ABSTRACT 
 
 
 
 
Title of Dissertation: EXPLORING ENDOGLYCOSIDASES FOR 
ANTIBODY GLYCOENGINEERING  
  
 Xin Tong, Doctor of Philosophy, 2019 
  
Dissertation directed by: Professor, Dr. Lai-Xi Wang, Department of 
Chemistry and Biochemistry 
 
 
Monoclonal antibodies (mAbs) comprise a rapidly growing class of 
therapeutics with great potential in treating infections, inflammations, cancers, and 
autoimmunities. The glycosylation of an antibody determines its functional efficacy 
and structural integrity, but stringent control of the intrinsically heterogeneous N-
glycans on an antibody remains a formidable challenge. Recent development of a 
chemoenzymatic glycosylation remodeling strategy using a family of carbohydrate-
modifying enzymes called endoglycosidases is emerging as an attractive method for 
producing homogeneous antibody glycoforms. The success of this method depends 
on the discovery of efficient endoglycosidases and glycosynthase mutants.  
Here, we describe mutagenesis studies on several endoglycosidases and their 
applications in expanding the current chemoenzymatic glycoengineering strategy to 
different antibodies. Five projects related to this effort are described here. First, 
glycosynthase variants of Endo-S were generated by mutagenesis to provide 
  
enhanced transglycosylation activities and diminished tendency to hydrolyze the 
product. Second, mutational studies on another endoglycosidase, Endo-S2, identified 
novel glycosynthase variants with broader substrate specificity and higher catalytic 
efficiency than Endo-S mutants. This work also provided the first kinetic studies for 
Endo-S and Endo-S2 mutants, with important mechanistic implications. Third, the 
unique properties of Endo-S allowed us to develop an alternative glycan remodeling 
strategy to synthesize several antibody glycoforms that were not easily accessible 
using the conventional glycoengineering approaches. This new method can be applied 
in a facile, one-pot fashion to modify antibody glycosylation without the need for 
purifying intermediates or switching enzymes. Fourth, to further expand the toolbox 
for antibody glycoengineering, the substrate specificity of a newly discovered 
endoglycosidase Endo-CC was characterized. The highly flexible selectivity of this 
enzyme for protein framework substrates paved the way for glycoengineering of 
additional antibody isotypes, particularly IgE, as a single N-glycan was found to be 
indispensable for the biological functions. Finally, remodeling of IgE glycosylation 
with alternative glycan structures was achieved for the first time by endoglycosidases. 
These novel IgE glycoforms displayed distinct binding properties for IgE receptors 
and revealed important new aspects of the structure-function relationship of IgE 
antibody glycosylation. Together, these studies should facilitate the development of 
novel antibody-based therapeutics that are optimized in their glycosylation patterns.   
 
 
 
  
 
 
 
 
 
 
EXPLORING ENDOGLYCOSIDASES FOR ANTIBODY GLYCOENGINEERING 
 
 
 
by 
 
 
Xin Tong 
 
 
 
 
 
Dissertation submitted to the Faculty of the Graduate School of the  
University of Maryland, College Park, in partial fulfillment 
of the requirements for the degree of 
Doctor of Philosophy 
2019 
 
 
 
 
 
 
 
 
 
 
Advisory Committee: 
Professor Lai-Xi Wang, Chair 
Professor Douglas Julin 
Professor Kwaku Dayie 
Professor Jason Kahn 
Professor Xiaoping Zhu 
 
 
 
 
 
  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
? Copyright by 
Xin Tong 
2019 
 
  
 
Acknowledgements 
I would like to express my sincere gratitude to everyone who helped me with 
my study and research at graduate school. I received a lot of help and support from 
people around me, and I would not have completed my degree without them. 
I would like to first thank my research advisor, Prof. Lai-Xi Wang, for his 
mentoring and guidance during all these years of research. Thank you for teaching me 
not only the knowledge and skills but also the scientific way of thinking and writing.  
I want to thank my committee members, Dr. Julin, Dr. LaRonde, Dr. Dayie, 
Dr. Kahn, and Dr. Zhu for their constructive criticism. Thank you all for helping me 
learn how to communicate my research with the general scientific community. 
I am extremely grateful to the current and past members of Wang group for 
their contribution to my research. I thank Dr. Tiezheng Li for being my mentor in the 
first few years of my graduate study. I also would like to thank Dr. Chao Li and Dr. 
Qiang Yang for their intellectual input in helping me face the challenges in several of 
my research projects. I enjoyed my time working as a graduate student in Wang lab. 
I am very grateful to all my friends, especially Jingheng Wang, who shared 
his valuable experience as a graduate student with me on a regular basis and helped 
me go through the final and most difficult stage of my degree.  
Finally, I would like to thank my parents for their long-standing support and 
encouragement. Without their understanding and financial help, I would never have 
the courage to travel so far to pursue education overseas. I love you very much!
 ii 
 
 
 
Table of Contents 
Acknowledgements ....................................................................................................... ii	
 ?
List of Tables ................................................................................................................ x	
 ?
List of Figures .............................................................................................................. xi	
 ?
List of Abbreviations ................................................................................................. xvi	
 ?
Chapter 1: Introduction ................................................................................................. 1	
 ?
A.	
 ? Introduction to antibodies .............................................................................. 1	
 ?
1.	
 ?Monoclonal antibody-based therapeutics ......................................................... 1	
 ?
2.	
 ?Overview of antibody structure and function ................................................... 4	
 ?
B.	
 ? Antibody glycosylation and its importance in biological functions ............ 9	
 ?
1.	
 ?Structure and functional diversity of antibody glycosylation ......................... 10	
 ?
2.	
 ?Glycosylation modulates effector functions of IgG and IgE antibodies ......... 11	
 ?
3.	
 ?Structural heterogeneity of antibody N-glycans ............................................. 14	
 ?
C.	
 ? Engineering antibody glycosylation by biosynthetic manipulation .......... 16	
 ?
1.	
 ?Biosynthetic pathways of N-glycans in eukaryotic expression systems ......... 17	
 ?
2.	
 ?Antibody glycan remodeling by biosynthetic pathway perturbation .............. 20	
 ?
D.	
 ? Antibody glycosylation remodeling by chemoenzymatic modifications .. 23	
 ?
1.	
 ?Introduction to endo-?-N-acetylglucosaminidase (ENGase) .......................... 25	
 ?
2.	
 ?Catalytic mechanism of ENGases and development of sugar oxazolines as 
activated glycan substrates for transglycosylation .......................................... 27	
 ?
3.	
 ?Glycosynthase mutants of ENGases for antibody glycan remodeling ............ 29	
 ?
4.	
 ?Examples of ENGase-mediated antibody glycoengineering .......................... 32	
 ?
 iii 
 
 
E.	
 ? Scope and limitations of current chemoenzymatic antibody remodeling 
strategy mediated by ENGases ............................................................................ 37	
 ?
1.	
 ?Efficiency of transglycosylation catalyzed by existing Endo-S glycosynthase 
mutants Endo-S D233Q and Endo-S D233A .................................................. 37	
 ?
2.	
 ?Substrate specificity of current ENGases for different N-glycans .................. 38	
 ?
3.	
 ?Glycosylation remodeling of antibodies beyond the IgG isotype ................... 38	
 ?
F.	
 ? Organization of the dissertation ................................................................... 40	
 ?
Chapter 2: Generation and comparative analysis of new glycosynthase mutants from 
endoglycosidase S (Endo-S) for antibody glycoengineering ...................................... 43	
 ?
A.	
 ? Introduction ................................................................................................... 43	
 ?
B.	
 ? Results ............................................................................................................ 51	
 ?
1.	
 ?Generation of additional glycosynthase mutants of Endo-S through site-
directed saturation mutagenesis at the catalytic residue Asp-233. ................. 51	
 ?
2.	
 ?Transglycosylation and hydrolysis activities of Endo-S mutants ................... 52	
 ?
3.	
 ?Enzyme kinetic analysis of Endo-S mutants D233M and D233A .................. 57	
 ?
4.	
 ?Chemical synthesis of complex glycan thiazoline inhibitor for Endo-S ......... 59	
 ?
5.	
 ?Inhibitory activity assays for the synthetic thiazoline-based inhibitor ........... 61	
 ?
6.	
 ?Investigation on non-enzymatic side reactions of glycan oxazoline .............. 62	
 ?
C.	
 ? Discussion ....................................................................................................... 66	
 ?
D.	
 ? Conclusion ..................................................................................................... 71	
 ?
E.	
 ? Materials and Methods ................................................................................. 72	
 ?
1.	
 ?Chemicals and biochemical materials ............................................................. 72	
 ?
2.	
 ?Mutagenesis, expression, and purification of Endo-S and mutants ................ 73	
 ?
 iv 
 
 
 
3.	
 ?Synthesis of deglycosylated antibody by wild-type Endo-S as acceptor 
substrate for transglycosylation ...................................................................... 74	
 ?
4.	
 ?Synthesis of S2G2F rituximab by mutant Endo-S D233Q as starting material 
for hydrolysis assays ....................................................................................... 75	
 ?
5.	
 ?Liquid Chromatography-Mass Spectrometry (LC-ESI-MS) analysis of 
antibody glycoforms ....................................................................................... 75	
 ?
6.	
 ?Quantification of hydrolysis and transglycosylation reaction products using 
internal standard and a single-point normalization factor ............................... 76	
 ?
7.	
 ?Transglycosylation and hydrolysis activity assays for Endo-S mutants ......... 76	
 ?
8.	
 ?Measurements of KM and Vmax of Endo-S mutants for SCT-oxa ................... 77	
 ?
9.	
 ?Measurements of KM and Vmax of Endo-S mutants for antibody .................... 78	
 ?
10.	
 ?Chemical synthesis of complex glycan-based thiazoline inhibitor ............... 78	
 ?
11.	
 ?Inhibition assays of thiazoline glycan inhibitor against ENGases ................ 79	
 ?
12.	
 ?Comparison of non-enzymatic glycosylations by enzyme treatment. .......... 80	
 ?
F.	
 ? Supporting Information ................................................................................ 82	
 ?
Chapter 3: Development of alternative antibody glycoengineering methods with 
relaxed substrate specificity for different glycan structures ....................................... 83	
 ?
A.	
 ? Introduction ................................................................................................... 83	
 ?
B.	
 ? Results ............................................................................................................ 87	
 ?
1.	
 ?Cloning, expression, and characterization of Endo-S2 ................................... 87	
 ?
2.	
 ?Generation of glycosynthase mutants of Endo-S2 .......................................... 87	
 ?
3.	
 ?Generation and comparative studies on the hydrolytic and transglycosylation 
activities of 19 glycosynthase mutants and wild-type enzyme of Endo-S2 .... 88	
 ?
 v 
 
 
 
4.	
 ?Enzyme kinetic parameters of Endo-S2 D184M and Endo-S2 D184A .......... 92	
 ?
5.	
 ?Endo-S2 D184M showed relaxed glycan substrate specificity ....................... 94	
 ?
6.	
 ?Transglycosylation of different N-glycans by Endo-S2 D184M .................... 97	
 ?
7.	
 ?Transglycosylation efficiencies of Endo-S2 and Endo-S mutants. ................. 98	
 ?
8.	
 ?Antibody transglycosylation using reduced amount of glycan oxazoline. ... 100	
 ?
9.	
 ?Endo-S hydrolysis of antibody glycoforms and N-glycan oxazolines .......... 102	
 ?
10.	
 ?Endo-S transglycosylation of antibody with additional glycans ................. 106	
 ?
11.	
 ?One-pot glycosylation remodeling of rituximab via wild-type Endo-S ...... 109	
 ?
12.	
 ?One-pot glycosylation remodeling of commercial antibody with modified 
glycan oxazolines by wild-type Endo-S ....................................................... 112	
 ?
C.	
 ? Conclusion ................................................................................................... 115	
 ?
D.	
 ? Materials and Methods ............................................................................... 118	
 ?
1.	
 ?Chemical and biochemical materials ............................................................ 118	
 ?
2.	
 ?Expression and purification of Endo-S WT and Endo-SD233Q .................. 118	
 ?
3.	
 ?Mutagenesis, expression, and purification of Endo-S2 and mutants ............ 119	
 ?
4.	
 ?Liquid Chromatography Mass Spectrometry (LC-ESI-MS) analysis ........... 120	
 ?
5.	
 ?HPAEC-PAD analysis of glycans and oxazolines ........................................ 121	
 ?
6.	
 ?Comparison of Endo-S hydrolysis against rituximab N-glycoforms ............ 122	
 ?
7.	
 ?Comparison of Endo-S hydrolysis towards different oxazolines ................. 122	
 ?
8.	
 ?Evaluation of Endo-S transglycosyaltion on deglycosylated rituximab using 
SCT-Oxa, HM-Oxa and Hyb-Oxa as glycosyl donors. ................................ 122	
 ?
 vi 
 
 
 
9.	
 ?Evaluation of transglycosylation activities of Endo-S WT and Endo-S D233Q 
towards deglycosylated rituximab antibody using HM-Oxa 5 and Hyb-Oxa 6 
as glycosyl donors. ........................................................................................ 123	
 ?
10.	
 ?Transglycosylation of rituximab with Hyb-Oxa and HM-Oxa by Endo-S WT 
in a one-pot manner. ..................................................................................... 124	
 ?
11.	
 ? Transglycosylation of rituximab with Man3-Oxa and AzideMan3-Oxa by 
Endo-S WT in a one-pot fashion. ................................................................. 124	
 ?
E.	
 ? Supporting Information .............................................................................. 125	
 ?
Chapter 4: Revisiting the substrate specificity of endoglycosidase Endo-CC and its 
mutants in hydrolysis and transglycosylation ........................................................... 126	
 ?
A.	
 ? Introduction ................................................................................................. 126	
 ?
B.	
 ? Results and Discussions .............................................................................. 131	
 ?
1.	
 ?Cloning, expression, and comparison of recombinant Endo-CC and Endo-CC 
fusion protein (CPD) ..................................................................................... 131	
 ?
2.	
 ?Generation of glycosynthase mutants from recombinant Endo-CC ............. 133	
 ?
3.	
 ?Substrate specificity of Endo-CC in hydrolysis of glycoproteins ................. 135	
 ?
4.	
 ?Studies on the substrate specificity of Endo-CC in transglycosylation ........ 140	
 ?
    C.	
 ?	
 ?	
 ?	
 ?Conclusion .................................................................................................... 146	
 ?
D.	
 ? Materials and Methods ............................................................................... 147	
 ?
1.	
 ?Chemical and biochemical materials ............................................................ 147	
 ?
2.	
 ?Mutagenesis, expression, and purification of recombinant Endo-CC .......... 148	
 ?
3.	
 ?Transglycosylation using wild-type Endo-CC and its mutants ..................... 149	
 ?
4.	
 ?Transglycosylation by Endo-CC N180H with different N-glycans .............. 149	
 ?
 vii 
 
 
5.	
 ?Enzymatic deglycosylation of ribonuclease B .............................................. 150	
 ?
6.	
 ?Enzymatic deglycosylation of sialylglycopeptide (SGP) with and without the 
core fucose .................................................................................................... 150	
 ?
7.	
 ?Differential deglycosylation of erythropoietin (EPO) containing a mixture of 
core-fucosylated and non-fucosylated N-glycans. ........................................ 151	
 ?
8.	
 ?Differential deglycosylation of antibody glycoforms with core-fucosylated, 
non-fucosylated complex type, and high-mannose N-glycans. .................... 151	
 ?
9.	
 ?Differential transglycosylation of deglycosylated antibody by wild-type Endo-
CC and Endo-CC N180H.............................................................................. 152	
 ?
E.	
 ? Supporting Information .............................................................................. 153	
 ?
Chapter 5: Glycosylation remodeling of IgE antibody and effects of Fc N-glycans on 
Fc?R binding ............................................................................................................. 154	
 ?
A.	
 ? Introduction ................................................................................................. 154	
 ?
B.	
 ? Results and Discussion ................................................................................ 158	
 ?
1.	
 ?Selective glycosylation remodeling of commercial and recombinant full-
length IgE antibodies and binding analyses on IgE N-glycoforms ............... 158	
 ?
2.	
 ?Generation of the mutant IgE-Fc containing only one oligomannose N-glycan 
with comparable Fc?RI-binding capacity to full-length IgE. ....................... 163	
 ?
3.	
 ?Generation of IgE-Fc oligomannose glycoforms using enzymatic modification 
and biosynthetic engineering and their binding properties to Fc?RI. ........... 164	
 ?
4.	
 ?Glycosylation remodeling of IgE-Fc-Asn394only showed limited impact on 
binding capacity to low-affinity IgE receptor, CD23 (Fc?RII) ..................... 174	
 ?
 viii 
 
 
5.	
 ?Chemoenzymatic glycosylation remodeling of IgE-Fc with complex N-glycan 
using the active glycosyl oxazoline and glycosynthase mutants .................. 179	
 ?
C.	
 ? Conclusion ................................................................................................... 187	
 ?
D.	
 ? Materials and Methods ............................................................................... 189	
 ?
1.	
 ?Mammalian cell culture: cell lines, strains and expression conditions ......... 189	
 ?
2.	
 ?Expression, purification and characterization of recombiant endoglycosidases 
Endo-CC, Endo-M, Endo-S and Endo-S2 and glycosynthase mutants ........ 189	
 ?
3.	
 ?Site-specific glycan remodeling of full-length IgE antibodies ..................... 190	
 ?
4.	
 ?Deglycosylation of high-mannose IgE-Fc-Asn394only-HM to afford IgE-Fc-
Asn394only-GlcNAc intermediate and IgE-Fc-Asp394 ............................... 190	
 ?
5.	
 ?Generation of glycoforms IgE-Fc-Asn394only-Man5 and IgE-Fc-Asn394only-
Man4-7 N-glycoforms by enzymatic modifications ....................................... 191	
 ?
6.	
 ?MALDI-TOF mass spectrometry .................................................................. 191	
 ?
7.	
 ?Surface Plasmon Resonance (SPR) .............................................................. 192	
 ?
8.	
 ?Assay for hydrolytic activities of Endo-CC, Endo-S, Endo-S2 and Endo-M 
against recombinant IgE-Fc-Asn394only-HM ............................................. 192	
 ?
9.	
 ?Glycan remodeling of deglycosylated IgE-Fc-Asn394only to a homogenous 
galactosylated complex type (G2-CT) N-glycoform .................................... 193	
 ?
E.	
 ? Supporting Information .............................................................................. 194	
 ?
Chapter 6:  Summary and future directions .............................................................. 195	
 ?
Bibliography ............................................................................................................. 204	
 ?
 
 
 
 ix 
 
 
List of Tables 
Table 1.  Comparison of specific transglycosylation efficiencies of Endo-S mutants 
measured by LC-ESI-MS using a commercial antibody rituximab as substrate ........ 56	
 ?
Table 2. Kinetic parameters of Endo-S mutants D233M and D233A for SCT-Oxa .. 58	
 ?
Table 3. Kinetic parameters of Endo-S mutants D233M and D233A for rituximab .. 58	
 ?
Table 4. Comparison of the specific hydrolytic activities of Endo-S2 D184 mutants 
using the synthetic complex type rituximab (S2G2-rituximab) as starting material. ... 90	
 ?
Table 5. Comparison of specific transglycosylation activity of Endo-S2 D184 mutants 
using glycan oxazoline as donor and deglycosylated rituximab as acceptor .............. 91	
 ?
Table 6. Kinetic parameters of Endo-S2 D184M and D184A for SCT-Oxa .............. 93	
 ?
Table 7. Kinetic parameters of Endo-S2 D184M and D184A for rituximab .............. 93	
 ?
Table 8. Kinetic parameters and binding affinities of commercial and recombinant 
anti-OVA IgE antiboides binding to Fc?RI? obtained from SPR analysis ............... 162	
 ?
Table 9. Kinetics and binding affinities of IgE-Fc N-glycoforms for Fc?RI. ........... 173	
 ?
Table 10. Kinetic parameters of IgE-Fc-Asn394only N-glycoforms with immobilized 
soluble domain of IgE-Fc low-affinity receptor CD23 measured by SPR. .............. 178	
 ?
 
 
 
 
 
 
 
 
 
 
 
 x 
 
 
List of Figures 
 
 
Figure 1. Market of monoclonal antibody therapeutics between 2002 and 2017 ......... 3	
 ?
Figure 2. A schematic representation of genetic and structural elements for a typical 
immunoglobulin molecul .............................................................................................. 5	
 ?
Figure 3. Overall structural arrangements of immunoglobulin classes and subclasses 
with glycosylation distributions highlighted ................................................................. 8	
 ?
Figure 4. A simplified representation of the five major classes of glycosidic linkages 
found in protein glycosylation ...................................................................................... 9	
 ?
Figure 5. A simplified 3D structure of a IgG antibody with N-glycans attached at a 
conserved glycosylation site Asn-297 in CH2 domain of IgG-Fc region .................... 12	
 ?
Figure 6. Selected N-glycoforms detected at Asn297 site of IgG?Fc domain. .......... 15	
 ?
Figure 7. Biosynthetic pathway of (a) N-glycoproteins in eukaryotic cells and (b) 
GDP-l-fucose in mammalian cells .............................................................................. 19	
 ?
Figure 8. Glycoengineering methods to produce both low-fucose or nonfucosylated 
monoclonal antibodies in vivo. ................................................................................... 22	
 ?
Figure 9. General concepts of ENGase-mediated glycosylation remodeling strategy of 
(a) glycoproteins and (b) glycopeptides ...................................................................... 25	
 ?
Figure 10. Crystal structure of endoglycosidase Endo-A (PDB: 3FHA) .................... 26	
 ?
Figure 11. Catalytic mechanisms of the ENGases showing key catalytic residues for 
some enzymes from: (a) family GH18 and (b) family GH85. .................................... 29	
 ?
Figure 12. Mechanism-based design of mutant glycosynthases of ENGases ............. 31	
 ?
Figure 13. Selected examples of antibody glycan remodeling using ENGases .......... 33	
 ?
 xi 
 
 
Figure 14. Glycosylation remodeling of IgG antibody rituximab using Endo-S and 
Endo-S glycosynthase mutants ................................................................................... 35 
Figure 15. Glycoengineering of a core-fucosylated therapeutic antibody rituximab.. 46 
Figure 16. Chemoenzymatic remodeling of core fucosylated antibody rituximab to a 
non-fucosylated G2 glycoform using ?-fucosidase and Endo-S mutants. .................. 47 
Figure 17. The general procedure of site-directed saturation mutagenesis of Endo-S 
using degenerate primers. ........................................................................................... 51	
 ?
Figure 18. Schematic presentation of hydrolysis and transglycosylation reactions by 
Endo-S and its mutants using antibody rituximab as starting material. ...................... 53	
 ?
Figure 19. Comparison of transglycosylation reactions catalyzed by Endo-S D233M 
and previously reported Endo-S D233Q and D233A mutants. ................................... 55	
 ?
Figure 20. Michaelis-Menten plots of Endo-S D233M and Endo-S D233A. ............. 59 
Figure 21. The chemical structure of the complex type N-glycan oxazoline (1) and the 
corresponding complex type N-glycan thiazoline (2). ................................................ 60	
 ?
Figure 22. Inhibition effects of Endo-S, Endo-S2, and Endo-M by CT-thiazoline .... 62	
 ?
Figure 23. Mass spectrometry analysis of antibody N-glycoforms generated by glycan 
remodeling of a therapeutic antibody trastuzumab (Herceptin). ................................ 64	
 ?
Figure 24. Non-enzymatic transfer of sialylated complex type glycan oxazoline to 
deglycosylated trastuzumab antibody in the absence of Endo-S enzymes ................. 65	
 ?
Figure 25. Molecular mechanism of Endo-S based on co-crystalized structure with the 
G2 glycan substrate.. ................................................................................................... 68 
Figure 26. Sequence alignment of Endo-S2 and Endo-S???????????.88 
 xii 
 
 
Figure 27. Schematic presentation of hydrolysis and transglycosylation by Endo-S2 
and its mutants using rituximab as substrate. .............................................................. 89	
 ?
Figure 28. Michaelis-Menten curves of Endo-S2 D184M and D184A ...................... 93	
 ?
Figure 29. Evaluation of substrate specificity of Endo-S2 D184M on different glycan 
structures (HM, CT, and Hyb type).. .......................................................................... 96 
Figure 30. Comparison of Endo-S2 D184M-catalyzed transglycosylation reactions 
with different N-glycans. ............................................................................................ 98	
 ?
Figure 31. Comparison of transglycosylation reactions catalyzed by selected mutants 
Endo-S2 D184Q, Endo-S2 D184M, and Endo-S D233Q. ........................................ 100	
 ?
Figure 32. Transglycosylation of antibody with reduced oxazoline substrate at 
different Endo-S D184M concentrations. ................................................................. 101	
 ?
Figure 33. Structures of rituximab N-glycoforms and N-glycan oxazolines used for 
the assessment of Endo-S activity. ............................................................................ 103	
 ?
Figure 34. Hydrolytic activity of Endo-S WT on different antibody N-glycoforms and 
oxazolines. ................................................................................................................ 105	
 ?
Figure 35. Endo-S-catalyzed transglycosylation reactions with oxazolines SCT-Oxa,  
HM-Oxa, Hyb-Oxa of deglycosylated rituximab ..................................................... 108	
 ?
Figure 36. One-pot glycosylation remodeling of commercial rituximab to give rise to 
hybrid and high-mannose rituximab N-glycoforms, Hyb-RTX and HM-RTX ........ 110	
 ?
Figure 37. Endo-S WT-catalyzed one-pot transglycosylation reactions with mannose 
(HM-Oxa 5) and hybrid (Hyb-Oxa 6). ...................................................................... 111	
 ?
Figure 38. Endo-S WT-catalyzed one-pot transglycosylation reactions with the  
chemically modified glycan oxazoline. .................................................................... 114	
 ?
 xiii 
 
 
Figure 39. Hydrolysis of glycoprotein RNase B by recombinant Endo-CC and Endo-
CC fusion protein (CPD).. ........................................................................................ 132	
 ?
Figure 40. Transglycosylation of deglycosylated RNase B by recombinant Endo-CC 
and Endo-CC (CPD).. ............................................................................................... 135	
 ?
Figure 41. Comparison of Endo-CC-catalyzed hydrolysis against fucosylated and 
non-fucosylated sialoglycoprotein (SGP) ................................................................. 137	
 ?
Figure 42. Selective deglycosylation of non-fucosylated N-glycans by Endo-CC from 
the heterogeneous N-glycoforms of erythropoietin (EPO). ...................................... 138	
 ?
Figure 43. Selective removal of non-fucosylated IgG N-glycans by Endo-CC.. ..... 139	
 ?
Figure 44. Comparison of transglycosylation efficiency of Endo-CC N180H for three 
major types of N-glycan oxazolines. ........................................................................ 141	
 ?
Figure 45. Comparison of transglycosylation reactions catalyzed by selected Endo-
CC mutants. .............................................................................................................. 143	
 ?
Figure 46. Transglycosylation reactions for deglycosylated antibody by Endo-CC and 
Endo-CC N180H enzymes at varied concentrations.. .............................................. 145	
 ?
Figure 47. Site-selective N-glycosylation remodeling of the full-length human IgE 
antibodies using endoglycosidase Endo-H. .............................................................. 160	
 ?
Figure 48. Surface plasmon resonance (SPR) representative sensorgrams of binding 
between full-length IgE antibodies and immobilized soluble domain of high-affinity 
IgE receptor Fc?RI .................................................................................................... 161	
 ?
Figure 49. Glycan remodeling of oligomannose IgE-Fc-Asn394only-HM (Man5-9) 
using a combination of glycan remodeling strategies ............................................... 164 
 xiv 
 
 
Figure 50. MALDI-TOF mass spectrometry identification of released N-glycans from 
IgE-Fc-Asn394only N-glycoforms generated through glycan remodeling .............. 166	
 ?
Figure 51. Schematic representation of the substrate specificity of ?1-2 and ?1-2, 3, 6 
mannosidase against differentially linked mannose residues of the high-mannose Asn-
linked oligosaccharides in glycoproteins .................................................................. 169 
Figure 52. Surface plasmon resonance (SPR) representative sensorgrams of binding 
events analyzed between remodeled IgE-Fc-Asn394only glycoforms and immobilized 
soluble domain of high-affinity IgE receptor Fc?RI ................................................. 172	
 ?
Figure 53. Representative surface plasmon resonance (SPR) sensorgrams of binding 
between homogeneous IgE-Fc-Asn394only N-glycoforms and immobilized soluble 
domain of IgE-Fc low-affinity receptor CD23/ Fc?RII. ........................................... 177	
 ?
Figure 54. Hydrolytic selectivities of ENGase Endo-S, Endo-S2, Endo-M and Endo-
CC against high-mannose IgE-Fc-Asn394only-HM (Man5-9) N-glycoforms.. ........ 181	
 ?
Figure 55. Comparison of enzymatic hydrolysis on native IgE-Fc-Asn394only-HM 
(Man5-9) N-glycoforms by Endo-CC, Endo-M, Endo-S, and Endo-S2. ................... 182	
 ?
Figure 56. Transglycosylation of IgE-Fc-Asn394only-GlcNAc with galactosylated 
complex type glycan oxazoline (CT-Oxa) using Endo-CC N180H. ........................ 184	
 ?
Figure 57. Synthesis of IgE-Fc-Asn394only-CT (G2) N-glycoform and its binding 
properties for high-affinity receptor Fc?RI ............................................................... 186	
 ?
Figure 58. Design for therapeutic intervention as IgE replacement therapy using a 
combination of anti-IgE antibodies and glycoengineered IgE-Fc. ........................... 203	
 ?
 
 
 
 xv 
 
 
List of Abbreviations 
3D Three-dimension 
ADC Antibody-drug conjugates 
ADCC Antibody-dependent cellular cytotoxicity 
CBP Carbohydrate-binding proteins 
CD Cluster of differentiation 
CDC Complement-dependent cytotoxicity 
CDG Congenital disorders of glycosylation 
CDR Complementarity-determining regions 
DiAcTridH 2,4,-diacetamido 2,4,6-trideoxyhexose 
EMA European Medicines Agency 
ENGase Endo-?-N-acetylglucosaminidases 
EPO Erythropoietin 
ESI Electrospray Ionization 
Fab Antigen-binding fragment 
Fc Crystallizable fragment 
Fc?R Fc alpha receptor 
Fc?R Fc gamma receptor 
FDA Food and Drug Administration 
Fmoc Fluorenylmethyloxycarbonyl chloride 
 xvi 
 
 
FUT ?1,6-fucosyltransferase 
GnT Glycosyltransferase 
GPI Glycosylphosphatidylinositol  
HM High-mannose 
HPAEC-PAD High-Performance Anion-Exchange Chromatography Coupled with Pulsed Amperometric Detection 
HPLC High-performance liquid chromatography 
Hyb Hybrid 
Ig Immunoglobulin 
MAb Monoclonal antibody 
MeCN Acetonitrile 
MS Mass spectrometry 
OST Oligosaccharyltransferase 
PBS Phosphate Buffered Saline 
RNase B Ribonuclease B 
RTX Rituximab 
SCT Sialylated complex type 
SGP Sialylated glycoprotein 
Tris Tris(hydroxymethyl)aminomethane 
 xvii 
 
 
Chapter 1: Introduction 
A.  Introduction to antibodies 
1. Monoclonal antibody-based therapeutics 
The monoclonal antibodies (mAb) represent one of the most important classes 
of biopharmaceutical treatments for a variety of health conditions.1 Since the 
licensing of the first therapeutic mAb muromonab-CD3 (under trade name 
Orthoclone OKT3) in 1986,2 over seventy mAb and mAb-based proteins, including 
the antibody fragments, antibody-drug conjugates and crystallizable fragment (Fc) 
fusion proteins, have been approved for clinic use by either the Food and Drug 
Administration (FDA) in the United States (US) or the European Medicines Agency 
(EMA) in the European Union (EU).1 The annual approval rate for mAb-related 
therapeutics has also significantly increased over the past decade (Figure 1A). Based 
on public reports released by the FDA and EMA, a total of 35 new mAbs were 
granted a first approval between 2012 and 2017 with an average of 5.8 approvals per 
year, whereas only an average of 1.6 mAb-based therapeutics were approved annually 
during 2002-2011.3  
The first therapeutic mAb, called muromonab-CD3, was prescribed as an anti-
rejection therapy to treat kidney failure caused by organ transplant rejection.2 The 
therapeutic mAbs are now applied to treat a diverse range of disease targets, including 
various types of malignancies/cancers, inflammatory and cardiovascular disorders, 
and infectious diseases (Figure 1B). The manufacturing systems for monoclonal 
antibodies have also undergone significant changes since 1986 to improve the safety 
 1 
 
 
 
and efficacy of antibody products, starting from the murine system to the chimeric 
and finally the fully humanized production.4 
Despite their successes in the clinic, the cost-effectiveness issue of antibody 
therapeutics has limited their development and prevented new beneficial drugs from 
being readily accessible to the general population worldwide.5 Recently, there has 
been a growing interest in the pharmaceutical industry in developing ?biosimilar? 
antibodies as affordable alternatives to reduce the costs and potentially improve the 
clinical performance of many antibody-based therapeutics. Recent research progress 
in protein engineering and post-translation modifications has driven new designs of 
?biobetter? antibodies that demonstrated enhanced therapeutic outcomes and reduced 
side effects when compared with their parent antibody molecules.5 Such efforts are 
represented by the ongoing development of antibody-drug conjugates (ADCs) and 
multi-specific antibodies against various cancers and viral infections.6 Improvements 
in the manufacturing process of antibodies, such as expression system and 
purification protocol, and more stringent quality controls are urgently needed to 
facilitate the development of more efficient antibody products. For instance, research 
into the degradation and recycling mechanisms of antibody proteins are contributing 
to the development of mAbs with prolonged serum half-life.7 In addition, recent 
studies on the site-specific modifications of antibodies by both chemical and 
enzymatic methods are currently being explored to improve the delivery and effector 
functions of several important classes of antibody-based therapeutics, which will be 
discussed in detail in the following sections.8 
 
 2 
 
 
 
 
Figure 1. Market analysis of monoclonal antibody-related therapeutics between 2002 
and 2017. (A) Trend of the number of mAb products approved annually by the FDA 
and EMA between 2002 and 2017.  (B) Distribution of disease indications targeted by 
antibody therapeutics. Therapeutic mAbs are now prescribed for a variety of health 
conditions. Oncology and hematology remain the disease areas most targeted by 
mAbs according to the public reported data by the FDA and EMA.3 
 3 
 
 
 
2. Overview of antibody structure and function 
Antibodies, also known as immunoglobulins (Igs), are soluble biomolecules 
generated by the immune system to specifically recognize and neutralize foreign 
pathogens and toxins during activation of an immune response.9 The immunoglobulin 
is typically a multi-subunit heterodimer that consists of four polypeptides, including 
two identical light (L) chains and two identical heavy (H) chains linked by multiple 
disulfide linkages (Figure 2). The structural components of each light (L) and heavy 
(H) chain are further functionally dividied into the ?variable? (V) and ?constant? (C) 
domains located at the NH2- and COOH-terminus of the heavy chain, respectively. 
The V domains from both H and L chains of antibody are responsible for recognizing 
the pathogens through binding to foreign structures called antigens with high affinity 
and selectivity, whereas the constant (C) domains from H chains mediate the effector 
functions of immunoglobulins by interacting with the other components of the host 
immune system.10  
 
2.1. Immunoglobulin fold: the basic building block of antibody structure 
The overall structural arrangement of immunoglobulin is highly conserved 
across a large protein family called the immunoglobulin superfamily (IgSF), and the 
conserved Ig domain can be found in many proteins whose functions are related to the 
immune system.11 These proteins typically adopt a common conformation known as 
the immunoglobulin fold, which is composed of a ?-sandwich framework connected 
by disulfide bonds and multiple loop structures. A pair of pleated ?-sheets consisting 
of antiparallel ?-strands form the hydrophobic core of the ?-sandwich. The unique 
 4 
 
 
 
conformation of the immunoglobulin fold contains several structural features that are 
necessary for antibody molecules to efficiently regulate the immune response, which 
are discussed in detail below.12  
 
 
Figure 2. A schematic representation of genetic and structural elements for a typical 
immunoglobulin molecule. The genetic segments of H and L chains are illustrated at 
the right section of the structure. The amino acid information such as disulfide bonds 
and functional regions (such as CDR, FR and pepsin-digestion sites) are shown at the 
left section of the antibody scheme.9  
 5 
 
 
 
2.2. Hypervariable (V) domain of antigen binding (Fab) fragment and antibody-
antigen recognition 
An antibody can be cleaved into three functional fragments by protease, which 
result in two antigen-binding fragments (Fab) that specifically recognize the non-self 
antigens and one crystallizable fragment (Fc) that modulate effector functions.13 The 
Fab consists of a complete L chain and a portion of H chain, whereas the Fc contains 
the constant domains from both H chains of antibody tetramer. Specific amino acid 
residues located within the connecting loops at the N-terminus of the Fab variable (V) 
domain form the binding pocket for antigen recognition. These regions, also known 
as the complementarity-determining regions (CDRs), are intrinsically hypervariable. 
The variability of V domains is the result of a series of genetic regulatory events, 
including gene segment organization and rearrangement, somatic hypermutation, and 
affinity maturation.14 This hypervariability provide the molecular basis for antibodies 
to recognize an extremely diverse range of antigens with high selectivity. The 
complementarity-determining regions of L and H chains together contribute to the 
antigen-binding site of each Fab domain of antibody.  
 
2.3. Constant (C) domain of crystallizable fragment (Fc) and antibody Fc receptors 
The isotypes and subclasses of an antibody are determined by constant 
domains of its heavy chains. Based on this classification, there are five antibody 
isotopes found in human circulation, including IgG, IgM, IgD, IgE, and IgA, each of 
which embodies unique structural properties and biological functions.15 As illustrated 
in Figure 3, in IgA, IgD and IgG, a flexible ?hinge region? with no defined secondary 
 6 
 
 
 
structure connects their Fc and Fab fragments, whereas this region is replaced by a 
more rigid Ig domain in IgE and IgM.16, 17 The Fc fragments of immunoglobulins 
modulate immune responses primarily through interactions with a variety of antibody 
Fc receptors found on the surface of cells associated with immune regulation.18 These 
receptors serve as the mediators between the humoral adaptive immune response and 
the innate immune machinery.19 The overall efficacy of an antibody-mediated 
immune response is largely dependent on the relative abundance of these Fc receptors 
on the immune cell surface, and an accurate regulation of the antibody/receptor 
interactions plays an important role in mounting an effective immune response 
against various types of infectious agents.20 Therefore, alteration in the structural 
compositions of antibody Fc domains can have profound impact on the 
immunoregulatory activity of antibodies and, as a result, the neutralization of foreign 
pathogens and infected cells. 
 
 7 
 
 
 
 
Figure 3. Overall structural arrangements of immunoglobulin classes and subclasses 
with glycosylation distributions highlighted. For IgM, a joint chain (J chain) protein is 
also marked in red. The J chain polypeptide is required for the IgM to be secreted into 
mucosa. The conserved glycosylation sites of each antibody isotype are marked at the 
amino acid residues where N-glycans are covalently attached.21 
 
 
 
 8 
 
 
 
B. Antibody glycosylation and its importance in biological functions 
The covalent linkages between sugars and amino acid residues comprise one 
of the most ubiquitous post-translation modifications of proteins in nature. A series of 
enzymatic modifications on the protein-associated carbohydrate structures lead to the 
biosynthesis of a large collection of glycoproteins with diverse biological functions.22 
These processes can be found across the entire phylogenetic spectrum, from bacteria 
and Archaea to higher eukaryotes. Depending on the specific amino acid residues and 
monosaccharides involved in forming the glycosidic linkages, more than 40 types of 
glycosidic bond can be found in proteins with unique anomeric configurations, which 
can be divided into different types of glycosylation: the N (Asn)-, O (Ser/Thr)- and C 
(Trp)-linked glycans, glypiation, and phosphoglycosylation (Figure 4).23 
 
Figure 4. A simplified representation of the five major classes of glycosidic linkages 
found in protein glycosylation. The DiAcTridH refers to the 2,4,-diacetamido 2,4,6-
trideoxyhexose, and the C-term refers to the carboxy-terminal group.24  
 9 
 
 
 
1. Structure and functional diversity of antibody glycosylation 
Antibodies are glycosylated proteins, and the location and relative abundance 
of oligosaccharides attached to the Fc and Fab regions of antibodies vary significantly 
among the immunoglobulin isotypes (Figure 3).25 Both N- and O-linked glycans can 
be found on the Fc, Fab and hinge regions of antibodies. Up to 15% of the molecular 
weight of an antibody (~150 kDa) can be contributed by multiple N-glycans (~ 2 kDa) 
attached to each of the heavy chains, whereas no conserved glycosylation sites can be 
found on the light chains. These associated oligosaccharides, if not contacted by the 
surrounding amino acid residues, are typically not rigid and adopt a dynamic structure 
due to the flexible N-glycosidic linkages.26  
The biological functions of antibody N-glycans are multifaceted, ranging from 
maintaining the structural integrity of antibody to participating in binding of antibody 
with other mediators such as lectins and immunoglobulin receptors.27-30 Some of the 
most important functions of antibody N-glycans include: extending the serum half-
life; facilitating the cellular transport, secretion and endocytosis; and regulating the 
effector functions by recruiting other immune components to the site of infections. In 
addition to the direct effects of N-glycosylation on antibody structure and function, 
the aberrant glycosylation pattern of certain serum immunoglobulins are associated 
with various underlying diseases, including rheumatoid arthritis, Berger's disease, and 
congenital disorders of glycosylation (CDGs).31 Other aspects of antibody functions, 
including the Fab-mediated antigen recognition, can also be affected by antibody 
glycosylation. For instance, the neutralization activity of certain anti-HIV antibodies 
was found to be associated with their glycosylation profiles at the Fab region, and a 
 10 
 
 
non-neutralizing anti-HIV antibody acquired potent antiviral activity when produced 
in an engineered expression system that optimized the glycosylation profile of the 
proteins produced. It was therefore suggested that the Fab-glycans of antibody could 
facilitate neutralization of HIV viruses by blocking their entry into the host cells.32 
 
2. Glycosylation modulates effector functions of IgG and IgE antibodies 
Currently, all approved antibody therapeutics are based on the IgG isotype, 
which accounts for approximately 70%-75% of the human antibodies secreted into 
the bloodstream daily.33 Human IgG antibodies typically contain the biantennary 
complex-type N-linked oligosaccharides based on a heptasaccharide core structure 
composed of mannoses (Man) and N-acetylglucosamines (GlcNAc) at the conserved 
Asn-297 residue within the CH?2 domains of the Fc region (Figure 5).34 Studies 
demonstrate that the N-glycans are crucial in maintaining the IgG conformation,35 and 
the carbohydrates attached at this site could be recognized as molecular epitopes for 
carbohydrate-binding proteins (CBPs) such as lectins and other regulatory factors 
associated with immune system.36 
Glycosylation patterns of IgG can have a profound impact on its interactions 
with receptors and factors involved in immune regulation. For instance, the binding 
affinity of IgG for the Fc gamma receptor Fc?R was significantly reduced when the 
conserved glycosylation site is removed by mutagenesis, which confirmed that this N-
glycan was necessary for the optimal binding between IgG and Fc?Rs.29 However, 
this phenomenon was not observed between IgA and the Fc alpha receptors (Fc?Rs), 
 11 
 
 
as recombinant IgA binds to Fc?Rs regardless of the presence of its conserved N-
glycans.37 
 
Figure 5. A simplified 3D structure of a typical IgG antibody with N-glycans attached 
at a conserved glycosylation site Asn-297 in the CH2 domain of the IgG-Fc region. 
The diagram shown here is based on the structure of a human IgG (PDB: 1HZH).  
The alpha backbone of heavy chains and light chains of antibodies are shown in cyan 
and purple, respectively, and the Fc glycans are colored in navy blue.38  
 
One of the major mechanisms by which antibodies control and neutralize 
pathogenic factors during the humoral immune response is antibody-dependent 
cellular cytotoxicity (ADCC). Through ADCC, the Fc fragments of IgG in complex 
with the antigen interact with the Fc?Rs on the surface of effector cells, such as 
natural killer (NK) cells and eosinophils, to promote the destruction of the antibody-
coated targets.39 Recent studies have further elaborated on the functional role of N-
glycans in the Fc?R-receptor binding.28, 29, 40, 41 For example, the removal of a core 
fucose monosaccharide from the conserved IgG-Fc N-glycans significantly enhanced 
 12 
 
 
the ADCC activity of several IgG antibodies, which was due to increased binding 
affinities of the modified IgG-Fc for the Fc?IIIa receptor.28 This important finding 
provided the molecular basis for the development of novel antibody therapeutics with 
a lower core-fucosylation level and better clinic performance in treating various types 
of malignancy.42  
Complement-dependent cytotoxicity (CDC) is another critical mechanism of 
action for antibodies to mediate cellular killing. The common pathway of CDC is 
initiated by interactions between IgG-Fc and the complement component factor C1q. 
A complex-type N-glycan consisting of N-acetylglucosamines, galactoses and sialic 
acids is typically required for the optimal binding between the Fc fragment and C1q 
as well as the subsequent activation of the rest of the complement machinery.43, 44 
The human intravenous immunoglobulin (IVIG), when administrated at high 
dosages, has been successfully applied to treat a number of inflammatory disorders.45 
In contrast to the proinflammatory activities of the IgG N-glycans mentioned above, 
the terminal ?2,6-sialic acid found in a small percentage of human IVIG preparations 
was discovered to be responsible for the anti-inflammatory effects in animal models46. 
It was later proposed that the interactions between the sialylated IgG N-glycoforms 
and certain sialic acid-binding factors could reduce inflammation by up-regulating the 
expression of certain inhibitory receptors Fc?Rs.47-49 
Unlike IgG antibodies, the functional relationship of N-glycosylation and IgE 
biology largely remains a question to be answered by future studies, even though IgE 
contains the highest level of N-glycosylation among all monomeric antibody isotypes, 
with seven conserved N-linked glycosylation sites located in the primary sequence of 
 13 
 
 
its heavy chain.50-52 However, a recent groundbreaking study on IgE N-glycosylation 
discovered a single oligomannose N-glycan attached at Asn-394 residue in Fc region 
of IgE antibody, which was found to be necessary for IgE to mediate the acute 
allergic response.53, 54 According to this study, genetic mutations and functional 
studies all suggested that this unique N-glycan located at Asn-394, which is 
homologous to the Asn-297 site of IgG antibodies, plays an obligatory role in 
determining the secondary structure of IgE-Fc region and, as a result, the ability of 
IgE to interact with its cognate IgE-Fc epsilon receptors (Fc?Rs) to mediate 
anaphylaxis. A more comprehensive introduction of IgE N-glycosylation will be 
given in chapter 5.  
3. Structural heterogeneity of antibody N-glycans 
Analysis of the N-glycan structures of antibodies either extracted from human 
serum samples or expressed as recombinant proteins have shown extensive structural 
variations. For purified IgG antibodies, according to X-crystallographic structures, the 
N-glycans located at the two identical Asn-297 residues of the IgG-Fc homodimer 
can be significantly different.55 As illustrated in Figure 6, each Asn-297 residue can 
be associated with a large collection of oligosaccharides that can be categorized into 
three major N-glycoforms, IgG-G2, -G1, and -G0. On the basis of the heptasaccharide 
core, the IgG-G0 contains an additional GlcNAc residue at the terminal position of 
each of the two branches of the complex N-glycan, which accounts for approximately 
35% of total IgG glycoforms found in human serum.56 Another relatively abundant 
IgG glycoform (~ 34% of the total serum IgG) is the IgG-G1 form, which has another 
terminal galactose (Gal) added to one of the two GlcNAc residues of the IgG-G0 
 14 
 
 
glycoform.57 Finally, the IgG-G2 glycoform contains two galactose residues located 
at the terminal ends of the IgG-G0 form, which also represents the least abundant IgG 
N-glycoform in the serum (~17%). The rest of the human serum IgG glycoforms 
(~14%) undergo further processing by having one or two terminal sialic acid residues 
linked to the galactose residues of the G1 or G2 forms. Additional variations also 
come from potential modifications such as core fucosylation (70% of the total IgG 
glycoforms) and the bisecting GlcNAc (30% of the total IgG glycoforms). 58 
 
 
Figure 6. Selected N-glycoforms detected at the Asn297 site of IgG?Fc domain. All 
symbols of monosaccharide comply with the standard symbol nomenclature for the 
representing glycans: galactose (Gal) (yellow circle); mannose (Man) (green circle); 
N-acetylglucosamine (GlcNAc) (blue square); N-acetylneuraminic acid (NeuAc) 
(purple diamond); fucose (Fuc) (red triangle).20  
 15 
 
 
C. Engineering antibody glycosylation by biosynthetic pathway manipulation 
Because the naturally existing and conventionally produced antibodies show a 
high level of heterogeneity in their N-glycosylation profiles, it has been extremely 
challenging to obtain the pure or enriched samples of antibody N-glycoforms under 
normal circumstances.59-61 As a result, methods to generate homogeneous antibody N-
glycoforms containing well-defined N-glycan structures are of high demand both in 
academia and biopharmaceutical industry.  
Therefore, in order to optimize the glycosylation of recombinant antibodies, a 
number of promising strategies have been developed with success. One approach is to 
manipulate the biosynthetic pathways of certain N-glycans associated with antibodies 
in vivo by exploiting the enzymes involved in this process.62 The success of this 
method has led to the development of monoclonal antibodies with improved clinic 
performance. However, limitations still exist in this method due to the extremely 
dynamic and complex cellular environment inherent in the expression systems62. 
Another promising concept for antibody glycoengineering is through in vitro 
chemoenzymatic modifications using various glycan-modifying enzymes combined 
with chemically modified substrates63. This section will highlight several recent 
examples of progress in antibody glycoengineering based on manipulation of the N-
glycan biosynthetic pathways in mammalian cells and other host systems. 
 
 
 
 
 
 
 16 
 
 
1. Biosynthetic pathways of N-glycans in eukaryotic expression systems 
Production of recombinant antibodies in well-established systems such as the 
Chinese Hamster Ovary (CHO) and Human Embryonic Kidney (HEK) cells often 
results in mixtures of different antibody N-glycoforms, which makes it difficult to 
carry out subsequent functional studies or therapeutic applications.59 Manipulation in 
the biosynthetic pathways of protein N-glycans provides an opportunity to control the 
structural variations of carbohydrate in some expression systems. Specifically, this 
strategy has been successfully applied to enhance the therapeutic efficacy of certain 
monoclonal antibody-based therapeutics by optimizing the N-glycosylation patterns.64  
One of the most successful applications of in vivo antibody glycoengineering 
is the development of antibody variants with reduced or completely abolished core 
fucosylation. Such studies have attracted tremendous attention in the past decade 
because of the functional relationship between the core fucosylation and the ADCC 
effects of IgG antibodies. As mentioned in the previous section, the removal of the 
core fucose from IgG antibodies can significantly enhance its ADCC activity and the 
neutralization efficacy against target cells.28 Perturbation to the biosynthesis of fucose 
in vivo by either genetic interference or manipulation in the expression medium can 
lead to glycoproteins with a lower core fucosylation level.65  
To understand the molecular mechanism of this biosynthetic glycoengineering 
approach, a brief introduction to the general aspects of N-glycoprotein biosynthesis in 
the eukaryotic system is provided here (Figure 7).26 The biosynthesis of all eukaryotic 
N-glycans initially takes place on the cytoplasmic side of the ER membrane, where 
fourteen monosaccharides are sequentially added to a lipid-phosphate (Dol-P) anchor 
 17 
 
 
to form an oligosaccharide precursor, namely the Dol-P-Glc3Man9GlcNAc2. The 
complete oligosaccharide is then transferred to an Asn residue within a conserved N-
glycosylation sequon (Asn-X-Ser/Thr) of a nascent polypeptide by the function of a 
crucial biosynthetic enzyme called the oligosaccharyltransferase (OST). The now-
attached N-glycan then undergoes further processing in the ER lumen and Golgi 
apparatus by a series of membrane-associated glycosidases and glycosyltransferases 
(Figure 7a). Certain glycan intermediates, such as mono-glucosylated and non-
glucosylated N-glycans, can be recognized by lectins to facilitate protein folding. In 
the case of fucosylation, an unique enzyme called the ?1,6-fucosyltransferase (FUT8) 
is responsible for adding the core fucose to the nascent glycoproteins. The FUT8 
recognizes the GlcNAcMan3GlcNAc2 intermediate as the glycosyl acceptor substrate 
and transfers the fucose from a nucleotide GDP-l-fucose as the glycosyl donor. The 
de novo biosynthetic pathway of GDP-l-fucose donor is also illustrated in Figure 7b. 
 
 
 
 
 18 
 
BI88CH26_Wang ARjats.cls March 19, 2019 12:34
 
Ribosome
a Alg 1
Alg 2
Alg 11
Alg 7 Alg 3 Alg 6 OST CNX/CRT
Alg 13 Alg 9 Alg 8 G-I folding
Alg 14 Alg 12 Alg 10 G-II cycle
Dol-P Dol-PP-Man9 Glycolipid Nascent Folded protein
precursor protein
Endoplasmic reticulum
Golgi apparatus
Translocation
Endoplasmic FUT8
reticulum GnT-II
Mns-I Mns-I GnT-I Mns-II GnT-V
GalT
SiaT
Protein Gal Phophate Fuc
Dol Man Neu5Ac GlcNAc
b
HO OH HO OH OH
HO O GMP O GMD
O
O FX protein OGDP
HO HO OO HO NADP+ HO OH
O P OH GTP PPi OGDP OGDP NADPH NADP+ HO
OH
OH
Mannose-1-phosphate GDP-D-mannose GDP-4-keto-6-deoxymannose GDP-L-fucose  
Figure 2
(a) BioFsyingthuersies o7f N. -Bglyicoospyrontetihnes tinice upkaarythotwic aceyll s.o(bf) D(ae)n oNvo-bgiolsyycntohpetricopteatihnwsa yionf GeDuPk-al-rfyucootsieci ncmealmlsm aalinandc e(llbs.) 
Abbreviations: Alg, asparagine-linked glycosylation enzyme; CNX/CRT, calnexin/calreticulin; Dol, dolichol; Fuc, fucose; FUT8,
?1,6-fucosyltransferase; FX, GDP-4-keto-6-D-deoxym2a6nnose epimerase/GDP-4-keto-6-liter-galactose reductase; G-I,?-glucosidase
I; G-IIG,?D-gPlu-col-sifduacseoIsI;eG ianl, gmalacmtosme;aGlailaTn, g caleacltloss.yltr aAnsbfebrarsee;vGiaDtPio, gnusan aosrien esdhipohwosnph:a Ate;lGgl,c NasApc,aNra-agceitnyelg-lulicnoskamedin e;
GMD, GDP-mannose 4,6-dehydratase; GMP, guanosine monophosphate; GnT,N-acetylglucosaminyltransferase; Man, mannose;
Mns,?g-mlyacnnoossyidlaaset;ioNneu Aecn,zNy-amceety;l nCeuraminic ais any amino acid except proline; OST, oligNosXacc/hCarR
ciTd;,N cXaSl/nTe, txheinc/ocnaselnrseutsitchuree amino acid sequence for N-glycosylation, where Xyltransferase; SiaT, sialyltransfelrianse;. Dol, dolichol; Fuc, fucose; 
FUT8, ?1,6-fucosyltransferase; FX, GDP-4-keto-6-D-deoxymannose epimerase; G-I, 
This GnTIII overexpression technology, acquired by Roche in 2005, was used to produce the
glycoeInIg, inIeIeIr,e d?a-ngtliu-CcDos2i0daansteib oId, y IoIb, inIuItIu;z uGmaabl,, wghaiclahcwtoasseap; proved by theDrug Administration in 2013 for treatment of non-Hodgkin?s lymphGomalaTa,n dgcahla
UctS Food androniocslyymltrpahno-sferase; GDP, 
cytic leukemia (49).
Whgiuleatnhoeseixnper esdsiiopnhoofspGhnaTtIeI;I dGidlcrNedAucce, thNe  fuaccoesetyllegvelul cofoIsgaGm, iint ed;id GnoMt aDbo, liGshDcoPr-emannose 4,6-
fucosylation. Attempting to completely blocking core fucosylation, Satoh and coworkers (27, 50)
produdcedhayndFraUtaTs8e;k nGocMkoPut, cgelulalinnoesuisnineg mseoquneonptihalohsopmhoaltoeg;o uGsnreTc,o macbientaytilognl.uTcohesaemngiin-yltransferase; 
neered cell line was capable of producing a nonfucosylated anti-CD20 antibody that showed a
100-foMldainnc,r eamseainnnAoDsCe;C aMctnivsit,y c?o-mmpaarnendowsitihdathseen; onNgleyucoAencg, inNee-raecdeatnytilbnoeduy.rTamhisinstiacb leacid; NXS/T, 
consensus amino acid sequence forw wNw.a-ngnluyalcreovisew
?
ysl.oargtionG,ly cowenghineererein gXof Ainsti boadniesy a2m6.5ino acid 
except proline; OST, oligosaccharyltransferase; SiaT, sialyltransferase. 
 19 
 
Annu. Rev. Biochem. 2019.88. Downloaded from www.annualreviews.org
 Access provided by University of Maryland - College Park on 03/29/19. For personal use only. 
NXS/T
 
2. Antibody glycan remodeling by biosynthetic pathway perturbation 
Based on the biosynthetic pathway of the core fucosylated N-glycans, several 
approaches to generate antibody N-glycoforms in vivo with reduced core fucose have 
been developed. The general concepts of these methods are summarized in Figure 8. 
The majority of these strategies take advantages of the key enzymes involved in the 
biosynthesis of core fucose and its precursors. For instance, down regulation of the 
genes expressing the GDP-mannose 4,6-dehydratase (GMD) and FUT8 has led to a 
lower presence of core fucose in the expressed antibodies in certain systems.65, 66 Two 
fucose-reduced mAbs, roledumab67 and ublituximab68, were produced in this fashion 
and are currently tested in phase II clinical trials. Other examples using the similar 
concept include the overexpression of GnTIII, a glycosyltransferase responsible for 
adding the bisecting GlcNAc to the N-glycans, which will block access of FUT8,69 
and complete knockout of FUT8, which prevents transfer of the GDP-l-fucose to the 
target glycoprotein. Using the GnTIII overexpression method, an anti-cancer antibody 
called obinutuzumab was developed and was approved by the FDA for treating non-
Hodgkin?s lymphoma and chronic lymphocytic leukemia.70 Furthermore, a FUT8-
knockout cell line was able to produce a non-fucosylated antibody with a 100-fold 
more potent ADCC activity than its unmodified counterpart. The resultant antibody 
glycovariant, mogamulizumab, was licensed in 2012 in Japan to combat the relapsed 
or refractory adult T cell leukemia/lymphoma.71 
Since GDP-l-fucose is the essential substrate for FUT8, several attempts were 
made to reduce core fucosylation by depleting the intracellular level of GDP-fucose 
or its precursors. A successful example was the generation of a GDP-4-keto-6-D-
 20 
 
 
deoxymannose epimerase GDP-4-keto-6-L-galactose reductase (FX)-knockout cell 
using CRISPR/Cas9 gene editing technology.72 Similar approaches also targeted the 
GDP-fucose transporter pathways by gene editing techniques to reduce the available 
GDP-fucose for FUT8.73 Alternatively, various monosaccharide analogs, including 2-
fluorofucose74, 5-alkynylfucose75, 6,6,6-trifluorofucose76, and arabinose77, were tested 
as potential biosynthetic inhibitors against the GDP-fucose. Finally, non-mammalian 
systems devoid of the endogenous FUT8 expression, such as insect78, plant79, fungi80 
and yeast81, were also exploited for glycoengineering of antibody core fucosylation.  
Because terminal sialylation82 and galactosylation43 have been associated with 
complement-dependent cytotoxicity and anti-inflammatory activity of IgG antibodies, 
respectively, there have been efforts to enrich these specific glycoforms for functional 
analysis and therapeutic studies. These efforts often involve genetic regulation of the 
enzymes, such as galactosyltransferases and sialyltransferases, responsible for adding 
the terminal sialic acid and galactose during the final processing steps. In addition, 
mutations at some amino acid sequences surrounding the glycosylation sites may also 
alter the terminal glycosylation level of certain antibodies due to steric hindrance.83, 84 
 21 
 
BI88CH26_Wang ARjats.cls March 19, 2019 12:34
 
a  GnTIII overexpression b  FUT8 knockout
h  Cell lines lacking FUT8
c  GMD knockout
0?2 0?2
Low-fucose or
afucosylated mAbs
g  Fucose analogs d  FX knockout
f  GDP-Fuc transporter knockout e  RMD overexpression  
Figure 3
GlycoFenigginuereeri n8g.m Getlhyodcsoteonprgoidnueceelroiwn-gfu cmoseeothr noodnsfu ctoos yplarteoddmuAcbes ilnovwivo-.f(ua)cOovseere xoprre snsioonfouf GconTsyIIlIainteCdH mO cAelbls.s 
(b) Knockout of FUT8 in CHO cells. (c) Knockout of GMD in CHO cells. (d) Knockout of FX in CHO cells. (e) Overexpression of
RMD in CHO cells. ( f ) Knockout of GDP-fucose transporter in CHO cells. (g) Use of fucose analogs that block GDP-fucose
produicnti ovni(vinoh.i b(itaio)n OofvGeMreDx).p(hr)eEsxspiroesnsi oonfi nGcenllTlinIeIsI thinat hCavHe lOow corenllose.n (dbog)e nKounsoFcUkTo8uetx porefs siFonU. ATb8br eivnia tCionHs:O 
CHO, Chinese hamster ovary; Fuc, fucose; FUT8,?1,6-fucosyltransferase; GMD, GDP-mannose 4,6-dehydratase; GnT,
N-acectyelglllusc.o s(acm)i nKyltnraonscfekroasue;tm oAfb s,GmMonoDcl oninal aCntHiboOdi esc;eRlMlsD. , (GdD)P -K6-ndeoocxyk-od-ulytx oo-hfe xFosX-4 -uinlo seC-4H-reOdu cctaeslel.s. (e) 
InOadvdeitrieoxnptoremsasmiomna lioanf cRellMlinDe  gilnyc oCenHgiOne ecreinlg, nonmammalian cell lines, including yeast(31), insect (60, 61), fungi (62) and plant (29, 63, 64), thaltsl.a c(kfe) nKdongeoncokuosuFUt Tof8 aGctDiviPty-wfuerceoaslseo transporter in 
explored for the production of nonfucosylated antibodies. While these nonmammalian cells lack
the bCioHsynOth ceteicllsp.a t(hgw)a yUtsoec orfe ffuucossyela taionna,lothgesy tphroadt ubcleomckan Gy nDoPnm-faumcmoasleia nprgolydcuocfotrimons  (inhibition of 
that could be immunogenic, such as the hypermannosylated glycoforms found in yeast and the
xyloseGaMndD?)-.1 ,3(h-f)u cEosxylpatreedssgilyocno foinrm cs efolulsn dthinapt lahnatsv. eT hleoswe l imoirt antioon sehnadvoegbeeennoaudsd reFsUsedT8 expression. 
by many studies that knock out the genes responsible for nonhuman glycoforms and overexpress
genesAthdadt ilteiaodntoalm oarbebmraetvurieathiuomnasn:i zeFdXN,- glGycDanPs -(341-,k6e5t?o7-06).-TDh-edmeoajxorygmenaentincoaspproachesfor glycoengineering to generate specific and enriched antibody glycoforms are summear izeepdiimn erase, RMD, 
Figure 3.
resistant maltodextrin.20  
Bios ynthetic Pathway Glycoengineering to Produce Antibodieswith Increased Galactose or Sialic Acid Content
 
Proba bly owing to steric hindrance, the Fc N-glycans of a typical recombinant IgG antibody usu-
ally c arry a low content of terminal galactose (approximately 80% are G0F and G1F glycoforms)and an even lower content (less than 10%) of terminal sialylation.Because terminal galactosylation
appea rs to be important for complement-dependent cytotoxicity (CDC) (71) and Fc sialylation is
respo nsible for the anti-inflammatory activity of IVIG (72), there have been efforts to enhance
the p roduction of these specific glycoforms for various functional studies and to develop more
efficie nt therapeutics.
 
 www.annualreviews.org ? Glycoengineering of Antibodies 26.7
 
 
 22 
 
Annu. Rev. Biochem. 2019.88. Downloaded from www.annualreviews.org
 Access provided by University of Maryland - College Park on 03/29/19. For personal use only. 
 
D. Antibody glycosylation remodeling by chemoenzymatic modifications   
Although the biosynthetic glycan engineering strategy has been successful in 
producing some important antibody N-glycoforms, there are several clear limitations 
associated with this approach.25, 31, 59 First, only a limited number of N-glycoforms 
can be efficiently generated using this method based on the existing biosynthetic 
pathways of N-glycans in conventional expression systems. Certain intermediate or 
aberrant glycoforms, such as the hybrid and triantennary N-glycans, might be difficult 
to obtain by manipulating the natural biosynthetic network.85 In addition to the target 
diversity issue, accurate regulation of antibody glycosylation is still challenging, as 
the levels of key enzymes involved in the production of antibody glycoforms are 
highly sensitive to the biochemical and physiological state of the cells used for 
expression, which can be easily affected by many factors during manufacturing.86 
An alternative strategy to optimize the N-glycosylation pattern of antibodies is 
to directly modify the native antibody N-glycans in vitro through the use of glycan 
processing enzymes and chemically modified sugar substrates. For example, partial 
enrichment for the terminal galactose and bisecting GlcNAc on an IgG antibody was 
achieved by treating the antibody with a recombinant GnTIII.87 Functional studies on 
these enriched antibody glycoforms revealed that the terminal galactose could affect 
the CDC activity of IgG regardless of the bisecting GlcNAc. In 2014, another study 
from the Bosques group further demonstrated that glycoengineered IVIG preparation 
with increased levels of galactosylation and sialylation exhibits better anti-
inflammatory activity than the primary IVIG.88 However, a broad application of this 
glycoengineering strategy is hindered by the limited availability of suitable starting 
 23 
 
 
antibody N-glycoforms. In addition, a large library of glycosidases and 
glycosyltransferases with different substrate specificities are needed for the 
construction of a full-length glycan chain by the step-wise addition or removal of 
single-sugar building blocks. As a result, the overall yield of this total synthesis might 
be problematic.4   
Instead of building an entire oligosaccharide chain with one sugar molecule at 
a time, an alternative strategy using a different family of glycan-modifying enzymes, 
called endo-?-N-acetylglucosaminidases or endoglycosidases (ENGase), seems to be 
much more convergent. In this method, the full-length or modified oligosaccharides 
are used as building blocks for ENGases to transfer to the glycoproteins, a process 
known as ?transglycosylation?.89 This strategy typically involves two major steps: (1) 
the release of the endogenous N-glycans from the heterogeneous antibody starting 
material (deglycosylation), and 2) the enzymatic formation of a new glycosidic 
linkage between the innermost GlcNAc residue of the deglycosylated protein and an 
activated N-glycan substrate using the catalytic activity of a mutant endoglycosidase 
(transglycosylation). The general concept of this strategy for glycoprotein and 
glycopeptide engineering is briefly illustrated in Figure 9.  
 24 
 
 
 
Figure 9. General concept of the ENGase-mediated glycosylation remodeling strategy 
of (a) glycoprotein and (b) glycopeptide. 89 
1. Introduction to endo-?-N-acetylglucosaminidase (ENGase)  
The endo-?-N-acetylglucosaminidases (ENGases) or endohexosaminidases 
comprise a family of glycoside hydrolases (EC 3.2.1.96)90 that catalyze the cleavage 
between the two GlcNAc residues within the diacetylchitobiose core of N-glycans.91 
The discovery of the first ENGase dates back to the early 1970's, and since then a 
large number of ENGases have been cloned from a series of organisms, such as 
bacteria, fungi, and higher organisms including mammals. Unlike other carbohydrate-
modifying enzymes, ENGases possess the unique ability to hydrolyze N-glycans 
attached to other molecular structures, such as glycoproteins and glycolipids. All 
ENGases found in the Carbohydrate-Active enZYmes Database (CAZY) can be 
divided into two sub-classes based on their primary sequence features: the glycoside 
hydrolase family GH18 and GH85,91 both sharing a conserved (?/?)8-TIM-barrel fold 
 25 
 
 
containing eight pairs of parallel ?-sheets and ?-helices. The crystal structure of an 
typical bacterial ENGase of GH85 family called Endo-A from Arthrobacter 
protophormia, is shown in Figure 10, where the key residues conserved in the loop 
area of the (?/?)8-TIM-barrel domain are responsible for the glycoside hydrolytic 
activity (highlighted by the arrow).92 
 
 
Figure 10. Crystal structure of endoglycosidase Endo-A (PDB: 3FHA) from two 
views. Parallel ?-sheets and ?-helices are highlighted in yellow and red, respectively. 
The catalytic domain (residues 2?350) of Endo-A contains a (?/?)8-TIM-barrel 
structure. The active site region is marked by an arrow.92 
 26 
 
 
2. Catalytic mechanism of ENGases and development of sugar oxazolines as 
activated glycan substrates for transglycosylation 
Although by definition the dominant enzymatic activity of ENGase is to 
release the N-linked oligosaccharides from the protein/peptide backbone, leaving only 
a single GlcNAc residue attached at the glycosylation site, some ENGases also retain 
the ability to re-attach N-glycans back to the exposed GlcNAc residue through a 
reverse reaction93. The dual functionalities of ENGases form the basic principles of 
the highly convergent chemoenzymatic glycan remodeling strategy mentioned in the 
previous section (Figure 9), and this method can be applied to synthesize both the 
homogeneous glycoproteins and glycopeptides with highly controlled regio- and 
stereo-selectivity.94  
Despite the diverse substrate specificities, all ENGases from the GH18 and 
GH85 families catalyze the hydrolysis of glycoproteins using two conserved acidic 
residues, a general acid-base and a nucleophile, in a two-step mechanism, as 
illustrated in Figure 11.89 For the GH18 members, the two key residues in the active 
site are typically a glutamic acid and an aspartic acid, whereas the aspartic acid is 
replaced by an asparagine at the same position in the GH85 enzymes. A major 
difference between the reaction mechanisms of ENGases and other ?-glycosidases is 
that the ENGases utilize neighboring group participation of the 2-acetamide from the 
second GlcNAc residue in catalysis.95 Based on this unique substrate-assisted 
mechanism, an oxazolinium ion or oxazoline was identified as a transition-state 
intermediate formed during the hydrolytic reaction of ENGases. A critical discovery 
in the field was the use of chemically synthesized N-glycan oxazolines as high-energy 
 27 
 
 
substrates to induce ENGases to catalyze the transglycosylation reaction. This is 
particularly useful as the activated glycan oxazolines may react with another GlcNAc 
residue of a different glycoprotein or glycopeptide to form a new glycosidic linkage 
in the final product.96 
Realizing the potential of glycan oxazoline in transglycosylation, in 2001 
Shoda and colleagues designed the first disaccharide oxazoline (Man?1-4GlcNAc) 
and used it as a probe to screen for potential transglycosylation activities of a library 
of ENGases.97 Since then, the disaccharide oxazoline was used as substrate for 
several ENGases, including Endo-A and Endo-M, to catalyze the synthesis of some 
relatively simple compounds such as trisaccharides.98 Although these earlier studies 
did not achieve efficient transglycosylation with more complicated glycosyl 
acceptors, such as glycopeptides or glycoproteins, they indeed laid the foundation for 
future use of more sophisticated oxazolines as the preferred substrates for the 
ENGase-mediated synthesis. One major advantage of this method was that the 
ENGases would not hydrolyze the product N-glycans containing truncated or 
modified sugar structures, which made it possible to achieve unidirectional synthesis 
of certain N-glycan conjugates.  
 28 
 
 
 
Figure 11. Catalytic mechanisms of ENGases showing key catalytic residues for some 
enzymes from: (a) family GH18; (b) family GH85.96  
 
3. Glycosynthase mutants of ENGases for antibody glycan remodeling 
Despite the fact that some modified N-glycans can be resistant to the intrinsic 
hydrolytic activity of ENGases, most full-length N-glycans of interest will still be 
rapidly hydrolyzed after they are transferred to the target glycoprotein-GlcNAc by the 
wild-type ENGases. This property greatly limited the application of ENGases in the 
general method of glycan remodeling. To solve this problem, Wang and Yamamoto 
investigated the possibility of using mutant ENGase generated by site-directed 
mutagenesis as ?glycosynthase? enzymes specialized in protein transglycosylation.99 
Based on the catalytic mechanisms mentioned earlier (Figure 11), a second catalytic 
 29 
 
 
residue known to be critical for the stabilization and deprotonation of the oxazolinium 
ion intermediate, an asparagine for GH85 or an aspartate for GH18, was selected as 
the site for mutation. Using this design, the first glycosynthase mutant of ENGases, 
developed together by Wang and Yamamoto, was generated by substituting the Asn-
175 residue of Endo-M with an alanine residue (Figure 12a). The generated mutant 
showed much higher transglycosylation activity than the wild-type Endo-M. 
Systematic mutagenesis at this site and functional assays demonstrated that an N175Q 
mutant was the most efficient variant in terms of synthesizing glycopeptides using 
sugar oxazolines, and this mutant later became a commercially available product.100  
Following this breakthrough, Wang and others applied this strategy to 
generate glycosynthase mutants of other members of GH18 family such as Endo-D101 
and Endo-A102, and these mutants proved to be useful for various synthetic purposes 
with distinct substrate specificities. Finally, studies by Wang and coworkers revealed 
that analogous mutations could be made at an aspartic acid in the active site of GH18 
ENGases such as Endo-S103 and Endo-F385 to generate novel glycosynthase mutants 
of this GH family.  According to previous studies on the relative activities of ENGase 
mutants, alanine or glutamine substitution seemed to yield the most effective 
glycosynthase mutants from a given ENGase.89 However, as a rapidly increasing 
number of new ENGases and mutants are being discovered and developed to expand 
the synthetic toolbox available, more systematic mutagenesis and functional studies 
are needed to verify whether other mutations at this catalytic N/D residue can yield 
more effective glycosynthase mutants than the previously reported A/Q substitutions. 
Finally, it is worth reiterating that these glycosynthase mutants of ENGases lack the 
 30 
 
 
activity to utilize the glycopeptides containing full-length N-glycans as substrates for 
transglycosylation, as the initial step of N-glycan hydrolysis from the glycopeptides 
cannot happen without the participation of the second catalytic residue.104 On the 
other hand, wild-type ENGases have the ability to catalyze such reactions but an 
enormous amount of glycopeptides are required to offset the dominant hydrolytic 
activity against the product.105 
 
 
 
Figure 12. Mechanism-based design of mutant glycosynthases of the ENGases: (a) 
mutations of the catalytic base (family GH18: D, family GH85: N) to alanine; (b) 
mutations of the catalytic base (family GH18: D, family GH85: N) to glutamine.  
 
 
 
 
 
 
 
 
 
 
 31 
 
 
4. Examples of ENGase-mediated antibody glycoengineering  
4.1. A brief history of antibody glycoengineering using ENGases and mutants 
 
In recent years the application of ENGase-mediated glycoengineering has 
made significant contributions to the synthesis of various antibody N-glycoforms, 
both naturally existing and artificially modified. One barrier to broad application of 
this strategy is that the majority of the IgG-Fc N-glycans produced in mammalian 
systems contains core fucose, which is not recognized by many ENGases. To 
circumvent this obstacle, Wang and colleagues in 2008 used a yeast expression 
platform to produce IgG-Fc fragment containing non-fucosylated oligomannose N-
glycans as the starting glycoforms, which could be easily hydrolyzed by many 
commonly used ENGase such as Endo-H.106 After deglycosylation, the truncated 
tetra- and hexa-saccharides were subsequently transferred to the exposed GlcNAc 
residue at Asn-297 by a bacterial ENGase Endo-A (Figure 13a). In this study, Endo-
A demonstrated the ability to transfer truncated or modified full-length N-glycans to 
IgG-Fc under relatively mild conditions, and the resulting IgG-Fc glycoform library 
was used to test the specific functional roles of N-glycans including the bisecting 
GlcNAc. However, limitations were realized: Endo-A could not transfer intact 
complex-type bi- or tri-antennary N-glycans to IgG-Fc, and the related studies on 
Endo-M revealed the same restricted substrate specificities.107 
The situation was improved when Wang and others discovered that another 
endoglycosidase called Endo-D from bacterium Streptococcus pneumoniae was able 
to use both the core fucosylated and non-fucosylated IgG-Fc fragments as substrates 
for transglycosylation reactions.101 Based on this discovery, native antibodies 
 32 
 
 
carrying the core-fucosylated N-glycans could be targeted for glycan remodeling. An 
example was demonstrated that the Endo-D N322Q mutant could transfer the simple 
core Man3GlcNAc glycan oxazoline to a commercial therapeutic antibody rituximab 
(Figure 13b). Nevertheless, neither Endo-A nor Endo-D could modify IgG-Fc with 
full-size complex type N-glycans. These limitations encouraged similar studies to be 
conducted on ENGases from the GH18 family such as Endo-S103 and Endo-F385, and 
very promising results were obtained with their glycosynthase mutants in terms of 
antibody glycoengineering. 
 
 
Figure 13. Selected examples of antibody glycan remodeling using ENGases: (a) 
Endo-A catalyzed attachment of a truncated core glycan to the non-fucosylated Fc 
region of human IgG; (b) Endo-D N322Q mutant catalyzed attachment of a truncated 
core glycan to the fucosylated Fc region of the mAb rituximab; (c) Endo-S D233Q 
mutant catalyzed attachment of a complex type biantennary glycan to the fucosylated 
intact rituximab.89 
 33 
 
 
4.2. Chemoenzymatic glycan remodeling of intact IgG antibodies by GH18 ENGases 
Endo-S and End-F3 
The first glycosynthase mutants of Endo-S, Endo-S D233A and D233Q, were 
developed by Wang and co-workers in 2012 and were subsequently applied to glycan 
remodeling of IgG antibodies with a series of complex biantennary N-glycans.103 The 
Endo-S mutants in combination with activated sugar oxazolinium substrates enabled a 
highly efficient glycan remodeling of a therapeutic IgG antibody, rituximab, 
originally carrying a mixture of the G0F, G1F, and, G2F glycans at the Fc region 
(Figure 14). The homogeneous N-glycoforms produced include a fully sialylated 
S2G2F form that was proposed to provide anti-inflammatory activity, a completely 
de-fucosylated G2 form that would display an enhanced ADCC activity versus its 
G2F form, and an artificial N-glycoform linked with an azido group that could be 
further coupled to other functional groups via orthogonal chemistry. 
Subsequent studies by other groups further demonstrated the applicability of 
Endo-S/Endo-S D233Q in preparing well-defined N-glycoforms of IgGs for various 
structural and functional purposes.42, 108, 109 However, Endo-S mutants could not act 
on the high-mannose or tri-antennary N-glycan oxazolines for transglycosylation. In 
this regard, the Wang group explored the intrinsic transglycosylation activity of a 
unique GH18 ENGase Endo-F3 that exhibited the ability to process the triantennary 
N-glycans in the presence of the core fucose. A glycosynthase mutant of Endo-F3, 
Endo-F3 D165A, was found to be able to transfer both bi- and tri-antennary complex 
type N-glycans to the fucosylated IgG acceptor in a highly efficient manner.85 
 
 34 
 
 
 
Figure 14. Glycosylation remodeling of IgG antibody rituximab using Endo-S and 
Endo-S glycosynthase mutants. Commercial therapeutic antibody rituximab was first 
deglycosylated using the wild-type Endo-S to generate fuc-?1,6-GlcNAc-rituximab, 
from which the core fucose could be removed by an ?-1,6-fucosidase to generate the 
GlcNAc-rituximab. To the deglycosylated antibody with and without fucose, different 
N-glycans were transferred to the core GlcNAc by Endo-S D233A or D233Q in the 
presence of glycan oxazolines, generating homogeneous rituximab glycoforms.103  
 
 
 
 
 35 
 
 
4.3. Chemoenzymatic glycan remodeling of intact IgG antibodies by Endo-CC 
A more recently discovered GH18 ENGase called Endo-CC from Coprinopsis 
cinerea was found to be able to act on both complex and high-mannose glycans.105, 
110-112 Inspired by the earlier studies, the glycosynthase mutants of Endo-CC were 
generated, and the mutant Endo-CC N180H was found to be most efficient in 
transferring the sialobiantennary complex N-glycan to various glycosyl acceptors. 
However, several aspects regarding the detailed catalytic properties of Endo-CC 
remain unclear. First, the substrates tested in Endo-CC transglycosylation were 
limited to the complex type N-glycans and glycoconjugates; therefore it is unknown 
whether Endo-CC and/or its glycosynthase mutants can react with the hybrid and 
high-mannose glycan structures. Secondly, current studies provided no experimental 
results concerning the tolerance of Endo-CC enzymes against the core fucose, which 
is a very important structural element in antibody glycosylation, as demonstrated in 
the previous section. Finally, the reported experiments related to the relative activities 
of Endo-CC mutants suggested that, among the 19 Endo-CC N180 mutants tested in 
the study, only two variants Endo-CC N180H and N180Q showed significant 
transglycosylation activities, which was not the case in many of the previously 
reported mutational studies of other ENGases. Additionally, it is still a question under 
debate whether the mutants of Endo-CC can independently catalyze the transfer of 
activated glycan oxazolines to the deglycosylated antibody.112 A more detailed 
discussion for Endo-CC in antibody glycoengineering is given in Chapter 4 of this 
dissertation.  
 36 
 
 
E. Scope and limitations of current chemoenzymatic antibody glycoengineering 
strategy mediated by ENGases 
Although the application of ENGases has proved very successful in producing 
homogeneous N-glycoproteins such as antibodies, there are still important functional 
limitations and unknown aspects associated with this glycoengineering strategy, 
which include the overall efficiency of the transglycosylation reactions by Endo-S, 
the restricted substrate specificities of Endo-S and other ENGases against additional 
N-glycans and acceptors, and the potential use of ENGases to remodel glycosylation 
of other antibody isotypes such as IgE and IgA.  
1. Efficiency of transglycosylation catalyzed by the existing Endo-S glycosynthase 
mutants Endo-S D233Q and Endo-S D233A 
First, regarding the efficiency of the transglycosylation reactions catalyzed by 
the Endo-S mutants, only two glycosynthase mutants of Endo-S, D233Q and D233A, 
were generated and tested so far for their glycosylation efficiencies against N-glycan 
oxazolines. Similar studies on other ENGases such as Endo-M raised the question of 
whether substitutions of the catalytic residue Asp-233 with additional amino acids 
may lead to new glycosynthase mutants with further enhanced transglycosylation 
efficiency for antibody remodeling. More importantly, there has been no mechanistic 
information, such as the kinetic parameters of Endo-S enzymes, available to help 
understand how Endo-S and its mutants catalyze the reactions differently. Finally, 
despite the relatively high efficiency of the Endo-S D233Q and D233A mutants, an 
excess amount of glycan oxazoline substrate (approximately 20 molar equivalents to 
the accepting antibody) is required to ensure a nearly complete transfer of the glycan. 
 37 
 
 
Novel ENGase mutants that can achieve a similar level of transglycosylation with less 
consumption of glycan substrates might be of great interest to facilitate more effective 
glycosylation remodeling of therapeutic antibodies. 
2. Substrate specificity of current ENGases for different N-glycans 
Although the Endo-S mutants could act on the biantennary complex type and 
truncated Man3GlcNAc glycans, they were unable to transfer the high-mannose and 
hybrid type N-glycans. The same restricted substrate specificity was observed for the 
Endo-F3 mutant D165A, which could not work on the non-fucosylated antibodies.85 
All these results indicated the need of a more versatile glycoengineering strategy with 
a relatively relaxed substrate selectivity. Adding to the limitations of the current 
glycan remodeling method is the fact that the antibody intermediate generated after 
each step of this multiple-reaction procedure needs to be purified from the reaction 
mixture before switching to the glycosynthase mutant. The concept of a potential 
?one-pot? method to optimize N-glycans of antibodies without the need of alternating 
enzymes and purifying intermediates might be worth pursuing to make this general 
strategy more effective.  
3. Glycosylation remodeling of antibodies beyond the IgG isotype  
So far all the ENGase-mediated glycan engineering efforts are focused on the 
IgG isotype of antibodies, and it remains unclear whether this general strategy using 
the chemoenzymatic modifications for glycan remodeling can be expanded to other 
antibody isotypes including IgA and IgE. This is becoming increasingly important as 
more studies are emerging to shed light on the functional significance of glycan 
 38 
 
 
structures of IgE antibodies. For example, a recent report suggested that a unique 
oligomannose N-glycan is critical for the IgE-induced allergic reactions.113 Access to 
synthetic IgE glycoforms with defined and modified glycan structures can potentially 
help us determine the IgE glycan participation in receptor binding and other cellular 
events related to anaphylaxis. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 39 
 
 
F. Organization of the dissertation 
The first chapter of this dissertation introduced the importance of antibody 
glycosylation and the basic concept of using chemoenzymatic glycoengineering to 
remodel the antibody N-glycans. A brief review of previous studies was also included 
to demonstrate how endoglycosidases were applied to synthesize homogeneous 
antibody glycoforms with well-defined glycan structures for various structural and 
functional studies.  
The second chapter focuses on the study that was published in Biochemistry 
in 2018.  In this work, several aspects of the ENGase-mediated glycan remodeling 
strategy were revisited to improve the overall efficiency of antibody glycosylation. 
Results of saturation mutagenesis on a previously identified ENGase Endo-S revealed 
additional glycosynthase mutants that could use glycan oxazolines as substrates for 
transglycosylation. A novel mutant Endo-S D233M was found to be more efficient in 
catalyzing the transfer of N-glycans than the previously identified Endo-S variants. 
Kinetic studies on selected Endo-S mutants provided new insight on the underlying 
mechanism of ENGases. A potent thiazoline-based inhibitor for Endo-S was also 
developed based on a unique substrate-assisted reaction mechanism. Finally, possible 
transglycosylation reactions using a reduced amount of glycan and the proposed non-
enzymatic side reactions were also investigated to help improve the effectiveness of 
our strategy. 
The third chapter describes our efforts in expanding the scope of the current 
glycosylation remodeling strategy beyond the restricted substrate specificity of some 
ENGases. Results of the mutational studies on another ENGase Endo-S2 identified 
 40 
 
 
new glycosynthase mutants that could act on all three major types of N-glycans with 
higher efficiency than the previously reported Endo-S mutants. Furthermore, results 
on the unique substrate specificity of the wild-type Endo-S promoted us to develop a 
one-pot glycoengineering strategy to synthesize homogeneous high-mannose or 
hybrid glycoforms of antibody in a more effective fashion. The work about Endo-S2 
mutants was published in the Journal of Biological Chemistry in 2016 and I made 
significant contribution to this work as a co-author. The work regarding the one-pot 
glycan remodeling of antibody was published in Bioorganic & Medicinal Chemistry 
in 2018.  
The fourth chapter presents our study on a recently discovered ENGase Endo-
CC. Previous studies on Endo-CC mutants indicated that only Endo-CC N180H and 
N180Q mutants showed significant transglycosylation activity with the complex-type 
glycan. Our results from substrate screening revealed that the Endo-CC mutants could 
transfer both the high-mannose and hybrid type glycans to the target glycopeptide. 
Core fucosylated substrates are resistant to the activity of both Endo-CC and mutants. 
In addition, comparison between the transglycosylation reaction time courses of 
different Endo-CC mutants demonstrated that, besides Endo-CC N180H and N180Q, 
other mutants of Endo-CC (N180Q, N180E and N180A) also retained the ability to 
use complex oxazoline for glycosylation reaction. Overall, these results improved our 
understanding of the catalytic properties of Endo-CC, which could be exploited for 
glycoengineering of glycopeptides and glycoproteins. The manuscript describing this 
work is in preparation.  
 41 
 
 
The fifth chapter detailed our attempt to perform glycosylation remodeling on 
another class of antibody, IgE. Recent mutagenesis and cellular studies on IgE 
glycans revealed a single oligomannose N-glycan that was indispensable for IgE 
biological activity. Biophysical analysis on the synthetic IgE-Fc N-glycoforms 
generated by a combined method of in vitro enzymatic modifications and biosynthetic 
engineering revealed a relatively small but significant reduction of binding affinity to 
the Fc?Rs upon deglycosylation. Glycosylation remodeling of IgE-Fc with the 
complex type N-glycan is currently ongoing with the aim of developing a competitive 
inhibitor based on a glycoengineered IgE-Fc fragment to treat allergic disorders like 
asthma.  
The final chapter summarizes the studies conducted in this research and their 
significance in optimizing the glycosylation states of monoclonal antibodies. Future 
studies regarding the mechanisms and applications of ENGases for glycoengineering 
will also be proposed in this chapter to further improve our understanding of ENGase 
catalysis and to expand the general concept of chemoenzymatic glycan remodeling 
strategy to other glycoproteins and glycopeptides. 
 
 
 
 
 
 
 
 42 
 
 
Chapter 2: Generation and comparative analysis of new glycosynthase 
mutants from endoglycosidase S (Endo-S) for antibody glycoengineering 
A. Introduction 
The cellular process of attaching carbohydrate entities to proteins and lipids, 
also known as glycosylation, gives rise to a highly diverse repertoire of functionally 
important glycan structures that are critical for the survival of many organisms.25 
More than 50% of intracellular proteins and 90% of secreted proteins are 
glycosylated.33 In particular, approximately 70% of the protein therapeutics currently 
on the market or in clinical trials are glycoproteins, and glycans are often found to be 
important for the pharmacokinetics and pharmacodynamics of therapeutic proteins, 
particularly monoclonal antibodies.69 Recombinant antibodies typically carry a 
mixture of different glycan structures at multiple glycosylation sites, collectively 
described as antibody glycoforms, which are difficult to isolate or enrich in large 
quantities.86 Because certain glycoforms of antibodies may confer unique biophysical 
properties and biological functions, such as mediating cellular killing and regulating 
anti-inflammatory activity, optimization of antibody glycosylation can potentially 
help develop more effective antibody therapeutics without altering the amino acid 
backbone.  
The endo-?-N-acetylglucosaminidase (ENGase)-catalyzed glycan remodeling 
mechanism has become an attractive strategy to optimize the heterogeneous 
glycosylation pattern of monoclonal antibodies.89 As the initial step of antibody 
glycoengineering, the variable N-glycans of the starting antibodies are trimmed off by 
an ENGase that cleaves between the two GlcNAc residues within the chitobiose core 
 43 
 
 
of N-linked oligosaccharides. Subsequent re-glycosylation or transglycosylation of 
antibody is achieved by the intrinsic ability of ENGase to catalyze the reverse 
reaction of glycan hydrolysis using a defined N-glycan in the form of activated 
oxazoline as substrate. The efficiency of this approach was significantly improved 
with the development of glycosynthase mutants of ENGases, such as End-M and 
Endo-D, which retained the transglycosylation activity for glycosyl oxazolines, with a 
much reduced tendency to hydrolyze the product, as discussed in the previous 
chapter.104 However, none of the commonly available ENGases were able to transfer 
intact complex N-glycans to a fucosylated antibody acceptor, which promoted the 
search for new ENGases with broader substrate specificity and higher efficiency for 
antibody glycoengineering.   
As part of our efforts to identify new ENGases for antibody glycosylation 
optimization, a family GH18 endoglycosidase Endo-S from the bacterial pathogen 
Streptococcus pyogenes was selected and studied for its ability to modify antibody N-
glycans.114 The intrinsic activity of this secreted Endo-S to specifically remove N-
linked glycans from the Fc region of human IgG antibodies significantly improves the 
survival of Streptococcus pyogenes by helping evade the host immune surveillance, 
as the N-glycans are required for the optimal immunoregulatory functions of IgGs, 
such as receptor binding and complement activation.115 Although Endo-S could 
process all four subclasses of IgG antibody (IgG1-4), it showed no significant 
hydrolytic activity against other antibody isotypes such as IgA and IgM, indicating 
that Endo-S recognizes specifically the protein structure of IgG antibodies for glycan 
hydrolysis.  In addition to its substrate selectivity for the IgG antibody, only the 
 44 
 
 
complex type N-glycoforms of IgGs are susceptible to the Endo-S-catalyzed 
hydrolysis, whereas the hybrid and high-mannose N-glycans of IgGs are unaffected 
by Endo-S treatment.114 The strict substrate specificity of Endo-S for both the IgG-Fc 
domain and complex N-glycans makes it a suitable candidate for chemoenzymatic 
glycoengineering of antibodies with different complex N-glycoforms.   
Inspired by previous studies on other ENGases, the first glycosynthase 
mutants of Endo-S, D233Q and D233A, were generated by Wang and colleagues 
through site-directed mutagenesis103. Using the activated glycan oxazoline as the 
glycosyl donor substrate, both mutants exhibited the remarkable activity to efficiently 
transfer a fully sialylated complex type glycan to the fucosylated intact IgG antibody 
without significant hydrolysis of the product (Figure 15). Such modifications could 
not be achieved by other ENGases such as Endo-A or Endo-D due to their restricted 
substrate specificities.104 In addition, this work also demonstrated that, following the 
removal of N-glycans by Endo-S, the core fucose linked to the exposed GlcNAc at 
the glycosylation site of IgG-Fc could be subsequently released by an ?-fucosidase. 
Using this combined enzymatic process, Endo-S D233Q and D233A mutants could 
then transfer glycans to the non-fucosylated antibody to afford new glycoforms with 
no core fucose attached (Figure 16). Finally, the ability of Endo-S mutants to transfer 
the truncated and modified glycans to the deglycosylated antibody in a site-specific 
manner enabled further functionalization of the antibody molecules through 
orthogonal ligation chemistry, which could lead to the development of a new class of 
antibody therapeutics called antibody-drug conjugates.6 
 45 
 
 
 
Figure 15. Glycoengineering of a core-fucosylated therapeutic antibody rituximab. 
Both natural and chemically modified N-glycan oxazolines were used to synthesize 
the homogeneous antibody N-glycoforms by Endo-S and its glycosynthase mutants 
Endo-S D233A and Endo-S D233Q. The detailed chemical structures of the glycan 
oxaozline substrates used in this procedure are shown here.103 
 
 
 46 
 
 
 
Figure 16. Chemoenzymatic remodeling of core fucosylated antibody rituximab to a 
non-fucosylated G2 glycoform using ?-fucosidase and Endo-S mutants. The chemical 
structure of G2 glycan oxaozline substrate used in this procedure is shown here.103 
 
Although a collection of homogeneous N-glycoforms of IgGs can now be 
prepared using the glycosynthase mutants of Endo-S, there are some clear limitations 
and unknown aspects associated with this glycoengineering strategy. Firstly, and 
perhaps most obviously, only two glycosynthase mutants of Endo-S, D233Q and 
D233A, were generated and tested so far for their transglycosylation activities against 
the antibody. It is unclear whether additional mutations at the Asp-233 residue of 
 47 
 
 
Endo-S can lead to design of new glycosynthase mutants with distinctive hydrolytic 
and transglycosylation properties. Secondly, no mechanistic aspects of the catalysis of 
Endo-S and its mutants, such as kinetic properties, have been studied to date to help 
illustrate the molecular mechanism behind the different catalytic efficiencies among 
Endo-S mutants. Although the crystal structure of a truncated form of Endo-S (amino 
acid residues 98-995) was recently reported, no co-crystallized ligands were included 
in this structure and only limited information was provided to address the molecular 
mechanism by which the mutant Endo-S catalyzes the glycosylation of N-glycans.116 
A more in-depth understanding of Endo-S catalysis is needed to further improve the 
efficiency of the current antibody glycoengineering strategy.  
Due to the important biological functions of glycoside hydrolases (GH), a 
large collection of the oligosaccharide-based small-molecule inhibitors have been 
developed as either molecular probes to dissect the cellular functions of GHs or as 
therapeutic interventions to target certain carbohydrate-related diseases.117 Based on 
the substrate-assisted mechanism of ENGases (Figure 10), a class of thiazoline-based 
inhibitors were designed by replacing the anomeric oxygen of the sugar oxazolinium 
ion intermediate with a sulfur atom.118-121 However, these inhibitors may possess 
promiscuous inhibitory activities against multiple ENGases or even other GH families 
within the cells due to the poor substrate selectivity. To date, there are no effective 
inhibitors that selectively target Endo-S over other ENGases. Therefore, the restricted 
substrate selectivity of Endo-S shown for the complex type N-glycans may provide a 
unique opportunity for the development of new selective thiazoline inhibitors for 
Endo-S and mutants.  
 48 
 
 
Finally, it is worth noting that the non-enzymatic chemical ligation between 
glycan oxazolines and certain lysine residues of IgG antibodies has been reported as a 
competing side reaction during the ENGase-catalyzed transglycosylation. Davis and 
colleagues suggested that the non-selective glycosylation side reactions could happen 
when a large excess of oxazoline was used for transglycosylation.122 Although Davis 
proposed a set of modified conditions to suppress these unwanted side reactions, the 
proposed method was relatively tedious and required multiple additions of oxazoline 
substrate in the presence of a high concentration of an Endo-S mutant. It is currently 
unclear whether these non-enzymatic glycosylation reactions may present a major 
issue for conventional remodeling of antibody glycosylation using Endo-S mutant and 
a reasonable amount of oxazoline (~10-20 molar equivalents to antibody), as typically 
used in our ENGase-mediated glycoengineering strategy.  
In this chapter we describe a systematic mutagenesis study at the Asp-233 
residue of Endo-S and evaluation of the relative catalytic activities of the resultant 
Endo-S mutants.123 We found that substitution of Asp-233 with the other 19 canonical 
amino acids led to varied functional effects on the hydrolysis and transglycosylation 
activities of Endo-S, with the Endo-S D233M possessing the highest overall catalytic 
efficiency among the mutants studied, which was defined by the transglycosylation-
to-hydrolysis activity ratio. Kinetic studies on the D233M mutant of Endo-S, as well 
as the previously identified D233A mutant, indicated that the enhanced overall 
catalytic efficacy of the Asp-to-Met mutant over that of the Asp-to-Ala mutant can be 
attributed mainly to two factors: the increased turnover number for the glycan 
oxazoline substrate and the enhanced affinity for the antibody substrate. These 
 49 
 
 
findings provide a plausible mechanistic explanation for the enhanced catalytic 
activity of Endo-S D233M over the other mutants. In addition, based on the unique 
mechanism of Endo-S and other GH18 ENGases, we describe here a facile synthesis 
of a complex type N-glycan thiazoline analog as a new mechanism-based inhibitor for 
Endo-S and other complex glycan-specific ENGases.124 Finally, investigation on the 
potential non-enzymatic side reactions between glycan oxazoline and antibody will be 
discussed.  Collectively, these results elucidate key mechanistic aspects of Endo-S 
catalysis and can significantly improve the overall efficiency of the current Endo-S-
mediated glycoengineering strategy.  
 
 
 
 
 
 
 
 
 
 
 
 
 50 
 
 
B. Results  
1. Generation of additional glycosynthase mutants of Endo-S through site-directed 
saturation mutagenesis at the catalytic residue Asp-233.  
To systematically screen for more efficient glycosynthase mutants of Endo-S, 
a library of Endo-S mutants was constructed by site-directed saturation mutagenesis 
at the previously identified catalytic residue Asp-233 (Figure 17). Previous studies 
have demonstrated that this aspartate residue is directly involved in stabilization and 
deprotonation of the oxazolinium ion intermediate during endoglycosidase-catalyzed 
hydrolysis. Replacement of this catalytic residue with a non-participating amino acid, 
could potentially render the mutant incapable of processing the intact glycoproteins or 
glycopeptides, but still allows the enzyme to retain the ability to transfer the glycan 
substrates to different acceptors if an activated sugar oxazoline is provided.  
 
Figure 17. The general procedure of site-directed saturation mutagenesis of Endo-S 
using the degenerate primers.  
 51 
 
 
For the subsequent comparative studies of their enzymatic activities, all Endo-
S mutants were expressed and purified under the same conditions that were used for 
the expression of the wild-type Endo-S. The expression yield of Endo-S mutants was 
15?30 mg/L of bacterial cell culture, which is comparable to the previously published 
expression yield of Endo-S and Endo-S D233A and D233Q mutants.103 Endo-S and 
mutants were expressed as fusion proteins with a Glutathione S-transferase (GST) tag 
located at the C terminus to facilitate protein folding and affinity purification. To 
ensure that the GST domain does not perturb the normal enzymatic activity of Endo-S 
enzymes, the GST tag was removed by thrombin protease cleavage.125 Comparison of 
the hydrolytic activities of the untagged and fusion proteins of Endo-S demonstrated 
the same level of enzymatic activity under previously reported reaction conditions.114 
2. Transglycosylation and hydrolysis activities of Endo-S mutants 
The specific activity of Endo-S and its glycosynthase mutants for hydrolysis 
of complex glycoforms of antibody was evaluated and quantified using a commercial 
therapeutic monoclonal antibody rituximab as model substrate. For assessing specific 
transglycosylation activities (Figure 18), the native N-glycan mixture (G0, G1 and G2) 
of rituximab was first removed by Endo-S, and the purified rituximab-GlcNAc was 
used as the glycosyl acceptor substrate for the subsequent transglycosylation assays. 
The full-sialylated complex type (S2G2) N-glycan was extracted from egg yolk and 
converted to the oxazoline donor substrate by chemical modifications.100 The results 
of both activity tests are summarized in Table 1.  
 52 
 
 
 
Figure 18. Schematic presentation of hydrolysis and transglycosylation reactions by 
Endo-S and its mutants using antibody rituximab as starting material. Complete 
deglycosylation of the original N-glycans of rituximab was confirmed by ESI-LS-MS 
analysis. The rituximab-GlcNAc intermediate was then purified via Protein-A affinity 
chromatography and was subsequently used for transglycosylation reactions catalyzed 
by the Endo-S mutants generated by site-directed mutagenesis.123 
 
Comparison of the hydrolytic activities demonstrated that all mutants of Endo-
S except Endo-S D233C showed significantly decreased tendency to hydrolyze the 
complex type N-glycans from the intact rituximab in comparison to the wild-type 
Endo-S. Specifically, the mutants D233F, D233G, D233H, D233K, D233R, D233L, 
D233M, D233P, D233W, and D233Y only possessed <1% of the intrinsic hydrolytic 
activity of the wild-type Endo-S. End-S D233C retained a significant capacity (~33% 
of the wild-type Endo-S) to hydrolyze the complex N-glycans attached at the Asn-297 
glycosylation site located in the IgG-Fc region. Besides the mutants D233N, D233E, 
D233S and D233T, which showed only marginal activities, the D233G mutant also 
showed a relatively high hydrolytic activity towards rituximab despite the fact that the 
simple hydrogen functional group of glycine residue is typically considered incapable 
of participating in hydrogen bonding and many other interactions involved in enzyme 
catalysis.126 
 53 
 
 
The transglycosylation assay revealed that most Endo-S mutants were still 
able to transfer the biantennary complex type glycan oxazoline to the deglycosylated 
rituximab with varied efficiencies. In particular, Endo-S D233C, D233E, D233G, 
D233M, D233N, and D233S exhibited the highest transglycosylation activities 
among the glycosynthase mutants tested here. However, via comparison of the overall 
transglycosylation efficiencies, most mutants with potent transglycosylation activities 
also possessed significant hydrolytic activities against antibody, a property less ideal 
for the glycoengineering application using ENGases. For instance, Endo-S D233C 
demonstrated the highest activities for both hydrolysis and transglycosylation among 
the 19 mutants studied here, but its overall catalytic efficiency for antibody glycoform 
remodeling (defined by the ratio of transglycosylation to hydrolysis, T/H) was found 
to be moderate. Interestingly, the Endo-S mutant D233M still retained a relatively 
high transglycosylation activity among the mutants with a remarkably low residual 
hydrolytic activity. Therefore, the highest overall catalytic efficiency makes D233M 
the most efficient glycosynthase mutant of Endo-S for antibody glycoengineering 
among all the mutants whose enzymatic activities were studied in this work. 
On the basis of specific transglycosylation and hydrolytic activities by Endo-S 
mutants against antibody (Table 1), Endo-S D233M mutant showed an approximately 
7- and 8-fold enhanced synthetic efficiency (T/H) versus those of D233Q and D233A, 
respectively. To directly demonstrate the enhanced catalytic efficiency of D233M for 
IgG glycosylation, we monitored the progression of three glycosylation reactions by 
mutants of Endo-S, D233M, D233Q, and D233A, under the same reaction condition. 
The parallel reactions were monitored by LC-ESI-MS (Figure 19). The comparison of 
 54 
 
 
the time courses of all three reactions indicated that the mutant D233M was more 
efficient in catalyzing the glycosylation of rituximab than the previously identified 
Endo-S D233A and D233Q.  According to this comparison, the D233M-catalyzed 
glycosylation reaction reached over 95% conversion of the rituximab N-glycoform 
within 30 min, while the Endo-S D233A- and D233Q-catalyzed reactions gave an 
approximately ?60% conversion yield at 60 min under the same condition. 
 
Figure 19. Comparison of transglycosylation reactions catalyzed by Endo-S D233M 
and previously reported Endo-S D233Q and D233A mutants. Fully sialylated 
complex glycan oxazoline SCT-Oxa was used as the glycosyl donor substrate. The 
reaction was conducted by incubating deglycosylated antibody (10 mg/mL) and SCT-
Oxa with each mutant enzyme (0.05 mg/mL) in PBS buffer with a pH of 7.4 at 30 C? . 
For each reaction, 20 molar equivalents of SCT-Oxa was added to the deglycosylated 
rituximab in the presence of Endo-S mutants. The data sets presented here are 
representative of three independent experiments.123 
 55 
 
 
Table 1.  Comparison of specific transglycosylation efficiencies of Endo-S mutants 
measured by LC-ESI-MS using a commercial antibody rituximab as substrate.123 
 
aThe specific hydrolysis activity of each enzyme was assayed by incubating the 
synthetic biantennary complex type rituximab with different mutants or wild-type 
Endo-S (0.01 mg/mL) in PBS pH 7.4 at 30  C? . To test the specific transglycosylation 
activity of each enzyme, the deglycosylated rituximab was used as the glycosyl 
acceptor and the complex type oxazoline (SCT) as the glycosyl donor substrate with a 
final concentration of enzymes at 0.01 mg/mL. The molar ratio of the glycosyl 
acceptor to donor substrate was 1:20. The experimental results shown here were 
collected from the average data of three independently repeated assays.  
 56 
 
 
3. Enzyme kinetic analysis of Endo-S mutants D233M and D233A  
The observed difference in the catalytic efficiency of mutants Endo-S D233M 
and D233A prompted us to perform kinetic measurements on the two mutants. Using 
a fully sialylated biantennary complex glycan oxazoline (SCT-Oxa) as the donor and 
the deglycosylated rituximab antibody as the acceptor substrate, we measured the 
kinetic profiles of Endo-S D233M and D233A mutants for both substrates. A mass 
spectrometry-based approach combined with an intact antibody internal standard was 
used to more accurately monitor the formation of the transglycosylation product in a 
quantitative manner. The results are summarized in Tables 2 and 3.  
For the biantennary complex-type glycan oxazoline, the catalytic efficiency of 
Endo-S D233M (kcat/KM = 0.03 min?1 ?M?1) showed a 3-fold increase in comparison 
with that of Endo-S D233A (kcat/KM = 0.01 min?1 ?M ?1) (Table 2 and Figure 20a). 
The apparent increase in the (kcat/KM) of Endo-S D233M versus that of D233A was a 
result of an approximately 4-fold increase in the rate of turnover (as measured by kcat) 
and a 2-fold reduced substrate affinity for the oxazoline estimated by the KM values, 
as a higher KM  is indicative of a lower substrate affinity. These results suggested that 
substitution with a methionine instead of an alanine at the Asp-233 residue enhanced 
the enzymatic turnover number of Endo-S for the glycan oxazoline substrate with a 
moderate decrease in the substrate affinity. 
For the glycosyl acceptor substrate, the deglycosylated antibody, Endo-S 
D233M exhibited a substrate affinity much higher than that of Endo-S D233A. This 
was clearly supported by the Michaelis?Menten curves generated for the two mutants 
(Figure 20b), which demonstrated that Endo-S D233M was much easier to saturate 
 57 
 
 
with the antibody substrate than Endo-S D233A was. The KM value of D233M for the 
deglycosylated antibody was estimated to be 30 ?M, while the KM value of D233A 
was projected to be higher than 200 ?M (Table 3). On the other hand, the catalytic 
turnover number (kcat) of Endo-S D233A for the antibody substrate could not be 
accurately measured due to the fact that the concentration of antibody required to 
saturate the enzyme exceeded the solubility of antibody under normal solution 
conditions.  
 
Table 2. Kinetic parameters of Endo-S mutants D233M and D233A for SCT-Oxa. 
 
Table 3. Kinetic parameters of Endo-S mutants D233M and D233A for rituximab. 
 
*The catalytic turnover number of Endo-S D233A could not be accurately 
determined because the concentration required for the antibody substrate saturation 
could not be reached in the experiments under normal conditions.123 
 
 
 58 
 
 
a b
 
Figure 20. Michaelis-Menten plots of Endo-S D233M and Endo-S D233A. Kinetics 
for the two substrates were shown here: a) for the glycosyl donor, a gradient of SCT-
Oxa (from 6.25 to 400 ?M) was tested with a constant concentration of the 
deglycosylated rituximab (20 ?M); b) for the glycosyl acceptor, a serial dilution of 
rituximab (0.32 to 20 ?M) was tested with 400 ?M of SCT-Oxa. For both Endo-S 
D233M and D233A, the enzyme concentration was fixed at 0.01 mg/mL (0.067 ?M). 
The data sets presented here are representative of three independent experiments.123 
4. Chemical synthesis of a complex glycan thiazoline inhibitor for Endo-S 
The designed complex type glycan-based thiazoline inhibitor compound 2 was 
derived from the naturally existing sialylated biantennary complex type free glycan 
SCT 1 (Figure 21). The SCT free glycan was first released from the sialoglycopeptide 
(SGP) by the endoglycosidase Endo-M. The original SGP peptide was harvested and 
purified from commercial egg yolk powder according to our previously published 
procedure.127 The released free SCT glycan then underwent a series of chemical 
modifications to afford the desired complex type glycan thiazoline 2.  
The chemical synthesis of the thiazoline inhibitor presented in this work was 
achieved in collaboration with Dr. Chao Li from Prof. Lai-Xi Wang?s group at the 
 59 
 
 
Department of Chemistry and Biochemistry, University of Maryland, College Park. 
The work was published in Journal of Carbohydrate Chemistry in 2017.124 Following 
our previously reported method Dr. Li was able to successfully convert the free SCT 
glycan 1 to the thiazoline-based inhibitor compound 2.120, 128 We performed the 
majority of chemical synthesis and purification experiments together, including the 
peracetylation of SCT glycan, synthesis of thiazoline derivative by the Lawesson?s 
reagent, and purification of the final product by size-exclusion chromatography. Dr. 
Li and I also conducted the targeted enymatic inhibition assays with the synthetic 
thiazoline inhibitor 2 together with assistance from other co-authors listed in the 
publication.124 
 
OH OH
HO HO HO HO
O O O O OO O O
HO OH HO NHAc HO OH HO NHAc
HO HO
HO O HO O
HO HO
O OH HO O OH HO
O
HO O O HO O
O
O
O HO N O HOO N S
HO HO
HO
OH O
HO O
HO HO HO HO
OH
HO HO
O O OO O
O O
O
HO OH HO NHAc HO OH HO NHAc
1 2  
Figure 21. The chemical structure of the complex type N-glycan oxazoline (1) and the 
corresponding complex type N-glycan thiazoline (2).124  
 
 
 60 
 
 
5. Inhibitory activity assays for the synthetic thiazoline-based inhibitor  
The inhibitory activity of the synthetic thiazoline inhibitor compound 2 was 
examined against several commonly studied ENGases. To probe the specific 
inhibition of compound 2 against Endo-S, a fully sialylated and fucosylated complex 
N-glycoform of antibody rituximab (S2G2F-rituximab) was synthesized for use as 
starting material, following the previously reported glycoengineering strategy.103 The 
relative abundance of the free glycans released by ENGases in the presence or 
absence of the inhibitor was monitored by high-performance anion-exchange 
chromatography coupled with pulsed amperometric detection instrument (Dionex 
HPAEC-PAD). It was found that glycan thiazoline 2 showed potent inhibitory 
activity against Endo-S, with an IC50 of 1.4 ?M (Figure 22A). To examine its 
inhibitory activity against Endo-M, sialoglycopeptide SGP was used as the substrate 
and the released glycans were also measured by HPAEC-PAD analysis. The results 
suggested that the glycan thiazoline 2 was also effective in inhibiting the hydrolase 
activity of Endo-M, and the IC50 was 0.27 ?M (Figure 22B). If the IC50 value was 
taken as an indicator for the inhibition potency, the thiazoline inhibitor would be 
more active against Endo-M than Endo-S by approximately 5- and 2-fold, 
respectively. In addition, in comparison with the previously reported high-mannose 
type glycan-based thiazoline (Man9GlcNAc), which exhibited an IC50 of 6 ?M against 
Endo-M, the synthetic complex type N-glycan thiazoline was 22-fold more active.14 
 
 
 
 61 
 
 
100  Endo-S 100
 Endo-S2
80 80
60 60
40 40
20 20
0 0
0.01 0.1 1 10 100 0.01 0.1 1 10
Concentration of 2 (?M) Concentration of 2 (?M)  
Figure 22. Inhibition effects of Endo-S, Endo-S2, and Endo-M by CT-thiazoline (2). 
A) Inhibition of Endo-S and Endo-S2 by 2 using synthetic S2G2F-rituximab as 
substrate; B) Inhibition of Endo-M by 2, using the protected glycan SCT-Fmoc as 
substrate.124 
6. Investigation on non-enzymatic side reactions of glycan oxazoline  
The formation of non-enzymatic glycosylation products between glycan 
oxazoline and antibody during transglycosylation was investigated here by carefully 
characterizing the chemical composition of the product antibody glycoform after each 
reaction. A typical transglycosylation reaction using Endo-S D233M and SCT-Oxa 
was carried out to validate the efficiency of this enzyme-assisted glycoengineering 
strategy. Non-enzymatic control experiments were conducted by directly incubating 
the deglycosylated antibody with a gradient of concentrations of glycan oxazolines in 
the absence of Endo-S enzymes. The nature of the glycosidic linkage formed after 
each reaction was verified by treating the product antibody with a unique amidase 
enzyme called PNGase F, which releases all N-linked glycans from glycoproteins 
including antibody.129 Any oligosaccharides that remain attached to the amino acid 
 62 
 
Inhibition (%)
Inhibition (%)
 
backbone after the PNGase F treatment are likely to be linked through a non-
ENGase-catalyzed chemical linkage formed by the side reactions.  
The standard transglycosylation reaction using the mutant Endo-S D233M and 
a reasonable amount of SCT-Oxa (20 molar equivalents to the antibody) achieved 
complete transglycosylation within 30-60 min (Figure 23). Analysis of the PNGase F-
treated therapeutic antibody trastuzumab revealed no additional products besides the 
expected complex type antibody N-glycoform. On the other hand, control reactions 
with different glycan oxazoline concentrations (20, 40, 80, and 160 molar equivalents 
relative to the antibody) demonstrated that non-specific chemical ligation between the 
glycan oxazoline and antibody acceptor indeed occurred, but only when a large 
excess of oxazoline was used during a prolonged reaction time (Figure 24). 
 
 
 
 
 63 
 
 
 
Figure 23. Mass spectrometry analysis of the antibody N-glycoforms generated by 
glycan remodeling of a therapeutic antibody trastuzumab (Herceptin). The reaction 
was performed by incubating deglycosylated antibody (10 mg/mL) and SCT-Oxa 
with a small amount of Endo-S D233M (0.05 mg/mL). The molar ratio of antibody to 
oxazoline was 1/20. Other reaction conditions followed the previously conducted 
transglycosylation reactions, as detailed in Figure 19.  LC-MS analysis of the product 
glycoform revealed a single mass peak corresponding to the complex type 
trastuzumab N-glycoform (B), which confirmed complete transfer of complex glycan 
oxazoline. The product antibody after transglycosylation was subject to the PNGase F 
treatment, and heavy chain (C) and light chain (D) of the final product were analyzed 
separately by ESI-LC-MS under reducing conditions. The mass spectra that matches 
the mass of the PNGase F-treated heavy chain and light chain of antibody indicated 
that no significant non-enzymatic transfer of the complex glycan oxazoline under our 
conventional reaction condition.103 
 64 
 
 
30 20 eqivalents SCT-Oxa
40 eqivalents SCT-Oxa
20 80 eqivalents SCT-Oxa
160 eqivalents SCT-Oxa
10
0
0 120 240 360 480 600 720
Time/min
 
Figure 24. Non-enzymatic transfer of the sialylated complex type glycan oxazoline to 
the deglycosylated trastuzumab antibody in the absence of the Endo-S enzymes. The 
parallel reactions were conducted by incubating the deglycosylated trastuzumab and 
SCT-Oxa in a series of concentrations (20, 40, 80 and 160 equivalents to antibody) in 
PBS buffer with a pH of 7.4 at 30 ?C. Other reaction conditions follow the previously 
stated transglycosylation reaction, as detailed in Figure 19 and experimental section. 
The product antibody of each control reaction was treated by PNGase F overnight, 
and the heavy chain and light chain of the final product were analyzed separately by 
ESI-LC-MS under reducing conditions. Selected raw data such as the mass spectra of 
antibody with the non-enzymatically transferred glycans are shown in the Supporting 
Information section. 
 
 
 
 
 
 
 
 65 
 
Non-enzymatic Glycosylation
 (%)
 
C. Discussion 
We have previously identified two glycosynthase mutants of Endo-S, D233A 
and D233Q, which could transfer the intact and modified complex glycan oxazolines 
to Fc-deglycosylated IgG antibodies to afford homogeneous N-glycoforms. However, 
systematic mutagenesis study on Endo-S at the critical Asp-233 site had not been 
performed, and mechanistic insight into the different catalytic activities of these 
mutants is currently lacking. 
In this study, we conducted a systematic mutagenesis analysis on Endo-S at 
Asp-233 and examined the hydrolytic and transglycosylation activities of the 
resulting mutants. The comparative studies revealed several glycosynthase mutants of 
Endo-S including D233M that demonstrated enhanced transglycosylation efficiencies 
compared to those of the previously identified Endo-S mutants. The kinetic data 
clearly indicated that the enhanced catalytic efficiency of the Endo-S D233M mutant 
versus that of the D233A mutant could be attributed to mainly two factors: the 
increased turnover number for the oxazoline donor substrate and the enhanced 
affinity for the antibody acceptor substrate. Nevertheless, the molecular basis behind 
this observed enhancement of the catalytic efficiency by the Asp-to-Met mutation 
remains to be elucidated by more detailed structural and mechanistic studies using 
additional biophysical approaches. 
The first crystal structure of Endo-S was reported in 2014 by Sundberg and 
colleagues,116 and more recently, the molecular structure of a full-length Endo-S 
bound with a complex free glycan was also solved by the same group.130 In this co-
crystallized structure, a functionally inactive double-mutant of Endo-S, in which both 
 66 
 
 
the general acid/base residue Glu-235 and the key catalytic reside Asp-233 were 
substituted by leucine and alanine, respectively, was generated to abolish the intrinsic 
hydrolytic activity of Endo-S and to facilitate the crystallization process. The crystal 
structure provides a possible binding mode for the sugar residues of the complex type 
G2 glycan substrate by Endo-S, including the first GlcNAc (+1) residue. Molecular 
docking simulations suggest additional amino acids that make contact with the 
glycan. The side chain of T281 interacts with the 06 atom of GlcNAc (+1) through a 
hydrogen bond, whereas the N-acetamido group of the GlcNAc is also stabilized by 
the main chains of two ?gating residues?, Q303 and Y305. In additional, hydrophobic 
interactions are implicated between GlcNAc (+1) and Y305 and W358 (Fig. 25b). 
Besides Y305, the structure also suggests that a D231 residue is involved in the 
interactions that stabilize the protonated D233-E235 network by providing a negative 
charge, and the transition state is stabilized by Y402 (Figure 25c).  
Although these reported structures and related molecular simulations revealed 
important information about Endo-S overall conformation and how the non-reducing 
end oligosaccharides of the free glycan interact with specific amino acids at the active 
site of Endo-S, these existing crystal structures with the ground state N-glycans may 
not accurately display how the oxazolinium ring is recognized by the glycosidase 
domain of Endo-S. One potential solution is to determine the structure of Endo-S in 
complex with a stable, non-hydrolyzable glycan oxazoline analog, for example a 
glycan thiazoline101, which we recently synthesized and tested.124 In addition to the 
glycan substrate, previous molecular simulations suggested that both Endo-S and IgG 
antibody would undergo significant conformational changes prior to the hydrolysis of 
 67 
 
 
N-glycans.116 Interestingly, kinetic analysis also indicated that the substrate affinity of 
Endo-S mutants for the IgG antibody is largely affected by the mutation at Asp-233 
(Figure 19 and Table 3). Therefore, it would be of great interest to determine the 
molecular structure of an antibody-bound Endo-S to help understand the mechanism 
behind the enhanced substrate affinity for IgG due to the methionine substitution. 
 
 
Figure 25. Molecular mechanism of Endo-S based on the co-crystalized structure with 
the G2 glycan substrate. a. Structure of the S2G2 glycan substrate based on molecular 
dynamic simulations. b. Structural alignment of the active site of Endo-F3 (gray) and 
Endo-S (orange). The residues of Endo-S are labeled with numbers. c. In the unbound 
form of Endo-S, no interactions can be observed between D233 and E235. In the first 
step of catalysis, structural displacement is found for both the GlcNAc (?1) reside and 
D233, making it possible for hydrogen bonding network between D233, E235 and the 
acetamido group. In the second step, E235 is protonated again and D233 was returned 
to its original position, facilitated by a deprotonated D231.130 
 68 
 
 
To further study the mechanistic aspects of Endo-S and its mutants, a facile 
synthesis of a complex N-glycan thiazoline compound was achieved starting from the 
sialoglycopeptide (SGP) that can be obtained easily from the chicken egg yolks in 
large quantities.131 This synthetic glycan thiazoline therefore represents a new type of 
mechanism-based inhibitor specially designed for the endoglycosidases with high 
selectivity for complex N-glycoproteins including antibodies. As mentioned above, 
this inhibitor may be useful as a transition state analog to form complexes with the 
ENGases during structural analysis. In addition, given the potential functional role of 
Endo-S in human infections by the pathogenic bacterium Streptococcus pyrogenes, 
this synthetic inhibitor may be further developed for its ability to control bacterial 
infections and pathogenesis as a potential therapeutic drug candidate.114 
Despite the effectiveness of the Endo-S mutants for antibody glycosylation 
remodeling, the potential non-enzymatic ligation reaction occurring between glycosyl 
oxazoline and protein residues has been suggested as a major issue for the application 
of glycan oxazoline in large-scale glycoengineering.122, 132 Previous studies using the 
tandem mass spectrometry (MS/MS) identified the amino groups (-NH2) of the lysine 
residues as potential chemical glycosylation sites for the activated glycan oxazolines.  
NMR spectroscopy also revealed that the oxazoline was chemically attached to the 
lysine residues of antibody through a different linkage from N-linked glycosylation, 
and this aberrant linkage is resistant to the PNGase F-catalyzed cleavage133. Here, by 
revisiting the nature of the glycosidic linkage formed in the product antibody 
glycoform, our results clearly demonstrated that, under our reported optimal 
conditions, glycoengineering of antibody using Endo-S D233M and SCT oxazoline 
 69 
 
 
could achieve complete transfer of glycans in a timely fashion without the formation 
of any non-enzymatic side products. By comparing the levels of the non-selective 
glycosylation under different oxazoline concentrations in the absence of enzyme, we 
further concluded that this non-selective chemical ligation reaction occurred only if a 
relatively high concentration of glycan oxazoline was used in the transglycosylation 
reaction over an extended time of incubation, which is normally not required for the 
complete transfer of glycan by Endo-S mutants. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 70 
 
 
D. Conclusion 
The glycosynthase mutants of Endo-S D233Q and D233A are important 
biocatalysts for remodeling the glycosylation patterns of antibodies. The use of 
endoglycosidases to engineer antibody glycosylation is highly selective and efficient 
because of controlled stereo- and regio-selectivity by enzymes. Therefore, generation 
of more efficient Endo-S mutants for the transglycosylation reaction could improve 
the current glycoengineering strategy of antibody glycoforms, which could potentially 
be applied to enhance the efficacy of antibody therapeutics in clinic applications. 
Here, we presented several novel Endo-S mutants such as Endo-S D233M that 
demonstrated potent transglycosylation activity with diminished product hydrolysis. 
Kinetic studies on the new mutants Endo-S D233M and the previously reported 
Endo-S D233A revealed important mechanistic information that could help us 
understand the different catalytic behaviors of the two Endo-S mutants studied here in 
catalyzing the transglycosylation of the IgG antibody. In addition, a new mechanism-
based inhibitor was designed for Endo-S and other ENGases that specifically target 
the complex type N-glycans. Further development of this inhibitor such as increasing 
its cell permeability could potentially lead to new therapeutic candidates for certain 
bacterial pathogenesis. Finally, the reported side reaction issue associated with the 
use of the activated glycan oxazoline was addressed in this chapter by characterizing 
the glycosylated antibody. In conclusion, the overall efficiency of transglycosylation 
process catalyzed by Endo-S glycosynthase mutants was significantly improved with 
the development of new glycosynthase mutants and the optimization of the reaction 
conditions.  
 71 
 
 
E. Materials and Methods 
1. Chemicals and biochemical materials 
All chemicals and biochemical materials were qualified as reagent grade or 
above. The reduced glutathione stock solution (Sigma-Aldrich) was prepared in 50 
mM Tris-HCl buffer at pH 8.0 and stored at 4 C? . The diluted glutathione elution 
buffer was prepared immediately before affinity purification to minimize oxidation.  
The monoclonal therapeutic antibody rituximab and trastuzumab were purchased 
from Genentech Inc. (San Francisco, CA). The antibody materials were purified by 
Protein-A affinity chromatography to remove any storage additives and stored in 
phosphate-buffered saline (PBS, containing 137 mM NaCl, 2.7 mM KCl, 8 mM 
Na2HPO4, and 2 mM KH2PO4, pH 7.4) at 4 C? . The fully sialylated complex-type 
glycan oxazoline was chemically synthesized according to the previously published 
procedure102. The concentration of antibody was measured on a NanoDrop 2000c by 
detecting absorbance at 280 nm using a molar extinction coefficient of 13700 M-1cm-1.  
High-performance anion-exchange chromatography with the pulsed amperometric 
detection (HPAEC-PAD) was performed on a Dionex ICS-5000 chromatography 
system (Fischer Scientific) equipped with an electrochemical detector (ED50) and an 
anion exchange column (PA200, 4 ? 250 mm). The mobile phase (flow rate, 0.5 
ml/min) was composed of 100 mM NaOH (eluent A) and 250 mM NaOAc in100 mM 
NaOH (eluent B). The gradient used was as follows: 0 min, 100% eluent A; 20 min, 
80% eluent A, 20% eluent B. The LC-ESI-MS was performed on an Exactive? Plus 
Orbitrap Mass Spectrometer (Thermo Scientific) equipped with a Waters XBridgeTM 
BEH300 C4 column (3.5 ?m, 2.1 ? 50 mm) and a C18 column (2.1 ? 10 mm, 1.5 ?m). 
 72 
 
 
2. Mutagenesis, expression, and purification of Endo-S and mutants 
The pGEX plasmid vector containing the coding sequence for the wild-type 
Endo-S was a gift from M. Collin (Lund University, Lund, Sweden). The Endo-S 
mutants were generated by PCR reactions using the Q5 site-directed mutagenesis kit 
(NEB) by the manufacturer?s instructions. For systematic mutagenesis at the Asp-233 
residue of Endo-S, two degenerate primers were used: 5? - CGTAAATTCGTGCTC 
AATNNNAATATCTAGTCCATCGACACCACGATCAGTT-3? (forward) and 5?- 
AACTGATCGTGGTGTCGATGGACTAGATATTNNNATTGAGCACGAATTTA
CG-3? (reverse). The corresponding mutations were verified by sequencing the 
coding sequence using T7 promoter and T7 terminator primers (Macrogen Inc., MD).  
Each of the wild-type and mutant Endo-S protein was expressed using a 
widely used T7 expression Escherichia coli strain BL21/DE3 competent cell (NEB). 
The transformed competent cells were grown in 20 mL of 2YT pre-medium with 100 
?g/mL carbenicillin antibiotic added. The pre-culture was incubated at 37 ?C and 
induced with 0.5 mM isopropyl ?-d-1-thiogalactopyranoside (IPTG) when the OD600 
reached 0.8?1.0. The induced cells were then incubated at 20 ?C overnight and 
harvested by high-speed centrifugation (8000 rpm for 30 min). The cell were lysed 
using the bacterial cell lysis buffer (Gold Biotechnology, Inc.) following the 
manufacturer?s instructions. After further centrifugation (2000 rpm for 30 min) the 
glutathione S-transferase (GST)-tagged proteins were purified using the GST affinity 
column GSTrap HP (GE Healthcare). The purified Endo-S proteins were diluted in 
phosphate-buffered saline (PBS) at pH 7.4 after buffer exchange using Amicon 
ultrafiltration units (10 kDa, Millipore). The purity of Endo-S proteins was shown to 
 73 
 
 
be >90% by sodium dodecyl sulfate?polyacrylamide gel electrophoresis (SDS-
PAGE), and the final concentration of each GST-tagged Endo-S variant was 
measured using a NanoDrop 2000c instrument at an absorbance of 280 nm using 
molar extinction coefficient of 12200 M-1cm-1. 
3. Synthesis of deglycosylated antibody by wild-type Endo-S as acceptor substrate 
for transglycosylation  
To generate the deglycosylated rituximab as the glycosyl acceptor substrate 
for the following transglycosylation assays, the commercial rituximab was treated 
with the wild-type Endo-S at a substrate/enzyme ratio of 1000/1 (w/w) for 30 min at 
37 ?C to release its native heterogeneous N-glycans. The complete release of the N-
glycans attached to the glycosylation site at Fc region of rituximab was confirmed by 
the liquid chromatography?mass spectrometry (LC?MS) analysis. The resulting 
deglycosylated rituximab containing only the fucosylated innermost GlcNAc residue 
was purified using protein-A affinity chromatography following manufacturer?s 
instructions (GE healthcare). The deconvoluted mass data (m/z) of the purified 
rituximab corresponds correctly with the calculated mass of (Fuc?1,6)GlcNAc-
rituximab molecule.	
 ? LC-MS results: calculated for the heavy chain of 
(Fuc?1,6)GlcNAc-rituximab, M = 49,420 Da; found (m/z), 49,412 (deconvoluted 
number).  
 74 
 
 
4. Synthesis of the S2G2F rituximab by mutant Endo-S D233Q as starting material 
for hydrolysis assays  
Because native commercial rituximab displays heterogeneous compositions of 
various N-glycans attached at the conserved N-glycosylation site at Asn-297, a 
homogeneous antibody N-glycoform is required both for the hydrolysis assay and as 
the internal standard used for the quantification of the reaction product. A reaction 
mixture of 69 ?M (Fuc?1,6)GlcNAc-rituximab and 1.38 mM SCT-oxa (20 eq.) was 
incubated with Endo-S2 D233Q (0.05 mg/mL) at 30 ?C in Tris buffer (0.1 M, pH 7.4) 
for 4 h. The LC-MS analysis indicated completion of the glycosylation reaction after 
2 h. The product glycoform S2G2F-rituximab was then purified by protein-A affinity 
chromatography. LC-MS results: calculated for the heavy chain of S2G2F-rituximab, 
M = 51,421 Da; found (m/z), 51,412 (deconvolution data).  
5. Liquid Chromatography-Mass Spectrometry (LC-ESI-MS) of antibodies  
The LC?MS characterization of antibody N-glycoforms was conducted on an 
Exactive Plus Orbitrap instrument (Thermo Scientific). The reduced antibody sample 
was analyzed with a Waters XBridgeTM BEH300 C4 column (3.5 ?m, 2.1 ? 50 mm). 
The program includes a 9 min linear gradient from 5 to 90% MeCN containing 0.1% 
formic acid at a flow rate of 0.4 mL/min. The original mass data of the heavy chain 
and light chain of antibody were then processed for deconvolution and integration 
using MagTran Software (Amgen). 
 75 
 
 
6. Quantification of hydrolysis and transglycosylation reaction products using 
internal standard and a single-point normalization factor 
To monitor hydrolysis and transglycosylation reactions in a more quantitative 
fashion, either of the following two glycan-modified antibody derivatives was used as 
the internal standards for the LC?MS analysis.134 One standard was the completely 
deglycosylated rituximab antibody that was generated by treating the antibody with 
the peptide-N-glycosidase F (PNGase F), which could remove all Fc N-glycans from 
the N-glycoproteins and simultaneously introduced an Asn-to-Asp mutation at the 
conserved N-glycosylation site; the other standard was the fully sialylated glycoform, 
S2G2F-rituximab, which was synthesized according to the previously described 
procedure.103 Serial dilutions containing a gradient of S2G2F-rituximab (0, 0.25, 0.5, 
0.75, 1.0, 1.25, and 1.5 ?M) were mixed with a stock solution containing 0.67 ?M 
internal standard. The relative intensity of the reaction product and the standard peaks 
were then analyzed by the selected ion monitoring (SIM) to calculate a single-point 
normalization factor. 135 For each glycan remodeling reaction, the amount of the 
reaction product could be deduced from its relative abundance to the internal standard 
following the normalized mathematical algorithm.136 
7. Transglycosylation and hydrolysis activity assays for Endo-S mutants 
The specific hydrolytic activity of each Endo-S mutant and wild-type (0.01 
mg/mL) was analyzed in Tris-HCl buffer (0.1 M, pH 7.4) at 30 ?C with the synthetic 
biantennary sialo-complex-type rituximab (1.0 mg/mL, 6.9 ?M) as starting material. 
Each reaction was terminated at 5 min, and an aliquot was taken from the reaction 
mixture and dissolved in 0.1% formic acid to quench the reaction. The relative 
 76 
 
 
quantity of the hydrolysis product was then calculated through deconvolution and 
integration of the selected MS peaks using the internal standard as a reference. The 
experimental details of product quantification using the LC?MS are described above. 
The specific transglycosylation activity of Endo-S mutants was measured in a similar 
fashion. The deglycosylated rituximab (10.0 mg/mL, 69 ?M) was incubated with 
SCT-Oxa (1.38 mM, 20 eq.) while being catalyzed by Endo-S enzymes (0.01 mg/mL) 
under the same condition that was used for the hydrolysis assay. The product S2G2F-
rituximab was analyzed by LC-MS as mentioned above. Both enzyme activity assays 
were repeated twice for each mutant and the wild-type Endo-S, and the average 
activity is recorded in the Results section. For the transglycosylation assay, at least 
three time points (t = 5, 10, and 15 min) were sampled to ensure that the specific 
activity was measured as the initial rate of the transfer reaction without any potential 
product hydrolysis. 
8. Measurements of KM and Vmax of Endo-S mutants for SCT-oxa 
To measure the kinetic parameters of Endo-S D233M and D233A, serial 
dilutions with a total volume of 20 ?L reaction were prepared with a gradient of the 
SCT-Oxa concentrations and a constant deglycosylated antibody concentration mixed 
in PBS buffer at pH 7.4. Each reaction was initiated by adding 1 ?L of Endo-S mutant 
to give the final concentration of 0.1 mg/mL, and the reaction was conducted at 30 ?C. 
The concentrations of SCT-Oxa were within the range of 6.25?400 ?M, whereas the 
concentration of the deglycosylated antibody was fixed at 20 ?M. For each reaction, 
three aliquots were taken at 1, 2, and 3 min of the same reaction, and the sampled 
reaction mixtures were immediately quenched in 100 ?L of 0.1% formic acid and 
 77 
 
 
frozen at -20 ?C. The quenched reaction samples were then characterized by ESI-LC-
MS, and the concentration of the product was calculated using the normalization 
factor and the internal standard, as shown above. A linear progression curve was 
confirmed for each reaction by sampling at multiple early time points and plotting 
them in the same graph, which emphasized that the V0 measured indeed represented 
the initial rate of each reaction. The KM and Vmax values were deduced by fitting the 
initial velocities against the SCT-Oxa concentrations using the software GraphPad 
Prism7. The experiments were conducted in triplicate to determine the errors in each 
measurement. A control reaction without the enzyme added was set up for each of the 
transglycosylation reactions to confirm that non-enzymatic transfer of SCT-Oxa to 
the antibody did not occur, and the results obtained here indeed represent the specific 
transglycosylation activity of each Endo-S mutant.  
9. Measurements of KM and Vmax of Endo-S mutants for antibody 
The KM and Vmax values of Endo-S mutant D233M and D233A for the 
deglycosylated antibody substrate were measured in a fashion similar to that used for 
the measurements of SCT-Oxa. In the serial reactions, the deglycosylated antibody 
concentrations ranged from 0.313 to 20 ?M, whereas the SCT-Oxa concentration was 
fixed at 400 ?M. A similar control reaction was also performed for each antibody 
concentration to monitor the non-enzymatic side reactions. 
10. Chemical synthesis of complex glycan-based thiazoline inhibitor  
The SGP glycopeptide starting material was extracted and purified from hen?s 
egg yolk according to the previously published method.127 The crude glycopeptide 
 78 
 
 
was enriched using the preparative HPLC to afford the pure SGP with a yield of 146 
mg from 90 eggs. ESI-MS results: calculated mass for SGP, M = 2865.8. Found mass: 
2866.7 [M+H]+. The sialylated N-glycan (50 mg, 50 mg/mL) was dissolved in buffer 
(pH 5.5, 5 mM CaCl2, 50 mM NaOAc). The sialic acid was removed by incubating 
the glycan with ?2?3,6,8-neuraminidase (80 U) at 37 ?C overnight. Complete release 
of sialic acid was confirmed by LC-MS, and the intermediate was purified by 
Sephadex G-15 size exclusion chromatography. The asialylated glycan was then 
subject to peracetylation reaction to protect the hydroxyl functional groups of the 
compound. A solution of the glycan (50 mg, 0.035 mmol) in pyridine (5 mL) and 
acetic anhydride (5 mL) was stirred at 40 ?C for 12 h until the TLC indicated that 
acetylation was complete. Finally, the peracetylated glycan was converted to 
thiazoline compound using the Lawesson?s reagent following previously reported 
method121, and the structure of the final product was confirmed by 1H NMR (CDCl3, 
400 MHz).  
11. Inhibition assays of thiazoline glycan inhibitor against ENGases 
The inhibitory activity of the thiazoline glycan compound against Endo-S was 
measured by testing the hydrolytic activity of Endo-S in the presence or absence of 
serial dilutions of the thiazoline inhibitor. A solution of synthetic S2G2F-rituximab 
(10 ?M) in PBS (100 ?L, 100 mM, pH 7.4) was incubated with the respective enzyme 
(5 ?M) in the presence of serial dilutions (0.02 to 50 ?M) of the N-glycan thiazoline. 
Reactions without any inhibitors added were carried out as a control. The mixture was 
incubated at 30 ?C for 10 min and the reaction was quenched by 0.1% TFA solution. 
The enzyme-released free N-glycans were quantitated by HPAEC-PAD, which was 
 79 
 
 
optimized for the analysis of oligosaccharides and related compounds. The inhibition 
assay against Endo-M was performed using the SGP glycopeptide as the substrate 
following the same procedure as that with Endo-S.  
12. Comparison of non-enzymatic glycosylation reactions by enzyme treatment.  
A standard transglycosylation reaction was carried out by incubating the 
deglycosylated (Fuc?1,6)GlcNAc-rituximab (69 ?M) with 1.38 mM SCT-Oxa in the 
presence of 0.05 mg/mL Endo-S D233M mutant. The reaction mixture was incubated 
at 30 ?C in Tris buffer (0.1 M, pH 7.4) for 60 min. The initial product was analyzed 
by ESI-LC-MS to confirm the corresponding mass for the S2G2F-rituximab and then 
purified from the reaction mixture by affinity chromatography. The fully glycosylated 
antibody was then reduced to the heavy chain and light chain fragments by incubating 
the antibody with 50 mM TCEP reducing agent overnight to break disulfide linkage. 
The light chain and heavy chain were then characterized independently by ESI-LC-
MS, as the two fragments could be separated on the C4 column (Waters XBridgeTM 
BEH300, 3.5 ?m, 2.1 ? 50 mm) prior to the mass analysis. To verify the glycosidic 
linkage, antibody fragments were diluted into 100 ?L of 1x Glycoprotein Denaturing 
Buffer (NEB) incubated with PNGase-F (20 mU, NEB) at 37?C overnight. Finally, 
the PNGase-F-treated antibody fragments were characterized by ESI-LC-MS as 
described above.  
To study the non-enzymatic chemical ligation between the glycan oxazoline 
and antibody, the same amount of deglycosylated rituximab (69 ?M) was incubated 
with a serial dilutions of SCT-Oxa (1.38 to 11.0 mM) at 30 ?C in Tris buffer (0.1 M, 
pH 7.4) for up to 12 h. Each reaction was sampled at 30, 60, 120, 240, 720 min. The 
 80 
 
 
same processing and analyzing procedures for the reacted antibody fragments were 
used to analyze non-enzymatic reactions as those for the standard transglycosylation 
reactions, which were mentioned in the previous experimental sections. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 81 
 
 
F. Supporting Information 
 
MDKHLLVKRTLGCVCAATLMGAALATHHDSLNTVKAEEKTVQVQKGLPSD
SLHYLSENSKKEFKEELSKAGQESQKVKEILAKAQQADKQAQELAKMKIPEK
IPMKPLHGPLYGGYFRTWHDKTSDPTEKDKVNSMGELPKEVDLAFIFHDWT
KDYSLFWKELATKHVPKLNKQGTRVIRTIPWRFLAGGDNSGIAEDTSKYPNT
PEGNKALAKAIVDEYVYKYNLDGLDVDVEHDSIPKVDKKEDTAGVERSIQVF
EEIGKLIGPKGVDKSRLFIMDSTYMADKNPLIERGAPYINLLLVQVYGSQGEK
GGWEPVSNRPEKTMEERWQGYSKYIRPEQYMIGFSFYEENAQEGNLWYDIN
SRKDEDKANGINTDITGTRAERYARWQPKTGGVKGGIFSYAIDRDGVAHQPK
KYAKQKEFKDATDNIFHSDYSVSKALKTVMLKDKSYDLIDEKDFPDKALRE
AVMAQVGTRKGDLERFNGTLRLDNPAIQSLEGLNKFKKLAQLDLIGLSRITK
LDRSVLPANMKPGKDTLETVLETYKKDNKEEPATIPPVSLKVSGLTGLKELD
LSGFDRETLAGLDAATLTSLEKVDISGNKLDLAPGTENRQIFDTMLSTISNHV
GSNEQTVKFDKQKPTGHYPDTYGKTSLRLPVANEKVDLQSQLLFGTVTNQG
TLINSEADYKAYQNHKIAGRSFVDSNYHYNNFKVSYENYTVKVTDSTLGTTT
DKTLATDKEETYKVDFFSPADKTKAVHTAKVIVGDEKTMMVNLAEGATVIG
GSADPVNARKVFDGQLGSETDNISLGWDSKQSIIFKLKEDGLIKHWRFFNDSA
RNPETTNKPIQEASLQIFNIKDYNLDNLLENPNKFDDEKYWITVDTYSAQGER
ATAFSNTLNNITSKYWRVVFDTKGDRYSSPVVPELQILGYPLPNADTIMKTVT
TAKELSQQKDKFSQKMLDELKIKEMALETSLNSKIFDVTAINANAGVLKDCIE
KRQLLK 
 
Supplementary Figure 1. Sequence of Endo-S (Genbank entry: AF296340).  
 
 82 
 
 
Chapter 3: Development of alternative antibody glycoengineering 
methods with relaxed substrate specificity for different glycan structures 
A. Introduction  
Although the glycosynthase mutants of Endo-S have been applied by many 
groups to synthesize homogeneous N-glycoforms of antibody for different research 
purposes,108, 111, 137 the currently known mutants of Endo-S, including D233M and 
D233Q, can only utilize the biantennary complex type or truncated Man3GlcNAc core 
glycan structures for antibody glycan remodeling. Other N-glycan structures such as 
the high-mannose or hybrid N-linked oligosaccharides were generally not susceptible 
to Endo-S-mediated transglycosylation. In terms of the narrow substrate specificity of 
the current glycoengineering strategy, a major breakthrough was made in 2016 by 
Wang and colleagues, who developed a new class of glycosynthase mutants for 
another GH18 family endoglycosidase F3 (Endo-F3). The glycosynthase mutants of 
Endo-F3, including Endo-F3 D165A, demonstrated the ability to transfer triantennary 
glycans to the deglycosylated antibody acceptor.85 However, Endo-F3 mutants only 
act on the core fucosylated glycosyl acceptors for transglycosylation and did not 
exhibit any transfer activities towards non-fucosylated substrates. These studies 
collectively suggest that the glycosynthase mutants currently used in antibody 
glycoengineering still have major limitations in terms of their substrate specificity 
and transglycosylation efficiency.89 
To further improve the current glycan remodeling strategy for monoclonal 
antibodies, Endo-S2, another GH18 family endoglycosidase also harvested from the 
 83 
 
 
pathogenic bacterium Streptococcus pyogenes of serotype M49 was investigated for 
its potential glycan-remodeling activities. Sequence alignment analysis revealed 
approximately 37% amino acid sequence identity between Endo-S and Endo-S2. 
Comparative hydrolytic assays on serum antibodies by the two enzymes suggested 
that Endo-S2 possessed a more relaxed substrate selectivity for the N-glycans 
attached at the Fc region of antibodies, although it remained a mystery whether Endo-
S2 was capable of transferring N-glycans to any glycosyl acceptors for 
transglycosylation reactions.138 By analogy to previous studies conducted on Endo-S, 
potential glycosynthase mutants of Endo-S2 could also be generated through site-
directed mutagenesis and studied for their transglycosylation activities based on a 
possible substrate-assisted mechanism. In particular, the functional variants of Endo-
S2 could potentially act on additional N-glycan oxazoline structures for 
transglycosylation, which might expand the scope of the current ENGase-mediated 
glycosylation method. Therefore, systematic mutational study was performed on 
Endo-S2 at the catalytic amino acid residue Asp-184, which was selected based on 
sequence alignment between Endo-S/S2 and other ENGases. The hydrolytic and 
transglycosylation activities of the generated Endo-S2 mutants were tested and 
compared in order to screen for more efficient glycosynthase mutants for antibody 
glycoengineering.139 The generation and comparison of Endo-S2 mutants was 
conducted using a similar experimental method to those on Endo-S, as described in 
Chapter 2. The results indicated that Endo-S2 did show both potent hydrolytic and 
transglycosylation activities, and saturation mutagenesis study successfully identified 
multiple glycosynthase mutants of Endo-S2, including Endo-S2 D184M and D184Q, 
 84 
 
 
which exhibited significant oxazoline-transferring activities with much diminished 
hydrolytic tendencies against the product antibody glycoforms. More importantly, the 
glycosynthase variants of Endo-S2 showed flexible substrate specificity using three 
different types of N-glycan oxazoline structures (complex, high-mannose, and hybrid) 
for transglycosylation onto a deglycosylated antibody. Comparative functional assay 
between Endo-S and Endo-S2 mutants revealed that the Endo-S2 D184M was more 
efficient than the previous Endo-S D233Q in catalyzing transglycosylation reactions. 
In addition, we demonstrated that a significantly reduced amount of glycan oxazoline 
(only 2.5 molar equivalents for each Fc monomer glycosylation site) was required to 
achieve complete glycosylation of the deglycosylated IgG-Fc fragment, which is 
significantly more efficient than the Endo-S D233M (10 molar equivalents each) in 
terms of substrate consumption and reaction completion time. Finally, similar to the 
mechanistic analyses on Endo-S mutants, the kinetic parameters of Endo-S2 D184M 
and D184A were measured and compared with those of Endo-S. Comparison between 
the kinetics of Endo-S2 D184M and D184A demonstrated similar patterns to Endo-S 
mutants. Interestingly, the combined results collectively suggested that the enhanced 
catalytic efficiencies of Endo-S2 D184M over Endo-S D233M for both glycan and 
antibody substrates were due to increased catalytic turnover numbers (kcat), whereas 
the substrate affinities (KM) did not change significantly.123 
Despite successful applications of Endo-S2 and mutants in glycoengineering, 
one drawback of this general method still persisted that the wild-type enzyme (either 
Endo-S or Endo-S2) was required to release the native heterogeneous N-glycans from 
the starting antibody prior to the transfer of new N-glycan structures. Therefore, the 
 85 
 
 
wild-type Endo-S/S2 must to be removed from the reaction mixture after the 
deglycosylation step in order to prevent the wild-type enzymes from hydrolyzing the 
final product, and an additional step of antibody purification was needed after the 
transglycosylation step. The necessary switching between the wild type and mutant 
enzymes in addition to the tedious antibody purification procedure would present a 
major issue for the cost-effectiveness of this method in large-scale industrial 
applications.109 To circumvent this problem, we re-evaluated the substrate specificity 
of Endo-S with a broader range of substrates including different N-glycan oxazolines 
and antibody N-glycoforms. Interestingly, our results clearly demonstrated that the 
wild-type Endo-S displayed promiscuous activities against multiple glycan structures 
for transglycosylation, which included hybrid and high-mannose glycan oxazolines. 
However, the product hybrid and high-mannose N-glycoforms of antibody are highly 
resistant to Endo-S-catalyzed hydrolysis. These selective catalytic properties of Endo-
S, combined with its highly efficient hydrolytic activity against the native N-glycans 
of conventional antibodies, allowed us to develop a one-pot glycoengineering method 
for IgG antibodies to synthesize additional target antibody N-glycoforms. Using this 
one-pot strategy, removal of the wild-type enzymes was no longer necessary since 
deglycosylation and transglycosylation of the antibody were catalyzed simultaneously 
by Endo-S. Finally, we applied this single-step method to the remodeling of a 
therapeutic antibody rituximab with a chemically modified N-glycan. As mentioned 
in Introduction section, an azido-functionalized N-glycoform of therapeutic antibody 
could be used as molecular handle for the bioorthogonal ligation chemistry to develop 
new classes of antibody-drug conjugates.140 
 86 
 
 
B. Results 
1. Cloning, expression, and characterization of Endo-S2 
The DNA sequence coding Endo-S2 (38?819) was cloned into a pET22b-CPD 
vector, which contains a C-terminal cysteine protease domain (CPD) derived from the 
Vibrio cholerae MARTX toxin with a 10-histidine tag.141 We recently reported that a 
high-level, soluble expression of Endo-F3 and its mutants could be achieved using 
this vector85. Following a similar strategy, Endo-S2 was successfully expressed in 
BL21 (DE3) Escherichia coli competent cells and was purified using immobilized 
metal ion affinity chromatography to obtain the soluble enzyme with a yield of more 
than 20 mg/liter. The recombinant Endo-S2 showed significant hydrolysis activity 
against the intact IgG antibody as demonstrated by its rapid deglycosylation of 
commercial rituximab, which was monitored by LC-MS analysis. 
2. Generation of glycosynthase mutants of Endo-S2 
We have previously generated glycosynthase mutants from endoglycosidases 
of both GH85 and GH18 families by site-directed mutation at a key residue that is 
responsible for promoting the formation of the oxazolinium ion intermediate during 
hydrolysis, which proceeds in a substrate-assisted mechanism. These include a key 
asparagine residue for endoglycosidases Endo-A (Asn-171), Endo-M (Asn-175), and 
Endo-D (Asn-322) of the GH85 family, or a key aspartic acid residue for the GH18 
family endoglycosidases Endo-S (Asp-233) and Endo-F3 (Asp-165). Sequence 
alignment of Endo-S2 and Endo-S revealed that the Asp-184 of Endo-S2 was the 
residue equivalent to the Asp-233 of Endo-S essential for promoting oxazolinium ion 
 87 
 
 
formation in hydrolysis (Figure. 26). To generate glycosynthase mutants from Endo-
S2, we systematically replaced the Asp-184 with other 19 natural amino acid residues 
using site-directed mutagenesis. The resulting 19 Asp-184 mutants of Endo-S2 were 
expressed as soluble proteins in pET22b-CPD vector and purified using immobilized 
metal ion affinity chromatography, in the same way as demonstrated for the wild-type 
Endo-S2. The expression of mutant Endo-S2 enzymes gave comparable yields (15?20 
mg/liters) as the expression of the wild type enzyme. 
 
Endo S2     WHDRASTGIDGKQQHPENTMAEVPKEVDILFVFHDHTASDSPFWSELKDSYVHKLHQQGT 133 
Endo S      WHDKTSDPT---EKDKVNSMGELPKEVDLAFIFHDWTKDYSLFWKELATKHVPKLNKQGT 177 
            ***::*      ::.  *:*.*:*****: *:*** * . * **.**  .:* **.:*** 
 
Endo S2     ALVQTIGVNELNGR-----TGLSKDYPDTPEGNKALAAAIVKAFVTDRGVDGLDIDIEHE 188 
Endo S      RVIRTIPWRFLAGGDNSGIAEDTSKYPNTPEGNKALAKAIVDEYVYKYNLDGLDVDVEHD 237 
             :::**  . * *      :  :..**:********* ***. :* .  :****:*:**: 
 
Endo S2     FTNKRTPEE----DARALNVFKEIAQLIGKNGSDKSKLLIMDTTLSVENNPIFKGIAEDL 244 
Endo S      SIPKVDKKEDTAGVERSIQVFEEIGKLIGPKGVDKSRLFIMDSTYMADKNPLIERGAPYI 297 
               *   :*      *:::**:**.:*** :* ***:*:***:*  .::**:::  *  : 
 
Figure 26. Sequence alignment of Endo-S2 and Endo-S. The aspartic acid residue 
(Asp 233 of Endo-S and Asp-184 of Endo-S2) critical for promoting the oxazolinium 
ion formation in hydrolysis and the catalytic general acid/base residue (Glu-235 of 
Endo-S and Glu-186 of Endo-S2) are highlighted. 
3. Generation and comparative studies on the hydrolytic and transglycosylation 
activities of the 19 glycosynthase mutants and wild-type enzyme of Endo-S2 
The hydrolysis activities on the IgG-Fc N-glycans in an intact antibody and 
the transglycosylation activities with glycan oxazoline were assessed following the 
scheme shown in Figure 27. The results indicate that most of the mutants at Asp-184 
residue led to significantly reduced or completely diminished hydrolysis activities on 
the IgG-Fc N-glycan. Among them, D184F, D184H, D184K, D184R, and D184W 
 88 
 
 
mutants were completely devoid of hydrolytic activity, whereas several other mutants, 
including D184C, D184E, D184G, D184N, D184S, D184Y, were found to still retain 
significant hydrolysis activity (Table 4). On the other hand, the evaluation of 
transglycosylation indicated that almost all the mutants possessed transglycosylation 
activity when using the deglycosylated rituximab as the acceptor and a biantennary 
complex type glycan oxazoline as the donor substrate, but the activities varied 
significantly among different mutants (Table 5). Among others, D184C, D184M, 
D184G, D184E, D184Y, D184S, and D184A were found to be the most active 
mutants. However, D184C, D184G, D184E, D184Y, D184S, and D184A also 
demonstrated significant residual hydrolytic activity. The most interesting mutant is 
Endo-S2 D184M, which retained only marginal hydrolytic activity but showed high 
transglycosylation activity (only second to D184C), making the best glycosynthase 
mutants to choose for the glycosylation remodeling of antibody. 
 
 
Figure 27. Schematic presentation of hydrolysis and transglycosylation by Endo-S2 
and its mutants using rituximab as the substrate. The fucosylated GlcNAc-rituximab, 
the deglycosylated rituximab carrying the core-fucosylated GlcNAc moiety at the 
glycosylation site; SCT-rituximab (S2G2), the sialylated complex type glycoform of 
rituximab. 
 
 89 
 
 
Table 4. Comparison of the specific hydrolytic activities of Endo-S2 D184 mutants 
using the synthetic complex type rituximab (S2G2-rituximab) as the starting material.a 
Specific hydrolysis Percentage of Mutant activity specific hydrolysis activity 
 *100 (umole/min/mg) % WT 11.185 100.000 
D184A 0.750 6.701 
D184C 1.386 12.392 
D184E 0.879 7.855 
D184F 0.000 0.000 
D184G 1.249 11.168 
D184H 0.000 0.000 
D184K 0.000 0.000 
D184R 0.000 0.000 
D184I 0.033 0.295 
D184L 0.045 0.405 
D184M 0.158 1.410 
D184N 0.842 7.526 
D184P 0.340 3.043 
D184Q 0.049 0.441 
D184S 1.387 12.398 
D184T 0.212 1.899 
D184V 0.409 3.659 
D184W 0.000 0.000 
D184Y 5.526 49.406 
 
aThe specific hydrolysis activity of each enzyme was assayed by incubating the 
synthetic biantennary complex type rituximab antibody with each of the mutants or 
the wild-type Endo-S2 (0.01 mg/mL) in PBS pH 7.4 at 30  ?C.  
 
 
 
 90 
 
 
Table 5. Comparison of specific transglycosylation activity of Endo-S2 D184 mutants 
using glycan oxazoline as donor and deglycosylated rituximab as acceptor.a 
Specific Percentage of 
Mutants transglycosylation specific transglycosylation 
activity activity 
 *10 (umol/min/mg) % WT 1.712 44.31 
D184A 1.023 26.46 
D184C 3.864 100.00 
D184E 2.006 51.907 
D184F 0.044 1.14 
D184G 2.270 58.73 
D184H 0.026 0.68 
D184K 0.021 0.54 
D184R 0.022 0.57 
D184I 0.066 1.71 
D184L 0.099 2.55 
D184M 2.866 74.18 
D184N 0.879 22.75 
D184P 0.160 4.13 
D184Q 0.383 9.92 
D184S 1.164 30.13 
D184T 0.828 21.42 
D184V 0.147 3.79 
D184W 0.046 1.20 
D184Y 1.198 30.99 
aTo test the specific transglycosylation activity of each enzyme, the deglycosylated 
rituximab antibody was used as the glycosyl acceptor and the SCT oxazoline as the 
glycosyl donor substrate with a final concentration of enzymes at 0.01 mg/mL. The 
molar ratio of the glycosyl acceptor to donor substrate was 1:20. The experimental 
results obtained here were representative from three independently conducted 
experiments. 
 91 
 
 
4. Enzyme kinetic parameters of Endo-S2 D184M and Endo-S2 D184A  
To evaluate the nature of the enhanced catalytic efficiency of Endo-S2 mutant, 
we measured the kinetic parameters of the D184M and D184A mutants of Endo-S2. 
For the sugar oxazoline, the Endo-S2 methionine mutation significantly enhanced its 
catalytic efficiency (kcat/KM) in comparison with that seen for the alanine mutation, 
and an approximately 10-fold increase in kcat/KM was observed for the D184M mutant 
over the D184A mutant of Endo-S2 (Table 6 and Figure 28a). In analogy to Endo-S, 
the increased kcat/KM value for D184M mutant was caused by a >15-fold increase in 
the catalytic turnover number (kcat) and a <2-fold decrease in the substrate affinity 
(KM) for the glycan oxazoline substrate. 
For the deglycosylated antibody, like the Endo-S D233M mutant, the Endo-S2 
D184M mutant also showed an affinity for the antibody substrate significantly higher 
than that of the D184A mutant. The KM of Endo-S2 D184A was estimated to be >20 
times higher than that of the Endo-S2 D184M mutant. Similar to the case of Endo-S, 
Michaelis?Menten curves indicated that the Endo-S2 D184A mutant showed no sign 
of saturation within the range of antibody concentrations tested. Therefore, the 
catalytic turnover number of Endo-S2 D184A also could not be accurately measured 
because of the difficulty of saturating the antibody substrate (Table 7 and Figure 28b).  
Interestingly, for both the sugar oxazoline and the antibody substrate, Endo-S2 
D184M demonstrated a catalytic turnover number significantly higher than that of 
Endo-S D233M, whereas substrate affinity for both substrates proved to be 
comparable between the two mutants. Therefore, the overall catalytic efficiency 
(kcat/KM) of Endo-S2 D184M was found to be higher than that of Endo-S D233M. 
 92 
 
 
Table 6. Kinetic parameters of Endo-S2 mutants D184M and D184A for SCT-Oxa. 
Enzyme	
 ? -??1 -??1 -??1k (min )	
 ? K (?M)	
 ? k /K (min 	
 ??M )	
 ?
cat	
 ? M cat M	
 ?	
 ?
Endo-??S2	
 ?D184M	
 ? 167.8?	
 ?20.5	
 ? 229.1	
 ??	
 ?34.8	
 ? 0.73	
 ?
Endo-??S2	
 ?D184A	
 ? 10.2	
 ??	
 ?2.4	
 ? 121.3	
 ??	
 ?17.1	
 ? 0.08	
 ?
Table 7. Kinetic parameters of Endo-S2 mutants D184M and D184A for rituximab. 
Enzyme	
 ? -??1 -??1 -??1k (min )	
 ? K (?M)	
 ?M k /K (min 	
 ??M )	
 ?cat	
 ? cat M	
 ?	
 ?
Endo-??S2	
 ?D184M	
 ? 29.8	
 ??	
 ?9.3	
 ? 32.6	
 ??	
 ?6.7	
 ? 0.91	
 ?
Endo-??S2	
 ?D184A	
 ? -??*	
 ? N.D.	
 ? -??*	
 ?
*The catalytic turnover number of Endo-S2 D184A was not determined because the 
concentration of antibody substrate required to saturate the enzyme could not be used 
in the experiments under normal conditions. 
a b
 
Figure 28. Michaelis-Menten curves of the Endo-S2 mutants (D184M and D184A): a) 
for glycosyl donor, a gradient of SCT-Oxa (6.25 to 400 ?M) was tested with a 
constant concentration of rituximab (20 ?M); b) for glycosyl acceptor, a serial 
dilution of antibody (0.32 to 20 ?M) was tested with 400 ?M of SCT. For both Endo-
S2 D184M and D184A, the enzyme concentration was controlled at 0.01 mg/mL 
(0.067 ?M). The data sets presented here are representative of three independent 
experiments. 
 93 
 
 
5. Endo-S2 D184M showed relaxed glycan substrate specificity 
Rituximab, a therapeutic monoclonal antibody, was selected as the model to 
examine the transglycosylation activity of the Endo-S2 Asp-184 mutants. The major 
N-glycans of the commercial rituximab are the core-fucosylated biantennary complex 
oligosaccharides containing 0?2 galactose residues, namely G0F, G1F, and G2F 
glycoforms, respectively. The glycosylation remodeling approach is shown in Fig. 29 
and the identity of antibody products were assessed by LC-MS analysis. Treatment of 
rituximab with wild-type Endo-S2 resulted in complete deglycosylation of the native 
rituximab. The Fuc-1,6GlcNAc-rituximab was then purified away from the wild-type 
endoglycosidase and the released glycans by Protein-A affinity chromatography and 
used as the acceptor in the following transglycosylation reactions. We demonstrated 
that Endo-S2 D184M was able to efficiently transfer the sialylated biantennary 
complex glycan to the acceptor antibody with 20 molar eq (i.e. 10 molar equivalents 
per monomeric Fc domain) of the glycan oxazoline. The reaction yield was estimated 
by LC-MS analysis to be over 95% as almost no starting material was detected, which 
confirmed the completion of the transglycosylation.  
In addition to the complex type N-glycan, the specificity of Endo-S2 was 
further tested with high-mannose type (HM) Man9GlcNAc oxazoline and the 
sialylated hybrid type (Hyb) Neu5AcGalGlcNAcMan5GlcNAc oxazoline as the donor 
substrate in the transglycosylation reactions. The reactions led to formation of 
corresponding antibody N-glycoforms. Under the same conditions, we found that the 
previously reported Endo-S mutants, including D233A and D233Q mutants of Endo-
S, showed only marginal transglycosylation activity with high-mannose and hybrid 
 94 
 
 
type N-glycans, although they could efficiently transfer biantennary complex type N-
glycan. In addition, the recently reported D165A mutant of Endo-F3 was unable to 
transfer the high-mannose or hybrid type N-glycans but could work on the bi- and 
triantennary complex type sugars. Thus, these Endo-S2-derived mutants represent the 
first glycosynthases that can efficiently transfer high-mannose and hybrid type N-
glycans to the core-fucosylated GlcNAc acceptor in an intact antibody. It should be 
mentioned that the Endo-A mutant (N171A and N171Q) could transfer high-mannose 
N-glycan to the GlcNAc-Fc domain, but they were unable to use core-fucosylated 
GlcNAc-Fc as acceptor. These studies also show that combined use of wild type 
Endo-S2 and Endo-S2 glycosynthase mutants provides an efficient glycosylation 
remodeling approach to homogeneous glycoforms of antibodies starting from a single 
precursor glycoform. 
 
 
 95 
 
 
 
Figure 29. Evaluation of substrate specificity of Endo-S2 D184M on different glycans 
(HM, CT, and Hyb type). GlcNAc-rituximab, the rituximab glycoform carry only first 
GlcNAc moiety at the glycosylation site; 1,6GlcNAc-rituximab, the deglycosylated 
rituximab carrying the core-fucosylated GlcNAc moiety at the glycosylation site; 
SCT-rituximab, the sialyl complex type glycoform of rituximab; HM-rituximab, the 
high-mannose type glycoform of rituximab; Hyb-rituximab, the hybrid type 
glycoform of rituximab. 
 
 
 
 
 96 
 
 
6. Transglycosylation of different N-glycan structures by Endo-S2 D184M  
To further characterize the N-glycan substrate preference of Endo-S2 D184M, 
parallel transglycosylation reactions were carried out with the N-glycan oxazolines of 
complex (SCT) type, high-mannose (HM) type, and hybrid (Hyb) type, respectively. 
The reaction progresses were monitored by LC-MS analysis of reaction aliquots taken 
at multiple time points and the results were summarized in Figure. 30. Under the 
same reaction conditions , the transglycosylation reaction with SCT-oxazoline (4) was 
completed within 20 min to give the full sialylated S2G2F-rituximab (3), whereas the 
transglycosylation with HM-oxazoline (7) and Hyb-oxazoline (5) were much slower. 
These results confirm that Endo-S2 D184M prefers the complex type over the high-
mannose and hybrid type N-glycans despite having a remarkably relaxed N-glycan 
specificity. 
The semipreparative synthesis of antibody N-glycoforms by Endo-S2 D184M 
was conducted in collaboration with Dr. Tiezheng Li from Prof. Lai-Xi?s group at the 
Department of Chemistry and Biochemistry, University of Maryland, College Park. 
This work was published in Journal of Biological Chemistry in 2016.139 The 
mutational studies and enzyme activity assays were performed by Xin Tong and Dr. 
Li and other authors. The large-scale glycosylation remodeling and purification of 
antibody glycoforms were achieved primarily by Dr. Li. 
 97 
 
 
 
Figure 30. Comparison of transglycosylation reactions by Endo-S2 D184M with 
different N-glycans. Transglycosylation reaction was carried out using deglycosylated 
rituximab 2 as the acceptor and different types of glycan oxazolines as the donor 
substrate under the catalysis of Endo-S2 D184M (0.05 mg/ml). The molar ratio of 
donor to acceptor was 20:1. The data presented here are representative of two 
independent experiments. 
 
7. Transglycosylation efficiencies of selected Endo-S2 and Endo-S mutants. 
We previously generated the first Endo-S glycosynthase mutants D233Q and 
D233A that transglycosylate rituximab efficiently with the complex type N-glycan 
oxazoline.103 The Endo-S Asp-233 mutants have been recently used by us and other 
groups to generate homogeneous monoclonal antibodies for structural and functional 
studies.20 To compare the transglycosylation efficiency of the glycosynthase mutants 
 98 
 
 
from both Endo-S2 and Endo-S, we selected to use the Endo-S D233Q mutant, the 
corresponding Endo-S2 D184Q mutant, and the Endo-S2 D184M mutant to catalyze 
three parallel transglycosylation reactions. The time course of each transglycosylation 
reaction was monitored by LC-MS analysis and summarized in Figure 31. Under the 
same reaction conditions, the D184M mutant of Endo-S2 showed remarkably potent 
transglycosylation activity, and reached completion of the glycan transfer within 10 
min. The other Endo-S2 mutant could also transfer the glycan smoothly and reached 
the completion within 1 h. However, the corresponding Endo-S mutant (D233Q) was 
much less efficient, reaching about 10% of transglycosylation at 1 h under the same 
conditions (Figure 31). As demonstrated in a separate experiment, more (10-fold) 
Endo-S D233Q mutant and a larger excess of the glycan oxazoline were required to 
achieve the same level of the transglycosylation catalyzed by the Endo-S2 D184Q 
mutant, and the D184M mutant of Endo-S2 was even much more efficient than the 
D184Q mutant. These studies suggest that some newly discovered Endo-S2 Asp-184 
mutants are superior to the previously reported Endo-S Asp-233 mutants for antibody 
glycosylation remodeling in both efficiency of reactions and the breadth of substrate 
diversity. 
 99 
 
 
 
Figure 31. Comparison of transglycosylation reactions catalyzed by selected mutants 
Endo-S2 D184Q, Endo-S2 D184M, and Endo-S D233Q. The transglycosylation was 
performed using deglycosylated rituximab 2 as the acceptor and SCT glycan 
oxazoline 4 as the donor substrate under the catalysis of different endoglycosidase 
mutants at a fixed concentration of 0.05 mg/ml. The molar ratio of donor to acceptor 
was 20:1. The data presented here are representative of two independent experiments. 
8. Antibody transglycosylation using a reduced amount of glycan oxazoline. 
To investigate whether it would be possible to achieve the same level of 
antibody glycosylation by Endo-S2 mutants with reduced oxazoline consumption, we 
performed the routine transglycosylation reactions by Endo-S2 D184M with only five 
molar equivalents of SCT-Oxa per antibody under conventional reaction conditions. 
As demonstrated in Figure 32, depending on the concentration of Endo-S2 D184M 
used, the glycosylation of antibody could be completed at different rates. It is worth 
 100 
 
 
mentioning that hydrolysis was observed when the enzyme concentration was high 
(0.25 mg/mL), emphasizing the importance of controlling the enzyme concentration 
of glycosynthase mutants used to minimize the product hydrolysis. Nevertheless, all 
three concentrations of Endo-S2 D184M reached complete transglycosylation within 
4 hours using only 5 equivalents of SCT-Oxa per antibody, which was significantly 
less than the previously reported amount. This discovery provides another advantage 
of using the End-S2 mutant such as Endo-S2 D184M for the glycan remodeling of 
antibody, particularly for the large scale manufacturing of antibody N-glycoforms. 
 
 
Figure 32. Transglycosylation of antibody with a reduced oxazoline substrate at 
different Endo-S D184M concentrations. Only five molar equivalents of SCT-Oxa 
substrate per antibody were used for the transglycosylation of antibody in the PBS 
buffer (pH 7.4) at 30 ?C. The transglycosylation yield (%) of each reaction was 
monitored by LC-MS and estimated using previously described methods.139 
 101 
 
 
9. Endo-S-catalyzed hydrolysis of antibody glycoforms and N-glycan oxazolines 
Previously, Collin and co-workers have reported that Endo-S and Endo-S2 
hydrolyze the Fc glycans of therapeutic antibodies with different selectivites.138 
While Endo-S2 can release all three major types of N-glycans from the Fc domain of 
several therapeutic antibodies including cetuximab, Endo-S is only efficient for 
releasing bi-antennary complex type N-glycans from the antibodies. To further 
confirm the substrate selectivity of the wild-type Endo-S, we performed side-by-side 
comparison of Endo-S-catalyzed hydrolysis reactions for the selected homogeneous 
glycoforms of rituximab, as well as various synthetic N-glycan oxazolines (Figure 33). 
We first tested three different homogeneous glycoforms of rituximab, namely, 
the sialylated complex type (1), the high-mannose (Man9GlcNAc2) type (2), and the 
hybrid type (3) (Figure 34). These homogeneous glycoforms (1-3) were synthesized 
following our recently published method, using the Endo-S2 glycosynthase mutant 
(EndoS2-D184M) as a key enzyme for transglycosylation.139 The enzymatic reactions 
were monitored by LC-ESI-MS analysis. We found that Endo-S could hydrolyze the 
biantennary complex form of rituximab (1) fast, resulting in over 90% hydrolysis 
within 2 h under the reaction condition. However, only marginal hydrolysis (less than 
5%) of the high-mannose and hybrid glycoforms (2 and 3) was observed under the 
same reaction condition (Figure 34A). A prolonged (overnight) incubation resulted in 
about 10% hydrolysis of the high mannose glycoform (2) and approximately 15% of 
the hybrid glycoform (3), respectively (data not shown). These data confirm that the 
high-mannose and hybrid type Fc glycoforms of antibodies are largely resistant to 
Endo-S-catalyzed hydrolysis although residual hydrolytic activity was detected, and 
 102 
 
 
the bi-antennary complex type glycoform is an excellent substrate for the wild-type 
Endo-S. 
 
Figure 33. Structures of different rituximab N-glycoforms and N-glycan oxazolines 
used for the assessment of Endo-S activity. 
 
Next, we compared the Endo-S WT-catalyzed hydrolysis of various synthetic 
N-glycan oxazolines. Synthetic glycan oxazolines, the mimics of the oxazolinium 
intermediate during the endoglycosidase-catalyzed reaction in a substrate-assisted 
mechanism, are the key donor substrates for the endoglycosidase-catalyzed 
transglycosylation used for glycoprotein remodeling. Thus, it is important to know 
whether the glycan oxazolines will be hydrolyzed and how fast they will be 
hydrolyzed by the enzyme of interest. Various glycan oxazolines were incubated with 
Endo-S WT in PBS buffer, and the hydrolysis products (free glycans) were monitored 
 103 
 
 
and quantitated by HPAEC-PAD (high performance anion exchange chromatography 
coupled with pulsed amperometric electrochemical detection). As illustrated in Figure 
33B, the complex type glycan oxazoline (4) was rapidly hydrolyzed by Endo-S WT. 
In contrast, the high-mannose type N-glycans oxazolines, including Man9 oxazoline 
(M9-Oxa 5) and the Man5 glycan oxazoline (Man5-Oxa 7), were shown to be rather 
resistant to Endo-S hydrolysis. The hybrid oxazoline (Hyb-Oxa 6) showed moderate 
hydrolysis by Endo-S. Interestingly, the Man3 oxazoline (M3-Oxa 8), corresponding 
to the shared N-glycan core, was also hydrolyzed very rapidly by Endo-S. However, 
chemical modifications at the two outer mannose moieties led to reduction of the 
enzymatic hydrolysis, as demonstrated by azide-Man3 glycan oxazoline (AzideM3-
Oxa 9) (Figure 34B). These results clearly indicate that wild-type Endo-S possesses 
distinct enzymatic activities towards different glycan oxazolines, as well as different 
glycoforms of antibodies. 
 
 
 104 
 
 
 
Figure 34. Hydrolytic activity of Endo-S WT on different rituximab N-glycoforms 
and oxazolines; A) Comparison of the hydrolytic activity of Endo-S WT towards the 
complex (1), hybrid (2) and high-mannose (3) type synthetic N-glycoforms of 
rituximab; B) Comparison of the hydrolytic activity of Endo-S WT towards various 
N-glycan oxazolines to the free N-glycans. The results shown here are representative 
of three independently conducted experiments. 
 105 
 
 
10. Endo-S-catalyzed transglycosylation of antibody with additional oxazolines 
The unique substrate specificity of Endo-S towards the different antibody N-
glycoforms and glycan oxazolines prompted us to examine the wild-type Endo-S-
catalyzed transglycosylation with the three major Fc N-glycans oxazolines, complex 
(SCT-Oxa 4), Man9 (HM-Oxa 5), and Hybrid (Hyb--Oxa 9), to the deglycosylated 
rituximab (GlcNAc-rituximab, 10) (Figure 35). The Fuc1,6GlcNAc-rituximab (10) 
was prepared by deglycosylation of rituximab with Endo-S, following our previously 
published procedure.103 We found that the wild-type Endo-S showed a significant 
transglycosylation activity with the complex type N-glycan oxazoline (4) to form the 
transglycosylation product (1) (Figure 35A). However, the product was quickly 
hydrolyzed by the enzyme, as the resultant complex type N-glycoform turned out to 
be an excellent substrate for Endo-S-catalyzed hydrolysis. This result was consistent 
with our previous observations.114 We also observed that, under the same condition, 
wild-type Endo-S was also able to transfer the high-mannose glycan oxazoline (5) 
and hybrid glycan oxazoline (6), but at a much lower reaction rate than that with the 
complex type glycan oxazoline (4). Nevertheless, since the transglycosylation 
products (2 and 3) were largely resistant to Endo-S catalyzed hydrolysis, the 
transglycosylation products could thus be accumulated. Previously we have generated 
novel Endo-S mutants, including EndoS-D233Q, which are glycosynthases that can 
use glycan oxazolines for transglycosylation but lack the activity to hydrolyze the 
product.103 The EndoS-D233Q mutant was efficient for transferring complex type N-
glycan oxazolines to the deglycosylated antibody; however, it was not clear how the 
glycosynthase mutant would act on the high-mannose and hybrid type N-glycan 
 106 
 
 
oxazolines for transglycosylation. Here we tested EndoS-D233Q with oxazolines 5 
and 6 for glycoengineering. Surprisingly, we found that the Endo-S mutant EndoS-
D233Q was unable to act on either the high-mannose glycan oxazoline (5) or the 
hybrid glycan oxazoline (6) for transglycosylation with the Fuc1,6GlcNAc-rituximab 
(10), while under same condition the wild-type Endo-S could achieve over 85% 
transglycosylation to form the corresponding antibody glycoforms (Figure 35B and 
Figure 35C). In comparison with the wild-type enzyme, the D233Q mutation reduced 
both the hydrolysis and transglycosylation activities of the enzyme, thus, resulting in 
the decreased efficiency toward the less favorable glycan oxazoline substrates, while 
it was still reactive enough towards the complex type N-glycan oxazoline (4), as 
shown in the previous publications.98 These new experimental data suggest that the 
wild-type Endo-S is complementary to the mutant Endo-S in terms of their substrate 
specificities, and the Endo-S WT will be particularly useful for the synthesis of high-
mannose and hybrid N-glycoforms of antibodies, which could not be achieved by the 
previously reported Endo-S mutants. 
 
 
 
 
 
 
 
 107 
 
 
 
 108 
 
 
Figure 35. Endo-S-catalyzed transglycosylation reactions with oxazolines SCT-Oxa 
4, HM-Oxa 5, Hyb-Oxa 6 of deglycosylated rituximab 10; A) Comparison of the 
transglycosylation activity of Endo-S WT towards oxazolines SCT-Oxa 4, HM-Oxa 
5, Hyb-Oxa 6; B) Comparison of the transglycosylation activity of Endo-S WT and 
Endo-S D233Q mutant towards the high mannose-type oxazoline (HM-Oxa 5); C) 
Comparison of transglycosylation activity of Endo-S WT and Endo-S D233Q mutant 
towards the hybrid-type oxazoline (Hyb-Oxa 6). The detailed reaction conditions are 
shown in Materials and Methods section. The results shown here are representative of 
three independently conducted experiments. 
11. One-pot glycosylation remodeling of rituximab via the wild-type Endo-S 
The potent transglycosylation activity of Endo-S WT with the hybrid and 
high-mannose type oxazolines, together with its efficient hydrolysis of the complex-
type N-glycans from the commercial antibody rituximab, prompted us to examine the 
possibility of remodeling antibody glycosylation in a one-pot fashion (Figure 36). 
Thus, a mixture of the commercial rituximab, glycan oxazoline (5 or 6), and the wild-
type Endo-S was incubated at room temperature. The reaction was monitored by LC-
ESI-MS analysis. As expected, the Endo-S-catalyzed deglycosylation of rituximab 
took place rapidly and was completed within 5 min under the reaction condition, even 
before any significant transglycosylation could occur (data not shown). Then the 
transglycosylation product (2 or 3) was steadily formed over time without apparent 
hydrolysis, reaching about 80% within 2 hours (Figure 37A and 37B). It was also 
observed that the transglycosylation with the hybrid type glycan oxazoline (6) went 
faster than that of the high-mannose type glycan oxazoline (5) under the same 
 109 
 
 
reaction condition. The mass spectrometry profiles of the antibody products formed 
after transglycosylation are shown in Figure 37C and 37D. The experimental data 
indicated that the free N-glycans in the medium released from the deglycosylation 
step did not interfere with the transglycosylation process, and the one-pot approach 
was able to transform the heterogeneous rituximab into homogenous high-mannose or 
hybrid glycoforms without the need of separating the deglycosylated intermediate or 
changing the catalyzing enzyme. It should be pointed out that this transformation 
could be optimized to push the reactions to completion by adding more glycan 
oxazolines and/or enzyme with prolonged incubation time. 
 
Figure 36. One-pot N-glycosylation remodeling of commercial rituximab to give rise 
to hybrid and high-mannose rituximab N-glycoforms, Hyb-RTX 2 and HM-RTX 3. 
 110 
 
 
 
Figure 37. Endo-S WT-catalyzed one-pot transglycosylation reaction with mannose 
(HM-Oxa 5) and hybrid (Hyb-Oxa 6): A) Transglycosylation with HM-Oxa 5 reached 
over 70% completion within 90 min; B) Transglycosylation with Hyb-Oxa 6 reached 
about 90% completion in 30 min; C) LC-MS spectrum of the transglycosylation 
product with HM-Oxa 5 showed two product species at 146542 and 148205; D) LC-
MS spectrum of transglycosylation product with Hyb-Oxa 6 showed two product 
species at 146551 and 148219. The results shown here are representative of two 
independently conducted experiments. 
 
 
 
 111 
 
 
12. One-pot glycosylation remodeling of commercial antibody with modified glycan 
oxazolines by wild-type Endo-S 
In addition to the hybrid and high mannose type N-glycans, we also tested the 
ability of Endo-S to act on the modified azido-tagged azide-Man3GlcNAc oxazoline 
(AzideM3-Oxa, 9) for Fc glycan remodeling of rituximab in a one-pot manner 
(Figure 38).  Following the same one-pot reaction protocol described above, we found 
that the conversion took place relatively slowly, but the transglycosylation product, 
the azide-tagged glycoform of rituximab was formed steadily. Within 6 h, nearly 40% 
conversion was achieved (Figure 38B). We also tested the Man3-glycan oxazoline 
(M3-Oxa 8). Interestingly, only marginal level of transglycosylation product was 
detected by LC-MS analysis in the case of using the Man3-glycan oxazoline (Figure 
38A). This result is consistent with our previous comparative studies on the Man3 
glycan oxazoline using the Endo-S D233Q mutant and the wild-type Endo-S, where 
only transitional formation of the transglycosylation product was observed by SDS-
PAGE analysis when the wild-type Endo-S was used, and significant conversion was 
only achieved using EndoS-D233Q mutant. Indeed, as shown in the oxazoline 
hydrolysis experiment described above (Figure 34B), Man3-glycan oxazoline was 
hydrolyzed rapidly by the wild-type Endo-S. In addition, the transglycosylation 
product was also hydrolyzed quickly by the wild-type Endo-S. Thus it was not a 
surprise that only transitional transglycosylation product could be detected during the 
reaction. It should be pointed out that Davis and co-workers previously reported the 
Endo-S-catalyzed transglycosylation with the same Man3-glycan oxazoline to the 
deglycosylated IgG antibody, but no detailed quantitative characterization of the 
 112 
 
 
transglycosylation product was reported except the MS analysis of the released N-
glycans. In the present case of the azide-Man3 glycan oxazoline (9), our enzymatic 
oxazoline hydrolysis experiment showed that the chemical modification increased to 
some extent its resistance toward the Endo-S-catalyzed hydrolysis (Figure 38B), 
which could partially explain the moderate transglycosylation activity using azide-
Man3-glycan oxazoline. It is expected that this one-pot enzymatic glycan remodeling 
method might be extended to other modified glycan oxazolines to prepare selectively 
functionalized antibody glycoforms for various structural and functional purposes 
such as development of antibody-drug conjugates. 
 
 
 113 
 
 
 
Figure 38. Endo-S WT-catalyzed one-pot transglycosylation reactions with 
chemically modified glycan oxazoline: A) transglycosylation with reached almost 40% 
completion within 360 min, whereas reaction with the Man3 core glycan (M3-Oxa 8) 
showed no detectable level of transglycosylation under the same condition; B) The 
LC-MS spectrum of the transglycosylation product with AzideM3-Oxa, 9 showed the 
two product species at mass 145618 and 146357, and a significant percentage of the 
starting material was still present in the reaction mixture after 360 min. The results 
shown here are representative of two independently conducted experiments. 
 114 
 
 
C. Conclusion 
We describe in this paper the discovery of a new class of glycosynthases 
derived from an endoglycosidase (Endo-S2) of the S. pyogenes M49 serotype, which 
showed broader substrate specificity and more potent transglycosylation activity for 
antibody glycosylation remodeling than the previously reported glycosynthases, such 
as those derived from Endo-A, Endo-M, and Endo-S. These findings were enabled by 
systematic mutagenesis at the critical residue Asp-184, coupled with comparative 
analysis of the hydrolysis and transglycosylation activities of the resulting 19 mutants. 
Our experimental data also revealed a remarkable difference in both the hydrolysis 
and transglycosylation activities among the 19 mutants, which would be difficult to 
predict without this comparative study. Several interesting mutants, including Endo-S 
D184M and D184Q, were identified that showed high transglycosylation activities for 
glycosylation remodeling but retained only marginal hydrolysis activity. Comparison 
of the hydrolysis and transglycosylation activities of Endo-S2 mutants reveals several 
interesting features. First, most of the mutants that showed high transglycosylation 
activities, including D184C, D184G, D184E, D184Y, D184S, and D184A mutants, 
also possessed relatively high residual hydrolysis activities. An exception is Endo-S2 
D184M, which demonstrated remarkable transglycosylation activity but retained a 
very low level of hydrolytic activity, making it the most efficient glycosynthase for 
glycosylation remodeling. Second, most of Endo-S2 mutants at the Asp-184 residue 
being replaced by amino acids with positively charged side chains (Lys, Arg, and His) 
or bulky hydrophobic side chains (Ile, Leu, Phe, and Trp) showed very low activities 
in both transglycosylation and hydrolysis. However, interestingly, the D184Y mutant 
 115 
 
 
retained the highest hydrolytic activity among all the mutants and also possessed a 
relatively high transglycosylation activity. A detailed understanding of the molecular 
mechanism behind these observations may rely on structural studies of the Endo-S2 
mutants and their complex with respective substrates. 
Another important discovery is the broader substrate specificity of Endo-S2-
derived glycosynthase mutants than the previously reported Endo-S glycosynthases. 
We found that the D184M and D184Q mutants, two notable Endo-S2 glycosynthases 
identified, were able to efficiently transfer all three major types of N-glycans 
including the high-mannose, complex, and hybrid type, in antibody glycosylation 
remodeling. In addition, the Endo-S2 glycosynthases could recognize both the core-
fucosylated and non-fucosylated GlcNAc moieties at the Fc domain as an acceptor for 
transglycosylation. These findings significantly expand the scope of the glycosylation 
remodeling strategy. For example, the previously reported Endo-S and Endo-F3 are 
specific for the complex type N-glycans and are unable to efficiently transfer the 
high-mannose and hybrid type N-glycans, and Endo-F3 is effective only for the core-
fucosylated GlcNAc acceptor.85 A direct comparison of the transglycosylation 
activities of the typical Endo-S2 and Endo-S glycosynthase mutants reveals that the 
Endo-S2 mutants are generally much more active than the corresponding Endo-S 
mutants. By an estimate of the initial rate, the D184Q mutant of Endo-S2 was at least 
10-fold more efficient than the corresponding D233Q mutant of Endo-S, and the most 
efficient Endo-S2 mutant, D184M, was estimated to be 100-fold better than Endo-S 
D233Q mutant in glycosylating the deglycosylated rituximab.  
 116 
 
 
It is expected that these highly efficient glycosynthases of Endo-S2 with the 
remarkably relaxed substrate specificity, will find a wide range of applications for 
producing various types of homogeneous glycoforms of antibodies in research and for 
developing more effective antibody-based therapeutics as well. 
We have demonstrated that the wild-type Endo-S possesses promiscuous 
hydrolysis and transglycosylation activities on different glycan oxazolines and 
antibody glycoforms, using therapeutic monoclonal antibody rituximab as a model 
system. The potent activity of Endo-S to deglycosylate the recombinant complex type 
N-glycoforms of antibodies, the promiscuous transglycosylation activity on the high-
mannose and hybrid type N-glycan oxazolines, and yet the lack of hydrolysis activity 
on the ?ground-state? high-mannose and hybrid type glycoforms collectively enabled 
us to develop an efficient one-pot N-glycan remodeling of heterogeneous glycoforms 
of rituximab to give homogenous high-mannose or hybrid type glycoforms, without 
the need of intermediate separation and enzyme switching. This one-pot enzymatic 
glycan remodeling method should be applicable to other therapeutic antibodies for 
obtaining specific N-glycoforms for various applications. 
 
 
 
 
 
 
 
 117 
 
 
D. Materials and Methods 
1. Chemical and biochemical materials  
Monoclonal antibody, rituximab, was purchased from Genentech Inc., (South 
San Francisco, CA). The complex (SCT) (4) and high-mannose type (HM) oxazoline 
(5) were harvested and synthesized from egg yolk and soybean flour, respectively, 
following previously described methods.127 The hybrid-type (Hyb) oxazoline (6) was 
converted from Man5GlcNAc structure (7) via a multi-step enzymatic treatment using 
a series of previously described enzymes, including ?-1,2-GlcNAc transferase, ?-1,4-
galactosyltransferase, and sialyltransferase. The Man3 (M3-Oxa 8) and azide-tagged 
Man3-(AizdeM3-Oxa 9) oxazolines were synthesized following previous publication. 
106 All oxazolines (4-9) were then converted from their corresponding free glycans 
using a one-pot transformation procedure.121 
2. Expression and purification of recombinant Endo-S WT and Endo-SD233Q  
The pGEX plasmid vector encoding the wild-type Endo-S was a gift from Dr. 
M. Collin (Lund University, Sweden). Site-directed mutagenesis was conducted as 
described in the previous publication to generate Endo-S D233Q mutant DNA.103 The 
plasmid DNA encoding both the wild-type Endo-S and the Endo-S D233Q mutant 
was transformed in E. coli BL21 (DE3) cells for overexpression. For purification of 
all Endo-S enzymes, the transformed cells were grown in 1L 2YT media with 100 
?g/mL Carbenicillin added. Cells were incubated at 37 ?C and induced with 0.5 mM 
isopropyl ?-D-1-thiogalactopyranoside (IPTG) when the cell density reached an 
OD600 of 0.8-1.0. The induced cells were incubated at 20 ?C overnight and harvested 
 118 
 
 
by high-speed centrifugation. The cell lysates were collected after treating with the 
bacterial cell lysis buffer (Gold Biotechnology, Inc.) following the manufacturer?s 
instructions. The GST-tagged proteins were purified using the GST resin (Thermo). 
The purified proteins were diluted into PBS at pH 7.4 after buffer exchange using 
Amicon ultra filtration units (10 kDa, Millipore). The purity of Endo-S proteins was 
over 90% on SDS-PAGE and the final concentration was recorded on a NanoDrop 
2000c at absorbance 280 nm. 
3. Mutagenesis, expression, and purification of recombinant Endo-S2 and mutants 
 The cDNA encoding amino acids 38 ? 819 of Endo-S2 from S. pyogenes 
NZ131 (serotype M49) was amplified by standard PCR and cloned into pCPD-Lasso 
vector (a pET22b-CPD derivative). For saturation mutagenesis of Asp-184 residue, a 
forward primer, 5?-CGTAAATTCGTGCTCAATNNNAATATCTAGTCCATCGA- 
CACCACGATCAGTT-3?, and a reverse primer, 5?-AACTGATCGTGGTGTCG- 
ATGGACTAGATATTNNNATTGAGCACGAATTTACG-3?, were used. Mutations 
were confirmed by DNA sequencing. The plasmids containing the mutated Endo-S2 
genes were transformed into E. coli BL21 (DE3). For simultaneous production of 20 
Endo-S2 Asp-184 variants, the transformants were cultured in 20 ml of 2YT broth 
media supplemented with 100 ?g/ml of carbenicillin. Cultures were grown at 37 ?C 
until cells reached an absorbance A600 of 0.8?1.0. Then 0.5 mM isopropyl-D-1-thio- 
galactopyranoside (IPTG) was added to the culture to induce protein overproduction 
at 20 ?C. After 24 h the cells were harvested by centrifugation. The cell pellets were 
then lysed by Bacterial Cell Lysis Buffer (Gold Biotechnology, Inc.) following the 
manufacturer?s instructions. The His10-tagged Endo-S2/CPD fusion proteins were 
 119 
 
 
purified with nickel-nitrilotriacetic acid spin columns (Qiagen). The purified Endo-S2 
proteins were desalted into PBS (pH 7.4) buffer using centrifugal diafiltration with 
Amicon Ultrafiltration (10 kDa, Millipore). The purity was confirmed by SDS-PAGE 
analysis and the concentration was measured on a NanoDrop 2000c using absorbance 
at 280 nm. For large-scale purification of selected Endo-S2 variants, 1 liter of culture 
was used. The cell lysate was applied to a HisTrap HP column (GE Healthcare) and 
washed with 0.5 M NaCl and 20 mM imidazole added (pH 7.4). Bound His-tagged 
protein was eluted with a gradient of 0 ?250 mM imidazole in PBS buffer. The eluted 
fractions containing Endo-S2 proteins were pooled, concentrated, and further purified 
by size exclusion chromatography using HiPrep 16/60 Sephacryl S-200 HR column 
(GE Healthcare).  
4. Liquid Chromatography Mass Spectrometry (LC-ESI-MS) analysis  
Antibody N-glycoforms were studied using an Exactive Plus Orbitrap 
instrument equipped with a HPLC system and an ESI source. The capillary 
temperature and the spray voltage were set according to previous publications. 134 135 
Approximately 25 ?L of the intact rituximab solution was injected and separated via a 
Waters XBridgeTM BEH300 C4 column (3.5 ?m, 2.1 x 50 mm) at a flow rate of 0.4 
mL/min. The program includes a 9 min linear gradient of 5-90% MeCN containing 
0.1 % formic acid. The starting material deglycosylated rituximab (m/z 144879) and 
transglycosylation products were measured using selected ion monitoring (SIM). The 
signal intensities of the target ions were optimized by automatic tuning function and 
the same optimized conditions were used in all LC-MS quantifications. The MagTran 
Software (Amgen) was used to analyze the data from the exported chromatogram 
 120 
 
 
results.  An average of scans was analyzed through deconvolution to give a spectrum 
list for each analyzed sample. Integration of the deconvoluted spectrum shows the 
relative abundance of the analyzed ions. Total peak intensities within 0.7 mass units 
surrounding the center of the target ion peak were used to determine the ratio of the 
starting material and the product species. This relative ratio can be used to calculate 
the reaction yield of each hydrolysis and transglycosylation reaction with the 
concentration of the starting material known.  
5. HPAEC-PAD analysis of glycans and oxazolines 
High-performance anion-exchange chromatography coupled with pulsed 
amperometric detection (HPAEC-PAD) analysis was conducted on a DIONEX 
chromatography system (DIONEX Corporation, Sunnyvale, CA) using an anion 
exchange column (CarboPac PA200, 3 ? 250 mm). The sample was delivered with a 
mobile phase consisting of deionized water (eluent A), 0.2 M NaOH (eluent B), 0.1M 
M NaOH (eluent C) and 0.1M M NaOH with 0.25M NaOAc (eluent D). The program 
for the complex and hybrid type glycans and oxazolines used a gradient as follows: 0 
min, 0% eluent A, 0% eluent B, 100% eluent C and 0% eluent D; 5.0 min, 0% eluent 
A, 0% eluent B, 100% eluent C and 0% eluent D; and 25.0 min, 0% eluent A, 0% 
eluent B, 60% eluent C and 40% eluent D at a flow rate of 1.0 mL/min. The program 
for oligomannose type glycans and oxazolines (Man9, Man5, Man3 and Azide-Man3) 
were analyzed using a different program: 0 min, 0% eluent A, 0% eluent B, 100% 
eluent C and 0% eluent D; 5.0 min, 0% eluent A, 0% eluent B, 100% eluent C and 0% 
eluent D; and 25.0 min, 0% eluent A, 0% eluent B, 80% eluent C and 20% eluent D 
 121 
 
 
with a flow rate of 1.0 mL/min. The relative abundance of each glycan and oxazoline 
was calculated from the integrated spectrum. 
6. Comparison of hydrolytic activity of Endo-S against rituximab N-glycoforms  
The rituximab N-glycoforms, SCT-RTX 1, HM-RTX 2 and Hyb-RTX 3 were 
synthesized according to previously published protocol.139 In three parallel reactions, 
each of the three N-glycoforms (0.01 mg, 6.9 ?M) was incubated with Endo-S WT (5 
mg/mL) at 30 ?C in 20 ?l of PBS buffer (pH 7.4) for a given amount of time. Each 
reaction was quenched by adding 20 ?l of 0.1% formic acid and the sample was 
placed at -20 ?C. The hydrolysis yield was measured by calculating the ratio of the 
hydrolyzed product versus the starting material using LC-ESI-MS as shown above.   
7. Comparison of hydrolytic activity of Endo-S WT towards different oxazolines 
In parallel reactions, SCT-Oxa 4, HM-Oxa 5, Hyb-Oxa 6, M5-Oxa 7, M3-Oxa 
8 and AzideM3-Oxa 9 oxazolines (1 mM) was incubated with Endo-S (0.05 mg/ml) at 
30 ?C in a total of 25 ?l of PBS buffer (pH 7.4). Aliquots of reaction mixture (1 ?l) 
was sampled at the same time intervals and diluted in 100 mM NaOH solution to 
quench the reaction. All reactions were monitored by HPAEC-PAD and the relative 
percentage of the hydrolyzed oxazolines (free glycans) was quantified from the 
integrated spectrum as detailed in previous experimental section.  
8. Evaluation of transglycosylation activity of Endo-S on deglycosylated rituximab 
using SCT-Oxa 4, HM-Oxa 5 and Hyb-Oxa 6 as glycosyl donors. 
The ability of Endo-S WT to transfer three different N-glycan oxazolines 
SCT-Oxa 4, HM-Oxa 5 and Hyb-Oxa 6 to the deglycosylated rituximab 10 in a 
 122 
 
 
conventional two-step transglycosylation reaction was compared under the same 
reaction conditions. First, the commercial rituximab underwent deglycosylation by 
incubation with a catalytic amount of Endo-S WT (0.01 mg/mL) at 30 ?C for a short 
period of time. Once the complete deglycosylation was confirmed, the deglycosylated 
rituximab antibody was subsequently purified through Protein-A chromatography for 
further modification. The following transglycosylation reaction was performed using 
the deglycosylated rituximab 10 (10 mg/mL) as the glycosyl acceptor and N-glycan 
oxazolines as glycosyl donors with a catalytic amount of Endo-S WT (0.05 mg/mL) 
at 30 ?C in 20 ?l of PBS buffer (pH 7.4). The molar ratio of the donor substrate to the 
glycosyl acceptor was set to be 40 : 1. 
9. Evaluation of transglycosylation activities of Endo-S WT and Endo-S D233Q 
towards deglycosylated rituximab antibody using HM-Oxa 5 and Hyb-Oxa 6 as 
glycosyl donors. 
The transglycosylation efficiency of Endo-S WT and Endo-S D233Q towards 
the deglycosylated rituximab was compared separately for HM-Oxa 5 and Hyb-Oxa 6. 
For reaction with either oxazoline, the deglycosylated rituximab (10 mg/mL) was 
incubated with each enzyme (0.5 mg/mL) at 30 ?C in 20 ?l of PBS buffer (pH 7.4). 
The molar ratio of the donor substrate to the glycosyl acceptor was also 40 : 1. The 
transglycosylation products were analyzed and quantified as mentioned previously.  
 123 
 
 
10. Transglycosylation of rituximab with Hyb-Oxa and HM-Oxa by Endo-S WT in a 
one-pot manner. 
Commercial Rituximab (0.1 mg, 69 ?M) was dissolved in PBS buffer (pH 7.4) 
and was mixed with 40 equivalents of hybrid and high mannose oxazolines HM-oxa 5 
(2.8 mM) and Hyb-oxa 6 (2.8 mM), respectively. The reactions were initiated by 
adding a catalytic amount of Endo-S WT (0.5 mg/ml) and incubated in PBS at 30?C 
(pH 7.4). For the product quantification, the concentration of the product glycoform 
carrying two N-glycans, 148205 Da for HM-RTX and 148219 Da for Hyb-RTX 
(deconvoluted mass), were counted twice as each antibody carries two glycosylation 
sites. The N-glycoform carrying only one N-glycan for each oxazoline was only 
counted once as the product formed. The same quantification method using LC-ESI-
MS was used for the similar experiments conducted in this study. 
11. Transglycosylation of commercial Rituximab with M3-Oxa and AzideM3-Oxa by 
Endo-S WT in a one-pot fashion. 
Commercial rituximab (0.1 mg, 69 ?M) was incubated in PBS at 30?C (pH 
7.4) with 40 equivalents of oxazolines M3-Oxa 8 (2.8 mM) and AzideM3-Oxa 9 (2.8 
mM), respectively. To initiate the reaction, a fixed amount of Endo-S WT (0.5 mg/ml) 
was added into the reaction mixture at t = 0 min. The reactions with the M3-Oxa and 
AzideM3-Oxa oxazolines were maintained for a prolonged period of time to allow the 
accumulation of transglycosylation product. Additional 40 equivalents of oxazolines 
were added to each reaction at t = 120 min. Transglycosylation with AzideM3-Oxa 9 
showed three species at 144879 Da (starting material), 145618 Da (antibody product 
 124 
 
 
with a single glycan attached) and 146357 Da (antibody product with two glycans 
attached).  The reaction yield was quantified as described above. 
E. Supporting Information 
MKTVQTGKTDQQVGAKLVQEIREGKRGPLYAGYFRTWHDRASTGIDGKQQHPENTMA
EVPKEVDILFVFHDHTASDSPFWSELKDSYVHKLHQQGTALVQTIGVNELNGRTGLS
KDYPDTPEGNKALAAAIVKAFVTDRGVDGLDIDIEHEFTNKRTPEEDARALNVFKEI
AQLIGKNGSDKSKLLIMDTTLSVENNPIFKGIAEDLDYLLRQYYGSQGGEAEVDTIN
SDWNQYQNYIDASQFMIGFSFFEESASKGNLWFDVNEYDPNNPEKGKDIEGTRAKKY
AEWQPSTGGLKAGIFSYAIDRDGVAHVPSTYKNRTSTNLQRHEVDNISHTDYTVSRK
LKTLMTEDKRYDVIDQKDIPDPALREQIIQQVGQYKGDLERYNKTLVLTGDKIQNLK
GLEKLSKLQKLELRQLSNVKEITPELLPESMKKDAELVMVGMTGLEKLNLSGLNRQT
LDGIDVNSITHLTSFDISHNSLDLSEKSEDRKLLMTLMEQVSNHQKITVKNTAFENQ
KPKGYYPQTYDTKEGHYDVDNAEHDILTDFVFGTVTKRNTFIGDEEAFAIYKEGAVD
GRQYVSKDYTYEAFRKDYKGYKVHLTASNLGETVTSKVTATTDETYLVDVSDGEKVV
HHMKLNIGSGAIMMENLAKGAKVIGTSGDFEQAKKIFDGEKSDRFFTWGQTNWIAFD
LGEINLAKEWRLFNAETNTEIKTDSSLNVAKGRLQILKDTTIDLEKMDIKNRKEYLS
NDENWTDVAQMDDAKAIFNSKLSNVLSRYWRFCVDGGASSYLGSGKILHNQNVNSWG
PITVTPTTDGGETRFDGQIIVQMENDPVVAKAAANLAGKHAESSVVVQLDSDGNYRV
VYGDPSKLDGKLRWQLVGHGRDHSETNNTRLSGYSADELAVKLAKFQQSFNQAENIN
NKPDHISIVGCSLVSDDKQKGFGHQFINAMDANGLRVDVSVRSSELAVDEAGRKHTK
DANGDWVQKAENNKVSLSWDAQLEGGSGGSGNSGHHHHHHHHHH 
 
Supplementary Figure 2: amino acid sequence of the recombinant Endo-S2 construct 
containing a CPD fusion tag at the C-terminus (GeneBank ID: KC155346). 
 
 
 
 
 
 
 
 
 125 
 
 
Chapter 4: Revisiting the substrate specificity of endoglycosidase Endo-
CC and its mutants in hydrolysis and transglycosylation 
A. Introduction 
The Endo-?-N-acetylglucosaminidase (ENGase) (EC 3.2.1.96) belongs to a 
family of glycoside hydrolases (GH) that catalyze the cleavage of the ?-1,4-glycosidic 
linkage within the innermost N,N?-diacetylchitobiose core of the asparagine-linked 
oligosaccharides.98 The unique ability of ENGases to process the N-glycan structures 
with stringent stereo- and regio-selectivity has made them important tools for analysis 
and synthesis of the homogeneous glycoproteins and oligosaccharides. Besides their 
hydrolytic activities, we and others have previously reported that many ENGases also 
possessed the capability to re-attach the released glycans to different acceptors by 
forming a new glycosidic linkage, a process also known as transglycosylation.89, 93, 94 
Several such ENGases, including Endo-A from Arthrobacter protophormiae102, 
Endo-M from Mucor hiemali99, and Endo-D from Streptococcus pneumonia101, 
exhibit the remarkable activity to transfer various N-linked oligosaccharides to the 
GlcNAc-containing peptides en bloc to afford new glycopeptides and glycoproteins. 
Importantly, those ENGases exhibit distinct substrate specificities. Whereas multiple 
ENGases such as Endo-A and Endo-H could act on the high mannose type N-glycans, 
Endo-M was found to be particularly interesting, as it has displayed relatively broad 
substrate specificity towards all three major types of N-glycans (complex, hybrid, and 
high-mannose type) for both hydrolysis and transglycosylation.100 However, the 
application of Endo-M in large-scale glycoengineering of glycoproteins is hindered 
by technical obstacles such as the difficulties to produce the active Endo-M enzymes 
 126 
 
 
in large quantities142. To overcome these problems with the existing ENGases, new 
unknown ENGases with broad substrate specificity and high catalytic efficiency are 
still being actively searched. 
Recently, the discovery of a new ENGase, namely Endo-CC from the fungus 
Coprinopsis cinerea, has attracted much attention in the glycoengineering field due to 
its ability to hydrolyze both the complex and high-mannose N-glycans.142 Using 
simple fluorescence-labeled oligosaccharides as standard substrates, Takegawa and 
colleagues demonstrated that Endo-CC was able to hydrolyze multiple types of 
Pyridylamino (PA)-labeled-glycans with relatively high efficiencies. Subsequently, 
multiple glycoproteins with different N-glycans, including the ribonuclease B (RNase 
B), human transferrin and fetuin, were tested with Endo-CC to study its substrate 
specificity for hydrolysis. The results suggested similar substrate selectivity between 
Endo-CC and Endo-M, both of which exhibited potent hydrolytic activities against 
the biantennary complex and high-mannose type N-glycans, although unlike Endo-M, 
Endo-CC displayed no activities against the bisecting GlcNAc. In addition, neither 
Endo-CC nor Endo-M were able to act on the triantennary or tetra-antennary N-linked 
oligosaccharides. However, it remains to be investigated whether the fucosylated 
glycoproteins and glycopeptides are subject to hydrolysis by Endo-CC.89 
Following similar studies on other ENGases, a library of glycosynthase 
mutants from Endo-CC were generated by site-directed mutagenesis at a previously 
identified catalytic key residue Asn-180. Functional studies further demonstrated that 
mutant Endo-CC N180H was the most efficient glycosynthase variant of Endo-CC to 
transfer the biantennary complex type N-glycan to the deglycosylated RNase B.110 
 127 
 
 
The reported comparative studies suggested that, unlike previously studied of Endo-
S123, Endo-S2139, and Endo-F385, among the 19 mutants generated at Asn-180 residue 
of Endo-CC, only two mutants Endo-CC N180H and N180Q showed significant 
transglycosylation activities using both the sialoglycopeptide (SGP) and the activated 
glycan oxazolines as the donor substrates. A following study revealed that Endo-CC 
N180H can transfer complex glycan to a range of acceptors including the monoclonal 
antibody, although a large excess amount of SGP instead of glycan oxazoline was 
used as donor substrate for the transglycosylation of antibody. In addition, Endo-CC 
N180H was also applied in combination with other ENGases such as Endo-S D233Q 
to improve the transglycosylation efficiency of antibody.111 
Despite successful application of Endo-CC enzymes in the chemoenzymatic 
synthesis of glycoproteins, there are still several unknown aspects regarding the 
substrate specificities and catalytic efficiency of Endo-CC and its mutants: A) The 
reported expression yield of the original Endo-CC construct was less than 0.5 mg/L of 
bacterial cell culture. It is unclear whether an alternative purification strategy utilizing 
any peptide or protein fusion tags could increase the production of Endo-CC and 
potentially enhance its stability; B) The glycosyl donor substrates tested for Endo-CC 
transglycosylation were limited to the complex type glycans only, thus it is currently 
unknown whether Endo-CC and/or its glycosynthase mutants could use other glycan 
oxazolines, such as the hybrid and the high-mannose glycans, for transglycosylation 
reactions; C) The previous studies provided no experimental data concerning the 
substrate selectivity of Endo-CC and its mutants for the core-fucosylated glycans and 
glycoconjugates; D) The reported experiments detailing transglycosylation activities 
 128 
 
 
of Endo-CC mutants did not show complete reaction progressions for Endo-CC N180 
mutants except for Endo-CC N180H and N180Q. Additional functional assays might 
be required to identify Endo-CC mutants that possess a different catalytic landscape 
from the known N180H and N180Q; E) Whether Endo-CC and/or its mutants could 
independently transfer activated glycan oxazoline instead of sialylglycopeptide (SGP) 
to the deglycosylated antibody is still a question under debate. 
To improve our understanding of the substrate specificity of Endo-CC and its 
mutants, we first created a new version of Endo-CC containing a C-terminal 6xHis-
tagged fusion protein called the ?Cysteine Protease Domain (CPD)? tag to improve 
solubility and expression yield143. The expression yield of the soluble Endo-CC using 
the new construct was estimated to be higher than 4 mg/L of E. coli culture and the 
CPD-tagged Endo-CC and Endo-CC N180H exhibited the same level of activity as 
the untagged enzymes. Next, we performed further substrate selectivity test on Endo-
CC and its glycosynthase mutants with a variety of substrates for both hydrolysis and 
transglycosylation. It was discovered that the Endo-CC N180H could potentially 
transfer additional glycan oxazolines, such as the high-mannose and hybrid type 
glycans, to afford new homogeneous glycopeptides and glycoproteins. Furthermore, 
side-by-side comparative studies on the fucosylated and non-fucosylated substrates, 
including SGP, Erythropoietin (EPO) and IgG antibodies, revealed that core-
fucosylation of glycoconjugates is not tolerated by Endo-CC and its mutants. Finally, 
we performed additional enzyme activity assays on the selected mutants of Endo-CC 
and demonstrated that several Endo-CC mutants besides N180H and N180Q 
exhibited significant transglycosylation activity towards a CD52-GlcNAc peptide 
 129 
 
 
using the complex glycan oxazoline as the donor substrate. However, the relative 
catalytic efficiency of these mutants varied significantly over prolonged reaction 
time. Interestingly, Endo-CC N180A, in contrast to several other Endo-CC mutants, 
demonstrated a relatively slow but persistent transglycosylation activity over a long 
period of reaction time. 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 130 
 
 
B. Results and Discussions 
1. Cloning, expression, and comparison of recombinant Endo-CC and Endo-CC 
fusion protein (CPD) 
The DNA sequence encoding Endo-CC was initially synthesized (GenScript) 
and cloned into a pET-41b (+) vector, which contains a 7-histidine tag at C terminus. 
It was previously reported that soluble expression of Endo-CC and mutants could be 
achieved using leaky expression without any induction.105 Following a similar 
expression protocol, Endo-CC was successfully expressed in Escherichia coli and 
purified via the immobilized metal ion affinity chromatography with the expected 
yield (~400 ?g/L). The Endo-CC DNA was also cloned into a pET22b-CPD vector, 
which contained an additional C-terminal fusion tag (CPD) harvested from the Vibrio 
cholerae MARTX toxin.141 Consistent with our previous studies on the expression of 
ENGases,85, 144 the CPD fusion protein significantly increased the expression yield of 
the soluble Endo-CC by approximately 10 folds (~ 4 mg/L). Using the ribonuclease B 
as the standard substrate, the hydrolytic activities of Endo-CC and CPD-tagged Endo-
CC were found to be comparable. Both versions of Endo-CC were able to rapidly 
hydrolyze the high-mannose N-glycans of RNase B under mild reaction conditions 
(Figure 39).  
 131 
 
 
 
 
Figure 39. Hydrolysis of RNase B by recombinant Endo-CC and Endo-CC (CPD). 
MS analysis of (A) RNase B and (B) Deglycosylated RNase B-GlcNAc. (C) 
Comparison of hydrolytic activities of untagged and CPD-tagged Endo-CC enzymes. 
A total of 10 mg/mL of RNase B was incubated with 0.1 mg/mL of both recombinant 
Endo-CC (+/- CPD) at 37 ?C in Tris buffer at pH 7.4. At t = 2 h, oligomannose 
glycans of RNase B were completely hydrolyzed by both versions of Endo-CC. The 
reactions were monitored by SDS-PAGE densitometry and the final products were 
 132 
 
 
confirmed by LC-MS (B). The schematic structure of RNase B was based on the 
published crystal structure (1RBB)145.  
2. Generation of glycosynthase mutants from recombinant Endo-CC 
Takegawa and co-workers have previously generated glycosynthase mutants 
of Endo-CC through site-directed mutagenesis at a key amino acid residue that is 
critical for the formation of the oxazolinium ion intermediate according to a substrate-
assisted mechanism.105 Based on previous studies on other endoglycosidases, such as 
Endo-M99, Endo-A92, Endo-S123 and Endo-S2139, several mutants of Endo-CC were 
generated via site-directed mutagenesis at the same key residue Asn-180, including 
N180A, N180H, N180M, N180Q and N180E, to study their transglycosylation 
activities in real-time reaction courses. These mutants were expressed and purified 
following the same protocol as for the wild-type Endo-CC with a comparable yield 
and purity, which were confirmed by SDS-PAGE.105 In particular, the CPD-tagged 
Endo-CC N180H mutant was also tested for its glycosylation activity in parallel to 
the original Endo-CC construct, and the results indicated that both versions of Endo-
CC N180H displayed similar transfer activities towards the deglycosylated RNase B 
using the sialylated complex glycan oxazoline (SCT-Oxa) as substrate (Figure 40). 
Nevertheless, like the wild-type Endo-CC, the expression yield of the CPD-tagged 
Endo-CC N180H (~ 5.0 mg/L) was significantly higher than that of the non-CPD 
mutant (0.5 mg/L).   
 133 
 
 
 
 
Figure 40. Transglycosylation of deglycosylated RNase B by recombinant Endo-CC 
and Endo-CC (CPD). LC?MS analysis of (A) N-glycoforms of RNase B-GlcNAc and 
(B) RNase B-SCT by Endo-CC N180H. (C) Comparison of the transglycosylation 
reactions on RNase B-GlcNAc by the untagged and CPD-tagged Endo-CC N180H 
mutants. A total concentration of 10 mg/mL of RNase B-GlcNAc was incubated with 
 134 
 
 
0.5 mg/mL of both recombinant Endo-CC N180H (+/- CPD) at 30 ?C in Tris buffer at 
pH 7.4. At t = 2 h, the sialylated complex form of RNase B was synthesized by both 
version of Endo-CC with similar catalytic efficiencies. The reactions were monitored 
by SDS-PAGE densitometry and the final products were confirmed by LC-MS (B). 
The schematic structure of RNase B was based on the published crystal structure 
(1RBB)145. 
3. Substrate specificity of Endo-CC in hydrolysis of glycoproteins: core-fucosylation 
of glycosyl acceptor substrates is not tolerated by Endo-CC 
Despite that sequence alignment between Endo-CC and Endo-M revealed 
significant sequence similarity (approximately 46%), it is currently unknown whether 
Endo-CC can process the core-fucosylated glycoproteins as suitable substrates for 
glycan hydrolysis.142 Here, a number of commonly studied glycoproteins in both their 
core-fucosylated and non-fucosylated forms were tested for their susceptibility to the 
Endo-CC-catalyzed hydrolysis of N-glycans, including the sialoglycoprotein (SGP), 
erythropoietin (EPO) and monoclonal IgG antibody N-glycoforms. As shown in 
Figure 41, the complex type N-glycans attached to the non-fucosylated SGP was 
completely released by Endo-CC after 4 h, whereas the core-fucosylated SGP peptide 
was resistant to Endo-CC-mediated hydrolysis after the overnight reaction.  
To more clearly reflect the substrate selectivity of Endo-CC against the core 
fucose, a glycoengineered protein erythropoietin (EPO), a therapeutic glycoprotein 
hormone that stimulates erythropoiesis, was prepared by an engineered mammalian 
expression system to carry a mixture of the core-fucosylated and non-fucosylated 
complex N-glycoforms (Figure 42A).146 Treatment of EPO N-glycoforms by Endo-
 135 
 
 
CC revealed that all non-fucosylated EPO N-glycans were hydrolyzed by Endo-CC in 
as early as t = 2 h, whereas the core-fucosylated EPO glycoforms remained 
unaffected after overnight incubation (Figure 42B).  
Similarly, both the core-fucosylated and non-fucosylated complex-type IgG 
N-glycoforms (namely IgG-S2G2F and IgG-S2G2) were synthesized and compared for 
their susceptibilities to the Endo-CC-catalyzed deglycosylation. As the comparative 
results demonstrated (Figure 43), although neither IgG glycoforms were completely 
hydrolyzed by Endo-CC after overnight, the IgG-S2G2F form was distinctively more 
resistant to Endo-CC hydrolysis in comparison with the IgG-S2G2. Interestingly, the 
high-mannose form of IgG (IgG-HM) was also synthesized139 and tested as a control 
glycoform. Under the same reaction conditions, the majority of the high-mannose N-
glycans of IgG-HM were released by Endo-CC with only a small portion of the 
mono-glycosylated species present on the mass spectrometry (MS) profile after the 
reaction. These unexpected results suggested that the high-mannose N-glycans of IgG 
are more susceptible to the Endo-CC-mediated hydrolysis in comparison to the 
fucosylated and non-fucosylated complex type N-glycans. 
 
 
 
 
 
 136 
 
 
 
Figure 41. Comparison of Endo-CC-catalyzed hydrolysis against fucosylated and 
non-fucosylated sialoglycoprotein (SGP).  The concentration of both forms of SGP 
was maintained at 50 mg/mL and samples were incubated with 0.1 mg/mL 
recombinant Endo-CC at 37 ?C in PBS buffer at pH 7.4. The hydrolytic reactions 
were monitored by HPAEC-PAD analysis. At t = 4 h, the non-fucosylated SGP 
(bottom curve) was completely hydrolyzed by Endo-CC, whereas core-fucosylated 
SGP (top curve) remained unchanged after overnight incubation, supported by the 
different retention times of SGP and the hydrolyzed SCT free glycan, respectively.  
 
 137 
 
 
 
Figure 42. Selective deglycosylation of non-fucosylated N-glycans by Endo-CC from 
the heterogeneous N-glycoforms of erythropoietin (EPO).A total concentration of 0.5 
mg/mL of EPO (A) with a mixture of fucosylated and non-fucosylated oligomannose 
N-glycans was incubated with 0.01 mg/mL Endo-CC at 37 ?C in PBS buffer at pH 
7.4. The reaction was quenched at t = 2 h and analyzed by LC-MS. (B) After 
deglycosylation, only the non-fucosylated N-glycans were completely released, while 
the fucosylated N-glycans remained unaffected. 
 138 
 
 
 
Figure 43. Selective removal of fucosylated and non-fucosylated IgG N-glycans by 
Endo-CC. A total concentration of 5 mg/mL of IgG N-glycoforms was incubated with 
0.05 mg/mL Endo-CC at 37 ?C in the Tris-HCl buffer at pH 7.4. For the fucosylated 
IgG-S2G2F glycoform (upper row), no hydrolysis was observed, whereas a small 
percentage of the non-fucosylated complex IgG-S2G2 underwent deglycosylation 
(middle row). After overnight incubation, nearly complete release of N-glycans was 
achieved by Endo-CC for the high-mannose IgG glycoform (IgG-HM, bottom row). 
The deglycosylation reactions were monitored by ESI-LC-MS. 
 139 
 
 
4. Studies on the substrate specificity of Endo-CC in transglycosylation: 
4.1. Transglycosylation of high-mannose, hybrid and complex type glycan oxazolines 
by mutant Endo-CC D180H  
To further understand the substrate specificity of Endo-CC N180H mutant 
against other types of N-glycans besides the complex type, three N-glycan oxazoline 
substrates derived from the complex (SCT), hybrid (Hyb) and high-mannose (HM) 
glycans were tested for transglycosylation by Endo-CC N180H towards the GlcNAc-
containing peptide (CD52-GlcNAc) to provide the side-by-side comparison for the 
substrate specificity of Endo-CC N180H mutant. The transfer reactions were assayed 
at multiple time points and the products were analyzed by the analytical HPLC and 
characterized by mass spectrometry. The results are summarized in Figure 44. Under 
the same reaction condition, the transglycosylation yield by Endo-CC N180H with 
HM-Oxa reached approximately 70% within 480 min to afford the CD52-HM, 
whereas the reactions with Hyb-Oxa and SCT-Oxa were relatively slower but still 
significant. These results demonstrated that mutant Endo-CC N180H possessed broad 
substrate selectivity against all three major N-glycan types for transglycosylation, 
despite that the transfer efficiency varied depending on the type of N-glycan, as the 
transfer for HM-Oxa was relatively more efficient than those for the SCT-Oxa and 
Hyb-Oxa by comparison. 
 
 
 
 
 140 
 
 
80
SCT-Oxa
Hyb-Oxa
60 HM-Oxa
40
20
0
0 60 120 180 240 300 360 420 480
Time/min
 
Figure 44. Comparison of transglycosylation efficiency of Endo-CC N180H for three 
major types of glycan oxazolines (high-mannose, hybrid and biantennary complex). 
The transglycosylation reactions were conducted using the purified CD52-GlcNAc (5 
mg/mL) as the glycosyl acceptor and different glycan oxazolines as the glycosyl 
donor substrates at 37 ?C in Tris-HCl buffer at pH 7.4. The final concentration of 
Endo-CC N180H was 0.125 mg/mL. The molar ratio of oxazoline to peptide was 5:1. 
The results presented here are representative of two independent sets of experiments. 
 
4.2. Comparative studies on the transglycosylation activities of Endo-CC mutants 
Previous systematic mutagenesis studies on other ENGases such as Endo-S123, 
Endo-S2139 and Endo-M100 suggested that substitution of the catalytic residue (D233, 
D184 and N175, respectively) with methionine (M), glutamine (Q), alanine (A) and 
glutamate (E) all led to generation of more efficient glycosynthase mutants for 
transglycosylation with glycan oxazolines. However, the previous mutational studies 
on Endo-CC indicated that only N180H and N180Q showed significant transfer 
 141 
 
Transglycosylation Yield (%)
 
activities against SGP and SCT-Oxa. Here, we performed similar mutagenesis on the 
Asn-180 residue of Endo-CC and expressed several Endo-CC mutants, including 
EnN180A, N180H, N180M and N180E and N180Q. The transglycosylation activities 
of these mutants were assayed using the CD52-GlcNAc glycopeptide as glycosyl 
acceptor substrate and the complex type glycan oxazoline (SCT-Oxa) as the glycosyl 
donor. The progression of each transfer reaction was monitored by the analytical 
HPLC followed by MS characterization (Figure 45). Side-by-side comparison of the 
progression patterns catalyzed by each mutant under the same reaction condition 
indicated that the N180H mutant indeed demonstrated the highest transglycosylation 
yield in the first 4 hours of the reaction, which was consistent with the previous 
publication110. However, the accumulation of the product by Endo-CC N180H 
stopped at t = 4 h, and the product CD52-SCT underwent significant hydrolysis 
during the rest of the reaction time. Similarly, the N180E mutant exhibited the second 
highest reaction yield within the first 8 hours before product hydrolysis. The 
behaviors of these mutants were consistent with several other ENGase mutants, as 
mentioned above. 
On the other hand, additional mutants, namely Endo-C N180M, N180A, and 
N180Q, were found to exhibit relatively slow but highly persistent product formation 
with no significant product hydrolysis observed in the first 24 hours of the reaction. 
Interestingly, the N180A mutant exhibited continuous accumulation of the transfer 
product for at least 24 hours without any sign of product hydrolysis. Therefore, at t = 
24 h, the overall transglycosylation yield of the N180A mutant was the highest among 
all the mutants tested here, as a large percentage of the transfer product formed by 
 142 
 
 
N180H and N180E was hydrolyzed. The wild-type Endo-CC was also included as a 
control to demonstrate the dominant hydrolytic activity of the wild-type enzyme and, 
as a result, no accumulation of the transfer product under the presented condition. 
Unlike the D233M and D184M mutant of Endo-S and Endo-S2, respectively, the 
Endo-CC N180M did not achieve a high level of transglycosylation in comparison 
with N180Q and N180A. However, the minimal tendency of Endo-CC N180M to 
hydrolyze the product was partially consistent with the behavior of Endo-S D233M 
Endo-S2 D184M.123, 139 
60
N180H
N180A
40 N180M
N180Q
N180E
20 WT
0
0 240 480 720 960 1200 1440
Time/min
 
Figure 45. Comparison of the transglycosylation reactions catalyzed by selected 
Endo-CC mutants. The reaction was carried out using CD52-GlcNAc (2 mg/mL) as 
the glycosyl acceptor and SCT-Oxa as the glycosyl donor substrate at 30 ?C in Tris-
HCl buffer at pH 7.4. The final concentration of mutant/wild-type enzyme was 0.125 
mg/mL. The molar ratio of SCT-Oxa to CD52-GlcNAc was 5: 1. The results 
presented here are representative of two independent sets of experiments.  
 143 
 
Transglycosylation Yield (%)
 
5. Glycosylation remodeling of a therapeutic antibody rituximab using Endo-CC and 
Endo-CC N180H with glycan oxazoline as glycosyl donor substrate 
Finally, glycoengineering of antibody was also attempted in this work using 
either the wild-type Endo-CC or Endo-CC mutant and the complex type glycan 
oxazoline (SCT-Oxa) as substrate. Similar studies have been conducted by Ito and 
others to generate the homogeneous antibody N-glycoforms using Endo-CC 
enzymes.111, 112 However, these studies involved a combination of two enzymes, 
Endo-CC and Endo-S mutants, and a large excess of SGP (approximately 25000 
molar equivalents to antibody) used as the glycosyl donor substrate. How Endo-CC 
and its mutants alone catalyze the transglycosylation of the deglycosylated antibody 
using the glycan oxazoline as donor substrate remains a question to be answered in 
more detail. Here, glycosylation remodeling was performed on rituximab, which has 
been prescribed for breast cancer treatment.63 The process of glycan remodeling of 
antibody was illustrated by a two-step chemoenzymatic strategy. First, the mixture of 
the native N-glycans of rituximab was removed by immobilized Endo-S2 to produce 
the deglycosylated antibody acceptor. The intermediate was then purified and used as 
glycosyl acceptor for the following transglycosylation reaction. The transfer reactions 
were conducted by incubating the deglycosylated antibody with a moderate amount 
of activated SCT-Oxa (20 molar equivalents to antibody) in the presence of varying 
concentrations of both Endo-CC and Endo-CC N180H. The relative reaction yield for 
transglycosylation by each enzyme was estimated by LC-MS following our 
previously described protocol.123 As shown in Figure 46, Endo-CC and Endo-CC 
N180H at concentration 0.05 mg/mL did not show significant transfer of SCT glycan 
 144 
 
 
oxazoline to the antibody after a prolonged incubation time. However, wild-type 
Endo-CC at a reduced concentration (0.01 mg/mL) indeed demonstrated nearly 40% 
conversion of antibody glycan at t = 4 h, although the same result was not observed 
for mutant Endo-CC N180H at the same low concentration. These results are 
supported by previous reports that a low concentration of the wild-type ENGase could 
be used for transglycosylation of certain substrates with moderate efficiency, 
although further optimization might be required to achieve a near-completion level of 
transglycosylation, especially with complex substrates like antibody.99 
 
50
Endo-CC N180H 0.05 mg/mL
40 Endo-CC WT 0.05 mg/mL
Endo-CC WT 0.01 mg/mL
30 Endo-CC N180H 0.01 mg/mL
20
10
0
0 60 120 180 240 300 360 420 480
Time/min
 
Figure 46. Transglycosylation reactions for deglycosylated antibody by Endo-CC and 
Endo-CC N180H enzymes at varied concentrations. The reactions were performed by 
adding 20 molar equivalents of SCT-Oxa to the deglycosylated trastuzumab (10 
mg/mL) in the presence of 0.01 mg/mL and 0.05 mg/mL Endo-CC or Endo-CC 
N180H at 30 ?C in 100 mM Tris-HCl at pH 7.4. The experimental data presented here 
are representative of three independent sets of experiments. 
 145 
 
Transglycosylation Yield (%)
 
C. Conclusion 
The substrate specificity of Endo-CC and mutants for hydrolysis and 
transglycosylation was re-evaluated in this work. The substrate requirement of Endo-
CC for the glycan composition was determined that it could not recognize the core-
fucosylated glycoproteins and glycopeptides for glycan hydrolysis. Furthermore, the 
transglycosylation experiments demonstrated that, besides complex N-glycans, the 
high-mannose and hybrid type glycan oxazolines could also serve as donor substrates 
for glycosynthase mutants of Endo-CC for transglycosylation. Finally, based on the 
previous mutagenesis studies, selected Endo-CC mutants, including N180H, N180Q, 
N180E, N180A and N180M, were tested for their transglycosylation activities. The 
functional assays revealed that all these mutants exhibited different but significant 
levels of abilities to transfer complex type oxazolines to the glycopeptide acceptor. In 
particular, Endo-CC N180A mutant exhibited highly persistent transfer activity over 
time in contrast to the other mutants. Moreover, the enzyme and its transfer mutants 
have been overexpressed with a good yield in E. coli as a CPD fusion protein. This 
study re-visits the substrate selectivity of Endo-CC and its mutants and provides 
important functional information for future work involving applications of Endo-CC 
for glycoengineering of various target glycopeptides and glycoproteins.     
 
 
 
 146 
 
 
D. Materials and Methods 
1. Chemical and biochemical materials  
All reagents, chemicals, and solvents used in this work were purchased from 
Sigma-Aldrich unless otherwise stated. Monoclonal antibody rituximab was 
purchased from Premium Health Services Inc. (Columbia, MD). The sialylated 
complex type glycan oxazoline (SCT-Oxa) was prepared as described in the 
previously published protocol.92 The hybrid type (Hyb) glycan was synthesized by 
sequential glycosylation reaction on a Man5GlcNAc intermediate using enzymes 
including ?1,2-GlcNAc transferase (GnT1), ?1,4-galactosyltransferase and ?2,6-
sialyltransferase.127 The details concerning the synthetic method are summarized 
elsewhere. The high-mannose type (HM) glycan was extracted from soybean flour 
based on a previously described method. All glycan oxazolines were synthesized 
according to the previously detailed one-pot conversion strategy.121 The 
endoglycosidase Endo-S2 used for deglycosylation was expressed and purified 
following our previous publication123. Analytical reverse-phase high-performance 
liquid chromatography (HPLC) was conducted using a Waters 626 HPLC instrument 
connected with an XBridge BEH130 C18 column (3.5 ?m, 4.6 ? 250 mm). The 
elution of glycopeptide was achieved using a linear gradient of acetonitrile (0?30%, 
v/v) in water containing TFA (0.1%) in 35 min at a flow rate of 0.5 mL/min under 
UV wavelength at 214 nm. The high-performance anion-exchange chromatography 
with pulsed amperometric detection (HPAEC-PAD) was performed on a Dionex ICS-
5000 chromatography system (Fischer Scientific) connected to an electrochemical 
detector (ED50) and an anion exchange column (CarboPac PA10 4 ? 250 mm). The 
 147 
 
 
elution on the PA10 column was achieved by a constant solvent containing 100 mM 
NaOH in water with a flow rate of 1.0 mL/min. Liquid chromatography-electrospray 
mass spectrometry (LC-ESI-MS) was used to characterize glycoproteins including 
ribonuclease B and antibody rituximab. MS analysis was conducted using an Exactive 
Plus Orbitrap Mass Spectrometer (Thermo Scientific) and a C18 column (proZap 
Expedite MS C18, 2.1 ? 10 mm, 1.5 ?m, P.J. Cobert Associates, Inc.) for analysis of 
glycopeptides, a C8 column (Poroshell 300SB-C8, 1.0 ? 75 mm, 5 ?m, Agilent) for 
glycoproteins such as RNase B (RNase B) and EPO, and an C-4 column (XBridge 
BEH300 C4, 2.1 ? 50 mm, 3.5 ?m, Waters) for antibody analysis. 
2. Mutagenesis, expression, and purification of recombinant Endo-CC 
The plasmid DNA encoding recombinant Endo-CC from Coprinopsis cinerea 
was synthesized (GenScript) and cloned into pET-41b (+) vector, which contains a 7 
histidine tag at C terminus. For site-directed mutagenesis at Asn-180 residue (N180H, 
N180Q, N180A, N180E, and N180M), a forward primer, 5??TTGGCTGCTC 
NNNGTTGAGATTG-3?, and a reverse primer, 5??CCATCGAATCCACGCTCG-3? 
were used. The site-directed mutagenesis was conducted using a Q5 Site-Directed 
Mutagenesis Kit (NEB, Inc.) Selected mutations were confirmed by DNA sequencing 
(Macrogen, MD). The plasmid was then transformed into E. Coli BL21 competent 
cell (DE3), and a 20 mL LB preculture was inoculated by the transformant containing 
100 ?g/mL of carbenicillin. The preculture was incubated 37?C for 12 h before 
inoculated into a 1L of LB and cultured at 30?C for another 12 h. The cells were 
harvested by centrifugation at 8000 x g for 20 min at 4?C, and the pelleted cells were 
lysed in 20 mL Bacterial Lysis Buffer (Gold Biotechnology, Inc.) following the 
 148 
 
 
manufacturer?s instructions. The His-tagged protein was purified from the centrifuged 
supernatant using HisTrap HP column (GE Healthcare) following the manufacturer's 
protocol. The resultant Endo-CC was buffer-exchanged into 1x PBS (pH 7.4) using 
dialysis tubing (Sigma). The purity of the protein was confirmed by SDS-PAGE and 
the concentration was estimated on a Nanodrop 2000c instrument at the absorbance 
280 nm.  
3. Transglycosylation using wild-type Endo-CC and its mutants 
The transglycosylation reactions catalyzed by each of the wild-type Endo-CC 
and mutants (0.125 mg/mL) was conducted in 50 mM Tris-HCl buffer (pH 7.4, 50 
?L) at 30 ?C using the synthetic CD52-GlcNAc peptide (2 mg/mL) as the starting 
material. Each reaction was initiated by adding 5 molar equivalents of SCT-Oxa, and 
the reaction was sampled at 0, 5, 15, 30, 60, 120, 240, 480, 1080 and 1440 min. For 
each time point, an aliquot of reaction mixture was quenched in 0.1% formic acid and 
frozen at ? 20 ?C. To monitor the reaction in a quantitative manner, each reaction 
mixture was subject to reverse phase HPLC for separation and MS for mass 
confirmation, and the transglycosylation yield was calculated from the integrated 
data. 
4. Transglycosylation by Endo-CC N180H with different N-glycan oxazolines 
Comparative studies on the transglycosylation efficiency of Endo-CC N180H 
against three major types of N-glycan oxazolines were carried out at 30 ?C in Tris-
HCl buffer at pH 7.4. The synthetic glycopeptide CD52-GlcNAc (5 mg/mL) was 
incubated with 7.5 equivalents of each of the glycan oxazolines. The concentration of 
 149 
 
 
Endo-CC N180H was 0.125 mg/mL. The reactions were sampled at 0, 30, 60, 120, 
240 and 480 min, and the quenched reaction was separated by reverse phase HPLC 
followed by MS analysis.  
5. Enzymatic deglycosylation of ribonuclease B 
The mutant enzyme Endo-CC N180H (0.1mg/mL) was incubated with 10 
mg/mL of the native ribonuclease B protein at 37 ?C in PBS buffer at pH 7.4. After 2 
hours of incubation, the enzyme-treated RNase B was subject to LC-ESI-MS analysis 
and the resultant mass spectrum was presented after deconvolution in Results.  
6. Enzymatic deglycosylation of sialylglycopeptide with and without core fucose 
The native sialylglycopeptide (SGP) was extracted from egg yolk power as 
described in the previous publication. The fucosylated form of SGP was synthesized 
by an ?1,6-fucosidase mutant from L.casei (AlfC). The details of the synthesis were 
reported in the previous report.147 During deglycosylation process, an equal amount 
of SGP (+/- core fucose, 50 mg/mL) was incubated with a relatively small amount of 
Endo-CC (0.1 mg/mL) at 37 ?C in PBS buffer at pH 7.4. The progress of hydrolysis 
was monitored by the high-performance anion-exchange chromatography coupled 
with pulsed electrochemical detection (HPAEC-PAD). A major shift of the peak 
representing the starting material (SGP) was detected after 4 hours of incubation, and 
a new glycan peak characteristic of the hydrolyzed SCT free glycan was observed for 
the non-fucosylated SGP, but not the fucosylated SGP. 
 150 
 
 
7. Differential deglycosylation of erythropoietin (EPO) containing a mixture of core-
fucosylated and non-fucosylated N-glycans. 
It has been reported that the core-fucosylated GlcNAc-EPO intermediate can 
be produced using a HEK293S GnT I knockout and FUT8 overexpressing cell line.146 
Using this intermediate, a sialylated glycoform of EPO containing a mixture of 
fucosylated and non-fucosylated HM glycans was prepared to test the substrate 
specificity of Endo-CC. A final concentration of 0.5 mg/mL EPO was treated by 0.01 
mg/mL recombinant Endo-CC. At t = 2 h, the reaction was quenched and injected to 
LC-ESI-MS for characterization.  
8. Differential deglycosylation of antibody glycoforms with core-fucosylated, non-
fucosylated complex type, and high-mannose N-glycans. 
Previous glycan analysis of commercial IgG antibodies has demonstrated that 
nearly all N-glycoforms of rituximab are core-fucosylated. The homogeneous non-
fucosylated S2G2 and HM rituximab glycoforms were generated based on a 
previously published protocol.139 For all three types of rituximab glycoforms, a final 
concentration of 5 mg/mL antibody was incubated with 0.05 mg/mL Endo-CC in PBS 
buffer (pH 7.4) at 37 ?C overnight. The glycosylation patterns of the resultant 
rituximab were analyzed by LC-ESI-MS.  
 
 151 
 
 
9. Differential transglycosylation of deglycosylated antibody by wild-type Endo-CC 
and Endo-CC N180H  
The transglycosylation activity of Endo-CC and its mutant Endo-CC N180H 
at different concentrations was measured by incubating the deglycosylated rituximab 
(10.0 mg/mL, 69 ?M) and the complex-type glycan oxazoline SCT-Oxa (1.38 mM, 
20 equivalents) in 50 mM Tris-HCl (pH 7.4, 20 ?L) at 30 ?C. The concentration of 
the wild-type Endo-CC was fixed at 0.01 mg/mL and 0.05 mg/mL. The concentration 
of the mutant Endo-CC N180H was 0.05 mg/mL. The reaction was sampled at 0, 5, 
15, 30, 60, 120, 240, and 480 min. For each sampled reaction mixture, an aliquot was 
quenched in 0.1% formic acid and subject to LC-ESI-MS analysis. To monitor the 
transglycosylation reaction in a quantitative manner, the reaction yield was estimated 
by LC-MS using an internal standard and a single-point normalization factor. SDS-
PAGE densitometry data were also obtained to support the reaction progression by 
the relative intensity of peaks.   
 
 
 
 
 
 
 
 
 
 152 
 
 
E. Supporting Information 
MPIAGKKFHPRALPEFWRTFREMDEWRATQTGPQARPAEGILKYVPRKIRPA
DIAGKGRLLVSHDYKGGYVEDPFSKSYSFNWWFSTDSFNYFAHHRITIPPPE
WINAAHRQGVPILGTIIFEGGSDEDILRMVIGKTPGSTSNFHAERNAEYTVPVS
SYYAELFADLAVERGFDGWLLNVEIGLQGGSEQARGLAAWVALLQQEVLK
KVGPHGLVIWYDSVTVRGDLWWQDRLNAFNLPFFLNSSGIFTNYWWYNDA
PQKQIDFLSRVDPNLTGQTAEPHQYNLQKTIQDIYIGVDVWGRGSHGGGGFG
AYKAIEHADPKGLGFSVALFAQGWTWETEEEKPGWNWAQFWDYDSKLWV
GPPGVVEAPDHTVKPGEYPCVHGPFQPISSFFLTYPPPDPLDLPFYTNFCPGIG
DAWFVEGKEVFRSETGWTDMDKQTTVGDLVWPRPKIYDLPSQNASQATLN
AAFNFNDAWNGGNSLQINLTVPGGATTYGAYWVPIQTFTFSSRRQYEASIVY
KPGLSGKTRFDAKYEVGIRTITGEDQGKIISNTTTEVGNGWRKVHILFEIETPV
EGGSIIVPSSIGLVIAVSNVSTTEQFEFPFLVGQITIHPHLPDRYKEFKPALLWLL
FTPSAGTNSLDGTLTWDVVAAIERPPPVEINNPDDAQIPWNLQPTKQEWFPDF
LYFNVYVLELLDGGGQGPPQWIGTTGYDGEKKRFFIYDESLPPTSGLRRFTFQ
IEGVLETGESTHWYDAPAAPSATAGGEQKRTRRTSLKSVLSPLRRKKSKGDIS
VAK 
Supplementary Figure 3: amino acid sequence of Endo-CC1 (XP_001839402). 
 
 
 
 153 
 
 
Chapter 5: Glycosylation remodeling of IgE antibody and effects of Fc 
N-glycans on Fc?R binding 
A. Introduction 
Although immunoglobulin E (IgE) represents the least abundant class of 
antibody found in human circulation,148 it provides an important immune defense 
mechanism against venom toxins and parasites.149 IgE is also the major antibody 
responsible for mediating various atopic and allergic disorders such as asthma and 
hayfever.150 The effector functions of IgE to activate allergic responses have attracted 
tremendous attention due to an increasing global onset of asthma and chronic 
allergies in childhood and adults.151 More recently, the unique immunoregulatory 
properties of IgE allowed development of novel IgE-based therapeutic interventions 
against cancers and other disease targets of interest. A first-in-class anticancer 
therapeutic IgE antibody called MOv18 was administered to patients with advanced 
solid tumors in a phase I clinic trail for the first time.152   
 The effector functions of IgE are mainly dependent on interactions with two 
principle IgE-Fc receptors (Fc?Rs), which include the high-affinity Fc?RI and the 
low-affinity Fc?RII/CD23.153 Compared with other antibodies and related receptors, 
IgE exhibited a remarkably high affinity for Fc?RI (KD ~ 0.1-1 nM). Thus, the 
majority of IgE is constantly bound to the surface of mast cells and basophiles, and a 
minimal amount of allergen is sufficient to initiate an acute hypersensitivity response 
by cross-linking the IgE-bound Fc?RIs and subsequently triggering cell 
degranulation.18 Although the affinity of IgE for the monomeric CD23 receptor (KD ~ 
 154 
 
 
1.0 - 10 ?M) is significantly lower than that for the high-affinity Fc?RI, a high avidity 
can be achieved between IgE-Fc and trimeric CD23 via multivalent interactions.154 
CD23 is involved in allergen presentation and transport of the IgE-captured allergen 
complexes, and functional interplay between IgE and CD23 on B cells plays an 
important role in regulating IgE homeostasis in mammalian systems.154-156  
Among all antibody isotypes, IgE is the most heavily glycosylated with seven 
potential N-glycosylation sites located in the heavy chain, and there is a considerable 
level of heterogeneity in the N-glycosylation patterns of IgE in terms of both the 
occupancy of the glycosylation sites and the specific glycan compositions.50, 53, 157 
Among the potential glycosylation sites of IgE, the Asn-394 site is highly conserved 
and is structurally homologous to the Asn-297 residue of IgG antibodies.158 Unlike 
other glycosylation sites located in the C?2 and C?3 domains, which all contain 
complex type N-glycans, a unique oligomannose N-glycan is associated with the Asn-
394 residue in the IgE Fc fragment (Figure 46). Structural analysis on IgE-Fc 
homodimers suggest that these heterogeneous high-mannose glycans at Asn-394 are 
involved in extensive non-covalent interactions with the C?3 and C?4 domains and 
may stabilize the two heavy chains by making contacts with each other.159-161 In 
contrast to IgG in which the complex N-glycan at Asn-294 is critical for binding to 
Fc?Rs and biological functions,25 it is less clear whether the N-glycosylation of IgE 
affects its binding activity towards the Fc?RI and CD23 receptors. Previous studies 
indicated that the non-glycosylated IgE-Fc fragments produced from bacterial and 
mammalian systems were able to bind IgE receptors with comparable affinities to the 
native IgE-Fc.162, 163 However, a recent study demonstrated that the single 
 155 
 
 
oligomannose N-glycan at Asn-394 is indispensable for IgE to elicit its biological 
activities in mediating anaphylaxis in vitro and in vivo.113 
To improve our understanding about the effect of N-glycosylation on IgE 
biology, we here describe a combined chemoenzymatic strategy to remodel the N-
glycosylation profile of IgE and the IgE-Fc fragment in a site-specific manner. For 
full-length IgE antibody, the oligomannose N-glycans were removed by Endo-H, an 
endoglycosidase that specifically releases the mannose-rich N-glycans but does not 
alter the fucosylated complex type glycan structures. Thus, other complex type N-
glycans remained unaffected and analysis on this partially deglycosylated IgE 
revealed a 2-4-fold reduction in binding affinity to the high-affinity IgE-Fc receptor 
Fc?RI, which may lead to more significant impact on the biological functions of IgE 
in vivo.  
According to the previous studies on IgE glycosylation, only the Asn-394 N-
glycan of IgE was found to be necessary for receptor binding and biological 
functions.113 Therefore, we generated a mutant IgE-Fc that lacks all N-glycosylation 
sites except for Asn-394 by the site-directed mutagenesis. This IgE-Fc variant known 
as IgE-Fc-Asn394only was then expressed in mammalian cells and the binding assays 
confirmed that the intact IgE and IgE-Fc-Asn394only exhibited similar levels of 
binding affinity for the immobilized Fc?RI. Using this IgE-Fc variant as a model 
protein, we applied a combined method of mammalian expression and enzymatic 
modifications to generate a panel of IgE-Fc N-glycoforms and analyzed their binding 
affinities to Fc?RI and CD23. The results clearly suggested that the deglycosylated 
IgE-Fc fragment showed a 4-fold reduction in its binding affinity to Fc?RI, whereas 
 156 
 
 
other oligomannose IgE-Fc glycoforms retained the same affinities as the native IgE-
Fc glycoform. Similarly, the IgE-Fc/CD23 interactions were also affected by the 
glycoengineering, and, like Fc?RI, a 2-3-fold decrease of IgE-Fc binding to CD23 
was observed upon the removal of the Asn-394 N-glycans. Finally, following a highly 
efficient chemoenzymatic glycan remodeling strategy developed by Wang and 
others,103 a complex N-glycan was transferred to the Asn-394 residue of IgE-Fc using 
IgE-Fc-Asn394only-GlcNAc as glycosyl acceptor and chemically activated glycosyl 
oxazoline (G2-CT-Oxa) as the donor substrate. The effective transglycosylation was 
catalyzed by a recently reported glycosynthase mutant Endo-CC N180H.105 The 
successful glycosylation of a complex glycan to IgE altered its binding properties to 
the high-affinity receptor Fc?RI, and this study clearly demonstrated the possibility of 
further glycosylation remodeling of IgE antibody with additional natural or artificial 
N-glycan structures for future structural and functional studies. 
The development of mammalian expression system for IgE-Fc glycoforms 
was assisted by Dr. Qiang Yang from Prof. Lai-Xi?s group at the Department of 
Chemistry and Biochemistry, University of Maryland, College Park. The enzymatic 
treatments of IgE-Fc N-glycoforms by multiple mannosidases were conducted by Xin 
Tong and Dr. Yang together. The mass spectrometry analysis of the released N-
glycans from IgE-Fc N-glycoforms was assisted by Dr. Zong and other co-authors 
involved in this study. 
 157 
 
 
B. Results and Discussion 
1. Selective glycosylation remodeling of commercial and recombinant full-length 
IgE antibodies and binding analyses on the product IgE N-glycoforms 
The full-length human IgE contains seven conserved N-glycosylation sites in 
its primary sequence.50 Among all potential sites, only the Asn-394 site is exclusively 
glycosylated with the oligomannose N-glycans, whereas other sites can be either 
unoccupied or associated exclusively with the complex type oligosaccahrides.157 
Interestingly, this unique oligomannose glycan in the C?3 domain of IgE was found 
by Shade and colleagues to be obligatory for Fc?RI receptor binding and the ability of 
IgE to promote anaphylaxis.113 However, in this study, binding between IgE and 
Fc?RI was analyzed in a cell-based functional assay using the fluorescence-activated 
cell sorting technique (FACS), and information regarding the in vitro binding kinetics 
of the glycoengineered full-length IgE antibodies with Fc?RI is lacking. 
To evaluate the functional effects of the previously studied N-glycans at the 
Asn-394 site, we first performed site-specific deglycosylation at Asn-394 of two full-
length IgE antibodies, a commercially available human IgE (Abcam-hIgE, Abcam, 
UK) and a recombinant ovalbumin (OVA)-specific human IgE antibody (OVA-hIgE). 
Ovalbumin is a well-established model allergen extracted from chicken eggs and has 
been frequently used to artificially induce the acute IgE-mediated pulmonary allergic 
response in experimental animals.153 The biological activity of the anti-OVA IgE to 
induce asthma are often studied using the OVA challenge model system.164 To 
selectively remove this oligomannose glycan from IgE without affecting the other 
complex N-glycans, the antibodies were treated with Endo-H, an endoglycosidase 
 158 
 
 
that specifically cleaves between the two GlcNAc residues in the diacetylchitobiose 
core of the high-mannose and hybrid N-linked oligosaccharides with no hydrolytic 
activities against the bystander complex N-glycans.165 Treatment of the intact IgE 
antibodies by Endo-H led to the release of the oligomannose N-glycans from IgE 
antibodies, as confirmed by sodium dodecyl sulfate polyacrylamide gel 
electrophoresis (SDS-PAGE) analysis (Figure 47).  
To assess the change in the binding affinity between IgE-Fc and Fc?RI caused 
by partial deglycosylation, we covalently immobilized the soluble domain of the 
high-affinity IgE receptor	
 ?(Fc?RIa) on a standard CM5 sensor chip, and then injected 
the site-selectively deglycosylated IgE antibodies through the channel to monitor the 
specific binding events between IgE and Fc?RIa using the Surface Plasmon 
Resonance (SPR) technique (Figure 48). As illustrated in Figure 48 and Table 8, for 
both the commercial and recombinant IgE antibodies, a roughly 4-fold reduction in 
binding affinity to the high-affinity receptor Fc?RI was observed after the 
oligomannose N-glycan was removed. In addition, all antibodies displayed binding 
affinities within the nM range to Fc?RI, which were consistent with the previous 
reported affinities of both the full-length IgE antibodies and IgE-Fc fragments to the 
high-affinity receptor.166 
In the previous study by Shade and others, the human IgE was treated with an 
alternative endoglycosidase F1 (EndoF1) that selectively removes the oligomannose 
N-glycans without affecting the other complex N-glycans.113 Their results indicated 
that Endo-F1 treatment abolished the IgE interactions with Fc?RI in cell-based assays 
using flow cytometry and saturation binding analysis. Comparing the two studies, it 
 159 
 
 
remains unclear why selective removal of the oligomannose N-glycan by Endo-H 
showed a less dramatic impact on Fc?RI binding in our in vitro experiments in 
comparison with those shown in the previous cell-based assays. It is worth noting that, 
in the previous publication, the IgE antibodies were treated with Endo-F1 at 37 ?C for 
72 hours prior to the cell-based binding assays. Whether a prolonged incubation of 
IgE at certain temperature could cause instability of antibody and loss of binding 
activity remains a question to be answered. 
 
 
Figure 47. Site-selective N-glycosylation remodeling of full-length human IgE 
antibodies using endoglycosidase Endo-H. The oligosaccharide composition of all 
occupied N-glycosylation sites are based on the previously published site-specific 
glycosylation analysis on IgE antibodies.50 Only the previously described Asn-394 
site contains an oligomannose N-glycan, whereas all the other N-linked carbohydrate 
chains attached to IgE are complex type N-glycans. The position of the key N-
glycosylation site Asn-394 was highlighted.  
 160 
 
 
(A)$Full(length$IgE$an1body$(Abcam)$$ (B)$Selec1vely$Deglycosylated$IgE$(Abcam)$
RU RU
1600 1600
1000 1000
400 400
-200 -200
-100 0 100 200 300 400 500 600 700 800 900 -100 0 100 200 300 400 500 600 700 800 900
Time s Time s
(C)$Full(length$An1(OVA$IgE$an1body$ (D)$Selec1vely$Deglycosylated$An1(OVA$IgE$
RU RU
1000 1000
600 600
200 200
-200 -200
-100 0 100 200 300 400 500 600 700 800 900 -100 0 100 200 300 400 500 600 700 800 900
Time s Time s 
Figure 48. Surface plasmon resonance (SPR) representative sensorgrams of binding 
between full-length IgE antibodies and immobilized soluble domain of high-affinity 
IgE receptor Fc?RI., (A) commercial IgE (Abcam, UK) and (B) selectively 
deglycosylated commercial IgE (IgE-Asn394-GlcNAc) (C) recombinant anti-OVA 
IgE, and (D) selectively deglycosylated recombinant anti-OVA IgE (IgE-Asn394-
GlcNAc). Eight concentrations of IgE antibodies (400 ? 3.125 nM) were tested in the 
standard SPR buffer, and a global monovalent fitting was applied to determine the 
kinetic parameters shown in Table 8. 
 
 
 
 
 
 
 161 
 
Response Response
Response Response
 
Table 8. Kinetic parameters and binding affinities of commercial IgE antibody and 
recombinant anti-OVA IgE binding to the immobilized Fc?RI? obtained from SPR 
analysisa 
Native Full-length IgE k a (1/Ms) k d (1/s) K D (M)
b
IgE (Abcam) (1.35	
 ??	
 ?0.07)	
 ?x	
 ?10+5 (2.41	
 ??	
 ?0.02)	
 ?x	
 ?10-??4 (1.78	
 ??	
 ?0.21)	
 ?x	
 ?10-??9
IgE-Asn394-GlcNAcc (0.78	
 ??	
 ?0.05)	
 ?x	
 ?10+5 (5.13	
 ??	
 ?0.04)	
 ?x	
 ?10-??4 (7.81	
 ??	
 ?0.17)	
 ?x	
 ?10-??9
Anti-OVA IgE (1.20	
 ??	
 ?0.04)	
 ?x	
 ?10+5 (2.96	
 ??	
 ?0.01)	
 ?x	
 ?10-??4 (2.46	
 ??	
 ?0.32)	
 ?x	
 ?10-??9
IgE-Asn394-GlcNAcd (0.91	
 ??	
 ?0.01)	
 ?x	
 ?10+5 (8.71	
 ??	
 ?0.02)	
 ?x	
 ?10-??4 (9.57	
 ??	
 ?0.09)	
 ?x	
 ?10-??9  
aKinetics for binding between native IgE antibodies and Fc?RI? were measured in 
Standard SPR buffer at 20 ?C and the obtained data were analyzed using global 
monovalent fitting model (BiaEvaluation). The parameters are average ? SD of 
experiments conducted in triplicate. The binding activity of the immobilized Fc?RI? 
was confirmed by standard IgE binding assay prior to each of the repeated binding 
experiments to ensure that the CM5 chip-immobilized Fc?RI? did not lose activity 
during storage.  
bThe dissociation constant KD was deduced as (kd/ka).  
IgE-Asn-394-GlcNAcc represents the selectively deglycosylated full-length IgE N-
glycoform generated from the commercial human native IgE (Abcam, UK), where 
only the oligomannose N-glycan at Asn-394 was removed by Endo-H treatment. 
IgE-Asn-394-GlcNAcd is the selectively deglycosylated full-length IgE N-glycoform 
derived from the recombinant anti-OVA IgE antibody, and the Asn-394 site contains 
only the innermost GlcNAc residue after the Endo-H-mediated hydrolysis.  
 
 
 162 
 
 
2. Generation of mutant IgE-Fc containing only one oligomannose glycosylation site 
at Asn-394 with comparable Fc?RI-binding capacity to full-length IgE antibodies. 
The ability of IgE to mediate allergic responses are primarily regulated 
through interactions with IgE receptors Fc?RI and Fc?RII/CD23.150 Structural studies 
and functional assays revealed that the Fc (crystallizable) domain of IgE consisting of 
C?2, C?3, and C?4 domains alone was sufficient to bind both IgE receptors with 
comparable affinities as full-length IgE antibodies.51 Furthermore, as described in a 
previous publication,113 only a single glycan at Asn-394 in the C?3 domain of IgE-Fc 
was required for IgE to interact with Fc?RI and to promote allergic reactions.113 
Based on these studies, we designed a construct of the IgE-Fc fragment selectively 
lacking all the conserved N-glycosylation sites except for Asn-394 by mutating 
asparagine (N) to glutamine (Q). This unique human IgE-Fc fragment (IgE-Fc-
Asn394only) was successfully expressed in HEK294 cell with only one glycosylation 
site preserved at the Asn394 residue of the C?3 domain. Mass spectrometry analysis 
on released glycans from this IgE-Fc antibody fragment confirmed that the majority 
of the N-linked glycans found were oligomannose oligosaccharides containing five (-
Man5, majority) to nine (-Man9, minority) mannose residues (Figure 50A), which was 
consistent with reported site-specific analysis of human IgE N-glycosylation.50 After 
confirming that this IgE-Fc variant (IgE-Fc-Asn394only-HM) bound to Fc?RI with 
similar kinetics and affinities to full-length IgE antibodies (Table 9), we used this 
IgE-Fc-Asn394only N-glycoform as starting material for further glycoengineering to 
study the impact of N-glycosylation remodeling on IgE binding with both the high-
affinity IgE receptor Fc?RI and the low-affinity IgE receptor Fc?RII/CD23.  
 163 
 
 
3. Generation of homogeneous IgE-Fc oligomannose glycoforms using enzymatic 
modifications and biosynthetic engineering and their binding properties to Fc?RI.  
To further dissect the functional role of oligosaccharide compositions at Asn-
394 of IgE-Fc, a panel of IgE-Fc-Asn394only N-glycoforms were generated through 
enzymatic modifications and biosynthetic inhibitor (Figure 49). The identity of the 
modified N-glycan associated with each IgE-Fc glycoform was confirmed by matrix-
assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometry 
analysis on the released glycans from IgE after treatment with amidase enzyme 
peptide-N-glycosidase F (PNGase F).  
 
 
 164 
 
 
Figure 49. Glycan remodeling of heterogeneous high-mannose N-glycoforms of IgE-
Fc-Asn394only-HM (Man5-9) using a combination of glycan remodeling strategies, 
including ?1-2 and ?1-2, 3, 6 mannosidases, endoglycosidase (Endo-H), peptide-N-
glycosidase (PNGase F), and biosynthetic N-glycan inhibitor kifunensine. For the N-
glycoforms that contain a mixture of glycans such as IgE-Fc-Asn394only-HM and 
IgE-Fc-Asn394only-Man4-7, only the most abundant species is illustrated here. 
 
3.1.Deglycosylation of IgE-Fc-Asn394only-HM N-glycoform by Endo-H reduced the 
Fc?RI-binding affinity to the same extent as for full-length IgE antibodies 
As demonstrated on the full-length IgE antibodies, the IgE-Fc-Asn394only-
HM N-glycoform was treated with the high-mannose-specific Endo-H to remove the 
native oligomannose glycans to afford deglycosylated IgE-Fc-Asn394only-GlcNAc 
N-glycoform, which was subsequently tested for its binding affinity to Fc?RI using 
the same conditions as in the previous binding studies. Interestingly, for the high-
affinity IgE receptor Fc?RI, the binding kinetics of IgE-Fc-Asn394only-HM and IgE-
Fc-Asn394only-GlcNAc displayed the same patterns as for the full-length glycoforms 
(Table 8 and Table 9), where the removal of the high-mannose N-glycans caused an 
approximately 4-fold reduction in binding affinity to Fc?RI. In addition, binding of 
IgE-Fc-Asn394only-HM were shown to be comparable to full-length IgE antibodies, 
which confirmed that IgE-Fc carrying only a single N-glycan at Asn-397 was 
sufficient to provide high-affinity binding to Fc?RI. Therefore, it would be plausible 
to use our mutant IgE-Fc fragment as model protein for following glycoengineering 
and functional studies. 
 165 
 
 
 
Figure 50. MALDI-TOF mass spectrometry identification of the released N-glycans 
from the IgE-Fc-Asn394only N-glycoforms generated through glycan remodeling: (A) 
IgE-Fc-Asn394only-HM produced under native expression conditions in HEK 293T 
cell line; (B) IgE-Fc-Asn394only-Man4-7 generated by the ?1-2, 3, 6 mannosidase 
processing; (C) IgE-Fc-Asn394only-Man9 expressed under the effect of 40 ?M of 
kifunensine in HEK 293T; (D) IgE-Fc-Asn394only-Man5 generated by the ?1-2 
mannosidase treatment. 
 
 166 
 
 
3.2. Biosynthetic glycoengineering of IgE-Fc-Asn394only-HM glycoform to generate 
homogeneous IgE-Fc-Asn394only-Man9.  
In parallel to enzymatic modification, the IgE-Fc-Asn394only fragment was 
expressed in a HEK293-derived cell line in the presence of mannosidase inhibitor 
kifunensine to minimize the processing of oligomannose N-glycans, which gave rise 
to a homogeneous high-mannose N-glycoform containing a total of nine mannose 
residues, IgE-Fc-Asn394only-Man9. Kifunensine is a monosaccharide-based inhibitor 
derived from plant alkaloid. This mannose analogue serves as a competitive inhibitor 
for Type-1 endoplasmic reticulum (ER) and Golgi ?-mannosidases during early stage 
of N-glycan biosynthesis and it can effectively disrupt the normal processing of 
Man9GlcNAc2 intermediate to Man5GlcNAc .1462  As a result, a high concentration of 
kifunensine could efficiently control the processing of N-glycans during biosynthesis 
and produce recombinant glycoproteins with homogeneous high-mannose N-glycans. 
Indeed, MALDI-TOF analysis revealed a single peak on mass spectrum 
corresponding to the Man9GlcNAc2 glycan (Figure 50C). Binding analysis further 
confirmed that the affinities of IgE-Fc-Asn394only-Man9 (KD = 2.13 ? 0.09 nM) and 
IgE-Fc-Asn394only-HM (KD = 2.41 ? 0.12 nM) to Fc?RI were nearly identical, 
which suggested that addition of mannose residues to the native high-mannose N-
glycans (Man5GlcNAc2 as the major N-glycoform) did not significantly alter the 
binding affinity between IgE-Fc and Fc?RI.  
 
 
 167 
 
 
3.3. Enzymatic modifications on IgE-Fc-Asn394only-HM to generate additional IgE-
Fc-Asn394only N-glycoforms. 
In order to remove the mannose residues off the IgE-Fc-Asn394only-HM N-
glycoform and to afford the homogeneous Man5GlcNAc2 N-glycan, the expressed 
IgE-Fc fragment was treated with ?1-2 mannosidase to remove all ?1-2 linked 
mannoses from the high-mannose N-glycans without affecting the ?1-3 and ?1-6 
linked mannoses, as shown in Figure 51. The glycan analysis confirmed that the 
released N-glycan from the resultant IgE-Fc-Asn394only-Man5 glycoform displayed 
the expected mass values according to the MALDI-TOF analysis (Figure 50D).   
To further trim IgE-Fc-Asn394only-Man5 to paucimannose N-glycoforms, 
which contain mainly Man3?4GlcNAc2 N-glycans, the starting Man5 N-glycoform was 
incubated with an alternative ?1-2, 3, 6 mannosidase harvested from the Jack Bean 
(Canavalia ensiformis) seeds, which possesses a broad substrate selectivity than the 
?1-2 mannosidase for all terminal ?1-2, ?1-3 and ?1-6 linked mannose residues of N-
glycans (Figure 51). Treatment of the IgE-Fc-Asn394only-Man5 glycoform by this 
mannosidase provided a paucimannose N-glycoform of IgE-Fc, which contained a 
mixture of Man4-7GlcNAc2 N-glycans, and the Man4GlcNAc2 N-glycan was found to 
be the most abundant species among the released N-glycans, as demonstrated by 
MALDI-TOF analysis. The resistance of the innermost ?1-4 and ?1-6 mannoses 
(Man1-3GlcNAc2) to enzymatic treatment was suggested in a previous structural study 
concerning N-linked carbohydrates of human IgE antibodies. The X-ray 
crystallographic structures indicated that the high-mannose N-glycans of IgE-Fc at 
Asn-394 was buried between the two C?3 domains and the adjacent C?2 domains 
 168 
 
 
were found to be blocking access of the intracellular mannosidases during IgE N-
glycan biosynthesis.155 This phenomenon was also observed for high-mannose N-
glycans of IgY-Fc found in avian species.167 Nevertheless, the kinetic parameters 
measured by SPR analysis (Table 9) suggested that the Man1-3GlcNAc2 core structure 
alone was able to maintain the same level of Fc?RI binding capacity as the original 
IgE-Fc-Asn394only-HM glycoform, since IgE-Fc-Asn394only-Man9, IgE-Fc-
Asn394only-Man5, IgE-Fc-Asn394only-Man4-7, and IgE-Fc-Asn394only-HM all 
exhibited similar binding affinities to the immobilized Fc?RI (~ 2 nM). Because only 
IgE-Fc-Asn394only-GlcNAc showed significantly reduced binding affinity (~ 9 nM) 
to Fc?RI, the innermost Man1-3GlcNAc2 glycan core is suggested to be the minimal 
structural requirement for the optimal binding of IgE-Fc fragment to Fc?RI. 
 
 
Figure 51. Schematic representation of the substrate specificity of ?1-2 and ?1-2, 3, 6 
mannosidase against differentially linked mannose residues of the high-mannose Asn-
linked oligosaccharides in glycoproteins. This figure is created based on the reported 
glycosidic linkage selectivities of commercial ?1-2 and ?1-2, 3, 6 mannosidase.168 
 
 
 
 
 169 
 
 
3.4. Complete removal of N-glycans from the native IgE-Fc glycoform by PNGase F 
caused further reduction of Fc?RI binding affinity. 
Finally, the native high-mannose type N-glycans of IgE-Fc-Asn394only were 
completely released from the conserved asparagine residue using a bacterial 
endoglycosidase-like amidase enzyme called peptide-N-glycosidase F (PNGase F). 
Unlike endoglycosidase, PNGase F catalyzes the cleavage between the asparagine 
residue and the innermost GlcNAc of all types of N-linked oligosaccharides and 
deaminiates the asparagine residue to an aspartate in the process.129 To maintain the 
structural integrity of the resultant IgE-Fc protein, the deglycosylation reaction was 
performed under the non-denaturing condition, and other glycoproteins including 
ribonuclease B (RNase B) and IgG antibody were also treated with PNGase F to 
provide side-by-side controls for the reaction.94 A complete deglycosylation of IgE-
Fc-Asn394only-HM was confirmed by SDS-PAGE, where a clear shift of protein 
bands was observed after PNGase F treatment.  
Similar to all the other IgE-Fc-Asn394only N-glycoforms, the binding affinity 
of resulting IgE-Fc-Asp394 protein to Fc?RI was measured and compared with other 
IgE-Fc N-glycoforms. The comparative data (Figure 52 and Table 9) clearly revealed 
a further reduction of IgE-Fc-Asp394 binding to Fc?RI [KD = (19.1 ? 0.31) ?10-9 M-1] 
in comparison to the Endo-H-treated IgE-Fc-Asn394only-GlcNAc form [KD = (8.64 
? 0.09) ?10-9 M-1]. These results may also suggest a functional role of the Man1-
3GlcNAc2 core structure for IgE-Fc binding to Fc?RI. However, it is important to take 
into consideration the fact that PNGase F treatment introduces an Asn-to-Asp 
mutation at the N-glycosylation site (Asn394) due to deamination. It remains unclear 
 170 
 
 
whether the decreased binding affinity could also be caused by the rearrangement of 
the local secondary structure of IgE as a result of glycan remodeling. Indeed, a 
previous study on murine IgG antibodies demonstrated that deglycosylation of the 
homologous complex type N-glycan from IgG-Fc at Asn-297 caused a significant 
conformational change that negatively impacts Fc?R binding.169 Furthermore, Shade 
and colleagues also investigated the contribution of Asn-394 N-glycans to the 
secondary structure of human IgE by circular dichroism (CD).113 In their study, a 
relatively small but significant shift in CD spectra was observed for Endo-F1-treated 
IgE antibody, which could indicate delicate changes in the overall conformation of 
human IgE causing variations in receptor binding properties. Therefore, further 
structural analysis such as CD and X-ray crystallography on the remodeled IgE-Fc N-
glycoforms described in this study are needed to shed light on the molecular 
mechanism behind the altered Fc?RI binding capacity due to N-glycoengineering. 
 
 
 
 171 
 
 
 
Figure 52. Surface plasmon resonance (SPR) representative sensorgrams of binding 
events analyzed between remodeled IgE-Fc-Asn394only glycoforms and immobilized 
soluble domain of high-affinity IgE receptor Fc?RI, (A) the native IgE-Fc-??
Asn394only-HM containing a mixture of Man5-9 oligomannose N-glycans, (B) Endo-
H-treated IgE-Fc-??Asn394only-GlcNAc, (C) IgE-Fc-??Asn394only-Man9, (D) IgE-Fc-
Asn394only-Man5, (E) IgE-Fc-Asn394only-Man4-7, and (F) PNGase F-treated IgE-
Fc-Asp394. Six concentrations of IgE antibodies (100?3.125 nM) were measured in 
the standard SPR buffer, and a global monovalent fitting was applied to determine the 
kinetic parameters shown in Table 9. 
 172 
 
 
Table 9. Kinetic parameters and binding affinities of IgE-Fc N-glycoforms for 
Fc?RI.a  
IgE-Fc N-glycoforms k (1/Ms) k (1/s) K (M)ba d D
IgE-Fc-Asn394only-HM (1.01	
 ??	
 ?0.01)	
 ?x	
 ?10+5 (2.79	
 ??	
 ?0.03)	
 ?x	
 ?10-??4 (2.76	
 ??	
 ?0.15)	
 ?x	
 ?10-??9
IgE-Fc-??Asn394only-Man +59 (1.75	
 ??	
 ?0.05)	
 ?x	
 ?10 (4.08	
 ??	
 ?0.04)	
 ?x	
 ?10-??4 (2.33	
 ??	
 ?0.09)	
 ?x	
 ?10-??9
IgE-Fc-Asn394only-Man5 (1.35	
 ??	
 ?0.04)	
 ?x	
 ?10+5 (3.46	
 ??	
 ?0.02)	
 ?x	
 ?10-??4 (2.56	
 ??	
 ?0.13)	
 ?x	
 ?10-??9
IgE-Fc-Asn394only-Man (1.57	
 ??	
 ?0.01)	
 ?x	
 ?10+54-7 (3.78	
 ??	
 ?0.02)	
 ?x	
 ?10-??4 (2.41	
 ??	
 ?0.21)	
 ?x	
 ?10-??9
IgE-Fc-??Asn394only-GlcNAc (1.05	
 ??	
 ?0.01)	
 ?x	
 ?10+5 (9.65	
 ??	
 ?0.02)	
 ?x	
 ?10-??4 (9.19	
 ??	
 ?0.18)	
 ?x	
 ?10-??9
IgE-Fc-Asp394only (5.21	
 ??	
 ?0.09)	
 ?x	
 ?10+4 (9.91	
 ??	
 ?0.11)	
 ?x	
 ?10-??4 (1.90	
 ??	
 ?0.32)	
 ?x	
 ?10-??8  
aThe data were measured by SPR from a panel of glycoengineered IgE-Fc glycoforms 
containing a single N-glycosylation site at Asn-394. These IgE-Fc glycoforms were 
generated by enzymatic modifications and biosynthetic glycan manipulation. Kinetic 
parameters for binding between Fc?RI? and glycoengineered IgE-Fc-Asn394only N-
glycoforms were obtained in Standard SPR buffer at 20 ?C and the obtained raw data 
were then analyzed using global monovalent 1 : 1 fitting model (BiaEvaluation). The 
parameters were means ? SD of experiments conducted in triplicate and the optimal 
binding capacity of the immobilized Fc?RI? domain was confirmed by standard IgE 
binding assay prior to each of the repeated binding experiments to ensure that the 
CM5 chip-immobilized Fc?RI? did not loss activity during storage.  
bThe dissociation constant KD was defined as (kd/ka) and calculated using the 
monovalent fitting model of BiaEvaluation software.  
 
 
 
 
 
 173 
 
 
4. Glycosylation remodeling of IgE-Fc-Asn394only showed no significant impact on 
binding capacity to low-affinity immunoglobulin E (IgE) receptor, CD23 (Fc?RII) 
While most studies regarding IgE glycosylation and receptor binding focus on 
the high-affinity IgE receptor, Fc?RI, there is limited discussion about the potential 
effects of N-glycans on the ability of IgE antibodies to interact with the low-affinity 
IgE receptor, CD23/Fc?RII. Because the CD23 signaling is critical for many allergy-
associated immune functions, including maintaining the IgE homeostasis, facilitating 
allergen recognition, and transporting allergens across airway and gut membranes51, 
interaction between IgE-Fc and CD23 has become a promising target for therapeutic 
intervention. An anti-CD23 antibody (lumiliximab) was developed as an anti-asthma 
therapy and exhibited the ability to effectively reduce the IgE level in human serum170. 
Regarding the functional relationship between IgE-Fc N-glycosylation and CD23-
binding capacity, an early study by Gould and others demonstrated that glycosylation 
was not necessary for IgE to bind B cells through CD23,171 as the enzymatically 
deglycosylated IgE-Fc ?-domain fragments displayed comparable or even increased 
binding to CD23.  However, only a series of recombinant ?-chain peptide fragments 
expressed in bacteria instead of mammalian systems were tested in their study and no 
detailed glycan analysis was performed on the resulting IgE-Fc fragments during the 
deglycosylation process. Given the controversial results about the exact glycosylation 
requirement for human IgE to bind Fc?RI and trigger allergic response, as discussed 
in previous sections, it is of great interest to investigate whether any alterations in the 
N-glycosylation profile of the IgE-Fc fragments produced in mammalian systems can 
also affect binding to CD23 and the associated B cell signaling pathways. 
 174 
 
 
Using the homogeneous IgE-Fc-Asn394only N-glycoforms described in the 
previous section, including IgE-Fc-??Asn394only-Man9, IgE-Fc-Asn394only-Man5, 
IgE-Fc-??Asn394only-GlcNAc, and IgE-Fc-Asp394, we compared the binding 
properties of these four IgE-Fc glycoforms to the monomeric soluble domain of 
CD23 immobilized on sensor surface using a similar experimental setup as for Fc?RIa. 
The representative sensorgrams and their binding kinetics are summarized in Figure 
53 and Table 10. Analogous to the results from Fc?RIa binding, comparison between 
binding kinetics of oligomannose (-Man9) and deglycosylated IgE N-glycoforms (-
GlcNAc and PNGase F-treated -Asp394) revealed an approximately 2-fold difference 
in binding affinities to CD23, which was mainly due to higher association rates of 
high-mannose N-glycoforms. As shown in Table 10, the IgE-Fc-??Asn394only-Man9 
and IgE-Fc-Asn394only-Man5 associates with CD23 faster than both IgE-Fc-??
Asn394only-GlcNAc and IgE-Fc-Asp394, which are supported by higher association 
constants [k = (6.91 ? 0.09) ? 104 Ms-1a  for IgE-Fc-??Asn394only-Man9 and ka = (2.35 ? 
0.01) ? 104 Ms-1 for IgE-Fc-Asp394]. A small difference in CD23 binding affinities 
was observed between the two mannose-rich N-glycoforms, including IgE-Fc-??
Asn394only-Man  [K -8 -19 D = (3.14 ? 0.08) ? 10  M ] and IgE-Fc-Asn394only-Man5 [KD 
= (5.89 ? 0.15) ? 10-8 M-1], which were due to variations in both ka and kd. Unlike 
previous studies regarding Fc?RI binding, no further reduction of IgE-Fc binding 
toward CD23 was observed after the complete removal of the N-glycans by PNGase 
F, as binding kinetics and affinities of IgE-Fc-??Asn394only-GlcNAc and IgE-Fc-
Asp394 were comparable [KD = (6.76 ? 0.07) ? 10-8 M-1 for IgE-Fc-??Asn394only-
GlcNAc and KD = (7.95 ? 0.14) ? 10-8 M-1 for IgE-Fc-Asp394].  
 175 
 
 
Although low-affinity IgE receptor CD23 is structurally categorized as a lectin, 
which represents a superfamily of carbohydrate-binding proteins, early structural 
study demonstrated that CD23 recognized IgE-Fc through protein-protein interactions 
that did not involve any carbohydrate determinants171. Adding to this point, a separate 
structural analysis on possible CD23/carbohydrate interactions suggested that soluble 
domain of CD23 did not show any carbohydrate-specific binding activities172. In this 
study, the abilities of simple carbohydrates such as glucose, galactose, lactose, 
mannose, and N-acetylglucosamine to bind to recombinant CD23 protein were tested 
using NMR chemical shift perturbation titrations. The results showed that none of 
these carbohydrates exhibited any specific binding to CD23 even at the highest 
concentrations. A recently published crystal structure of IgE-Fc in complex with 
CD23 showed that IgE binding sites for CD23 and Fc?RI were located at the opposite 
faces of C?3 domain and the Asn-394 N-glycan was found to be in proximity to 
Fc?RI binding site but not to the CD23 binding site, which suggested that a change of 
IgE glycosylation may only impact CD23 binding through long-range interactions or 
allosteric effects154. Interestingly, comparison of the IgE-Fc structures in complex 
with either CD23 or Fc?RI demonstrated that binding events with the two receptors 
displayed an allosteric effect and were mutually exclusive173. Since several structural 
studies showed that removal of Asn-394 N-glycans did result in small conformational 
changes to the IgE-Fc domain and reduced its binding to Fc?RI, the same molecular 
mechanism may also contribute to the observed decrease of CD23-binding activity 
caused by deglycosylation. Nevertheless, similar to the previous studies on Fc?RI, 
 176 
 
 
further structural and biophysical studies on IgE-Fc N-glycoforms in complex with 
CD23 are needed to help explain the varied binding kinetics observed in this study. 
 
 
Figure 53. Representative surface plasmon resonance (SPR) sensorgrams of 
interactions between the homogeneous IgE-Fc-Asn394only N-glycoforms and the 
immobilized soluble domain of the IgE-Fc low-affinity receptor CD23/ Fc?RII, (A) 
IgE-Fc-??Asn394only-Man9, (B) IgE-Fc-Asn394only-Man5, (C) Endo-H-released IgE-
Fc-??Asn394only-GlcNAc, and (D) PNGase F-released IgE-Fc-Asp394. A total of eight 
IgE-Fc concentrations (400 ? 3.125 nM) were tested in the standard SPR buffer at 
20 ?C, and a global monovalent fitting model was used to measure the kinetic 
parameters shown in Table 10. 
 
 
 177 
 
 
Table 10. Kinetic parameters of IgE-Fc-Asn394only glycoforms with the 
immobilized soluble domain of IgE-Fc low-affinity receptor CD23 measured by SPRa. 
IgE-Fc N-glycoforms ka(1/Ms) kd(1/s) KD(M)b
IgE-Fc!Asn394only-Man9 (6.91'?'0.09)'x'10+4 (2.17'?'0.01)'x'10!3 (3.14'?'0.08)'x'10!8
IgE-Fc-Asn394only-Man5 (5.78'?'0.04)'x'10+4 (3.40'?'0.05)'x'10!3 (5.89'?'0.15)'x'10!8
IgE-Fc!Asn394only-GlcNAc (2.86'?'0.01)'x'10+4 (1.93'?'0.02)'x'10!3 (6.76'?'0.07)'x'10!8
IgE-Fc-Asp394only (2.35'?'0.11)'x'10+4 (1.87'?'0.17)'x'10!3 (7.95'?'0.14)'x'10!8 	
 ?
aKinetic parameters for binding events between CD23 and glycoengineered IgE-Fc-
Asn394only N-glycoforms were measured in Standard SPR buffer at 20 ?C and the 
obtained kinetic parameters were then analyzed using a global monovalent fitting 
algorithm (BiaEvaluation). The experiments were performed in triplicate and the 
optimal binding capacity of the immobilized CD23 receptor was confirmed by the 
standard IgE binding assay prior to each of the binding experiments to ensure that the 
CM5 chip-immobilized CD23 did not loss its binding capacity during storage at 4 ?C.  
bThe dissociation constant KD was defined as (kd/ka) and calculated using the 
monovalent fitting model of BiaEvaluation software.  
 
 
 
 
 
 
 
 
 178 
 
 
5. Chemoenzymatic N-glycosylation remodeling of IgE-Fc with complex N-glycan 
using the active glycosyl oxazoline and glycosynthase mutants 
The glycoengineering studies in previous sections with the heterogeneous 
IgE-Fc-Asn394only-HM were successful in producing the uniformly glycosylated 
variants IgE-Fc-Asn394only-Man9 and IgE-Fc-Asn394only-Man5 by enzymatic 
modifications and biosynthetic inhibitor. However, the oligomannose core of the 
native N-glycans attached at Asn-394 of IgE-Fc makes it intrinsically difficult to 
convert the original IgE N-glycans to a completely different glycoform such as IgE-
Fc with complex type N-glycan. As discussed extensively in the previous chapters, a 
conserved complex type N-glycan associated with Asn-297 residue in the IgG-Fc 
region can critically modulate the biological functions of different subclasses of IgG 
antibodies.20 Taking advantage of the different substrate specificities of various 
endoglycosidases, we and other groups have developed a collection of 
endoglycosidases and corresponding glycosynthases for glycoengineering of 
glycoproteins, including Endo-S, Endo-S2, and Endo-CC. These enzymes in 
combination with the activated glycosyl oxazolines have been highly successful in 
producing many glycopeptides and glycoproteins with desired N-glycan structures 
including IgG antibodies. Therefore, it will be of great interest to apply this strategy 
on IgE antibody with the goal of generating novel IgE-Fc N-glycoforms that could 
not be obtained using the existing techniques. Generation of such atypical fully-
processed complex type N-glycoforms of IgE at Asn-394 may lead to IgE 
glycovariants with distinct structural and functional properties from the wild type 
 179 
 
 
IgE-Fc, which can be exploited as potential therapeutic interventions against asthma 
and other allergic diseases.156, 174 
 
5.1.Selection of suitable endoglycosidases for IgE-Fc glycosylation remodeling  
In order to screen for potential endoglycosidases that can efficiently transfer 
the complex type N-glycans to IgE-Fc, we used the engineered IgE-Fc-Asn394only-
HM as starting material to test hydrolytic activities of a number of endoglycosidases 
that have been previously studied for this purpose, including Endo-M100, Endo-S103, 
Endo-S2139, and Endo-CC110 (Figure 54). Our initial results in Figure 55 indicated 
that treatment of IgE-Fc-Asn394only-HM with Endo-M and Endo-S2 led to only 
marginal hydrolysis of the IgE-Fc N-glycans even when a high concentration of 
enzyme and an extended incubation time were applied. In addition, enzyme Endo-S 
did not show any hydrolytic activities against native IgE-Fc glycoforms after an 
overnight reaction. These results are consistent with our previous observations that 
conserved Asn-394 glycosylation site of IgE-Fc was not largely accessible by 
enzymatic modifications such as ?-mannosidase.162 The inability of Endo-S to 
process high-mannose N-glycans from IgE was expected due to its restricted substrate 
selectivity for the complex type N-glycans only.130 Although Endo-S2 indeed 
possesses a broad substrate specificity for all three major types of N-glycans found in 
IgG antibodies (complex, hybrid, and high-mannose) and is efficient in processing the 
IgG N-glycans, many structural and functional studies suggested that both Endo-S 
and Endo-S2 were highly selective for the IgG class of antibody and exhibited only 
marginal hydrolytic activities against other commonly studied glycoproteins such as 
 180 
 
 
ribonuclease B.138 In contrast to Endo-S/S2, Endo-M was able to process the high-
mannose N-glycans of a large number of glycopeptides and glycoproteins, but the 
enzyme did not demonstrate any enzymatic activities toward the IgG antibody, as 
illustrated in previous study.89 Interestingly, the recently discovered endoglycosidase 
Endo-CC from fungus	
 ?Coprinopsis cinerea, as characterized in Chapter 4, was found 
to able to efficiently remove the indigenous N-glycans from the IgE-Fc-Asn394only-
HM form (Figure 55). This was supported by the results shown in Chapter 4 that 
Endo-CC displayed a relatively relaxed substrate selectivity for both the carbohydrate 
structures and protein framework of various glycoconjugates112.  
 
 
 
Figure 54. Hydrolytic selectivity of Endo-S, Endo-S2, Endo-M and Endo-CC against 
the high-mannose IgE-Fc-Asn394only-HM (Man5-9) N-glycoforms. For Endo-CC, 
rapid deglycosylation was observed against IgE-Fc N-glycoform, whereas Endo-S/S2 
and Endo-M did not show significant hydrolysis of the native N-glycans associated 
with the heterogeneous IgE-Fc fragment. Comparative hydrolysis assay is shown in 
Figure 54. 
 
 181 
 
 
100
Endo-CC
80 Endo-M
Endo-S
60 Endo-S2
40
20
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Time (h)
 
Figure 55. Comparison of enzymatic hydrolysis on native IgE-Fc-Asn394only-HM 
(Man5-9) N-glycoforms by the recombinant Endo-CC, Endo-M, Endo-S, and Endo-S2.	
 ?
A total concentration of 2 mg/mL of original IgE-Fc starting material was incubated 
with 0.1 mg/mL of all four enzymes 37 ?C in the Tris buffer at pH 7.4. The reaction 
was checked by SDS-PAGE and the relative hydrolysis of the IgE-Fc was quantified 
using densitometry analysis of the protein bands corresponding to the native and the 
deglycosylated IgE-Fc fragments. Each reaction was conducted in duplicate. 
 
5.2.Generation of a non-fucosylated complex type IgE-Fc-Asn394only N-glycoform 
using mutant Endo-CC N180 glycosynthase-catalyzed transglycosylation and CT 
(G2) glycan oxazoline 
Having confirmed that Endo-CC was able to efficiently hydrolyze the native 
N-glycans associated with IgE-Fc, we then performed transglycosylation reactions on 
the IgE-Fc-Asn394only-GlcNAc acceptor using the Endo-CC glycosynthase mutant 
Endo-CC N180H and a fully galactosylated glycan oxazoline (G2-CT-Oxa) as the 
 182 
 
Native IgE-Fc Glycoform (%)
 
donor substrate (Figure 56). A number of reaction conditions such as oxazoline molar 
equivalent and enzyme concentration were tested to maximize the transglycosylation 
yield and to minimize any non-enzymatic side reactions (data not shown). The final 
results demonstrated that treatment of IgE-Fc-Asn394only-GlcNAc with a total of 50 
molar equivalents of G2-CT-Oxa (25 equivalents per each glycosylation site of IgE-
Fc monomer) and 0.1 mg/mL of CPD-tagged recombinant Endo-CC N180H for 4 
hours resulted in a homogeneous IgE-Fc-CT N-glycoform that appeared as a single 
band on the SDS-PAGE (Figure 57A). The molecular weight of the product band was 
estimated to be about 2 kDa (approximately the size of a G2-CT glycan per one IgE-
Fc monomer) larger than the starting material. The band of a lower-molecular-weight 
species corresponding to the starting material was not seen after the reaction, 
suggesting a complete conversion of IgE-Fc. Again, MALDI-TOF mass spectrometry 
analysis on the PNGase F-released N-glycans of the reaction product revealed a major 
mass species of 1663.3 Da, which corresponds to the released G2-CT glycan 
(calculated M.W. 1664, Figure 57B). A minor species with a M.W. of 1257.4 was 
also present in the released glycan profile possibly due to incomplete deglycosylation 
of the native high-mannose N-glycans from the previous step. The product was 
subsequently treated with both Endo-CC and PNGase F to confirm that the 
transferred complex type N-glycan was not linked to IgE-Fc through non-enzymatic 
chemical ligation.122 Therefore, a highly efficient and regioselective chemoenzymatic 
glycosylation modeling of the IgE antibody with the novel complex type glycan 
oxazoline was achieved.  
 
 183 
 
 
 
Figure 56. Transglycosylation of IgE-Fc-Asn394only-GlcNAc with the galactosylated 
complex type glycan oxazoline (CT-Oxa) using glycosynthase Endo-CC N180H.  
 
5.3.Glycoengineered complex type IgE-Fc-Asn394only-CT displayed distinct binding 
properties from those of the oligomannose IgE-Fc N-glycoforms   
The binding kinetics of the remodeled IgE-Fc glycoform for the high-affinity 
IgE receptor Fc?RI was measured by surface plasmon resonance (SPR). Following 
the previously described procedure, serial dilutions of both the product glycoform 
IgE-Fc-Asn394only-CT and starting glycoform IgE-Fc-Asn394only-GlcNAc were 
analyzed as analytes and compared for their binding kinetics. As expected, the 
deglycosylated IgE-Fc exhibited an expected binding affinity for Fc?RI [KD = (9.47 ? 
0.08) ? 10-9 M-1] prior to transglycosylation. On the other hand, the KD value of the 
complex IgE-Fc glycoform (IgE-Fc-Asn394only-CT) to Fc?RI was found to slightly 
higher than that of the deglycosylated IgE-Fc [KD =  (1.28 ? 0.01) ? 10-8 M-1] (Figure 
57C and 57D), although the difference was not dramatic statistically. However, the 
kinetic parameters of the two IgE-Fc glycoforms studied here revealed to be 
significantly different. The association rate constant ka of the glycoengineered IgE-
Fc-Asn394only-CT [ka =  (6.79 ? 0.06) ? 105 Ms-1] was approximately 5-fold higher 
 184 
 
 
than that of the deglycosylated glycoform [ka = (1.25 ? 0.01)? 105 Ms-1], whereas the 
relationship was reversed for the dissociation rate constant, as the IgE-Fc-
Asn394only-CT [kd = (8.69 ? 0.11)? 104 s-1] dissociates from Fc?RI faster than IgE-
Fc-Asn394only-GlcNAc [kd = (1.18 ? 0.03)? 103 s-1]. These results suggest a 
significant change of IgE-Fc binding kinetics to Fc?RI caused by the remodeling of 
the complex type N-glycan. In addition to the CT-Oxa, previous studies on Endo-CC 
N180H substrate specificities showed that the mutant was also able to transfer the 
fully sialylated complex type glycan oxazoline (SCT-Oxa) to different glycosyl 
acceptors including the therapeutic IgG antibody112, which suggested that other 
complex glycan structures could potentially be transferred to IgE antibody as well. 
Therefore, this chemoenzymatic glycoengineering strategy discussed here can 
provide additional homogeneous IgE N-glycoforms that might be useful for structural 
and functional studies as well as development for novel anti-allergic therapies.  
 
 
 
 185 
 
 
 
Figure 57. Glycan remodeling of IgE-Fc-Asn394only-CT (G2) N-glycoform and its 
binding properties for the high-affinity receptor Fc?RI. (A) SDS-PAGE analysis to 
monitor transglycosylation of IgE-Fc by Endo-CC N180H: Lane 1, IgE-Fc-
Asn394only-GlcNAc; Lane 2, reaction at t = 1 h; Lane 3, reaction at t = 2 h; Lane 3, 
reaction at t = 4 h and final product of IgE-Fc-Asn394only-CT. (B) MALDI-TOF MS 
analysis of the released N-glycans from the remodeled IgE-Fc-Asn394only-CT. (C) 
and (D) SPR sensorgrams of the binding between Fc?RI and (C) IgE-Fc-Asn394only-
GlcNAc or (D) IgE-Fc-Asn394only-CT, respectively. Eight concentrations of IgE-Fc 
N-glycoform (400?3.125 nM) were injected as analytes and the kinetic data was 
analyzed using a monovalent fitting model. 
 186 
 
 
C. Conclusion 
A combined glycosylation remodeling strategy of enzymatic modification and 
biosynthetic perturbation for intact IgE antibodies and IgE-Fc fragments is described 
in this work. This method enabled the generation of a series of homogeneous IgE-Fc 
N-glycoforms, and binding properties of these IgE glycoforms to both the high- and 
low-affinity IgE receptors Fc?RI and CD23 were characterized. To our knowledge, 
this study represents the first investigation on the functional role of N-glycans in 
regulating the binding between IgE-Fc expressed in mammalian cells and the low-
affinity CD23 receptor, which is important for mediating IgE homeostasis in human. 
Although the sugar composition did not display any dramatic impact on IgE binding 
to either receptor, the presence of the native N-glycans and, for the first time, the 
remodeling of a novel complex N-glycan at the critical Asn-394 residue did show 
changes to the binding properties of IgE-Fc glycoforms. These adequately combined 
approaches greatly expand the scope of the chemoenzymatic glycoengineering 
strategy beyond the IgG antibody and opened the avenue for further remodeling of 
other human antibodies such as IgA and IgE, using endoglycosidases and additional 
oxazoline structures, such as the sialylated (S2G2-CT) and agalactosylated (G0-CT) 
glycoforms, which are difficult to produce using current techniques. In addition, 
recent structural studies suggested that anti-IgE antibodies such as omalizumab175 and 
8D6155 provide effective amelioration of asthma-like allergic diseases by targeting a 
mixed protein-carbohydrate epitope including Asn-394 N-glycan at IgE-Fc. Therefore, 
glycoengineering of IgE-Fc N-glycans may further our understanding of the structural 
 187 
 
 
mechanism behind IgE/anti-IgE recognition and potentially facilitate the development 
of novel therapeutic interventions against allergic disorders.  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 188 
 
 
D.  Materials and Methods 
1. Mammalian cell culture: cell lines, strains and expression conditions 
The plasmid DNA expressing the high-mannose IgE-Fc-Asn394only-HM 
fragment was designed with asparagine-to-glutamine (N-to-Q) mutations at all 
conserved N-glycosylation sites except for Asn-394 (Genscript). The IgE-Fc DNA 
was transfected into the HEK293T cells by transient transfection with 20 ?g/ml of 
kifunensine (Cayman Chemical, Ann Arbor, Michigan) added as an inhibitor for 
mannosidases. The mammalian HEK297T cells were grown in FreeStyle F17 
Expression Medium in the presence of 2 mM GlutaMax (Thermo Fisher Scientific). 
All cell lines were subject to 150 rpm shaking at 37 ?C with 8% CO2. The detailed 
protocol for transfection and purification of the IgE-Fc protein was described in a 
previous publication.176 
2. Expression, purification and characterization of recombinant endoglycosidases 
Endo-CC, Endo-M, Endo-S and Endo-S2 and their glycosynthase mutants in 
bacterial cell culture 
The detailed protocol for the expression and purification of endoglycosidases 
and their related glycosynthase mutants, including Endo-CC (Chapter 4), Endo-S 
(Chapter 1), and Endo-S2 (Chapter 3), were described in the experimental sections of 
previous Chapters. The recombinant CPD-tagged Endo-M was generated following a 
previous published protocol99.   
 
 189 
 
 
3. Site-specific glycan remodeling of full-length IgE antibodies 
To selectively release the oligomannose N-glycan at Asn-394 without 
removing the other complex N-glycans attached at the heavy chain of the full-length 
IgE antibodies, 50 ?g of both the commercial human IgE antibody (Abcam, UK) and 
the recombinant ovalbumin (OVA)-specific IgE antibody were treated with 1000 
units of recombinant Endo-Hf (New England Biolabs, NEB) in a total volume of 50 
?l PBS buffer (pH 7.4) at 37?C for overnight incubation. The commonly used 
Glycoprotein Buffer 3 (New England Biolabs, NEB) was not used in this study 
because the deglycosylation were found to be efficient without adding any additional 
buffer components. Removal of N-glycans from IgE were monitored and confirmed 
by SDS-PAGE analysis with a shift of bands corresponding to the differentially 
glycosylated heavy chains of IgE antibody. The anti-OVA IgE antibody was provided 
by Dr. Robert Anthony?s group at Harvard University, MA.  
4. Deglycosylation of starting high-mannose IgE-Fc-Asn394only-HM to afford IgE-
Fc-Asn394only-GlcNAc acceptor intermediate and IgE-Fc-Asp394 intermediate 
The expressed heterogeneous IgE-Fc-Asn394only-HM N-glycoforms were 
dialyzed into PBS at pH 7.4 and purified using an anion exchange chromatography 
(GE Healthcare) following the manufacturer?s instruction. For the deglycosylation of 
IgE-Fc by endoglycosidase Endo-H, a total of 500 ?g IgE-Fc-Asn394only-HM was 
treated with 1000 units of commercial Endo-Hf enzyme (New England Biolabs, NEB) 
in 50 ?l of PBS buffer (pH 7.4) at 37?C for 4 hours. The completely deglycosylated 
IgE-Fc-Asp394 was prepared in a similar fashion by incubating 200 ?g IgE-Fc-
Asn394only-HM with 20 units of PNGase F (New England Biolabs, NEB) in 100 ?l 
 190 
 
 
of PBS buffer (pH 7.4) at 37?C for overnight reaction. Removal of N-glycans were 
monitored and confirmed by SDS-PAGE densitometry followed by MALDI-TOF 
mass spectrometry characterization of Fc-Asn394only-GlcNAc and IgE-Fc-Asp394.  
5. Generation of N-glycoforms IgE-Fc-Asn394only-Man5 and IgE-Fc-Asn394only-
Man4-7 N-glycoforms by enzymatic modifications 
For each N-glycoform, 500 ?g IgE-Fc-Asn394only-HM was treated with 20 
units (RU) of either commercial ? 1-2 or ? 1-2, 3, 6 mannosidase (New England 
Biolabs, NEB) in 50 ?l of PBS buffer (pH 7.4) at 37?C for overnight reaction. One 
overnight of ? 1-2, 3, 6 mannosidase showed little progress of mannose removal and 
additional 20 units of ? 1-2, 3, 6 mannosidase was added to the reaction mixture and 
incubated for additional 12 h. The final glycoforms of modified IgE-Fc were analyzed 
by SDS-PAGE and MALDI-TOF analysis on the released N-glycans by PNGase F.  
6. MALDI-TOF mass spectrometry 
The N-glycans were released from the IgE-Fc proteins by overnight 
incubation with a small amount of commercial PNGase F (20 units) at 37 ?C for up to 
12 hours. The released glycans were then characterized using a Bruker 
UltrafleXtreme MALDI-TOF/TOF mass spectrometer. The instrument was set to the 
positive reflectron mode. A solution of 2,5-dihydroxybenzoic acid (DHB, Sigma-
Aldrich) was formulated by adding 100mg of DHB into 1mL of 50% acetonitrile 
(ACN) and used as the MALDI-TOF matrix for released N-glycans. To improve the 
detection sensitivity for some N-glycans studied here, N,N-dimethylaniline (DMA, 
Sigma-Aldrich) was also added to the sample mixture prior to analysis.  
 191 
 
 
7. Surface Plasmon Resonance (SPR) 
The interactions between different N-glycoforms of IgE-Fc-Asn394only 
proteins and both the high- and low-affinity IgE receptors, Fc?RI and CD23, were 
studied using a BIAcore T200 system (GE Healthcare Life Sciences) at 25 ?C. The 
soluble domain of both Fc?RI and CD23 were immobilized on the CM5 sensor chip 
(GE Healthcare Life Sciences) using the amine coupling kit according to the 
manufacturer?s instruction. For each IgE receptor, a 1000 response unit (RU) of 
immobilization was achieved by pH scouting. A reference channel was also prepared 
by blocking the active functional groups on the CM5 chip. The IgE-Fc analytes were 
prepared in serial dilutions in HBS-P buffer (GE Healthcare Life Sciences). A total of 
6-8 concentrations of each IgE-Fc samples were injected across the sensor surface at 
a flow rate of 40 ?L/min, and the CM5 surface was regenerated after each analysis 
using 20 mM HCl solution at pH 2.4 (GE Healthcare Life Sciences). The obtained 
binding parameters were analyzed using a global monovalent fitting model 
(BiaEvaluation) to calculate the binding affinity constant (KD). The results were 
collected from three independently conducted binding experiments.   
 
8. Assay for hydrolytic activities of Endo-CC, Endo-S, Endo-S2 and Endo-M 
against recombinant IgE-Fc-Asn394only-HM 
For each hydrolysis reaction, a total of 100 ?g IgE-Fc-Asn394only-HM was 
incubated with a final concentration of 0.5 mg/mL of each enzyme in 50 ?l PBS 
buffer (pH 7.4) at 37?C. The reaction was sampled at multiple time points up to t = 12 
 192 
 
 
h. Hydrolysis of high-mannose N-glycans from IgE-Fc was estimated by SDS-PAGE 
densitometry analysis. The experiments were performed in duplicate.  
9. Glycan remodeling of the deglycosylated IgE-Fc-Asn394only to a homogenous 
galactosylated complex type (G2-CT) N-glycoform 
In the first step of glycan remodeling, 100 ?g of IgE-Fc-Asn394only-HM was 
treated with Endo-H to provide the deglycosylated IgE-Fc glycosyl acceptor. 75 ?g of 
this IgE-Fc-Asn394only-GlcNAc intermediate was mixed with 3 ?g of CPD-tagged 
recombinant Endo-CC N180H, and 50 ?g of G2-Oxa, in 20 mM Tris-HCl buffer, at 
pH 7.4, which was reported to be the optimal reaction condition for Endo-CC N180H 
transglycosylation.110 The reaction temperature was set to 30 ?C.89 Additional 50 ?g 
of G2-Oxa and 1?g of Endo-CC N180H were added to increase the reaction yield. 
Aliquot was collected at the 1, 2, and 4 h time points to monitor the transfer of the 
complex N-glycan by both SDS-PAGE and MALDI-TOF analysis. After 4 hours, the 
transglycosylation reaction was complete and further incubation did not show any 
improvements of total reaction yield.  The biantennary galactosylated complex glycan 
oxazoline (G2-Oxa) was prepared following the previously reported protocol.99  
 
 
 
 
 
 
 
 
 193 
 
 
E. Supporting Information 
GTMGWSCIILFLVATATGVHSSVCSRDFTPPTVKILQSSCDGGGHFPPTIQLLC
LVSGYTPGTIQITWLEDGQVMDVDLSTASTTQEGELASTQSELTLSQKHWLS
DRTYTCQVTYQGHTFEDSTKKCADSNPRGVSAYLSRPSPFDLFIRKSPTITCLV 
VDLAPSKGTVQLTWSRASGKPVQHSTRKEEKQRNGTLTVTSTLPVGTRDWIE
GETYQCRVTHPHLPRALMRSTTKTSGPRAAPEVYAFATPEWPGSRDKRTLAC
LIQNFMPEDISVQWLHNEVQLPDARHSTTQPRKTKGSGFFVFSRLEVTRAEW
EQKDEFICRAVHEAASPSQTVQRAVSVNPGK 
Supplementary Figure 4: Amino acid sequence of the mutated IgE-Fc fragment IgE-
Fc-Asn394only. A CD5 signal peptide is inserted at the beginning of the sequence. 
The Asn394 is highlighted in red. All glycosylation sites (N) of this IgE-Fc construct 
are mutated to glutamine (Q) except for Asn-394. 
 
 
 194 
 
 
Chapter 6:  Summary and future directions 
 
The major focus of this dissertation is to develop and improve the current 
chemoenzymatic remodeling method for antibody glycosylation using various 
endoglycosidases and glycosynthases. The glycosylation patterns of monoclonal 
antibodies, both naturally existing and recombinantly expressed, are typically 
complex and highly heterogeneous in their carbohydrate compositions and can 
critically determine the structural and biological properties of antibody molecules.4, 10, 
16, 34 To more effectively control and optimize the structural diversity of antibody 
glycans, an efficient in vitro chemoenzymatic glycoengineering strategy was 
developed based on the unique hydrolysis and transglycosylation activities of 
endoglycosidases and glycosynthases, respectively.20, 99, 103, 106  
In Chapter 2, systematic site-directed mutagenesis analysis at the key catalytic 
residue Asp-233 of Endo-S (Streptococcus pyogenes) identified many glycosynthase 
mutants of Endo-S including Endo-S D233M that exhibited higher transglycosylation 
efficiencies than the previously reported D233Q and D233A.123 Quantitative 
determination of antibody glycosylation was achieved using the electrospray 
ionization (ESI) mass spectrometry and an antibody-based internal standard.136 
Kinetic analyses of Endo-S D233M and D233A indicated that the Asp-to-Met 
mutation at Asp-233 led to the increased enzyme turnover number for the glycan 
donor substrate and enhanced substrate affinity for the glycan acceptor substrate. The 
potential occurrence of the non-enzymatic chemical glycosylation of glycan 
oxazolines to random residues of antibody was investigated by comparing the 
reaction conditions and characterizing the chemical composition of the antibody 
 195 
 
 
products formed after glycoengineering. The comparative results suggested that such 
unwanted side reactions did not occur under the optimized reaction conditions 
commonly used for the antibody glycosylation by endoglycosidases. Potential 
chemical ligation of glycan oxazolines with antibody acceptor residues only became 
significant when a large excess of glycan substrate was used.109, 122, 132 
In Chapter 3, similar to studies on glycosynthase mutants of Endo-S, 
systematic mutagenesis on another endoglycosidase Endo-S2 led to the first antibody-
specific glycosynthases that could efficiently catalyze transglycosylation of IgG 
antibodies with additional N-glycans such as the N-linked hybrid and high-mannose 
oligosaccharides. Kinetic studies on the Endo-S2 mutants D184M and D184A 
revealed highly similar enzyme behaviors to glycosynthase mutants of Endo-S, 
including an enhanced affinity for antibody and a faster turnover number for glycan 
oxazoline. The superior catalytic activities of Endo-S2 D184M over other ENGases 
promoted the synthesis of antibody N-glycoforms with more diverse sugar structures. 
In collaboration with my colleague Dr. Tiezheng Li in Prof. Lai-Xi Wang?s lab and 
Prof.	
 ? Jeffrey Ravetch from the Rockefeller University, a comprehensive library of 
IgG N-glycoforms were generated using mutant Endo-S2 D184M, and both in vitro 
and in vivo functional assays on the resulting library identified the core fucosylation 
as a major determinant of the antibody activity in modulating the immune 
response.177 
In Chapter 4, substrate specificity analyses of the wild-type Endo-S revealed 
promiscuity for additional N-glycan substrates, and demonstrated that, unlike Endo-S 
mutants, the wild-type Endo-S was capable of recognizing both the high-mannose and 
 196 
 
 
hybrid glycans for antibody glycosylation with a minimal level of product hydrolysis. 
A highly efficient ?one-pot? glycan remodeling method was developed based on the 
relaxed substrate selectivity of Endo-S to modify different antibody glycoforms and 
glycan oxazolines, which also included a chemically modified azido-tagged N-glycan. 
Because of its ability to convert antibody N-glycosylation without the need of 
purifying the deglycosylated intermediate or switching enzymes, this method could 
be potentially useful in large-scale antibody manufacturing in terms of lowered cost 
due to reduced producing steps.178 
The scope and limitations related to the catalytic properties of Endo-CC and 
its glycosynthase mutants were investigated using a series of glycoprotein and N-
glycan oxazolines as substrates. Parallel transglycosylation reactions indicated that 
the mutant Endo-CC N180H could potentially utilize several types of N-glycan 
substrates for transglycosylation in addition to complex oxazoline. Competitive 
functional assays confirmed that core fucosylation was not recognized by Endo-CC in 
terms of glycosyl acceptor substrate selectivity. In addition, the relative activities of 
Endo-CC mutants were re-evaluated in a quantitative assay, which revealed 
additional mutants that showed significant transglycosylation activities over a long 
period of reaction time. This study introduced an alternative endoglycosidase with 
broad substrate selectivity for the protein acceptor part in the glycan remodeling 
procedure.  
Finally, a combined strategy of both enzymatic modifications and biosynthetic 
inhibition was applied to another isotype of antibody known as IgE. Surface binding 
assays using these generated IgE N-glycoforms revealed significant change of 
 197 
 
 
receptor binding affinities caused by alteration in IgE glycan structures attached at Fc 
region. Furthermore, chemoenzymatic glycosylation remodeling of IgE successfully 
gave rise to a novel complex type IgE-Fc glycoform with a distinct receptor binding 
profile from the deglycosylated and native glycoforms of IgE antibody. The results 
collectively suggested that further remodeling of IgE antibodies with additional 
glycan structures, such as terminal sialylation and core fucosylation, is possible 
through glycosynthase-catalyzed transglycosylation reactions using different 
activated glycosyl oxazolines. 
Together, studies in this dissertation greatly improved the scope of the current 
chemoenzymatic glycan remodeling strategy of antibody and other glycoconjugates. 
Nevertheless, there are still limitations and challenges associated with the application 
of this method in other systems. Future work is needed to improve our understanding 
of the catalytic mechanisms of endoglycosidases and to help us design more efficient 
glycan-remodeling enzyme tools for various structural and functional purposes.  
First, the molecular mechanism behind the different behaviors of Endo-S and 
Endo-S2 and their glycosynthase mutants in glycan remodeling of antibody is still not 
clear. Structural analyses reported by other groups on Endo-S and Endo-S2 both in 
their native forms and in complex with free glycans revealed some molecular 
information on how the branches of N-linked oligosaccharide were recognized by a 
number of loops located in the glycoside hydrolase (GH) domain conserved in both 
enzymes.116, 130, 179 However, based on the general catalytic mechanism of 
endoglycosidases, the ground-state glycan in the crystal structure may not accurately 
reflect the interactions between catalytic residues and oxazolinium ion intermediates 
 198 
 
 
during glycosylation. In order to explain the different turnover numbers of Endo-S 
and Endo-S2 mutants for the glycan oxazoline substrate observed in our kinetic 
results,123, 139 a non-participating thiazoline glycan analogue124 can be applied in the 
future crystallographic analysis of Endo-S and Endo-S2 to help understand how the 
oxazoline is recognized by both enzymes. Our kinetic analyses on glycosynthase 
mutants of both enzymes also clearly indicated that a single mutation at Asp-233 and 
Asp-184 of Endo-S and Endo-S2, respectively, exhibited profound effects on 
substrate affinities for the antibody acceptor. Therefore, structural analysis on Endo-
S/S2 mutant interacting with a deglycosylated Fc fragment of IgG antibody by either 
co-crystallization or molecular simulations can be helpful in understanding the 
mechanisms behind this observation. Together, these structural and functional studies 
can help design possible chimeric endoglycosidase or glycosynthase variants with 
more customizable glycan remodeling abilities in glycoengineering. 
Another structural feature of the complex and hybrid N-glycans that has not 
been addressed in this work is the ?bisecting? N-acetylglucosamine (GlcNAc) residue 
that is associated with core ?-mannose of N-glycans. Unlike other GlcNAc structures 
found in N-glycans, the bisecting GlcNAc demonstrated several unique functional 
properties in N-glycan biosynthesis and has been implicated in neurodegenerative 
disorders such Alzheimer?s disease.180 In terms of antibody N-glycosylation, the 
bisecting GlcNAc-containing N-glycan has been regarded as one of the critical 
quality attributes (CQA) for therapeutic antibodies181 and has been associated with 
the physiological state of individuals including human longevity.182 The unique 
properties of bisecting N-glycoforms of antibody made it of interest to study the	
 ?
 199 
 
 
structural-functional relationships of this specific sugar residue in antibody. However, 
like other N-glycans, it has been difficult to isolate or enrich the pure bisecting N-
glycoforms of antibody from natural sources given the relatively low abundance of 
such antibodies in animal serum.36 To address this problem, the chemoenzymatic 
glycan modeling method can be tested for its compatibility with the bisecting glycan 
substrates. The glycosynthase mutant Endo-S2 D180M can be a promising candidate 
enzyme for transferring the bisecting N-glycans to the antibody acceptor because a 
previous study already demonstrated that the wild-type Endo-S2 was able to 
efficiently remove a wide range of N-glycans from human antibodies including the 
bisecting glycans.183 Generation of the homogeneous bisecting N-glycoforms of 
antibodies may further our understanding of the disease-associated variations related 
to antibody glycosylation. 
Finally, considering the universal distribution of N-glycans across all antibody 
isotypes59, it might be interesting to investigate the potential application of this 
method in other antibody isotypes such as IgE and IgA. Binding analyses on IgE N-
glycoforms revealed distinct interaction patterns between IgE-Fc and its related 
immune receptors. However, the available IgE-Fc glycoforms studied in this work are 
still limited to the oligomannose and basic complex forms without any additional 
modifications such as terminal sialic acid or core fucose. Since extensive studies have 
suggested that both terminal sialylation and core fucosylation can dramatically affect 
IgG interactions with immune receptors, we hypothesize that glycoengineering of IgE 
antibody with these structural features might lead to novel IgE N-glycoforms with 
interesting biophysical and immunoregulatory properties. Glycan remodeling with 
 200 
 
 
sialylated N-glycans can be experimented using the same Endo-CC N180H mutant, as 
our results in Chapter 4 clearly indicated that Endo-CC could recognize a wide range 
of N-glycan substrates including the sialylated complex type. For the core fucosylated 
N-glycans, a recently reported mutant of an ?1,6-fucosidase can be used to attempt 
direct core fucosylation on IgE-Fc or intact IgE antibodies. The ability of this mutant 
to efficiently transfer the modified fucose substrates to a variety of glycopeptides and 
glycoproteins has been demonstrated in several previous studies.94, 147, 184 In 
conclusion, generation of possible IgE-Fc N-glycoforms with improved binding 
affinities to both IgE receptors Fc?RI and CD23 may allow us to design novel 
competitive inhibitors for the anaphylactic IgE antibodies in vivo and facilitate the 
development of new therapeutics for allergic disorders such as asthma.  
Glycosylation remodeling of IgE antibodies can also affect the recognition of 
allergic IgE antibodies by anti-IgE therapeutics. Interestingly, structural and 
functional studies suggest that IgE-neutralizing IgG antibodies such as	
 ?8D6155 and 
MEDI4212185 target the pathogenic IgE antibodies through a mixed protein-
carbohydrate epitope, which involves the critical oligomannose glycan located at 
Asn-394 that we discussed in Chapter 5. Removal or modifications of this N-glycan 
from IgE-Fc domain is likely to affect recognition of IgE by these anti-IgE antibodies. 
However, such modifications might not dramatically affect normal binding of IgE-Fc 
with the low-affinity receptor CD23, as indicated by our results of our IgE 
glycoengineering and binding studies in Chapter 5. These studies may contribute to 
the development of new anti-IgE therapies, as one of the major problems associated 
with the current anti-IgE therapeutics is the disruption of normal IgE antibody 
 201 
 
 
homeostasis due to lack of IgE/CD23 signaling.155 Therefore, a potential 
glycoengineered IgE-Fc fragment with no functionality to trigger the allergic 
response due to lack of the allergen-binding Fab domain may serve as a therapeutic 
intervention in combination with the conventional anti-IgE antibodies to treat long-
term allergic disorders (Figure 58). Such IgE-Fc N-glycoforms may help maintain the 
normal homeostasis of circulating IgE antibodies through binding to the CD23 
receptors on B cell surface, yet it may be less likely to be recognized and sequestered 
by the anti-IgE antibodies due to its changed glycan epitope at Fc region. Similar IgE 
combination therapies have already been experimented in previous studies using 
alternative strategies to disrupt anti-IgE/IgE recognition such as mutagenesis.186  
 
 
 
 202 
 
 
 
Figure 58. Design for therapeutic intervention as IgE replacement therapy using a 
combination of anti-IgE antibodies and glycoengineered IgE-Fc. Briefly, the native 
allergic IgE antibody is sequestered from the immune cell surface by high-affinity 
anti-IgE antibodies. The receptors are then bound by IgE-Fc N-glycoforms with 
similar binding affinities to maintain the homeostasis of serum IgE level. 
Modifications on IgE-Fc N-glycans at Asn-394 and/or other glycosylation sites may 
reduce the recognition by anti-IgE antibodies on the cell surface. 
 
 
 
 
 203 
 
 
Bibliography 
[1] Kaplon, H., and Reichert, J. M. (2018) Antibodies to watch in 2018, Mabs-Austin 
10, 183-203. 
[2] Burk, M. L., and Matuszewski, K. A. (1997) Muromonab-CD3 and antithymocyte 
globulin in renal transplantation, The Annals of pharmacotherapy 31, 1370-
1377. 
[3] Grilo, A. L., and Mantalaris, A. (2019) The increasingly human and profitable 
monoclonal antibody market, Trends in Biotechnology 37, 9-16. 
[4] Mastrangeli, R., Palinsky, W., and Bierau, H. (2019) Glycoengineered antibodies: 
towards the next-generation of immunotherapeutics, Glycobiology 29, 199-
210. 
[5] Elgundi, Z., Reslan, M., Cruz, E., Sifniotis, V., and Kayser, V. (2017) The state-
of-play and future of antibody therapeutics, Advanced Drug Delivery Reviews 
122, 2-19. 
[6] Su, D., Kozak, K. R., Sadowsky, J., Yu, S. F., Fourie-O'Donohue, A., Nelson, C., 
Vandlen, R., Ohri, R., Liu, L., Ng, C., He, J., Davis, H., Lau, J., Del Rosario, 
G., Cosino, E., Cruz-Chuh, J. D., Ma, Y., Zhang, D., Darwish, M., Cai, W., 
Chen, C., Zhou, H., Lu, J., Liu, Y., Kaur, S., Xu, K., and Pillow, T. H. (2018) 
Modulating antibody-drug conjugate payload metabolism by conjugation site 
and linker modification, Bioconjugate Chemistry 25, 21-50. 
[7] Almagro, J. C., Daniels-Wells, T. R., Perez-Tapia, S. M., and Penichet, M. L. 
(2018) Progress and challenges in the design and clinical development of 
antibodies for cancer therapy, Frontiers in Immunology 8, 31-45. 
[8] Elvin, J. G., Couston, R. G., and van der Walle, C. F. (2013) Therapeutic 
antibodies: market considerations, disease targets and bioprocessing, 
International Journal of Pharmaceutics 440, 83-98. 
[9] Schroeder, H. W., Jr., and Cavacini, L. (2010) Structure and function of 
immunoglobulins, The Journal of Allergy and Clinical Immunology 125, S41-
S52. 
[10] Jennewein, M. F., and Alter, G. (2017) The Immunoregulatory Roles of 
Antibody Glycosylation, Trends in Immunology 38, 358-372. 
[11] Williams, A. F., and Barclay, A. N. (1988) The immunoglobulin superfamily--
domains for cell surface recognition, Annual Review of Immunology 6, 381-
405. 
[12] Harpaz, Y., and Chothia, C. (1994) Many of the immunoglobulin superfamily 
domains in cell-adhesion molecules and surface-receptors belong to a new 
structural set which is close to that containing variable domains, Journal of 
Molecular Biology 238, 528-539. 
[13] Smith, K. A., Nelson, P. N., Warren, P., Astley, S. J., Murray, P. G., and 
Greenman, J. (2004) Demystified ... recombinant antibodies, Journal of 
Clinical Pathology 57, 912-917. 
[14] Matthias, P., and Rolink, A. G. (2005) Transcriptional networks in developing 
and mature B cells, Nature Reviews. Immunology 5, 497-508. 
[15] Jazwinska, E. C., Dunckley, H., Propert, D. N., Gatenby, P. A., and Serjeantson, 
S. W. (1988) Gm typing by immunoglobulin heavy-chain gene RFLP 
analysis, American Journal of Human Genetics 43, 175-181. 
 204 
 
 
[16] Cymer, F., Beck, H., Rohde, A., and Reusch, D. (2018) Therapeutic monoclonal 
antibody N-glycosylation - Structure, function and therapeutic potential, 
Biologicals : Journal of the International Association of Biological 
Standardization 52, 1-11. 
[17] Normansell, D. E. (1987) Human immunoglobulin subclasses, Diagnostic and 
Clinical Immunology 5, 115-128. 
[18] Kraft, S., and Kinet, J. P. (2007) New developments in Fc epsilon RI regulation, 
function and inhibition, Nature Reviews Immunology 7, 365-378. 
[19] Torres, M., and Casadevall, A. (2008) The immunoglobulin constant region 
contributes to affinity and specificity, Trends in Immunology 29, 91-97. 
[20] Wang, L. X., Tong, X., Li, C., Giddens, J. P., and Li, T. (2019) 
Glycoengineering of antibodies for modulating functions, Annual Review of 
Biochemistry 88, 433-459. 
[21] Schroeder, H. W., and Cavacini, L. (2010) Structure and function of 
immunoglobulins, Journal of Allergy and Clinical Immunology 125, S41-S52. 
[22] Kurosu, M. (2019) Structure-based drug discovery by targeting N-glycan 
biosynthesis, dolichyl-phosphate N-acetylglucosaminephosphotransferase, 
Future Medicinal Chemistry, in press.  
[23] Piirainen, M. A., Salminen, H., and Frey, A. D. (2019) Tailoring N-Glycan 
biosynthesis for production of therapeutic proteins in Saccharomyces 
cerevisiae, Methods in Molecular Biology 1923, 227-241. 
[24] Bieberich, E. (2014) Synthesis, Processing, and Function of N-glycans in N-
glycoproteins, Advances in Neurobiology 9, 47-70. 
[25] Arnold, J. N., Wormald, M. R., Sim, R. B., Rudd, P. M., and Dwek, R. A. (2007) 
The impact of glycosylation on the biological function and structure of human 
immunoglobulins, Annual Review of Immunology 25, 21-50. 
[26] Stanley, P., Taniguchi, N., and Aebi, M. (2015) N-Glycans, In Essentials of 
Glycobiology (rd, Varki, A., Cummings, R. D., Esko, J. D., Stanley, P., Hart, 
G. W., Aebi, M., Darvill, A. G., Kinoshita, T., Packer, N. H., Prestegard, J. 
H., Schnaar, R. L., and Seeberger, P. H., Eds.), pp 99-111, Cold Spring 
Harbor (NY). 
[27] Liu, H., Bulseco, G. G., and Sun, J. (2006) Effect of posttranslational 
modifications on the thermal stability of a recombinant monoclonal antibody, 
Immunology Letters 106, 144-153. 
[28] Natsume, A., Wakitani, M., Yamane-Ohnuki, N., Shoji-Hosaka, E., Niwa, R., 
Uchida, K., Satoh, M., and Shitara, K. (2005) Fucose removal from complex-
type oligosaccharide enhances the antibody-dependent cellular cytotoxicity of 
single-gene-encoded antibody comprising a single-chain antibody linked the 
antibody constant region, Journal of Immunological Methods 306, 93-103. 
[29] Sazinsky, S. L., Ott, R. G., Silver, N. W., Tidor, B., Ravetch, J. V., and Wittrup, 
K. D. (2008) Aglycosylated immunoglobulin G1 variants productively engage 
activating Fc receptors, Proceedings of the National Academy of Sciences of 
the United States of America 105, 20167-20172. 
[30] Siberil, S., de Romeuf, C., Bihoreau, N., Fernandez, N., Meterreau, J. L., 
Regenman, A., Nony, E., Gaucher, C., Glacet, A., Jorieux, S., Klein, P., 
Hogarth, M. P., Fridman, W. H., Bourel, D., Beliard, R., and Teillaud, J. L. 
 205 
 
 
(2006) Selection of a human anti-RhD monoclonal antibody for therapeutic 
use: Impact of IgG glycosylation on activating and inhibitory Fc gamma R 
functions, Clinical Immunology 118, 170-179. 
[31] Albert, H., Collin, M., Dudziak, D., Ravetch, J. V., and Nimmerjahn, F. (2008) 
In vivo enzymatic modulation of IgG glycosylation inhibits autoimmune 
disease in an IgG subclass-dependent manner, Proceedings of the National 
Academy of Sciences of the United States of America 105, 15005-15009. 
[32] Miranda, L. R., Duval, M., Doherty, H., Seaman, M. S., Posner, M. R., and 
Cavacini, L. A. (2007) The neutralization properties of a HIV-specific 
antibody are markedly altered by glycosylation events outside the antigen-
binding domain, Journal of Immunology 178, 7132-7138. 
[33] Clerc, F., Reiding, K. R., Jansen, B. C., Kammeijer, G. S. M., Bondt, A., and 
Wuhrer, M. (2016) Human plasma protein N-glycosylation, Glycoconjugate 
Journal 33, 309-343. 
[34] Giorgetti, J., D'Atri, V., Canonge, J., Lechner, A., Guillarme, D., Colas, O., 
Wagner-Rousset, E., Beck, A., Leize-Wagner, E., and Francois, Y. N. (2018) 
Monoclonal antibody N-glycosylation profiling using capillary electrophoresis 
- Mass spectrometry: Assessment and method validation, Talanta 178, 530-
537. 
[35] Deisenhofer, J. (1981) Crystallographic refinement and atomic models of a 
human Fc fragment and its complex with fragment B of protein A from 
Staphylococcus aureus at 2.9- and 2.8-A resolution, Biochemistry 20, 2361-
2370. 
[36] Malhotra, R., Wormald, M. R., Rudd, P. M., Fischer, P. B., Dwek, R. A., and 
Sim, R. B. (1995) Glycosylation changes of IgG associated with rheumatoid 
arthritis can activate complement via the mannose-binding protein, Nature 
Medicine 1, 237-243. 
[37] Gomes, M. M., Wall, S. B., Takahashi, K., Novak, J., Renfrow, M. B., and Herr, 
A. B. (2008) Analysis of IgA1 N-glycosylation and its contribution to 
FcalphaRI binding, Biochemistry 47, 11285-11299. 
[38] Saphire, E. O., Parren, P. W., Pantophlet, R., Zwick, M. B., Morris, G. M., Rudd, 
P. M., Dwek, R. A., Stanfield, R. L., Burton, D. R., and Wilson, I. A. (2001) 
Crystal structure of a neutralizing human IgG against HIV-1: a template for 
vaccine design, Science 293, 1155-1159. 
[39] Lewis, G. K., Pazgier, M., Evans, D. T., Ferrari, G., Bournazos, S., Parsons, M. 
S., Bernard, N. F., and Finzi, A. (2017) Beyond viral neutralization, AIDS 
Research and Human Retroviruses 33, 760-764. 
[40] Kanda, Y., Yamada, T., Mori, K., Okazaki, A., Inoue, M., Kitajima-Miyama, K., 
Kuni-Kamochi, R., Nakano, R., Yano, K., Kakita, S., Shitara, K., and Satoh, 
M. (2007) Comparison of biological activity among nonfucosylated 
therapeutic IgG1 antibodies with three different N-linked Fc oligosaccharides: 
the high-mannose, hybrid, and complex types, Glycobiology 17, 104-118. 
[41] Chen, C. L., Hsu, J. C., Lin, C. W., Wang, C. H., Tsai, M. H., Wu, C. Y., Wong, 
C. H., and Ma, C. (2017) Crystal Structure of a homogeneous IgG-Fc 
glycoform with the N-Glycan designed to maximize the antibody dependent 
cellular cytotoxicity, ACS Chemical Biology 12, 1335-1345. 
 206 
 
 
[42] Tsai, T. I., Li, S. T., Liu, C. P., Chen, K. Y., Shivatare, S. S., Lin, C. W., Liao, S. 
F., Lin, C. W., Hsu, T. L., Wu, Y. T., Tsai, M. H., Lai, M. Y., Lin, N. H., Wu, 
C. Y., and Wong, C. H. (2017) An effective bacterial fucosidase for 
glycoprotein remodeling, ACS Chemical Biology 12, 63-72. 
[43] Quast, I., Keller, C. W., Maurer, M. A., Giddens, J. P., Tackenberg, B., Wang, L. 
X., Munz, C., Nimmerjahn, F., Dalakas, M. C., and Lunemann, J. D. (2015) 
Sialylation of IgG Fc domain impairs complement-dependent cytotoxicity, 
The Journal of Clinical Investigation 125, 4160-4170. 
[44] Arnold, J. N., Wormald, M. R., Suter, D. M., Radcliffe, C. M., Harvey, D. J., 
Dwek, R. A., Rudd, P. M., and Sim, R. B. (2005) Human serum IgM 
glycosylation: identification of glycoforms that can bind to mannan-binding 
lectin, The Journal of Biological Chemistry 280, 29080-29087. 
[45] Campbell, I. K., Miescher, S., Branch, D. R., Mott, P. J., Lazarus, A. H., Han, 
D., Maraskovsky, E., Zuercher, A. W., Neschadim, A., Leontyev, D., 
McKenzie, B. S., and Kasermann, F. (2014) Therapeutic effect of IVIG on 
inflammatory arthritis in mice is dependent on the Fc portion and independent 
of sialylation or basophils, Journal of immunology 192, 5031-5038. 
[46] Bruckner, C., Lehmann, C., Dudziak, D., and Nimmerjahn, F. (2017) Sweet 
SIGNs: IgG glycosylation leads the way in IVIG-mediated resolution of 
inflammation, International Immunology 29, 499-509. 
[47] Hachiya, A., Kobayashi, N., Matsuzaki, S., Takeuchi, Y., Akazawa, Y., 
Shigemura, T., Motoki, N., Masumoto, J., and Agematsu, K. (2018) Analysis 
of biomarker serum levels in IVIG and infliximab refractory Kawasaki disease 
patients, Clinical Rheumatology 37, 1937-1943. 
[48] Seite, J. F., Cornec, D., Renaudineau, Y., Youinou, P., Mageed, R. A., and 
Hillion, S. (2010) IVIg modulates BCR signaling through CD22 and promotes 
apoptosis in mature human B lymphocytes, Blood 116, 1698-1704. 
[49] Schwab, I., Seeling, M., Biburger, M., Aschermann, S., Nitschke, L., and 
Nimmerjahn, F. (2012) B cells and CD22 are dispensable for the immediate 
antiinflammatory activity of intravenous immunoglobulins in vivo, European 
Journal of Immunology 42, 3302-3309. 
[50] Plomp, R., Hensbergen, P. J., Rombouts, Y., Zauner, G., Dragan, I., Koeleman, 
C. A., Deelder, A. M., and Wuhrer, M. (2014) Site-specific N-glycosylation 
analysis of human immunoglobulin E, Journal of Proteome Research 13, 536-
546. 
[51] Sutton, B. J., and Davies, A. M. (2015) Structure and dynamics of IgE-receptor 
interactions: Fc epsilon RI and CD23/Fc epsilon RII, Immunological Reviews 
268, 222-235. 
[52] Mackay, G. A., Hulett, M. D., Cook, J. P. D., Trist, H. M., Henry, A. J., 
McDonnell, J. M., Beavil, A. J., Beavil, R. L., Sutton, B. J., Hogarth, P. M., 
and Gould, H. J. (2002) Mutagenesis within human Fc epsilon RI alpha 
differentially affects human and murine IgE binding, Journal of Immunology 
168, 1787-1795. 
[53] Shade, K. T. C., Platzer, B., Washburn, N., Mani, V., Bartsch, Y. C., Conroy, 
M., Pagan, J. D., Bosques, C., Mempel, T. R., Fiebiger, E., and Anthony, R. 
 207 
 
 
M. (2015) A single glycan on IgE is indispensable for initiation of 
anaphylaxis, Journal of Experimental Medicine 212, 457-467. 
[54] Shade, K. T., and Anthony, R. (2018) Breaking the allergic cascade by 
modulating IgE glycosylation, Glycobiology 28, 1060-1060. 
[55] Butler, M., Quelhas, D., Critchley, A. J., Carchon, H., Hebestreit, H. F., Hibbert, 
R. G., Vilarinho, L., Teles, E., Matthijs, G., Schollen, E., Argibay, P., Harvey, 
D. J., Dwek, R. A., Jaeken, J., and Rudd, P. M. (2003) Detailed glycan 
analysis of serum glycoproteins of patients with congenital disorders of 
glycosylation indicates the specific defective glycan processing step and 
provides an insight into pathogenesis, Glycobiology 13, 601-622. 
[56] Parekh, R., Roitt, I., Isenberg, D., Dwek, R., and Rademacher, T. (1988) Age-
related galactosylation of the N-linked oligosaccharides of human serum IgG, 
Journal of Experimental Medicine 167, 1731-1736. 
[57] Jefferis, R., Lund, J., Mizutani, H., Nakagawa, H., Kawazoe, Y., Arata, Y., and 
Takahashi, N. (1990) A comparative study of the N-linked oligosaccharide 
structures of human IgG subclass proteins, The Biochemical Journal 268, 529-
537. 
[58] Wormald, M. R., Rudd, P. M., Harvey, D. J., Chang, S. C., Scragg, I. G., and 
Dwek, R. A. (1997) Variations in oligosaccharide-protein interactions in 
immunoglobulin G determine the site-specific glycosylation profiles and 
modulate the dynamic motion of the Fc oligosaccharides, Biochemistry 36, 
1370-1380. 
[59] Zauner, G., Selman, M. H., Bondt, A., Rombouts, Y., Blank, D., Deelder, A. M., 
and Wuhrer, M. (2013) Glycoproteomic analysis of antibodies, Molecular & 
Cellular Proteomics : MCP 12, 856-865. 
[60] Stadlmann, J., Weber, A., Pabst, M., Anderle, H., Kunert, R., Ehrlich, H. J., 
Schwarz, H. P., and Altmann, F. (2009) A close look at human IgG sialylation 
and subclass distribution after lectin fractionation, Proteomics 9, 4143-4153. 
[61] 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., and Deelder, A. M. (2007) Glycosylation 
profiling of immunoglobulin G (IgG) subclasses from human serum, 
Proteomics 7, 4070-4081. 
[62] Clausen, H., Wandall, H. H., Steentoft, C., Stanley, P., and Schnaar, R. L. (2015) 
Glycosylation Engineering, In Essentials of Glycobiology (rd, Varki, A., 
Cummings, R. D., Esko, J. D., Stanley, P., Hart, G. W., Aebi, M., Darvill, A. 
G., Kinoshita, T., Packer, N. H., Prestegard, J. H., Schnaar, R. L., and 
Seeberger, P. H., Eds.), pp 713-728, Cold Spring Harbor (NY). 
[63] Giddens, J. P., and Wang, L. X. (2015) Chemoenzymatic Glyco-engineering of 
monoclonal antibodies, Glyco-Engineering: Methods and Protocols 1321, 
375-387. 
[64] Yamane-Ohnuki, N., and Satoh, M. (2009) Production of therapeutic antibodies 
with controlled fucosylation, Mabs-Austin 1, 230-236. 
[65] von Horsten, H. H., Ogorek, C., Blanchard, V., Demmler, C., Giese, C., Winkler, 
K., Kaup, M., Berger, M., Jordan, I., and Sandig, V. (2010) Production of 
non-fucosylated antibodies by co-expression of heterologous GDP-6-deoxy-
D-lyxo-4-hexulose reductase, Glycobiology 20, 1607-1618. 
 208 
 
 
[66] Shinkawa, T., Nakamura, K., Yamane, N., Shoji-Hosaka, E., Kanda, Y., 
Sakurada, M., Uchida, K., Anazawa, H., Satoh, M., Yamasaki, M., Hanai, N., 
and Shitara, K. (2003) The absence of fucose but not the presence of galactose 
or bisecting N-acetylglucosamine of human IgG1 complex-type 
oligosaccharides shows the critical role of enhancing antibody-dependent 
cellular cytotoxicity, Journal of Biological Chemistry 278, 3466-3473. 
[67] Yver, A., Homery, M. C., Fuseau, E., Guemas, E., Dhainaut, F., Quagliaroli, D., 
Beliard, R., and Prost, J. F. (2012) Pharmacokinetics and safety of roledumab, 
a novel human recombinant monoclonal anti-RhD antibody with an optimized 
Fc for improved engagement of FCgammaRIII, in healthy volunteers, Vox 
Sang 103, 213-222. 
[68] Sharman, J. P., Farber, C. M., Mahadevan, D., Schreeder, M. T., Brooks, H. D., 
Kolibaba, K. S., Fanning, S., Klein, L., Greenwald, D. R., Sportelli, P., 
Miskin, H. P., Weiss, M. S., and Burke, J. M. (2017) Ublituximab (TG-1101), 
a novel glycoengineered anti-CD20 antibody, in combination with ibrutinib is 
safe and highly active in patients with relapsed and/or refractory chronic 
lymphocytic leukaemia: results of a phase 2 trial, British Journal of 
Haematology 176, 412-420. 
[69] Ferrara, C., Brunker, P., Suter, T., Moser, S., Puntener, U., and Umana, P. (2006) 
Modulation of therapeutic antibody effector functions by glycosylation 
engineering: influence of Golgi enzyme localization domain and co-
expression of heterologous beta1, 4-N-acetylglucosaminyltransferase III and 
Golgi alpha-mannosidase II, Biotechnology and Bioengineering 93, 851-861. 
[70] Ratner, M. (2014) Genentech's glyco-engineered antibody to succeed Rituxan, 
Nature Biotechnology 32, 6-7. 
[71] Ureshino, H., Kamachi, K., and Kimura, S. (2019) Mogamulizumab for the 
Treatment of Adult T-cell Leukemia/Lymphoma, Clinical Lymphoma, 
Myeloma & Leukemia, in press. 
[72] Louie, S., Haley, B., Marshall, B., Heidersbach, A., Yim, M., Brozynski, M., 
Tang, D., Lam, C., Petryniak, B., Shaw, D., Shim, J., Miller, A., Lowe, J. B., 
Snedecor, B., and Misaghi, S. (2017) FX knockout CHO hosts can express 
desired ratios of fucosylated or afucosylated antibodies with high titers and 
comparable product quality, Biotechnology and Bioengineering 114, 632-644. 
[73] von Horsten, H. H., Ogorek, C., Blanchard, V., Demmler, C., Giese, C., Winkler, 
K., Kaup, M., Berger, M., Jordan, I., and Sandig, V. (2010) Production of 
non-fucosylated antibodies by co-expression of heterologous GDP-6-deoxy-
D-lyxo-4-hexulose reductase, Glycobiology 20, 1607-1618. 
[74] Okeley, N. M., Alley, S. C., Anderson, M. E., Boursalian, T. E., Burke, P. J., 
Emmerton, K. M., Jeffrey, S. C., Klussman, K., Law, C. L., Sussman, D., 
Toki, B. E., Westendorf, L., Zeng, W., Zhang, X., Benjamin, D. R., and 
Senter, P. D. (2013) Development of orally active inhibitors of protein and 
cellular fucosylation, Proceedings of the National Academy of Sciences of the 
United States of America 110, 5404-5409. 
[75] Allen, J. G., Mujacic, M., Frohn, M. J., Pickrell, A. J., Kodama, P., Bagal, D., 
San Miguel, T., Sickmier, E. A., Osgood, S., Swietlow, A., Li, V., Jordan, J. 
B., Kim, K. W., Rousseau, A. C., Kim, Y. J., Caille, S., Achmatowicz, M., 
 209 
 
 
Thiel, O., Fotsch, C. H., Reddy, P., and McCarter, J. D. (2016) Facile 
Modulation of Antibody Fucosylation with Small Molecule Fucostatin 
Inhibitors and Cocrystal Structure with GDP-Mannose 4,6-Dehydratase, ACS 
Chemical Biology 11, 2734-2743. 
[76] McKenzie, N. C., Scott, N. E., John, A., White, J. M., and Goddard-Borger, E. 
D. (2018) Synthesis and use of 6,6,6-trifluoro-L-fucose to block core-
fucosylation in hybridoma cell lines, Carbohydrate Research 465, 4-9. 
[77] Hossler, P., Chumsae, C., Racicot, C., Ouellette, D., Ibraghimov, A., Serna, D., 
Mora, A., McDermott, S., Labkovsky, B., Scesney, S., Grinnell, C., Preston, 
G., Bose, S., and Carrillo, R. (2017) Arabinosylation of recombinant human 
immunoglobulin-based protein therapeutics, Mabs-Austin 9, 715-734. 
[78] Iizuka, M., Ogawa, S., Takeuchi, A., Nakakita, S., Kubo, Y., Miyawaki, Y., 
Hirabayashi, J., and Tomita, M. (2009) Production of a recombinant mouse 
monoclonal antibody in transgenic silkworm cocoons, The FEBS Journal 276, 
5806-5820. 
[79] Cox, K. M., Sterling, J. D., Regan, J. T., Gasdaska, J. R., Frantz, K. K., Peele, C. 
G., Black, A., Passmore, D., Moldovan-Loomis, C., Srinivasan, M., Cuison, 
S., Cardarelli, P. M., and Dickey, L. F. (2006) Glycan optimization of a 
human monoclonal antibody in the aquatic plant Lemna minor, Nature 
Biotechnology 24, 1591-1597. 
[80] Ward, M., Lin, C., Victoria, D. C., Fox, B. P., Fox, J. A., Wong, D. L., 
Meerman, H. J., Pucci, J. P., Fong, R. B., Heng, M. H., Tsurushita, N., 
Gieswein, C., Park, M., and Wang, H. (2004) Characterization of humanized 
antibodies secreted by Aspergillus niger, Applied and Environmental 
Microbiology 70, 2567-2576. 
[81] Wildt, S., and Gerngross, T. U. (2005) The humanization of N-glycosylation 
pathways in yeast, Nature Reviews. Microbiology 3, 119-128. 
[82] Kaneko, Y., Nimmerjahn, F., and Ravetch, J. V. (2006) Anti-inflammatory 
activity of immunoglobulin G resulting from Fc sialylation, Science 313, 670-
673. 
[83] Lund, J., Takahashi, N., Pound, J. D., Goodall, M., and Jefferis, R. (1996) 
Multiple interactions of IgG with its core oligosaccharide can modulate 
recognition by complement and human Fc gamma receptor I and influence the 
synthesis of its oligosaccharide chains, Journal of Immunology 157, 4963-
4969. 
[84] Chung, C. Y., Wang, Q., Yang, S., Ponce, S. A., Kirsch, B. J., Zhang, H., and 
Betenbaugh, M. J. (2017) Combinatorial genome and protein engineering 
yields monoclonal antibodies with hypergalactosylation from CHO cells, 
Biotechnology and Bioengineering 114, 2848-2856. 
[85] Giddens, J. P., Lomino, J. V., Amin, M. N., and Wang, L. X. (2016) Endo-F3 
Glycosynthase Mutants Enable Chemoenzymatic Synthesis of Core-
fucosylated Triantennary Complex Type Glycopeptides and Glycoproteins, 
The Journal of biological chemistry 291, 9356-9370. 
[86] Beck, A., Wagner-Rousset, E., Bussat, M. C., Lokteff, M., Klinguer-Hamour, C., 
Haeuw, J. F., Goetsch, L., Wurch, T., Van Dorsselaer, A., and Corvaia, N. 
(2008) Trends in glycosylation, glycoanalysis and glycoengineering of 
 210 
 
 
therapeutic antibodies and Fc-fusion proteins, Current Pharmaceutical 
Biotechnology 9, 482-501. 
[87] Hodoniczky, J., Zheng, Y. Z., and James, D. C. (2005) Control of recombinant 
monoclonal antibody effector functions by Fc N-glycan remodeling in vitro, 
Biotechnology Progress 21, 1644-1652. 
[88] Washburn, N., Schwab, I., Ortiz, D., Bhatnagar, N., Lansing, J. C., Medeiros, A., 
Tyler, S., Mekala, D., Cochran, E., Sarvaiya, H., Garofalo, K., Meccariello, 
R., Meador, J. W., 3rd, Rutitzky, L., Schultes, B. C., Ling, L., Avery, W., 
Nimmerjahn, F., Manning, A. M., Kaundinya, G. V., and Bosques, C. J. 
(2015) Controlled tetra-Fc sialylation of IVIg results in a drug candidate with 
consistent enhanced anti-inflammatory activity, Proceedings of the National 
Academy of Sciences of the United States of America 112, E1297-1306. 
[89] Fairbanks, A. J. (2017) The ENGases: versatile biocatalysts for the production of 
homogeneous N-linked glycopeptides and glycoproteins, Chemical Society 
Reviews 46, 5128-5146. 
[90] Lombard, V., Golaconda Ramulu, H., Drula, E., Coutinho, P. M., and Henrissat, 
B. (2014) The carbohydrate-active enzymes database (CAZy) in 2013, 
Nucleic Acids Research 42, D490-495. 
[91] Henrissat, B., and Davies, G. (1997) Structural and sequence-based classification 
of glycoside hydrolases, Current Opinion in Structural Biology 7, 637-644. 
[92] Yin, J., Li, L., Shaw, N., Li, Y., Song, J. K., Zhang, W., Xia, C., Zhang, R., 
Joachimiak, A., Zhang, H. C., Wang, L. X., Liu, Z. J., and Wang, P. (2009) 
Structural basis and catalytic mechanism for the dual functional endo-beta-N-
acetylglucosaminidase A, Plos One 4, e4658. 
[93] Wang, L. X. (2008) Chemoenzymatic synthesis of glycopeptides and 
glycoproteins through endoglycosidase-catalyzed transglycosylation, 
Carbohydrate Research 343, 1509-1522. 
[94] Li, C., and Wang, L. X. (2018) Chemoenzymatic Methods for the Synthesis of 
Glycoproteins, Chem Rev 118, 8359-8413. 
[95] Tews, I., vanScheltinga, A. C. T., Perrakis, A., Wilson, K. S., and Dijkstra, B. 
W. (1997) Substrate-assisted catalysis unifies two families of chitinolytic 
enzymes, Journal of the American Chemical Society 119, 7954-7959. 
[96] Li, C., and Wang, L. X. (2016) Endoglycosidases for the synthesis of 
polysaccharides and glycoconjugates, Advances in Carbohydrate Chemistry 
and Biochemistry 73, 73-116. 
[97] Fujita, M., Shoda, S., Haneda, K., Inazu, T., Takegawa, K., and Yamamoto, K. 
(2001) A novel disaccharide substrate having 1,2-oxazoline moiety for 
detection of transglycosylating activity of endoglycosidases, Biochimica Et 
Biophysica Acta-General Subjects 1528, 9-14. 
[98] Wang, L. X. (2011) The Amazing Transglycosylation Activity of Endo-beta-N-
acetylglucosaminidases, Trends in Glycoscience and Glycotechnology : TIGG 
23, 33-52. 
[99] Umekawa, M., Huang, W., Li, B., Fujita, K., Ashida, H., Wang, L. X., and 
Yamamoto, K. (2008) Mutants of Mucor hiemalis endo-beta-N-
acetylglucosaminidase show enhanced transglycosylation and glycosynthase-
like activities, Journal of Biological Chemistry 283, 4469-4479. 
 211 
 
 
[100] Umekawa, M., Li, C., Higashiyama, T., Huang, W., Ashida, H., Yamamoto, K., 
and Wang, L. X. (2010) Efficient glycosynthase mutant derived from Mucor 
hiemalis endo-beta-N-acetylglucosaminidase capable of transferring 
oligosaccharide from both sugar oxazoline and natural N-glycan, Journal of 
Biological Chemistry 285, 511-521. 
[101] Abbott, D. W., Macauley, M. S., Vocadlo, D. J., and Boraston, A. B. (2009) 
Streptococcus pneumoniae endohexosaminidase D, structural and mechanistic 
insight into substrate-assisted catalysis in family 85 glycoside hydrolases, 
Journal of Biological Chemistry 284, 11676-11689. 
[102] Huang, W., Yang, Q., Umekawa, M., Yamamoto, K., and Wang, L. X. (2010) 
Arthrobacter endo-beta-N-acetylglucosaminidase shows transglycosylation 
activity on complex-type N-glycan oxazolines: one-pot conversion of 
ribonuclease B to sialylated ribonuclease C, Chembiochem : a European 
Journal of Chemical Biology 11, 1350-1355. 
[103] Huang, W., Giddens, J., Fan, S. Q., Toonstra, C., and Wang, L. X. (2012) 
Chemoenzymatic glycoengineering of intact IgG antibodies for gain of 
functions, Journal of the American Chemical Society 134, 12308-12318. 
[104] Wang, L. X., and Amin, M. N. (2014) Chemical and chemoenzymatic synthesis 
of glycoproteins for deciphering functions, Chemistry & Biology 21, 51-66. 
[105] Eshima, Y., Higuchi, Y., Kinoshita, T., Nakakita, S., and Takegawa, K. (2015) 
Transglycosylation activity of glycosynthase mutants of endo-beta-N-
acetylglucosaminidase from Coprinopsis cinerea, Plos One 10, e0132859. 
[106] Wei, Y., Li, C., Huang, W., Li, B., Strome, S., and Wang, L. X. (2008) 
Glycoengineering of human IgG1-Fc through combined yeast expression and 
in vitro chemoenzymatic glycosylation, Biochemistry 47, 10294-10304. 
[107] Higuchi, Y., Eshima, Y., Huang, Y., Kinoshita, T., Sumiyoshi, W., Nakakita, 
S., and Takegawa, K. (2017) Highly efficient transglycosylation of sialo-
complex-type oligosaccharide using Coprinopsis cinerea endoglycosidase and 
sugar oxazoline, Biotechnology Letters 39, 157-162. 
[108] Goodfellow, J. J., Baruah, K., Yamamoto, K., Bonomelli, C., Krishna, B., 
Harvey, D. J., Crispin, M., Scanlan, C. N., and Davis, B. G. (2012) An 
endoglycosidase with alternative glycan specificity allows broadened 
glycoprotein remodelling, Journal of the American Chemical Society 134, 
8030-8033. 
[109] Lin, C. W., Tsai, M. H., Li, S. T., Tsai, T. I., Chu, K. C., Liu, Y. C., Lai, M. Y., 
Wu, C. Y., Tseng, Y. C., Shivatare, S. S., Wang, C. H., Chao, P., Wang, S. Y., 
Shih, H. W., Zeng, Y. F., You, T. H., Liao, J. Y., Tu, Y. C., Lin, Y. S., 
Chuang, H. Y., Chen, C. L., Tsai, C. S., Huang, C. C., Lin, N. H., Ma, C., Wu, 
C. Y., and Wong, C. H. (2015) A common glycan structure on 
immunoglobulin G for enhancement of effector functions, Proceedings of the 
National Academy of Sciences of the United States of America 112, 10611-
10616. 
[110] Higuchi, Y., Eshima, Y., Huang, Y., Kinoshita, T., Sumiyoshi, W., Nakakita, S. 
I., and Takegawa, K. (2017) Highly efficient transglycosylation of sialo-
complex-type oligosaccharide using Coprinopsis cinerea endoglycosidase and 
sugar oxazoline, Biotechnology Letters 39, 157-162. 
 212 
 
 
[111] Iwamoto, M., Sekiguchi, Y., Nakamura, K., Kawaguchi, Y., Honda, T., and 
Hasegawa, J. (2018) Generation of efficient mutants of endoglycosidase from 
Streptococcus pyogenes and their application in a novel one-pot 
transglycosylation reaction for antibody modification, Plos One 13, e0193534. 
[112] Manabe, S., Yamaguchi, Y., Abe, J., Matsumoto, K., and Ito, Y. (2018) 
Acceptor range of endo-beta-N-acetylglucosaminidase mutant endo-CC 
N180H: from monosaccharide to antibody, Royal Society open science 5, 
171521. 
[113] Shade, K. T., Platzer, B., Washburn, N., Mani, V., Bartsch, Y. C., Conroy, M., 
Pagan, J. D., Bosques, C., Mempel, T. R., Fiebiger, E., and Anthony, R. M. 
(2015) A single glycan on IgE is indispensable for initiation of anaphylaxis, 
The Journal of Experimental Medicine 212, 457-467. 
[114] Collin, M., and Olsen, A. (2001) EndoS, a novel secreted protein from 
Streptococcus pyogenes with endoglycosidase activity on human IgG, The 
EMBO Journal 20, 3046-3055. 
[115] Allhorn, M., Briceno, J. G., Baudino, L., Lood, C., Olsson, M. L., Izui, S., and 
Collin, M. (2010) The IgG-specific endoglycosidase EndoS inhibits both 
cellular and complement-mediated autoimmune hemolysis, Blood 115, 5080-
5088. 
[116] Trastoy, B., Lomino, J. V., Pierce, B. G., Carter, L. G., Gunther, S., Giddens, J. 
P., Snyder, G. A., Weiss, T. M., Weng, Z., Wang, L. X., and Sundberg, E. J. 
(2014) Crystal structure of Streptococcus pyogenes EndoS, an 
immunomodulatory endoglycosidase specific for human IgG antibodies, 
Proceedings of the National Academy of Sciences of the United States of 
America 111, 6714-6719. 
[117] Wadood, A., Ghufran, M., Khan, A., Azam, S. S., Jelani, M., and Uddin, R. 
(2018) Selective glycosidase inhibitors: A patent review (2012-present), 
International Journal of Biological Macromolecules 111, 82-91. 
[118] Knapp, S., Vocadlo, D., Gao, Z. N., Kirk, B., Lou, J. P., and Withers, S. G. 
(1996) NAG-thiazoline, an N-acetyl-beta-hexosaminidase inhibitor that 
implicates acetamido participation, Journal of the American Chemical Society 
118, 6804-6805. 
[119] Macdonald, J. M., Tarling, C. A., Taylor, E. J., Dennis, R. J., Myers, D. S., 
Knapp, S., Davies, G. J., and Withers, S. G. (2010) Chitinase Inhibition by 
Chitobiose and Chitotriose Thiazolines, Angewandte Chemie-International 
Edition 49, 2599-2602. 
[120] Reid, C. W., Blackburn, N. T., Legaree, B. A., Auzanneaua, F. I., and Clarke, 
A. J. (2004) Inhibition of membrane-bound lytic transglycosylase B by NAG-
thiazoline, FEBS Letters 574, 73-79. 
[121] Li, B., Takegawa, K., Suzuki, T., Yamamoto, K., and Wang, L. X. (2008) 
Synthesis and inhibitory activity of oligosaccharide thiazolines as a class of 
mechanism-based inhibitors for endo-beta-N-acetylglucosaminidases, 
Bioorganic & Medicinal Chemistry 16, 4670-4675. 
[122] Parsons, T. B., Struwe, W. B., Gault, J., Yamamoto, K., Taylor, T. A., Raj, R., 
Wals, K., Mohammed, S., Robinson, C. V., Benesch, J. L., and Davis, B. G. 
 213 
 
 
(2016) Optimal synthetic glycosylation of a therapeutic antibody, Angewandte 
Chemie-International Edition 55, 2361-2367. 
[123] Tong, X., Li, T., Li, C., and Wang, L. X. (2018) Generation and comparative 
kinetic analysis of new glycosynthase mutants from Streptococcus pyogenes 
endoglycosidases for antibody glycoengineering, Biochemistry 57, 5239-5246. 
[124] Li, C., Zhu, S. L., Tong, X., Ou, C., and Wang, L. X. (2017) A facile synthesis 
of a complex type N-glycan thiazoline as an effective inhibitor against the 
antibody-deactivating endo--N-acetylglucosaminidases, Journal of 
Carbohydrate Chemistry 36, 336-346. 
[125] Zhao, X. Y., Li, G. S., and Liang, S. F. (2013) Several affinity tags commonly 
used in the chromatographic purification, Journal of Analytical Methods in 
Chemistry 440, 83-98. 
[126] Yan, B. X., and Sun, Y. Q. (1997) Glycine residues provide flexibility for 
enzyme active sites, The Journal of Biological Chemistry 272, 3190-3194. 
[127] Seko, A., Koketsu, M., Nishizono, M., Enoki, Y., Ibrahim, H. R., Juneja, L. R., 
Kim, M., and Yamamoto, T. (1997) Occurence of a sialylglycopeptide and 
free sialylglycans in hen's egg yolk, Biochimica et Biophysica Acta 1335, 23-
32. 
[128] Li, B., Takegawa, K., Suzuki, T., Yamamoto, K., and Wang, L. X. (2008) 
Synthesis and inhibitory activity of oligosaccharide thiazolines as a class of 
mechanism-based inhibitors for endo-beta-N-acetylglucosaminidases, 
Bioorganic & Medicinal Chemistry 16, 4670-4675. 
[129] Huang, Y., and Orlando, R. (2017) Kinetics of N-Glycan release from human 
immunoglobulin G (IgG) by PNGase F: all glycans are not created equal, 
Journal of Biomolecular Techniques : JBT 28, 150-157. 
[130] Trastoy, B., Klontz, E., Orwenyo, J., Marina, A., Wang, L. X., Sundberg, E. J., 
and Guerin, M. E. (2018) Structural basis for the recognition of complex-type 
N-glycans by Endoglycosidase S, Nature Communications 9, 1874. 
[131] Liu, L., Prudden, A. R., Capicciotti, C. J., Bosman, G. P., Yang, J. Y., Chapla, 
D. G., Moremen, K. W., and Boons, G. J. (2019) Streamlining the 
chemoenzymatic synthesis of complex N-glycans by a stop and go strategy, 
Nature Chemistry 11, 161-169. 
[132] Liu, C. P., Tsai, T. I., Cheng, T., Shivatare, V. S., Wu, C. Y., Wu, C. Y., and 
Wong, C. H. (2018) Glycoengineering of antibody (Herceptin) through yeast 
expression and in vitro enzymatic glycosylation, Proceedings of the National 
Academy of Sciences of the United States of America 115, 720-725. 
[133] Wang, N., Seko, A., Daikoku, S., Kanie, O., Takeda, Y., and Ito, Y. (2016) 
Non-enzymatic reaction of glycosyl oxazoline with peptides, Carbohydrate 
Research 436, 31-35. 
[134] Ge, X., Sirich, T. L., Beyer, M. K., Desaire, H., and Leary, J. A. (2001) A 
strategy for the determination of enzyme kinetics using electrospray ionization 
with an ion trap mass spectrometer, Analytical Chemistry 73, 5078-5082. 
[135] Danan, L. M., Yu, Z., Hoffhines, A. J., Moore, K. L., and Leary, J. A. (2008) 
Mass spectrometric kinetic analysis of human tyrosylprotein sulfotransferase-
1 and -2, Journal of the American Society for Mass Spectrometry 19, 1459-
1466. 
 214 
 
 
[136] Pi, N., Armstrong, J. I., Bertozzi, C. R., and Leary, J. A. (2002) Kinetic analysis 
of NodST sulfotransferase using an electrospray ionization mass spectrometry 
assay, Biochemistry 41, 13283-13288. 
[137] Guo, W. L., Tang, F., Qin, K., Zhou, M., Le, Z. P., and Huang, W. (2017) 
Glycoengineering and glycosite-specific labeling of serum IgGs from various 
species, Carbohydrate Research 446, 32-39. 
[138] Sjogren, J., Cosgrave, E. F., Allhorn, M., Nordgren, M., Bjork, S., Olsson, F., 
Fredriksson, S., and Collin, M. (2015) EndoS and EndoS2 hydrolyze Fc-
glycans on therapeutic antibodies with different glycoform selectivity and can 
be used for rapid quantification of high-mannose glycans, Glycobiology 25, 
1053-1063. 
[139] Li, T., Tong, X., Yang, Q., Giddens, J. P., and Wang, L. X. (2016) 
Glycosynthase mutants of endoglycosidase S2 show potent transglycosylation 
activity and remarkably relaxed substrate specificity for antibody 
glycosylation remodeling, The Journal of Biological Chemistry 291, 16508-
16518. 
[140] Tong, X., Li, T., Orwenyo, J., Toonstra, C., and Wang, L. X. (2018) One-pot 
enzymatic glycan remodeling of a therapeutic monoclonal antibody by 
endoglycosidase S (Endo-S) from Streptococcus pyogenes, Bioorganic & 
Medicinal Chemistry 26, 1347-1355.  
[141] Shen, A., Lupardus, P. J., Morell, M., Ponder, E. L., Sadaghiani, A. M., Garcia, 
K. C., and Bogyo, M. (2009) Simplified, enhanced protein purification using 
an inducible, autoprocessing enzyme tag, Plos One 4, e8119. 
[142] Eshima, Y., Higuchi, Y., Kinoshita, T., Nakakita, S. I., and Takegawa, K. 
(2015) Transglycosylation activity of glycosynthase mutants of endo-beta-N-
acetylglucosaminidase from Coprinopsis cinerea, Plos One 10, e0132859. 
[143] Biancucci, M., Dolores, J. S., Wong, J., Grimshaw, S., Anderson, W. F., 
Satchell, K. J., and Kwon, K. (2017) New ligation independent cloning 
vectors for expression of recombinant proteins with a self-cleaving 
CPD/6xHis-tag, BMC Biotechnology 17, 1. 
[144] Li, T., Tong, X., Yang, Q., Giddens, J. P., and Wang, L. X. (2016) 
Glycosynthase mutants of endoglycosidase S2 show potent transglycosylation 
activity and remarkably relaxed substrate specificity for antibody 
glycosylation remodeling, The Journal of Biological Chemistry 291, 16508-
16518. 
[145] Williams, R. L., Greene, S. M., and Mcpherson, A. (1987) The Crystal-
Structure of Ribonuclease-B at 2.5-a Resolution, Journal of Biological 
Chemistry 262, 16020-16031. 
[146] Yang, Q., An, Y. M., Zhu, S. L., Zhang, R. S., Loke, C. M., Cipollo, J. F., and 
Wang, L. X. (2017) Glycan Remodeling of Human Erythropoietin (EPO) 
Through Combined Mammalian cell engineering and chemoenzymatic 
transglycosylation, ACS Chemical Biology 12, 1665-1673. 
[147] Li, C., Zhu, S., Ma, C., and Wang, L. X. (2017) Designer alpha1,6-fucosidase 
mutants enable direct core fucosylation of intact N-glycopeptides and N-
glycoproteins, Journal of the American Chemical Society 139, 15074-15087. 
 215 
 
 
[148] Platts-Mills, T. A., Heymann, P. W., Commins, S. P., and Woodfolk, J. A. 
(2016) The discovery of IgE 50 years later, Annals of Allergy, Asthma & 
Immunology : Official Publication of the American College of Allergy, 
Asthma, & Immunology 116, 179-182. 
[149] Mukai, K., Tsai, M., Starkl, P., Marichal, T., and Galli, S. J. (2016) IgE and 
mast cells in host defense against parasites and venoms, Seminars in 
Immunopathology 38, 581-603. 
[150] Gould, H. J., Sutton, B. J., Beavil, A. J., Beavil, R. L., McCloskey, N., Coker, 
H. A., Fear, D., and Smurthwaite, L. (2003) The biology of IgE and the basis 
of allergic disease, Annual Review of Immunology 21, 579-628. 
[151] Cockcroft, D. W. (2018) Epidemic thunderstorm asthma, The Lancet. Planetary 
Health 2, e236-e237. 
[152] Karagiannis, S. N., Josephs, D. H., Bax, H. J., and Spicer, J. F. (2017) 
Therapeutic IgE Antibodies: Harnessing a macrophage-mediated immune 
surveillance mechanism against cancer, Cancer Research 77, 2779-2783. 
[153] Gould, H. J., and Sutton, B. J. (2008) IgE in allergy and asthma today, Nature 
Reviews. Immunology 8, 205-217. 
[154] Dhaliwal, B., Pang, M. O., Keeble, A. H., James, L. K., Gould, H. J., 
McDonnell, J. M., Sutton, B. J., and Beavil, A. J. (2017) IgE binds 
asymmetrically to its B cell receptor CD23, Scientific Reports 7, 45533. 
[155] Chen, J. B., Ramadani, F., Pang, M. O. Y., Beavil, R. L., Holdom, M. D., 
Mitropoulou, A. N., Beavil, A. J., Gould, H. J., Chang, T. W., Sutton, B. J., 
McDonnell, J. M., and Davies, A. M. (2018) Structural basis for selective 
inhibition of immunoglobulin E-receptor interactions by an anti-IgE antibody, 
Scientific Reports 8, 11548. 
[156] Tu, Y., Salim, S., Bourgeois, J., Di Leo, V., Irvine, E. J., Marshall, J. K., and 
Perdue, M. H. (2005) CD23-mediated IgE transport across human intestinal 
epithelium: inhibition by blocking sites of translation or binding, 
Gastroenterology 129, 928-940. 
[157] Arnold, J. N., Radcliffe, C. M., Wormald, M. R., Royle, L., Harvey, D. J., 
Crispin, M., Dwek, R. A., Sim, R. B., and Rudd, P. M. (2004) The 
glycosylation of human serum IgD and IgE and the accessibility of identified 
oligomannose structures for interaction with mannan-binding lectin, J 
Immunol 173, 6831-6840. 
[158] Fridriksson, E. K., Beavil, A., Holowka, D., Gould, H. J., Baird, B., and 
McLafferty, F. W. (2000) Heterogeneous glycosylation of immunoglobulin E 
constructs characterized by top-down high-resolution 2-D mass spectrometry, 
Biochemistry 39, 3369-3376. 
[159] Dore, K. A., Davies, A. M., Drinkwater, N., Beavil, A. J., McDonnell, J. M., 
and Sutton, B. J. (2017) Thermal sensitivity and flexibility of the Cepsilon-3 
domains in immunoglobulin E, Biochimica et biophysica acta. Proteins and 
proteomics 1865, 1336-1347. 
[160] Wan, T., Beavil, R. L., Fabiane, S. M., Beavil, A. J., Sohi, M. K., Keown, M., 
Young, R. J., Henry, A. J., Owens, R. J., Gould, H. J., and Sutton, B. J. (2002) 
The crystal structure of IgE Fc reveals an asymmetrically bent conformation, 
Nature Immunology 3, 681-686. 
 216 
 
 
[161] Holdom, M. D., Davies, A. M., Nettleship, J. E., Bagby, S. C., Dhaliwal, B., 
Girardi, E., Hunt, J., Gould, H. J., Beavil, A. J., McDonnell, J. M., Owens, R. 
J., and Sutton, B. J. (2011) Conformational changes in IgE contribute to its 
uniquely slow dissociation rate from receptor FcepsilonRI, Nature Structural 
& Molecular Biology 18, 571-576. 
[162] Basu, M., Hakimi, J., Dharm, E., Kondas, J. A., Tsien, W. H., Pilson, R. S., Lin, 
P., Gilfillan, A., Haring, P., Braswell, E. H., Nettleton, M. Y., and Kochan, J. 
P. (1993) Purification and characterization of human recombinant IgE-Fc 
fragments that bind to the human high-affinity IgE receptor, Journal of 
Biological Chemistry 268, 13118-13127. 
[163] Hunt, J., Beavil, R. L., Calvert, R. A., Gould, H. J., Sutton, B. J., and Beavil, A. 
J. (2005) Disulfide linkage controls the affinity and stoichiometry of IgE Fc-
epsilon3-4 binding to FcepsilonRI, The Journal of Biological Chemistry 280, 
16808-16814. 
[164] Zosky, G. R., and Sly, P. D. (2007) Animal models of asthma, Clinical and 
Experimental Allergy : Journal of the British Society for Allergy and Clinical 
Immunology 37, 973-988. 
[165] Taner, S. B., Pando, M. J., Roberts, A., Schellekens, J., Marsh, S. G. E., 
Malmberg, K. J., Parham, P., and Brodsky, F. M. (2011) Interactions of NK 
cell receptor KIR3DL1*004 with chaperones and conformation-specific 
antibody reveal a functional folded state as well as predominant intracellular 
retention, Journal of Immunology 186, 62-72. 
[166] Sutton, B. J., Davies, A. M., Bax, H. J., and Karagiannis, S. N. (2019) IgE 
antibodies: from structure to function and clinical translation, Antibodies 8, 
19. 
[167] Taylor, A. I., Fabiane, S. M., Sutton, B. J., and Calvert, R. A. (2009) The 
crystal structure of an avian IgY-Fc fragment reveals conservation with both 
mammalian IgG and IgE, Biochemistry 48, 558-562. 
[168] Aikawa, J., Takeda, Y., Matsuo, I., and Ito, Y. (2014) Trimming of 
glucosylated N-glycans by human ER alpha1,2-mannosidase I, Journal of 
Biochemistry 155, 375-384. 
[169] Feige, M. J., Nath, S., Catharino, S. R., Weinfurtner, D., Steinbacher, S., and 
Buchner, J. (2009) Structure of the murine unglycosylated IgG1 Fc fragment, 
Journal of Molecular Biology 391, 599-608. 
[170] Rosenwasser, L. J., and Meng, J. F. (2005) Anti-CD23, Clinical Reviews in 
Allergy & Immunology 29, 61-72. 
[171] Vercelli, D., Helm, B., Marsh, P., Padlan, E., Geha, R. S., and Gould, H. (1989) 
The B-Cell binding-site on human immunoglobulin-E, Nature 338, 649-651. 
[172] Hibbert, R. G., Teriete, P., Grundy, G. J., Beavil, R. L., Reljic, R., Holers, V. 
M., Hannan, J. P., Sutton, B. J., Gould, H. J., and McDonnell, J. M. (2005) 
The structure of human CD23 and its interactions with IgE and CD21, The 
Journal of Experimental Medicine 202, 751-760. 
[173] Dhaliwal, B., Yuan, D. P., Pang, M. O. Y., Henry, A. J., Cain, K., Oxbrow, A., 
Fabiane, S. M., Beavil, A. J., McDonnell, J. M., Gould, H. J., and Sutton, B. J. 
(2012) Crystal structure of IgE bound to its B-cell receptor CD23 reveals a 
mechanism of reciprocal allosteric inhibition with high affinity receptor Fc 
 217 
 
 
epsilon RI, Proceedings of the National Academy of Sciences of the United 
States of America 109, 12686-12691. 
[174] Epp, A., Sullivan, K. C., Herr, A. B., and Strait, R. T. (2016) Immunoglobulin 
glycosylation effects in allergy and immunity, Current Allergy and Asthma 
Reports 16, 79. 
[175] Davies, A. M., Allan, E. G., Keeble, A. H., Delgado, J., Cossins, B. P., 
Mitropoulou, A. N., Pang, M. O. Y., Ceska, T., Beavil, A. J., Craggs, G., 
Westwood, M., Henry, A. J., McDonnell, J. M., and Sutton, B. J. (2017) 
Allosteric mechanism of action of the therapeutic anti-IgE antibody 
omalizumab, The Journal of Biological Chemistry 292, 9975-9987. 
[176] Yang, Q., and Wang, L. X. (2016) Mammalian-1,6-Fucosyltransferase (FUT8) 
Is the Sole Enzyme Responsible for the N-Acetylglucosaminyltransferase I-
independent Core Fucosylation of High-mannose N-Glycans, The Journal of 
Biological Chemistry 291, 11064-11071. 
[177] Li, T., DiLillo, D. J., Bournazos, S., Giddens, J. P., Ravetch, J. V., and Wang, 
L. X. (2017) Modulating IgG effector function by Fc glycan engineering, 
Proceedings of the National Academy of Sciences of the United States of 
America 114, 3485-3490. 
[178] Tong, X., Li, T., Orwenyo, J., Toonstra, C., and Wang, L. X. (2017) One-pot 
enzymatic glycan remodeling of a therapeutic monoclonal antibody by 
endoglycosidase S (Endo-S) from Streptococcus pyogenes, Bioorganic & 
Medicinal Chemistry 26, 1347-1355.  
[179] Klontz, E. H., Trastoy, B., Deredge, D., Fields, J. K., Li, C., Orwenyo, J., 
Marina, A., Beadenkopf, R., Gunther, S., Flores, J., Wintrode, P. L., Wang, L. 
X., Guerin, M. E., and Sundberg, E. J. (2019) Molecular basis of broad 
spectrum N-Glycan specificity and processing of therapeutic IgG monoclonal 
antibodies by endoglycosidase S2, ACS Central Science 5, 524-538. 
[180] Kizuka, Y., and Taniguchi, N. (2018) Neural functions of bisecting GlcNAc, 
Glycoconjugate Journal 35, 345-351. 
[181] Reusch, D., and Tejada, M. L. (2015) Fc glycans of therapeutic antibodies as 
critical quality attributes, Glycobiology 25, 1325-1334. 
[182] Ruhaak, L. R., Uh, H. W., Beekman, M., Koeleman, C. A. M., Hokke, C. H., 
Westendorp, R. G. J., Wuhrer, M., Houwing-Duistermaat, J. J., Slagboom, P. 
E., and Deelder, A. M. (2010) Decreased levels of bisecting GlcNAc 
glycoforms of IgG are associated with human longevity, Plos One 5(9), 
e12566. 
[183] Sjogren, J., Struwe, W. B., Cosgrave, E. F. J., Rudd, P. M., Stervander, M., 
Allhorn, M., Hollands, A., Nizet, V., and Collin, M. (2013) EndoS(2) is a 
unique and conserved enzyme of serotype M49 group A Streptococcus that 
hydrolyses N-linked glycans on IgG and alpha(1)-acid glycoprotein, 
Biochemical Journal 455, 107-118. 
[184] Giddens, J. P., Lomino, J. V., DiLillo, D. J., Ravetch, J. V., and Wang, L. X. 
(2018) Site-selective chemoenzymatic glycoengineering of Fab and Fc 
glycans of a therapeutic antibody, Proceedings of the National Academy of 
Sciences of the United States of America 115, 12023-12027. 
 218 
 
 
[185] Sheldon, E., Schwickart, M., Li, J., Kim, K., Crouch, S., Parveen, S., Kell, C., 
and Birrell, C. (2016) Pharmacokinetics, pharmacodynamics, and safety of 
MEDI4212, an anti-IgE monoclonal antibody, in subjects with atopy: a Phase 
I study, Advances in Therapy 33, 225-251. 
[186] Pennington, L. F., Tarchevskaya, S., Brigger, D., Sathiyamoorthy, K., Graham, 
M. T., Nadeau, K. C., Eggel, A., and Jardetzky, T. S. (2016) Structural basis 
of omalizumab therapy and omalizumab-mediated IgE exchange, Nature 
Communications 7, 11610. 
 
 219