Waters Application Notes - Glycans

Waters Application Notes 

Glycans

There are a variety of complementary techniques practiced to get the 
complete story about a glycoprotein. Each technique varies in complexity 
and provides a different layer of information.

This application notebook highlights a body of work that has been developed 
to detail applications of innovative chemistry solutions that provide 
complimentary techniques used to answer the what and why questions.  

RELEASED 

GLYCAN 

ANALYSIS

MONOSACCHARIDE 

SIALIC ACID 

COMPOSITION

M

et

ho

d c

om

pl

ex

it

y

Increasing information

MIDDLE UP/DOWN -  

SUBUNIT ANALYSIS

GLYCOPEPTIDE 

MAPPING

INTACT PROTEIN 

ANALYSIS

Stephan M. Koza, Ph.D.
Stephan joined Waters Corporation in 
2011. His research, development, and 
applications group has a primary focus on 
the use of UPLC, HPLC, LC-MS, and sample 
preparation technologies for the analysis 
of biomolecules. He received his B.A. in 
Chemistry from The Massachusetts College of Liberal Arts 
and his Ph.D. in Analytical Chemistry from The University of 
New Hampshire Prior to joining Waters he had nearly twenty 
years of experience with Genetics Institute, Wyeth, and Pfizer 
where he was primarily involved with biopharmaceutical 
characterization and analytical method development.

Matthew Lauber, Ph.D.
Matthew is a Principal Applications 
Chemist within the Consumables Business 
Unit at Waters Corporation. For the last 
4 years, he has applied his expertise in 
protein chemistry and LC-MS based protein 
characterization methods toward the 
development and application of state-of-the-art reagents and 
separation chemistries. Starting with his graduate studies at 
Indiana University, Matthew has shown a penchant to pursue 
new methodologies for puzzling analytical challenges. Most 
recently, Matthew has helped lead investigations into the 
rapid and sensitive analyses of released N-glycans as well as 
the development of new strategies for characterizing protein 
glycosylation that are based on the use of a wide-pore HILIC 
column technology.

Kenneth J. Fountain III, M.S.
Kenneth received his B.S. in Biotechnology 
from the Worcester Polytechnic Institute and 
his M.S. in Chemistry from Tufts University. 
He began his career with Waters Corporation 
in 2001 as a Bioseparations Chemist. 
In this role, he developed LC and LC-MS 
methods for oligonucleotide, peptide, and intact protein 
analysis. In 2004, he joined the Analytical R&D group at 
Genzyme Corporation, where he implemented the use of 
UPLC Technology for the analysis of small molecule drugs 
and biomaterials. Ken returned to Waters Corporation in 
2007, where he spent 7 years in the Consumables Business 
Unit managing the Chemistry Applied Technology Group and 
more recently, the Technology Advancement Group. Both 
of these groups focus on the evaluation, development, and 
application of chromatography columns, reagents, standards, 
and sample preparation devices for the pharmaceutical, 
biopharmaceutical, and food and environmental market 
areas. In January 2015, Ken accepted a position as the 
Director of the Biopharmaceutical Business Development 
Group in the worldwide marketing organization at Waters.

Erin E. Chambers, Ph.D.
As a principal scientist, Erin has been 
working almost exclusively on peptide and 
protein bioanalysis for the last seven years, 
while managing small and large molecule 
bioanalysis and clinical research applications 

for Waters’ Consumables Business Unit. A conversation 
with her college dean helped put her on the path of scientific 
study. Ultimately, Erin graduated from Yale University with 
a degree in chemistry and earned her doctorate from Kings 
College London. Erin loves the dynamic nature of her job, 
helping to develop new methods and analytical systems that 
have a profound influence on illnesses like diabetes and 
Alzheimer’s disease.

Scott McCall, M.S.
Scott is a Research Chemist within the 
Consumables Business Unit at Waters 
Corporation. During his 8 year career 
at Waters he has played a role in the 
development of numerous types of column 
chemistries. Most recently, his focus has 
been toward investigating biomolecule separations including 
peptides, proteins, and glycans. Scott has created methods 
for profiling released high mannose and complex N-glycan 
structures from monoclonal antibodies, and continues to work 
to answer complex questions regarding glycan analysis. 

Rajiv Bharadwaj, Ph.D.
Functioning in the role of Application 
Specialist at Waters India Applications 
Laboratory, Rajiv is instrumental in the 
biopharmaceuticals and proteomics operations. 
His area of focus includes high resolution 
mass spectrometry based solutions, technical 
support, and application development. As a day-dreamer, he 
is influenced by fast technological turnover. Motivated by his 
profound experience with R&D’s of Novozymes and Biocon, he 
aspires to create benchmarks in analytical sciences. He has been 
actively participating in scientific conferences/publications at 
organizations like Proteomics Society of India and IIT Bombay, 
to name a few. Trained at Indian Institute of Science (IISc), 
Bangalore, he also holds a university rank for post graduation in 
biotechnology. Being a gourmet food-enthusiast, Rajiv enjoys 
bringing new flavors to his dishes. 

Jennifer Fournier, M.S. 
Jennifer joined Waters Corporation in 
2004. She has worked in many different 
parts of the organization. She started in Life 
Science Research and Development, then 
moved into the manufacturing group for the 
MassPREP line of standards, and now is a 
Product Manager for the same products she used to develop 
and manufacture. She received her B.S. in Biotechnology and 
M.S. in Biology from Worcester Polytechnic Institute. Prior to 
joining Waters she taught high school biology and also worked 
in manufacturing for a small pharmaceutical company that 
manufactures Albuterol for inhalers. In her free time, Jennifer 
enjoys spending time with her children and coaching. 
 
 
 
 

Our Scientists

[ 1 ]

Michael F. Morris M.S. 
As a Senior Scientist with Waters Synthesis 
Group, Michael has worked on particle 
synthesis and chromatographic support 
development for the last 10 years. 
Mike has brought 30 years experience 
developing novel separations reagents 
for pharmaceutical, biopharma, and diagnostic markets to 
Waters’ Consumables Business Unit. Prior to joining the 
synthesis group in 2006 he worked with VICAM, now a 
division of Waters Corp, developing diagnostic products for 
the food, agriculture, and dairy markets. Having grown up 
in Pennsylvania farm country, he was initially interested in 
studying Forestry but was convinced as an undergrad that 
Chemistry offered a better fit to his curiosity about biological 
systems. He went on to earn an M.S. in Biochemistry from 
Purdue University. Mike has always been an avid outdoors 
man and still spends most of his weekends in the woods and 
on the water.

Darryl Brousmiche, Ph.D.
Darryl is a Principal Scientist and Manager 
of the Synthesis Group in the Consumables 
Business Unit at Waters Corporation. 
Since joining Waters in 2003, he has 
been engaged in developing new sample 
prep and stationary phase materials for 
chromatographic separations, as well as novel tagging 
reagents. Darryl began his career studying Organic 
Chemistry at the University of Ottawa before completing his 
Ph.D. in Organic Photochemistry at the University of Victoria. 
He then spent two years as a postdoc at the University of 
California, Berkeley investigating light harvesting materials. 
In his free time, Darryl enjoys camping, hiking, and skiing.

Pamela Iraneta, B.S. Chemistry, B.S.E.E.
Pamela has worked for Waters Corporation 
for over 25 years. Her interest in all 
forms of chromatography started while 
characterizing silica sols using Sedimentation 
Field Flow Fraction (SF3) – a technique 
based on some of the same principles used 
in chromatographic separations. She has experience in the 
following chromatographic techniques: mix-mode ion exchange, 
reversed-phase, HILIC, inverse-size exclusion, size exclusion 
chromatography (SEC), gel permeation chromatography (GPC), 
and supercritical fluid chromatography for both small molecule 
and large biomolecule separations. Throughout her career at 
Waters, she has made significant contributions to the Oasis 
family of solid phase extraction (SPE) products, as well as the 
Ostro protein precipitation and phospholipid removal plates, 
and PoraPak Rxn cartridges. For column chemistries, she has 
contributed to the success of ACQUITY BEH, XBridge, XSelect, 
UPC2, BEH, to CORTECS, among others. Pam is a co-author on 
26 peer-reviewed publications and inventor on 14 patents. 

The patents are diverse in their application and have included 
antimicrobial chromatographic systems, chromatographic 
materials using charged surface technology, hybrid monoliths, 
96-well SPE extraction devices, dried biological matrix 
carriers and extraction devices, and chromatographic 
materials for the separation of glyco-proteins and -peptides, 
among others.

Sean M. McCarthy, Ph.D.
Sean joined Waters Corporation in 2008, 
and currently serves as a Principal 
Scientist and Senior Manager within the 
Biopharmaceutical Business Organization. 
During his tenure at Waters, he has led 
applications development teams for a 
variety of bioseparations application areas including protein 
and peptide chromatography, oligonucleotide analysis, and 
analytical method development/transfer. Sean received his 
Ph.D. in Chemistry from the University of Vermont in 2005, 
and prior to joining Waters completed postdoctoral training 
within the Department of Pathology at the same institution, 
where he focused on environmental oxidative stress 
related diseases using a variety of biochemical and mass 
spectrometric techniques.

Robert E. Birdsall, Ph.D.
As a Senior Scientist, Robert has an active 
interest in the application of instrumentation 
in the characterization of biopharmaceuticals. 
Robert has spearheaded several key 
initiatives since joining the Pharmaceutical 
Business Operations at Waters. He 
attributes his success, in the field of bioseparations, to his 
academic advisors who fueled his passion and curiosity in the 
sciences, the culmination of which resulted in a Doctorate in 
Analytical Chemistry from Purdue University. One of Robert’s 
favorite aspects of his job is developing novel applications 
using cutting-edge instrumentation that can assist the 
biopharmaceutical industry in bringing safe and efficacious 
therapeutic products to market.

Ying Qing Yu, Ph.D.
Ying Qing is a Senior Science Manager 
in the Pharmaceutical Group at Waters 
Corporation. She joined Waters Corporation 
in 2001, shortly after she received her 
Ph.D. in Analytical Chemistry from Purdue 
University. Her group’s focus is on protein 
biotherapeutics characterization using the UPLC/QTof MS 
platform. The projects she is currently working on range from 
glycan profiling, peptide mapping, and lately, the hydrogen/
deuterium exchange mass spectrometry for higher order 
structural protein analysis. She has extensive experience 
in mass spectrometry, gas-phase ion chemistry, and LC 
separation techniques.

Our Scientists (Continued)

[ 2 ]

Table of Contents

[ INTRODUCTION ]

A Review of Glycan Analysis Requirements (Published in BioPharm International) ................5

[ INTACT PROTEINS ]

Developing High Resolution HILIC Separations of Intact Glycosylated Proteins  
Using a Wide-Pore Amide-Bonded Stationary Phase .............................................................. 11

Measuring the Glycan Occupancy of Intact mAbs Using HILIC and  
Detection by Intrinsic Fluorescence ........................................................................................... 23

Intact Protein Characterization: Waters Biopharmaceutical System Solution with UNIFI ....... 26

Streamlining Compliant and Non-Compliant Intact Mass Analysis of  
Biotherapeutic mAbs with the Biopharmaceutical Platform Solution with UNIFI ................... 28

[ GLYCOPEPTIDE MAPPING AND SUBUNIT ANALYSIS ]

HILIC Glycopeptide Mapping with a Wide-Pore Amide Stationary Phase ................................ 37

Mapping IgG Subunit Glycoforms Using HILIC and a Wide-Pore Amide Stationary Phase ..... 46

Structural Comparison of Infliximab and a Biosimilar via Subunit Analysis  
Using the Waters Biopharmaceutical Platform with UNIFI ....................................................... 56

[ RELEASED GLYCANS ]

Rapid Preparation of Released N-Glycans for HILIC Analysis Using a  
Novel Fluorescence and MS-Active Labeling Reagent .............................................................. 65

Robustness of RapiFluor-MS N-Glycan Sample Preparations and  
Glycan BEH Amide HILIC Chromatographic Separations .......................................................... 76

GlycoWorks HILIC SPE Robust Glycan Sample Preparation ...................................................... 88

Quality Control and Automation Friendly GlycoWorks RapiFluor-MS  
N-Glycan Sample Preparation .................................................................................................... 91

Profiling Released High Mannose and Complex N-Glycan Structures from  
Monoclonal Antibodies Using RapiFluor-MS Labeling and  
Optimized Hydrophilic Interaction Chromatography ................................................................ 97

[ 3 ]

Table of Contents (Continued)

Enhancing the Peak Capacity of High Molecular Weight N-Glycan  
HILIC Separations with a Wide-Pore Amide Bonded Stationary Phase ................................. 106

Exploiting RapiFluor-MS Labeling to Monitor Diverse N-Glycan Structures via  
Fluorescence and Mass Detection ........................................................................................... 109

New Capabilities for Monitoring Released N-Glycans through the Combined Use of  
RapiFluor-MS Labeling, ACQUITY UPLC H-Class Bio System, and  
Serial Fluorescence/ACQUITY QDa Mass Detection ............................................................... 113

Rapidly Monitoring Released N-Glycan Profiles During Process Development Using  
RapiFluor-MS and the ACQUITY QDa Detector...................................................................... 118

Applying a Novel Glycan Tagging Reagent, RapiFluor-MS, and an  
Integrated UPLC-FLR/QTof MS System for Low Abundant N-Glycan Analysis .................... 123

Released N-linked Glycan Analysis Using the Glycan Application Solution with UNIFI ..... 128

A Holistic Workflow for Acquisition, Processing, and Reporting Fluorescent-Labeled  
Glycans Using the Biopharmaceutical Platform Solution with UNIFI ................................... 131

N-Linked Glycan Characterization and Profiling: Combining the Power of Accurate Mass,  
Reference Glucose Units, and UNIFI Software for Confident Glycan Assignments .............. 139

Transferring RapiFluor-MS Labeled N-Glycans between UPLC and HPLC ............................ 149

[ RELEASED N- AND O-LINKED GLYCANS ]

Characterization of EPO N-Glycans Using RapiFluor-MS and HILIC Profiling ...................... 156

Comprehensive Characterization of the N and O-Linked Glycosylation of a  
Recombinant Human EPO ....................................................................................................... 159

[ MONOSACCHARIDES ]

Future Proofing the Biopharmaceutical QC Laboratory: Chromatographic Scaling  
of HPLC Monosaccharide Analyses Using the ACQUITY UPLC H-Class Bio System ........... 169

[ 4 ]

32   

BioPharm International  www.biopharminternational.com  October 2015

C

a

l

l

is

t

a 

im

a

g

e

s

/g

e

t

t

y i

m

a

g

e

s

M

ore  than  t wo -thirds  of 
recombinant  biopharma-
ceutical  products  on  the 
market  are  glycoproteins, 

and  every  stage  of  their  manufacture  is 
carefully  monitored  and  tested  to  ensure 
consistency  in  quality,  safety,  and  effec-
tiveness  (1).  Of  the  various  aspects  of 
biopharmaceutical  production  (such  as 
yield,  protein  folding,  and  post-trans-
lational  modifications),  the  host  cell’s 
biosynthesis  of  attached  oligosaccharides 
(glycans)  is  often  the  most  difficult  to 
control.  Selected  expression  systems  and 
even slight changes in process conditions 
can  alter  the  synthesis  of  glycans  and 
as  a  consequence,  the  physicochemical 
properties  (e.g.,  serum  half-life),  safety, 
efficacy,  and  immunogenicity  of  the 
end  product.  Regulatory  agencies  world-

wide  require  state-of-the-art  glycan  anal-
yses  and  the  demands  placed  on  these 
methods  have  steadily  increased  as  bet-
ter  technologies  have  been  developed. 
Ultimately,  robust,  information-rich,  and 
reproducible methods for glycan analysis 
must  be  included  in  regulatory  filings 
for  glycoprotein-based  biotherapeutics  to 
ensure accuracy and consistency. Method 
simplification  and  standardization  will 
provide additional assurance that the gly-
can-analysis methods used are transferra-
ble between testing sites both within and 
outside  (e.g.,  contract  research  organiza-
tions) of the organization, ensuring better 
quality and efficiency in manufacturing.

Glycans face new scrutiny
By  2008,  the  biotechnology  company 
Genzyme  had  developed  and  marketed 

a review of Glycan 

analysis requirements

Jennifer Fournier

The author 

explores 

the basic 

rationale and 

requirements for 

standardized 

glycan analysis.

Jennifer Fournier is product marketing 

manager, consumables business 

unit-asr, at waters corporation.

Glycosylation

ES684993_BP1015_032.pgs  10.05.2015  23:09    ADV  

black

yellow

magenta

cyan

October 2015 

www.biopharminternational.com  BioPharm International    33

the  drug  Myozyme  (alglucosidase 
alfa)  for  the  treatment  of  Pompe 
disease,  a  rare  and  progressively 
debilitating  disorder  characterized 
by deficiency of lysosomal enzyme 
alpha-glucosidase  (GAA).  The  com-
pany  was  preparing  to  expand  the 
targeted treatment population from 
primarily  children  to  adults.  Its 
160-L production facility was work-
ing  at  capacity,  so  $53  million  was 
invested  to  build  a  2000-L  facility 
for  Myozyme  in  Allston,  MA  (2). 
The company was ready to launch, 
but  FDA  rejected  Genzyme’s  appli-
cation  to  sell  the  drug  from  the 
2000-L  plant.  According  to  regu-
lators,  the  version  made  in  the 
2000 -L  tank  was  no  longer  the 
same  drug  as  the  one  produced 
in  the  160 -L  tank.  FDA  argued 
that  the  differences  in  glycosyl-
ation—specifically in this case, the 
composition  of  mannose-6-phos-
phate—meant  that  the  drug  was 
no  longer  the  biological  equivalent 
of  the  original  material  produced 
in  the  160-L  bioreactor,  and  may 
in fact introduce unknown clinical 
variables.  Genzyme  argued  that  it 
had  already  conducted  a  clinical 
trial  on  the  larger  batch  material, 
demonstrating safety and effective-
ness.  Ultimately,  Genzyme  had  to 

market the product from the larger 
bioreactor under a different name.

The  incident  was  a  watershed 

moment  in  the  biopharmaceutical 
industry,  marking  the  emergence 
of  new  challenges  (1).  First,  regula-
tory  authorities  were  beginning  to 
scrutinize  the  glycan  structures  of 
biopharmaceutical  products  more 
carefully  based  on  established  tech-
nical  guidelines  (e.g.,  ICH  Q5E,  ICH 
Q6B, and FDA’s Guidance for Industry, 
PAT—A  Framework  for  Innovative 
P h a r m a c e u t i c a l   D e v e l o p m e n t , 
M a n u f a c t u r i n g ,   a n d   Q u a l i t y 
Assurance),  yet  there  remained 
inconsistencies  in  how  FDA,  the 
European  Medicines  Agency  (EMA), 
and  Japanese  regulators  determined 
what  is  “biosimilar”.  Second,  prod-
ucts with complex glycosylation pat-
terns have the potential to easily fall 
out  of  specification  with  changes 
in  biomanufacturing  processes  and 
scale-up, so to meet the new regula-
tory demands, manufacturers had to 
start  carefully  characterizing  prod-
uct  glycosylation  and  its  relation  to 
the biological and clinical activity of 
a medication, and begin monitoring 
relevant  glycan  characteristics  dur-
ing production (3, 4, 5).

In the years following FDA’s deci-

sion  on  Myozyme,  the  attention 

given  to  glycan  structure  in  bio-
pharmaceuticals has only increased, 
reflecting  improvement  in  analyti-
cal technology and a greater under-
standing of the role these structures 
play  in  the  physical  characteristics, 
stability,  biological  activity,  and 
the  clinical  safety  and  effective-
ness of a drug (6, 7). The  technical 
guidelines  for  characterizing  and 
monitoring  glycans  have  changed 
little  since  20 08;  manufact ur-
ers  refer  mainly  to  International 
Conference  on  Har monization 
(ICH)  documents  Q5E  and  Q6B 
(3,  4).  These  documents  list  the 
following  recommendations  on 
characterizing glycans:

“For  glycoproteins,  the  carbo-

hydrate  content  (neutral  sugars, 
amino  sugars,  and  sialic  acids)  is 
determined. In addition, the struc-
ture  of  the  carbohydrate  chains, 
the oligosaccharide pattern (anten-
nary profile), and the glycosylation 
site(s)  of  the  polypeptide  chain  is 
analyzed, to the extent possible.”

O t her  g u idel i nes  e x ist,  set-

t i n g   e x p e c t at ion s   for   g lyc a n 
analysis,  such  as  FDA’s  Guidance 
f o r   I n d u s t r y,   I m m u n o g e n i c i t y 
Assessment  for  Therapeutic  Protein 
Products,  and  EMA’s  2007  mono-
graph  on  the  characterization  of 

Glycosylation

Table I: 

Common sample preparation methods in glycan analysis.

Sample preparation

Description

Application

PNGase F

Peptide-N-glycosidase F

Release of N-glycan chain except those 

with (∝1,3)-linked core fucose

Release of complex, hybrid, and 

oligomannose N-glycans

PNGase A

Peptide-N-glycosidase A

Release of N-glycan chain containing 

(∝1,3)-linked core fucose

Release of complex, hybrid, and 

oligomannose N-glycans

Proteolysis

The use of a protease to generate peptides 

(including glycopeptides) from a glycoprotein

The peptides are often analyzed to investigate 

glycosylation sites and occupancy

Alkaline beta 
elimination/
hydrazinolysis

Chemical cleavage of O-linked glycans from 

polypeptide chains

Primarily used in the analysis of O-linked 

glycans 

Permethylation

The methylation of oligosaccharide hydroxyl 

groups to make glycans more amenable to 

mass spectrometric (MS) analysis

MS-based characterization of glycans 

including linkage analysis

Amine/
glycosylamine 
labeling

Modification of glycans to facilitate 

fluorescence detection

Detection of glycans and glycopeptides when 

native detection is not available. Increases 

options for chromatographic separation 

methods. May also enhance MS analysis

ES684985_BP1015_033.pgs  10.05.2015  23:08    ADV  

black

yellow

magenta

cyan

[ INTRODUCTION ]
Written by Jennifer Fournier, Product Marketing Manager at Waters, and published in BioPharm International.

[ 5 ]

34   

BioPharm International  www.biopharminternational.com  October 2015

monoclonal  antibodies  (8).  The 
monog raph  says  the  follow ing 
on glycans:

“Glycan  structures  should  be 

characterized, and particular atten-
tion should be paid to their degree 
of  mannosylation,  galactosylation, 
fucosylation,  and  sialylation.  The 
distribution  of  the  main  glycan 
structures  present  (often  G0,  G1, 
and G2) should be determined.”

T hese  doc u ments,  however, 

present  few  details  on  how  to  set 
specification  limits  on  glycans, 
or  recommend  technologies  and 
procedures  for  consistent  analyti-
cal  results.  The  consequences  for 
this  long-standing  ambiguity  are 
that  manufacturers  and  regulators 
sometimes  end  up  with  different 
ideas  as  to  what  constitutes  a  nec-
essary  specification  for  a  glycan 
structure.  Furthermore,  compa-
nies  submit  reports  to  regulatory 
authorities  with  widely  different 
analytical  approaches.  Procedures 
may  vary  even  within  the  same 
organization,  potentially  leading 
to  inconsistent  results,  analytical 
testing  failures,  and  ultimately, 
regulatory delays.

Quality by DesiGn  

vs. Quality in practice
In 2002, in response to an increas-
ing  burden  on  FDA  of  regulating 
product  manufacturing,  and  a 
perception  among  companies  that 
regulatory  requirements  were  lim-
iting  flexibility  in  process  optimi-
zation,  FDA  implemented  changes 
through  its  Pharmaceutical  cGMP 
21st  Century  Initiative  and  the 
release  of  FDA’s  process  analyti-
cal  technology  guidance  (PAT)  (5). 
The  new  approach  placed  greater 
responsibility  on  the  manufac-
turers  to  monitor  quality  control 
through  timely  measurements  and 
corrections during processing.

A round  the  same  time,  ICH 

published  t wo  g uidance  doc u-
ments:  ICH  Q8  Phar maceutical 
Development  (7),  ICH  Q9  Quality 
Risk  Management  (8),  and  ICH 
Q10  Quality  Systems  Approach  to 
Pharmaceutical cGMP Regulations (9, 
10,  11).  These  documents  helped 
to further define current scientific 
and risk-based approaches to phar-
maceutical quality control.

The  concept  of  quality  by  design 

(QbD)  was  incorporated  into  FDA 

review  in  2004,  which  together 
with the aforementioned guidelines, 
emphasized a greater understanding 
of  the  product  and  its  manufactur-
ing  process,  and  designing  quality 
control into the process, rather than 
testing  it  after  the  fact  (12).  This 
approach  is  particularly  well-suited 
to  glycan  analysis,  which  is  typi-
cally  associated  with  a  complex  set 
of  critical  quality  attributes  (CQAs) 
(such  as  sialylation,  antennary 
structure,  or  glycan  structure  het-
erogeneity)  that  are  important  to 
the  biological  or  clinical  activity  of 
the drug. The CQAs must be identi-
fied, measured during process devel-
opment,  and  maintained  within 
required parameters (i.e., the design 
space) during production.

In  the  case  of  glycans,  the  mea-

surement  itself  may  introduce 
uncertainty and risk, due to a high 
variability of outcomes when char-
acterizing  oligosaccharide  chains. 
An  interlaboratory  study  present-
ing  11  industrial,  regulatory,  and 
academic  labs  with  the  same  set 
of  four  released  N-glycans  demon-
strated that results were not consis-
tent between the laboratories when 

Glycosylation

Table II: 

Common separation methods in glycan analysis.

Separation method

Description

Application

HPAEC

High-pH anion-exchange 

chromatography

Liquid chromatographic separation of 
negatively charged (acidic) molecules 

carried out at high pH 

Separation, identification, and quantification 

of glycans and glycopeptides. Analysis of 

monosaccharide and/or sialic acid composition. 

Often coupled with pulsed amperometric detection 

(PAD) for detection of underivatized molecules

HPCE/CZE

High-performance capillary 

electrophoresis/capillary-

zone electrophoresis

Separation of molecules by charge using an 

electric field in a narrow capillary channel

Separation, identification, and quantification of 

charged glycans. Analysis and quantification of 

sialylation

HILIC

Hydrophilic-interaction, 

high-performance liquid 

chromatography

A variation of high-performance liquid 

chromatography (HPLC) that separates 

molecules using a hydrophilic stationary 
phase and an organic-yet-water-miscible 

liquid phase

Separation, identification, and quantification of 

glycans and glycopeptides 

WAX–HPLC

Weak-anion exchange–

high-performance liquid 

chromatography

Separates anionic molecules based on 

their degree of charge

Separation, identification, and quantification of 

glycans and glycopeptides

RP–HPLC

Reverse-phase–high-

performance liquid 

chromatography

Separates molecules on the basis of 

differences in the strength of their interaction 

with a hydrophobic stationary phase

Separation, identification, and quantification of 

glycans and glycopeptides

ES684990_BP1015_034.pgs  10.05.2015  23:09    ADV  

black

yellow

magenta

cyan

October 2015 

www.biopharminternational.com  BioPharm International    33

the  drug  Myozyme  (alglucosidase 
alfa)  for  the  treatment  of  Pompe 
disease,  a  rare  and  progressively 
debilitating  disorder  characterized 
by deficiency of lysosomal enzyme 
alpha-glucosidase  (GAA).  The  com-
pany  was  preparing  to  expand  the 
targeted treatment population from 
primarily  children  to  adults.  Its 
160-L production facility was work-
ing  at  capacity,  so  $53  million  was 
invested  to  build  a  2000-L  facility 
for  Myozyme  in  Allston,  MA  (2). 
The company was ready to launch, 
but  FDA  rejected  Genzyme’s  appli-
cation  to  sell  the  drug  from  the 
2000-L  plant.  According  to  regu-
lators,  the  version  made  in  the 
2000 -L  tank  was  no  longer  the 
same  drug  as  the  one  produced 
in  the  160 -L  tank.  FDA  argued 
that  the  differences  in  glycosyl-
ation—specifically in this case, the 
composition  of  mannose-6-phos-
phate—meant  that  the  drug  was 
no  longer  the  biological  equivalent 
of  the  original  material  produced 
in  the  160-L  bioreactor,  and  may 
in fact introduce unknown clinical 
variables.  Genzyme  argued  that  it 
had  already  conducted  a  clinical 
trial  on  the  larger  batch  material, 
demonstrating safety and effective-
ness.  Ultimately,  Genzyme  had  to 

market the product from the larger 
bioreactor under a different name.

The  incident  was  a  watershed 

moment  in  the  biopharmaceutical 
industry,  marking  the  emergence 
of  new  challenges  (1).  First,  regula-
tory  authorities  were  beginning  to 
scrutinize  the  glycan  structures  of 
biopharmaceutical  products  more 
carefully  based  on  established  tech-
nical  guidelines  (e.g.,  ICH  Q5E,  ICH 
Q6B, and FDA’s Guidance for Industry, 
PAT—A  Framework  for  Innovative 
P h a r m a c e u t i c a l   D e v e l o p m e n t , 
M a n u f a c t u r i n g ,   a n d   Q u a l i t y 
Assurance),  yet  there  remained 
inconsistencies  in  how  FDA,  the 
European  Medicines  Agency  (EMA), 
and  Japanese  regulators  determined 
what  is  “biosimilar”.  Second,  prod-
ucts with complex glycosylation pat-
terns have the potential to easily fall 
out  of  specification  with  changes 
in  biomanufacturing  processes  and 
scale-up, so to meet the new regula-
tory demands, manufacturers had to 
start  carefully  characterizing  prod-
uct  glycosylation  and  its  relation  to 
the biological and clinical activity of 
a medication, and begin monitoring 
relevant  glycan  characteristics  dur-
ing production (3, 4, 5).

In the years following FDA’s deci-

sion  on  Myozyme,  the  attention 

given  to  glycan  structure  in  bio-
pharmaceuticals has only increased, 
reflecting  improvement  in  analyti-
cal technology and a greater under-
standing of the role these structures 
play  in  the  physical  characteristics, 
stability,  biological  activity,  and 
the  clinical  safety  and  effective-
ness of a drug (6, 7). The  technical 
guidelines  for  characterizing  and 
monitoring  glycans  have  changed 
little  since  20 08;  manufact ur-
ers  refer  mainly  to  International 
Conference  on  Har monization 
(ICH)  documents  Q5E  and  Q6B 
(3,  4).  These  documents  list  the 
following  recommendations  on 
characterizing glycans:

“For  glycoproteins,  the  carbo-

hydrate  content  (neutral  sugars, 
amino  sugars,  and  sialic  acids)  is 
determined. In addition, the struc-
ture  of  the  carbohydrate  chains, 
the oligosaccharide pattern (anten-
nary profile), and the glycosylation 
site(s)  of  the  polypeptide  chain  is 
analyzed, to the extent possible.”

O t her  g u idel i nes  e x ist,  set-

t i n g   e x p e c t at ion s   for   g lyc a n 
analysis,  such  as  FDA’s  Guidance 
f o r   I n d u s t r y,   I m m u n o g e n i c i t y 
Assessment  for  Therapeutic  Protein 
Products,  and  EMA’s  2007  mono-
graph  on  the  characterization  of 

Glycosylation

Table I: 

Common sample preparation methods in glycan analysis.

Sample preparation

Description

Application

PNGase F

Peptide-N-glycosidase F

Release of N-glycan chain except those 

with (∝1,3)-linked core fucose

Release of complex, hybrid, and 

oligomannose N-glycans

PNGase A

Peptide-N-glycosidase A

Release of N-glycan chain containing 

(∝1,3)-linked core fucose

Release of complex, hybrid, and 

oligomannose N-glycans

Proteolysis

The use of a protease to generate peptides 

(including glycopeptides) from a glycoprotein

The peptides are often analyzed to investigate 

glycosylation sites and occupancy

Alkaline beta 
elimination/
hydrazinolysis

Chemical cleavage of O-linked glycans from 

polypeptide chains

Primarily used in the analysis of O-linked 

glycans 

Permethylation

The methylation of oligosaccharide hydroxyl 

groups to make glycans more amenable to 

mass spectrometric (MS) analysis

MS-based characterization of glycans 

including linkage analysis

Amine/
glycosylamine 
labeling

Modification of glycans to facilitate 

fluorescence detection

Detection of glycans and glycopeptides when 

native detection is not available. Increases 

options for chromatographic separation 

methods. May also enhance MS analysis

ES684985_BP1015_033.pgs  10.05.2015  23:08    ADV  

black

yellow

magenta

cyan

34   

BioPharm International  www.biopharminternational.com  October 2015

monoclonal  antibodies  (8).  The 
monog raph  says  the  follow ing 
on glycans:

“Glycan  structures  should  be 

characterized, and particular atten-
tion should be paid to their degree 
of  mannosylation,  galactosylation, 
fucosylation,  and  sialylation.  The 
distribution  of  the  main  glycan 
structures  present  (often  G0,  G1, 
and G2) should be determined.”

T hese  doc u ments,  however, 

present  few  details  on  how  to  set 
specification  limits  on  glycans, 
or  recommend  technologies  and 
procedures  for  consistent  analyti-
cal  results.  The  consequences  for 
this  long-standing  ambiguity  are 
that  manufacturers  and  regulators 
sometimes  end  up  with  different 
ideas  as  to  what  constitutes  a  nec-
essary  specification  for  a  glycan 
structure.  Furthermore,  compa-
nies  submit  reports  to  regulatory 
authorities  with  widely  different 
analytical  approaches.  Procedures 
may  vary  even  within  the  same 
organization,  potentially  leading 
to  inconsistent  results,  analytical 
testing  failures,  and  ultimately, 
regulatory delays.

Quality by DesiGn  

vs. Quality in practice
In 2002, in response to an increas-
ing  burden  on  FDA  of  regulating 
product  manufacturing,  and  a 
perception  among  companies  that 
regulatory  requirements  were  lim-
iting  flexibility  in  process  optimi-
zation,  FDA  implemented  changes 
through  its  Pharmaceutical  cGMP 
21st  Century  Initiative  and  the 
release  of  FDA’s  process  analyti-
cal  technology  guidance  (PAT)  (5). 
The  new  approach  placed  greater 
responsibility  on  the  manufac-
turers  to  monitor  quality  control 
through  timely  measurements  and 
corrections during processing.

A round  the  same  time,  ICH 

published  t wo  g uidance  doc u-
ments:  ICH  Q8  Phar maceutical 
Development  (7),  ICH  Q9  Quality 
Risk  Management  (8),  and  ICH 
Q10  Quality  Systems  Approach  to 
Pharmaceutical cGMP Regulations (9, 
10,  11).  These  documents  helped 
to further define current scientific 
and risk-based approaches to phar-
maceutical quality control.

The  concept  of  quality  by  design 

(QbD)  was  incorporated  into  FDA 

review  in  2004,  which  together 
with the aforementioned guidelines, 
emphasized a greater understanding 
of  the  product  and  its  manufactur-
ing  process,  and  designing  quality 
control into the process, rather than 
testing  it  after  the  fact  (12).  This 
approach  is  particularly  well-suited 
to  glycan  analysis,  which  is  typi-
cally  associated  with  a  complex  set 
of  critical  quality  attributes  (CQAs) 
(such  as  sialylation,  antennary 
structure,  or  glycan  structure  het-
erogeneity)  that  are  important  to 
the  biological  or  clinical  activity  of 
the drug. The CQAs must be identi-
fied, measured during process devel-
opment,  and  maintained  within 
required parameters (i.e., the design 
space) during production.

In  the  case  of  glycans,  the  mea-

surement  itself  may  introduce 
uncertainty and risk, due to a high 
variability of outcomes when char-
acterizing  oligosaccharide  chains. 
An  interlaboratory  study  present-
ing  11  industrial,  regulatory,  and 
academic  labs  with  the  same  set 
of  four  released  N-glycans  demon-
strated that results were not consis-
tent between the laboratories when 

Glycosylation

Table II: 

Common separation methods in glycan analysis.

Separation method

Description

Application

HPAEC

High-pH anion-exchange 

chromatography

Liquid chromatographic separation of 
negatively charged (acidic) molecules 

carried out at high pH 

Separation, identification, and quantification 

of glycans and glycopeptides. Analysis of 

monosaccharide and/or sialic acid composition. 

Often coupled with pulsed amperometric detection 

(PAD) for detection of underivatized molecules

HPCE/CZE

High-performance capillary 

electrophoresis/capillary-

zone electrophoresis

Separation of molecules by charge using an 

electric field in a narrow capillary channel

Separation, identification, and quantification of 

charged glycans. Analysis and quantification of 

sialylation

HILIC

Hydrophilic-interaction, 

high-performance liquid 

chromatography

A variation of high-performance liquid 

chromatography (HPLC) that separates 

molecules using a hydrophilic stationary 
phase and an organic-yet-water-miscible 

liquid phase

Separation, identification, and quantification of 

glycans and glycopeptides 

WAX–HPLC

Weak-anion exchange–

high-performance liquid 

chromatography

Separates anionic molecules based on 

their degree of charge

Separation, identification, and quantification of 

glycans and glycopeptides

RP–HPLC

Reverse-phase–high-

performance liquid 

chromatography

Separates molecules on the basis of 

differences in the strength of their interaction 

with a hydrophobic stationary phase

Separation, identification, and quantification of 

glycans and glycopeptides

ES684990_BP1015_034.pgs  10.05.2015  23:09    ADV  

black

yellow

magenta

cyan

[ INTRODUCTION ]

[ 6 ]

October 2015 

www.biopharminternational.com  BioPharm International    33

the  drug  Myozyme  (alglucosidase 
alfa)  for  the  treatment  of  Pompe 
disease,  a  rare  and  progressively 
debilitating  disorder  characterized 
by deficiency of lysosomal enzyme 
alpha-glucosidase  (GAA).  The  com-
pany  was  preparing  to  expand  the 
targeted treatment population from 
primarily  children  to  adults.  Its 
160-L production facility was work-
ing  at  capacity,  so  $53  million  was 
invested  to  build  a  2000-L  facility 
for  Myozyme  in  Allston,  MA  (2). 
The company was ready to launch, 
but  FDA  rejected  Genzyme’s  appli-
cation  to  sell  the  drug  from  the 
2000-L  plant.  According  to  regu-
lators,  the  version  made  in  the 
2000 -L  tank  was  no  longer  the 
same  drug  as  the  one  produced 
in  the  160 -L  tank.  FDA  argued 
that  the  differences  in  glycosyl-
ation—specifically in this case, the 
composition  of  mannose-6-phos-
phate—meant  that  the  drug  was 
no  longer  the  biological  equivalent 
of  the  original  material  produced 
in  the  160-L  bioreactor,  and  may 
in fact introduce unknown clinical 
variables.  Genzyme  argued  that  it 
had  already  conducted  a  clinical 
trial  on  the  larger  batch  material, 
demonstrating safety and effective-
ness.  Ultimately,  Genzyme  had  to 

market the product from the larger 
bioreactor under a different name.

The  incident  was  a  watershed 

moment  in  the  biopharmaceutical 
industry,  marking  the  emergence 
of  new  challenges  (1).  First,  regula-
tory  authorities  were  beginning  to 
scrutinize  the  glycan  structures  of 
biopharmaceutical  products  more 
carefully  based  on  established  tech-
nical  guidelines  (e.g.,  ICH  Q5E,  ICH 
Q6B, and FDA’s Guidance for Industry, 
PAT—A  Framework  for  Innovative 
P h a r m a c e u t i c a l   D e v e l o p m e n t , 
M a n u f a c t u r i n g ,   a n d   Q u a l i t y 
Assurance),  yet  there  remained 
inconsistencies  in  how  FDA,  the 
European  Medicines  Agency  (EMA), 
and  Japanese  regulators  determined 
what  is  “biosimilar”.  Second,  prod-
ucts with complex glycosylation pat-
terns have the potential to easily fall 
out  of  specification  with  changes 
in  biomanufacturing  processes  and 
scale-up, so to meet the new regula-
tory demands, manufacturers had to 
start  carefully  characterizing  prod-
uct  glycosylation  and  its  relation  to 
the biological and clinical activity of 
a medication, and begin monitoring 
relevant  glycan  characteristics  dur-
ing production (3, 4, 5).

In the years following FDA’s deci-

sion  on  Myozyme,  the  attention 

given  to  glycan  structure  in  bio-
pharmaceuticals has only increased, 
reflecting  improvement  in  analyti-
cal technology and a greater under-
standing of the role these structures 
play  in  the  physical  characteristics, 
stability,  biological  activity,  and 
the  clinical  safety  and  effective-
ness of a drug (6, 7). The  technical 
guidelines  for  characterizing  and 
monitoring  glycans  have  changed 
little  since  20 08;  manufact ur-
ers  refer  mainly  to  International 
Conference  on  Har monization 
(ICH)  documents  Q5E  and  Q6B 
(3,  4).  These  documents  list  the 
following  recommendations  on 
characterizing glycans:

“For  glycoproteins,  the  carbo-

hydrate  content  (neutral  sugars, 
amino  sugars,  and  sialic  acids)  is 
determined. In addition, the struc-
ture  of  the  carbohydrate  chains, 
the oligosaccharide pattern (anten-
nary profile), and the glycosylation 
site(s)  of  the  polypeptide  chain  is 
analyzed, to the extent possible.”

O t her  g u idel i nes  e x ist,  set-

t i n g   e x p e c t at ion s   for   g lyc a n 
analysis,  such  as  FDA’s  Guidance 
f o r   I n d u s t r y,   I m m u n o g e n i c i t y 
Assessment  for  Therapeutic  Protein 
Products,  and  EMA’s  2007  mono-
graph  on  the  characterization  of 

Glycosylation

Table I: 

Common sample preparation methods in glycan analysis.

Sample preparation

Description

Application

PNGase F

Peptide-N-glycosidase F

Release of N-glycan chain except those 

with (∝1,3)-linked core fucose

Release of complex, hybrid, and 

oligomannose N-glycans

PNGase A

Peptide-N-glycosidase A

Release of N-glycan chain containing 

(∝1,3)-linked core fucose

Release of complex, hybrid, and 

oligomannose N-glycans

Proteolysis

The use of a protease to generate peptides 

(including glycopeptides) from a glycoprotein

The peptides are often analyzed to investigate 

glycosylation sites and occupancy

Alkaline beta 
elimination/
hydrazinolysis

Chemical cleavage of O-linked glycans from 

polypeptide chains

Primarily used in the analysis of O-linked 

glycans 

Permethylation

The methylation of oligosaccharide hydroxyl 

groups to make glycans more amenable to 

mass spectrometric (MS) analysis

MS-based characterization of glycans 

including linkage analysis

Amine/
glycosylamine 
labeling

Modification of glycans to facilitate 

fluorescence detection

Detection of glycans and glycopeptides when 

native detection is not available. Increases 

options for chromatographic separation 

methods. May also enhance MS analysis

ES684985_BP1015_033.pgs  10.05.2015  23:08    ADV  

black

yellow

magenta

cyan

34   

BioPharm International  www.biopharminternational.com  October 2015

monoclonal  antibodies  (8).  The 
monog raph  says  the  follow ing 
on glycans:

“Glycan  structures  should  be 

characterized, and particular atten-
tion should be paid to their degree 
of  mannosylation,  galactosylation, 
fucosylation,  and  sialylation.  The 
distribution  of  the  main  glycan 
structures  present  (often  G0,  G1, 
and G2) should be determined.”

T hese  doc u ments,  however, 

present  few  details  on  how  to  set 
specification  limits  on  glycans, 
or  recommend  technologies  and 
procedures  for  consistent  analyti-
cal  results.  The  consequences  for 
this  long-standing  ambiguity  are 
that  manufacturers  and  regulators 
sometimes  end  up  with  different 
ideas  as  to  what  constitutes  a  nec-
essary  specification  for  a  glycan 
structure.  Furthermore,  compa-
nies  submit  reports  to  regulatory 
authorities  with  widely  different 
analytical  approaches.  Procedures 
may  vary  even  within  the  same 
organization,  potentially  leading 
to  inconsistent  results,  analytical 
testing  failures,  and  ultimately, 
regulatory delays.

Quality by DesiGn  

vs. Quality in practice
In 2002, in response to an increas-
ing  burden  on  FDA  of  regulating 
product  manufacturing,  and  a 
perception  among  companies  that 
regulatory  requirements  were  lim-
iting  flexibility  in  process  optimi-
zation,  FDA  implemented  changes 
through  its  Pharmaceutical  cGMP 
21st  Century  Initiative  and  the 
release  of  FDA’s  process  analyti-
cal  technology  guidance  (PAT)  (5). 
The  new  approach  placed  greater 
responsibility  on  the  manufac-
turers  to  monitor  quality  control 
through  timely  measurements  and 
corrections during processing.

A round  the  same  time,  ICH 

published  t wo  g uidance  doc u-
ments:  ICH  Q8  Phar maceutical 
Development  (7),  ICH  Q9  Quality 
Risk  Management  (8),  and  ICH 
Q10  Quality  Systems  Approach  to 
Pharmaceutical cGMP Regulations (9, 
10,  11).  These  documents  helped 
to further define current scientific 
and risk-based approaches to phar-
maceutical quality control.

The  concept  of  quality  by  design 

(QbD)  was  incorporated  into  FDA 

review  in  2004,  which  together 
with the aforementioned guidelines, 
emphasized a greater understanding 
of  the  product  and  its  manufactur-
ing  process,  and  designing  quality 
control into the process, rather than 
testing  it  after  the  fact  (12).  This 
approach  is  particularly  well-suited 
to  glycan  analysis,  which  is  typi-
cally  associated  with  a  complex  set 
of  critical  quality  attributes  (CQAs) 
(such  as  sialylation,  antennary 
structure,  or  glycan  structure  het-
erogeneity)  that  are  important  to 
the  biological  or  clinical  activity  of 
the drug. The CQAs must be identi-
fied, measured during process devel-
opment,  and  maintained  within 
required parameters (i.e., the design 
space) during production.

In  the  case  of  glycans,  the  mea-

surement  itself  may  introduce 
uncertainty and risk, due to a high 
variability of outcomes when char-
acterizing  oligosaccharide  chains. 
An  interlaboratory  study  present-
ing  11  industrial,  regulatory,  and 
academic  labs  with  the  same  set 
of  four  released  N-glycans  demon-
strated that results were not consis-
tent between the laboratories when 

Glycosylation

Table II: 

Common separation methods in glycan analysis.

Separation method

Description

Application

HPAEC

High-pH anion-exchange 

chromatography

Liquid chromatographic separation of 
negatively charged (acidic) molecules 

carried out at high pH 

Separation, identification, and quantification 

of glycans and glycopeptides. Analysis of 

monosaccharide and/or sialic acid composition. 

Often coupled with pulsed amperometric detection 

(PAD) for detection of underivatized molecules

HPCE/CZE

High-performance capillary 

electrophoresis/capillary-

zone electrophoresis

Separation of molecules by charge using an 

electric field in a narrow capillary channel

Separation, identification, and quantification of 

charged glycans. Analysis and quantification of 

sialylation

HILIC

Hydrophilic-interaction, 

high-performance liquid 

chromatography

A variation of high-performance liquid 

chromatography (HPLC) that separates 

molecules using a hydrophilic stationary 
phase and an organic-yet-water-miscible 

liquid phase

Separation, identification, and quantification of 

glycans and glycopeptides 

WAX–HPLC

Weak-anion exchange–

high-performance liquid 

chromatography

Separates anionic molecules based on 

their degree of charge

Separation, identification, and quantification of 

glycans and glycopeptides

RP–HPLC

Reverse-phase–high-

performance liquid 

chromatography

Separates molecules on the basis of 

differences in the strength of their interaction 

with a hydrophobic stationary phase

Separation, identification, and quantification of 

glycans and glycopeptides

ES684990_BP1015_034.pgs  10.05.2015  23:09    ADV  

black

yellow

magenta

cyan

[ 7 ]

36   

BioPharm International  www.biopharminternational.com  October 2015

Glycosylation

comparing  analyses  of  sialylation 
and  antennary  structure  (13).  This 
particular  study  did  not  address 
the  potentially  added  variability 
caused  by  sample  preparation.  The 
variability  in  outcomes  may  be 
due  in  part  to  the  availability  of 
numerous  analytical  approaches 
and  differences  between  labs  as 
to  the  selection  of  approach  and 
limitations of available equipment. 
Inconsistent  levels  of  training  and 
expertise  in  glycan  analysis  may 
also have had an impact.

basic reQuirements  

for stanDarDizeD protocols  

of Glycan analysis
The  establishment  of  a  robust  pro-
tocol  for  glycan  analysis  can  help 
extract  the  maximum  benefit  from 
QbD  practice;  give  manufacturers 
greater  control  over  product  quality 
and  comparability  between  batches 
and  process  modifications;  and 
ensure consistency and quality in reg-
ulatory  submissions.  Such  a  protocol 
should have the following features.

Well-characterized  
reference standards
A  selection  of  known  glycopro-
teins,  glycopeptides,  released  gly-
cans,  and  monosaccharides  will 
help calibrate and validate any sys-
tem  of  glycoprofiling  used  in  the 
initial characterization of the prod-
uct  or  monitoring  of  the  manufac-
turing process.

Well-characterized  
sample standards
Isolated  product  with  a  known 
clinical  safety  and  efficacy  profile 
provides a reference point for com-
paring  glycan  structure  of  batch 
products  under  different  process 
conditions and times.

Comprehensive identification  
of critical glycan attributes
Structural  features  of  glycans  have 
been  linked  to  circulating  half-life 
of  the  glycoprotein  in  the  blood 
(sialylation);  placental  transport 
(galactosylation);  antibody-depen-
dent  cell-mediated  cytotoxicity 

(core  fucosylation);  and  a  wide 
range  of  effector  functions,  bio-
availability,  and  safety  character-
istics  (14,  15).  Critical  attributes 
may include:
•  Antennary profile
•   Sia lylat ion  st ate,  i nc lud i ng 

degree  and  linkage  type  (∝2-3 

vs. ∝2-6)

•  Site-specific  glycosylation  pro-

files and occupancy

•  Fucosylation
•  Galactosylation
•  N-acetyl-lactosamine repeats
•  High mannose residues composi-

tion

•  Absence of immunogenic elements 

such  as  N-glycolylneuraminic 
acid  ( Neu5Gc),  deacet ylated 
N-acetylneuraminic acid (Neu5Ac), 
and Gal∝(1-3)Gal.

Variations  in  these  CQAs  intro-

duced  by  manufacturing  can  orig-
inate  from  selection  of  cell  line, 
bioreactor conditions such as nutri-
ent levels, pH or oxygen content, as 
well  as  inadvertent  modifications 
during downstream purification.

Table III: 

Common detection methods in glycan analysis.

Detection method

Description

Application

PAD

Pulsed amperometric 

detection

Permits detection without fouling 

electrodes

Detection of non-derivatized glycans 

(most sugars do not absorb UV). 

Frequently linked to high-pH anion-

exchange chromatography (HPAEC)

FD

Fluoresence detection

Selective fluorescent labeling and 

detection, which may use derivatizing 

agents 

For the analysis of derivatized glycans 

and glycopeptides when native 

detection does not offer sufficient 

sensitivity. Increases options for 

chromatographic separation methods. 
May also enhance mass spectrometric 

(MS) analysis

ESI–MS

Electrospray ionization–

mass spectrometry

Mass measurement of gas-phase 
ionized molecular species, where 

ions are generated by applying a high 

voltage to a liquid to create an aerosol, 

with little fragmentation of molecules. 

Can be directly integrated with liquid 

chromatography

Mass mapping of glycans and 
glycopeptides (including non-

derivatized) for identification of 

sequence, antennary pattern, 

modifications, and heterogeneity, etc.

MALDI–TOF MS

Matrix-assisted laser 

desorption ionization–

time-of-flight mass 

spectrometry

Mass measurement of gas-phase 

ionized molecular species, where ions 

are generated by embedding molecules 

in a solid matrix, and releasing them as 

ions via laser ablation

Mass mapping of glycans and 
glycopeptides (including non-

derivatized) for identification of 

sequence, antennary pattern, 

modifications, and heterogeneity, etc.

ES684992_BP1015_036.pgs  10.05.2015  23:09    ADV  

black

yellow

magenta

cyan

[ INTRODUCTION ]

[ 8 ]

October 2015 

www.biopharminternational.com  BioPharm International    37

Glycosylation

Establishment of ranges  
of acceptable variation in  
complex glycosylation patterns
M a ny  g lycoprote i n s,  p a r t ic u -
larly  those  with  multiple  glyco-
sylation  sites,  do  not  exist  as  a 
single species, but as a mixture of 
glycoforms. The natural complex-
ity  and  heterogeneity  of  glycan 
str uct ures  can  have  impor tant 
functional  relevance  for  a  pro-
tein,  and  even  minor,  low-abun-
dance  glycoform  species  can  be 
c r uc ia l.  For  clinica l  pu r poses, 
each  product  may  have  a  differ-
ent  tolerance  or  requirement  for 
glycoform  distribution.  In  par-
ticular,  clarity  on  the  extent  to 
which  low-abundance  glycoforms 
should  be  identified  and  moni-
tored is essential.

Adherence to best practices  
in sample preparation
Selecting  the  most  appropriate 
method  from  the  wide  range  of 
published  and  commercial  sample 
preparation methods can be daunt-
ing.  For  example,  purification  of 
glycans  after  release  from  protein 
may  be  performed  by  solvent  pre-
cipitation,  solid-phase  extraction, 
or  size-exclusion,  hydrophobic-
interaction,  or  hydrophilic-inter-
ac t ion  ch romatog raphy.  Some 
methods  may  lead  to  a  non-stoi-
chiometric recovery of oligosaccha-
rides, skewing the results of glycan 
profiling.  Recent  developments  in 
sample  preparation  have  allowed 
for  a  reduction  in  preparation 
times  and  improved  quantitative 
yields of both high- and low-abun-
dance glycoforms (16).

Selection of glycoanalysis  
technologies, methods, and strategy
There  is  a  wide  array  of  technolo-
gies  that  can  be  applied  to  glycan 
analysis  (see  Tables  I–III).  A  series 
of  detailed  optimal  work f lows 
and  best  practices  would  help 
to  harmonize  analytical  proce-
dures  between  and  within  orga-

nizations  that  submit  regulatory 
reports.  Workflows  would  cover 
initial  characterization  through 
to  routine  monitoring  and  quality 
control.  Considerations  should  be 
made  with  respect  to  the  simplic-
ity  and  time  of  analysis,  as  long 
as  the  required  levels  of  accuracy 
and  reproducibility  are  not  com-
promised.
  T he  u s e  of  or t hogon a l  a nd 
c o m p l e m e n t a r y   m e t h o d s   o f 
analysis  help  compensate  for  sys-
tematic  errors  in  measurement. 
These  methods  typically  isolate 
molecules  and  their  fragments 
based  on  different  physical  prop-
er t ies  (e.g.,  h igh-per for ma nce 
capillary  electrophoresis  [HPCE] 
vs.  hydrophilic  interaction  liq-
uid  chromatography  [HILIC])  or 
analytical  treatment  (e.g.,  electro-
spray  ionization–mass  spectrom-
etry  [ESI–MS]  vs.  matrix-assisted 
l a s e r   d e s o r p t i o n / i o n i z a t i o n –
time-of-flight  mass  spectrometry 
[MALDI–TOF–MS]),  and  are  com-
pared  to  compensate  for  potential 
bias  introduced  by  each  analytical 
method.

conclusion
Pharmaceutical  regulatory  agen-
cies  worldwide  have  laid  out  the 
general  principles  of  quality  con-
trol  and  risk  management  in  bio-
pharmaceutical  manufacturing. 
Of  the  many  CQAs  that  require 
consideration, the variation of the 
N-linked  and  O-linked  glycosyl-
ation  profiles  of  biotherapeutic 
glycoproteins  is  one  of  the  most 
complex to assess. Currently, there 
are  numerous  methods  used  to 
elucidate  these  structures  with 
varying  degrees  of  accuracy  and 
precision.  In  addition,  the  use  of 
these  somewhat  disparate  meth-
odologies makes it not always pos-
sible  to  directly  compare  results 
bet ween  laborator ies.  To  meet 
regulatory  requirements  for  con-
sistent  process  and  quality  con-
trol,  it  would  be  benef icial  to 

establish  more  specific  and  stan-
dardized  g uidelines  for  glycan 
analysis  performance  with  respect 
to  reproducibility,  accuracy,  and 
sensitivity for the characterization 
and  routine  monitoring  of  critical 
glycoforms, including those of low 
abundance. While such guidelines 
are  within  purview  of  national 
reg ulator y  bodies  and  interna-
tional  consensus  organizations 
(such  as  ICH),  no  such  guidelines 
have  been  released  to  date.  The 
requirements  for  glycan  analy-
sis  described  in  this  article  could 
address  many  of  the  issues  related 
to  process  and  quality  control  in 
glycoprotein manufacturing.

references
  1.   G. Walsh, Nature Biotech. 28 (9), pp. 

917–924 (2010).

  2.   G. Mack, Nature Biotech. 26 (6), pp. 

592 (2008).

  3.   ICH, Q5E Specifications: Test 

Procedures and Acceptance Criteria for 
Biotechnological/Biological Products

, 

EMA Document CPMP/ICH/5721/03 
(Geneva, 2003).

  4.   ICH Q6B Specifications: Test Procedures 

and Acceptance Criteria for 
Biotechnological/Biological Products

, 

EMA Document CPMP/ICH/365/96 
(Geneva, 1999).

  5.   FDA, Guidance for Industry, PAT—A 

Framework for Innovative 
Pharmaceutical Development, 
Manufacturing, and Quality Assurance 
(Rockville, MD, Sept. 2004).

  6.   R. Jefferis, Biotechnol. Prog. 21, pp. 

11–16 (2005).

  7.   S.A. Brooks, Mol. Biotechnol. 28 (3), pp. 

241–255 (2004).

  8.   EMA, Guideline on Development, 

Production, Characterization and 
Specifications for Monoclonal Antibodies 
and Related Products

, EMEA/CHMP/

BWP/157653/2007 (London, 2007).

  9.   FDA, Guidance for Industry, Q8 

Pharmaceutical Development 

(Rockville, 

MD, May 2006).

 10.   FDA, Guidance for Industry, Q9 Quality 

Risk Management

, (Rockville, MD, June 

2006).

 11.   A.S. Rathore, A. Sharma, and D. Chillin, 

BioPharm Int.

 19, pp. 48–57 (2006).

 12.   A.S. Rathore, Trends Biotechnol. 27 (9), 

pp. 546–553 (2009).

 13.   S. Thobhani et al., Glycobiology, 19 (3), 

pp. 201–211 (2009).

 14.   T. Kibe et al., J. Clin. Biochem. Nutr. 21 

(1), pp. 57–63 (1996).

 15.   A. Okazaki et al., J. Mol. Biol. 336 (5), 

pp. 1239–1249 (2004).

 16.   M.A. Lauber et al., Anal. Chem. 87 (10), 

pp. 5401–5409 (2015). ◆

ES684991_BP1015_037.pgs  10.05.2015  23:09    ADV  

black

yellow

magenta

cyan

[ 9 ]

INTACT PROTEINS

[ 10 ]

WAT E R S S O LU T IO NS
ACQUITY UPLC® Glycoprotein BEH Amide,  
300Å Column

Glycoprotein Performance Test Standard

ACQUITY UPLC H-Class Bio System

Xevo® G2 QTof Mass Spectrometer

K E Y W O R D S
ACQUITY UPLC H-Class Bio System,  
BEH Amide 300Å, glycans,  
glycosylated proteins, glycosylation, HILIC

A P P L I C AT IO N  B E N E F I T S 

■

■

Improved HILIC separations of intact  
protein glycoforms through optimization  
of stationary phase (bonded phase and pore 
size), ion pairing, column pressurization,  
and injection approaches.

■

■

MS-compatible HILIC to enable detailed 
investigations of sample constituents.

■

■

Orthogonal selectivity to conventional 
reversed-phase (RP) separations  
for enhanced characterization of  
glycoprotein samples.

■

■

Glycoprotein BEH Amide, 300Å,  
1.7 µm stationary phase is QC tested via  
a glycoprotein separation to ensure 
consistent batch to batch reproducibility.

I N T RO DU C T IO N
Hydrophilic interaction chromatography (HILIC) has been widely adopted as a 
tool for separating highly polar compounds. In fact, it has become a relatively 
widespread technique for small molecule separations. By comparison, the 
application of HILIC to large biomolecules has been comparatively limited 
even though there are instances in which the separation selectivity of HILIC 
would be highly valuable, for example during the characterization of protein 
glycosylation. A standard approach to the analysis of glycans involves their 
enzymatic or chemical release from their counterpart protein followed by their 
chromatographic separation using HILIC. UPLC®-based separations founded 
upon an optimized, sub-2-µm amide-bonded stationary phase has transformed 
HILIC separations of released glycans by facilitating faster, higher resolution 
separations.1-2 Although released glycan analysis is a gold-standard approach, it 
has historically required lengthy and at times cumbersome sample preparation 
techniques. And while the recent introduction of the GlycoWorks™ RapiFluor-
MS™ N-Glycan Kit alleviates many of these shortcomings,3 alternative means 
of characterizing protein glycosylation must sometimes be investigated,4-6 for 
instance when it is of interest to elucidate sites of modification.7

To enable the complementary analysis of glycans as they are still attached to 
their counterpart proteins, we present an optimized HILIC stationary phase and 
corresponding methods for resolving the glycoforms of intact and digested 
glycoproteins. A wide-pore (300Å) amide-bonded, organosilica (ethylene 
bridged hybrid; BEH)8 stationary phase is employed along with rigorously 
developed methods to achieve unprecedented separations of the glycoforms of 
intact proteins ranging in mass from 10 to 150 kDa. 

Developing High Resolution HILIC Separations of Intact Glycosylated Proteins 

Using a Wide-Pore Amide-Bonded Stationary Phase 
Matthew A. Lauber, Scott A. McCall, Bonnie A. Alden, Pamela C. Iraneta, and Stephan M. Koza

Waters Corporation, Milford, MA, USA

[ 11 ]

E X P E R IM E N TA L 

LC conditions
LC system:  

ACQUITY UPLC H-Class Bio System

Sample temp.:  

5 °C

Analytical 
column temp.:  

30 °C  
(unless noted otherwise in the caption)

UV detection:  

214/280 nm, 2 Hz

Fluorescence detection: Ex 280/Em 320 nm, 10 Hz 

Flow rate: 

0.2 mL/min

Injection volume:  

≤1 µL (aqueous diluents). Note: It might 
be necessary to avoid high organic 
diluents for some samples due to the 
propensity for proteins to precipitate 
under ambient conditions. A 2.1 mm I.D. 
column can accommodate up to a 1.2 µL 
aqueous injection before chromatographic 
performance is negatively affected. 

Columns:  

ACQUITY UPLC Glycoprotein BEH 
Amide, 300Å, 1.7 µm, 2.1 x 150 mm 
(p/n 176003702, with Glycoprotein 
Performance Test Standard);

ACQUITY UPLC Glycoprotein BEH 
Amide 300Å, 1.7 µm, 2.1 x 100 mm 
(p/n 176003701, with Glycoprotein 
Performance Test Standard);

ACQUITY UPLC BEH HILIC, 130Å, 1.7 µm, 
2.1 x 150 mm (p/n 186003462)

XBridge BEH HILIC, 130Å, 5 µm,  
2.1 x 150 mm (p/n 186004446)

ACQUITY UPLC Glycan BEH Amide, 130Å, 
1.7 µm, 2.1 x 150 mm (p/n 186004742);

ACQUITY UPLC Glycan BEH Amide, 130Å, 
1.7 µm, 2.1 x 100 mm (p/n 186004741);

Sample description
Glycoprotein Performance Test Standard (a formulation of bovine 
RNase A and RNase B, p/n 186008010) and RNase B (Sigma R7884) 
were reconstituted in 18.2 MΩ water to a concentration  
of 2 mg/mL. Trastuzumab was diluted with water from its formulated 
concentration of 21 mg/mL to a concentration of 2 mg/mL. 

For column conditioning, the components of a vial of Glycoprotein 
Performance Test Standard (100 µg) were dissolved in 25 µL of 
0.1% trifluoroacetic acid (TFA), 80% acetonitrile (ACN) to create  
a 4 mg/mL protein solution.

To investigate the resolution of glycan occupancy isoforms, Intact 
mAb Mass Check Standard (p/n 186006552) was deglycosylated 
using the following techniques. The glycoprotein (15 µg) was 
reconstituted to a concentration of 0.52 mg/mL into a 28.2 μL 
solution of 1% (w/v) RapiGest™ SF Surfactant and 50 mM HEPES 
(pH 7.9). This solution was heated to 90 °C over 3 minutes, 
allowed to cool to 50 °C, and mixed with 1.2 μL of GlycoWorks 
Rapid PNGase F solution. Deglycosylation was completed by 
incubating the samples at 50 °C for 5 minutes. To produce 
partial deglycosylation, Intact mAb Mass Check Standard was 
deglycosylated using only a 5 minute, 50 °C incubation with 
PNGase F without a heat-assisted pre-denaturation.

Method conditions 
(unless otherwise noted) 

Column conditioning
New (previously unused) ACQUITY UPLC Glycoprotein BEH 
Amide, 300Å, 1.7 µm Columns should be conditioned, before 
actual test sample analyses, via two sequential injections and 
separations of 40 µg Glycoprotein Performance Test Standard 
(10 µL injections of 4 mg/mL in 0.1% TFA, 80% ACN) or with 
equivalent loads of a test sample for which the column has been 
acquired. The separation outlined in Figure 2 can be employed for 
conditioning with the Glycoprotein Performance Test Standard. 

[ 12 ] Developing High Resolution HILIC Separations of Intact Glycosylated Proteins Using a Wide-Pore Amide-Bonded Stationary Phase

Competitor columns:   PolyHYDROXYETHYL A™, 300Å,  

3 µm, 2.1 x 100 mm; 

 Glycoplex® A, 3 µm, 2.1 x 100 mm; 

ZORBAX® RRHD 300-HILIC, 300Å,  
1.8 µm, 2.1 x 100 mm; 

 Halo® PentaHILIC, 90Å,  

2.7 µm, 2.1 x 100 mm;

 SeQuant® ZIC-HILIC, 200Å,  

3.5 µm, 2.1 x 100 mm; 

Accucore™ Amide, 150Å,  
2.6 µm, 2.1 x 100 mm;

 TSKgel® Amide-80, 80Å,  

3 µm, 2.0 x 100 mm

Column connector 
(for coupling 
150 mm columns): 

0.005 x 1.75 mm UPLC SEC Connection 
Tubing (p/n 186006613)

Vials:  

Polypropylene 12 x 32 mm , 300 μL  
Screw Neck Vial, (p/n 186002640)

Gradient used to demonstrate the progression of HILIC separation 
technologies (Figure 1): 

Column dimension: 

2.1 x 150 mm

Mobile phase A:  

0.1% (v/v) TFA, water

Mobile phase B:  

0.1% (v/v) TFA, ACN

  Time 

%A 

%B 

Curve 

 0  20.0  80.0  6 
 20  80.0  20.0  6 
 21  20.0  80.0  6 
 30  20.0  80.0  6

Focused gradient for RNase B HILIC separations (Figures 2 and 5)

Column dimension: 

2.1 x 150 mm

Mobile phase A:  

0.1% (v/v) TFA, water

Mobile phase B:  

0.1% (v/v) TFA, ACN

  Time  %A 

%B  Curve 

  0  20.0 80.0  6 
  1  34.0 66.0  6 
 21  41.0  59.0  6 
 22  100.0  0.0  6 
 24  100.0 0.0  6 
 25  20.0 80.0  6 
 35  20.0 80.0  6

Gradient for benchmarking/evaluations (Figure 3)

Column dimension: 

2.1 x 100 mm

Mobile phase A:  

0.1% (v/v) TFA, water

Mobile phase B:  

0.1% (v/v) TFA, ACN

  Time 

%A 

%B 

Curve 

  0.0  20.0 80.0  6 
 0.7  30.0  70.0 

6 

 29.3  45.0  55.0 

6 

 30.0  80.0  20.0 

6 

 31.3  80.0  20.0 

6 

 32.0  20.0  80.0 

6 

 40.0  20.0  80.0 

6

Gradient employed to select a mobile phase additive (Figure 4):

Column dimension: 

2.1 x 150 mm

Mobile phase A:  

0.1% (v/v) TFA, water or 50 mM  
ammonium formate, pH 4.4 or  
0.5% (w/v) formic acid, water

Mobile phase B:  

ACN

Time 

%A 

%B 

Curve 

  0  20.0 80.0  6 
 20 80.0  20.0  6 
 21 20.0 80.0  6 
 30 20.0 80.0  6

Focused gradient for reversed phase of RNase B (Figure 6):

Column dimension: 

2.1 x 150 mm

Mobile phase A:  

0.1% (v/v) TFA, water

Mobile phase B:  

0.1% (v/v) TFA, ACN

Time 

%A 

%B 

Curve 

 0 95.0  5.0  6 
 1 74.5 25.5 6 
 21 67.5  32.5  6 
 22 10.0  90.0  6 
 24 10.0  90.0  6 
 25 95.0  5.0  6 
 35 95.0  5.0  6

[ 13 ]

Developing High Resolution HILIC Separations of Intact Glycosylated Proteins Using a Wide-Pore Amide-Bonded Stationary Phase

Focused gradient for intact trastuzumab (Figures 7 and 8)

Column dimension: 

2.1 x 150 mm, with varying lengths  
25 µm I.D. PEEK post-column tubing

Or two coupled 2.1 x 150 mm columns

Mobile phase A:  

0.1% (v/v) TFA, water

Mobile phase B:  

0.1% (v/v) TFA, ACN

Time 

%A 

%B 

Curve 

  0  20.0 80.0  6 
  1  30.0 70.0  6 
 21 37.0 63.0  6 
 22 70.0 30.0  6 
 24 70.0 30.0  6 
 25 20.0 80.0  6 
 45 20.0 80.0  6

Conditions for resolving glycan occupancy isoforms  
of an IgG (Figure 9):

Column dimension: 

Two coupled 2.1 x 150 mm or a single  
2.1 x 150 mm 

Column temp.:  

80 °C

Mobile phase A:  

0.1% TFA, 0.3% HFIP in water 

Mobile phase B:  

0.1% TFA, 0.3% HFIP in ACN 

Time 

%A 

%B 

Curve 

 0.0  20  80  6 
 10.0  50  50 

6 

 11.0 100  0 

6 

 14.0 100  0 

6 

 15.0  20  80 

6 

 25.0  20  80 

6

MS conditions
MS system:  

Xevo G2 QTof

Ionization mode:  

ESI+

Analyzer mode:  

Resolution (~20 K)

Capillary voltage:  

3.0 kV

Cone voltage:  

45 V

Source temp.:  

150 °C

Desolvation temp.:  

350 °C

Desolvation gas flow:   800 L/Hr

Calibration:  

NaI, 2 µg/µL from 100–2000 m/z

Acquisition:  

500–4000 m/z, 0.5 sec scan rate 

Data management: 

MassLynx® Software (v4.1)

R E SU LT S A N D D IS C U S S IO N

Progression of HILIC technology for  
glycoprotein separations
HILIC originated in the early 1990s as a separation technique 
to resolve highly polar molecules using mobile phases adapted 
from reversed phase chromatography.9 The HILIC separation 
mechanism is largely believed to be dependent on a polar 
stationary phase that adopts an immobilized water layer.9 
Hydrophilic analytes partition into this immobilized water layer 
and undergo interaction with the phase via a combination of 
hydrogen bonding, dipole-dipole, and ionic interactions. In this 
way, hydrophilic analytes will be retained on the HILIC phase 
under apolar initial mobile phase conditions and later eluted  
by increasing polar mobile phase concentration via use of  
an LC gradient.9

Numerous HILIC or HILIC-like stationary phases have been 
developed in the last two decades. Many based solely on 
unbonded silica particles are widely available, so too are 
HILIC phases based on polyalcohol bondings or charge bearing 
surfaces, such as those with zwitterionic bondings. For the 
enhanced retention and selectivity of glycans, amide bonded 
phases have become increasingly popular. The ACQUITY UPLC 
Glycan BEH Amide stationary phase found in Waters Glycan 
Column has, for instance, found wide-spread use for high 
resolution released glycan separations. 

As mentioned before, HILIC has, however, not seen wide-spread 
use in intact large molecule applications. Concerns that high 
organic solvent concentrations can result in protein precipitation 
have most likely discouraged many from attempting to develop 
HILIC-based, protein separation methods. Endeavoring beyond 
these perceptions, we have developed a new amide-bonded 
stationary phase based on a wide-pore, organosilica (ethylene 
bridged hybrid; BEH) particle that was specifically designed to 
facilitate large molecule separations. It exhibits a porous network 
accessible to most proteins and an average pore diameter that 
does not impart significant peak broadening due to restricted 
diffusion, which can occur when protein analytes are too close  
in size to the average pore diameter of a stationary phase  
(e.g. within a factor of 3). 

[ 14 ] Developing High Resolution HILIC Separations of Intact Glycosylated Proteins Using a Wide-Pore Amide-Bonded Stationary Phase

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

5

6

7

8

9

10

11

12

13

14

15

A

21

4

Time (min)

Unbonded BEH 

1.7  m 
130Å 

Amide-Bonded BEH 

1.7  m 
130Å 

Amide-Bonded BEH 

1.7  m 
300Å 

Improved  

Resolution 

Unbonded BEH 

5  m 
130Å 

Glycoprotein BEH Amide,  

300Å, 1.7  m 

Figure 1. Progression of HILIC stationary-phase technologies for intact glycoprotein separations. 
Separation of 1 µg of RNase B using 2.1 x 150 mm columns packed with stationary phases  
ranging from HPLC-size unbonded organosilica (XBridge BEH HILIC, 130Å, 5 µm) to  
sub-2-µm amide-bonded organosilica 300Å, 1.7 µm particles (ACQUITY UPLC Glycoprotein  
BEH Amide 300Å, 1.7 µm). 

Figure 2. Separations of the Glycoprotein Performance Test Standard (RNase A + RNase B 
glycoforms) using a Glycoprotein BEH Amide 300Å, 1.7 µm Column versus a BEH Amide, 130Å, 
1.7 µm Column. The reported resolution values were calculated using the half-height peak widths of 
species 1 and 2 (RNase A and RNase B Man5 glycoforms, respectively). Fluorescence detection at Ex 
280 nm and Em 320 nm and a column temperature of 45 °C were employed in this example.

EU

0.00

20.00

40.00

60.00

80.00

100.00

120.00

140.00

160.00

5.00 

10.00

15.00

20.00

EU

0.00

20.00

40.00

60.00

80.00

100.00

120.00

140.00

160.00

Minutes

5.00 

10.00

15.00

20.00

Rs,(1/2) 

21.2 

Glycoprotein BEH 

Amide 

300Å, 1.7 m 

Rs,(1/2) 

17.1 

ACQUITY UPLC

Glycan BEH 

Amide 

130Å, 1.7 m 

1 

2 

3 

4 

5 

6 

Glycoprotein Performance 

Test Standard 

300Å 

130Å 

Peak 

Species 

Rs

Rs

1 

RNase A 

–

–

2 

RNase B (+Man 5) 

21.2 

17.1 

3 

RNase B (+Man 6) 

3.5 

2.8 

4 

RNase B (+Man 7) 

2.7 

2.3 

5 

RNase B (+Man 8) 

2.6 

2.3 

6 

RNase B (+Man 9) 

3.1 

2.4 

The progression of HILIC technology culminating 
in this new stationary phase is remarkable. The 
emerging technology of large molecule HILIC can be 
captured by separations of bovine ribonuclease B 
(RNase B), a 13 kDa protein comprised of several high 
mannose (Man5 to Man9) glycoforms. Figure 1  
shows RNase B separated by several different 
stationary phases. From bottom to top, increasingly 
better separations of RNase B were achieved as 
increasingly newer chromatographic technologies 
were adopted, from 5 µm to 1.7 µm particles, 
from unbonded to amide bonded particles, and 
from standard pore diameter (130Å) to wide-pore 
diameter (300Å) particles. It is with BEH Amide, 
300Å, 1.7 µm particles that RNase B glycoforms are 
best separated. The use of a wide-pore stationary 
phase plays a significant role in achieving optimal 
resolution. This is highlighted in Figure 2 wherein 
benchmarking results are presented from the use of 
a newly developed test mixture, called Glycoprotein 
Performance Test Standard, which contains bovine 
RNase B, its corresponding glycoforms and its  
aglycosylated isoform (RNase A). Example 
separations are provided for this standard wherein  
a focused gradient has been used with the wide-pore 
(300Å) BEH Amide as well as the standard pore size 
(130Å) BEH Amide stationary phase. Notice that the 
widepore amide column affords a measurable (24%) 
increase in the resolution between the aglycosylated 
RNase A isoform and the Man5 glycoform of  
RNase B, in addition to sizeable increases in 
resolution throughout the separation.

[ 15 ]

Developing High Resolution HILIC Separations of Intact Glycosylated Proteins Using a Wide-Pore Amide-Bonded Stationary Phase

Time

8.00

10.00

12.00

14.00

16.00

18.00

20.00

22.00

5.0e-2

1.0e-1

8.00

10.00

12.00

14.00

16.00

18.00

20.00

22.00

5.0e-2

1.0e-1

BEH Amide

1.7  m 

130Å

BEH Amide

1.7  m

300Å

A 

B 

0.00

10.00

20.00

30.00

0.0

2.0e-1

4.0e-1

0.00

10.00

20.00

30.00

0.0

2.0e-1

4.0e-1

0.00

10.00

20.00

30.00

0.0

2.0e-1

4.0e-1

0.00

10.00

20.00

30.00

0.0

2.0e-1

4.0e-1

0.00

10.00

20.00

30.00

0.0

2.0e-1

4.0e-1

Poor 

Retentivity

Poor Recovery

ZIC-HILIC

3.5  m

200Å

Halo 

PentaHILIC

2.7  m

90Å

PolyGLYCOPLEX A

3  m

PolyHYDROXYETHYL A

3  m

300 Å

Zorbax 300 HILIC

1.8  m

300 Å

0.00

10.00

20.00

30.00

0.0

2.0e-1

4.0e-1

0.00

10.00

20.00

30.00

0.0

2.0e-1

4.0e-1

Amide-80

3  m

80Å

AccucoreAmide

2.6  m

150Å

BEH Amide

1.7  m

130Å

BEH Amide

1.7  m

300Å

Desirable Retentivity

Poor 

Resolution

0.00

10.00

20.00

30.00

0.0

2.0e-1

4.0e-1

0.00

10.00

20.00

30.00

0.0

2.0e-1

4.0e-1

Figure 3. Evaluation of commercially available HILIC columns for intact glycoprotein separations. (A) UV chromatograms obtained for RNase B using 10 different 
stationary phases. (B) Zoomed HILIC UV chromatograms for the highest resolution separations.

The significance of these recent developments becomes more apparent when benchmarked against other 
commercially available HILIC phases. RNase B separations resulting from an evaluation of 10 different 
HILIC stationary phases are shown in Figure 3. It can be seen that 6 out of the 10 evaluated materials showed 
undesirable characteristics, including poor recovery and poor retention. It was only with the amide bonded 
stationary phases and particle technologies based on 100Å or greater pore diameters that reasonable 
separations of RNase B glycoforms could be achieved.

Mobile phase optimization, MS compatibility, and orthogonality to reversed phase
High resolution HILIC separations of protein glycoforms require that mobile phase selection be given 
significant consideration. Most HILIC separations have been developed so as to rely on ammonium  
salts (formate or acetate) to mitigate significant ionic interactions and to control mobile phase pH.  
The suitability of such mobile phase systems to glycoproteins was evaluated using RNase B.  

[ 16 ] Developing High Resolution HILIC Separations of Intact Glycosylated Proteins Using a Wide-Pore Amide-Bonded Stationary Phase

Figure 4. Optimization of mobile phase conditions for separations of intact and digested glycoproteins. (A) UV chromatograms obtained for RNase B when using  
various mobile phases and a Glycoprotein BEH Amide, 300Å, 1.7 µm, 2.1 x 150 mm Column. (B) Schematic portraying the utility of ion pairing for glycoprotein  
HILIC separations. Reduced hydrophilicity imparted via ion pairing with a hydrophobic, strong acid is displayed with shading. [PDB:1RBB] 

Figure 4 shows the corresponding RNase B chromatogram obtained when 0.1% TFA is used as the mobile phase 
modifier instead of 50 mM ammonium formate or 0.5% formic acid, two mobile phase compositions more 
frequently used for HILIC separations.2,7 It is with 0.1% TFA that glycoforms are best resolved. Along with 
enhancing glycoform resolution, the TFA-modified mobile phase reduced the retention of RNase B. Together 
these observations highlight the significance of acidic, ion pairing mobile phases to being able to achieve high 
resolution glycoprotein separations using HILIC. It is proposed that the acidic condition imparted by the TFA 
ensures that acidic residues of the protein are fully protonated and thus present in their more hydrophobic state. 
In addition, the ion pairing of the TFA counter ion to basic residues, ensures that cationic residues will also be 
separated in a more hydrophobic form. In this way, retention of a glycoprotein onto a HILIC phase is  
driven primarily by the glycans and a separation more selective to resolving differences in the glycan 
modification is achieved.  

0.005

0.005

0.015

0.025

8

9

10

11

12

13

14

15

16

17

18

A

280

Time (min)

50 mM Ammonium Formate

pH 4.4

0.5% FA

0.1% TFA

HOOC-

-COO

H  

-NH3

+ 

-N

H

3 + 

HO

OC-

+H

3

N-

+

H3N- 

F3C-COO- 

-

OOC-CF3 

F3C-COO- 

-

OOC-CF3 

-NH3

+ 

-

OO

O C-CF

H

-N

3 + 

+

-

OOC-CF3

+H

3

N-

F3C--COO

CO -

-COO

H

HO

OC-

HOOC-

+

H3N-

F3C-COO-

TFA 

Ion Pairing 

A 

B 

Fortunately, TFA-modified mobile phases can be readily coupled to ESI-MS, due to their volatility. This  
aspect of the developed HILIC methods enables on-line characterization of the resolved glycoforms and 
presents a new option for profiling a sample containing glycosylated protein. To this end, the peaks resolved 
from RNase B using a BEH Amide, 300Å, 1.7 µm column were subjected to interrogation by ESI-MS.  

[ 17 ]

Developing High Resolution HILIC Separations of Intact Glycosylated Proteins Using a Wide-Pore Amide-Bonded Stationary Phase

Figure 5. HILIC-MS of RNase B. (A) UV (bottom) and TIC (top) chromatograms obtained for RNase B when using a focused gradient and a Glycoprotein BEH Amide,  
300Å, 1.7 µm, 2.1 x 150 mm Column. (B) Deconvoluted mass spectra obtained for each labeled peak along with corresponding glycoform identifications. 

Figure 5 shows both a UV chromatogram and a corresponding total ion chromatogram (TIC) obtained 
when separating RNase B. By summing and deconvoluting (MaxEnt™ 1) the mass spectra obtained for the 
six labeled peaks, it was confirmed that RNase B glycoforms were being detected. In fact, the observed 
deconvoluted masses support identifications of aglycosylated RNase B (RNase A) along with RNase B 
modified by Man5 through Man9. 

0

20000

40000

60000

80000

100000

120000

0

0.02

0.04

0.06

0.08

0.1

0.12

5

10

15

20

TIC Intensit

y

A

214

Time (min)

Blank

0E+0

5E+3

1E+4

2E+4

2E+4

3E+4

3E+4

13500

14000

14500

15000

15500

16000

Intensit

y

m/z

1

RNase B

aglycosylated

2

3

4

5

6

1)

2)

+Man5

3)

+Man6

4)
+Man7

5)

+Man8

6)

+Man9

=162 Da

=1217 Da

= 1217 Da, 2HexNAc/5Hex

  = 162 Da, 1Hex

MWavg, theoretical

MWavg, observed

RNase B (4 S-S)

13682.3

13682.5

+ Man5

14899.4

14899.7

+ Man6

15061.6

15061.8

+ Man7

15223.7

15223.9

+ Man8

15385.8

15385.9

+ Man9

15548.0

15548.3

A 

B 

[ 18 ] Developing High Resolution HILIC Separations of Intact Glycosylated Proteins Using a Wide-Pore Amide-Bonded Stationary Phase

Figure 6. Orthogonality of reversed phase with BEH C4, 300Å, 
1.7 µm and HILIC with BEH Amide, 300Å, 1.7 µm Columns. (A) 
Separation of RNase B (1 µg) using an ACQUITY UPLC Protein 
BEH C4, 300Å, 1.7 µm, 2.1 x 150 mm Column. (B) Separation of 
RNase B (1 µg) using an ACQUITY UPLC Glycoprotein BEH Amide, 
300Å, 1.7 µm, 2.1 x 150 mm Column. 

Finally, it should be pointed out that the newly developed stationary phase and the demonstrated 
methodologies provide new separation selectivity, one that is orthogonal and complementary to 
conventional reversed phase separations. Figure 6A shows that RNase B can, for instance, be separated by 
reversed-phase chromatography using a BEH C4, 300Å, 1.7 µm column so as to produce a high resolution 
separation of aglyocosylated RNAse B (RNase B) from its glycosylated isoforms. By reversed phase, 
however, none of high mannose glycoforms of RNase B can be resolved from one another. In contrast, a BEH 
Amide, 300Å, 1.7 µm column yields baseline resolution of each major glycoform (Figure 6B). 

0

0.09

0.18

0.27

A

214

0

0.02

0.04

0.06

6

8

10

12

14

16

18

20

A

214

Time (min)

+Man5

+Man6

+Man7

+Man8

+Man9

RNase B

aglycosylated

RNase B

aglycosylated

+Man5

+Man6

+Man7

+Man8
+Man9

BEH Amide

300Å 1.7 m

BEH C4

300Å 1.7 m

RNase A 

RNase A 

Separation of the heterogeneous glycoforms of an intact mAb
To explore the limits of this new technology, we have investigated the capabilities of resolving the 
glycoforms of intact mAbs. Specifically, separations of trastuzumab have been explored. These experiments 
required special considerations regarding sample injection, primarily because trastuzumab and numerous 
other glycoproteins are not readily soluble in high organic concentrations. In fact, 70–80% ACN is 
generally a solution condition that initiates the precipitation of proteins, such as an IgG. Accordingly, 
conditions for the optimal injection of aqueous diluents were developed. It has been found that a 2.1 mm I.D. 
column can accommodate an injection of aqueous sample up to 1 µL. From a 2 mg/mL aqueous sample of 
trastuzumab, appropriate sample mass loads could thus be injected and HILIC separations of the IgG could 
be performed. It should be mentioned that high ACN diluents can be used in intact protein HILIC, but care 
must be taken to enhance the solubility of the protein sample through either the use of TFA ion pairing at 
concentrations between 0.2–1.0%, the combined application of TFA and hexafluoroisopropanol (HFIP), or by 
use of co-solvents, such as dimethylsulfoxide (DMSO) (data not shown).

[ 19 ]

Developing High Resolution HILIC Separations of Intact Glycosylated Proteins Using a Wide-Pore Amide-Bonded Stationary Phase

As shown in Figure 7, trastuzumab can indeed be separated into multiple chromatographic peaks using a 
BEH Amide, 300Å, 1.7 µm column and an injection from a simple 100% aqueous diluent. However, at the 
backpressures produced from just a 150 mm length column, a noticeably tailing profile was observed. MS 
analysis indicated that the first set of peaks could be accurately assigned as the G0F/G0F, G0F/G1F, G1F/G1F, 
and G1F/G2F glycoforms of intact trastuzumab. An intact IgG is a dimeric structure, with a minimum of two 
N-glycan sites on two heavy chains, explaining the observation of combinatorially formed glycoforms. This  
is consistent with observations by intact mass analysis of IgGs.10 The tailing component of the chromatographic 
profile was in contrast found by MS to correspond to multiple, co-eluting trastuzumab glycoforms. With this 
result, we proposed that on-column aggregation was occurring and that increased column pressure could be a 
solution to HILIC of intact immunoglobulins, specifically since it had previously been reported that ultrahigh 
pressures can be beneficial to limiting carryover and ghosting during reversed phase of intact proteins.11 The 
effects of introducing additional column pressure was investigated by means of introducing varying lengths of 
narrow I.D., post-column PEEK tubing. Figure 7 (darker traces) displays the effects of introducing increasingly 
higher column pressure. By doubling the column pressure so that trastuzumab would elute under conditions 
of approximately 7,500 psi, the putative, aggregate peaks in the chromatographic profile were eliminated. It 
is encouraging that under these conditions the resulting chromatographic profile is represented by 5 major 
glycoforms, which again is consistent with ESI-MS of intact trastuzumab.10 It is interesting to additionally 
note that retention decreases as column pressure increases. This is a phenomenon that has been described 
previously for HILIC separations of monosaccharides.12 It has been proposed that increasingly higher pressures 
result in less coordination of water to the analyte and in turn reduced retention, an opposite effect to that 
observed during reversed phase chromatography.12 

0

0.1

0.2

10

15

20

A

21

4

Time (min)

Increasing 

Pressure

3200 psi

4500 psi

7300 psi

*Pressure at retention time of the mAb

Increasing Column Pressure

Minimizes On-Column Aggregation 

N-Linked Glycans

Fab

Fc

2 N-Glycans 

Figure 7. Effect of column pressure on the HILIC separation of an IgG. Trastuzumab (1 µg) was separated on Glycoprotein BEH Amide, 300Å, 1.7 µm, 2.1 x 150 mm 
Column with and without flow restriction. [PDB:1IGT]

[ 20 ] Developing High Resolution HILIC Separations of Intact Glycosylated Proteins Using a Wide-Pore Amide-Bonded Stationary Phase

Given that intact IgGs benefit from separations at 
ultrahigh pressures, we pursued separations based on 
the use of two BEH amide 300Å, 1.7 µm,  
2.1 x 150 mm columns coupled with a low volume, 
high pressure column connector. The separation for 
intact trastuzumab obtained with these coupled 
columns is displayed in Figure 8, along with 
extracted ion chromatograms that provide evidence 
to achieving separations of the glycoforms. This  
300 mm configuration provided the requisite column 
pressures for an optimal HILIC separation and 
additionally produced greater resolution between 
glycoforms. Clearly, additional theoretical plates 
are therefore advantageous during HILIC of even very 
high molecular weight species, which supports the 
significance of partitioning for such separations. 

An LC method for glycan occupancy
A UPLC HILIC separation of an intact IgG can be used 
for more than just an attempt to separate individual 
glycoforms. Equally interesting is the use of these 
new separation capabilities to resolve information 
about glycan occupancy. To this end, we evaluated 
the capabilities of the BEH Amide, 300Å column 
to assess the glycan occupancy of an IgG. This was 
exemplified by a study of reaction products resulting 
from various PNGase F deglycosylation treatments. 
Using an elevated 80 °C column temperature, TFA 
ion pairing, and an HFIP mobile phase additive, we 
have been successful in enhancing the solubility 
of IgGs and collapsing the fine structure otherwise 
captured for the individual, heterogenous intact  
IgG glycoforms (i.e. G0F/G0F versus G0F/G1F). 
Figure 9 presents HILIC fluorescence chromatograms 
resulting from such a separation of native Intact 
mAb Mass Check Standard (a murine IgG1 mAb) and 
its partially as well as completely deglycosylated 
isoforms. As can be seen, HILIC fluorescence  
profiles for these three samples are dramatically 
different. On-line mass spectrometric detection  
has confirmed that the peaks in these profiles 
correspond to different states of glycan occupancy.  

0

0.1

0.2

15

16

17

18

19

20

21

22

23

24

25

A

21

4

Time (min)

0

200

400

600

800

1000

1200

15

16

17

18

19

20

21

22

23

24

25

XI

C 

In

te

ns

it

y

Time (min)

(G0F)2

3023 ±0.6 m/z, 49+

G0F/G1F

3026 ±0.6 m/z, 49+ (G1F)2 or

G0F/G2F

3029 ±0.6 m/z, 49+

G1F/G2F

3033 ±0.6 m/z, 49+

(G2F)2

3062 ±0.6 m/z, 49+

Figure 8. Separation of intact trastuzumab glycoforms using coupled ACQUITY UPLC Glycoprotein 
BEH Amide, 300Å, 1.7 µm, 2.1 x 150 mm Columns. A UV chromatogram and extracted ion 
chromatograms for each of the major heterogenous glycoforms of trastuzumab are displayed.  
The column pressure at the retention time of the mAb was approximately 7,000 psi.

Figure 9. Assaying glycan occupancy and deglycosylation by intact protein HILIC-FLR-MS. HILIC 
fluorescence profiles obtained for three different samples are shown: (A) native, (B) partially 
deglycosylated, and (C) completely deglycosylated Intact mAb Mass Check Standard. Samples  
of this mAb (1.5 µg) were separated using two coupled Glycoprotein BEH Amide, 300Å, 1.7 µm, 
2.1 x 150 mm Columns. HILIC fluorescence profiles of partially deglycosylated Intact mAb Mass 
Check Standard using a (D) ACQUITY UPLC Glycan BEH Amide, 130Å, 1.7 µm, 2.1 x 150 mm 
Column versus a (E) Glycoprotein BEH Amide, 300Å, 1.7 µm, 2.1 x 150 mm Column.

145000

150000

145000

150000

146786

146952

145000

150000

148240

148560

Intact

-1 Glycan

-2 Glycans

PNGase F

148.4 kDa

145.3 kDa

146.8 kDa

9

14 min

PNGase F

Untreated 

Native mAb 

Partially PNGase F 

Deglycosylated mAb 

Completely PNGase F 

Deglycosylated 

A 

B 

C 

Time

8.00

10.00

12.00

14.00

EU x 10

e4

0.000

1000000.063

2000000.125

3000000.250

4000000.250

5000000.500

6000000.500

7000000.500

D 

E 

Glycan BEH Amide 

130Å, 1.7  m 

Glycoprotein BEH Amide 

300Å, 1.7  m 

[ 21 ]

Developing High Resolution HILIC Separations of Intact Glycosylated Proteins Using a Wide-Pore Amide-Bonded Stationary Phase

Waters Corporation 
34 Maple Street 
Milford, MA 01757 U.S.A. 
T: 1 508 478 2000 
F: 1 508 872 1990 
www.waters.com

Waters, The Science of What’s Possible, ACQUITY UPLC, Oasis, and Empower are registered trademarks of Waters Corporation.  
All other trademarks are the property of their respective owners.

©2015 Waters Corporation. Produced in the U.S.A.  April 2015  720005380EN  AG-PDF

The most strongly retained species, represented by the native mAb sample, 
corresponds to the doubly (fully) glycosylated form of the intact mAb. The 
partially deglycosylated mAb sample meanwhile yielded several additional peaks 
with lower HILIC retention, two of which with corresponding detected molecular 
weights that are indicative of once deglycosylated and fully deglycosylated mAb 
species and a third with a corresponding detected molecular weight consistent 
with PNGase F. In contrast, the completely deglycosylated mAb sample presented 
a homogenous fluorescence profile along with an observed molecular weight 
for the mAb that is in agreement with the predicted molecular weight of the 
deglycosylated mAb (145.3 kDa). It is worth noting that when attempting to use 
the BEH Amide, 130Å, 1.7 µm stationary phase, none of the above peaks could 
be resolved (Figures 9D and 9E). So indeed, the widepore phase facilitates the 
development of previously unobtainable separations.

In our hands, the above assay has been used to develop rapid enzymatic 
deglycosylation protocols.3 However, it is natural to suggest that these same 
methods could be applied to measure the glycan occupancy of an intact 
therapeutic mAb, in which case the relative abundance of aglycosylated forms 
(-2 and -1 N-glycans) could potentially be monitored by fluorescence and 
corroborated by LC-MS.

C O N C LU S IO NS
HILIC of small molecules has garnered wide-spread attention and use. In 
contrast, the application of the technique to large biomolecule separations has 
been limited. With the development of the above mentioned amide-bonded, 
wide-pore HILIC stationary phase and corresponding methods, it is now 
possible to resolve the glycoforms of intact glycosylated proteins, as has 
been exemplified by the resolution of the heterogenous glycoforms on intact 
trastuzumab. Alternatively, the described techniques can be applied to studies 
of glycan occupancy. Just as reversed phase separations are employed for 
resolving protein isoforms that have varying hydrophobicities, HILIC separations 
with BEH Amide 300Å can be explored for resolving protein isoforms that 
exhibit varying hydrophilicities, such as isoforms differing with respect to 
glycan occupancy. With the availability of these new separation capabilities, 
it will be possible to perform more detailed characterization of intact 
glycoproteins, whether by means of combining HILIC with optical detection or 
with ESI-MS. 

References
1.  Ahn, J.; Yu, Y. Q.; Gilar, M., UPLC-FLR Method Development 

of 2-AB Labeled Glycan Separation in Hydrophilic 
Interaction Chromatography (HILIC). Waters Appication Note 
720003238EN 2010.

2.  Ahn, J.; Bones, J.; Yu, Y. Q.; Rudd, P. M.; Gilar, M., Separation 

of 2-aminobenzamide labeled glycans using hydrophilic 
interaction chromatography columns packed with 1.7 microm 
sorbent. J Chromatogr B Analyt Technol Biomed Life Sci 2010, 
878 (3–4), 403–8.

3.  Lauber, M. A.; Brousmiche, D. W.; Hua, Z.; Koza, S. M.; Guthrie, 

E.; Magnelli, P.; Taron, C. H.; Fountain, K. J., Rapid Preparation 
of Released N-Glycans for HILIC Analysis Using a Novel 
Fluorescence and MS-Active Labeling Reagent. Waters 
Application Note 720005275EN 2015.

4.  Wang, B.; Tsybovsky, Y.; Palczewski, K.; Chance, M. R., 

Reliable Determination of Site-Specific In Vivo Protein 
N-Glycosylation Based on Collision-Induced MS/MS and 
Chromatographic Retention Time. J Am Soc Mass Spectrom 
2014, 25 (5), 729–41.

5.  Shah, B.; Jiang, X. G.; Chen, L.; Zhang, Z., LC-MS/MS peptide 

mapping with automated data processing for routine profiling 
of N-glycans in immunoglobulins. J Am Soc Mass Spectrom 
2014, 25 (6), 999–1011.

6.  Houel, S.; Hilliard, M.; Yu, Y. Q.; McLoughlin, N.; Martin, S. M.; 

Rudd, P. M.; Williams, J. P.; Chen, W., N- and O-glycosylation 
analysis of etanercept using liquid chromatography and 
quadrupole time-of-flight mass spectrometry equipped with 
electron-transfer dissociation functionality. Anal Chem 
2014, 86 (1), 576–84.

7.  Gilar, M.; Yu, Y. Q.; Ahn, J.; Xie, H.; Han, H.; Ying, W.; Qian, 

X., Characterization of glycoprotein digests with hydrophilic 
interaction chromatography and mass spectrometry.  
Anal Biochem 2011, 417 (1), 80–8.

8.  O’Gara, J. E.; Wyndham, K. D., Porous Hybrid 

Organic-Inorganic Particles in Reversed-Phase Liquid 
Chromatography.  
J Liq Chromatogr Relat Technol. 2006, 29, 1025–1045.

9.  Alpert, A. J., Hydrophilic-interaction chromatography for 

the separation of peptides, nucleic acids and other polar 
compounds. J Chromatogr 1990, 499, 177–96.

10. Xie, H.; Chakraborty, A.; Ahn, J.; Yu, Y. Q.; Dakshinamoorthy, 

D. P.; Gilar, M.; Chen, W.; Skilton, S. J.; Mazzeo, J. R., Rapid 
comparison of a candidate biosimilar to an innovator 
monoclonal antibody with advanced liquid chromatography 
and mass spectrometry technologies. MAbs 2010, 2 (4).

11. Eschelbach, J. W.; Jorgenson, J. W., Improved protein 

recovery in reversed-phase liquid chromatography by the use 
of ultrahigh pressures. Anal Chem 2006, 78 (5), 1697–706.

12. Neue, U. D.; Hudalla, C. J.; Iraneta, P. C., Influence of 

pressure on the retention of sugars in hydrophilic interaction 
chromatography. J Sep Sci 2010, 33 (6–7), 838-40.

[ 22 ] Developing High Resolution HILIC Separations of Intact Glycosylated Proteins Using a Wide-Pore Amide-Bonded Stationary Phase

Authors: Matthew A. Lauber and Stephan M. Koza

Figure 1. HILIC analysis of intact trastuzumab. (A) HILIC fluorescence chromatograms for native 
and partially deglycosylated trastuzumab. (B) Raw and deconvoluted ESI mass spectra for 
the major species resolved in the native trastuzumab sample. (C) Raw and deconvoluted ESI 
mass spectra for the low abundance species resolved in the native trastuzumab sample. HILIC 
was performed on 1 µL aqueous injections of 5 mg/mL trastuzumab using an ACQUITY UPLC 
Glycoprotein BEH Amide 300Å 1.7 µm Column heated to 80˚C. Sample was separated at a flow 
rate of 0.4 mL/min with aqueous (A) and acetonitrile (B) mobile phases containing 0.1% (v/v) TFA 
and 0.3% (v/v) HFIP. A linear gradient was applied as follows:  hold at 20% A for 0.5 min, 20 to 
25% in 0.5 min, 25% to 40% A in 9 min, 40% to 100% A in 0.5 min, hold at 100% A for 0.5 
min, 100% to 20% A in 0.5 min, and hold at 20% A for 3.5 min for re-equilibration. Eluting 
proteins were detected by intrinsic fluorescence (Ex 280 nm/Em 320 nm).     

Species 

% 

2 N-Glycans 

68.9 

1 N-Glycan 

26.6 

0 N-Glycans 

4.5 

Species 

% 

2 N-Glycans 

75.4 

1 N-Glycan 

21.5 

0 N-Glycans 

3.1 

Species 

% 

2 N-Glycans 

99.0 

1 N-Glycan 

1.0 

0 N-Glycans 

0.0 

Trastuzumab  

Native 

5.0 

10.0 min 

2 N-Glycans 

1 N-Glycan 

2 N-Glycans 

1 N-Glycan 

0 N-Glycans 

2 N-Glycans 

1 N-Glycan 

0 N-Glycans 

Trastuzumab 

PNGase F  

3 min 23˚C 

Trastuzumab 

PNGase F  

3 min 23˚C 

1 min 50˚C 

A 

C 

B 

Trastuzumab (2 N-Glycans) 

Trastuzumab (1 N-Glycan) 

FA2/FA2 

148,057 Da 

FA2/FA2G1 

148221 Da 

FA2 

146,614 Da 

FA2G1 

146,773 Da 

1,000 

5,000  

m/z 

1,000 

5,000  

m/z 

145  

149  

kDa 

145  

149  

kDa 

Intrinsic Fluorescence Detection 

G OA L
To demonstrate the use of HILIC with an 
ACQUITY UPLC® Glycoprotein BEH Amide 
300Å Column to assay the glycan occupancy 
of intact mAbs.

BAC KG RO U N D
Monoclonal antibodies (mAbs) have emerged 
as some of the most important therapeutics on 
the market. These mAbs that are prescribed 
for therapeutic use are most often expressed 
from eukaryotic cell lines, such as CHO, and, as 
a result, are N-glycosylated at two consensus 
site asparagine residues in the Fc portion of 
their heavy chains. Since glycosylation can be a 
measure of efficacy, safety and manufacturing 
conditions, it is often critical to characterize 
and routinely monitor the N-glycan profile of a 
mAb drug substance.1-2 The recent introduction 
of the GlycoWorks™ RapiFluor-MS™ N-Glycan 
Kit has made it significantly easier to perform 
highly detailed HILIC-based, released 
N-glycan analyses and to thereby elucidate the 
heterogeneity of N-glycosylation.3-4 However, 
it is also critical to determine the  extent to 
which the asparagine linkage sites are occupied 
with N-glycans, particularly since incomplete 
glycosylation dramatically changes the 
effector functions of a mAb.5 Traditionally, 
this assessment has been performed by a 
sized-based Capillary Electrophoresis-Sodium 
Dodecyl Sulfate (CE-SDS) separation of the 
heavy chains resulting from reduction of a 
mAb.6,7 To instead directly assess glycan 

Unprecedented Hydrophilic Interaction 
Chromatography (HILIC) separations of intact mAb 
glycan occupancy variants using the ACQUITY UPLC 
Glycoprotein BEH Amide 300A 1.7µm Column.

Measuring the Glycan Occupancy  

of Intact mAbs using HILIC and Detection  

by Intrinsic Fluorescence  

[ 23 ]

A Single Column Chemistry for 

Multiple Glycoprotein Analyses

The A

CQUIT

Y UPLC

®

 Glycopr

ote

in B

EH Am

ide, 300Å, 1.7 µm C

olum

n offers: 

■ 

Optim

ized wid

e-pore, HI

LIC sta

tiona

ry p

hase fo

r resolv

ing g

lycoforms fr

om

inta

ct or diges

ted glyco

prote

ins

■ 

Gener

atio

n of dom

ain s

pecific g

lycan l

inka

ges wit

h or w

ithou

t MS

■ 

Elucidatio

n of site s

pecific g

lycan occu

pancy

■ 

High reso

lutio

n glyco

peptid

e mappi

ng wit

hout l

imita

tions due to p

eptide/

glyc

an 

size o

r com

positio

n

■ 

Impr

oved reso

lutio

n in sepa

ratio

ns of large, re

leased 

N-g

lycans (

EPO, F

acto

r IX

)

■ 

QC tes

ted wit

h Water

s Glycopr

ote

in P

erform

ance T

est S

tanda

rd

©2015 W

ater

s Cor

pora

tion. W

ater

s, The S

cience of W

hat’s P

ossib

le,  

and A

CQUIT

Y UPLC a

re registered tr

adem

arks of W

ater

s Cor

pora

tion.

wat

ers

.co

m/gl

yco

pro

tein

NEW

A Single Column Chemistry for 

Multiple Glycoprotein Analyses

he ACQU

ITY U

PLC

®

 Glycopr

ote

in B

EH Am

ide, 300Å, 1.7 µm C

olum

n offers: 

■ 

Optim

ized wid

e-pore, HI

LIC sta

tiona

ry p

hase fo

r resolv

ing g

lycoforms fr

om

inta

ct or diges

ted glyco

prote

ins

■ 

Gener

atio

n of dom

ain s

pecific g

lycan l

inka

ges wit

h or w

ithou

t MS

■ 

Elucidatio

n of site s

pecific g

lycan occu

pancy

■ 

High reso

lutio

n glyco

peptid

e mappi

ng wit

hout l

imita

tions due to p

eptide/

glyc

an 

size o

r com

positio

n

■ 

Impr

oved reso

lutio

n in sepa

ratio

ns of large, re

leased 

N-g

lycans (

EPO, F

acto

r IX

)

■ 

QC tes

ted wit

h Water

s Glycopr

ote

in P

erform

ance T

est S

tanda

rd

©2015 W

ater

s Cor

pora

tion. W

ater

s, The S

cience of W

hat’s P

ossib

le,  

and A

CQUIT

Y UPLC a

re registered tr

adem

arks of W

ater

s Cor

pora

tion.

wat

ers

.co

m/gl

yco

pro

tein

Figure 2. HILIC analysis of four intact mAbs. Bevacizumab, NIST IgG1κ candidate reference 
material, and trastuzumab were injected without any sample preparation except for dilution to  
5 mg/mL from their respective formulations. Prior to HILIC analysis, Intact mAb Mass Check 
Standard (1 mg) was reconstituted in 500 µL of 6M guanidine HCl, filtered  via 3 passes with a 
100KDa MWCO polyethersulfone membrane (GE Healthcare Life Sciences, Vivaspin 500), and 
concentrated to 5 mg/mL.   

Bevacizumab 

IgG1   

Reference 

Material 

Trastuzumab 

5.0 

10.0   5.0 

10.0 min 

Intrinsic Fluorescence Detection 

1 N-Glycan 

4.9% 

1 N-Glycan 

0.7% 

1 N-Glycan 

1.3% 

1 N-Glycan 

1.0% 

Murine IgG1

(Intact mAb

Mass Check Std)

occupancy variants for intact mAbs, we have 
developed an LC separation based on HILIC. To 
achieve these unprecedented separations, a new, 
purposefully designed HILIC column was employed. 
This new column, the ACQUITY UPLC Glycoprotein 
BEH Amide 300Å 1.7 µm Column, contains a wide-
pore amide bonded stationary phase that ensures 
that mAb species have access to the porous network 
of the stationary phase and are not significantly 
impaired by restricted diffusion.8

T H E  S O LU T IO N
A high-throughput, high resolution HILIC separation 
was established for intact mAbs using a 2.1 x 
150 mm wide-pore BEH amide column, a 0.4 
mL/min flow rate, and a column temperature of 
80 ˚C. In addition, two mobile phase additives, 
0.1% trifluoroacetic acid (TFA) and 0.3% 
hexafluoroisopropanol (HFIP) were employed to 
improve the solubility of intact mAbs in the high 
organic, initial mobile phase conditions of the 
HILIC gradient. To enhance the sensitivity of this 
LC method, the intact proteins were detected by 
means of their intrinsic fluorescence.  Excitation 
and emission wavelengths of 280 and 320 nm 
were found to provide optimal signal-to-noise 
and consistently flat chromatographic baselines 
that are desirable for peak integration.  A 
representative set of chromatograms resulting 
from this 15 minute LC method is shown in Figure 
1A. Three chromatograms are displayed. The top 
chromatogram shows trastuzumab as injected 
from a dilution of its formulation, while the other 
two chromatograms show samples of trastuzumab 
after being subjected to partial PNGase F 
deglycosylation.  Deglycosylated samples of 
trastuzumab clearly showed three distinct peaks in 
their HILIC profiles, as was predicted if the glycan 
occupancy variants of a mAb were to be resolved. 
The unadulterated sample of trastuzumab contained 
measurable levels of only the fully occupied and 
singly occupied forms (1%) as confirmed by online 
mass analysis. The deconvoluted mass spectrum 
corresponding to the main LC peak exhibited several 
masses, such as 148,057 Da and 148,221 Da, 

that are within 2 Da of the theoretical molecular weights for the predominant 
glycoforms of trastuzumab (Figure 1B).9 In contrast, the deconvoluted 
mass spectrum for the smaller, less strongly retained peak showed reduced 
heterogeneity and masses that were lighter by approximately 1000 to 2000 
Da, consistent with the single occupancy form (Figure 2B). Fully aglycosylated 
species were not detected in this sample. An interesting observation from these 
data is that the levels of fully deglycosylated forms are lower than would be 
predicted if both sites were deglycosylated at the same rate. In these examples, 
the observed level of fully deglycosylated forms is approximately one-third 
lower than would be predicted. This may indicate that either the digestion rate 
for one of the N-glycans is slower than the other, or that upon removal of the 
first N-glycan, the digestion rate of the remaining N-glycan is reduced.

To assess the applicability of this technique to other mAbs, we analyzed three 
additional IgG samples.  Results on trastuzumab, bevacizumab, a candidate 
IgG1

κ reference material, and a murine IgG1 (Intact mAb Mass Check Standard, 

p/n 186006552) are shown in Figure 2. Integrations on the HILIC-fluorescence 
chromatograms indicated that these samples contain 1 N-glycan (incomplete 
glycosylation) variants at relative abundances ranging from 0.7 to 4.9%. These 
observations suggest that this HILIC method could be an attractive technique for 
assaying incomplete glycosylation of mAbs down to levels well below 1% for  
the 1 N-glycan form.  Detection limits for the 0 N-glycan form may perhaps be 
even lower. Of these samples, bevacizumab might be predicted to have the  
highest abundance of the 0 N-glycan form. Indeed, the bevacizumab profile 
presented a peak, albeit very minor, with a retention time consistent with  
a 0 N-glycan species, the area of which would contribute to the overall profile 

[ 24 ] Measuring the Glycan Occupancy of Intact mAbs Using HILIC and Detection by Intrinsic Fluorescence

Waters Corporation 
34 Maple Street 
Milford, MA 01757 U.S.A. 
T: 1 508 478 2000 
F: 1 508 872 1990 
www.waters.com

A Single Column Chemistry for 

Multiple Glycoprotein Analyses

The A

CQUIT

Y UPLC

®

 Glycopr

ote

in B

EH Am

ide, 300Å, 1.7 µm C

olum

n offers: 

■ 

Optim

ized wid

e-pore, HI

LIC sta

tiona

ry p

hase fo

r resolv

ing g

lycoforms fr

om

inta

ct or diges

ted glyco

prote

ins

■ 

Gener

atio

n of dom

ain s

pecific g

lycan l

inka

ges wit

h or w

ithou

t MS

■ 

Elucidatio

n of site s

pecific g

lycan occu

pancy

■ 

High reso

lutio

n glyco

peptid

e mappi

ng wit

hout l

imita

tions due to p

eptide/

glyc

an 

size o

r com

positio

n

■ 

Impr

oved reso

lutio

n in sepa

ratio

ns of large, re

leased 

N-g

lycans (

EPO, F

acto

r IX

)

■ 

QC tes

ted wit

h Water

s Glycopr

ote

in P

erform

ance T

est S

tanda

rd

©2015 W

ater

s Cor

pora

tion. W

ater

s, The S

cience of W

hat’s P

ossib

le,  

and A

CQUIT

Y UPLC a

re registered tr

adem

arks of W

ater

s Cor

pora

tion.

wat

ers

.co

m/gl

yco

pro

tein

NEW

at a level of only 0.05%. Such an observation 
suggests that the 0 N-glycan level is lower 
than would be statistically predicted; however, 
rigorous determination of the quantitative limits 
of this analysis and the identity of this putative 
0 N-glycan peak would need to be evaluated to 
confirm that this is indeed the case.

This strategy for measuring mAb glycan occupancy 
is most appealing in that it requires minimal, 
if any, sample preparation. We have observed 
that some interferences can be encountered that 
are due to partially fragmented and/or reduced 
mAb species.  In which case, as with the Intact 
mAb Mass Check Standard, a simple, centrifugal 
filtration clean-up step with a 100kDa MWCO 
polyethersulfone membrane was sufficient to 
minimize such interferences. Given the unique 
selectivity of the HILIC separation, it should also 
be possible to address potential interferences 
by performing offline or online 2D-LC, wherein 
a size exclusion or reversed phase separation 
could be coupled to the wide-pore amide HILIC 
separation. Future investigations could also 
include reducing glycan heterogeneity (via a 
sialidase or other exoglycosidase) to allow the mAb 
glycan occupancy variants to be more discretely 
resolved. Similarly, there is an opportunity to use 
this separation in combination with enzymes that 
generate Fc subunits.

Waters, The Science of What’s Possible, and ACQUITY UPLC are registered trademarks of Waters Corporation. GlycoWorks and  
RapiFluor-MS are trademarks of Waters Corporation. All other trademarks are the property of their respective owners.

©2015 Waters Corporation. Produced in the U.S.A.  June 2015  720005435EN  AO-PDF

S UMM A RY
In addition to profiling the heterogeneity of glycosylation, it is also critical 
to assay glycan occupancy. Here, we demonstrate that a Glycoprotein BEH 
Amide column, purposefully designed for large molecule HILIC separations, 
can be used to directly quantify incomplete N-glycan occupancy in intact 
mAb samples. Unlike a conventional CE-SDS separation of reduced mAbs, this 
technique provides a non-inferred assessment on the nature (i.e. 1 N-glycan 
versus 0 N-glycans) of glycan occupancy for the intact mAb. The proposed HILIC 
methodology is also MS-compatible, making it possible to readily confirm the 
assignments of observed peaks. 

References

1.  Beck, A.; Wagner-Rousset, E.; Ayoub, D.; Van Dorsselaer, A.; Sanglier-Cianferani, S., 

Characterization of therapeutic antibodies and related products. Anal Chem 2013, 85 (2), 
715–36.

2.  Dalziel, M.; Crispin, M.; Scanlan, C. N.; Zitzmann, N.; Dwek, R. A., Emerging principles for the 

therapeutic exploitation of glycosylation. Science 2014, 343 (6166), 1235681.

3.  Lauber, M. A.; Yu, Y. Q.; Brousmiche, D. W.; Hua, Z.; Koza, S. M.; Magnelli, P.; Guthrie, E.; Taron, 

C. H.; Fountain, K. J., Rapid Preparation of Released N-Glycans for HILIC Analysis Using a 
Labeling Reagent that Facilitates Sensitive Fluorescence and ESI-MS Detection. Anal Chem 
2015, DOI: 10.1021/acs.analchem.5b00758.

4.  Lauber, M. A.; Koza, S. M., Enhancing the Peak Capacity of High Molecular Weight N-Glycan 

HILIC Separations with a Wide-Pore Amide Bonded Stationary Phase. Waters Tech Brief 
720005381EN 2015.

5.  Beck, A.; Wagner-Rousset, E.; Ayoub, D.; Van Dorsselar, A.; Sanglier-Cianferani, S., 

Characterization of Therapeutic Antibodies and Related Products. Analytical Chemistry 2013,  
(85), 715–736.

6.  Chen, G.; Ha, S.; Rustandi, R. R., Characterization of glycoprotein biopharmaceutical products 

by Caliper LC90 CE-SDS gel technology. Methods Mol Biol 2013, 988, 199–209.

7.  Lacher, N. A.; Roberts, R. K.; He, Y.; Cargill, H.; Kearns, K. M.; Holovics, H.; Ruesch, M. N., 

Development, validation, and implementation of capillary gel electrophoresis as a replacement 
for SDS-PAGE for purity analysis of IgG2 mAbs. Journal of Separation Science 2010, 33 (2), 
218–227.

8.  Renkin, E. M., J. Gen. Physio. 1954,  (38), 225.

9.  Damen, C. W.; Chen, W.; Chakraborty, A. B.; van Oosterhout, M.; Mazzeo, J. R.; Gebler, J. C.; 

Schellens, J. H.; Rosing, H.; Beijnen, J. H., Electrospray ionization quadrupole ion-mobility 
time-of-flight mass spectrometry as a tool to distinguish the lot-to-lot heterogeneity in 
N-glycosylation profile of the therapeutic monoclonal antibody trastuzumab. J Am Soc Mass 
Spectrom 2009, 20 (11), 2021-33.

[ 25 ]

Measuring the Glycan Occupancy of Intact mAbs Using HILIC and Detection by Intrinsic Fluorescence 

GOAL
To demonstrate the capabilities of the integrated 
UPLC®-MS analysis of an intact monoclonal anti-
body with a comprehensive platform for accurate 
mass measurement, data processing, and reporting 
with UNIFI® Scientific Information System.  

BACKGROUND
The growing biotherapeutic pipeline 
means that the efficient characterization 
of monoclonal antibodies (mAb) is of 
growing importance, both to regulatory 
authorities and to pharmaceutical 
companies. Being able to perform 
acquisition and processing within  
the same platform, complete with an  
audit trail, is an important goal for 
regulated environments.

Accurately identifying post-translational 
modifications such as protein 
glycosylation is required as part  
of guidelines as they play several key 
roles in biological systems. Fast and 
accurate analysis of the glycoproteins  
is required in order to ensure the safety 
and efficacy of the biotherapeutic.

The ACQUITY UPLC® H-Class Bio’s  
high-resolution bioseparations combined  
with high mass accuracy mass spectrometry  
detection with the Xevo® G2 Tof provides  
routine UPLC-MS applications for 
biopharmaceutical laboratories. 

This UNIFI-based platform addresses previous 
limitations with a comprehensively integrated platform 
for data acquisition by chromatography and mass 
spectrometry, with automated reporting.

Intact protein characterization:
Waters Biopharmaceutical 
System Solution with UNIFI

Figure 1. UNIFI’s intact protein analysis report 
of an intact mAb, Trastuzumab. The report 
shows, from top to bottom, the integrated 
chromatographic peak, the charge state 
distribution, the continuum deconvoluted data, 
the integrated and processed data, and finally 
the summary table assigning major glycoforms 
according to mass.

[ 26 ]

Waters Corporation 
34 Maple Street 
Milford, MA 01757 U.S.A. 
T: 1 508 478 2000 
F: 1 508 872 1990 
www.waters.com

Waters, The Science of What’s Possible, ACQUITY UPLC, Xevo, UNIFI, and UPLC are registered trademarks of 
Waters Corporation. All other trademarks are the property of their respective owners.

©2011–2016 Waters Corporation. Produced in the U.S.A.  January 2016  720003843EN  LM-PDF

The intact protein analysis report demonstrates the report objects, which can be 
entirely configured by the user: TIC summarized chromatogram; raw, deconvoluted, 
and centroid mass spectra; and tabulated summary of the interpreted LC-MS data 
(Figure 1). This detailed view shows an example of a deconvoluted spectrum within  
a specified mass range and parameter settings defined in the method. Deconvolution 
reveals several core glycosylated species which match the number of glucose 
residues and level of fucosylation. Another report object is a table with mass 
measurement of the intact mAb and accurately assigned mAb glycan variants 
(Figure 2). Mass errors were reported for each Trastuzumab MS peak with  
a corresponding retention time entry from the TIC chromatogram. 

Such an integrated LC-MS approach provides the user flexibility to work with both 
raw and processed data followed by quick and efficient data management. 

SUMMARY 
The capabilities of the Biopharmaceutical Platform Solution with UNIFI have been 
successfully demonstrated with the example of an intact biotherapeutic mAb.

Modern instrumentation and evolving analytical techniques extend the limits  
of the biopharmaceutical industry and consequently impose strict control of  
manufacturing processes. 

Highly efficient and cost-effective integrated UPLC-MS approaches with the UNIFI 
Scientific Information System for data processing and reporting satisfies regulatory 
requirements and facilitates intact protein characterization. This technology covers 
the range from detailed structural protein characterization to sophisticated data 
management with UPLC-MS platforms.

There is a large set of data generated during each 
mAb analysis requiring interpretation of a variety of 
glycosylated forms and comprehensive characterization 
of the final product. This step sets productivity limits 
to otherwise high-throughput procedures and hinders 
automation of the process.  

The UNIFI-based platform addresses these limitations 
with a comprehensively integrated platform for data 
acquisition by chromatography and mass spectrometry, 
with automated reporting. 

THE SOLUTION 
To solve the problem of time-consuming data analysis 
and facilitate data processing of therapeutic mAb,  
the Biopharmaceutical Platform Solution with UNIFI 
was configured for the study of intact proteins.  
This represents a holistic approach of UPLC-MS data 
acquisition followed by automatic processing and 
annotation of the data in a high-throughput manner, 
which are further exported for data management. 

UPLC-MS analysis of the mAb Trastuzumab was 
performed automatically. Aqueous solutions of 0.1% 
FA and 0.1% FA solution in acetonitrile were used as 
eluents A and B, respectively. Column temperature 
set to 80 °C is critical for successful chromatographic 
separation. The system included an ACQUITY UPLC 
H-Class Bio, an ACQUITY UPLC Protein BEH C4 
Column, and a Xevo G2 Tof. The UNIFI Scientific 
Information System for acquisition, data processing, 
and reporting completes this comprehensive 
Biopharmaceutical Platform Solution.  

Figure 2. Zoomed in view of the table in Figure 1 summarizing mass measurement 
of the intact mAb and accurately assigned mAb glycan variants.

[ 27 ]

Intact Protein Characterization: Waters Biopharmaceutical System Solution with UNIFI

WAT E RS SO LU T IONS
Biopharmaceutical Platform Solution 
with UNIFI

ACQUITY UPLC® H-Class Bio System

Xevo® G2 QTof

ACQUITY UPLC Protein BEH C4 Column

Xevo G2-S QTof

ACQUITY UPLC Tunable UV Detector

K E Y WO R DS
Intact mass analysis, mAb,  
biotherapeutic characterization

A P P LIC AT ION BEN E FIT S 
The Biopharmaceutical Platform Solution  
with UNIFI® enables a fully integrated workflow 
for intact mass analysis, including acquisition, 
processing, and reporting, for organizations  
in early development and those operating  
under regulatory compliant environments.  
The ability to automate and standardize  
intact mass analysis enables laboratories  
to deploy their scientific resources with  
greater efficiency and effect. 

INT RO DUC T ION
Intact mass analysis is a rapid and convenient method for confirming protein 
identity and profiling product-related variants. In conjunction with other 
analytical techniques, such as peptide mapping and released glycan analysis, 
intact mass analysis can help determine if the biomolecule had been correctly 
cloned, expressed, purified, and formulated during the biopharmaceutical drug 
development process. 

Intact mass analysis can provide a semi-quantitative view of product 
heterogeneity and is often employed to determine relative composition of product 
glycoforms. As a lot release test, intact protein mass analysis often provides a 
quick identity test using the mass of a major variant, sometimes in conjunction 
with a purity test with defined product variation for peaks corresponding 
to variants displaying critical product attributes. Demonstration of process 
consistency through such comparability exercises is critical to obtain initial 
regulatory approval and for later process improvement studies. 

Data processing and report generation often become productivity-limiting 
tasks for organizations responsible for biotherapeutic protein characterization 
and analysis. It is still common for LC-MS intact protein data to be manually 
processed, an inefficient process that lacks standardization and is prone to human 
error. Further inefficiency and sources of error result from scientists having to 
reformat results into graphical and tabular formats suitable for communicating 
information to their organizations. 

The ability to automate and standardize data acquisition, processing, and reporting 
for intact mass analysis allows laboratories to deploy their scientific resources 
with greater efficiency and effect. The Waters® UNIFI Scientific Information 
System enables these benefits, as well as regulatory compliance, to be realized 
throughout discovery, development, and quality management organizations. 

In this application note, an integrated and compliant-ready solution for intact 
mass analysis is described. The combination of UPLC® separations, optimized 
application-tested protein column chemistries, the Xevo G2-S QTof for mass 
detection, all used under control of the UNIFI Scientific Information System, 
achieves the goal of total workflow automation and standardization. 

Streamlining Compliant and Non-Compliant Intact Mass Analysis of 

Biotherapeutic mAbs with the Biopharmaceutical Platform Solution with UNIFI 
Henry Shion and Weibin Chen

Waters Corporation, Milford, MA, USA

[ 28 ]

E X P E RIM ENTA L 

LC conditions
System: 

ACQUITY UPLC H-Class Bio System

Detector: 

ACQUITY UPLC Tunable UV Detector

Column:  

ACQUITY UPLC Protein BEH C4 Column, 
300Å, 1.7 µm, 2.1 mm X 50 mm  
(p/n 186004495)

Column temp.:  

80 °C

Mobile phase A:  

Water

Mobile phase B:  

Acetonitrile

Mobile phase C:  

1% formic acid 

Optical detection:  

UV 280 nm

LC gradient table:

  Time  

Flow  

 (min) (mL/min)  %A 

%B 

%C 

%D  Curve

 Initial 

0.40 

85.0 

5.0  10.0 

0  Initial 

 1.00 

0.40  85.0 

5.0  10.0  0 

6 

 1.01 

0.20  85.0 

5.0  10.0  0 

6 

 3.50 

0.20 

5.0 

95.0  0.0 

0 

6 

 3.70 

0.40 

5.0 

95.0  0.0 

0 

6 

 4.00 

0.40 

10.0 

80.0  10.0  0 

6 

 4.50 

0.40 

10.0 

80.0  10.0  0 

6 

 5.00 

0.40  85.0 

5.0  10.0  0 

6 

 5.50 

0.40  85.5 

5.0  10.0  0 

6

Total run time: 6.5 min

MS conditions
Mass spectrometer: 

Xevo G2-S QTof

Capillary:  

2.5 kV 

Sampling cone:  

80 V 

Extraction cone:  

4 V   

Source temp.:  

150 °C 

Desolvation temp.:  

350 °C 

Cone gas flow:  

0 L/h 

Desolvation gas flow:   800 L/h

Informatics
UNIFI Scientific Information System

Results derived from an intact IgG1 mAb mass analysis are 
used to illustrate how this integrated system solution can help 
the biopharmaceutical laboratories to streamline a common 
analytical workflow, shown in Figure 1, and more quickly and 
efficiently communicate key information needed to bring better 
molecules to market faster.

Biopharmaceutical System Solution with UNIFI

■

■

ACQUITY UPLC H-Class Bio System

■

■

Protein Separation Technology (PrST) Columns 

■

■

Xevo G2-S QTof with an ACQUITY UPLC Tunable UV Detector

■

■

UNIFI Scientific Information System

Sample preparation
Waters Intact mAb Mass Check Standard (p/n 186006552) was 
analyzed by solubilizing the standard (10 mg/mL or 67 μM, 
100 µL DI water to standard vial, 5 min sonication), and diluting 
20X (Final 3.3 μM, 0.50 μg/μL) with eluent A for Xevo G2 QTof 
analysis or 200X (0.33 μM, 0.05 μg/μL) for Xevo G2-S analysis.

Analysis Method
A holistic UNIFI method contains information 

sufficient for LC-MS acquisition, data processing, 

reporting, and report sign-off.

Analysis
Data is acquired once the sample list reaches 

the top of the sample queue, is subjected 

to automatic post-acquisition processing, and 

the assigned reporting templates are executed.    

Report
Standard report elements have been optimized 

for displaying antibody results. Custom calculations 

and filters efficiently summarize overall findings.   

Figure 1. Intact mass analysis workflow with the UNIFI Scientific  
Information System.

[ 29 ]

Streamlining Compliant and Non-Compliant Intact Mass Analysis of Biotherapeutic mAbs 

with the Biopharmaceutical Platform Solution with UNIFI 

R E SU LT S AND DIS CUSSION
An automated mAb LC-MS analysis set of 11 injections was automatically acquired, processed, and reported as 
specified in a single UNIFI method. Data are representative of a simple method development set, where the goal 
of the researcher is to assess the extent of product glycovariation and determine analytical reproducibility.

For the processed results, a single injection is represented in the review panel of the UNIFI analysis center, 
shown in Figure 2. This panel is configured to convey chromatographic information (integrated total ion 
chromatogram), the MaxEnt™ deconvoluted MS spectrum corresponding to the summed spectra under the 
detected peak, and a component summary window filtered to display the top five most intensely assigned 
glycoforms (G0F/G0F, G0F/G1F, G1F/G1F, G1F/G2F, or G2F/G2F). 

This combined panel enables a researcher to assess chromatographic quality, the quality of MS data processing, 
and the quality of glycoform assignments in a single display. Closer examination also reveals the relative 
abundance of each glycoform was automatically calculated as part of the processing.

Figure 2. The review panel in UNIFI displays automatically processed experimental results.

[ 30 ]

Streamlining Compliant and Non-Compliant Intact Mass Analysis of Biotherapeutic mAbs  
with the Biopharmaceutical Platform Solution with UNIFI 

Having designated one sample as the reference enables a researcher to select the comparative mode display 
of the review tab. The binary comparison display, shown in Figure 3, provides a means to visually examine 
the differences between the two samples, thus revealing the extent of variation between samples. In this 
display, comparative chromatograms and spectra (A280 and summed m/z spectra) are depicted, along with the 
component summary, now reformatted to address comparative questions. Since both injections were from the 
same sample, minimal experimental result differences are predictably observed.

Figure 3. UNIFI’s review panel (compare mode) displays automatically processed experiment results, with a focus on identifying similarities and differences between  
a reference sample and unknown samples.

[ 31 ]

Streamlining Compliant and Non-Compliant Intact Mass Analysis of Biotherapeutic mAbs 

with the Biopharmaceutical Platform Solution with UNIFI 

The summary plot tool within the review tab enables researchers to quickly compare trends and differences 
within the larger data set. The variation of mAb glycoform MS response, as shown in Figure 4, would be a 
common application of this capability, as would comparisons of observed retention time or mass error across 
the sample set. The consistent MS response of glycoforms across all injections illustrates the expected 
reproducibility of the intact mass analysis of replicated injections.

Figure 4. summary plot view of MS response for 5 mAb glycoforms (n = 11 injections).

UNIFI reporting
The reporting functionality within UNIFI Software is powerful, addressing one of the common bottlenecks 
encountered by organizations when generating and managing large volumes of complex scientific data.  
The ability to customize common report objects by means of filters, formatting, and the use of custom fields 
and calculations enables report content to be automatically generated by an entire organization with high 
quality on a consistent basis. Based on the analytical objectives, one or more report templates can be  
attached to the analysis method.

[ 32 ]

Streamlining Compliant and Non-Compliant Intact Mass Analysis of Biotherapeutic mAbs  
with the Biopharmaceutical Platform Solution with UNIFI 

The first page of a typical intact mass analysis experimental report contains a summary of sample information 
and acquisition status, as shown in Figure 5. More detailed experimental results (such as TUV and TIC 
chromatograms, raw and deconvoluted MS spectra, and identified component response summary table) are 
often grouped for each injection, as shown in Figure 6. 

In the case of mAbs, generic report objects were tuned to account for the rapid desalting LC-MS method that 
was used, the acquisition of UV and MS data, and the typical input m/z and output mass ranges encountered 
during antibody ESI mass analysis.

Figure 5. Typical first page of an intact protein LC-MS report in UNIFI Software that summarizes sample and acquisition details.

[ 33 ]

Streamlining Compliant and Non-Compliant Intact Mass Analysis of Biotherapeutic mAbs 

with the Biopharmaceutical Platform Solution with UNIFI 

Figure 6. Example report object grouping (TUV and TIC chromatograms, raw and deconvoluted MS spectra, and identified component response summary table) from a 
single injection within the analysis. 

[ 34 ]

Streamlining Compliant and Non-Compliant Intact Mass Analysis of Biotherapeutic mAbs  
with the Biopharmaceutical Platform Solution with UNIFI 

Waters Corporation 
34 Maple Street 
Milford, MA 01757 U.S.A. 
T: 1 508 478 2000 
F: 1 508 872 1990 
www.waters.com

Waters, The Science of What’s Possible, ACQUITY UPLC, UPLC, Xevo, and UNIFI are registered trademarks of Waters Corporation.  
MaxEnt is a trademark of Waters Corporation. All other trademarks are the property of their respective owners.

©2013–2016 Waters Corporation. Produced in the U.S.A.  January 2016  720004617EN  AG-PDF

CONC LUSIONS
The intact mass analysis workflows within the Biopharamceutical Platform 
Solution with UNIFI enable automated data acquisition, processing, and reporting 
of a typical method validation sample set. This demonstrates UNIFI Software’s 
ability to facilitate robust glycoform profiling of a recombinant mAb, removes  
the necessity of manual data processing, and improves the process of data review 
and reporting. The implementation of such highly automated workflows should 
enable biotherapeutic development and quality organizations to handle larger 
volumes of sample requests with the same resources, while improving the quality 
of the information they provide.

In addition, the ability to automate reporting summary results across the sample sets eliminates the use of 
external software for data aggregation, as shown in Figure 7. This not only greatly increases the timeliness of 
communicating results, but avoids the human errors and validation efforts that cost analytical organizations 
time and money. In the case of this typical method validation injection set, the precision of MS response and 
mass accuracy is reported for one of the observed glycoforms.

Figure 7. The report object summarizes MS response and mass accuracy/precision across all injections within the sample set in table and bar chart formats.

[ 35 ]

Streamlining Compliant and Non-Compliant Intact Mass Analysis of Biotherapeutic mAbs 

with the Biopharmaceutical Platform Solution with UNIFI 

GLYCOPEPTIDE MAPPING AND SUBUNIT ANALYSIS

[ 36 ]

WAT E R S S O LU T IO NS
ACQUITY UPLC® Glycoprotein BEH 
Amide 300Å Column (patent pending)

Glycoprotein Performance Test Standard

ACQUITY UPLC H-Class Bio System

Waters SYNAPT® G2-S HDMS

K E Y W O R D S
ACQUITY UPLC H-Class Bio System,  
BEH Amide 300Å, glycans,  
glycosylated proteins, glycosylation, 
HILIC, mab, glycopeptide, glycoprotein

A P P L I C AT IO N  B E N E F I T S 

■

■

Orthogonal selectivity to conventional 
reversed phase (RP) peptide mapping 
for enhanced characterization of 
hydrophilic protein modifications, such as 
glycosylation

■

■

Class-leading HILIC separations  
of IgG glycopeptides to interrogate  
sites of modification

■

■

MS compatible HILIC to enable detailed 
investigations of sample constituents

■

■

Enhanced glycan information that 
complements RapiFluor-MS released 
N-glycan analyses

■

■

Glycoprotein BEH Amide 300Å  
1.7 µm stationary phase is QC tested  
via a glycoprotein separation to ensure  
consistent batch to batch reproducibility

I N T RO DU C T IO N
Peptide mapping of biopharmaceuticals has longed been used as a tool 
for identity tests and for monitoring residue-specific modifications.1-2 In a 
traditional analysis, peptides resulting from the use of high fidelity proteases, 
like trypsin and Lys-C, are separated with very high peak capacities by reversed 
phase (RP) separations with C18 bonded stationary phases using ion-pairing 
reagents. Separations such as these are able to resolve peptides with single 
amino acid differences such as asparagine; and the two potential products of 
asparagine deamidation, aspartic acid and isoaspartic acid.3-4 

Nevertheless, not all protein modifications are so easily resolved by RP 
separations. Glycosylated peptides, in comparison, are often separated with 
relatively poor selectivity, particularly if one considers that glycopeptide 
isoforms usually differ in their glycan mass by about 10 to 2,000 Da. So, 
while RP separations are advantageous for generic peptide mapping, they are 
limited in their ability to resolve hydrophilic modifications. Previous studies 
have demonstrated that hydrophilic interaction chromatography (HILIC) 
with an amide-bonded stationary phase can provide complementary and 
highly resolving separations of glycosylated peptides.5-6 These studies have 
demonstrated that amide-bonded stationary phases are particularly effective 
for these separations, because they afford high retentivity as a consequence of 
their hydrophilicity and propensity for hydrogen bonding.7 

Expanding upon this technology, we have developed an amide-bonded 
stationary phase with a nominally larger pore diameter, a so-called “wide-
pore” material, such that amide HILIC separations can be universally applied 
to separating the glycoforms of both intact and digested glycoproteins. This 
stationary phase found in ACQUITY UPLC Glycoprotein BEH Amide 300Å 1.7 
μm Columns ensures that glycopeptides, regardless of their size, will have 
access to the majority of the porous network and be less prone to restricted 
diffusion.8-9 In previous work, we have demonstrated the use of this HILIC column 
to assay the glycan occupancy of an intact monoclonal antibody (mAb),10 to 
map the domain-specific glycosylation of IgG subunits,11 and to improve the 
resolution of tri- and tetra-antennary GlycoWorks,™ RapiFluor-MS™ labeled 
N-glycans.12 Here, we explore the use of the Glycoprotein BEH Amide 300Å 1.7 
μm Column to produce high resolution HILIC separations of glycopeptides from 
three different monoclonal antibodies: trastuzumab, cetuximab and an IgG1K 
candidate reference material from NIST.

HILIC Glycopeptide Mapping with a Wide-Pore Amide Stationary Phase 
Matthew A. Lauber and Stephan M. Koza

Waters Corporation, Milford, MA, USA

[ 37 ]

E X P E R IM E N TA L 

Sample description

Lys-C digest of trastuzumab and NIST candidate  
reference material 
An adaptation of a previously published single reaction vial, 
overnight (16+ hours) procedure4 was employed to prepare  
non-reduced Lys-C digests of trastuzumab and a IgG1K monoclonal 
antibody candidate reference material obtained from NIST 
(#8670, lot# 3F1b). TFA quenched digests were stored at -80 °C 
until analyzed. In preparation for HILIC chromatography, aqueous 
digests were diluted with 4 parts acetonitrile and 0.1 parts 
dimethylsulfoxide and were then centrifuged at 16 x 1000 g for  
10 minutes to remove any insoluble composition. Supernatant  
from the centrifuged digest was thereafter injected.

Lys-C/tryptic digest of cetuximab
Reduced and alkylated cetuximab was digested with a combination 
of Achromobacter protease I (Lys-C) and trypsin. Formulated 
cetuximab was concentrated to 10 mg/mL and buffer exchanged 
with a 10 kDa MWCO centrifugal filter (Millipore, Billerica, MA) into 
a solution of 6 M GuHCl, 50 mM DTT, and 0.2 M phosphate (pH 8.1), 
then incubated at 37 °C for 2 hours. Thereafter, the sample was 
diluted with a solution of iodoacetamide, bringing the antibody 
concentration to 8 mg/mL and the buffer composition to 4.8 M 
GuHCl, 40 mM DTT, 50 mM iodoacetamide, and 0.17 M phosphate 
(pH 8.1). Alkylation with iodoacetamide was allowed to proceed 
under these conditions for 10 min in the dark at 37 °C, before  
being quenched by the addition of cysteine, diluted with a urea- 
containing buffer, and mixed with Achromobacter protease I (Lys-C) 
at a 4:1 w/w ratio. The resulting digest solution of 0.8 mg/mL 
cetuximab, 0.5 M GuHCl, 3 M Urea, 40 mM NH2OH, 4 mM DTT,  
5 mM iodoacetamide, 6 mM cysteine, and 0.1 M phosphate (pH ~7.1) 
was incubated at 37 °C. After 2 hours of incubation, this digest 
solution was diluted two fold with water and an aliquot of trypsin 
(Sigma T6567), such that the protein:trypsin ratio was 4:1 (w/w). 
After incubation at 37 °C for another 2 hours, the digest solution 
was again diluted two fold with water and a fresh aliquot of trypsin. 
With a total protein:trypsin ratio of 2:1 (w/w), the digest was left to 
incubate at 37 °C for 16 hours. Following this incubation, the digest 
was quenched by acidification with TFA and stored at -80 °C until 
analyzed. In preparation for HILIC chromatography, aqueous digests 
were diluted with 4 parts acetonitrile and 0.1 parts dimethylsulfoxide 
and were then centrifuged at 16 x 1000 g for 10 minutes to remove 

any insoluble composition. Supernatant from the centrifuged digest 
was thereafter injected.

Method conditions 
(unless otherwise noted): 

Column conditioning
ACQUITY UPLC Glycoprotein BEH Amide 300Å 1.7 µm Columns 
(as well as other amide columns intended for glycoprotein 
or glycopeptide separations) should be conditioned via two 
sequential injections and separations of 40 µg Glycoprotein 
Performance Test Standard (p/n 186008010; 10 µL injections 
of 4 mg/mL in 0.1% TFA, 80% ACN) or with equivalent loads 
of a test sample for which the column has been acquired. The 
separation outlined by the following method can be employed for 
conditioning with the Glycoprotein Performance Test Standard.

Column conditioning gradient
2.1 x 150 mm

Mobile phase A: 

0.1% (v/v) TFA, H2O

Mobile phase B: 

0.1% (v/v) TFA, ACN

  Time 

%A 

%B 

Curve 

 0.0  15.0  85.0 

6 

 0.5  15.0  85.0 

6 

 1.0  33.0  67.0 

6 

 21.0  40.0 

60.0 

6 

 22.0  100.0 

0.0 

6 

 24.0  100.0 

0.0 

6 

 25.0  15.0 

85.0 

6 

 35.0  15.0 

85.0 

6

LC conditions for LC-UV-MS of mAb glycopeptides  
(Figures 1–6):
LC system: 

ACQUITY UPLC H-Class Bio System

Sample temp.:  

10 °C

Analytical 
column temp.:  

30 °C (trastuzumab Lys-C digest  
HILIC separations)

60 °C (cetuximab Lys-C/tryptic digest  
HILIC separations)

60 °C (trastuzumab Lys-C reversed  
phase separations)

[ 38 ] HILIC Glycopeptide Mapping with a Wide-Pore Amide Stationary Phase

Flow Rate: 

0.2 mL/min

Mobile phase A: 

0.1% (v/v) TFA, H2O

Mobile phase B: 

0.1% (v/v) TFA, ACN

HILIC injection volume:   100–250 µL (Aqueous digests were 

diluted with 4 parts acetonitrile and  
0.1 parts dimethylsulfoxide to obtain  
a miscible, HILIC compatible diluent.) 

Reversed phase 
injection volume: 

24.2 µL (Aqueous digest)

Columns:  

ACQUITY UPLC Glycoprotein BEH 
Amide 300Å 1.7 µm, 2.1 x 150 mm 
(p/n 176003702, with Glycoprotein 
Performance Test Standard)

ACQUITY UPLC Peptide BEH C18 300 Å  
1.7 µm, 2.1 x 150 mm (p/n 186003687)

Vials:  

Polypropylene 12 x 32 mm Screw Neck 
Vial, 300 µL Volume (p/n 186002640)

Gradient used for reversed phase separations of trastuzumab  
Lys-C digests (Figure 1A):
  Time 

%A 

%B 

Curve 

 0.0  98.0 

2.0 

6 

 96.0  50.0 

50.0 

6 

 99.0  20.0 

80.0 

6 

 101.0  20.0 

80.0 

6 

 102.0  98.0 

2.0 

6 

 113.0  98.0 

2.0 

6

Gradient used for HILIC separations of trastuzumab  
Lys-C digests and Lys-C/tryptic digests of cetuximab 
(Figures 1B-6):
  Time 

%A 

%B 

Curve 

 0.0  20.0  80.0 

6 

 60.0  50.0 

50.0 

6 

 61.0  80.0 

20.0 

6 

 63.0  80.0 

20.0 

6 

 64.0  20.0 

80.0 

6 

 75.0  20.0 

80.0 

6

MS conditions for IgG subunit separations
MS system:  

SYNAPT G2-S HDMS

Ionization mode:  

ESI+

Analyzer mode:  

Resolution (~20 K)

Capillary voltage:  

3.0 kV

Cone voltage:  

25 V

Source temp.:  

120 °C

Desolvation temp.:  

350 °C

Desolvation gas flow:   800 L/Hr
Calibration:  

NaI, 1 µg/µL from 100–2000 m/z

Lockspray: 

300 fmol/µL Human Glufibrinopeptide B  
in 0.1% (v/v) formic acid,  
70:30 water/acetonitrile  
every 90 seconds

Acquisition:  

50–2500 m/z, 0.1 sec scan rate 

Data management: 

MassLynx Software (V4.1) / UNIFI V1.7

LC Conditions for a Glycopeptide Mapping of an IgG1K  
with Fluorescence Detection (Figure 7):
LC system:  

ACQUITY UPLC H-Class Bio System

Sample temp.:  

10 °C

Analytical 
column temp.:  

45 °C 

Fluorescence detection: Ex 280/Em 320 nm  

(10 Hz scan rate, Gain =1) 

Injection volume:  

100 µL (DMF/ACN diluted sample)

Mobile phase A:  

0.1% TFA in water

Mobile phase B: 

0.1% TFA in ACN

Columns:  

ACQUITY UPLC Glycoprotein BEH Amide 
300Å 1.7 µm, 2.1 x 50 mm  
(p/n 176003702, with Glycoprotein 
Performance Test Standard)

Other columns:  

Column A: 2.6 µm, 2.1 x 150 mm

Column B: 1.8 µm, 2.1 x 150 mm 

Vials:  

Polypropylene 12 x 32mm Screw Neck 
Vial, 300 µL Volume (p/n 186002640)

Gradient (Figure 7):
 Time 
 (min)  

%A 

%B 

Curve 

 0.0  15.0  85.0 

6 

 0.5  15.0  85.0 

6 

 1.0  30.0  70.0 

6 

 21.0  37.0 

63.0 

6 

 22.0  100.0 

0.0 

6 

 24.0  100.0 

0.0 

6 

 25.0  15.0 

85.0 

6 

 35.0  15.0 

85.0 

6

Data management: 

UNIFI v1.7

[ 39 ]

HILIC Glycopeptide Mapping with a Wide-Pore Amide Stationary Phase 

R E SU LT S A N D D IS C U S S IO N

Orthogonal and complementary glycopeptide mapping separations
To demonstrate a conventional approach to peptide mapping, we first performed LC-UV-MS analysis on a 
Lys-C digest of a mAb using a RP chromatographic separation with a wide-pore C18 bonded stationary phase 
(Peptide BEH C18 300Å 1.7 μm). Trastuzumab was selected for this study, given its prominence as a first 
generation mAb drug product and a potential target for biosimilar development.13 Figure 1A shows a UPLC 
chromatogram that is typical for a Lys-C digest of trastuzumab, wherein peptides are broadly resolved 
across a separation with a gradient corresponding to a change of 0.5% acetonitrile per minute. The non-
glycosylated peptides of the digest spread across the extremes of the chromatogram while the glycopeptides 
elute in an approximately one minute wide window at a retention time of about 60 minutes. The conditions 
to produce this high resolution separation involve the use of mobile phases modified with trifluoroacetic acid 
(TFA); the same mobile phases that have proven to be optimal for HILIC of proteinaceous analytes.10-11 

Accordingly, an orthogonal method to the RP separation can be achieved via HILIC by simply reversing the 
gradient and using the newly developed wide-pore amide bonded stationary phase (Glycoprotein BEH Amide 
300Å 1.7 μm). An example of a chromatogram obtained from a column packed with this wide-pore amide 
material and a gradient ramp of 0.5% acetonitrile per minute is shown in Figure 1B. Here, the peptides from the 
Lys-C digested trastuzumab are very clearly segregated into early and late eluting species, corresponding to 
the non-glycosylated and glycosylated species, respectively. The use of TFA ion pairing facilitates obtaining 
this separation, as it masks the hydrophilicity of the peptide residues and provides improved selectivity for the 
hydrophilic modifications. Also note that the glycopeptides have not only been class separated with the amide 
column, but the selectivity of the peptide glycoforms is remarkably improved over the analogous RP separation. 

0.0

0.5

1.0

0

5

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

90

A

214

0.0

0.5

1.0

0

5

10

15

20

25

30

35

40

45

A

214

Time (min)

Peptide BEH C18  

300Å 1.7 

m 

Glycoprotein BEH Amide  
300Å 1.7 

m 

Non-glycosylated

Peptides  

Non-glycosylated

Peptides  

Glycopeptides 

Lys-C 

Glycopeptides 

A 

B 

Figure 1. Lys-C glycopeptide mapping of trastuzumab. (A) A traditional reversed phase separation of 
the Lys-C digest using a 2.1 x 150 mm ACQUITY UPLC Peptide BEH C18 300Å 1.7 µm Column.  

(B) A HILIC separation of the Lys-C digest using a 2.1 x 150 mm ACQUITY UPLC Glycoprotein BEH 
Amide 300Å 1.7 µm Column. In each analysis, 9.2 µg of the Lys-C digest was separated using  
the same gradient slope and injecting sample from a diluent comprised of either approximately 
0.2% TFA in 80:20 ACN/water (HILIC) or 100% water (reversed phase). 

[ 40 ] HILIC Glycopeptide Mapping with a Wide-Pore Amide Stationary Phase

By focusing on the strongly retained peaks, one can begin to interrogate the glycosylation of the 
trastuzumab molecule (Figure 2A). In particular, MS data acquired from online mass detection and a total 
ion chromatogram (TIC) can be applied to identify the peptide species and its corresponding glycoforms, as 
shown in Figure 2B. This Lys-C glycopeptide map presents a 29 amino acid residue peptide (K16) from the Fc 
domain of trastuzumab. From an analysis of the MS data, many biantennary structures typical found on mAbs 
in relatively high abundance can be readily identified (Figure 3A). Further interrogation of the MS data, also 
shows that low abundance N-glycan species can likewise be detected. Figure 3B, for instance, provides  
MS data supporting the identification of monosialylated and disialylated glycoforms at retention times of 
approximately 34.7 minutes and 36.4 minutes, respectively. These identifications correlate extremely well 
with the released N-glycan profiles of trastuzumab that have been previously reported.14-15 

0.01

0.04

0.07

25

26

27

28

29

30

31

32

33

34

35

36

37

38

A

214

0E+0

1E+6

2E+6

25

26

27

28

29

30

31

32

33

34

35

36

37

38

In

te

ns

it

y

Time (min)

TIC 

A1 

M5 

FA2G1 

FA2G2 

FA1G1 

A2G1 

A2 

FA1 

FA2 

-17.03 Da 

FA2G1 

-17.03 Da 

UV 

FA2G1’ 

FA2 

A 

B 

HC:K16 

TKPREEQYNSTYRVVSVLTVLHQDWLNGK 

Glycoforms 

33

34

35

36

37

38

FA2G2 

1380 1381 1382 1383 1384

m/z

FA2G2S1 
1380.8834, 4+ 
-4.1 ppm 

1453 1454 1455 1456 1457

m/z

FA2G2S2 
1453.6593, 4+ 
-2.5 ppm 

min 

Species  Modification  Calculated 

Mass (Da) 

RT 

(Min) 

Intensity 

Error 

(ppm) 

HC:K16 

A1 

4555.1865  26.4 

15461 

-1.0 

HC:K16 

A2 

4758.2656  27.7 

219414 

-1.1 

HC:K16 

FA1 

4701.2441  27.7 

17931 

-3.3 

HC:K16 

FA2 

4904.3237  28.8 

1502968 

-1.7 

HC:K16 

M5 

4676.2129  29.3 

51057 

-0.7 

HC:K16 

A2G1 

4920.3188  29.7 

89714 

-0.9 

HC:K16 

A2G1 

4920.3188  30.3 

33516 

-1.9 

HC:K16 

FA1G1 

4863.2974  30.4 

17005 

-1.1 

HC:K16 

FA2G1 

5066.3765  30.8 

1156785 

-0.7 

HC:K16 

FA2G1’ 

5066.3765  31.4 

312270 

-1.0 

HC:K16 

FA2G2 

5228.4292  33.2 

226743 

-2.0 

A 

B 

Figure 2. Lys-C glycopeptide mapping of trastuzumab 
with HILIC and an ACQUITY UPLC Glycoprotein BEH 
Amide 300Å 1.7 µm Column. (A) UV chromatogram for 
the Lys-C glycopeptide retention window. (B) Total ion 
chromatogram (TIC) for the same retention window. 

Figure 3. Mass spectrometric data supporting trastuzumab  
Lys-C glycopeptide identifications. (A) Retention times,  
MS intensities and mass errors for assignments labeled in 
Figure 2. (B) MS spectra supporting the identification of low 
abundance Lys-C glycopeptides modified with mono and 
di-sialylated N-glycans. 

[ 41 ]

HILIC Glycopeptide Mapping with a Wide-Pore Amide Stationary Phase

Lot-to-lot analysis of trastuzumab glycosylation via HILIC-UV glycopeptide mapping
HILIC-MS based glycopeptide mapping clearly yields information-rich data. However, these HILIC glycopeptide 
mapping separations also lend themselves to methods based only on optical detection. We have, for example, 
applied a HILIC-UV method to perform lot-to-lot analysis of trastuzumab glycosylation for two drug product 
samples. Representative HILIC chromatograms for glycopeptide K16 obtained from two different lots of 
trastuzumab are shown in Figure 4A. Previous released glycan analyses on these lots have shown there to 
be differences in glycosylation.14 Through comparison of peak areas across the glycopeptide profile, we have 
found that these two lots of trastuzumab indeed differ with respect to their glycosylation. Specifically, these 
lots of trastuzumab appear to have different extents of terminal galactosylation, as can be seen in the differing 
abundances of FA2, FA2G1 and FA2G2 glycoforms (Figure 4B). This observation was consistent with data 
obtained from previous released glycan analyses and previous HILIC based profiling of trastuzumab subunits.11 

1 

2 

3 

4 5  6 

7 

8 

9 

1 

2 

4 5  6 

7 

8 

9 

3 

25.0 

0.02 

0.02 

0.08 

0.08 

A

214 

A

214 

Lot 1 

Lot 2 

0

10

20

30

40

50

1 2 3 4 5 6 7 8 9

% Amount

Component

Batch 1
Batch 2

37.5 min 

FA2 

FA2G1 

FA2G2 

A 

B 

Figure 4. Lot-to-lot profiling of trastuzumab Lys-C peptide glycoforms. (A) HILIC chromatograms of trastuzumab Lys-C glycopeptides 
from two different lots of drug product. (B) Relative abundances of the major sample components. Analyses were performed in 
triplicate using a 2.1 x 150 mm ACQUITY UPLC Glycoprotein BEH Amide 300Å 1.7 µm Column. 

Complementing GlycoWorks RapiFluor-MS N-glycan analyses with domain and peptide specific 
information about mAb glycosylation
An appealing aspect of glycopeptide mapping is that it can be applied to the elucidation of domain and  
peptide specific information. By inference or ETD fragmentation analyses, or both, glycopeptide mapping  
can also be used to detail the exact sites of glycosylation.16 As we have noted before,11 IgGs contain one 
conserved N-glycosylation site at Asn297 of the heavy chain, meaning they will be modified with  
two glycans in their Fc subunit. In addition, some IgGs and even some mAb IgG therapeutics exhibit  
multi-domain glycosylation. Cetuximab, for instance, is glycosylated in both its Fc and Fab domains,17  
making it a very interesting case study for this work. 

[ 42 ] HILIC Glycopeptide Mapping with a Wide-Pore Amide Stationary Phase

A HILIC glycopeptide map of a Lys-C/tryptic digest of cetuximab provides a clear indication of the 
complicated glycan profile of this molecule (Figure 5). In the presented chromatogram, approximately thirty 
chromatographic peaks are observed. Furthermore, a cursory analysis of the MS data has shown there to be at a 
minimum twenty five different glycoform species with rather high relative abundances of greater than 1–2%. 
Figure 6 provides the MS data supporting these assignments. As can be seen, 9 unique glycoforms could be 
assigned to tryptic peptide T22 from the Fc domain, while the other 16 glycoforms could be assigned to tryptic 
peptide T8 from the Fab domain of cetuximab. It is interesting to note that the majority of the Fab domain (T8) 
glycans contain immunogenic epitopes, such as non-human alpa-1,3-galactose or non-human N-glycolyl-
neuraminic acid moieties.18 In previous work, these glycan species were identified through complementary 
subunit mapping and RapiFluor-MS released N-glycan analyses.11 With these results on glycopeptide mapping,  
we show yet another complementary technique for assessing protein glycosylation. 

Species 

Modification 

Calculated 

Mass (Da)  RT (Min)  Intensity 

Error 

(ppm) 

HC:T22 

FA1 

2429.9592 

32.2 

9047 

1.5 

HC:T22 

FA2 

2633.0386 

33.6 

263523 

1.4 

HC:T22 

M5 

2404.9275 

35.2 

78988 

0.0 

HC:T22 

FA2G1 

2795.0913 

36.1 

186156 

1.2 

HC:T22 

FA2G1' 

2795.0913 

36.7 

57185 

0.7 

HC:T22 

[Hex5HexNAc3DHex1] 

2754.0647 

37.8 

7783 

1.0 

HC:T22 

M6 

2566.9805 

38.1 

5432 

-1.8 

HC:T22 

FA2G2 

2957.1443 

38.9 

38257 

1.1 

HC:T22 

[Hex6HexNAc3DHex1] 

2916.1177 

40.5 

12246 

-1.3 

HC:T8 

FA2 

3350.3689 

29.4 

5688 

4.9 

HC:T8 

M5 

3122.2578 

30.4 

6437 

2.7 

HC:T8 

FA2G1 

3512.4216 

31.7 

8925 

3.2 

HC:T8 

FA2G2 

3674.4746 

34.1 

11695 

8.2 

HC:T8 

FA2G2Ga1 

3836.5273 

36.7 

9657 

-0.6 

HC:T8 

FA2G2Ga1' 

3836.5273 

37.0 

15605 

0.3 

HC:T8 

FA2G2Sg1 

3981.5649 

37.5 

25532 

1.9 

HC:T8 

FA2G2Ga2 

3998.5801 

39.1 

212644 

-0.7 

HC:T8 

A2G2 

3471.3950 

39.1 

21542 

2.0 

HC:T8 

[Hex7HexNAc5DHex1] 

4144.6377 

39.2 

6944 

-10.8 

HC:T8 

FA2G2Ga1Sg1 

4143.6177 

39.6 

90905 

-0.7 

HC:T8 

FA2G2Sg2 

4288.6553 

40.2 

7428 

-2.5 

HC:T8 

[Hex8HexNAc5DHex1] 

4363.7124 

41.8 

5280 

1.7 

HC:T8 

[Hex7HexNAc5DHex1NGNA1] 

4508.7498 

42.4 

7114 

-3.6 

HC:T8 

[Hex9HexNAc5DHex1] 

4525.7651 

43.8 

20104 

4.3 

HC:T8 

[Hex8HexNAc5DHex1NGNA1] 

4670.8027 

44.2 

7640 

1.2 

*Hex: hexose, HexNAc: N-acetylate hexosamine, DHex: deoxyhexose, NGNA: N-glycoyl neuraminic acid / (‘): structural isomer 

0.015

0.040

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

A

21

4

Time (min)

T8 

+FA2G1 

T22 

+FA1 

T22 

+FA2 

T8 

+FA2G2 

T22 

+M5 

T22 

+FA1G1 

T22 

+FA2G1 

T22 

+FA2G1’ 

T8 

+FA2G2Ga1 

T8 

+FA2G2Sg1 

T22 

+Hex5HexNAc3DHex1 

T22 

+M6 

T22 

+FA2G2 

T8 

+Hex8HexNAc5DHex1NGNA1 

T8 

+Hex9HexNAc5DHex1 

T8 

+Hex8HexNAc5DHex1 

T8 

+Hex7HexNAc5DHex1NGNA1 

T8 

+FA2G2Sg2 

T22 

+Hex6HexNAc3DHex1 

T22 

+FA2G2Ga1 

T8 (deamidated) 

+FA2G2Ga2 

T8  

+FA2G2Ga2 

T8  

+FA2G2Ga1Sg1 

HC:T22 

EEQYNSTYR 

Glycoforms 

HC:T8 

MNSLQSNDTAIYYC(am)AR  

Glycoforms 

Figure 5. Assaying the N-linked glycan sites of cetuximab and their microheterogeneity using a combined Lys-C/tryptic digest. 
Assignments for glycoforms of peptide T22 from the Fc domain of cetuximab are shown in dark gray, while assignments to the 
glycoforms of peptide T8 from the Fab domain are shown in red. Analyses were performed on 9.2 µg of Lys-C/tryptic digest using  
a 2.1 x 150 mm ACQUITY UPLC Glycoprotein BEH Amide 300Å 1.7 µm Column and a sample diluent of approximately 0.2% TFA  
in 80:20 ACN/water. 

Figure 6. Mass spectrometic data supporting 
glycopeptide assignments from the HILIC-UV-MS 
analysis of Lys-C/trypsin digested cetuximab. From 
a cursory analysis of MS data, 9 unique glycoforms 
could be assigned to the Fc domain, and 16 unique 
glycoforms could be assigned to the Fab domain.

[ 43 ]

HILIC Glycopeptide Mapping with a Wide-Pore Amide Stationary Phase

Benchmarking the Capabilities of the 
Glycoprotein BEH Amide 300Å 1.7 μm Column
The peak capacities obtained in these example 
glycopeptide separations are particularly 
noteworthy when a comparison is made to otherwise 
available column technologies. To benchmark 
the performance of the Glycoprotein BEH Amide 
300Å 1.7 μm Column, we have analyzed a Lys-C 
digest of a NIST candidate reference material, an 
IgG1K mAb. In this testing, a focused gradient was 
used along with intrinsic peptide fluorescence 
instead of low wavelength UV detection so that 
higher signal-to-noise could be achieved in the 
obtained chromatograms. The glycopeptide that 
originates from the Fc domain of a mAb will contain 
a tryptophan residue upon Lys-C cleavage, which in 
large part makes this detection mechanism feasible. 
Three fluorescence chromatograms obtained for 
the Lys-C glycopeptides from the NIST IgG1K are 
presented in Figure 7. These three chromatograms 
were obtained from the use of the ACQUITY UPLC 
Glycoprotein BEH Amide 300Å 1.7 µm as well as 
two commercially available alternatives (Columns 
A and B). Peak capacities have been measured for 
each specific column using the retention windows 
demarcated by the most extreme glycopeptide 
retention times (*) and the half-height peak widths 
of the K16+FA2, K16+FA2G1, K16+FA2G1', 
K16+FA2G2, and K16+FA2G2Ga1 peaks. This 
analysis shows that these columns exhibit strikingly 
different resolving power. With an effective peak 
capacity of 72.8, the Glycoprotein BEH Amide 
column shows a superior peak capacity and 
performance increases over the alternative  
amide column technologies of 40 and 96%. 

Glycoprotein BEH Amide 300Å 1.7 m 

Column B: 1.8 µm, 2.1 x 150 mm

Column A: 2.6 µm, 2.1 x 150 mm

HC:K16 

+FA2 

HC:K16 

+FA2G2Ga1 

Pc* = 51.9 

HC:K16 

+FA2G1 

HC:K16 

+FA2G1’  HC:K16 

+FA2G2 

* 

* 

* 

* 

* 

* 

Pc* = 37.1 

Pc* = 72.8 

Figure 7. Lys-C glycopeptide mapping of an IgG1K using fluorescence detection and 
various 2.1 x 150 mm columns packed with amide bonded stationary phase:  
a Competitor Column A: 150Å 2.6 µm, 2.1 x 150 mm (Top), a Competitor Column B:  
1.8 µm, 2.1 x 150 mm (Middle), and on an ACQUITY UPLC Glycoprotein BEH Amide 
300Å 1.7 µm column (Bottom). Peak capacities were calculated based on the half-height 
peak widths of the labeled glycopeptides and the retention window established by the 
earliest and latest eluting glycopeptide species, marked with asterisks(*). Comparative 
separations may not be representative in all applications.

[ 44 ] HILIC Glycopeptide Mapping with a Wide-Pore Amide Stationary Phase

Waters Corporation 
34 Maple Street 
Milford, MA 01757 U.S.A. 
T: 1 508 478 2000 
F: 1 508 872 1990 
www.waters.com

Waters, ACQUITY UPLC, SYNAPT and The Science of What's Possible are registered trademarks of Waters Corporation. GlycoWorks  
and RapiFluor-MS are trademarks of Waters Corporation. All other trademarks are the property of their respective owners.

©2015 Waters Corporation. Produced in the U.S.A.  September 2015  720005409EN  AG-PDF

C O N C LU S IO NS
Glycopeptide mapping of glycoproteins presents a highly effective technique 
that can be used to elucidate both domain and peptide-specific glycosylation. 
In this work, we have demonstrated the use of an ACQUITY UPLC Glycoprotein 
BEH Amide 300Å 1.7 μm Column to obtain HILIC separations of glycopeptides 
that complement the chromatographic information afforded by a reversed phase 
separation. In addition, our results indicate that these HILIC separations provide 
exemplary peak capacity in comparison to other commercially available amide 
column technologies. That the HILIC separation is MS-compatible means that 
information-rich data can be readily acquired to characterize a glycopeptide 
map. For instance, this work shows that it can be a relatively straightforward 
exercise to characterize multidomain protein glycosylation, such as the Fc 
and Fab domain glycosylation of cetuximab. Combined with other recently 
developed strategies, such as HILIC subunit mapping and GlycoWorks RapiFluor-
MS released N-glycan analyses, glycopeptide mapping with the ACQUITY UPLC 
Glycoprotein BEH Amide Column shows significant promise for facilitating the 
characterization of protein glycosylation to unprecedented levels of detail.

References

1.  Xie, H.; Gilar, M.; Gebler, J. C., Characterization of protein impurities and site-specific 

modifications using peptide mapping with liquid chromatography and data independent 
acquisition mass spectrometry. Anal Chem 2009, 81 (14), 5699–708.

2.  Witze, E. S.; Old, W. M.; Resing, K. A.; Ahn, N. G., Mapping protein post-translational 

modifications with mass spectrometry. Nat Methods 2007, 4 (10), 798–806.

3.  Huang, H. Z.; Nichols, A.; Liu, D., Direct identification and quantification of aspartyl 

succinimide in an IgG2 mAb by RapiGest assisted digestion. Anal Chem 2009, 81 (4), 
1686–92.

4.  Lauber, M. A.; Koza, S. M.; McCall, S. A.; Alden, B. A.; Iraneta, P. C.; Fountain, K. J.,  

High-Resolution Peptide Mapping Separations with MS-Friendly Mobile Phases and  
Charge-Surface-Modified C18. Anal Chem 2013, 85 (14), 6936–44.

5.  Gilar, M.; Yu, Y. Q.; Ahn, J.; Xie, H.; Han, H.; Ying, W.; Qian, X., Characterization of glycoprotein 

digests with hydrophilic interaction chromatography and mass spectrometry. Anal Biochem 
2011, 417 (1), 80–8.

6.  Martin Gilar; Ying-Qing Yu; Joomi Ahn; Xie, H., Analysis of Glycopeptide Glycoforms in 

Monoclonal Antibody Tryptic Digest using a UPLC HILIC Column. Waters Application Note 
720003363en 2010.

7. Ahn, J.; Bones, J.; Yu, Y. Q.; Rudd, P. M.; Gilar, M., Separation of 2-aminobenzamide labeled 

glycans using hydrophilic interaction chromatography columns packed with 1.7 microm sorbent. J 

Chromatogr B Analyt Technol Biomed Life Sci 2010, 878 (3–4), 403–8.

8.  Gustavsson, P.-E.; Larsson, P.-O., Support Materials 

for Affinity Chromatography. In Handbook of Affinity 
Chromatography, Hage, D., Ed. Taylor & Francis: Boca Raton, 
FL, 2006; pp 15–33.

9.  Renkin, E. M., J. Gen. Physio. 1954, (38), 225.

10. Lauber, M. A.; Koza, S. M., Developing High Resolution  

HILIC Separations of Intact Glycosylated Proteins using 
a Wide-Pore Amide-Bonded Stationary Phase Waters 
Application Note 720005380en 2015.

11. Lauber, M. A.; Koza, S. M., Mapping IgG Subunit Glycoforms 

using HILIC and a Wide-Pore Amide Stationary Phase Waters 
Application Note 720005385en 2015.

12. Lauber, M. A.; Koza, S. M., Enhancing the Peak Capacity of 

High Molecular Weight N-Glycan HILIC Separations with a 
Wide-Pore Amide Bonded Stationary Phase. Waters Tech 
Brief 720005381en 2015.

13. Beck, A.; Sanglier-Cianferani, S.; Van Dorsselaer, A., 

Biosimilar, biobetter, and next generation antibody 
characterization by mass spectrometry. Anal Chem 2012,  
84 (11), 4637–46.

14. Yu, Y. Q.; Ahn, J.; Gilar, M., Trastuzumab Glycan Batch-to-

Batch Profiling using a UPLC/FLR/MS Mass Spectrometry 
Platform. Waters Appication Note 720003576en 2010.

15. Xie, H.; Chakraborty, A.; Ahn, J.; Yu, Y. Q.; Dakshinamoorthy, 

D. P.; Gilar, M.; Chen, W.; Skilton, S. J.; Mazzeo, J. R., Rapid 
comparison of a candidate biosimilar to an innovator 
monoclonal antibody with advanced liquid chromatography 
and mass spectrometry technologies. MAbs 2010, 2 (4).

16. Houel, S.; Hilliard, M.; Yu, Y. Q.; McLoughlin, N.; Martin, 

S. M.; Rudd, P. M.; Williams, J. P.; Chen, W., N- and 
O-glycosylation analysis of etanercept using liquid 
chromatography and quadrupole time-of-flight mass 
spectrometry equipped with electron-transfer dissociation 
functionality. Anal Chem 2014, 86 (1), 576–84.

17. Qian, J.; Liu, T.; Yang, L.; Daus, A.; Crowley, R.; Zhou, Q., 

Structural characterization of N-linked oligosaccharides 
on monoclonal antibody cetuximab by the combination of 
orthogonal matrix-assisted laser desorption/ionization 
hybrid quadrupole-quadrupole time-of-flight tandem  
mass spectrometry and sequential enzymatic digestion.  
Anal Biochem 2007, 364 (1), 8–18.

18. Arnold, D. F.; Misbah, S. A., Cetuximab-induced anaphylaxis 

and IgE specific for galactose-alpha-1,3-galactose.  
N Engl J Med 2008, 358 (25), 2735; author reply 2735–6.

[ 45 ]

HILIC Glycopeptide Mapping with a Wide-Pore Amide Stationary Phase

WAT E R S  S O LU T IO NS
ACQUITY UPLC® Glycoprotein BEH 
Amide, 300Å Column

Glycoprotein Performance Test Standard

GlycoWorks™ RapiFluor-MS™ N-Glycan Kit

ACQUITY UPLC H-Class Bio System

Xevo® G2 QTof Mass Spectrometer

SYNAPT® G2-S HDMS

K E Y W O R D S
ACQUITY UPLC H-Class Bio System,  
BEH Amide 300Å, Glycans, Glycosylated 
Proteins, Glycosylation, HILIC, IdeS

A P P L I C AT IO N  B E N E F I T S 

■

■

Improved HILIC separations of IgG  
subunit glycoforms. 

■

■

MS-compatible HILIC to enable detailed 
investigations of sample constituents.

■

■

Orthogonal selectivity to conventional 
reversed-phase (RP) separations for 
enhanced characterization of hydrophilic 
protein modifications.

■

■

Domain-specific glycan information that 
complements profiling glycosylation by 
RapiFluor-MS released N-glycan analyses. 

■

■

Glycoprotein BEH amide, 300Å,  
1.7 µm stationary phase is QC tested via  
a glycoprotein separation to ensure 
consistent batch to batch reproducibility.

I N T RO DU C T IO N
Without question, the most successfully exploited protein modality for therapeutic 
applications has been monoclonal antibodies (mAbs), which currently account 
for nearly half of the biopharmaceutical market.1 An intriguing characteristic of 
mAbs, in particular IgG-based mAbs, is that they are formed by the linking of two 
identical light chains and two identical heavy chains through disulfide bonding 
and non-covalent interactions. Moreover, the resulting mAb structure exhibits 
functionally significant subunits, for instance one crystallizable fragment  
(Fc domain) and two equivalent antigen binding fragments (Fab domains).  
In what is commonly referred to as a middle-up or middle-down analysis,2-5 native 
mAbs can be proteolyzed into these and other related subunits enzymatically, 
as a means to perform cell-based studies and to facilitate characterization. One 
increasingly popular way to produce subunit digests of mAbs is via the IdeS 
protease (Immunoglobulin Degrading Enzyme of S. pyogenes).2,6 IdeS cleaves 
with high fidelity at a conserved sequence motif in the hinge region of humanized 
mAbs to cleanly produce, upon reduction, three 25 kDa mAb fragments that are 
amenable to mass spectrometry and useful for localizing different attributes of 
therapeutic mAbs (Figure 1).3 IdeS digestion combined with reversed-phase (RP) 
chromatography has, in fact, been proposed as a simple identity test for mAbs 
and fusion proteins, because IdeS produced subunits from different drug products 
will exhibit diagnostic RP retention times.3 Additionally, RP techniques have been 
shown to be useful in assaying and obtaining domain specific information about 
oxidation, since RP retention can be dramatically affected by the oxidation of 
protein residues, such as methionine.3

Mapping IgG Subunit Glycoforms Using HILIC and a Wide-Pore  

Amide Stationary Phase 
Matthew A. Lauber and Stephan M. Koza

Waters Corporation, Milford, MA, USA

Cleavage  

Site 

-G—G- 

IdeS 

Digestion 

(Fab )2 

2x Light Chain 

2x Fd

2x Fc/2 

Denaturation 

Reduction 

2x Fc/2 

Figure 1. IdeS digestion and reduction scheme for preparing IgG LC, Fd', and Fc/2 subunits. 

[ 46 ]

E X P E R IM E N TA L 

Sample description

IdeS digestion and reduction of mAbs:
Formulated trastuzumab was diluted 7 fold into 20 mM 
phosphate (pH 7.1) and incubated at a concentration of 3 mg/mL  
with IdeS (Promega, Madison, WI) for 30 minutes at 37 °C at  
a 50:1 w/w ratio of trastuzumab to IdeS. The resulting 
IdeS-digested antibody was denatured and reduced by the 
addition of 1M TCEP (tris(2-carboxyethyl)phosphine) and solid 
GuHCl (guanidine hydrochloride). The final buffer composition for 
the denaturation/reduction step was approximately 6 M GuHCl, 
80 mM TCEP, and 10 mM phosphate (pH 7.1). IdeS-digested 
trastuzumab (1.5 mg/mL) was incubated in this buffer at 37 °C 
for 1 hour. An IdeS digested, reduced sample of an IgG1K mAb 
obtained from NIST as candidate reference material #8670  
(lot #3F1b) was prepared in the same manner.

Cetuximab IdeS/carboxypeptidase B digestion  
and reduction:
Prior to digestion with IdeS,10 cetuximab was treated with 
carboxypeptidase B to complete the partial removal of the 
lysine-C-terminal residues that is typical of the antibody.4 
Formulated cetuximab was mixed with carboxypeptidase B 
(223 µ/mg, Worthington, Lakewood, NJ) at a ratio of 100:1 
(w/w), diluted into 20 mM phosphate (pH 7.1), and incubated 
at a concentration of 1.8 mg/mL for 2 hours at 37 °C. The 
carboxypeptidase B treated cetuximab was then added to 
100 units of IdeS and incubated for 30 minutes at 37 °C. The 
resulting IdeS digest was denatured and reduced by the addition 
of 1 M TCEP and solid GuHCl. The final buffer composition for 
the denaturation/reduction step was approximately 6 M GuHCl, 
80 mM TCEP, and 10 mM phosphate (pH 7.1). IdeS-digested 
cetuximab (0.9 mg/mL) was incubated in this buffer at 37 °C 
for 1 hour.

Preparation of RapiFluor-MS Labeled N-Glycans  
from Cetuximab:
RapiFluor-MS labeled N-glycans were prepared from  
cetuximab using a GlycoWorks RapiFluor-MS N-Glycan Kit  
(p/n 176003606) according to the guidelines provided in  
its Care and Use Manual (715004793).

Method conditions 
(unless otherwise noted)

Column conditioning
ACQUITY UPLC Glycoprotein BEH Amide, 300Å, 1.7 µm columns 
(as well as other amide columns intended for glycoprotein 
separations) should be conditioned via two sequential injections 
and separations of 40 µg Glycoprotein Performance Test 
Standard (p/n 186008010; 10 µL injections of 4 mg/mL in 
0.1% trifluoroacetic acid [TFA], 80% acetonitrile [ACN]) or with 
equivalent loads of a sample for which the column has been 
acquired. The separation outlined by the following method can  
be employed for conditioning with the Glycoprotein Performance 
Test Standard.

Column conditioning gradient:
Column dimensions:  2.1 x 150 mm
Mobile phase A:  

0.1% (v/v) TFA, water

Mobile phase B:  

0.1% (v/v) TFA, ACN 

Time 

 (min) 

%A 

%B 

Curve 

 0.0  15.0 85.0  6 
 0.5  15.0 85.0  6 
 1.0  33.0 67.0  6 
 21.0  40.0  60.0  6 
 22.0  100.0  0.0 

6 

 24.0  100.0  0.0 

6 

 25.0  15.0  85.0  6 
 35.0  15.0  85.0  6

LC conditions for IgG subunit separations
LC system:  

ACQUITY UPLC H-Class Bio System

Sample temp.:  

5 °C

Analytical 
column temp.:  

45 °C (trastuzumab and NIST IgG1K  
subunit HILIC separations)

60 °C (cetuximab subunit HILIC 
separations)

80 °C (trastuzumab reversed phase (RP) 
subunit separations)

Flow rate: 

0.2 mL/min

Mobile phase A: 

0.1% (v/v) TFA, water

Mobile phase B: 

0.1% (v/v) TFA, ACN

UV detection:  

214 nm, 10 Hz

[ 47 ]

Mapping IgG Subunit Glycoforms Using HILIC and a Wide-Pore Amide Stationary Phase

Injection volume:  

≤1.2 µL (aqueous diluents). Note: It might 
be necessary to avoid high organic 
diluents for some samples due to the 
propensity for proteins to precipitate 
under ambient conditions. A 2.1 mm 
I.D. column can accommodate up to 
a 1.2 µL aqueous injection before 
chromatographic performance is 
negatively affected. 

Waters columns:  

ACQUITY UPLC Glycoprotein BEH 
Amide, 300Å, 1.7 µm, 2.1 x 150 mm 
(p/n 176003702, with Glycoprotein 
Performance Test Standard);

ACQUITY UPLC Glycan BEH Amide, 
130Å, 1.7 µm, 2.1 x 150 mm  
(p/n 186004742); 

ACQUITY UPLC Protein BEH C4, 300Å, 
1.7 µm, 2.1 x 150 mm (p/n 186004497) 

Other columns:  

Agilent® AdvanceBio Glycan Mapping, 
1.8 μm, 2.1 x 150 mm;

Thermo Scientific® Accucore™ Amide  
150 HILIC, 150Å 2.6 μm, 2.1 x 150 mm 

Vials:  

Polypropylene 12 x 32 mm Screw Neck 
Vial, 300 µL volume (p/n 186002640)

Gradient used for reversed-phase (RP) separations of 
trastuzumab subunits (Figure 2A):

Time 

 (min) 

%A 

%B 

Curve 

 0.0  95.0  5.0  6 
 1.0  66.7 33.3  6 
 21.0  59.7  40.3  6 
 22.0  20.0  80.0  6 
 24.0  20.0  80.0  6 
 25.0  95.0  5.0 

6 

 35.0  95.0  5.0 

6

Gradient used for HILIC separations of IgG subunits (Figures 2–7):

  Time 
 (min)  %A 

%B  Curve 

  0.0  20.0 80.0  6 
  1  30.0 70.0  6 
  21  37.0 63.0  6 
  22  100.0 0.0  6 
  24  100.0 0.0  6 
  25  20.0 80.0  6 
  35  20.0 80.0  6

MS conditions for IgG subunit separations
MS system:  

Xevo G2 QTof or SYNAPT G2-S HDMS

Ionization mode:  

ESI+

Analyzer mode:  

Resolution (~20 K)

Capillary voltage:  

3.0 kV

Cone voltage:  

45 V

Source temp.:  

150 °C

Desolvation temp.:  

350 °C

Desolvation gas flow:   800 L/Hr
Calibration:  

NaI, 2 µg/µL from 500–5000 m/z

Acquisition:  

500–4000 m/z, 0.5 sec scan rate 

Data management: 

MassLynx® Software (v4.1)/UNIFI V1.7 

LC conditions for RapiFluor-MS Released N-Glycan  
HILIC separations:
LC system:  

ACQUITY UPLC H-Class Bio System

Sample temp.:  

10 °C

Analytical 
column temp.:  

60 °C 

Fluorescence detection: Ex 265/Em 425 nm (RapiFluor-MS) 

(5 Hz scan rate [50 mm column], Gain =1) 

Injection volume:  

10 µL (DMF/ACN diluted sample)

Mobile phase A:  

50 mM ammonium formate, pH 4.4  
(LC-MS grade; from a 100x concentrate,  
p/n 186007081)

Mobile phase B:  

ACN (LC-MS grade)

Columns:  

ACQUITY UPLC Glycan BEH Amide, 130Å, 
1.7 µm, 2.1 x 50 mm (p/n 186004740) 

Vials:  

Polypropylene 12 x 32mm, 300 μL,  
Screw Neck Vial, (p/n 186002640)

Gradient used for RapiFluor-MS N-Glycan HILIC Separations  
(Figure 7B):

  Time  Flow Rate 
  (min)   (mL/min)  %A 

%B  Curve

 0.0 

0.4  25  75  6 

 11.7 

0.4 

46  54 

6 

 12.2 

0.2  100  0 

6 

 13.2 

0.2  100  0 

6 

 14.4 

0.2 

25  75 

6 

 15.9 

0.4 

25  75 

6 

 18.3 

0.4 

25  75 

6

[ 48 ] Mapping IgG Subunit Glycoforms Using HILIC and a Wide-Pore Amide Stationary Phase

MS conditions for RapiFluor-MS N-Glycan HILIC 
separations
MS system:  

SYNAPT G2-S HDMS

Ionization mode:  

ESI+

Analyzer mode:  

TOF MS, resolution mode (~20 K)

Capillary voltage:  

3.0 kV

Cone voltage:  

80 V

Source temp.:  

120 °C

Desolvation temp.:  

350 °C

Desolvation gas flow:   800 L/Hr

Calibration:  

NaI, 1 µg/µL from 500–2500 m/z

Lockspray 
(ASM B-side):  

100 fmol/µL Human Glufibrinopeptide B 
in 0.1% (v/v) formic acid,  
70:30 water every 90 seconds 

Acquisition:  

500–2500 m/z, 1 Hz scan rate 

Data management: 

MassLynx Software (v4.1)

It should, however, be kept in mind that many IgG modifications 
more strongly elicit changes in the hydrophilicity of a mAb 
along with its capacity for hydrogen bonding. A very obvious 
example of this type of modification is glycosylation. Glycans 
released from a mAb are very often profiled by hydrophilic 
interaction chromatography (HILIC), in which case an amide 
bonded stationary phase has historically been used, because it 
affords high retentivity as a consequence of its hydrophilicity and 
propensity for hydrogen bonding.7 Here, we propose that HILIC 
with an amide bonded stationary phase also be considered for 
IgG subunit separations. For such an application, a stationary 
phase with a wide average pore diameter is critical, so that large 
subunit structures will have access to the majority of the porous 
network and be less prone to restricted diffusion while eluting 
through a column.8-9 Through the development of a sub-2-μm 
wide-pore amide stationary phase, we have facilitated a novel 
and complementary workflow to RP based subunit analyses. In 
this application note, we demonstrate the use of a glycoprotein 
BEH amide, 300Å, 1.7 μm column to develop LC-MS and LC-UV 
techniques that can be used to rapidly profile domain specific 
information about the N-linked glycosylation of IgG molecules. 

R E SU LT S A N D D IS C U S S IO N

Orthogonal, complementary IgG subunit separations
To demonstrate a conventional approach to IgG subunit mapping, 
we first analyzed a reduced\IdeS digest of an IgG1 mAb using 
a RP chromatographic separation with a wide-pore C4 bonded 
stationary phase (Protein BEH C4, 300Å, 1.7 μm). The IgG1 mAb 
selected for this work was trastuzumab, given its prominence as 
a first generation mAb drug product and a potential target for 
biosimilar development.11 Figure 2A shows a UPLC chromatogram 
that is typical for reduced, IdeS-digested trastuzumab, wherein 
three peaks are near equally spaced with an elution order 
corresponding to the Fc/2, LC and Fd' subunits, respectively. The 
conditions to produce this high resolution separation entail the 
use of TFA for ion-pairing. Interestingly, the same mobile phases 
have proven to be optimal for protein HILIC, as they reduce 
the hydrophilicity of protein residues by masking them via a 
hydrophobic ion pair. This, in turn, leads to improved selectivity 
for hydrophilic modifications.12 That is, an orthogonal method to 
the RP separation can be achieved via HILIC by simply reversing 
a gradient and using a newly developed wide-pore amide bonded 
stationary phase (glycoprotein BEH Amide, 300Å, 1.7 μm).  

[ 49 ]

Mapping IgG Subunit Glycoforms Using HILIC and a Wide-Pore Amide Stationary Phase

An example of a chromatogram obtained from a 
column packed with this wide-pore amide material 
is shown in Figure 2B. Here, the same reduced, 
IdeS digested trastuzumab is separated into 
approximately 10 peaks. The first two eluting 
peaks correspond to the Fd' and LC subunits, while 
the remaining, more strongly retained peaks 
correspond to the glycoforms of the Fc/2 subunit. 
By focusing on the more strongly retained peaks, 
an analyst can elucidate information about the 
heterogeneity of glycosylation (Figure 3A). Given 
that this is a method with volatile mobile phases, 
the glycoform peaks can be readily interrogated 
by ESI-MS. Deconvoluted mass spectra and 
molecular weights corresponding to species in the 
glycoform profile are presented in Figures 3B and 
3C. In Figure 3, chromatographic peaks are labeled 
with the same color as their corresponding mass 
spectra. Notice that this HILIC separation facilitates 
producing deconvoluted mass spectra for individual 
glycoforms with limited interference between 
similar molecular weight species, for instance the 
Fc/2+A2G1 versus the Fc/2+FA2 species (orange 
versus blue spectrum). In a first pass analysis, all 
glycan species from trastuzumab that are known 
to be present at a relative abundance greater than 
2% are readily detected.13 It should be noted that 
lower abundance species, such as Fc/2+M5 (Man5), 
are also detected and can be observed by extracted 
ion chromatograms (XICs). This indicates there is a 
possibility to perform selected reaction monitoring 
(SRM) MS analyses when and if there is a need to 
monitor particular low abundance structures. While 
it is not resolved under these conditions, the M5 
Fc/2 glycoform is resolved in a different example 
separation (see below, Figure 7A). 

0.00

0.04

0.08

0.12

0.16

6

7

8

9

10

11

12

13

14

15

16

A

214

Time (min)

0.00

0.06

0.12

0.18

0.24

6

7

8

9

10

11

12

13

14

15

16

A

214

Time (min)

Fc/2 Glycoforms 

Fd’ 

LC 

Protein BEH C4, 300Å, 1.7 µm 

LC 

Fd’ 

Glycoprotein BEH Amide, 300Å, 1.7 µm 

Fc/2 

A 

B 

+Glycans 

Figure 2. Trastuzumab subunit separations. (A) 1 µg of reduced, IdeS digested separated using 
an ACQUITY UPLC Protein BEH C4, 300Å, 1.7 µm Column (0.7 µL aqueous injection). (B) 1 µg of 

reduced, IdeS digested separated using an ACQUITY UPLC Glycoprotein BEH Amide, 300Å, 1.7 µm 
Column (0.7 µL aqueous injection). 

Figure 3. Profiling trastuzumab Fc/2 subunit glycoforms. (A) Retention window from Figure 2B 
corresponding to the glycoform separation space. (B) Deconvoluted ESI mass spectra for the 
HILIC chromatographic peaks. Chromatographic peaks are labeled with the same color as their 
corresponding mass spectra. (C) Molecular weights for the observed trastuzumab subunits. 

Species 

MWAvg 

Theoretical 

(Da) 

MWAvg 

Observed 

(Da) 

Fd  

25383.6 

25383.3 

LC 

23443.1 

23443.1 

Fc/2+A2 

25090.2 

25091.0 

Fc/2+FA2 

25236.3 

25236.9 

Fc/2+A2G1 

25252.3 

25248.7 

Fc/2+FA2G1 

25398.5 

25398.5 

Fc/2+FA2G1  

25398.5 

25399.1 

Fc/2+FA2G2 

25560.6 

25561.8 

B 

C 

0.01

0.02

0.03

0.04

0.05

A

21

4

Time (min)

Fc/2

+FA2

Fc/2

+FA2G1

Fc/2

+FA2G2

Fc/2

+FA2G1’

Fc/2

+A2G1

Fc/2

+A2

A 

24900

25300

25700

Molecular Weight Da

25091.0 Da 

25236.9 Da 

25248.7 Da 

25398.5 Da 

25399.1 Da 

255561.8 Da 

[ 50 ] Mapping IgG Subunit Glycoforms Using HILIC and a Wide-Pore Amide Stationary Phase

Batch-to-batch analysis of trastuzumab  
Fc/2 glycosylation by HILIC-UV profiling
Clearly, data generated by subunit-level HILIC-MS 
are very information-rich. Optical detection 
based subunit HILIC separations can be equally 
informative. To this end, we have applied a HILIC-UV 
method to perform batch-to-batch analysis of 
trastuzumab Fc/2 glycosylation, as exemplified 
in Figure 4. Two example HILIC chromatograms for 
Fc/2 glycoforms obtained from two different lots 
of trastuzumab are shown in Figure 4A. Previous 
testing on these lots has demonstrated differences 
in glycosylation at the released glycan level.14 
Here, by integration of peaks across the profile, 
we have found that the two lots of trastuzumab 
indeed differ with respect to their Fc domain 
glycosylation profiles, in ways consistent with the 
mentioned released glycan assays. In particular, 
these lots of trastuzumab differ with respect to their 
extents of terminal galactosylation, as estimated 
from the abundances of FA2, FA2G1, and FA2G2 
Fc/2 subunits (Figure 4B). This is an informative 
observation, since the extent of galactosylation can 
affect complement-dependent cytotoxicity (CDC).15

Lifetime testing of glycoprotein BEH amide 
300Å, 1.7μm columns for profiling IgG  
subunit glycoforms
The ability of a BEH amide, 300Å, 1.7 μm column  
to robustly deliver the above mentioned separations 
over time was tested by performing a series of 
experiments involving a single column being 
subjected to 300 sequential injections of a reduced, 
IdeS digested trastuzumab sample. This was a 
potentially challenging use scenario given that 
the reduced, IdeS digested mAb sample contains 
both high concentrations of guanidine denaturant 
and TCEP reducing agent. Total ion chromatograms 
corresponding to the 20th, 180th, and 300th 
injections of this experiment are displayed in Figure 
5A. In these analyses, particular attention was paid 
to the half-height resolution of the Fc/2+A2 and 
Fc/2+FA2 species, which was assessed every 20th 
separation using extracted ion chromatograms (XICs). 

Batch 1 

Batch 2 

0 

10 

20 

30 

40 

50 

1 

2 

3 

4 

5 

6 

7 

8 

% Amount

Component 

Batch 1 

Batch 2 

FA2 

FA2G1 

FA2G2 

A 

B 

Figure 4. Batch-to-batch profiling of trastuzumab Fc/2 subunit glycoforms. (A) HILIC 
chromatograms of trastuzumab Fc/2 subunit glycoforms from two different lots of drug product.  
(B) Relative abundances of the major sample components. Analyses were performed in triplicate 
using an ACQUITY UPLC Glycoprotein BEH Amide, 300Å, 1.7 µm, 2.1 x 150 mm Column. 

Figure 5. Lifetime testing of an ACQUITY UPLC Glycoprotein BEH Amide, 300Å, 1.7 µm,  
2.1 x 150 mm Column for sequential injections of reduced, IdeS digested trastuzumab. (A) Total 
ion chromatograms (TICs) from the 20th, 180th, and 300th injections. Example extracted ion 
chromatograms (XICs) for Fc/2+A2 and Fc/2+FA2 that were used to measure resolution.  
(B) Chromatographic parameters observed across the 300 injection lifetime test. Each panel  
shows results for each 20th injection, including retention time (RT) of the FA2 glycoform,  
Rs between A2 and FA2 glycoforms, maximum pressure across the run, and % carryover as 

measured by a repeat gradient and XICs.

3000

4000

5000

6000

7000

0

1

2

3

4

5

6

2

4

6

8

10

0.0

0.1

0.2

0.3

0.4

0

100

200

300

Injection

Fc/2+FA2 

1010.0-1011.0  

m/z 

Fc/2+A2 

1004.2-1005.2  

m/z 

XIC 

Rs 2.17 

20th Injection 

A 

B 

RT

(Fc/2+FA2

, min)

R

s 

(Fc/2+

A2

/FA

2) 

Max Pressur

e

(psi) 

% Carryove

r  

(Fc/2+FA2

) 

Fc/2 

+A2 

Fc/2 

+FA2 

Fc/2 

+A2 

Fc/2 

+FA2 

Fc/2 

+A2 

Fc/2 

+FA2 

[ 51 ]

Mapping IgG Subunit Glycoforms Using HILIC and a Wide-Pore Amide Stationary Phase

In this testing, several additional chromatographic parameters were also monitored, including the retention 
time of the Fc/2+FA2 species, the maximum system pressure observed during the chromatographic run, and 
the percent (%) carryover of the most abundant glycoform, the Fc/2+FA2 species (Figure 5B). Plots of these 
parameters underscore the consistency of the subunit separation across the lifetime of the column. With 
noteworthy consistency, the column produced relatively stable retention times, a consistent resolution of the 
A2 and FA2 glycoforms (Rs≈2), a maximum system pressure consistently at only ~6 Kpsi, and a remarkably 
low carryover between 0.1 and 0.2%. This latter aspect of the HILIC separations is particularly noteworthy 
since it indicates that carryover with these methods is almost an order of magnitude lower than analogous  
C4 based RP methods. 

Benchmarking the capabilities of the glycoprotein BEH amide, 300Å, 1.7 μm column
We have benchmarked the performance of this new wide-pore column technology against not only its standard 
pore diameter analog but also its two most closely related, commercially-available alternatives. Figure 6 
presents chromatograms obtained for a reduced, IdeS digested sample of an IgG1K mAb acquired from NIST 
using these different column technologies. In a visual comparison, it is clear that the glycoprotein BEH amide 
300Å column significantly outperforms the other three columns. To quantify this assessment, peak-to-
valley ratios were calculated for the separation of the FA2 glycoform away from the FA2G1 glycoform. The 
glycoprotein BEH amide 300Å column was found to demonstrate improvements of 48%, 152%, and 261% 
over the 130Å glycan BEH amide, the Accucore amide, and AdvanceBio glycan mapping columns. This mAb 
sample also has a particularly interesting attribute in that it has a reasonably high relative abundance of an 
immunogenic alpha-1,3-galactose containing glycan (an FA2G2Ga1 structure).16-17 As shown in Figure 6, this 
Fc/2+FA2G2Ga1 species can be readily visualized with the wide-pore amide column. This represents a sizeable 
improvement in the peak capacity of large molecule HILIC separations for this emerging application.

0E+0

2E+8

7.5

12.5

In

te

ns

it

y

7.0

12.0 6.5

11.5 9.5

14.5

Glycoprotein BEH Amide,  

300Å, 1.7  m  

Glycan BEH Amide,  

130Å, 1.7  m  

Accucore Amide,  

150Å, 2.6  m  

AdvanceBio  

Glycan Mapping, 

1.8 µm 

Fc/2 

+ FA2 

Fc/2 

+ FA2G1 

p/v 

8.3 

p/v 

5.6 

p/v 

3.3 

p/v 

2.3 

Fc/2 

+ FA2G2Ga1 

Figure 6. Subunit glycoform profiles of an IgG1K obtained with various 2.1 x 150 mm columns: ACQUITY UPLC Glycoprotein BEH 
Amide, 300Å, 1.7 µm Column, ACQUITY UPLC Glycan BEH Amide, 130Å, 1.7 µm Column, Thermo Scientific Accucore Amide,  
150Å, 2.6 µm column, and Agilent AdvanceBio Glycan Mapping, 1.8 µm column. Peak-to-valley (p/v) ratios for the Fc/2+FA2 
versus FA2G1 glycoforms are provided. An alpha gal containing Fc/2+FA2G2Ga1 is readily visualized with the glycoprotein BEH 
amide, 300Å, 1.7 µm column.

[ 52 ] Mapping IgG Subunit Glycoforms Using HILIC and a Wide-Pore Amide Stationary Phase

Figure 7. HILIC Profiling of  
cetuximab glycosylation. (A) 
HILIC-UV chromatogram of reduced, 
IdeS/carboxy peptidase B-digested 
cetuximab obtained using an ACQUITY 
UPLC Glycoprotein BEH Amide, 300Å,  
1.7 µm, 2.1 x 150 mm Column. 
Species corresponding to Fc/2 and 
Fd' subunits are labeled in gray and 
red, respectively. Subunit glycan 
assignments based on deconvoluted 
mass spectra are provided. (B) HILIC- 
fluorescence chromatograms of 
RapiFluor-MS labeled N-glycans  
from cetuximab obtained using an  
ACQUITY UPLC Glycan BEH Amide,  
130Å, 1.7 µm, 2.1 x 50 mm Column.  
Mass spectral data supporting the 
assignments of the RapiFluor-MS 
labeled N-glycans are provided. 

0.01

0.02

0.03

0.04

0.05

0.06

8

9

10

11

12

13

14

15

16

17

18

A

21

4

Time (min)

Fc/2 

+FA2 

Fc/2  

+FA2G2 

Fc/2 

Glycosylated 

Fc/2 

+M5 

Fd’   pE 

+ (FA2G2Ga2) 

Fd’   pE 

+ (FA2G2Ga1Sg1) 

Fd’   pE 

+ (Hex9HexNAc5DHex1) 

Fd’  

Glycosylated 

N-term pE

Fd’   pE 

+ (FA2G2Ga1) 

Fd’   pE 

+ (FA2G2Sg1) 

Fc/2 

+FA2G1 

0E+0

4E+6

4

5

6

7

8

9

10

11

12

EU

Time (min)

M5 

FA2 

FA2G1 

FA2G2 

Hex9HexNAc5DHex1 

FA2G2Ga1 

FA2G2Ga2 

FA2G2Ga1Sg1 

FA2G2Sg1 

Domain-Specific  

Glycan  

Information 

High Resolution  

High Sensitivity  

Released  

N-Glycan Profile 

A 

B 

Species 

MW

Avg 

Theoretical 

(Da) 

MWAvg

Observed 

(Da) 

Mass 

Error 

(Da) 

LC 

23427.0 

23427.1 

0.1 

Fc/2-K  

+ FA2 

25236.3 

25237.4 

1.1 

Fc/2-K  

+ M5 

25008.1 

25008.8 

0.7 

Fc/2-K  

+ FA2G1 

25398.5 

25399.8 

1.3 

Fc/2-K  

+ FA2G2 

25560.6 

25562.0 

1.4 

Fd' pE  

+ FA2G2Ga1 

27385.5 

27386.8 

1.3 

Fd' pE  

+ FA2G2Sg1 

27530.6 

27531.8 

1.2 

Fd' pE  

+ FA2G2Ga2 

27547.6 

27548.2 

0.6 

Fd' pE + 

FA2G2Ga1Sg1 

27692.7 

27693.1 

0.4 

Fd' pE + 

Hex9HexNAc5DHex1 

28075.1 

28075.3 

0.2 

Species 

MW

Mono 

Theoretical 

(Da) 

MW

Mono 

Observed 

(Da) 

Mass 

Error 

(ppm) 

FA2 

1545.6080 

1545.6136 

3.6 

M5 

1773.7190 

1773.7242 

2.9 

 FA2G1 

1935.7719 

1935.7834 

5.9 

FA2G2 

2097.8247 

2097.8136 

-5.3 

FA2G2Ga1 

2259.8775 

2259.8860 

3.8 

FA2G2Sg1 

2404.9150 

2404.9150 

0.0 

FA2G2Ga2 

2421.9303 

2421.9320 

0.7 

FA2G2Ga1Sg1 

2566.9678 

2566.9792 

4.4 

Hex9HexNAc5DHex1  2949.1154 

2949.1424 

9.2 

Complementing RapiFluor-MS N-glycan analyses with 
domain specific information about mAb glycosylation
One of the key advantages to profiling IgG subunits by HILIC 
is being able to elucidate domain specific information about 
glycosylation. In an IgG structure, there exists one conserved 
N-glycosylation site at Asn297 of the heavy chain. As a 
consequence, most IgGs will be modified with two glycans in 
the CH2 domains (constant heavy chain 2 domains) of the Fc 
subunit. However, it is estimated that 20% of human IgGs are also 
modified in their CH1 domains, which reside in the Fab subunits, 
and more specifically the IdeS generated Fd' subunit.18-19 For 
example, it is known that cetuximab, a chimeric mAb expressed 
from a murine cell line, is glycosylated in both its CH1 and 
CH2 domains.20 Characterization of this mAb has thus proven 
to be an interesting case study for the application of our newly 
developed techniques. HILIC separations obtained for a reduced, 
IdeS digested sample of carboxypeptidase B treated cetuximab 
showed only one weakly retained subunit species, which could 
be easily assigned to the LC subunit by online ESI-MS (data not 
shown). Furthermore, and as shown in Figure 7A, the glycoform 
retention window for cetuximab was populated with twice as 
many peaks as had been observed for trastuzumab and its 

glycosylated Fc/2 subunit. Deconvoluted ESI-MS data from these 
HILIC-MS separations confirmed that the first grouping of peaks 
(labeled in gray) corresponded to Fc/2 glycoforms and typical mAb 
glycan species, such as FA2, FA2G1, M5, and FA2G2. Meanwhile, 
the second grouping of peaks were found to be distinctively 
related to glycoforms of the Fd' subunit given their unique masses. 
Curiously, each of the identified Fd’ glycoforms (labeled in red) are 
immunogenic in nature, containing either non-human alpha-1,3-
galactose or non-human N-glycolyl-neuraminic acid epitopes.21 

The identification of these glycan species has been confirmed 
through released N-glycan analyses. Using the newly developed 
GlycoWorks RapiFluor-MS N-Glycan Kit,22 cetuximab N-glycans 
were rapidly prepared and labeled with the novel fluorescence and 
MS-active labeling reagent, RapiFluor-MS. The resulting labeled 
N-glycans were subsequently separated using a glycan BEH amide, 
130Å, 1.7 μm column and detected by fluorescence and positive 
ion mode ESI-MS, as portrayed in Figure 7B. The sensitivity gains 
afforded by the RapiFluor-MS label facilitated making confident 
assignments of the released N-glycan structures. The species 
that have been assigned as a result of both this released glycan 
analysis as well as the subunit HILIC-UV-MS method are supported 
by previous reports on cetuximab glycosylation.6,20 

[ 53 ]

Mapping IgG Subunit Glycoforms Using HILIC and a Wide-Pore Amide Stationary Phase

With the combination of released glycan and subunit-derived glycan 
information, cetuximab glycosylation has been characterized with significant 
detail. With the RapiFluor-MS released glycan analysis, a very high resolution 
separation has been achieved with an LC-MS compatible method in which 
glycans can even be subjected to detailed MS/MS analyses. With an equally 
MS-compatible subunit HILIC separation, domain-specific glycan information 
has been readily obtained with minimal sample preparation. Each method has 
therefore provided complementary information on the glycosylation of the 
mAb. Nevertheless, the widepore amide HILIC method stands out as a useful 
technique for rapidly screening mAbs for multidomain glycosylation.

C O N C LU S IO NS
Subunit analyses of mAbs represent a useful strategy for rapidly investigating 
domain-specific modifications. The combination of high fidelity IdeS proteolysis 
with high resolution LC-UV-MS has presented a new approach to mAb identity 
testing and assaying oxidation.3 The current subunit mapping strategies have 
exclusively relied upon reverse phase chromatography. However, since N-linked 
glycosylation of IgG proteins elicit dramatic changes in hydrophilicity and 
hydrogen bonding characteristics, a separation by hydrophilic interaction 
chromatography (HILIC) can be effectively used for this application or as a 
complementary method to reversed-phase separations since the same mobile 
phases can be employed. For this reason, we have proposed the use of HILIC with 
an amide bonded stationary phase that has been optimized for large molecule 
separations, the wide-pore glycoprotein BEH amide, 300Å, 1.7 μm stationary 
phase. Along with new developments in released N-glycan analysis afforded 
by RapiFluor-MS,22 the glycoprotein BEH amide, 300Å, 1.7 μm column enables 
new possibilities for routine monitoring and detailed characterization of mAb 
glycosylation, including elucidation of domain-specific glycan information.

[ 54 ] Mapping IgG Subunit Glycoforms Using HILIC and a Wide-Pore Amide Stationary Phase

Waters Corporation 
34 Maple Street 
Milford, MA 01757 U.S.A. 
T: 1 508 478 2000 
F: 1 508 872 1990 
www.waters.com

Waters, The Science of What’s Possible, ACQUITY UPLC, Oasis, MassLynx, SYNAPT, Xevo, and Empower are registered trademarks  
of Waters Corporation. GlycoWorks and RapiFluor-MS are trademarks of Waters Corporation. All other trademarks are the property  
of their respective owners.

©2015 Waters Corporation. Produced in the U.S.A.  April 2015  720005385EN  AG-PDF

References

1.  Aggarwal, R. S., What’s fueling the biotech engine – 2012 to 2013.  

Nat Biotechnol 2014, 32 (1), 32–9.

2.  Wang, B.; Gucinski, A. C.; Keire, D. A.; Buhse, L. F.; Boyne, M. T., 2nd, 

Structural comparison of two anti-CD20 monoclonal antibody drug products 
using middle-down mass spectrometry. Analyst 2013, 138 (10), 3058–65.

3.  Gucinski, A. C., Rapid Characterization and Comparison of Stressed 

anti-CD20 Drugs using Middle Down Mass Spectrometry. In 61st ASMS 
Conference on Mass Spectrometry and Allied Topics, Minneapolis, MN, 2013.

4.  Ayoub, D.; Jabs, W.; Resemann, A.; Evers, W.; Evans, C.; Main, L.; Baessmann, 

C.; Wagner-Rousset, E.; Suckau, D.; Beck, A., Correct primary structure 
assessment and extensive glyco-profiling of cetuximab by a combination 
of intact, middle-up, middle-down and bottom-up ESI and MALDI mass 
spectrometry techniques. MAbs 2013, 5 (5), 699–710.

5.  Alain Beck; Hélène Diemer; Daniel Ayoub; François Debaene; Elsa Wagner-

Rousset; Christine Carapito; Alain Van Dorsselaer; Sanglier-Cianférani, S., 
Analytical characterization of biosimilar antibodies and Fc-fusion proteins. 
Trends in Analytical Chemistry 2013, 48, 81–95.

6.  Janin-Bussat, M. C.; Tonini, L.; Huillet, C.; Colas, O.; Klinguer-Hamour, C.; 

Corvaia, N.; Beck, A., Cetuximab Fab and Fc N-glycan fast characterization 
using IdeS digestion and liquid chromatography coupled to electrospray 
ionization mass spectrometry. Methods Mol Biol 2013, 988, 93–113.

7.  Ahn, J.; Bones, J.; Yu, Y. Q.; Rudd, P. M.; Gilar, M., Separation of 

2-aminobenzamide labeled glycans using hydrophilic interaction 
chromatography columns packed with 1.7 microm sorbent. J Chromatogr B 
Analyt Technol Biomed Life Sci 2010, 878 (3–4), 403–8.

8.  Gustavsson, P.-E.; Larsson, P.-O., Support Materials for Affinity 

Chromatography. In Handbook of Affinity Chromatography, Hage,  
D., Ed. Taylor & Francis: Boca Raton, FL, 2006; pp 15–33.

9.  Renkin, E. M., J. Gen. Physio. 1954, (38), 225.

10. Tetaz, T.; Detzner, S.; Friedlein, A.; Molitor, B.; Mary, J. L., Hydrophilic 

interaction chromatography of intact, soluble proteins. J Chromatogr A 
2011, 1218 (35), 5892–6.

11. Beck, A.; Sanglier-Cianferani, S.; Van Dorsselaer, A., Biosimilar, biobetter,  

and next generation antibody characterization by mass spectrometry.  
Anal Chem 2012, 84 (11), 4637–46.

12. Lauber, M. A.; Koza, S. M., Waters Application Note ****to be updated 2015.

13. Xie, H.; Chakraborty, A.; Ahn, J.; Yu, Y. Q.; Dakshinamoorthy, D. P.; Gilar, 

M.; Chen, W.; Skilton, S. J.; Mazzeo, J. R., Rapid comparison of a candidate 
biosimilar to an innovator monoclonal antibody with advanced liquid 
chromatography and mass spectrometry technologies. MAbs 2010, 2 (4).

14. Yu, Y. Q.; Ahn, J.; Gilar, M., Trastuzumab Glycan Batch-to-Batch Profiling 

using a UPLC/FLR/MS Mass Spectrometry Platform. Waters Application Note 
720003576en.

15. Raju, T. S.; Jordan, R. E., Galactosylation variations in marketed therapeutic 

antibodies.  
MAbs 2012, 4 (3), 385–91.

16. Schiel, J.; Wang, M.; Formolo, T.; Kilpatrick, L.; Lowenthal, M.; Stockmann, 

H.; Phinney, K.; Prien, J. M.; Davis, D.; Borisov, O. In Biopharmaceutical 
Characterization: Evaluation of the NIST Monoclonal Antibody Reference 
Material, 62nd Conference on Mass Spectrometry and Allied Topics, 
Baltimore, MD, Baltimore, MD, 2014.

17. Bosques, C. J.; Collins, B. E.; Meador, J. W., 3rd; Sarvaiya, H.; Murphy, J. L.; 

Dellorusso, G.; Bulik, D. A.; Hsu, I. H.; Washburn, N.; Sipsey, S. F.; Myette, 
J. R.; Raman, R.; Shriver, Z.; Sasisekharan, R.; Venkataraman, G., Chinese 
hamster ovary cells can produce galactose-alpha-1,3-galactose antigens on 
proteins. Nat Biotechnol 2010, 28 (11), 1153–6.

18. Jefferis, R., Glycosylation of Recombinant Antibody Therapeutics.  

Biotechnol Prog 2005, (21), 11–16.

19. Huang, L.; Biolsi, S.; Bales, K. R.; Kuchibhotla, U., Impact of variable  

domain glycosylation on antibody clearance: an LC/MS characterization.  
Anal Biochem 2006, 349 (2), 197–207.

20. Qian, J.; Liu, T.; Yang, L.; Daus, A.; Crowley, R.; Zhou, Q., Structural 

characterization of N-linked oligosaccharides on monoclonal antibody 
cetuximab by the combination of orthogonal matrix-assisted laser  
desorption/ionization hybrid quadrupole-quadrupole time-of-flight tandem 
mass spectrometry and sequential enzymatic digestion. Anal Biochem 2007, 
364 (1), 8–18.

21. Arnold, D. F.; Misbah, S. A., Cetuximab-induced anaphylaxis and IgE specific 

for galactose-alpha-1,3-galactose. N Engl J Med 2008, 358 (25), 2735; 
author reply 2735–6.

22. Lauber, M. A.; Brousmiche, D. W.; Hua, Z.; Koza, S. M.; Guthrie, E.; Magnelli, 

P.; Taron, C. H.; Fountain, K. J., Rapid Preparation of Released N-Glycans for 
HILIC Analysis Using a Novel Fluorescence and MS-Active Labeling Reagent. 
Waters Application Note 720005275EN 2015.

[ 55 ]

Mapping IgG Subunit Glycoforms Using HILIC and a Wide-Pore Amide Stationary Phase

WAT E RS SO LU T IONS
Biopharmaceutical Platform Solution  
with UNIFI®

ACQUITY UPLC® H-Class Bio System

Xevo® G2-S QTof

ACQUITY UPLC Tunable UV Detector

UNIFI Scientific Information System

ACQUITY UPLC Protein BEH C4 Column

K E Y WO R DS
Biosimilar, intact mass analysis, intact 
mass subunit analysis, light chain, 
heavy chain, glycosylation, glycoprofile, 
infliximab, mAb, biotherapeutic 
characterization

A P P LIC AT ION BEN E FIT S 
A streamlined workflow within an integrated 
UPLC®-MS/MS system solution that features 
automated data acquisition, processing, and 
reporting, and that is deployable to both 
regulated and non-regulated laboratories, 
enables efficient structural analysis and 
comparative analysis of multiple batches of an 
innovator product and its biosimilar candidate 
within a comparability study.

INT RO DUC T ION
The expiration of patents and other intellectual property rights for originator 
biologics over the next decade opens up ample opportunities for biosimilars to 
enter the market and push industry competition to a high level.1-4 

Compared to small molecule drugs, biopharmaceuticals have much more 
complex structures and are more expensive to develop. The complexity of the 
biopharmaceutical molecular entity puts greater challenges on organizations 
seeking to manufacture safe and effective biosimilar products for patients. 
Regulatory bodies such as the U.S. FDA and EMA5-8 require a demonstration of 
comprehensive characterization for the drug substance: Confirming primary 
sequence and identifying post-translational modifications (PTMs), establishing 
biophysical and functional comparability for the innovator and candidate 
biosimilar, and performing studies that establish expected variation within an 
innovator biotherapeutic.

Infliximab (Remicade) is a monoclonal antibody (mAb) used to treat autoimmune 
diseases; it was first approved by the FDA for the treatment of Crohn’s disease in 
1998, and in 2013 two biosimilars have been submitted for approval in Europe.

In this application note, we characterize infliximab and a biosimilar candidate, 
produced in a different cell line, using Waters® Biopharmaceutical Platform 
Solution with UNIFI Scientific Information System. The objective is to screen 
multiple lots of both the innovator and biosimilar products at the subunit level 
(light chain (LC) and heavy chain (HC)) to establish comparability at this higher 
level of structure. Lot-to-lot and batch-to-batch comparisons will show product 
variation, illustrating the range of quality attributes to be considered in a 
candidate biosimilar. 

Structural Comparison of Infliximab and a Biosimilar via Subunit Analysis 

Using the Waters Biopharmaceutical Platform with UNIFI
Henry Shion and Weibin Chen

Waters Corporation, Milford, MA, USA

[ 56 ]

E X P E RIM ENTA L 
Biopharmaceutical Platform Solution with UNIFI

■

■

ACQUITY UPLC H-Class Bio System

■

■

Xevo G2-S QTof

■

■

ACQUITY UPLC Tunable UV Detector

■

■

UNIFI Scientific Information System 

Intact protein LC-MS conditions
Column:  

ACQUITY UPLC Protein BEH C4 Column, 
300Å, 1.7 µm, 2.1 mm x 50 mm  
(p/n 186004495)

Column temp.:  

80 °C

Mobile phase A:  

Water

Mobile phase B:  

Acetonitrile

Mobile phase C:  

Not used

Mobile phase D:  

0.5% TFA (in water)

Detection:   

UV 280 nm

LC gradient table:

 Time 

Flow 

 (min) (mL/min)  %A 

%B 

%C 

%D 

Curve

 Initial 

0.20 

65.2 

29.8 

0 

5.0 

Initial 

 12.0 

0.20 

63.5 

31.5 

0 

5.0 

6 

 14.0 

0.20 

63.5 

31.5 

0 

5.0 

6 

 14.1 

0.20 

10.0  85.0  0 

5.0 

6 

 15.1 

0.20 

10.0  85.0  0 

5.0 

6 

 15.2 

0.20 

65.2  29.8 

0 

5.0 

6 

 18.0 

0.20 

65.2  29.8 

0 

5.0 

6

Total run time: 20.0 min

MS conditions
Capillary:  

3.0 kV 

Sampling cone:  

80 V

Extraction cone:  

4 V

Source temp.:  

125 °C

Desolvation temp.:  

350 °C

Cone gas flow:  

0 L/Hr 

Desolvation gas flow:   800 L/Hr

Data acquisition and processing
MaxEnt™ 1 for MS spectra deconvolution

UNIFI Scientific Information System

Sample preparation
Three batches of innovator infliximab were acquired from  
Jenssen Biotech, Inc. (Horsham, PA, USA). The batches were 
produced by the SP2/0 mouse cell line. Three batches of  
candidate biosimilar infliximab produced by an alternative 
mammalian cell line (Chinese hamster ovary (CHO) were  
obtained from a third-party collaborator. All of the samples  
were stored at -80 °C before analysis.

A reduction buffer solution containing 25 mM NaCl, 25 mM Tris,  
1 mM EDTA (pH 8.0) was made to prepare mAb subunits. For  
each of the six batches, 10 μL of formulated mAb solution  
(21 mg/mL, the commonly used concentration level for patient 
injection) was mixed with 180 μL of reduction buffer in a  
1.5 mL Eppendorf tube for a protein concentration of 1.0 mg/mL. 
A concentrated dithiothreitol (DTT) solution (100 mM in H2O) was 
then added to each solution to obtain a final DTT concentration  
of 1.0 mM. The samples were incubated at 37 °C for 20 minutes. 
The samples were briefly centrifuged, then 105 μL of each sample 
was mixed with an equal volume of aqueous solution containing 
3% acetonitrile and 0.1% formic acid. The final concentration of 
the mAb was about 0.5 mg/mL. Triplicate injections of each sample 
were made onto an ACQUITY UPLC Protein BEH C4 Column,  
300Å, 1.7 µm, 2.1 mm x 50 mm (p/n 186004495)  LC-MS 
analysis of the mAb subunit.

[ 57 ]

Structural Comparison of Infliximab and a Biosimilar via Subunit Analysis Using the Waters Biopharmaceutical Platform with UNIFI

R E SU LT S AND DIS CUSSION

Figure 1. Reversed-phase (C4) chromatograms of the innovator infliximab (top) and a biosimilar 

infliximab (bottom). The signal trace is the Total Ion Current (TIC) from the mass spectrometer.

Figure 2. The combined raw MS spectra and deconvoluted spectra in mirror image plots. The MS 
spectra of the light chain (eluting around 3.5 min in Figure 1) from the innovator sample (top)  
and the biosimilar sample (bottom) are displayed.

Subunit characterization for Infliximab  
from two cell lines
Figure 1 shows the reversed-phase LC-MS 
chromatograms from the analysis of reduced 
infliximab from both the innovator and biosimilar 
products. There are two major components to 
each chromatogram, a peak at ~3.5 minutes 
and a complex set of peaks at ~10 minutes. The 
chromatographic peaks eluting around 3.5 minutes 
have ESI-MS measurements of 23434.0 Da, 
respectively, in full agreement with the calculated 
mass of the light chain of infliximab (23434.0 Da). 
The complex peak eluting at ~10 minutes  
is comprised of several species with MW in  
the 51,000 Da range, corresponding to the  
glycosylated heavy chain. 

Figure 2 shows the comparison of the light chain 
spectra in a mirror plot using UNIFI Scientific 
Information System’s software, with the combined 
raw MS spectra shown on the left panel (as 
demonstrated by multiple charged spectrum 
envelopes) and the MaxEnt 1 deconvoluted spectra 
displayed on the right. 

The results indicate that there is only one isoform 
and no noticeable difference in the light chains 
between the innovator and biosimilar samples.  
This observation is consistent with other IgG1 
biosimilar studies9 that show little or no  
post-translational modifications of LC subunits. 

The chromatographic profile of the heavy chains of 
infliximab was more complicated than that of the 
light chain. A cluster of peaks is observed around  
10 minutes in Figure 1, corresponding to different 
isoforms of the heavy chains, and they exhibit 
significantly different chromatographic behavior 
from that of the light chain. Similarly, as shown 
by Figure 1 and 3, the heavy chains of infliximab 
from the two cell lines show quite distinct 
chromatographic and spectral differences. 

[ 58 ] Structural Comparison of Infliximab and a Biosimilar via Subunit Analysis Using the Waters Biopharmaceutical Platform with UNIFI

The two major chromatographic peaks (at 9.5 min and 10.5 min) from the analysis of reduced innovator 
infliximab (Figure 1A) come from the heavy chains and appear to have multiple isoforms (Figure 3C).  
Mass spectrometry analysis of these peaks (Figure 3) shows that variation in both the polypeptide sequence 
(+/- lysine) and glycosylation contribute to the heterogeneity of the innovator HC.

This is in contrast to the biosimilar sample, which displays a more homogeneous peak at 9.8 minutes  
(Figure 1B) and fewer mass variants (Figure 3D). 

Several major glycoforms (e.g. G0, G0F, G1F, G2F, and Man5) were identified for the innovator heavy chain as 
shown in Figure 4, demonstrating a high degree of heterogeneity of the innovator infliximab. The biosimilar has 
three major glycoforms (G0, G0F, and G1F) and no apparent amino acid variations. All of the MS peaks in the 
deconvoluted spectrum can be automatically identified in UNIFI Scientific Information System’s software based 
on the mAb’s reported sequence and the suspected PTM, and annotated, as displayed in Figure 4.

Figure 3. The combined raw MS 
spectra mirror image comparison 
between the heavy chains (eluted 
around 10 min in Figure 1) for the 
innovator and the biosimilar.

Figure 4. The innovator heavy chain 
deconvoluted spectra reveals the 
presence of glycoforms G0, G0F, 
G1F, G2F, and Man5 as well as lysine 
variations. Incomplete removal of 
lysine from the C-terminus of the HC 
is a known variant for IgG1.

[ 59 ]

Structural Comparison of Infliximab and a Biosimilar via Subunit Analysis Using the Waters Biopharmaceutical Platform with UNIFI

Assessment of batch-to-batch variability
The analysis of reduced IgG is a straightforward, high-sensitivity method that provides valuable information 
on the identity and amount of related variants of mAb structure. Analysis of the reduced infliximab indicates 
that its structural heterogeneity resides within the heavy chain of the antibody, and includes variation in both 
glycosylation and amino acid sequence. The incomplete removal of C-terminal lysine residues is a known 
structural variant, so it can be surmised that this PTM is occurring in the innovator infliximab. 

As demonstrated by the spectra of the HC (Figure 4), the biantennary oligosaccharides G0F, G1F, and G2F, along 
with smaller amounts of the high mannose forms, are the major glycoforms of infliximab. Since there is only 
one N-glycosylation site on the HC, the intensity of peaks for the various oligosaccharide structures can be 
used to quantify the relative abundance of the various glycoforms. The MaxEnt 1 algorithm used for generating 
the deconvoluted spectra preserves the intensity information from the raw spectra, for quantitative assessment 
of structural variation. 

This measurement establishes a foundation upon which structural comparison for multiple batches of infliximab 
can be performed, thus making the analysis at the subunit level an attractive approach to establishing 
development requirements for biosimilars. 

On the basis of the analysis of reduced infliximab subunits, we compared the structure differences among 
multiple batches of infliximab from the two cell lines. Regulatory guidelines for biosimilar development 
recommend that any analytical characterization first establish the structural variation range of the reference 
product. As such, analysis of multiple lots of reference products (infliximab from SP2/0 cells) as well as 
biosimilar products (inflixmab from CHO cells) is necessary to establish the range of values for critical 
structural features. In the meantime, replicate analysis is also performed for each sample to demonstrate the 
reproducibility of the LC-MS method itself. 

The analysis of multiple samples in triplicate helps establish a vigorous analytical procedure to provide 
sound analytical support for biosimilar development. However, this approach generates a high volume of data 
that requires efficient informatics tools to process data and produce meaningful results. The UNIFI Scientific 
Information System automatically acquires and processes the data and generates reports on the results, 
demonstrating the great power and flexibility available for such data analysis tasks. 

Next, we demonstrate how UNIFI Scientific Information System’s software can be utilized to streamline the 
structural comparison of reduced infliximab from two cell lines.

Structural comparison
Figure 5A displays the MS response summary plot for glycoform G0 in percentage. This UNIFI Scientific 
Information System plot offers a simple and direct view to demonstrate the variation in relative abundance 
of the G0 glycoform across the injections of innovator and biosimilar batches. This functionality removes the 
scientific and compliance burden of summarizing reports of such data in Excel or other data analysis tools that 
are not core features of the instrument’s software. By including both automated processing statistical reporting 
within UNIFI, the software also prevents human transcription errors that may require significant time and effort 
to identify and correct. Similar plots can be readily generated within UNIFI Scientific Information System for 
other glycoforms identified in the analysis, such as G1F and Man5, as shown in Figure 5B and 5C. 

[ 60 ] Structural Comparison of Infliximab and a Biosimilar via Subunit Analysis Using the Waters Biopharmaceutical Platform with UNIFI

The triplicate analysis for each sample shows a highly reproducible measurement. There is some minor batch-
to-batch variability, notably in the abundance of G1F in the innovator (5B) as well as the Man5 content in the 
biosimilar (5C). On the other hand, it appears that the biosimilar, produced in CHO cells, has approximately 10 
times more non-fucosylated G0 glycoform compared to that of the innovator (SP2/0 cell line) product. It is also 
observed that there are about twice as many G1F glycoforms (by percentage) in the biosimilar batches than in 
the innovator, and there is about 30% more Man5 glycoform in the innovator batches than in the biosimilar 
sample batches.

Figure 5. Relative abundance of the G0 (A), G1F (B), and  Man5 (C) glycoforms in infliximab HC, from all the injections of the innovator (left, blue) and biosimilar  
(right, Red) batches.

As this example shows, the glycoforms of infliximab from two cell lines can be readily analyzed and 
information on the glycosylation variation can be quickly obtained via UNIFI Scientific Information System 
software’s automated workflow covering data acquisition, processing, and reporting. Additionally, the 
workflow can be deployed in both non-regulated and regulated environments, so a common analytical 
platform can be employed and consistent information acquired across the entire development process.

[ 61 ]

Structural Comparison of Infliximab and a Biosimilar via Subunit Analysis Using the Waters Biopharmaceutical Platform with UNIFI

Another major source of HC heterogeneity is lysine variants. Depending on the cell line and other production 
conditions, a lysine residue may remain on the C-terminus of the polypeptide chain. Figure 6 displays the 
percentage of clipped-lysine variants, automatically calculated in UNIFI Scientific Information System 
software, for both the innovator and biosimilar batches. As can be seen, the percentage was much smaller 
for the biosimilar sample batches (from CHO cell line) as compared to that observed in the innovator (SP2/0 
cell line) batches. This experimental result confirms that there was a much lower level of C-terminal lysine in 
antibodies derived from the CHO cell line, and the lysine content is more consistent from batch-to-batch. The 
innovator infliximab has a lower overall abundance for variants with the complete removal of lysine, and the 
amount does vary from batch to batch.

Figure 6. The percentage of clipped-lysine (0K) variants, automatically calculated in UNIFI Scientific Information System, is shown for 
the innovator (left, blue) and biosimilar (right, red) batches.

[ 62 ] Structural Comparison of Infliximab and a Biosimilar via Subunit Analysis Using the Waters Biopharmaceutical Platform with UNIFI

Waters Corporation 
34 Maple Street 
Milford, MA 01757 U.S.A. 
T: 1 508 478 2000 
F: 1 508 872 1990 
www.waters.com

Waters, The Science of What’s Possible, ACQUITY UPLC, UPLC, Xevo, and UNIFI are registered trademarks of Waters Corporation.  
MaxEnt is a trademark of Waters Corporation. All other trademarks are the property of their respective owners.

©2013–2016 Waters Corporation. Produced in the U.S.A.  January 2016 720004796EN  AG-PDF

CONC LUSIONS
In this work, the extent of comparability was established between multiple batches 
of innovator and candidate biosimilar infliximab, using an integrated analytical 
platform with capabilities for automated data processing and reporting. The 
Biopharmaceutical Platform Solution with UNIFI Scientific Information System 
was applied to study these samples at the level of reduced heavy and light chain 
subunits, and to report on several biotherapeutic structural differences between 
these preparations. 

Overall, the innovator molecule exhibited more heterogeneity with respect to 
PTM’s (glycosylation and C-terminal lysine) compared to the candidate biosimilar. 
Potentially significant differences were found between the innovator and the 
biosimilar samples, particularly in regard to the presence of fucosylated glycans.  
We found that the biosimilar had a much higher abundance of the non-fucosylated 
glycoform G0, and less of the fucosylated G1F, in comparison to the innovator. 
Some batch-to-batch variability was observed among both the innovator batches 
and the biosimilar batches. 

The power to universally deploy high resolution analytics to address these 
important questions, combined with the ability to quickly communicate these 
results, enables organizations to make rapid and confident decisions in the race  
to market with safe and effective innovator and biosimilar therapeutics.

References

1.  Biosimilars and Follow-On Biologics Report: The Global 

Outlook 2010–2025, Visiongain Ltd. 2010.

2.  Lawrence, S. Billion dollar babies: biotech drugs as 

blockbusters. Nature Biotech. 2007; 25, 380–382. 

3.  Pharmaceutical Research & Manufacturers of America. 

Medicines in Development: Biologics. 2013. http://www.
phrma.org/sites/default/files/pdf/biologics2013.pdf

4.  Erickson, BE. Untangling biosimilars. Chem. Eng. News, 

2010; 88, 25–27.

5.  Guidance for Industry, Quality Considerations in 

Demonstrating Biosimilarity to a Reference Protein Product, 
FDA website [online]. 2012. http://www.fda.gov/downloads/
Drugs/GuidanceComplianceRegulatoryInformation/
Guidances/UCM291134.pdf 

6.  European Medicines Agency. Guideline on similar biological 

medicinal products containing biotechnology-derived 
proteins as an active substance: quality issues. EMA 
website [online]. 2011. http://www.ema. europa.eu/docs/
en_GB/document_library/Scientific_ guideline/2009/09/
WC500003953.pdf 

7.  Chakraborty A. Chen W. Gebler J. Characterization of reduced 

monoclonal antibody by on-line UPLC-UV/ESI-TOF MS, 
Waters Application note. 2009; 720002919en. 

8.  European Medicines Agency, European Medicines Agency 

recommends approval of first two monoclonal-antibody 
biosimilar. [online]. 06/2013. http://www.ema.europa.eu/
docs/en_GB/document_library/Press_release/2013/06/
WC500144941.pdf

9.  Xie H. Chakraborty A. Ahn J. Yu Y. Dakshinamoorthy D. Gilar 

M. Chen W. Skilton SJ. Mazzeo JR. Rapid comparison of a 
candidate biosimilar to an innovator monoclonal antibody 
with advanced liquid chromatography and mass spectrometry 
technologies, MAbs. 2010; Jul–Aug; 2(4): 379–394.

[ 63 ]

Structural Comparison of Infliximab and a Biosimilar via Subunit Analysis Using the Waters Biopharmaceutical Platform with UNIFI

RELEASED N-GLYCANS

[ 64 ]

WAT E R S  S O LU T IO NS
GlycoWorks™ RapiFluor-MS™  
N-Glycan Kit

GlycoWorks HILIC µElution Plate

ACQUITY UPLC® Glycan BEH Amide 130Å 
Column

ACQUITY UPLC H-Class Bio System

ACQUITY® QDa® Mass Detector

Xevo® G2-XS QTof MS

SYNAPT® G2-S HDMS

K E Y W O R D S
GlycoWorks, RapiFluor-MS,  RapiGest™ SF,  
Rapid Tagging, PNGase F, Deglycosylation, 
ACQUITY UPLC H-Class Bio System, BEH 
Amide 130Å, Glycans, Glycoproteins, 
Glycosylation, HILIC, Fluorescence

A P P L I C AT IO N  B E N E F I T S

■

■

Preparation of labeled N-glycans  
(from glycoprotein to analysis ready sample)  
in 30 minutes

■

■

Complete deglycosylation to produce 
unbiased results

■

■

Simple, streamlined protocol provided with 
the GlycoWorks RapiFluor-MS N-Glycan Kit

■

■

Unprecedented sensitivity for labeled 
N-glycans with at least 2 and 100 fold 
increases to fluorescence and MS  
detection, respectively

■

■

Accurate profiling based on robust SPE  
for neutral to tetrasialylated N-glycans

I N T RO DU C T IO N
The N-glycan profile of a biopharmaceutical is commonly defined as a critical 
quality attribute, since it can be a measure of efficacy, safety, and manufacturing 
conditions.1-2 Therefore, it is important that approaches for the glycan analysis of 
clinical and commercial biotherapeutic formulations exhibit high sensitivity and 
facilitate detailed characterization. Additionally, it would be highly advantageous 
if such an analysis could also be performed with rapid turnaround times and high 
throughput capacity to expedite product development. Most analytical strategies 
for evaluating N-glycans from glycoproteins involve deglycosylation via PNGase 
F and the labeling of the resulting N-glycans with a chemical moiety that imparts 
a detectable attribute. In one, highly effective approach, labeled glycans are 
separated by hydrophilic interaction chromatography (HILIC) and detected by 
fluorescence (FLR) and sometimes mass spectrometry (MS).3-10  

Unfortunately, conventional approaches to the preparation of N-glycans for 
HILIC-FLR-MS are either laborious, time-consuming, or require compromises in 
sensitivity.11 For instance, a conventional deglycosylation procedure requires 
that a glycoprotein sample be incubated for about 1 hour, while many analysts 
generically employ an overnight (16 hour) incubation. Combined with this 
process is a lengthy, 2 to 3 hour labeling step that relies on reductive amination 
of reducing, aldehyde termini that form on N-glycans only after they hydrolyze 
from their glycosylamine forms. And in the case of one of the most frequently 
employed labeling compounds, 2-aminobenzamide (2-AB), the resulting glycans 
can be readily detected by fluorescence but are rather challenging to detect by 
electrospray ionization mass spectrometry (ESI-MS).  

Variations to conventional approaches for N-glycan sample preparation have been 
explored, but have not, as of yet, presented a solution that combines the desired 
attributes of simplicity, high MS sensitivity, and high throughput. Alternative 
labeling reagents, for example procainamide, that have functional groups to 
enhance electrospray ionization efficiency have been used,12 but this does not 
address the cumbersome, time consuming nature of relying on a reductive 
amination labeling step. Rapid tagging procedures that yield labeled glycans in a 
matter of minutes have consequently been investigated. In fact, two rapid tagging 
glycan labels were recently introduced, including a rapid tagging analog 

Rapid Preparation of Released N-Glycans for HILIC Analysis Using  

a Novel Fluorescence and MS-Active Labeling Reagent
Matthew A. Lauber,1 Darryl W. Brousmiche,1 Zhengmao Hua,1 Stephan M. Koza,1 Ellen Guthrie,2 Paula Magnelli,2  

Christopher H. Taron,2 Kenneth J. Fountain1
1

 Waters Corporation, Milford, MA, USA

2

 New England BioLabs, Ipswich, MA, USA

[ 65 ]

E X P E R IM E N TA L

Method conditions (unless otherwise noted): 

LC conditions
LC system:  

ACQUITY UPLC H-Class  
Bio System

Sample temp.:  

5 °C

Analytical column 
temp.:  

60 °C 

Flow rate: 

0.4 mL/min

Fluorescence detection: Ex 265/Em 425 nm 

(RapiFluor-MS)

Ex 278/Em 344 nm 
(Instant AB)

Ex 330/Em 420 nm (2-AB)

(2 Hz scan rate [150 mm 
column]/5 Hz scan rate 
[50 mm column], Gain =1) 

Injection volume:  

≤1 µL (aqueous diluents 
with 2.1 mm I.D. columns)

≤30 µL (DMF/ACN diluted 
samples with 2.1 mm  
I.D. columns)

Columns:  

ACQUITY UPLC Glycan 
BEH Amide 130Å, 1.7 µm,  
2.1 x 50 mm  
(p/n 186004740)

ACQUITY UPLC Glycan 
BEH Amide 130Å, 1.7 µm,  
2.1 x 150 mm  
(p/n 186004742)

Sample collection/
vials:  

Sample Collection Module 
(p/n 186007988)

Polypropylene 12 x 32 mm 
Screw Neck Vial, 300 µL 
volume (p/n 186002640)

Gradient used with 2.1 x 50 mm columns:
Mobile phase A:  

50 mM ammonium 
formate, pH 4.4 (LC-MS 
grade; from a 100x 
concentrate,  
p/n 186007081)

Mobile phase B:  

ACN (LC-MS grade)

Time 

Flow rate 

%A 

%B 

Curve 

 (mL/min) 
0.0 

0.4  25 75  6 

11.7 

0.4  46 54  6 

12.2  0.2  100 0  6 
13.2  0.2  100 0  6 
14.4  0.2  25 75  6 
15.9  0.4  25 75  6 
18.3  0.4  25 75  6

Gradient used with 2.1 x 150 mm columns:
Mobile phase A:  

50 mM ammonium 
formate, pH 4.4 (LC-MS 
grade; from  
a 100x concentrate,  
p/n 186007081)

Mobile phase B:  

ACN (LC-MS grade)

Time 

Flow rate 

%A 

%B 

Curve 

 (mL/min) 
0.0 

0.4  25 75  6 

35.0  0.4  46 54  6 
36.5  0.2  100 0  6 
39.5  0.2  100 0  6 
43.1  0.2  25 75  6 
47.6  0.4  25 75  6 
55.0  0.4  25 75  6

[ 66 ] Rapid Preparation of Released N-Glycans for HILIC Analysis Using a Novel Fluorescence and MS-Active Labeling Reagent

of aminobenzamide (AB).13 In a rapid reaction, the precursor glycosylamines 
of reducing, aldehyde terminated glycans are modified via a urea linked 
aminobenzamide. Although such a rapid tagging reagent accelerates the 
labeling procedure, it does not provide the enhanced ionization efficiencies 
needed in modern N-glycan analyses.

To address the above shortcomings, we have developed a sample preparation 
solution that enables unprecedented FLR and MS sensitivity for glycan detection 
while also improving the throughput of N-glycan sample preparation. A novel 
labeling reagent has been synthesized that rapidly reacts with glycosylamines 
upon their release from glycoproteins. Within a 5 minute reaction, N-glycans 
are labeled with RapiFluor-MS, a reagent comprised of an N-hydroxysuccinimide 
(NHS) carbamate rapid tagging group, an efficient quinoline fluorophore, and 
a highly basic tertiary amine for enhancing ionization. To further accelerate the 
preparation of N-glycans, rapid tagging has been directly integrated with a  
Rapid PNGase F deglycosylation procedure involving  RapiGest SF surfactant  
and a HILIC µElution SPE clean-up step that provides highly quantitative 
recovery of the released and labeled glycans with the added benefit of not 
requiring a solvent dry-down step prior to the LC-FLR-MS analysis of samples.

SAM P L E D E S C R I P T IO N
N-glycans from Intact mAb Mass Check Standard (p/n 186006552), bovine 
fetuin (Sigma F3004), and pooled human IgG (Sigma I4506) were prepared 
according to the guidelines provided in the GlycoWorks RapiFluor-MS N-Glycan 
Kit Care and Use Manual (715004793).  

To compare the response factors of Instant AB™ and RapiFluor-MS labeled 
glycans, labeling reactions were performed with equivalent molar excesses of 
reagent, and crude reaction mixtures were directly analyzed by HILIC-FLR-MS in 
order to avoid potential biases from SPE clean-up procedures. Response factors 
were determined as ratios of the FA2 N-glycan (Oxford notation) chromatographic 
peak area to the mass of glycoprotein from which the glycan originated. 

To compare the response factors of 2-AB labeled versus RapiFluor-MS labeled 
glycans, equivalent quantities of labeled N-glycans from pooled human IgG 
were analyzed by HILIC-FLR-MS. Column loads were calibrated using external 
quantitative standards of 2-AB labeled triacetyl chitotriose and RapiFluor-MS 
derivatized propylamine (obtained in high purity; confirmed by HPLC and  
1H NMR).  Response factors were determined as ratios of the FA2 
chromatographic peak area to the mole quantity of glycan.

The procedure for extracting labeled RapiFluor-MS glycans after derivatization 
was evaluated using a test mixture containing N-glycans released and labeled 
from a 1:1 mixture (by weight) of pooled human IgG and bovine fetuin. The test 
mixture was prepared and then reconstituted in a solution equivalent in 

MS conditions

MS system:  

SYNAPT G2-S HDMS

Ionization mode:  

ESI+

Analyzer mode:  

TOF MS, resolution mode 
(~20 K)

Capillary voltage:  

3.0 kV

Cone voltage:  

80 V

Source temp.:  

120 °C

Desolvation temp.:  

350 °C

Desolvation gas flow:   800 L/Hr

Calibration:  

NaI, 1 µg/µL from 
500–2500 m/z

Lockspray 
(ASM B-side):  

100 fmol/µL Human 
Glufibrinopeptide B in 
0.1% (v/v) formic acid, 
70:30 water/acetonitrile 
every 90 seconds 

Acquisition:  

500–2500 m/z,  
1 Hz scan rate 

Data management: 

MassLynx Software (V4.1)

[ 67 ]

Rapid Preparation of Released N-Glycans for HILIC Analysis Using a Novel Fluorescence and MS-Active Labeling Reagent

composition to the solution glycans are subjected to when following the protocol of the  RapiFluor-MS 
N-Glycan Kit. All other sample preparation techniques are described in the GlycoWorks  RapiFluor-MS 
N-Glycan Kit Care and Use Manual (715004793).

R E SU LT S A N D D IS C U S S IO N

Rational design of a new N-Glycan labeling reagent 
A new labeling reagent for facilitating N-glycan analysis has been synthesized based on rational design 
considerations (Figure 1) that would afford rapid labeling kinetics, high fluorescence quantum yield, 
and significantly enhanced MS detectability. Conventional N-glycan sample preparation is dependent 
on reductive amination of aldehyde terminated saccharides, a process that requires glycans to undergo 
multiple chemical conversions and a lengthy high temperature incubation step.11 Moreover, glycans must be 
reductively aminated in anhydrous conditions in order to minimize desialylation. Sample preparations are 
therefore burdened with transitioning a sample from aqueous to anhydrous conditions. For these reasons, the 
newly designed labeling reagent foregoes reductive amination and instead takes advantage of an aqueous 
rapid tagging reaction. To this end, Waters has drawn upon its experience with rapid fluorescence labeling 
of amino acids to develop a new reagent that meets the needs of modern, N-glycan analysis. More than 20 
years ago, Waters introduced a rapid tagging labeling reagent, known as AccQ•Fluor™, that is now widely 
use to accurately profile the amino acid composition of protein samples via fluorescence detection.14-15 

O

NH

O

N

O

NH

N

N

O

O

Rapid

Fluorescence

MS

NHS Carbamate Rapid

Tagging Group

Quinolinyl

Fluorophore

Tertiary Amine

Charge Tag

Figure 1. RapiFluor-MS Molecular Structure. 
Features of the chemical structure that enable 
rapid tagging, efficient fluorescence, and 
enhanced ionization efficiency are highlighted.

Figure 2. Reaction Schematic for RapiFluor-MS 
Derivatization of an N-glycosylamine. The 
pathway on the left shows the derivatization of 
a glycosylamine, which produces an N-glycan 
with a urea (NH-CO-NH) linked RapiFluor-MS 
label. Hydrolysis of RapiFluor-MS is shown in 
the pathway on the right.

+

+

H2O

+

+ CO2

[ 68 ] Rapid Preparation of Released N-Glycans for HILIC Analysis Using a Novel Fluorescence and MS-Active Labeling Reagent

AccQ•Fluor possesses two important chemical characteristics: an 
NHS-carbamate rapid tagging reactive group and a highly efficient 
quinolinyl fluorophore. These features of AccQ•Fluor form the basis 
of the new glycan labeling reagent. The NHS-carbamate reactive 
group of this reagent enables glycosylamine bearing N-glycans 
to be rapidly labeled following their enzymatic release from 
glycoproteins. Within a 5 minute reaction, N-glycans are labeled 
with the new reagent under ambient, aqueous conditions to yield a 
highly stable urea linkage (Figure 2). In addition to rapid tagging 
capabilities, the new labeling reagent also supports high sensitivity 
for both MS and flourescence detection. A quinoline fluorophore 
serves as the central functionality of the new reagent that, as with 
AccQ•Fluor, facilitates high sensitivity fluorescence detection. 
In addition to AccQ•Fluor, however, the new reagent has been 
synthesized with a tertiary amine side chain as a means to enhance 
MS signal upon positive ion mode electrospray ionization (ESI+). 
In summary, the resulting N-glycan labeling reagent is built upon 
our expertise in chemical reagents and three important chemical 
attributes, a rapid tagging reactive group, an efficient fluorophore, 
and a highly basic MS active group. To describe these noteworthy 
attributes, the new labeling reagent has accordingly been  
named  RapiFluor-MS.

300x zoom 

4.5

5

5.5

6

6.5

7

4.5

5

5.5

6

6.5

7

FLR 

MS 

(BPI) 

FLR 

MS 

(BPI) 

342.8

179.9

233.7

0.3

0

100

200

300

400

Compound 4

Compound 1

A 

FA2 

FA2 

FA2 

FA2 

B 

C 

FLR 

MS 

(BPI) 

Re

sponse 

Fa

ctors 

(F

A2 

Pe

ak Area per Sample of N-Glycans 

from 1 

g of Intact mAb Mass Check Standard/1000)

Rapi Fluor-MS Labeled N- Glycans  

from Intact mAb Mass Check Standard (0.4 µg) 

Instant AB Labeled N-Glycans  

from Intact mAb Mass Check Standard (0.4 µg) 

min 

min 

0.0E+0

2.6E+6

1.7E+6

0.0E+0

2.6E+6

0.0E+0
1.7E+6

0.0E+0

RapiFluor-MS 

Labeled 

Instant AB 

Labeled 

NH

O

NH

N

NH

O

N

Glycan

NH

O

NH

Glycan

O

NH2

RapiFluor-MS enables high sensitivity detection
RapiFluor-MS N-glycan labeling has been extensively studied. In 
particular, the response factors of   RapiFluor-MS labeled glycans 
have been benchmarked against those observed for glycans labeled 
with alternative reagents. The most closely related, commercially 
available alternative to  RapiFluor-MS is an NHS carbamate analog 
of aminobenzamide, known as Instant AB.13 Figures 3A and 3B 
present HILIC fluorescence and base peak intensity (BPI) MS 
chromatograms for equivalent quantities of N-glycans released from 
a murine monoclonal antibody (Intact mAb Mass Check Standard, 
p/n 186006552) and labeled with RapiFluor-MS and Instant AB, 
respectively. Based on the observed chromatographic peak areas, 
response factors for fluorescence and MS detection were determined 
for the most abundant glycan in the IgG profile, the fucosylated, 
biantennary FA2 glycan (Figure 3C). Our results for the FA2 glycan 
indicate that RapiFluor-MS labeled glycans produce 2 times higher 
fluorescence signal and, more astoundingly, 780 times greater  
MS signal than N-glycans labeled with Instant AB. 

Figure 3. HILIC-FLR-MS of (A) RapiFluor-MS 
and (B) Instant AB Labeled N-Glycans from 
Intact mAb Mass Check Standard. Fluorescence 
(FLR) chromatograms are shown in orange and 
base peak intensity (BPI) MS chromatograms 
are shown in blue. Labeled glycans (from 0.4 
µg of glycoprotein, 1 µL aqueous inection) 
were separated using a ACQUITY UPLC BEH 
Amide 130Å, 1.7 µm, 2.1 x 50 mm Column. 
(C) Response factors for RapiFluor-MS and 
Instant AB labeled glycans (measured as 
the FA2 peak area per sample of N-glycans 
resulting from 1 µg of Intact mAb Mass Check 
Standard). Fluorescence (FLR) and MS (base 
peak intensity) response factors are shown in 
orange and blue, respectively. Analyses were 
performed in duplicate.

[ 69 ]

Rapid Preparation of Released N-Glycans for HILIC Analysis Using a Novel Fluorescence and MS-Active Labeling Reagent

In a similar fashion,  RapiFluor-MS labeling has also been compared to conventional 2-AB labeling. To draw such a comparison, N-glycans 
released from pooled human IgG and labeled with either  RapiFluor-MS or 2-AB were analyzed by HILIC-FLR-MS at equivalent mass loads 
(Figures 4A and 4B, respectively). Given that rapid tagging and reductive amination are performed by significantly different procedures, 
external calibrations were established using quantitative standards in order to determine the amounts of FA2 glycan loaded and eluted 
from the HILIC column. Response factors determined using these calibrated amounts of FA2 glycan are provided in Figure 4C. Again, it 
is found that  RapiFluor-MS labeled glycans are detected with significantly higher signal (14 times higher fluorescence and 160 times 
greater MS signal versus 2-AB labeled glycans.)

3

4

5

6

7

8

9

4.0E+6

3

4

5

6

7

8

9

3.3E+6

100x zoom 

FA2 

(2.17 pmol) 

FA2 

FA2  

(2.61 pmol) 

FA2 

105.8

7.4

126.9

0.8

0

50

100

150

A 

B 

RapiFluor-MS 

Labeled 

FLR 

MS 

(BPI) 

FLR 

MS 

(BPI) 

2-AB 

Labeled 

C 

RapiFluor-MS Labeled N-Glycans from Pooled Human IgG 

FLR 

MS 

(BPI) 

Response Factors 

(Peak Area per 

pmol of Labeled FA2 Glyca

n 

/ 1000)

2-AB Labeled 

N -Glycans from Pooled Human IgG 

min 

min 

0.0E+0

0.0E+0

4.0E+6

3.3E+6

0.0E+0

0.0E+0

NH

O

NH

N

NH

O

N

Glycan

NH

Glycan

NH2

O

Figure 4. HILIC-FLR-MS of (A) RapiFluor-MS and (B) 2-AB 
Labeled N-Glycans from Pooled Human IgG. Fluorescence (FLR) 
chromatograms are shown in orange and base peak intensity (BPI) 
MS chromatograms are shown in blue. Labeled glycans  
(~14 pmol total glycan, 1 µL aqueous injection) were separated 
using a ACQUITY UPLC Glycan BEH Amide 130Å, 1.7 µm,  
2.1 x 50 mm Column. The quantities of FA2 glycan were calibrated 
via two-point external calibrations with quantitative standards.  
(C) Response factors for RapiFluor-MS and 2-AB labeled glycans 
(measured as the FA2 peak area per picomole of FA2 determined 
by the external calibration). Fluorescence (FLR) and MS (BPI) 
response factors are shown in orange and blue, respectively. 
Analyses were performed in duplicate.

Figure 5. Relative Performance of Glycan Labels. Response 
factors normalized to the fluorescence and MS response 
factors of RapiFluor-MS labeled N-glycans. (*) Comparative 
result extrapolated from a published comparison of N-glycans, 
wherein it was found that procainamide provided comparable 
fluorescence and up to 50 fold greater ESI-MS sensitivity when 
compared to 2-AB (Klapoetke et al. 2010).