smo... Max. 5.0e6 cps.2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44Time, min0.02.0e54.0e56.0e58.0e51 61.2e61.4e61.6e61.8e62.0e62.2e62.4e626 62.8e63.0e63.2 63.4e63.6e63.8e64.0e64 64.4e64.6e64.8 65.0e6Intensity, cps18.1923.40Bas  Pe kChrom. of -MS2 (1057.10) CE (-30): from Sample 5 (20120709  Bronchial Asthma Serum MSMS 05) of 20120709 Glycans .... Max. 5.3e5 cps.2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44Time, min2.0e4468.1 52.4e682.05. e3.24 56.8e40.5.2Intensity, cps16.8726.85H6N5S3 m/z = 959.4 H5N4S2 m/z =1111.1H6N5F1S3 m/z =1008.3 H6N5F2S3 m/z =1057.4(A) (B) (C) (D) 132  from human serum. Three aliquots of a control sample of human serum was processed following the protocol shown in Figure 5.1 and the resulting N-glycan mixture was then analyzed using HPC-coated capillaries. The intra-batch reproducibility was evaluated by performing multiple injections (n=6) from a single aliquot in three consecutive days. The migration time percent relative standard deviations (%RSD) for the 10 most abundant components was less than 1% (0.28-0.45%), while the %RSD for the relative abundance was less than 10% (4.1-7.3%). Reproducibility of migration times and relative abundance of the 10 most abundant N-glycans from the three aliquots was evaluated in an analogous manner to the single aliquot analysis. The inter-batch reproducibility was determined by performing multiple injections (n=3) from the three aliquots in two consecutive days. The migration time %RSD for the ten components was also less than 1% (0.48-0.65%), while the relative abundance %RSD was less than 10% (3.7-8.6%). These variations are probably due to the enzymatic release and SPE purification prior to CE-ESI-MS analysis and are typical values for CE-MS without standards.328  5.3.6 Variations of glycosylation in different types of human serum samples In order to evaluate the applicability of the CE-ESI-MS method to study abnormal glycosylation for different pathologic conditions, human serum of patients suffering prostate carcinoma (late-stage, III) and bronchial asthma were evaluated. Samples were processed following the methodology presented in Figure 5.1. Differences between the sera sample against a control sample were found in terms of the type and abundance of the N-glycans present.  In order to obtain a holistic perspective of the serum glycome and to simplify the comparison among the different types of samples, compositional glycoprofiling was initially performed. Observed N-glycans were classified according to their composition into one of the 133  following groups: (1) high-mannose glycans; (2) non-fucosylated and non-sialylated complex/hybrid glycans; (3) fucosylated (but non-sialylated) complex/hybrid glycans; (4) sialylated (but non-fucosylated) complex/hybrid glycans; and (5) fucosylated and sialylated complex/hybrid glycans. Figure 5.5A shows the relative abundances of the most abundant glycans found in common for the control, prostate cancer and bronchial asthma human sera. Relative abundances were calculated based on the total glycan abundance observed for each sample type. Although many glycans were observed in common, differences in terms of the relative abundance and glycan type was noticeable. Pie charts based upon relative abundances were used to approximate the proportions of the different types of N-glycans observed in each type of sample (Figure 5.5B, C, D). According to Figures 5.5B, C and D, sialylated and fucosylated complex/hybrid glycans were found to be the most abundant N-glycans in human serum: 35.29% in control serum, 40.35% in prostate carcinoma serum and 40.68% in bronchial asthma serum. High abundances were also found for fucosylated complex/hybrid glycans (25.50% in control serum; 26.32% in prostate carcinoma serum, and 23.73% in bronchial asthma serum), as well as for sialylated complex/hybrid glycans (21.57% in control serum; 21.05% in prostate carcinoma serum, and 22.03% in bronchial asthma serum). In general, lower relative abundances were found for non-fucosylated and non-sialylated complex/hybrid glycans (11.76% in control serum; 7.02% in prostate carcinoma serum, and 6.78% in bronchial asthma serum) and high-mannose glycans (5.88% in control serum; 5.26% in prostate carcinoma serum, and 6.78% in bronchial asthma serum).     134  (A)      Figure 5.5 (A) Comparison of the relative abundances of all the N-glycans found in common among the human sera samples. Pie charts illustrating the N-glycans types and relative abundances present in (B) normal human serum, (C) prostate carcinoma (stage III) serum and (D) asthma serum obtained using the optimized CE-ESI-MS conditions.   051015202530354045Relative abundance (%)Normal Serum  Prostate Cancer Serum Asthma Serum 5.88%11.76%25.50%21.57%35.29%5.26%7.02%26.32%21.05%40.35%6.78%6.78%23.73%22.03%40.68%(B) (C) (D)5.88%11.76%25.49%21.5635.29%5.26%7.01%26.31%21.05%40.35%6.77%6.77%23.72%22.03%40.67%Normal Human SerumProstate Cancer (stageIII) SerumAsthma SerumComplex/Hybrid (non-fucosylated, non-sialylated)High Mannoseo plex/ ybrid (fucosylated)Complex/Hybrid (sialylated)Complex/Hybrid (fucosylated,sialylated)5.88%11.76%25.49%21.5635.29%5.26%7.01%26.31%21.05%40.35%6.77%6.77%23.72%22.03%40.67%Normal Human SerumProstate Cancer (stageIII) SerumAsthma SerumComplex/Hybrid (non-fucosylated, non-sialylated)High Mannosel / ri  (fuc sylated)Complex/Hybrid (sialylated)Complex/Hybrid (fucosylated,sialylated)135  Previous studies on glycosylation profiles of human serum underline the correlation between altered glycosylation of serum glycoproteins and diverse pathophysiologies, and the potential of aberrant glycans as promising biomarkers.316, 329-331  In that regard, Ryden et al.332 have introduced the α1-acid glycoprotein (AGP) fucosylation index to characterize the degree of fucosylation of glycans present in serum of patients with demonstrated liver cirrhosis. Imre et al.316 have developed a fucosylation index and a branching index to characterize the degree of fucosylation and branching of AGP-derived glycans in patients with lymphoma and with ovarian tumor. Despite some correlation has been found between the degree of fucosylation and the degree of branching on the pathophysiology, the methods focused only in changes occurring at the AGP level.316, 332 Potentially, other glycans derived from human serum glycoproteins may have also a predictive power to discriminate abnormal vs. healthy individual samples. Therefore, here we propose a fucosylation-sialylation index to compare and characterize all the glycans present in human serum, given our results of increased relative abundances of fucosylated and sialylated glycans in abnormal serum samples. The fucosylation-sialylation index indicates the average abundance of fucosylated and sialylated complex/hybrid glycans according to the relative abundance of glycans containing fucose and sialic acid units and it is given by the following expression:  𝑭𝒖𝒄𝒐𝒔𝒚𝒍𝒂𝒕𝒊𝒐𝒏  𝑺𝒊𝒂𝒍𝒚𝒍𝒂𝒕𝒊𝒐𝒏 𝑰𝒏𝒅𝒆𝒙 (𝑭𝒖𝒄  𝑺𝒊𝒂 𝒊𝒏𝒅𝒆𝒙)= 1  𝑔𝑙   𝑛  1𝑆1 + 2  𝑔𝑙   𝑛  1𝑆2      2𝑆2   3  𝑔𝑙   𝑛  1𝑆3      3𝑆3 +  4  𝑔𝑙   𝑛  1𝑆4 (    4𝑆4) 𝑔𝑙   𝑛  Where glycan F1S1 indicates the sum of relative abundances of glycans containing one fucose and one sialic acid; glycan F1S2 indicates the sum of relative abundances of glycans containing one fucose and two sialic acids or two fucoses and two sialic acids; glycan F1S3 accounts for the 136  relative abundance of glycans containing one fucose and three sialic acids or three fucoses and three sialic acids; glycan F1S4 is the sum of relative abundances of glycans containing one fucose and four sialic acids or four fucoses and four sialic acids, while  𝑔𝑙   𝑛  indicates the sum of abundances of all glycan peaks that due to normalization correspond to 100. The Fuc-Sia index calculated for control, prostate carcinoma and bronchial asthma sera were 0.298, 0.387 and 0.405, respectively. These numbers indicate that the index is low for healthy individuals and is approximately 30 and 35% higher in abnormal serum samples associated with prostate cancer and bronchial asthma, respectively. Despite that the Fuc-Sia index is a simple parameter to evaluate the occurrence of fucosylated-sialylated glycans it shows that the degree of fucosylation-sialylation is different for the glycans associated with diverse abnormalities. In general, the index provides a holistic perspective of the glycosylation changes associated with prostate carcinoma and bronchial asthma in comparison to normal serum. A broader study, including more samples types should be carried out to further evaluate the practical values of the index as an indicator of pathophysiology. However, the results presented here demonstrate the potential of CE coupled to mass spectrometry to provide valuable data for characterizing pathophysiologies in terms of glycan analysis in human samples.   5.4 CONCLUDING REMARKS  In this study, we have demonstrated the potential of a simple CE-ESI-MS method using a flow-through microvial interface for the analysis of enzymatically released serum N-glycans. The CE-MS allowed characterizing heterogeneous complex mixtures of serum glycans with a potential for biomarker discovery. The use of hydrophilic HPC-coated capillaries an acidic buffer conditions facilitated the glycoprofiling of both neutral and acidic N-glycans under the 137  same electrophoretic run without the need to remove sialic acid residues. Underivatized N-glycans were employed in this study, because the labeling procedure often affects the original glycoprofile and may lead to desialylation. Migration order of the complicated glycan mixtures revealed different mechanisms for the separation based not only on charge-to-size ratio but also on a combination of hydrodynamic volume, size, tri-dimensional structure and extension of the glycan chain length.  The separation power of CE and the high resolution and mass accuracy of the TOF-MS made it possible to unravel the complexity of the sample and allowed the identification of more than 60 distinct glycan compositions in normal human serum. In addition, the baseline separation of isomeric species achieved is of note, because it demonstrates the potential of the CE-ESI-MS/MS method for identification of individual species that could potentially be used to detect specific markers for disease diagnosis. Further identification of the structural or linkage isomers observed would require the use of a cocktail of glycosidases but was not possible at this stage. Glycoprofiling of sera was performed in terms of N-glycan composition and relative abundances of undecorated, decorated complex/hybrid (fucosylated, sialylated and fucosylated-sialylated) and high-mannose type glycans. Although the method was applied to a reduced sample set, the fucosylation-sialylation index revealed variability in the fucosylation and sialylation levels of complex/hybrid type glycans between normal and diseased-related serum samples. Although, there is a considerable increment of about 30% in the Fuc-Sia index between normal and abnormal serum samples, larger sample sets that allow studies with high specificity and sensitivity would be required to fully reveal the potential our CE-MS method for identification of potential biomarkers of clinical use.   138  Chapter 6: Concluding remarks and future work  6.1 CONCLUDING REMARKS The material presented in this thesis provides a new analytical platform required for the study of complex N-glycosylation of glycoproteins present in animal serum, human serum and recombinant therapeutic drugs by a flexible strategy that combines capillary electrophoresis separation with mass spectrometry detection. The current state of CE-MS research involves the use of both home-made and commercially available interfaces. Our goal was to test the applicability of the interfacing strategy developed earlier in our research group to study protein glycosylation of relevant biological samples and recombinant therapeutic proteins. However, given the flexibility of the technique and the diversity of CE modes, it is applicable to a wider selection of analytes compared to other analytical separation techniques.  The flow-through microvial design is demonstrated to provide maximum flexibility in the choice of conditions and separation modes of capillary electrophoresis operations. While much work has been done in this development, there are two fundamental difficulties: the limit in electrical current during separation and the choice of BGE components. These limits made it more challenging to the development of methods for analyzing intact glycoproteins, enzymatically released N-glycans, and glycopeptides. Although these restrictions can compromise the efficiency of the ESI process and could potentially affect the performance of the mass spectrometer, they are not unique to the flow through microvial design.   A comprehensive strategy based on CE-MS for the analysis of glycosylated proteins not only has demonstrated the ability of our ESI-MS interface to separate and detect intact protein glycoforms, but also showed the compatibility of the interface with different types of capillary 139  coatings. The modifier solution allowed this type of system configuration by providing the solution requirements of the CE separation satisfy the conditions for generating a stable electrospray. Glycopeptides enzymatically released from the intact model glycoprotein were used to identify the glycosylation site on the protein, by subtracting the masses of the glycans from the masses of the glycopeptides. Investigating protein glycosylation at the level of glycopeptide is as important as investigating the released N-glycans, and reveals the potential of our methodology to perform comprehensive glycoprotein characterization.   CE-MS analysis of enzymatically released N-glycans from model glycoproteins, recombinant therapeutic drugs, animal and human serum were successfully performed with the flow-through microvial interface. Enzymatically released N-glycans from our model protein were derivatized in a one-step reaction with a recently reported labeling reagent to improve their separation efficiency and detection sensitivity by ESI-MS. Derivatized glycans were successfully baseline resolved. Further studies of enzymatically released N-glycans were performed almost exclusively on unlabeled glycans in order to preserve the structural identity of the glycans, and to minimize sample handling that could negatively impact their detection. To summarize, a simple and effective methodology to characterize heterogeneous mixtures of N-glycans from human IgG and rHuEPO has been developed. The methodology revealed the ability of our CE-MS system to determine a high number of glycoforms with high mass accuracy. It was possible to simultaneously glycoprofile underivatized sialylated and neutral glycans, without the need to label the glycans or to remove sialic acid residues from acidic glycans. A small overimposed pressure and interaction of neutral glycans with buffer components allowed these species to be detected in the same electrophoretic run along with acidic glycans. It was possible to unravel the 140  complexity of the released carbohydrates and to show detailed information about minor glycan modifications that were not detected when a complementary approach based on LC-MS/MS was employed.  Furthermore, a successful CE-ESI-MS/MS methodology was developed to analyze glycosylation of structurally similar glycan components present in fish serum and to perform structural analysis of acetylated glycan species. Tandem MS operated in selected reaction monitoring (SRM) provided useful structural information of native of isomeric species. Fragmentation spectra revealed the identity of isomeric glycan species, which were electrophoretically separated at baseline. The methodology presented is useful not only for the characterization of acetylation of sialic acids but to study alteration of sialylation patterns involved in human cancer progression. The approach could be potentially used to study other types of glycan modifications, as well as to allow determination of overall glycan composition in glycoproteins.  In order to preserve glycan structure, an optimized deglycosylation analysis procedure has been developed and applied to the enzymatic release of N-glycans from human serum to monitor glycan alterations in the context of malignancies. The capability of the CE-ESI-MS method to perform global glycan profiling of human serum, including isomer separation, demonstrates its potential for the discovery of glycan disease markers with high selectivity and specificity. The method could be an ideal complement to established techniques for glycan characterization of therapeutic glycoproteins, as it provides an overall glycosylation profile of their carbohydrate composition and their associated species.   141  6.2 FUTURE RESEARCH DIRECTIONS  6.2.1 Exoglycosidase digestion Although we were able to perform structural analysis of complicated glycan mixtures using tandem mass spectrometry, there is a great deal more that can be done in this area. The majority of the results obtained with MS/MS techniques can be summarized as efforts to achieve glycan fragments to ascertain monomer sequence with branching information. However, the challenges of analytical glycobiology involve not only the determination of the sugar sequence and branching, but also elucidation of the interglycosidic linkages and anomeric configuration. Detailed knowledge of the glycan structure will assist with correlating glycan distributions in glycoproteins to their respective function. In that regard, a full range of exoglycosidases enzymes are often require to achieve complete structural analysis. These enzymes are specific to the stereochemistry, the anomeric configuration of the monosaccharide being released, and its linkage site regarding the remainder of the glycan chain. Exoglycosidase digestions can be applied both to mixtures of glycans and to a specific glycan. Given the complexity of the sample mixtures after endoglycosidase digestion, further improvements on reproducibility and separation efficiency would have to be attained. First of all, higher stability of the capillary surface coating in the presence of different buffer components to attain better glycan separation is an important issue to be investigated for improving analyte separation. Secondly, the use of enzyme cocktails for detailed glycan analysis would require the use of extremely rigorous clean up protocols before MS analysis is performed on glycan fragments. Thirdly, because the currently used procedures with enzymes are not straightforward, failure to degrade a sample may be the result of enzyme’s activity loss or misdosing the enzymes, and should be detected by parallel control samples. For cases in which glycans are released from bacterial proteins, the 142  exoglycosidase strategy can still be useful, although the PNGase F will not be capable of hydrolyzing the carbohydrate-peptide linkage. In such case, a combination of proteolytic digestion should have to be attempted.      6.2.2 High-throughput glycan analysis for biomarker discovery The potential of the flow-through microvial interface has been demonstrated for CE-ESI-MS analysis of complicated human serum glycans, but it only represents an early development stage application. The long-term goal is to adapt the conditions that have been already established for glycan composition profiling, to high-throughput analysis of human serum glycans related to diverse pathophysiologies. Difficulties in analyzing glycan patterns in human serum will be overcome by using CE-MS-based techniques which provide two dimensions of separation and a wealth of information related to glycan structure— useful in connecting glycosylation and pathophysiology. Our future challenge will be to evaluate the results using the most suitable bioinformatics method to fully discriminate glycan patterns in large volumes of data and to find differences, if any, among sample classes (i.e. normal vs. abnormal human serum). Furthermore, as the intrinsic complexity of glycosylation allows the discovery of biomarkers at various levels of structural analysis, optimum CE-ESI-MS methodologies could be beneficial for the study of altered glycosylation of plasma proteins which may lead to promising biomarker discovery. Future CE-ESI-MS methodologies will not only allow structural characterization of glycans, but also assessment of quantitative changes in their glycosylation patterns to attempt selection of putative biomarkers.    143  6.2.3 Use of microfluidic platforms with flow-through microvial interface for the analysis of glycoproteins and their derivatives While the development of microfluidic instrumentation in our research group is not as mature as the flow-through microvial technology, significant progress has being made in the last years that would allow coupling microfluidic platforms to our CE-MS interface. The integration of microfluidcs to ESI-MS detection is highly advantageous in the glycoproteomics field in terms of analytical and economical reasons. Improved detection limits, short analysis time, reduce sample and reagent consumption, and high throughput analysis, are just some of the features that microfluidics-MS can offer in the challenging area of glycoprotemics. In addition, interfacing the flow-through microvial technology with microfluidcs separation possesses the potential to handle tedious sample processing operations on a chip. Therefore, our future challenge will be to adapt the conditions already develop for glycoproteins and their derivatives, to perform adequate separation and ionization using microfluidics-MS technologies. Future attempts in terms of glycan analysis will involve performing glycan release with immobilized PNGase F, removal of deglycosylated proteins, capture of the released glycans, glycan separation on a microfluidic chip, and glycan ionization and MS detection using the flow-through microvial interface. To further enhance the analysis of glycoconjugates on a chip, mapping sequencing and structural elucidation of glycoconjugates derived from relevant biological samples can be attempted. The future implementation of microfluidic devices with MS detection in glycomics will further expand high-throughput analysis to correlate carbohydrate content and glycoprotein functionality.      144  Bibliography  (1) Hjertén, S. Chromatogr. Rev. 1967, 9, 122-219. (2) Jorgenson, J. W.; Lukacs, K. D. Anal. Chem. 1981, 53, 1298-1302. (3) Olivares, J. A.; Nguyen, N. T.; Yonker, C. R.; Smith, R. D. Anal. Chem. 1987, 59, 1230-1232. (4) Hommerson, P.; Khan, A. M.; de Jong, G. J.; Somsen, G. W. Mass Spectrom. Rev. 2011, 30, 1096-1120. (5) Spiro, R. G. Glycobiology 2002, 12, 43R-56R. (6) Varki, A.; Cummings, R. D.; Esko, J. D.; Freeze, H. H.; Stanley, P.; Bertozzi, C. R.; Hart, G. W.; Etzler, M. E., (Eds.), Essentials of Glycobiology; Cold Spring Laboratory Press: NY, 2009. (7) Landers, J. Handbook of Capillary Electrophoresis, ; CRC Press Inc.: Florida U.S.A, 1997. (8) Horvath, J.; Dolnik, V. Electrophoresis 2001, 22, 644-655. (9) Doherty, E. A. S.; Meagher, R. J.; Albarghouthi, M. N.; Barron, A. E. Electrophoresis 2003, 24, 34-54. (10) Lucy, C. A.; Mac Donald, A. M.; Gulcev, M. D. J. Chromatogr. A 2008, 1184, 81-105. (11) Righetti, P. G.; Gelfi, C.; Verzola, B.; Castelletti, L. Electrophoresis 2001, 22, 603-611. (12) Robb, C. S. J. Liq. Chromatogr. 2007, 30, 729-759. (13) Huhn, C.; Ramautar, R.; Wuhrer, M.; Somsen, G. W. Anal. Bioanal. Chem. 2010, 396, 297-314. (14) Ballard, J. N. M.; Lajoie, G. A.; Yeung, K. K. C. J. Chromatogr. A 2007, 1156, 101-110. (15) Larsson, M.; Lutz, E. S. M. Electrophoresis 2000, 21, 2859-2865. (16) Dolnik, V. Electrophoresis 2006, 27, 126-141. (17) Bonnici, P. J.; Damen, M.; Waterval, J. C. M.; Heck, A. J. R. Anal. Biochem. 2001, 290, 292-301. (18) Moini, M.; Demars, S. M.; Huang, H. Anal. Chem. 2002, 74, 3772–3776. (19) Katayama, H.; Ishihama, Y.; Asakawa, N. Anal. Chem. 1998, 70, 5272-5277. (20) Lai, C. C.; Her, G.-R. Rapid Commun. Mass Spectrom. 2000, 14, 2012-2018. (21) Haselberg, R.; de Jong, G. J.; Somsen, G. W. J. Sep. Sci. 2009, 32, 2408-2415. (22) Shen, Y.; Smith, R. D. J. Microcolumn Separations. 2000, 12, 135-141. (23) Fanali, S.; D’Orazio, G.; Foret, F.; Kleparnik, K.; Aturki, Z. Electrophoresis 2006, 27, 4666-4673. (24) Larsson, M.; Sundberg, R.; Folestad, S. J. Chromatogr. A 2001, 934, 75-85. (25) Gennaro, L. A.; Salas-Solano, O.; Ma, S. Anal. Biochem. 2006, 355, 249-258. (26) Jayo, R. G.; Li, J.; Chen, D. D. Y. Anal. Chem. 2012, 84, 8756-8762. (27) Cole, R. Electrospray Ionization Mass Spectrometry; John Wiley & Sons Inc.: New York, 1997. (28) Dunn, W. B.; Ellis, D. I. Trac-Trend Anal. Chem. 2005, 24, 285-294. (29) Han, X. L.; Gross, R. W. Mass Spectrom. Rev 2005, 24, 367-412. (30) Heck, A. J. R.; van den Heuvel, R. H. H. Mass Spectrom. Rev 2004, 23, 368–389. (31) Bailey, A. G. Electrostatic spraying of liquids; Wiley: New York, 1988. (32) Dole, M.; Mack, L. L.; Hines, R. L.; Mobley, R. C.; Ferguson, L. D.; Alice, M. B. J. Chem. Phys. 1968, 49, 2240-2249. 145  (33) Yamashita, M.; Fenn, J. B. J. Phys. Chem. 1984, 88, 4451-4459. (34) Cech, N.; Enke, C. Mass Spectrom. Rev 2001, 20, 362-387. (35) Kebarle, P. J. Mass Spectrom. 2000, 35, 804-817. (36) Kebarle, P.; Peschke, M. Anal. Chim. Acta 2000, 406, 11-35. (37) Gaskell, S. J. Mass Spectrom. 1997, 32, 677-688. (38) Cole, R. J. Mass Spectrom. 2000, 35, 763-772. (39) Iribarne, J. V.; Thomson, B. A. J. Chem. Phys. 1976, 64, 2287-2295. (40) Thomson, B. A.; Iribarne, J. V. J. Chem. Phys. 1979, 71, 4451-4463. (41) Terabe, S.; Otsuka, K.; Ando, T. Anal. Chem. 1985, 57, 834-841. (42) Smith, R. D.; Barinaga, C. J.; Udseth, H. R. Anal. Chem. 1988, 60, 1948-1952. (43) Smith, R. D.; Olivares, J. A.; Nguyen, N. T.; Udseth, H. R. Anal. Chem. 1988, 60, 436-441. (44) Preisler, J.; Hu, P.; Rejtar, T.; Moskovets, E.; Karger, B. L. Anal. Chem. 2002, 74, 17-25. (45) Michalke, B.; Schramel, O.; Kettrup, A. Fresen. J. Anal. Chem. 1999, 363, 456-459. (46) Takada, Y.; Sakairi, M.; Koizumi, H. Anal. Chem. 1995, 67, 1474-1476. (47) Mol, R.; de Jong, G. J.; Somsen, G. W. Electrophoresis 2005, 26, 146–154. (48) Taichrib, A.; Pioch, M.; Neusüß, C. Electrophoresis 2012, 33, 1356-1366. (49) Haselberg, R.; de Jong, G. J.; Somsen, G. W. Electrophoresis 2013, 34, 99-112. (50) Guzman, N. A.; Majors, R. E. In LC·GC Europe 2001, pp 1-9. (51) Hernandez-Borges, J.; Neusüß, C.; Cifuentes, A.; Pelzing, M. Electrophoresis 2004, 25, 2257-2281. (52) Liu, C. C.; Zhang, J. Z.; Dovichi, N. J. Rapid Commun. Mass Spectrom. 2005, 19, 187-192. (53) Zamfir, A. D. J. Chromatogr. A 2007, 1159, 2-13. (54) Mokaddem, M.; Gareil, P.; Belgaied, J.-E.; Varenne, A. Electrophoresis 2009, 30, 1692-1697. (55) Lee, E. D.; Muck, W.; Henion, J. D.; Covey, T. R. Biomed. Environ. Mass Spectrom. 1989, 18, 844-850. (56) Niessen, V. M. A.; Tjaden, U. R.; Vandergreef, J. J. Chromatogr. 1993, 636, 3-19. (57) Pleasance, S.; Thibault, P.; Kelly, J. J. Chromatogr. 1992, 591, 325-339. (58) Maxwell, E. J.; Chen, D. D. Y. Anal. Chim. Acta 2008, 627, 25-33. (59) Nilsson, S.; Klett, O.; Svedberg, M.; Amirkhani, A.; Nyholm, L. Rapid Commun. Mass Spectrom. 2003, 17, 1535–1540. (60) Chen, Y.-R.; He, G.-R. Rapid Commun. Mass Spectrom. 2003, 17, 437–441. (61) Zamfir, A. D.; Dinca, N.; Sisu, E.; Peter-Katalinić, J. J. Sep. Sci. 2006, 29, 414–422. (62) Bendahl, L.; Hansen, S. H.; Olsen, J. Rapid Commun. Mass Spectrom. 2002, 16, 2333–2340. (63) Zhu, X.; Thiam, S.; Valle, B. C.; Warner, I. M. Anal. Chem. 2002, 74, 5405-5409. (64) Nilsson, S.; Wetterhall, M.; Bergquist, J.; Nyholm, L.; Markide, K. E. Rapid Commun. Mass Spectrom. 2001, 15, 1997–2000. (65) Dahlin, A. P.; Wetterhall, M.; Liljegren, G.; Bergström, S. K.; Andrén, P.; Nyholm, L.; Markides, K. E.; Bergquist, J. Analyst 2005, 130, 193-199. (66) Fang, L.; Zhang, R.; Williams, E. R.; Zare, R. N. Anal. Chem. 1994, 66, 3696-3701. (67) Cao, P.; Moini, M. J. Am. Soc. Mass Spectrom. 1997, 8, 561-564. (68) Moini, M. Anal. Chem. 2001, 73, 3497–3501. 146  (69) Moini, M. Anal. Chem. 2007, 79, 4241-4246. (70) Petersson, M. A.; Hulthe, G.; Fogelqvist, E. J. Chromatogr. A 1999, 854, 141-154. (71) Whitt, J. T.; Moini, M. Anal. Chem. 2003, 75, 2188–2191. (72) Figeys, D.; van Oostveen, I.; Ducret, A.; Aebersold, R. Anal. Chem. 1996, 68, 1822–1828. (73) Severs, J. C.; Smith, R. D. Anal. Chem. 1997, 64, 2154–2158. (74) Ishihama, Y.; Katayama, H.; Asakawa, N.; Oda, Y. Rapid Commun. Mass Spectrom. 2002, 16, 913–918. (75) Haselberg, R.; Ratnayake, C. K.; de Jong, G. J.; Somsen, G. W. J. Chromatogr. A. 2010, 1217, 7605-7611. (76) Maxwell, J. E.; Zhong, X.; Zhang, H.; van Zeijl, N.; Chen, D. D. Y. Electrophoresis 2010, 31, 1130-1137. (77) Douglas, D. J. Mass Spectrom. Rev. 2009, 28, 937-960. (78) Du, Z.; Douglas, D. J.; Konenkov, N. J. Anal. At. Spectrom. 1999, 14, 1111-1119. (79) Steel, C.; Henchman, M. J. Chem. Educ. 1998, 75, 1049-1054. (80) Morelle, W.; Michalski, J.-C. Curr. Anal. Chem. 2005, 1, 29-57. (81) March, R. E. J. Mass Spectrom. 1997, 32, 351–369. (82) Cooks, R. G.; Kaiser Jr., R. E. Acc. Chem. Res. 1990, 23, 213-219. (83) Hoffmann, E. D.; Stroobant, V. Mass spectrometry: principles and applications 3rd. Edition ed.; John Wiley: Chichester, England, 2007. (84) Sheeley, D. M.; Reinhold, V. N. Anal. Chem. 1998, 70, 3053-3059. (85) Viseux, N.; de Hoffmann, E.; Domon, B. Anal.Chem. 1998, 70, 4951-4959. (86) Delaney, J.; Vouros, P. Rapid Commun. Mass Spectrom. 2001, 15, 325-334. (87) Weiskopf, A. S.; Vouros, P.; Harvey, D. J. Rapid Commun. Mass Spectrom. 1997, 11, 1493-1504. (88) Weiskopf, A. S.; Vouros, P.; Harvey, D. J. Anal. Chem. 1998, 70, 4441-4447. (89) Hoffmann, E. D.; Stroobant, V. Mass spectrometry: principles and applications, 3rd. ed ed.; John Wiley: Chichester, England, 2007. (90) Haselberg, R.; de Jong, G. J.; Somsen, G. W. Electrophoresis 2011, 32, 66-82. (91) El Rassi, Z., Ed. Carbohydrate Analysis: High Performance Liquid Chromatography and Capillary Electrophoresis; Vol 58, Elsevier Amsterdam, 1995. (92) Apweiler, R.; Hermjakob, H.; Sharon, N. biochim. Biophys. Acta, Gen. Subj. 1999, 1473, 4-8. (93) Thibault, P., Honda, S. , Ed. Capillary Electrophoresis of Carbohydrates; Humana Press, 2003. (94) Dube, D. H.; Bertozzi, C. R. Nat. Rev. Drug Discov. 2005, 4, 477-4889. (95) Kyselova, Z.; Mechref, Y.; Al Bataineh, M. M.; Dobrolecki, L. E.; Hickey, R. J.; Vinson, J.; Sweeney, C. J.; Novotny, M. V. J. Proteom. Res. 2007, 6, 1822-1832. (96) de Leoz, M. L.; An, H. J.; Kronewitter, S.; Kim, J.; Beecroft, S.; Vinall, R.; Miyamoto, S.; de Vere White, R.; Lam, K. S.; Lebrilla, C. B. Dis. Markers 2008, 25, 243-258. (97) Norton, P. A.; Comunale, M. A.; Krakover, J.; Rodemich, L.; Pirog, N.; D'Amelio, A.; Philip, R.; Mehta, A. S.; Block, T. M. J. Cell. Biochem. 2008, 104, 136-149. (98) Winchester, B. Glycobiology 2005, 15, 1R-15R. (99) Rillahan, C.; Paulson, J. Annu. Rev. Biochem. 2011, 80, 797-823. (100) Mariño, K.; Bones, J.; Kattla, J. J.; Rudd, P. M. Nat. Chem. Biol. 2010, 6, 713-723. 147  (101) Pabst, M.; Altmann, F. Proteomics 2011, 11, 631-643. (102) Mechref, Y. Electrophoresis 2011, 32, 3467-3481. (103) An, H. J.; Kronewitter, S. R.; de Leoz, M. L. A.; Lebrilla, C. B. Curr. Opin. Chem. Biol. 2009, 13, 601-607. (104) Abd Hamid, U. M.; Royle, L.; Saldova, R.; Radcliffe, C. M.; Harvey, D. J.; Storr, S. J.; Pardo, M.; Antrobus, R.; Chapman, C. J.; Zitzmann, N.; Robertson, J. F.; Dwek, R. A.; Rudd, P. M. Glycobiology 2008, 18, 1105-1118. (105) Pierce, A.; Saldova, R.; Abd Hamid, U. M.; Abrahams, J. L.; McDermott, E. W.; Evoy, D.; Duffy, M. J.; Rudd, P. M. Glycobiology 2010, 20, 1283-1288. (106) Storr, S. J.; Royle, L.; Chapman, C. J.; Abd Hamid, U. M.; Robertson, J. F.; Murray, A.; Dwek, R. A.; Rudd, P. M. Glycobiology 2008, 18, 456-462. (107) Guile, G. R.; Rudd, P. M.; Wing, D. R.; Prime, S. B.; Dwek, R. A. Anal. Biochem. 1996, 240, 210-226. (108) Zauner, G.; M.;, D. A.; Wuhrer, M. Electrophoresis 2011, 32, 3456-3466. (109) Wuhrer, M.; Deelder, A. M.; Hokke, C. H. J. Chromatogr. B 2005, 825, 124-133. (110) Pabst, M.; Kolarich, D.; Poltl, G.; Dalik, T.; Lubec, G.; Hofinger, A.; Altmann, F. Anal. Biochem. 2009, 384, 263-273. (111) Chen, X.; Flynn, G. C. Anal. Biochem. 2007, 370, 147-161. (112) Ciucanu, I.; Kerek, F. Carbohydr. Res. 1984, 131, 209-217. (113) Davies, M. J.; Smith, K. D.; Carruthers, R. A.; Chai, W.; Lawson, A. M.; Hounsell, E. F. J. Chromatogr. 1993, 646, 317-326. (114) Chu, C. S.; Niñonuevo, M. R.; Clowers, B. H.; Perkins, P. D.; An, H. J.; Yin, H.; Killeen, K.; Miyamoto, S.; Grimm, R.; Lebrilla, C. B. Proteomics 2009, 9, 1939-1951. (115) Costello, C. E.; Contado-Miller, J. M.; Cipollo, J. F. J. Am. Soc. Mass Spectrom. 2007, 18, 1799-1812. (116) Ruhaak, L. R.; Deelder, A. M.; Wuhrer, M. Anal. Bioanal.Chem. 2009, 394, 163-174. (117) Palmisano, G.; Larsen, M. R.; Packer, N. H.; Thaysen-Andersen, M. RSC Adv. 2013, 3, 22706–22726. (118) Koizumi, K. J. Chromatogr. A 1996, 720, 119-126. (119) El Rassi, J. J. Chromatogr. A 1996, 720, 93-118. (120) Ludwiczak, P.; Brando, T.; Monsarrat, B.; Puzo, G. Anal. Chem. 2001, 73, 2323-2330. (121) Chen, F.-T.; Evangelista, R. Electrophoresis 1998, 19, 2639-2644. (122) Chiesa, C.; O'Neill, R. A. Electrophoresis 1994, 15, 1132-1140. (123) Frisch, E.; Kaup, M.; Egerer, K.; Weimann, A.; Tauber, R.; Berger, M.; Blanchard, V. Electrophoresis 2011, 32, 3510-3515. (124) Hutterer, K. M.; Birrell, H.; Camilleri, P.; Jorgenson, J. W. J. Chromatogr. B Biomed. Sci. Appl. 2000, 745, 365-372. (125) Kakehi, K.; Ueda, M.; Suzuki, S.; Honda, S. J. Chromatogr. 1993, 630, 141-146. (126) Kakehi, K.; Kinoshita, M.; Hayase, S.; Oda, Y. Anal. Chem. 1999, 71, 1592-1596. (127) Guttman, A. Electrophoresis 1997, 18, 1136-1141. (128) Nakajima, K.; Kinoshita, M.; Matsushita, T.; Urashima, M.; Suzuki, M.; Suzuki, S.; Kakehi, K. Anal. Biochem. 2006, 348, 105-114. (129) Scheinthal, B. M.; Bettelheim, F. A. Carbohydr. Res. 1968, 6, 257-265. (130) Zaia, J. OMICS 2010, 14, 401-418. (131) Zaia, J. Mass Spectrom. Rev. 2004, 23, 161-227. 148  (132) Novotny, M. V.; Mechref, Y. J. Sep. Sci. 2005, 28, 1956-1968. (133) Huang, L.; Riggin, R. M. Anal. Chem. 2000, 72, 3539-3546. (134) Que, A. H.; Novotny, M. V. Anal. Chem. 2002, 74, 5184-5191. (135) Nakano, M.; Kakehi, K.; Taniguchi, N.; Kondo, A. In Capillary Electrophoresis of Carbohydrates: From Monosaccharides to Complex Polysaccharides; Volpi, N., Ed.; Humana Press: New York 2011, pp 205-235. (136) Mechref, Y.; Novotny, M. Mass Spectrom. Rev. 2009, 28, 207-222. (137) Cortes, D. F.; Kabulski, J. L.; Lazar, A. C.; Lazar, I. M. Electrophoresis 2011, 32, 14-29. (138) Campa, C.; Coslovi, A.; Flamigni, A.; Rossi, M. Electrophoresis 2006, 27, 2027-2050. (139) Krenkova, J.; Foret, F. Proteomics 2012, 12, 2978-2990. (140) Zhao, S. S.; Zhong, X.; Tie, C.; Chen, D. D. Y. Proteomics 2012, 12, 2991-3012. (141) Huhn, C.; Selman, M. H.; Ruhaak, L. R.; Deelder, A. M.; Wuhrer, M. Proteomics 2009, 9, 882-913. (142) Morelle, W.; Michalski, J.-C. Nat. Protoc. 2007, 2, 1585-1602. (143) Thaysen-Andersen, M.; Mysling, S.; Højrup, P. Anal. Chem. 2009, 81, 3933-3943. (144) Chen, R.; Jiang, X.; Sun, D.; Han, G.; Wang, F.; Ye, M.; Wang, L.; Zou, H. J. Proteome Res. 2009, 8, 651-661. (145) Larsen, M. R.; Hojrup, P.; Roepstorff, P. Mol. Cell. Proteomics 2005, 4, 107-119. (146) Wu, S. L.; Huhmer, A. F.; Hao, Z.; Karger, B. L. J. Proteome Res. 2007, 6, 4230-4244. (147) Ruhaak, L. R.; Zauner, G.; Huhn, C.; Bruggink, C.; Deelder, A. M.; Wuhrer, M. Anal. Bioanal. Chem. 2010, 397, 3457–3481. (148) Lim, J. M.; Aoki, K.; Angel, P.; Garrison, D.; King, D.; Tiemeyer, M.; Bergmann, C.; Wells, L. J. Proteome Res. 2009, 8, 673-680. (149) O'Neil, R. A. J. Chromatogr. A 1996, 720, 201-215. (150) Karlsson, N. G.; Packer, N. H. Anal. Biochem. 2002, 305, 173-185. (151) Huang, Y.; Mechref, Y.; Novotny, M. V. Anal. Chem. 2001, 73, 6063-6069. (152) El Rassi, Z., Ed. Carbohydrate Analysis by Modern Chromatography and Electrophoresis; Vol. 66, Elsevier: Amsterdam, 2002. (153) Lamari, F. N.; Kuhn, R.; Karamanos, N. K. J. Chromatogr. B 2003, 793, 15-36. (154) Hanisch, F. G.; Jovanovic, M.; Peter-Katalinic, J. Anal. Biochem. 2001, 290, 47-59. (155) Chai, W.; Feizi, T.; Yuen, C.; Lawson, A. M. Glycobiology 1997, 7, 861-872. (156) Patel, T.; Bruce, J.; Merry, A.; Bigge, C.; Wormald, M. R.; Jaques, A.; Parekh, R. Biochemistry 1993, 32, 679-693. (157) Deguchi, K.; Takegawa, Y.; Ito, H.; Miura, N.; Yoshioka, S.; Nagai, S.; Nakagawa, H.; Nishimura, S. Rapid Commun. Mass Spectrom. 2006, 20, 412-418. (158) Royle, L.; Campbell, M. P.; Radcliffe, C. M.; White, D. M.; Harvey, D. J.; Abrahams, J. L.; Kim, Y. G.; Henry, G. W.; Shadick, N. A.; Weinblatt, M. E.; Lee, D. M.; Rudd, P. M.; Dwek, R. A. Anal. Biochem. 2008, 376, 1-12. (159) Wuhrer, M.; Deelder, A. M. Anal. Chem. 2005, 77, 6954-6959. (160) Kang, P.; Mechref, Y.; Klouckova, I.; Novotny, M. V. Rapid Commun. Mass Spectrom. 2005, 19, 3421-3428. (161) You, J.; Sheng, X.; Ding, C.; Sun, Z.; Suo, Y.; Wang, H.; Li, Y. Anal. Chim. Acta 2008, 609, 66-75. (162) Lattova, E.; Perreault, H. J. Chromatogr. B 2003, 793, 167-179. 149  (163) Bigge, J. C.; Patel, T. P.; Bruce, J. A.; Goulding, P. N.; Charles, S. M.; Parekh, R. B. Anal. Biochem. 1995, 230, 229-238. (164) Evangelista, R. A.; Guttman, A.; Chen, F. T. Electrophoresis 1996, 17, 347-351. (165) Anumula, K. R. Anal. Biochem. 1994, 220, 275-283. (166) Liu, J.; Shirota, O.; Novotny, M. V. J. Chromatogr. A 1991, 559, 223-235. (167) Anumula, K. R. Anal. Biochem. 2006, 350, 1-23. (168) Shilova, N. V.; Bovin, N. V. Russ. J. Bioorg. Chem. 2003, 29, 309-324. (169) Nashabeh, W.; El Rassi, J. J. Chromatogr. 1992, 600, 279-287. (170) Kamoda, S.; Ishikawa, R.; Kakehi, K. J. Chromatogr. A 2006, 1133, 332-339. (171) Anumula, K. R.; Dhume, S. T. Glycobiology 1998, 8, 685-694. (172) Guttman, A.; Starr, C. Electrophoresis 1995, 16, 993-997. (173) Klockow, A.; Amado, R.; Widmer, H. M.; Paulus, A. Electrophoresis 1996, 17, 110-119. (174) Ma, S.; Nashabeh, W. Anal. Chem. 1999, 71, 5185-5192. (175) Guttman, A.; Chen, F.-T.; Evangelista, R. A. Electrophoresis 1996, 17, 412-417. (176) Laroy, W.; Contreras, R.; Callewaert, N. Nat. Protoc. 2006, 1, 397-405. (177) Kamoda, S.; Kakehi, K. Electrophoresis 2008, 29, 3595-3604. (178) Guttman, A.; Chen, F.-T.; Evangelista, R. A.; Cooke, N. Anal. Biochem. 1996, 233, 234-242. (179) Evangelista, R. A.; Chen, F.-T.; Guttman, A. J. Chromatogr. A 1996, 745, 273-280. (180) Hermentin, P.; Doenges, R.; Witzel, R.; Hokke, C. H.; Vliegenthart, J. F. G.; Kamerling, J. P.; Conradt, H. S.; Nimtz, M.; Brazel, D. Anal. Biochem. 1994, 221, 29-41. (181) Guttman, A.; Pritchett, T. Electrophoresis 1995, 16, 1906-1911. (182) Hase, S. J. Chromatogr. A 1996, 720, 173-182. (183) Tomiya, N.; Awaya, J.; Kurono, M.; Endo, S.; Arata, Y.; Takahashi, N. Anal. Biochem. 1988, 171, 73-90. (184) Takahashi, N.; Nakagawa, H.; Fujikawa, K.; Kawamura, Y.; Tomiya, N. Anal. Biochem. 1995, 226, 139-146. (185) Doubet, S.; Albersheim, P. Glycobiology 1992, 2, 505. (186) Raman, R.; Raguram, S.; Venkataraman, G.; Paulson, J. C.; Sasisekharan, R. Nat. Methods 2005, 2, 817-824. (187) Ranzinger, R.; Herget, S.; Wetter, T.; von der Lieth, C. W. BMC Bioinformatics 2008, 9, 384-397. (188) Lutteke, T.; Bohne-Lang, A.; Loss, A.; Goetz, T.; Frank, M.; von der Lieth, C. W. Glycobiology 2006, 16, 71R-81R. (189) Cooper, C. A.; Joshi, H. J.; Harrison, M. J.; Wilkins, M. R.; Packer, N. H. Nucleic Acids Res. 2003, 31, 511-513. (190) Hashimoto, K.; Goto, S.; Kawano, S.; Aoki-Kinoshita, K. F.; Ueda, N.; Hamajima, M.; Kawasaki, T.; Kanehisa, M. Glycobiology 2006, 16, 63R-70R. (191) Domon, B.; Costello, C. E. Glycoconj. J. 1988, 5, 397-409. (192) Harvey, D. J. J. Mass Spectrom. 2000, 35, 1178-1190. (193) Morelle, W.; Faid, V.; Chirat, F.; Michalski, J.-C. Methods Mol. Biol. 2009, 534, 5-21. (194) Harvey, D. J. Mass Spectrom. Rev 2006, 25, 595-662. (195) North, S. J.; Hitchen, P. G.; Haslam, S. M.; Dell, A. Curr. Opin. Struct. Biol. 2009, 19, 498-506. 150  (196) Balaguer, E., Demelbauer, U., Pelzing, M., Sanz-Nebot, V., Barbosa, J., Neusüß, C. Electrophoresis 2006, 27, 2638-2650. (197) Ongay, S.; Neusüß, C. Anal. Bioanal. Chem. 2010, 398, 845-855. (198) Leymarie, N.; Zaia, J. Anal. Chem. 2012, 84, 3040-3048. (199) Wheeler, S. F.; Domann, P.; Harvey, D. J. Rapid Commun. Mass Spectrom. 2009, 23, 303-312. (200) Andre, S.; Kozar, T.; Kojima, S.; Unverzagt, C.; Gabius, H. J. Biol. Chem. 2009, 390, 557-565. (201) Pioch, M.; Bunz, S.-C.; Neusüß, C. Electrophoresis 2012, 33, 1517-1530. (202) Liu, T., Li, J., Zeng, R., Shao, X., Wang, K., Xia, Q. Anal. Chem. 2001, 73, 5875-5885. (203) Alley, W. R.; Mechref, Y.; Novotny, M. Rapid Commun. Mass Spectrom. 2009, 23, 495-505. (204) Kelly, J. F.; Locke, S. J.; Ramaley, L.; Thibault, P. J. Chromatogr. A 1996, 720, 409-427. (205) Neusüß, C.; Pelzing, M. In Methods in Molecular Biology, 2009; Vol. 492, pp 201-213. (206) Demelbauer, U. M.; Plematl, A.; Kremser, L.; Allmaier, G.; Josic, J.; Rizzi, A. Electrophoresis 2004, 25, 2026-2032. (207) Balaguer, E., Neusüß, C. Anal. Chem. 2006, 78, 5384-5393. (208) Balaguer, E., Neusüß, C. Chromatographia 2006, 64, 351-357. (209) Neusüß, C., Demelbauer, U., Pelzing, M. Electrophoresis 2005, 26, 1442-1450. (210) Robinson, L. N.; Artpradit, C.; Raman, R.; Shriver, Z. H.; Ruchirawat, M.; Sasisekharan, R. Electrophoresis 2012, 33, 797-814. (211) Bindila, L.; Peter-Katalinic, J.; Zamfir, A. Electrophoresis 2005, 26, 1488-1499. (212) Amon, S.; Plematl, A.; Rizzi, A. Electrophoresis 2006, 27, 1209-1219. (213) An, H. J.; Peavy, T. R.; Hedrick, J. L.; Lebrilla, C. B. Anal. Chem. 2003, 75, 5628-5637. (214) Henning, S.; Peter-Katalinic, J.; Pohlentz, G. J. Mass. Spectrom. 2007, 42, 1415-1421. (215) Neue, K.; Mormann, M.; Peter-Katalinic, J.; Pohlentz, G. J. Proteome Res. 2011, 10, 2248-2260. (216) Tie, C.; Zhang, X.-X. Anal. Methods. 2012, 4, 357-359. (217) Zhong, X.; Maxwell, J. E.; Chen, D. D. Y. Anal. Chem. 2011, 83, 4916-4923. (218) Zhong, X.; Maxwell, J. E.; Ratnayake, C.; Mack, S. T.; Chen, D. D. Y. Anal. Chem. 2011, 83, 8748-8755. (219) Maxwell, E. J.; Ratnayake, C.; Jayo, R.; Zhong, X.; Chen, D. D. Y. Electrophoresis 2011, 32, 2161-2166. (220) Joao, H.; Dwek, R. Eur. J. Biochem. 1993, 218, 239-244. (221) Sottani, C.; Fiorentino, M.; C.;, M. Rapid Commun. Mass Spectrom 1997, 11, 907-913. (222) Prien, J. M.; Ashline, D. J.; Lapadula, A. J.; Zhang, H.; Reinhold, V. N. J. Am. Soc. Mass Spectrom. 2009, 20, 539-556. (223) Hua, S.; Nwosu, C. C.; Strum, J. S.; Seipert, R. R.; An, H. J.; Zivkovic, A. M.; German, J. B.; Lebrilla, C. B. Anal. Bioanal. Chem. 2012, 403, 1291-1302. (224) Schneider, B. B.; Douglas, D. J.; Chen, D. D. Y. J. Am. Soc. Mass Spectrom. 2001, 12, 772-779. (225) Alvarez-Manilla, G.; Atwood III, J.; Guo, Y.; Warren, N. L.; Orlando, R.; Pierce, M. J. Proteome Res. 2006, 5, 701-708. (226) Fan, J.-Q.; Kondo, A.; Kato, I.; Lee, Y., C.; Anal. Biochem. 1994, 219, 224-229. 151  (227) Jensen, P. H.; Karlsson, N. G.; Kolarich, D.; Packer, N. H. Nat. Protoc. 2012, 7, 1299-1310. (228) Bones, J.; McLoughlin, N.; Hilliard, M.; Wynne, K.; Karger, B. L.; Rudd, P. M. Anal. Chem. 2011, 83, 4154-4162. (229) Houel, S.; Hilliard, M.; Yu, Y. Q.; MacLoughlin, N.; Martin, S. M.; Rudd, P. M.; Williams, J. P.; Chen, W. Anal. Chem. 2014, 86, 576-584. (230) Delobell, A.; Cantais, F.; Catrain, A.; Dereux, E.; Van Vyncht, G. Methods Mol. Biol. 2013, 988, 115-143. (231) Yuen, C.-T.; Zhou, Y.; Wang, Q.-Z.; Hou, J.-F.; Bristow, A.; Wang, J.-Z. Biologicals 2011, 39, 396-403. (232) Shahrokh, Z.; Royle, L.; Saldova, R.; Bones, J.; Abrahams, J. L.; Artemenko, N. V.; Flatman, S.; Davies, M.; Baycroft, A.; Sehgal, S.; Heartlein, M. W.; Harvey, D. J.; Rudd, P. M. Mol. Pharmaceutics 2011, 8, 286-296. (233) Reusch, D.; Haberger, M.; Kailich, T.; Heidenreich, A. K.; Kampe, M.; Bulau, P.; Wuhrer, M. mAbs 2014, 6, 185-196. (234) Ahn, J.; Bones, J.; Yu, Y.-Q.; Rudd, P. M.; Gilar, M. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 2010, 878, 403-408. (235) Pabst, M.; Grass, J.; Toegel, S.; Liebminger, E.; Strasser, R.; Altmann, F. Glycobiology 2012, 22, 389-399. (236) Higel, F.; Demelbauer, U.; Seidl, A.; Friess, W.; Sörgel, F. Anal. Bioanal. Chem. 2013, 405, 2481-2493. (237) Prater, B. D.; Connelly, H. M.; Qin, Q.; Cockrill, S. L. Anal. Biochem. 2009, 385, 69-79. (238) Anumula, K. R. Anal. Biochem. 2006, 220, 1-23. (239) Wada, Y.; Parastoo, A.; Costello, C. E.; Dell, A.; Dwek, R. A.; Geyer, H.; Geyer, R.; Kakehi, K.; Karlsson, N. G.; Kato, K.; Kawasaki, N.; Khoo, K. H.; Kim, S.; Kondo, A.; Lattova, E.; Mechref, Y.; Miyoshi, E.; Nakamura, K.; Narimatsu, H.; Novotny, M.; Packer, N. H.; Perreault, H.; Katalinic, J. P.; Gottfried, P.; Vermon, R. N.; Rudd, P. M.; Suzuki, A.; Tariguchi, N. Glycobiology 2007, 17, 411-422. (240) Zhao, S. S.; Chen, D. D. Y. Electrophoresis 2014, 35, 96-108. (241) Nakano, M.; Higo, D.; Arai, E.; Nakagawa, T.; Kakehi, K.; Taniguchi, N.; Kondo, A. Glycobiology 2009, 19, 135-143. (242) Gennaro, L. A.; Salas-Solano, O. Anal. Chem. 2008, 80, 3838-3845. (243) Maxwell, E. J.; Zhong, X.; Chen, D. D. Y. Anal. Chem. 2010, 82, 8377-8381. (244) Soliman, L. C.; Hui, Y.; Hewavitharana, A. K.; Chen, D. D. Y. J. Chromatogr. A 2012, 1267, 162-169. (245) Lindenburg, P. W.; Ramautar, R.; Jayo, R. G.; Chen, D. D. Y.; Hankemeier, T. Electrophoresis 2013. (246) Pabst, M.; Bondili, J. S.; Stadlmann, J.; Mach, L.; Altmann, F. Anal. Chem. 2007, 79, 5051-5057. (247) Wormald, M. R.; Rudd, P. M.; Harvey, D. J.; Chang, S.-C.; Scragg, I. G.; Dwek, R. A. Biochemistry 1997, 36, 1370-1380. (248) Kobata, A. Biochim. Biophys. Acta 2008, 1780, 472-478. (249) Wuhrer, M.; Stam, J. C.; van de Geijn, F. E.; Koeleman, C. A. M.; Verrips, C. T.; Dolhain, R. J. E. M.; Hokke, C. H.; Deelder, A. M. Proteomics 2007, 7, 4070-4081. 152  (250) Gahoual, R.; Burr, A.; Busnel, J. M.; Kuhn, L.; Hammann, P.; Beck, A.; François, Y. N.; Leize-Wagner, E. MAbs 2013, 5, 479-490. (251) Galuska, S. P.; Geyer, H.; Weinhold, B.; Kontou, M.; R hrich, R. C.; Bernard, U.; Gerardy-Schahn, R.; Reutter, W.; M nster-K hnel, A.; Geyer, R. Anal. Chem. 2010, 82, 4591-4598. (252) Royle, L.; Radcliffe, C. M.; Dwek, R. A.; Rudd, P. M. In Methods in Molecular Biology: Glycobiology Protocols Brockhausen, I., Ed.; Humana Press Inc.: Totowa, NJ., 2006; Vol. 347, pp 125-143. (253) Pucic, M.; Knezevic, A.; Vidic, J.; Adamczyk, B.; Novokmet, M.; Polasek, O.; Gornik, O.; Supraha-Goreta, S.; Wormald, M. R.; Redzic, I.; Campbell, H.; Wright, A.; Hastie, N. D.; Wilson, J. F.; Rudan, I.; Wuhrer, M.; Rudd, P. M.; Josicab, D.; Lauc, G. Mol. Cell. Proteomics 2011, 10, 1-15. (254) Stadlmann, J.; Pabst, M.; Kolarich, D.; Kunert, R.; Altmann, F. Proteomics 2008, 8, 2858-2871. (255) Rudd, P. M.; Morgan, B. P.; Wormald, M. R.; Harvey, D. J.; van den Berg, C. W.; Davis, S. J.; Ferguson, M. A.; Dwek, R. A. J. Biol. Chem. 1997, 272, 7229-7244. (256) Sasaki, H.; Bothner, B.; Dell, A.; Fukuda, M. J. Biol. Chem. 1987, 262, 12059-12076. (257) Hokke, C. H.; Bergwerff, A. A.; Van Dedem, G. W. K.; Kamerling, J. P.; Vliegenthart, J. F. G. Eur. J. Biochem. 1995, 228, 981-1008. (258) Kawasaki, N.; Ohta, M.; Hyuga, S.; Hyuga, M.; Hayakawa, T. Anal. Biochem. 2000, 285, 82-91. (259) Chen, F.-T.; Dobashi, T. S.; Evangelista, R. A. Glycobiology 1998, 8, 1045-1052. (260) Watson, E.; Bhide, A.; van Halbeek, H. Glycobiology 1994, 4, 227-237. (261) Llop, E.; Gutierrez-Gallego, R.; Belalcazar, V.; Gerwig, G.; Kamerling, J. P.; Segura, J.; Pascual, J. A. Proteomics 2007, 7, 4278-4291. (262) Rice, K. G.; Takahashi, N.; Namiki, Y.; Tran, A. D.; Lisi, P. J.; Lee, Y. C. Anal. Biochem. 1992, 206, 278-287. (263) Harvey, D. J. Proteomics 2001, 1, 311-328. (264) Shahrokh, Z.; Royle, L.; Saldova, R.; Bones, J.; Abrahams, J. L.; Artemenko, N. V.; Flatman, S.; Davies, M.; Baycroft, A.; Sehgal, S.; Heartlein, M. W.; Harvey, D. J.; Rudd, P. M. Mol. Pharmaceutics 2011, 8, 286-296. (265) Llop, E.; Gutiérrez-Gallego, R.; Segura, J.; Mallorqui, J.; Pascual, J. A. Anal. Biochem. 2008, 383, 243-254. (266) Fukuda, M. N.; Sasaki, H.; Lopez, L.; Fukuda, M. Blood 1989, 73, 84-89. (267) Storring, P. R.; Tiplady, R. J.; Gaines Das, R. E.; Stenning, B. E.; Lamikanra, A.; Rafferty, B.; Lee, J. Br. J. Haematol. 1998, 100, 79-89. (268) Tsuda, E.; Goto, M.; Murakami, A.; Akai, K.; Ueda, M.; Kawanishi, G.; Takahashi, N.; Sasaki, R.; Chiba, H.; Ishihara, H.; Mori, M.; Tejima, S.; Endo, S.; Arata, Y. Biochemistry 1988, 27, 5646-5654. (269) Nimtz, M.; Martin, W.; Wray, V.; Kloppel, K.-D.; Augustin, J.; Conradt, H. S. Eur. J. Biochem. 1993, 213, 39-56. (270) Irie, A.; Koyama, S.; Kozutsumi, Y.; Kawasaki, T.; Suzuki, A. J. Biol. Chem. 1998, 273, 15866-15871. 153  (271) Chou, H. H.; Takematsu, H.; Diaz, S.; Iber, J.; Nickerson, E.; Wright, K. L.; Muchmore, E. A.; Nelson, D. L.; Warren, S. T.; Varki, A. Proc. Natl. Acad. Sci. U.S.A 1998, 95, 11751–11756. (272) Hokke, C. H.; Bergwerff, A. A.; Van Dedem, G. W. K.; Oostrum, J. V.; Kamerling, J. P.; Vliegenthart, J. F. G. FEBS Letters 1990, 275, 9-14. (273) Tangvoranuntakul, P.; Gagneux, P.; Diaz, S.; Bardor, M.; Varki, N.; Varki, A.; Muchmore, E. PNAS 2003, 100, 12045-12050. (274) Le Floch, F.; Tessier, B.; Chenuet, S.; Guillaume, J.-M.; Cans, P.; Marc, A.; Georgen, J.-L. Biotecnol. Prog. 2004, 20, 864-871. (275) Skibeli, V.; Nissen-Lie, G.; Torjesen, P. Blood 2001, 98, 3626-3634. (276) Everest-Dass, A. V.; Jin, D.; Thaysen-Andersen, M.; Nevalainen, H.; Kolarich, D.; Packer, N. H. Glycobiology 2012 22, 1465-1479. (277) Nakano, M.; Saldanha, R.; Göbel, A.; Kavallaris, M.; Packer, N. H. Mol. Cell. Proteomics 2011, 10, M111 009001. (278) Zhang, Z.; Zhang, F.; Liu, Y. J. Chromatogr. Sci. 2013, 51, 666-683. (279) Ullsten, S.; Zuberovic, A.; Wetterhall, M.; Hardenborg, E.; Markides, K. E.; Bergquist, J. Electrophoresis 2004, 25, 2090-2099. (280) Varki, A., Diaz, S. Anal. Biochem. 1984, 137, 236-247. (281) Varki, A. Glycobiology 1992, 2, 25-40. (282) Schauer, R. Glycobiology 1991, 1, 449-452. (283) Vandammme-Feldhaus, V., Schauer, R. J. Biochem. 1998, 124, 111-121. (284) Klein, A.; Krishna, M.; Varki, N. M.; Varki, A. Proc. Natl. Acad. Sci. USA 1994, 91, 7782-7786. (285) Schauer, R. Advan. Carb. Chem. Biochem. 1982, 40, 131-234. (286) Varki, A.; Diaz, S. J. Biol. Chem. 1983, 258, 12465-12471. (287) Orskov, F.; Orskov, I.; Sutton, A.; Schneerson, R.; Lin, W.; Egan, W.; Hoff, G.; Robbins, J. J. Exp. Med. 1979, 149, 669-685. (288) Rogers, G. N.; Herrier, G.; Paulson, J. C.; Klenk, H. D. J. Biol. Chem. 1986, 261, 5947-5951. (289) Kamerling, J., Schauer, R., Shukla, A., Stoll, S., Van Halbeek, H., Vliegenthart, J. Eur. J. Biochem. 1987, 162, 601-607. (290) Varki, A.; Kornfeld, S. J. Exp. Med. 1980, 152, 532-544. (291) Srinivasan, G., Schauer, R. Glycoconj. J. 2009, 26, 935-944. (292) Kohla, G., Stockfleth, E., Schauer, R. Neurochem Res. 2002, 7-8, 583-592. (293) Butor, C., Diaz, S., Varki, A. J. Biol. Chem. 1993, 268, 10197-10206. (294) Ylönen, A., Kalkkinen, N., Saarinen, J., Bøgwald, J., Helin, J. Glycobiology 2001, 11, 523-531. (295) Ylönen, A., Helin, J., Bøgwald, J., Jaakola, A., Rinne, A., Kalkkinen, N. Eur. J. Biochem. 2002, 269, 2639-2646. (296) Liu, X., Afonso, L., Altman, E., Johnson, S., Brown, L., Li, J. Proteomics 2008, 8, 2849-2857. (297) Du, J.; Meledeo, M. A.; Wang, Z.; Khanna, H. S.; Paruchuri, V. D. P.; Yarema, K. J. Glycobiology 2009, 19, 1382-1401. (298) Klein, A., Diaz, S., Ferreira, I., Lamblin, G.,  Roussel, P., Manzi, A. Glycobiology 1997, 7, 421-432. 154  (299) He, Y., Lacher, N., Hou, W., Wang, Q., Isele, C., Starkey, J., Ruesch, M. Anal. Chem. 2010, 82, 3222-3230. (300) Sandra, K., Beeumen,J., Stals, I., Sandra,P., Claeyssens, M., Devreese, B. Anal. Chem. 2004, 76, 5878-5886. (301) Chen, F., Evangelista, R. Anal. Biochem. 1995, 230, 273-280. (302) Suzuki, H., Müller, O., Guttman, A., Karger, B. Anal. Chem. 1997, 69, 4554-4559. (303) Kabel, M., Heijnis, W., Bakx, E., Kuijpers, R., Voragen, A., Schols, H. J. Chromatogr. A 2006, 1137, 119-126. (304) Guttman, A. Nature 1996, 380, 461-462. (305) Higa, H. H., Manzi, A., Varki, A. J. Biol. Chem. 1989, 264, 19435-19442. (306) Zimmer, G.; Reuter, G.; Schauer, R. Eur. J. Biochem. 1992, 204, 209-215. (307) Packer, N. H.; von der Lieth, C. W.; Aoki-Kinoshita, K. F.; Lebrilla, C. B.; Paulson, J. C.; Raman, R.; rudd, P. M.; Sasisekharan, R.; Taniguchi, N.; York, W. Proteomics 2008, 8, 8-20. (308) Zhang, X. L. Curr. Med.Chem. 2006, 13, 1141-1147. (309) Guan, F.; Uboh, C. E.; Soma, L. R.; Birks, E.; Chen, J.; You, Y.; Rudy, J.; Li, X. Anal. Chem. 2008, 80, 3811-3817. (310) Alley, W. R. J.; Madera, M.; Mechref, Y.; Novotny, M. V. Anal. Chem. 2010, 82, 5095-5106. (311) Hongsachart, P.; Huang-Liu, R.; Sinchaikul, S.; Pan, F. M.; Phutrakul, S.; Chuang, Y. M.; Yu, C. J.; Chen, S. T. Electrophoresis 2009, 30, 1206-1220. (312) Zhao, J.; Patwa, T. H.; Qiu, W.; Shedden, K.; Hinderer, R.; Misek, D. E.; Anderson, M. A.; Simeone, D. M.; Lubman, D. M. J. Proteom. Res. 2007, 6, 1864-1874. (313) Scott, D. W.; Chen, J.; Chacko, B. L.; Traylor, J. G. J.; Orr, A. W.; Patel, R. P. Arterioscler. Thromb. Vasc. Biol. 2012, 32, 51-59. (314) Dube, R.; Rook, G. A.; Steele, J.; Brealey, R.; Dwek, R. A.; Rademacher, T. W.; Lennard-Jones, J. Gut. 1990, 31, 431-434. (315) Arnold, J. N.; Saldova, R.; Abd Hamid, U. M.; Rudd, P. M. Proteomics 2008, 8, 3284-3293. (316) Imre, T.; Kremmer, T.; Héberger, K.; Molnár-Sz llősi, E.; Ludányi, K.; Pócsfalvi, G.; Malorni, A.; Drahos, L.; Vékey, K. J. Proteomics 2008, 71, 186-197. (317) Kronewitter, S. R.; de Leoz, M. L. A.; Peacock, K. S.; McBride, K. R.; An, H. J.; Miyamoto, S.; Leiserowitz, G. S.; Lebrilla, C. B. J. Proteome Res. 2010, 9, 4952-4959. (318) Isailovic, D.; Kurulugama, R. T.; Plasencia, M. D.; Stokes, S. T.; Kyselova, Z.; Goldman, R.; Mechref, Y.; Novotny, M. V.; Clemmer, D. E. J. Proteome Res. 2008, 7, 1109-1117. (319) Ruhaak, L. R.; Uh, H. W.; Beekman, M.; Hokke, C. H.; Westendorp, R. G.; Houwing-Duistermaat, J.; Wuhrer, M.; Deelder, A. M.; Slagboom, P. E. J. Proteom. Res. 2011, 10, 1667-1674. (320) Bones, J.; Mittermayr, S.; O’Donoghue, N.; Guttman, A.; Rudd, P. M. Anal. Chem. 2010, 82, 10208-10215. (321) Ni, W.; Bones, J.; Karger, B. L. Anal. Chem. 2013, 85, 3127–3135. (322) Hua, S.; An, H. J.; Ozcan, S.; Ro, G. S.; Soares, S.; De Vere-White, R.; Lebrilla, C. B. Analyst 2011, 136, 3663-3671. (323) Vanderschaeghe, D.; Laroy, W.; Sablon, E.; Halfon, P.; Van Hecke, A.; Delanghe, J.; Callewaert, N. Mol. Cell Proteomics 2009, 5, 986-949. 155  (324) Dwek, R. A. Chem. Rev. 1996, 96, 683-720. (325) Aldredge, D.; An, H. J.; Tang, N.; Waddell, K.; Lebrilla, C. B. J. Proteome Res. 2012, 11, 1958-1968. (326) Ruhaak, L. R.; Huhn, C.; Waterreus, W.-J.; de Boer, A. R.; Neusüss, C.; Hokke, C. H.; Deelder, A. M.; Wuhrer, M. Anal. Chem. 2008, 80, 6119-6126. (327) Zhang, J.; Lamotte, L.; Dodds, E. D.; Lebrilla, C. B. Anal. Chem. 2005, 77, 4429-4438. (328) Ohnesorge, J.; Neusüß, C.; Wätzig, H. Electrophoresis 2005, 26, 3973-3987. (329) Brinkman-Van der Linden, E. C. M.; Havenaar, E. C.; Van Ommen, E. C. R.; Van Kamp, G. J.; Gooren, L. J. G.; Van Dijk, W. Glycobiology 1996, 6, 407-412. (330) Brinkman-Van der Linden, E. C.; Van Ommen, E. C.; Van Dijk, W. Glycoconj. J. 1996, 13, 27-31. (331) Turner, G. A. Clin. Chim. Acta 1992, 208, 149-171. (332) Ryden, I.; Pahlsson, P.; Lundblad, A.; Skogh, T. Clin. Chim. Acta 2002, 317, 221-229. (333) Rouse, J., Strang,A., Yu, W., Vath, J. Anal. Biochem. 1998, 256, 33-46.         156  Appendices Appendix A  A.1     CE-MS spectra of the glycoforms of RNAse B    02072012 rnase b cems 12 #1098 RT: 10.60 AV: 1 NL: 1.26E5F: + p NSI Full ms [1500.00-2000.00]1550 1600 1650 1700 1750 1800 1850 1900 1950 2000m/z05101520253035404550556065707580859095100Relative Abundance1864.131657.601876.201956.401843.931902.531998.271755.201710.531806.931763.871559.531641.331528.801598.00[M+8H]8+[M+9H]9+RNase B-Man 5(A) 02072012 rnase b cems 12 #1119 RT: 10.80 AV: 1 NL: 1.80E5F: + p NSI Full ms [1500.00-2000.00]1550 1600 1650 1700 1750 1800 1850 1900 1950 2000m/z0510520530540550556065707580859095100Relative Abundance1864.601884.931657.671675.671904.531924.271978.531755.401850.731776.601713.471560.671639.131611.931542.00[M+8H]8+[M+9H]9+RNase B-Man 6(B) 157     RNase B-Man 702072012 rnase b cems 12 #1133 RT: 10.93 AV: 1 NL: 1.38E5F: + p NSI Full ms [1500.00-2000.00]1550 1600 1650 1700 1750 1800 1850 1900 1950 2000m/z05101520253035404550556065707580859095100Relative Abundance1864.801884.601905.271657.601675.671924.871693.201710.801982.331943.601757.531848.001782.801643.131560.931596.471506.93[M+8H]8+[M+9H]9+(C) 02072012 rnase b ce s 12 #1144 RT: 11.04 AV: 1 NL: 1.15E5F: + p NSI Full s [1500.00-2000.00]1550 1600 1650 1700 1750 1800 1850 1900 1950 2000m/z05101520253035404550556065707580859095100Relative Abundance1865.001884.531925.471904.871711.801657.931675.731945.671964.131729.801972.001816.601775.931614.201576.131535.93[M+8H]8+[M+9H]9+RNase B-Man 8(D) 158   Figure A. 1CE-MS spectra for the glycoforms of RNase B. In all cases [M+8H]8+ is the most abundant charge state observed. CE-MS spectrum corresponding to (A) RNase B-Man5, (B) RNase B-Man6, (C) RNase B-Man7, (D) RNase B-Man8 and (E) RNase B-Man9. Data was acquired using the LCQ*Duo ion trap MS.    A.2     CE-MS spectra for T3-labeled glycans derived from RNase B    RNase B-Man 902072012 rnase b cems 12 #1155 RT: 11.15 AV: 1 NL: 8.26E4F: + p NSI Full ms [1500.00-2000.00]1550 1600 1650 1700 1750 1800 1850 1900 1950 2000m/z05101520253035404550556065707580859095100Relative Abundance1864.401885.131925.331945.331905.131712.271675.871729.671954.801990.401844.201779.131817.271614.671576.131553.80[M+8H]8+[M+9H]9+(E)  +Q1: 31.947 to 32.248 min from Sample 6 (02162012 RNase B PNGase F 06) of 02162012 RNase B PNGase F.wiff (Turbo Spray) Max. 3.5e6 cps.900 950 1000 1050 1100 1150 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650 1700 1750 1800 1850 1900m/z, amu2.0e54.0e56.0e58.0e51.0e61.2e61.4e61.6e61.8e62.0e62.2e62.4e62.6e62.8e63.0e63.1e6Intensity, cps985.01471.11147.01309.11493.1162.0162.1162.0[T3-HexNAc2Hex5+H]+[T3-HexNAc2Hex4+H]+[T3-HexNAc2Hex3+H]+[T3-HexNAc2Hex2+H]+[T3-HexNAc2Hex5 + Na]+(A)159      +Q1: 32.916 to 33.217 min from Sample 6 (02162012 RNase B PNGase F 06) of 02162012 RNase B PNGase F.wiff (Turbo Spray) Max. 2.8e6 cps.900 950 1000 1050 1100 1150 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650 1700 1750 1800 1850 1900m/z, amu5.0e41.0e51.5e52.0e52.5e53.0e53.5e54.0e54.5e55.0e55.5e56.0e56.5e57.0e57.5e58.0e58.5e59.0e59.5e51.0e61.1e61.1e61.2e61.2e61.3e61.3e61.4e6Intensity, cps1147.0985.11309.01633.3162.0161.9162.3162.0[T3-HexNAc2Hex6+H]+[T3-HexNAc2Hex5+H]+[T3-HexNAc2Hex4+H]+[T3-HexNAc2Hex3+H]+[T3-HexNAc2Hex2+H]+(B) + 1: 29.909 to 30.043 in fro  Sa ple 8 (02162012 RNase B PNGase F 08) of 02162012 RNase B PNGase F.wiff (Turbo Spray), su... Max. 1.6e6 cps.900 950 1000 1050 1100 1150 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650 1700 1750 1800 1850 1900m/z, amu1. 52.03.0e54.0e55.67.0e58.0e59.010 61.1e61.2e61.3e641.5e61.6e6Intensity, cps906.91047.6920.5987.91050.11188.21795.41121.2931.41304.61226.51034.11436.71469.81406.21879.3942.3 1778.61629.11586.91080.3 1657.41838.71545.3[T3-HexNAc2Hex7+2H]2+(C)160     Figure A. 2 Mass spectra corresponding to (A) T3-labeled GlcNAc2Man5 identified as [T3-HexNAc2Hex5+H]+ at m/z 1471.1; (B) T3-labeled GlcNAc2Man6 identified as [T3-HexNAc2Hex6+H]+ at m/z 1633.3; (C) T3-labeled GlcNAc2Man7 identified as [T3-HexNAc2Hex7+2H]2+ at m/z 906.9;(D) T3-labeled GlcNAc2Man8  identified as [T3-HexNAc2Hex8+2H]2+ at m/z 987.9 and (E) T3-labeled GlcNAc2Man9  identified as [T3-HexNAc2Hex9+2H]2+ at m/z 1069.0. Data was acquired using the API 4000 triple-quadrupole MS.    +Q1: 30.778 to 30.878 min from Sample 8 (02162012 RNase B PNGase F 08) of 02162012 RNase B PNGase F.wiff (Turbo Spray), su... Max. 2.5e6 cps.900 950 1000 1050 1100 1150 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650 1700 1750 1800 1850 1900m/z, amu1.0e52.0e53.0e54.0e55.0e56.0e57.0e58.0e59.0e51.0e61.1e61.2e61.3e61.4e61.5e61.6e61.7e61.8e61.9e62.0e62.1e62.2e62.3e62.4e62.5e6Intensity, cps987.91128.81001.4979.61269.3998.91127.71069.2923.11410.3 1495.61142.1 1255.9 1395.91284.11628.21297.5 1789.51715.3 1859.01572.5 1644.6981.51424.5[T3-HexNAc2Hex8+2H]2+(D) +Q1: 31.379 to 31.480 min from Sample 8 (02162012 RNase B PNGase F 08) of 02162012 RNase B PNGase F.wiff (Turbo Spray), su... Max. 1.4e6 cps.900 950 1000 1050 1100 1150 1200 1250 1300 1350 1400 1450 1500 1550 1600 1650 1700 1750 1800 1850 1900m/z, amu5.0e41.0e51.5e52.0e52.5e53.0e53.5e54.0e54.5e55.0e55.5e56.0e56.5e57.0e57.5e58.0e58.5e59.0e59.5e51.0e61.1e61.1e61.2e61.2e61.3e61.3e61.4e6Intensity, cps1069.01072.31209.91012.91082.41212.0980.41350.81093.8902.21006.21136.41365.7 1509.41791.51075.81663.31474.51243.8 1531.9 1586.01715.51821.31252.0[T3-HexNAc2Hex9+2H]2+(E)161  Appendix B B.1     CE-MS analysis of N-glycans from rHuEPO using the maXisTM ultra-high resolution TOF MS  Table B.1 Monosaccharide composition, theoretical mass and observed mass (Da) for the total number of N-glycans observed from rHuEPO. This table includes acetylation, Neu5Ac/Neu5Gc variation and the occurrence of LacNAc repeats. Glycans were observed as [M-2H]2-, [M-3H]3- and/or [M-4H]4-species. Data was obtained using the maXisTM ultra high-resolution TOF MS.  Glycan Composition Theor.Mass (Da)Obs.Mass (Da)Err. (ppm)Observed m/zH N F SAcetylationNeu5Ac/NeuGc [M-2H]2-[M-3H]3-[M-4H]4-5 4 1 2 2370.1597 2370.1618 -0.89 1184.07295 4 1 2 1 2386.1689 2386.1706 -0.71 1192.07735 4 1 2 1 2412.1523 2412.1506 0.7 1205.06736 5 1 1 2444.2302 2444.2282 0.82 1221.10615 4 1 2 2 2454.1528 2454.1547 -0.77 1226.06946 5 1 2 2735.4983 2735.5005 -0.8 1366.74236 5 1 2 1 2777.4977 2777.5001 -0.86 1387.74216 5 1 2 1 2751.4803 2751.4789 0.51 1374.73156 5 1 2 1 1 2793.4899 2793.4914 -0.54 1395.73776 5 1 2 1 2 2809.4867 2809.4891 -0.85 1403.73666 5 1 2 2 2 2851.4823 2851.4795 0.98 1424.73186 5 1 3 3026.7469 3026.7491 -0.73 1512.3666 1007.90846 5 1 3 1 3042.7318 3042.7343 -0.82 1013.23686 5 1 3 1 3068.7477 3068.7501 -0.78 1021.90877 6 1 2 3100.8256 3100.8269 -0.42 1549.4055 1032.6017 6 1 2 1 3142.8277 3142.8291 -0.45 1570.4066 1046.60178 7 1 1 3174.9091 3174.9101 -0.31 1586.4471 1057.29548 7 1 1 1 3216.9127 3216.9155 -0.87 1607.4498 1071.29728 7 1 1 1 1 3232.8901 3232.8887 0.43 1615.43647 6 1 3 3391.0954 3391.0923 0.91 1694.5082 1129.35617 6 1 3 1 3410.0878 3410.0856 0.65 1135.68727 6 1 3 1 3433.1035 3433.101 0.73 1715.5425 1143.3597 6 1 3 1 1 3450.0675 3450.0648 0.78 1149.01368 7 1 2 3466.1691 3466.1702 -0.32 1154.38217 6 1 3 2 3475.1122 3475.1097 0.72 1736.5469 1157.36198 7 1 2 1 3482.1582 3482.1601 -0.55 1159.7127 6 1 3 2 1 3492.0714 3492.0688 0.74 1163.01498 7 1 2 1 3508.1528 3508.1556 -0.8 1168.37727 6 1 3 3 3517.1144 3517.1127 0.48 1757.5484 1171.36299 8 1 1 3540.2456 3540.2483 -0.76 1769.1162 1179.07488 7 1 2 2 3550.1589 3550.1573 0.45 1182.37787 6 1 3 4 3559.1104 3559.1112 -0.22 1185.36249 8 1 1 1 3582.2469 3582.2447 0.61 1193.07369 8 1 1 2 3624.2397 3624.2419 -0.61 1207.07267 6 1 4 3682.1862 3682.1836 0.71 1226.7718 919.53797 6 1 4 1 3699.184 3699.182 0.54 1232.105 923.78757 6 1 4 1 3725.1978 3725.1964 0.38 930.2911162        Nomenclature of the glycan monosaccharide composition as indicated in Figure 3.2.     Glycan Composition Theor.Mass (Da)Obs.Mass (Da)Err. (ppm)Observed m/zH N F SAcetylationNeu5Ac/NeuGc [M-2H]2-[M-3H]3-[M-4H]4-7 6 1 4 1 1 3741.3406 3741.338 0.69 1246.1046 934.32658 7 1 3 3757.4278 3757.43 -0.61 1877.7071 1251.46877 6 1 4 2 3767.352 3767.354 -0.42 1254.7765 940.83048 7 1 3 1 3773.4196 3773.422 -0.66 1256.79947 6 1 4 2 1 3783.3489 3783.352 -0.93 1260.1094 944.83018 7 1 3 1 3799.411 3799.409 0.61 1265.46167 6 1 4 3 3809.3425 3809.34 0.76 951.32697 6 1 4 3 1 3825.3324 3825.33 0.73 1274.1019 955.32449 8 1 2 3831.4901 3831.488 0.44 1276.15488 7 1 3 2 3841.4123 3841.415 -0.68 1279.46367 6 1 4 4 3851.3319 3851.33 0.49 961.82457 6 1 4 4 1 3867.3311 3867.328 0.7 1288.1015 965.82419 8 1 2 1 3873.4927 3873.49 0.75 1290.15537 6 1 4 5 3893.3311 3893.329 0.59 972.324210 9 1 1 3905.5779 3905.581 -0.87 1300.85247 6 1 4 6 3935.3379 3935.34 -0.64 1310.7721 982.827110 9 1 1 1 3947.5712 3947.573 -0.33 1314.84957 6 1 4 7 3977.3318 3977.334 -0.65 1324.7702 993.32568 7 1 4 4048.307 4048.31 0.64 1348.429 1011.06948 7 1 4 1 4063.2887 4063.292 -0.71 1353.423 1014.81498 7 1 4 1 4089.3209 4089.317 0.9 1021.32138 7 1 4 1 1 4108.29 4108.287 0.68 1368.421 1026.06389 8 1 3 4122.7563 4122.754 0.58 1373.2433 1029.688 7 1 4 2 4131.3243 4131.322 0.46 1376.099 1031.82268 7 1 4 2 1 4148.2723 4148.27 0.65 1381.749 1036.05949 8 1 3 1 4164.7503 4164.749 0.38 1387.24168 7 1 4 3 4174.3309 4174.333 -0.55 1390.436 1042.57538 7 1 4 3 1 4191.3245 4191.3212 0.79 1395.018 1046.82238 7 1 4 4 4216.3376 4216.335 0.66 1404.5453 1053.07578 7 1 4 5 4257.3447 4257.343 0.35 1418.106 1063.32788 7 1 4 6 4299.3361 4299.333 0.74 1432.103 1073.82529 8 1 4 4414.0199 4414.018 0.5 1470.3312 1102.49610 9 1 3 4488.0996 4488.103 -0.78 1495.026410 9 1 4 4779.3589 4779.363 -0.92 1592.1131163  B.2      Analysis of glycosylation of rHuEPO by LC-MS/MS Table B. 2 Identification and annotation of native N-glycans of rHuEPO observed with PGC LC-MS/MS. Glycans were observed as [M-2H]2-, [M-3H]3- and/or [M-4H]4- ion species. MS Data was acquired using the HCT 3-D ion trap.     Nomenclature of the glycan monosaccharide composition as indicated in Figure 3.2. Glycan Composition Theo. Mass (Da)Obs. Mass (Da)Observed m/zH N F S Acetylation [M-2H]2-[M-3H]3-[M-4H]4-5 4 1 1787.6 1785.6 892.34 4 1 1 1916.7 1914.6 956.84 4 1 1 1 1958.7 1956.7 977.85 4 1 1 2078.8 2076.7 1037.85 4 1 1 1 2120.8 2118.7 1058.86 5 1 2152.9 2150.7 1074.85 4 1 2 2370.1 2367.8 1183.46 5 1 1 2444.2 2441.8 1220.47 6 1 2518.3 2515.9 1257.46 5 1 2 2735.4 2732.9 1365.9 910.37 6 1 1 2809.5 2807.0 1403.0 934.97 6 1 1 1 2851.5 2849.0 1424.0 949.06 5 1 3 3026.7 3024.0 1511.5 1007.37 6 1 2 3100.8 3098.0 1548.5 1032.037 6 1 2 1 3142.8 3140.1 1569.5 1046.07 6 1 2 2 3184.8 3182.1 1590.5 1060.07 6 1 3 3392.0 3389.1 1694.0 1129.0 846.57 6 1 3 1 3434.0 3431.2 1715.1 1143.0 857.07 6 1 3 2 3476.0 3473.2 1157.08 7 1 2 3466.1 3463.2 1153.78 7 1 2 1 3508.1 3505.2 1167.77 6 1 4 3683.3 3680.2 1226.0 919.37 6 1 4 1 3725.3 3722.3 1240.1 929.88 7 1 3 3757.4 3754.3 1250.7 937.88 7 1 3 1 3799.4 3796.3 1264.7 948.38 7 1 3 2 3841.4 3838.3 1278.7 958.89 8 1 3 4122.7 4119.4 1372.4 1029.1164  Appendix C C.1      CE-LIF of Glucose ladders with MS-amenable buffers Several initial experiments were conducted in order to examine buffer effects on CE-LIF separations of labeled carbohydrates for further development of CE-MS methodologies. For this study, we used a standard sample of APTS-labeled glucose ladder. The glucose ladder consisted of a mixture of oligomers with lengths in the range of 1 to more than 20 glucose units (G1 to G20).  Derivatization with a sulfonated label, such as APTS, introduces strongly acidic groups to the glycans which thereby remain negatively charged even at low pH.  CE-LIF was performed on polyacrylamide-based neutral capillaries, using a selection of MS-friendly background electrolytes (BGE’s). Because electroosmotic flow (EOF) is minimal in polyacrylamide-based neutral capillaries, reverse polarity was used and detection of the glucose ladder components was made at the anode, with smaller species migrating faster than larger ones. Four MS-friendly BGE’s were tested: ammonium acetate, ammonium formate, acetic acid and formic acid in the concentration range of 10 to 30 mM. Although most BGE systems provided acceptable separation, ammonium acetate at low concentration (10 mM, pH 4.73) resulted in baseline resolution for each ladder component and was used in further analysis. Figure B.1 illustrates the CE-LIF electropherogram of APTS-labeled glucose ladder using the ammonium acetate buffer. Peaks migrating between 5 and 10 minutes correspond to the excess of APTS, which is not completely removed by the acetone precipitation process. The subsequent peaks correspond to the labeled oligomers from G1 to G20. The high sensitivity of the LIF allows observing up to 24 glucose units (G24) with good resolution and peak shape. The high resolution of the CE-LIF method suggested that the method could resolve a mixture of N-glycans having 165  very small structural differences, as for positional isomers, for example. For that reason, CE-LIF was applied for the analysis of APTS-labeled N-glycans of control fish sera.            However, direct application of the method resulted in no separation for the fish serum glycan components showing only one peak for the glycan mixture (data not shown). Several parameters were varied in order to optimize the electrophoretic conditions for the APTS-labeled N-glycans separation, including pH, buffer concentration, and separation voltage. Ammonium acetate buffers (pH 3.1-5.0) with a concentration from 15 to 50 mM were tested. At pH > 3.5 the migration time of the analytes increased as the buffer pH increased without giving any separation. Reduction in pH from 3.5 to 3.1 provided a slight improvement in separation, showing a series of co-migrating peaks, with pH 3.1 providing the better separation. Buffer concentration in the range from 15 to 50 mM (pH 3.1) was tested, with 15 and 20 mM giving similar separation. Buffer concentrations higher than 30 mM showed a slight improvement in separation with increasing buffer concentration; however, baseline resolution was not achieved 0 10 20 30 40 50 6005101520253035Time(min)RFUFigure C. 1 Capillary electrophoresis separation of APTS-labeled glucose ladder standard with laser- induced fluorescence detection. 166  even at 50 mM. Moreover, buffer concentrations higher than 30 mM gave high CE currents (> 20 µA) that would not allow them to be used in future CE-MS analysis with our interface arrangement even at low CE voltages.76, 217-219 Further addition of methanol to the BGE’s considerably reduced the CE currents but resulted in longer migration times without improving the separation. Optimum buffer concentration was chosen at 25 mM as it was found to provide a good compromise among separation, peak shape, electrical current and analysis time. Finally, the separation voltage was also optimized, and a value of -20 kV was selected. C.2      CE-MS of APTS-labeled N-glycans of fish serum   Following our recent findings for the CE separation and ESI-MS detection of standard glycans under reverse EOF conditions,219 we initially evaluate the ionization properties of  the APTS-labeled N-glycans of fish serum using the LCQ*Duo ion trap MS. The conditions used for the CE-LIF analysis were slightly modified to allow ESI-MS detection of the APTS-labeled N-glycans: 25 mM ammonium acetate buffer (pH 3.1)/ 20% methanol. Although the labeled glycans were detected under these conditions, the MS signals were found to be up to one order of magnitude stronger when the concentration of the ammonium acetate buffer was reduced to 10 mM. The reason for this could be the lower background redox currents obtained with the 10 mM buffer which improved analyte ionization leading to stronger signals in the MS. Figure C.2 shows the base peak electropherogram for the mixture of APTS-labeled N-glycans. Only three ions can be observed at m/z 1355.01, 1376.09 and 1397.21 ([M-2H]2-), which is not in agreement with the number of glycan species detected by the CE-LIF (Fig. 4.1) and which will be discussed later. These three ions correspond to APTS-labeled N-glycans of bi-antennary oligosaccharides 167  with the addition of different numbers of acetyl groups present in control fish sera, as reported by Liu et al.296              In terms of migration time and resolution, differences were observed between CE-LIF and CE-ESI-MS electropherograms. The reasons for this are the modified electrophoretic conditions: reduced buffer concentration, organic solvent present in the buffer, increased injection amount required for MS detection, differences in capillary length and extra-column effects caused by the modifier solution while performing CE-ESI-MS. From the point of view of separation, the better resolution observed in CE-LIF is due to two intimately related factors. First, the buffer concentration used for LIF detection was higher than that used for ESI-MS detection (25 mM and 10 mM, respectively); high buffer concentration in ESI-MS generated high CE currents and led to corona discharge and unstable operations. In order to reduce the CE currents methanol was added to the BGE, which also made the BGE similar in composition to 27 28 29 30 31 32020406080100  Relative Abundance (%)Time (min)29.231355.0129.501376.1129.611397.21Figure C. 2 Base peak electropherogram for the CE-MS analysis of APTS-labeled N-glycans of control fish serum. BGE: 10 mM ammonium acetate (pH 3.1)/ 20% MeOH. Data acquired using the LCQ* Duo ion trap MS. 168  that of the modifier solution. However, this further increased the migration time of the glycans resulting in broader peaks, leading to a reduced separation. In order to evaluate the effect of the type of electrolyte on the separation, BGE’s consisting of acetic acid and formic acid (0.2-2%) with 10-30% methanol were also tested. Separations using acetic acid as the primary electrolyte component gave poor resolution for the mixture of APTS-labeled N-glycans. Different methanol content (i.e., 10%, 20%, and 30%) or concentrated acetic acid solutions did not lead to better results. Use of increasing amounts of formic acid resulted in increased analysis time due to the high content of methanol required to reduce the CE current, without improving the separation. Of the conditions tested, the BGE composed of 1% formic acid and 30% methanol allowed obtaining stronger signal intensities for the APTS-labeled N-glycans but separation was not improved in comparison to that shown in Figure C.2.  As discussed before, APTS-labeling enables sensitive CE-LIF detection of O-acetylated N-glycans in samples of control fish sera.  However, APTS derivatization is not mandatory when ESI-MS detection is used, since these analytes can be detected in negative ESI mode, which indeed turns out to be superior to positive ESI detection due to the almost complete dissociation of the sialic acids at pH 3.1 (pKa sialic acid ≈ 2.0-2.8).129 Moreover, introduction of extra negative charges could be detrimental for CE separation with ESI-MS detection since APTS derivatization imparts three additional charges on these molecules, resulting in highly negative iso-charged species with the same migration velocity. The CE separation mechanism is based on different analyte mobility expressed in terms of charge-to-size ratio of the analytes. Considering this, the acetyl groups located at the sialic acids are too small to significantly alter the hydrodynamic volumes of large glycans having similar structure. As a result, the charge-to-size 169  ratio of the APTS-labeled N-glycans will be almost the same and there will be small discrepancies in the electrophoretic mobilities of the labeled glycans, yielding almost no separation, if any.   C.3     Optimization of CE-MS conditions for native N-glycan analysis   Initially, typical MS-compatible acidic BGE’s based on ammonium acetate, formic acid and acetic acid were tested. The latter two, did not provide successful separations but a slight enhancement was achieved when the ammonium acetate buffer was used. However, a considerable improvement in separation was obtained when ε-Aminocaproic acid was employed as BGE. ε-Aminocaproic acid is a zwitterionic buffer that has been used to effectively suppress protein adsorption to the capillary wall and to perform glycoform characterization of proteins.196, 207, 208 The pH of the BGE (pH 3.03) was chosen to be slightly above the pKa of sialic acid to promote its dissociation, allowing the analysis of negatively charged species without compromising spray stability and signal intensity. BGE concentration was varied in the range of 25 to 50 mM to find the optimum concentration that would give acceptable resolution at a reasonable peak shape. Figure C.3 illustrates the effect of variable concentrations of ε-Aminocaproic acid on the resolution, ESI response and electrophoretic profile of the native O-acetylated fish serum glycans. It was observed that a gradual increase in BGE concentration resulted in improved separation efficiency and resolution for the glycan mixture. Higher BGE concentrations (> 50 mM) were not employed due to large ESI currents already generated with the 50 mM BGE. Higher buffer concentrations would negatively impact the spray stability and the ESI response. The optimum buffer concentration was chosen to be 40 mM (Fig.4.2), as it 170  provided the best resolution for the glycan components without compromising analysis time, signal abundances or ESI currents.                          46 48 50 52 54 56 58 600204060801001153.02  52.251132.052.6253.091111.0453.451132.0353.761153.07(A)Relative Abundance (%)Time (min)50 52 54 56 58 60 62 64020406080100Relative Abundance (%)  56.471132.0256.781 53.0357.391111.0157.791132.1058.231153.00BTime (min)66 68 70 72 74 76 78 80020406080100Relative Abundance (%)   72.631 32.0173.16153.0374.0011 1.0074.491132.0375.051153.00(C)Ti e ( in)Figure C.3 Base peak electropherogram of underivatized O-acetylated N-glycans of fish serum at variable concentrations of the BGE. (A) 25 mM, (B) 30 mM and (C) 50 mM ε-Aminocaproic acid. All BGE’s contain 20% methanol. Separations were performed at -20 kV.  In all cases, data was acquired using the LCQ* Duo ion trap MS. 171  C.4      Potential identification of isomeric N-glycans in O-acetylated bi-antennary oligosaccharides   Figure C.4 shows the CE-MS/MS spectrum for an O-acetylated bi-antennary oligosaccharide at m/z 1195.0 whose isomers have been baseline separated. It can be noticed that the same fragments are present in the three isomers but with different intensity. Those fragments are due to glycosidic cleavages at m/z 373.7 Da, cross-ring cleavages at m/z 331.9 Da and loss of water adducts at m/z 272 Da and 314.0 Da, which correspond to di-O-acetylated Neu5Ac, mono-O-acetylated Neu5Ac and neutral losses from Neu5Ac and mono-O-acetylated Neu5Ac, respectively. Consequently, at this stage we are not able to identify the location of the O-acetyl groups in the sialic acids. However, given the structural characteristics of the highly O-acetylated oligosaccharides, it can be hypothesized that di-O-acetylated species comprise the three possible structures, carrying O-acetyl groups in 7-, 8-, and 9-OH of Neu5Ac.  As stated by Rouse et al.,333 identification of glycan isomers is possible only if the proper isomeric standards are available. At this stage, due to the lacks of individual standards for O-acetylated Neu5Ac, we determined the identity of the separated isomers based on the relative intensities of the fragment ions corresponding to the di-O-acetyl Neu5Ac in the CE-ESI-MS/MS spectrum. Figure C.4 shows that fragment ions at nominal m/z 374.0 have relative intensities of (A) 1.15 e3, (B) 9.0 e3 and (C) 2.1 e4, approximately which indicate their precursor di-O-acetyl Neu5Ac oligosaccharide. As previously discussed, Neu5,9Ac2 is the most common sialic acid derivative and it often appears to be accompanied by the 7-O-acetyl analogue in variable amounts.289, 298, 305 It has been observed, that O-acetyl groups at the 7-OH of Neu5Ac can undergo spontaneous migration to the 9-OH of Neu5Ac under physiological conditions increasing the concentration of Neu5,9Ac2.289 172     -MS2 (1195.00) CE (-70): 21.295 to 21.696 min from Sample 1 (FSG 15 week 0 MS MS 1195.0) of DJLRJ110829B10.wiff (Ion Spray),... Max. 4329.8 cps.200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400m/z, amu100200300400500600700800900100011001200130014001500160017001800190020002100Intensity, cps314.0331.9373.7272.6202.0363.5262.2254.1218.0211.2398.6242.4225.0249.5390.0283.0322.2293.0301.4405.2350.0325.9358.0232.0305.1297.7367.0339.0 402.0276.1258.0 280.0352.0228.0 376.0294.8204.0 266.0239.0 380.0246.0205.8 393.2285.6(A)Neu5Ac –H2ONeu5,7, 8 Ac3Neu5,7 Ac2Neu5,7 Ac2-H2O% Intensitym/z -  ( . )  (-70): 23.167 to 23.5 8 min from Sample 1 (FSG 15 week 0 MS MS 1195.0) of DJLRJ110829B10.wiff (Ion Spray),... Max. 9087.9 cps.200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400m/z, amu5100015002000250030354000450050005500606700075008000850090Intensity, cps373.9314.0332.0272.0.391.6340.4327.5219.5255.2320.5903371.0381.8261.9 298.3351.5283.0243.0226.2 405.6343.0357.8210.396.0259.03 6.4.9(B)Neu5Ac –H2ONeu5,8, 9 Ac3Neu5,9 Ac2Neu5,9 Ac2-H2O% Intensitym/z173   Figure C. 4 CE-MS/MS spectrum acquired using the API 3000 triple quadrupole MS for an O-acetylated bi-antennary oligosaccharide at m/z 1195.0. Glycan was present in samples taken from week 2 of the stress experiment. Proposed glycan structure at nominal m/z 374.0 is (A) Neu5,7,8 Ac3, (B) Neu5,8,9 Ac3 and (C) Neu5,7,9 Ac3.                         -MS2 (1195.00) CE (-70): 22.331 to 22.598 min from Sample 1 (FSG 15 week 0 MS MS 1195.0) of DJLRJ110829B10.wiff (Ion Spray), ... Max. 2.1e4 cps.200 210 220 230 240 250 260 270 280 290 300 310 320 330 340 350 360 370 380 390 400m/z, amu1000.02000.03000.04000.05000.06000.07000.08000.09000.01.0e41.1e41.2e41.3e41.4e41.5e41.6e41.7e41.8e41.9e42.0e42.1e4Intensity, cps374.0332.0314.0271.5253.4211.1225.9390.0262.6382.0406.0287.6 363.4358.4248.0339.0269.0 298.1246.0234.0 345.4308.3 322.1370.0276.0325.9(C)Neu5Ac –H2ONeu5,7, 9 Ac3Neu5,9 Ac2Neu5,9 Ac2-H2O% Intensitym/z174  C.5       Potential N-glycan isomers  Table C.5. Observed m/z corresponding to bi- and tri-antennary N-glycans depending on the period of the stress experiment. N-glycans were present in control (week 0) and stress fish serum samples (week 1 to 4).         Type of GlycansWeek Bi-antennary glycans Tri-antennary glycansObs. m/z Charge state Obs. m/z Charge state01111.10[M-2H]2-1132.041153.0411111.10[M-2H]2-959.73[M-3H]3-1132.01 973.401153.02 987.131173.90 1001.271194.80 1015.5021111.05[M-2H]2-987.13[M-3H]3-1132.00 1001.301153.02 1015.501173.80 1029.401195.00 1043.4031111.06[M-2H]2-959.40[M-3H]3-1132.00 973.401153.001173.801195.0041111.10[M-2H]2-973.40[M-3H]3-1132.02 987.101153.03 1001.301173.801195.00175  Appendix D D.1     Optimization of denaturation temperature and time for enzymatic N-glycan release The effect of the denaturation temperature was evaluated by analyzing in triplicate each of the temperature conditions for the enzymatic release of glycans from a control sample of human serum. 20 µL of human serum (adjusted to 50 µL with 10 mM Tris-HCl at pH 8.0) was mixed with 0.1 % sodium dodecyl sulfate and 50 mM β-mercaptoethanol at room temperature, 60  C and 80  C for 10 minutes.After the enzymatic digestion with PNGase F, samples were purified by SPE and analyzed by CE-ESI-MS. The relative abundances of the ten most abundant glycans were used as an approximation to compare glycan yields. Relative abundances for glycans incubated at room temperature were 25% lower in average than those incubated at 60  C, while abundances for glycans incubated at 60  C and 80  C were found to be similar (Figure D.1A). For that reason, 60  C was chosen as the optimum incubation temperature and was applied in the following experiments.  In order to optimize the denaturation time, two control serum samples were analyzed in triplicate. The denaturation of serum glycoproteins, in the presence of sodium dodecyl sulfate and β-mercaptoethanol, was performed at 60  C for 10 and 30 min. After the enzymatic digestion with PNGase F, samples were purified by SPE and analyzed by CE-ESI-MS. The relative abundances of the ten most abundant glycans were found to be almost the same for incubations performed at 10 and 30 min, and incubation times of 10 min were chosen for further experiments (Figure D.1B). The amount PNGase F was used according to the manufacturer’s specifications.    176  (A)    (B)    Figure D.1. (A) Comparison of the average yield of the ten most abundant N-glycans o tained  y denaturation of serum glycoproteins for  0 minutes at room temperature,  0  C and  0  C. (B) Comparison of the a erage yield of the same glycans o tained after denaturation of serum glycoproteins at  0  C for 10 minutes and 30 minutes. The average relative abundance was calculated based on the CE-MS response of three samples per reaction condition. The tables below each graph show the peak area of the glycans, and their identity is as follows: peak 1, H3N5F1; peak 2, H5N4S1; peak 3, H5N4F1S1; peak 4, H5N4S2; peak 5, H5N5F1S1; peak 6, H5N4F1; peak 7, H5N5F1S2; peak 8, H6N5S2; peak 9, H6N5F1S2; peak 10, H6N5F1S3, 05101520253035401 2 3 4 5 6 7 8 9 10Temp = 60˚ C 0.96 17.56 5.80 36.20 2.87 3.91 1.85 3.60 0.90 0.69Temp = 80˚ C 0.91 17.32 5.54 35.35 3.86 3.89 2.22 2.57 1.14 0.63Rom Temp 0.49 12.15 4.35 28.07 2.34 2.89 1.43 2.11 0.96 0.85Relative  abundance (%)05101520253035401 2 3 4 5 6 7 8 9 1010 min 0.96 17.56 5.80 36.20 2.87 3.91 1.85 3.60 0.90 0.6930 min 0.90 17.00 5.01 36.41 2.14 4.56 1.67 3.48 1.00 0.49Relative abundance (%)177  where H= hexose, N= N-acetyl hexosamine, F= fucose and S= sialic acid. The numbers next to the monosaccharides above indicates the amount of each monosaccharide in the glycan composition.    D.2     Representative CE-MS spectra of isomeric N-glycans from human serum  Figure D.2 CE-MS spectra of some acidic N-glycans released from human serum. Glycan composition as follows: (A) H5N4S2 observed at m/z 1110.9, (B) H6N5S3 observed at m/z 959.0, (C) H6N5F1S3 observed at m/z 1008.1 and (D) H6N5F2S3 observed at m/z 1057.0. For glycans composition, the following notation is used hexose (H), N-acetylhexosamine (N), fucose (F) and N-acetylneuraminic acid (S).   -MS2 (1111.10) CE (-40): 20.994 to 21.428 min from Sample 2 (20120707 Prostate Cancer Serum MSMS 1111 02) of 20120707 Glyca... Max. 9.2e6 cps.600 650 700 750 800 850 900 950 1000 1050 1100 1150 1200m/z, amu5.0e51.0e61.5e62.0e62.5e63.0e63.5e64.0e64.5e65.0e65.5e66.0e66.5e67.0e67.5e68.0e68.5e69.0e6Intensity, cps1110.91060.41101.9(A)H5N4S2m/z= 1110.9 -MS2 (959.40) CE (-30): 16.909 to 17.140 min from Sample 7 (20120707 Prostate Cancer Serum MSMS 959  07) of 20120707 Glycans... Max. 1.4e6 cps.200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950 1000m/z, amu5.0e41.0e51.5e52.0e52.5e53. 53.5e54.0e54.5e55.0e55. 56.0e56.5e570 57.5e58.0e58 59.0e59.5e51.0e61.1e61.1e61.2e61.2e61.31.3e61.4e61.4e6Intensity, cps959.0925.3953.0891.3(B)H6N5S3m/z= 959.0 -MS2 (1 08.30) CE (-30): 17.024to 17.293 min from Sample 12(20120707 Prostate Cancer Serum MSMS 1008  12) of 20120707 Glyc... Max. 3.1e6 cps.200 250 300 3 400 450 500 550 6 650 700 750 800 8 900 950 1000 1050 11 0 1150 1200m/z, amu2.0e54.0e56.0e58.0e51.0e61.2e61.4e61.6e61.8e62.0e62.2e62.4e62.6e62.8e63.0e63.1e6Intensity, cps1008.1974.21002.1(C)H6N5F1S3m/z= 1008.1 -MS2 (1057.40) CE (-30): 17.101 to 17.409 min from Sample 9(20120707 Prostate Cancer Serum MSMS 1057  09) of 20120707 Glyca... Max. 2.2e5 cps.200 250 300 350 400 450 500 550 6 0 650 700 750 800 850 900 950 1000 1050 11 0 1150 1200m/z, amu1.0e42.0e43.0e44.0e45.0e46 47.0e48.0e49 41.0e51.1e51.2e51.3e51.4e51.5e5.6 51.7e51.8e519 52.0e52.1e52.2e5Intensity, cps1057.01023.61051.0290.01008.4960.1(D)H6N5F2S3m/z= 1057.0

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0166079/manifest

Comment

Related Items