Methods in

Molecular Biology 2414

Fadil Bidmos
Janine Bossé
Paul Langford Editors

Bacterial
Vaccines
Methods and Protocols
METHODS IN MOLECULAR BIOLOGY

Series Editor
John M. Walker
School of Life and Medical Sciences
University of Hertfordshire
Hatfield, Hertfordshire, UK

For further volumes:

http://www.springer.com/series/7651
For over 35 years, biological scientists have come to rely on the research protocols and
methodologies in the critically acclaimed Methods in Molecular Biology series. The series was
the first to introduce the step-by-step protocols approach that has become the standard in all
biomedical protocol publishing. Each protocol is provided in readily-reproducible step-by-
step fashion, opening with an introductory overview, a list of the materials and reagents
needed to complete the experiment, and followed by a detailed procedure that is supported
with a helpful notes section offering tips and tricks of the trade as well as troubleshooting
advice. These hallmark features were introduced by series editor Dr. John Walker and
constitute the key ingredient in each and every volume of the Methods in Molecular Biology
series. Tested and trusted, comprehensive and reliable, all protocols from the series are
indexed in PubMed.
 Bacterial Vaccines

Methods and Protocols

Edited by

Fadil Bidmos, Janine Bossé, and Paul Langford

Department of Infectious Disease, Imperial College London, London, United Kingdom
Editors
Fadil Bidmos Janine Bossé
Department of Infectious Disease Department of Infectious Disease
Imperial College London Imperial College London
London, United Kingdom London, United Kingdom

Paul Langford
Department of Infectious Disease
Imperial College London
London, United Kingdom

ISSN 1064-3745 ISSN 1940-6029 (electronic)

Methods in Molecular Biology
ISBN 978-1-0716-1899-8 ISBN 978-1-0716-1900-1 (eBook)
https://doi.org/10.1007/978-1-0716-1900-1
© The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Science+Business Media, LLC, part
of Springer Nature 2022
This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and
retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter
developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply,
even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations
and therefore free for general use.
The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to
be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty,
expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been
made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

This Humana imprint is published by the registered company Springer Science+Business Media, LLC, part of Springer
Nature.
The registered company address is: 1 New York Plaza, New York, NY 10004, U.S.A.
Preface

Vaccination is considered the second most effective public health intervention available, only
following the provision and access to clean water for disease prevention. Compared to the
peak number of deaths prior to introduction of bacterial vaccines against diphtheria, tetanus,
pertussis (whooping cough), and disease caused by Haemophilus influenzae type b, it is
estimated that c. 1.3 million lives, predominantly of children, have been saved annually. The
associated reduction in morbidity by prevention of disease through herd immunity, and
reduction in secondary infections that complicate vaccine-preventable diseases, is estimated
to save 386 million life years and 96 million disability-adjusted life years (DALYs) globally.
In addition to reduction of mortality and morbidity, the use of vaccines has substantial social
benefits. The latter include equity of healthcare, strengthening of health and social care
infrastructure, increased social mobility, improved life expectancy, and empowerment of
women. Economically, in lower- and middle-income countries, an investment of $34 billion
in vaccines resulted in savings of $586 billion from the cost associated with direct illness.
Cost-effectiveness or cost/benefit analyses do not typically consider prevention of long-
term morbidity following acute infection, such as hearing loss or amputation of limbs as
sequelae of meningococcal disease, and the calculated figures can be considered
underestimates.
Given their benefits, it seems a paradox that historically pharmaceutical manufacturers
have been wary of investing in vaccines because of concerns about legal liability and a
comparative low return on investment compared to pharmaceuticals. That situation has
recently started to change in the bacterial arena because of the threats posed by long-
standing diseases where there is still a need for improved vaccines (e.g., for tuberculosis).
In addition, with an aging worldwide population, vaccination is increasingly part of a life-
course strategy to meet the needs of the elderly. However, the main driver is the increase in
worldwide antimicrobial resistance (AMR). As Dame Sally Davies, the UK Chief Medical
Officer, stated: “The world is facing an antibiotic apocalypse. Unless action is taken to halt the
practices that have allowed antimicrobial resistance to spread and ways are found to develop
new types of antibiotics, we could return to the days when routine operations, simple wounds or
straightforward infections could pose real threats to life.” In particular, there is currently much
research focus on vaccine development for the so-called bacterial ESKAPE pathogens:
Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter bauman-
nii, Pseudomonas aeruginosa, and Enterobacter spp. Furthermore, new preventions and/or
antibiotics are urgently needed for other World Health Organization-designated critical or
high priority pathogens such as Neisseria gonorrhoeae, where strains have already been
isolated that are resistant to current recommended treatment regimes. The solution to
AMR will also involve improvement in and introduction of novel bacterial vaccines to reduce
antimicrobial use in animals to prevent transfer of resistance genes through natural transfor-
mation, transconjugation, transduction, and vesiduction in zoonotic pathogens that cause
disease in humans.
Thus, a book on bacterial vaccine methods is extremely timely. We have organized the
chapters into three sections: (1) vaccine antigen discovery; (2) vaccine production and
delivery; and (3) immunology of vaccine candidates. The first two chapters (Ong and He;
Leow et al.) describe reverse vaccinology methods—the use of bioinformatic approaches to

v
vi Preface

identify vaccine candidates, which have been successfully applied in development of the
commercially available Neisseria meningitidis serogroup B Bexsero vaccine. With over
200,000 bacterial genomes publicly available, e.g., through NCBI and EMBL, the use of
reverse vaccinology approaches can be a valuable starting point for immunogenic antigen
and/or epitope discovery, and crucially is not hypothesis-driven. Four further chapters
describe different proteomic approaches for vaccine antigen discovery. These include classic
methods to identify potential immunogenic epitopes, such as immunoprecipitation
(Reglinski), enzymatic surface shaving (Luu and Lan), two-dimensional electrophoresis
combined with western blotting (Obradovic and Wilson), as well as pan-proteomic array
technology (Campo and Oberai) which is also a bottom-up approach starting with genome
data, like reverse vaccinology. Nine chapters are devoted to vaccine production and delivery
systems, illustrating the range and ingenuity of platforms available. These include a relatively
simple yet elegant bacterin inactivation mechanism involving low-energy electron irradia-
tion (Fertey et al.), and bioengineering approaches such as those described for Gram-
negative metal ion transporters (Chauduri et al.), Gram-positive extracellular membrane
vesicles (Stentz et al.), and biological conjugation (Terra and Kay). Different methods for
preparation of outer membrane vesicles/generalized modules for membrane antigens are
presented, including those for secretory production of heterologous antigens (Kawamoto
and Kurihara), for reduction of potential toxicity (Hirayama and Nakao), and for ensuring
quality and stability (Micoli et al.). In addition, two chapters describe platforms that can be
used to present peptide subunits (trimethyl chitosan-based polyelectrolyte complexes)
(Zhao et al.) or multi-epitope fusion antigens (MEFA) (Li et al.), respectively. The final
set of chapters provide methods for assessment of antigens at the discovery, delivery, and
post-licensure phases, including monitoring of immune responses to inhaled antigens
(Ashhurst et al.) and in infants following maternal immunization (Rice and Holder). The
production of a universal human complement source (Alexander et al.) is described, which
could be used to ensure reproducibility of the functional opsonophagocytic activity and
serum bactericidal assays detailed by Semchenko et al. and Wagstaffe et al. Furthermore,
informative infection models, including the ex vivo organ co-perfusion model (Hames et al.)
and controlled human infection model (Dale et al.) described, as well as techniques such as
multi-color flow cytometry and high-dimensional data analysis for analyses of vaccine
responses (Cole et al.) will improve evaluation of prospective vaccines and enable prioritiza-
tion of candidates for field testing, thus speeding up the time to market.
We hope that researchers will not only find specific chapters of interest but also take the
time to review others detailed in the book, as overall they give a snapshot of the variety of
methods available and the extremely high innovative nature of the bacterial vaccine field. We
would like to express our thanks to all of the authors for their valuable contributions,
meeting deadlines, and their understanding through the editing process, and to the series
editor, John Walker, for his excellent guidance.

London, UK Fadil Bidmos

Janine Bossé
Paul Langford
Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix

1 Vaccine Design by Reverse Vaccinology and Machine Learning . . . . . . . . . . . . . . . 1

Edison Ong and Yongqun He
2 Application of Reverse Vaccinology and Immunoinformatic Strategies
for the Identification of Vaccine Candidates Against Shigella flexneri . . . . . . . . . . 17
Chiuan Yee Leow, Candy Chuah, Abu Bakar Abdul Majeed,
Norazmi Mohd Nor, and Chiuan Herng Leow
3 Purification of Prospective Vaccine Antigens from Gram-Positive
Pathogens by Immunoprecipitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
Mark Reglinski
4 Rapid Surface Shaving for Proteomic Identification of Novel
Surface Antigens for Vaccine Development. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
Laurence Don Wai Luu and Ruiting Lan
5 Two-Dimensional Electrophoresis Coupled with Western Blot as a
Method to Detect Potential Neutralizing Antibody Targets from
Gram-Negative Intracellular Bacteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Milan Obradovic and Heather L. Wilson
6 Panproteome Analysis of the Human Antibody Response to Bacterial
Vaccines and Challenge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
Joseph J. Campo and Amit Oberai
7 Low-Energy Electron Irradiation (LEEI) for the Generation of
Inactivated Bacterial Vaccines. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97
Jasmin Fertey, Bastian Standfest, Jana Beckmann, Martin Thoma,
Thomas Grunwald, and Sebastian Ulbert
8 Design and Production of Hybrid Antigens for Targeting Integral
Outer Membrane Proteins in Gram-Negative Bacteria . . . . . . . . . . . . . . . . . . . . . . . 115
Somshukla Chaudhuri, Nikolas F. Ewasechko, Luisa Samaniego-Barron,
Jamie E. Fegan, and Anthony B. Schryvers
9 Preparation of Trimethyl Chitosan-Based Polyelectrolyte Complexes
for Peptide Subunit Vaccine Delivery . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
Lili Zhao, Sahra Bashiri, Istvan Toth, and Mariusz Skwarczynski
10 Multiepitope Fusion Antigen: MEFA, an Epitope- and Structure-Based
Vaccinology Platform for Multivalent Vaccine Development. . . . . . . . . . . . . . . . . . 151
Siqi Li, Kuo Hao Lee, and Weiping Zhang
11 Production, Isolation, and Characterization of Bioengineered
Bacterial Extracellular Membrane Vesicles Derived from
Bacteroides thetaiotaomicron and Their Use in Vaccine Development . . . . . . . . . . 171
Régis Stentz, Ariadna Miquel-Clopés, and Simon R. Carding

vii
viii Contents

12 Membrane Vesicles Produced by Shewanella vesiculosa HM13 as

a Prospective Platform for Secretory Production of Heterologous
Proteins at Low Temperatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
Jun Kawamoto and Tatsuo Kurihara
13 Glycine Induction Method: Effective Production of Immunoactive
Bacterial Membrane Vesicles with Low Endotoxin Content . . . . . . . . . . . . . . . . . . 207
Satoru Hirayama and Ryoma Nakao
14 Methods for Assessment of OMV/GMMA Quality and Stability . . . . . . . . . . . . . 227
Francesca Micoli, Renzo Alfini, and Carlo Giannelli
15 Production of Vaccines Using Biological Conjugation . . . . . . . . . . . . . . . . . . . . . . . 281
Emily J. Kay and Vanessa S. Terra
16 Immunological Assessment of Lung Responses to Inhalational
Lipoprotein Vaccines Against Bacterial Pathogens. . . . . . . . . . . . . . . . . . . . . . . . . . . 301
Anneliese S. Ashhurst, Cameron C. Hanna, Richard J. Payne,
and Warwick J. Britton
17 Determination of Maternal and Infant Immune Responses to
Pertussis Vaccination in Pregnancy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 325
Thomas Rice and Beth Holder
18 Generation of a Universal Human Complement Source by
Large-Scale Depletion of IgG and IgM from Pooled Human Plasma . . . . . . . . . . 341
Frances Alexander, Emily Brunt, Holly Humphries, Breeze Cavell,
Stephanie Leung, Lauren Allen, Rachel Halkerston, Elodie Lesne,
Elizabeth Penn, Stephen Thomas, Andrew Gorringe, and Stephen Taylor
19 Assessment of Serum Bactericidal and Opsonophagocytic Activity
of Antibodies to Gonococcal Vaccine Targets. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
Evgeny A. Semchenko, Freda E.-C. Jen, Michael P. Jennings,
and Kate L. Seib
20 Opsonophagocytic Killing Assay to Measure Anti–Group A
Streptococcus Antibody Functionality in Human Serum . . . . . . . . . . . . . . . . . . . . . . 373
Helen Wagstaffe, Scott Jones, Marina Johnson, and David Goldblatt
21 Neisseria lactamica Controlled Human Infection Model. . . . . . . . . . . . . . . . . . . . . 387
Adam P. Dale, Diane F. Gbesemete, Robert C. Read, and Jay R. Laver
22 Analyzing Macrophage Infection at the Organ Level . . . . . . . . . . . . . . . . . . . . . . . . 405
Ryan G. Hames, Zydrune Jasiunaite, Joseph J. Wanford,
David Carreno, Wen Y. Chung, Ashley R. Dennison,
and Marco R. Oggioni
23 Multicolor Flow Cytometry and High-Dimensional Data Analysis
to Probe Complex Questions in Vaccinology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 433
Megan E. Cole, Yanping Guo, Hannah M. Cheeseman,
and Katrina M. Pollock

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 449
Contributors

ABU BAKAR ABDUL MAJEED • Faculty of Pharmacy, Universiti Teknologi MARA, Kuala
Selangor, Selangor, Malaysia
FRANCES ALEXANDER • Public Health England, Porton Down, UK
RENZO ALFINI • GSK Vaccines Institute for Global Health, Siena, Italy
LAUREN ALLEN • Public Health England, Porton Down, UK
ANNELIESE S. ASHHURST • School of Chemistry, The University of Sydney, Sydney, NSW,
Australia; Tuberculosis Research Program Centenary Institute, The University of Sydney,
Camperdown, NSW, Australia; School of Medical Sciences, Faculty of Medicine and Health,
The University of Sydney, Camperdown, NSW, Australia
SAHRA BASHIRI • School of Chemistry & Molecular Biosciences, The University of Queensland,
St Lucia, QLD, Australia
JANA BECKMANN • Fraunhofer Institute for Organic Electronics, Electron Beam and Plasma
Technology FEP, Dresden, Germany
WARWICK J. BRITTON • Tuberculosis Research Program Centenary Institute, The University of
Sydney, Camperdown, NSW, Australia; School of Medical Sciences, Faculty of Medicine and
Health, The University of Sydney, Camperdown, NSW, Australia; Department of Clinical
Immunology, Royal Prince Alfred Hospital, Camperdown, NSW, Australia
EMILY BRUNT • Public Health England, Porton Down, UK
JOSEPH J. CAMPO • Antigen Discovery, Inc., Irvine, CA, USA
SIMON R. CARDING • Quadram Institute, Norwich, UK
DAVID CARRENO • Department of Genetics and Genome Biology, University of Leicester,
Leicester, UK
BREEZE CAVELL • Public Health England, Porton Down, UK
SOMSHUKLA CHAUDHURI • Department of Microbiology, Immunology, and Infectious
Diseases, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada
HANNAH M. CHEESEMAN • Department of Infectious Disease, Imperial College London,
London, UK
CANDY CHUAH • Faculty of Medicine and Health Sciences, Universiti Tunku Abdul
Rahman, Kajang, Selangor, Malaysia
WEN Y. CHUNG • Department of Hepatobiliary and Pancreatic Surgery, University
Hospitals of Leicester, Leicester, UK
MEGAN E. COLE • The Pirbright Institute, Surrey, UK
ADAM P. DALE • Clinical and Experimental Sciences, University of Southampton,
Southampton, UK
ASHLEY R. DENNISON • Department of Hepatobiliary and Pancreatic Surgery, University
Hospitals of Leicester, Leicester, UK
NIKOLAS F. EWASECHKO • Department of Microbiology, Immunology, and Infectious Diseases,
Cumming School of Medicine, University of Calgary, Calgary, AB, Canada
JAMIE E. FEGAN • Department of Molecular Genetics, University of Toronto, Toronto, ON,
Canada
JASMIN FERTEY • Fraunhofer-Institute for Cell Therapy and Immunology IZI, Leipzig,
Germany

ix
x Contributors

DIANE F. GBESEMETE • Clinical and Experimental Sciences, University of Southampton,

Southampton, UK; NIHR Clinical Research Facility, University Hospital Southampton
NHS Foundation Trust, Southampton, UK
CARLO GIANNELLI • GSK Vaccines Institute for Global Health, Siena, Italy
DAVID GOLDBLATT • Great Ormond Street Institute of Child Health, University College
London, London, UK
ANDREW GORRINGE • Public Health England, Porton Down, UK
THOMAS GRUNWALD • Fraunhofer-Institute for Cell Therapy and Immunology IZI, Leipzig,
Germany
YANPING GUO • Cancer Research UK Flow Cytometry Translational Technology Platform,
Cancer Institute, UCL, London, UK
RACHEL HALKERSTON • Public Health England, Porton Down, UK
RYAN G. HAMES • Department of Genetics and Genome Biology, University of Leicester,
Leicester, UK
CAMERON C. HANNA • School of Chemistry, The University of Sydney, Sydney, NSW, Australia
YONGQUN HE • Center of Computational Medicine and Bioinformatics, Unit for Laboratory
Animal Medicine, Department of Microbiology and Immunology, University of Michigan
Medical School, Ann Arbor, MI, USA
SATORU HIRAYAMA • Department of Bacteriology I, National Institute of Infectious Diseases,
Tokyo, Japan; Division of Microbiology and Infectious Diseases, Niigata University
Graduate School of Medical and Dental Sciences, Niigata, Japan
BETH HOLDER • Institute of Reproductive and Developmental Biology (IRDB), Department
of Metabolism, Development and Reproduction (MDR), Imperial College London, London,
UK
HOLLY HUMPHRIES • Public Health England, Porton Down, UK
ZYDRUNE JASIUNAITE • Department of Genetics and Genome Biology, University of Leicester,
Leicester, UK
FREDA E.-C. JEN • Institute for Glycomics, Griffith University, Gold Coast, QLD, Australia
MICHAEL P. JENNINGS • Institute for Glycomics, Griffith University, Gold Coast, QLD,
Australia
MARINA JOHNSON • Great Ormond Street Institute of Child Health, University College
London, London, UK
SCOTT JONES • Great Ormond Street Institute of Child Health, University College London,
London, UK
JUN KAWAMOTO • Institute for Chemical Research, Kyoto University, Uji, Kyoto, Japan
EMILY J. KAY • Faculty of Infectious and Tropical Diseases, Department of Infection Biology,
London School of Hygiene & Tropical Medicine, Keppel Street, London, UK
TATSUO KURIHARA • Institute for Chemical Research, Kyoto University, Uji, Kyoto, Japan
RUITING LAN • School of Biotechnology and Biomolecular Sciences, University of New South
Wales, Sydney, NSW, Australia
JAY R. LAVER • Clinical and Experimental Sciences, University of Southampton,
Southampton, UK
KUO HAO LEE • Computational Chemistry and Molecular Biophysics Unit, National
Institute of Health/National Institute on Drug Abuse, Baltimore, MD, USA
CHIUAN HERNG LEOW • Institute for Research in Molecular Medicine, Universiti Sains
Malaysia, Penang, Malaysia
CHIUAN YEE LEOW • School of Pharmaceutical Sciences, Universiti Sains Malaysia, Penang,
Malaysia
 Contributors xi

ELODIE LESNE • Public Health England, Porton Down, UK

STEPHANIE LEUNG • Public Health England, Porton Down, UK
SIQI LI • Department of Pathobiology, University of Illinois at Urbana-Champaign,
Urbana, IL, USA
LAURENCE DON WAI LUU • School of Biotechnology and Biomolecular Sciences, University of
New South Wales, Sydney, NSW, Australia
FRANCESCA MICOLI • GSK Vaccines Institute for Global Health, Siena, Italy
ARIADNA MIQUEL-CLOPÉS • Quadram Institute, Norwich, UK
NORAZMI MOHD NOR • School of Health Sciences, Universiti Sains Malaysia, Kubang
Kerian, Kelantan, Malaysia
RYOMA NAKAO • Department of Bacteriology I, National Institute of Infectious Diseases,
Tokyo, Japan
AMIT OBERAI • Antigen Discovery, Inc., Irvine, CA, USA
MILAN OBRADOVIC • Faculté de médecine vétérinaire/Faculty of Veterinary Medicine,
University of Montreal, St-Hyacinthe, QC, Canada
MARCO R. OGGIONI • Department of Genetics and Genome Biology, University of Leicester,
Leicester, UK; Department of Pharmacy and Biotechnology, University of Bologna,
Bologna, Italy
EDISON ONG • Department of Computational Medicine and Bioinformatics, University of
Michigan Medical School, Ann Arbor, MI, USA; GlaxoSmithKline Vaccines, Rixensart,
Belgium
RICHARD J. PAYNE • School of Chemistry, The University of Sydney, Sydney, NSW, Australia;
Australian Research Council Centre of Excellence for Innovations in Peptide and Protein
Science, The University of Sydney, Sydney, NSW, Australia
ELIZABETH PENN • Public Health England, Porton Down, UK
KATRINA M. POLLOCK • Department of Infectious Disease, Imperial College London, London,
UK
ROBERT C. READ • Clinical and Experimental Sciences, University of Southampton,
Southampton, UK; NIHR Clinical Research Facility, University Hospital Southampton
NHS Foundation Trust, Southampton, UK; NIHR Southampton Biomedical Research
Centre, Southampton, UK
MARK REGLINSKI • Division of Molecular Microbiology, School of Life Sciences, University of
Dundee, Dundee, UK
THOMAS RICE • Centre for Endocrinology, William Harvey Research Institute, Barts and The
London School of Medicine and Dentistry, Queen Mary University of London, London, UK
LUISA SAMANIEGO-BARRON • Department of Microbiology, Immunology, and Infectious
Diseases, Cumming School of Medicine, University of Calgary, Calgary, AB, Canada
ANTHONY B. SCHRYVERS • Department of Microbiology, Immunology, and Infectious Diseases,
Cumming School of Medicine, University of Calgary, Calgary, AB, Canada
KATE L. SEIB • Institute for Glycomics, Griffith University, Gold Coast, QLD, Australia
EVGENY A. SEMCHENKO • Institute for Glycomics, Griffith University, Gold Coast, QLD,
Australia
MARIUSZ SKWARCZYNSKI • School of Chemistry & Molecular Biosciences, The University of
Queensland, St Lucia, QLD, Australia
BASTIAN STANDFEST • Fraunhofer Institute for Manufacturing Engineering and
Automation IPA, Stuttgart, Germany
RÉGIS STENTZ • Quadram Institute, Norwich, UK
STEPHEN TAYLOR • Public Health England, Porton Down, UK
xii Contributors

VANESSA S. TERRA • Faculty of Infectious and Tropical Diseases, Department of Infection

Biology, London School of Hygiene & Tropical Medicine, Keppel Street, London, UK
MARTIN THOMA • Fraunhofer Institute for Manufacturing Engineering and Automation
IPA, Stuttgart, Germany
STEPHEN THOMAS • Public Health England, Porton Down, UK
ISTVAN TOTH • School of Chemistry & Molecular Biosciences, The University of Queensland, St
Lucia, QLD, Australia; Institute for Molecular Bioscience, The University of Queensland,
St. Lucia, QLD, Australia; School of Pharmacy, The University of Queensland,
Woolloongabba, QLD, Australia
SEBASTIAN ULBERT • Fraunhofer-Institute for Cell Therapy and Immunology IZI, Leipzig,
Germany
HELEN WAGSTAFFE • Great Ormond Street Institute of Child Health, University College
London, London, UK; Department of Infectious Disease, Imperial College London,
London, UK
JOSEPH J. WANFORD • Department of Genetics and Genome Biology, University of Leicester,
Leicester, UK
HEATHER L. WILSON • Department of Veterinary Microbiology and the School of Public
Health, Vaccine & Infectious Disease Organization (VIDO), University of Saskatchewan,
Saskatoon, SK, Canada
WEIPING ZHANG • Department of Pathobiology, University of Illinois at Urbana-
Champaign, Urbana, IL, USA
LILI ZHAO • School of Chemistry & Molecular Biosciences, The University of Queensland, St
Lucia, QLD, Australia
 Chapter 1

Vaccine Design by Reverse Vaccinology and Machine

Learning
Edison Ong and Yongqun He

Abstract
Reverse vaccinology (RV) is the state-of-the-art vaccine development strategy that starts with predicting
vaccine antigens by bioinformatics analysis of the whole genome of a pathogen of interest. Vaxign is the first
web-based RV vaccine prediction method based on calculating and filtering different criteria of proteins.
Vaxign-ML is a new Vaxign machine learning (ML) method that predicts vaccine antigens based on extreme
gradient boosting with the advance of new technologies and cumulation of protective antigen data. Using a
benchmark dataset, Vaxign-ML showed superior performance in comparison to existing open-source RV
tools. Vaxign-ML is also implemented within the web-based Vaxign platform to support easy and intuitive
access. Vaxign-ML is also available as a command-based software package for more advanced and custom-
izable vaccine antigen prediction. Both Vaxign and Vaxign-ML have been applied to predict SARS-CoV-
2 (cause of COVID-19) and Brucella vaccine antigens to demonstrate the integrative approach to analyze
and select vaccine candidates using the Vaxign platform.

Key words Vaccine, Antigen, Reverse vaccinology, Machine learning, Vaxign, Vaxign-ML, Vaxitop

1 Introduction

The conventional vaccine development process, which goes from

one wet-lab research to another is time-consuming. The advances
in high-throughput sequencing technologies have provided the
foundation for Reverse Vaccinology (RV) to predict potential vac-
cine candidates from the whole pathogen genome using bioinfor-
matics methods. The first RV study, pioneered by Rappuoli et al.
successfully selected and verified 28 immunogenic proteins using
bioinformatics analyses followed by experimental validation
[1]. Eventually, 5 out of these 28 protein candidates were formu-
lated in Bexsero®, licensed in the USA and Europe [2, 3].
In the first-generation RV technology development, a general
workflow included the prediction of: (a) subcellular localization,
(b) transmembrane helix, (c) adhesin probability, (d) signaling pep-
tide, (e) protein function, (f) conserved domain, and (g) similarity

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_1,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

1
2 Edison Ong and Yongqun He

Table 1
Vaxign system components

Vaxign system
component Included analysis
Vaxign Subcellular localization
Transmembrane helix
Adhesin probability
Similarity to host (human, mouse, and pig)
Vaxign-ML Extreme gradient boosting machine learning model to predict vaccine
candidate
Vaxitop Vaccine epitope prediction and analysis system based on the principle of reverse
vaccinology
Postprediction analysis IEDB epitope search
IEDB population coverage
EggNOG function
EggNOG orthologs
Pangenome analysis Orthologs’ phylogeny
Multiple sequence alignment

to host (Table 1). These properties were shown to correlate to

vaccine protection significantly [4], and several programs
(NERVE [5], Vaxign [6], and VacSol [7]) utilized these properties
as the filtering criteria for vaccine candidate prediction. On the
other hand, Doytchinova and Flower developed the first machine
learning-based RV program, VaxiJen, which applied discriminant
analysis by partial least square trained on the auto cross-covariance
transformed physicochemical properties of the pathogen protein
sequences [8]. By adopting and expanding the VaxiJen training
samples, Bowman et al., and later on, Heinson et al. applied a
support vector machine to predict vaccine candidates
[9, 10]. Another major difference between the VaxiJen program
and the Bowman-Heinson method was that the former used phys-
icochemical features while the latter used biological parts derived
from the pathogen protein sequences. A thorough benchmarking
study was conducted by Dalsass et al. to systematically evaluate the
RV prediction tools/methods [11], but the best performer’s accu-
racy was still suboptimal and there is room for improvement.
Motivated by the advance of machine learning in other bio-
medical research fields, including vaccine candidate prediction, the
Vaxign program was extended to include a machine learning com-
ponent, Vaxign-ML, in 2020 [12]. Vaxign-ML is a supervised
machine learning classification tool to predict protective antigens.
To identify the best machine learning method with optimized
conditions, five machine learning algorithms (logistic regression,
support vector machine, k-nearest neighbors, random forest, and
 Vaccine Design by Reverse Vaccinology and Machine Learning 3

extreme gradient boosting) were tested with biological and

physicochemical features extracted from the manually annotated
Protegen protective vaccine antigen database [13]. Nested fivefold
cross-validation and leave-one-pathogen-out validation were used
to ensure unbiased performance assessment and the capability to
predict vaccine candidates for a new emerging pathogen. The Vax-
ign-ML (extreme gradient boosting trained on all Protegen data)
was the best performing model compared to three publicly available
reverse vaccinology programs with a high-quality benchmark data-
set (see Note 1) and showed superior performance in predicting
protective antigens.
This article illustrates the procedures of Vaxign and Vaxign-ML
implementation. Furthermore, we have applied Vaxign and Vaxign-
ML to predict vaccine candidates for two pathogens, SARS-CoV-
2 (cause of COVID-19) and Brucella abortus (cause of zoonotic
brucellosis).

2 Vaxign-ML

The Vaxign-ML models (individual models for bacteria, virus, and

parasite) were trained on the protective antigens available in Prote-
gen [13]. All protective antigens stored in Protegen are manually
collected and curated using the in-house semiautomatic annotation
system [13]. For each specific protective antigen, the Protegen
database contains the detailed reference citation information from
PubMed and extracted general information of protective antigens
from the NCBI databases. The current Protegen database
contained 1432 protective antigens from 44 bacteria, 40 viruses,
19 parasites, and other noninfectious diseases (cancer and allergy)
(Table 2), which is near double the number since the initial release
in 2010. The Vaxign-ML tool was also applied to predict SARS-
CoV-2 vaccine candidate [14].

2.1 Vaxign-ML The most convenient and straightforward way to use Vaxign-ML
Stand-Alone Web Tool for vaccine candidate prediction is to access the Vaxign-ML stand-
alone web tool http://www.violinet.org/vaxign2/vaxign-ml. A
typical Vaxign-ML web interface is provided in Fig. 1. The standard
procedure for running Vaxign-ML is as follows:
1. Input the protein sequence(s). The following formats are sup-
ported: (a) FASTA format, (b) UniProtKB Protein ID,
(c) NCBI Protein ID, (d) NCBI Protein RefSeq, (e) NCBI
Gene ID, and (f) FASTA File Download Link. Users can also
upload the protein sequence(s) in FASTA format instead of
typing into the input text field.
4 Edison Ong and Yongqun He

Table 2
Protegen protective antigen statistics

Pathogen name # Protective antigens

Gram-positive bacteria 195
Bacillus anthracis (14)
Clostridium botulinum (10)
Listeria monocytogenes (2)
Mycobacterium tuberculosis (26)
Staphylococcus aureus (29)
Streptococcus agalactiae (22)
Streptococcus equi (17)
Streptococcus pneumoniae (26)
Streptococcus pyogenes (44)
Gram-negative bacteria 465
Actinobacillus pleuropneumoniae (20)
Bordetella pertussis (11)
Borrelia burgdorferi (11)
Brucella spp. (26)
Burkholderia pseudomallei (13)
Campylobacter jejuni (19)
Chlamydia muridarum (15)
Chlamydia trachomatis (14)
Chlamydophila abortus (12)
Chlamydophila pneumoniae (14)
Coxiella burnetii (12)
Edwardsiella tarda (22)
Escherichia coli (50)
Francisella tularensis (12)
Haemophilus influenzae (16)
Haemophilus parasuis (28)
Helicobacter pylori (23)
Leptospira spp. (11)
Neisseria meningitidis (20)
Pseudomonas aeruginosa (15)
Rickettsia spp (14)
Shigella (11)
(continued)
 Vaccine Design by Reverse Vaccinology and Machine Learning 5

Table 2
(continued)

Pathogen name # Protective antigens

Treponema pallidum (17)
Yersinia pestis (26)
Viruses 417
African Swine Fever Virus (7)
Bovine Herpesvirus (7)
Dengue Virus (11)
Ebola Virus (19)
Foot-and-Mouth Disease Virus (8)
Hantavirus (10)
Hepatitis B Virus (8)
Hepatitis C Virus (3)
Herpes Simplex Virus type 1 and 2 (12)
Human Immunodeficiency Virus (41)
Influenza Virus (50)
Japanese Encephalitis Virus (10)
Marburg Virus (9)
Measles Virus (5)
Mumps Virus (2)
Porcine Respiratory and Reproductive Syndrome Virus (7)
Pseudorabies Virus (8)
Rotavirus (9)
SARS-CoV (3)
Vaccinia Virus (12)
Western Equine Encephalomyelitis Virus (6)
Yellow Fever Virus (5)
Parasites 184
Leishmania donovani (15)
Leishmania infantum (8)
Leishmania major (13)
Neospora caninum (10)
Plasmodium spp. (33)
Toxoplasma gondii (18)
Trypanosoma cruzi (19)
(continued)
6 Edison Ong and Yongqun He

Table 2
(continued)

Pathogen name # Protective antigens

Fungi 9
Cancer 73
Other diseases (e.g., allergy, arthritis, and diabetes) 28
Total 1371

Fig. 1 Vaxign-ML web interface

2. Input pathogen organism type. The current Vaxign-ML sup-

ports “Gram positive bacterium,” “Gram negative bacterium,”
and “Virus.”
3. Click the submit button.
4. A Vaxign-ML prediction status page will be displayed, which
will be automatically refreshed every 30 s.
5. Once the result is ready, the overall result summary page will
appear, where the Vaxign-ML Protegenicity score(s) will be
displayed, and the result is also available for download in
Excel, CSV, and PDF formats. Users can perform additional
analyses to aid the vaccine candidate selection (see Subheading
3.4). The Protegenicity score represents the level of protective
antigenicity, that is, the extent of protective antigenicity a
protein candidate can stimulate in vivo [12].
 Vaccine Design by Reverse Vaccinology and Machine Learning 7

2.2 Vaxign-ML The stand-alone Vaxign-ML command-line tool is executable in

Stand-Alone Docker containers (tested on Docker version 1.13.1 and Docker
Command-Line Tool API version 1.26). Docker can be downloaded from https://docs.
docker.com/get-docker/, and installed with the appropriate plat-
form. The source code of the command-line tool is also available in
GitHub: https://github.com/VIOLINet/Vaxign-ML-docker.
Here is how the stand-alone Vaxign-ML can be conducted.
1. Download the prediction script from https://raw.
githubusercontent.com/VIOLINet/Vaxign-ML-docker/mas
ter/VaxignML.sh, and allow this script to be executed by the
system. Then, execute the downloaded VaxignML.sh script
with the following parameters: (a) input FASTA file,
(b) output directory to store Vaxign-ML prediction result,
and (c) pathogen organism type (currently support Gram-pos-
itive and -negative bacterium, and virus). There is also an
optional fourth parameter to specify the custom Vaxign-ML
model (see next step for the instruction to create custom
Vaxign-ML model).
2. (Optional) Instead of using the pretrained Vaxign-ML model,
users can also create a custom Vaxign-ML model using their
own dataset. This customized model can be created by first
downloading the Train.sh shell script from https://raw.
githubusercontent.com/VIOLINet/Vaxign-ML-docker/mas
ter/Train.sh, and allow this script to be executed by the system.
Second, prepare two protein sequences files in FASTA format
for the positive (protective antigens) and negative (nonprotec-
tive antigens). Execute the downloaded Train.sh script with the
following parameters: (a) positive data FASTA file; (b) negative
data FASTA file; (c) output directory to store the custom
Vaxign-ML model; and (d) pathogen organism type (currently
support gram-positive and -negative bacterium, and virus).

2.3 SARS-CoV-2 The Vaxign-ML was used to predict COVID-19 vaccine candidates
Example from the SARS-CoV-2 proteome [14]. As expected, the top candi-
date indicated by Vaxign-ML was the SARS-CoV-2 spike protein
(Table 3), which has been the primary target of many vaccines
currently in clinical trials. However, Vaxign-ML also predicted
several nonstructural proteins as a potential vaccine candidate.
Among which, the nsp3 protein was extensively studied in the
Vaxign-ML COVID-19 vaccine prediction study. A separate study
reported that the PL-PRO, as part of the nsp3, was associated with
innate immunity and virulence factor [15]. Overall, Vaxign and
Vaxign-ML offer a valuable resource to select vaccine candidates
and prioritize them for experimental verification.
8 Edison Ong and Yongqun He

Table 3
Predicted SARS-CoV-2 vaccine candidates using Vaxign-ML

SARS-CoV-2 Protein Vaxign-ML Score

Surface glycoprotein 97.623*
Nonstructural protein 3 (including PL-pro domain) 95.283*
Nonstructural protein 8 90.349*
Nonstructural protein 2 89.647
Nonstructural protein 4 89.647
Proteinase 3CL-PRO 89.647
Nonstructural protein 7 89.647
Nonstructural protein 9 89.647
Nonstructural protein 10 89.647
RNA-directed RNA polymerase 89.647
Helicase 89.647
Uridylate-specific endoribonuclease 89.647
0
2 -O-methyltransferase 89.647
Nucleocapsid phosphoprotein 89.647
Guanine-N7 methyltransferase 89.629
Nonstructural protein 6 89.017
Membrane glycoprotein 84.102
Host translation inhibitor 79.312
ORF3a 66.925
ORF6 33.165
ORF8 31.023
Envelope protein 23.839
ORF7a 11.199
ORF10 6.266
*Vaxign-ML predicted vaccine antigen candidates

3 Vaxign with Vaxign-ML Integration

The Vaxign-ML has been fully integrated into the Vaxign system.
The overall Vaxign system framework includes the primary Vaxign
analyses, Vaxign-ML vaccine candidate prediction, Vaxitop epitope
prediction, postprediction analysis, and pangenome analysis
(Fig. 2).
 Vaccine Design by Reverse Vaccinology and Machine Learning 9

Fig. 2 Vaxign dynamic analysis web interface

3.1 Vaxign Main 1. From the Vaxign main page (http://www.violinet.org/

Interface vaxign2), users can submit a protein sequence(s) for prediction.
The following formats are supported: (a) FASTA format,
(b) UniProtKB Protein ID, (c) NCBI Protein ID, (d) NCBI
Protein RefSeq, (e) NCBI Gene ID, and (f) FASTA File Down-
load Link. Users can also upload the protein sequence(s) in
FASTA format instead of typing into the input text field.
2. Select the pathogen organism type. The current Vaxign-ML
supports “Gram positive bacterium”, “Gram negative bacte-
rium”, “Virus”, and “Parasite”.
3. Select whether the primary Vaxign analyses include
(a) subcellular localization, (b) transmembrane helix,
(c) adhesion probability, and (d) similar to host (human,
mouse, and pig) proteins.
4. Users can choose whether the Vaxign-ML is to be included in
the pipeline.
5. Besides the primary Vaxign and Vaxign-ML analyses, the sys-
tem also supports epitope prediction via the in-house Vaxitop
program.
6. Click the “submit” button. Users can optionally provide their
email address and be notified to the email address once the
result is ready.
10 Edison Ong and Yongqun He

7. A status page will be displayed, which will be automatically

refreshed every 30 s.
8. Once the result is ready, the overall result summary page will
appear. The selected analysis results of the input protein
sequences will be displayed, which is also available for down-
load in Excel, CSV, and PDF formats. Users can perform
additional analyses to aid the vaccine candidate selection (see
Subheading 3.4).

3.2 Vaxign Analyses In addition to the prediction analyses described above, Vaxign also
After Vaccine includes additional analyses, including epitope predictions using
Candidate Prediction Vaxitop epitope prediction, EggNOG Functions and Orthologs,
and pangenome orthologs’ phylogeny. The following is the proce-
dure of how such analyses are performed.
1. Following step 5 in Subheading 3.1 or step 8 in Subheading
3.3, users can perform additional analyses to aid their selection
of vaccine candidates.
2. There are two tabs in the upper panel of the overall result
summary: “Filter Results” and “Analysis.”
3. For the “Filter Results” tab, users can apply different filters to
narrow down the candidate list. These filters include subcellu-
lar localization, number of transmembrane helices, adhesin
probability, and similarity to human/mouse/pig proteins.
Besides, users can also include or exclude candidates with
orthologs in specific pathogen strain(s). (For details of obtain-
ing the ortholog analysis, see Subheading 3.3)
4. For the “Analysis” tab, users can browse the predicted epitopes
and the corresponding population coverage of the restricted
epitopes if the Vaxitop option is selected in Subheading 3.3
step 5 (see Note 2). Users can also view the complete orthologs
table of the pangenome analysis for specific pathogen strain(s).
(For details of obtaining the pangenomic analysis, see Subhead-
ing 3.3)
5. Users can also click the individual input protein, and a web
page will be opened with detailed information of the protein,
including basic information, Vaxitop prediction (if not
selected, a new Vaxitop analysis will be performed automati-
cally), IEDB epitope (see Note 3), IEDB population coverage
(see Note 2), EggNOG Functions and Orthologs (see Note 4),
and pangenome orthologs’ phylogeny and multiple sequence
alignment. (For details of obtaining the pangenomic analysis,
see Subheading 3.3)

3.3 Vaxign 1. The Vaxign pangenome analysis can be accessed http://www.

Pangenome Analysis violinet.org/vaxign2/project.
2. For each pangenome analysis, create a new project.
 Vaccine Design by Reverse Vaccinology and Machine Learning 11

3. Enter into the project summary page, and click “Start Vaxign
Dynamic Analysis” button.
4. Users can submit protein sequence(s) for prediction, and the
following formats are supported: (a) FASTA format,
(b) UniProtKB Protein ID, (c) UniProt Proteome ID,
(d) NCBI Protein ID, (e) NCBI Protein RefSeq, (f) NCBI
Gene ID, (g) NCBI BioProject ID, (h) NCBI Nucleotide ID,
and (i) FASTA File Download Link. In addition, users can also
upload the protein sequence(s) in FASTA format instead of
typing into the input text field.
5. Enter the genome group or pathogen name (e.g., Mycobacte-
rium tuberculosis), and genome or strain name (e.g., Mycobac-
terium tuberculosis H37Rv).
6. Select the pathogen organism type, the current Vaxign-ML
supports “Gram positive bacterium”, “Gram negative bacte-
rium”, “Virus”, and “Parasite”.
7. Select whether the basic Vaxign analyses includes (a) subcellular
localization, (b) transmembrane helix, (c) adhesin probability,
and (d) similarity to host (human, mouse, and pig) proteins.
8. Users can choose whether Vaxign-ML is to be included in the
pipeline.
9. Besides the basic Vaxign and Vaxign-ML analyses, the system
also supports epitope prediction via the in-house Vaxitop
program.
10. Click the submit button. Users will be directed to the project
summary page with the Vaxign analysis status displayed.
11. Repeat steps 3–10 for the pathogen strains to be included in
the pangenome analysis.
12. Once the protein sequences of all the pathogen strains are
submitted and the corresponding Vaxign analyses are finished,
users can click the “Run Vaxign Ortholog Analysis” for the
pangenome analysis.
13. Once the analysis is completed, users can click on the “result”
button for each submitted query. Then follow the same proce-
dures (see Subheading 3.4) to include or exclude candidates
with orthologs in specific pathogen strain(s).
14. To facilitate easier Vaxign usage, a set of precomputed queries
were created, which can be accessed from the home page
(or http://www.violinet.org/vaxign2/precompute). The pre-
computed Vaxign results from either the protein level or
genome level. Users are prompted to set up preferred query
criteria; the output data are then provided. The query of pre-
computed Vaxign results is fast.
12 Edison Ong and Yongqun He

3.4 Prediction Brucella is a facultative intracellular bacterium that causes brucello-

of Brucella Vaccine sis, one of the most common zoonotic diseases in humans and
Antigens Using Vaxign various domestic and wildlife animals [16]. B. abortus,
and Vaxign-ML B. melitensis, and B. suis can cause brucellosis in cattle, goat, and
pigs, respectively. These three Brucella strains can also cause bru-
cellosis in humans. Although several live attenuated B. abortus
vaccines such as RB51 and S19 are available to protect cattle against
brucellosis, there is still no safe and effective vaccine against human
brucellosis. A major reason is related to the safety of live attenuated
whole organism vaccines. These vaccines may still cause human
brucellosis. We have previously applied Vaxign to predict Brucella
vaccine targets (Ref. PMC1539029). This study used an updated
Vaxign tool and the new Vaxign-ML tool to support Brucella
vaccine antigen prediction.
Figure 3 illustrates the results of the Brucella vaccine prediction
using Vaxign and Vaxign-ML. In this case, virulent Brucella abortus
strain 2308 has 3023 proteins. From this proteome, Vaxign-ML
predicted 482 proteins being protective vaccine antigens. Vaxign
identified 44 out of the 3033 proteins as being outer membrane
(OM) or extracellular proteins; similarly, 29 out of the 482 Vaxign-
ML predicted vaccine antigens were also identified as OM or extra-
cellular proteins. Since proteins with more than one transmem-
brane helices are often difficult to express recombinantly [6], we
further selected 43 out of the 44 OM proteins from the Vaxign
prediction with less than or equal to one transmembrane helix.
None of the 29 Vaxign-ML OM vaccine antigens predicted as
vaccine antigens had more than one transmembrane protein.
Given the adhesin criterion, 31 out of the 43 Vaxign OM proteins,

Fig. 3 B. abortus 2308 Vaxign and Vaxign-ML analyses. Both Vaxign and Vaxign-
ML were applied to analyze the genome of B. abortus strain 2308. Five different
filtering criteria were used for the filtering analysis
 Vaccine Design by Reverse Vaccinology and Machine Learning 13

B.abortus 2308 proteins (total: 3023) Vaxign-ML Predicted results (total: 482)

Unknown Unknown
Cytoplasmic
29% 15%
45%
Cytoplasmic
54%
Periplasmic
Periplasmic 12%
3%
Outer Membrane
Outer Membrane 5%
1%
Extracellular Extracellular
0.4% 3%

Cytoplasmic Cytoplasmic
Membrane Membrane
22% 11%

Fig. 4 Comparison of subcellular locations of all proteins of B. abortus strain 2308 and the 482 vaccine
candidates predicted by Vaxign-ML

and 23 out of the 29 Vaxign-ML OM vaccine antigens were pre-

dicted to be adhesins. Eventually, after filtering with no host pro-
tein similarity, we found 22 proteins that meet all filtering criteria
and ML cutoff (Fig. 3).
Figure 4 illustrates a comparison of subcellular locations of all
the 3023 proteins in the proteome and the 482 vaccine candidates
predicted by Vaxign-ML. Out of the 3023 proteins, only 0.4% are
extracellular proteins, 1% are OM proteins, and 3% are periplasmic
proteins. In comparison, out of the 482 Vaxign-ML vaccine anti-
gens, 3% are extracellular proteins, 5% are OM proteins, and 12%
are periplasmic proteins. It is clear that the OM proteins, periplas-
mic proteins, and extracellular proteins are significantly enriched in
the Vaxign-ML predicted set compared to the whole proteome. It is
also noted that there are considerably more percentages of cyto-
plasmic membrane proteins and proteins with unknown functions.
Table 4 lists all the 22 proteins predicted by Vaxign and Vaxign-
ML. This list includes three flagella related proteins (FlgE, flagellin,
and FlgK), two TonB-dependent receptors (BAB_RS22470 and
BAB_RS31825), and eight porin family proteins. The porin family
proteins form the primary group of proteins being protective anti-
gens for Brucella abortus vaccine development. The flagella-related
and TonB-dependent proteins are also favorable candidates based
on our Vaxign and Vaxign-ML analyses.

4 Notes

1. The high-quality benchmark dataset was collected and curated

[8–11]. Any duplicated data presented in the Protegen data-
base (which is used as the training data of Vaxign-ML) was
removed from this dataset to ensure a nonbiased performance
evaluation of the Vaxign-ML. As a quality check of the negative
14 Edison Ong and Yongqun He

Table 4
Predicted B. abortus vaccine candidates using Vaxign and Vaxign-ML

Vaxign-ML
Gene name Protein name score Localization
BAB_RS31580 Flagellar hook protein FlgE 94.8 EC
BR_RS15470 Flagellin 98.6 EC
BAB_RS25510 Hypothetical protein 95.7 EC
BR_RS08885 AprI/Inh family metalloprotease inhibitor, omp19 93.6 OM
flgK Flagellar hook-associated protein FlgK 91.7 OM
BAB_RS19300 LPS-assembly protein LptD 97.5 OM
BOV_RS07445 OmpW family protein 94.2 OM
BAB_RS26700 Outer membrane beta-barrel protein 91.9 OM
BR_RS06570 Outer membrane protein assembly factor BamD 92.5 OM
BAB1_0907 Peptidoglycan-binding LysM:Peptidase M23/M37 96.8 OM
BAB_RS22155 Porin 93.9 OM
BOV_RS00570 Porin family protein 91.1 OM
BR_RS00555 Porin family protein 92.3 OM
BAB_RS27875 Porin family protein 95.2 OM
BR_RS03235 Porin family protein, OMP_b-brl domain-containing 96.4 OM
protein
BAB_RS23735 Porin family protein, OMP_b-brl domain-containing 94 OM
protein
BAB_RS19090 Porin Omp2a 94.9 OM
BAB_RS19100 Porin Omp2b 96.9 OM
ybgF Tol-pal system protein YbgF 96.8 OM
BAB_RS22470 TonB-dependent receptor, Heme transporter BhuA 97.9 OM
BAB_RS31825 TonB-dependent receptor, Heme transporter BhuA 99.3 OM
BAB_RS31975 YadA-like family protein 98.1 OM
Note: The genome of B. abortus strain 2308 was used for this analysis

samples, a manual search of all negative samples in the literature

found no reported experimental evidence for the negative
samples to induce immune responses.
2. The population coverage is calculated using the IEDB popula-
tion coverage too l [17].
3. The full-length input protein will be searched against the IEDB
epitope database for reported T cell and B cell epitopes.
4. The EggNOG functions and orthologs are predicted using
EggNog mapper [18, 19].
 Vaccine Design by Reverse Vaccinology and Machine Learning 15

Acknowledgments

This work was supported by NIH-NIAID grants 1R01AI081062

and 1UH2AI13293.

References

1. Pizza M, Scarlato V, Masignani V, Giuliani tumour antigens and subunit vaccines. BMC
MM, Aricò B, Comanducci M, Jennings GT, Bioinformatics 8:4. https://doi.org/10.
Baldi L, Bartolini E, Capecchi B, Galeotti CL, 1186/1471-2105-8-4
Luzzi E, Manetti R, Marchetti E, Mora M, 9. Bowman BN, McAdam PR, Vivona S, Zhang
Nuti S, Ratti G, Santini L, Savino S, JX, Luong T, Belew RK, Sahota H, Guiney D,
Scarselli M, Storni E, Zuo P, Broeker M, Valafar F, Fierer J, Woelk CH (2011) Improv-
Hundt E, Knapp B, Blair E, Mason T, ing reverse vaccinology with a machine
Tettelin H, Hood DW, Jeffries AC, Saunders learning approach. Vaccine 29:8156–8164.
NJ, Granoff DM, Venter JC, Moxon ER, https://doi.org/10.1016/j.vaccine.2011.07.
Grandi G, Rappuoli R (2000) Identification 142
of vaccine candidates against serogroup B 10. Heinson AI, Gunawardana Y, Moesker B, Den-
meningococcus by whole-genome sequencing. man Hume CC, Vataga E, Hall Y, Stylianou E,
Science 287:1816–1820. https://doi.org/10. McShane H, Williams A, Niranjan M, Woelk
1126/science.287.5459.1816 CH (2017) Enhancing the biological relevance
2. Vernikos G, Medini D (2014) Bexsero H of machine learning classifiers for reverse vacci-
chronicle. Pathog Glob Health 108:305–311. nology. Int J Mol Sci 18:312. https://doi.org/
https://doi.org/10.1179/2047773214Y. 10.3390/ijms18020312
0000000162 11. Dalsass M, Brozzi A, Medini D, Rappuoli R
3. Folaranmi T, Rubin L, Martin SW, Patel M, (2019) Comparison of open-source reverse
MacNeil JR (2015) Use of serogroup B menin- vaccinology programs for bacterial vaccine anti-
gococcal vaccines in persons aged 10 years at gen discovery. Front Immunol 10:1–12.
increased risk for serogroup B meningococcal https://doi.org/10.3389/fimmu.2019.
disease: recommendations of the advisory com- 00113
mittee on immunization practices, 2015. 12. Ong E, Wang H, Wong MU, Seetharaman M,
MMWR Morb Mortal Wkly Rep 64:608–612 Valdez N, He Y (2020) Vaxign-ML: supervised
4. Ong E, Wong MU, He Y (2017) Identification machine learning reverse vaccinology model
of new features from known bacterial protec- for improved prediction of bacterial protective
tive vaccine antigens enhances rational vaccine antigens. Bioinformatics 36:3185–3191.
design. Front Immunol 8:1–11. https://doi. https://doi.org/10.1093/bioinformatics/
org/10.3389/fimmu.2017.01382 btaa119
5. Vivona S, Bernante F, Filippini F (2006) 13. Yang B, Sayers S, Xiang Z, He Y (2011) Prote-
NERVE: new enhanced reverse vaccinology gen: a web-based protective antigen database
environment. BMC Biotechnol 6:35. https:// and analysis system. Nucleic Acids Res
doi.org/10.1186/1472-6750-6-35 39:1073–1078. https://doi.org/10.1093/
6. He Y, Xiang Z, Mobley HLT (2010) Vaxign: nar/gkq944
the first web-based vaccine design program for 14. Ong E, Wong MU, Huffman A, He Y (2020)
reverse vaccinology and applications for vaccine COVID-19 coronavirus vaccine design using
development. J Biomed Biotechnol reverse vaccinology and machine learning.
2010:1–15. https://doi.org/10.1155/2010/ Front Immunol 11:1581
297505 15. Shin D, Mukherjee R, Grewe D, Bojkova D,
7. Rizwan M, Naz A, Ahmad J, Naz K, Obaid A, Baek K, Bhattacharya A, Schulz L, Widera M,
Parveen T, Ahsan M, Ali A (2017) VacSol: a Mehdipour AR, Tascher G, Geurink PP,
high throughput in silico pipeline to predict Wilhelm A, van der Heden van Noort GJ,
potential therapeutic targets in prokaryotic Ovaa H, Müller S, Knobeloch K-P,
pathogens using subtractive reverse vaccinol- Rajalingam K, Schulman BA, Cinatl J,
ogy. BMC Bioinformatics 18:1–7. https:// Hummer G, Ciesek S, Dikic I (2020) Papain-
doi.org/10.1186/s12859-017-1540-0 like protease regulates SARS-CoV-2 viral
8. Doytchinova IA, Flower DR (2007) VaxiJen: a spread and innate immunity. Nature 587
server for prediction of protective antigens,
16 Edison Ong and Yongqun He

(7835):657–662. https://doi.org/10.1038/ DR, Letunic I, Rattei T, Jensen LJ, Von

s41586-020-2601-5 Mering C, Bork P (2019) EggNOG 5.0: a
16. Xiang Z, Zheng W, He Y (2006) BBP: Brucella hierarchical, functionally and phylogenetically
genome annotation with literature mining and annotated orthology resource based on 5090
curation. BMC Bioinformatics 7:347. https:// organisms and 2502 viruses. Nucleic Acids Res
doi.org/10.1186/1471-2105-7-347 47:D309–D314. https://doi.org/10.1093/
17. Bui HH, Sidney J, Dinh K, Southwood S, nar/gky1085
Newman MJ, Sette A (2006) Predicting popu- 19. Huerta-Cepas J, Forslund K, Coelho LP,
lation coverage of T-cell epitope-based diag- Szklarczyk D, Jensen LJ, Von Mering C, Bork
nostics and vaccines. BMC Bioinformatics P (2017) Fast genome-wide functional annota-
7:1–5. https://doi.org/10.1186/1471-2105- tion through orthology assignment by
7-153 eggNOG-mapper. Mol Biol Evol
18. Huerta-Cepas J, Szklarczyk D, Heller D, Her- 34:2115–2122. https://doi.org/10.1093/
nández-Plaza A, Forslund SK, Cook H, Mende molbev/msx148
 Chapter 2

Application of Reverse Vaccinology and Immunoinformatic

Strategies for the Identification of Vaccine Candidates
Against Shigella flexneri
Chiuan Yee Leow , Candy Chuah , Abu Bakar Abdul Majeed ,
Norazmi Mohd Nor , and Chiuan Herng Leow

Abstract
Reverse vaccinology (RV) was first introduced by Rappuoli for the development of an effective vaccine
against serogroup B Neisseria meningitidis (MenB). With the advances in next generation sequencing
technologies, the amount of genomic data has risen exponentially. Since then, the RV approach has widely
been used to discover potential vaccine protein targets by screening whole genome sequences of pathogens
using a combination of sophisticated computational algorithms and bioinformatic tools. In contrast to
conventional vaccine development strategies, RV offers a novel method to facilitate rapid vaccine design and
reduces reliance on the traditional, relatively tedious, and labor-intensive approach based on Pasteur”s
principles of isolating, inactivating, and injecting the causative agent of an infectious disease. Advances in
biocomputational techniques have remarkably increased the significance for the rapid identification of the
proteins that are secreted or expressed on the surface of pathogens. Immunogenic proteins which are able
to induce the immune response in the hosts can be predicted based on the immune epitopes present within
the protein sequence. To date, RV has successfully been applied to develop vaccines against a variety of
infectious pathogens. In this chapter, we apply a pipeline of bioinformatic programs for identification of
Shigella flexneri potential vaccine candidates as an illustration immunoinformatic tools available for RV.

Key words Antigens, Bacteria, Epitope, Immunoinformatics, Outer membrane protein (OMP),
Shigella flexneri, Reverse vaccinology (RV), Vaccine

1 Introduction

Shigellosis is an acute inflammatory bowel disease caused by the

gram-negative intracellular enterobacterial genus, Shigella. Food
and water sources that are contaminated with human or animal
waste contribute primarily toward the transmission of Shigella
[1]. Shigella spp. are increasingly exhibiting resistance against cur-
rently available antibiotics, and it was recently reported that tetra-
cycline and trimethoprim–sulfamethoxazole are no longer used for

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_2,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

17
18 Chiuan Yee Leow et al.

treatment of severe diarrhea and dysentery in some areas [2]. Thus,

development of a protective vaccine for the prevention of Shigella
infection, especially in epidemic areas, is a high priority. As yet, no
practical vaccine for shigellosis has been licensed. Vaccine candi-
dates currently in clinical trial are either not adequately attenuated
or have proven to be less immunogenic in the host [3]. In order to
develop an effective vaccine for long term protection, it is essential
to accurately identify the immunogenic protective antigens. Devel-
opment of vaccines through conventional techniques can take dec-
ades. With the increasing availability of whole genome sequences,
in silico analysis of genomic data can be used to identify potential
vaccine targets in a process termed “reverse vaccinology” (RV) [4].
In the RV strategy, candidate antigens likely to elicit protective
immune responses are identified at the gene level. Following
genome-wide analysis, a high-throughput cloning and expression
platform is used to generate and purify recombinant proteins pre-
dicted as outer membrane, invasion, and other virulence-related
antigens. The purified recombinant antigens are then evaluated
in vivo and in vitro to shortlist protective antigens for vaccine
research. The RV approach was firstly introduced by Rappuoli and
his colleagues to develop a safe and protective vaccine against
serogroup B Neisseria meningitidis (MenB) [5]. With an expanding
wealth of genomic data, the RV approach has been used for the
discovery of potential vaccine candidates in Streptococcus pneumo-
niae [6], Campylobacter jejuni [7], Shigella flexneri [8], and
COVID-19 [9], among others.
In 2008, a genome-based computational vaccine discovery tool
known as Vaxign was introduced by He and colleagues [10]. Vaxign
is a vaccine target prediction and analysis system developed based
on the principles of RV and is composed of a sophisticated compu-
tational pipeline that utilizes bioinformatic technology to find
potential antigenic proteins from genome data for vaccine develop-
ment [10, 11]. The major predicted features include identification
of subcellular location of proteins, transmembrane domains, adhe-
sion probability, sequence similarity to host proteome, and MHC
class I and II epitope binding. Among these features, subcellular
localization is considered as one of the main criteria for target
prediction [10, 11]. Vaxign is a part of web-based system called
Vaccine Investigation and Online Information Network (VIOLIN,
http://www.violinet.org) and has been widely used for the identifi-
cation of vaccine targets against various pathogens including Bru-
cella spp. [12, 13], Acinetobacter baumannii [14], and
Mycobacterium tuberculosis [15].
The immunogenicity of an antigen is associated with its ability
to interface with the humoral (B-cell) and cellular (T-cell) immune
systems. A vaccine consisting of both B- and T-cell epitopes is of
importance to effectively elicit strong immune responses for long term
protection [16, 17]. The integration of genomic, proteomic, and
 Reverse Vaccinology for S. flexneri Antigen Discovery 19

bioinformatic strategies has paved way for the development of

immunoinformatics. To date, a number of B- and T-cell epitope
prediction tools have been developed using various computational
algorithms. This has led us toward the evolution of designing
modern vaccines based on epitope-focused recombinant antigens
[18, 19].
In this chapter, we describe our pipeline for identification of the
most conserved and immunogenic outer membrane proteins
(OMPs) from the S. flexneri genome, which involves identification
of vaccine targets using Vaxign (see Subheading 3.1), analysis of
conserved identity with other strains of the same species (see Sub-
heading 3.2) and human homologs of the predicted proteins (see
Subheading 3.3), analysis of antigenicity using VaxiJen v2.0 (see
Subheading 3.4), prediction of linear B-cell epitopes using
BCPREDS (see Subheading 3.5), and prediction of HLA Class I
and Class II T-cell epitopes using HLApred (see Subheading 3.6).
An overview of the reverse vaccinology steps adopted in this chapter
is summarized in Fig. 1.

Shigella flexneri 2a strain 2457T

Proteome
Vaxign (4061 proteins)
Protein subcellular localization
Transmembrane helices
Adhesin probability Outer membrane protein
(36 proteins)
Vaxijen v2.0
Antigenicity prediction
Antigenicity
(29 proteins)
BLASTP (NCBI)

Conservation among Shigella strains

(5 proteins)

BLASTP (UniProt)

Identification of human homologs

(5 proteins)

Five (5) OMPs selected as potential vaccine candidates

NP_836212 NP_837825 NP_838270 NP_838556 NP_838365

B-cell epitope prediction

BCPreds

T-cell epitope prediction

HLAPred

Fig. 1 Flowchart of the use of reverse vaccinology approach for the prediction of vaccine candidates against
Shigella flexneri
20 Chiuan Yee Leow et al.

2 Materials

1. Windows or Mac computer.

2. Google Chrome or other web browser (see Note 1).
3. Vaxign: Vaccine Design Pipeline [10, 11].http://www.violinet.
org/vaxign/
4. BLASTP: Basic Local Alignment Search Tool for Protein
Sequences (NCBI) [20, 21].https://blast.ncbi.nlm.nih.gov/
Blast.cgi?PAGE¼Proteins
https://www.uniprot.org/blast/
5. VaxiJen v2.0: Prediction of Protective Antigens and Subunit
Vaccines [22].http://www.ddg-pharmfac.net/vaxijen/
VaxiJen/VaxiJen.html
6. BCPREDS: B-cell Epitope Prediction Server [23, 24].http://
ailab-projects1.ist.psu.edu:8080/bcpred/
7. HLAPred: Identification and Prediction of HLA Class I and
Class II T-cell Epitope [25, 26].http://crdd.osdd.net/
raghava/hlapred/

3 Methods

3.1 Identification 1. Point the web browser to the Vaxign website at http://www.
of Vaccine Targets violinet.org/vaxign/. The Vaccine Design section of the Vax-
Using Vaxign ign page appears (see Note 2).
3.1.1 Select a Genome 2. In the “Vaxign Query” section (Fig. 2a), first go to “Select a
Genome Group.” Select <Shigella (5) > in the drop-down
menu. The number (5) means that Vaxign contains five gen-
omes corresponding to Shigella species (see Note 3).
3. Next, go to “Select a Genome.” Select <Shigella flexneri 2a
strain 2457T genome> in the drop-down menu (see Note 4).
4. Keep the default settings for “Keywords” and “Sort by”
options. Proceed to “Filter Options.”

3.1.2 Select Filter 1. Set subcellular localization by selecting <Outer Membrane> in

Options the scrollable selection menu (see Notes 5 and 6).
2. Restrict the number of transmembrane helices. The default
setting for protein transmembrane number is less than or
equal to 1. Check the box to include proteins having less than
one transmembrane helix (see Note 7).
3. For adhesin probability analysis, the default cutoff is 0.51.
Check the box to include proteins having an adhesin probabil-
ity score equal or more than 0.51 (see Note 8).
 Reverse Vaccinology for S. flexneri Antigen Discovery 21

Fig. 2 Vaxign query submission for genome-based vaccine candidate prediction. (a) Vaxign query web
interface. (b) An example of potential vaccine candidates predicted using Vaxign pipeline

4. The remaining Filter Options (i.e., options 4–9 shown in

Fig. 2a) for analysis of protein orthologue comparisons,
sequence similarity to host proteome, and epitope prediction
were not included in this case and they were substituted using
other online software described in the following section (see
Notes 9–13).
5. Click the “Submit” button and wait. The result of analysis is
shown in Fig. 2b.
22 Chiuan Yee Leow et al.

6. In the result page, check one or more box at the left-most side
of the table to select protein(s) for further display.
7. Click link(s) in the “Protein Accession” for detailed informa-
tion, which will be displayed on a new page. The desired
FASTA protein sequences can be saved to appropriate files, or
copied and pasted, as required into subsequent analysis plat-
forms, such as BLASTP, below (see Note 14).
8. If biological function of the protein is of interest, click the
“Run Vaxign COG analysis for all records” link above the
Table. A new page will be displayed. If functional orthologues
of the protein are of interest, click “Show Ortholog Table” link
above the Table. A new page will be displayed.
9. Click the link above the table for “Export all records to MS
Excel file” (for all proteins are selected) or “Export selected to
MS Excel file” (if only certain proteins are selected).
10. Save the MS Excel file on a working directory.

3.2 Identification 1. Point the internet browser to the NCBI BLASTP website at
of Conserved Identity https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE¼Proteins
with Other Shigella 2. In the “Enter Query Sequence” box (Fig. 3a),
Strains (a) Copy and paste a either a protein accession number or the
protein sequence (either raw sequence or FASTA-
formatted sequences of proteins identified in Subheading
3.1.2, step 7) into the form field.
(b) Alternatively, the user can upload a protein sequence file
saved in FASTA-format.
(c) Entering a “Job Title” is optional.
(d) Ignore checking “Align two or more sequences” (see Note
15).
(e) Proceed to next section.
3. In the “Choose Search Set” section,
(a) Select the <non-redundant protein sequences (nr) > data-
base in the drop-down menu.
(b) Enter organism(s) to search against. In this case, <Shigella
sonnei (taxid:624)>; <Shigella dysenteriae (taxid:622)>;
and < Shigella boydii (taxid:621) > were selected (see
Note 16).
(c) Do not check the “Exclude” boxes.
(d) Proceed to next section.
4. For “Program Selection,” use the default blastp (protein–pro-
tein BLAST) program.
5. Click the “BLAST” button at the bottom of the page. The
results will appear as in Fig. 3b (see Note 17).
 Reverse Vaccinology for S. flexneri Antigen Discovery 23

Fig. 3 An example of BLASTP protein sequence similarity analysis among Shigella spp. proteomes. (a) Protein
sequence submission interface. Outcome of BLASTP analysis. (b) The result showing NP_838365 of S. flexneri
is orthologous in S. dysenteriae
24 Chiuan Yee Leow et al.

3.3 Identification 1. Point the internet browser to the BLAST function within the
of Human Homologs UniProt website at https://www.uniprot.org/blast [27].
2. Copy and paste a protein sequence (raw data or FASTA-
formatted sequences of proteins identified in Subheading
3.1.2, step 7) into the form field (Fig. 4a).

Fig. 4 An example of protein sequence similarity analysis against human whole proteome using BLAST tool in
UniProt. (a) Protein sequence submission interface. (b) Outcome of BLAST analysis showing NP_838365 has
no similar protein sequence found in human proteome
 Reverse Vaccinology for S. flexneri Antigen Discovery 25

3. To identify the most related human proteins, change the pro-

gram parameters in the drop-down menus under the form as
follows.
(a) Target database: Human.
(b) E-Threshold: 0.1.
(c) Matrix: Auto.
(d) Filtering: None.
(e) Gapped: No.
(f) Hits: 250.
4. Click the “Run BLAST” button at the bottom of the page. The
results appear as in Fig. 4b (see Note 18).

3.4 Antigenicity 1. Point the internet browser to the VaxiJen v2.0 website at
Analysis http://www.ddg-pharmfac.net/vaxijen/VaxiJen/VaxiJen.
html
2. At “Enter a PROTEIN sequence here,” copy and paste a
protein sequence (plain format) in the form field. Alternatively,
a file containing multiple protein sequences saved in FASTA-
format can be submitted using the “Choose file” button
(Fig. 5a).
3. At “Select a TARGET ORGANISM,” select <Bacteria> in the
scrollable selection menu.
4. Keep default setting for the “THRESHOLD” (cutoff at 0.4)
(see Note 19).
5. Check the “Sequence Output” box. Check the other provided
boxes if user would like to learn more results for the submitted
protein sequence (see Note 20).
6. Click the “Submit” button at the bottom of the page. The
results appear as in Fig. 5b.
7. Five outer membrane proteins predicted to be potential vaccine
candidates against S. flexneri were selected for further analysis,
and the relevant results for these are shown in Table 1. Each
protein was subsequently analyzed to identify potential B- and
T-cell epitopes.

3.5 Prediction 1. Point the internet browser to the predictions page of the
of Linear B-Cell BCPREDS website at http://ailab-projects1.ist.psu.
Epitopes edu:8080/bcpred/predict.html
2. Copy and paste a protein sequence (plain format) in the form
field (Fig. 6a).
3. At “Methods,” select the “Fixed length epitope prediction”
option <BCPred> with a window (epitope length) of
20 amino acids, as set by the default parameters (see Note 21).
4. At “Specificity,” select <75% > in the drop-down menu (see
Note 22).
26 Chiuan Yee Leow et al.

Fig. 5 VaxiJen v2.0 query submission for protein antigenicity prediction. (a) VaxiJen v2.0 query web interface.
(b) An example of result page. The protein (NP_838365) was predicted to be antigenic. Threshold for this
model is 0.4; overall antigen prediction is 0.6201 (predicted to be probable ANTIGEN)

5. Check the box for “report only non-overlapping epitopes.”

6. Click the “Submit query” button at the bottom of the page.
The result of analysis is shown in Fig. 6b.

3.6 Prediction of HLA 1. Point the internet browser to the HLApred website at http://
Class I and Class II crdd.osdd.net/raghava/hlapred/ (see Note 23).
T-Cell Epitope 2. Copy and paste a protein sequence (plain format) into the
“Paste Your Protein Sequence” form field, or use the “Upload
Sequence File” option to choose a saved sequence file (Fig. 7).
 Reverse Vaccinology for S. flexneri Antigen Discovery 27

Table 1
List of potential vaccine candidates against S. flexneri predicted using RV approach

Amino acid sequence homology

percentage in Shigella strains
Antigenicity score Human
Shigella Shigella Shigella (Vaxijen) proteome
No. Protein name boydii dysentery sonnei Threshold > 0.4 homology
1 NP_836212 99% 99% 99% 0.7454 No significant
Outer membrane similarity
receptor found
2 NP_837825 93% 93% 95% 0.6986 No significant
Outer membrane similarity
porin protein C found
3 NP_838270 99% 100% 99% 0.7117 No significant
Lipoprotein NlpD similarity
found
4 NP_838556 99% 99% 99% 0.5494 No significant
Outer membrane similarity
channel protein found
5 NP_838365 99% 99% 99% 0.6168 No significant
Lipoprotein similarity
found

3. At “Choose Format of input sequence,” select <Amino acids in

single letter code>. Alternatively, standard formats (e.g.,
EMBL, PIR, FASTA, GCG) are optionally chosen if the pro-
tein sequence submitted is based on FASTA-format.
4. At “Select HLA allele,” select multiple alleles by pressing the
“Alt” (Window OS) or “cmd” (MACS OS) key followed by
“mouse click.” Alternatively, select “ALL” if the entire set of
alleles are to be analyzed.
5. To include HLA-I and HLA-II binders in the analysis, click the
Class I and Class II alleles through “Both” option in
selection box.
6. At “Method,” select the <Predict Binder> as set by default.
7. At “Prediction Parameter,” select <3% > in the “Threshold”
drop-down menu; select <HTML Mapping> in the “Display
Format”; and “Display Top <4 > Peptides” (see Note 24).
8. Click the “Submit sequence” button at the bottom of the page.
The result of analysis for HLA class I and HLA class II binders
are shown in Figs. 8 and 9, respectively.
28 Chiuan Yee Leow et al.

Fig. 6 BCPreds server for B-cell linear epitope prediction. (a) Cover page of BCPreds. (b) An example of result
page for the prediction of B-cell epitope performed using BCPreds. Table showing seven 20-mers B-cell
epitopes were predicted for NP_838365. Peptide (20 mers) with score >0.75 were predicted to be potential
B-cell linear epitope
 Reverse Vaccinology for S. flexneri Antigen Discovery 29

Fig. 7 A web interface of HLAPred. A server for prediction and identification of HLA class I and class II T-cell
epitope using quantitative matrix-based prediction method

4 Notes

1. We have tested and confirmed all the programs are compatible

with Google Chrome; Internet Explorer or Mozilla Firefox. We
have chosen to use Google Chrome in our protocol. Further-
more, the type of computer and operating system do not affect
the online programs used in the protocol.
2. Two methods are available for running Vaxign: (a) Vaxign
query is for precomputed results (b) Dynamic analysis is for
use of own sequence input (sequence format can be chosen in
the drop-down menu).
3. The current genomes available in the Vaxign are limited to 350.
If a genome of interest is not currently available in the pre-
computed Vaxign database, the user may request addition of
the genome by contacting the Vaxign server developer on the
provided page (http://www.violinet.org/contact_us.php).
30 Chiuan Yee Leow et al.

Fig. 8 An example of result page (HTML-II view) for HLA class I T-cell epitope predicted using HLAPred.
Predicted binders (9-mer) are displayed as blue colored region, with P1 anchor or the starting residue of each
predicted binding frame as red colored

Fig. 9 An example of result page (HTML-II view) for HLA class II T-cell epitope predicted using HLAPred.
Predicted binders (9-mer) are displayed as blue colored region, with P1 anchor or the starting residue of each
predicted binding frame as red colored
 Reverse Vaccinology for S. flexneri Antigen Discovery 31

4. Sequenced genomes used in Vaxign were retrieved from NCBI

RefSeq database and used in Vaxign vaccine prediction
pipeline.
5. Subcellular localization provides important information for
diagnostics, drug, and vaccine design. PSORTb (v.3.0) [28],
a subcellular localization prediction tool, is included in Vaxign
vaccine prediction pipeline in order to identify the protein
functional annotation based on the protein localization in the
bacteria.
6. If the user is interested to explore the total proteomes in the
chosen genome, the user can choose “Any Localization” in the
scrollable selection menu. Alternatively, the user can select
more than two options in the drop-down menu by pressing
the “Ctrl” key + mouse left click.
7. The prediction of transmembrane helices and topology of pro-
teins in Vaxign pipeline is performed using HMMTOP [29]. A
protein comprising more than one transmembrane helix is
considered to have difficulty in protein production and purifi-
cation when it is expressed as a recombinant protein. It is wise
to exclude those proteins with multiple transmembrane span-
ning regions from the selection process.
8. The prediction of adhesion characteristic of the protein is
performed using SPAAN software [30]. Proteins predicted to
have adhesin capacities are known to be involved in host-
bacterial interaction. These proteins play vital role in bacterial
invasion. The adhesin-like proteins appear to be potential vac-
cine candidates since they can be recognized by immune cells in
the host [31].
9. Analysis of protein orthologue comparisons and sequence sim-
ilarity to host proteome were not included in the protocol
described in this chapter because the Vaxign server only pro-
vides qualitative results if these options are chosen. In order to
analyze the antigen similarity with higher stringency settings
for Expectation (E) values and percent identity, it is recom-
mended to replace this step using alternative online servers
such as BLASTP on NCBI or UniProt, as described in this
chapter. If stringency or quantitation of the sequence similarity
comparison is not required, user can proceed with the built-in
options in Vaxign.
10. Epitope prediction option at Vaxign is only limited to MHC
class I and MHC class II analysis. This option was not used in
our current protocol, where the shortlisted antigens are sub-
jected to B-cell epitope prediction followed by T-cell epitope
prediction using alternative online software. However, the user
can use the built-in option in Vaxign if only T-cell epitopes are
to be analyzed.
32 Chiuan Yee Leow et al.

11. “Have Orthologs In” is used for identifying protein conserved

regions among a selected list of strains corresponding to the
specific genome, for example, Shigella as in this case.
12. “No Similarity to Host Proteins” is used to exclude those
protein targets that present either in humans, mouse and pigs.
13. “MHC class I and II epitope prediction” is performed using
Vaxitop, a built-in vaccine epitope prediction and analysis sys-
tem based on the principle of reverse vaccinology. This feature
is available after initial protein filtering.
14. Even though the Vaxign pipeline saves time for users by short-
listing the proteins of interest, the user still needs to analyze the
shortlisted proteins manually for best results. Compared to the
total proteomes which are likely made up of >3000 proteins,
dealing with selected proteins in the result page is compara-
tively time saving and manageable. We highly recommend the
user to click into the details of each protein before proceeding
to the next step.
15. Check this box if alignment of query with a known subject is
required. This option is also known as BLAST 2 Sequences.
16. Initially, a single box is available to enter choice of organism. If
more organisms are to be selected, the user can add in addi-
tional rows by clicking the “+” button on the right.
17. Conserved immunogenic antigens are promising vaccine can-
didates provided they are able to induce broad-spectrum pro-
tection against bacterial pathogens [32]. BLAST is a promising
tool for the comparative analysis of the sequence similarity
qualitatively and quantitatively. In general, the similarity
between the sequences corresponds to the E value. The lower
the E value, the more likely the match is to be significant. E
values between 0.1 and 10 are generally more stringent (less
false positive), whereas over 10 are unlikely to have biological
significance. In this case, E value <0.1 was chosen. This option
can be found in the parameter setting at the bottom of the
page. In the result page, identity percentages equal to 80% or
higher are considered to be orthologues. Proteins sharing
identity percentages <80% are considered not conserved and
are eliminated from the dataset.
18. The proteins having no significant identity (<35%) are consid-
ered to be sufficiently distant to the human proteome and will
not interfere with normal host immune mechanism when used
as a vaccine candidate [8].
19. For the antigenicity prediction by VaxiJen, the default thresh-
old set for bacterial antigen analysis is 0.4. The percentage
threshold parameter allows the user to select for different
 Reverse Vaccinology for S. flexneri Antigen Discovery 33

stringency levels. As the threshold window increases (i.e., >

0.4), the specificity increases and the sensitivity decreases.
20. VaxiJen is a server designed to analyze protective antigens
using alignment-independent prediction. The server contains
models derived by auto cross covariance (ACC) preprocessing
of amino acids properties. It has been trained to identify poten-
tial antigenic epitopes for bacterial, viral, tumor, parasite, and
fungal origin. By default, the serve displays only the Sequence
Output. In this mode, the threshold value, protein sequence
and prediction result are displayed. If the user would like to
learn more about the antigen ACC value, ACC Output can be
selected. Alternatively, user can choose the Summary Output.
In this mode, only the prediction result is provided.
21. BCPREDS server allows the user to select among three predic-
tion methods: (a) BCPred, (b) AAP method, and (c) FBCPred.
It is recommended to use at least two of these methods, as
identification of overlapping epitopic regions identified using
multiple B-cell epitope prediction algorithms enhances the
accuracy of prediction.
22. Increasing specificity percentage corresponds to a high strin-
gency prediction.
23. HLApred is a quantitative matrix-based method that allows the
prediction of HLA binding sites in an antigenic sequence for
51 HLA class-I alleles and 36 HLA class-II alleles. Instead of
choosing all alleles, user can also focus on the superalleles
which are reported to cover more than 99% of the human
population. For HLA Class I, supertype HLA alleles are
“HLA-A*01:01,” “HLA-A*02:01,” “HLA-A*03:01,”
“HLA-A*24:02,” “HLA-B*07:02,” and “HLA-B*44:03”
whereas for HLA class II, supertype HLA alleles are “HLA-
DRB1*01:01,” “HLA-DRB1*03:01,” “HLA-DRB1*04:01,”
“HLA-DRB1*07:01,” “HLA-DRB1*08:01,” “HLA-
DRB1*11:01,” “HLA-DRB1*13:01,” and “HLA-
DRB1*15:01” [33]. The promiscuous epitopes are those
which bind with many HLA alleles. For instance, in the exam-
ple protein (NP_838365) shown in Fig. 7, region 59 YRISRT
TGT67 has binding affinity for most of the HLA class II alleles
used in the query (see results shown in Fig. 9), suggesting it
can be a potential T-cell epitope candidate. However, due to
discrepancies between different algorithms, users are recom-
mended to employ >2 alternative epitope prediction algo-
rithms to increase the likelihood of the predicted epitope(s).
24. The percentage threshold parameter allows the user to select
for different stringency levels; a lower threshold corresponds to
a high stringency prediction.
34 Chiuan Yee Leow et al.

Acknowledgments

The authors acknowledge the funding support provided for this

work by USM Research University (Individual) Grant (no.: 1001.
CIPPM.8011078) and The Malaysian Ministry of Higher Educa-
tion of the Higher Institutions Centre of Excellence Program
under Grant (no: 311/CIPPM/4401005). The authors would
like to thank Associate Professor Dr. Oliver He Yongqun at Uni-
versity of Michigan Medical School for providing his excellent
guidance on the Vaxign Vaccine Design platform.

References

1. Kotloff KL, Riddle MS, Platts-Mills JA et al potential cellular- and antibody-dependent

(2018) Shigellosis. Lancet 391 vaccine candidates. Clin Exp Vaccine Res 9
(10122):801–812. https://doi.org/10.1016/ (1):15–25. https://doi.org/10.7774/cevr.
S0140-6736(17)33296-8 2020.9.1.15
2. Banga Singh KK, Ojha SC, Deris ZZ, Rahman 9. Ong E, Wong MU, Huffman A, He Y (2020)
RA (2011) A 9-year study of shigellosis in COVID-19 coronavirus vaccine design using
Northeast Malaysia: antimicrobial susceptibil- reverse vaccinology and machine learning.
ity and shifting species dominance. J Public Front Immunol 11:1581. https://doi.org/
Health 19(3):231–236. https://doi.org/10. 10.3389/fimmu.2020.01581
1007/s10389-010-0384-0 10. He Y, Xiang Z, Mobley HL (2010) Vaxign: the
3. Ashkenazi S, Cohen D (2013) An update on first web-based vaccine design program for
vaccines against Shigella. Ther Adv Vaccines 1 reverse vaccinology and applications for vaccine
(3):113–123. https://doi.org/10.1177/ development. J Biomed Biotechnol
2051013613500428 2010:297505. https://doi.org/10.1155/
4. Tettelin H (2009) The bacterial pan-genome 2010/297505
and reverse vaccinology. In: de Reuse H, Beres- 11. Xiang Z, He Y (2013) Genome-wide predic-
will S (eds) Microbial pathogenomics, Genome tion of vaccine targets for human herpes sim-
Dyn, vol 6. Karger, Basel, pp 35–47. https:// plex viruses using Vaxign reverse vaccinology.
doi.org/10.1159/000235761 BMC Bioinformatics 14 Suppl 4:S2. https://
5. Pizza M, Scarlato V, Masignani V et al (2000) doi.org/10.1186/1471-2105-14-S4-S2
Identification of vaccine candidates against ser- 12. He Y (2012) Analyses of Brucella pathogenesis,
ogroup B meningococcus by whole-genome host immunity, and vaccine targets using sys-
sequencing. Science 287(5459):1816–1820. tems biology and bioinformatics. Front Cell
https://doi.org/10.1126/science.287.5459. Infect Microbiol 2:2. https://doi.org/10.
1816 3389/fcimb.2012.00002
6. Mamede LD, de Paula KG, de Oliveira B et al 13. Gomez G, Pei J, Mwangi W, Adams LG, Rice-
(2020) Reverse and structural vaccinology Ficht A, Ficht TA (2013) Immunogenic and
approach to design a highly immunogenic invasive properties of Brucella melitensis 16M
multi-epitope subunit vaccine against Strepto- outer membrane protein vaccine candidates
coccus pneumoniae infection. Infect Genet Evol identified via a reverse vaccinology approach.
85:104473. https://doi.org/10.1016/j. PLoS One 8(3):e59751. https://doi.org/10.
meegid.2020.104473 1371/journal.pone.0059751
7. Gupta N, Kumar A (2020) Designing an effi- 14. Garg N, Singh R, Shukla G, Capalash N,
cient multi-epitope vaccine against Campylo- Sharma P (2016) Immunoprotective potential
bacter jejuni using immunoinformatics and of in silico predicted Acinetobacter baumannii
reverse vaccinology approach. Microb Pathog outer membrane nuclease, NucAb. Int J Med
147:104398. https://doi.org/10.1016/j. Microbiol 306(1):1–9. https://doi.org/10.
micpath.2020.104398 1016/j.ijmm.2015.10.005
8. Leow CY, Kazi A, Hisyam Ismail CMK et al 15. Hossain MS, Azad AK, Chowdhury PA,
(2020) Reverse vaccinology approach for the Wakayama M (2017) Computational identifi-
identification and characterization of outer cation and characterization of a promiscuous
membrane proteins of Shigella flexneri as T-cell epitope on the extracellular protein 85B
 Reverse Vaccinology for S. flexneri Antigen Discovery 35

of Mycobacterium spp. for peptide-based sub- 25. Singh H, Raghava GP (2003) ProPred1: pre-
unit vaccine design. BioMed Res Int diction of promiscuous MHC class-I binding
2017:4826030. https://doi.org/10.1155/ sites. Bioinformatics 19(8):1009–1014.
2017/4826030 https://doi.org/10.1093/bioinformatics/
16. He Y, Rappuoli R, De Groot AS, Chen RT btg108
(2010) Emerging vaccine informatics. J 26. Singh H, Raghava GP (2001) ProPred: predic-
Biomed Biotechnol 2010:218590. https:// tion of HLA-DR binding sites. Bioinformatics
doi.org/10.1155/2010/218590 17(12):1236–1237. https://doi.org/10.
17. Moise L, Cousens L, Fueyo J, De Groot AS 1093/bioinformatics/17.12.1236
(2011) Harnessing the power of genomics and 27. UniProt C (2019) UniProt: a worldwide hub
immunoinformatics to produce improved vac- of protein knowledge. Nucleic Acids Res 47
cines. Expert Opin Drug Discov 6(1):9–15. (D1):D506–D515. https://doi.org/10.
https://doi.org/10.1517/17460441.2011. 1093/nar/gky1049
534454 28. Yu NY, Wagner JR, Laird MR et al (2010)
18. Kazi A, Chuah C, Majeed ABA et al (2018) PSORTb 3.0: improved protein subcellular
Current progress of immunoinformatics localization prediction with refined localization
approach harnessed for cellular- and antibody- subcategories and predictive capabilities for all
dependent vaccine design. Pathog Glob Health prokaryotes. Bioinformatics 26
112(3):123–131. https://doi.org/10.1080/ (13):1608–1615. https://doi.org/10.1093/
20477724.2018.1446773 bioinformatics/btq249
19. Kazi A, Hisyam Ismail CMK, Anthony AA et al 29. Tusnady GE, Simon I (2001) The HMMTOP
(2020) Designing and evaluation of an transmembrane topology prediction server.
antibody-targeted chimeric recombinant vac- Bioinformatics 17(9):849–850. https://doi.
cine encoding Shigella flexneri outer mem- org/10.1093/bioinformatics/17.9.849
brane antigens. Infect Genet Evol 80:104176. 30. Sachdeva G, Kumar K, Jain P, Ramachandran S
https://doi.org/10.1016/j.meegid.2020. (2005) SPAAN: a software program for predic-
104176 tion of adhesins and adhesin-like proteins using
20. Mount DW (2007) Using the Basic Local neural networks. Bioinformatics 21
Alignment Search Tool (BLAST). CSH Proto- (4):483–491. https://doi.org/10.1093/bioin
cols 2007:pdb top17. https://doi.org/10. formatics/bti028
1101/pdb.top17 31. Raynes JM, Young PG, Proft T et al (2018)
21. Altschul SF, Gish W, Miller W, Myers EW, Lip- Protein adhesins as vaccine antigens for
man DJ (1990) Basic local alignment search Group A Streptococcus. Pathog Dis 76(2).
tool. J Mol Biol 215(3):403–410. https:// https://doi.org/10.1093/femspd/fty016
doi.org/10.1016/S0022-2836(05)80360-2 32. Nagy G, Emody L, Pal T (2008) Strategies for
22. Doytchinova IA, Flower DR (2007) VaxiJen: a the development of vaccines conferring broad-
server for prediction of protective antigens, spectrum protection. Int J Med Microbiol 298
tumour antigens and subunit vaccines. BMC (5–6):379–395. https://doi.org/10.1016/j.
Bioinformatics 8:4. https://doi.org/10. ijmm.2008.01.012
1186/1471-2105-8-4 33. De Groot AS, Ardito M, McClaine EM et al
23. El-Manzalawy Y, Dobbs D, Honavar V (2008) (2009) Immunoinformatic comparison of
Predicting linear B-cell epitopes using string T-cell epitopes contained in novel swine-origin
kernels. J Mol Recognit 21(4):243–255. influenza A (H1N1) virus with epitopes in
https://doi.org/10.1002/jmr.893 2008-2009 conventional influenza vaccine.
24. El-Manzalawy Y, Dobbs D, Honavar V (2008) Vaccine 27(42):5740–5747. https://doi.org/
Predicting flexible length linear B-cell epitopes. 10.1016/j.vaccine.2009.07.040
Comput Syst Bioinformatics Conf 7:121–132
 Chapter 3

Purification of Prospective Vaccine Antigens from

Gram-Positive Pathogens by Immunoprecipitation
Mark Reglinski

Abstract
Immunoprecipitation is an affinity purification technique that exploits the highly specific interactions
formed between antibodies and their cognate antigens to purify molecules of interest from complex
biological solutions. The generation of an effective humoral response provides protection against a wide
range of gram-positive pathogens, and thus immunoprecipitation using antibodies purified from immune
humans or animals provides a simple but effective means of isolating prospective vaccine antigens from
fractionated bacterial cells for downstream identification. The commercial availability of antibody prepara-
tions from donated human plasma, containing antibodies against many common gram-positive pathogens,
allows the protocol to be performed in the absence of bespoke vaccination experiments. Thus, immuno-
precipitation has the potential to reduce the number of animals used in vaccine studies by allowing an initial
screen for promising antigens to be conducted in vitro.

Key words Immunoprecipitation, Vaccine antigen, Gram-positive, Affinity purification, IVIG

1 Introduction

Immunoprecipitation (or immunoaffinity purification, see Note 1)

is a technique that uses immobilized antibodies to purify specific
proteins from complex biological solutions [1]. Today immunopre-
cipitation is commonly employed to isolate proteins that have been
genetically engineered to include specific peptide tags for which
commercial immunoprecipitation resins are available [2]. However,
the technique was initially developed to isolate naturally occurring
antibodies or antigens implicated in the pathophysiology of
infection, allergy, or autoimmunity for downstream characterisa-
tion [3–5].
The protocol below provides a simple framework for applying
immunoprecipitation to the identification of prospective vaccine
antigens from gram-positive pathogens and can be performed
using standard laboratory equipment and consumables. The

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_3,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

37
38 Mark Reglinski

protocol as described uses pooled human immunoglobulin (Intra-

venous Immunoglobulin G, IVIG), a clinical antibody infusion
prepared from the plasma of more than 3000 blood donors [6];
however, it could equally be performed using antiserum from
experimentally vaccinated animals. As both IVIG and animal anti-
serum will contain antibodies raised in response to many patho-
gens, antibodies targeted toward the pathogen selected for study
are first purified to produce a pathogen-reactive antibody pool
(PRAP) which is then coupled to a commercially available resin.
The resulting immunoprecipitation resin is then used to purify
prospective vaccine antigens from bacterial preparations (e.g., cell
wall extracts or cell lysates) for downstream characterisation.

2 Materials

2.1 Preparation of 1. 30% w/v raffinose (see Note 2): Dissolve 30 g of raffinose in
Gram-Positive Cell 80 ml of deionized water using a magnetic hotplate and a stir
Wall Extracts bar. Add deionized water to 100 ml and autoclave. Store at
room temperature (RT).
2. 1 M Tris–HCl (pH 8): Dissolve 12.1 g of tris base (NH2C
(CH2OH)3) in 80 ml of deionized water and adjust to pH 8
with concentrated HCl. Add deionized water to 100 ml and
store at RT.
3. 10 kU/ml mutanolysin: Add 1 ml of molecular grade water to
a vial containing 10,000 U of mutanolysin. Vortex vigorously
to ensure that the powder has completely dissolved and store at
20  C in 100 μl aliquots.
4. 100 mg/ml lysozyme: Dissolve 100 mg of lysozyme in 1 ml of
molecular grade water. Vortex vigorously to ensure that the
powder has completely dissolved and store at 20  C in 100 μl
aliquots.
5. Protease Inhibitor Cocktail Set III (Millipore).
6. Cell wall extraction buffer: 30% raffinose, 10 mM Tris–HCl,
100 U/ml mutanolysin, 1 mg/ml lysozyme, 1% protease
inhibitor cocktail. Combine 960 μl of 30% raffinose with
10 μl of Tris–HCl, mutanolysin, lysozyme, and protease inhib-
itor cocktail solutions to give 1 ml of cell wall extraction buffer
(see Note 3).
7. 20 ml polypropylene syringes with Luer slip.
8. 18-gauge, 1-in. beveled needles.
9. Slide-A-Lyzer Dialysis Cassettes, 20 K MWCO, 12 ml
(Thermo-Fisher).
10. Vivaspin 20 centrifugal concentrators, MWCO 10 kDa
(Sartorius).
 Immunoprecipitation of Vaccine Antigens 39

11. Millex-GP Syringe Filter Units, 0.22 μm (Millipore).

12. Colorimetric protein quantification assay reagents. For exam-
ple, Pierce BCA Protein Assay Kit or Pierce Coomassie (Brad-
ford) Protein Assay Kit.

2.2 Coupling of 1. Coupling buffer (see Note 4): 0.1 M sodium bicarbonate,
Proteins to Cyanogen 0.5 M sodium chloride (pH 8.3). Dissolve 8.4 g of sodium
Bromide Activated bicarbonate (NaHCO3) and 29 g of sodium chloride (NaCl) in
Sepharose 800 ml of deionized water. Adjust to pH 8.3 with 5 M sodium
hydroxide (NaOH, see Note 5) and add deionized water to 1 l.
Store at RT.
2. 1 mM hydrochloric acid (HCl, see Note 6): Dilute 83 μl of
concentrated HCl in 1 L of deionized water. Store at 4  C.
3. 0.2 M glycine: Dissolve 7.5 g of glycine (NH2CH2COOH) in
400 ml of deionized water. Add deionized water to 500 ml and
store at RT.
4. Wash buffer: 0.1 M acetate, 0.5 M sodium chloride (pH 4).
Dissolve 4.2 g of anhydrous sodium acetate (C2H3O2Na) and
14.5 g of NaCl in 400 ml of deionized water. Add 2.2 ml of
glacial acetic acid and adjust to pH 4 using concentrated HCl
(see Note 5). Add deionized water to 500 ml and store at RT.
5. Cyanogen bromide-activated Sepharose (CBr-Sepharose), lyo-
philized powder (Sigma-Aldrich).
6. Econo-Column® Chromatography Column, 1.5  5 cm
(Bio-Rad).
7. Horizontal tube rotator capable of supporting 15 ml centrifuge
tubes.
8. Sodium azide: 1% in aqueous solution, purchased premade
from commercial vendors.

2.3 Affinity 1. Loading buffer: 0.1 M sodium phosphate monobasic, 0.15 M

Purification of sodium chloride (pH 7). Dissolve 12 g of anhydrous sodium
Pathogen-Reactive phosphate monobasic (NaH2PO4) and 8.7 g of NaCl in 800 ml
Antibodies and of deionized water. Adjust to pH 7 with 5 M NaOH. Add
Immunoprecipitation deionized water to 1 L and store at RT.
of Prospective Vaccine 2. 1 M acetic acid: Dilute 5.74 ml of glacial acetic acid in 94.26 ml
Antigens of deionized water. Store at RT.
3. Neutralization buffer: 3 M Tris–HCl (pH 8.8). Dissolve 36.3 g
of Tris base in 80 ml of deionized water and adjust to pH 8.8
with concentrated HCl. Add deionized water to 100 ml and
store at RT.
4. Pooled human immunoglobulin (e.g., Privigen 100 mg/ml
solution for infusion from CSL Behring UK, see Note 7).
5 ml at 5 mg/ml in loading buffer.
40 Mark Reglinski

3 Methods

3.1 Preparation of 1. Culture the target bacterium to the desired growth phase
Gram-Positive Cell under appropriate conditions in 500 ml aliquots (see Note 8).
Wall Extracts 2. Centrifuge the culture at 4000  g for 10 min and resuspend
the cell pellet in 10 ml of 30% raffinose.
3. Centrifuge the cell suspension at 4000  g for 10 min and
resuspend the cell pellet in 10 ml of cell wall extraction buffer.
4. Incubate at 37  C for 3 h with occasional agitation.
5. Pellet the protoplasts at 14,000  g for 10 min and aspirate the
supernatant (cell wall extract).
6. Pass the cell wall extract through a 0.22 μm syringe filter and
transfer to a rehydrated 20 kDa MWCO cassette using a 20 ml
syringe and 18-gauge needle (see Note 9).
7. Dialyze the sample against 2 l of coupling buffer for at least 4 h
at 4  C.
8. Discard the spent dialysate and replace with 2 l of fresh cou-
pling buffer. Dialyze the sample overnight at 4  C.
9. Transfer the cell wall extract to a centrifugal filter unit and
concentrate to approximately 1 ml through multiple rounds
of centrifugation at 3000  g in a refrigerated centrifuge (see
Note 10).
10. Measure protein concentration of cell wall extract using a
colorimetric protein assay (see Note 11) and store at 20  C.

3.2 Coupling the Cell 1. Gradually add 250 mg of CBr-Sepharose to 50 ml of ice cold
Wall Proteins to CBr- 1 mM HCl and incubate at RT for 30 min with occasional
Sepharose agitation (see Note 12).
2. Transfer the resin suspension to a 1.5  5 cm chromatography
column and allow HCl to flow through the column.
3. Wash the resin bed with 5 ml of deionized water then with 5 ml
of coupling buffer, and immediately resuspend the resin in
2–4 ml of coupling buffer containing approximately 1 mg of
cell wall extract (see Note 13).
4. Incubate the resin slurry overnight at 4  C using a horizontal
tube rotator to maintain the liquid–solid suspension (see
Note 12).
5. Allow the resin to settle and collect the flow through for
downstream analysis (see Note 14).
6. Block the unreacted groups with 5 ml of 0.2 M glycine for 2 h
at RT using a horizontal tube rotator to maintain the liquid–
solid suspension.
 Immunoprecipitation of Vaccine Antigens 41

7. Allow the blocking solution to flow through the column and

wash the resin with 10 ml of coupling buffer then with 10 ml of
wash buffer. Complete this wash cycle of high and low pH
buffer solutions four more times to ensure removal of all
unreacted ligand.
8. Wash the column with 10 ml of loading buffer and resuspend
the affinity purification resin in 2 ml of loading buffer supple-
mented with 100 μl of 1% sodium azide solution (final concen-
tration 0.05%). Store at 4  C.

3.3 Affinity 1. Wash the affinity purification resin extensively with loading
Purification of buffer to remove the sodium azide.
Pathogen-Reactive 2. Apply 5 ml of pooled human immunoglobulin (25 mg of IgG)
Antibodies in loading buffer and incubate at RT for 2 h using a horizontal
tube rotator to maintain the liquid–solid suspension.
3. Allow resin to settle and collect flow-through for downstream
PRAP purifications (see Note 15).
4. Wash the resin with 20 ml of loading buffer and elute the
bound antibody (PRAP) directly into 8 ml of neutralization
buffer using 8 ml of 1 M acetic acid (see Note 16).
5. Wash the column with 10 ml of loading buffer and strip the
remaining antibody using 10 ml of 0.5 M NaOH (see
Note 17).
6. Wash the column with loading buffer until the pH of the flow-
through is neutral.
7. Resuspend the resin in 2 ml of loading buffer supplemented
with 100 μl of 1% sodium azide solution (final concentration
0.05%). Store at 4  C.
8. Transfer the eluted PRAP to a rehydrated 20 kDa MWCO
cassette using a 20 ml syringe and 18-gauge needle.
9. Dialyze the sample against 2 l of coupling buffer for at least 4 h
at 4  C.
10. Discard the spent dialysate and replace with 2 l of fresh cou-
pling buffer. Dialyze the sample overnight at 4  C.
11. Transfer the PRAP to a centrifugal filter unit and concentrate
to 1 mg/ml through multiple rounds of centrifugation at
3000  g in a refrigerated centrifuge (see Note 10).
12. Measure protein concentration of the PRAP and store at 4  C.
42 Mark Reglinski

3.4 Coupling of 1. Couple 1 mg of the PRAP to 1 ml of swollen CBr-Sepharose in

Pathogen-Reactive a 1.5  5 cm chromatography column according to the proto-
Antibodies to CBr- col above to produce the immunoprecipitation resin.
Sepharose and 2. Resuspend the immunoprecipitation resin in 5 ml of concen-
Immunoprecipitation trated cell wall extract supplemented with 0.1% protease inhib-
of Prospective Vaccine itor cocktail and incubate for 2 h at RT using a horizontal tube
Antigens rotator to maintain the liquid–solid suspension.
3. Allow the immunoprecipitation resin to settle and collect the
flow-through (containing unbound surface proteins) for
downstream analysis.
4. Wash the immunoprecipitation resin with 20 ml of loading
buffer and elute the bound protein fraction (containing pro-
spective vaccine antigens) directly into 8 ml of neutralization
buffer using 8 ml of 1 M acetic acid (see Note 16).
5. Wash the immunoprecipitation resin with 10 ml of loading
buffer and strip the remaining protein using 10 ml of 0.5 M
NaOH (see Note 17).
6. Wash the immunoprecipitation resin with loading buffer until
the pH of the flow through is neutral.
7. Resuspend the immunoprecipitation resin in 2 ml of loading
buffer supplemented with 100 μl of 1% sodium azide solution
(final concentration 0.05%). Store at 4  C.
8. Transfer the immunoprecipitated protein fraction to a rehy-
drated 20 kDa MWCO cassette using a 20 ml syringe and
18-gauge needle.
9. Dialyze the sample against 2 l of loading buffer for at least 4 h
at 4  C.
10. Discard the spent dialysate and replace with 2 l of fresh loading
buffer. Dialyze the sample overnight at 4  C.
11. Transfer the immunoprecipitated protein fraction to a centrif-
ugal filter unit and concentrate to approximately 50 μl through
multiple rounds of centrifugation at 3000  g in a refrigerated
centrifuge (see Note 10).
12. Measure the protein concentration of the immunoprecipitated
protein fraction using a colorimetric protein assay (see Note
11) and store at 20  C for downstream characterisation.

4 Notes

1. Immunoprecipitation is a technique that arose from a series of

classical immunological protocols such as the capillary precipi-
tin and double diffusion tests. Such assays rely on antibody-
mediated cross-linking of high molecular weight branching
 Immunoprecipitation of Vaccine Antigens 43

antigens such as capsular polysaccharides or crude cellular pre-

parations to generate insoluble “precipitates” that could be
visualized in solution or in specialized agarose gels [7]. How-
ever, the purification of antigens using immobilized antibodies
(or vice versa) does not cause antigen precipitation and thus
while the term “immunoprecipitation” has been widely
adopted, it does not accurately describe the procedure.
2. Less expensive sugars such as sucrose can be used for osmotic
stabilization of the bacterial protoplasts generated following
digestion of the cell wall. However, high molecular weight
polysaccharides such as raffinose have demonstrated superior
protoplast stabilizing capacities in comparative studies [8].
3. The lysozyme/mutanolysin based cell wall extraction buffer
can be modified for species specific applications. For example,
affinity purification resins have been prepared from S. aureus
using 1 mg/ml of lysostaphin in place of lysozyme [9]. Affinity
purification resins have also been prepared using whole
S. pneumoniae lysates in place of cell wall extracts [10]. While
no direct comparison has been made, it is likely that affinity
purification resins produced using whole cell lysates are less
efficient than resins produced using cell wall extracts for identi-
fication of prospective vaccine antigens. Where whole cell
lysates are used, much of the binding capacity of the resin will
be lost to prominent intracellular proteins that may have less
contact with the immune response. This may reduce both the
level of circulating antibodies against such proteins in the IVIG
and the likelihood that their recognition by antibodies will
promote opsonophagocytic killing of the bacterium in down-
stream vaccination studies.
4. While other buffers can be used, amine containing reagents
such as Tris and glycine must be avoided as these will react with
the active groups on the resin, reducing its protein binding
capacity.
5. As the coupling and acetate buffers are supplemented 500 mM
of NaCl, adjusting the pH with HCl and NaOH will not affect
the ionic strength of the solution appreciably.
6. 1 mM HCl is used to swell the resin in place of deionized water
to prevent hydrolysis of the reactive groups which occurs at an
alkaline pH. The HCl dilution outlined is based on the use of
37% HCl which has a molarity of approximately 12. The con-
centration of commercial, concentrated HCl can vary from
36 to 38% (approximately 11.8 M to 12.5 M); however, achiev-
ing an exact HCl concentration of 1 mM is not critical to the
success of the coupling protocol.
7. Intravenous immunoglobulin G, a clinical preparation of
pooled human immunoglobulin, is a prescription only
44 Mark Reglinski

medication that is not available for purchase through normal

laboratory suppliers. In previous publications IVIG has been
obtained from clinical suppliers (e.g., Privigen 100 mg/ml
solution for infusion from CSL Behring UK) via a centralized
hospital pharmacy. IgG from human serum can be purchased
from Sigma-Aldrich (catalog number I4506); however, purifi-
cation of PRAPs from this reagent has not been tested.
8. Previously cell wall extracts from cells grown to logarithmic
phase (defined as an A600 of 0.4–0.8) in standard laboratory
broth have been successfully used to purify opsonic PRAPs
from several species [9, 11].
9. The dialysis cassettes will swell substantially due to the osmotic
pressure generated by the 30% raffinose solution. The cassettes
have previously been filled with 20 ml of cell wall extract with
no adverse effects. Dialysis cassettes with a molecular weight
cut off of 20 kDa or higher are recommended to facilitate
removal of residual lysozyme and mutanolysin.
10. The amount of time and number of rounds of centrifugation
required to concentrate the samples will vary. An initial centri-
fugation time of 20 min is recommended. The filtrate reser-
voirs of Vivaspin centrifugal filter units can be removed and
emptied allowing volumes greater than 20 ml to be concen-
trated using a single unit over multiple spins. For large cell wall
extract volumes, tangential flow filtration may be useful for
both raffinose removal and protein concentration.
11. The protein concentration of the resulting cell wall extract will
vary from species to species and the protocol may need to be
scaled up considerably to reach the 1 mg of material recom-
mended for affinity resin preparation. In previous studies, cell
wall extracts from multiple strains of the same species have been
pooled to increase the repertoire of reactive antigens available
for IgG recognition [9, 11]. Prior to CBr-Sepharose coupling
nonspecific IgG binding proteins (e.g., Sbi and protein A from
S. aureus) that predominantly bind the Fc region of IgG can be
removed using a column prepared using a cleaved IgG prepa-
ration [9]. Suitable antibody fragments can also be purchased
directly from several commercial suppliers including Sigma-
Aldrich (e.g., catalog number AG714). Quantification of the
total protein present in the cell wall extracts by ultraviolet
absorption (e.g., using a Nanodrop) is not recommended as
the complexity of the samples complicates the selection of an
appropriate extinction coefficient.
12. 250 mg of lyophilized CBr-Sepharose will swell to give a resin
bed of approximately 1 ml. Pouring the HCl onto the dehy-
drated CBr-Sepharose will result in irreversible clumping and
should be avoided. End over end rotation or the use of a
 Immunoprecipitation of Vaccine Antigens 45

magnetic stir bar to maintain the liquid–solid suspension

should also be avoided as this may adversely affect the integrity
of the beads.
13. A resin-to-buffer ratio of 1:2 to 1:4 is required to maintain the
liquid–solid suspension during tube rotation.
14. The protein concentration of the cell wall extract solution can
be measured before and after resin incubation to confirm that
the immobilization has been successful.
15. The available data suggests that several PRAPs targeting sur-
face antigens from different bacterial species can be sequentially
purified from a single pool of human immunoglobulin
[9]. However, it should be noted that antibodies recognizing
conserved epitopes may be absent from, or drastically reduced
within the secondary and tertiary preparations.
16. The bound antibodies/antigens are eluted directly into 3 M
Tris–HCl (pH 8.8) to neutralize the acetic acid before irrevers-
ible denaturation occurs. The neutralization buffer should be
agitated to ensure that the eluant is rapidly incorporated.
17. The column can be stripped and reused at least five times.

References

1. Wofsy L, Burr B (1969) The use of affinity 7. Mehta PD (1972) Immunological techniques.
chromatography for the specific purification Indian J Ophthalmol 20(2):49–54
of antibodies and antigens. J Immunol 103 8. Marquis RE (1965) Osmotic stability of bacte-
(2):380–382 rial protoplasts related to molecular size of sta-
2. DeCaprio J, Kohl TO (2019) Tandem Immu- bilizing solutes. Biochem Biophys Res
noaffinity purification using anti-FLAG and Commun 20(5):580–585. https://doi.org/
anti-HA antibodies. Cold Spring Harb Protoc 10.1016/0006-291x(65)90438-9
2019(2). https://doi.org/10.1101/pdb. 9. Reglinski M, Sriskandan S (2019) Treatment
prot098657 potential of pathogen-reactive antibodies
3. Del Villano BC, Defendi V (1973) Characteri- sequentially purified from pooled human
zation of the SV40 T antigen. Virology 51 immunoglobulin. BMC Res Notes 12(1):228.
(1):34–46. https://doi.org/10.1016/0042- https://doi.org/10.1186/s13104-019-4262-
6822(73)90363-2 8
4. Alspaugh MA, Talal N, Tan EM (1976) Differ- 10. Wilson R, Cohen JM, Reglinski M, Jose RJ,
entiation and characterization of autoantibo- Chan WY, Marshall H, de Vogel C,
dies and their antigens in sjögren’s syndrome. Gordon S, Goldblatt D, Petersen FC,
Arthritis Rheum 19(2):216–222. https://doi. Baxendale H, Brown JS (2017) Naturally
org/10.1002/art.1780190214 acquired human immunity to pneumococcus
5. Ohman JL, Lowell FC, Bloch KJ (1974) Aller- is dependent on antibody to protein antigens.
gens of mammalian origin. J Immunol 113 PLoS Pathog 13(1):e1006137. https://doi.
(6):1668 org/10.1371/journal.ppat.1006137
6. Seite JF, Shoenfeld Y, Youinou P, Hillion S 11. Reglinski M, Gierula M, Lynskey NN, Edwards
(2008) What is the contents of the magic RJ, Sriskandan S (2015) Identification of the
draft IVIg? Autoimmun Rev 7(6):435–439. Streptococcus pyogenes surface antigens recog-
https://doi.org/10.1016/j.autrev.2008.04. nised by pooled human immunoglobulin. Sci
012 Rep 5:15825. https://doi.org/10.1038/
srep15825
 Chapter 4

Rapid Surface Shaving for Proteomic Identification of Novel

Surface Antigens for Vaccine Development
Laurence Don Wai Luu and Ruiting Lan

Abstract
The bacterial cell surface (surfaceome) is the first site encountered by immune cells and is thus an important
site for immune recognition. As such, the characterization of bacterial surface proteins can lead to the
discovery of novel antigens for potential vaccine development. In this chapter, we describe a rapid 5-min
surface shaving proteomics protocol where live bacterial cells are incubated with trypsin and surface
peptides are “shaved” off. The shaved peptides are subsequently identified with liquid chromatography–-
tandem mass spectrometry (LC-MS/MS). Several checkpoints, including colony forming unit (CFU)
counts, flow cytometry, and a false positive unshaved control, are introduced to ensure cell viability/
membrane integrity are maintained and that proteins identified are true surface proteins. The protein
topology of shaved peptides can be bioinformatically confirmed for surface location. Surface shaving
facilitates identification of surface proteins expressed under different conditions, by different strains as
well as highly abundant essential and immunogenic bacterial surface antigens for potential vaccine
development.

Key words Surface shaving, Surfaceome, Surface proteins, Antigens, Vaccines, Proteomics, Cell
surface, Mass spectrometry, LC-MS/MS, Epitopes

1 Introduction

The bacterial cell surface or “surfaceome” is the site where many

important biological processes and host-pathogen interactions
occurs including nutrient transport, cellular adhesion, delivery of
cytotoxic effectors and immune recognition. As a result, this site is
rich with many potential vaccine antigens. The advancement of
proteomics and liquid chromatography–tandem mass spectrometry
(LC-MS/MS), along with different surface extraction and enrich-
ment techniques, has facilitated the characterization of these sur-
face proteins.
Initially, surface protein identification was performed using
subcellular fractionation methods which involve either differential
lysis of different cellular compartments and/or differential

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_4,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

47
48 Laurence Don Wai Luu and Ruiting Lan

centrifugation in various buffers to separate surface proteins from

other compartments [1]. However, these methods often had cyto-
plasmic contamination and were long and tedious to perform. As a
result, an alternate method for bacterial surface protein identifica-
tion was developed, initially using Streptococcus pyogenes, called
“surface shaving” [2]. This method involves the incubation of live
whole bacterial cells with trypsin (or other proteases) to “shave”
peptides off the cell surface without lysing the cell. This method is
comparatively easier to perform and provides important informa-
tion on protein topology and surface exposed epitopes. Knowledge
of surface exposed epitopes can be used to identify peptide regions
accessible to immune recognition and inform vaccine design. How-
ever, one concern from surface shaving is the potential for inadver-
tent cell lysis, during incubation with trypsin, leading to the release
of contaminating cytoplasmic proteins [3]. As a result, surface
shaving has been more commonly used in gram-positive bacteria,
which have a thicker peptidoglycan cell wall to withstand shaving,
compared to gram-negatives [4].
To address the question of how to determine whether proteins
identified by surface shaving are truly surface localized, or released
as a result of cell lysis during incubation, Solis et al. [5] developed
an “unshaved control” strategy to detect peptides released in the
absence of trypsin digestion and eliminate false-positive prediction
of these as surface proteins. To improve surface shaving for use in
gram-negative bacteria and surface antigen discovery, we further
refined the technique by reducing the amount of trypsin needed
and the shaving incubation time, as well as incorporating colony
forming unit (CFU) counts and flow cytometry checkpoints to
ensure cell viability/membrane integrity. We have demonstrated
the feasibility of using rapid surface shaving in the gram-negative
bacterium Bordetella pertussis, which causes whooping cough, to
compare surface protein expression between two representative
strains and to identify novel surface antigens for vaccine
development [6].
In this chapter, we detail the steps involved for our rapid surface
shaving protocol of live bacterial cells for proteomic identification
and quantification of surface antigens (Fig. 1). Although B. pertussis
is used as an example, rapid surface shaving can be similarly applied
to both gram-negative and gram-positive bacteria, as well as eukar-
yotes [7]. The protocol involves preparation of cells (see Subhead-
ing 3.1), rapid surface shaving and establishment of a false positive
unshaved control (see Subheading 3.2), determining cell viability/
membrane integrity after surface shaving (see Subheading 3.3) and
preparing peptides for proteomic identification with LC-MS/MS
and subsequent analysis to identify highly expressed, essential, and
immunogenic surface antigens for further vaccine development (see
Subheadings 3.4 and 3.5).
 Rapid Bacterial Surface Shaving to Identify Surface Antigens 49

Fig. 1 Overview of the rapid surface shaving protocol for proteomic identification of surface antigens from live
cells. Live whole cells are incubated with trypsin for 5 min to shave off surface peptides. An unshaved control
is incubated under identical conditions without trypsin to identify peptides released during incubation which
could lead to false positive prediction of surface exposure. Cell viability/membrane integrity is assessed using
flow cytometry with propidium iodide (PI) and CFU counts. Shaved peptides are desalted and concentrated
with C18 StageTips and analysed with LC-MS/MS.

2 Materials

2.1 Preparation 1. B. pertussis culture (see Notes 1 and 2).

of B. pertussis Cells 2. Bordet–Gengou (BG) agar: 14 g of Bordet–Gengou agar base,
340 mL of ultrapure water, 6 mL of 50% glycerol (v/v).
Autoclave-sterilize and cool to 50
C. Add 25 mL of sterile
defibrinated horse blood and pour into sterile Petri dishes.
3. Thalen–IJssel (THIJS) medium: 3.319 g NaCl, 0.107 g
NH4Cl, 0.5 g KH2PO4, 0.5 g KCl, 0.1 g MgCl2.6H2O,
1.525 g Tris, and 1.87 g Na glutamate.H2O. Add
50 Laurence Don Wai Luu and Ruiting Lan

994.169 mL of ultrapure water, 3.76 mL of L-lactate (40%

w/v), and 2.071 mL of 5 M NaOH. Autoclave-sterilize and
store at 4
C for up to 6 months (see Note 3).
4. 100x THIJS supplement: 0.2 g L-cystine, 0.13 g CaCl2.2H2O,
0.5 g L-glutathione reduced, 0.05 g FeSO4.7H2O, 0.02 g
nicotinic acid, and 0.1 g L-ascorbic acid. Add 44 mL of ultra-
pure water and 6 mL of 1 M HCl. Filter-sterilize with a 0.22-μ
m filter and store as 1 mL aliquots at -20
C for up to 1 year.
5. 100 heptakis (also known as (2,6-di-O-
methyl)-β-cyclodextrin): 1 g of heptakis, 10 mL of ultrapure
water. Filter-sterilize with a 0.22-μm filter and store as 1 mL
aliquots at -20
C for up to 1 year.
6. TPP TubeSpin bioreactor tube (50 mL) (see Note 4).

2.2 Rapid Surface 1. 1 Phosphate buffered saline (PBS), pH 7.4.

Shaving 2. 10 mM HEPES, pH 7.4: 0.238 g HEPES, 80 mL of ultrapure
water. Adjust pH to 7.4 with NaOH and add ultrapure water to
a final volume of 100 mL. Filter-sterilize with a 0.22-μm filter
and store at 4
C.
3. 1μg/μL trypsin: Reconstitute 20 μg of lyophilized trypsin
(Sequencing Grade Modified) in 20 μL of ultrapure water.
Store at 80
C.
4. Charcoal blood agar (CBA): Dissolve 25 g of charcoal agar base
in 450 mL of ultrapure water. Autoclave-sterilise and cool to
50
C. Add 50 mL of sterile defibrinated horse blood and pour
into sterile Petri dishes.
5. Formic acid, LC-MS grade.

2.3 Flow Cytometry 1. FACS buffer (1% BSA and 0.1% sodium azide in 1 PBS): Add
0.5 g BSA and 0.05 g sodium azide to 50 mL of 1 PBS. Filter-
sterilise with a 0.22-μm filter and store at 4
C (see Note 5).
2. 0.5 mg/mL propidium iodide (PI) (see Note 6).

2.4 C18 Clean Up 1. Pierce 200 μL C18 StageTips from Thermo Fisher to desalt
and Liquid and concentrate peptides for LC-MS/MS.
Chromatography– 2. Centrifuge adapter from GL Sciences to hold the StageTips in
Tandem Mass place in 1.5 mL microcentrifuge tubes (Fig. 2).
Spectrometry 3. 0.2% heptafluorobutyric acid (HFBA), HPLC grade.
(LC-MS/MS)
4. 80% acetonitrile, 5% formic acid: 800 μL of acetonitrile (HPLC
grade), 50 μL of formic acid (LC-MS grade), and 150 μL of
ultrapure water.
3. 0.1% formic acid.
 Rapid Bacterial Surface Shaving to Identify Surface Antigens 51

Fig. 2 Set up of 200 μL C18 StageTips in the centrifuge adapter and 1.5 mL microcentrifuge tube

2.5 Bioinformatic 1. Mascot Daemon (v.2.5.1)

Analysis 2. Scaffold (v4.4.5)
and Identification
3. R package (qvalue v2.24.0)
of Surface Proteins
4. PSORTb v3.0 [8]
5. SignalP v5.0 [9]
6. TMHMM v2.0 [10]
7. Phobius [11]
8. Protter v1.0 [12]
9. KofamKOALA [13]
10. eggNOG-mapper v2 [14]
11. DAVID v6.8 [15]
12. IEDB [16]
13. DEG [17]

3 Methods

3.1 Preparation 1. Inoculate one loopful of Bvg+ B. pertussis grown on BG agar

of B. pertussis Cells into a TPP TubeSpin bioreactor tube with 20 mL of THIJS
medium supplemented with 1 THIJS supplement and 1x
heptakis (see Note 7).
2. Incubate the culture for 24 h at 37
C with shaking at 180 rpm.
3. Measure the OD600 with 1 mL of the 24 h culture.
52 Laurence Don Wai Luu and Ruiting Lan

4. Adjust the starting OD600 to 0.05 in a total of 20 mL of fresh

THIJS medium (supplemented with 1 THIJS supplement
and 1 heptakis) in a new TPP TubeSpin bioreactor tube.
5. Incubate the culture for 12 h at 37
C with shaking at 180 rpm.
6. Optional: record the OD600 after 12 h incubation (see Notes
8 and 9).

3.2 Rapid Surface 1. Centrifuge the 12 h culture at 3300  g for 15 min at 4
C and
Shaving with Trypsin discard the supernatant (see Note 10).
2. Gently wash cells with 1 mL of ice-cold 1 PBS and centrifuge
at 3300  g for 15 min at 4
C. Discard supernatant and repeat
wash step for a total of 3 washes (see Note 11).
3. Gently resuspend cells in 2 mL of 10 mM HEPES (see
Note 12).
4. Split the sample into two by aliquoting 1 mL of the sample into
a TPP TubeSpin bioreactor tube labeled “shaved” and another
labeled as “unshaved control” (see Notes 13 and 14).
5. Add 1 μL of 1 μg/μL trypsin to the “shaved” sample (see
Note 15).
6. Incubate both “shaved” and “unshaved control” samples for
5 min at 37
C with gentle agitation (see Notes 16 and 17).
7. Transfer both samples into two new, appropriately labeled,
1.5 mL microcentrifuge tubes.
8. Optional: aliquot 10 μL of each sample for colony forming unit
(CFU) counts (see Note 18).
9. Pellet cells in both samples by centrifugation at maximum
speed (20,238  g) for 1 min at 4
C (see Note 19).
10. For both the “shaved” and “unshaved control” samples, trans-
fer the respective supernatants into new, appropriately labeled,
1.5 mL microcentrifuge tubes. The cell pellets from each sam-
ple are kept on ice until processed for flow cytometry (see
Subheading 3.3).
11. Add 1 μL of 1 μg/μL trypsin to the supernatant of the
“unshaved control” sample and incubate for 5 min at 37
C.
Keep the “shaved” sample supernatant on ice to prevent fur-
ther trypsin digestion while the “unshaved control” is under-
going trypsin digestion.
12. Add 1% of formic acid (10 μL) to both samples to inactivate
trypsin digestion (see Note 20).

3.3 Determining Cell 1. Prepare a positive control for cell death by resuspending 1 loop-
Viability After Rapid ful of Bvg+ B. pertussis grown on BG agar in 1 mL of 1x PBS
Surface Shaving Using and heat at 70
C for 30 min. Centrifuge at maximum speed
Propidium Iodide (PI). (20,238  g) for 1 min at 4
C. Discard supernatant and
resuspend cells in 900 μL of FACS buffer (see Note 21).
 Rapid Bacterial Surface Shaving to Identify Surface Antigens 53

2. Prepare an unstained control by resuspending 1 loopful of

Bvg+ B. pertussis grown on BG agar in 1 mL of 1 PBS.
3. Gently resuspend the “shaved” and “unshaved control” cells
(from Subheading 3.2, step 9) in 1 mL ice-cold 1x PBS to
wash. Centrifuge at maximum speed (20,238  g) for 1 min at
4
C. Discard supernatant. Repeat wash step once for a total of
2 washes.
4. Gently resuspend cells in 900 μL of FACS buffer.
5. Add 100 μL of PI (0.5 mg/mL) to the “shaved” and
“unshaved control” cells, as well as to the “positive control
for cell death” sample. Incubate on ice for 5 min in the dark.
6. Analyse cell viability by flow cytometry (see Note 22).

3.4 C18 Peptide 1. Prepare the “shaved” and “unshaved control” supernatant
Clean-Up samples (from Subheading 3.2, step 12) for C18 clean-up by
and LC-MS/MS mixing 50 μL of sample with 20 μL of 0.2% HFBA in appropri-
ately labeled 1.5 mL microcentrifuge tubes (see Note 23).
2. For each sample, place a C18 StageTip into a centrifuge adapter
and insert into an appropriately labeled 1.5 mL microcentri-
fuge tube (Fig. 2).
3. Initialize the C18 StageTips by adding 20 μL of 80% acetoni-
trile, 5% formic acid. Centrifuge at 200  g for 2 min (see
Note 24).
4. Reequilibrate the C18 StageTips with 20 μL of 0.2% HFBA
and centrifuge at 200  g for 2 min. Discard flowthrough.
5. Load each 70 μL sample (from Subheading 3.4, step 1) into a
prepared C18 StageTip and centrifuge at 200  g for 2 min (see
Notes 25 and 26).
6. Wash the samples on the C18 StageTips with 20 μL of 0.2%
HFBA and centrifuge at 200  g for 2 min.
7. Transfer the C18 StageTips with bound sample, along with
their centrifuge adaptors, to new appropriately labeled 1.5 mL
microcentrifuge tubes and elute samples with 20 μL of 80%
acetonitrile, 5% formic acid and centrifuge at 200  g for 2 min.
8. Vacuum-dry samples (see Note 27).
9. Resuspend dried peptides in 10 μL of 0.1% formic acid and
perform LC-MS/MS (see Note 28).

3.5 Bioinformatics 1. Load the resulting LC-MS/MS raw datafiles into Mascot
Analysis Daemon (v.2.5.1) for protein identification with the following
and Identification search parameters: instrument type ¼ ESI-type; peptide toler-
of Surface Proteins ance ¼ 4 ppm; MS/MS tolerance ¼ 0.4 Da; variable modifica-
tions ¼ Carbamidomethyl (C) and Oxidation (M); enzyme
54 Laurence Don Wai Luu and Ruiting Lan

specificity ¼ trypsin; max number of missed cleavages ¼ 1; and

search database ¼ custom B. pertussis database (see Note 29).
2. Load the Mascot output into Scaffold (v4.4.5), which imple-
ments the ProteinProphet algorithm to validate protein identi-
fications. Choose “spectrum counting” as the quantitative
technique. Biological replicates (and shaved or unshaved con-
trol samples) can be grouped together by providing the same
sample category name (e.g., Sample category 1: B. pertussis
shaved, Sample category 2: B. pertussis unshaved control). Set
the threshold for identification in Scaffold as: protein thresh-
old ¼ 99%, peptide threshold ¼ 95% and minimum number of
peptides per protein ¼ 1 (Fig. 3a) (see Notes 30 and 31).
3. Remove the peptides identified in the “unshaved control”
dataset from matched false-positive predicted surface peptides
in the “shaved” dataset by unticking the valid box of the
matched peptide sequence in the Proteins tab (Fig. 3b) (see
Note 32).

Fig. 3 (a) Example screenshot of the Samples tab in Scaffold illustrating the number of proteins and peptides
identified in the unshaved control (1191_C in the purple box outline) and shaved samples (1191_T in the red
box outline). The settings for protein threshold, peptide threshold, and minimum number of peptides per
protein are shown in the green box outline. (b) The highlighted peptide (MLDTTVALMSAK) from the cytoplasmic
protein glyceraldehyde-3-phosphate dehydrogenase was identified in the unshaved control sample and is
removed from the shaved samples by unticking the “valid” box (black box outline) in the Proteins tab
 Rapid Bacterial Surface Shaving to Identify Surface Antigens 55

4. Perform relative quantitation analysis using normalized total

spectral counts in Scaffold’s quantitative analysis setup page.
Normalized total spectral counts can be used to quantify and
identify significant surface protein expression changes between
two strains or conditions. Select the “shaved” samples as input
with the following settings: statistical test ¼ T-test; multiple
test correction ¼ no correction; significance level ¼ p < 0.05;
fold change ¼ fold change by category; use normalization ¼ yes;
minimum value ¼ 0.0; and quantitative method ¼ Total spec-
tra. In the Export menu bar, click “Samples Report” to export
the samples report to excel for further analysis.
5. Perform multiple test correction using the Storey–Tibshirani
method in R by importing Scaffold p values and calculating the
q value using the qvalue R package (see Notes 33 and 34).
6. Calculate the normalized spectral abundance factor (NSAF) for
each surface protein to identify the most abundant surface
proteins (see Note 35).

Normalised total spectral counts

Normalised spectral abundance factor ðNSAFÞ ¼ :
Amino acid length of protein
7. Other web-based bioinformatic analyses that can be performed
include: PSORTb v3.0 to predict the protein cellular location;
SignalP v5.0 to identify proteins with signal peptides for trans-
location across membranes; TMHMM v2.0 to identify trans-
membrane helices; Phobius to predict transmembrane
topology; and Protter v1.0 to visualize the topology of shaved
peptides in surface proteins (Fig. 4). Functional category anno-
tation of surface proteins can be performed with tools such as
KofamKOALA, eggNOG-mapper v2 and/or DAVID v6.8.
IEDB can be used to identify known and predicted immune
epitopes of surface proteins/epitopes identified through rapid
surface shaving. DEG can be used to identify which surface
antigens are essential for “survival”.

4 Notes

1. B. pertussis is a human respiratory pathogen. To limit the risk of

infection, all work involving live cells should be performed in a
class II biological safety cabinet and all personnel working with
B. pertussis should be immunized against whooping cough.
2. B. pertussis is grown on BG agar at 37
C for 4 days. Ensure that
all B. pertussis colonies on the plate are in the Bvg+ phase when
virulence factors are expressed. Bvg+ colonies are small (1 mm)
and hemolytic, while Bvg- colonies are larger and
nonhemolytic.
56 Laurence Don Wai Luu and Ruiting Lan

Fig. 4 Topology of peptides identified by rapid surface shaving (in green) were mapped to their proteins using
Protter v1.0. (a-e) display 5 well-characterized B. pertussis autotransporter proteins that are either current
antigens found in the pertussis acellular vaccine, or proposed to be good candidates as future vaccine
antigens (A. Pertactin (Prn); B. Tracheal colonization factor A (TcfA); C. Bordetella resistance to killing A (BrkA);
D. Virulence associated gene 8 (Vag8); and E. Subtilisin (SphB1). For A-E, N-signal peptides are shown in red,
autotransporter passenger domains, which are known to be surface exposed, are colored purple while channel
domains, which are embedded into the membrane, are shown in grey. In (f), the Bordetella Bvg-intermediate
phase protein (BipA) is displayed. BipA is an outer membrane ligand binding protein and another proposed
potential vaccine antigen. The shaved peptides (green) from BipA are mapped to the predicted surface
exposed area from Phobius. The bacterial outer membrane is depicted as an orange bar and the numbers
1 and 2 are the predicted BipA transmembrane domains that anchor the protein to the outer membrane.
(Reprinted from Vaccine, 38 (3), Luu LDW, Octavia S, Aitken C, Zhong L, Raftery MJ, Sintchenko V and Lan R,
Surfaceome analysis of Australian epidemic Bordetella pertussis reveals potential vaccine antigens, 539–548,
2020, with permission from Elsevier)

3. THIJS is a chemically defined medium that was modified from

the original Stainer-Scholte (SS) medium [18]. THIJS is
designed for the optimal growth of B. pertussis and is prepared
according to Thalen et al. [19]. Our study found that THIJS
promotes the production of virulence factors in B. pertussis and
therefore is the growth medium of choice for the liquid culture
of B. pertussis [20]. However, SS medium may also be used.
4. B. pertussis is a strict aerobe and a respiratory pathogen, there-
fore 50 mL TPP TubeSpin bioreactor tubes, which have lids
fitted with 0.22-μm filters, are used for liquid cultures. Other
vessels can be used, provided that there is enough oxygen
exchange and the required safety standards are met.
5. BSA is difficult to dissolve. To dissolve BSA in 1 PBS, lightly
dust BSA over the surface of the liquid and allow it to sink and
dissolve over time. Do not stir or vortex.
6. PI is a dye which binds DNA and cannot translocate across
intact membranes. It can be used to determine cell viability and
membrane integrity after rapid surface shaving. PI is resus-
pended in 1 PBS and stored in the dark at 4
C.
7. 200 μL of THIJS supplement and 200 μL of heptakis are added
to 20 mL of THIJS medium immediately prior to each use.
 Rapid Bacterial Surface Shaving to Identify Surface Antigens 57

8. To obtain enough power, a minimum of 6 biological replicates

for each strain/condition is needed if relative quantitation of
surface proteins is performed for B. pertussis.
9. From our previous studies [21], the B. pertussis strains used are
known to be in the early exponential growth phase after 12 h
incubation in THIJS. The OD600 for B. pertussis clinical strains
L1423 and L1191 should be 0.4–0.6 [21].
10. Rapid surface shaving steps 1–4 should be completed as fast as
possible (within 90 min) to minimize cell lysis during the
sample preparation stage. After incubation, the cultures should
always be kept on ice throughout the experiment (unless oth-
erwise indicated) to minimize cell lysis.
11. Ensure cells are gently washed and resuspended throughout
the experiment to minimize cell lysis from vigorous pipetting.
It is also recommended to avoid pipetting liquid directly onto
the cells.
12. 10 mM HEPES was found to minimize cell lysis during surface
shaving by Walters et al. [3].
13. An “unshaved control” sample is established to identify false
positive predictions of surface peptides, as described in Solis
et al. [5]. The control is incubated under identical conditions
as the “shaved” sample, except in the absence of trypsin. The
purpose of the “unshaved control” is to identify proteins/
peptides falsely predicted to be surface exposed (likely released
due to secretion, cell lysis and/or membrane leakage during
rapid surface shaving), and to remove them from the proteins/
peptides identified in the shaved sample.
14. B. pertussis is a strict aerobe, therefore, to maintain oxygen
exchange and minimize the chance of cell death, surface
shaving is performed in the 50 mL TPP TubeSpin bioreactor
tubes; however, other vessels may be used for other organisms
(e.g., 1.5 mL microcentrifuge tubes).
15. We previously tested different trypsin concentrations (1 μg,
2.5 μg, and 5 μg) and digestion times (5, 10, and 15 min) to
optimize surface shaving parameters for B. pertussis and obtain
enough shaved peptides for identification, without
compromising membrane integrity. We assessed cell/mem-
brane integrity using CFU counts and flow cytometry with
PI, as well as by determining the number of known outer
membrane peptides/proteins identified in shaved samples and
the number of cytoplasmic peptides/proteins identified in the
unshaved control samples. Rapid surface shaving using 1 μg of
trypsin for 5 min identified the highest proportion of known
outer membrane proteins with the least amount of cell lysis in
B. pertussis [6]. For other organisms or strains, digestion con-
ditions may need to be optimized.
58 Laurence Don Wai Luu and Ruiting Lan

16. Samples can be incubated in a 37
C water bath with gentle

agitation at 60 rpm. Due to the short incubation period, a
water bath is recommended over an incubator as the water
ensures efficient heat transfer.
17. Longer incubation times may be used for other organisms,
however they must be validated with CFU counts and flow
cytometry (see Subheading 3.3) to ensure cell viability and
membrane integrity are not compromised.
18. Ensure tubes are placed immediately on ice to prevent further
trypsin digestion. CFU counts are performed in triplicate on
charcoal blood agar (CBA) using the drop plate method [22],
with incubation at 37
C for 5–7 days, or until colonies appear.
19. Quick centrifugation at maximum speed for 1 min is used to
pellet cells quickly and prevent further trypsin digestion.
20. Flow cytometry to assess cell viability should be performed
first. Peptides in the “shaved” and “unshaved control” sample
supernatants can be stored in -80
C (following addition of 1%
of formic acid) before proceeding to C18 clean-up and mass
spectrometry. Ideally, C18 clean-up should be performed on
the day of shaving or on the next day, as storage of peptides in
solution should not be longer than a few days at 80
C.
21. The positive control for cell death can be prepared up to several
days in advance and stored at 4
C.
22. Cell viability was analysed on the BD FACSCanto II using a
blue laser 488 nm excitation and a 670 LP bandpass filter. Ten
thousand events were acquired for each sample and data ana-
lysed using FlowJo software. Ensure that there is minimal cell
lysis and disruption to the membrane as a result of rapid surface
shaving; cell viability should be at least 98% (Fig. 5).
23. For C18 peptide clean-up, three 1.5 mL tubes will be required
per “shaved” and “unshaved control” sample: one for the
supernatant samples to be loaded into the C18 StageTip (see
Subheading 3.4, step 1); one to initialize the C18 StageTips
and for flowthrough collection after supernatant samples are
loaded (see Subheading 3.4, steps 2–6); and one for the elution
of samples (see Subheading 3.4, step 7). These can be
organized and labeled ahead of time with simple labels such
as “SL” and “UL” for load, “SFT” and “UFT” for flow-
through collection and “SE” and “UE” for elution, for the
respective shaved (S) and unshaved (U) samples. If there are
multiple strains and/or biological replicates being processed at
one time, the strain name and biological replicate (indicated by
a hyphen and number) can also be added to the labels, for
example, “L1423–1 SL” and “L1191–1 SL”.
 Rapid Bacterial Surface Shaving to Identify Surface Antigens 59

105

104

Foward Scatter A
103

102

–102

0 103 104 105

Propidium lodide

Fig. 5 Flow cytometry dot plots of propidium iodide stained cells indicating intact
membrane integrity and minimal cell lysis from cells subjected to rapid surface
shaving protocol (purple dots). Unshaved control cells (incubated in the absence
of trypsin) are illustrated as pink dots, while the heated treated cells (positive
control for cell death) are shown as red dots. (Reprinted from Vaccine, 38 (3),
Luu LDW, Octavia S, Aitken C, Zhong L, Raftery MJ, Sintchenko V and Lan R,
Surfaceome analysis of Australian epidemic Bordetella pertussis reveals
potential vaccine antigens, 539–548, 2020, with permission from Elsevier)

24. For each centrifugation step, check that all liquid has flowed
through the StageTip. If there is remaining liquid, increase
centrifugal force by 100  g and centrifuge again until all liquid
has passed.
25. Centrifuge samples at maximum speed for 1 min prior to
loading onto the StageTips to remove precipitates which may
block liquid from passing through.
26. Keep the flowthrough from the load and wash steps in case
peptides have not bound and store it at 80
C. If after
LC-MS/MS of the C18-eluted samples (see Subheading 3.4,
step 9), no/minimal peptides are detected especially in the
“shaved” samples, this may indicate that the peptides did not
bind to the C18 StageTips. If this is the case, the flowthrough
can be reloaded into a new C18 StageTip that has been initi-
alized and reequilibrated (see Subheading 3.4, steps 3–5) and
cleaned up again.
60 Laurence Don Wai Luu and Ruiting Lan

27. Dried peptides can be stored at 80
C before proceeding with

mass spectrometry and should remain stable long-term
(>6 months) if proper storage and handling conditions are
followed [23]. However, we have not directly assessed the
effect of long-term storage on results.
28. LC-MS/MS was performed using the LTQ-Orbitrap Velos
with 0.1 μL of samples separated over a 30 min gradient, as
detailed in Luu et al. [20]. Due to the low quantity of peptides
released by rapid surface shaving, it is difficult to determine the
concentration of proteins/peptides (e.g., with Qubit) loaded
for LC-MS/MS.
29. Generic databases, such as NCBI or SWISS-PROT, can be
searched for sequence matches. However, to reduce search
time, a specific custom database can be created. In this case,
we used a custom B. pertussis database containing amino acid
FASTA files for Tohama I (NC_002929.2), CS
(NC_017223.1), B1917 (CP009751.1) and B1920
(CP009752.1) strains.
30. The protein identification threshold for number of peptides
identified per protein is set to 1, as rapid digestion may not
result in a large number of peptides [5].
31. We typically obtain an average of ~250 unique peptide identi-
fications in the shaved samples and an average ~20 peptides
identified in the false-positive unshaved control for B. pertussis
(Fig. 3a). If there is a large amount of cytoplasmic peptides/
proteins identified in the unshaved control, it may be a sign of
cell lysis/compromised membrane integrity.
32. Peptides corresponding to those identified in the “unshaved
control,” representing false-positive predictions of surface
exposure, can be removed from the matched “shaved” sample,
or from the entire dataset of “shaved” samples. We were more
conservative and removed the predicted false-positive peptides
for the whole dataset.
33. Multiple test correction can also be performed in Scaffold by
setting multiple test correction as Benjamini–Hochberg
(or other methods).
34. Upregulated surface proteins can be defined as those showing
>1.2 fold change and downregulated surface proteins as those
showing <0.8 fold change, with significance set at p value
<0.05 and q value <0.05.
35. Normalized spectral abundance factor (NSAF) [24] normalizes
spectral count to amino acid length to quantify relative protein
abundance within a sample. Identification of highly abundant
surface antigens as targets of potential vaccines is important
since they may have a greater potential for immune recognition
and antibody binding [25].
 Rapid Bacterial Surface Shaving to Identify Surface Antigens 61

References
1. Molloy MP, Herbert BR, Slade MB, application to complete genomes. J Mol Biol
Rabilloud T, Nouwens AS, Williams KL, Goo- 305(3):567–580
ley AA (2000) Proteomic analysis of the 11. K€all L, Krogh A, Sonnhammer EL (2004) A
Escherichia coli outer membrane. Eur J Bio- combined transmembrane topology and signal
chem 267(10):2871–2881 peptide prediction method. J Mol Biol 338
2. Rodrı́guez-Ortega MJ, Norais N, Bensi G, (5):1027–1036
Liberatori S, Capo S, Mora M, Scarselli M, 12. Omasits U, Ahrens CH, Müller S, Wollscheid B
Doro F, Ferrari G, Garaguso I (2006) Charac- (2013) Protter: interactive protein feature visu-
terization and identification of vaccine candi- alization and integration with experimental
date proteins through analysis of the group A proteomic data. Bioinformatics 30
Streptococcus surface proteome. Nat Biotechnol (6):884–886
24(2):191–197 13. Aramaki T, Blanc-Mathieu R, Endo H,
3. Walters MS, Mobley HL (2009) Identification Ohkubo K, Kanehisa M, Goto S, Ogata H
of uropathogenic Escherichia coli surface pro- (2020) KofamKOALA: KEGG ortholog
teins by shotgun proteomics. J Microbiol assignment based on profile HMM and adap-
Methods 78(2):131–135 tive score threshold. Bioinformatics 36
4. Olaya-Abril A, Jiménez-Munguı́a I, Gómez- (7):2251–2252
Gascón L, Rodrı́guez-Ortega MJ (2014) Sur- 14. Huerta-Cepas J, Forslund K, Coelho LP,
fomics: shaving live organisms for a fast pro- Szklarczyk D, Jensen LJ, Von Mering C, Bork
teomic identification of surface proteins. J P (2017) Fast genome-wide functional annota-
Proteome 97:164–176 tion through orthology assignment by
5. Solis N, Larsen MR, Cordwell SJ (2010) eggNOG-mapper. Mol Biol Evol 34
Improved accuracy of cell surface shaving pro- (8):2115–2122
teomics in Staphylococcus aureus using a false- 15. Sherman BT, Lempicki RA (2009) Systematic
positive control. Proteomics 10 and integrative analysis of large gene lists using
(10):2037–2049 DAVID bioinformatics resources. Nat Protoc 4
6. Luu LDW, Octavia S, Aitken C, Zhong L, Raf- (1):44
tery MJ, Sintchenko V, Lan R (2019) Surfa- 16. Vita R, Mahajan S, Overton JA, Dhanda SK,
ceome analysis of Australian epidemic Martini S, Cantrell JR, Wheeler DK, Sette A,
Bordetella pertussis reveals potential vaccine Peters B (2019) The immune epitope database
antigens. Vaccine 38(3):539–548 (IEDB): 2018 update. Nucleic Acids Res 47
7. Bark SKN, Ahmad R, Dantzler K, Lukens AK, (D1):D339–D343
De Niz M, Szucs MJ, Jin X, Cotton J, 17. Luo H, Lin Y, Gao F, Zhang C-T, Zhang R
Hoffmann D, Bric-Furlong E (2018) Quanti- (2014) DEG 10, an update of the database of
tative proteomic profiling reveals novel plasmo- essential genes that includes both protein-
dium falciparum surface antigens and possible coding genes and noncoding genomic ele-
vaccine candidates. Mol Cell Proteomics 17 ments. Nucleic Acids Res 42(D1):D574–D580
(1):43–60 18. Thalen M, van den IJssel J, Jiskoot W,
8. Nancy YY, Wagner JR, Laird MR, Melli G, Zomer B, Roholl P, de Gooijer C, Beuvery C,
Rey S, Lo R, Dao P, Sahinalp SC, Ester M, Tramper J (1999) Rational medium design for
Foster LJ (2010) PSORTb 3.0: improved pro- Bordetella pertussis: basic metabolism. J Bio-
tein subcellular localization prediction with technol 75(2):147–159
refined localization subcategories and predic- 19. Thalen M, van der Ark A, van den Ijssel J, van
tive capabilities for all prokaryotes. Bioinfor- Straaten I, Jansen D, Beuvery C, Martens D,
matics 26(13):1608–1615 Tramper J (2008) Improving the cellular per-
9. Armenteros JJA, Tsirigos KD, Sønderby CK, tussis vaccine: increased potency and consis-
Petersen TN, Winther O, Brunak S, von tency. Vaccine 26(5):653–663. https://doi.
Heijne G, Nielsen H (2019) SignalP 5.0 org/10.1016/j.vaccine.2007.11.047
improves signal peptide predictions using 20. Luu LDW, Octavia S, Zhong L, Raftery M,
deep neural networks. Nat Biotechnol 37 Sintchenko V, Lan R (2017) Characterisation
(4):420–423 of the Bordetella pertussis secretome under dif-
10. Krogh A, Larsson B, Von Heijne G, Sonnham- ferent media. J Proteome 158:43–51
mer EL (2001) Predicting transmembrane pro- 21. Luu LDW, Octavia S, Zhong L, Raftery MJ,
tein topology with a hidden Markov model: Sintchenko V, Lan R (2018) Proteomic
62 Laurence Don Wai Luu and Ruiting Lan

adaptation of Australian epidemic Bordetella 24. Paoletti AC, Parmely TJ, Tomomori-Sato C,
pertussis. Proteomics 18(8):1700237 Sato S, Zhu D, Conaway RC, Conaway JW,
22. Herigstad B, Hamilton M, Heersink J (2001) Florens L, Washburn MP (2006) Quantitative
How to optimize the drop plate method for proteomic analysis of distinct mammalian
enumerating bacteria. J Microbiol Methods 44 mediator complexes using normalized spectral
(2):121–129 abundance factors. Proc Natl Acad Sci U S A
23. Hoofnagle AN, Whiteaker JR, Carr SA, 103(50):18928–18933
Kuhn E, Liu T, Massoni SA, Thomas SN, 25. Keasey S, Pugh C, Tikhonov A, Chen G,
Townsend RR, Zimmerman LJ, Boja E Schweitzer B, Nalca A, Ulrich RG (2010) Pro-
(2016) Recommendations for the generation, teomic basis of the antibody response to mon-
quantification, storage, and handling of pep- keypox virus infection examined in
tides used for mass spectrometry–based assays. cynomolgus macaques and a comparison to
Clin Chem 62(1):48–69 human smallpox vaccination. PLoS One 5
(12):e15547
 Chapter 5

Two-Dimensional Electrophoresis Coupled with Western

Blot as a Method to Detect Potential Neutralizing Antibody
Targets from Gram-Negative Intracellular Bacteria
Milan Obradovic and Heather L. Wilson

Abstract
Antigen selection is a critical step in subunit vaccine design, especially if the goal is to identify antigens that
can be bound by neutralizing antibodies to prevent invasion of cells by intracellular bacteria. Here, we
describe a method involving two dimensional gel electrophoresis (2-DE) coupled with western blotting
(WB) and mass spectrometry (MS) to identify bacterial proteins that: (1) interact with the host target cell
proteins, and (2) are targeted by antibodies from sera from infected animals. Subsequent steps would be
performed to validate that the bacteria are targeted by neutralizing antibodies to prevent invasion of the
eukaryotic cells.

Key words Lawsonia intracellularis, Two-dimensional gel electrophoresis, Neutralizing antibodies,

Antigen, Western blotting, Gram-negative bacteria

1 Introduction

2-DE coupled with an immunoblotting technique and MS is an

efficient, robust, and proven methodology to detect antigens from
bacteria, cancer cells, and fungi that are recognized by the human
immune system [1–5]. Multiple bacterial species have been ana-
lyzed using immunoproteomics which led to important discoveries
for Helicobacter pylori, uropathogenic Escherichia coli, Staphylococ-
cus aureus, Streptococcus thermophilus, and other bacteria research
[6–9]. Bacterial cells have lower complexity compared to mamma-
lian cells which readily allow for successful protein separation and
analysis using 2-DE and immunoproteomics [1]. Although 2-DE
coupled with western blotting and MS has been used for decades,
there are still limitations in the precise reproduction of gels and
difficulties in separating proteins that are hydrophobic, very acidic
or basic [3]. One of the solutions to mitigate these limitations is to
lower the complexity of the sample by analyzing specific regions of

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_5,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

63
64 Milan Obradovic and Heather L. Wilson

the bacterial or secreted bacterial products. We have developed a

protein selection procedure that, coupled with 2-DE, WB, and MS,
was able to identify proteins from intracellular bacteria recognized
by antibodies from hyperimmune serum and to determine which of
these immunogenic proteins bind to host cells [10].
Lawsonia intracellularis is an important pig pathogen that only
causes disease when it invades intestinal epithelial cells. The identi-
fication of L. intracellularis proteins that bind to and/or interact
with pig intestinal epithelial cell proteins will help identify neutra-
lizing antibody targets that, once bound by hyperimmune sera
antibodies, will prevent invasion. The proteins of L. intracellularis
detected with this method were further validated experimentally by
cloning the genes from proteins of interest to produce and purify
recombinant proteins. These recombinant proteins were then used
to generate protein-specific hyperimmune sera antibodies. We then
assessed whether L. intracellularis coated with each antigen-
specific hyperimmune sera failed to invade the intestinal pig epithe-
lial cells (IPECs). These data helped us identify which antigens were
targeted by neutralizing antibodies. These antigens were then used
for subunit vaccine development that was validated using in vitro
and in vivo studies [10, 11]. This immunoproteomic methodology
was efficient and practical in the selection of immunogenic proteins
from intracellular bacteria and their utilization as subunit vaccine
antigens.

2 Materials

All reagents are analytical grade unless otherwise specified and solu-
tions were made using deionized water. All reagents, solutions and gels
were used at room temperature, unless specified otherwise. All steps
are performed in 1.5 mL Eppendorf tubes unless specified otherwise.

2.1 Selection 1. Cy5 dye.

of L. intracellularis 2. Sodium bicarbonate buffer (NaHCO3): 100 mM NaHCO3
Proteins that Interact buffer consisting of 8.401 g in 1 L ddH2O adjusted to pH 8.2.
with Intestinal Pig
3. 3 kilodalton (kDa) molecular weight cutoff (3 K MWCO)
Epithelial Cells (IPEC) Amicon filters, 15 mL volume.
4. Bicinchoninic acid protein assay kit.
5. IPEC medium: Dulbecco’s Modified Eagle Medium
(DMEM)/F-12 with 5% Fetal bovine serum (FBS), insulin
(10 μg/mL), transferrin (5.5 μg/mL), selenium (5 ng/mL)
and 5 ng/mL of epidermal growth factor (EGF).
6. Radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris
pH 8, 150 mM sodium chloride, 0.1% SDS, 1% deoxycholic
acid, 1% Nonidet P-40 substitute, distilled water), complete
with 0.1 mM phenylmethylsulfonyl fluoride (PMSF) in
isopropanol.
 Identification of Lawsonia intracellularis Neutralizing Antibody Targets. . . 65

2.2 Rehydration 1. Immobilized pH gradient (IPG strip; Immobiline™ DryStrip,

of Immobilized pH pH 4–7, 13 cm, GE Healthcare) (see Note 1).
Gradient (IPG) Strips 2. Rehydration buffer, 10 mL volume: 9 M urea, 2% CHAPS
(Fisher BioReagents), 1% Dithiothreitol (DTT), 2% Pharma-
lyte pH 5–8 (GE Healthcare) (see Notes 2 and 3),
bromophenol blue.
3. Dry Strip Cover fluid (GE Healthcare).

2.3 Isoelectric 1. Dry Strip Cover fluid (GE Healthcare).

Focusing (IEF) 2. IPGphor device (GE Healthcare-Amersham Biosciences).
of Rehydrated Strips
3. Ettan IPGphor rehydration strip apparatus (Pharmacia
Using IPGphor Device
Biotech).

2.4 10% SDS 1. Sodium dodecyl sulfate (SDS) equilibration buffer 2 210 mL:
PAGE Gel 6 M urea, 75 mM Tris–HCl pH 8.8, 29.3% glycerol, 2% SDS,
0.002% bromophenol blue (can be stored at 20  C in 20 mL
aliquots).
2. DTT, 100 mg per 210 mL equilibration buffer (see Note 4).
3. 2.5% iodoacetamide, 250 mg per 210 mL equilibration buffer
(see Note 4).
4. Resolving gel buffer (total 25.32 mL for two gels; see Note 5):
25.32 mL double-distilled water (ddH2O), 21.36 mL 30%
acrylamide mix, 16.04 mL 1.5 M Tris pH 8.8, 0.64 mL 10%
SDS, 0.64 mL 10% ammonium persulphate (APS), and
25.6 μL tetramethylethylenediamine (TEMED).
5. Silver stain kit.

2.5 Semidry Transfer 1. Nitrocellulose membranes.

2. Immerse Blot Absorbent Filter paper.
3. Towbin transfer buffer: 25 mM Tris, 192 mM glycine, pH 8.3
with 20% methanol (v/v).
4. Plastic container.
5. Bio-Rad Trans-Blot SD semidry transfer cell.

2.6 Immunoblotting 1. Nitrocellulose membranes.

2. Plastic containers.
3. 10 Tris-buffered saline (TBS): 1.5 M NaCl, 100 mM Tris–
HCl, pH 7.4.
4. TBS containing 0.05% Tween 20 (TBST).
5. Blocking solution: 5% milk in TBS.
6. Diluent solution: 5% milk in TBST.
66 Milan Obradovic and Heather L. Wilson

7. Rabbit hyperimmune serum against L. intracellularis 1:500 in

5% milk in TBST (acquired from naı̈ve rabbits before immuni-
zation [negative sera] and from rabbits immunized with whole
inactivated bacteria) as primary antibodies.
8. Anti-rabbit IR 800 antibody (1 μg/mL).
9. Odyssey scanner (LI-COR®).

3 Methods

3.1 Selection 1. L. intracellularis propagation in McCoy cells, lysis, and protein

of L. intracellularis extraction is out of the scope of this chapter and further infor-
Proteins that Interact mation is available in previously published protocols [10, 12].
with the IPEC Cells 2. Add Cy5 dye to L. intracellularis proteins resuspended in
and Their Preparation NaHCO3 buffer in dye/proteins molar ratio of 8:1. Incubate
for 2-DE at room temperature for 4 h in the dark with gentle shaking on
the nutator.
3. After the 4-h incubation, add Cy5-stained bacterial proteins
into the 3 K MWCO 15 mL volume filters and centrifuge at
4700  g for 15 min. Wash with ddH2O three times using the
same settings (see Note 6).
4. After the final wash, carefully resuspend the Cy5-stained pro-
teins with ddH2O, pipetting up/down from the walls of the
filter chamber. Transfer resuspended Cy5-L. intracellularis
proteins to a new Eppendorf tube.
5. Determine the concentration of Cy5-labeled proteins using a
BCA protein assay kit.
6. Add 700 μg of Cy5-L. intracellularis proteins into the antibi-
otic- and FBS-free IPEC medium with 106 IPEC cells.
7. Incubate for 3 h with gentle shaking on a nutator at 4  C in
the dark.
8. After the 4-h incubation, centrifuge the mixture at 500  g for
10 min to pellet the cells. Discard the supernatant (see Note 7).
9. Add RIPA buffer with PMSF to the cell pellet to lyse the IPEC
cells. Freeze/thaw the cells using liquid nitrogen and a water
bath at 37  C.
10. Centrifuge 10,000  g for 10 min and preserve the
supernatant.
11. Add 4 volumes of ice-cold acetone to the supernatant, vor-
tex, and store at 20  C for 1 h.
12. Centrifuge at 14,000  g for 10 min. Resuspend the pellet in
NaHCO3 buffer. Perform the BCA protein assay and store at
20  C (see Note 8).
 Identification of Lawsonia intracellularis Neutralizing Antibody Targets. . . 67

13. For analytical gels, add 250 μg IPEC + Cy-5 labeled L. intra-
cellularis proteins to rehydration buffer up to the 250 μL
volume. For MS preparatory gels, add 600 μg of L. intracellu-
laris proteins in rehydration buffer up to the 250 μL volume.

3.2 Rehydration 1. Pipet 250 μL of protein sample in rehydration buffer into one
of Two 13 Cm pH 4–7 lane of the Ettan IPGphor rehydration strip apparatus. Slowly
IPG Strips (See Note 9) discharge the sample into the rack lane about the same length
of the strip. Avoid creating bubbles (see Note 10).
2. Remove strip from a 20  C freezer.
3. Carefully peel off the protective backing from the strip.
4. Gently place the strip (gel side down) onto the sample, being
careful not to create bubbles and ensuring the entire strip is in
contact with the sample.
5. Cover strip with 3–4 mL of Dry Strip Cover fluid.
6. Place the lid on a rack and leave on a level bench for at least 10 h
or overnight, in the dark.

3.3 IEF of Rehydrated 1. Place strip in a strip holder, gel side up—positive end of gel at
Strips Using IPGphor the pointed end of the holder and negative end all the way to
the flat side.
2. Place 1–1.5 cm long paper strips (presoaked with water and
blotted almost completely dry) on each end of the strip ensur-
ing some overlap with the gel.
3. Place electrodes on the ends of a filter paper not in contact with
the gel.
4. Completely cover with Dry Strip Cover fluid (around 4.5 mL).
5. Ensure that pointed ends of holders (with the positive end of
the strip) are in the positive area of the apparatus.
6. Run focusing: 150 V step and hold for 3 h.
300 V step and hold for 1200 Volt hours (Vh).
1000 V gradient for 3900 Vh.
8000 V gradient for 13,500 Vh.
8000 V step and hold for 25,000 Vh.
7. After focusing, strips are placed in a tube, gel side up, and
stored at 80  C until ready for the next step (see Note 11).

3.4 10% SDS-PAGE 1. Run the second dimension on a medium size 10% SDS-PAGE
Gel, Using Bio-Rad gel, using Bio-Rad PROTEAN II xi cell apparatus connected to
Protean II Xi Cell a circulating water bath to maintain temperature at around
Apparatus 10  C.
2. Thaw the required number of strips at room temperature
(RT) for 15–30 min.
68 Milan Obradovic and Heather L. Wilson

3. For each strip, prepare 2 10 mL portions of the SDS equili-

bration buffer in 15 mL tubes.
4. Label one tube as “DTT in equilibration buffer” and add in
10 mg DTT to 10 mL of equilibration buffer.
5. Label second tube as “Iodoacetamide in equilibration buffer”
and add in 25 mg iodoacetamide to 10 mL equilibration buffer.
6. Add each strip to the tube labeled “DTT in equilibration
buffer” and gently rock for 15 min at RT.
7. Remove strip and add to the tube labeled “Iodoacetamide in
equilibration buffer” and gently rock for 15 min at RT.
8. Prepare the resolving buffer in a 100 mL glass flask. Add a
magnetic stir bar and de-gas with a vacuum while stirring the
buffer for 10 min, being careful not to aspirate the buffer. Add
APS and TEMED, as described in Subheading 2.4 above, and
mix while avoiding the creation of bubbles. Using a 50 mL
glass pipet, slowly pour the buffer within a 20 cm  21 cm  1 mm
gel glass cassette to cast the gel. Leave space for a stacking gel
and gently overlay with water or 0.1% SDS solution to prevent
excessive drying of the upper part of the gel.
9. Dip a strip briefly into SDS PAGE running buffer and apply to
the well with the positive end farthest from the protein ladder
well and with the gel facing the shorter glass plate. Gently push
the strip to the bottom of the well with a comb, pipette tip or
forceps.
10. Overlay with 1–1.5 mL of 0.5% agarose in running buffer for
each strip using the minimum volume to submerge the strip.
Allow the agarose 1 min to cool and solidify.
11. Add protein ladder to marker lane and run the gel at 110 V for
16 h. Alternatively, run the gel at 90 V overnight.
12. After electrophoresis, carefully separate the gel plates with the
use of a plastic spatula. The gel should remain intact on one of
the glass plates. Use gloves and avoid touching the center of
the gel in order to prevent contamination. Wash the analytical
gel with ddH2O and transfer it carefully to a container with a
western blot transfer buffer.
13. Transfer the preparative gel to another container for silver
staining (see Note 12).
14. Perform silver staining of the preparative gel using a Silver stain
kit (Sigma-Aldrich) and following the manufacturer’s protocol.
After the silver staining, the preparative gel should be covered
with 1% acetic acid and stored at 4  C, overnight.

3.5 Semidry Transfer

 Identification of Lawsonia intracellularis Neutralizing Antibody Targets. . . 69

1. Cut a nitrocellulose membrane and absorbent filter paper to

the size of the analytical gel and immerse in Towbin transfer
buffer for 30 min at RT (see Note 13).
2. Place the presoaked filter paper at the middle of the plate
surface of a semidry transfer cell.
3. Place the precut nitrocellulose membrane, from step 1 above,
on top of the filter paper and ensure that two surfaces match
in size.
4. Carefully place the analytical gel above the nitrocellulose mem-
brane to avoid creating bubbles or wrinkles.
5. Place the filter paper over the analytical gel. Use a glass tube or
roller to gently roll over the upper filter paper and remove
bubbles or excessive buffer. Close the lid of the machine.
6. Semidry transfer for this size of the gel is achieved using the
Bio-Rad Trans-Blot SD semidry transfer cell at 15 V for 60 min
(see Note 14).

3.6 Immunoblotting 1. Block the nitrocellulose membrane with 50 mL of 5% milk in

TBS for 1 h.
2. Remove 5% milk in TBS and wash 3 times for 10 min
with TBST.
3. Add 50 mL of a 1:500 dilution of hyperimmune rabbit serum
in 5% milk with TBST.
4. Incubate with gentle shaking on a nutator overnight at 4  C.
5. Decant milk (see Note 15) and wash 3 times for 10 min
with TBST.
6. Add secondary antibody IR800 (1:10,000 dilution) to 50 mL
of 5% milk in TBST and incubate with gentle shaking, for
30 min to 1 h, covered to protect from the light.
7. Following incubation, remove the secondary antibody and
wash 3 times for 10 min with TBST.
8. Scan with LI-COR® Odyssey Scanner in IR700 and IR800
channels (see Fig. 1a).
9. Save the digital copy of the image for further analysis or print it
on A4 paper (see Note 16).

3.7 Excising Gel 1. Excise gel with silver-stained dots which correspond to IR-800
Plugs for Mass labeled proteins detected by WB analysis using a sterile biopsy
Spectrometry punch (3 mm diameter) in a biological cabinet to avoid con-
tamination of gel samples with environmental proteins
(Fig. 1b).
2. Collect gel plugs and store them in Eppendorf tube with
100 μL ultrapure water at 20  C. Send gel plug samples for
MS analysis (see Notes 17 and 18).
70 Milan Obradovic and Heather L. Wilson

Fig. 1 Protein separation by 2-dimensional electrophoresis and selection of

spots for mass spectrometry, cropped image. Cy5-labeled Lawsonia
intracellularis proteins that bound to IPEC1 cells were subjected to isoelectric
focusing using IPG strip 4–7 (horizontal plane) followed by SDS-PAGE using 10%
SDS-PAGE gel (vertical plane). Molecular weight markers are indicated (kDa). (a)
Proteins were transferred to a nitrocellulose membrane and incubated with
hyperimmune rabbit serum as primary antibody and anti-rabbit IR800
secondary antibody. Proteins visible in IR700 channel are red and indicate all
Cy5-labeled bacterial proteins. The proteins visible in IR800 channel are green
and indicate proteins bound by rabbit antibodies from rabbits immunized with
whole-cell L. intracellularis. Note, all green proteins are also red and therefore
should appear yellow in colour but they are overwhelmed by the green
fluorescence. (b) A replicate gel was stained with PROTSil-1 silver stain kit.
Position and numbering of gel spots are indicated by red circles. Gel plug
samples 1.4, 2.3, 3.1, 3.2, and 4 were submitted for mass spectrometry.
(Reproduced from [10] with permission from Elsevier)
 Identification of Lawsonia intracellularis Neutralizing Antibody Targets. . . 71

4 Notes

1. The adequate pH and type of the IPG strip should be evaluated

experimentally before the main experiment.
2. If the protocol requires IPG strips pH 3–10, then Pharmalyte
pH 3–7 should be used.
3. DTT and Pharmalyte should be added last, just before mixing
with the protein sample. Store DTT at 20  C and Pharmalyte
at 4  C.
4. Equilibration buffer with DTT or iodoacetamide is aliquoted
in 20 mL volume, in 50 mL conical tubes, and stored at
20  C. Use of aliquots are suggested to avoid frequent
freeze-thaw cycles, which could damage the properties of the
buffer.
5. We ran two 10% SDS PAGE gels (analytical and preparatory) at
the same time to have consistent results.
6. Washing of Cy5-labeled proteins is important to remove the
unbound Cy5 dye from the buffer. The remaining Cy5-labeled
proteins are visible as a blue pellet at the bottom of the filter
chamber. The size of the filter is based on the size of the
proteins that you want to extract and analyze downstream.
7. The speed and the time of centrifugation depend on the type of
cells and should be experimentally standardized. Targeted
Cy5-labeled L. intracellularis proteins bind to the IPEC cells
and the unbound proteins are left in the supernatant.
8. To further concentrate the protein sample use 3 kDa centrifu-
gal filters (14,000  g for 15 min).
9. Strip 1 is for the analytical gel and strip 2 is for the
preparatory gel.
10. For gels, use 250 μg proteins for analytical gel and 600 μg
proteins for preparatory gel in rehydration solution, total
250 μL volume, for each per IPG strip.
11. The strip can be imaged after the IEF to confirm that IEF
worked well as evidenced by sharp bands. We used the Odyssey
scanner (Li-COR) with IR 700 channel since Cy5 is fluorescent
at 700 nm.
12. The preparatory gel was left on one of the glass plates during
silver staining and storage to ensure that the gel did not break
and to allow easier cutting of gel plugs with a biopsy puncher.
13. Readers can perform the semidry transfer using their own
lab kits.
72 Milan Obradovic and Heather L. Wilson

14. Shorter time could be used but with higher voltage such as
25 V for 30 min. The time and voltage should be established
based on the size of the gel and semidry apparatus protocol.
15. From our experience, hyperimmune rabbit serum in 5% milk in
TBST could be used again up to 4 times if stored at 20  C.
16. There are multiple image software and projectors available for
processing of immunoblot image and correlation of protein
spots to the preparative gel. This is useful if larger and complex
proteome is analyzed. Although we did not have this equip-
ment, we performed selection and Cy5 staining of proteins (see
Subheading 3.1), thus reducing the number of proteins on the
membrane and increasing their detection. We were able to
compare the printed image of the immunoblot to the silver
stained gel and select gel plugs for MS with great accuracy.
17. Gel plug samples (annotated as 1.4, 2.3, 3.1, 3.2, and
4 (Fig. 1b)) were sent to Plateforme Protéomique Centre de
Recherche du CHU de Québec CHUL, Québec, Canada for
MS analysis.

Acknowledgments

This chapter is published with permission of the director of VIDO

as journal series #916.

References
1. Rabilloud T, Strub JM et al (2001) A compari- immunogenic proteins. Electrophoresis 20
son between Sypro ruby and ruthenium II tris (4–5):1001–1010
(bathophenanthroline disulfonate) as fluores- 6. Haas G, Karaali G et al (2002) Immunopro-
cent stains for protein detection in gels. Prote- teomics of Helicobacter pylori infection and
omics 1(5):699–704 relation to gastric disease. Proteomics 2
2. Lahner E, Bernardini G et al (2011) Immuno- (3):313–324
proteomics of Helicobacter pylori infection in 7. Hagan EC, Mobley HLT (2007) Uropatho-
patients with atrophic body gastritis, a predis- genic Escherichia coli outer membrane antigens
posing condition for gastric cancer. Int J Med expressed during urinary tract infection. Infect
Microbiol 301(2):125–132 Immun 75(8):3941–3949
3. Havlasova J, Hernychova L et al (2005) Pro- 8. Enany S, Yoshida Y et al (2013) Two dimen-
teomic analysis of anti-Francisella tularensis sional electrophoresis of the exo-proteome
LVS antibody response in murine model of produced from community acquired methicil-
tularemia. Proteomics 5(8):2090–2103 lin resistant Staphylococcus aureus belonging to
4. Kellner R, Lichtenfels R et al (2002) Targeting clonal complex 80. Microbiol Res 168
of tumor associated antigens in renal cell carci- (8):504–511
noma using proteome-based analysis and their 9. Salzano AM, Novi G et al (2013) Mono-
clinical significance. Proteomics 2 dimensional blue native-PAGE and
(12):1743–1751 bi-dimensional blue native/urea-PAGE
5. Pitarch A, Pardo M et al (1999) or/SDS-PAGE combined with nLC-ESI-LIT-
Two-dimensional gel electrophoresis as analyt- MS/MS unveil membrane protein heteromeric
ical tool for identifying Candida albicans
 Identification of Lawsonia intracellularis Neutralizing Antibody Targets. . . 73

and homomeric complexes in Streptococcus 11. Obradovic MR, Wilson HL (2020) Immune
thermophilus. J Proteome 94:240–261 response and protection against Lawsonia
10. Obradovic M, Pasternak JA et al (2019) Immu- intracellularis infections in pigs. Vet Immunol
noproteomic analysis of Lawsonia intracellu- Immunopathol 219:109959
laris identifies candidate neutralizing antibody 12. Lawson GH, McOrist S et al (1993) Intracel-
targets for use in subunit vaccine development. lular bacteria of porcine proliferative enteropa-
Vet Microbiol 235:270–279 thy: cultivation and maintenance in vitro. J Clin
Microbiol 31(5):1136–1142
 Chapter 6

Panproteome Analysis of the Human Antibody Response

to Bacterial Vaccines and Challenge
Joseph J. Campo and Amit Oberai

Abstract
High-density protein microarray is an established technology for characterizing host antibody profiles
against entire pathogen proteomes. As one of the highest throughput technologies for antigen discovery,
proteome microarrays are a translational research tool for identification of vaccine candidates and biomar-
kers of susceptibility or protection from microbial challenge. The application has been expanded in recent
years due to increased availability of bacterial genomic sequences for a broader range of species and strain
diversity. Panproteome microarrays now allow for fine characterization of antibody specificity and cross-
reactivity that may be relevant to vaccine design and biomarker discovery, as well as a fuller understanding of
factors underlying themes of bacterial evolution and host–pathogen interactions. In this chapter, we present
a workflow for design of panproteome microarrays and demonstrate statistical analysis of panproteomic
human antibody responses to bacterial vaccination and challenge. Focus is particularly drawn to the
bioinformatics and statistical tools and providing nontrivial, real examples that may help foster hypotheses
and rational design of panproteomic studies.

Key words Protein microarray, Panproteome, Antigenic diversity, Antibody profiling, Differential
analysis

1 Introduction

Advances in proteomics tools have been drivers of vaccine antigen

discovery. High throughput gene cloning and cell-free in vitro
transcription and translations (“IVTT”) of hundreds to thousands
of proteins at a time allows proteome-wide analysis of specific
antibody binding in a single assay. Expressed proteins are immobi-
lized as spots in a 2D grid format on a planar surface, such as
nitrocellulose-coated glass microscope slides. These protein arrays
are typically fabricated using robotic microarray printers utilizing
contact print pins or contactless droplet jet printers. Antibody-
containing biological specimens can be probed over the printed
arrays, and antibody binding to specific, known proteins can be
revealed using fluorescent or colorimetric secondary antibody

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_6,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

75
76 Joseph J. Campo and Amit Oberai

detection systems. Proteome microarray is now an established

technology. For specific details about the laboratory protocols, we
refer the readers to a comprehensive chapter in a previous volume
of this series by Driguez et al. for schistosome protein microarrays
[1]. The same principles described by Driguez et al. apply for
bacterial proteomes, as well, and we follow a standard development
pipeline for protein microarray development and antibody profiling
(see Fig. 1).
Genomic sequence information is readily accessible for numer-
ous pathogens, often for multiple laboratory strains and clinical
isolates encompassing broad species and strain coverage of bacterial
populations. This advance in microbial genomics is driven by
improvements in whole genome sequencing and reduction in
costs of sequencing a bacterial genome [2]. Characterization of
bacterial genomic diversity allows for the definition of a pangenome
and, similarly, the protein-coded features that constitute a panpro-
teome. However, the definition of a panproteome with regards to
microarray development varies depending on the scientific ques-
tions addressed and availability of genomic sequence data. At the
species level, a panproteome may encompass all genes that are
common between species of interest, as well as the genes that are
unique to each species. Within species, at the strain level, a panpro-
teome may encompass genes encoding the common “core prote-
ome,” those displaying allelic diversity and unique genes in the
accessory genome. And at the individual protein level, focus can
be drawn to the repertoire of protein fragments or epitopes repre-
senting genetic polymorphisms. Broadly defined, the panproteome
used for microarray development comprises protein sequences
providing enough coverage of genomic diversity to address the
scientific questions.
Using whole genome sequencing data on Streptococcus pneu-
moniae clinical isolates from pediatric nasopharyngeal swabs [3],
we developed our first panproteome microarray for profiling pneu-
mococcal antibodies in participants of a trial of a pneumococcal
whole cell vaccine [4, 5]. The array contents included proteins
present in nearly all isolates sequenced, constituting the core pneu-
mococcal proteome, as well as genes present in at least 20% of
isolates. Also included in the array was an expanded repertoire of
variants of zinc metalloproteases ZmpA and ZmpB and pneumo-
coccal surface proteins A (PspA) and C (PspC), constituting the
“diverse core loci” (Fig. 2). This approach has been applied for
development of panproteome microarrays for other pathogens,
such as Brucella spp. and Treponema pallidum subsp. pallidum
strains. It has also highlighted the complexities of targeting highly
diverse bacteria such as Clostridium spp. commonly found as
 Panproteome Analysis of Human Antibody Responses 77

Fig. 1 Proteome microarray development and antibody profiling workflow (reproduced from https://
antigendiscovery.com/adi-proteome-microarray-technology/ with permission from Antigen Discovery, Inc.)

Fig. 2 Selection of S. pneumoniae genes for inclusion in panproteome microarray. Among 616 sequenced
pediatric clinical isolates, genes present in at least 20% of the bacterial population were included in an array
containing 4296 proteins

commensal bacteria in the human gut microbiome. This chapter

describes the process of developing panproteome microarrays from
concept to protein selection and the specific bioinformatics ana-
lyses. Wet lab protocols have been published previously [1] and will
not be detailed herein, but statistical methods pertinent to
panproteome-wide analysis are presented.
78 Joseph J. Campo and Amit Oberai

2 Materials

2.1 Genomic A prerequisite for panproteome microarray development is access

Information to whole genome sequencing information; at least, two genomes
are required (see Note 1). The source of data can be public data-
bases or privately curated datasets.
1. Genbank (GCA) is an annotated collection of all publicly avail-
able nucleotide and protein sequences derived from public
sequence repositories and, thereby containing potential
redundancies.
2. Refseq (GCF) provides a complete set of nonredundant, exten-
sively cross-referenced and annotated nucleotide and protein
sequences for each species and strain, with assembly records
maintained by the NCBI.

2.2 Data Analysis 1. BLAST+ (version 2.11.0 or later) https://ftp.ncbi.nlm.nih.

Software (See Note gov/blast/executables/blast+/LATEST/
2 for Descriptions) 2. PSORTb (version 3.0.2 or later) [6] https://www.psort.org/
psortb/
3. TMHMM (version 2.0 or later) [7] http://www.cbs.dtu.dk/
services/TMHMM/
4. SignalP (version 5.0 or later) http://www.cbs.dtu.dk/ser-
vices/SignalP/
5. NCBI web CD-Search tool [8] https://www.ncbi.nlm.nih.
gov/Structure/bwrpsb/bwrpsb.cgi
6. R (version 4.0.3 or later) https://www.r-project.org/
7. R studio (version 1.3 or later) https://www.rstudio.com/ (see
Note 3).
8. R packages for analysis of differential reactivity.
(a) limma (see Note 4) [9].
(b) Lme4 (see Note 5).
(c) ROCR.
(d) Rtsne (see Note 6) [10].
9. R packages for data visualization.
(a) ggplot2 (https://ggplot2.tidyverse.org/),
(b) gplots,
(c) ComplexHeatmap https://jokergoo.github.io/Com
plexHeatmap-reference/book/
 Panproteome Analysis of Human Antibody Responses 79

3 Methods

3.1 Select Pathogen Identify an organism or set of organisms from which to perform
genome-wide homology analysis prior to designing the contents of
a panproteome microarray. Base the selection of organisms on the
research questions, but balance factors such as limitation of array
space (i.e., number of proteins), and quality and source of genomic
information. See Note 7 for considerations in selecting organisms.

3.2 Identify The goal is to capture species or strain diversity by identifying the
Conserved and Diverse panproteome, or the highly diverse proteins, for example those
Proteins under immune pressure. The steps involved in capturing the pan-
proteome are as follows, using an example of 3 selected species:
1. Download whole proteome sequences for the group of species
or strains of the pathogens of interest that will constitute the
panproteome microarray. The fasta format files “cds_from_ge-
nomic.fna” and “translated_cds.faa” (nucleotide and protein
files, respectively) are downloaded from NCBI Genbank or
Refseq, or they may be obtained from private curated whole
genome sequencing data (see Note 8).
2. Run BLASTp for all vs. all sequence homology comparison for
the selected species or strains. The stand-alone BLAST is run in
the windows command prompt (cmd) application. The first
step in running BLASTp is to format each of the proteome
fasta files for BLAST using the “makeblastdb” command. For
comparing 3 species A, B, and C, with proteome sequence files
A_translated_cds.faa, B_translated_cds.faa, and C_transla-
ted_cds.faa, the makeblast db command would be as follows:

makeblastdb -in A_translated_cds.faa -dbtype prot

makeblastdb -in B_translated_cds.faa -dbtype prot
makeblastdb -in C_translated_cds.faa -dbtype prot

3. In BLASTp, align all sequences in each species to all other

species in both directions, that is,
– A is compared to itself, B and C;
– B is compared to itself, A and C; and.
– C is compared to itself, A and B.
There will therefore be nine output alignment files gener-
ated for a three species comparison:
– with respect to A: AA, AB, AC,
– with respect to B: BB, BA, BC,
– with respect to C: CC, CA, CB,
80 Joseph J. Campo and Amit Oberai

The BLASTp command used for one of these comparisons,

for example A vs B (the entire command should be typed in one
continuous line):

"C:\Program Files\NCBI\blast-2.11.0+\bin\blastp" -query

A_translated_cds.faa -subject B_translated_cds.faa -outfmt "6
qseqid sseqid pident qstart qend qlen sstart send slen evalue"
-max_hsps 1 -max_target_seqs 1 -evalue 1e-10 > AB.out

In the BLASTp command line, -outfmt 6 refers to tabular

alignment output format. The parameters chosen here for
output are:
– qseqid (query sequence ID),
– sseqid (subject sequence ID),
– pident (percent identical matches in the aligned region),
– qstart (query protein start position in the alignment),
– qend (query protein end position in the alignment),
– qlen (length of the query protein sequence),
– sstart (subject protein start position in the alignment),
– send (subject protein end position in the alignment),
– slen (length of the subject protein sequence),
– max_hsps 1 (show only the best high scoring pair or align-
ment for a query-subject pair),
– max_target_seqs 1 (show only one aligned query-subject
pair),
– e-value (a statistical score for the alignment equal to the
number of expected hits of similar quality (score) that could
be found just by chance—when e-value<0.01 it is almost
identical to p-value), and,
– AB.out (is the name of the output alignment file in tab
delimited format).
4. Open the R program and find the working drive by typing the
getwd() command into the command line, or set the working
drive using the setwd() command.
5. Move the homology search output A_vs_B.out into the R
working drive and import the data into R using the read.
table() command as follows:

> blastAB.df <- read.table(“AB.out”, sep = "\t", stringsAs-

Factors = FALSE, check.names = FALSE)
 Panproteome Analysis of Human Antibody Responses 81

6. Calculate an overall percent identity for each alignment as:

> overall_pid_query <- blastAB.df$pident * (blastAB.df$qend –

blastAB.df$qstart + 1) / blastAB.df$qlen
> overall_pid_subject <- blastAB.df$pident * (blastAB.df
$send – blastAB.df$sstart + 1) / blastAB.df$slen

7. Group the sequences that are common to all species as follows.

A threshold percent identity value is set (typically 80% or 90%,
see Note 9) using the overall percent identity of aligned pro-
teins returned by BLASTp, and the aligned protein pair is
required to have identity above that threshold value for consid-
eration to be added to the group.
The hits are first sorted into categories or groups as:
– For Species A: A-B-C, else A-B, else A-C else A-A.
– For Species B: A-B-C, else A-B, else B-C, else B-B.
– For Species C: A-B-C, else A-C, else B-C, else C-C.
Note that the red bold text (above) is intended to highlight
the reference proteome with respect to which comparison is
being done. For example, for species A, A-B-C is a protein that
has a blast hit between species A and B, and between species A
and C, above the required threshold percent identity. If not,
then the protein is sorted at the next level A-B which means the
protein has a blast hit between A and B, and if not then at the
next level A-C for a blast hit between A and C, and finally A-A.
A-A is a singleton which did not have any hits above the
required threshold with species B or C. The same is done for
species B and for species C. In this way, all hits for the 3 species
are sorted into one of these shown categories. Redundancy of
repeating a protein between different category comparisons is
avoided by removing the protein from subsequent consider-
ation for grouping once it has been counted into one of the
groups or categories. Note that the complexity of groupings
increases factorially with number of species selected. An exam-
ple of setting the threshold identity for grouping in R is coded
as follows:

> perc <- 80 #threshold for homology

> blastAB_grp.df <- blastAB.df[overall_pid_query >= perc &
overall_pid_subject >= perc, ] #80%ID threshold over full
length of query sequence and subject sequence
> blastABq.vec <- blastAB_grp.df[, 1] #AB is the query IDs for
hits with species 2
> blastABs.vec <- blastAB_grp.df[, 2] #sseqid ABs is the
subject ID’s
82 Joseph J. Campo and Amit Oberai

3.3 Classify Proteins Classify proteins by intersections between groups, here using the
in the Core 3 species example homology analysis in Subheading 3.2.
and Panproteome
1. The proteins that intersect in groups A-B-C, A-B-C and A-B-C
are the protein sequences that are common to all species. This
group of proteins forms the core proteome. Define vectors
group_abc.vec as the vector containing proteins sequence
IDs from proteins in group A-B-C, group_bac.vec as the
vector containing protein sequence IDs from proteins in group
A-B-C, and group_cab.vec as vector containing the protein
sequence IDs from proteins in group A-B-C, as follows:

> group_abc.vec <- c()

> for(idx in 1:length(blastAA.vec)) {
if ((blastAA.vec[idx] %in% blastAB.vec) & (blastAB.vec[idx] %
in% blastAC.vec)) { ##all 3 overlap 123
idx2 <- grep(blastAA.vec[idx], blastABq.vec)
idx3 <- grep(blastAA.vec[idx], blastACq.vec)
x3a <- paste(blastAA.vec[idx], blastABs.vec[idx2], blastACs.
vec[idx3], sep = "~") ##all 3 sequence ID’s in the hit with
species 1
group_abc.vec <- c(group_abc.vec, x3a) }}

2. Calculate the intersections of all group vectors, as follows:

> intersect_all.vec <- intersect(intersect(group_abc.vec,

group_bac.vec), group_cab.vec)

3. For proteins not common to all species, the largest group of

species to which they are common is identified. This process of
identifying the largest group of species to which a protein is
found in common continues as proteins common to groups of
2 species are defined. If we define vectors group_ab.vec as
the vector containing protein sequence IDs from proteins in
group A-B, group_ba.vec as the as the vector containing
protein sequence IDs from proteins in group A-B,
group_ac.vec as the vector containing protein sequence
IDs from proteins in group A-C, group_ca.vec as the vector
containing protein sequence IDs from proteins in group A-C,
group_bc.vec as the vector containing protein sequence IDs
from proteins in group B-C, and group_cb.vec as the vector
containing protein sequence IDs from proteins in group B-C,
then:
 Panproteome Analysis of Human Antibody Responses 83

> intersect_ab.vec <- intersect(group_ab.vec, group_ba.vec)

> intersect_ac.vec <- intersect(group_ac.vec, group_ca.vec)
> intersect_bc.vec <- intersect(group_bc.vec, group_cb.vec)

4. The remaining proteins are singletons, or those that are not

common to any other species. These groups of proteins that are
outside the core proteome form the panproteome. Tabulate
the hits for each intersection. For example, during our devel-
opment of a panproteome microarray for Brucella spp., five
organisms were selected and assessed for core and panpro-
teomes, with the results tabulated in Table 1 (see Note 7).
5. For larger selections of organisms, for example capturing strain
diversity among numerous clinical isolates, the sequences of
proteins encoded by all unique clusters of orthologous genes
(COGs) in the sequenced bacterial population are queried
against each isolate genome, and the number of genome hits
above the percent identity threshold for each protein is tabu-
lated. COGs that are to be included in the panproteome micro-
array are filtered by prevalence in the bacterial population. For
example, a threshold of 20% was used to gate for genes
included in our S. pneumoniae panproteome microarray
(Fig. 2), that is, all COGs present in at least 20% of pediatric
clinical isolates were included in the array design (see Note 10)
[4]. Additionally, highly diverse proteins ZmpA, ZmpB, PspA,
and PspC present in all isolates were assessed for divergence in
order to select an expanded repertoire of these “diverse core
loci” (Fig. 3).
While the intended goal is to capture as much species and strain
diversity as possible on the panproteome microarray, a major con-
sideration is the number of available spots on the array. In the case
of close homology between the included species, the core proteome
may define the bulk of the proteins involved, as in the case of the
Nichols and SS14 strains of syphilis (Treponema pallidum subsp.
pallidum, Table 2). In this case, the diversity (or the pan-proteome)
will be small, and capturing the diversity on the array will be easier.
However, in the case where there is low homology between species
leading to failure to identify a core proteome, a large fraction of the
proteomes will constitute the panproteome. Consequently, it may
not be possible to be as inclusive of the diversity due to the limita-
tion of the number of spots on the array. An example of such a
failure is shown in our analysis of the pathogen and gut commensal
Clostridium spp. (Table 3), where the pairwise distribution of
shared genes greater than a permissive 70% sequence identity
threshold was low between species (Fig. 4).
84 Joseph J. Campo and Amit Oberai

Table 1
Brucella species homology analysis for classification of the core proteome and panproteome
 Panproteome Analysis of Human Antibody Responses 85

Fig. 3 Defining variants for the four core variable Streptococcus pneumoniae antigens. Each histogram shows
the distribution of pairwise similarities between representatives of (a) PspA, (b) PspC, (c) ZmpA, and (d) ZmpB.
The vertical red lines on each plot represent the thresholds that were used as a cut-off to define distinct
variants: 0.525 for PspA; 0.675 for PspC; 0.575 for ZmpA, and 0.475 for ZmpB (reproduced from [4] with
permission from PNAS)

3.4 Down-Select Given limitations in the number of proteins on the microarray, we

Proteins from the Core can downselect the proteins in the panproteome using a set of
and Panproteome if criteria to allow as much inclusivity as possible to capture maximum
Necessary diversity, while limiting the proteins to the number of available
spots. Firstly, the defined core proteome will constitute a part of
the array. Secondly, the remaining spots that are available are
assigned to proteins using a scoring scheme (see Note 11) for top
86 Joseph J. Campo and Amit Oberai

Table 2
T. pallidum subsp. pallidum (T.p.p.) Nichols and SS14 strain homology analysis for classification of
the core proteome and panproteome

ranked proteins that is made from tabulating results of protein

feature prediction software, including PSORTb (see Note 12),
TMHMM, SignalP, and the NCBI web CD-Search tool. Addition-
ally, a dictionary of keywords (see Note 13) is supplemented to
these criteria for scoring the top antigens and picking out antigens
missed by the above scoring method. The process is executed as
follows:
1. Annotate proteins using protein feature prediction software.
(a) PSORTb: Open a web browser and navigate to the
PSORTb webserver (see Subheading 2.2, item 2). Select
options from scroll down menus: (1) bacteria or archaea;
(2) Gram stain negative or Gram stain positive with sub-
option of with or without outer membrane; (3) output
format long or short; and (4) show results by email or on
the web. Upload the sequence file in fasta format or paste
fasta format sequences in the box and click “Submit.” Use
the “Localization” header in the output file for scoring
proteins by priority, depending on the design of the study
(e.g., “Cell Membrane” > “Cytoplasmic”). Set the
weighting of scores depending on the objectives of the
study.
l Use Deeploc 1.0 for Eukaryote subcellular localization
prediction http://www.cbs.dtu.dk/services/
DeepLoc/
(b) TMHMM: Open a web browser and navigate to the
TMHMM webserver (see Subheading 2.2, item 3).
Upload the sequence file in fasta format or paste fasta
format sequences in the box and click “Submit.” Use the
“PredHel” field in the output file, which indicates the
number of predicted transmembrane domains, for scoring
 Panproteome Analysis of Human Antibody Responses 87

Table 3
Clostridium species homology analysis for classification of the core proteome (yellow highlighted for
3 or more species) and panproteome

proteins by priority, depending on the design of the study

(e.g., PredHel >0 yields a score of 1 for a study prioritiz-
ing any transmembrane domains).
(c) SignalP: Open a web browser and navigate to the SignalP
webserver (see Subheading 2.2, item 4). Upload the
sequence file in fasta format or paste fasta format
sequences in the box. Select the organism group from
options: Eukarya, Gram-positive, Gram-negative, or
Archaea. Select output format from options: Long output
and Short output (preferred). Click “Submit.” Click the
“Download” button to download the “Prediction sum-
mary” as a text file. Use the “Prediction” field in the
88 Joseph J. Campo and Amit Oberai

Fig. 4 Distribution of percent identity between Clostridium species. The histograms show for pairwise BLASTp
results for (a) C. difficile vs. C. perfringens and (b) C. scindens vs. P. bifermentans showing that most genes
are poorly conserved between species

output file to score proteins by presence or absence of any

of the signal peptides: Sec signal peptide (Sec/SPI), Lipo-
protein signal peptide (Sec/SPII) or Tat signal peptide
(Tat/SPI). If “Other” has the highest likelihood, then
there are no predicted signal peptides.
(d) CDSearchTool: Open a web browser and navigate to the
CD Search Tool webserver (see Subheading 2.2, item 5).
Upload the sequence file in fasta format or paste fasta
format sequences in the box and click submit. You may
change the “Adjust search options” if needed or use the
default settings. The retrieved ontologies are given e-value
scores. Select an e-value threshold acceptable for the study
 Panproteome Analysis of Human Antibody Responses 89

(e.g., <1  10 10) and score proteins based on significant

ontologies that are included in a keyword list, defined by
the study objectives (see Note 13).
2. After running all protein feature predictions, tabulate results by
adding the weighted scores for each protein and sorting pro-
teins by highest score.
3. Use the ranked list of proteins to fabricate panproteome micro-
arrays using the desired methods for expressing and printing
proteins, or following the published protocols [1].

3.5 Statistical Typical methods for univariate differential reactivity analysis have
Analysis been reported extensively for proteome microarrays (see https://
of the Panproteome antigendiscovery.com/publications/), which also apply for pan-
Antibody Response proteome analyses. Examples are given from our studies of panpro-
to Vaccines teome antibody responses to a S. pneumoniae whole cell vaccine,
and Challenge showing responses to both the core proteome and diverse core
loci [5].
– To investigate homology between whole cell vaccine strains and
proteins present on the panproteome microarray, follow the
methods described in Subheading 3.2, steps 1–6 to perform
pairwise BLASTp using amino acid sequences of each protein on
the array as queries against the vaccine strain genome as the
subject. This provides a distribution of homology (% sequence
identity) or, conversely, divergence (100 – (% sequence iden-
tity)) from the vaccine strain (Fig. 5).
– To investigate panproteomic immunogenicity profiles, differen-
tial reactivity analysis can be performed using standard para-
metric and nonparametric methods such as t-tests, Wilcoxon’s
rank-sum tests, empirical Bayes models or multivariable linear
models, as appropriate for the study design. Example code for
performing empirical Bayes moderated t-tests on array data and
example output in a study of immunogenicity of a S. pneumoniae
whole cell vaccine (Fig. 6):

> library(limma)
> # sample.classes is a vector of classes for group comparison
> # int.data.df is a data frame containing the matrix of data
with rows as proteins and columns as samples
> trt <- as.factor(sample.classes)
> design <- model.matrix(~trt) # Create a design matrix to fit
limma model
> model <- lmFit(int.data.df, design = design) # Fit ANOVA
model and build contrasts
> result <- eBayes(model)
> ebayes <- topTable(result, coef = 2, number = nrow(int.data.
df),
90 Joseph J. Campo and Amit Oberai

Fig. 5 Distribution of percent identity between S. pneumoniae panproteome microarray proteins and the
S. pneumoniae whole cell vaccine strain RM200 showing high level sequence identity for most proteins
constituting the core proteome and greater diversity for a subset of proteins on the microarray (reproduced
from [5] under a CC-BY license)

Fig. 6 Volcano plot showing the statistical and biological significance of changes in IgG binding following
S. pneumoniae whole cell vaccination, showing increase in antibodies targeting diverse proteins in the
panproteome. Points corresponding to variants of diverse core loci PspA, PspC, ZmpA, or ZmpB are colored
in purple, blue, green and orange, respectively (reproduced from [5] under a CC-BY license)

genelist = row.names(int.data.df),
sort.by = "none", p.value = 1.1)

– Proteome array data is commonly visualized with heat maps.

Multiple R packages facilitate the rendering of heat maps using
the data organized in matrix format with signal intensity values
arranged in columns by sample and rows by protein
(or transposed to invert rows and columns). Order of rows and
columns can be ordered manually or by clustering algorithms.
Heat maps can be annotated with additional data to highlight
associations. For example, to show the association of amino acid
sequence similarity of the diverse core loci of a panproteome
microarray with a strain used for vaccination or challenge,
 Panproteome Analysis of Human Antibody Responses 91

Fig. 7 Changes in IgG binding to S. pneumoniae diverse core loci following vaccination with a whole cell
vaccine. The variants with higher similarity to the RM200 vaccine strain tended to have greater immunoge-
nicity for ZmpA, showing a strain-transcendent immune response (reproduced from [5] under a CC-BY license)

proteins can be ordered by pairwise BLASTp results, sorting by

percent sequence identity (Fig. 7).
– All p-values should be reported with appropriate adjustment,
such as the false discovery rate described by Benjamini and
Hochberg (see Note 14) [11].

4 Notes

1. There is no fast rule for the genomes selected for panproteome

microarray development and will depend on the unique char-
acteristics of the taxa and study questions for the organism of
interest. For example, the genomic structure of T. pallidum
subsp. pallidum, the causal bacteria of syphilis, can be divided
into two genetic subclusters that can be represented by the
Nichols and SS14 strains (Table 2) [12, 13]. Thus, our pan-
proteome approach for syphilis selected only two genomes. In
contrast, a total of 91 clinical isolates of S. pneumoniae were
92 Joseph J. Campo and Amit Oberai

selected for development of a panproteome microarray out of a

total of 616 sequenced isolates [4].
2. Definitions/descriptions of the software tools from Subhead-
ing 2.2 (Data Analysis Software):
– BLAST+ is a suite of programs for local sequence alignment,
which can be installed from the NCBI download site at
https://ftp.ncbi.nlm.nih.gov/blast/executables/blast+/
LATEST/. The program used in this chapter is protein
BLAST (BLASTp).
– PSORTb is a web based software for prediction of subcellu-
lar localization of bacterial proteins, accessible at https://
www.psort.org/psortb/.
– TMHMM is a web based and stand-alone software for
prediction of transmembrane helices in proteins. Multipass
membrane proteins are given a higher score than membrane
proteins with a single predicted transmembrane helix. It is
accessible at http://www.cbs.dtu.dk/services/TMHMM/
.
– NCBI web CD-Search tool is a web based software that
provides domain and function annotations based on several
domain defining software, such as Pfam, COG, and TIGR-
FAM (includes Gene Ontology “GO” terminology). The
CD-Search conserved domain database (CDD) search
returns the domain defined region of the query protein,
along with their domain annotations, as well as an associated
bit score and e-value. It is accessible at https://www.ncbi.
nlm.nih.gov/Structure/bwrpsb/bwrpsb.cgi.
– R is a free software environment for statistical computing
and graphics, which can be downloaded from the official
website of the R Project for Statistical Computing (https://
www.r-project.org/).
– R studio is an integrated development environment (IDE)
for R (https://www.rstudio.com/). This is an optional soft-
ware that provides a user-friendly interface that for some
users of other statistical software will provide greater famil-
iarity than the standard R interface (also see Note 2).
– R packages are provided by the R environment in addition
to base R functions. Several R packages provide convenient
functions for groupwise comparison of antibody measure-
ments on proteome microarrays.
– Limma is a library built for the analysis of gene expression
data, limma contains useful functions for performing empir-
ical Bayesian methods for differential analysis, providing
variance shrinkage and robustness for small sample sizes on
large arrays (also see Note 3).
 Panproteome Analysis of Human Antibody Responses 93

– Lme4 is a package for fitting linear and generalized linear

mixed-effects models, useful for analysis of longitudinal
datasets (e.g., vaccine/challenge clinical trials) where sub-
jects have repeated measurements (also see Note 4).
– ROCR is a package for analyzing and visualizing perfor-
mance of classifiers, particularly useful in calculating the
area under the receiver operating characteristic (“ROC”)
curve.
– Rtsne is a package for performing t-distributed stochastic
neighbor embedding (t-SNE) [10], which has been
employed in proteome microarray studies for dimension
reduction and visualization (also see Note 5).
– R packages for data visualization are in addition to base R
plotting functions. Several R packages provide comprehen-
sive, modular graphics functions that can be used for visual-
ization of panproteome microarray data.
– ggplot2 is a comprehensive system of creating graphics with
a learning curve, but well-documented help and examples of
graphical parameterization (https://ggplot2.tidyverse.org/
).
– gplots contains the function “heatmap.2” which is useful for
producing heat maps of array data.
– ComplexHeatmap is a package for more complex datasets.
This package includes capabilities for overlaying protein
annotation and sample phenotype metadata. Taking full
advantage of heat map customization requires a steep
learning curve, but detailed instructions and examples are
available on the website (https://jokergoo.github.io/Com
plexHeatmap-reference/book/).
3. Other programs that facilitate script-writing include Sublime
Text (https://www.sublimetext.com/) and Notepad++
(https://notepad-plus-plus.org/).
4. For most purposes, a Student’s t-test with p-value adjustment
for the false discovery rate will perform as well as empirical
Bayes moderated t-tests, given that the goal in most protein
array studies for vaccine antigen discovery is to identify the top
N antigens.
5. There is ambiguity about the calculation of p-values for linear
mixed effects regression, and the lme4 package in R does not
compute p-values purposefully (discussion by the package
author shown here: https://stat.ethz.ch/pipermail/r-help/
2006-May/094765.html). To approximate the likelihood of
an effect estimate being nonzero, null models and full models
are fit, and likelihood ratios are tested with an ANOVA test.
94 Joseph J. Campo and Amit Oberai

6. Other dimension reduction methods such as principal compo-

nent analysis (PCA) can be substituted and is a generally more
familiar method. t-SNE is a nonlinear embedding technique
that has shown good clustering of longitudinal data from mul-
tiple time points for individuals’ proteome-wide antibody pro-
files [5, 14].
7. Selection of organisms for the panproteome is not trivial in that
much depends on the research question, desired breadth of
coverage and practical matters such as space for proteins on
the array and costs. For example, one of our research questions
for characterizing the antibody response to a S. pneumoniae
whole cell vaccine was whether vaccination induced strain-
transcendent antibodies [5]. Thus, we used genomic informa-
tion from 616 S. pneumoniae pediatric clinical isolates for
development of a panproteome microarray targeted to
within-species diversity. On the other hand, our work with
B. melitensis asked whether cross-reactive and specific antibody
profiles exist between multiple species of Brucella, thus we used
reference strains of 5 Brucella species for homology analysis.
8. The preference for using Genbank or Refseq is subjective to the
task. In our experience, the latter provides greater annotation
and removes redundancy, while the former is more often refer-
enced in literature and more readily cross-referenced with gene
IDs from other studies.
9. There is no fast rule for a percent identity threshold. A lower
threshold will permit more divergent proteins to be classified as
“core.” The decision is subjective to the characteristics of the
bacterial population. In the examples presented in Tables 1–3,
three different threshold were used (90%, 95%, and 70%,
respectively).
10. There is no fast rule for the threshold of prevalence of a protein
in the bacterial population (present in % of isolates) for selec-
tion. In the case of a U-shaped distribution, as seen in Fig. 2, a
cutoff can be empirically defined, assuming that the lower
prevalence genes include rare genes in the bacterial population
and also sequencing misreads. However, distributions as seen
in Clostridium spp. (Fig. 4) may require cutoffs based on
practical considerations such as availability of spots on the array.
11. The score system used in our down-selection of antigens uses
an arbitrary point system, whereby points are tallied based on
results of predictive models such as PSORTb, TMHMM, and
SignalP, as well as keyword searches in the protein annotations.
The overarching goal in our studies is to enrich selection for
surface-exposed proteins.
12. PSORTb scoring hierarchy (with SignalP prediction) used in
order of highest to lowest score is (1) Extracellular + predicted
 Panproteome Analysis of Human Antibody Responses 95

signal peptide, (2) Extracellular, (3) Outer membrane/peri-

plasmic + predicted signal peptide (or cell wall in case of Gram
positive), (4) Outer membrane/periplasmic (or cell wall in case
of Gram positive), (5) Inner membrane or cytoplasmic/mem-
brane + predicted signal peptide, (6) Inner membrane or cyto-
plasmic/membrane, (7) Cytoplasmic + predicted signal
peptide, and (8) Cytoplasmic.
13. Keywords are selected with respect to the organism; for exam-
ple, “flagella” is a keyword for bacteria with flagellar proteins.
Example keywords include flagellin, flagella, flagellar, Fli, fim-
brial, flag, membrane, secreted, adhesin, cell wall, surface,
hemagglutinin, transport, cilia, pilus, pili, Pil, porin, holin,
Ton, Omp, receptor, transfer protein, toxin, antigen, trigger
factor, protease, lysin, and adhesin.
14. Consideration should be given to the independence assump-
tion of measurements. If an outsized proportion of the pan-
proteome microarray content is variants of the same protein, it
follows that the measurements are not independent and that
the degree of false discovery may be overestimated by the
Benjamini–Hochberg method. P-value adjustment methods
should follow best practices for the type of dataset analyzed.

References
1. Driguez P, Doolan DL, Molina DM et al 6. Yu NY, Wagner JR, Laird MR et al (2010)
(2015) Protein microarrays for parasite antigen PSORTb 3.0: improved protein subcellular
discovery. Methods Mol Biol 1201:221–233. localization prediction with refined localization
https://doi.org/10.1007/978-1-4939-1438- subcategories and predictive capabilities for all
8_13 prokaryotes. Bioinformatics 26:1608–1615.
2. Köser CU, Ellington MJ, Cartwright EJP et al https://doi.org/10.1093/bioinformatics/
(2012) Routine use of microbial whole btq249
genome sequencing in diagnostic and public 7. Krogh A, Larsson B, von Heijne G, Sonnham-
health microbiology. PLoS Pathog 8: mer EL (2001) Predicting transmembrane pro-
e1002824. https://doi.org/10.1371/journal. tein topology with a hidden Markov model:
ppat.1002824 application to complete genomes. J Mol Biol
3. Croucher NJ, Finkelstein JA, Pelton SI et al 305:567–580. https://doi.org/10.1006/
(2013) Population genomics of post-vaccine jmbi.2000.4315
changes in pneumococcal epidemiology. Nat 8. Marchler-Bauer A, Derbyshire MK, Gonzales
Genet 45:656–663. https://doi.org/10. NR et al (2015) CDD: NCBI’s conserved
1038/ng.2625 domain database. Nucleic Acids Res 43:
4. Croucher NJ, Campo JJ, Le TQ et al (2017) D222–D226. https://doi.org/10.1093/nar/
Diverse evolutionary patterns of pneumococcal gku1221
antigens identified by pangenome-wide immu- 9. Smyth GK (2004) Linear models and empirical
nological screening. Proc Natl Acad Sci 114: bayes methods for assessing differential expres-
E357–E366. https://doi.org/10.1073/pnas. sion in microarray experiments. Stat Appl
1613937114 Genet Mol Biol 3:1–69
5. Campo JJ, Le TQ, Pablo JV et al (2018) 10. van der Maaten L, Hinton G (2008) Visualiz-
Panproteome-wide analysis of antibody ing high-dimensional data using t-SNE. J
responses to whole cell pneumococcal vaccina- Mach Learn Res 9:2579–2605
tion. eLife 7:e37015. https://doi.org/10. 11. Benjamini Y, Hochberg Y (1995) Controlling
7554/eLife.37015 the false discovery rate: a practical and powerful
96 Joseph J. Campo and Amit Oberai

approach to multiple testing. J Royal Stat Soc 13. Arora N, Schuenemann VJ, J€ager G et al
Ser B 57:289–300 (2016) Origin of modern syphilis and emer-
12. Pětrošová H, Pospı́šilová P, Strouhal M et al gence of a pandemic Treponema pallidum clus-
(2013) Resequencing of Treponema pallidum ter. Nat Microbiol 2:16245. https://doi.org/
ssp. pallidum strains Nichols and SS14: correc- 10.1038/nmicrobiol.2016.245
tion of sequencing errors resulted in increased 14. Camponovo F, Campo JJ, Le TQ et al (2020)
separation of syphilis Treponeme subclusters. Proteome-wide analysis of a malaria vaccine
PLoS One 8:e74319. https://doi.org/10. study reveals personalized humoral immune
1371/journal.pone.0074319 profiles in Tanzanian adults. Elife 9:e53080.
https://doi.org/10.7554/eLife.53080
 Chapter 7

Low-Energy Electron Irradiation (LEEI) for the Generation of

Inactivated Bacterial Vaccines
Jasmin Fertey, Bastian Standfest, Jana Beckmann, Martin Thoma,
Thomas Grunwald, and Sebastian Ulbert

Abstract
Vaccines consisting of whole inactivated bacteria (bacterins) are generated by incubation of the pathogen
with chemicals. This is a time-consuming procedure which may lead to less immunogenic material, as
critical antigenic structures can be altered by chemical modification. A promising alternative approach is
low-energy electron irradiation (LEEI). Like other types of ionizing radiation, it mainly acts by destroying
nucleic acids but causes less damage to structural components like proteins. As the electrons have a limited
penetration depth, LEEI is currently used for sterilization of surfaces. The inactivation of pathogens in
liquids requires irradiation of the culture in a thin film to ensure complete penetration. Here, we describe
two approaches for the irradiation of bacterial suspensions in a research scale. After confirmation of
inactivation, the material can be directly used for vaccination, without any purification steps.

Key words Bacterial inactivation, Low-energy electron irradiation, LEEI, Bacterial vaccine, Bacterin,
Electron beam

1 Introduction

Prophylactic vaccines provide an alternative to the use of antibiotics

by preventing and controlling infectious diseases in humans and
animals. Some bacterial vaccines contain only parts of the pathogen
(e.g., polysaccharide structures or other recombinantly expressed
proteins), or inactivated bacterial toxins. Currently used veterinary
vaccines mainly include attenuated live vaccines and inactivated
vaccines [1]. Live attenuated vaccines provide protection through
a limited infection of a living organism which elicits an immune
response, similar to that of a natural infection. However, they have
the very rare potential to revert to a pathogenic form leading to
disease. Inactivated or killed vaccines are safe and show similar
protection against systemic infections and disease. They consist of
killed whole bacteria, which are usually produced by chemical

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_7,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

97
98 Jasmin Fertey et al.

treatment by incubating the pathogens with toxic substances, for

example binary ethyleneimine or formaldehyde. This is a time-
consuming step and requires days or even weeks of incubation,
depending on the organism. In many cases, the chemicals need to
be inactivated or removed before the material can be administered
as a vaccine. Additionally, the inactivation process often results in
chemical modifications of critical antigenic structures and therefore
leads to lower vaccine efficacy.
As an alternative strategy to chemical inactivation, ionizing
radiation has been used to generate bacterial vaccines, but these
approaches are still in the experimental stage, despite promising
results [2–4]. Compared to chemical inactivation, that mostly acts
by crosslinking nucleic acids and proteins, the major inactivation
mechanism of ionizing irradiation is the damage to nucleic acids,
while other structural components such as proteins remain largely
intact [5–7]. As a rule of thumb, the dose which is required for
inactivation is dependent on the genome size, with smaller gen-
omes being more resistant to the effects. Larger genomes are more
sensitive toward irradiation because the number of induced DNA
double-strand breaks per cell at a given dose is proportional to the
size of the genome. Therefore, more irreparable strand breaks
occur upon irradiation in larger genomes, hence for the inactivation
of bacteria (except for extremely resistant species, e.g., Deinococcus
radiodurans, or dormant bodies like spores) generally lower doses
are required than for the inactivation of viruses [8–10]. The inacti-
vation doses are in the kilogray (kGy) range and currently used
irradiation technologies (i.e., gamma-, X-rays, or high-energy elec-
tron beam) generate considerable amounts of radiation, either
directly or as a side product. Therefore complex concrete and lead
shielding constructions for its absorption are required to protect
personnel and environment [11]. This has so far prevented the
application of ionizing radiation for vaccine manufacturing pro-
cesses. In contrast, low-energy electron irradiation (LEEI) uses a
lower voltage, up to 300 kilo electron volts (keV), instead of high-
energy electron irradiation that operates in the range of 1–10 mega
electron volts (MeV). LEEI generates lower amounts of secondary
radiation (Bremsstrahlung) than high-energy electrons and there-
fore requires only a few centimeters of protecting lead. This enables
the use of LEEI in normal laboratory settings, including potential
integration into GMP processes.
Due to high dose rates, LEEI acts very fast and the applied
doses can be exactly adjusted via controlling the acceleration volt-
age. This leads to more reproducible results than, for example, with
gamma irradiation, where the activity of the radioactive source
decreases because of the constant decay [12]. Electron-beam treat-
ment is already an FDA-approved sterilization procedure for
food and pharmaceutical products. LEEI sources require low main-
tenance and devices are already commercially available (i.e.,
 Bacterial Inactivation by Low-Energy Electron Irradiation 99

https://www.ebeamtechnologies.com/en/products). However,
they are currently not suited for irradiating liquids and the tech-
nique is mainly used for surface sterilization. The major challenge
for LEEI in inactivating liquid solutions is the limited penetration
depth of the electrons. To overcome this limitation, the liquid has
to be irradiated in a thin film, and manufacturing of larger volumes
in a multiliter scale with this technique is challenging. With the
currently available commercial devices specified for surface sterili-
zation, this can be achieved by using Petri dishes, followed by
covering the liquid with a piece of polypropylene foil [13, 14]. How-
ever, amounts that are required for larger study groups or larger
animals are difficult to produce with this small-scale approach.
Further development of irradiation technologies for larger volumes
is therefore required.
We have recently described the development of the first
research-scale prototype for LEEI of liquids in higher throughput
[8], and this development will form the basis for commercially
available devices. We have also shown that LEEI can be used to
generate efficient vaccines against viruses, parasites, and gram-neg-
ative bacteria [13, 15, 16]. In this chapter, we demonstrate both
small-scale (see Subheading 3.2) and large-scale (see Subheading
3.3) methods to generate LEEI-inactivated bacterial vaccines and
to evaluate their antigenic components (see Subheading 3.4) before
administration into animals. As an example, we chose the gram-
negative bacterium Rodentibacter pneumotropicus to describe the
procedures for small-scale and larger-scale irradiation experiments.
R. pneumotropicus belongs to the family Pasteurellaceae and is
related to other human and veterinary pathogens such as Haemo-
philus influenzae, Actinobacillus pleuropneumoniae, and Pasteurella
multocida. It is an opportunistic pathogen commonly found in the
microbiome of the respiratory tract of mice and rats. Primary or
secondary infections can result in severe diseases, and make
R. pneumotropicus one of the most prevalent infectious agents in
laboratory rodents [17].

2 Materials

Prepare and store all reagents at room temperature (unless indi-

cated otherwise).

2.1 Buffers, 1. Brain Heart Infusion broth: 37.0 g per 1 L distilled water.
Reagents and Media Autoclave and store at 4
C (see Note 1).
2. Phosphate buffered saline (PBS), 10 stock: 1.37 M NaCl,
270 mM KCl, 100 mM Na2PO4, 20 mM KH2PO4 pH 7.4 (see
Note 2).
100 Jasmin Fertey et al.

3. ELISA coating buffer: 15 mM Na2CO3, 35 mM NaHCO3,

pH 9.6. Filter through 0.22 μm membrane filter.
4. PBS–Tween: PBS containing 0.05% Tween 20.
5. Antibodies: polyclonal mouse serum against R. pneumotropicus
and HRP-conjugated anti-mouse antibody (store at 4
C or at
20
C for long-term storage) (see Note 3).
6. Blocking solution: 5% skim milk in PBS–Tween (store at 4
C).
7. Detection substrate: 3,30 ,5,50 -tetramethylbenzidine (TMB)
(see Note 4).
8. H2SO4: 1 M.
9. 70% ethanol.

2.2 Equipment 1. 37
C orbital shaker

2. 37
C static incubator
3. Tabletop centrifuge with rotor for 15 mL reaction tubes.
4. Microcentrifuge with rotor for 1.5 mL reaction tubes.
5. Biological safety cabinet (BSC).
6. Spectrophotometer for checking optical density at 600 nm
(OD600).
7. Cuvettes for absorbance measurement.
8. 0.22 μm membrane filters
9. Polyethylene terephthalate/polyethylene (PET/PE) material
with 12 cm width (see Note 5).
10. Small round pieces (6 cm in diameter) of oriented polypropyl-
ene (OPP)-foil (see Notes 5 and 6).
11. Impulse sealer with a length of 80–100 cm.
12. Impulse sealer with a length of 30 cm (optional; see Note 7).
13. ELISA reader.
14. 1.5 mL tubes
15. Pipets.
16. Pipetting aid.
17. 15 mL and 50 mL tubes
18. Sterile scissors.
19. Petri dishes (10 cm diameter; see Note 8).
20. ELISA plates (96-well) (see Note 9).

2.3 Specialized 1. Electron irradiation device (Fig. 1a) with module for auto-
Equipment mated bag transport (Fig. 2), for bag irradiation experiments.
 Bacterial Inactivation by Low-Energy Electron Irradiation 101

Fig. 1 (a) Photograph of the custom-built LEEI device. The picture shows the closed device with the electron
beam source and the leaden irradiation chamber in the middle. Feed lines with flexible hoses for the in- and
outlet of liquids (e.g., for cooling inside the irradiation chamber) are guided through the metal box followed by
baffles to shield the environment from occurring Bremsstrahlung. The movable module holder can be rolled
out of the irradiation chamber and carries the modules (shown in Fig. 3) for liquid handling. The irradiation
chamber was constructed as a research-scale prototype and different experimental modules designed for
liquid handling can be inserted completely [8]. This increases the overall footprint of the electron beam device.
(b) Photograph of the commercially available EB-Lab 200 (https://www.ebeamtechnologies.com/en/eblab).
The picture shows the opened chamber, where the sample holder for Petri plates is placed, and the cover is
closed before irradiation. The sample is transported to the left for irradiation where the electron beam source
is located

2. EB Lab 200 or EB Lab 300 (https://www.ebeamtechnologies.

com/en/eblab) or comparable electron irradiation device
(Fig. 1b), for small-scale experiments.

3 Methods

This section explains the preparation of the irradiated bacterial

sample for use in animal vaccination experiments. Two irradiation
procedures, either in a 230 μL volume (small-scale in Petri dishes,
covered with plastic foil; Fig. 3) or in a 20 mL volume (larger scale
in disposable bags; Figs. 4 and 5) are described, followed by testing
the antigenicity by ELISA.
All experimental steps involving living R. pneumotropicus must
be performed in a laboratory authorized for the work with BSL-2
organisms.

3.1 Preparing the 1. Inoculate bacteria in 50 mL growth medium and incubate over
Bacterial Solution for night at 37
C, 180 rpm (see Note 10).
Irradiation 2. Harvest bacteria by centrifugation (3452  g, 4
C, 10 min)
(see Note 11).
102 Jasmin Fertey et al.

Motor-driven transport roller

Braking roller

Bag filled with inactivated

pathogens in solution

Bag with
pathogens
in solution

Conveyor Belts

Fig. 2 Schematic drawing of the module. The disposable bag containing the
pathogen solution is placed on the left side between the two conveyor belts
(marked by the red line). The roller transports the bag through the irradiation
zone (marked by radiation warning sign). The bag is stretched by the braking
roller to limit the liquid height and to enable the low-energy electrons to
penetrate the bag completely. The bag with the irradiated solution (marked in
green) is transported to the right and can be taken out of the module after the
irradiation is completed

3. Discard supernatant and resuspend bacteria in 50 mL PBS.

Centrifuge again at 3452  g, 4
C, 10 min (see Note 12).
4. Repeat step 3 twice and resuspend in a final volume of 50 mL
PBS (see Note 13).
5. Check the OD600 in a spectrophotometer and adjust with PBS
to a maximum value between 1 and 2 (see Note 14).
6. Store washed bacteria on ice, or at 4
C, until irradiation (see
Note 15).
 Bacterial Inactivation by Low-Energy Electron Irradiation 103

Fig. 3 (a) Sample preparation for LEEI in a small-scale experimental setup. 230 μL of bacteria are pipetted to
the center of a Petri dish and covered with a piece of round, sterilized oriented polypropylene (OPP)-foil
forming the thin liquid film. (b) Photograph of sample holder with prepared Petri dishes, placed in the center of
the sample holder. The picture shows the sample holder already inserted into the irradiation device

3.2 Irradiation in a 1. Take the sample holder out of the irradiation device and put
Small-Scale into a biological safety cabinet (BSC). Discard lids from the
Experimental Setup Petri dishes, as they are not needed anymore. A maximum of six
dishes can be placed on the sample holder and processed at the
same time.
2. For each sample, pipet 230 μL of washed bacteria to the center
of a Petri dish and cover the drop with a piece of sterilized
OPP-foil (Fig. 3a). Be careful to avoid bubbles during pipet-
ting and spreading the liquid, since they might affect the height
of the liquid film (see Note 16).
3. Put the Petri dishes in the center of the sample holder and
insert the sample holder in the device (Fig. 3b) (see Note 17).
4. Close the cover of the irradiation chamber, set the desired
parameters and start the irradiation process (see Note 18).
5. Take the sample holder out of the irradiation device and put it
back into the BSC.
6. Recover the liquid from Petri dishes by carefully tapping the
edge of the dish on the surface of the BSC and collecting the
liquid that runs down to the edge using a pipet. Transfer the
sample to a sterile 1.5 mL tube.
7. For inactivation testing, inoculate 10% of the recovered sample
into fresh growth medium and incubate at 37
C overnight.
Include a nonirradiated sample as positive control for growth
(see Note 19).
8. Save 10–20% of the recovered sample for checking the antige-
nicity (see Subheading 3.4). Ensure to save the same amount of
the nonirradiated sample, since this will be the reference for the
antigenicity testing.
9. Store the residual sample at 80
C until use.
104 Jasmin Fertey et al.

12 cm

15 cm

10 cm

39 cm 65 cm

11 cm

Fig. 4 Dimensions of PET/PE-bags for irradiation of 20 mL bacterial solution in

bags. The area containing the liquid (pale green) is 39 cm  10 cm. Although the
bag could contain more than 20 mL, this volume ensures the optimal liquid
height for complete penetration by LEEI

3.3 Irradiation 1. Prepare irradiation bags by placing two layers of PET/PE film
in Bags carefully over one another. Seal the long (65 cm) sides with the
80–100 cm impulse sealer, taking care that the two layers are
properly aligned on top of each other. Then seal the bottom of
the bag, creating the seam about 10–11 cm from the bottom
edge (Fig. 4), using the 30 cm impulse sealer (see Note 7).
Make sure that all three sides are completely sealed (Figs. 4 and
5) and that the sealing seams are intact to ensure that no
infectious content spills out (see Note 20). Leave the top of
the bag open for filling (Fig. 5).
2. In a biological safety cabinet (BSC), fill 20 mL of washed
bacteria in PBS or other buffer into a prepared irradiation bag
using a 10 mL or 20 mL pipet and a pipetting aid (Fig. 5). Use
one bag per irradiation dose to be tested and include an extra
 Bacterial Inactivation by Low-Energy Electron Irradiation 105

sealed seam
(both sides)

sealed seam
(top)

removal of
excessive air

sealed seam
(bottom)

PET/PE bag 20 ml washed sealed bag with

bacteria in buffer bacteria in buffer

Fig. 5 sample preparation for LEEI in disposable bags. 20 mL of washed bacteria are transferred to PET/PE-
bags (see Fig. 1 for dimensions). Air is removed from the bag and the top is sealed using an impulse sealing
machine

bag/sample for the nonirradiated control (i.e., sample that

receives no irradiation, but is processed). If processing effects
are observed in downstream experiments, include an
“untreated control” without processing as well for optimiza-
tion of the irradiation process.
3. Remove excessive air from the bag before sealing by pulling the
bag top down over the rim of the BSC to squeeze out residual
air, then seal the upper part of the bag to close it (see Note 21).
Disinfect the closed bag by wiping or spraying with disinfectant
before transporting to the irradiation module, in case infec-
tious content spilled out of the bag.
4. Insert the bag in the irradiation module between the conveyor
belts and fix the bag between the two rollers (Fig. 2).
5. Close the irradiation chamber, set the desired parameters and
start the irradiation process (see Note 22).
6. After irradiation, remove the processed bag from the device by
loosening the rollers that hold the bag in place and pull the bag
out of the module.
106 Jasmin Fertey et al.

7. Store irradiated bags on ice or cool packs until ready to recover

the sample (see Note 23).
8. Recover the bacterial solution in a BSC by squeezing the liquid
to the lower part of the bag and cutting the upper part of the
bag with sterile scissors, then pipet the content into a fresh,
sterile vessel.
9. For inactivation testing, inoculate 10% of the recovered sample
into fresh growth medium and incubate over night at 37
C (see
Note 24). Always include a nonirradiated sample as growth
control.
10. Save 10% of the recovered sample into a 1.5 mL centrifuge tube
for antigenicity testing. Ensure to save the same amount of the
nonirradiated sample, since this will be the reference for the
antigenicity testing. Centrifuge the sample in a microcentrifuge
for 30 s at 24,100  g.
11. Carefully record (and mark directly on the tube) the volume of
the 10% sample for antigenicity testing and discard the super-
natant. Use this sample directly for antigenicity testing (see
Subheading 3.4), or store pellet at 80
C until use.
12. Transfer the remaining recovered sample into a 15 mL tube.
The bacteria can be stored at 80
C, either as a pellet or in
buffer. This depends on the intended administration or formu-
lation (i.e., which buffer, concentration, etc.). If you wish to
pellet the bacteria, use a tabletop centrifuge (3452  g, 4
C,
10 min), record the initial volume as described in step 10, and
store pellet at 80
C until use.

3.4 Determining the 1. In a BSC, resuspend pelleted bacteria in the initially noted
Antigenicity of the volume of buffer (PBS or coating buffer).
Irradiated Sample by 2. Always include the nonirradiated sample as positive control for
ELISA the ELISA experiment. Make sure that the nonirradiated sam-
ple is diluted in the same way as the irradiated sample. Include
the buffer (e.g., PBS) used for resuspending the bacteria before
irradiation in the experiment as background negative control
(see Note 25).
3. Dilute samples in coating buffer and transfer 100 μL of each
dilution per well in duplicates or triplicates to the ELISA plate.
Usually, 1–5 μL of sample (prepared from washed bacteria with
an OD600 of 1–2) in 100 μL coating buffer per well are suffi-
cient for obtaining optimal signals (see Note 26). Incubate the
plate on a shaker over night at 4
C.
4. On the next day, wash the plate 3 times with 200 μL per well
PBS-T (see Note 27).
 Bacterial Inactivation by Low-Energy Electron Irradiation 107

5. Block the plate by adding 100 μL blocking solution (5% skim

milk in PBS-T) per well and incubate for 1–2 h at room
temperature.
6. Remove blocking solution (see Note 28).
7. Dilute primary antibody in blocking solution and add 100 μL
per well to the plate. Incubate for 2 h at room temperature (see
Note 29).
8. Wash plate 3 times with 200 μL per well PBS-T.
9. Dilute secondary antibody in blocking solution and add
100 μL per well to the plate. Incubate for 1 h at room
temperature.
10. Prepare fresh TMB solution (see Note 4) and add 100 μL per
well. Incubate in the dark for 15 to max. 30 min (see Note 30).
11. Stop the reaction by adding 50 μL H2SO4 per well. You should
observe a color change from blue to yellow.
12. Read absorbance at 450 nm in an ELISA reader.
13. Set the signal of the nonirradiated control as 100% and calcu-
late the percentage of signal of the irradiated samples (see
Notes 31 and 32).

4 Notes

1. BHI is the recommended growth medium for

R. pneumotropicus, for other bacteria appropriate culture
media should be used for growth and inactivation testing.
2. The buffer used for irradiation can be replaced by other buffer-
ing solutions, for example, Tris-based or Hepes-based buffers,
if nonirradiated or untreated bacteria show loss of viability after
storage in PBS. If concentrated stock solutions of buffers are
prepared (e.g., 10x PBS), ensure to dilute them to 1 working
concentration before use.
3. For our ELISA-experiments we used serum from a lab mouse
experimentally infected with the homologous bacterial
strain [17].
4. We use a commercially available TMB Substrate Set (e.g., Cat.
No. 421101 from BioLegend) which contains TMB Substrate
A and TMB Substrate B. Equal volumes of each substrate are
mixed immediately before use. For example, for one 96-well
plate, mix 5.5 mL TMB Substrate A with 5.5 mL of TMB
Substrate B in a clean container. After mixing the reagents
together, TMB substrate working solution should be colorless
or very faint blue. The mixed substrate is not stable for a long
108 Jasmin Fertey et al.

time, only the amount needed for each assay run should be
prepared.
5. We recommend using polyethylene terephthalate/polyethyl-
ene (PET/PE) material for the bags and oriented polypropyl-
ene (OPP) for the round plastic-foils for irradiation [18]. In
general, PET/PE and round pieces of OPP foil are available at
specialized packaging companies that provide packaging solu-
tions for medical or food products (e.g., sicht-pack HAGNER
GmbH or tbs-pack GmbH, both in Germany). We recommend
ordering PET/PE film rolls with the intended bag width (i.e.,
12 cm), and enough length to generate the desired number of
bags, each 65 cm long (Fig. 4).
6. For the small-scale irradiation experiments, it is crucial to avoid
shadowing effects during irradiation, otherwise there is the
chance that the sample is not evenly irradiated. Therefore, the
OPP foil must have a smaller diameter than the Petri plates. It is
recommended to sterilize the round OPP-foils before use by
putting them into 70% ethanol for 15 min, then allow to dry on
paper towels in a BSC. This treatment also removes any nonor-
ganic contaminations.
7. It is possible to use the 80–100 cm impulse sealer for all sides.
However, we prefer to have a smaller (30 cm) impulse sealer in
the BSC to directly seal the bags after filling. This makes the
disinfection (by spraying or wiping off the bag) much easier
because no infectious content spills out.
8. We tried different Petri dishes and identified Corning® Pri-
maria™ to be the best choice. The very hydrophilic surface
treatment ensures that the liquid is evenly distributed under the
OPP-foil, providing a homogenous liquid layer. In untreated
Petri dishes the liquid stays more or less as a drop on the
surface, making it difficult to achieve a liquid layer with an
even height. We therefore recommend using specially treated
Petri dishes (comparable to Corning® Primaria™).
9. In most cases, Nunc PolySorp® surface coated ELISA plates
worked well for coating the irradiated samples; however, there
have been some exceptions where other surfaces performed
better. If weak ELISA-signals are observed due to insufficient
coating, checking other plates (such as Nunc MaxiSorp® or
MultiSorp®) might improve the results.
10. We have also performed experiments with other bacteria, for
example, Escherichia coli and Bacillus cereus. We usually store
the bacteria at 80
C as cryostocks that can be used directly
for inoculation. We observe optimal growth if the bacteria are
precultured in a lower volume, such as 5 mL, at 37
C,
180 rpm on the day before the inoculation for the experiment
and use 0.5 mL of this culture for inoculation of the final
culture.
 Bacterial Inactivation by Low-Energy Electron Irradiation 109

11. Depending on the bacterial strain used, the centrifugation time

might vary, depending on the consistency of the resulting
pellet. Resuspending the bacteria should not be too difficult.
12. If decrease in bacterial viability is observed in nonirradiated
control or after storage of the bacteria in PBS, the buffer used
for washing and during irradiation should be exchanged. Fur-
thermore, glycerol (up to a final concentration of 10%) can be
added to the chosen buffer if handling stress is observed during
antigenicity testing (marked by a strong decrease in ELISA
signal intensity in the untreated vs. the nonirradiated sample;
see Notes 25 and 26). The untreated sample is a sample which
is just washed with buffer and stored, the nonirradiated sample
is a sample that underwent all steps, except the irradiation
(pipetting, filling, etc.). This optimization step should be
done before the irradiation experiment. So far, the best results
we have obtained were using PBS only. Irradiation in culture
media is not recommended, as we have observed that the
inactivation is less reproducible. It is possible that components
in the media absorb the irradiation and bacterial aggregates
may form as well.
13. It is crucial to have a homogenous resuspension to ensure
complete inactivation. If clumps are visible, wash the bacterial
pellet until no clumps are visible anymore. Note that many
clinical isolates (especially Pasteurellaceae species) form strong
biofilms. These cultures might require more washing and
resuspension steps. Also note that some species, such as
A. pleuropneumoniae, may lose this phenotype following mul-
tiple passage in broth culture.
14. Irradiation is possible with higher concentrations of bacteria,
but it is difficult to ensure homogeneity in very dense solu-
tions. We tested solutions with an OD600 of up to 3 with good
inactivation results after irradiation. However, the readable
range for some spectrophotometers may have a high uncer-
tainty above 1.5. In this case it is better to dilute the suspended
bacteria 1:2 or 1:5 prior to measuring the OD600, and then
back calculate the value for the undiluted stock. Adjusting the
solution to a defined OD600 helps when comparing the input
material between different irradiation experiments. Other
assays for quantification like colony forming units per mL
(CFU/mL) can be used as well, but require an additional
overnight incubation step.
15. It is recommended to store the bacteria for no longer than 2 or
3 h at this stage, longer storage might lead to reduced growth
in the positive controls afterward, which increases the risk of
discrepant results when checking for CFU/mL and antigenic-
ity in the ELISA.
110 Jasmin Fertey et al.

16. Ensure that the liquid sample is in the center of the dish.
Otherwise, shadowing effects from the side of the Petri dish
might lead to incomplete inactivation.
17. Regardless of the number of Petri dishes that are used for each
irradiation run, always ensure that they are in the center of the
sample holder (Fig. 4b). If they are placed at the side, shadow-
ing effects may occur, and the sample is not completely
irradiated.
18. The small-scale setup is optimal for determining the required
inactivation dose and for measuring the antigenicity of the
inactivated sample. It can be used as a small-scale test before
the actual experiment. In this case, several doses should be
tested, and CFU/mL should be determined afterward in addi-
tion to liquid culture for checking the inactivation kinetics. In
this way, a killing curve and the minimal dose required for
inactivation can efficiently be established.
19. The nonirradiated control consists of an extra Petri dish with
sample that is treated in the same way (covering with foil) like
the samples, except without the irradiation process. The sam-
ples can stay in the BSC, while the other samples are irradiated.
20. Make sure that the sealer is not too hot and the seams are
tightly sealed.
21. Removing air is crucial for processing of the bag and efficient
inactivation. There should be as little air as possible in the
closed bag. During irradiation, degassing of the buffer solution
might be observed, in this case air bubbles will be visible in the
irradiated bag. It is crucial that the thin liquid film that is
generated by stretching the bag between two rollers (Fig. 2)
is less than 150 μm thick to ensure complete penetration of the
electrons to the bottom part of the bag. It is recommended
that extra sealed seams (i.e., more than one) are added on the
top of the bag to ensure that infectious content is not released
when the bag is squeezed.
22. Depending on the velocity with which the bag is transported,
the irradiation process takes approx. 2–5 min per bag. Since the
applied dose is a product of applied energy and velocity of the
sample, these parameters have to be determined empirically
and controlled using commercially available, adjusted dosime-
ter films (e.g., Risø B3, Risø High Dose Reference Laboratory,
Denmark). For commercially available irradiation devices, as
used for the small-scale experiments, the dose application is
precalibrated and usually no dosimetry is needed.
23. Cooling the irradiated solution might help to slow down resid-
ual hydrolysis reactions in the liquid after irradiation and
improve antigenicity due to less formation of free radicals.
 Bacterial Inactivation by Low-Energy Electron Irradiation 111

24. Bacterial growth should be drastically reduced after irradiation;

however, there might be residual viability in the sample that is
not visible after overnight incubation. Therefore, it is recom-
mended to incubate the culture for longer periods (up to
3 days, or—depending on the doubling time of the organ-
ism—even longer).
25. We recommend testing each sample (including the nonirradi-
ated sample, the resuspension buffer and the coating buffer) in
triplicate. To check for handling effects during the irradiation
procedure, it might be useful to include an untreated sample as
well (see Note 26). It is not necessary to distribute replicates of
samples and blanks in different parts of the plate. When testing
only a few samples, it is possible to use only parts of the plate
for the measurement. This is recommended to save reagents
(such as primary or secondary antibodies) which may be expen-
sive or in limited supply. The coating of the replicates (row by
row or in columns) can be determined by the user. If only a part
of the plate is coated and an ELISA-washer is used, it makes
sense to coat the plate in a way that the washer has to wash only
this part and not the whole plate. Also, if washing by hand is
performed, it is useful to use a multichannel pipet (8-channel
or 12-channel) and coat the plate in a way, that pipetting steps
are reduced.
26. The amount of material to be coated per well has to be pre-
determined in a separate ELISA experiment and is dependent
on the antibody that is used. It is recommended to initially use
untreated, fresh bacteria in different amounts to find the linear
range for the signal intensity. Signals should be in an OD450
range between 0.8 and 1.5. Signals above OD450 ¼ 1.8 are
saturated, signals below OD450 ¼ 0.5 are often too close to the
background, especially when working with animal sera where
the antibodies have not been purified.
27. Washing can be performed by hand or by using an ELISA
washer. It is recommended to tap plates on a paper towel to
remove residual washing solution from the wells before adding
antibodies to prevent dilution effects.
28. After blocking, plates can be washed 3 times with PBS-T, dried
and sealed in a bag for storage. Blocked plates can be stored for
several weeks in a cool and dry place until use.
29. It is recommended to use a polyclonal serum or specific anti-
bodies for the detection of surface proteins to check the integ-
rity of surface antigens. For testing disruption of the bacteria,
antibodies raised against internally located antigens can be used
to estimate the ratio of disrupted vs. intact bacteria. Working
dilutions must be optimized for each antibody before the
actual experiment.
112 Jasmin Fertey et al.

30. Development of a blue color should be visible after 30 min. If

the color development occurs directly after adding the sub-
strate, check plate every 5 min and stop the reaction after
15 min.
31. In our hands, values of at least 70% were obtained for the
irradiated, compared to nonirradiated, samples. This material
resulted in a robust immune response after vaccination of
animals.
32. The irradiated material can be directly used for vaccination or
formulation with adjuvant. We found that, for gram-negative
bacteria, it is not necessary to use adjuvants since the LPS
structure is better conserved than in chemically treated bacteria
and possibly acts as immune stimulant [15].

References
1. Hoelzer K, Bielke L, Blake DP et al (2018) 8. Fertey J, Thoma M, Beckmann J et al (2020)
Vaccines as alternatives to antibiotics for food Automated application of low energy electron
producing animals. Part 1: challenges and irradiation enables inactivation of pathogen-
needs. Vet Res 49:1–10. https://doi.org/10. and cell-containing liquids in biomedical
1186/s13567-018-0560-8 research and production facilities. Sci Rep
2. Babb R, Chen A, Hirst TR et al (2016) Intra- 10:12786. https://doi.org/10.1038/
nasal vaccination with γ-irradiated Streptococcus s41598-020-69347-7
pneumoniae whole-cell vaccine provides 9. Espinosa AC, Jesudhasan P, Arredondo R et al
serotype-independent protection mediated by (2012) Quantifying the reduction in potential
B-cells and innate IL-17 responses. Clin Sci health risks by determining the sensitivity of
130:697–710. https://doi.org/10.1042/ poliovirus type 1 chat strain and rotavirus
CS20150699 SA-11 to electron beam irradiation of iceberg
3. Bordin AI, Pillai SD, Brake C et al (2014) lettuce and spinach. Appl Environ Microbiol
Immunogenicity of an electron beam inacti- 78:988–993. https://doi.org/10.1128/
vated Rhodococcus equi vaccine in neonatal AEM.06927-11
foals. PLoS One 9:e105367. https://doi.org/ 10. Skowron K, Paluszak Z, Olszewska H et al
10.1371/journal.pone.0105367 (2014) Effectiveness of high energy electron
4. Jwa MY, Jeong S, Ko EB et al (2018) Gamma- beam against spore forming bacteria and
irradiation of Streptococcus pneumoniae for the viruses in slurry. Radiat Phys Chem
use as an immunogenic whole cell vaccine. J 101:36–40. https://doi.org/10.1016/j.
Microbiol 56:579–585. https://doi.org/10. radphyschem.2014.04.001
1007/s12275-018-8347-1 11. IAEA Safety standards (2010) Radiation safety
5. Alizadeh E, Orlando TM, Sanche L (2015) of gamma, electron and x-ray irradiation facil-
Biomolecular damage induced by ionizing ities. IAEA Safety standards series. https://
radiation: the direct and indirect effects of www-pub.iaea.org/MTCD/publications/
low-energy electrons on DNA. Annu Rev PDF/Pub1454_web.pdf
Phys Chem 66:379–398. https://doi.org/10. 12. Silindir M, Özer AY (2009) Sterilization meth-
1146/annurev-physchem-040513-103605 ods and the comparison of E-beam sterilization
6. Brown F (1995) Formaldehyde as an inacti- with gamma radiation sterilization. FABAD J
vant. Vaccine 13:231. https://doi.org/10. Pharm Sci 34:43–53
1016/0264-410x(95)93142-v 13. Fertey J, Bayer L, Grunwald T et al (2016)
7. Delrue I, Verzele D, Madder A et al (2012) Pathogens inactivated by low-energy-electron
Inactivated virus vaccines from chemistry to irradiation maintain antigenic properties and
prophylaxis: merits, risks and challenges. induce protective immune responses. Viruses
Expert Rev Vaccines 11:695–719. https:// 8:319. https://doi.org/10.3390/v8110319
doi.org/10.1586/erv.12.38 14. Bayer L, Fertey J, Ulbert S et al (2018) Immu-
nization with an adjuvanted low-energy
 Bacterial Inactivation by Low-Energy Electron Irradiation 113

electron irradiation inactivated respiratory syn- Vet Parasitol 266:18–26. https://doi.org/10.

cytial virus vaccine shows immunoprotective 1016/j.vetpar.2019.01.001
activity in mice. Vaccine 36:1561–1569. 17. Fornefett J, Krause J, Klose K et al (2018)
https://doi.org/10.1016/j.vaccine.2018.02. Comparative analysis of humoral immune
014 responses and pathologies of BALB/c and
15. Fertey J, Bayer L, K€ahl S et al (2020) C57BL/6 wildtype mice experimentally
Low-energy electron irradiation efficiently infected with a highly virulent Rodentibacter
inactivates the gram-negative pathogen Roden- pneumotropicus (Pasteurella pneumotropica)
tibacter pneumotropicus-A new method for the strain. BMC Microbiol 18:45. https://doi.
generation of bacterial vaccines with increased org/10.1186/s12866-018-1186-8
efficacy. Vaccines 8:113. https://doi.org/10. 18. Gotzmann G, Portillo J, Wronski S et al (2018)
3390/vaccines8010113 Low-energy electron-beam treatment as alter-
16. Thabet A, Schm€aschke R, Fertey J et al (2019) native for on-site sterilization of highly functio-
Eimeria tenella oocysts attenuated by low nalized medical products – a feasibility study.
energy electron irradiation (LEEI) induce pro- Radiat Phys Chem 150:9–19. https://doi.org/
tection against challenge infection in chickens. 10.1016/j.radphyschem.2018.04.008
 Chapter 8

Design and Production of Hybrid Antigens for Targeting

Integral Outer Membrane Proteins in Gram-Negative
Bacteria
Somshukla Chaudhuri, Nikolas F. Ewasechko, Luisa Samaniego-Barron,
Jamie E. Fegan, and Anthony B. Schryvers

Abstract
Metal ion transporters in the outer membrane of gram-negative bacteria that are responsible for acquiring
iron and zinc are attractive vaccine targets due to their essential function. The core function is mediated by
an integral outer membrane TonB-dependent transporter (TBDT) that mediates the transport of the metal
ion across the outer membrane. Some TBDTs also have a surface lipoprotein (SLP) that assists in the
efficient capture of the metal ion-containing host protein from which the metal ion is extracted. The
challenges in producing the integral outer membrane protein for a commercial subunit vaccine prompted us
to develop a hybrid antigen strategy in which surface loops of the TBDT are displayed on the lipoprotein,
which can readily be produced as a soluble protein. The focus of this chapter will be on the methods for
production of hybrid antigens and evaluating the immune response they elicit.

Key words Hybrid antigens, TonB-dependent transporter, Surface lipoprotein, Outer membrane
proteins, ELISA, Transferrin binding proteins

1 Introduction

Our hybrid antigen approach was originally developed to target the

surface epitopes of the integral outer membrane protein (OMP),
transferrin binding protein A (TbpA) [1], and has since been used
to target surface regions of other TonB-dependent transporters
(TBDTs) [2]. TBDTs are a family of integral membrane proteins
in the outer membrane of gram-negative bacteria and are primarily
involved in the acquisition of metal ions or metal ion complexes
[3]. TBDTs consist of a C-terminal 22-strand beta-barrel and a
N-terminal plug region that interacts with TonB from an inner
membrane complex that provides energy derived from ATP hydro-
lysis to drive the transport process [4]. The principles underlying
the hybrid antigen approach could also be applied more generally

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_8,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

115
116 Somshukla Chaudhuri et al.

Fig. 1 TonB-dependent transporters (TBDTs) such as transferrin binding protein A (TbpA) and surface
lipoproteins (SLP) such as TbpB both have loops (red) that are anchored by antiparallel beta-strands
(green). Loops on the soluble SLP can be replaced with loops from TBDTs to generate novel hybrid antigens

to other integral OMPs in gram-negative bacteria. Integral OMPs

consist of beta-barrels that have connecting extracellular surface
“loop” regions that are potential targets for the immune response.
These integral OMPs are typically not suitable for commercial
vaccine production due to poor solubility. Thus, displaying the
extracellular surface loops on soluble forms of surface lipoproteins
(SLPs) that also possess beta-barrel type structures provide a poten-
tially commercially viable means of antigen production (Fig. 1).
The first step in the design of hybrid antigens is the selection of
an appropriate integral OMP to target for vaccine development.
Ideally, the OMP is present in all strains of the bacterial pathogen
that will be targeted with the vaccine and is required for survival in
their mammalian or other vertebrate host. TBDTs are ideal targets
due to the essential role they play in the survival of the bacteria and
their expression under most conditions in their host [3, 5]. Several
groups of gram-negative bacteria that reside exclusively in the
upper respiratory or genitourinary tracts of their host possess
TBDTs that are utilized for acquiring iron from host transferrin
or lactoferrin [6]. These TBDTs have been shown to be essential
for survival and disease pathogenesis [7–9] and thus fit the criteria
for ideal vaccine targets.
The other requirement for the hybrid antigen approach is an
SLP with the appropriate structural features for loop display—a
characteristic of proteins transported to the surface by the surface
lipoprotein assembly module (SLAM) system [10], which is present
in most gram-negative species of interest. Receptors involved in
 Design and Production of Hybrid Antigens 117

acquisition of iron from host transferrin or lactoferrin have a

bilobed SLP component, transferrin or lactoferrin binding protein
B (TbpB or LbpB), that can readily serve as the loop display
scaffold. In the first hybrid antigen study, a loopless C-lobe
(LCL), derived from the TbpB of the human pathogen Neisseria
meningitidis, and the intact TbpB from the porcine pathogen
Glaesserella parasuis were used as display scaffolds [1]. Using the
detailed structural information for the scaffold proteins [11, 12]
and the TBDT, TbpA from Neisseria [13], single, modest-sized
loop regions were selected for display. The successful production,
using conventional systems, of soluble hybrid antigens in this study
was likely a function of displaying single, modest-sized TbpA loop
regions as opposed to using multiple, larger loops. One conse-
quence of using the LCL in this study was the need to use a
Neisseria strain deficient in TbpB to assess the contribution of the
displayed TbpA loops to the functional properties of the immune
response induced by the hybrid antigen. It is important to appreci-
ate that most hybrid antigens that are designed for commercial use
would employ a scaffold from the targeted bacterial pathogen and
thus would resemble the Neisseria TbpA/B hybrid antigens
described in this study.
The N. meningitidis LCL [1] was also used as the scaffold in a
follow-up study for display of the surface loops of the zinc transport
protein, ZnuD, from Acinetobacter baumannii [2]. Since a series of
hybrid antigens displaying individual loops (or combinations of
loops) were planned, an N-terminal maltose binding protein
(MBP) was included to facilitate proper folding of the hybrid
antigen. However, the expression from a T7 promoter resulted in
the production of insoluble protein for most of the hybrid antigens,
requiring solubilization with 8 M urea and refolding with a gradient
of decreasing urea concentrations to obtain a soluble hybrid anti-
gen [2]. In this study, the hybrid antigen was not purified from the
N-terminal MBP after cleavage with tobacco etch virus (TEV)
protease, but rather the entire recombinant protein complex was
used in the immunization and challenge experiments. Fortunately,
the control LCL scaffold with N-terminal MBP protein did not
induce a protective immune response against A. baumannii in a
mouse sepsis model, indicating that the protection induced by
hybrid antigens could be attributed to the individual loop regions
present on the scaffold.
One of the challenges with the hybrid antigen approach is the
ability to determine the proportion of the antibody response that is
induced against the specific loop being displayed, and particularly
the titer of antibody directed against conformations of the loop
present in the native TBDT/OMP. Although whole-cell enzyme-
linked immunosorbent assays (ELISAs) using the target bacterium
can provide some assessment of the titer of the relevant antibodies,
118 Somshukla Chaudhuri et al.

results are influenced by numerous factors including the level of

expression of the targeted TBDT/OMP, antigenicity of the targets
(which could be affected by the method of inactivating or killing
the bacterium for coating plates), and the presence of preexisting
antibodies in tested sera that could react with a variety of antigens
present on the bacterium.
In this chapter, we describe a method for preparing ELISA
plates with the native TBDT that is designed to overcome these
limitations and provide the titer of antibody directed against native
conformation of the loops present in the hybrid antigen. The use of
the N-terminal streptavidin binding peptide (SBP) and
streptavidin-coated plates eliminates the need to isolate the
TBDT from a crude extract and can be exploited to readily evaluate
the cross-reactivity of antisera against heterologous TBDT variants.
We have developed this protocol by modifying our high-
throughput, nonbiased ELISA method which uses streptavidin-
coated plates to capture and immobilize soluble antigens from
crude Escherichia coli lysates in which the recombinant protein of
interest is biotinylated in vivo during expression [14]. The exten-
sion of this approach with TbpA protein was feasible due to the
ability to assess functionally folded protein by binding of labeled
transferrin. Although this assay may not be available for other
TBDTs, such as siderophore receptors, it could be performed in
parallel or by coexpression of TbpA and the TBDT. For hybrid
antigens displaying multiple loops, the sera would have to be pre-
treated with hybrid antigens displaying individual loops, or the
single-loop hybrids would have to be included in the incubation
mixtures to determine the titer against individual loops.
In this chapter, we provide detailed methods for designing
hybrid antigens and producing soluble and insoluble proteins
with a conventional T7 expression system that clearly has many
limitations. Alternative expression systems are being explored and
the reader should not hesitate to use alternate systems for produc-
tion and purification of the hybrid antigens. The method for prep-
aration of ELISA plates for evaluating the antibody response
against the loops displayed in the hybrid antigen can be implemen-
ted with a representative set of variant TBDTs to evaluate the cross-
reactivity of the antibody response that would likely correlate with
the cross-protective properties of the immune response. We hope
that the additional information in the Notes regarding the design
and production of hybrid antigens and assessment of the immune
response will provide sufficient insight for successful implementa-
tion of this approach for other OMPs being considered as candidate
vaccine antigens.
 Design and Production of Hybrid Antigens 119

2 Materials

Prepare all solutions using double-distilled water. Prepare and store

all solutions at room temperature, unless indicated otherwise. Any
sterilization procedures, if applicable, will be listed in the following
descriptions of each reagent.

2.1 Autoinduction 1. ZY media: 1% tryptone, 0.5% yeast extract. Weigh out 15 g

Medium Components tryptone and 7.5 g yeast extract and transfer to a 4 L Erlen-
meyer flask. Add 1.5 L of distilled water and place a piece of
aluminum foil over the mouth of the flask. Autoclave, then
store at room temperature.
2. 20
NPS: 0.5 M (NH4)2SO4, 1 M KH2PO4, 1 M Na2HPO4.
Weigh out 132 g (NH4)2SO4, 272 g KH2PO4, and 284 g
Na2HPO4 and transfer to a large beaker. Add 1400 mL of
distilled water, dissolve solids with agitation, then transfer
solution to a graduated cylinder and make up volume to 2 L
with water. Transfer to a 2-L glass bottle, autoclave, then store
at room temperature.
3. 50
5052: 2.5% glucose, 10% lactose, 25% glycerol. Weigh out
50 g glucose, 200 g lactose, and 500 g (396.42 mL) glycerol
and transfer to a large beaker. Add 1400 mL distilled water,
dissolve solids with agitation, then transfer solution to a
graduated cylinder and make up volume to 2 L with water.
Transfer to a 2 L glass bottle, autoclave, then store at room
temperature.
4. 1 M MgSO4: Weigh out 61.62 g of MgSO4 heptahydrate and
transfer to a beaker. Add 200 mL of distilled water, dissolve
solids with agitation, then transfer to a graduated cylinder and
make up volume to 250 mL with water. Transfer to a 250 mL
glass bottle, autoclave, then store at room temperature.

2.2 Other Media 1. LB liquid media: Dissolve 25 g of commercial premixed solids

Used for Cultivating in 1 L of distilled water by shaking manually or with a magnetic
Bacteria stir bar. Autoclave, then store at room temperature.
2. LB agar plates: Dissolve 32 g of commercial premixed solids in
1 L of distilled water by shaking manually or with a magnetic
stir bar. Autoclave, let cool, and then add the appropriate
antibiotic (if desired). Mix using a magnetic stir bar, then
pour into Petri dish plates (~20–25 mL in each).

2.3 Stock Solutions 1. 1 M Tris pH 8.0: Weigh out 242.28 g of Tris and transfer to a
of Components large beaker. Add 1400 mL of distilled water, dissolve solids
of Protein Purification with agitation, then transfer solution to a graduated cylinder
Buffers and make up volume to ~1950 mL to leave room for the
120 Somshukla Chaudhuri et al.

addition of acid to adjust to the desired pH. Filter-sterilize,

adjust pH to 8.0 using concentrated HCl, then store at room
temperature in a 2 L glass bottle.
2. 5 M NaCl: Weigh out 584.4 g of NaCl and transfer to a large
beaker. Add 1400 mL of distilled water. Partially dissolve solids
using heat and agitation, then gradually add more water until
all solids have dissolved. If necessary, transfer solution to a
graduated cylinder and make up volume to 2 L with water
(see Note 1). Filter-sterilize, then store at room temperature
in a 2-L glass bottle.
3. 1 M imidazole pH 7.4: Weigh out 68.08 g of imidazole and
transfer to a large beaker. Add 700 mL of distilled water,
dissolve solids with agitation, then transfer solution to a
graduated cylinder and make up volume to ~950 mL to leave
room for the addition of acid to adjust to the desired
pH. Filter-sterilize, adjust pH to 7.4 using concentrated HCl,
then store at room temperature in a 1 L glass bottle wrapped in
aluminum foil to shield solution from exposure to light.
4. 10
phosphate-buffered saline (PBS) pH 7.4: Weigh out 80 g
NaCl, 2 g KCl, 14.4 g Na2HPO4, 2.4 g KH2PO4 and transfer
to a large beaker. Add 700 mL of distilled water, dissolve solids
with agitation, then transfer solution to a graduated cylinder
and make up volume to ~950 mL to leave room for the
addition of acid to adjust to the desired pH. Filter-sterilize,
adjust pH to 7.4 using concentrated HCl, then store at room
temperature in a 1 L glass bottle.

2.4 Protein 1. Resuspension Buffer: 50 mM Tris pH 8.0, 300 mM NaCl,

Purification Buffers 10 mM imidazole pH 7.4. Mix 50 mL of 1 M Tris pH 8.0
stock solution, 60 mL of 5 M NaCl stock solution, and 10 mL
of 1 M imidazole pH 7.4 stock solution in a graduated cylinder.
Make up volume to 1 L using distilled water. Filter-sterilize,
then store at room temperature.
2. Wash Buffer: 50 mM Tris pH 8.0, 1 M NaCl, 20 mM imidazole
pH 7.4. Mix 50 mL of 1 M Tris pH 8.0 stock solution, 200 mL
of 5 M NaCl stock solution, and 20 mL of 1 M imidazole
pH 7.4 stock solution in a graduated cylinder. Make up volume
to 1 L using distilled water. Filter-sterilize, then store at room
temperature.
3. Elution Buffer: 50 mM Tris pH 8.0, 300 mM NaCl, 300 mM
imidazole pH 7.4. Mix 50 mL of 1 M Tris pH 8.0 stock
solution, 60 mL of 5 M NaCl stock solution, and 300 mL of
1 M imidazole pH 7.4 stock solution in a graduated cylinder.
Make up volume to 1 L using distilled water. Filter-sterilize,
then store at room temperature.
 Design and Production of Hybrid Antigens 121

4. Exchange Buffer: 50 mM Tris pH 8.0, 600 mM NaCl. Mix

100 mL of 1 M Tris pH 8.0 stock solution with 120 mL of 5 M
NaCl stock solution in a graduated cylinder. Make up volume
to 1 L using distilled water. Filter-sterilize, then store at room
temperature.
5. Denaturing Lysis Buffer: 50 mM NaH2PO4, 300 mM NaCl,
6 M urea, pH 8.0. Weigh out 6.9 g NaH2PO4 monohydrate
and 360.36 g urea and transfer to a large beaker. Add 60 mL of
5 M NaCl stock solution and 640 mL distilled water. Dissolve
solids with agitation, then transfer to a graduated cylinder and
make up volume to ~950 mL, leaving room for the addition of
acid to adjust the pH. Adjust pH to 8.0 using
concentrated HCl.
6. Denaturing Wash Buffer: 50 mM NaH2PO4, 300 mM NaCl,
6 M urea, 5 mM imidazole, pH 8.0. Weigh out 6.9 g
NaH2PO4 monohydrate and 360.36 g urea and transfer to a
large beaker. Add 60 mL of 5 M NaCl stock solution, 5 mL of
1 M imidazole pH 7.4 stock solution, and 635 mL of distilled
water. Dissolve solids with heat and agitation, then transfer to a
graduated cylinder and make up volume to ~950 mL, leaving
room for the addition of acid to adjust the pH. Adjust pH to
8.0 using concentrated HCl.
7. Refolding Wash Buffer: 50 mM NaH2PO4, 300 mM NaCl,
5 mM imidazole, pH 8.0. Weigh out 6.9 g NaH2PO4 mono-
hydrate and transfer to a large beaker. Add 60 mL of 5 M NaCl
stock solution, 5 mL of 1 M imidazole pH 7.4 stock solution,
and 635 mL of distilled water. Dissolve solids with agitation,
then transfer to a graduated cylinder and make up volume to
~950 mL, leaving room for the addition of acid to adjust to the
pH. Adjust pH to 8.0 using concentrated HCl.
8. Refolding Elution Buffer: 50 mM NaH2PO4, 300 mM NaCl,
400 mM imidazole, pH 8.0. Weigh out 6.9 g NaH2PO4
monohydrate and transfer to a large beaker. Add 60 mL of
5 M NaCl stock solution, 400 mL of 1 M imidazole pH 7.4
stock solution, and 240 mL of distilled water. Dissolve solids
with agitation, then transfer to a graduated cylinder and make
up volume to ~950 mL, leaving room for the addition of acid
to adjust the pH. Adjust pH to 8.0 using concentrated HCl.

2.5 Other Materials 1. Protease inhibitor tablets: store at 4  C or according to the

for Protein Production manufacturer’s instructions until ready to use.
and Purification 2. Lysozyme: prepare 10 mg/mL solution by dissolving solids in
ddH2O and storing in 6-mL aliquots at 20  C until ready
to use.
3. Deoxyribonuclease (DNase) I: prepare 5 mg/mL solutions by
dissolving solids in ddH2O and storing in 200-μL aliquots at
20  C until ready to use.
122 Somshukla Chaudhuri et al.

4. Nickel-nitrilotriacetic acid (Ni-NTA) resin: prepare according

to the manufacturer’s instructions and store at 4  C until ready
to use (see Note 2).
5. Dialysis tubing, 6–8 kDa molecular weight cutoff (MWCO):
Store at room temperature until ready to use. Moisten the
membrane with ddH2O immediately prior to use.
6. Centrifugation concentrator, 50 kDa MWCO (see Note 3):
Prime with 20 mL ddH2O followed by 20 mL of the appropri-
ate buffer immediately prior to use.

2.6 Materials 1. DNA and protein sequences of the SLP scaffold and the TBDT.
for Hybrid Antigen 2. Access to programs/software for protein modeling, protein
Design visualization, and gene cloning and visualization (see Note 4).

2.7 Materials 1. 50 mM Tris pH 8.0: Add 50 mL of 1 M Tris pH 8.0 stock

for TBDT Extraction solution to a 1-L graduated cylinder. Make up volume to 1 L
using distilled water. Filter-sterilize, adjust pH to 8.0 with HCl,
transfer to a 1-L glass bottle, then store at room temperature.
2. Elugent detergent (see Note 5).

2.8 ELISA Reagents 1. PBST: Add 100 mL of 10
PBS to a 1-L graduated cylinder.
Make up volume to 1 L using distilled water. Add 0.5 mL of
Tween-20 (final concentration: 0.05%) using a wide-bore 1 mL
pipette tip. Pipette up and down numerous times to ensure that
all the viscous detergent is mixed into the PBS. Shake the bottle
to mix well. Store at room temperature. This can be prepared
several days in advance.
2. Blocking solution: Add 2.5 g of skim milk powder to 25 mL of
PBST in a 50-mL conical tube and vortex to mix. Make up
volume to 50 mL with more PBST. This can be prepared a day
in advance and stored at 4  C. Bring to room temperature prior
to use. Alternatively, 5% w/v bovine serum albumin (BSA) can
also be used as a blocking reagent: Add 50 g of BSA to 1 L of
PBS and dissolve with stirring. Once dissolved, filter-sterilize
and store at 4  C.
3. Diluent solution: Add 1.25 g of skim milk powder to 25 mL of
PBST in a 50-mL conical tube and vortex to mix. Make up
volume to 50 mL with more PBST. This can be prepared a day
in advance and stored at 4  C. Bring to room temperature prior
to use.
4. Transferrin solution: Using a stock solution of 0.5 mg/mL
transferrin conjugated to horseradish peroxidase (Tf-HRP),
prepare a 1:1000 dilution for a working solution in 2.5% skim
milk solution in PBST just prior to use. Vortex to mix well.
 Design and Production of Hybrid Antigens 123

5. 3,30 ,5,50 -Tetramethylbenzidine (TMB) substrate solution: Pre-

pare according to manufacturer’s instructions and store at 4  C
until use.
6. 4 N HCl quenching solution: Make this solution in a fume
hood. Measure out 200 mL of distilled water in a graduated
cylinder, then transfer to a clean 500 mL glass bottle. Measure
out 100 mL of 12 N HCl in a clean glass graduated cylinder.
Gently and slowly pour the HCl into the water. Stir with a clean
glass rod. Let sit for 2 min. Store at room temperature.
7. Serum samples from animals immunized with TbpA, TbpA-
based hybrid antigens, or other TBDT-based antigens. Store at
20  C until use. Thaw on ice for 20–30 min and vortex to mix
prior to use.
8. Secondary antibody solution: HRP-conjugated antibody spe-
cific for detection of IgG of species immunized. Prepare
according to the manufacturer’s instructions, divide into
10-μL aliquots, and store at 20  C until use. For use, thaw
an aliquot on ice for 20 min and prepare working solution to
1:10,000 (or according to manufacturer’s recommendations)
in diluent solution just prior to use.
9. Other miscellaneous materials: streptavidin- or neutravidin-
coated 96- or 384-well ELISA plates, multichannel pipettors,
adhesive cover slips for ELISA plates, reagent boats/basins for
containing ELISA solutions.

3 Methods

Carry out all procedures at room temperature unless stated other-

wise. All techniques with live cells are to be performed in an aseptic
manner either next to a Bunsen burner or in a biosafety cabinet.

3.1 Designing 1. Acquire published structures of the SLP scaffold and the
and Preparing Hybrid TBDT. If structures are not available, generate structural mod-
Antigen Genes els of the scaffold and the TBDT in silico. Programs such as
I-TASSER, RaptorX, or Phyre2 can all be used.
2. Visualize the models using PyMOL.
3. Identify the anchoring residue sites of loops on the SLP scaf-
fold and the TBDT (see Note 6). Locate the antiparallel beta-
strands in each model and label the residues in alternating color
leading up to the “loop,” as shown in Fig. 2 for the example
with TBDT (A. baumannii ZnuD) and SLP scaffold (deriva-
tive of TbpB from N. meningitidis). The splicing site of the
SLP scaffold is determined by selecting anchoring residues that
are next to each other (parallel to each other) and have side
chains that face the same direction on the scaffold (“out” and
124 Somshukla Chaudhuri et al.

Fig. 2 Example of hybrid antigen design. (a) Computer generated model of the TBDT, ZnuD, from Acinetobacter
baumannii with a loop highlighted (red) where the residues of the anchoring beta-barrel are colored differently
depending on whether their side chains are oriented toward the inside of the barrel (orange) or the outside/
away from the barrel (cyan). (b) The scaffold, a derivative of the SLP TbpB C-lobe from Neisseria meningitidis
strain M982 (PDB 5KKX) [1], is labeled similarly to a, with alternating residues colored. (c) The loop from the
ZnuD is determined by selecting parallel anchoring residues (side chains facing the same direction) on the
beta-barrel. If the anchoring residues of the loop are facing in (black arrow in c), then the anchoring residues
of the scaffold face out (black arrow in d) to maintain the antiparallel beta-strand formation

labeled with arrow in Fig. 2d). These anchoring residues are

included in the SLP scaffold (see Note 6).
4. Determine the splicing site of the loop by selecting parallel
anchoring residues that face the opposite direction to that of
the SLP scaffold on the TBDT (“in” and labeled with arrow
Fig. 2c). These anchoring residues are included in the loop.
5. Using an in silico program for DNA cloning, identify the gene
segments corresponding to the structural attributes of the SLP
scaffold and TBDT identified in steps 3 and 4.
6. Generate the DNA sequence of the novel hybrid antigen by
combining the different gene segments from both the SLP
scaffold and the TBDT. Translate the gene sequence in silico
to ensure the resulting protein is in frame.
7. The novel hybrid gene can either be synthesized by a commer-
cial vendor or can be generated in the lab via splicing-by-
overlap/extension polymerase chain reaction (SOE-PCR)
using primers that anneal to the appropriate regions of the
templates.
 Design and Production of Hybrid Antigens 125

Fig. 3 A schematic of the His-Bio-MBP T7 expression vector (pE5770) with a

gene of interest (hybrid antigen) cloned into the expression locus. The
components of the vector (see Note 7) that are relevant to the protein
production methods discussed in this manuscript are illustrated in the figure

8. Clone the hybrid gene fragment into plasmid pE5770 or any

equivalent plasmid with similar elements (see Note 7 and
Fig. 3).
9. In preparation for transformation of the newly constructed
plasmid into E. coli TOP10 cells for long-term storage, thaw
an aliquot of chemically competent TOP10 cells on ice.
10. Pipet 50–100 ng of the plasmid encoding the hybrid antigen
gene into the thawed aliquot and incubate on ice for 30 min
(see Notes 8 and 9).
11. Heat-shock the cells at 42  C for 30–60 s. Promptly place the
tube back on ice for 2 min.
12. Add 400 μL of cold sterile LB medium and then transfer the
tube to a 37  C shaking incubator for 1 h.
13. Plate 100 μL of transformation mixture onto an LB agar plate
supplemented with 100 μg/mL ampicillin and incubate at
37  C overnight. Colonies generated with the pE5770 plasmid
can be streaked onto plates supplemented with 30 μg/mL
kanamycin to screen for loss of the original insert.
14. Confirm by colony PCR or PCR of plasmid DNA prepared
from positive clones and sequencing of insert or by restriction
digestion of plasmid.
15. Isolate and store plasmid at 20  C for subsequent use.
16. Inoculate 5-mL of LB medium supplemented with 100 μg/
mL ampicillin with a colony confirmed to contain the desired
plasmid. Incubate overnight (~16 h) at 37  C in a shaking
incubator.
17. The next day, create a 16% glycerol stock by mixing 800 μL of
the broth culture with 200 μL of sterile 80% glycerol. Store at
80  C.
126 Somshukla Chaudhuri et al.

3.2 Expressing 1. Thaw out the plasmid encoding the desired hybrid protein
Hybrid Protein designed in Subheading 3.1, as well as the competent cells
Antigens in E. coli derived from the strain of E. coli to be used for protein expres-
Using a T7 Expression sion (see Note 10). The plasmid concentration should be
Vector ~100 ng/μL, and competent cells should be stored in 50- or
100-μL aliquots. After removing the competent cells from the
80  C freezer, let the cells sit on ice for 20 min prior to
transformation.
2. Add 1–2 μL of plasmid to the competent cells and let the cells
sit on ice for 30 min. Next, transfer to a prechilled 14-mL
round-bottom culture tube.
3. Prepare a 42  C water bath or heat block, then “heat-shock”
the cells at this temperature for 30–60 s, followed by a 2-min
incubation on ice. Promptly add 700 μL of LB medium, then
incubate the cells with shaking at 37  C for 1 h.
4. Add 6 mL of LB containing the appropriate antibiotic (i.e., add
6 μL of 100 mg/mL ampicillin if you are using ampicillin
resistance to select for the presence of your plasmid) to the
culture and incubate with shaking at 37  C for another 4 h.
5. Approximately 1–1.5 h before the end of the 4-h incubation
period, prepare the autoinduction medium by aseptically add-
ing 75 mL NPS, 30 mL 5052, and 1.5 mL MgSO4 to 1.5 L of
ZY media (see Note 11). Four 4-L flasks containing 1.5 L of
media in each are used in a typical protein production run, but
the culture volume can be adjusted depending on the expected
yield of the protein being produced. Add the appropriate anti-
biotic at half the normal concentration (i.e., use 50 μg/mL
instead of 100 μg/mL if using ampicillin). Prewarm the media
in 37  C shaking incubator at a shaking speed of ~50 rpm (see
Note 12).
6. At the end of the 4-h incubation mentioned in step 4, inocu-
late 1.5 L of autoinduction media with 1.5 mL of the starter
culture. Repeat as needed, depending on the number of flasks
being used.
7. Incubate the flasks at 37  C with vigorous shaking (175 rpm)
for 18 h, then adjust the temperature to 20  C and incubate for
another 24 h (see Note 13).
8. Harvest cells by centrifugation at 5000
g for 25 min at 4  C.
9. Prepare Resuspension Buffer (25 mL per L of culture) and add
1 protease inhibitor tablet, 6 mL of 10 mg/mL lysozyme, and
200 μL of 5 mg/mL DNase I (increase the amounts of these
reagents as necessary if the culture volume is greater than 6 L).
Mix in an appropriate-sized beaker, keeping in mind that the
cell pellets will subsequently be transferred to this beaker.
 Design and Production of Hybrid Antigens 127

10. After centrifugation, decant the supernatant and resuspend the

cell pellet(s) in the Resuspension Buffer mixture prepared in
step 9 (see Note 14). For optimal lysis, add a magnetic stir bar
to the mixture and let stir at slow speed for 30 min at 4  C to
get rid of cell clumps.
11. Lyse the cells by passing the sample through a cell homoge-
nizer four times (see Note 15).
12. After cell lysis, centrifuge at 35,000
g for 90 min at 4  C to
separate out the cell debris.
13. Collect the supernatant and discard the pelleted cell debris.
Filter the supernatant through a 0.2-μm filter (see Note 16).
14. Determine whether the fusion protein is present in the super-
natant by performing sodium dodecyl sulfate-polyacrylamide
gel electrophoresis (SDS-PAGE) using a 10-μL aliquot of the
supernatant. If a band corresponding to the molecular weight
of the fusion protein of interest is detected in the SDS-PAGE
gel, proceed to Subheading 3.3. If the protein of interest is not
detected, proceed to Subheading 3.4 (see Note 17).

3.3 Purification 1. Prepare 1 L each of Resuspension Buffer, Wash Buffer, and

of Soluble Protein Elution Buffer.
Antigens 2. Wash and equilibrate a 5-mL Ni-NTA column with Resuspen-
sion Buffer prior to use (see Note 18).
3. Using either a peristaltic pump or a fast-purification liquid
chromatography (FPLC) system, continuously circulate the
lysate prepared in Subheading 3.2 through the Ni-NTA col-
umn overnight (16–18 h) at 4  C. If using a peristaltic pump,
circulate lysate at ~0.5 mL/min (see Note 19).
4. Wash the column with at least 75 mL Wash Buffer, or until the
protein concentration of the wash fractions is <0.1 mg/mL.
Take one or more samples of the wash fractions for SDS-PAGE
analysis.
5. Elute the protein bound to the column using high-imidazole
Elution Buffer in 2-mL aliquots until the protein concentration
reaches <0.1 mg/mL.
6. Take samples of the eluted fractions for SDS-PAGE to verify
the presence of the desired protein antigen and determine
which fractions to collect and pool for further processing
(Fig. 4a). If desired, include the wash fractions as well to
confirm that the protein is present in the eluted fraction and
not the wash fractions and therefore is binding to the Ni-NTA
resin.
7. Determine the amount of protein present in the pooled
sample-containing fraction using a spectrophotometer or
NanoDrop device (or equivalent method of determining pro-
tein concentration).
128 Somshukla Chaudhuri et al.

Fig. 4 Representative SDS-PAGE results for purification of a soluble hybrid antigen using the TbpB derived
from Haemophilus influenzae as an example. (a) SDS-PAGE results for 10-μL aliquots of fractions from a
Ni-NTA (IMAC—immobilized metal affinity chromatography) column. FT1 indicates a sample from the pooled
flow through buffer containing proteins that did not bind to the column. Fractions A3-D9 represent the specific
wells of the 96-well collection block that collected the fractions eluted with the imidazole gradient. Fractions
A3–D9 were pooled based on the presence of the 130-kDa MBP-TbpB fusion protein for subsequent TEV
cleavage. (b) SDS-PAGE results for a 10-μL aliquot of the MBP-TbpB fusion protein following cleavage with
TEV protease. The gel confirms that TEV cleavage was complete as there is no MBP-TbpB band present with
TbpB, MBP and TEV protease as the main protein bands. Applying this sample to a Ni-NTA column removes the
MbpB and TEV, leaving TbpB as the main protein (similar to the purified LbpB example shown in Fig. 5)

8. In preparation for cleavage of the fusion protein using tobacco

etch virus (TEV) protease—for which high-salt, low-imidazole
conditions are optimal—transfer the eluted protein sample to
dialysis tubing, ensure the tube is fastened at both ends, and
place the tube in a beaker containing 2 L of Exchange Buffer.
Place a magnetic stir bar in the beaker and let stir at a slow
speed overnight (16–18 h) at 4  C.
9. Remove the dialysis tubing from the 2-L beaker, open at one
end of the tube, and add TEV protease (see Note 20). Then,
place the tubing back in the Exchange Buffer (this can be
carried over from step 8; there is no need to prepare fresh
Exchange Buffer). Let stir overnight at 4  C.
10. The next day, run a sample of the cleaved protein on an SDS-
PAGE to determine the cleavage efficiency (Fig. 4b). If cleav-
age is incomplete—typically ascertained by the presence of the
intact fusion protein in the SDS-PAGE gel—add more TEV
protease.
11. In preparation for a second round of Ni-NTA chromatography
to remove the cleaved MBP fusion partner and TEV protease
from the protein sample, dialyze the sample overnight at 4  C
against 2 L of Resuspension Buffer. This step serves to replen-
ish the imidazole that was removed from the sample during
dialysis against Exchange Buffer in step 8.
 Design and Production of Hybrid Antigens 129

Fig. 5 Representative SDS-PAGE results following the separation of MBP and TEV
protease from the recombinant protein of interest using a Ni-NTA column. A
modified Neisseria gonorrhoeae lactoferrin binding protein B (LbpB) is provided
as an example. Lane 1: flow through containing the LbpB; lane 2: pooled eluted
fractions (MBP and TEV protease are both present here); lane 3: concentrated
sample of flow through (lane 1). Each lane was loaded with a 10-μL aliquot

12. Circulate the cleaved protein through the Ni-NTA column as

described in step 3 (see Note 21). The polyhistidine-tagged
MBP and TEV protease will bind to the resin and will thus
separate these contaminants from the desired protein antigen.
13. After circulation through the Ni-NTA column, take a sample of
the protein (which should not have bound to the column) and
run an SDS-PAGE gel to confirm that the MBP and TEV
protease have been separated from the desired protein antigen
(Fig. 5).
14. Exchange the buffer containing the purified protein by dialysis
against PBS overnight (16–18 h) at 4  C, then concentrate the
protein to a final concentration of 1–10 mg/mL by centrifuga-
tion using a concentrator with the appropriate MWCO.
15. Promptly store the protein in 100-μL aliquots at 80  C until
ready to use for animal immunizations (see Note 22).

3.4 Purification 1. Perform steps 1–8 described under Subheading 3.2.

of Insoluble Protein 2. Prepare Denaturing Lysis Buffer (60 mL per L of culture).
Antigens After centrifugation (Subheading 3.2, step 8), decant the
supernatant and resuspend the cell pellet using Denaturing
Lysis Buffer.
130 Somshukla Chaudhuri et al.

3. Lyse the cells by sonicating for two 5 min periods on ice using
conditions that do not result in overheating the samples. These
generally involve short pulses with cooling periods of 2–4 times
longer. We use a QSonica sonicator (pulse 03, amplitude 65%).
4. Centrifuge the lysate for 35 min at 20,440
g at 4  C.
5. Collect the supernatant and filter through a 0.2-μm filter.
6. Add 3 mL of Ni-NTA resin for every 50 mL of filtered lysate
(see Note 23), then mix gently in 50-mL conical tubes using a
rotary shaker overnight (~16 h) at room temperature.
7. Prepare four buffers with decreasing concentrations of urea
(3 M, 1.5 M, 0.75 M, 0.375 M) by starting with Denaturing
Wash Buffer and subsequently performing serial dilutions,
using Refolding Wash Buffer as a diluent.
8. Centrifuge the lysate–resin mixture for 10 min at 3220
g at
room temperature, then decant the supernatant. Next, using
18 mL (or 6 volumes of Ni-NTA resin) of the initial buffer in
the aforementioned serial dilution (50 mM NaH2PO4,
300 mM NaCl, 3 M urea, 5 mM imidazole, pH 8.0), resuspend
the Ni-NTA resin and incubate the resulting mixture on a
rotary shaker for 30 min at room temperature.
9. Repeat step 8 for each of the subsequent buffers in the serial
dilution series prepared in step 7.
10. Resuspend the Ni-NTA resin in 18 mL Refolding Wash Buffer,
then transfer the mixture to a gravity column.
11. Elute any proteins bound to the Ni-NTA using 18 mL of
Refolding Elution Buffer (see Note 24). The eluted volume
can be collected in 1-mL fractions.
12. Verify the presence of the desired hybrid antigen using SDS-
PAGE, then perform steps 8–12 described under Section.

3.5 Cloning of Tbdt 1. Identify mature sequence of TBDT using the online software
in Fusion SignalP (see Note 25).
with Streptavidin 2. Design primers to amplify the selected gene or synthesize the
Binding Protein (SBP) sequence.
3. Clone the TBDT gene into plasmid pE5771 or an alternate
expression plasmid containing SBP (see Note 26).
4. Transform the newly constructed plasmid into E. coli TOP10
cells for long-term storage, as described in Subheading 3.1,
steps 9–15.

3.6 Preparation 1. Thaw plasmid DNA and an aliquot of chemically competent

of Detergent-Extracted E. coli C43 cells on ice for 20 min (see Note 27 and Fig. 6).
SBP-TBDT Fusion 2. Pipette 50–100 ng of SBP-TBDT plasmid into the thawed
Protein aliquot and incubate on ice for 30 min.
 Design and Production of Hybrid Antigens 131

Fig. 6 A schematic of the His-SBP-TBDT T7 expression vector (pE5771) with a

gene of interest (a TBDT to be used in ELISA assays) cloned into the expression
locus. The components of the vector (see Note 26) that are relevant to the
protein production methods discussed in this manuscript are illustrated in the
figure

3. Heat-shock the cells at 42  C for 30–60 s. Promptly place the

tube back on ice for 2 min.
4. Add 400 μL of cold sterile LB medium.
5. Gently pipette out the entire mixture from the 1.5 mL tube and
into a 50-mL conical tube. Incubate in a shaking incubator at
37  C for 1 h.
6. Add 14.5 mL of autoinduction medium supplemented with
100 μg/mL ampicillin and incubate in a shaking incubator at
37  C overnight.
7. The following day, centrifuge the culture at 3220
g at 4  C
for 10 min.
8. Decant the supernatant, add 15 mL of cold 50 mM Tris
pH 8.0, and resuspend the cell pellet.
9. Centrifuge again as described in step 7, decant the supernatant,
then add 2 mL of cold 50 mM Tris pH 8.0 and 40 μL of
Elugent detergent (final concentration: 1%, commercial Elu-
gent is a 50% solution). Shake overnight at 4  C (see Note 28).
10. The following day, centrifuge the samples at 16,100
g at 4  C
for 30 min.
11. Carefully remove the supernatant containing the detergent
extracted SBP-TBDT and pipette slowly into a 15-mL conical
tube containing 8 mL of cold filtered PBST with 0.125%
Elugent (0.25% of commercial preparation, final protein dilu-
tion: 1 in 5).
12. Gently rotate the tube by hand several times to ensure that it
mixes well. Keep on ice until ready to use (see Notes 29
and 30).
132 Somshukla Chaudhuri et al.

3.7 Coating 1. Pour the SBP-TBDT–PBST mixture into a reagent boat, and,
Streptavidin ELISA using a multichannel pipette, pipette out 100 μL into each well
Plates with Detergent- of a streptavidin-coated 96-well plate (see Note 31).
Extracted SBP-TBDT 2. Incubate at room temperature for 1 h, then wash with 250 μL/
well of PBST three times (see Notes 32 and 33).
3. Add 250 μL of the prepared 5% blocking solution to each well
after the wash step, incubate at room temperature again for 1 h,
then wash with 250 μL/well of PBST three times.

3.8 Assessing 1. During the blocking step (Subheading 3.7, step 3), add 1 μL
Conformation of TbpA Tf-HRP to 1 mL of diluent solution (i.e., final dilution of
by Its Ability to Bind 1/1000).
to Transferrin 2. Add 100 μL of the Tf-HRP solution to each of four wells (i.e.,
two test and two control wells), incubate for 1 h, then wash
with 250 μL of PBST three times.
3. Add 50 μL of TMB to the four wells and develop in the dark
(a cupboard/drawer is adequate) for 20 min, then add 25 μL of
4 N HCl to stop the reaction. Measure the optical density at
450 nm (see Note 34 and Fig. 7).

3.9 Assessing 1. During the blocking step (Subheading 3.7, step 3), thaw
antibody Titers serum samples on ice for at least 20 min and vortex to mix well.
in Serum Samples 2. Add 100 μL of diluent solution to each well. Add additional
from Immunized diluent solution in the first well of each row (A to H). The
Animals amount of additional diluent solution will depend on the
desired starting dilution for the serum used in the assay, such
that the final volume of diluent with added serum is 200 μL.

Fig. 7 Schematic of TbpA conformation assay. Detergent-extracted TbpA fused to

streptavidin binding peptide (SBP) is immobilized on streptavidin-coated wells.
Transferrin conjugated to horseradish peroxidase (Tf-HRP) is then added to the
wells. If the immobilized TbpA is folded properly, this allows Tf binding to occur,
which results in a signal from the HRP once the developing substrate is added
 Design and Production of Hybrid Antigens 133

3. Add the desired amount of each serum sample to the appropri-

ate wells containing diluent solution to achieve the desired
initial serum dilution in a total volume of 200 μL in each well
in the first column. Gently pipet up and down 10 times to mix.
4. Adjust a multichannel pipettor to 100 μL, then mix the serum
solutions in the first column of the plate by pipetting up and
down 5 times.
5. Remove 100 μL from the first column and transfer it the next
column. Pipette up and down 5 times. Repeat with the next
column.
6. Continue performing twofold serial dilutions until column
11 (see Note 35). Discard the leftover 100 μL.
7. Incubate at room temperature for 1 h, then wash with 250 μL
of PBST three times.
8. Prepare the secondary antibody solution by adding 5 μL of
anti-rabbit IgG or anti-mouse IgG (as appropriate for the
animals in which the immune sera were generated) antibody
conjugated to HRP to a 50-mL diluent solution (final antibody
dilution: 1 in 10,000). Vortex to mix well.
9. Add 100 μL to each well, except for column 12.
10. In column 12, add 100 μL of diluent solution (no antibody) in
wells 12A, 12B, and 12C. These are negative controls.
11. In wells 12D, 12E, and 12F, add 100 μL of the secondary
antibody solution. These are the controls to check for back-
ground absorbance caused by the secondary antibody binding
to TBDT.
12. If using TbpA, then in wells 12G and 12H, add 100 μL of the
1:1000 Tf-HRP solution described in Subheading 3.8, steps 1
and 2. These are the controls that ensure that the TbpA is
properly folded and coating the wells at similar levels.
13. Incubate at room temperature for 1 h, then wash with 250 μL/
well of PBST three times.
Add 50 μL/well of TMB and develop in the dark
(a cupboard/drawer is adequate) for 20 min, then add 25 μL
of 4 N HCl to stop the reaction. Measure the optical density at
450 nm.

4 Notes

1. 5 M is close to the upper limit of solubility of NaCl in water;

hence, when making the 5 M NaCl stock solution, in addition
to the use of vigorous stirring and heat, it is necessary to add
close to the full volume of water before all solids can be
dissolved.
134 Somshukla Chaudhuri et al.

2. The preference of our group is to use either free Ni-NTA resin

in a gravity column or a 5-mL affinity chromatography column
containing Ni-NTA resin (i.e., the 5-mL HisTrap High Perfor-
mance column from Cytiva) installed as part of an FPLC system
or attached to a peristaltic pump.
3. Use a centrifugation concentrator with a lower MWCO if the
protein of interest has a lower molecular weight than 50 kDa.
Since our protein antigens are usually ~60–70 kDa, the protein
concentrator most often used by our group is Cytiva’s Vivaspin
20 with a MWCO of 50 kDa. The MWCO should ideally be
just below the molecular weight of the desired protein antigen
to facilitate the removal of any lower molecular weight con-
taminants that may be present in the sample.
4. Our group prefers to use the online platform I-TASSER for
better models or more challenging projects (lower sequence
identity) (https://zhanglab.ccmb.med.umich.edu/I-
TASSER/) or use Phyre2 (http://www.sbg.bio.ic.ac.uk/
~phyre2/html/page.cgi?id¼index) for more rapid results in
protein structure prediction/modelling. PyMOL is the pre-
ferred program for visualizing the models, and either Gene
Construction Kit (http://www.textco.com/gene-construc
tion-kit.php) or SnapGene (https://www.snapgene.com/) is
used for in silico cloning.
5. We use a commercially available Elugent detergent (50% Solu-
tion) that is readily available from a variety of commercial
suppliers. We have only used this method for TbpA and LbpA
that enabled us to monitor isolation of functional protein by a
solid-phase binding assay. However, this method should work
effectively for other TBDTs.
6. We recommend selecting multiple loops of the TBDT to dis-
play on the scaffold as there is no empirical method for deter-
mining yet which loop will successfully fold with the scaffold
and be sufficiently immunogenic and protective. For the scaf-
fold, loops that are large can be selected to remove and then
replace with a loop from a TBDT. We believe such positions are
likely to accommodate TBDT loops of various sizes. We sug-
gest designing and synthesizing various single and multiple
loop hybrids using different loops in different positions initi-
ally. The number of hybrids can be narrowed down based on
which constructs yield stable proteins for immunizations. Out
of those hybrids, the one(s) that elicit a protective immune
response can be used for future applications.
7. The custom expression vector pE5770 is available from
Addgene (www.addgene.org) but any commercial or other
available plasmids with similar components would be suitable.
 Design and Production of Hybrid Antigens 135

The vector contains the T7 phage promoter for high level

expression by the T7 RNA polymerase. It also contains an
N-terminal polyhistidine (6 His residues) tag to allow for the
purification of the fusion protein by Ni-NTA chromatography,
a biotin acceptor peptide to facilitate the in vivo biotinylation
by the E. coli biotin ligase BirA, MBP to promote the solubility
and proper folding of the fusion protein and a cleavage site
recognized by TEV protease to allow for the removal of the
aforementioned components in the N-terminal fusion tag. The
pE5770 plasmid has BamH1 and Xba1/HindIII restriction
sites flanking a kanamycin resistance cassette that can be used
to confirm successful cloning of the gene of interest (absence of
kanamycin resistance). This expression system requires an
E. coli strain (e.g., ER2566) that contains a chromosomal
copy of the gene encoding T7 RNA polymerase under the
control of the lac promoter and preferably is deficient in both
lon and ompT proteases. In the presence of glucose, catabolite
repression inhibits lactose transport so that expression of T7
RNA polymerase is inhibited by binding of the lacI repressor to
the lac promoter. When the glucose in the autoinduction
media is depleted, lactose transport is no longer inhibited,
leading to activation of T7 RNA polymerase expression, and
thereby triggers the expression of the protein of interest
encoded by the T7 expression vector (Fig. 3). The rise in
cAMP levels upon glucose depletion also increases expression
by a cAMP-CAP complex binding to the CAP site upstream of
the lac promoter.
8. After the addition of plasmid DNA into the tube containing
chemically competent cells, we recommend flicking the tube
gently to mix well.
9. Place the LB medium aliquot on ice and set the water bath/
heat block at 42  C immediately after starting the 30-min
incubation on ice to ensure that the correct temperature is
reached in time for the subsequent step.
10. Our preference is to use the E. coli strain ER2566; however,
any strain that fits the criteria outlined in Note 7 would suffice.
11. Due to problems encountered with strains containing the T7
polymerase gene, the transformed mixture is used directly for
expression rather than initial plating and selection of colonies
for expression experiments. An autoinduction medium with-
out any antibiotics can be prepared beforehand in 4-L flasks
and 1-L bottles and stored at room temperature for up to
1 month for use. For making 1-L of medium add 1 mL of
MgSO4, 50 mL of 20
NPS, and 20 mL of 50
5052 to
929 mL of sterile ZY media in a sterile 1-L bottle.
136 Somshukla Chaudhuri et al.

12. This step is not always necessary; however, we find that pre-
warming may enhance bacterial growth and protein yields
compared to inoculation at room temperature.
13. The yield of protein after growth at 37  C appears to predict
whether an additional period of growth at low temperature will
lead to an increased yield. The expression plasmids with lower
yields after growth at 37  C (perhaps unstable or prone to
degradation) do not achieve enhanced yields with an additional
incubation period at low temperature.
14. For efficient resuspension of the cell pellets, scoop out each cell
pellet using a spatula and transfer it to the beaker containing
the Resuspension Buffer mixture. Then, transfer ~20 mL of the
Resuspension Buffer mixture into each centrifugation bottle
using a serological pipettor to resuspend any remaining bacte-
ria. Finally, transfer each of the resulting cell suspensions to the
beaker containing the cell pellets and the remaining Resuspen-
sion Buffer mixture.
15. Perform four discrete passes through the homogenizer and
avoid recirculating the sample as it results in uneven lysis—
that is, collect the lysate in a beaker after each independent pass
through the homogenizer instead of continuously directing
the sample back through the homogenizer.
16. Use appropriate size filtration devices for filtering the superna-
tant. When using syringe filters, limit the volume to 25–50 mL
of lysate, depending on viscosity. If the sample cannot be
filtered easily using a 0.2-μm syringe filter, use a 0.45-μm filter
first before passing the lysate through the 0.2-μm filter. If the
lysate still cannot be easily filtered, an additional 200 μL of
DNase I (5 mg/mL) can be added and mixed in with the lysate
for 30 min using a magnetic stir bar prior to filtration.
17. If the protein antigen of interest is readily detected in the
supernatant, this suggests that the recombinant protein is suf-
ficiently soluble to allow for the purification of the protein
using the protocol outlined in Subheading 3.3. If the SDS-
PAGE gel indicates the absence of the protein in the superna-
tant, or “soluble fraction,” this suggests that constitutive
expression of the protein in the cytoplasm of E. coli has resulted
in aggregation and misfolding of the protein. Thus, due to
the apparent insolubility resulting from improper folding of
the protein of interest, the protein is likely present alongside
the cell debris in the pellet following centrifugation (Subhead-
ing 3.2, step 12). In this circumstance we recommend restart-
ing the protein production run (beginning at the start of
Subheading 3.2) and proceeding directly to Subheading 3.4
immediately following Subheading 3.2, step 8.
 Design and Production of Hybrid Antigens 137

18. A 5-mL column is ideal for culture sizes of 6 L or less. If larger

than 6 L, use two 5-mL columns in tandem. Our preference is
to use the 5-mL HisTrap High Performance column from
Cytiva. If a chromatography system is not available, the purifi-
cation can be performed with Ni-NTA resin in a gravity fed
column, by applying wash and elution buffers manually.
19. Circulation can be done at room temperature if 0.02% sodium
azide is included in the lysate to prevent any microbial growth.
For continuous circulation overnight, set up the column and
any attached tubing such that the sample exiting the column is
collected the same receptacle as the sample entering the
column.
20. In lieu of commercial TEV protease we routinely use a TEV
protease with an N-terminal polyhistidine tag produced
in-house in E. coli using a T7 expression vector. When
expressed with a polyhistidine tag, it can be purified using a
Ni-NTA column as described in Subheading 3.2, concentrated
to a final concentration of 1–10 mg/mL, mixed with sterile
100% glycerol to a final concentration of 50%, and stored at
20  C. This glycerol mixture—typically containing ~5 mg of
TEV protease for up to 100 mg of fusion protein—can then be
added directly to the dialysis tubing containing the entirety of
the protein antigen sample as part of the TEV cleavage step.
Cleavage of the fusion protein by TEV protease is carried out
efficiently during the overnight (16–18 h) incubation period at
4  C—there is no need to extract the protein sample from the
dialysis tubing for the cleavage step.
21. A single circulation would suffice and would reduce the risk of
losing any uncleaved fusion protein in the process. If just
circulating once, set up the column and any attached tubing
such that the sample exiting the column is collected in a sepa-
rate receptacle from the sample entering the column. Then,
circulate the sample until the entire sample has passed through
the column.
22. Our experience in preparation of antigens by these methods
have not resulted in situations where lipopolysaccharide (LPS)
toxicity was an issue or where levels measured by the limulus
amoebocyte lysate coagulation (LAL) assay were of concern.
Thus, aliquots were normally promptly stored at 80  C;
however, it may be prudent to test protein preparations with
the LAL assay when using new strains or new proteins to
ensure LPS removal is not required. If endotoxin removal is
deemed necessary, the methods our group prefers include the
MonoQ column from Cytiva, the CHT I or CHT II columns
from Bio-Rad, or the Pierce High Capacity Endotoxin
Removal Spin Columns.
138 Somshukla Chaudhuri et al.

23. Before adding 3 mL of Ni-NTA resin, wash the resin to elimi-

nate the ethanol, and use 3 mL of Denaturing Lysis Buffer to
equilibrate the resin.
24. Depending on the protein purification yield, a greater volume
of Refolding Elution Buffer may be needed. Verify the protein
concentration at the end of the elution step to confirm that all
of the column-bound protein has been eluted.
25. As it is possible to assay SBP-TbpA for proper folding using the
transferrin binding assay described in Subheading 3.8, we rec-
ommend cloning an SBP-TbpA fusion construct alongside
cloning of the desired target SBP-TBDT construct (Subhead-
ing 3.5) as a control for steps in Subheadings 3.6 and 3.7.
26. The custom expression vector pE5771 is available from
Addgene (www.addgene.org). Vector pE5771 was designed
by our group for expression of TBDTs such as TbpA and
LbpA. It has a kanamycin resistance cassette in the expression
locus between BamH1 and XbaI sites so that colonies gener-
ated from cloning the gene of interest into the expression locus
can be screened on plates containing kanamycin to confirm
insertion (no growth on kanamycin-containing plates). This
plasmid encodes the pelB signal sequence that results in sub-
stantial levels of periplasmic proteins being secreted into the
periplasm of E. coli protein expression strains such as BL21.
The signal sequence is followed by four amino acids preceding
a polyhistidine (6
His) tag, a TEV protease cleavage site, SBP,
and a second TEV cleavage site, upstream of the mature tbdt
sequence (Fig. 6). This plasmid can also be used for large-scale
TBDT protein preparations, where detergent-extracted TBDT
membrane extracts can be purified using a Ni-NTA column
and cleaved with TEV protease to generate purified TBDT
protein.
27. We recommend using a cell line like E. coli C43, a derivative of
the commonly used BL21 expression cell line. This cell line has
the wild-type lac promoter in the chromosome instead of the
mutated lacUV5 promoter present in the parent. The result is
reduced expression of the T7 RNA polymerase and subsequent
reduction in expression of the target gene. Cell lines like C43
or the similar C41, also known as the Walker cell lines, have
been shown to be more appropriate for expression of mem-
brane proteins and toxins. Chemically competent cells can be
premade and stored at 80  C in 50- or 100-μL aliquots in
1.5-mL Eppendorf tubes until use. If the intention is to screen
antisera for reactivity against different variants of the TBDT, it
is advised to consider performing the procedure in Subheading
3.6 in parallel with plasmids encoding the different variants.
 Design and Production of Hybrid Antigens 139

28. We typically remove the cell–Tris–Elugent mixture from the

50 mL conical tube and transfer it into a 2-mL Eppendorf tube
and use a rotary shaker in the refrigerator overnight. A specific
rpm is not required as long as it is gently shaking overnight.
29. We recommend using the TBDT–PBST mixture for coating
ELISA plates the same day as we have not evaluated the impact
of storage of the mixture with TbpA which would have enabled
us to evaluate maintenance of native conformation by a binding
assay.
30. The proportions described in this protocol amount to just
enough TBDT for one 96-well ELISA plate coated with 100
μL/well of TBDT-containing solution. This can be scaled up
or down based on how many plates or wells the user needs. We
recommend making more than what is required in case some
protein is lost in the process of coating the plates.
31. When using TbpA as the TBDT, either on its own or as a
control for correct conformation of the detergent extracted
TBDT, it is recommended to first perform a separate ELISA
to test for correct conformation of TbpA prior to assessing
antibody titers from immunized animals. For this conforma-
tion ELISA, only one column of a streptavidin-coated plate is
needed, adding SBP-TbpA/PBST to only two wells (but
blocking all wells in the column) following the instructions in
Subheading 3.7, steps 1–3, and then proceeding to Subhead-
ing 3.8. If a weak signal is observed in this assay, anti-TBDT
titres assessed in a subsequent ELISA should be interpreted
with caution as the TBDT used to coat the plate may be
misfolded. If a ligand known to bind to the TBDT under
investigation is identified, this TBDT-ligand interaction can
be exploited to give the user more confidence that the TBDT
in question is properly folded (see Note 34).
32. When an entire plate is not needed, (e.g., when determining
the ability of the TbpA to bind transferrin, as described in
Subheading 3.8, see Note 31) only coat the wells to be used.
Cover the unused portion of the plate, which can be used for a
later experiment, with an adhesive slip. We typically use
streptavidin-coated Greiner Bio-One plates in our experi-
ments; however, other commercially available streptavidin/
neutravidin-coated plates can also be used. Plates can also be
coated with streptavidin or neutravidin in-house [14].
33. To wash, discard the PBST in the wells by inverting the 96-well
plate over the sink and expelling the liquid with a sharp flick of
the wrist. Following this, vigorously tap the plate face-down on
a stack of paper towels to remove the remaining liquid. An
automated plate washer may also be used for these steps, if
available.
140 Somshukla Chaudhuri et al.

34. The TbpA/Tf-HRP ELISA is used to assess whether the TbpA

is properly folded on the plate. For TBDTs that bind metal-
containing proteins (hemoglobin, hemoglobin–haptoglobin,
calprotectin), a similar approach could be used, with an addi-
tional step to add appropriate target protein, followed by HRP-
conjugated antibody specific for the target protein; however,
alternative approaches will be required for TBDTs that directly
bind metal ions or metal ions bound by compounds (heme,
siderophore–iron complexes).
35. We recommend performing twofold serial dilutions across the
columns labeled 1–11, leaving the last column for any perti-
nent controls; however, if no controls are needed, up to 12 dif-
ferent dilutions can be used.

References

1. Fegan JE, Calmettes C, Islam EA, Ahn SK, factors in Actinobacillus pleuropneumoniae
Chaudhuri S, Yu RH, Gray-Owen SD, Moraes serotype 7 infection. FEMS Microbiol Lett
TF, Schryvers AB (2019) Utility of hybrid 209(2):283–287
transferrin binding protein antigens for protec- 9. Anderson JE, Hobbs MM, Biswas GD, Spar-
tion against pathogenic Neisseria species. Front ling PF (2003) Opposing selective forces for
Immunol 10:247. https://doi.org/10.3389/ expression of the gonococcal lactoferrin recep-
fimmu.2019.00247 tor. Mol Microbiol 48(5):1325–1337. https://
2. Qamsari MM, Rasooli I, Chaudhuri S, Astaneh doi.org/10.1046/j.1365-2958.2003.
SDA, Schryvers AB (2020) Hybrid antigens 03496.x
expressing surface loops of ZnuD from Acine- 10. Hooda Y, Lai CCL, Moraes TF (2017) Identi-
tobacter baumannii is capable of inducing pro- fication of a largefamily of Slam-dependent sur-
tection against infection. Front Immunol face lipoproteins in gram-negative bacteria.
11:1–13. https://doi.org/10.3389/fimmu. Front Cell Infect Microbiol 7:207. https://
2020.00158 doi.org/10.3389/fcimb.2017.00207
3. Noinaj N, Guillier M, Barnard TJ, Buchanan 11. Frandoloso R, Martinez-Martinez S,
SK (2010) TonB-dependent transporters: reg- Calmettes C, Fegan J, Costa E, Curran D,
ulation, structure, and function. Annu Rev Yu R, Gutierrez-Martin CB, Rodriguez-Ferri
Microbiol 64:43–60. https://doi.org/10. EF, Moraes TF, Schryvers AB (2015) Non-
1146/annurev.micro.112408.134247 binding site-directed mutants of transferrin
4. Postle K, Kadner RJ (2003) Touch and go: binding protein B enhances their immunoge-
tying TonB to transport. Mol Microbiol 49 nicity and protective capabilities. Infect Immun
(4):869–882 83(3):1030–1038
5. Hood MI, Skaar EP (2012) Nutritional immu- 12. Calmettes C, Alcantara J, Yu R-H, Schryvers
nity: transition metals at the pathogen-host AB, Moraes TF (2012) The structural basis of
interface. Nat Rev Microbiol 10(8):525–537. transferrin sequestration by transferrin binding
https://doi.org/10.1038/nrmicro2836 protein B. Nat Struct Mol Biol 19
6. Morgenthau A, Pogoutse A, Adamiak P, Moraes (3):358–360. https://doi.org/10.1038/
TF, Schryvers AB (2013) Bacterial receptors for nsmb.2251
host transferrin and lactoferrin: molecular 13. Noinaj N, Buchanan SK, Cornelissen CN
mechanisms and role in host-microbe interac- (2012) The transferrin-iron import system
tions. Future Microbiol 8(12):1575–1585 from pathogenic Neisseria species. Mol Micro-
7. Cornelissen CN, Kelley M, Hobbs MM, biol 86(2):246–257. https://doi.org/10.
Anderson JE, Cannon JG, Cohen MS, Sparling 1111/mmi.12002
PF (1998) The transferrin receptor expressed 14. Fegan JE, Yu R-H, Islam EA, Schryvers A
by gonococcal strain FA1090 is required for (2021) Development of a non-biased, high-
the experimental infection of human male throughput ELISA for the rapid evaluation of
volunteers. Mol Microbiol 27(3):611–616 immunogenicity and cross-reactivity. J Immu-
8. Baltes N, Hennig-Pauka I, Gerlach GF (2002) nol Methods 493:113027
Both transferrin binding proteins are virulence
 Chapter 9

Preparation of Trimethyl Chitosan-Based Polyelectrolyte

Complexes for Peptide Subunit Vaccine Delivery
Lili Zhao, Sahra Bashiri, Istvan Toth, and Mariusz Skwarczynski

Abstract
A variety of delivery vehicles have been explored as adjuvanting/delivery platforms for peptide-based
subunit vaccines. Polysaccharide-based systems have been found to be especially attractive due to their
immune stimulating properties, biodegradability, biocompatibility, and low toxicity. Among them, chitosan
and its derivatives are the most common cationic nanocarriers used for the delivery of antigens. Trimethyl
chitosan (TMC) is a partially quaternized, water-soluble, and mucoadhesive derivative of chitosan. This
chapter describes the preparation of a TMC-based polyelectrolyte complex as a delivery system for peptide
subunit vaccines.

Key words Trimethyl chitosan, Alginate, Polyelectrolyte complexes, Peptide subunit vaccine,
Nanoparticles

1 Introduction

Peptide subunit vaccines, which contain only the minimum neces-

sary components required to stimulate immune responses, have
been extensively investigated over the past few decades. Since pep-
tides are mostly nonimmunogenic on their own, employing an
adjuvant and/or a delivery system is vital to the efficacy of peptide
vaccines [1, 2]. Delivery systems, such as lipid-based formulations
[3], polymers [4], dendrimers [5], and inorganic nanoparticles [6],
have been widely explored for peptide subunit vaccine delivery.
Efficient delivery systems enable long-lasting release of peptide
antigen through a depot effect or enhanced trafficking to the
lymph nodes through the lymphatic system [7, 8]; targeting anti-
gen presenting cells [9, 10], enhanced mucosa adhesion and mucus
penetration of peptide antigen at mucosa sites [11, 12]; antigen
protection from degradation [13]; and ultimately, stimulation of
the desired immune responses.

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_9,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

141
142 Lili Zhao et al.

Fig. 1 Schematic illustration of the preparation of the LCP-1-loaded polyelectrolyte complex: LCP-1/alginate/
TMC. (a) Primary complex (LCP-1/alginate), (b) Ternary PEC (LCP-1/alginate/TMC)

Polysaccharides composed of repeating monosaccharide units

have been widely used for vaccine delivery due to their biodegrad-
ability, biocompatibility, and low toxicity [14, 15]. They can form
polyelectrolyte complexes (PECs) with counterparts possessing
opposite charges through electrostatic interactions
[16, 17]. Among them, chitosan and its derivatives are the most
common cationic nanocarriers used for the delivery of antigens
[16, 18]. Since chitosan has very poor water solubility, its soluble
derivative, trimethyl chitosan (TMC) is typically used as a compo-
nent of self-adjuvanting delivery systems [19, 20]. TMC is partially
quaternized, water-soluble, and mucoadhesive.
Various methods have been employed for the preparation of
PECs, such as polyelectrolyte titration, jet mixing, and ionic gela-
tion [21]. Titration is usually favored, as it is simple and proceeds
under mild conditions [17]. This protocol describes the prepara-
tion of PECs comprised of peptide antigen (LCP-1), negatively
charged alginate, and positively charged TMC (Fig. 1) [22]. The
protocol includes two parts: (1) synthesis of TMC; and (2) formu-
lation of LCP-1/alginate/TMC PECs.

2 Materials

2.1 Synthesis of TMC 1. Low molecular weight chitosan (75–85% deacetylated).

2. Sodium iodide.
3. N-Methyl-2-pyrrolidone (NMP).
4. Aqueous sodium hydroxide solution, 15% (w/w).
5. Diethyl ether and ethanol mixture (50:50 v/v).
6. Sodium hydroxide pellets.
7. Aqueous sodium chloride (NaCl) solution, 10% (w/v).
8. Deuterium water (D2O).
9. Acetone.
 Chitosan-Based Nanovaccine 143

2.2 Formulation and 1. Stock solution of alginate sodium in Milli-Q water (2 mg/mL).
Characterization of 2. Stock solution of trimethyl chitosan (TMC) in Milli-Q water
PECs (LCP-1/Alginate/ (2 mg/mL).
TMC)
3. Stock solution of LCP-1 in Milli-Q water (1 mg/mL).
4. Phosphotungstic acid (0.5% w/v).

2.3 Equipment 1. Magnetic stirrer with hot plate, magnetic stir bar, and ice bath.
2. Thermometer.
3. Condenser.
4. Freeze dryer.
5. Nuclear magnetic resonance (NMR) spectrometer (300 MHz
or higher).
6. Probe-type sonicator.
7. Malvern Zetasizer Nano-ZS (Malvern Instruments Ltd., UK or
equivalent) to measure dynamic light scattering (DLS).
8. Transmission electron microscope.
9. Dialysis bag (3500 Da).
10. Balance, microbalance.
11. Laboratory glassware and apparatus.

3 Methods

This titration method was developed for the preparation of

LCP-1/alginate/TMC PECs; however, it can also be used for
other PEC nanoparticles (e.g., LCP-1/chondroitin sulfate/TMC,
LCP-1/dextran/TMC, LCP-1/hyaluronic acid/TMC, and
LCP-1/heparin/TMC [22]). Although, different quantities of
charged polymers are required to formulate the specific PEC
nanoparticles.

3.1 Synthesis of TMC 1. To 250 mL round bottom flask add low molecular weight
chitosan (1 g) and sodium iodide (2.4 g), add 40 mL of NMP.
2. Stir the mixture until fully dissolved.
3. Add 6 mL of 15% aqueous NaOH solution and stir the mixture
for 20 min. The temperature should be kept at 60
C (see
Note 1).
4. Add 6 mL of the methylation agent, methyl iodide, to the
reaction mixture. Attach a reflux condenser to the flask and
stir the mixture for 1 h at 60
C.
5. Slowly pour the reaction mixture into 50 mL of diethyl ether–
ethanol solution to precipitate the product (see Note 2).
144 Lili Zhao et al.

6. Filter the precipitate using a Buchner funnel under vacuum.

Wash the product with diethyl ether (2  5 mL washes) (see
Note 3).
7. Add 2.4 g of sodium iodide to the 250 mL round bottom flask
containing the product from step 6, then add 40 mL of NMP.
Attach a reflux condenser to the flask (see Note 4).
8. Stir the mixture at 60
C until fully dissolved.
9. Add 5.5 mL of 15% aqueous NaOH to the solution and stir the
mixture for 20 min.
10. Add 3.5 mL of methyl iodide to the reaction mixture and stir
for 45 min at 60
C.
11. Add another 1 mL of methyl iodide, and 0.6 g of NaOH pellets
to the mixture (see Note 5).
12. Stir the mixture for 1 h at 60
C.
13. Repeat steps 5 and 6.
14. Transfer the resultant precipitate into the 100 mL round bot-
tom flask and add 50 mL of 10% (w/v) NaCl. Stir for at least
18 h to replace I ions with Cl ions.
15. Transfer the solution into the dialysis bag (3500 Da) and
dialyze it at room temperature against deionized water for
3 days, changing the water at least two times a day.
16. Freeze the purified solution in a dry ice-acetone cold bath, then
lyophilize the frozen compound using a freeze dryer.
17. Analyze the product by NMR spectroscopy. Typical spectra are
as follows: (D2O, 300 MHz): ppm 5.25–5.70 (m, H(1)),
3.55–4.75 (m, H(2), H(3), H(4), H(5), 2H(6)), 3.43 (s,
OCH3), 3.35 (s, OCH3), 3.25 (bs, N-(CH3)3), 2.39 (bs,
N-(CH3)2), and 1.98 (bs, C(O)CH3). The degree of quater-
nization (DQ), degree of demethylation (DM), degree of 3-
and 6-O-methylation (DOM-3 and DOM-6, respectively), and
degree of acetylation (DA) of TMC were calculated from 1H
NMR spectra, according to the following equations [23, 24]:


DQ% ¼ ðCH3 Þ3 = 9  Hð1Þ  100:


DM% ¼ ðCH3 Þ2 = 6  Hð1Þ  100:


DOM  3% ¼ ðOCH3 Þ= 3  Hð1Þ  100:


DOM  6% ¼ ðOCH3 Þ= 3  Hð1Þ  100:


DA% ¼ ðCðOÞCH3 Þ= 3  Hð1Þ  100:
(CH3)3, (CH3)2, (OCH3), and (C(O)CH3) are the inte-
grals of the hydrogens of the trimethylated amino groups at
3.25 ppm, the demethylated amino groups at 2.39 ppm, the
methylated hydroxyl groups at either 3.35 (DOM-6) or 3.43
 Chitosan-Based Nanovaccine 145

Fig. 2 1H NMR spectra of trimethyl chitosan

(DOM-3) ppm, and the acetylated groups at 1.98 ppm, respec-

tively. H(1) is the integral of peaks between 5.25 and 5.70 ppm:
the signal related to the carbohydrate hydrogen atom bound to
the C-1’s of the TMC molecule (Fig. 2).

3.2 Preparation of 1. Dilute the alginate stock solution (2 mg/mL) to 0.1, 0.2, 0.4,
PECs (LCP-1/Alginate/ 0.6, 0.8, and 1.0 mg/mL, with Milli-Q water (see Note 6).
TMC) 2. Transfer 100 μL of LCP-1 solution (100 μg) into a 2 mL
3.2.1 Prepare the centrifuge tube; dilute 10 times with Milli-Q water.
Primary Complexes (LCP-1/ 3. Add 50 μL of the diluted alginate solution (0.1 mg/mL, con-
Alginate) taining 5 μg alginate) into the centrifuge tube and mix
thoroughly.
4. Sonicate the mixture for 4 min (2  2 min) using a probe-type
sonicator at 120 W (duration 2 min, duty cycle 50%) in an
ice bath.
5. Transfer the mixture to a scintillation vial and stir it continu-
ously for 1 h at room temperature to stabilize the complexes.
6. Measure the particle size, polydispersity index (PDI), and zeta
potential using dynamic light scattering (DLS).
146 Lili Zhao et al.

Fig. 3 Optimization of the ratio of LCP-1 to alginate for the preparation of primary complexes (LCP-1/alginate):
the optimum ratio is indicated by white shading. (a) The bars show size (nm), while the line represents
polydispersity index (PDI). (b) The bars show zeta potential (mV). LCP-1/alginate ¼ 10:4, size 177  2 nm, PDI
0.205  0.01, zeta potential 37.1  0.7 mV

7. Repeat steps 3–6 five times, except in step 3, add 50 μL of

alginate solution at concentrations of 0.2, 0.4, 0.6, 0.8, and
1.0 mg/mL (instead of 50 μL of the diluted alginate solution)
to prepare primary complexes with the mixing ratios of LCP-1:
alginate of 10:0.5, 10:1, 10:2, 10:3, 10:4, and 10:5, respec-
tively (Fig. 3) (see Note 7).

3.2.2 Preparation of 1. Dilute TMC stock solution (2 mg/mL) to 0.6, 1.0, 1.2, and
Ternary PECs (LCP-1/ 1.6 mg/mL with Milli-Q water.
Alginate/TMC) 2. Add 50 μL of TMC solution (0.6 mg/mL, containing 30 μg
TMC) dropwise to freshly prepared LCP-1/alginate complex
suspension (containing 100 μg of LCP-1, 40 μg of alginate).
3. Stir the solution for 1 h at room temperature to produce the
final PEC nanoparticles.
4. Measure the particle size, PDI, and zeta potential using DLS.
5. Repeat steps 2–4 five times, except at step 2, add 50 μL of
TMC solution at concentrations of 0.6, 1.0, 1.2, 1.6, and
2.0 mg/mL (instead of 50 μL of the diluted TMC solution)
to produce ternary PEC nanoparticles at ratios of 10:4:3,
10:4:5, 10:4:6, 10:4:8, and 10:4:10, respectively (see Note 8).
6. Measure the particle size, PDI, and zeta potential of the pri-
mary complexes and final ternary PEC nanoparticles using
DLS. The expected particle size and zeta potential for different
concentration are shown in Fig. 4 (see Note 9).
7. Examine the morphology of the PEC nanoparticles using
transmission electron microscopy. Drop PEC nanoparticles
(dilution might be required) onto a carbon-coated copper
grid and leave it for 1 min to allow the nanoparticles to deposit
 Chitosan-Based Nanovaccine 147

Fig. 4 Optimization of PEC nanoparticle (LCP-1/alginate/TMC) preparation using DLS monitoring: the optimum
mixing ratio is indicated by white shading. (a) The bars show size (nm), while the line represents PDI. (b) The
bars show zeta potential (mV). LCP-1/alginate/TMC ¼ 10:4:8, size 237  4 nm, PDI 0.199  0.01, zeta
potential 29.3  1.5 mV

Fig. 5 Transmission electron microscopy images of LCP-1-loaded PEC

nanoparticles (LCP-1/alginate/TMC)

onto the grid. Then, drop 0.5% (w/v) phosphotungstic acid

onto the grid and leave it for 30 s. Air dry the grids before
observation. The individual particles may be directly observed
and measured by TEM (Fig. 5).
148 Lili Zhao et al.

4 Notes

1. To increase the temperature of the solution to 60
C, use a hot

plate and silicone oil bath. Assemble the thermometer into the
oil bath using the holder. Fix the round bottom flask onto the
oil bath using a retort stand.
2. The temperature of the diethyl ether–methanol solution
should be low (~4
C) to improve precipitation/isolation of
the product. The solution should be precooled in the refriger-
ator or immersed in an ice bath.
3. The chitosan derivative will precipitate in diethyl ether–ethanol
solution, so should be filtered immediately as the precipitate
tends to aggregate in larger lumps as time goes on. At this
stage, the TMC degree of quaternization is around 20% [25].
4. Follow steps 7–13 to synthesize TMC with a DQ of 60–90%.
5. This step helps to produce TMC with a higher DQ (80–90%).
6. It takes a few hours, to overnight, for polymers to swell and
dissolve in water.
7. The optimum mixing ratio of LCP-1 to alginate is determined
as the minimal amount of alginate required to produce uniform
(PDI < 0.3), small-sized nanoparticles (<200 nm) with a
surface charge of around 30 mV, and where further addition
of alginate does not decrease the charge. The ratio of LCP-1:
alginate of 10:4 was used to prepare the primary complexes.
8. PEC nanoparticles of higher concentration can be produced;
however, they may require longer stirring time to allow poly-
mer complexation and particle stabilization.
9. According to the change in charge, the ratio of LCP-1, algi-
nate, and TMC should be 10:4:10. However, as particle size
and PDI both increased at this ratio, we instead determined
10:4:8 to be the optimum ratio.

References
1. Nevagi RJ, Toth I, Skwarczynski M (2018) 5. Fujita Y, Taguchi H (2011) Current status of
Peptide-based vaccines. In: Koutsopoulos S multiple antigen-presenting peptide vaccine
(ed) Peptide applications in biomedicine, bio- systems: application of organic and inorganic
technology and bioengineering. Woodhead nanoparticles. Chem Cent J 5(1):48
Publishing, Oxford 6. Skwarczynski M, Toth I (2014) Recent
2. Skwarczynski M, Toth I (2016) Peptide-based advances in peptide-based subunit nanovac-
synthetic vaccines. Chem Sci 7(2):842–854 cines. Nanomedicine 9:2657–2669
3. Bartlett S, Skwarczynski M, Toth I (2020) 7. Reddy ST, Van Der Vlies AJ, Simeoni E et al
Lipids as activators of innate immunity in pep- (2007) Exploiting lymphatic transport and
tide vaccine delivery. Curr Med Chem 27 complement activation in nanoparticle vac-
(17):2887–2901 cines. Nat Biotechnol 25(10):1159–1164
4. Nevagi RJ, Skwarczynski M, Toth I (2019) 8. Henriksen-Lacey M, Christensen D, Bramwell
Polymers for subunit vaccine delivery. Eur VW et al (2011) Comparison of the depot
Polym J 114:397–410 effect and immunogenicity of liposomes based
 Chitosan-Based Nanovaccine 149

on dimethyldioctadecylammonium (DDA), 18. Li X, Min M, Du N et al (2013) Chitin, chit-

3β-[N-(N’,N’-Dimethylaminoethane)carbo- osan, and glycated chitosan regulate immune
myl] cholesterol (DC-Chol), and 1,2-Dio- responses: the novel adjuvants for cancer vac-
leoyl-3-trimethylammonium propane cine. Clin Dev Immunol 2013:387023
(DOTAP): prolonged liposome retention. 19. Nevagi RJ, Dai W, Khalil ZG et al (2019)
Mol Pharm 8(1):153–161 Structure-activity relationship of group a strep-
9. Yang JR, Luo YC, Shibu MA et al (2019) Cell- tococcus lipopeptide vaccine candidates in tri-
penetrating peptides: efficient vectors for vac- methyl chitosan-based self-adjuvanting
cine delivery. Curr Drug Deliv 16(5):430–443 delivery system. Eur J Med Chem
10. Hussein WM, Liu TY, Skwarczynski M et al 179:100–108
(2014) Toll-like receptor agonists: a patent 20. Marasini N, Ghaffar KA, Giddam AK et al
review (2011 - 2013). Expert Opin Ther Pat (2017) Highly immunogenic trimethyl
24(4):453–470 chitosan-based delivery system for intranasal
11. Zheng J, Gao S, Cui X et al (2019) Adjuvants lipopeptide vaccines against Group A Strepto-
and delivery systems based on polymeric nano- coccus. Curr Drug Deliv 14(5):701–708
particles for mucosal vaccines. Int J Pharm 21. Jeong Y, Lee D, Choe K et al (2017)
572:118731 Polypeptide-based polyelectrolyte complexes
12. Marasini N, Skwarczynski M, Toth I (2017) overcoming the biological barriers of oral insu-
Intranasal delivery of nanoparticle-based vac- lin delivery. J Indust Eng Chem 48:79–87
cines. Ther Deliv 8(3):151–167 22. Zhao L, Jin W, Cruz JG et al (2020) Develop-
13. Malonis RJ, Lai JR, Vergnolle O (2020) ment of polyelectrolyte complexes for the
Peptide-based vaccines: current progress and delivery of peptide-based subunit vaccines
future challenges. Chem Rev 120 against group A Streptococcus. Nanomaterials
(6):3210–3229 (Basel) 10:823
14. Sedaghat B, Stephenson R, Toth I (2014) Tar- 23. Marasini N, Giddam AK, Ghaffar KA et al
geting the mannose receptor with mannosy- (2016) Multilayer engineered nanoliposomes
lated subunit vaccines. Curr Med Chem 21 as a novel tool for oral delivery of lipopeptide-
(30):3405–3418 based vaccines against group A Streptococcus.
15. Li P, Wang F (2015) Polysaccharides: candi- Nanomedicine 11(10):1223–1236
dates of promising vaccine adjuvants. Drug 24. Polnok A, Borchard G, Verhoef JC et al (2004)
Discov Ther 9(2):88–93 Influence of methylation process on the degree
16. Zhao L, Skwarczynski M, Toth I (2019) of quaternization of N-trimethyl chitosan chlo-
Polyelectrolyte-based platforms for the deliv- ride. Eur J Pharm Biopharm 57(1):77–83
ery of peptides and proteins. ACS Biomater 25. Verheul RJ, Amidi M, Van Der Wal S et al
Sci Eng 5(10):4937–4950 (2008) Synthesis, characterization and in vitro
17. Meka VS, Sing MKG, Pichika MR et al (2017) biological properties of O-methyl free N,N,N-
A comprehensive review on polyelectrolyte trimethylated chitosan. Biomaterials 29
complexes. Drug Discov Today 22 (27):3642–3649
(11):1697–1706
 Chapter 10

Multiepitope Fusion Antigen: MEFA, an Epitope-

and Structure-Based Vaccinology Platform for Multivalent
Vaccine Development
Siqi Li, Kuo Hao Lee, and Weiping Zhang

Abstract
Vaccines are regarded as the most cost-effective countermeasure against infectious diseases. One challenge
often affecting vaccine development is antigenic diversity or pathogen heterogeneity. Different strains
produce immunologically heterogeneous virulence factors, therefore an effective vaccine needs to induce
broad-spectrum host immunity to provide cross-protection. Recent advances in genomics and proteomics,
particularly computational biology and structural biology, establishes structural vaccinology and highlights
the feasibility of developing effective and precision vaccines. Here, we introduce the epitope- and structure-
based vaccinology platform multiepitope-fusion-antigen (MEFA), and provide instructions to generate
polyvalent MEFA immunogens for vaccine development. Conceptually, MEFA combines epitope vaccinol-
ogy and structural vaccinology to enable a protein immunogen to present heterogeneous antigenic domains
(epitopes) and to induce broadly protective immunity against different virulence factors, strains or diseases.
Methodologically, the MEFA platform first identifies a safe, structurally stable and strongly immunogenic
backbone protein and immunodominant (ideally neutralizing or protective) epitopes from heterogeneous
strains or virulence factors of interest. Then, assisted with protein modeling and molecule dynamic
simulation, MEFA integrates heterogeneous epitopes into a backbone protein via epitope substitution for
a polyvalent MEFA protein and mimics epitope native antigenicity. Finally, the MEFA protein is examined
for broad immunogenicity in animal immunization, and assessed for potential application for multivalent
vaccine development in preclinical studies.

Key words MEFA (multiepitope fusion antigen), Vaccinology platform, Multivalent vaccine, Epi-
topes, Structural vaccinology

1 Introduction

Infectious diseases are a leading cause of death and a major threat to

global health [1, 2]. Vaccines are considered the most cost-effective
countermeasure to reduce human mortality and morbidity, to
eliminate pandemic or epidemic risk, and to lower antibiotic use.
Success in vaccine and vaccination against infectious diseases
includes the eradication of smallpox and the control of polio,

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_10,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

151
152 Siqi Li et al.

measles, mumps, and diphtheria. However, major challenges

remain in developing vaccines for other diseases including human
immunodeficiency virus (HIV), diarrhea caused by viral, bacterial,
or parasitic pathogens, tuberculosis (TB) and malaria which con-
tinuously claim millions of human lives annually [3, 4]. While a lack
of clear understanding of host pathogen interactions and difficulties
in identifying immune biomarkers or immune correlates of protec-
tion are commonly major challenges in vaccine development, anti-
genic diversities and immunological heterogeneity among
pathogen strains or virulence factors have also severely hampered
development of effective vaccines against certain infectious dis-
eases. Strategies including conserved antigen and cocktail vaccines
have been attempted to overcome antigenic diversity or immuno-
logical heterogeneity, but they often encounter bottlenecks at
improving protective efficacy. Different from conventional
approaches, we introduce a new vaccinology platform called multi-
epitope fusion antigen, MEFA [5, 6], to construct polyvalent
immunogens and to develop broadly protective vaccines. Benefit-
ing from recent advance in genomics and proteomics, particularly
computational biology and structural biology, MEFA combines
epitope vaccinology and structural vaccinology concepts to gener-
ate epitope- and structure-based polyvalent protein immunogens
for developing cross protective vaccines against heterogeneous
strains, pathogens, or diseases [7].
The MEFA platform enables integration of foreign epitopes,
ideally protective or neutralizing epitopes (epitopes induce protec-
tive immune responses) of various pathogenic strains or virulence
factors, into a backbone protein and mimicking of epitope native
antigenicity, thus constructing a polyvalent immunogen for broadly
protective immunity. In principle, a protective epitope, if it retains
native antigenicity after being integrated into a backbone protein,
induces protective host immunity against the representing viru-
lence factor or strain. Therefore, an MEFA immunogen carrying
multiple protective epitopes from different strains serves as a poly-
valent antigen for a cross protective vaccine. By substituting back-
bone protein epitopes with protective epitopes (of heterogeneous
strains or virulence factors) of interest, we produce a chimeric
polyvalent immunogen (Fig. 1), and by applying protein modeling
and molecular dynamic simulation, we mimic epitope native anti-
genicity. Mechanistically, the MEFA vaccinology platform com-
bines novel in silico analyses with conventional in vivo empirical
studies to accelerate development of broadly protective multivalent
vaccines. First, an MEFA backbone immunogen and epitopes of
interest (from heterogeneous strains or virulence factors) are iden-
tified, and an MEFA immunogen is constructed in silico, followed
by synthesis and cloning of an MEFA gene into an expression
vector. Then, the MEFA protein is expressed, extracted and used
for animal immunization studies. Finally, the MEFA protein-
 MEFA for Multivalent Vaccines 153

Fig. 1 Illustration of constructing a polyvalent MEFA immunogen by using the MEFA vaccinology platform.
Surface-exposed epitopes of a backbone immunogen are substituted with protective or immunodominant
epitopes from heterogeneous virulence factors or strains of interest to construct an MEFA immunogen

induced broad-spectrum immune responses are measured, and

application of an MEFA immunogen for multivalent vaccine devel-
opment is evaluated in preclinical investigation prior to human
subject safety, immunogenicity, and efficacy studies.

2 Materials

2.1 Computation MEFA backbone protein selection, epitope identification, and pro-
and Programs tein modeling are carried out on standard desktop or laptop com-
puters through open-access servers. Atomic molecular dynamics
simulation requires a local server and remote access to license-
based CHARMM server or open-source GROMACS-5.0.7 MD
server.
Programs or servers that are applied for in silico analyses are as
follows.
1. IEDB (http://www.iedb.org/) or BepiPred (http://www.cbs.
dtu.dk/services/BepiPred/).
2. ExPASy (https://www.expasy.org/).
3. PyMol (https://www.pymol.org/).
4. Phyre2 (http://www.sbg.bio.ic.ac.uk/~phyre2/html/page.
cgi?id¼index).
5. GROMACS (http://www.gromacs.org/).
6. Or CHARMM (https://www.charmm.org/), and MMTSB
toolset (http://blue11.bch.msu.edu/mmtsb/Main_Page).

2.2 A Backbone The MEFA platform requires a backbone protein and epitopes of
Protein heterogeneous strains or virulence factors of interest to construct
and Heterogeneous an MEFA immunogen. A backbone protein should be safe and
Epitopes of Interest structurally stable, preferably a mutant key virulence factor which
becomes avirulent after epitope substitution and induces protective
154 Siqi Li et al.

immunity (housekeeping proteins may not be desirable since they

potentially induce host immune tolerance or immunity against
normal flora organisms). A backbone protein needs to be strongly
immunogenic, possess multiple well-separated continuous epi-
topes, and ideally exhibits systemic and mucosal adjuvanticity.
Additionally, a backbone protein can be easily expressed in an
expression vector and an Escherichia coli strain and extracted at a
great purity and in high yield. In the case there is no suitable
backbone proteins available from a target pathogen, mutant cholera
toxin (CT) of Vibrio cholerae and heat-labile toxin (LT) of entero-
toxigenic E. coli, which have been demonstrated with little or low
enterotoxicity but strong immunogenicity and mucosal and sys-
temic adjuvanticity [8–11], can serve as alternative backbone
proteins.
Epitopes from heterogeneous strains or virulence factors of
interest are presented on an identified backbone protein. Preferably,
one epitope represents one strain or one virulence factor and
induces protective immune responses against its representing strain
or virulence factor. With several protective epitopes integrated into
a backbone immunogen, an MEFA immunogen induces broad
immunity against multiple heterogeneous strains or virulence fac-
tors. Neutralizing or functionally protective epitopes which have
been characterized are available at databases including Immune
Epitope Database—IEDB (www.iedb.org). However, in many
cases, protective epitopes likely have not been identified. Alterna-
tively, immunodominant epitopes which are identified in silico can
be used initially. Eventually, as we recently described [12–15],
immunodominant epitopes are screened in empirical studies and
protective epitopes are selected for MEFA immunogen
construction.

2.3 MEFA 1. pET28a or other expression vectors, to clone an MEFA immu-

Immunogen nogen gene.
Construction, 2. E. coli strain BL21-CodonPlus (DE3) or other E. coli strains, to
Expression, host MEFA plasmids for MEFA protein expression.
and Extraction 3. LB agar plates and 2 YT medium broth.
4. Kanamycin, 30 μg/mL in final (for pET28a).
5. IPTG (isopropyl-β-D-thiogalactopyranoside).
6. B-PER, Bacterial Protein Extraction Reagent (in phosphate
buffer; solution contains a proprietary, nonionic detergent in
20 mM Tris–HCl; pH 7.5).
7. Ni-NTA Agarose, Ni-NTA columns, wash buffer (50 mM
NaH2PO4, 300 mM NaCl, 20 mM imidazole, 0.05% Tween
20, pH 8.0), elution buffer (50 mM NaH2PO4, 300 mM
NaCl, 250 mM imidazole, 0.05% Tween 20, pH 8.0), to
extract 6 His-tagged recombinant protein (see Note 1).
 MEFA for Multivalent Vaccines 155

8. Spectrum Spectra/Por dialysis membrane tubing, at suitable

pore sizes depending on MEFA protein molecular mass
(5–10 kDa less than the MEFA protein molecular mass).
9. 1x IB solubilization buffer: 50 mM CAPS, pH 11.0, supple-
mented with 1% N-lauroylsarcosine and 1 mM DTT.
10. Protein refolding and dialysis buffer: 20 mM Tris–HCl,
pH 8.5, supplemented with 0.1 mM DTT or without DTT.

2.4 Protein 1. SDS-PAGE, for MEFA protein Coomassie blue stain or West-
Biochemical Property ern blot with specific antibodies.
and Immunogenicity 2. Coomassie blue stain buffer (0.125% Coomassie blue R350,
Characterization 50% methanol, 10% acetic acid), and destain buffer (20% meth-
anol, 10% acetic acid, in ddH2O).
3. Protein immunogenicity studies: MEFA protein-induced T cell
and/or B cell immune responses are examined in animal immu-
nization studies. Antigen specific adaptive immunity and
in vitro protective activities, which vary and are specific to the
pathogen of interest, are measured to assess MEFA immuno-
gen broad immunity (B cell epitopes are used for MEFA immu-
nogen construction in this protocol).
Animal immunization: 8-week-old mice, 25-gauge needle and
1-mL syringe, alum or incomplete Freund’s adjuvant; mouse fecal
and serum samples collected for measuring antigen-specific anti-
body responses.
Antibody titration: 96-well microtiter plates, ELISA coating
antigens specific to the virulence factors targeted by MEFA protein,
HRP-conjugated goat anti-mouse IgG and IgA secondary antibo-
dies, 3,30 ,5,50 -tetramethybenzidine (TMB) Microwell Peroxidase
Substrate, a 96-well plate reader.
Antibody in vitro protection: assays to evaluate in vitro protec-
tion of MEFA-induced antibodies vary and are specific to patho-
gens or virulence factors of interest.
Preclinical challenge studies: if animal challenge models are
available, immunizing animals with an MEFA protein to assess
epitope-specific immunogenicity, then challenging the immunized
animals with pathogen strains to measure cross protection and to
evaluate potential application of an MEFA immunogen for multi-
valent vaccine development.

3 Methods

3.1 Backbone Online programs that predict epitopes, locate epitope positions,
Protein Selection image protein secondary structure, and examine protein stability
are applied to select a suitable backbone protein. B cell epitopes on
a backbone protein, including epitope amino acid sequence, posi-
tion, and antigenic score, are predicted and analyzed by using
156 Siqi Li et al.

Fig. 2 Illustration of in silico identifying B cell immunodominant epitopes with IEDB program. Left: a diagram to
show in silico predicted epitopes, with antigenic scores and locations, from a backbone protein. Right: a table
to indicate epitope position, sequence and length

immune epitope database and analysis resource server IEDB

[16]. Backbone protein half-life and stability are evaluated with
ExPASy ProtParam [17]. Protein homology/analogy recognition
engine version 2.0—Phyre2 [18] and PyMol are used to generate
backbone protein 3D structure images and to show epitope presen-
tation on a backbone protein (see Note 2).

3.1.1 To Predict 1. Open IEDB www.iedb.org/.

Backbone Immunogen 2. Select B-cell Tools.
Epitopes
3. Click “Prediction of linear epitopes from protein sequence” for
B cell epitope prediction.
4. Enter protein Swiss-Prot ID or input a backbone protein amino
acid sequence, select “Bepipred Linear Epitope Prediction 2.0”
for B-cell epitopes, and submit to proceed with epitope
prediction.
5. Identified epitopes are presented in a graph and tables (Fig. 2)
(see Note 3).

3.1.2 To Examine 1. Open ProtParam tool link at ExPASy (https://web.expasy.

Backbone Protein Stability org/ProtParam/).
2. Enter backbone protein Swiss-Prot ID or accession number
(if protein is deposited in ExPASy database); alternatively,
input protein amino acid sequence as directed.
3. Press “computer parameters” command to proceed
computation.
4. Retrieve data of protein molecular weight, extinction coeffi-
cients, estimated protein half-life, and protein instability index
(see Note 4).

3.1.3 To Generate 1. Open Phyre2 site http://www.sbg.bio.ic.ac.uk/~phyre2/

Backbone Protein Model html/page.cgi?id¼index.
and to Illustrate Protein 3D 2. Type in your email address (to receive analysis updates) in a
Images and Epitopes designated box shown on the window.
 MEFA for Multivalent Vaccines 157

3. Input protein amino acid sequence and name your optional job
description.
4. Select “Intensive” as the modeling and analysis mode.
5. Press “Phyre Search.”
6. Receive email notices of computation updates, results in a “xxx.
pdb file format” (xxx is the name of your optional job descrip-
tion; the confidence in the intensive model should be greater
than 90%), and a link to view your protein model and 3D
structure image.
7. Download your “xxx.pdb” file from Phyre2 site, view protein
structure, and further analyze “xxx.pdb” data for 3D structure
with PyMol/ EdiPyMol.
8. Download PyMol (https://www.pymol.org/).
9. Open your Phyre2 file “xxx.pdb” through PyMol program by
using “File”!“Open”!“Downloads” by selecting your “xxx.
pdb file.
10. Select the “Display” tab on the top to view protein amino acid
sequence.
11. Select the “S” tab to show “surface” at the right for display
type of the secondary structure in a 3D model, and the “C” for
a uniform color for the backbone protein model (Fig. 3a).
12. Mark epitope amino acid sequences in different colors to image
the location and surface exposure of each epitope (Fig. 3b) (see
Note 5).

Fig. 3 Illustration of a backbone protein 3D model (a) and position of a backbone epitope (b), by PyMol
program
158 Siqi Li et al.

3.2 Identification With a desirable backbone immunogen selected, epitopes from the
of Epitopes from heterogeneous strains or virulence factors of interest must be iden-
Target Heterogeneous tified. Protective epitopes are ideal for MEFA immunogen con-
Strains or Virulence struction. Having epitope native antigenicity retained, an MEFA
Factors immunogen carrying multiple protective epitopes of various strains
or virulence factors induces broadly protective immune responses.
As pointed out in Subheading 3.1, protective epitopes from well
characterized pathogens are deposited at IEDB (www.iedb.org)
and can be used directly for MEFA immunogen construction. For
strains or virulence factors without protective epitopes being iden-
tified, immunodominant epitopes in silico identified can be used
first; eventually protective epitopes selected by screening immuno-
dominant ones in empirical studies are included in an MEFA
immunogen.

3.2.1 To Predict Epitopes Immunodominant linear epitopes from each virulence factor or
of Heterogeneous Strains strain are predicted with IEDB, the same as backbone protein
or Virulence Factors epitope prediction described in Subheading 3.1.1, with the only
of Interest difference is the input of different amino acid sequences.

3.2.2 To Identify Immunodominant epitopes identified in silico may not necessarily

Protective Epitopes induce protective immune responses. Practically, empirical studies
are followed to screen immunodominant epitopes and to identify
the protective epitopes. A panel of immunodominant epitopes from
a virulence factor or strain is selected (based on antigenic scores),
and each epitope is genetically fused to a carrier protein for epitope
fusions (to mimic epitope native antigenicity). With each epitope
fusion gene cloned in an expression vector, individual epitope
fusion proteins are extracted and used for mouse immunization.
As we described [12–14], epitope fusion-induced immune
responses are measured in ELISAs with specific coating antigens
to confirm immune dominance, and then examined for protection
against the virulence factor or disease with in vitro assays. Epitopes
in the fusion proteins that induce protective immune responses are
determined as protective.

3.3 MEFA Having a backbone protein and protective (or immunodominant)

Immunogen epitopes from heterogeneous strains or virulence factors identified,
Construction we are ready in silico to construct and to optimize an MEFA
immunogen. Substitution of backbone epitopes with selecting pro-
tective epitopes (or immunodominant epitopes if protective ones
are not identified) is carried out either sequentially, substituting one
epitope at a time, or with all epitopes replaced at once and modified
afterward.
To substitute epitopes sequentially, we select an epitope of
interest that possesses a similar length and antigenic score as a
backbone epitope, proceed the substitution, construct a chimeric
protein, and examine the resultant protein comparatively with the
 MEFA for Multivalent Vaccines 159

backbone protein for protein stability in ExPASy as in Subheading

3.1.2 and for epitope presentation with Phyre2 and PyMol as in
Subheading 3.1.3. If the top-ranked epitope from a virulence factor
or strain does not match well with a backbone epitope, we select the
second or the third ranked epitope whichever shows the better
match with the backbone epitope. With assessment of no major
alteration in protein stability and secondary structure, we proceed
to substitution of the next epitope and continue the process until all
epitopes are substituted. Alternatively, based on epitope length and
antigenic score, we pairwise epitopes of interest with backbone
epitopes and complete epitope substitution at once. The resultant
MEFA protein is examined with ExPASy, Phyre2, and PyMol for
MEFA protein stability, protein secondary structure, and epitope
presentation. If protein stability and structure are altered signifi-
cantly, we adjust epitope pairing and optimize epitope substitution.
A MEFA protein is further characterized with molecular
dynamic simulation programs from GROMACS or CHARMM as
we described previously [6]. With comparative analyses of protein
dynamic properties, including protein structure stability, conforma-
tional flexibility and solvent-accessible surface, of an MEFA protein
with a backbone protein, MEFA protein overall structure and
stability similarity is examined. With analyses of dynamic simulation
focusing on an epitope on the MEFA versus its native format on a
virulence factor protein, mimicking of epitope native antigenicity is
evaluated.

3.3.1 GROMACS 1. Download and install SSH client from link https://www.
to Characterize an MEFA netsarang.com/en/all-downloads/ (Xshell for Windows OS);
Protein log in from your desktop as an SSH client to connect remote
server (Linux OS) and execute commands (note: more infor-
mation at https://netsarang.atlassian.net/wiki/spaces/
ENSUP/pages/31555780/Getting+Started).
2. Download and install GROMACS (http://www.gromacs.org/
) and FFTW (version 3.3.8) (http://www.fftw.org/download.
html).
3. Input “MEFA protein (or backbone or each virulence factor).
pdb” file into Xshell generated from Phyre2 under your server
account.
4. Generate an MEFA protein topology form the input pdb.file,
with command gmx pdb2gmx -f MEFA protein.pdb -o
mefa_processed.gro -water spce, with the selection of GRO-
MOS96 53a6 force field.
5. Select periodic boundary conditions (PBC) at “cubic” shape
and “1.0 nm” parameters with gmx editconf -f mefa_pro-
cessed.gro -o mefa_newbox.gro -c -d 1.0 -bt cubic.
160 Siqi Li et al.

6. Enter solvent water into the simulation box with gmx solvate
-cp mefa_newbox.gro -cs spc216.gro -o mefa_solv.gro -p
topol.top.
7. Conduct system neutralization with command gmx grompp -f
ions.mdp -c mefa_solv.gro -p topol.top -o ions.tpr, add
counter ions through “genion” with gmx genion -s ions.tpr
-o mefa_solv_ions.gro -p topol.top -pname NA -nname CL
-np 9.
8. Minimize energy with commands gmx grompp -f minim.
mdp -c mefa_solv_ions.gro -p topol.top -o em.tpr and
gmx mdrun -v -deffnm em.
9. Equilibrate NVT with gmx grompp -f nvt.mdp -c em.gro -p
topol.top -o nvt.tpr and gmx mdrun -deffnm nvt.
10. Equilibrate NPT with command gmx grompp -f npt.mdp -c
nvt.gro -t nvt.cpt -p topol.top -o npt.tpr and gmx mdrun
-deffnm npt.
11. Carry out molecule dynamics (MD) simulations, with gmx
grompp -f md.mdp -c npt.gro -t npt.cpt -p topol.top -o
md_0_1.tpr and gmx mdrun -deffnm md_0_1.
12. Calculate Cα root-mean-square deviation (RMSD) and root-
mean-square fluctuations (RMSF) (Fig. 4), with commands
gmx trjconv -s md_0_1.tpr -f md_0_1.xtc -o
md_0_1_noPBC.xtc -pbc mol -ur compact, gmx rms -s
md_0_1.tpr -f md_0_1.trr -o rmsd.xvg, and gmx rmsf -s
md_0_1.tpr -f md_0_1.xtc -o rmsf.xvg, respectively.

Fig. 4 GROMOS molecule dynamic simulation to show root-mean-square deviation—RMSD (a) and root-
mean-square fluctuation—RMSF (b) of a backbone protein and an MEFA immunogen protein. RMSD and
RMSF are used to assess structural and antigenic similarity between a backbone protein and an MEFA protein,
or between an epitope on the MEFA and the epitope on the native virulence factor protein. Boxed is a backbone
protein epitope and a substituting epitope on an MEFA protein
 MEFA for Multivalent Vaccines 161

Fig. 5 GROMOS molecule dynamic simulation to show solvent average surface area (SASA) of a backbone
protein (or an MEFA immunogen protein, or the epitope of interest on an MEFA protein or native virulence
factor protein). SASA is used to assess structural and antigenic similarity between a backbone protein and an
MEFA protein, or between an epitope on the MEFA and the epitope on the native virulence factor protein

13. Calculate solvent-accessible surface areas (SASA) (Fig. 5) with

gmx sasa -s md_0_1.tpr -f md_0_1.trr -o area.xvg -or -oa.
14. Compare results of the MEFA versus backbone, and each
epitope on the MEFA versus the native epitope on each viru-
lence factor, to evaluate MEFA protein construction in silico
(if overall protein structure or epitope native antigenicity is
significantly altered, MEFA is optimized by modifying epitope
position or selecting a different epitope from the representing
virulence factor) (see Note 6).

3.3.2 CHARMM Alternatively, an MEFA protein is examined with license based

to Characterize MEFA CHARMM, by constructing protein comparative modeling using
Protein Rosetta [19–21], carrying out molecular dynamics
(MD) simulation, and analyzing protein secondary structure.
1. Download and install MMTSB toolset (http://blue11.bch.
msu.edu/mmtsb/Main_Page).
2. Using “convpdb.pl” from MMTSB toolset to clean up the
PDB file with convpdb.pl -segnames model_pdb.pdb > ms.
pdb. Here ms.pdb is the final clean PDB file for CHARMM.
3. Download and install CHARMM (https://www.charmm.org/
).
4. Prepare protein structure in CHARMM format charmm
pdbid¼ms < mkprotein.inp.
5. Prepare water box for protein solvation charmm Water-
Depth¼8 target¼protein cubic¼1 < mkwaterbox.inp.
162 Siqi Li et al.

6. Set protein solvation and neutralizing charge charmm tar-

get¼protein < solvation.inp.
7. Apply 5.0 ns equilibrium in NVT, harmonic positional restrains
on heavy atoms but gradually reduced charmm < nvt_e-
quil.5 ns.inp > npt_equil.5ns.log.
8. Conduct the first 50 ns production simulation charmm-
gpu < nvt_prod.0.inp > nvt_prod.0.log (see Note 7).
9. Get protein dcd out of water box charmm < get-dcd-
nvt_prod.0.inp and charmm cycle¼1 < get-dcd-nvt_prod.
restart.inp for additional cycle.
10. Calculate protein backbone and Cα RMSD charmm dcdin¼0
< anal-rms-asa.inp, and charmm dcdin¼1 < anal-rms-asa.
inp for additional dcd file.
11. Calculate accessible surface areas (ASA) charmm dcdin¼0 <
anal-cs*.inp, and charmm dcdin¼1 < anal-cs*.inp for addi-
tional dcd file. Here the anal-cs*.inp files are predefined. It is
based on different protein segments (i.e., a range of residue id
for epitope region).
12. Normalize ASA into the relative ASA. The calculation using the
ASA of each epitope region divided by total ASA (see Note 8).

3.4 MEFA Protein After protein structure stability and epitope antigenicity have been
Expression, Extraction, assessed and verified in silico, we synthesize an MEFA immunogen
and Immunogenicity gene and clone the gene for protein expression and extraction. The
Characterization MEFA gene can be inserted into the pET28a expression vector
directly at synthesis, or with two restriction enzyme sites flanked
at the MEFA gene for cloning into the pET28a vector, for expres-
sion of 6xHis-tagged or tag-less recombinant MEFA protein in
E. coli strain BL21 (DE3), by following standard molecular biology
protocols [22]. MEFA protein expression and extraction are ver-
ified by SDS PAGE with Coomassie blue staining and western
blotting with specific antibodies; MEFA protein broad immunoge-
nicity is confirmed in mouse immunization.

3.4.1 MEFA Protein Recombinant 6xHis-tag MEFA protein is purified with Ni-NTA
Expression and Extraction agarose in a column (www.qiagen.com), whereas tag-less MEFA
protein expressed in inclusion body is extracted with B-PER™
Bacterial Protein Extraction Reagent (https://www.thermofisher.
com).
1. Pick a colony from overnight growth on a LB/kan agar plate,
culture in 5–10 ml 2 YT medium supplemented with kana-
mycin (30 μg/mL) at 37  C overnight on a shaker (200 rpm).
2. Transfer 2–3 mL bacterial growth to 200 mL 2 YT medium
supplemented with kanamycin (30 μg/mL) in a 500 mL flask,
culture at 37  C on a shaker (200 rpm) till the OD600 reaches
0.5–0.7.
 MEFA for Multivalent Vaccines 163

3. Add IPTG (0–1 mM; optimal concentration varies for different

proteins), culture for 4 more hours, and transfer into a 250 mL
high-speed centrifuge bottle.
4. Harvest bacteria by centrifugation for 15 min at 17,000  g
(12,000 rpm with rotor F21-8x50Y), discard supernatant.
5. Freeze the pellet in a 80  C freezer overnight.
6. Thaw the frozen pellet and add 10 mL B-PER reagent (in pho-
sphate buffer to lyse bacteria. Vortex or pipet up and down
until the suspension is homogenous, incubate for 30 min at
room temperature (RT) on the shaker (100 rpm).
7. Sonicate bacterial lysate for 5 min (15 s on, 15 s off, at 30%
amplitude; on ice).
8. Centrifuge bacterial lysate at 17,000  g at 4  C for 15 min,
resuspend pellet in 10–20 mL 1 phosphate buffered saline
(PBS, 137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4,
1.8 mM KH2PO4).
9. To extract 6xHis tag MEFA protein, gently mix the suspension
with 5–10 mL Ni-NTA agarose resin by rotating at 4  C
overnight.
10. Apply the suspension to a 100 mL column, flow through, wash
4 with 100 mL wash buffer, elute 4 with 100 mL elution
buffer to collect 6  His tag MEFA protein, run samples (20
μL) on SDS-PAGE. Use Amico or equivalent spin column (at a
suitable pore size), or spectrum tubing with polyethylene gly-
col (PEG) at 4  C, to concentrate the MEFA protein, and store
at 80  C (ready for immunization).
11. To extract tagless MEFA inclusion body protein, homogenize
suspension from step 8 by passing up and down with 20 mL
syringe (use 18G needle).
12. Add 100 mL PBS, vortex, centrifuge at 17,000  g at 4  C for
15 min, collect pellet.
13. Repeat step 12 twice to wash protein pellet.
14. Dissolve the purified inclusion body protein pellet in
5–10 mL PBS.
15. Transfer resuspended protein to a 1.5 mL micro centrifuge
tube (1 mL per tube), centrifuge with a bench top microcen-
trifuge at 16,200  g at RT for 10 min, discard supernatant.
16. (To refold purified inclusion body protein) Suspend and solu-
bilize protein pellet with 1 mL 1 IB solubilization buffer
(freshly made) by vigorously pipetting and vortex; incubate
suspension on a shaker (100 rpm) at RT for 1–2 h.
17. Centrifuge at 16,200  g at RT for 10 min and collect
supernatant.
164 Siqi Li et al.

18. Check protein with SDS-PAGE, proceed to one-step dialysis

and refolding.
19. To dialyze and refold protein, transfer all of the supernatant
into a dialysis tubing membrane with a pore size 5–10 kDa less
than the MEFA protein molecular mass.
20. Prepare dialysis buffer for a minimum of two buffer exchanges,
for a volume of greater than 50 times of the protein suspension
sample volume (for an example, 500 mL dialysis and refolding
buffer for 10 mL protein suspension).
21. Dialyze for 4 h at 4  C and continue with two additional buffer
exchanges using dialysis buffer without DTT; If visible precip-
itation occurs following dialysis, centrifuge the dialysis protein
solution at 16,200  g for 10 min at 4  C.
22. Collect clear supernatant containing refolded MEFA protein,
aliquot, measure protein concentration, and store MEFA pro-
tein at 80  C.
23. Characterize refolded protein in ELISAs, SDS-PAGE Coomas-
sie blue stain (incubation for 1 h in stain buffer, and then in
destain buffer overnight; room temperature (50 rpm shaker),
and western blot with MEFA protein specific antibodies.

3.4.2 MEFA Protein Eight-week-old female BALB/c mice, 8–10 per group, are com-
Immunization monly used for immunization with MEFA protein to examine
and MEFA-Specific broad immunogenicity (see Note 9).
Antibody Titration
1. Mix and emulsify 50 μg MEFA protein (in 50 μL PBS) with an
equal amount of incomplete Freund’s adjuvant (IFA); 50 μL
PBS mixed with IFA as a negative control.
2. Inject each mouse with MEFA protein (or PBS for the control
group) intraperitoneally (IP) [or intramuscularly (IM), subcu-
taneously (SC), or intradermally (ID)].
3. Two booster injections (of the same dose and route of the
primary) are followed at the interval of 2 weeks.
4. Collect mouse serum and fecal samples 2 weeks after the final
booster (see Note 10).
5. Store serum and fecal supernatant at 80  C until the use for
antibody titration and antibody neutralization studies.
6. To titrate antigen-specific antibody responses, coat 96-well
microtiter 2HB ELISA plates with epitope specific antigen
(50–100 ng of each virulence factor protein in 100 μL
0.05 M carbonate-bicarbonate buffer, pH 9.6; per well) over-
night at 4  C.
7. Discard coating reagent, wash wells three times with PBST
(PBS with 0.05% Tween 20), 150 μL PBST per well.
 MEFA for Multivalent Vaccines 165

8. Block uncoated sites with 5% nonfat milk (in PBST), 150 mL

per well, for 1 h at 37  C.
9. Discard blocking buffer, wash 3 with PBST.
10. Add twofold serial dilution of mouse serum samples (1:200
initial dilution) or fecal suspension samples (1:50 initial dilu-
tion), 100 μL per well, in triplicate, incubate 1 h at 37  C.
11. Wash wells with PBST three times and one time with PBS, add
HRP-conjugated goat anti-mouse IgG (1:5.000 dilution) or
IgA (1:3000 dilution), 100 μL per well, incubate for 1 h at
37  C.
12. Wash wells with PBST three times, add 3,30 ,5,5-
0
-tetramethylbenzidine (TMB; 0.4 g/L in an organic base)
Microwell Peroxidase Substrate System, 100 μL per well, incu-
bate at room temperature for 25 min.
13. Read optical density at OD650 in a plate reader.
14. Calculate OD readings into log10, by multiplying the highest
dilution that produces OD readings greater than 0.3 (after
subtraction of the background readings) by the adjusted OD
and converting to a log10 scale (see Note 11).

3.4.3 MEFA MEFA-induced antibody protective activities are examined with

Protein-Induced Antibody in vitro assays, or in vivo if animal challenge models are available.
In Vitro and In Vivo These assays vary and are pathogen or disease specific.
Protective Activities

4 Notes

The protocols for in silico epitope identification, protein stability

assessment, protein 3D image creation and molecule dynamic sim-
ulation provided in this chapter are basic and simplified. Detailed
instructions to run programs and to analyze data are available from
individual program home pages. While most programs are user
friendly, molecule dynamic simulation from CHARMM and
GROMCS is comprehensive and requires specialized computer-
related skills. Other online programs can be used alternatively or
synergistically to identify consensus epitopes and to enhance pro-
tein structure and stability characterization. Additionally, programs
in MacOS and WinOS versions can be slightly different. Native
antigenicity of epitopes on MEFA are assessed comparatively with
their original format or individual representing virulence factor
proteins, based on epitope dynamic property, structure conforma-
tional flexibility of residues, accessible surface areas, epitope anti-
genic scores as well as visual assessment of epitope conformation
based on 3D structure. Eventually, in silico constructed MEFA
166 Siqi Li et al.

immunogens are validated in empirical studies for broad immuno-

genicity and potential application of multivalent vaccine
development.
Mice are used here to initially confirm MEFA protein broad
immunogenicity. Other animal species, if they are relevant to the
disease of interest, will be preferred for immunogenicity studies.
Since B cell epitopes are discussed, only MEFA-induced antibody
responses, but not T cell epitopes and T cell responses, are included
in this chapter. Additionally, in vitro antibody neutralization assays
are specific to each virulence factor or disease, whereas suitable
animal challenge models may not exist for many pathogens and
diseases, thus no protocols for antibody functional assays and ani-
mal challenge studies are provided. A suitable animal challenge
model uses the animal species that are naturally susceptible to the
pathogen and mimics disease progression, and after infection
develop similar disease outcomes and similar levels of immune
responses as humans [23–26]. Therefore, vaccine efficacy can be
evaluated preclinically by immunizing the animals and then infect-
ing the immunized animals with a pathogen. Eventually, vaccine
candidacy is examined in human subject studies. Control human
infection models (CHIM) [27], which directly test immunogenic-
ity, safety and efficacy of vaccine candidates in volunteers, however,
enables the use of animal challenge models to be skipped and
acceleration of vaccine development.
1. Other wash and elution buffers can be found at the QIAGEN
home page.
2. Protein secondary structure can be generated with CHARMM
and ExPASy as well.
3. IEDB predicts T-cell epitopes as well; instructional details for
epitope prediction are at the IEDB webpage www.iedb.org/.
4. A protein is classified stable if an instability index is less than 40.
Instructions for protein stability assessment are available at
ExPASy home page.
5. Phyre2 and PyMol program updates and online assistance are at
Phyre2 and PyMol webpages.
6. A constructed MEFA protein is expected to show protein
structure and stability similar to the backbone protein; epitopes
of interest on the MEFA protein display similar structural
conformation and solvent-accessible surface areas as they are
on virulence factor proteins.
7. Depending on convergence, one may increase the simulation
length, for another 50 ns production simulation with charmm-
gpu cycle¼1 < nvt_prod.restart.inp > nvt_prod.1.log. The
index number can be changed to extend another 50 ns until the
simulation is converged.
 MEFA for Multivalent Vaccines 167

8. Backbone or MEFA model was first solvated in a cubic box of

TIP3P water, with the box size was set at 69 Å. Sodium ions
were added to neutralize the total charge of the system. Energy
minimization was first performed to remove improper molecu-
lar arrangement. Followed by 5.0 nanosecond
(ns) equilibrium, harmonic positional restrains were applied
on heavy atoms but gradually reduced to equilibrate the pro-
tein structures and water orientation. Finally, the production
simulation was performed for a total simulation length of
350 ns. In the production run, NPT simulation was performed
in Langevin dynamics with the constant temperature of 298 K
and the constant pressure at 1.0 atm. Timestep 2.0 fs was used
with the SHAKE algorithm applied to all hydrogen-containing
bonds. Particle mesh Ewald was utilized for electrostatics cut-
off of 13 Å. The cutoff of van der Waals interactions is 13 Å,
with a switching function between 12 and 13 Å was used. The
time evolution of the backbone or MEFA RMSD was first
calculated to benchmark the convergence of simulation. The
solvent accessible surface area (ASA) was calculated using 1.4 Å
as the size of water probe and was further normalized into the
relative ASA by calculating ASA for each epitope region. The
relative ASA is a normalized calculation using each epitope
region divided by total ASA.
9. To avoid gender bias, both sexes, in equal distribution,
are used.
10. Fecal pellets are suspended in fecal reconstitution buffer:
10 mM Tris, 100 mM NaCl, 0.05% Tween-20, 5 mM sodium
azide, pH 7.4, supplemented with protease inhibitor phenyl-
methylsulfonyl fluoride (0.5 mM final concentration); 1 g fecal
pellet in 5 ml buffer; spin at 13,000  g at room temperature
for 15 min to collect supernatant.
11. Other adjuvants including double mutant heat-labile toxin of
enterotoxigenic E. coli (dmLT) can also be used for mouse
immunization under parenteral or mucosal routes. The
amount of each coating antigen and secondary dilutions are
optimized in standard checkboard test.

Acknowledgments

This work is supported by NIH R01AI121067-01A1 and Univer-

sity of Illinois at Urbana-Champaign. We also thank Shuangqi
Wang, Ti Lu, and Ipshita Upadhyay for technical assistance.
168 Siqi Li et al.

References
1. Collaborators GD, Collaborators GH (2017) https://doi.org/10.1016/j.vaccine.2004.11.
Global, regional, and national disability- 010
adjusted life-years (DALYs) for 333 diseases 12. Huang JC, Duan QD, Zhang WP (2018) Sig-
and injuries and healthy life expectancy nificance of enterotoxigenic Escherichia coli
(HALE) for 195 countries and territories, (ETEC) heat-labile toxin (LT) enzymatic sub-
1990-2016: a systematic analysis for the global unit epitopes in LT enterotoxicity and immu-
burden of disease study 2016. Lancet 390 nogenicity. Appl Environ Microbiol 84(15):
(10100):1260–1344 e00849-00818. https://doi.org/10.1128/
2. Global health estimates 2016: Deaths by cause, AEM.00849-18
age, sex, by country and by region, 2000–2016 13. Lu T, Moxley RA, Zhang W (2019) Mapping
(2018). https://www.who.int/healthinfo/ the neutralizing epitopes of enterotoxigenic
global_burden_disease/estimates/en/index1. Escherichia coli (ETEC) K88 (F4) fimbrial
html. Accessed 21 Jan 2021 adhesin and major subunit FaeG. Appl Environ
3. Nabel GJ (2013) Designing tomorrow’s vac- Microbiol 85(11):e00329–e00319. https://
cines. New Engl J Med 368(6):551–560. doi.org/10.1128/AEM.00329-19
https://doi.org/10.1056/NEJMra1204186 14. Lu T, Seo H, Moxley RA, Zhang W (2019)
4. Mascola JR, Ahmed R (2015) Vaccines: vac- Mapping the neutralizing epitopes of F18 fim-
cines for cancer and infectious diseases. Curr brial adhesin subunit FedF of enterotoxigenic
Opin Immunol 35:V–Vii. https://doi.org/10. Escherichia coli (ETEC). Vet Microbiol
1016/j.coi.2015.07.009 230:171–177. https://doi.org/10.1016/j.
5. Ruan X, Sack DA, Zhang W (2015) Genetic vetmic.2019.02.015
fusions of a CFA/I/II/IV MEFA (multiepi- 15. Lu T, Moxley RA, Zhang W (2020) Applica-
tope fusion antigen) and a toxoid fusion of tion of a novel epitope and structure
heat-stable toxin (STa) and heat-labile toxin vaccinology-assisted fimbria-toxin multiepi-
(LT) of enterotoxigenic Escherichia coli tope fusion antigen of enterotoxigenic Escher-
(ETEC) retain broad anti-CFA and antitoxin ichia coli for multivalent vaccine development
antigenicity. PLoS One 10(3):e0121623 against porcine post-weaning diarrhea. Appl
6. Duan Q, Lee KH, Nandre RM, Garcia C, Environ Microbiol 86(24):e00274-20.
Chen J, Zhang W (2017) MEFA (multiepitope https://doi.org/10.1128/AEM.00274-20
fusion antigen)-novel technology for structural 16. Vita R, Mahajan S, Overton JA, Dhanda SK,
vaccinology, proof from computational and Martini S, Cantrell JR et al (2019) The
empirical immunogenicity characterization of immune epitope database (IEDB): 2018
an enterotoxigenic Escherichia coli (ETEC) update. Nucleic Acids Res 47(D1):
Adhesin MEFA. J Vaccines Vaccin 8(4):367. D339–D343. https://doi.org/10.1093/nar/
https://doi.org/10.4172/2157-7560. gky1006
1000367 17. Artimo P, Jonnalagedda M, Arnold K,
7. Seo H, Duan Q, Zhang W (2020) Vaccines Baratin D, Csardi G, de Castro E et al (2012)
against gastroenteritis, current progress and ExPASy: SIB bioinformatics resource portal.
challenges. Gut Microbes 11(6):1486–1517. Nucleic Acids Res 40(W1):W597–W603.
https://doi.org/10.1080/19490976.2020. https://doi.org/10.1093/nar/gks400
1770666 18. Kelley LA, Mezulis S, Yates CM, Wass MN,
8. Lycke N, Holmgren J (1986) Strong adjuvant Sternberg MJE (2015) The Phyre2 web portal
properties of cholera toxin on gut mucosal for protein modeling, prediction and analysis.
immune responses to orally presented antigens. Nat Protoc 10(6):845–858. https://doi.org/
Immunology 59(2):301–308 10.1038/nprot.2015.053
9. Clements JD, Hartzog NM, Lyon FL (1988) 19. Chivian D, Baker D (2006) Homology model-
Adjuvant activity of Escherichia coli heat-labile ing using parametric alignment ensemble gen-
enterotoxin and effect on the induction of oral eration with consensus and energy-based
tolerance in mice to unrelated protein antigens. model selection. Nucleic Acids Res 34(17):
Vaccine 6(3):269–277 e112. https://doi.org/10.1093/nar/gkl480
10. Lycke N (2005) Targeted vaccine adjuvants 20. Raman S, Vernon R, Thompson J, Tyka M,
based on modified cholera toxin. Curr Mol Sadreyev R, Pei JM et al (2009) Structure pre-
Med 5(6):591–597 diction for CASP8 with all-atom refinement
11. Freytag LC, Clements JD (2005) Mucosal using Rosetta. Proteins Struct Funct Genet
adjuvants. Vaccine 23(15):1804–1813.
 MEFA for Multivalent Vaccines 169

77:89–99. https://doi.org/10.1002/prot. development and testing. ILAR J 56

22540 (1):53–62. https://doi.org/10.1093/ilar/
21. Song JS, Liu YQ, Liu CZ, Xie JP, Ma LX, Wang ilv009
LP et al (2013) Cumulative analgesic effects of 25. Swearengen JR (2018) Choosing the right ani-
EA stimulation of sanyinjiao (SP 6) in primary mal model for infectious disease research. Ani-
dysmenorrhea patients: a multicenter rando- mal Model Exp Med 1:100–108. https://doi.
mized controlled clinical trial. Zhen Ci Yan org/10.1002/ame2.12020
Jiu 38(5):393–398 26. Ramin Sedaghat Herati EJW (2018) What is
22. Ausubel FM, Brent R, Kingston RK, Moore the predictive value of animal models for vac-
DD, Seidman JG, Smith JA, Struhl K (1999) cine efficacy in humans. Cold Spring Harb Per-
Short protocols in molecular biology, 4th edn. spect Biol 10:a031583. https://doi.org/10.
Wiley, New York, NY 1101/cshperspect.a031583
23. Spira WM, Sack RB, Froehlich JL (1981) Sim- 27. Gordon SB, Rylance J, Luck A, Jambo K, Fer-
ple adult rabbit model for Vibrio cholerae and reira DM, Manda-Taylor L et al (2017) A
enterotoxigenic Escherichia coli diarrhea. Infect framework for controlled human infection
Immun 32(2):739–747. https://doi.org/10. model (CHIM) studies in Malawi: report of a
1128/IAI.32.2.739-747.1981 Wellcome Trust workshop on CHIM in low
24. Gerdts V, Wilson HL, Meurens F, van Drunen income countries held in Blantyre, Malawi.
Littel-van den Hurk S, Wilson D, Walker S et al Wellcome Open Res 2:70. https://doi.org/
(2015) Large animal models for vaccine 10.12688/wellcomeopenres.12256.1
 Chapter 11

Production, Isolation, and Characterization

of Bioengineered Bacterial Extracellular Membrane Vesicles
Derived from Bacteroides thetaiotaomicron and Their Use
in Vaccine Development
Régis Stentz, Ariadna Miquel-Clopés, and Simon R. Carding

Abstract
Bacterial extracellular vesicles (BEVs) possess features that make them well suited for the delivery of
therapeutics and vaccines. This chapter describes methods for engineering the commensal human intestinal
bacterium Bacteroides thetaiotaomicron (Bt) to produce BEVs carrying vaccine antigens and accompanying
methods for isolating and purifying BEVs for mucosal vaccination regimens.

Key words Bacteroides thetaiotaomicron, Bacterial extracellular vesicles, Crossflow ultrafiltration,

Vaccines, Immunization

1 Introduction

Conventional vaccines based on the use of attenuated or inactivated

forms of the target pathogen have successfully eradicated smallpox
and rinderpest as well as significantly reducing the burden of many
other infectious diseases throughout the past century. However,
the time needed to identify vaccine targets, the high cost of vaccine
development and manufacture, and the limited production capac-
ity, make these traditional approaches less than optimal in the rapid
response to epidemics and pandemics [1]. Furthermore, these
vaccines are usually delivered parenterally via injection, which
makes mass immunization costly particularly in resource-poor
developing countries [2]. There is, therefore, a need for the devel-
opment of new vaccines that are versatile, cost-effective, safe, and
enable global immunization. To this end, various new vaccination
technologies have emerged including the use of synthetic protein
and peptide antigens [3]. Protein subunit vaccines are attractive
because of their inherent safety although they can suffer from poor

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_11,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

171
172 Régis Stentz et al.

immunogenicity and high manufacturing costs [4]. To address

these constraints, nanoparticle-based delivery technologies have
been developed which includes nanoparticle-sized extracellular
vesicles naturally produced by bacteria.
Bacterial extracellular vesicles (BEVs) are spherical nanostruc-
tures composed of membrane-derived lipid bilayers with a diameter
of between 20 and 400 nm. BEVs generated by gram-negative
bacteria primarily consist of vesicles derived from the outer mem-
brane containing phospholipids, outer membrane proteins, lipopo-
lysaccharides, and capsular polysaccharides with their lumen
principally filled with periplasmic content [5]. These components,
which include microbe-associated molecular pattern molecules,
confer inherent and potent adjuvanticity on BEVs which together
with their natural temperature, chemical resistance, and straightfor-
ward isolation [6–8] makes them well suited as vaccine delivery
vehicles capable of enhancing the immunogenicity of protein/pep-
tide antigens without the need for chemical adjuvants [9]. The
ability of BEVs to interact with, and be acquired by, mucosal
epithelial and immune cells [10–12] further enhances their suitabil-
ity for mucosal administration and the generation of local and
systemic immunity [13].
We have engineered the gram-negative bacterium Bacteroides
thetaiotaomicron (Bt), a prominent member of the intestinal micro-
biota of all animals [14, 15], to incorporate virus-, bacteria-, and
human-derived proteins into its BEVs [8, 16]. These engineered Bt
BEVs have been used to protect the gastrointestinal or respiratory
tracts against infection, tissue inflammation and injury. Here, we
describe the methods to implement secretion of vaccine antigens
and other proteins into Bt BEVs for mucosal delivery, which are
outlined in Fig. 1.

2 Materials

Unless stated otherwise, all solutions are prepared with double-

distilled water (ddH2O) with all reagents being stored at ambient
(room) temperature (~20
C). Most methods require the use of
piston pipettes and sterile single-use pipette tips.

2.1 Synthetic Gene 1. Protein sequence (e.g., stalk of the hemagglutinin antigen of
Design influenza A virus strain H5N1 [8, 17]).
2. Secretion sequence signal (e.g., N-terminal sequence of Bt
outer membrane protein A [OmpA] (BT_3852)).
3. Optional additional sequences (e.g., His-tag or FLAG-tag
sequences).
4. Escherichia coli/Bacteroides shuttle expression vector sequence
(e.g., pGH90 [18]) (see Note 3).
 Production, Isolation, and Characterization of Bioengineered Bacterial. . . 173

Fig. 1 Overview of the production and extraction of bioengineered bacterial extracellular vesicles (BEVs)
derived from B. thetaiotaomicron for the development of novel vaccines. Immunization and sample collection
for preclinical development and evaluation are described in Notes 1 and 2

2.2 Cloning The required materials for the cloning of the conjugative plasmid
of Synthetic Gene harboring the gene of interest are as follows:

2.2.1 Generation 1. E. coli/Bacteroides shuttle expression vector.

of Recombinant DNA 2. Restriction enzymes.
3. Ligation buffer and enzyme.

2.2.2 Transformation 1. Crushed ice for thawing the MAX Efficiency® DH10B compe-
of Competent Cells tent cells.
2. S.O.C. medium: commercially available.
3. Luria-Bertani (LB) agar plates: prepare according to the man-
ufacturer’s instructions.
4. Shaking incubator.
5. LB broth for the dilution of the bacterial cell suspension:
prepare according to the manufacturer’s instructions.
6. Agar plates containing ampicillin (working concentration 100
μg/mL).
7. 1.5 mL centrifuge tubes.
8. Water bath.

2.2.3 Screening 1. Shaking incubator.

of Recombinant Bacteria 2. Benchtop centrifuge.
3. Small-scale plasmid isolation kit.
4. Gel electrophoresis chamber and electrophoresis power supply.
174 Régis Stentz et al.

5. Stained agarose gel: 1% (w/v) agarose, e.g., with EvaGreen®

Dye, Biotium.
6. DNA loading buffer.
7. DNA ladder.

2.3 Conjugative 1. E. coli donor strain containing shuttle vector carrying gene of
Transfer of Shuttle interest.
Vector into Bt 2. E. coli helper strain, e.g., J53/R751 [19].
3. 5 mL LB bottle with ampicillin 100 μg/mL (or another suit-
able antibiotic).
4. 5 mL LB bottle with trimethoprim 200 μg/mL (or another
suitable antibiotic)
5. Brain heart infusion (BHI) agar supplemented with hemin and
antibiotics (gentamicin 200 μg/mL and/or erythromycin 5
μg/mL, or another suitable antibiotic): Dissolve 18.5 g of
BHI powder and 7.5 g of agar in 0.5 liter of deionized water
and add 0.75 μM of hemin (BHIH). Autoclave the medium
and leave for a minimum of 24 h in an anaerobic cabinet to fully
deoxygenate.
6. Filter disc 0.45 μm pore size, 25 mm.
7. 50 mL conical centrifuge tubes.
8. Tweezers.
9. Sterile 25 mL wide neck universal glass bottle.

2.4 Assessing 1. BHIH agar plates containing gentamicin 200 μg/mL and
Protein Expression erythromycin 5 μg/mL.
and Secretion into BEV 2. Sterile inoculation loop(s).
2.4.1 Culture of Bt 3. 20 mL BHIH containing erythromycin 5 μg/mL in Universal
Transconjugants bottles.
4. Anaerobic cabinet.
5. 50 mL conical centrifuge tubes.
6. Refrigerated centrifuge.
7. 20
C freezer.

2.4.2 Cell Total Protein 1. 0.2 M Tris–HCl, pH 7.2.

Extraction 2. Sonicator.
3. Refrigerated centrifuge.
4. Bradford reagent.
5. 96-well microplate.
6. Bovine serum albumin (BSA).
 Production, Isolation, and Characterization of Bioengineered Bacterial. . . 175

2.4.3 BEV Total Protein 1. 20 mL syringe.

Extraction 2. 0.22 μm pore-size polyethersulfone (PES) membranes
(Sartorius).
3. Centrifugal concentrator, 100 kDa molecular weight cutoff.
4. Refrigerated centrifuge.
5. 0.2 M Tris–HCl, pH 7.2.
6. Sonicator.
7. Bradford reagent.
8. 96-well microplate.
9. BSA.

2.4.4 Protein Western 1. Gel electrophoresis equipment.

Blotting/Antigen 2. Protein gel.
Immunodetection
3. SDS sample loading buffer.
4. Reducing agent.
5. Antioxidant for protein electrophoresis.
6. Running buffer.
7. Blotting equipment.
8. Western blotting membrane.
9. Tweezers.
10. Tris-Glycine transfer buffer (25): Dissolve 18.2 g Tris Base
and 90.0 g of glycine in 450 mL of deionized water. Mix well
and adjust the volume to 500 mL with deionized water. The
pH of the buffer is 8.3. Store the buffer at 20
C. The buffer is
stable for 6 months at 25
C.
11. Methanol.
12. Orbital shaker.
13. Tris-buffered saline (TBS) buffer: 50 mM Tris–HCl, 150 mM
NaCl, pH 7.5.
14. TBST buffer: TBS buffer with 0.05% Tween 20.
15. Blocking buffer: TBST with 5% non-fat dry milk.
16. Chemiluminescent substrate.
17. Primary antibody (e.g., 6-His Tag monoclonal antibody)
used at a working concentration recommended by the manu-
facturer or by other providers if the antibody is not commer-
cially available (see Note 4).
18. Secondary antibody labelled with horse radish peroxidase
(HRP) (see Note 5).
19. Imaging System.
176 Régis Stentz et al.

2.5 Bacteria 1. BHIH medium.

Medium-Scale Culture 2. Anaerobic cabinet.
2.5.1 BHIH Bacterial 3. Magnetic stirrer.
Culture 4. Sterile magnetic stirring bar.

2.5.2 BDM+ Bacterial 1. Bacteroides defined medium Plus (BDM+): To prepare

Culture 500 ml, dissolve 2.61 g of KH2PO4 and 7.03 g of K2HPO4 *
3 H2O into 481 mL of deionized water, add the following
solutions to a final concentration of 15 mM NaCl and 8.5 mM
(NH4)2SO4, adjust the pH to 7.4 using 5 M NaOH, autoclave
and place in the anaerobic cabinet to equilibrate for a minimum
of 24 h. Add the rest of the solutions to a final concentration of
30 mM of glucose, 0.2 mM L-histidine, 100 nM vitamin B12,
6 μM vitamin K3 (menadione), 0.1 mM MgCl2, 50 μM CaCl2,
4.1 mM L-cysteine and 1.4 μM FeSO4 * 7 H2O. Leave the
medium for a minimum of 24 h in the anaerobic cabinet to fully
deoxygenate. Add 2 μM of Protoporphyrin IX freshly made
before using the media. All stock solutions in exception of
Protoporphyrin IX are prepared in advance and autoclaved or
filter sterilized, as below:
l 1.5 M NaCl (autoclaved).
l 0.85 M (NH4)2SO4 (autoclaved).
l 3 M Glucose (filtered).
l 0.2 M L-histidine, store at 4
C (filtered).
l 1 mM vitamin B12, store at 4
C (filtered).
l 6 mM vitamin K3 (menadione), dissolve in ethanol and
store at 20
C (filtered).
l 0.1 M MgCl2 * 6 H2O, store in anaerobic cabinet
(autoclaved).
l 50 mM CaCl2 * 2 H2O, store in anaerobic
cabinet (autoclaved).
l 0.41 M L-cysteine hydrochloride, store in anaerobic cabinet
(filtered).
l 2 mM Protoporphyrin IX.
l 1.4 mM FeSO4 * 7 H2O, store in anaerobic cabinet
(filtered).
2. Anaerobic cabinet.
3. Magnetic stirrer.
4. Sterile magnetic stirring bar.

2.5.3 Supernatant 1. Spectrophotometer (wavelength: 600 nm) with cuvette holder

Collection and cuvettes.
 Production, Isolation, and Characterization of Bioengineered Bacterial. . . 177

2. Refrigerated high speed floor centrifuge, including appropriate

rotor with 500 mL sealable centrifuge bottles.
3. 0.2 μm PES bottle top filter unit (500 mL).
4. Membrane vacuum pump.
5. Sterile 500 mL bottles.

2.6 EV Isolation 1. Filtration cassette Vivaflow 50 R (100,000 MWCO, Hydro-

stat, model VF05H4, Sartorius).
2. Peristaltic pump for running the Vivaflow-unit.
3. Sterile phosphate-buffered saline (PBS), pH 7.4.
4. 0.22 μm PES syringe filters.
5. 5 mL sterile syringes.
6. 15 mL conical centrifuge tubes.
7. 1.5 mL sterile low-bind microcentrifuge tubes.
8. Deionised water.
9. 0.5 M NaOH solution.
10. 10% (v/v) ethanol.

2.7 EVs Purification In this part of the chapter, the use of size-exclusion chromatogra-
phy is described for the removal of remaining proteins. Two
options are proposed: 2.7.1 and 2.7.2 for an increased resolution.

2.7.1 Routine Purification 1. qEVoriginal/35 nm SEC columns (IZON).

2. Support to maintain column in a vertical position.
3. 1.5 mL lo-bind microcentrifuge tubes.
4. Amicon Ultra 0.5 mL centrifugal filters (RC, 10 kDa MWCO).
5. Sterile PBS, pH 7.4.
6. 0.22 μm PES membrane syringe filter.
7. 1 mL sterile syringes.
8. LB and BHIH agar plates.

2.7.2 High-Resolution 1. CL2-B Sepharose.

Fractionation 2. Sterile PBS, pH 7.4.
3. Chromatography column (120 cm  1 cm) (e.g., Econo-Col-
umn® Chromatography Column) in PBS.
4. Chromatography fraction collector.
5. UV spectrophotometer.
6. Vivaspin 20 centrifugal concentrator (100 kDa molecular
weight cutoff).
7. 0.22 μm PES membrane syringe filter.
178 Régis Stentz et al.

2.8 EVs Size 1. Nanoparticle Analyzer (ZetaView TWIN Particle Tracking

and Concentration Analyzer instrument or equivalent).
Analysis 2. Particle-free deionized water.
3. 1 or 10 mL sterile syringes.

2.9 Antigen 1. Proteinase K.

Localization 2. Phenylmethanesulfonyl fluoride (PMSF).
with Proteinase K
3. Water bath.
Assay
4. Sodium Dodecyl Sulphate (SDS).
5. See also materials in Subheadings 2.4.3 and 2.4.4.

2.10 Antigen 1. Recombinant antigen.

Quantification 2. See also materials in Subheadings 2.4.3 and 2.4.4.

3 Methods

3.1 Synthetic Gene The gene can be synthesized de novo using commercial gene syn-
Design thesis services. The N-terminus of the protein of interest is fused in
frame to the signal peptide of the product of BT_3852 (OmpA of
Bt); MKKILMLLAFAGVASVASA. The chimeric protein sequence
is tested in silico for cleavage of the OmpA signal sequence using
http://www.cbs.dtu.dk/services/SignalP/. If unsuccessful,
change or add amino acids as appropriate to the N-terminus of
the gene of interest, downstream from the signal peptide sequence.
To facilitate immunodetection and/or purification of the protein in
downstream applications, a fusion tag can be added to the gene. It
is important that the coding sequence of the desired protein incor-
porates codon usage optimization for expression in Bt, which is
usually provided as part of gene synthesis services. The desired
target sequence is then integrated into an acceptor vector.

3.2 Cloning of Gene 1. Digest the plasmid containing the synthetic gene with restric-
of Interest tion enzymes to excise the gene from the vector carrying the
synthetic gene (e.g., pEX-K168).
3.2.1 Generation
of Recombinant DNA 2. Digest the E. coli/Bacteroides shuttle vector pGH90 (see Sub-
heading 2.1) (see Note 6).
3. Ligate the gene into the digested pGH090 expression vector to
allow translational fusion (see Note 7).

3.2.2 Ransformation 1. Prepare LB agar plates containing ampicillin.

of Competent Cells 2. Thaw one vial of competent MAX Efficiency® DH10B on ice.
3. Gently add 1–5 μL of ligation mixture to the MAX Efficiency®
DH10B.
 Production, Isolation, and Characterization of Bioengineered Bacterial. . . 179

4. Place the vial with the bacteria suspension on ice for 30 min.
5. Induce a heat-shock at 42
C for 30 s.
6. Incubate the bacteria on ice for an additional 5 min.
7. Add 950 μL S.O.C. medium.
8. Incubate at 37
C with agitation (250 rpm).
9. Seed 100 μL of the bacterial suspension onto agar plates con-
taining ampicillin.
10. Incubate the plates for 16–18 h at 37
C.

3.2.3 Screening 1. Pick individual colonies from agar plates and add each colony
of Cloned to a tube containing 5 mL LB medium and ampicillin.
Recombinant DNA 2. Incubate the liquid cultures at 37
C and 250 rpm for 16–18 h.
3. Isolate plasmid DNA from each culture using a small-scale
isolation kit (e.g., QIAprep Spin Miniprep Kit) according to
the manufacturer’s instructions.
4. Digest the plasmid DNA using the appropriate restriction
enzymes (see Note 8).
5. Resolve digested DNA on a TBE gel.
6. Confirm the identity of the plasmids containing the insert of
the expected size by DNA sequencing using appropriate
primers.

3.3 Conjugative 1. Prepare BHIH agar plates either with or without gentamycin
Transfer of Shuttle and erythromycin.
Vector into Bt 2. Grow cultures of the E. coli donor strain with ampicillin (con-
3.3.1 Triparental Mating taining the plasmid with the correct inserted sequence, see
Procedure Subheading 3.2.3) and the E. coli helper strain with trimetho-
prim in 10 mL of LB, at 37
C, under agitation for 16–18 h. In
parallel, grow culture of the Bt recipient strain in BHIH in an
anaerobic cabinet at 37
C for 16–18 h.
3. Inoculate 10 mL of LB with 100 μL of the E. coli donor strain
and the helper strain (no antibiotics added) cultures and incu-
bate for 2 h at 37
C with agitation (e.g., 200 rpm). In parallel,
inoculate in 30 mL BHIH contained in a 50 mL tube with 800
μL of the Bt culture and incubate for 2 h at 37
C in an
anaerobic cabinet.
4. Add donor and helper cultures to the Bt recipient in the 50 mL
tube, mix by vortexing briefly and centrifuge at 2000  g for
15 min at 20
C.
5. Remove the supernatant and resuspend cells in 100 μL of
BHIH. Transfer cell suspension to the surface of a sterile 0.45
μm filter placed on a BHIH agar plate. Incubate the plate
aerobically for 16–18 h at 37
C.
180 Régis Stentz et al.

6. Transfer the filter to a sterile wide-necked Universal bottle and

add 1 mL of BHIH and resuspend the bacterial conjugation
mixture by vortexing thoroughly.
7. Make serial dilutions and plate 100 μL of each dilution and the
non-diluted cell suspensions onto BHIH agar plates containing
gentamycin (to prevent E. coli growth) and erythromycin
(selection of Bt transconjugants).

3.4 Assessing 1. Pick 4 individual colonies and restreak each one on separate
Protein Expression BHIH agar plates containing gentamicin and erythromycin.
and Secretion into 2. Incubate the plates anaerobically at 37
C for 48 h.
BEVs
3. Inoculate bottles containing 20 mL of BHIH with each of the
3.4.1 Culture of Bt 4 isolated clones.
Transconjugants 4. Incubate the bottles anaerobically at 37
C for 48 h.
5. Centrifuge the 20 mL of culture in 50 mL tubes, at 6000  g
for 15 min, at 4
C.
6. Collect the supernatant.
7. Wash cell pellets once in PBS before storing at 20
C prior to
analysis.

3.4.2 Cell Total Protein 1. Resuspend thawed cell pellets in 250 μL of 0.2 M Tris–HCl
Extraction (pH 7.2).
2. Disrupt the cells via sonication using eight 10-s pulses (ampli-
tude, 6 μm), with 30-s pauses on ice between each pulse.
3. Cell extracts are obtained after centrifugation at 14,000  g for
30 min at 4
C and harvesting the supernatant.
4. Measure total protein concentration of each supernatant/sam-
ple using the Bio-Rad protein assay according to the manufac-
turer’s instructions using BSA to generate a standard curve.

3.4.3 BEV Total Protein 1. Filter the 20 mL supernatants through 0.22 μm pore-size PES
Extraction membranes to remove debris and cells.
2. Concentrate the supernatants by ultrafiltration (100 kDa
molecular weight cutoff, Vivaspin 20) to a final volume of
250 μL.
3. Discard the filtrate.
4. Rinse the retentate with 20 mL of 0.2 M Tris–HCl, pH 7.2 and
concentrated to 250 μL.
5. Collect the retentate and disrupt the vesicles via sonication
using eight 10-s pulses (amplitude, 6 μm), with 30-s pauses
on ice between each pulse.
6. Measure the total protein content and concentration using the
Bio-Rad protein assay according to the manufacturer’s instruc-
tions using BSA to generate a standard curve.
 Production, Isolation, and Characterization of Bioengineered Bacterial. . . 181

3.4.4 Protein Western 1. Add BEV and cell extracts obtained in Subheading 3.4.3 to
Blotting/Antigen loading buffer containing 0.4 M freshly prepared dithiothreitol
Immunodetection (DTT).
2. Load 7 μg of the total protein onto a 12% precast gel and
separate by electrophoresis at 180 V for 40 min.
3. Transfer the proteins from the gel onto a polyvinylidene
difluoride (PVDF) membrane using the XCell II™ Blot Mod-
ule or equivalent (according to the manufacturer’s instruc-
tions) at 25 V for 2 h in a solution containing Tris-Glycine
transfer buffer and methanol 20% (v/v).
4. Incubate the membrane with blocking buffer by gently shaking
for 30 min at 20
C using an orbital shaker.
5. Discard the blocking solution and incubate the membrane for
16–18 h at 4
C in blocking buffer containing primary anti-
body (Usually 1:1000 to 1:10000).
6. After washing 3 times with TBST, membranes are incubated
with HRP-conjugated secondary antibody in blocking buffer
for 1 h at 20
C.
7. After 3 washes with TBST, Enhanced Chemiluminescent sub-
strate (ECL) is added to detect bound antibody.

3.4.5 Medium-Scale 1. Inoculate 10 mL of BHIH with a frozen stock for ~16 h.

Bacterial Culture 2. Inoculate 500 mL of BHIH with 0.5 mL of the preinoculum
and Harvesting Conditioned for 17 h (starting OD600 ~ 0.005) with mild stirring.
mediaBHIH Bacterial
Culture

3.4.6 BDM+ Bacterial Initially BHIH was used as a standard medium for generating Bt
Culture BEVs [8, 16]. However, considering the need to reduce and even
exclude animal-derived products from medical/therapeutic formu-
lations to be used in humans, we have modified a chemically defined
Bacteroides growth media (BDM) [20] (see Subheading 2.5.2) for
Bt BEV production.
1. Inoculate 10 mL of BHIH with a frozen stock for ~16 h.
2. Inoculate 10 mL of BDM+ with 100 μL BHIH culture from
step 1 for ~8 h.
3. Inoculate 500 mL of BDM+ with 0.5 mL of the preinoculum
for 17 h (starting OD600 ~ 0.005) with gentle stirring.

3.4.7 Supernatant 1. Collect bacterial cultures at final OD600 1.5–2.5.

Collection 2. Precool centrifuge and canisters for 5 min at 4
C.
3. Decant the culture into two 2 Nalgene™ PPCO centrifuge
bottles and centrifuge at 6037  g for 30 min at 4
C.
182 Régis Stentz et al.

4. Filter-sterilize the supernatant with a 0.22 μm bottle-top filter

unit and transfer the filtrate into a sterile 500 mL bottle (see
Note 9).
5. Samples are stored at 4
C prior to BEVs isolation for up to
24 h.

3.5 Isolation of BEVs During the filtration process, the membrane will retain any mole-
cules >100 kDa including BEVs (retentate) and concentrate them
in the reservoir, whereas molecules <100 kDa will be removed in
the flow-through (filtrate) directly to the waste. BEVs isolation
should be performed at ambient (20–22
C) temperature.
Three procedures are used in a step wise manner for isolating
BEVs, which are illustrated in Fig. 2.

3.5.1 Module Rinsing 1. Set up the system as illustrated in Fig. 2 Option 2.

2. Place 400 mL of deionized water in the reservoir and pump the
liquid through the system at an initial speed setting of 2 to
remove air pockets and then increase the speed setting to 4–5
(200–300 mL/min) until 400 mL has been run through the
system. Collect filtrate in waste bin. Check for any leaks.

3.5.2 Sample 1. Set up the module as illustrated in Fig. 2 Option 1.

Concentration 2. Fill the reservoir with 500 mL filtrate.
3. Pump sample through the system. The initial recirculation
speed setting is 2 for at least 1 cycle and then adjusted to setting
4–5 for sampling (200–300 mL/min). Maximum recirculation
speed setting is 5. Reduce the speed for lower volumes to avoid
foaming.
4. Concentrate the sample until there is ~5 mL left in the system.
5. Switch off the pump.
6. Pour 500 mL of PBS pH 7.4 into the reservoir and pump
through the system at an initial speed setting of 2 to remove
any air pockets and then increase the speed to setting 4–5.
7. Reduce the recirculation speed setting to 1–2 (20–40 mL/
min) to avoid foaming.
8. Switch off the pump with 1–4 mL remaining in the system.
9. Set up the module as illustrated in Fig. 2 Option 3.
10. Start the pump at a speed setting of 1–2 and carefully collect
the concentrated samples in a 15 mL tube or microcentrifuge
tube (use a bigger tube than the volume to be collected to
account for foam).
11. When no more concentrated sample emerges from the tubing,
switch off the pump.
 Production, Isolation, and Characterization of Bioengineered Bacterial. . . 183

Fig. 2 BEV isolation module system. Option 1. Sample concentration. Option 2. Module rinsing/decontamina-
tion. Option 3. Collection of concentrated samples

12. Place 0.5–1 mL of PBS in the reservoir and pump through the
system at speed setting 1–2 to flush out remaining BEVs.
Collect in the original 15 mL tube or microcentrifuge tube.
Final volume collected depends on concentration of BEVs
needed.
13. Switch off the pump.
14. Centrifuge at 15,000  g for 20 min at 4
C to remove any
precipitate.
15. Filter-sterilize the supernatant using a 0.22 μm syringe filter,
collecting the filtrate in sterile 1.5 mL lo-bind microcentrifuge
tubes or 15 mL tube.

3.5.3 Module 1. Set up the system as illustrated in Fig. 2 Option 2.

Decontamination 2. Place 400 mL of deionized water in the reservoir and pump
and Washing liquid through the system at an initial speed setting of 2 to
remove any air pockets and increase the speed setting to 4–5
(200–400 mL/min). Collect filtrate in waste bin.
3. When all the water has been collected, switch off the pump.
4. Set up the system as illustrated in Fig. 2 Option 1.
184 Régis Stentz et al.

5. Place 250 mL of decontamination solution (250 mL 0.5 M

NaOH) and pump it through the system at a speed setting of
3–4 (50–100 mL/min). Allow to recirculate for a minimum of
20 min, then switch off the pump.
6. Set up the system as illustrated in Fig. 2 Option 2.
7. Place 400 mL of deionized water in the reservoir and pump
liquid through the system at an initial speed setting of 2 to
purge any air pockets and increase the speed setting to 4–5
(200–400 mL/min).
8. Switch off the pump.
9. Set up the system as illustrated in Fig. 2 Option 2.
10. Place 250 mL of 10% Ethanol in the reservoir and pump liquid
through the system at an initial speed setting of 2 to purge any
air pockets and increase the speed setting to 3–4 (50–100 mL/
min).
11. When half of the 10% Ethanol has been filtered, switch off the
pump and dismantle the system module.
12. Store cassette membrane in 10% Ethanol at 4
C to avoid
contamination.
13. Rinse the reservoir with water and let it air dry. Reservoir and
tubing are left to dry at 20
C.
BEV isolation waste is discarded in the sink drain.

3.6 BEVs Purification Contaminants (e.g., proteins) of BEV preparations can be removed
by size exclusion chromatography (SEC). We recommend using the
method described in Subheading 3.7, step 1 for routine prepara-
tions and in Subheading 3.7, step 2 if increased resolution is
needed, for instance, to size fractionate BEVs.

3.6.1 Routine Purification 1. Bring the BEVs preparation, column buffer (PBS pH 7.4) and
qEVoriginal/35 nm SEC column(s) to ~20
C. Do not remove
the column caps until operational temperature is reached.
2. Secure the column in a vertical position using a stand.
3. Carefully remove the column top-cap.
4. Attach a column reservoir (if available) and add 1.5 column
volume of buffer (PBS 1, 15 mL).
5. Remove the bottom column cap and allow the buffer to run
under gravity to waste.
6. If any buffer other than PBS is used, flush with at least 3 column
volumes of the buffer (>30 mL).
7. The column will stop flowing when the buffer has entered the
loading frit.
8. Load 0.5–1 mL of isolated BEVs onto the loading frit.
 Production, Isolation, and Characterization of Bioengineered Bacterial. . . 185

9. Collect the void volume (3 mL) fraction(s) into 1.5 mL lo-bind

microcentrifuge tubes.
10. Allow the sample to completely run into the column.
11. Top up the reservoir with 15 mL buffer (PBS 1) and collect
0.5–1 mL elution fractions.
12. For a loading volume of 1 mL of BEVs and collecting 0.5 mL
elution fractions, BEVs will elute in fractions 7–12 with pro-
teins eluting in fractions 10–20.
13. After the eluted fractions have been collected, flush the column
with 1.5 volumes of buffer (PBS 1, 15 mL) before loading
another sample.
14. If storing the column, flush with buffer containing 20% ethanol
or 0.05% sodium azide.
15. Store the column at 4
C to avoid contamination.
16. Filter-sterilize the BEVs using a 0.22 μm syringe filter, collect-
ing the filtrate in sterile 1.5 mL lo-bind microcentrifuge tubes
or 15 mL tube.
17. Pool BEVs elution fractions and concentrate to desired volume
using Amicon Ultra 0.5 mL centrifugal filters (RC,
10 kDa MWCO).

3.6.2 High-Resolution See Note 10.

Size Fractionation

3.7 BEVs Size Size and concentration of isolated BEVs suspension is determined
and Concentration by nanoparticle tracking analysis (NTA) using a suitable NTA
Analysis instrument. The protocol described below is for the ZetaView
PMX-220 TWIN instrument from Particle Metrix GmbH.
1. Prepare instrument set up according to the manufacturer’s
instructions.
2. Inject 10–20 mL particle-free deionized water using a 10 mL
syringe to perform cell quality check; avoid injecting air
bubbles.
3. Inject 5–10 mL of 1:25,000 100 nm polystyrene standard bead
suspension to perform focus auto-alignment.
4. Inject 10–20 mL particle-free deionized water using a 10 mL
syringe to rinse instrument; avoid injecting air bubbles.
5. Dilute aliquots of BEVs suspension in 1:1000 to 1:20,000 in
particle-free deionized water.
6. Inject 1 mL sample with a syringe; avoid injecting air bubbles.
186 Régis Stentz et al.

7. Acquire size distribution video data using the following

settings.
– temperature: 25
C;
– frames: 60;
– duration: 2 s;
– cycles: 2;
– positions: 11;
– camera sensitivity: 80; and,
– shutter value: 100.
8. Analyse data using the ZetaView NTA software (version
8.05.12) with the following post acquisition settings:
– minimum brightness: 20;
– max area: 2000;
– min area: 5; and,
– trace length: 30.

3.8 Proteinase K To establish if heterologous proteins are expressed in the lumen or

Assay at the surface of BEVs we use broad-spectrum proteinase
K. Proteinase K digests proteins exposed at the surface of BEVs
but not in the lumen. Extracts of BEVs obtained in the presence or
absence of Proteinase K samples and analyzed by Western blotting
using antibodies specific for the heterologous protein makes it
possible to distinguish between surface- and lumen-expressed anti-
gens (see Note 11).
1. Add 100 mg/L of proteinase K into intact 1011 BEVs/mL or
solubilized (in 1% SDS) and incubate for 1 h in a water bath at
37
C.
2. The activity of proteinase K is stopped by addition of
1 mM PMSF.
3. Load samples onto a 12% polyacrylamide gel and perform a
Western blot following steps described in Subheading 3.4.4.

3.9 Antigen The amount of antigen expressed in BEVs is readily determined by

Quantification Western blotting using serial dilutions of the recombinant antigen
and comparing the intensity of the bands visualized on the blot to
estimate the concentration of antigen in the BEVs using Image Lab
Software (Bio-Rad).
1. Prepare serial dilutions of the recombinant antigen and the
isolated BEVs.
2. Load samples onto a 12% polyacrylamide gel and perform a
Western blot (Subheading 3.4.4).
3. Analyse the image of the blot using the quantity tool of Image
Lab software (Bio Rad).
 Production, Isolation, and Characterization of Bioengineered Bacterial. . . 187

4 Notes

1. Immunization: Details of immunization protocols are provided

in Carvalho et al. [8, 16]. In short, animals receive a primary
immunization of filter-sterilized BEVs (see Note 12) via the
nasal or oral route with booster immunizations carried out
7–14 days late with an infectious challenge following after a
further 7–10 days. At necrosis body fluids and tissues are
harvested for downstream analyses of antibody and immune
cell profiles and histopathology.
2. Sample collection and antibody profiles: Serum bronchoalveo-
lar lavage and saliva samples are routinely used in ELISAs to
identify and quantify antigen-specific IgA and IgG antibodies
as described in Carvalho et al. [8, 16]. In short, the ELISAs
include coating the plate with recombinant protein for 16 h at
4
C. After washing and incubating with blocking solution,
serial dilutions of samples are added and incubated for 16 h at
4
C. After washing, a secondary antibody conjugated to HRP
is added for 1 h at 20
C. A chromogenic substrate is then
added and absorbance at 450 nm is recorded using a spectro-
photometer. Tissue homogenates (e.g., salivary glands and
lungs) can also be used in this antibody detection assay.
3. One possibility is to use the plasmid-borne inducible gene
expression system developed for Bt that is based upon a Bt
endogenous mannan-inducible promoter [21]. This system
allows to create translational fusions and generate protein pro-
ducts with the possibility of adding a C-terminal
polyhistidine tag.
4. A 6x-His Tag antibody should be used if the chimeric protein
contains a His-tag sequence. An alternative is to use antibodies
raised against epitopes of the antigenic protein either commer-
cially available or from other sources.
5. The secondary antibody can be conjugated to with various
enzyme, fluorescent proteins, biotin or to polymers. The sec-
ondary anti-Ig antibody must have specificity for the immuno-
globulins present in the species in which the primary antibody
was raised.
6. The digestion of the E. coli/Bacteroides shuttle vector pGH090
with NcoI and EcoRI is provided as an example. In the case of
DNA sequence constraints or if the EcoRI restriction site
cannot be removed from the internal sequence of the synthetic
gene, other restrictions sites can be used for the design of the
30 -end of the fragment (e.g., BamHI or SmaI) that are located
downstream from the NcoI site of pGH090 [18].
188 Régis Stentz et al.

7. If necessary, the DNA fragment containing the gene of interest

can be extracted after excision of the band at the expected size
from an agarose gel after electrophoresis using the QIAquick
Gel Extraction Kit (Qiagen) or equivalent, following the man-
ufacturer’s instructions.
8. As an example, we describe the cloning of the gene encoding
H5F the highly conserved stalk region of the hemagglutinin
molecule of IAV strain H5N1 (VN/04:A/ VietNam/1203/
04 [17]) into the Bacteroides expression vector pGH090
[8]. The synthetic gene fused at its 50 -end with a signal peptide
sequence was designed to contain a BspHI restriction site at its
50 -end and an EcoRI restriction site at its 30 -end to enable
DNA cloning. BspHI was chosen because a lysine residue
follows the first amino acid (methionine) in the sequence of
the signal peptide. Therefore, the lysine AAA/G codon which
starts with an A and follows the ATG start codon is included in
the BspHI (TCATGA) restriction sequence. BspHI restriction
enzyme generates a cohesive end compatible with the NcoI
cohesive end of the restricted pGH090 vector to allow transla-
tion fusion of the synthetic gene. As a result, the BspHI site was
lost and a combination of EcoRI and a site located within the
vector and outside of the gene of interest was used to digest
recombinant plasmids.
9. BEV sterility is confirmed by checking for the growth of any
contaminating bacterial cells. Spread 100 μL of the filter-
sterilized solution on BHIH agar plates, incubate in an anaero-
bic cabinet for 48 h at 37
C and confirm the absence of
colonies.
10. To increase the resolution of the SEC procedure and obtain a
better separation of vesicles of different sizes, the SEC can be
performed using a 120 cm  1 cm column (Econo-Column®
Chromatography Columns, Bio-Rad) filled with 90 mL of
CL2-B Sepharose (Sigma-Aldrich) [22]. The absorbance of
the fractions is measured at 280 nm and the first fractions
displaying an absorbance peak are pooled. Pooled fractions
are concentrated to 1 mL with a Vivaspin 20 centrifugal con-
centrator (100 kDa molecular weight cutoff, Sartorius) and the
retentate is filtered through a 0.22 μm PES membrane (Sarto-
rius). The concentration of the vesicles can then be determined
using nanoparticle tracking analysis as described in Subheading
3.7, step 1.
11. If the protein is expressed on the surface of the BEV Proteinase
K will degrade it and the band will be absent on the immuno-
blot. If the protein is expressed in the lumen, the band will still
be evident. SDS-treatment of vesicles makes their luminal con-
tents accessible to Proteinase K and serves as control for
enzyme activity and confirmation of the protein being
expressed in the lumen of BEVs.
 Production, Isolation, and Characterization of Bioengineered Bacterial. . . 189

12. The sterility of BEV suspensions stored at 4
C is examined by

checking for growth of any contaminating bacterial cells prior
to immunization. Add 90 μL of sterile BHIH broth to 10 μL of
BEV suspension and spread the 100 μL onto BHIH agar
plates. Incubate in both aerobic and anaerobic conditions for
48 h at 37
C and check for the absence of colony.

Acknowledgments

This work was supported in part by the UK Biotechnology and

Biological Sciences Research Council (BBSRC) under grant num-
bers BB/J004529/1, BB/R012490/1, and BBS/E/
F000PR10355. We thank Dr. Emily Jones for her contribution in
setting up methods for SEC purification and the use of the Zeta-
view NTA, and Dr. Rokas Juodeikis for contributing to the devel-
opment of BDM+ medium.

References

1. Rauch S, Jasny E, Schmidt KE et al (2018) Release 286:167–178. https://doi.org/10.

New vaccine technologies to combat outbreak 1016/j.jconrel.2018.07.035
situations. Front Immunol 9:1963. https:// 8. Carvalho AL, Fonseca S, Cross K et al (2019)
doi.org/10.3389/fimmu.2018.01963 Bioengineering commensal bacteria-derived
2. Wallis J, Shenton DP, Carlisle RC (2019) outer membrane vesicles for delivery of biolo-
Novel approaches for the design, delivery and gics to the gastrointestinal and respiratory
administration of vaccine technologies. Clin tract. J Extracell Vesicles 8:1632100. https://
Exp Immunol 196:189–204. https://doi. doi.org/10.1080/20013078.2019.1632100
org/10.1111/cei.13287 9. Chen DJ, Osterrieder N, Metzger SM et al
3. Francis MJ (2018) Recent advances in vaccine (2010) Delivery of foreign antigens by engi-
technologies. Vet Clin North Am Small Anim neered outer membrane vesicle vaccines. Proc
Pract 48:231–241. https://doi.org/10.1016/ Natl Acad Sci U S A 107:3099–3104. https://
j.cvsm.2017.10.002 doi.org/10.1073/pnas.0805532107
4. Shin MD, Shukla S, Chung YH et al (2020) 10. Stentz R, Carvalho AL, Jones EJ et al (2018)
COVID-19 vaccine development and a poten- Fantastic voyage: the journey of intestinal
tial nanomaterial path forward. Nat Nanotech- microbiota-derived microvesicles through the
nol 15:646–655. https://doi.org/10.1038/ body. Biochem Soc Trans 46:1021–1027.
s41565-020-0737-y https://doi.org/10.1042/BST20180114
5. Schwechheimer C, Kuehn MJ (2015) Outer- 11. Cecil JD, Sirisaengtaksin N, O’Brien-Simpson
membrane vesicles from gram-negative bacte- NM et al (2019) Outer membrane vesicle-host
ria: biogenesis and functions. Nat Rev Micro- cell interactions. Microbiol Spectr 7. https://
biol 13:605–619. https://doi.org/10.1038/ doi.org/10.1128/microbiolspec.PSIB-
nrmicro3525 0001-2018
6. Arigita C, Jiskoot W, Westdijk J et al (2004) 12. Jones EJ, Booth C, Fonseca S et al (2020) The
Stability of mono- and trivalent meningococcal uptake, trafficking, and biodistribution of Bac-
outer membrane vesicle vaccines. Vaccine teroides thetaiotaomicron generated outer
22:629–642. https://doi.org/10.1016/j.vac membrane vesicles. Front Microbiol 11:57.
cine.2003.08.027 https://doi.org/10.3389/fmicb.2020.00057
7. Kanojia G, Raeven RHM, van der Maas L et al 13. Miquel-Clopés A, Bentley EG, Stewart JP et al
(2018) Development of a thermostable spray (2019) Mucosal vaccines and technology. Clin
dried outer membrane vesicle pertussis vaccine Exp Immunol 196:205–214. https://doi.org/
for pulmonary immunization. J Control 10.1111/cei.13285
190 Régis Stentz et al.

14. Smith CJ, Rocha ER, Paster BJ (2006) The 19. Shoemaker NB, Getty C, Gardner JF et al
medically important Bacteroides spp. in health (1986) Tn4351 transposes in Bacteroides spp.
and disease. In: Dworkin M, Falkow S, and mediates the integration of plasmid R751
Rosenberg E, Schleifer KH, Stackebrandt E into the Bacteroides chromosome. J Bacteriol
(eds) The prokaryotes. Springer, New York, 165:929–936. https://doi.org/10.1128/jb.
NY. https://doi.org/10.1007/0-387-30747- 165.3.929-936.1986
8_14 20. Bryant WA, Stentz R, Le Gall G et al (2017) In
15. Human Microbiome Project Consortium silico analysis of the small molecule content of
(2012) Structure, function and diversity of outer membrane vesicles produced by Bacter-
the healthy human microbiome. Nature oides thetaiotaomicron indicates an extensive
486:207–214. https://doi.org/10.1038/ metabolic link between microbe and host.
nature11234 Front Microbiol 8:2440. https://doi.org/10.
16. Carvalho AL, Miquel-Clopés A, Wegmann U 3389/fmicb.2017.02440
et al (2019) Use of bioengineered human com- 21. Horn N, Carvalho AL, Overweg K et al (2016)
mensal gut bacteria-derived microvesicles for A novel tightly regulated gene expression sys-
mucosal plague vaccine delivery and immunisa- tem for the human intestinal symbiont Bacter-
tion. Clin Exp Immunol 196:287–304. oides thetaiotaomicron. Front Microbiol
https://doi.org/10.1111/cei.13301 7:1080. https://doi.org/10.3389/fmicb.
17. Valkenburg SA, Mallajosyula VV, Li OT et al 2016.01080
(2016) Stalking influenza by vaccination with 22. Durant L, Stentz R, Noble A et al (2020)
pre-fusion headless HA mini-stem. Sci Rep Bacteroides thetaiotaomicron-derived outer
6:22666. https://doi.org/10.1038/ membrane vesicles promote regulatory den-
srep22666 dritic cell responses in health but not in inflam-
18. Wegmann U, Horn N, Carding SR (2013) matory bowel disease. Microbiome 8:88.
Defining the Bacteroides ribosomal binding https://doi.org/10.1186/s40168-020-
site. Appl Environ Microbiol 79:1980–1989. 00868-z
https://doi.org/10.1128/AEM.03086-12
 Chapter 12

Membrane Vesicles Produced by Shewanella vesiculosa

HM13 as a Prospective Platform for Secretory Production
of Heterologous Proteins at Low Temperatures
Jun Kawamoto and Tatsuo Kurihara

Abstract
Extracellular membrane vesicles (EMVs) produced by Gram-negative bacteria are useful as a vaccine
platform. During growth in broth at 18
C, Shewanella vesiculosa HM13 produces a large number of
EMVs that contain a 49-kDa major cargo protein, named P49. Enhanced green fluorescent protein fused to
the C-terminus of P49 is delivered to EMVs, suggesting that P49 is useful as a carrier to target foreign
proteins to EMVs for production of artificial EMVs with desired functions. This method is potentially
useful for the preparation of designed vaccines and is described in detail in this chapter.

Key words Extracellular membrane vesicles, Protein fusion, Recombinant protein production, Pro-
tein secretion, Protein transport

1 Introduction

Gram-negative bacteria produce extracellular membrane vesicles

(EMVs), which play various physiological roles such as intercellular
communication, defense, pathogenicity, and catalysis [1–3]. Given
their high immunogenicity and nonreplicating nature, these vesi-
cles are useful as vaccines [4, 5]. By modifying the composition of
EMVs, it is possible to develop artificial EMVs with desired immu-
nogenic properties [6, 7]. A method for heterologous protein
transport to EMVs would contribute to the development of such
engineered EMVs.
Shewanella vesiculosa HM13 is a Gram-negative bacterium
isolated from the intestine of horse mackerel and is a prospective
host for secretory production of heterologous proteins at low
temperatures, which may be beneficial for improving stability
and/or decreasing toxicity of some expressed proteins [8–
11]. S. vesiculosa HM13 produces a large number of EMVs that
contain a 49-kDa protein, named P49, of unknown function as

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_12,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

191
192 Jun Kawamoto and Tatsuo Kurihara

Fig. 1 EMVs and their cargo proteins from S. vesiculosa HM13. (a) Field emission-scanning electron
microscopy image of S. vesiculosa HM13. Blebs on the cell surface and vesicles secreted into the extracellular
milieu were observed. White triangles indicate blebbing of the outer membrane to produce EMVs. Bar
indicates 100 nm. (b) Transmission electron microscopy (TEM) observation of EMVs. TEM images demonstrate
that S. vesiculosa HM13 produces EMVs with a uniform diameter of approximately 50 nm. The inset
represents a magnified image of the boxed area. Bar indicates 500 and 50 nm, respectively. (c) SDS-PAGE
gel image of PVF and purified EMVs. The EMVs of S. vesiculosa HM13 carry a single major cargo protein, P49.
(Parts of this figure have been reproduced/modified from [8], with permission from Frontiers Media SA)

their single major cargo protein (Fig. 1). P49 is expected to be

useful as a carrier to transport heterologous proteins to EMVs to
obtain engineered EMVs with desired properties. As a proof of
concept, we fused enhanced green fluorescent protein (eGFP) to
the C-terminus of P49 and expressed it in S. vesiculosa HM13
[8]. We found that the fusion protein was delivered to EMVs.
Although there is room for improvement in yield and purity, this
system may serve as a new platform for the preparation of functio-
nalized EMVs. Here, we describe the experimental details of the
construction of the recombinant S. vesiculosa HM13 strain produc-
ing the P49-eGFP fusion protein, preparation of EMVs, and analy-
sis of the localization of the fusion protein.

2 Materials

2.1 Equipment 1. Cryopreservation beads (see Note 1).

2. 1.5 mL tubes.
3. 50 mL tubes.
4. 500 mL flasks.
5. 0.22-μm polyethersulfone (PES) syringe filters.
6. 0.45-μm PES syringe filters.
7. Syringes.
8. Temperature-controlled shaker with cooling function.
 Low Temperature Production of Engineered Vesicles 193

9. Reciprocal shaker incubator.

10. UV-visible spectrophotometer.
11. Ultracentrifuge with a 12  10 mL angle rotor Ti-50.
12. Polycarbonate ultracentrifuge tubes, 16  76 mm.
13. Low-speed centrifuge.
14. Ultrasonic probe sonicator.
15. Polyacrylamide gel electrophoresis (SDS-PAGE)-associated
instruments.
16. Semidry blotter.
17. Filter paper.
18. Hydrophobic polyvinylidene fluoride (PVDF) transfer
membrane.
19. Benchtop reciprocal shaker.
20. Chemiluminescence image scanner.
21. Polymerase chain reaction (PCR) machine.
22. Agarose gel electrophoresis apparatus.
23. Heat block incubator.

2.2 Solutions 1. Rifampicin stock solution: 50 mg/mL in DMSO. Store at

and Reagents 30
C under dark conditions.
2. Kanamycin stock solution: 50 mg/mL in water. Filter with a
0.22-μm PES syringe filter. Store at 30
C.
3. EMV suspension buffer (DPBSS): Dulbecco’s phosphate-
buffered saline (2.7 mM KCl, 8.9 mM Na2HPO4·7H2O,
1.5 mM KH2PO4, and 135.9 mM NaCl, pH 7.2) with addi-
tional 0.2 M NaCl. Filter with a 0.22-μm syringe filter.
4. Trichloroacetic acid (TCA) solution (100% w/v).
5. Acetone (prechilled at 30
C).
6. Precast 5–20% gradient polyacrylamide gels.
7. 2 SDS-sample buffer: 125 mM Tris–HCl (pH 6.8), 4% SDS,
20% glycerol, 0.04% bromophenol blue, and 10%
2-mercaptoethanol.
8. 10 SDS-running buffer: 250 mM Tris base, 1.9 M glycine,
and 1% SDS.
9. Protein transfer anode buffer 1: 300 mM Tris-base, 0.05%
SDS, and 20% methanol.
10. Protein transfer anode buffer 2: 25 mM Tris-base, 0.05% SDS,
and 20% methanol.
11. Protein transfer cathode buffer: 25 mM Tri-base, 0.05% SDS,
40 mM 6-aminohexanoic acid, and 20% methanol.
194 Jun Kawamoto and Tatsuo Kurihara

12. Tris-buffered saline (TBS): 137 mM NaCl, 2.68 mM KCl,

25 mM Tris–HCl (pH 7.4).
13. TBS-T: 0.05% (w/v) Tween 20 in TBS.
14. Blocking solution: 5% (w/v) skim milk in TBS-T.
15. Primary antibody: rabbit anti-GFP.
16. Secondary antibody: goat anti-rabbit IgG (H + L)-HRP
conjugate.
17. Chemiluminescent substrate for horseradish peroxidase (HRP)
enzyme (see Note 2).

2.3 Bacterial Strains, 1. S. vesiculosa HM13-Rifr (see Note 3).

Plasmid, 2. S. vesiculosa P49-eGFP.
and Culture Media
3. Competent Escherichia coli S17–1/λpir (see Note 4).
4. pKNOCK plasmid (see Note 5).
5. pGreen plasmid (see Note 6).
6. Luria Bertani (LB) medium: Dissolve 10 g tryptone, 5 g yeast
extract, and 10 g NaCl in 1 L water and adjust the pH to 7.0
with NaOH. Autoclave and store at room temperature (RT).
7. LB agar: Mix 15 g agar powder with 1 L of LB liquid medium,
prepared as described above. Autoclave, allow to cool to 56
C,
supplement as required, and pour into sterile Petri plates. Store
at 4
C (see Note 7).

3 Methods

3.1 Construction of a All PCRs are performed in 50 μL reactions using a high-fidelity

P49-eGFP Fusion proof reading polymerase (see Note 8), according to manufac-
Protein Expression turer’s protocol. Details for PCR primers used, including indication
Strain of complementary sequences to allow fusion of the amplicons, are
shown in Table 1.
3.1.1 Construction
of pKP49eGFP 1. Amplify a 523 bp-fragment of the 30 -terminal region of the P49
and Transformation into gene (without stop codon) using genomic DNA from
Conjugal Donor Strain S. vesiculosa HM13 and primers P49C-fwd/P49C-rev. With
the addition of 15 bp overhangs on both primers to allow
directional fusion to amplified eGFP and pKNOCK sequences,
the total amplicon size is 553 bp.
2. Amplify the entire eGFP gene from the pGreen plasmid using
primers GFP-fwd/GFP-rev. With the addition of a 15 bp over-
hang to the GFP-rev primer to allow fusion to the linear
pKNOCK sequence, the total amplicon size is 732 bp. No
additional overhang is added to the GFP-fwd primer, but the
first 15 bp are complementary to the overhang on the P49C-
rev primer.
 Low Temperature Production of Engineered Vesicles 195

Table 1
Primers used in PCR reactions

a
Complementary sequences used to fuse amplicons for construction of pKP49eGFP are indicated, where appropriate, by
matched text color

3. Amplify the complete 2,096 bp linear sequence of the

pKNOCK plasmid by inverse-PCR using primers pKNOCK-
fwd/pKNOCK-rev.
4. Check the correct amplification of PCR products by analyzing
5 μL each on a 1% agarose gel, visualized with 0.5 mg/L
ethidium bromide solution.
5. Purify each specific amplicon using a silica column PCR
cleanup kit and elute in 25 μL RNase-free water (see Note 9).
6. Assemble the DNA fragments using a ligation-free DNA
assembly kit (see Note 10) to generate pKP49eGFP (Fig. 2).
7. After fusion reaction, add 50 μL of competent E. coli S17–1/λ
pir cells.
8. Incubate on ice for 30 min.
9. Incubate at 42
C in a heat block incubator for 30 s to 1 min.
10. Immediately put the tube on ice and incubate for 3 min.
11. Add 300 μL LB medium and incubate at 37
C and 200 rpm
for 1 h on a reciprocal shaker.
12. Spread the 300 μL culture on LB agar containing 50 μg/mL
kanamycin.
13. Incubate at 37
C for overnight.
14. Isolate single colonies and confirm transformants by colony
PCR using primers P49C-fwd/GFP-rev, with correct clones
amplifying a 1,270 bp fragment.
196 Jun Kawamoto and Tatsuo Kurihara

Kanr
Kanr
oriT

pKNOCK-Km pKP49eGFP
oriT R6K
Plasmid eGFP -ori
construction 3 -region of P49-
R6K-ori coding gene w/o
stop codon

Conjugative transformation
& homologous recombination

S. vesiculosa HM13-Rifr
genome stop

Selection with Rif and Kan

P49-eGFP
genome Kanr
stop stop
P49-gene 3 - pKNOCK-specific primer
region primer (pK-check-rev)
(P49C-fwd)

Fig. 2 Schematic illustration of the construction of the eGFP-fused P49 expression strain of S. vesiculosa
HM13. A 523 bp sequence corresponding to the 30 end of the P49 gene was fused to the 717 bp eGFP gene,
and cloned into the pir-dependent suicide plasmid, pKNOCK. Integration of the plasmid into the genome by
homologous recombination was confirmed by PCR with a set of primers annealing to the P49 gene and the
pKNOCK-specific region

15. For further confirmation, the plasmids were extracted from the
transformants and analyzed by sequencing using primers
pK-check-fwd/pK-check-rev.
16. Store glycerol stocks of confirmed culture at 80
C until use
as plasmid-donor cells.

3.1.2 Conjugal Transfer 1. Inoculate a cryopreservation bead of S. vesiculosa HM13-Rifr

of pKP49eGFP from E. coli (plasmid-recipient) onto an LB agar plate containing
S17–1/λpir into 50 μg/mL rifampicin. Incubate at 18
C until single colonies
S. vesiculosa HM13-Rifr are formed.
2. At the same time, plate out the plasmid donor strain E. coli
S17–1/λpir carrying pKP49eGFP, prepared in Subheading
3.1.1, onto LB agar containing 50 μg/mL kanamycin. Incu-
bate overnight at 37
C.
3. Inoculate a single colony of S. vesiculosa HM13-Rifr into 5 mL
LB containing 50 μg/mL rifampicin. Incubate overnight at
18
C and 180 rpm until the OD600 reaches 1–2.
 Low Temperature Production of Engineered Vesicles 197

4. At the same time, inoculate a single colony of E. coli S17–1/λ

pir carrying pKP49eGFP into 5 mL LB medium containing
50 μg/mL kanamycin. Incubate overnight at 37
C and
185 rpm.
5. For both donor and recipient strains, subculture 50 μL of the
overnight growth into fresh 5 mL aliquots of LB containing the
appropriate antibiotic and continue to incubate at the appro-
priate temperature for each. For the culture of E. coli S17–1/λ
pir carrying pKP49eGFP, incubate at 37
C and 180 rpm until
the OD600 reaches 0.5–1 (3–4 h). For the recipient S. vesiculosa
HM13-Rifr culture, incubate at 18
C and 180 rpm until the
OD600 reaches about 1 (approx. 12 h).
6. Mix 100 μL aliquots of donor and recipient cultures by pipet-
ting in a 1.5 mL tube and centrifuge for 5 min at 5,000  g and
18
C to pellet cells.
7. Remove the supernatant and add 150 μL of fresh LB to gently
resuspend the cell pellet.
8. Drop the mixture onto the center of an LB agar plate contain-
ing no antibiotics and allow it to dry in a Class II safety cabinet
for about 30 min without spreading.


9. Incubate at 18 C for overnight for conjugative plasmid
transfer.
10. Transfer the cell lawn to a 1.5 mL tube containing 1 mL LB
and suspend the cells.
11. Inoculate 100–200 μL of the cell suspension onto an LB agar
plate containing 50 μg/mL rifampicin and 50 μg/mL
kanamycin.
12. Incubate at 18
C until single colonies are formed.
13. Isolate representative colonies for verification of correct clones.

3.1.3 Verification 1. Prepare a standard Taq polymerase PCR reaction mixture con-
of Plasmid Integration into taining a P49C-fwd and pK-check-rev (Fig. 2), allowing
the S. vesiculosa 10–20 μL per colony to be tested.
HM13-Rifr Chromosome 2. Aliquot 10–20 μL PCR reaction mixture, and mix in single
colonies to be tested, in respective PCR tubes. Amplify
30 cycles using manufacturer’s recommendations for times
and temperatures for each step.
3. Analyze 10 μL of each PCR product by 1% agarose gel
electrophoresis.
4. Check the size of amplified gene fragments, which should be
1,304 bp for correct clones.
5. Isolate single colonies for which targeted plasmid integration
was detected and preserve confirmed S. vesiculosa P49-eGFP
cultures using cryopreservation beads at 80
C until use.
198 Jun Kawamoto and Tatsuo Kurihara

3.2 Bacterial Culture 1. Inoculate a cryopreservation bead of S. vesiculosa P49-eGFP

for EMV Preparation onto an LB agar containing 50 μg/mL rifampicin and 50 μg/
mL kanamycin.
2. Incubate at 18
C until single colonies are formed.
3. Inoculate a single colony into 5 mL LB containing 50 μg/mL
rifampicin and 50 μg/mL kanamycin.
4. Incubate at 18
C and 180 rpm until the OD600 reaches 1–2
(about 16 h).
5. Inoculate 50 μL of the seed culture into fresh 5 mL LB contain-
ing 50 μg/mL rifampicin and 50 μg/mL kanamycin (see
Note 11).
6. Incubate at 18
C and 180 rpm until the OD600 reaches 2–3
(about 24 h) (see Note 12).
7. Subject the cell culture to the subsequent EMV preparation.

3.3 EMV Preparation 1. Transfer the 5 mL culture to 1.5 mL tubes and pellet the
bacterial cells by centrifugation at 6,800  g and 4
C for
15 min.
2. Carefully transfer the supernatants to new 1.5 mL tubes.
3. Collect the cell pellets into a single 1.5 mL tube and wash the
combined pellet three times with 1 mL of fresh LB medium,
discarding the supernatant after each wash (see Note 13).
4. Keep cells on ice and use them for soluble and insoluble protein
preparation described in Subheading 3.4.1, or store at 80
C
until use.
5. Centrifuge the supernatants from step 2 at 13,000  g and
4
C for 15 min to remove possible contaminants such as
broken cell debris in EMV-containing fraction.
6. Filter the supernatants through a 0.45-μm PES syringe filter to
remove residual cellular materials.
7. Transfer the filtrates into a 16  76 mm polycarbonate ultra-
centrifuge tube, prepare a weight-matched balance tube, and
ultracentrifuge at 100,000  g and 4
C for 2 h (see Note 14).
8. Transfer the supernatant (see Note 15) to new 1.5 mL tubes
and store as a post-vesicle fraction (PVF) at 80
C until use.
9. Add 1 mL DPBSS to the ultracentrifuge tubes and suspend the
pellet (see Note 16).
10. Collect the EMV suspension into 1.5 mL tubes (see Note 17)
for Western blotting, or store at 80
C until use.
 Low Temperature Production of Engineered Vesicles 199

3.4 Localization 1. Suspend the cells (prepared in Subheading 3.3, step 4) in 10
Analysis of P49-Fusion volume of DPBSS (typically 300–500 μL for cells from a 5 mL
Protein culture sample).
3.4.1 Preparation
2. Sonicate cells on ice for 3 min (0.5 s ON, 1 s OFF, at 20%
of Cellular Protein Samples amplitude).
3. Transfer the suspension to a new 1.5 mL tube and pellet
undisrupted cells at 20,500  g and 4
C for 15 min.
4. Transfer the supernatant into a 16  76 mm polycarbonate
ultracentrifuge tube, add 5 mL DPBSS, and prepare a weight-
matched balance tube (see Note 14).
5. Ultracentrifuge at 100,000  g and 4
C for 2 h.
6. Collect the supernatant as soluble protein fraction and store at
80
C for Western blot analysis.
7. Add 50 μL DPBSS into the ultracentrifuge tube to resuspend
the pellet by pipetting or gently vortexing.
8. Transfer the suspension to a new 1.5 mL tube.
9. Store the suspension as insoluble protein fraction at 80
C for
Western blot analysis.

3.4.2 Sample 1. Add 1/10 volume of 100% TCA to each of the four samples
Preparation for SDS-PAGE (i.e., soluble and insoluble cell protein fractions, PVF,
and Western Blotting and EMVs).
2. Vortex and incubate on ice for 30 min.
3. Centrifuge at 20,500  g and RT for 30 min.
4. Discard supernatants.
5. Add 20 μL prechilled acetone to each tube and wash by
pipetting.
6. Centrifuge at 20,500  g and RT for 30 min.
7. Discard supernatants and repeat the wash step twice.
8. Air-dry the pellets for 5 min.
9. Add 40 μL of 1 SDS-sample buffer to each tube and suspend
samples by pipetting.
10. Boil the suspensions at 100
C for 5 min and cool at RT.

3.4.3 SDS-PAGE 1. For each sample prepared in Subheading 3.4.2, mix 2.4 μL
and Western Blotting with 27.6 μL of 1 SDS-sample buffer and load 10 μL
(corresponding to 100 μL culture) to respective lanes of a
precast 5–20% gradient polyacrylamide gel.
2. Run the gel at 120 V in 1 SDS-running buffer until the dye
front reaches the bottom of the gel.
3. For semidry Western-blot transfer of the proteins (see Note
18), cut the PVDF membrane to the size of the separation
200 Jun Kawamoto and Tatsuo Kurihara

gel, wash in methanol and soak in 30–50 mL Protein transfer

cathode buffer for 10 min.
4. Soak the gel in 30–50 mL of Protein transfer cathode buffer for
10 min.
5. Prepare three sheets of gel-sized, thick absorbent filter paper
(see Note 19) by soaking one in Protein transfer anode buffer
1, another in Protein transfer anode buffer 2, and the third in
Protein transfer cathode buffer, each for 10 min.
6. Place the filter paper soaked in Protein transfer anode buffer
1 on a semidry blotter, followed by the filter paper soaked in
Protein transfer anode buffer 2.
7. Place the PVDF membrane on the filter papers on the blotter,
followed by the gel and then the filter paper incubated in
Protein transfer cathode buffer.
8. Run the transfer at 0.23 A for 30 min.
9. Following protein transfer, immerse the membrane in 40 mL
blocking solution for 1 h on a reciprocal benchtop shaker at RT
(or overnight at 4
C).
10. Rinse the membrane three times in 40 mL TBS-T for 10 min
each at RT.
11. Incubate the membrane in 20 mL primary antibody solution,
that is, rabbit anti-GFP antibody diluted 1:10,000 in blocking
solution (see Note 20), for 1 h on a reciprocal benchtop shaker
at RT.
12. Rinse the membrane three times in 40 mL TBS-T for 10 min
each at RT.
13. Incubate the membrane in 50 mL secondary antibody solu-
tion, that is, goat anti-rabbit IgG (H + L)-HRP conjugate
diluted 1:50,000 in blocking solution (see Note 20), for 1 h
on a reciprocal benchtop shaker at RT.
14. Rinse the membrane three times in 40 mL TBS-T for 10 min
each at RT.
15. Transfer the membrane onto a sheet of plastic wrap.
16. Spread 700 μL chemiluminescent HRP substrate, on the sur-
face of the membrane and incubate for 5 min at RT.
17. Wrap the membrane in a plastic wrap and place it facing down
on chemiluminescence image scanner.
18. Expose the membrane to obtain the band corresponding to the
fusion protein (Fig. 3).
 Low Temperature Production of Engineered Vesicles 201

P49-eGFP
Cell

l.
l. F Vs
so EM
So In PV

75 P49-fused
eGFP

30
P49-free
eGFP
(Mass/kDa)

Fig. 3 Detection of eGFP fused to the C-terminus of P49 in S. vesiculosa HM13

EMVs. Soluble (Sol.) and insoluble (Insol.) cell fractions as well as PVF and EMVs
were subjected to SDS-PAGE and Western blotting. The fusion protein was
detected with an anti-GFP antibody. A fusion protein of about 75 kDa was
detected in both cellular fractions and EMVs, but not PVF, indicating that P49
transports the fusion partner to EMVs. P49-free eGFP (~29 kDa) was also
detected in these fractions. Proteolytic cleavage of the fusion protein probably
occurs after transport of the fusion protein to EMVs (see Note 21). (This figure
has been reproduced/modified from [8], with permission from Frontiers Media
SA)

4 Notes

1. For long-term storage of S. vesiculosa HM13 and its derivatives,

it is better to use cryopreservation beads rather than glycerol
stocks because the latter method often causes inefficient recov-
ery of the bacterial strains. For efficient recovery of the bacterial
strains from cryopreservation beads, it is better to use new
bacterial colonies (smooth pink colonies) for preparation of
the stock.
2. Any commercial HRP conjugate substrate kit, such as Chemi-
Lumi One Ultra (Nacalai Tesque, Kyoto, Japan), may be used
according to manufacturer’s protocol.
3. S. vesiculosa HM13-Rifr is a spontaneous rifampicin-resistant
mutant of a hypervesiculating psychrotrophic bacterium,
S. vesiculosa HM13 (Fig. 1), and the parent strain of
S. vesiculosa P49-eGFP. It is necessary to have a resistance
marker in the plasmid recipient strain that is different from
202 Jun Kawamoto and Tatsuo Kurihara

that of the plasmid being transferred. This allows selection of

transconjugants by growth in the presence of both antibiotics
(Fig. 2). S. vesiculosa HM13-Rifr was generated by spreading
the wild-type bacterial suspension on an LB plate containing
50 μg/mL rifampicin and incubating at 18
C for several days
until single colonies were formed. The colonies were isolated
and inoculated onto a new culture plate containing 50 μg/mL
rifampicin. Sequence analysis of the rpoB gene coding for RNA
polymerase ß-subunit demonstrated that Ser532 of the wild-
type enzyme was replaced with Tyr in the Rifr mutant
[8, 12]. The mutation of rpoB does not affect the growth
characteristics and cargo loading of P49 to the EMVs.
4. E. coli S17-1/λpir [13] is used as a conjugal donor strain for
pir-dependent R6K plasmids, such as pKNOCK [14]. Compe-
tent S17-1/λpir cells, prepared as described by Inoue et al.
[15], allow heat shock transformation with the desired plasmid,
which can then be conjugally transferred to a target
recipient cell.
5. pKNOCK [14] is a plasmid harboring an RP4 oriT, a pir-
dependent R6K ori, and a kanamycin resistance (Kanr) gene
(Fig. 2), which has been shown to work as a suicide vector in
S. vesiculosa [8].
6. pGreen is a plasmid carrying the eGFP-coding gene [16] and
was used for construction of the eGFP-integration plasmid,
pKP49eGFP.
7. Because rifampicin should be protected from light, rifampicin-
containing plates must be stored under dark conditions at 4
C.
8. A high-fidelity proofreading polymerase, such as Q5 High-
Fidelity DNA Polymerase (New England Biolabs Japan Inc.,
Tokyo, Japan) or similar, is required to reduce the possible
introduction of mutations during amplification of the
sequences.
9. Any PCR clean up kit, such as Wizard SV Gel and PCR Clean-
Up System (Promega, Madison, WI, USA) or similar, may be
used. If required (i.e., more than one amplicon band visible),
specific amplicons may be gel purified using an appropriate kit.
10. Different commercial kits are available for directional ligation-
free DNA assembly. We used NEBuilder HiFi DNA Assembly
Master Mix (New England Biolabs) according to the manufac-
turer’s instruction.
11. Detailed characterization of EMVs such as proteomics, lipido-
mics, and analysis of surface carbohydrates [9, 11] requires
higher amounts of purified EMVs. For these purposes,
mid-scale (30 mL) or large-scale (1 L) cultures are required,
with concentration using tangential flow filtration (TFF) with a
 Low Temperature Production of Engineered Vesicles 203

100-kD cut-off filter, as described by Chutkan et al. [17],

rather than ultracentrifugation. After TFF concentration,
EMVs can be collected by ultracentrifugation as described in
Subheading 3.3.
12. The maximum yield of EMVs carrying high purity P49 is
usually obtained from cultures in the mid- to late-stationary
phase (OD600 ¼ 2–3).
13. To remove EMVs remaining at the cell surface, the cell pellet
should be washed at least three times. Avoid carrying over the
supernatant. If the cell pellet is loose, leave some supernatant at
the bottom of the tube, add 500 μL LB medium into the tube,
centrifuge at 2,000  g at 18
C for 15 min and carefully pipet
to remove all the supernatant including the remaining EMVs
from the cell pellet.
14. The tubes must be balanced to within 0.01 g on a scale.
15. After ultracentrifugation, the EMV pellet was visible (Fig. 4).
To prevent pellet dispersion, immediately collect the superna-
tant (PVF) from the ultracentrifuge tube using a syringe nee-
dle, without touching the pellet.
16. After ultracentrifugation, EMVs often adhere to the bottom of
the centrifuge tube, and it is difficult to resuspend the EMV
pellet with DPBSS. To avoid yield loss, add 100–200 μL
DPBSS at a time, and gently pipet up and down without
directly touching the pellet. Repeat resuspension with the
remaining DPBSS until the pellet is completely dispersed.
17. Store purified EMVs of S. vesiculosa HM13 at 80
C. Long-
term storage (more than 4 days) of EMVs at 4
C often affects
the size of the EMVs and stability of the cargo protein, proba-
bly due to unwanted aggregation of EMVs and proteolysis of

Fig. 4 EMV pellet obtained by ultracentrifugation. EMVs purified from 5 mL of the

culture supernatant of S. vesiculosa HM13 were visible as a pale pink pellet after
ultracentrifugation
204 Jun Kawamoto and Tatsuo Kurihara

the cargo. However, EMVs can be stored at 4
C for use within

2–3 days for morphological studies by transmission electron
microscopy (TEM), nanoparticle tracking analysis, and
dynamic light scattering. For proteome analysis, fresh EMVs
or EMVs stored at 80
C should be used.
18. For Western blotting, we used a semidry method described by
Komatsu [18] with modification.
19. Extra-thick blot filter paper can be purchased precut (e.g.,
Bio-Rad Blot-Absorbent Filter Paper 7.5 cm  10 cm 
2.45 mm) to match the size of precast gels.
20. Dilution may be different depending on source of the
antibodies.
21. Analysis of secretory production of eGFP fused to P49 demon-
strated the loading of P49-free eGFP, in addition to the fusion
protein, onto EMVs. The appearance of P49-free eGFP might
be due to the proteolytic cleavage of the fusion protein after
being transported to EMVs. Regulation of this putative pro-
teolytic activity could expand the utility of this secretory pro-
tein production system, for example, for production of
P49-free foreign protein by facilitating cleavage or production
of fusion protein with higher purity by suppressing cleavage.

Acknowledgments

This work was supported in part by JSPS KAKENHI

(JP17H04598, JP18K19178, and JP20K20570 to TK and
JP16K14885 and JP20K05786 to JK) and a grant from the Insti-
tute for Fermentation, Osaka (L-2019-2-012 to TK). TEM obser-
vations were performed in collaboration with the Analysis and
Development System for Advanced Materials (ADAM) at the
Research Institute for Sustainable Humanosphere, Kyoto
University.

References
1. Schwechheimer C, Kuehn MJ (2015) Outer- 3. Nagakubo T, Nomura N, Toyofuku M (2020)
membrane vesicles from Gram-negative bacte- Cracking open bacterial membrane vesicles.
ria: biogenesis and functions. Nat Rev Micro- Front Microbiol 10:3026. https://doi.org/
biol 13:605–619. https://doi.org/10.1038/ 10.3389/fmicb.2019.03026
nrmicro3525 4. Anand D, Chaudhuri A (2016) Bacterial outer
2. Toyofuku M, Tashiro Y, Hasegawa Y et al membrane vesicles: new insights and applica-
(2015) Bacterial membrane vesicles, an over- tions. Mol Membr Biol 33:125–137. https://
looked environmental colloid: biology, envi- doi.org/10.1080/09687688.2017.1400602
ronmental perspectives and applications. Adv 5. McCarthy PC, Sharyan A, Sheikhi Moghad-
Colloid Interf Sci 226:65–77. https://doi. dam L (2018) Meningococcal vaccines: current
org/10.1016/j.cis.2015.08.013 status and emerging strategies. Vaccines (Basel)
 Low Temperature Production of Engineered Vesicles 205

6:12. https://doi.org/10.3390/ 12. Taniguchi H, Aramaki H, Nikaido Y et al

vaccines6010012 (1996) Rifampicin resistance and mutation of
6. Gerritzen MJH, Martens DE, Wijffels RH et al the rpoB gene in Mycobacterium tuberculosis.
(2017) Bioengineering bacterial outer mem- FEMS Microbiol Lett 144:103–108. https://
brane vesicles as vaccine platform. Biotechnol doi.org/10.1111/j.1574-6968.1996.
Adv 35:565–574. https://doi.org/10.1016/j. tb08515.x
biotechadv.2017.05.003 13. Hoffmann A, Thimm T, Dröge M et al (1998)
7. Irene C, Fantappiè L, Caproni E et al (2019) Intergeneric transfer of conjugative and mobi-
Bacterial outer membrane vesicles engineered lizable plasmids harbored by Escherichia coli in
with lipidated antigens as a platform for Staph- the gut of the soil microarthropod Folsomia
ylococcus aureus vaccine. Proc Natl Acad Sci U S candida (Collembola). Appl Environ Micro-
A 116:21780–21788. https://doi.org/10. biol 64:2652–2659. https://doi.org/10.
1073/pnas.1905112116 1128/AEM.64.7.2652-2659.1998
8. Chen C, Kawamoto J, Kawai S et al (2020) 14. Alexeyev MF (1999) The pKNOCK series of
Isolation of a novel bacterial strain capable of broad-host-range mobilizable suicide vectors
producing abundant extracellular membrane for gene knockout and targeted DNA insertion
vesicles carrying a single major cargo protein into the chromosome of Gram-negative bacte-
and analysis of its transport mechanism. Front ria. Biotechniques 26:824–826, 828. https://
Microbiol 10:3001. https://doi.org/10. doi.org/10.2144/99265bm05
3389/fmicb.2019.03001 15. Inoue H, Nojima H, Okayama H (1990) High
9. Casillo A, Di Guida R, Carillo S et al (2019) efficiency transformation of Escherichia coli
Structural elucidation of a novel lipooligosac- with plasmids. Gene 96:23–28. https://doi.
charide from the cold-adapted bacterium org/10.1016/0378-1119(90)90336-P
OMVs producer Shewanella sp. HM13. Mar 16. Hellens RP, Edwards EA, Leyland NR et al
Drugs 17:34. https://doi.org/10.3390/ (2000) pGreen: a versatile and flexible binary
md17010034 Ti vector for Agrobacterium-mediated plant
10. Kamasaka K, Kawamoto J, Chen C et al (2020) transformation. Plant Mol Biol 42:819–832.
Genetic characterization and functional impli- https://doi.org/10.1023/a:1006496308160
cations of the gene cluster for selective protein 17. Chutkan H, MacDonald I, Manning A, Kuehn
transport to extracellular membrane vesicles of MJ (2013) Quantitative and qualitative pre-
Shewanella vesiculosa HM13. Biochem Biophys parations of bacterial outer membrane vesi-
Res Commun 526:525–531. https://doi.org/ cles. Methods Mol Biol 966:259–272.
10.1016/j.bbrc.2020.03.125 https://doi.org/10.1007/978-1-62703-245-
11. Di Guida R, Casillo A, Yokoyama F et al (2020) 2_16
Detailed structural characterization of the 18. Komatsu S (2015) Western blotting using
lipooligosaccharide from the extracellular PVDF membranes and its downstream applica-
membrane vesicles of Shewanella vesiculosa tions. Methods Mol Biol 1312:227–236.
HM13. Mar Drugs 18:231. https://doi.org/ https://doi.org/10.1007/978-1-4939-2694-
10.3390/md18050231 7_24
 Chapter 13

Glycine Induction Method: Effective Production

of Immunoactive Bacterial Membrane Vesicles with Low
Endotoxin Content
Satoru Hirayama and Ryoma Nakao

Abstract
Bacteria are known to release nanometer scale proteoliposomes termed bacterial membrane vesicles (MVs),
and it is considered that native and bioengineered MVs would be applicable for development of acellular
vaccines and novel drug delivery systems in medical settings. However, important considerations for
manufacturing purposes include the varied productivity of MV among bacterial species and strains, as
well as endotoxicity levels due to the lipopolysaccharide component. The method for MV induction using
glycine described here is simple and provides a solution to these problems. Glycine weakens bacterial
peptidoglycans and significantly increases bacterial MV formation, while the relative endotoxin activity of
glycine-induced MVs is extremely reduced as compared to that of noninduced MVs. Nevertheless, glycine-
induced MVs elicit strong immune responses at levels nearly equivalent to those of noninduced MVs. Taken
together, the present method for induction by glycine is convenient for research studies of bacterial MVs
and has potential for use in medical applications including vaccine development.

Key words Membrane vesicles, Glycine, Escherichia coli, Endotoxin, Adjuvant

1 Introduction

Bacterial membrane vesicles (MVs) are spherical nanostructural

bodies, released from the cells under various growth conditions.
Bacterial MVs generally contain various antigens and have adjuvant
activities; thus, their application to vaccines has been widely stud-
ied, though amounts produced vary among species and strains.
When targeting bacteria with a small amount of MV formation,
the resultant low yield is a major obstacle for development of
medical applications, and no general method for massive produc-
tion of MVs has been established.
Various MV formation mechanism models have been pro-
posed, some of which involve bacterial peptidoglycans (PGs).
Investigation of a model utilizing PG endopeptidases found that

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_13,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

207
208 Satoru Hirayama and Ryoma Nakao

MV production is likely enhanced due to accumulation of sub-

stances such as PG fragments in the periplasm [1], while another
study showed that MV formation in Escherichia coli was increased
by deletion of genes involved in PG biosynthesis [2]. Furthermore,
it was recently reported that the action of endolysin, a
PG-degrading enzyme derived from phage, which causes explosive
cell lysis, was shown to induce formation of MVs [3, 4]. Given this
background, it is considered that a mechanism that acts on bacterial
PGs is likely an effective means to induce MVs.
Very recently, we reported a simple method for adding glycine
during bacterial culture to increase bacterial MV production [5]. In
E. coli, PGs consist of alternating repeating units of two amino
sugars, N-acetylmuramic acid and N-acetylglucosamine, with L-
alanine, D-glutamic acid, meso-2,6-diaminopimelic acid, and D-
alanine internally bridging between them [6] (Fig. 1). Glycine has
a function to weaken PGs by competing for and replacing DL-
alanine contained in them [7]. Addition of a large amount of
glycine (e.g., ~2.0%) was found to severely suppress bacterial
growth, whereas bacterial MV production was significantly
increased by adjusting the added amount. In studies of an E. coli
probiotic strain (Nissle 1917) [8–10], addition of 1.0% glycine
increased the yield of MVs by approximately 70-fold regarding
the protein amount and approximately 50-fold regarding the lipid
amount as compared to no addition [5]. Our recent findings

Fig. 1 Structure of E. coli peptidoglycan. The peptidoglycan (PG) is formed from

two alternating amino sugars, a straight chain of N-acetylglucosamine (GlcNAc)
and N-acetylmuramic acid (MurNAc) with L-alanine (L-Ala), D-glutamic acid (D-
Glu), meso-2,6-diaminopimelic acid (m-Dpm), and D-alanine (D-Ala) internally
bridging between them. Glycine weakens PG by competing for and replacing DL-
alanine contained within
 Low Endotoxin Glycine-Induced Membrane Vesicles 209

Fig. 2 Noninduced and glycine-induced MVs of E. coli. Noninduced MVs (a) and glycine-induced MVs (b) of a
flagellar-deficient derivative of E. coli Nissle 1917 were purified by ultracentrifugation, then observed by
transmission electron microscopy. The mean of diameter of glycine-induced MVs was apparently greater than
that of noninduced MVs. No flagella were observed as the flagella master regulator gene flhD was deleted in
both preparations [5], while fimbrial structures (black arrow heads) associated with MVs were noted in both.
Scale bars represent 100 nm

showed that glycine-induced MVs had increased average particle

size (Fig. 2) and altered protein composition. Furthermore, endo-
toxin activity was reduced in glycine-induced MVs by approxi-
mately eightfold, whereas the cytokine-inducing activity in mouse
macrophage-like cells J774.1 in vitro and mucosal adjuvant activity
in vivo were comparable to those of noninduced MVs [5]. These
differences in properties are thought to be related to the fact that
the mode of MV formation is also changed. As a result, it is
considered that glycine induction can be used as a method to easily
and significantly induce formation of bacterial MVs that would be
suitable for application to vaccines as well as adjuvants. The most
effective amount of added glycine may vary depending on bacterial
species and target strain. Nevertheless, a glycine induction method
is considered to be at least practically applicable for increased MV
yield in E. coli [5] and Acinetobacter baumannii (unpublished
findings by Nakao R et al.).
In the present chapter, a glycine induction method for prepa-
ration of large amounts of MVs with low endotoxin content is
described (using the E. coli Nissle 1917 strain as an exemplar),
along with methods for assessing cytokine-inducing activity and
mucosal adjuvanticity of the MVs. Because both MV productivity
and safety are important considerations in the context of potential
future clinical applications, it is proposed that the present method
will be useful for various medical research fields, especially those
involved in drug delivery research and vaccinology.
210 Satoru Hirayama and Ryoma Nakao

2 Materials

2.1 Induction 1. A flagella-deficient derivative of E. coli Nissle 1917 strain (DSM

of E. coli MV 6601, serotype O6:K5:H1) (see Note 1).
Production by Addition 2. LB broth: 10 g/L tryptone, 5 g/L yeast extract, and 10 g/L
of Glycine NaCl in distilled water. Autoclave and store at room tempera-
ture or 4
C.
3. LB agar plates: LB broth added 15 g/L agar. After autoclaving,
dispense into Petri dishes. Store at room temperature or 4
C.
4. Culture tubes and flasks.
5. Incubators with and without shaker.
6. 20% glycine solution: Dissolve 20 g of glycine in 80 mL dis-
tilled water then bring the volume up to 100 mL (see Note 2).
Sterilize using a 0.22-μm filter and store at room temperature
(see Note 3).
7. Centrifuge.
8. 400-mL polypropylene centrifuge bottles with screw caps.
9. Bottle top PVDF membrane filter units with pore size of
0.45 μm (see Note 4).
10. Aspirator with a vacuum pump (see Note 4).
11. Ultracentrifuge with a fixed-angle rotor.
12. 70-mL polycarbonate ultracentrifuge bottles with aluminum
screw caps.
13. Phosphate-buffered saline (PBS), 10  stock: NaCl 80 g/L,
KCl 2 g/L, Na2HPO4 14.4 g/L, KH2PO4 2.4 g/L in distilled
water. Mix solid with ca. 900 mL distilled water, adjust pH to
7.4, then bring final volume to 1 L and autoclave to sterilize.
For 1 PBS, dilute 1/10 with distilled water, and sterilize by
filtration through 0.22-μm filter (see Notes 5 and 6).
14. 20 mM Tris–HCl buffer (pH 8.0): Prepare 1 M Tris–HCl
buffer (pH 8.0) by weighing 12.1 g of Tris base and dissolving
it in 80 mL distilled water. Adjust pH to 8.0 with concentrated
HCl then bring the volume up to 100 mL and autoclave to
sterilize. Dilute the 1 M Tris–HCl (pH 8.0) 1/50 with distilled
water, and sterilize by filtration through 0.22-μm filter (see
Notes 5 and 6).

2.2 Quantification 1. Bovine serum albumin (BSA) solution: Dissolve BSA in dis-
of MV Components tilled water to obtain 10 mg/mL stock solution. Dispense
aliquots into microtubes and store at 20
C.
2.2.1 Quantification
of the Amount of Total 2. 96-well microtiter plates (see Note 7).
Protein in MVs
 Low Endotoxin Glycine-Induced Membrane Vesicles 211

3. Coomassie brilliant blue (CBB) G-250 dye-based reagent (i.e.,

Bradford assay reagent, e.g., Bio-Rad Protein Assay, Bio-Rad,
Hercules, CA, USA).
4. 96-well microplate reader.

2.2.2 Quantification 1. Linoleic acid solution: Dissolve water-soluble linoleic acid in

of the Amount of Total Lipid distilled water to obtain 1 mg/mL stock solution. Dispense
in MVs aliquots into microtubes and store at 20
C.
2. FM4-64 dye solution: Dissolve FM4-64 dye (N-(3-triethylam-
moniumpropyl)-4-(6-(4-(diethylamino) phenyl)hexatrienyl)
pyridinium dibromide) in distilled water to prepare a
0.5 mg/mL stock solution. Dispense aliquots into brown
microtubes and store at 20
C protected from light.
3. 96-well black microtiter plate.
4. Fluorescence microplate reader.

2.2.3 Quantification 1. Lipopolysaccharide (LPS) solution: Dissolve commercial LPS

of Endotoxin in MVs in distilled water to obtain 1 mg/mL solution (see Note 8).
Dispense aliquots into microtubes and store at 20
C.
2. Endotoxin-free water.
3. Endotoxin-free pipette tips.
4. Endotoxin-free 96-well microtiter plate.
5. Limulus amebocyte lysate (LAL) assay kit (e.g., Endospecy
ES-50M Set, Seikagaku Co, Tokyo, Japan).
6. Incubator.
7. 0.6 M acetic acid in endotoxin-free water (see Note 9).
8. Fluorescence microplate reader.

2.3 Evaluation In 1. Mouse macrophage-like cell line J774.1 (see Note 10).
Vitro: 2. Supplemented RPMI 1640 medium (S-RPMI): RPMI 1640
Cytokine-Inducing supplemented with 10% heat-inactivated fetal calf serum,
Activity of MVs 100 units/mL penicillin G, and 0.1 mg/mL streptomycin
2.3.1 Addition of MVs
sulfate.
to Macrophage-like Cells 3. 90-mm nontreated culture dishes (see Note 10).
and Sample Preparation 4. CO2 incubator.
5. Microscope.
6. Centrifuge.
7. 50-mL polypropylene centrifuge tubes.
8. Trypan blue solution: 0.4% (w/v) trypan blue dissolved in
distilled water.
9. Cell counting system.
212 Satoru Hirayama and Ryoma Nakao

10. 24-well microtiter plates.

11. 1.5-mL microtubes.

2.3.2 Quantitative 1. RNA purification kit suitable for use with cultured cells (e.g.,
Real-Time PCR RNeasy Mini Kit, QIAGEN, Hilden, Germany).
2. Centrifuge.
3. 15-mL or 50-mL polypropylene centrifuge tubes.
4. Real-time PCR cDNA synthesis kit (e.g., ReverTra Ace qPCR
RT Master Mix with gDNA Remover, TOYOBO, Tokyo,
Japan).
5. 96-well PCR plates with plate seals.
6. Thermal cycler.
7. Forward and reverse primers and TaqMan probes for detection
of the expression of the target cytokines and β-actin internal
control (Table 1).
8. Enzyme and buffer for real-time PCR (e.g., Premix Ex Taq
[Probe qPCR], TaKaRa Bio, Shiga, Japan).
9. TE buffer: 10 mmol/L Tris-HCl (pH 8.0), 1 mmol/L EDTA
(pH 8.0), autoclaved.
10. DNA standard for absolute quantification: Prepare template
DNA of interest (e.g., plasmid cloned target sequences or
synthesized gene). Adjust to 0.5  1010 copies/μL with TE
buffer.
11. Real-time PCR instrument.

2.3.3 Sandwich ELISA 1. 96-well ELISA plate.

for Cytokines 2. Plastic wrap.
3. Capture antibody for target cytokine molecules (e.g., anti-
murine IL-6 antibody).
4. PBS (see Subheading 2.1).
5. PBST: 0.05% Tween-20 in PBS.
6. Blocking buffer: 1% BSA in PBS.
7. Recombinant protein of target cytokine molecules (e.g., IL-6).
8. Diluent: 0.1% BSA in PBST.
9. Detection antibody for target cytokine molecules (e.g., bioti-
nylated anti-murine IL-6 antibody).
10. Conjugated avidin: Avidin-alkaline phosphatase
(AP) conjugate or avidin-horseradish peroxidase (HRP)
conjugate.
 Low Endotoxin Glycine-Induced Membrane Vesicles 213

Table 1
Primers and probes for quantitative real-time PCR

Target Sequence (50 –30 )

IL-4 Forward CGCCATGCACGGAGATG
Reverse CGAGCTCACTCTCTGTGGTGTT
Probea TGCCAAACGTCCTCACAGCAACGA
IL-6 Forward CCAGAAACCGCTATGAAGTTCCT
Reverse CACCAGCATCAGTCCCAAGA
Probea TCTGCAAGAGACTTCCATCCAGTTGCC
IL-12p40 Forward AGCTCGCAGCAAAGCAAGAT
Reverse TGGAGACACCAGCAAAACGA
Probea TGTCCTCAGAAGCTAACCATCTCCTG
TNF-α Forward AGACCCTCACACTCAGATCATCTTC
Reverse CCTCCACTTGGTGGTTTGCTA
Probea CAAAATTCGAGTGACAAGCCTGTAGCCC
IFN-α Forward CTGCTAGTGATGAGCTACTGGTCAA
Reverse GGGTCAAGGCTCTCTTGTTCCT
Probea CTGCTCCCTAGGATGTGACCTGCCTCA
IFN-β Forward GCTCCTGGAGCAGCTGAATG
Reverse TCCGTCATCTCCATAGGGATCT
Probea TCAACCTCACCTACAGGGCGGACTTC
IFN-γ Forward AGCCAGATTATCTCTTTCTACCTCAGA
Reverse GCAATACTCATGAATGCATCCTTT
Probea CAGGCCATCAGCAACAACATAAGGGTC
β-Actin Forward CACCGATCCACACAGAGTACTTG
Reverse CAGTGCTGTCTGGTGGTACCA
Probea CAGTAATCTCCTTCTGCATCCTGTCAGCAA
Sequences other than those for TNF-α have been previously published [11]
a
Probes are labeled with FAM (6-carboxyfluorescein) at 50 -end and TAMRA (6-carboxytetramethylrhodamine) at 30 -end
of the sequences

11. Substrate solution: For AP reaction, prepare 3 g/L para-nitro-

phenyl phosphate in diethanolamine buffer (9.7 mL/L dietha-
nolamine, 0.1 g/L MgCl2·6H2O, 0.2 g/L NaN3 in distilled
water, pH adjusted to 9.6). For HRP reaction, prepare
1.25 mmol/L 3,30 ,5,50 -tetramethylbenzidine and
2.21 mmol/L hydrogen peroxide, and less than 1% dimethyl
sulfoxide in a 0.08 mol/L acetate buffer at pH 4.9.
12. 96-well microplate reader.
214 Satoru Hirayama and Ryoma Nakao

2.4 Evaluation In 1. 5-week-old female BALB/c mice (see Note 11).

Vivo: Mucosal 2. PBS (see Subheading 2.1).
Adjuvanticity of MVs
3. Ovalbumin (OVA) stock solution: Dissolve OVA in PBS to
2.4.1 Intranasal obtain 2 mg/mL solution. Dispense aliquots into microtubes
Immunization of Mice and store at 20
C.
4. Isoflurane.
5. Anesthesia inhalation device (see Note 12).

2.4.2 Sample Collection 1. PBS (see Subheading 2.1).

from Mice 2. Parasympathetic stimulant solution: Prepare 0.8 mg/mL iso-
proterenol in PBS. Separately, prepare 0.2 mg/mL pilocarpine
in PBS. Mix equal volumes of the two solutions and filter the
solution through a 0.22-μm filter (see Note 13).
3. 1 mL syringes.
4. 26G needles.
5. 21G nonbeveled needles (see Note 14).
6. Scissors.
7. Tweezers.
8. PBS containing 0.1% BSA: Dissolve BSA at 0.1% concentration
(w/v) in PBS.
9. 1.5-mL microtubes.

2.4.3 ELISA 1. 96-well ELISA plate.

for Evaluation of Mucosal 2. Plastic wrap.
Adjuvanticity of MVs
3. OVA solution (see Subheading 2.4.1).
4. ELISA coating buffer: 1.59 g/L Na2CO3, 2.93 g/L NaHCO3,
0.2 g/L NaN3 in distilled water, pH adjusted to 9.6.
5. PBST (see Subheading 2.3.3).
6. 1% skim milk in PBST.
7. 0.5% skim milk in PBST.
8. Incubator.
9. Enzyme-linked detection antibody (e.g., rabbit AP-labeled
anti-murine IgG (H + L) antibody, goat HRP-labeled anti-
murine IgA (alpha) antibody).
10. Substrate solution (see Subheading 2.3.3).
11. 96-well microplate reader.
 Low Endotoxin Glycine-Induced Membrane Vesicles 215

3 Methods

Perform all procedures at room temperature, unless otherwise

specified.

3.1 Induction 1. For seed culture, pick a single colony of E. coli from an over-
of E. coli MV night LB agar plate culture and inoculate a small amount (e.g.,
Production by Addition 2 mL) of LB broth. Incubate at 37
C for 8 h with shaking at
of Glycine ca. 150 rpm.
2. Inoculate 1/200 volume of E. coli culture into 80 mL of LB
broth without glycine (noninduction control) (see Note 15),
and LB broth with glycine added to a final concentration of
~2.0% (see Note 16). Incubate at 37
C for 16 h with shaking at
ca. 150 rpm (see Note 17).
3. Centrifuge bacterial cultures at 7,200  g for 30 min at 4
C
and collect the culture supernatants.
4. Filter culture supernatants using 0.45-μm PVDF filters to
completely remove remaining bacterial cells (see Note 4).
5. Ultracentrifuge culture supernatants at 103,800  g for 2 h at
4
C to obtain MV pellets.
6. Suspend MV pellets in small amount (e.g., 1.5 mL) of PBS or
20 mM Tris-HCl (pH 8.0) (see Notes 6, 18 and 19).

3.2 Quantification The amounts of total protein and total lipid in MVs can be quanti-
of MV Components fied by the Bradford and FM4-64-based methods, respectively. We
routinely perform the Bradford method for quality control of MVs,
and sometimes use the FM4-64-based method as an alternative
quality control technique. The endotoxin activity of MVs can be
also measured by the Limulus assay and normalized by the amount
of protein or lipid contained in MVs.

3.2.1 Quantification 1. Make a two-fold serially diluted solution (e.g., 0.1–1000 μg/
of the Amount of Total mL) of standard BSA with distilled water.
Protein in MVs 2. Dilute MV samples two-fold serially with distilled water (see
Note 20).
3. Pipette 10 μL each of standard BSA and MV samples into
appropriate wells of a 96-well microtiter plate (see Notes 7
and 21).
4. Add appropriate volume (e.g., 200 μL/well) of dye reagent for
Bradford protein assay (see Note 22).
5. Incubate the plate for 5–15 min at room temperature.
6. Measure absorbance at 595 nm.
7. Estimate the amount of protein contained in the MV samples
from the calibration curve created by the BSA standards.
216 Satoru Hirayama and Ryoma Nakao

3.2.2 Using FM4-64, which is a lipophilic dye, and water-soluble linoleic

Quantification of theAmount acid, the amount of lipid contained in MVs is calculated as the
of Total Lipid in MVs linoleic acid equivalent.
1. Make a twofold serially diluted solution (e.g., 0.5–500 μg/mL)
of standard linoleic acid with distilled water.
2. Dilute MV samples twofold serially with distilled water (see
Note 23).
3. Dilute FM4-64 dye solution to 1/200 (i.e., 2.5 μg/mL) with
distilled water.
4. Pipette 5 μL each of standard linoleic acid and MV samples into
appropriate wells of a 96-well black microtiter plate (see
Note 21).
5. Add 100 μL/well of 2.5 μg/mL FM4-64 dye.
6. Incubate the plate for 10 min at room temperature protected
from light.
7. Detect fluorescence from FM4-64 dye with a fluorescence plate
reader with the excitation and emission wavelengths at 535 nm
and 625 nm, respectively.
8. Estimate the amount of lipid contained in the MV samples
from the calibration curve created by the linoleic acid
standards.

3.2.3 Quantification It is recommended to first make a rough estimate of the required

of Endotoxin in MVs dilutions for standards and samples, to determine the readable
range, and then narrow the dilution range to quantify samples.
1. Make a tenfold serially diluted solution of standard LPS and
MV samples with endotoxin-free water.
2. Pipette 25 μL each of standard LPS and MV samples into
appropriate wells of an endotoxin-free 96-well microtiter
plate (see Note 21).
3. Add 25 μL/well of LAL reagent, incubate at 37
C for 30 min.
4. Add 100 μL/well of 0.6 M acetic acid.
5. Measure absorbance at 405 nm.
6. Estimate the required dilution rates of LPS standards and MV
samples.
7. Make a twofold serially diluted solution of standard LPS and
MV samples within measurable range using endotoxin-free
water.
8. Repeat steps 2 through 5.
9. Estimate the endotoxin unit (EU) of the MV samples from the
calibration curve created by the LPS standards (see Note 8).
 Low Endotoxin Glycine-Induced Membrane Vesicles 217

3.3 Evaluation In This section presents two methods for assessing MV immunoactiv-
Vitro: ity using a mouse macrophage-like cell line; quantification of intra-
Cytokine-Inducing cellular cytokine mRNA expression levels and determination of the
Activity of MVs amount of cytokines into the medium. Different cytokines (and/or
their specific mRNA expression profiles) may be evaluated, affect-
ing the choice of reagents used for the real-time PCR and ELISA
methods described. We normally assay for IL-6, IL-12 and TNF-α,
and β-actin for internal control.

3.3.1 Addition of MVs 1. Incubate J774.1 cells in 50 mL of S-RPMI in 5 separate non-

to Macrophage-like Cells treated culture dishes (10 mL/dish) at 37
C in a humidified
and Sample Preparation atmosphere containing 5% CO2 until cells are confluent (see
Notes 10 and 24).
2. Harvest the cells from dishes and transfer them to a
centrifuge tube.
3. Disperse cells well, then remove an aliquot of cell suspension
(e.g., 20 μL) and mix with 0.4% (w/v) trypan blue
solution 1:1.
4. Count live cells (cells not stained blue) and estimate total
number of live cells in the suspension.
5. Collect cells by centrifugation at 300  g for 10 min at room
temperature and discard the supernatant.
6. Resuspend cells in an appropriate volume S-RPMI 1640
prewarmed at 37
C to adjust number of live cells to
7.3  105/mL.
7. Dispense 900 μL of cell suspension (i.e., 6.6  105 cells) into
each of 21 wells of a 24-well microtiter plate (see Note 24).
8. Standardize glycine-induced and noninduced MVs to desired
concentrations (e.g., 0.1, 1, and 10 ng-protein/mL) with PBS.
9. Add 100 μL of each MV sample to appropriate wells of the
24-well plate containing 900 μL of cell suspension and mix. For
the negative control wells, add and mix 100 μL of PBS alone.
10. Incubate for 2 h (for samples used in quantitative real-time
PCR) or 12 h (for samples used in cytokine ELISA) at 37
C in
humidified atmosphere containing 5% CO2.
11. Pipette cells and MV mixture into a 1.5-mL microtube and
centrifuge at 300  g for 10 min at room temperature. Col-
lected cells can be used for analysis of cytokine mRNA expres-
sion by quantitative real-time PCR (see Subheading 3.3.2),
while the supernatant can also be used for determining the
amounts of secreted cytokines by ELISA (see Subheading
3.3.3).
218 Satoru Hirayama and Ryoma Nakao

3.3.2 Quantitative 1. Prepare total RNA from collected cells prepared in Subheading
Real-Time PCR 3.3.1 using an RNA purification kit appropriate for culture
cells, according to the manufacturer’s instructions.
2. Synthesize cDNA from total RNA using the cDNA synthesis
kit for real-time PCR, according to the manufacturer’s
instructions.
3. Prepare a master mixture for quantitative real-time PCR con-
taining 0.04 μM each of forward and reverse primers and
probe, specific for the cytokine-encoding mRNA to be quanti-
fied, and the appropriate concentration of enzyme and buffer
(20 μL/PCR reaction).
4. Dispense 18 μL/well of the master mixture into a 96-well real-
time PCR plate.
5. Dilute the appropriate standard plasmid tenfold serially (e.g.,
100–109 copies per 2 μL) with sterile water or TE buffer.
6. Add 2 μL each of the serially diluted standard plasmid and the
cDNA samples to appropriate wells (Fig. 3) containing the
PCR mixture dispensed in step 4.
7. Seal the plate and perform real-time PCR reaction (1 cycle of
95
C for 30 s, 40 cycles consisting of 95
C for 5 s and 60
C
for 34 s). Set to measure the fluorescence of the sample at
60
C.
8. Create a calibration curve using Ct (threshold cycle) values of
standards and plasmid copy numbers. Use the calibration curve
to determine the copy number of the target gene contained in
the cDNA sample (see Note 25).

3.3.3 Cytokine ELISA 1. To coat an ELISA plate with a capture antibody specific for the
cytokine to be quantified, dilute the antibody with PBS and
dispense 100 μL/well into a 96-well microtiter plate.
2. Seal the plate with plastic wrap and incubate overnight at room
temperature.
3. Remove contents and wash the plate with 300 μL/well PBST
four times.
4. Dispense 300 μL/well blocking buffer into the plate.
5. Seal the plate with plastic wrap and incubate for 1 h at room
temperature.
6. Remove contents and wash the plate with 300 μL/well PBST
four times.
7. Twofold serially dilute the recombinant cytokine protein stan-
dard (e.g., ranging from 28 to 21 ng/mL) with PBST. Also,
dilute culture supernatants prepared in Subheading 3.3.1 with
PBST (see Note 26).
 Low Endotoxin Glycine-Induced Membrane Vesicles 219

Fig. 3 Example 96-well plate layout for real-time PCR detection of cytokine expression. Dilutions of standards
are tested in duplicate, whereas each of the 21 cDNA samples from stimulated J774.1 cells are tested in
triplicate. NC ¼ negative control

Fig. 4 Example 96-well plate layout for detection of cytokine concentration by ELISA. Dilutions of recombinant
cytokine standards are tested in duplicate, whereas each of the 21 culture supernatant samples from
stimulated J774.1 cells may be tested either in triplicate for a single dilution, or in three different concentra-
tions (serial twofold dilutions). NC ¼ negative control

8. Add 100 μL/well each of the diluted standards and samples to

appropriate wells (Fig. 4) of the plate.
9. Seal the plate with plastic wrap and incubate at room tempera-
ture for 2 h.
10. Remove contents and wash the plate with 300 μL/well PBST
four times.
11. Add 100 μL/well of the appropriate biotinylated cytokine-
specific detection antibody, diluted with diluent to the concen-
tration recommended by the manufacturer.
12. Seal the plate with plastic wrap and incubate at room tempera-
ture for 2 h.
220 Satoru Hirayama and Ryoma Nakao

Fig. 5 Immunization timeline. Immunize mice initially at 6 weeks of age, and

again 3 and 6 weeks later. Collect samples (saliva, serum, and nasal wash) at
2 weeks after final immunization

13. Remove contents and wash the plate with 300 μL/well PBST
four times.
14. Add 100 μL/well of enzyme-conjugated avidin, appropriately
diluted with diluent, as recommended by the manufacturer.
15. Seal the plate with plastic wrap and incubate at room tempera-
ture for 30 min.
16. Remove contents and wash the plate with 300 μL/well PBST
four times.
17. Add 100 μL/well of substrate solution into the plate.
18. Sequentially (e.g., 10 min, 30 min, 60 min, and 120 min)
measure the absorbance of 405 nm (for AP) or 650 nm (for
HRP) (see Note 27).

3.4 Evaluation In Intranasally immunize mice with OVA as an antigen and bacterial
Vivo: Mucosal MVs as an adjuvant to evaluate the adjuvant activity of MVs. The
Adjuvanticity of MVs immunization schedule is shown in Fig. 5.

3.4.1 Nasal 1. Purchase 4 mice of 5-week-old female BALB/c per experimen-

Immunization to Mice tal group (see Note 28). In this case, 16 mice were divided into
the following 4 groups: (1) PBS (mock, negative control);
(2) OVA (antigen alone); (3) OVA + glycine-induced MVs;
(4) OVA + noninduced MVs.
2. Maintain the mice for 1 week to acclimate to environment.
3. On inoculation day (i.e., day 0 when mice are 6 weeks old),
prepare the following samples for administration to each
group: (1) 50 μL of PBS; (2) 50 μL of OVA (0.5 μg/μL) in
PBS; (3) 50 μL of mixture of OVA (0.5 μg/μL) + glycine-
induced MVs (0.1 μg/μL) in PBS; (4) 50 μL of mixture of
OVA (0.5 μg/μL) + noninduced MVs (0.1 μg/μL) in PBS.
4. Anesthetize mice with isoflurane by inhalation (see Note 29).
 Low Endotoxin Glycine-Induced Membrane Vesicles 221

5. Using a pipette, inoculate 5 μL per nostril (total 10 μL per

mouse) with the appropriate sample for each group. For test
samples, this equates to 5 μg per mouse OVA +/ 1 μg per
mouse of either glycine-induced or noninduced MVs (see Note
30).
6. At 21 days after first immunization (9 weeks old), nasal immu-
nization is performed again by repeating steps 3–5.
7. At 21 days after second immunization (12 weeks old), nasal
immunization is performed again by repeating steps 3–5.
8. At 14 days after the third immunization (14 weeks old), collect
saliva, serum, and nasal wash samples, as described in Subhead-
ing 3.4.2.

3.4.2 Sample Collection 1. To collect saliva samples, inject 200 μL of the parasympathetic
from Mice stimulant solution into the abdominal cavity. When saliva secre-
tion is induced, collect saliva from the mouth with a pipette.
Store at 80
C (see Note 31).
2. For serum samples, collect whole blood from anesthetized
mice by cardiac puncture through the diaphragm with 1-mL
syringe and 26G needle (see Notes 32 and 33). Centrifuge
blood samples at 300  g for 10 min to precipitate blood
cells, collect the serum in new microtubes. Store at 20
C.
3. For nasal wash samples, remove the head of the deceased
mouse after exsanguination and dissect out the lower jaw.
Insert a syringe needle (21G nonbeveled) into the nasal cavity
from the posterior opening, and flush with 1 mL of PBS con-
taining 0.1% BSA. Collect the outflow from the nostrils in 1.5-
mL microtube. Repeat the flushing step three times, pool the
replicates and centrifuge at 300  g for 10 min to remove cell
debris. Store at 20
C.

3.4.3 ELISA The samples collected in Subheading 3.4.2 may be assayed for
for Evaluation of Mucosal OVA-specific IgG, IgA, IgM, and IgE antibodies, as desired, by
Adjuvanticity of MVs using the appropriate enzyme-linked detection antibody. Each type
of sample (i.e., saliva, serum, nasal wash) is tested separately, but a
single ELISA plate can be used to assay for two types of
OVA-specific antibodies, for example, IgG and IgA (Fig. 6).
1. To coat ELISA plates, dispense 100 μL/well of 0.1 μg/μL
OVA in ELISA coating buffer (i.e., 10-μg OVA per well).
2. Seal the plate with plastic wrap and incubate overnight at 4
C.
3. Remove contents and wash the plate with 300 μL/well PBST
three times.
4. Dispense 150 μL/well of 1% skim milk in PBST to block.
222 Satoru Hirayama and Ryoma Nakao

Fig. 6 Example 96-well plate layout for samples assayed for OVA-specific IgG and IgA antibodies. Sets of
samples (saliva, serum, or nasal wash) are tested in duplicate (each at a single dilution) for OVA-specific IgG
(top half of the plate) and IgA (bottom half of the plate) by using the appropriate enzyme-linked detection
antibody

5. Seal the plate with plastic wrap and incubate for 2 h at 37
C.

6. Remove contents and wash the plate with 300 μL/well PBST
three times.
7. Appropriately dilute the samples prepared in Subheading 3.4.2
with 0.5% skim milk in PBST (see Note 34).
8. Add 100 μL/well of each of diluted sample to appropriate wells
on the plate (Fig. 6).
9. Seal the plate with plastic wrap and incubate at 37
C for 1 h.
10. Remove contents and wash the plate with 300 μL/well PBST
three times.
11. Add 100 μL/well of appropriately enzyme-linked detection
antibody, diluted in 0.5% skim milk in PBST to the concentra-
tion recommended by the manufacturer.
12. Seal the plate with plastic wrap and incubate at 37
C for 1 h.
13. Remove contents and wash the plate with 300 μL/well PBST
three times.
14. Add 100 μL/well of substrate solution into the plate.
15. Seal the plate with plastic wrap and incubate at 37
C.
16. Sequentially (e.g., 10 min, 30 min, 60 min) measure the absor-
bance of 405 nm (for AP) or 650 nm (for HRP) (see Note 27).
 Low Endotoxin Glycine-Induced Membrane Vesicles 223

4 Notes

1. Nissle 1917 is a representative probiotic E. coli strain. The

flagella-deficient derivative is used as an MV producer, in
order to completely exclude flagella from the MV preparation
without any sample loss [5].
2. If it is difficult to dissolve, heat the solution at 50
C.
3. If crystals form during storage, they will dissolve when heated
at 50
C.
4. Filtration using bottle top filter unit and aspirator is convenient
for rapid filtration of bacterial supernatant. Depending on the
bacterial species and strains, a 0.22-μm PVDF filter may also be
required. Check for sterility of flowthrough by plating on LB
agar plate. Resterilize using 0.45- or 0.22-μm filters as needed.
5. Filtration is recommended to remove contaminants in buffer.
6. As a solvent for suspending MV pellets, PBS, Tris–HCl, dis-
tilled water, or another buffer can be used, depending on the
purpose of the individual study.
7. It is preferable to use a 96-well plate with low protein
adsorption.
8. Refer to the product standard certificate for each lot to calcu-
late the EU of LPS.
9. Different solutions may be required depending on the LAL
assay kit used.
10. We recommend that J774.1 cells are cultured on nontreated
culture surfaces and passaged by simple dilution method (e.g.,
1:10 split), because J774.1 cells are very sticky on tissue culture
treated surface. Even if treated with trypsin-EDTA, it is diffi-
cult to efficiently detach J774.1 cells from the tissue culture
treated surface. Cell scrapping (e.g., by cell scraper) is not
recommended as it is known to affect cell viability.
11. Purchase animals at least 1 week before starting experiments.
12. An anesthesia bottle can be used instead.
13. Prepare and mix the solutions of isoproterenol and pilocarpine
just prior to intraperitoneally administrating the mixture
to mice.
14. It is also possible to cut the tip of the 21G beveled needle to
make it nonbeveled.
15. Due to the low MV yield of the noninduced control, multiple
bacterial culture supernatants (e.g., from 5–8 flasks of 80-mL
cultures) are required to measure the amount of protein and
lipid in later step.
224 Satoru Hirayama and Ryoma Nakao

16. Effective glycine concentrations may vary by bacterial species

and strain. It is important to consider the optimum glycine
concentration for the target strain.
17. If necessary, monitor the degree of increase in turbidity of the
bacterial culture over time.
18. Glycine-induced MV pellets are large. Therefore, be careful to
not create foam when pipetting to uniformly mix them in the
buffer. In this case, it is useful to control flow speed/volume of
pipetting by cutting the end of pipet-tips. On the other hand,
due to the low MV yield from noninduced culture (see Note
15), the MV pellets must be collected together from several
flask cultures and suspended with a small amount of buffer. If
necessary, to prepare MVs with greater purity, perform further
purification by repetitive washing or density gradient
centrifugation [12].
19. In general, antigens carried by MVs are more stable against
degradation by heat or enzymes, as compared with those in
free-form. Nevertheless, the suitable storage conditions of
MVs must be determined by preliminary experiments, because
stability of MVs may not only differ among individual strains,
but also vary due to the preparation method. We usually pre-
pare MVs in an appropriate solution, store at 20
C, and use
for testing within 1 year after isolation. For longer storage
(e.g., ~2 years), MVs prepared in an appropriate solvent can
be also lyophilized by freeze-dryer apparatus then stored at
4
C. Prior to use for experiment, lyophilized MVs must be
reconstituted with the same volume of distilled water as before
lyophilizing.
20. In most cases, optimal results will be obtained by testing each
sample without dilution or by testing different serial dilutions
(e.g., 2, 4 and eightfold).
21. We recommend measuring standards and samples in duplicate
or triplicate.
22. Depending on the MV sample, it may aggregate, making pro-
tein measurement difficult. In that case, use dye reagent for
BCA method [13] instead.
23. The number of twofold serial dilutions required depends on
the yield of MVs and the amount of buffer used for suspension
of pellet. We suggest that 1/5 diluted sample is used as starting
material, and further make six twofold serial dilutions for
testing.
24. In this case, 5 culture dishes of 10-mL S-RPMI (i.e., 50 mL
total) will yield more than 2  107 J774.1 cells. This is suffi-
cient for performing an assay using 21 wells in a 24-well plate,
accommodating triplicate samples of the following 7 groups:
 Low Endotoxin Glycine-Induced Membrane Vesicles 225

glycine-induced MVs at three different concentrations (0.01,

0.1, and 1 ng-protein/mL), noninduced MVs at three differ-
ent concentrations (0.01, 0.1, and 1 ng-protein/mL), and a
negative control (PBS alone).
25. The copy number of the target gene should be normalized by
dividing it by the copy number of the internal control (e.g.,
β-actin).
26. A single dilution of each supernatant sample may be tested in
triplicate, or each sample may be tested in three different dilu-
tions (serial twofold dilutions). How much the sample needs to
be diluted depends on the cytokine being tested and may need
to be determined empirically. Depending on the cytokine
being tested, a dilution of ca. ~16 fold may be required.
27. When using HRP-linked detection antibody and the
corresponding substrate, 100 μL/well of 1 N HCl can be
added at the appropriate time to stop the HRP reaction and
measure the absorbance at 450 nm. As the solution changes
from blue to yellow, its absorbance increases.
28. It is mandatory that approval by the appropriate ethical com-
mittee as well as related government agency be obtained before
performing any animal experiments. All mice must be housed
and maintained in a suitable animal facility according to all
regulations/laws applicable to animal welfare.
29. Mouse inhalation anesthesia can be performed with isoflurane
at a concentration of 2% and flow rate of 2 L/min.
30. Insufficient anesthesia may cause mice to sneeze. It is best to
wait 1 min after inoculation of one nasal cavity before inoculat-
ing the other so that the first dose does not block the airway,
thus lowering the possibility of sneezing.
31. Storage at 80
C reduces viscosity. The storage is recom-
mended to easily handle saliva samples.
32. Slowly drain the blood to prevent heart collapse.
33. Approximately 1 mL can be collected.
34. In most cases, optimal results will be obtained by testing 1000-
fold dilution for serum IgG, and 100-fold dilutions for serum
IgA, IgM, IgE, salivary S-IgA, and nasal wash S-IgA.

Acknowledgments

The authors wish to thank Michiyo Kataoka for technical assistance

with the transmission electron microscope observations. This study
was supported by JSPS KAKENHI (JP18K15160, JP19K22644,
JP20K18492, JP20H03861) and the Japan Agency for Medical
Research and Development (AMED) (JP18fk0108124).
226 Satoru Hirayama and Ryoma Nakao

References
1. Schwechheimer C, Kuehn MJ (2015) Outer- 8. Rembacken B, Snelling A, Hawkey P,
membrane vesicles from gram-negative bacte- Chalmers D, Axon A (1999) Non-pathogenic
ria: biogenesis and functions. Nat Rev Micro- Escherichia coli versus mesalazine for the treat-
biol 13(10):605–619 ment of ulcerative colitis: a randomised trial.
2. McBroom AJ, Johnson AP, Vemulapalli S, Lancet 354(9179):635–639
Kuehn MJ (2006) Outer membrane vesicle 9. Kruis W, Frič P, Pokrotnieks J, Lukáš M,
production by Escherichia coli is independent Fixa B, Kaščák M, Kamm M, Weismueller J,
of membrane instability. J Bacteriol 188 Beglinger C, Stolte M (2004) Maintaining
(15):5385–5392 remission of ulcerative colitis with the probiotic
3. Turnbull L, Toyofuku M, Hynen AL, Escherichia coli Nissle 1917 is as effective as
Kurosawa M, Pessi G, Petty NK, Osvath SR, with standard mesalazine. Gut 53
Cárcamo-Oyarce G, Gloag ES, Shimoni R (11):1617–1623
(2016) Explosive cell lysis as a mechanism for 10. Behnsen J, Deriu E, Sassone-Corsi M, Raffa-
the biogenesis of bacterial membrane vesicles tellu M (2013) Probiotics: properties, exam-
and biofilms. Nat Commun 7(1):1–13 ples, and specific applications. Cold Spring
4. Toyofuku M (2019) Bacterial communication Harb Perspect Med 3(3):a010074
through membrane vesicles. Biosci Biotechnol 11. Ichinohe T, Watanabe I, Ito S et al (2005)
Biochem 83(9):1599–1605 Synthetic double-stranded RNA poly(I:C)
5. Hirayama S, Nakao R (2020) Glycine signifi- combined with mucosal vaccine protects
cantly enhances bacterial membrane vesicle against influenza virus infection. J Virol 79
production: a powerful approach for isolation (5):2910–2919
of LPS-reduced membrane vesicles of probiotic 12. Hirayama S, Nakao R (2021) Intranasal vaccine
Escherichia coli. Microb Biotechnol 13 study using Porphyromonas gingivalis mem-
(4):1162–1178 brane vesicles: isolation method and applica-
6. Vollmer W, Blanot D, De Pedro MA (2008) tion to a mouse model. Methods Mol Biol
Peptidoglycan structure and architecture. 2210:157–166. https://doi.org/10.1007/
FEMS Microbiol Rev 32(2):149–167 978-1-0716-0939-2_15
7. Hammes W, Schleifer K, Kandler O (1973) 13. Smith PK, Krohn RI, Hermanson GT, Mallia
Mode of action of glycine on the biosynthesis AK, Gartner FH, Provenzano M, Fujimoto
of peptidoglycan. J Bacteriol 116 EK, Goeke NM, Olson BJ, Klenk D (1985)
(2):1029–1053 Measurement of protein using bicinchoninic
acid. Anal Biochem 150(1):76–85
 Chapter 14

Methods for Assessment of OMV/GMMA Quality and Stability

Francesca Micoli, Renzo Alfini, and Carlo Giannelli

Abstract
Outer membrane vesicles (OMV) represent a promising platform for the development of vaccines against
bacterial pathogens. More recently, bacteria have been genetically modified to increase OMV yield and
modulate the design of resulting particles, also named generalized modules for membrane antigens
(GMMA). OMV/GMMA resemble the bacterial surface of the pathogen, where key antigens to elicit a
protective immune response are and contain pathogen-associated molecular patterns (e.g., lipopolysacchar-
ides, lipoproteins) conferring self-adjuvanticity. On the other hand, OMV/GMMA are quite complex
molecules and a comprehensive panel of analytical methods is needed to ensure quality, consistency of
manufacture and to follow their stability over time. Here, we describe several procedures that can be used
for OMV/GMMA characterization as particles and for analysis of key antigens displayed on their surface.

Key words Outer membrane vesicles, OMV, GMMA, Vaccines, Analytical methods

1 Introduction

During the last years outer membrane vesicles (OMV) have

received great attention as platform for the development of vaccines
against bacterial pathogens [1]. Gram-negative bacteria, during
growth, naturally release small bilayered membrane structures
from the cell surface with formation of native OMV (nOMV), a
process that has been associated with several biological functions
[2]. OMV release often happens at levels that are too low to
support vaccine manufacture and different ways have been identi-
fied to increase OMV shedding. In particular, vesicle-like aggre-
gates of insoluble outer membrane proteins can be chemically
extracted from whole bacteria using detergents (e.g., deoxycholate)
resulting in detergent-extracted OMV (dOMV), or bacteria can be
genetically manipulated to increase blebbing resulting in the
so-called mutant-derived OMV (mdOMV) or GMMA (generalized
modules for membrane antigens) [3, 4]. By this way, vesicle integ-
rity is preserved and GMMA present outer membrane antigens in
their native environment and conformation. Additional mutations

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_14,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

227
228 Francesca Micoli et al.

can be introduced to reduce OMV endotoxicity, more often

through modification of the lipid A structure [5]. OMV producer
strains can also be engineered for expression of specific protein or
polysaccharide antigens, eventually supporting the development of
multicomponent vaccines [6]. Simplicity of manufacture make
OMV an attractive technology for affordable vaccines [7]. Further-
more, OMV resemble the bacterial surface of the pathogen, where
key antigens to elicit a protective immune response are and contain
pathogen-associated molecular patterns, such as lipopolysacchar-
ides (LPS) or lipoproteins, that confer self-adjuvanticity [5]. OMV
have optimal size for cellular uptake and display antigens in multi-
plicity, favoring B cells activation, all factors contributing to make
OMV a promising vaccine platform [1, 6, 8]. On the other hand,
OMV are quite complex molecules (Fig. 1) and development of a
comprehensive panel of analytical methods is needed to ensure
quality, consistency of manufacture and to follow OMV stability
over time. Here, we describe a number of procedures (Table 1) that
can be used for OMV characterization as particles and for analysis of
key antigens displayed on their surface.

Outer membrane proteins

LPS including OAg, core and Lipid A
Phospholipids
Periplasmic proteins

lumen

LPS OM

PP PG
IM

Cytoplasm

Abbreviations
LPS: Lipopolysaccharides
OAg: O-antigen
OM: Outer membrane
PG: Peptidoglycan
PP: Periplasm
IM: Inner membrane

Fig. 1 Schematic representation of OMV formation and their main components

 OMV/GMMA Analytics 229

Table 1
Methods for OMV/GMMA characterization

Attribute Method Performed on References

Identity Dot blot/western blot GMMA –
Particle size DLS/NTA/HPLC-SEC MALS GMMA [9]
Purity: Soluble proteins HPLC-SEC (fluorescence GMMA [9]
emission)
Purity: DNA content HPLC-SEC (ABS260/ABS280) GMMA [9]
Protein pattern/composition SDS-PAGE/MS GMMA [10, 11]
Total protein content Micro BCA/Lowry/amino GMMA [12]
acidic analysis
Total sugar content HPAEC-PAD/Dische/cELISA GMMA or extracted [13–15]
OAg/core
Lipid A structure MALDI-MS Isolated lipid A [16]
Lipid A amount HPLC RP-QqQ GMMA –
OAg molecular size HPLC-SEC/semicarbazide Extracted OAg/core [14]
distribution and molar ratio
of OAg of different length
1
OAg O-acetylation level H NMR/Hestrin Extracted OAg [17]
Key proteins quantification SRM/cELISA GMMA [10, 18]
OAg: O-antigen portion of lipopolysaccharide (LPS) molecules

2 Materials

Prepare all solutions using ultrapure water (grade 1, >18 MΩ-cm at

25
C; prepared by purifying deionized water) and analytical grade
reagents. Prepare all reagents at room temperature (RT), unless
otherwise specified.

2.1 Dot 1. Phosphate buffered saline with 0.05% Tween 20 (PBS/T):

Blot/Western Blot weigh 0.5 g of Tween 20 in a weighing boat and transfer into
a 1000 mL graduated cylinder by washing weighting boat with
PBS. Add PBS up to 1000 mL and mix on a magnetic stirrer
until the solution is homogeneous.
2. PBS/T with 3% bovine serum albumin (PBS/T + 3% BSA):
weigh in a weighing boat 3 g of BSA and transfer in a 100 mL
graduated cylinder; add PBS/T solution up to 100 mL. Mix on
a magnetic stirrer.
3. PBS/T with 0.1% BSA (PBS/T + 0.1% BSA): Transfer 87 mL
of PBS/T in a 100 mL glass bottle and add 3 mL of PBS/T+
3% BSA. Mix on a magnetic stirrer.
230 Francesca Micoli et al.

4. IgG Mouse anti-Salmonella Typhimurium primary antibody

1:3000 in PBS/T + 0.1% BSA: Transfer 9 mL of PBS/T + 0.1%
BSA in a 15 mL Falcon tube; add 3 μL of mouse anti-S.
Typhimurium primary antibody (Abcam, code AB8274) and
mix by vortex. Prepare this solution just before use.
5. Goat anti-mouse IgG secondary antibody (AP conjugate)
1:6000 in PBS/T+ 0.1% BSA: Transfer 18 mL of PBS/T+
0.1% BSA in a 50 mL Falcon tube; add 3 μL of goat anti-
mouse IgG AP conjugate secondary antibody (Sigma, code
A3438) and mix by vortex. Prepare this solution just
before use.
6. Developing solution: In a 50 mL Falcon tube, solubilize
2 tablets of SIGMA Fast BCIP with 20 mL of water by vigor-
ously vortexing and keep it at 4
C protected from light until its
use (to be prepared just before use).

2.2 Dynamic Light 1. DLS instrument (i.e., Zetasizer Nano—Malvern Panalytical).

Scattering (DLS)

2.3 Size Exclusion 1. Tosoh TSK gel 6000 PW column (30 cm  7.5 mm; cat.
High-Performance 805765) connected in series with a Tosoh TSK gel 4000 PW
Liquid column (30 cm  7.5 mm; cat. 805763) and with Tosoh TSK
Chromatography gel PWH guard column (7.5 cm  7.5 mm; cat. 806732).
with Multiangle Light 2. HPLC system equipped with UV, fluorimeter, and multiangle
Scattering static light scattering (MALS) detectors (e.g., DAWN
(HPLC-SEC/MALS) HELEOS II, Wyatt).
3. Filter PBS eluent through 0.22 μm filter before use.

2.4 Nanoparticle 1. NS300 NanoSight instrument (Malvern) equipped with a

Tracking Analysis CMOS camera and a 488 nm monochromatic laser. Data
(NTA) acquisition and processing with NTA software.
2. Low-bind tubes.
3. Low-bind pipette tips.

2.5 SDS-Page 1. Prestained molecular weight standard.

2. 1 M (1,4-Dithioerythritol) (DTT) solution.
3. 10% acrylamide gels, Bis-Tris buffer, for SDS-PAGE.
4. Sample loading buffer (4) containing lithium dodecyl sulfate
and Coomassie G250.
5. Brilliant Blue G—Colloidal Concentrate stain: Dilute the con-
tent to the working concentration in the original container
according to the dye datasheet. Shake the working concentra-
tion dye container before use.
6. 3-(N-morpholino)propanesulfonic acid (MOPS) running
buffer: Dilute 25 mL of 20 MOPS running buffer with
475 mL of water in a graduate glass cylinder.
 OMV/GMMA Analytics 231

7. Fixing solution: Under chemical hood mix in a 100 mL

graduated cylinder 20 mL methanol (MeOH), 2 mL acetic
acid and then add water to reach 50 mL.

2.6 Micro BCA 1. Micro BCA Protein Assay Kit (Thermo Fisher, cat. 23235).

2.7 Ultracentri- 1. Ultracentrifuge equipment with appropriate rotor (see Note 1).
fugation

2.8 Lowry 1. 0.8 N sodium hydroxide (NaOH): Weigh 1.6 g of NaOH,

transfer it in a 50 mL Falcon tube and dissolve into a final
volume of 50 mL water.
2. 20% sodium carbonate: Weigh 4 g of sodium carbonate, trans-
fer it in a 100 mL cylinder and dissolve into a final volume of
20 mL water.
3. 0.4% Potassium Tartrate/0.2% cupric sulfate: Weigh 200 mg of
(CHOHCOOK)2 ½ H2O, transfer it in a 50 mL Falcon tube
and add 25 mL of water. Weigh 100 mg of CuSO4 5 H2O,
transfer it in the same Falcon tube. Dissolve it by vortex and
add 25 mL of water to reach a final volume of 50 mL.
4. 10% sodium dodecyl sulfate (SDS): Weigh 5 g of SDS, transfer
it in a 100 mL cylinder and dissolve into a final volume of
50 mL water.
5. Copper–tartrate–carbonate (CTC) solution: In a 15 mL Falcon
tube, add 3 mL of 0.4% potassium tartrate–0.2% cupric sulfate
solution and 3 ml of 20% sodium carbonate. Vortex the solu-
tion, wait 10 min and vortex the solution again before use.
6. Reagent A: In 50 mL Falcon tube, add 6 mL CTC Solution,
6 mL 10% SDS, 6 mL 0.8 N NaOH, 6 mL water and mix well
with the vortex mixer (see Note 2).
7. Folin reagent: In a 50 mL Falcon tube covered with aluminum
foil (to protect from light) add 5 mL water and 1 mL Folin–
Ciocalteu’s solution (Merck, 1.09001) (see Note 2).
8. 120 μg/mL BSA standard: Weigh 120 μL of 2 mg/mL BSA in
a Falcon tube. Calculate the amount of water to add using the
following formula:
Water ðmgÞ ¼ 16:67  2 mg=mL BSA weight ðmgÞ  2 mg=mL BSA weight ðmgÞ

2.9 Amino Acid 1. Hydrolysis tubes (Waters, cat. WAT079007 or WAT007571).

Analysis 2. Protein low-bind tubes, 2 mL.
3. Total Recovery HPLC vials (Waters, cat. 186000384c).
4. Kit AccQTag Ultra (Waters, cat. 186003836).
232 Francesca Micoli et al.

5. Hydrolyzed Amino Acids Standard (Waters, cat.

WAT088122).
6. UPLC BEH C18 130 Å, 1.7 μm, 2.1 mm 100 mm column
(Waters, cat. 186,003,837).
7. UHPLC system equivalent to Waters ACQUITY UPLC
H-CLASS, equipped with UV detector.
8. Column preheater (Waters, cat. 205000730).
9. Post Column connector PEEK id 0.0025 (Waters, cat.
430001783).
10. AccQTag Ultra Eluent B/10 (eluent line B): Filter 0.2 μm 1 L
water. With a 1 L graduated cylinder measure accurately
900 mL of filtered water and transfer it in an appropriate bottle
for eluent. With a 100 mL graduated cylinder measure accu-
rately 100 mL of AccQTag Ultra Eluent B and transfer in the
eluent bottle previously filled with water. Mix accurately.
11. HPLC Needle/seal washing solution: Prepare in a bottle a
solution with 50% water/acetonitrile (ACN).
12. 6 N Hydrochloric acid (HCl) in single-use ampules.
13. Device to perform dry hydrolysis in HCl vapor phase: Alfatech
Hydrosmart 2 with hydrolysis vessels, cat. 20001001.
14. 1% Phenol in 6 N HCl: Under the chemical hood transfer a
maximum of 40 mg of phenol powder inside a weighed vial,
close the vial and weigh it (this procedure is needed to avoid
phenol breathing). Determine the amount of 6 N HCl for
phenol dissolution using the following formula:
μL 6 N HCl ¼ 100  mg Phenol
Transfer the required volume of 6 N HCl into the vial
containing the weighed phenol. Close the screw cap vial and
vortex until complete dissolution.
15. 100 mM HCl: In a 25 mL graduated cylinder place 418 μL 6 N
HCl and add water up to reach 25 mL total volume.
16. 6 N HCl, 0.1% phenol solution: In a 2 mL glass vial mix 630 μL
of 6 N HCl and 70 μL of 1% Phenol solution in 6 N HCl.
Prepare it just before use.

2.10 High-Pression 1. 11.25 μg/mL Rhamnose (Rha), Galactose (Gal), Glucose

Anion Exchange (Glc), Mannose (Man) sugars mix (neutral sugar) standard
Chromatography solution: In different 2 mL vials, weigh accurately 9 mg of
Coupled with Pulsed each standard and dissolve them in a suitable amount of water
Amperometric added by weighing on balance, in order to obtain solutions at
Detection 4.5 mg/mL. To calculate the amount of water to be added to
(HPAEC-PAD) Analysis each vial, use the following formulas:
(Neutral Sugars)
 OMV/GMMA Analytics 233

μg sugar ¼ weight μg  purity%  (sugar MW without hydra-

tion water)/(sugar MW with hydration water),
(If the sugar has no hydration water, the last ratio present in the
formula is equal to 1).
μL water ¼ (μg sugar)/4.5
Weigh an empty 500 mL glass bottle and transfer 1 mL of
each sugar solution prepared above; then add water to a total
weight of 400 g in order to obtain a solution with a concentra-
tion of 11.25 μg/mL for each standard.
2. 4  250 mm CarboPac PA10 column (Thermo Fisher, cat.
046110) coupled with a 4  50 mm PA10 guard column
(Thermo Fisher, cat. 046115).
3. 8 M TFA solution: Under chemical hood, with the glass cylin-
der, measure 30 mL of TFA and transfer it into a glass bottle.
Add in the bottle 20.5 mL of water and gently mix.
4. 4 M TFA solution: Dilute 10 mL of 8 M TFA adding 10 mL of
water.
5. 50 mM NaOH: Fill the eluent bottle with 2 L of water using a
volumetric flask. Degas for 15 min by bubbling helium. Add
5.2 mL of 50% NaOH to the bottle. Degas for 10 min more.
6. 500 mM NaOH: Fill the eluent bottle with 2 L of water using
the volumetric flask. With a pipette, remove 26 mL of water
from the bottle. Degas for 15 min by bubbling helium. Add
26 mL of 50% NaOH to the bottle. Degas for 10 min more.
7. 1 M sodium acetate (AcONa) 100 mM NaOH: Degas 1.5 L of
water for 15 min by bubbling helium. Dissolve the whole
content of preweighed AcONa (Thermo Fisher, cat. 059326)
by adding 500 mL of degassed water directly into the AcONa
bottle. Transfer the solution to a 1 L volumetric flask. Wash
twice the AcONa bottle with about 100 mL more of degassed
water and add to the solution in the 1 L volumetric flask. Fill
the 1 L volumetric flask with degassed water, up to reach 1 L of
volume for the AcONa solution. Place a 0.22 μm vacuum-cup
filter on an empty eluent 2 L bottle, connect it to the vacuum
system and filter the prepared 1 L solution. Degas the solution
for 10 min by bubbling Helium. Add 5.2 mL of 50% NaOH to
the bottle. Degas with Helium for 10 min more.

2.11 HPAEC-PAD 1. 4  50 mm CarboPac PA1 guard column (Thermo Fisher, cat.

Analysis (Amino 043096), 4  250 mm CarboPac PA1 column (Thermo Fisher,
Uronic Acid) cat. 035391).
2. TFA-HCl Mixture: According to the volume needed (3 mL for
each sample and 14 mL for the calibration curve), prepare the
quantity of solution needed in a glass bottle according to the
volumes reported in Table 2.
234 Francesca Micoli et al.

Table 2
Preparation of acid solution “TFA-HCl Mixture”

TFA HCl
Final volume mL mL mL
20 2.6 17.4
25 3.3 21.7
30 3.9 26.1
35 4.6 30.4
40 5.2 34.8
45 5.9 39.1
50 6.5 43.5
55 7.2 47.8
60 7.8 52.2
65 8.5 56.5

2.12 Dische 1. Disposable polystyrene cuvette with polypropylene sealing cap.

Colorimetric Method 2. Sulfuric acid: In 50 mL glass bottle add 40 mL of sulfuric acid.
Close the bottle and cool in ice for at least 1 h before use.
3. 1 M cysteine: inside an Eppendorf tube, weigh cysteine hydro-
chloride to achieve a weight of 157.6 mg. Add water and mix
by vortexing. Calculate the volume of water to add by using the
following formula:
mg cysteine hydrochloride
water to be added ðmLÞ ¼
157:6
Keep cooled in ice before usage. This solution has to be
prepared and used the same day.

2.13 High- 1. Luna 3u C8(2), 50  2 mm, 100A (Phenomenex, cat.

Performance Liquid 00B-4248-B0).
Chromatography Mass 2. SPE online cartridge holder, MercuryMS 20 mm cartridge
Spectrometry holder (Phenomenex, cat. CH0–5845).
(HPLC-MS) (See 3. Strata-X 25 μm On-Line Extraction, Cartridge 20  2.0 mm
Note 3) (Phenomenex, cat. 00M-S033-B0-CB).
4. 50% Isopropyl alcohol (IPA) solution (v/v): Using a 100 mL
graduated cylinder, measure 40 mL of IPA and transfer it in a
bottle. Measure then 40 mL of water and add to the bottle,
mixing with the IPA.
5. 5 nmol/mL (1220 ng/mL) 3-hydroxymyristic acid
(3OH-My) in IPA/Water 1:1 (v/v): Weigh 10 mg of
3OH-My in a low-bind tube and dissolve them in a suitable
amount of IPA added by weight on balance, in order to obtain
 OMV/GMMA Analytics 235

a 10 mg/mL solution. To calculate the amount of IPA to be

added use the following formula:
weight μg
mg IPA ¼  purity%  0:785
10
Then, to obtain a 1.22 mg/mL 3OH-My solution in IPA,
in a low-bind tube, weigh accurately 150 mg of the solution
prepared and add by weight the quantity of IPA calculated with
the following formula:


weight mg
mg IPA ¼  1230  weight mg
150
Weigh (tare) then an empty 1 L glass bottle and transfer
1000 μL of this solution, add IPA up to achieve a total weight
of 392.5 g (500 mL density 0.785 g/mL). Add then 500 g of
water to achieve a total weight of 892.5 g (500 mL) and mix.
1 mL aliquots of the standard solution in suitable cryotubes can
be stored at 80
C.
6. 5 nmol/mL (1080 ng/mL) 3-hydroxylauric acid (3OH-La)
IPA/Water 1:1 (v/v): Weigh about 10 mg of 3OH-La in a
low-bind tube and dissolve them in a suitable amount of IPA
added by weight on balance, in order to obtain a 10 mg/mL
solution. To calculate the amount of IPA to be added use the
following formula:
weight μg
mg IPA ¼  purity%  0:785
10
Then, to obtain a 1.08 mg/mL 3OH-La solution in IPA,
in a low-bind tube, weigh accurately 150 mg of the solution
prepared and add by weight the quantity of IPA calculated with
the following formula:


weight mg
mg IPA ¼  1389  weigh mg
150
Weigh (tare) an empty 1 L glass bottle and transfer 1000
μL of this solution, add then IPA up to achieve a total weigh of
392.5 g (500 mL density 0.785 g/mL). Add then 500 g of
water to achieve a total weight of 892.5 g (500 mL) and mix.
1 mL aliquots of the standard solution in suitable cryotubes can
be stored at 80
C.
7. 1.5 M NaOH: In a 50 mL Falcon tube add 500 μL of 50%
NaOH to 5.9 mL of water.
8. Eluent A (40%MeOH, 0.05% formic acid (FA)): In a 200 mL
graduated cylinder measure 80 mL MeOH, add 100 μL FA and
finally add water up to 200 mL. Transfer into a 250 mL glass
bottle.
236 Francesca Micoli et al.

9. Eluent B (70% ACN, 0.05% FA): In a 200 mL graduated

cylinder measure 140 mL of ACN, add 100 μL FA and finally
add up to 200 mL of water. Transfer into a 250 mL glass bottle.
10. Eluent C (0.1% TFA): With a graduated cylinder measure
100 mL of water, transfer into a glass bottle and add 100 μL
of TFA.
11. Eluent D (33% ACN in ISP, 0.1% TFA): With a graduated
cylinder measure 33 mL of ACN, transfer into a glass bottle.
Measure then with a graduated cylinder 67 mL of isopropyl
alcohol and mix it with the ACN in the glass bottle. Add 100
μL of TFA and mix.
12. Eluent E (65% ACN): In a 100 mL graduated cylinder measure
65 mL of ACN and add water to reach 100 mL. Transfer in a
100 mL glass bottle.

2.14 Matrix-Assisted 1. 50% ACN: In a 15 mL Falcon tube, add 1 mL of water, then

Laser 1 mL of ACN.
Desorption/Ionization 2. 200 mg/mL Super-DHB matrix solution: Weigh 5 mg of
Time-of-Flight Mass Super-DHB matrix in a 500 μL Eppendorf tube, then dissolve
Spectrometry in 25 μL of 50% Water/ACN and vortex. The solution has to
(MALDI-TOF MS) be used the same day of preparation (mass spectrum acquisition
day).
3. Standard MW peptide: Fill an Eppendorf tube with 100 μL of
water and add an aliquot (5 μL) of Peptide calibration standard,
mix by vortexing.
4. 5% acetic acid (v/v): In a 15 mL Falcon tube, put 1.9 mL of
water and add 100 μL of acetic acid.
5. 4:1 chloroform–MeOH (v/v): In a 2 mL glass screw cap vial,
mix 200 μL of MeOH and 800 μL of chloroform.

2.15 Competitive- 1. 10 Sample Dilution buffer (SDB) (10 PBS containing 1.0%
ELISA (cELISA) BSA and 0.5% Tween 20): Add 10 PBS into a graduated
cylinder up to 1.8 L. Add gently 10.0 mL of Tween 20 into
the cylinder followed by 20.0 g BSA. Stir for 1 h and if needed
continue stirring until the BSA is completely dissolved. When
the BSA is completely dissolved, add 10 PBS up to 2 L and
stir again the solution for few min. Check the pH of the
solution (which should be in the range 6.5  0.5) using a pH
meter or a pH paper. Filter-sterilize the solution into two 1 L
bottles using two bottle-top 0.22 μm filter units.
2. SDB working solution: In a graduated cylinder add water up to
the volume needed to have a 1:10 dilution, according to the
total volume to be prepared. Add 10 SDB (e.g., one 50 mL
aliquot 10 SDB in 450 mL water). Stir for few min with a
magnetic stirrer. Filter-sterilize the solution into 1 L bottles
using a bottle-top 0.22 μm filter unit.
 OMV/GMMA Analytics 237

3. Blocking buffer (5% fat-free milk in PBS): Add 800 mL of 1

PBS into a graduated cylinder. Add 50 g of skim milk powder
to the cylinder containing 1 PBS. Stir for 30 min. Check to
ensure that no lumps of powdered milk remain, especially on
the top surface. If so, break up the lumps and continue stirring
for 30 min. When the milk is dissolved, add 1 PBS after
checking pH (in the range of 7.4  0.3) up to 1 L total volume
and stir again for few min. Transfer the solution in a sterile glass
bottle.
4. 20 Washing Buffer (20 PBS containing 1% Tween 20): Add
into a graduated cylinder around 1.8 L of 20 PBS and 20 mL
of Tween 20. Stir for 1 h. When the Tween 20 is completely
mixed, add 20 PBS up to 2 L and stir again for few min.
Check the pH of the solution using a pH meter or a pH paper,
that needs to be in the range of 6.5  0.5. Transfer the solution
in a sterile glass bottle.
5. Washing Buffer solution: Take one 2 L cylinder, put 100 mL of
neat solution and add Water up to 2.0 L total volume. Stir for
few min with a magnetic stirrer and transfer in a bottle.
6. Carbonate coating buffer solution (0.05 M Carbonate buffer,
pH 9.6): Use the Carbonate Buffer capsules from Sigma-
Aldrich C3041-100CAP to prepare the volume of coating
buffer as needed. Put inside the bottle a magnetic bar and
pour the contents of one capsule for every 100 mL of buffer
to be prepared. Add water up to the final volume. Stir until the
tablets are dissolved. When the salts are dissolved, filter the
solution and check its pH using a pH meter or a pH paper. The
pH needs to be in the range 9.6  0.3.
7. 10x Phosphate coating buffer (500 mM Sodium Phosphate
Buffer, pH 7.0): dissolve 29.2 g of NaH2PO4.H2O and
77.3 g Na2HPO4.7H2O in 1 L of water.
8. Phosphate coating buffer: dilute with water 1:10 the 10
Phosphate coating buffer, stir for few min with a magnetic
stirrer and filter 0.22 μm the solution.

2.16 OAg Extraction 1. 10 Acetate buffer (final 100 mM pH 3.9 after ten-fold dilu-
tion): Weigh 1099 mg of AcONa and transfer it in a 100 mL
graduated cylinder. Add, in sequence, 50 mL of water and
4950 μL of acetic acid. Add water up to 100 mL, insert a
magnetic stirring bar and homogenize the solution.

2.17 Size Exclusion 1. Tosoh TSK gel G3000 PWXL column (30 cm  7.8 mm; cat.
High-Performance 808021) with a Tosoh TSK gel PWXL guard column
Liquid (4.0 cm  6.0 mm; cat. 808033).
Chromatography, 2. Eluent 0.1 M NaCl, 0.1 M NaH2PO4, 5% ACN, pH 7.2:
Derivatization Weigh 5.85 g of NaCl, 3.22 g of NaH2PO4 and 10.37 g of
with Semicarbazide Na2HPO4 and transfer the powders inside a 1 L glass
(HPLC-SEC SCA) graduated cylinder. Add water up to 850 mL in the 1 L
238 Francesca Micoli et al.

graduated cylinder. Put a magnetic stir bar inside the cylinder

and solubilize the powder placing the cylinder on a magnetic
stirrer. Measure, in a 50 mL graduated cylinder, 50 mL ACN
and add them in the 1 L cylinder with the buffer solution. Add
water to bring the total volume to 1 L. Leave the cylinder on
the stirrer until the solution is completely homogeneous.
Filter-sterilize with a 0.22 μm filter.
3. λ-DNA Molecular Weight Marker III 0.12–21.2 kb (Roche;
cat. 10528552001): dilute 20 with HPLC eluent before
injection.
4. 200 mg/L Sodium azide (NaN3) solution.
5. Semicarbazide solution: dissolve 100 mg Semicarbazide hydro-
chloride and 90.5 mg of AcONa anhydrous in 10 mL of water.
6. 40 μg/mL 3-Deoxy-D-manno-oct-2-ulosonic acid (KDO)
ammonium salt solution (the concentration value is referred
to ammonium salt).

2.18 Nuclear 1. 4 M Sodium deuteroxide (NaOD): Weigh 160 mg of NaOH

Magnetic Resonance and transfer in a tube. Calculate with the following formula the
(NMR) quantity of deuterium hydroxide (D2O) needed to achieve 4 M
solution and add to the powder.
sodium hydroxide ðmgÞ
Total mL of water ¼
160
Close the tube and mix on a vortex. Keep the tube closed
until use in order to limit carbonation.

2.19 Hestrin To prepare solutions from powders, use magnetic stir bars and a
Colorimetric Method magnetic stirrer to accelerate the dissolution.
1. 1 M Acetic acid: Dilute in a 100 mL glass graduate cylinder
5.7 mL of acetic acid up to 100 mL with water.
2. 1 mM AcONa solution, pH 4.5: Weigh 8.2 mg of AcONa
anhydrous and dissolve it in a 100 mL glass graduated cylinder
with 80 mL of water. The pH is adjusted to 4.5 with acetic acid
1 M solution. Calculate the total water volume of the solution
needed to get the correct concentration using the following
equation:
sodium acetate weight ðmgÞ
Total mL of water ¼
0:082
Add water to the cylinder in order to reach the calculated
volume.
3. 30 mg/mL acetylcholine: Resuspend the contents of a 150 mg
vial of acetylcholine chloride using 5 mL of AcONa solution
1 mM, pH 4.5. Prepare just prior use.
 OMV/GMMA Analytics 239

4. 2 M Hydroxylamine hydrochloride: Weigh 6.9 g of hydroxyl-

amine hydrochloride and dissolve it in a 50 mL glass gruated
cylinder with water up to 50 mL.
5. 3.5 M NaOH solution: Weigh 7.0 g of NaOH and dissolve it in
a 50 mL glass graduated cylinder with water up to 50 mL.
6. 4 M HCl: In a 50 mL Falcon tube dilute 10 mL of 37% HCl by
adding 20 mL of water.
7. 0.1 M HCl: In a 100 mL glass graduate cylinder, dilute 830 μL
of 37% HCl up to 100 mL with water. Transfer and store it in a
glass bottle.
8. 0.37 M Iron chloride solution in 0.1 M HCl: Weigh 3.0 g of
iron (III) chloride and dissolve it in a 50 mL glass graduated
cylinder with 0.1 M HCl up to 50 mL. Transfer the solution in
a glass bottle together with a stir bar that will be useful to
resuspend iron precipitate (if present) before use.
9. Basic hydroxylamine solution: In a 50 mL Falcon tube, dilute
10 mL of 2 M hydroxylamine solution with 10 mL of 3.5 M
NaOH solution. Mix by vortex.

3 Methods

3.1 Dot Dot blot/western blot are used for verifying OMV/GMMA iden-
Blot/Western Blot tity. Specific primary antibodies are used according to the key
antigen(s) present on the samples. In the Materials section we
have reported as an example the use of a primary antibody (and
corresponding secondary antibody) specific for S. Typhimurium
OAg, key component of S. Typhimurium GMMA [14].
1. Prepare a positive control (+ CTRL) by diluting the
OMV/GMMA sample to use as positive control to 8 μg/mL
(protein concentration) with PBS. Use PBS as negative control
( CTRL).
2. Dilute each OMV/GMMA sample to be analyzed to 8 μg/mL
(protein concentration) with PBS.
3. Cut the 0.2 μm polyvinylidene difluoride (PVDF) membrane
(other membrane materials could be more suitable, depending
on the product type) in order to have a resulting area large
enough to cover the Dot-blot apparatus area.
4. Wet the PVDF membrane in a plastic case with a suitable
volume of methanol for about 1 min, manually shaking the
case in order to make the membrane entirely wet.
5. Transfer the membrane in another case filled with PBS and
leave the membrane with PBS for 3 min, shaking manually
the case in order to facilitate the entire wetting of the
240 Francesca Micoli et al.

membrane with PBS. Perform all this preparation immediately

before placing the PVDF membrane in the Dot-Blot apparatus
(see Note 4).
6. Place the PBS-wet PVDF membrane in the Dot-Blot appara-
tus. Cut a small portion of the PVDF right bottom corner in
order to recognize the layout order at the end of the procedure.
Quickly close the Dot-Blot apparatus.
7. Spot 100 μL of samples/+ CTRL/ CTRL in each well (each
sample and controls are analyzed in duplicate) (see Note 5).
8. After deposition, switch on the vacuum system at about 150
mBar and keep it turned on until all the loaded samples are
dried (typically 1 min). Wait about 30 s more and then switch
off the vacuum. Slowly open the Dot-Blot apparatus and use
tweezers to get the membrane.
9. Transfer the membrane to a plastic case and block with
PBS/T + 3% BSA for 30 min under moderate agitation.
10. While blocking, prepare the primary antibody solution. At the
end of 30 min, transfer the membrane to a container contain-
ing 5 mL of primary antibody solution. Place the container on
a rocker at 4
C and leave under moderate agitation overnight
11. After the overnight incubation, place the container at RT for
10–15 min to let the temperature rise.
12. Remove the solution from the container and wash the mem-
brane three times with 5 mL of PBS/T for 10 min/wash,
under agitation.
13. Prepare the secondary antibody solution and incubate the
membrane with 5 mL of this solution for 1 h under moderate
agitation.
14. Wash the membrane three times with PBS/T for 10 min/
wash, under agitation, as before.
15. Prepare the developing solution during the washing time.
16. Incubate the membrane with 6 mL of the freshly prepared
developing solution and leave under moderate agitation until
the spots are clearly visible (typically 1–2 min).
17. Stop the reaction by washing the membrane three times with
5 mL of water for 5 min/wash, under agitation. Leave the
membrane in water. For the assay to be considered valid,
there should be no visible spot for the - CTRL (PBS), whereas
spots should be seen for the OMV/GMMA + CTRL, and both
replicates of the tested samples (see Note 6).

3.2 Dynamic Light DLS is one of the methods that can be used for OMV/GMMA
Scattering particle size determination [9]. This technique allows determina-
tion of a mean hydrodynamic diameter (Z-average diameter) and a
 OMV/GMMA Analytics 241

Size Distribution by Intensity

15
Intensity (Percent)

10

0
0.1 1 10 100 1000 10000
Size (d.nm)

Fig. 2 Z-average diameter determination by DLS for a GMMA sample as an example

polydispersity index (PDI), describing the amplitude of the distri-

bution. Hydrodynamic diameter is calculated using the Sto-
kesEinstein equation, obtaining the diffusion coefficient by
measuring intensity fluctuations of scattered light produced by
particles as they undergo Brownian motion.
The hydrodynamic diameter of the particles is expressed by a
Z-average value (see Note 7), providing also a polydispersity index
(PDI) of the size values calculated.
1. Adjust the settings on the Zetasizer Nano ZS as follows: Mea-
surement type ¼ Size, Material ¼ Protein, Dispersant ¼ buffer
in which the sample is diluted (see Note 8), General
options ¼ Mark-Houwink parameters (set as default), Temper-
ature ¼ 25
C, Equilibration time ¼ 120 s (see Note 9),
Measurement angle ¼ 173
Backscatter (NIBS default), Mea-
surement duration ¼ Automatic, Measurements ¼ 3, Delay
between measurement ¼ 30 s, Data processing ¼ Analysis
model General Purpose (normal resolution).
2. Dilute the sample to 50 μg/mL protein with an appropriate
buffer and transfer to a DLS cuvette.
3. Insert the cuvette in the cell holder, place the thermal cap on
top of the cuvette and close the lid.
4. Start the measurement. Figure 2 reports Z-average diameter
determination for a GMMA sample as an example.

3.3 HPLC-SEC/MALS HPLC-SEC analysis can be used to assess OMV/GMMA purity:

presence of soluble proteins and relevant amounts of DNA can be
detected. Use of MALS allows particle size determination too.
Presence of soluble proteins is assessed using the fluorescence
channel detecting presence of eventual peaks at different (higher)
elution times with respect to particles. For a quantitative estimate of
protein impurities, the ratio of the area of the OMV/GMMA peak
and of the area of soluble proteins can be used.
242 Francesca Micoli et al.

Presence of nucleic acids is assessed based on the ABS260/

ABS280 ratio on the UV channels. DNA content can be accurately
quantified by other methods like Threshold.
GMMA radii are evaluated with the MALS detector. The radius
is calculated for each point of the particle peak and the average
values, obtained using averaging formulas with different weights
(Rn, Rw and Rz), are reported. The ratio Rw/Rn indicates whether
the peak is homogenous with respect to radius: for a monodisperse
sample the average radius is independent of the averaging method
and the ratio is equal to 1.
1. Set the HPLC instrument as follows: column compartment
temperature ¼ 30
C, autosampler compartment tempera-
ture ¼ 4
C, UV detector as first detector after column with
acquisition channels 260 nm and 280 nm, fluorimeter detector
as second detector after column with excitation wavelength
280 nm/emission wavelength 336 nm, multi angle light scat-
tering detector as third detector after column.
2. Gently mix the sample in order to homogenize the content,
dilute with PBS or with an appropriate buffer to 150 μg/mL of
protein (sample concentration may be adapted depending on
sample type).
3. Run the sample injecting 80 μL eluting with PBS at 0.5 mL/
min flow rate and with 70 min run time (see Note 10). In
Fig. 3a the chromatogram (fluorescence channel) of a purified
S. Enteritidis GMMA sample (as an example) containing
neglectable quantity of soluble proteins is reported. In
Fig. 3b the chromatogram of a protein standards mixture
(BioRad cod 151-1901: 670-1.35 kDa) is reported, showing
the retention time range in which soluble proteins are
expected. Figure 3c represents the sample of GMMA spiked
with the protein standard mixture, clearly indicating ability of
the method to distinguish OMV/GMMA particles from solu-
ble proteins of different size. In Fig. 4 UV ABS chromatograms
at 260 and 280 nm are overlapped for a GMMA sample to
estimate the presence of nucleic acids. For sample containing
very low/no DNA impurities, the ratio of the areas of the two
chromatograms relative to the OMV/GMMA peak (at 260 and
280 nm respectively) is expected to be close to 1. Ratio > 1
reflects presence of DNA impurities. DNA impurities can be
detected also at lower size compared to OMV/GMMA with
the typical A260/A280 ratio.
4. Elaborate the MALS data using the following setting in the
Wyatt Astra software: Despiking ¼ heavy, Baseline ¼ auto find
baselines (check, selecting each detector, if the corresponding
found baseline is appropriate. Correct the starting/ending
points if needed (see Note 11), LS analysis ¼ Sphere as model
 OMV/GMMA Analytics 243

Fig. 3 HPLC-SEC analyses (fluorescence emission profiles) of a purified S. Enteritidis GMMA sample (a), of a
protein standards mixture (b) and of the GMMA sample spiked with such standard (c), showing ability of this
method to separate GMMA particles from soluble proteins of different size

Fig. 4 HPLC-SEC analysis (UV detection at 260 and 280 nm), purified S. Enteritidis GMMA

with fit degree 1 (see Note 12). On the 90
detector (LS11)
chromatogram, select as peak for the average radius calculation
the time range corresponding to 20% height in OMV/GMMA
peak leading and tailing (see Note 13). Check the fitting of the
244 Francesca Micoli et al.

Fig. 5 LS Detector 11 (90
) with radius, purified S. Enteritidis GMMA

data and eventually disable some detectors removing them

from the fitting. Chromatogram in Fig. 5 is reported as an
example.

3.4 Nanoparticle 1. Set the instrument as follows: minimum track length ¼ auto-
Tracking Analysis matic settings, minimal expected particle size and blur
(NTA) setting ¼ applied, viscosity settings for water with automatic
correction for the temperature, syringe pump speed 20.
2. Dilute the sample just before the analysis to 10–100 ng/mL of
protein (the concentration needs to be adjusted depending on
the sample) using low-bind material (pipette tips and tubes)
and acquire five replicate videos of 30 s at 25 frames per second,
generating five replicate histograms to be averaged.
3. Analyze particles movement with camera level ¼ 16, slider
shutter ¼ 1300 and slider gain ¼ 512. Test and adjust detec-
tion threshold value for the sample appearance. Figure 6 shows
a typical NTA graph for a GMMA sample.

3.5 SDS-Page SDS-PAGE analysis is used to look at protein pattern profile of

OMV/GMMA preparations. Total protein composition can be also
determined by MS [10, 11].
1. Dilute a GMMA sample to 300 μg/mL protein concentration.
2. In 1.5 mL Eppendorf tubes, add for each sample the following
quantities: 4.5 μL of LDS sample buffer (4), 1.5 μL of 1 M
DTT and 10 μL of Sample solution.
3. Spin down the liquids, vortex and spin down again.
4. Place the tubes in a block heater at 100
C for 10 min; then let
them cool at RT and spin down the liquid.
 OMV/GMMA Analytics 245

Concentration

1000
900
800
700
600
500
Inte 400
nsit 300
SEn BPR 60~28.nano
y (a 200
.u.) )
100 n(nm
Size SEn BPR~15-40-42
0

Fig. 6 NTA graph of S. Enteritidis GMMA (as an example), in which particle size, particle concentration, and
relative intensity are plotted together

5. Insert the precast gel in the electrophoresis cell after having

removed the strip present in the bottom of the gel.
6. Transfer MOPS buffer into the middle part of the cell up to
cover the gel wells and more up to 2 cm from the front bottom
of the cell.
7. Remove the wells protection from the gel, use the first well to
load 10 μL of the marker, load the samples in the remaining
wells.
8. Close the cell and connect it to the power supply.
9. Set up the power supply at 40 mA (constant) with 200 V as
maximum limit and run the gel (typically 75 min). Wait until
the blue marker of the wells reaches the bottom of the gel.
Open the cell and remove the gel from its holder using the gel
knife.
10. To fix the gel, place 50 mL of water into the staining container
and place the gel in it.
11. Remove the water and add 50 mL of fixing solution. Leave the
container closed on a rocker (set at 35 rpm) in the chemical
hood for 30 min at RT.
12. Remove the fixing solution from the container and discard into
the appropriate waste container.
13. Mix 40 mL of the dye solution and 10 mL of MeOH in a
50 mL Falcon tube, then add the staining solution in the
container with the gel and leave it (closed) on a rocker (set at
35 rpm) overnight at RT.
246 Francesca Micoli et al.

191 kDa

97 kDa

64 kDa

51 kDa

39 kDa

28 kDa

19 kDa
14 kDa

Fig. 7 SDS-PAGE analysis of S. flexneri 2a GMMA

14. Remove the staining solution from the container.

15. Wash the gel by adding to the container 50 mL of water and
leave it on a rocker (set at 35 rpm) for 10 min. Repeat this
operation twice.
16. Add 35 mL water and then 15 mL MeOH to the container
with the gel and leave it (closed) on a rocker (set at 35 rpm)
under chemical hood 1 h at RT.
17. Remove the solution from the container, then add 50 mL
water and leave it (closed) on a rocker (set at 35 rpm) under
chemical hood for 2 h at RT.
18. Collect a picture of the gel with a camera or scanner, as example
in Fig. 7 a gel of S. flexneri 2a GMMA is reported.

3.6 Micro BCA Micro BCA analysis is used to quantify OMV/GMMA total protein
content.
1. Set the spectrophotometer wavelength at 562 nm.
2. Set a thermostatic bath to 60
C.
3. To prepare 20 μg/mL BSA solution, open a BSA Standard
ampoule from the Micro BCA kit, mix gently. Transfer 100
μL of BSA 2 mg/mL in a Falcon tube and weigh. Calculate the
amount of water to add using the following formula:

Water ðmgÞ ¼ 100  2 mg=mL BSA weight ðmgÞ  2 mg=mL BSA weight ðmgÞ
 OMV/GMMA Analytics 247

Table 3
Dilutions for preparing calibration curve for Micro BCA analysis

Std conc. 20μg/mL BSA Water

μg/mL μL μL
0 0 500
1.8 45 455
2.4 60 440
5.0 125.0 375.0
10.0 250.0 250.0
15.0 375.0 125.0
20.0 500.0 0

4. In 5 mL tubes, prepare in duplicate the diluted solutions from

the 20.0 μg/mL BSA solution, using the volumes reported in
Table 3.
5. For each OMV/GMMA sample, first gently mix in order to
homogenize, then prepare three replicates for analysis. The
volume of sample, or diluted sample, required for each replicate
is 500 μL (see Notes 14 and 15).
6. Prepare a reagent mixture by combining the reagents from the
Micro BCA kit as follows: in a 50 mL Falcon tube, add 0.5 mL
of Reagent C and 12.0 mL of Reagent B, and mix well with the
vortex. Then add 12.5 mL of Reagent A and mix well with
vortex.
7. To every tube containing standards and samples, add 500 μL of
the freshly prepared reagent mixture.
8. Vortex each tube, cover with aluminum foil or Parafilm, and
incubate in the preheated thermostatic bath for 1 h.
9. Allow the tubes to cool at RT for 10 min before starting sample
readings.
10. Read each sample as per the following sequence: First read the
“blank” cuvette (i.e., containing water only), if a single beam
instrument is used (in order to subtract its absorbance from
samples/standards absorbances). If a double beam instrument
is used, leave the “blank” containing cuvette in the reference
slot during all sample/standard readings. Next, read the BSA
standard solutions (each in duplicate), in order of increasing
concentration, and finally the samples (each in triplicate).
11. Calculate a quadratic fitting regression between the ABS
measured and BSA concentrations (in the range 1.8–20 μg/
mL) and calculate on it the concentration of the sample
248 Francesca Micoli et al.

analyte. If the concentration of a sample is outside the range of

the calibration curve, dilute the sample with water and repeat
the analysis.

3.7 Ultracentri- The procedure is applied to separate OMV/GMMA particles from

fugation the solution, that is, to quantify soluble proteins eventually present
in the sample. By ultracentrifugation, OMV/GMMA (centrifuged
pellet) are separated from soluble proteins (centrifuge supernatant)
(see Note 16).
1. Dilute the sample to a maximum concentration of 1 mg/mL
(protein based).
2. Pellet OMV/GMMA by ultracentrifugation 4
C, 300 ,
110 k rpm, using a rotor with K-factor 15 (see Note 1).
3. For each sample tube, immediately (see Note 17) after ultra-
centrifugation ends, carefully transfer the supernatant with
pipette, without touching nor resuspending the
OMV/GMMA pellet, to a fresh tube for further analysis in
step 6 below.
4. Resuspend the pelleted OMV/GMMA fraction, after superna-
tant removal (see Note 18) and store the tube closed overnight
at 4
C.
5. The following day, pipet several times the solution in the tube,
resuspending and homogenizing the pellet.
6. Analyze the protein content of the supernatant and resus-
pended pellet samples, using either the Micro BCA (see Sub-
heading 3.6) or Lowry (see Subheading 3.8) method. The
percentage of soluble proteins can be expressed as ratio
between the protein content determined in the ultracentrifu-
gation supernatant and the total protein content of the sample
prior to ultracentrifugation.

3.8 Lowry Lowry is a colorimetric method that can be used to quantify

OMV/GMMA total protein content.
1. Set the spectrophotometer wavelength at 750 nm.
2. In 5 mL tubes, prepare (in duplicate) diluted solutions from
the 120 μg/mL BSA standard solution, using the volumes
reported in Table 4.
3. For each sample prepare three replicates for the analysis; the
volume of sample, or diluted sample, required for each replicate
is 400 μL (see Notes 14 and 19).
4. Add 400 μL of Reagent A to each tube containing standards/
samples and mix well with the vortex. Add 200 μL of Folin
reagent mixture to each tube containing standards/samples
 OMV/GMMA Analytics 249

Table 4
Dilutions for preparing calibration curve for Lowry analysis

Std conc. 120 μg/mL BSA PBS

μg/mL μL μL
0 0 400
9.0 30.0 370.0
21.0 70.0 330.0
42.0 140.0 260.0
60.0 200.0 200.0

and 400 μL of Reagent A. Vortex each tube. Cover the tubes

with aluminum foil or Parafilm, incubate them at RT and start a
75 min countdown.
5. Read samples/standards following the sequence: First read the
“blank” containing cuvette, if a single beam instrument is used
(in order to subtract its absorbance from samples/standards
absorbances). If a double beam instrument is used, leave the
“blank” containing cuvette in the reference slot during all
sample/standard readings. Next, read the BSA standard solu-
tions (each in duplicate) in order of increasing concentration,
and finally the samples (each in triplicate). For each tube,
transfer its content in a quartz cuvette, place it in the spectro-
photometer, read the ABS value and discard the cuvette con-
tent. Wash the cuvette with PBS after calibration curve readings
(before starting the first sample reading) and between readings
of different samples.
6. Calculate a linear regression between the ABS and BSA con-
centrations (in the range 9–60 μg/mL) and calculate on it the
concentration of the sample analyte. If the concentration of a
sample is outside the range of the calibration curve, dilute the
sample with PBS and repeat the analysis.

3.9 Amino Acid Amino acid analysis is another method to quantify total protein
Analysis content of OMV/GMMA samples. In contrast to the Micro BCA
and Lowry described above, Amino Acid Analysis quantifies actual
content of each amino acid present in the sample, instead of assum-
ing similar behavior of OMV/GMMA sample as a BSA standard in
the colorimetric assays.
MS analysis can also be used to quantify specific proteins in
OMV/GMMA that are relevant for vaccine efficacy. Such methods
need to be set up and optimized specifically for the protein of
interest. Examples can be found in [10, 18]. Before amino acid
content determination, the time required to hydrolyze the sample
has to be defined with a hydrolysis kinetic study (i.e., hydrolysis for
250 Francesca Micoli et al.

16, 24, 48, 72 h) in order to optimize conditions to achieve best

amino acids recovery for the specific sample being analyzed.
1. For the HCl vapor phase hydrolysis (see Notes 20 and 21),
before starting operations, treat glass hydrolysis tubes at
400
C for 4 h, to avoid sample contamination (i.e., Gly, Ser
Pro from skin, Asp, Glu from paper).
2. Switch on the Hydrosmart 2 at 114
C in order to preheat the
oven prior to inserting hydrolysis vessels.
3. For each sample, prepare hydrolysis tubes in triplicate by
weight. Place the hydrolysis tube on the analytical balance,
tare and add the sample to hydrolyze (see Notes 20–22).
4. Dry the hydrolysis tubes filled with sample solution in the
centrifugal evaporator.
5. With the help of forceps, put the dried hydrolysis tubes inside
the hydrolysis vessels (up to 12 hydrolysis tubes for each hydro-
lysis vessel; the oven can contain up to 3 hydrolysis vessels for a
total of 36 tubes corresponding to 12 samples in triplicate).
6. Add at the bottom of the vessel (outside the hydrolysis tubes)
200 μL of 6 N HCl, 0.1% Phenol solution.
7. Cap the vessel, perform 4–5 vacuum–nitrogen cycles and close
the cap valve leaving the vessels under nitrogen.
8. Place the vessels in their oven compartments, cover them with
oven’s cap and start the hydrolysis.
9. After the hydrolysis time is complete, extract carefully the
vessels from the oven and let them cool for about 10 min
under chemical hood, maintaining the samples under inert
atmosphere.
10. Under a chemical hood, release the cap valve, unscrew the cap
and take out with forceps hydrolysis tubes from the vessels.
11. To remove the eventual HCl residue drops, place sample
hydrolysis tubes in a Speedvac concentrator for about 30 min.
12. Resuspend each sample in 100 μL 100 mM HCl and vortex
well, as partial redissolution of the hydrolyzed sample is one of
the main causes of low recovery (see Note 22).
13. Transfer the entire content of the hydrolysis tube in low-bind
Eppendorf tube.
14. Prepare, in the low-bind tubes, two replicates of each dilution
of the Standard Curve Solution starting from the 2.5 mM
(2500 nmol/mL) standard solution of each amino acid (except
for Cystine that has a concentration of 1.25 mM) as reported in
Table 5, vortex all the standard dilutions tubes (see Note 23).
15. To derivatize the samples and standards directly in HPLC vials,
transfer 70 μL of borate buffer AccQTag Ultra in each vial.
 OMV/GMMA Analytics 251

Table 5
Dilutions for preparing curve for amino acidic analysis

Amino acid 250 nmol/mL amino acid standard Water 2500 nmol/mL amino acid standard
nmol/mL solution (μL) (μL) solution (μL)
0 0 500 0
15 30 470 0
25 50 450 0
62.5 125 375 0
125 250 250 0
187.5 375 125 0
250 0 1800 200

Then to appropriate vials, add 10 μL of sample solution from

step 13 and each standard solution by inserting the tip in the
buffer and pipetting 3–4 times. Vortex each vial for 10 s. Add
20 μL of reconstituted Reagent AccQTag Ultra (see Note 24)
to each vial, releasing the reagent from the pipette by inserting
the tip in the sample buffered solution and pipetting 3–4 times.
Place all the vials at 55
C for 10 min (see Note 25).
16. To run the amino acid analysis by HPLC-RP, set the HPLC
instrument configuration as follows: column compart-
ment ¼ 49
C, autosampler compartment ¼ 20
C, UV detec-
tor acquisition at 260 nm ABS, Flow rate ¼ 0.7 mL/min, Run
time ¼ 10.2 min, Sample injection volume ¼ 1 μL, Gradient
Program (Fig. 8) (see Note 26).
A representative chromatogram is reported in Fig. 9.
17. Calculate a linear regression for each amino acid standard (peak
area vs concentration in the range 15–250 nmol/mL) and
calculate on it the concentration of each amino acid in the
sample analyte. For each standard/sample injection, check if
the chromatogram 6-aminoquinolone (AMQ) area is higher
with respect to the 80% of the average AMQ in the blanks. This
check is performed to ensure that the derivatizing agent quan-
tity is enough for the reaction (at least 5 times the amino
groups). If the test fails, the corresponding standard/sample
is not suitable for the analysis (see Note 27).
Calculate the total protein of the sample by sum of the
weight of each quantified amino acid, considering the conden-
sation water using the following formula:
X
Total protein ¼ n¼Amino acid
ðamino acidn molecular weight  18Þ
 amino acidn concentrationðnmol=mLÞ
252 Francesca Micoli et al.

Fig. 8 Gradient program for amino acid analysis by HPLC-RP. A -AccQTag Ultra
Eluent A; B -AccQTag Ultra Eluent B/10; C -Water (Filtered 0.22 μm); D -AccQTag
Ultra Eluent B

Lys - 6.908
A

Val - 7.229
Tyr - 6.991

Leu - 7.919
Phe - 8.027
lle - 7.845
Met - 7.115
Cys - 6.836
Drv - 6.680
Pro - 5.778
Ala - 5.141
Thr - 4.735
Glu - 4.339
NH3 - 1.764

Arg - 3.222
Gly - 3.399
Ser - 3.076
His - 2.141

3.798

Brv 6.675

B
Lys - 6.907
NH3 - 1.763

Asp - 3.801

Leu - 7.928
Ala - 5.144

Val - 7.234

Phe - 8.035
Glu - 4.339
Gly - 3.403

Thr - 4.739

Tyr - 6.991

lle - 7.853
Pro - 5.774

Met - 7.119
Ser - 3.077

3.559
Arg - 3.217
His - 2.136

7.615
7.759
7.412
6.155
5.403
5.050
2.531

1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00 6.50 7.00 7.50 8.00
Minutes

Fig. 9 Chromatograms (260 nm profiles) of standard amino acids mixture (a) and of a purified S. Typhimurium
GMMA sample after hydrolysis and derivatization for amino acid analysis (b)

3.10 Total Sugar LPS is one of the main constituents of OMV (naturally released)/
Quantification GMMA membranes (Fig. 1). LPS molecules are constituted of
three main portions: Lipid A is linked to the 3-deoxy-D-manno-
octulosonic acid (KDO) terminus of the core region, which in turn
is attached to the OAg chain (Fig. 10). Capsular polysaccharides
can also be present on OMV/GMMA surface depending on the
 OMV/GMMA Analytics 253

Fig. 10 Schematic representation of LPS molecules, constituted by lipid A, core region and repeating units of
the OAg chain

specific type of OMV/GMMA [19]. Such sugars can be key target

for protection as it is the case of Salmonella [20] and Shigella [21],
and in-depth analysis of LPS/capsular polysaccharides can be
needed.
According to the specific sugars that constitute the LPS/cap-
sular polysaccharide molecules present, different analytical methods
can be used. Here we report details on some methods that can be
performed at this scope, starting from methods for total sugar
determination. Such methods can be performed directly on
OMV/GMMA particles avoiding any step of sugar chains isolation
before measurement.

3.11 HPAEC-PAD This method allows quantification of neutral sugars, that are usually
(Neutral Sugars) [14] found as constituents of LPS core regions or OAg chains (such as in
S. Typhimurium, S. Enteritidis, Salmonella Paratyphi A, Shigella
flexneri). Concentrations of Rha, Gal, Glc, Man can be estimated
through this analysis.
1. In 2 mL screw cap vials, prepare in duplicate the dilutions of
calibration curve standard solutions, starting from the 11.25
μg/mL neutral sugar standard mix, as indicated in Table 6.
2. For triplicate analysis, prepare three 2 mL screw cap vials con-
taining 450 μL of the diluted sample.
3. To each vial containing standards or samples, add 150 μL 8 M
TFA (final concentration 2 M), close and vortex each vial for
few seconds.
4. Place all vials in a rigid cardboard rack and incubate in a pre-
heated oven at 100
C for 4 h.
5. After this time, allow all the vials to cool at 2–8
C for 30 min.
6. Remove the caps and dry the standards and samples overnight
in centrifugal evaporator at RT in order to remove the TFA.
254 Francesca Micoli et al.

Table 6
Dilutions for preparing calibration curve for HPAEC-PAD neutral sugars analysis

μg/mL 11.25 μg/mL Neutral sugar standard solution Water

(each sugar) μL μL
0 0 450
0.5 20 430
1.0 40 410
2.5 100 350
5.0 200 250
7.5 300 150
10 400 50

7. Redissolve the content of each vial adding 450 μL of water and

vortex for few seconds.
8. Filter the content of each vial into the sampler polypropylene
vial using 0.45 μm nylon syringe filters (4 mm diameter).
9. Set up the instrument with CarboPac PA10 guard column,
CarboPac PA10 column connected in series, Column/detector
compartment set at 35
C, Autosampler compartment set at
10
C, Electrochemical detector equipped with gold working
electrode and potential sets with standard carbohydrate
4-potential waveform.
10. Run the analysis of standards and samples with an injection
volume of 25 μL in full loop mode with 50 min run time and
1 mL/min eluent flow rate using the following eluent
program:
20 min 18 mM NaOH (36% of eluent 50 mM NaOH),
10 min 28 mM NaOH 100 mM AcONa (36% 50 mM NaOH;
10% 1 M AcONa with 100 mM NaOH), 20 min 18 mM
NaOH (36% of eluent 50 mM NaOH) (see Note 28).
11. For each sugar standard, calculate a linear regression between
the peak areas and monomer concentrations (in the range
0.5–10 μg/mL) and calculate on it the concentration of each
sugar monomer in the sample (see Fig. 11). If the concentration
of a sample is outside the range of the calibration curve, dilute
the sample with water and repeat the analysis.

3.12 HPAEC-PAD This method allows quantification of amino uronic acids, present
(Amino Uronic Acid) for example in Shigella sonnei OAg and capsular polysaccharide
[13] displayed on OMV/GMMA surface.
1. In 2 mL screw cap vials, prepare in duplicate the dilutions of
calibration curve standard solutions, starting from the 2.56
 OMV/GMMA Analytics 255

Fig. 11 HPAEC-PAD profile of standard neutral sugar mixture (a) and of a purified S. Typhimurium GMMA
sample (b)

μg/mL and 0.512 μg/mL OAg standard solutions, as indi-

cated in Table 7 (see Note 29) .
2. For triplicate analysis, prepare three 2 mL screw cap vials con-
taining 300 μL of the diluted sample.
3. To each vial containing standard or sample, add 1 mL of
TFA-HCl Mixture, close and vortex each vial for few seconds.
4. Place all vials in a preheated block heater at 80
C for 4.5 h.
5. After this time, allow all the vials to cool at 2–8
C for 15 min.
6. Remove the caps and dry the standards and samples under
nitrogen flux in order to remove the major part of solvent/
HCl.
7. Finish drying the standards and samples overnight in centrifu-
gal evaporator at RT.
8. Redissolve the content of each vial adding 300 μL of water and
vortex accurately for few seconds.
9. Filter the content of each vial into the sampler polypropylene
vial using 0.45 μm nylon syringe filters (4 mm diameter).
10. Set up the instrument with CarboPac PA1 guard column,
CarboPac PA1 column connected in series, Column/detector
compartment set at 25
C, Autosampler compartment set at
10
C, Electrochemical detector equipped with gold working
electrode and potential sets with standard carbohydrate
4-potential waveform.
256 Francesca Micoli et al.

Table 7
Dilutions for preparing calibration curve for HPAEC-PAD amino uronic acid analysis

S. sonnei OAg 2.56 μg/mL 0.512 μg/mL Water

μg/mL standard solution (μL) standard solution (μL) (μL)
0 – – 300
0.160 – 94 206
0.321 – 188 112
0.649 76 – 224
1.280 150 – 150
2.560 300 – 0

11. Run the analysis of samples and standards with an injection

volume of 25 μL in full loop mode with 15 min run time and
1.5 mL/min eluent flow rate using 400 mM NaOH (see
Note 30).
12. Calculate a linear regression between the areas of the peak of
interest and the concentrations of the OAg standard (in the
range 0.16–2.56 μg/mL) and calculate on it the concentration
of the OAg in the sample (see Fig. 12, analyte peak at 8.9 min).
If the concentration of a sample is outside the range of the
calibration curve, dilute the sample with water and repeat the
analysis.

3.13 Dische Dische is a colorimetric method, that allows quantification of

Colorimetric Method methyl pentose (6-deoxyhexose) content in samples (i.e., rhamnose
and fucose) and can be often used as a simple and rapid method for
sugar quantification in OMV/GMMA samples.
This method is based on the following steps: in the first step
concentrated sulfuric acid at 100
C hydrolyzes the polysaccharide
and allows the monosaccharides to form methyl furfural (from
methyl pentoses) or hydroxymethyl furfural (from hexoses). Both
species react with sulfur present in cysteine. Compounds that derive
from methyl furfural have a characteristic maximum in absorption
spectrum at 396 nm, while compounds that derive from hydroxy-
methyl furfural have a maximum at 415 nm. In the single com-
pound absorption spectra, considering the bell shape close to the
maximum, the absorption at 396 nm due to chromophores deriv-
ing from hexoses is numerically equal to their absorption at
427 nm; the absorption at 427 nm due to chromophores deriving
from methyl pentoses is close to zero. Consequently, performing
the whole reaction on a mixture of hexoses and methyl pentoses, it
is possible to calculate the methyl pentose contribution to 396 nm
ABS by subtracting the 427 nm ABS.
 OMV/GMMA Analytics 257

Fig. 12 HPAEC-PAD profile of S. sonnei standard OAg (a) and of a S. sonnei GMMA sample (b) after acid
hydrolysis resulting in the formation of 2-amino-2-deoxy-α-L-altropyranuronic acid quantified in the analysis

The chromophore species is the same regardless the stereo-

chemistry of the methylpentose sugars, that is, with a calibration
curve of fucose standard it is possible to perform an accurate
determination of sample containing rhamnose.
The procedure has been tested for the following interferences
in methylpentose determination: hexoses (mannose, galactose, glu-
cose), 2-aminohexoses: N-acetylglucosamine, uronic acid (glu-
curonic acid), 3,6-dideoxy hexoses (tyvelose), KDO (3-Deoxy-D-
manno-oct-2-ulosonic acid), proteins (BSA up to 100 μg/mL with
Rha concentration at 5 μg/mL), and DNA.
All substances above have no interference with methyl pentose
determination, provided the ABS values of the sample at 396 and
427 nm remain under 1.2. (i.e., Rha can be quantified at 5 μg/mL
in presence of 500 μg/mL glucose but Rha at 30 μg/mL in
presence of 500 μg/mL Glc results in an ABS that exceeds this
limit and needs a predilution).
1. Prepare a single set of the diluted solutions for a calibration
curve from a fucose standard solution (50 μg/mL) directly in
2 mL Eppendorf tubes, following volumes reported in Table 8.
2. Gently mix samples before dilution in order to homogenize the
vial content. Samples are analyzed in duplicate. The (diluted)
sample quantity needed for each replicate is 500 μL (see
Note 31).
258 Francesca Micoli et al.

Table 8
Dilutions for preparing calibration curve for Dische determination

Fucose 50 μg/mL fucose Water

μg/mL μL μL
2 20 480
4 40 460
8 80 420
16 160 340
32 320 180

3. To each standard and sample (500 μL), add 1050 μL of ice

cooled sulfuric acid.
4. Vortex each tube.
5. Place the tubes in a thermoblock and heat for 5 min at 100
C.
6. Cool the tubes in ice for 10 min.
7. Use an empty cuvette to set the spectrophotometer ABS to
zero reading at 427 nm and 396 nm.
8. For each standard and sample, transfer 1 mL from the cooled
tube into a cuvette and read the ABS 427 nm and 396 nm
(to be used as “pre cys” values for the calculations).
9. Without removing the contents from the cuvettes, add then
32 μL of cysteine 1 M into each cuvette.
10. Close each cuvette with a cap and mix well by vortexing.
11. Wait 10 min at RT.
12. Read the ABS 427 nm and 396 nm of each cuvette (to be used
as “post cys” values for the calculations). (see Notes 32 and
33).
13. For each standard and sample, calculate ΔABS using the fol-
lowing formula:
   
Post Cys Pre Cys Post Cys Pre Cys
ABS ¼ ABS396  ABS396  ABS427  ABS427

Calculate a linear regression between the ΔABS and fucose

standard concentrations (in the range 2–32 μg/mL) and cal-
culate on it the concentration of methyl pentoses in the sample.
If the concentration of a sample is outside the range of the
calibration curve, dilute the sample with water and repeat the
analysis.

3.14 HPLC-MS The lipid A present in the sample is quantified through the
3-hydroxymyristic or 3-hydroxylauric acids present as primary
esters in its structure (in Fig. 13 the lipid A structures of S. sonnei
 OMV/GMMA Analytics 259

Fig. 13 Lipid A structures of S. sonnei (a) and N. meningitidis (b) reported as examples with the 3-hydroxy fatty
acids (in red) quantified by HPLC-MS after the hydrolysis of ester bonds

(A) and Neisseria meningitidis (B) are represented as examples with

the 3-hydroxy fatty acids quantified in this assay after the ester
bonds hydrolysis evidenced in red). The procedure consists in an
initial hydrolysis of the sample to release the 3-hydroxy-fatty acids,
cleaving quantitatively the ester bonds from the glucosamine O-3
present in the lipid A.
The 3-hydroxy-fatty acid is then separated by RP-HPLC and
quantified by MS detector, that is, triple quadrupole by selected
reaction monitoring (SRM) or Orbitrap mass spectrometer by
parallel reaction monitoring (PRM). Both samples/standards are
assayed using an on-line SPE hydrophobic cartridge that allows to
concentrate and desalt the analytes. The analysis, depending on the
hydroxy-fatty acid type present as ester in the lipid A structure of
OMV/GMMA, is performed to quantify only 3-hydroxymyristic
acid (3OH-My) or 3-hydroxylauric acid (3OH-La) (hydroxy-fatty
acids present as amide are not released in the hydrolysis
conditions used).
An alternative method for quantification is reported in litera-
ture by Lyngby et al. [22], while Limulus amebocyte lysate (LAL)
based assays are not deemed appropriate for this type of sample (see
Note 34).
The parameters reported in the following paragraphs are for a
system equipped with Thermo Quantum Access triple quadrupole.
1. Perform a M/z accuracy check in the region used (see Note 35).
2. Set up the MS instrument with the parameters as follows: Mass
polarity ¼ negative, ESI gas nitrogen, ESI spray voltage ¼ 4000,
260 Francesca Micoli et al.

Table 9
Summary of the method time events

Flow Eluent A Eluent B Divert Valve

Time mL/min % % position Comment
0.00 1.00 100 0 Load Injection, SPE wash
1.00 1.00 100 0 Load End SPE wash
1.01 0.50 0 100 Load Start dead volume compensation
1.75 0.50 0 100 Load End dead volume compensation
1.76 0.25 0 100 Inject Injection in column/elution
7.00 0.25 0 100 Inject End elution
7.01 1.00 100 0 Load SPE reequilibration
8.00 1.00 100 0 Load SPE reequilibration

ESI Sheath gas pressure ¼ 25, ESI Aux gas pressure ¼ 10,
Capillary temperature ¼ 300
C, Tube lens offset ¼ 70,
Skimmer Offset ¼ 10, q cell collision gas Argon ¼ 1 mTorr,
Collision energy ¼ 16 V, Data type ¼ centroid.
3. Set up the MS scan event (time segment 0–10 min) with the
following parameters: Q1 with 215 (for 3OH-La or 243 for
3OH-My)  0.02 M/z as precursor, Q3 with 59  0.02 M/z as
product using scan time 0.5 s (see Note 36).
4. Program the time events of the HPLC instrumental method as
reported in Table 9.
5. Set up the instrument for the analysis connecting MS Divert
valve, SPE cartridge, RP-column and tubes as shown in Fig. 14
closing valve position 4 with a screw cap fitting. Position the
ESI cone as shown in Fig. 15 (B level).
6. To calculate the dead volume (needed only for first time setup;
see Note 37) assemble the HPLC-QqQ equipment as reported
in Fig. 14 connecting the ESI-MS directly to position 5 without
the C8 column. Program the HPLC pump, divert valve and
QqQ spectrometer as reported in Table 10. Program the MS
spectrometer to measure 3OH-La or 3OH-My according to
the specific sample to be analyzed. Inject 1 μL of a standard
dilution (no subjected to hydrolysis treatment) and collect the
chromatogram. In Fig. 16 examples of chromatograms
acquired by this way are reported.
Consider the time needed for starting the leading of the
first peak (i.e., 2.5 min in the Fig. 16) and calculate the eluent B
volume pumped at that time with its flow rate (in the example:
2.5–1 ¼ 1.5 min; 1.5 min * 0.25 mL/min ¼ 375 μL). This is
the dead volume of eluent B to be pumped after SPE wash step
with eluent A before commute the divert valve. In terms of
 OMV/GMMA Analytics 261

Fig. 14 SPE, RP-column and tubes connections in the two different valve
positions

Fig. 15 Detail of ESI cone setup

time, it has to be divided by the flow rate 0.5 mL/min of the

dead volume compensation step (0.75 min in the example).
Adjust the time of the dead volume compensation step
reported in Table 9 according to this value.

Lasting ð min Þ of dead volume compensation step at flow rate of 0:5 mL= min ¼
ðfirst peak leading time ð min Þ  1Þ  0:25
¼
0:5
262 Francesca Micoli et al.

Table 10
Instrument setting for dead volume calculation

Flow Eluent A Eluent B Divert Valve

Time mL/min % % position ESI-QqQ settings
0.00 1.00 100 0 Load –
1.00 1.00 100 0 Load –
1.01 0.25 0 100 Inject 3OH-La/3OH-my
7.00 0.25 0 100 Inject –
7.01 1.00 100 0 Load –
8.00 1.00 100 0 Load –

7. Perform a system equilibration setting the divert valve in inject

position and fluxing the SPE cartridge and the column with
20 column volumes (about 3.2 mL) of eluent B. Then set the
divert valve in load position and flux the SPE cartridge with
4 mL of eluent A.
8. Using low-bind tubes, prepare duplicate samples for a standard
curve, starting from a vial containing 3-hydroxy-fatty acid stan-
dard (5 nmol/mL), following the scheme reported in
Table 11.
9. For each sample to be analyzed, gently mix the sample in order
to homogenize the content (see Note 38).
10. Using low-bind tubes, dilute the sample with water, in three
independent replicates up to half of the desired dilution factor
(i.e., if final dilution factor needed is 6, dilute only with dilu-
tion factor 3).
11. In low-bind tubes, transfer 200 μL of the previous dilution
replicates and add 200 μL of IPA (dilution factor 2), homoge-
nize the samples.
12. Transfer 200 μL of the previous replicates (in 50% IPA) in
low-bind tubes and proceed with the following treatments.
13. To each tube containing standards and samples, add 40 μL of
1.5 M NaOH, close it, and mix by vortex.
14. Warm the tubes at 40
C in a preequilibrated water bath or
oven for 2 h.
15. Chill the tubes at 4
C for 15 min.
16. Transfer the standards and samples in the HPLC low-bind
vials.
17. Analyze standards and samples with 1 μL injections creating a
sample list with the following order: six injections of the
5 nmol/mL standard (to be used just as system check; usually
 OMV/GMMA Analytics 263

Fig. 16 Examples of chromatograms (HPLC-MS for lipid A quantification) for

dead volume calculation

Table 11
Dilutions for preparing calibration curve for HPLC-MS lipid A quantification analysis

5 nmol/mL STD 50% IPA

nmol/mL μL μL
0.5 20 180
1.0 40 160
1.5 60 140
2.5 100 100
5.0 200 0
264 Francesca Micoli et al.

the area of the first injection is lower than the other replicates),
one injection for each calibration standard in order of increas-
ing concentration, one injection for each sample replicate, one
injection for each calibration standard in order of increasing
concentration. At the end of sample and standard list runs,
after the last chromatographic analysis is completed, store the
column with 10 column volumes (1.6 mL) of eluent E. After
several analyses, Column Cleaning may be needed (see Note
39). In Fig. 17 are reported chromatograms related to
3OH-La (RT 4.71 min) and 3OH-My (RT 5.57 min).

Fig. 17 Examples of chromatograms related to 3OH-La (RT 4.71 min) (a) and 3OH-My (RT 5.57 min) (b)
quantification by HPLC-MS.
 OMV/GMMA Analytics 265

18. Calculate a linear regression for the 3-hydroxy-fatty acid stan-

dard between peak areas and concentrations and calculate on it
the concentration of the sample analyte. Considering the struc-
ture of lipid A (that contains 2x 3OH-FA esters per molecule),
lipid A amount is the half of 3OH-FA molar amount measured
in the analysis. The result is expressed as nmol/mL of lipid A.

3.15 MALDI-TOF MS 1. Dilute the OMV/GMMA sample with PBS to a final concen-
tration close to 1 mg/mL of protein.
2. In a Wheaton 2 mL screw cap vial, add 200 μL of sample and 50
μL of acetic acid 5% solution.
3. Close the container and keep it at 100
C in a preheated oven
for 2 h (see Note 40).
4. Cool the vial, vortex it, and transfer the whole content in a
2 mL Eppendorf tube.
5. Centrifuge at 14,000  g, 10
C, for 15 min and discard the
supernatant.
6. Add 1.5 mL of water to the pellet and resuspend it.
7. Centrifuge at 14,000  g, 10
C, for 15 min, discard the
supernatant.
8. Add 1.5 mL of water to the pellet and resuspend it.
9. Centrifuge at 14,000  g, 10
C, for 15 min, discard the
supernatant.
10. Dry the pellet overnight in a centrifugal evaporator.
11. Just before deposition on MALDI plate, dissolve the whole
dried pellet in 100 μL of 4:1 chloroform–MeOH by vortexing.
12. For each standard and sample to be assayed, pipet 2 μL into a
0.5 mL tube, add 2 μL of Super-DHB solution, and mix them
carefully pipetting.
13. Withdraw 2 μL of the solution and load it on the target plate.
14. Let the spot dry at RT, possibly with gentle air flux (i.e., plate
left under chemical hood near the front border).
15. To acquire MS Spectra, set up the spectrometer in order to
work in negative, reflectron mode and acquire MS in the range
1000–3000 Da.
16. Calibrate the spectrometer by shooting with laser at the pep-
tide standard spot coordinate.
17. For each sample, record the mass of the sample by shooting
with the laser at the sample spot coordinate (see Note 41).
Shoot with a laser intensity suitable for obtaining resolution of
isotopic cluster peaks (usually less than 50% of its maximum),
in different positions within the sample spot, summing the
result spectrum until clear peaks appear over the baseline
noise (Fig. 18) (see Note 42).
266 Francesca Micoli et al.

A 1585.592 B 1585.499
1000 1250
Penta- Penta-acylated
800 acylated 1000

Intensity a.u.
Intensity a.u.

600 Penta-
Penta- 750
acylated+P
acylated+P
1505.448
400 500
1665.542

200 250
1279.175
0 0
1000 1200 1400 1600 1800 1200 1400 1600 1800
m/z m/z

Fig. 18 MALDI-MS spectra of pentaacylated lipid A of S. Enteritidis GMMA isolated by mild hydrolysis in
acetate buffer pH 4.5 with 3% N-Octyl-β-D-glucopyranoside preserving the lipid A structure (a), compared to
acetic acid hydrolysis (b)

3.16 cELISA [15] The cELISA working principle is based on the competition
between the coating antigen and the specific antigen to quantify
in OMV/GMMA samples for the antigen-specific primary anti-
body. The more antigen is present on OMV/GMMA, the less
primary antibody can bind to the coating antigen, and the less
signal can be detected by ELISA. The antigen of interest on
OMV/GMMA is quantified by comparing the ELISA signal
obtained with a standard curve. The standard curve is built by
spiking the primary antibody with known amounts of
OMV/GMMA displaying the antigen of interest.
MS analysis can also be used for quantifying specific protein
antigens, as well as to determine entire OMV/GMMA protein
composition [10].
1. Prepare 10 sequential dilutions (2- or 3-fold), named as Std01-
Std10, of a freshly prepared OMV/GMMA sample solution,
starting from a concentration previously set, based on the
specific OMV/GMMA sample and antibodies used. To prepare
350 μL of each standard (sufficient for 50 μL each, plated in
duplicate on each of three replicate plates), begin by weighing
the appropriate volume of undiluted standard (Std01) in a tube
and add the appropriate volume of dilution buffer. To prepare
Std02 to Std10, weigh in the corresponding tubes the appro-
priate volume of previous standard and add the appropriate
volume of dilution buffer.
2. Transfer 350 μL of each standard dilution point into successive
wells in a row of a 1 mL deep-well plate. Use the diluted
standards within 1 day.
3. Add 350 μL dilution buffer to each of two blank wells in the
same row as the standards.
4. For positive controls, prepare three test dilutions of the sample
used to prepare the standards, each a total of 0.8 mL in SDB in
2 mL tubes.
 OMV/GMMA Analytics 267

5. For each sample to be tested, prepare three test dilutions, each

a total of 0.8 mL in SDB in 2 mL tubes.
6. Prepare 32 mL of coating solution by diluting the antigen to be
used for coating the plate. Concentration depends on the
specific antigen used, usually in the range 1–10 μg/mL.
7. Dispense 100 μL of coating solution in each well of 3 Nunc
Maxisorp ELISA plates.
8. Seal the plates and incubate overnight at 4
C.
9. For each of the three ELISA plates, aspirate the coating solu-
tion using a plate washer and add 200 μL of blocking buffer in
each well. Wait 3 min between each of the three plates.
10. Incubate the plates at 25
C for 1 h.
11. Before the end of the blocking step, prepare a 2 concentration
primary antibody solution in SDB (dilution depends on the
specific antibody used) which will be spiked with either the
standard curve or the positive control or sample(s).
12. Add 350 μL of the 2 concentration of primary antibody
solution to each standard dilution point into the 1 mL deep-
well plate.
13. Pipette four/five times to mix the suspension.
14. Add to each of the three dilutions of the positive control, and
each of the three dilutions of sample(s) to be tested, 0.8 mL of
the 2 concentration of primary antibody solution and vortex
the 2 mL tubes.
15. For each positive control and sample tube, dispense 400 μL/
well into each of 4 wells of the deep-well plate containing the
standards.
16. Incubate the deep-well plate at 4
C in a Mixmate plate shaker.
Minimum incubation time depends on the specific antigen
being measured. If an overnight incubation is needed, prepare
the standard, the positive control and sample dilutions the day
before running the assay.
17. At the end of the blocking step (i.e., step 10), wash the ELISA
plates.
18. Transfer 100 μL of each dilution point of the standard curve
and the positive control and sample(s) from the deep-well plate
to each of the three the ELISA plates. Maintain the same order
in processing the three ELISA plates and allow an interval of
execution time of at least 1.5 min.
19. Incubate the plates at 25
C for 2 h.
20. Remove primary antibody solution and wash the ELISA plates
with washing buffer solution.
268 Francesca Micoli et al.

21. Add 100 μL of secondary antibody solution to each well fol-

lowing the correct timing.
22. Incubate the plates at 25
C for 1 h.
23. Approximately 10 min before the secondary antibody step
ends, prepare the substrate solution by dissolving 2 tablets of
Tris and p-nitrophenyl phosphate (pNPP) in 40 mL of water
(substrate can be different according to the secondary
antibody used).
24. As soon as the incubation of the secondary antibody times up,
wash the plates and add 100 μL of substrate solution in each
well, maintaining the timing between the three plates.
25. Incubate plates at 25
C for 1 h.
26. Read the absorbances at 405 and 490 nm and calculate the
difference between them (OD405nm–OD490nm).
27. Interpolate the standard curve using a four-parameter nonlin-
ear (4PL) regression (see Note 43); the OD values on which
the 4PL curve is linear represent the quantification range of the
standard curve. For each plate, check if the positive control
sample is within 30% of the expected value, otherwise exclude
the entire plate data from the analysis. Consider the assay valid
if at least two of the three plates are valid.
28. Calculate the amount of test samples by the average of all the
wells (four replicates at each of the three dilutions tested and
from the three different plates) with OD within the acceptable
range of the curve.

3.17 OAg Extraction OMV/GMMA treatment at low pH and high temperature results
in cleavage of the labile linkage between KDO, at the proximal end
of the core oligosaccharide, and lipid A, releasing the OAg-core
chains or core alone in the supernatant (Fig. 10) and causing
OMV/GMMA proteins and lipids to precipitate. The extracted
sugars reflect the composition of LPS molecules on
OMV/GMMA. They can be constituted by OAg chains attached
to the core region (from LPS) and/or core region only (from
lipooligosaccharides). OAg-core (simply indicated as OAg from
here after) and core chains can be easily isolated (e.g., by size
exclusion chromatography) and characterized in depth. Hydrolysis
can be performed with acetate buffer at pH 3.9, or in 1% acetic acid,
based on the OAg stability. The advantage to use 1% acetic acid is
that the supernatant can be directly dried for further analysis with-
out any desalting step (that can be eventually performed by the use
a Cytiva PD10).
1. For polysaccharide extraction with acetate buffer, in a screw cap
vial, add 0.9 mL of OMV/GMMA sample and 100 μL of 10
acetate buffer.
 OMV/GMMA Analytics 269

2. Keep the vial in an oven preheated at 100
C for 100 min.

3. Transfer the whole vial content in an Eppendorf tube and
centrifuge at 14,000  g for 10 min.
4. Recover the supernatant that contains extracted polysacchar-
ides (i.e., OAg, core).
5. For polysaccharide extraction with 1% acetic acid, dilute the
OMV/GMMA sample with water to an appropriate
concentration.
6. In a screw cap vial add 1 mL of diluted OMV/GMMA sample
and 10 μL of acetic acid.
7. Keep the vial in an oven preheated at 100
C for 100 min.
8. Transfer the whole vial content in an Eppendorf tube and
centrifuge at 14,000  g for 10 min.
9. Recover the supernatant that contains extracted polysacchar-
ides (i.e., OAg, core).

3.18 HPLC-SEC SCA OAg/core samples are analyzed by HPLC-SEC. A refractive index
detector can be used to estimate apparent molecular size using
dextrans (in the range 5–150 kDa) to run a calibration curve and
GPC software. The samples are analyzed after derivatization with
Semicarbazide to quantify the KDO present at the reducing end of
OAg/core chains. This reaction is performed as a slight modifica-
tion of the Semicarbazide assay for α-ketoacid determination [23]
and allows calculation of molar concentration of the chains and
quantification of molar ratio of populations at different size when
present.
1. Using water, prepare in tubes, the dilutions of KDO ammo-
nium salt 40 μg/mL standard solution in duplicate
(as indicated in Table 12). Vortex all standard dilution tubes
for few seconds.
2. Mix the OMV/GMMA sample in order to homogenize the
content.

Table 12
Dilutions for preparing calibration curve for HPLC-SCA analysis

KDO (nmol/mL) 40 μg/mL KDONH4 (μL) Water (μL)

15.7 10 90
31.4 20 80
54.9 35 65
78.4 50 50
156.8 100 0
270 Francesca Micoli et al.

3. Transfer 100 μL of the sample into each of four different

Eppendorf tubes (two to be derivatized with SCA and two to
be used as blank replicates).
4. To each blank sample tube, add 100 μL of water. Vortex all
tubes for few seconds.
5. To each standard tube and remaining sample tubes add 100 μL
of Semicarbazide solution. Vortex all tubes for few seconds.
6. Keep all the tubes in a 50
C preheated bath for 50 min.
7. Chill the tubes in a 2–8
C fridge for 15 min.
8. Vortex all tubes for few seconds.
9. Transfer samples, blank samples and standards in HPLC vials.
10. Set up the instrument with G3000 PWXL guard columns
connected in series, column compartment temperature set at
30
C, autosampler compartment set at 4
C and UV detector
acquisition channel at 252 nm ABS.
11. Run the sample/standard analysis with 80 μL injection vol-
ume, 0.5 mL/min flow rate, 35 min run time, using isocratic
conditions with eluent 0.1 M NaCl, 0.1 M NaH2PO4, 5%
ACN, pH 7.2 (see Note 44). After the last chromatographic
analysis is completed, store the system in NaN3 0.02% preser-
vative solution.
12. To quantify the KDO (nmol/mL) in the samples, correct the
peak area corresponding to OAg/core in samples derivatized
with Semicarbazide by subtracting the area of the
corresponding blank. Quantify the amount of KDO using the
calibration curve built with the peak areas of derivatized KDO
standard at 252 nm (see Fig. 19). If the concentration of a
sample is outside the range of the calibration curve, dilute the
sample with water and repeat the analysis.
1
3.19 NMR H NMR analysis is performed to confirm the identity of the
OAg/core samples by detecting typical signals of the OAg chain
and/or core region, confirming the presence of the characteristic
sugars and to quantify the O-acetylation level of the polysaccharide,
if present.
1. Record a 1H NMR spectrum with the sample polysaccharide
dissolved in 650 μL D2O (see Note 45).
2. Add directly in the tube, after having recorded the first spec-
trum, 35 μL 4 M NaOD in order to obtain a final 200 mM
concentration. Keep the sample at 37
C for 2 h to achieve a
complete de-O-acetylation.
3. Record a 1H NMR spectrum of the de-O-acetylated sample.
Figure 20 shows 1H NMR spectra of S. flexneri 6 capsular
polysaccharide isolated from GMMA.
 OMV/GMMA Analytics 271

Fig. 19 HPLC-SEC chromatogram of OAg/core extracted from S. Typhimurium GMMA (Fig. 10): detection at
252 nm after derivatization with semicarbazide (a); detection in differential refractive index (dRI) of the
underivatized sample (b)
272 Francesca Micoli et al.

0.15 [rel]
a NAc of GalNAc
OAc on
Rhaˡˡ4Ac when Rha are
NAc of not Ac
OAc on GalNAc
Rhaˡˡ3Ac

0.10
0.05
0.00
5 4 3 2 [ppm]

[rel]
b NAc of GalNAc
when Rha are
not Ac

12
10
8
6
Free
acetate

4
2
0
5 4 3 2 [ppm]

Fig. 20 1H NMR spectra of S. flexneri 6 capsular polysaccharide isolated from GMMA pre (a) and post (b) de-O-
acetylation in NaOD

4. Calculate the O-acetylation level by comparing acetate anion

signal (released after treatment with NaOD, at 1.91 ppm) and a
known signal of the OAg repeating unit structure. Eventually
correct the value for the acetate anion already present before
the basic treatment.

3.20 Hestrin The O-acetyl ester content can also be measured by the Hestrin
Colorimetric Method colorimetric method. This method is based on the reaction of the
ester groups with hydroxylamine in a basic media to form hydro-
xamic acid which, at low pH, generates a complex with Fe3+ with a
maximum absorbance at 540 nm. The Hestrin method suffers from
interference by common salts like phosphates; for this reason, OAg
extracted samples often need to be desalted for it to work properly.
O-acetylation is expressed as molar ratio (%) between acetic esters
and OAg repeating units (see Note 46).
 OMV/GMMA Analytics 273

Table 13
Dilutions for preparing calibration curve for O-acetyl determination by Hestrin colorimetric method

Acetylcholine (μmol/mL) 3 mg/mL std. acetylcholine (μL) Water (μL)

0 (blank) 0 1000
0.413 25 975
0.826 50 950
1.65 100 900
2.48 150 850
3.30 200 800
4.13 250 750

1. In a 15 mL Falcon tube, dilute 0.5 mL of 30 mg/mL acetyl-

choline adding 4.5 mL of water in order to obtain the standard
solution at 3 mg/mL.
2. In 5 mL glass tubes prepare, in duplicate, the dilutions of
acetylcholine standard solution (3.0 mg/mL) using volumes
reported in Table 13.
3. In 5 mL glass tubes, add (in sequence and vortexing each time)
200 μL of diluted samples/standard curve solutions and 400
μL of the basic hydroxylamine solution.
4. Wait 3 min before proceeding to the next step.
5. Add 200 μL of 4 M HCl, 200 μL of the 0.37 M iron chloride
solution and 100 μL of isopropanol.
6. Transfer all standards/samples to disposable cuvettes.
7. Read ABS of standards/samples in the following sequence with
spectrophotometer wavelength set at 540 nm: first read the
“blank” containing cuvette (if a single beam instrument is used
this step is necessary in order to subtract the blank absorbance
from samples/standards absorbances. If a double beam instru-
ment is used, leave the “blank” containing cuvette in the
reference slot during all sample/standard readings). Next,
read the acetylcholine standard curve solutions (each in dupli-
cate) from lower to higher concentrations. Finally, read the
samples (each in triplicate).
8. Calculate a linear regression between the ABS and acetylcho-
line standard concentrations (in the range 0.413–4.13 nmol/
mL) and calculate on it the concentration of the sample ester
groups. If the concentration of a sample is outside the range of
the calibration curve, dilute the sample with water and repeat
the analysis.
274 Francesca Micoli et al.

4 Notes

1. With respect to the conditions reported in the text, calculate

the centrifuge run time using the K-factor of the rotor in use
and the maximum speed achievable (calculation tools are avail-
able in each rotor manufacturer website).
2. The reagent has to be prepared just before use.
3. All samples containing lipid A and fatty acid standards dilutions
have to be placed and pipetted into low-bind tubes with
low-bind pipette tips and HPLC low-bind vials have to be used.
4. Be sure that no air bubbles are present between paper and
gasket and between paper and membrane and avoid complete
membrane drying in all steps from this point.
5. Avoid loading border wells.
6. As an alternative to the dot blot, western blot analysis can be
used. The OMV/GMMA sample is run by SDS-PAGE as
described in Subheading 3.5. The gel is transferred on a mem-
brane by using the appropriate apparatus and the recognition
with a specific primary antibody is performed as described for
dot blot.
7. Size distribution by intensity is preferred to measurements by
number or by volume to have more reproducible results
(Z-average) and because the refractive index (RI) values of
particles are not known.
8. For data analysis, the viscosity and RI of the sample buffer
solution (at 25
C) needs to be indicated, the parameters for
the main buffers are present in the software and a tool to
calculate them for new buffer composition is present.
9. It is essential that the sample reaches the correct temperature
before reading. For optimal measurement, temperatures of
sample, cuvette and cuvette holder must be stable and homog-
enous during the reading.
10. Whenever the 260 or 280 baselines show a trend during equil-
ibration of the column, wash for 200 min the stationary phase
with a PBS solution containing 1 M NaCl. If needed, a stron-
ger column wash (with a suitable titanium pump instrument)
can be performed, with 0.1 M NaOH for 200 min. Be careful,
after cleaning, to lower the pH with a phosphate buffer before
storage of the column in the preservative solution.
11. MALS elaboration, for each point of the chromatogram, uses
the 18 detector intensities with respect to the respective base-
lines to calculate the radius (peak area is not used). Be careful to
choose the baseline parallel with an eventual signal evident
drift.
 OMV/GMMA Analytics 275

12. Use a configuration with up-to-date light scattering detector

normalization parameters (BSA monomer is injected as recom-
mended by the instrument manufacturer to calculate these
parameters).
13. This is needed, as the LS signals of many detectors are too noisy
under that threshold.
14. For concentrated samples (i.e., >1 mg/mL protein), to avoid
dilution accuracy errors, the sample is diluted by weight on
analytical balance. Using as tare the empty dilution tube, regis-
ter the weight of the sample and of added water in order to
calculate the real dilution factor for each sample analyzed.
15. Check sample matrix compatibility as reported on the Thermo
kit instructions.
16. Use the centrifugation conditions reported as starting point, to
define the optimal conditions for a specific product a centrifu-
gation kinetic study needs to be performed finding the condi-
tions that allow to pellet quantitatively OMV/GMMA and not
soluble proteins. The ultracentrifugation supernatant of the
kinetic study can be for example analyzed by HPLC-SEC
with a fluorescence detector as reported in Subheading 3.3. It
is important to standardize in the centrifugation protocol also
the tube filling volume.
17. OMV/GMMA tend to spontaneously resuspend with time.
18. This step can be used to change the buffer of the
OMV/GMMA and to concentrate the sample by resuspension
in a lower volume with respect to the starting one.
19. In case of samples containing TBS, dilute them with PBS in
order to have a TBS final concentration of at maximum 1 mM
to avoid interference in the assay.
20. In general, to run an amino acid determination, for each
hydrolysis tube to match the amino acid calibration curves
ranges, about 40 μg protein for GMMA samples are required.
21. During acid hydrolysis, tryptophan normally decomposes; cys-
teine and cystine, tyrosine, threonine, and serine partially
decompose; methionine is partially oxidized; isoleucine and
valine have lower hydrolysis rate, and therefore the recovery
might be lower; asparagine and glutamine are hydrolyzed to
aspartic acid and glutamic acid.
22. Once the total amino acid concentration has been obtained
from the analysis, calculation of the amount of protein in the
original sample is done using the following formula
(to correlate the 100 μL resuspension volume to the initial
dried sample volume):

protein conc: found in amino acid analysis  100

Sample protein μg=mL ¼
Dried Sample initial volume ðμLÞ
276 Francesca Micoli et al.

23. Before taking up the standard solution for transfer, rinse the tip
with it by pipetting up and down multiple times.
24. Before taking up the reagent for transfer, rinse the tip with it by
pipetting up and down multiple times.
25. Do not add derivatizing agent to several vials and vortex them
all at the end but proceed to capping and vortexing immedi-
ately for each vial.
26. If a binary UPLC system is used for the analysis, refer to the
Waters Kit instruction material for the appropriate eluent
scheme and gradient program.
27. Being the AMQ the product of derivatizing reagent reaction
with water, if a degraded (i.e. hydrolyzed for humidity) reagent
is used this test doesn’t work.
28. After the last chromatographic analysis is completed, store the
column in 18 mM NaOH solution.
29. In this case, a standard OAg sample is used for building the
calibration curve, as the monomer (2-amino-2-deoxy-α-L-
altropyranuronic acid) resulting from sample hydrolysis and
quantified by the chromatographic analysis is not commercially
available.
30. After the last chromatographic analysis is completed, store the
column in 100 mM NaOH solution.
31. It is advisable, due to the short incubation time, to proceed in
parallel with less than 15 standard/sample tubes. If the number
of samples is higher, proceed with a cluster of 10 tubes for each
5 min incubation.
32. Both ABS have to be lower than 1.2 in order to be considered
acceptable.
33. Do not leave concentrated sulfuric acid solution in polystyrene
cuvettes but appropriately discard the solution just after usage
(cuvettes are slowly corroded).
34. LAL is not deemed appropriate, as samples contain such high
endotoxin units that they typically require a dilution factor of
million-billion.
35. M/z accuracy check needs to be done at first implementation of
the procedure on a new instrument or after a calibration of the
MS instrument. This is required, as accuracy issues in M/z,
especially for 59 Da fragment, were experienced. The check is
performed by direct infusion of 3-OH-lauric acid/3-OH-myr-
istic acid standard diluted in 70% ACN 0.05% FA to about
0.5–1 nmol/mL. Perform direct infusion of the standards
and acquire/visualize in real time with the following para-
meters: First set Q1 in full scan mode and note the 215 M/z
closest peak and/or the 243 M/z closest peak: each of these
 OMV/GMMA Analytics 277

two M/z values has to be set for the corresponding analyte

precursor ion. Set QqQ with Q1 in order to filter the
3-OH-lauric acid or 3-OH-myristic acid ion using M/z value
found in previous step, Q2 as collision cell with Argon pressure
at 1 mTorr and a collision energy of 16 V, Q3 in full scan mode.
Look at the 59 M/z closest peak: this M/z value has to be set for
the analyte product ion (acetate). In the described method
wherever the values 243, 215 and 59 Da are reported consider
the actual M/z values found for analyte precursor ions and
product ion.
36. The ESI and collision cell settings depend on the MS instru-
ment, in this chapter are reported settings for the Thermo
Quantum Access instrument. In the case of a different MS
instrument, automatic optimization of ESI and collision cell
parameters is needed: perform it by direct infusion of the
standard diluted in eluent B.
37. The dead volume calculation is needed for this analysis because
if the diverter valve commutes just after the SPE washing step,
a considerable amount of eluent A is injected into the C8 RP
column that results in no more equilibration in eluent B.
38. Samples and reference standards need to be in 50% IPA (v/v)
to ensure complete hydrolysis of ester bonds. For this reason,
the last dilution of the sample needs to be a two-fold dilution
by volume in IPA.
39. Cleaning should be performed after every 15–30 samples
assayed in triplicate. Run twice the program reported in
Table 14 for the binary pump at 250 μL/min to wash the
column. At the end of washing step, store the column with
10 column volumes (1.6 mL) of eluent E (65% ACN).
40. Depending on the purpose, a milder hydrolysis can be per-
formed to preserve the lipid A structure (i.e., pyrophosphate
groups): OMV/GMMA at 1 mg/mL in protein are kept in a
screw-cap vial at 100
C for 1 h in presence of 40 mM acetate
buffer pH 4.5 and 3% N-Octyl-β-D-gluco-pyranoside.

Table 14
Column cleaning procedure

Eluent C Eluent D
Min % %
0 75 25
2 75 25
32 0 100
35 75 25
278 Francesca Micoli et al.

41. The highest intensity spectra are usually recovered far from big
crystal formations.
42. Using the MALDI instrument in reflectron mode, it is likely to
find as highest peak exact m/z + 1 or exact m/z + 2 for the
isotopic abundance of the elements present in the structure.
During hydrolysis or matrix desorption, lipid A can lose one
phosphate (80 m/z) or other substituents so more than one
peak can be observed for one product. Note that +22 m/z/
+38 m/z peaks, respectively, represent +Na and + K adducts.
43. Y ¼ ((b/(a  Log10(X)))ˆ(d)  1)/c, where a, b, c, and
d represent values of the curve parameters (i.e., respectively:
a ¼ minimum asymptote, b ¼ Hill’s slope, c ¼ inflection point,
and d ¼ maximum asymptote), Y the amount determined, and
X the absorbance reading.
44. Perform system equilibration with elution buffer before start-
ing the analysis. This equilibration step lasts 70 min if the
column system has to be equilibrated with a different eluent;
60 min or more are needed also to warm up the PDA lamps
before the analysis.
45. The first 1H NMR spectrum is recorded to ensure the absence
of impurities or other signals at the same chemical shift of the
acetate anion released after de-O-acetylation of the sample that
would interfere with the quantification of the O-acetyl content.
46. Considering the multiple steps to prepare the sample for the
O-acetylation measurement (OAg extraction, desalting), prior
to calculating the O-acetyl ester/repeating unit ratio, it is
recommended to measure the OAg quantity on the same trea-
ted sample prepared for ester measurements and use this value
for the further calculation.

Acknowledgments

All authors were involved in drafting the book chapter and

approved the final version. The development of this article was
sponsored by GlaxoSmithKline Biologicals SA. The authors declare
the following interests: FM, CG, and RAL are employees of the
GSK group of companies.

References
1. Micoli F, MacLennan CA (2020) Outer mem- 74(1):81–94. https://doi.org/10.1128/
brane vesicle vaccines. Semin Immunol mmbr.00031-09
50:101433. https://doi.org/10.1016/j. 3. Gerke C, Colucci AM, Giannelli C et al (2015)
smim.2020.101433 Production of a Shigella sonnei vaccine based
2. Ellis TN, Kuehn MJ (2010) Virulence and on generalized modules for membrane anti-
immunomodulatory roles of bacterial outer gens (GMMA), 1790GAHB. PLoS One 10
membrane vesicles. Microbiol Mol Biol Rev
 OMV/GMMA Analytics 279

(8):e0134478. https://doi.org/10.1371/jour 14. De Benedetto G, Alfini R, Cescutti P et al

nal.pone.0134478 (2017) Characterization of O-antigen deliv-
4. van de Waterbeemd B, Streefland M, van der ered by generalized modules for membrane
Ley P et al (2010) Improved OMV vaccine antigens (GMMA) vaccine candidates against
against Neisseria meningitidis using genetically nontyphoidal Salmonella. Vaccine 35
engineered strains and a detergent-free purifi- (3):419–426. https://doi.org/10.1016/j.vac
cation process. Vaccine 28(30):4810–4816. cine.2016.11.089
https://doi.org/10.1016/j.vaccine.2010.04. 15. Rossi O, Aruta MG, Acquaviva A et al (2020)
082 Characterization of competitive ELISA and
5. Mancini F, Rossi O, Necchi F, Micoli F (2020) formulated alhydrogel competitive ELISA
OMV vaccines and the role of TLR agonists in (FAcE) for direct quantification of active ingre-
immune response. Int J Mol Sci 21(12):4416. dients in GMMA-based vaccines. Methods
https://doi.org/10.3390/ijms21124416 Protoc 3(3):62. https://doi.org/10.3390/
6. van der Pol L, Stork M, van der Ley P (2015) mps3030062
Outer membrane vesicles as platform vaccine 16. Rossi O, Caboni M, Negrea A et al (2016)
technology. Biotechnol J 10(11):1689–1706. Toll-like receptor activation by generalized
https://doi.org/10.1002/biot.201400395 modules for membrane antigens from lipid A
7. Kis Z, Shattock R, Shah N, Kontoravdi C mutants of Salmonella enterica serovars typhi-
(2019) Emerging technologies for low-cost, murium and enteritidis. Clin Vaccine Immunol
rapid vaccine manufacture. Biotechnol J 14 23(4):304–314. https://doi.org/10.1128/
(1):e1800376. https://doi.org/10.1002/ CVI.00023-16
biot.201800376 17. Micoli F, Ravenscroft N, Cescutti P et al (2014)
8. Li M, Zhou H, Yang C et al (2020) Bacterial Structural analysis of O-polysaccharide chains
outer membrane vesicles as a platform for bio- extracted from different Salmonella typhimurium
medical applications: an update. J Control strains. Carbohydr Res 385:1–8. https://doi.
Release 323:253–268. https://doi.org/10. org/10.1016/j.carres.2013.12.003
1016/j.jconrel.2020.04.031 18. Biagini M, Spinsanti M, De Angelis G et al
9. De Benedetto G, Cescutti P, Giannelli C, (2016) Expression of factor H binding protein
Rizzo R, Micoli F (2017) Multiple techniques in meningococcal strains can vary at least
for size determination of generalized modules 15-fold and is genetically determined. Proc
for membrane antigens from Salmonella typhi- Natl Acad Sci U S A 113(10):2714–2719.
murium and Salmonella enteritidis. ACS https://doi.org/10.1073/pnas.1521142113
Omega 2(11):8282–8289. https://doi.org/ 19. Raso MM, Gasperini G, Alfini R et al (2020)
10.1021/acsomega.7b01173 GMMA and glycoconjugate approaches com-
10. Tani C, Stella M, Donnarumma D et al (2014) pared in mice for the development of a vaccine
Quantification by LC-MS(E) of outer mem- against Shigella flexneri serotype 6. Vaccine 8
brane vesicle proteins of the Bexsero® vaccine. (2):160. https://doi.org/10.3390/
Vaccine 32(11):1273–1279. https://doi.org/ vaccines8020160
10.1016/j.vaccine.2014.01.011 20. MacLennan CA, Martin LB, Micoli F (2014)
11. Maggiore L, Yu L, Omasits U et al (2016) Vaccines against invasive Salmonella disease:
Quantitative proteomic analysis of Shigella flex- current status and future directions. Hum Vac-
neri and Shigella sonnei generalized modules cin Immunother 10(6):1478–1493. https://
for membrane antigens (GMMA) reveals doi.org/10.4161/hv.29054
highly pure preparations. Int J Med Microbiol 21. Mani S, Wierzba T, Walker RI (2016) Status of
306(2):99–108. https://doi.org/10.1016/j. vaccine research and development for Shigella.
ijmm.2015.12.003 Vaccine 34(26):2887–2894. https://doi.org/
12. Micoli F, Alfini R, Di Benedetto R et al (2020) 10.1016/j.vaccine.2016.02.075
GMMA is a versatile platform to design effective 22. Lyngby J, Olsen LH, Eidem T, Lundanes E,
multivalent combination vaccines. Vaccine 8 Jantzen E (2002) Quantification of lipopoly-
(3):540. https://doi.org/10.3390/ saccharides in outer membrane vesicle vaccines
vaccines8030540 against meningococcal disease. High-
13. Giannelli C, Raso MM, Palmieri E et al (2020) performance liquid chromatographic determi-
Development of a specific and sensitive nation of the constituent 3-hydroxy-lauric acid.
HPAEC-PAD method for quantification of vi Biologicals 30(1):7–13. https://doi.org/10.
polysaccharide applicable to other polysacchar- 1006/biol.2001.0285
ides containing amino uronic acids. Anal Chem 23. Macgee J, Doudoroff M (1954) A new phos-
92(9):6304–6311. https://doi.org/10.1021/ phorylated intermediate in glucose oxidation. J
acs.analchem.9b05107 Biol Chem 210(2):617–626
 Chapter 15

Production of Vaccines Using Biological Conjugation

Emily J. Kay and Vanessa S. Terra

Abstract
The production of conjugate vaccines within an E. coli (Escherichia coli) host provides an inexhaustible
supply without the need for culture of pathogenic organisms. The machinery for expression of glycan and
acceptor protein, as well as the coupling enzyme, are all housed within the E. coli chassis, meaning that there
are no additional steps required for individual purification and chemical conjugation of components. In
addition, there are far fewer purification steps necessary to obtain a purified glycoconjugate for use in
vaccine testing. Here we describe production and purification of a HIS-tagged Campylobacter jejuni AcrA
protein conjugated to Streptococcus pneumoniae serotype 4 capsule.

Key words Vaccine, Biological conjugation, Streptococcus pneumoniae, PGCT, PglB

1 Introduction

The use of biological conjugation, or protein glycan coupling

technology (PGCT) was first exploited after the discovery of the
PglB enzyme encoded within the genome sequence of Campylo-
bacter jejuni [1] and its subsequent expression in Escherichia coli
nearly 20 years ago [2]. In the intervening years, multiple improve-
ments and refinements have been made to the initial conjugation
protocols. These improvements include: introduction of a PglB
consensus sequence (sequon) into the protein to be glycosylated
[3], chromosomal integration of the various cloned components
required for the reaction [4, 5], optimized culture conditions
[6, 7], and genetic engineering of PglB to increase efficiency of
glycan transfer [8]. In addition, alternative conjugation enzymes
have been discovered such as PglL from Neisseria meningitidis
[9, 10], PglS from Acinetobacter species [11] and Ngt from
Actinobacillus pleuropneumoniae [12]. PglL and PglS are both
O-linked oligosaccharyltransferases that are more promiscuous in
terms of substrate specificity than PglB, although PglL has never

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_15,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

281
282 Emily J. Kay and Vanessa S. Terra

been demonstrated to transfer sugars with glucose at the reducing

end, whereas PglS has [11]. PglL glycosylates a handful of proteins
including PilE [9] and more recently an extended glycosylation
recognition motif (MOOR) has been defined and used as a fusion
to other proteins [10]. PglS glycosylates a single protein, ComP,
and although the glycosylation site is known, the minimum section
of protein required for glycosylation has yet to be defined [11]. Ngt
is an N-linked oligosaccharyltransferase which acts in the cyto-
plasm, adding glucose directly from nucleotide activated sugar to
protein [12].
There are many ways to achieve biological conjugation but
some basic principles apply. In the example presented in this chap-
ter, conjugation using CjPglB will be considered. The requirements
for this enzyme are a glycan with an N-acetylated reducing end
sugar that is built on undecaprenol pyrophosphate, and a host
strain that is free of competing, compatible glycans, such as endog-
enous O-antigen. In addition, the host strain must possess all the
necessary sugar precursors to assemble the glycan or extra biosyn-
thetic pathways need to be added. A modified acceptor protein
must be co-expressed, containing at least one accessible consensus
sequon of (D/E-X-N-X-S/T; where X is any amino acid except for
proline) [13], as well as the introduction of a signal peptide that
targets the protein to the periplasm where the transfers occurs. The
consensus sequon can either be engineered within the protein in an
accessible loop, providing there is structural information for
the protein in question, or added to the N- and/or C- terminus.
The addition of a signal peptide sequence targets the protein to the
periplasm. Though optimization is often required to find the best
leader sequence, common options are those from proteins PelB and
DsbA [14]. Finally, the conjugating enzyme, CjPglB, must be
expressed concurrently with the glycan and acceptor protein
[15]. Any of the components can be expressed from compatible
plasmids or integrated into the chromosome of the host strain.
Again, optimization of both plasmid backbone and inducer systems
is often required.
Once all components are assembled and expressed within
E. coli, the optimum growth and induction conditions must be
determined [6]. Here we demonstrate the steps following large-
scale growth, where the glycoconjugate must be harvested, purified
via an affinity tag on the acceptor protein, and subjected to addi-
tional clean-up steps to remove contaminating proteins and endo-
toxin (Fig. 1).
In this example, AcrA, a C. jejuni protein with an added PelB
leader sequence and two native internal glycosylation sequons, is
conjugated to Streptococcus pneumoniae serotype 4 capsule (SP4),
recombinantly expressed in E. coli, via the oligosaccharyl transferase
enzyme, CjPglB [16]. The conjugate is purified first using a
His-Trap column on an AKTA chromatography system (Fig. 2),
 Production of Vaccines Using Biological Conjugation 283

Fig. 1 Glycoengineering approach to the production and purification of glycoconjugate vaccines. An E. coli cell
is transformed with the components needed to generate the glycoconjugate protein, shown here encoded on
plasmids for (1) glycan expression; (2) target protein with leader sequence and glycosylation sequons; (3) the
enzyme responsible for coupling the glycan to the protein. Any of the components could be inserted into the
chromosome, as in the protocol described in this chapter where PglB (the coupling enzyme is inserted onto the
chromosome. The resulting glycoconjugate must then undergo several rounds of purification to remove
contaminants and unconjugated glycan. (This figure was created with BioRender.com)

followed by subsequent steps to remove endotoxin, and finally

anion exchange chromatography to further purify the glycosylated
protein (Fig. 3).

2 Materials

All buffers used for chromatography should be passed through

0.2 μm filter and degassed by spinning with a magnetic stir bar
whilst applying vacuum for at least 30 min.

2.1 Cell Growth 1. E. coli W311B containing plasmids pWA2 and pB4 (see
Note 1).
2. Growth medium SSOB: 2% tryptone, 0.5% yeast extract,
171 mM NaCl, 2.5 mM KCl, 10 mM MgCl2, 10 mM
MgSO4. Weigh 40 g tryptone, 10 g yeast extract, 20 g NaCl,
372 mg KCl, 1.904 g MgCl2 and 4.928 g MgSO4·7H2O and
284 Emily J. Kay and Vanessa S. Terra

Fig. 2 (a) AKTA trace of AcrA-SP4 glycoconjugate eluted from the His-Trap
column with imidazole. The green line denotes increasing concentration of
imidazole, the x axis denotes the volume of liquid passed through the column,
and the Y axis denotes the UV absorbance at 280 nm, as an indication of eluted
protein concentration. Fractions are labeled by their position in the AKTA
collection tray, which accommodates 15 tubes per row. The red lines indicate
the 1 mL fractions collected between A1 and B14. (b) Selected eluate fractions
separated on SDS-PAGE gel followed by Western blot with anti-His tag fluores-
cent antibody (red) and anti-SP4 capsule fluorescent antibody (green), as
visualized by a LiCor Odyssey digital imager. M molecular weight marker
PageRuler Plus

dissolve in 1.6 L distilled water. Once dissolved, make the

volume up to 2 L with distilled water and autoclave at 121
C
with 15 psi of steam for 20 min.
 Production of Vaccines Using Biological Conjugation 285

Fig. 3 (a) AKTA trace of AcrA-SP4 glycoconjugate eluted from the Resource S column with NaCl. The green line
denotes increasing concentration of NaCl, the x axis denotes the volume of liquid passed through the column,
and the Y axis denotes the UV absorbance at 280 nm, as an indication of eluted protein concentration.
Fractions are labeled by their position in the AKTA collection tray, which accommodates 15 tubes per row. The
red lines indicate the 1 mL fractions collected between A1 and B6. Selected eluate fractions were separated
on SDS-PAGE gel followed by either (b) Western blot with fluorescent anti-His tag antibody (red) and
fluorescent anti-SP4 capsule antibody (green), or (c) Coomassie staining, both visualized by a LiCor Odyssey
digital imager. S1–S4 denote high salt fractions collected after the salt concentration was increased to 1 M
NaCl. M molecular weight marker PageRuler Plus

3. Tetracycline: 10 mg/mL in 70% ethanol, filter-sterilized and

stored in aliquots at 20
C.
4. Ampicillin: 100 mg/mL in distilled water, filter-sterilized and
stored in aliquots at 20
C.
5. Inducer: 1 M IPTG, in distilled water, filter-sterilized and
stored in aliquots at 20
C.
6. PglB cofactor: 1 M MnCl2, in distilled water and filter-
sterilized (see Note 2).
286 Emily J. Kay and Vanessa S. Terra

2.2 Cell Harvesting 1. Centrifuge with rotors capable of accommodating 500 mL

and Lysis containers and spinning at 6000  g and 30 mL containers at
10,000  g.
2. High pressure cell homogenizer (French press or similar).
3. Lysis buffer: 50 mM NaH2PO4, 0.3 M NaCl, pH 7.4. Weigh
3.45 g of NaH2PO4 and 8.77 g of NaCl and dissolve in 400 mL
distilled water. Adjust the pH to 7.4 with concentrated NaOH
before making the volume up to 500 mL and passing through a
0.2 μm filter membrane (see Note 3).
4. Lysozyme: 100 mg/mL in 10 mM Tris–HCl, pH 8. First
prepare 1 M Tris–HCl, pH 8 by weighing 2.11 g of Tris base
and dissolving it in 80 mL distilled water. Adjust the pH to
8 with concentrated HCl before making the volume up to
100 mL and passing through a 0.2 μm filter membrane, or
autoclave at 121
C with 15 psi of steam for 20 min. Next,
prepare 10 mL of 10 mM Tris–HCl pH 8 by diluting 1 M
solution 1:100 with distilled water. Weigh out 1 g of lysozyme
and dissolve in a 10 mL final volume of 10 mM Tris–HCl,
pH 8. The 100 mg/mL lysozyme solution is then passed
through a 0.2 μm filter membrane before storage in 500 μL
aliquots at 20
C.
5. Benzonase nuclease high strength (250 U/μL).
6. 0.2 μm filter units.
7. Vacuum pump.

2.3 AKTA 1. AKTA Chromatography system.

High-Performance 2. 20% ethanol (1 L).
Liquid
3. 0.5 N NaOH (500 mL).
Chromatography
4. Distilled, degassed water (1 L).
5. Peristaltic pump.
6. 0.2 μm filter units.
7. Magnetic stirrer, stir bar, and vacuum pump for degassing
buffers.

2.4 HisTrap 1. HisTrap 1 mL column (e.g., HisTrap FF column, Cytiva, or

Purification similar product).
2. Wash buffer I: 50 mM NaH2PO4, 0.3 M NaCl, 0.1% Triton
X-114, pH 7.4. Weigh 1.725 g of NaH2PO4 and 4.385 g of
NaCl and dissolve in 200 mL distilled water. Adjust the pH to
7.4 with NaOH before making the volume up to 249.75 mL
and adding 250 μL of Triton X-114. Pass through a 0.2 μm
filter membrane and store at 4
C.
3. Wash buffer II: 50 mM NaH2PO4, 0.3 M NaCl, 20 mM
imidazole, pH 7.4. Weigh 3.45 g of NaH2PO4, 8.77 g of
 Production of Vaccines Using Biological Conjugation 287

NaCl and 0.68 g of imidazole, and dissolve in 400 mL distilled

water. Adjust the pH to 7.4 with NaOH before making the
volume up to 500 mL and passing through a 0.2 μm filter
membrane.
4. Elution buffer: 50 mM NaH2PO4, 0.3 M NaCl, 300 mM
imidazole, pH 7.4. Weigh 1.725 g of NaH2PO4, 4.385 g of
NaCl and 5.1 g of imidazole, and dissolve in 200 mL distilled
water. Adjust the pH to 7.4 with HCl before making the
volume up to 250 mL and passing through a 0.2 μm filter
membrane.

2.5 Resource S 1. Resource S 1 mL columns prepacked with ion exchange resin

Purification (e.g., SOURCE15S resin, Cytiva) (see Note 4).
2. Start buffer: 20 mM MES, pH 6. Weigh 1.952 g MES and
dissolve in 400 mL distilled water. Adjust pH to 6 with NaOH
and pass through a 0.2 μm filter membrane.
3. Elution buffer: 20 mM MES, 1 M NaCl, pH 6. Weigh 976 mg
MES and 14.61 g NaCl in 250 mL. Adjust with NaOH to pH 6
before passing through a 0.2 μm filter membrane.
4. 2 M NaCl (250 mL).
5. 1 M NaOH (250 mL).
6. Storage buffer: 20% ethanol–0.2 M sodium acetate. 50 mL
ethanol, 4.1 g sodium acetate, to 250 mL with water.

2.6 PD-10 Desalting 1. PD-10 desalting columns—containing Sephadex G-25 resin.

and Endotoxin 2. High-capacity endotoxin removal spin columns.
Removal
3. Centrifugal concentrator columns, such as Vivaspin, with a
10 kDa molecular weight cut off (MWCO).
4. PD-10 equilibration buffer: 50 mM NaH2PO4, 25 mM NaCl,
pH 7.4. Weigh 1.725 g of NaH2PO4 and 0.365 g of NaCl and
dissolve in 200 mL distilled water. Adjust the pH to 7.4 before
making the volume up to 250 mL and passing through a
0.2 μm filter membrane.
5. Endotoxin removal buffer: 50 mM NaH2PO4, 0.1 M NaCl,
pH 7.4. Weigh 1.725 g of NaH2PO4 and 1.461 g of NaCl and
dissolve in 200 mL distilled water. Adjust the pH to 7.4 before
making the volume up to 250 mL and passing through a
0.2 μm filter membrane.
6. Tube rotator with “Ferris Wheel” disk for end over end
rotation.
7. Lyophilizer.

2.7 SDS-Page 1. Bis-Tris acrylamide gels.

2. Power pack.
288 Emily J. Kay and Vanessa S. Terra

3. Electrophoresis chamber and corresponding gel holder.

4. 10 MOPS running buffer: To 800 mL of distilled water add
41.86 g of MOPS free acid, 4.1 g of sodium acetate and 3.72 g
of Na2EDTA. Adjust solution to pH 7.7 using NaOH and
make the total volume up to 1 L with distilled water.
5. 4 Laemmli sample buffer: 240 mM Tris–HCl, pH 6.8, 40%
glycerol, 8% SDS, 0.02% Bromophenol blue. To 12 mL of
distilled water add 12 mL of 1 M Tris, pH 6.8, 20 mL 100%
glycerol, 4 g of SDS, 10 mg of Bromophenol blue, before
making the final volume up to 50 mL. Add 100 μL of
2-mercaptoethanol per 900 μL (see Note 5).
6. Heat block that can accommodate 1 mL tubes and heat to
100
C.

2.8 Western Blot 1. Nitrocellulose membrane.

2. Semidry transfer apparatus.
3. Power pack.
4. Tweezers.
5. Filter paper.
6. Towbin buffer: 25 mM Tris, 192 mM glycine, pH 8.3, 20%
(v/v) methanol. To 50 mL distilled water add 303 mg Tris
base, 1.44 g glycine and 20 mL methanol before adjusting the
volume to 100 mL with distilled water (see Note 6).
7. Phosphate buffer saline (PBS): NaCl: 137 mM KCl: 2.7 mM
Na2HPO4: 10 mM KH2PO4: 1.8 mM. Using a fine balance
measure 8 g of NaCl, 200 mg of KCl, 1.44 g of Na2HPO4 and
245 mg of KH2PO4. Add all salts to 800 mL of distilled water.
Adjust the pH to 7.4 with concentrated HCl before making the
final volume up to 1 L with distilled water.
8. PBS-T (PBS 0.1% Tween): To 500 mL of PBS add 500 μL of
Tween-20.
9. PBS-M: 2% milk powder in PBS, prepared freshly before use.
10. Primary antibodies: mouse anti-6-His antibody and rabbit
anti-SP4 antibody (Statens serum Institute, Denmark).
11. Secondary antibodies: IRDye® 680RD goat anti-mouse and
IRDye® 800CW goat anti-rabbit conjugates.
12. Rocker or orbital shaking platform.
13. Digital imaging system for detection of fluorescence at 680 nm
and 800 nm (see Note 7).

2.9 Coomassie 1. Staining tray big enough to hold SDS-PAGE gel and around
Staining 30 mL liquid.
 Production of Vaccines Using Biological Conjugation 289

2. Coomassie staining solution: 50% MeOH, 10% HoAC, 40%

H2O, 0.25% Coomassie Blue R-250. To 80 mL distilled water
add 100 mL methanol, 20 mL acetic acid, and 500 mg Coo-
massie Blue R-250. Stir the solution for 3–4 h and then filter
through Whatman filter paper. Store at room temperature.
3. De-stain solution: 5% methanol, 7.5% acetic acid, 87.5% dis-
tilled water. To 875 mL distilled water add 50 mL methanol
and 75 mL acetic acid.

3 Methods

All equipment should be appropriately cleaned according to the

manufacturer’s specifications and local equipment usage
regulations.

3.1 Cell Growth 1. In a 5 L flask, inoculate 2 L SSOB containing appropriate

and Expression supplements (20 μg/mL tetracycline; 100 μg/mL ampicillin,
1 mM IPTG, and 4 mM MnCl2) with a preinduced starter
culture of E. coli W311B+ pWA2 + pB4 (i.e., 50 mL of SSOB
with 0.5 mM IPTG, 20 μg/mL tetracycline, and 100 μg/mL
ampicillin, inoculated with a single colony from a plate, incu-
bated overnight at 28
C) to an OD600nm 0.03 (see Note 8).
2. Incubate at 28
C, shaking at 110 rpm for a total of 24 h.
3. Harvest cells by centrifugation at 5400  g for 1 h at 4
C in
500 mL centrifuge bottles.
4. Discard the supernatant and freeze pellets from each of the
centrifuge buckets at 80
C after weighing them (see Note 9).

3.2 Cell Lysis 1. Thaw pellet by submerging storage tube in cold water for
15 min.
2. Resuspend cell pellet (3–4 g wet weight) in 100 mL lysis buffer
+ 1 mg/mL lysozyme (1.2 mL 100 mg/mL) + 10 μL benzo-
nase (2.5 KU) (see Note 10).
3. Lyse with a Stansted High Pressure cell homogenizer or similar
high-pressure cell homogenizer. Prerinse the homogenizer
with 70% ethanol, distilled water and lysis buffer before use.
Sample should be passed through the high-pressure cell
homogenizer as many times as required for the resuspended
pellet suspension to significantly clear.
4. After lysis, centrifuge sample at 3000  g for 15 min at 4
C to
pellet large debris.
5. Supernatant should be decanted into fresh centrifuge tubes and
insoluble cell debris removed by centrifugation at 7800  g for
1 h at 4
C.
290 Emily J. Kay and Vanessa S. Terra

6. At this stage, the supernatant should be passed through a

0.2 μm filter into a suitable storage container for loading
onto the HisTrap column (see Note 11).

3.3 Chromatography 1. Prepare a 1 mL HisTrap-FF column by flushing through with:

Purification 5 column volumes (CV) distilled water, followed by 5 CV lysis
with HisTrap Columns buffer (see Note 12).
2. Cleared and filtered cell lysate is loaded onto the prepared 1 mL
HisTrap-FF column using a peristaltic pump at a rate of 1 mL/
min.
3. After loading, wash the column for 1 h (60 mL) at 4
C with
Wash I (contains 0.1% Triton-X114) (see Note 13).
4. Prepare AKTA by washing the pumps and feed lines with
around 30 mL 20% ethanol and then with water, followed by
5 mL each of wash buffer II and elution buffer for pump A
and B, respectively, until baseline levels.
5. Load the sample-containing HisTrap-FF column onto the
AKTA, with wash buffer II running at 1 mL/min through
the feed line so that you can attach the column drop-to-drop
to avoid introducing air into the column.
6. Wash the column with 30 mL wash buffer II, until UV baseline
is steady (collect wash fraction of UV peaks to analyze by SDS-
PAGE/Western blot).
7. Elute protein from column using an imidazole gradient at
1 mL/min (feed line A ¼ wash buffer II, feed line B ¼ elution
buffer. Change from 0–100% B over 30 mL, with 1 mL frac-
tions collected) (see Note 14).
8. Analyze selected fractions showing increased OD280 (Fig. 2a)
by SDS-PAGE with Coomassie staining and by Western blot
for detection of His and glycan (Fig. 2b) (see Note 15).
9. Stored HisTrap columns in 20% ethanol (see Note 16).
10. Clean the AKTA with 30 mL of 0.5 M NaOH, water, and 20%
ethanol before storage.

3.4 SDS-Page 1. Prepare each of the selected eluted fractions separately in

Laemmli buffer (i.e., 5 μL sample, 10 μL PBS, 5 μL 4
Laemmli buffer containing mercaptoethanol) in 1 mL tubes
and heat at 95
C for 10 min. If both Coomassie staining and
Western blot are desired, double the quantity of sample can be
prepared and run on two identical gels.
2. Centrifuge heated samples for 3 min at maximum speed in a
benchtop centrifuge to pellet any debris.
3. Dilute the 10 MOPS by adding 100–900 mL distilled water.
 Production of Vaccines Using Biological Conjugation 291

4. Assemble the gel in the electrophoresis chamber, fill the inner

portion between the gel(s) and the gel holder with 1 MOPS
running buffer. Pour the remaining 1 MOPS running buffer
into the outer chamber.
5. Load 18 μL of each sample supernatant into appropriate wells
of a NuPAGE 10% Bis-Tris Gel Novex® (or similar SDS-
PAGE gel).
6. Run a protein ladder alongside your samples. Confirm that the
protein ladder is compatible with the imaging system used.
7. Attach the electrodes to the power pack and run the gel at
100 V, surrounded by ice, for 2 h 25 min or until the dye front
has just run off the bottom of the gel.
8. Remove the gel from the electrophoresis chamber and rinse
with distilled water before proceeding with either Western blot
or Coomassie staining.

3.5 Western Blot 1. Cut nitrocellulose membrane to the size of the gel and soak in
MilliQ water for 10 min, then transfer to Towbin buffer (see
Note 17).
2. Cut six pieces of filter paper to the size of the gel and soak them
in Towbin buffer.
3. Assemble the transfer stack in the semi-dry transfer unit as
follows, using tweezers to handle filter paper and membrane:
on the anode, place three Towbin-soaked filter papers followed
by the nitrocellulose membrane. Submerge the rinsed gel in
Towbin buffer, then carefully place it on the membrane being
careful not to move it once it has touched the membrane. At
this point air bubbles should be removed by gently rolling a
clean object over the top. The remaining three Towbin-soaked
filter papers should be placed on top of the stack and air
bubbles removed. The cathode is then placed on top and the
electrodes connected to the power pack.
4. The transfer is run at 1 mA/cm2 for 1 h before removing the
membrane to a suitable clean washing tray.
5. Wash the membrane once with PBS to remove residual Towbin
buffer and then incubate with PBS-M for 1 h at room
temperature.
6. Pour off the PBS-M and wash the membrane three times with
around 30 mL PBS-T. Gentle agitation should be applied
during incubation of the wash steps by using a rocker or orbital
shaker at room temperature.
7. Incubate the membrane with a mix of the two primary anti-
bodies diluted in 10 mL PBS-M (1:10000 mouse anti-His
antibody and 1:1000 rabbit anti-SP4 antibody) for 1 h at
room temperature for simultaneous detection of protein
(anti-His) and glycan (anti-SP4) (see Note 18).
292 Emily J. Kay and Vanessa S. Terra

8. Wash the membrane as in step 6.

9. Incubate the membrane with a mix of the two secondary anti-
bodies, both at 1:10000 in 10 mL PBS-M, for 45 min at room
temperature. The IRDye® 680RD goat anti-mouse allows for
detection of recombinant proteins and IRDye® 800CW Goat
anti-rabbit detection of glycosylation (see Note 19).
10. Wash the membrane as in step 6 with a final wash in PBS only
to wash the tween off the membrane before imaging.
11. Detect the fluorescent signal emitted by the secondary antibo-
dies using a digital imager capable of fluorescence detection at
680 nm and 800 nm.

3.6 Coomassie Stain 1. Place the rinsed gel in a staining tray with around 30 mL of
Coomassie stain. Place on a rocker or orbital shaker for for
2–4 h, until the gel is a uniform blue colour.
2. Pour off the Coomassie stain and de-stain using de-staining
solution for 4–24 h, changing the de-stain solution periodically
until background is clear and protein bands are prominent.
3. Image the gel using any image detection equipment or camera
against a white background.

3.7 Desalting 1. Prepare an endotoxin removal column the day before use.
and Endotoxin Twist off bottom closure, loosen cap and place in 50 mL
Removal centrifuge tube. Spin at 500  g for 1 min. Discard solution.
Plug the bottom with the plug provided with the column and
add 8 mL 0.2 N NaOH, replace cap and invert several times.
Incubate overnight at room temperature.
2. Pool selected HisTrap eluate fractions from Subheading 3.3
that contain target protein, as determined by Western blot and
Coomassie-stained gel in Subheadings 3.4–3.6 (Fig. 2) (see
Note 20).
3. Prewash a 10 kDa MWCO centrifuge filter column (e.g.,
Vivaspin) twice by adding the maximum fill volume of MilliQ
water at centrifuging at 4000  g for 1 min, then load pooled
sample onto the column and concentrate until sample volume
is no more than 2.5 mL (see Note 21).
4. Use PD-10 desalting column gravity protocol. Remove storage
buffer by pouring it off the column. Drip through 25 mL
PD-10 equilibration buffer, then apply a maximum of 2.5 mL
sample (see Note 22).
5. Discard flow through and then elute sample with 3.5 mL
equilibration buffer, collecting the eluate in 1 mL fractions.
6. Clean the PD-10 column with 25 mL equilibration buffer,
then store in 20% ethanol (see Note 23).
 Production of Vaccines Using Biological Conjugation 293

7. Pool the first 2.375 mL of eluate from PD-10 column and add
125 μL 2 M NaCl to increase the salt concentration to 0.1 M.
8. To prepare an endotoxin removal column, remove bottom
plug and cap and centrifuge column at 500  g 1 min, then
wash with 8 mL each of 2 M NaCl, followed by water, and
3 rounds of endotoxin removal buffer, with centrifugation at
500  g 1 min for each wash.
9. Plug the endotoxin removal column and add 2.5 mL sample
(salt adjusted PD-10 eluate from step 6), replace cap and invert
several times. Incubate the column at 4
C with end-over-end
mixing, on a tube rotator fitted with a “Ferris Wheel” disk,
overnight.
10. Remove bottom plug and cap and centrifuge the column at
500  g 1 min to elute sample off the endotoxin removal
column into a fresh collection tube (see Note 24).
11. Remove salt before anion exchange by using PD-10 desalting
columns as in steps 4–6.

3.8 Cation Exchange 1. Prepare the Resource S column (see Note 25) by running
Chromatography through 5 CV each of MilliQ water, start buffer, elution buffer,
and start buffer before loading sample.
2. The desalted sample in PD-10 equilibration buffer must be
diluted in start buffer before loading onto the column (e.g.,
3 mL sample diluted up to 35 mL with start buffer).
3. After loading, the column is attached to the AKTA and washed
with 10–20 mL start buffer until the UV baseline is steady (see
Note 26).
4. The sample is eluted from the column using a salt gradient up
to 0.5 M NaCl (feed line A ¼ start buffer, feed line B ¼ elution
buffer. Change from 100% A to 50% B over 20 mL, with 1 mL
fractions collected).
5. The eluted fractions can be analyzed at this stage by SDS-
PAGE with Coomassie staining and by Western blot as in
Subheadings 3.4–3.6 (Fig. 3).
6. Clean the Resource S column on the AKTA using a high salt
wash of 10 mL 1 M NaCl (this can be achieved by using only
feed line B, that is, increase from 50% B to 100% B for 10 min)
(see Note 27).
7. The AKTA is cleaned with 30 mL 0.5 M NaOH, distilled water,
and 20% EtOH.

3.9 Buffer Exchange 1. Pool the desired fractions from Resource S column (see Note
and Storage 28), concentrate to around 2.5 mL and buffer exchange using
Vivaspin 10 kDa MWCO centrifuge filter column, as described
in Subheading 3.7, step 3 (again, prewash filter twice with
MilliQ water before loading sample).
294 Emily J. Kay and Vanessa S. Terra

2. Exchange the buffer using Vivaspin 10 kDa MWCO centrifuge

column by washing three times with PBS (see Note 29).
3. Determine protein concentration using any acceptable protein
quantification method such as BCA (Pierce). It is also advisable
to quantify the glycan (see Note 30).
4. Lyophilize in ready to use aliquots for long term storage (see
Note 31).

4 Notes

1. pWA2 [17] is a derivative of pBR322 [18], which has a pMB1

origin of replication and an ampicillin resistance cassette,
encoding AcrA, the protein to be glycosylated. pB4 [19]
encodes the SP4 polysaccharide locus cloned into
pBBR1MCS-3, which contains a tetracycline resistance cassette
and a pBBR1 origin of replication, compatible with IncC, IncP,
IncQ, and IncW group plasmids [20]. W311B is Escherichia
coli W3110 [21] with PglB under an IPTG inducible promoter,
integrated into the chromosome using a transposon [16]. It is
essential that the plasmids are from different incompatibility
groups [22, 23]. Even with different antibiotic resistance selec-
tion, incompatible plasmids will not be maintained stably
within E. coli leading to variable copy number and even plasmid
loss. This is a major issue for consistent and reproducible
glycoconjugate production.
2. MnCl2 must be made fresh each time of use or stored in ready
to use 1 mL aliquots at 20
C for up to 6 months as it is
unstable. MnCl2 is a cofactor for PglB and was shown to have a
binding pocket in the active site of PglB [24, 25]. Addition of
MnCl2 was shown to improve glycosylation at a concentration
of 4 mM [26].
3. Lysis buffer, wash buffer II and elution buffer are used at RT,
wash buffer I is used at 4
C, but all His-purification buffers
may be stored at 4
C for up to 1 month.
4. Other cation or anion exchange resin may be used but the
buffers described in this protocol are specific for the Resource
S columns (Cytiva). Other resins may require different buffers
for binding and elution.
5. Instead of mercaptoethanol, dithiothreitol (DTT or Cleland’s
reagent) may be added to a final 1 concentration of 50 mM as
a reducing agent. Sample buffer should not be stored with
reducing agent; instead, it should be added to a working ali-
quot just before use.
 Production of Vaccines Using Biological Conjugation 295

6. Towbin buffer with methanol should not be stored. 100 mL is

enough for transfer of 2 midi gels of size 8.7 cm  13.3 cm.
10 Towbin buffer without methanol can be made in larger
quantities and stored at room temperature, the buffer should
be discarded once the salts visibly precipitate from solution.
7. A LI-COR Odyssey® (LI-COR Biosciences, UK) image detec-
tion system was used, but any comparable system capable of
detecting the secondary antibodies may be used as appropriate.
8. If facility for incubating 2 L flasks is unavailable, then the
culture may be split into smaller flasks as long as the ratio of
culture to flask capacity does not exceed 2:5.
9. At this stage the pellet can be transferred to an appropriate tube
and stored at 80
C until purification.
10. If the pellet weighs more than 3–4 g wet weight, then add in a
proportional amount of lysis buffer with lysozyme and
benzonase.
11. It is best to load the sample onto the column as soon as possible
rather than storing, as the protein is liable to degrade without
the addition of protease inhibitors, which may then affect
downstream applications.
12. It is advisable to keep a separate column for each glycoconju-
gate to be purified, but each column can be used multiple
times. If the column has been previously used and stored
stripped, then it must be recharged before use as follows:
5 CV distilled water, recharge with 1 mL 0.1 M NiSO4, 5 CV
water, 5 CV lysis buffer.
13. An on-column Triton wash is used to significantly reduce the
amount of contaminating endotoxin [27]. It is necessary to
perform the wash at 4
C due to the cloud point of Triton
X-114. Endotoxin removal is necessary if a functional
O-antigen ligase is present and a further anionic exchange
purification step is needed. The O-antigen ligase will attach
recombinant glycan to the lipid A core which will then
co-purify with the glycoconjugate protein. If the endotoxin
removal step is not taken, the lipid-linked glycan will block
the anion exchange column. Following the purification, a Lim-
ulus amebocyte lysate (LAL) assay may be used to determine
how much endotoxin remains in the sample. There are many
commercial kits available for this.
14. The gradient can be altered according to the protein to be
eluted. It is best to collect at least 5 mL of eluate before the
protein elutes. This allows for contaminating proteins to be run
off the column before elution of the protein of interest. The
gradient is achieved by increasing the % of elution buffer run-
ning through the column up to 100%.
296 Emily J. Kay and Vanessa S. Terra

15. As the protein AcrA has two glycosylation sites there should be
two bands visible above the unglycosylated protein, denoting
short repeat units attached at one or both glycosylation sites. In
addition, glycan bands should be visible above the protein,
showing polymerized glycan attached to the protein. Control
samples of protein only and glycan only can be included in
Western blots for comparison. Any glycan visible at the bottom
of the gel will be glycan that is not attached to the target
protein.
16. It is recommended to store columns stripped and only recharge
with Ni2+ before use. A full cleaning of columns is done using
5–10 CV water, 5–10 CV stripping buffer, 60 CV 1 M NaOH
(contact time at least 1 h), 5–10 CV wash II buffer and 5–10
CV 20% ethanol (protocol can be found in the manufacturer’s
manual). Stripping Buffer: 20 mM NaH2PO4, 0.5 M NaCl,
and 50 mM EDTA at pH 7.4. To 200 mL of distilled water add
0.69 g of NaH2PO4, 7.305 g of NaCl, and 3.653 g of EDTA.
Adjust to pH 7.4 with NaOH and make up to volume to
250 mL with distilled water. The EDTA will not dissolve
until the pH is adjusted.
17. We recommend use of Hybond™-C Extra nitrocellulose mem-
brane (Amersham Biosciences, UK) for better results.
18. A 10 mL volume of primary antibody solution in PBS-M is
generally enough to completely cover a membrane in a tray
11  14 cm. If antibody is scarce then a heat-sealable bag,
the size of the membrane, may be used to reduce the volume of
the primary antibody solution required to completely cover the
membrane. The aim is to cover the membrane with solution at
all times without air bubbles crossing the surface. In the exper-
iment described in this chapter, both primary antibodies are
added for simultaneous detection, which requires a detection
system that can detect two different secondary signals at once
(e.g., fluorescence). If this is not available, the primary anti-
bodies should be added separately.
19. There are different detection systems that can be used for
Western blotting. For example, horseradish peroxidase,
HRP-conjugated secondary antibodies may be used, followed
by an enhanced chemiluminescence (ECL) detection kit. The
membrane can then be exposed to X-ray film in the dark and
then developed. The disadvantage is that only one primary
antibody can be used at a time, so either two gels for each
sample could be run, or the membrane stripped and reprobed.
20. Different fractions may be pooled according to downstream
application or desired ratio of glycosylated to unglycosylated
protein. As the sample will be further purified by ion exchange
chromatography, it is preferable to pool as many fractions as
 Production of Vaccines Using Biological Conjugation 297

possible that contain visible polymerized glycan (i.e., visible

green banding above the protein). Where a shift in protein is
seen this denotes addition of smaller repeats added at either
one or two of the available glycosylation sites. With reference
to Fig. 2b, fractions A10-B2 would be pooled.
21. If the sample volume exceeds the maximum fill volume for the
column then it may be added after the initial spin has reduced
the volume, however, it is preferable to use a column with a
larger capacity to start with. The sample will remain in the
concentrator filter so the eluate may be discarded or kept for
analysis as desired. A Vivaspin column at 4
C will concentrate
very slowly. If concentration time exceeds 3 h, the temperature
can be increased from 4
C to 10
C. If the protein concentra-
tion is high as determined by Coomassie staining of the frac-
tions before concentration, then a larger capacity concentrator
column may be used and concentration stopped at a higher
volume. If the sample is over concentrated then aggregation
and precipitation of the protein may occur.
22. If the sample volume is less than 2.5 mL, first add the sample
(noting the volume) to the column and wait until it has entered
the resin bed, then add an amount of equilibration buffer so
that the total volume of sample plus buffer added to the
column is 2.5 mL. This avoids diluting the sample.
23. The PD-10 columns can be reused several times but it is
recommended to use them with the same sample to avoid
cross-contamination.
24. The endotoxin removal columns can be reused at least five
times. To clean them, regenerate with 8 mL 0.2 N NaOH
overnight. Elute by centrifugation at 500  g for 1 min, then
wash with 8 mL 2 M NaCl, followed by 8 mL water, centrifu-
ging as indicated after each wash. Columns are stored at 4
C,
after adding 8 mL 20% ethanol to the column and plugging the
column with the provided caps for the top and bottom.
25. Choice of ion exchange column is dependent on the pI of the
protein to be purified, use Resource Q (anion) for proteins
with a pI of 1–11 and Resource S (cation) for proteins with a pI
of 3–14, according to the manufacturer’s instructions. Addi-
tionally, the pH of the start and elution buffers will vary
depending on protein. In the case of AcrA, the pI is 7, so
Resource S was used with buffers at pH 6. AcrA was not stable
and precipitated in Resource Q buffer at pH 8. As a starting
point, Resource S buffers are used at 1 pH unit below the pI of
the protein and Resource Q buffers at 1 pH unit above the pI
of the protein. From the starting point, the pH can be further
optimized to achieve the best yield versus purity of your sam-
ple, as appropriate.
298 Emily J. Kay and Vanessa S. Terra

26. The maximum pressure for the Resource S column is 1.5 MPa,
therefore flow rates should be adjusted so as not to exceed this
pressure.
27. The Resource S column is cleaned with 5–10 CV each of 2 M
NaCl, 1 M NaOH, 2 M NaCl, distilled water, start buffer,
distilled water, and 20% ethanol–0.2 M sodium acetate. If the
column is blocked, as indicated by no flow through the column
at the maximum pressure tolerance for the column, then inject
with 1 mg/mL pepsin in a solution of 0.5 M NaCl and 0.1 M
acetic acid, and leave overnight before flushing through with
5–10 CV of distilled water and cleaning as above. Once the
column is in 20% ethanol–0.2 M sodium acetate it may be
stored at 4
C until next use.
28. As with the initial His-purified sample, different fractions may
be pooled according to downstream application or desired
ratio of glycosylated to unglycosylated protein. In this example,
fractions A9–A15 (Fig. 3) were pooled. This was to capture as
much protein as possible that is glycosylated while minimizing
the proportion of contaminating protein (i.e., protein bands
that appear on the Coomassie stained gel that are absent from
the Western blot).
29. Vivaspin columns can be used for buffer exchange by diluting
the concentrated sample up to the original volume with PBS
and concentrating again. Repeating the process three times. If
preferred, alternative buffer exchange methods may be used
(e.g., dialysis cassettes).
30. Sugar content can be quantified using, for example, hot phe-
nol–sulfuric acid method [28] or high-performance anion-
exchange chromatography with pulsed amperometric detec-
tion (HPAEC–PAD) on a Dionex system [29].
31. Lyophilizing glycoconjugates suspended in PBS will result in
salts being retained in the sample, therefore water should be
used for reconstitution. Buffer exchange into 50 mM ammo-
nium bicarbonate buffer may be preferable, because it sublimes
into ammonia, water, and carbon dioxide upon lyophilization,
often without causing measurable harm to the protein. A
50 mM ammonium bicarbonate buffer, pH 7.8, is made by
dissolving 4 g NH4HCO3 in 800 mL distilled water, then
making the final volume up to 1 L. If ammonium bicarbonate
buffer is used then the sample should be reconstituted in PBS.
If a lyophilizer is not available, then the purified glycoconjugate
may be stored at 80
C, but freeze–thaw cycles should be
avoided.
 Production of Vaccines Using Biological Conjugation 299

References
1. Parkhill J, Wren BW, Mungall K et al (2000) 10. Pan C, Sun P, Liu B et al (2016) Biosynthesis of
The genome sequence of the food-borne path- conjugate vaccines using an O-linked glycosyl-
ogen Campylobacter jejuni reveals hypervari- ation system. mBio 7(2):e00443–e00416.
able sequences. Nature 403(6770):665–668. https://doi.org/10.1128/mBio.00443-16
https://doi.org/10.1038/35001088 11. Harding CM, Nasr MA, Scott NE et al (2019)
2. Wacker M, Linton D, Hitchen PG et al (2002) A platform for glycoengineering a polyvalent
N-linked glycosylation in Campylobacter jejuni pneumococcal bioconjugate vaccine using
and its functional transfer into E. coli. Science E. coli as a host. Nat Commun 10(1):891.
298(5599):1790–1793. https://doi.org/10. https://doi.org/10.1038/s41467-019-
1126/science.298.5599.1790 08869-9
3. Fisher AC, Haitjema CH, Guarino C et al 12. Naegeli A, Neupert C, Fan Y-Y et al (2014)
(2011) Production of secretory and extracellu- Molecular analysis of an alternative
lar N-linked glycoproteins in Escherichia coli. N-glycosylation machinery by functional trans-
Appl Environ Microbiol 77(3):871–881. fer from Actinobacillus pleuropneumoniae to
https://doi.org/10.1128/aem.01901-10 Escherichia coli. J Biol Chem 289
4. Yates LE, Natarajan A, Li M et al (2019) (4):2170–2179. https://doi.org/10.1074/
Glyco-recoded Escherichia coli: jbc.M113.524462
Recombineering-based genome editing of 13. Kowarik M, Young NM, Numao S et al (2006)
native polysaccharide biosynthesis gene clus- Definition of the bacterial N-glycosylation site
ters. Metab Eng 53:59–68. https://doi.org/ consensus sequence. EMBO J 25
10.1016/j.ymben.2019.02.002 (9):1957–1966. https://doi.org/10.1038/sj.
5. Strutton B, Jaffé SRP, Pandhal J, Wright PC emboj.7601087
(2018) Producing a glycosylating Escherichia 14. Georgiou G, Segatori L (2005) Preparative
coli cell factory: the placement of the bacterial expression of secreted proteins in bacteria: sta-
oligosaccharyl transferase pglB onto the tus report and future prospects. Curr Opin
genome. Biochem Biophys Res Commun 495 Biotechnol 16(5):538–545. https://doi.org/
(1):686–692. https://doi.org/10.1016/j. 10.1016/j.copbio.2005.07.008
bbrc.2017.11.023 15. Cuccui J, Wren B (2015) Hijacking bacterial
6. K€ampf MM, Braun M, Sirena D et al (2015) In glycosylation for the production of glycocon-
vivo production of a novel glycoconjugate vac- jugates, from vaccines to humanised glycopro-
cine against Shigella flexneri 2a in recombinant teins. J Pharm Pharmacol 67(3):338–350.
Escherichia coli: identification of stimulating https://doi.org/10.1111/jphp.12321
factors for in vivo glycosylation. Microb Cell 16. Herbert JA, Kay EJ, Faustini SE et al (2018)
Factories 14(1):12. https://doi.org/10. Production and efficacy of a low-cost recombi-
1186/s12934-015-0195-7 nant pneumococcal protein polysaccharide
7. Caillava AJ, Ortiz GE, Melli LJ et al (2019) conjugate vaccine. Vaccine 36
Improving bioreactor production of a recom- (26):3809–3819. https://doi.org/10.1016/j.
binant glycoprotein in Escherichia coli: effect of vaccine.2018.05.036
specific growth rate on protein glycosylation 17. Feldman MF, Wacker M, Hernandez M et al
and specific productivity. Biotechnol Bioeng (2005) Engineering N-linked protein glycosyl-
116(6):1427–1438. https://doi.org/10. ation with diverse O antigen lipopolysaccharide
1002/bit.26953 structures in Escherichia coli. Proc Natl Acad
8. Ihssen J, Haas J, Kowarik M et al (2015) Sci U S A 102(8):3016–3021. https://doi.
Increased efficiency of Campylobacter jejuni org/10.1073/pnas.0500044102
N-oligosaccharyltransferase PglB by structure- 18. Sutcliffe JG (1978) Nucleotide sequence of the
guided engineering. Open Biol 5 ampicillin resistance gene of Escherichia coli
(4):140227–140227. https://doi.org/10. plasmid pBR322. Proc Natl Acad Sci U S A
1098/rsob.140227 75(8):3737–3741. https://doi.org/10.1073/
9. Faridmoayer A, Fentabil MA, Haurat MF et al pnas.75.8.3737
(2008) Extreme substrate promiscuity of the 19. Kay EJ, Yates LE, Terra VS et al (2016)
Neisseria oligosaccharyl transferase involved in Recombinant expression of Streptococcus pneu-
protein O-glycosylation. J Biol Chem 283 moniae capsular polysaccharides in Escherichia
(50):34596–34604. https://doi.org/10. coli. Open Biol 6(4):150243. https://doi.org/
1074/jbc.M807113200 10.1098/rsob.150243
300 Emily J. Kay and Vanessa S. Terra

20. Kovach ME, Elzer PH, Hill DS et al (1995) an inhibitory peptide. Sci Rep 8(1):16297.
Four new derivatives of the broad-host-range https://doi.org/10.1038/s41598-018-
cloning vector pBBR1MCS, carrying different 34534-0
antibiotic-resistance cassettes. Gene 166 26. Garcia-Quintanilla F, Iwashkiw JA, Price NL
(1):175–176. https://doi.org/10.1016/ et al (2014) Production of a recombinant vac-
0378-1119(95)00584-1 cine candidate against Burkholderia pseudomal-
21. Bachmann BJ (1996) Derivations and geno- lei exploiting the bacterial N-glycosylation
types of some mutant derivatives of Escherichia machinery. Front Microbiol 5:381. https://
coli K-12. In: Neidhardt FC (ed) Escherichia doi.org/10.3389/fmicb.2014.00381
coli and Salmonella: cellular and molecular 27. Zimmerman T, Petit Frère C, Satzger M et al
biology, 2nd Ed edn. ASM Press, (2006) Simultaneous metal chelate affinity
Washington, DC, pp 2460–2488 purification and endotoxin clearance of recom-
22. Thomas CM (2014) Plasmid incompatibility. binant antibody fragments. J Immunol Meth-
In: Bell E (ed) Molecular life sciences: an ency- ods 314(1):67–73. https://doi.org/10.1016/
clopedic reference. Springer New York, j.jim.2006.05.012
New York, NY, pp 1–3. https://doi.org/10. 28. Masuko T, Minami A, Iwasaki N et al (2005)
1007/978-1-4614-6436-5_565-2 Carbohydrate analysis by a phenol–sulfuric acid
23. Carattoli A (2009) Resistance plasmid families method in microplate format. Anal Biochem
in Enterobacteriaceae. Antimicrob Agents Che- 339(1):69–72. https://doi.org/10.1016/j.
mother 53(6):2227–2238. https://doi.org/ ab.2004.12.001
10.1128/AAC.01707-08 29. Talaga P, Vialle S, Moreau M (2002) Develop-
24. Lizak C, Gerber S, Numao S et al (2011) X-ray ment of a high-performance anion-exchange
structure of a bacterial oligosaccharyltransfer- chromatography with pulsed-amperometric
ase. Nature 474(7351):350–355. https://doi. detection based quantification assay for pneu-
org/10.1038/nature10151 mococcal polysaccharides and conjugates. Vac-
25. Napiórkowska M, Boilevin J, Darbre T et al cine 20(19):2474–2484. https://doi.org/10.
(2018) Structure of bacterial oligosaccharyl- 1016/S0264-410X(02)00183-4
transferase PglB bound to a reactive LLO and
 Chapter 16

Immunological Assessment of Lung Responses

to Inhalational Lipoprotein Vaccines Against Bacterial
Pathogens
Anneliese S. Ashhurst, Cameron C. Hanna, Richard J. Payne,
and Warwick J. Britton

Abstract
Lipopeptides or lipoproteins show potential as safe and effective subunit vaccines for protection against
bacterial pathogens. Provided suitable adjuvants are selected, such as the TLR2-stimulating molecules
Pam2Cys and Pam3Cys, these may be formulated as inhalational vaccines to optimize localized pulmonary
immune responses. Here, we present methods to assess antigen-specific memory lymphocyte responses to
novel vaccines, with a focus on immune responses in the lung tissue and bronchoalveolar space. We describe
detection of T-cell responses via leukocyte restimulation, followed by intracellular cytokine staining and
flow cytometry, enzyme-linked immunosorbent spot assay (ELISpot), and sustained leukocyte restimula-
tion for detection of antigen-specific memory responses. We also detail assessment of antibody responses to
vaccine antigens, via enzyme-linked immunosorbent assay (ELISA)-based detection. These methods are
suitable for testing a wide range of pulmonary vaccines.

Key words Tuberculosis, Inhalational vaccination, Lungs, Bronchoalveolar lavage, Lipoprotein, ELI-
Spot, ELISA, Intracellular cytokine staining, Flow cytometry, Leukocyte restimulation

1 Introduction

The concept of vaccination has been one of the most transformative

and impactful innovations in the history of medicine. Vaccines have
eliminated the devastating smallpox virus from our communities
and have drastically reduced the incidence of many other bacterial
and viral diseases that previously caused mass morbidity and mor-
tality. The global pandemic caused by the novel severe acute respi-
ratory syndrome coronavirus 2 (SARS-CoV-2), that emerged in
late 2019, has for many brought into perspective our critical reli-
ance on vaccines to control infectious disease [1]. There remain,
however, many infectious diseases for which we do not yet have safe
and highly effective vaccines. This includes the bacterial pathogen

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_16,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

301
302 Anneliese S. Ashhurst et al.

Mycobacterium tuberculosis, the etiological agent of tuberculosis

(TB), which remains the single biggest infectious disease killer
throughout history [2].
To date, there is only one licensed vaccine for TB, Mycobacte-
rium bovis Bacille Calmette-Guérin (BCG), which has been widely
used in TB control programs since 1921. While it has proven
efficacy in reducing severe childhood forms of TB, including extra-
pulmonary manifestations such as TB meningitis, it fails to provide
reliable protection against pulmonary TB in adolescents and adults.
In addition, as a live attenuated mycobacterial vaccine, BCG is not
suitable for use in immunocompromised individuals, limiting its
use [3, 4]. There is an urgent need to continue vaccine develop-
ment to provide safe and more effective immunizations against this
dangerous pathogen.
One promising approach that offers improved safety compared
to live attenuated or viral-vectored vaccines is subunit vaccination,
including vaccines that utilize antigenic peptide epitopes or pro-
teins as antigens. These require coadministration of an immune-
activating adjuvant in order to provoke a sufficient protective
immune response. In order to optimize the effect of antigen–
adjuvant coadministration, one strategy involves covalent conjuga-
tion of these components to form a self-adjuvanting vaccine. Con-
ceptually, a self-adjuvanting vaccine optimizes the timing of
antigen–adjuvant delivery to antigen presenting cells (APCs). This
is beneficial for providing APCs the requisite secondary costimula-
tion at the same moment they are receiving and processing antigen
for presentation to T-cells. Moreover, this strategy is thought to
enhance phagocytic uptake of vaccine antigens, as pattern recogni-
tion receptor (PRR) ligands can induce receptor-mediated phago-
cytosis [5]. This strategy has been demonstrated extensively by us
and others for indications such as cancer [6–12], allergy [13], and
bacterial infections [14–17]. For example, Rai et al. reported the
use of a multistage biepitope vaccine conjugated to the TLR2
agonist, Pam2Cys, which elicited significant CD4+ and CD8+
T-cell immunity and reduced M. tuberculosis burden in mice to a
greater extent than the BCG vaccine [16].
We have recently reported the synthesis, immunological and
efficacy testing of lipopeptide and lipoprotein-based self-adjuvant-
ing vaccines, containing M. tuberculosis proteins or T cell epitopes,
in preclinical murine models. The antigens were synthetically pro-
duced by solid-phase peptide synthesis and conjugated covalently
to the TLR2-targeting adjuvants Pam2Cys and Pam3Cys using a
peptide ligation strategy, providing high purity self-adjuvanted vac-
cine constructs [18, 19]. We opted for a totally synthetic strategy as
opposed to recombinant protein expression in order to avoid con-
taminants and enable access to highly pure products. Moreover,
most adjuvants are hydrophobic which creates difficulties when
fusing them to highly hydrophilic peptides and proteins. A chemical
Immunological Assessment of Inhalational Lipoprotein Vaccines 303

synthesis approach can alleviate these issues as conjugation of the

adjuvant to the peptide can take place in organic solvent.
We have also demonstrated immunological advantages of deliv-
ering vaccines such as these to the pulmonary mucosa via inhalation
to provide a greater degree of localized protection [18, 20]. Grow-
ing evidence indicates that pulmonary or mucosal vaccination stra-
tegies will enhance localized lung immune responses against
M. tuberculosis and other respiratory bacterial or viral infections
[15, 21–24], and this is a rapidly developing field of vaccine
research. We encourage investigators to explore different routes
of immunization when developing novel vaccines, and consider
matching the route of immunization to the route of transmission
of the pathogen. However, not all types of vaccines are suitable for
delivery by all routes, so careful selection is required. Lipopeptide
or protein-based vaccines may be particularly suitable for mucosal
delivery, as they do not carry the risks associated with live vaccines
and can be used in individuals with compromised immune systems.
Importantly, unlike live or viral vectored vaccines, protein-based
vaccines may be given repeatedly to boost immunity and may also
be readily formulated as dry powders to facilitate delivery as an
inhalational vaccination [20].
In this chapter, we detail commonly used methods to perform
assessment of vaccine-induced peptide or protein-specific memory
immune responses, focusing on responses in the lung. These must
be examined in an experimental animal model, with mice being the
most frequently used. As an exemplar, we present assessment of
peptide or protein-specific responses in murine models of pulmo-
nary vaccination against M. tuberculosis. The materials and methods
may, however, be adapted to assess the immune response against
any peptide or protein vaccine antigen of interest. We recommend
analyzing both localized and systemic antigen-specific immune
responses to vaccination, that is, the immune response at the site
of immunization, as well as the immune response in blood and
secondary or peripheral lymphoid organs. Isolation of leukocytes
from nonlymphoid tissue presents more difficulty than the standard
methods used to process blood and lymphoid tissue. Therefore, we
present here methods to isolate cells from bronchoalveolar lavage
(BAL) and lung tissue to assess localized responses to inhalational
vaccines.
To understand the antigen-specific memory immune response
induced by a vaccine, it is of value to assess both cell-mediated and
humoral responses. For intracellular bacterial pathogens such as
M. tuberculosis, the critical protective role of T-cells is well estab-
lished [25] and assessing these is the focus of the protocols detailed
here. We describe detection of vaccine-induced antigen-specific
T-cell responses via several methods: leukocyte restimulation fol-
lowed by intracellular cytokine staining (ICS) and flow cytometry,
enzyme-linked immunosorbent spot (ELISpot) assay, and
304 Anneliese S. Ashhurst et al.

sustained leukocyte restimulation for detection of antigen-specific

memory responses. The different methods provide a means to
obtain a wide range of data on the T-cell responses to vaccine
antigens, and also take into consideration that the investigator
may be limited by the equipment or reagents available. Consider-
ation should be given to which method is most suitable, and a
summary of the major reportable outcomes, advantages, and dis-
advantages of the techniques is provided in Table 1. While not
necessarily critical for defense against M. tuberculosis, the role of
vaccine-induced humoral immunity is of great importance for
defenses against other bacterial pathogens. Therefore, we also detail
assessment of antibody responses to vaccine antigens, using
isotype-specific enzyme-linked immunosorbent assays (ELISA).
Finally, while assessment of the immune response generated by
a vaccine is required to progress a vaccine toward clinical develop-
ment, ultimately, it is critical to determine whether a vaccine can
prevent or limit infection with the pathogen of interest in a suitable
preclinical model. This is particularly the case if the immune corre-
lates of protection for a particular pathogen are not well-defined, as
is the case against M. tuberculosis [26]. While providing detailed
methods for efficacy testing is beyond the scope of this chapter, it is
important to identify immune responses correlating to protection
in preclinical animal models by performing similar immunological
assessment of the response generated by the vaccine prior to and
postinfectious challenge.

2 Materials

Prepare all solutions using ultrapure water and store at room tem-
perature (~22  C), unless indicated otherwise. Standard cell culture
consumables and plasticware, including 70 μm filters, should be
sterile. Follow all institutional biosafety and chemical hazard reg-
ulations, including when disposing of waste.

2.1 Leukocyte All reagents should be sterile. Saline buffer may be autoclaved,
Preparation otherwise filter buffers or stock solutions through a nonpyrogenic
0.22 μm membrane filter.
1. 1 phosphate buffered saline (PBS), pH 7.4.
2. Fluorescence activated cell sorting (FACS) buffer (keep at
4  C): PBS, 2% v/v fetal bovine serum (FBS) or fetal calf
serum (FCS), 2 mM ethylenediaminetetraacetic acid (EDTA).
3. Complete cell culture media (keep at 4  C): Roswell Park
Memorial Institute (RPMI) 1640 Medium (containing L-glu-
tamine), 10% FBS or FCS, 0.05 mM 2-Mercaptoethanol,
100 U/ml penicillin–streptomycin.
4. PBS/Heparin (keep at 4  C): PBS, 20 U/ml Heparin.
 Immunological Assessment of Inhalational Lipoprotein Vaccines 305

Table 1
Comparison of key methods for assessing T-cell cytokine responses to peptide or protein vaccine
antigens

Major reportable
Method outcome(s) Key advantages Disadvantages
Leukocyte Proportion of antigen- Flexibility—Simultaneous Cannot provide
restimulation, ICS responsive cytokine single-cell level quantitation of the
and flow positive T-cells, detection of multiple cytokines released
cytometry including single-cell different phenotypes Requires access to a
level analysis of High sensitivity—Large multiparameter flow
multifunctionality and numbers of cells can be cytometer and analysis
phenotype analyzed to allow software
detection of rare Depending on
antigen-responsive experiment design,
populations may be costly
ELISpot Enumeration of antigen- High sensitivity Limited to detection of
specific cells capable of Cost-effective one cytokine only
releasing the detected High-throughput (unless variant
cytokine methods such as dual-
color [40] or
FluoroSpot [41] are
utilized)
Cannot determine the
phenotype of cells
releasing the detected
cytokine
Sustained leukocyte Quantitative detection of High-sensitivity, Takes longer
restimulation— cytokine released particularly useful for Cannot determine the
Recall of memory Can measure multiple detection of memory phenotype of cells
antigen-specific cytokines if multiple T-cell responses (not releasing the detected
T-cell responses ELISAs are performed recently activated) cytokine
on cell supernatants, or Cost-effective Assessment of cellular
if multiplex methods High-throughput proliferation by
such as cytokine bead Cell supernatants can be radiological or
array are used stored frozen and fluorescent methods
Can also examine antigen- cytokines quantitated may be costly
specific cell proliferation when convenient
in the same assay, either Offers several ways to
simply by enumerating determine cellular
cells, or by radioisotope proliferation, which
or fluorescent methods can be selected based
on the resources
available

5. Ammonium-Chloride-Potassium (ACK) lysis buffer: 150 mM

ammonium chloride (NH4Cl), 1 mM potassium hydrogen
carbonate (KHCO3), 100 mM EDTA (pH 7.3).
6. Tissue digestive enzymes, collagenase IV and DNase I, in PBS.
7. Trypan blue 0.01% w/v in 1 PBS.
306 Anneliese S. Ashhurst et al.

2.2 Antigen 1. Nonspecific T-cell stimulant, such as anti-CD3 and anti-CD28

Restimulation monoclonal antibodies.
of Leukocyte 2. Purified peptide or protein vaccine antigens of interest.
Suspensions and Flow
3. Brefeldin A.
Cytometry
4. Fluorescently labeled monoclonal antibodies for cell surface or
intracellular immunostaining.
5. Intracellular staining buffer kit, such as BD Cytofix/Cyto-
perm™ Fixation/Permeabilization Kit.

2.3 ELISpot 1. Nonspecific T-cell stimulant, such as Concanavalin A.

2. ELISpot plate (PVDF membrane preferable).
3. Cytokine-specific primary coating antibody.
4. Complementary cytokine-specific biotinylated secondary
antibody.
5. PBS-Tween (PBST): 1 PBS, 0.1% v/v Tween 20.
6. PBS/bovine serum albumin (BSA): 1 PBS, 0.5% w/v BSA.
7. Avidin Alkaline Phosphatase (AAP) enzyme conjugate.
8. Alkaline Phosphatase Substrate.

2.4 ELISA 1. Purified peptide or protein vaccine antigens of interest.

2. Carbonate/bicarbonate coating buffer 50 mM, pH 9.6.
3. Blocking buffer (store 4  C): PBS, 1% w/v BSA.
4. Wash buffer: PBS, 0.05% v/v Tween 20.
5. Detection antibody (e.g., anti-mouse IgA), horseradish perox-
idase (HRP)-conjugated (optimize concentration for each
detection antibody).
6. Tetramethylbenzidine (TMB) substrate.
7. Stop solution: 2 M HCl.

3 Methods

Carry out all procedures at room temperature (RT; ~22  C), unless
indicated otherwise.

3.1 Isolation of BAL Leukocytes should be collected and processed using aseptic tech-
and Lung Leukocytes niques, with sterile reagents and equipment. Keep tissue or cells at
from Vaccinated Mice 4  C or on wet ice as much as is practical, to slow down metabolic
processes that alter cell phenotype and decrease the rate of cell
death. Keep FACS buffer and complete cell culture media at 4  C.

1. Euthanize the mouse as per the requirements of your institu-

tional animal ethics committee. Avoid cervical dislocation, as
 Immunological Assessment of Inhalational Lipoprotein Vaccines 307

3.1.1 Bronchoalveolar this is likely to cause injury to the trachea, chest cavity, and
Lavage Collection and Cell blood vessels, resulting in difficulty performing the lavage.
Isolation 2. Immediately after confirmation of euthanasia, spray 70% etha-
nol on the neck to sterilize and make an incision in the skin.
Incise away the muscles to expose the trachea. Use two pairs of
curved forceps to gently tease away the membrane covering the
trachea.
3. Slide one pair of curved forceps under the trachea, to provide
stability during intubation.
4. Very carefully, with the bevel facing upward, punch a hole in
one side of the trachea with a 21 GA needle.
5. Insert a ~1.8 in. 20 GA flexible cannula, attached to a 1 ml
syringe containing 1 ml sterile PBS, into the hole created by the
needle to 0.5–1.5 cm depth. Do not exceed this, as it may
damage the lungs. With a second pair of fine forceps, pinch
the trachea to seal the cannula firmly in place. Remove the pair
of forceps under the trachea.
6. Perform lavage: slowly inject the 1 ml PBS into the lungs—you
should see the lungs inflate and the chest expand as the lungs
fill with PBS. There should be no leakage of PBS from the
trachea or lungs. If this occurs, it is likely the seal with the
forceps on the trachea is not sufficiently firm, or, there has been
a rupture in the trachea.
7. Pull back on the syringe to collect as much fluid from the lungs
as possible, usually 700–900 μl. Dispense into an Eppendorf or
Falcon tube. In a healthy naı̈ve mouse, there should be no
obvious evidence of blood in the bronchoalveolar lavage fluid
(BALF); however, this may occur in mice with inflammation or
infection in the lungs.
8. Centrifuge the BAL for 5 min (300–500  g, 4  C). Collect the
supernatant—this is the BALF and can be used in assays to
quantitate the presence of cytokines/chemokines or antibodies
in the airways by ELISA. If not used immediately, this should
be stored at 80  C to 30  C. It may be beneficial to add
protease inhibitors to prevent degradation of protein analytes.
9. Resuspend the cells in ~200 μl FACS buffer or complete cell
culture media by gently pipetting up and down. Count viable
cells by trypan blue exclusion. The number of cells will be
variable, but in a healthy mouse expect ~1  105 or less, and
up to 1  106 following pulmonary immunization or if inflam-
mation or infection is present.

3.1.2 Lung Leukocytes 1. Euthanize the mouse as per the requirements of your institu-
tional animal ethics committee, avoiding cervical dislocation.
308 Anneliese S. Ashhurst et al.

2. Immediately after confirmation of euthanasia, expose the lungs

and heart, using blunt edged scissors to cut away the rib cage
and surrounding tissue, being careful not to pierce the lungs.
3. If you wish to remove circulating blood from the lung tissue,
immediately perfuse. Holding the heart with tweezers, inject
into the apex or right atrium 10 ml PBS/Heparin. The lung
should inflate and clear of blood until they are white. If lungs
do not inflate, try pulling back on the needle a little to make
sure you are not blocking the major blood vessels between the
lung and heart. If you pierce the lungs while removing the rib
cage, you will have difficulty getting the lungs to inflate.
4. Remove the lungs in pieces and place in 1.5 ml complete cell
culture media in a 24-well tissue culture plate or 5 ml flat
bottom tube.
5. Using a pair of fine pointed small scissors, chop the lungs into
small pieces (1–2 mm in diameter). In between samples, dip the
scissors in 70% ethanol, then sterile PBS, to prevent cross-
contamination. Alternatively, use a scalpel and a Petri dish to
slice-up the lung. The aim is to increase the surface area of the
tissue for digestive enzymes.
6. Once the lung is chopped, place the lung pieces and associated
media into a tube using a sterile transfer pipette. Use one
pipette per experimental group. Add RPMI only to make up
to a total volume of 5 ml.
7. Add enzymes from sterile stock solutions to achieve a final
concentration of 50 U/ml collagenase and 13 μg/ml DNAse I.
8. Incubate for 30–45 min at 37  C on a gentle shaker, rocker or
rotating wheel. Alternatively use a 37  C water bath and gently
invert samples every 10–15 min to mix.
9. Pass media and tissue (using a 5 ml syringe plunger or similar)
through a 70 μm mesh sieve to disperse clumps and rupture
larger pieces of tissue, into a new V-bottom 10 or 15 ml tube.
Use 1 sieve and syringe for every 2 to 3 samples, changing
between groups. Wash through thoroughly with complete
media.
10. Centrifuge for 10 min (500  g, 4  C). Discard supernatant.
Whenever discarding supernatants, do not invert tube multiple
times, as this will resuspend the cell pellet and result in inadver-
tent loss of cells.
11. If all the blood was removed by perfusion, the cell pellet will be
white. If blood remains, lyse erythrocytes with ACK lysis
buffer. Resuspend the pellet by flicking the tube, then add
~2 ml ACK lysis buffer for 1–2 min (this time will need to be
optimized for each batch of ACK lysis buffer—do not incubate
for longer than necessary to lyse red blood cells, as this will
 Immunological Assessment of Inhalational Lipoprotein Vaccines 309

increase cell death). Dilute with cold (4  C) PBS to at least five

times the volume to stop lysis.
12. Filter again through a 70 μm sieve—one sieve and syringe per
experimental group. Spin cells down for 5 min (500  g, 4  C).
13. Resuspend pellet by flicking tube. Add 500 μl complete cell
culture media and then resuspend cells by gently pipetting up
and down.
14. Count viable cells by trypan blue exclusion. A normal lung will
yield between 1 and 3 million cells; inflamed or infected lungs
will yield between 30 and 40 million cells.

3.2 Assessment Flow cytometric methods provide enormous scope to determine

of Antigen-Specific the pattern of T-cell responses to the vaccine and the complex
T-cell Responses Via phenotypes of T-cells, at a single-cell level. For instance, cell surface
Ex vivo Antigen immunostaining can be utilized to enumerate vaccine-specific α/β
Restimulation, TcR CD4+ and CD8+ T-cells, using epitope-specific MHC Class II
Intracellular Cytokine and I-tetramer complexes [27, 24]. Their location in the lung
Immunostaining vasculature or parenchyma may be determined by prior intravascu-
and Flow Cytometry lar labeling. This can be achieved by intravenous injection of 5 μg
anti-CD45-APC/Cy7 (or another fluorophore), 3–5 min prior to
euthanasia, to distinguish labeled vascular leukocytes from unla-
beled lung parenchymal cells [28–31]. Lung resident memory T
cells stimulated by pulmonary vaccines can also be identified within
the lung parenchyma by cell surface markers, which may include
CD69, CD103, CXCR3, and CXCR6 [31, 32]. Other T-cell popu-
lations may be stimulated by the vaccine adjuvant, such as donor-
unrestricted T cells including γ/δ TcR T cells and MAIT cells [33],
and these may also be enumerated with fluorochrome-labeled spe-
cific mAbs. Phenotyping of these various T-cell subtypes should be
performed by immunostaining freshly collected leukocyte suspen-
sions that have been kept at 4  C, or on wet ice, from the time of
collection. The prolonged incubation period at 37  C required to
assess antigen-specific cytokine responses, as in the following pro-
tocol, may change the surface expression of these memory or
phenotypic markers.
To assess CD4+ and CD8+ T-cell cytokine responses specific to
a vaccine peptide or protein antigen, single-cell suspensions of
leukocytes isolated from mice (e.g., from the lungs, blood, spleen,
lymph nodes) must first be restimulated in culture with the antigen.
A reagent that blocks intracellular protein transport processes, such
as Brefeldin A, is added so that cytokine from responding T-cells
accumulates within the Golgi complex. Immunostaining of cell
surface and intracellular markers is then performed with
fluorophore-labeled antibodies, and data are acquired by flow cyto-
metric analysis.
This method provides the most flexibility for detection of
different types of T-cell cytokine responses, including
310 Anneliese S. Ashhurst et al.

multifunctional responses, but is limited to providing the propor-

tion of cytokine positive responsive cells. That is, it cannot provide a
quantitative report of the amount of cytokine released in response
to antigen.
Subheading 3.2.1 of the protocol must be performed under
aseptic conditions to reduce the risk of T-cells responding to con-
taminants; however, the steps from Subheading 3.2.2 onward need
not be sterile.

3.2.1 Restimulation 1. Add 2–4  106 leukocytes (see Note 1) from each sample into a
of Antigen-Specific T-cells well of a 96-well round- or V-bottom plate. Use one well for
in Ex vivo Culture each antigen you wish to restimulate the sample with.
2. Centrifuge the plate to pellet cells (3–5 min, 300–500  g) and
discard the supernatant (see Note 2). Resuspend the cells, by
gently pipetting up and down (see Note 3), in 200 μl per well
complete culture media containing the peptide or protein vac-
cine antigen of interest (5–10 μg/ml is generally suitable). Also
prepare negative and positive control samples (see Note 4).
3. Incubate the plate at 37  C, 5% CO2 to allow the antigen to be
processed for presentation on MHCI/II to T-cells. This incu-
bation time may require optimization. As a guide: for peptide
restimulation 1–2 h; for protein restimulation, 3–4 h.
4. Add Brefeldin A to each well (10 μg/ml; this will block trans-
port of proteins from the Golgi) and pipette gently to resus-
pend the cells. Incubate (humidified) at 37  C, 5% CO2 for
4–6 h to allow intracellular accumulation of the cytokines (see
Note 5).

3.2.2 Immunostaining The number and type of cell surface or intracellular markers
Cells for Detection selected for immunostaining will vary greatly depending on the
of Intracellular Cytokine experimental design. In addition, the monoclonal antibody panels
will vary between laboratories, accounting for the configuration of
the cytometer available for data acquisition, as well as more practical
considerations such as antibody stocks already within the laboratory
and the budget available. We recommend the following papers to
assist in the critical steps of selecting antibody combinations for use
in flow cytometry, optimization of antibody dilutions and tips for
data acquisition on the cytometer [34, 35]. The antibody clones
and fluorophores utilized to generate the sample data in Fig. 1 are
included in Note 6, but this is by no means restrictive.
From this step onward, keep the cells on wet ice or at 4  C to
prevent any further metabolic changes or increasing cell death. As
vaccine experiments often involve testing many samples, to improve
efficiency and consistency, we strongly recommend performing cell
staining in a 96-well plate, rather than staining cells in individual
FACS tubes.
 Immunological Assessment of Inhalational Lipoprotein Vaccines 311

Fig. 1 Gating strategy for analysis of intracellular stained flow cytometry data, to quantitate the proportion of
antigen-specific cytokine producing T-cells and determine polyfunctional T-lymphocyte responses to immu-
nization. Debris and dead cells are excluded, CD4+ or CD8+ T-cells selected, cytokine positive cells gated,
followed by Boolean gating to enumerate the frequency of polyfunctional populations. A representative
immunostained lung sample from a Pam2Cys-ESAT61–20-TB10.43–11 intranasally immunized mouse is
shown after ex vivo stimulation in the presence of Brefeldin A (10μg/ml) with (a) ESAT61–20 peptide (10μg/
ml) recall, (b) TB10.43–11 peptide (10μg/ml) recall and (c) no peptide recall, used as a negative control to set
gates for cytokine producing cells
312 Anneliese S. Ashhurst et al.

1. Centrifuge the cells to pellet (3–5 min, 300–500  g), and

discard supernatant as before. Add 200 μl cold (4  C) PBS to
each well and pipette gently to resuspend, washing the cells,
and centrifuge as before.
2. To allow identification of live cells, stain dead cells with a fixable
cell viability dye (e.g., LIVE/DEAD Fixable Blue Dead Cell
Stain Kit, for UV excitation, available from Life Tech), diluted
as per the manufacturer’s instructions (see Note 7). Also
include Fc blocking reagent (e.g., Mouse BD Fc Block—pur-
ified rat anti-mouse CD16/CD32, clone 2.4G2, at 1:100
dilution) to prevent nonspecific binding of antibodies used in
the immunostaining steps to Fc receptors on leukocytes.
Pipette gently to resuspend cells in 50 μl of a stock solution
(i.e., for consistency, prepare enough staining stock to provide
sufficient volume for every sample). Incubate on ice for
20–30 min, or with gentle agitation on a plate shaker, at 4  C.
3. To halt staining with the viability dye, add 200 μl cold FACS
wash directly into each well, then centrifuge the cells to pellet
(3–5 min, 300–500  g).
4. To perform immunostaining of cell surface markers, prepare a
stock staining solution consisting of FACS buffer and your
selection of appropriately diluted monoclonal antibodies and
gently resuspend cells in 50 μl of the stain. Incubate on ice for
20–30 min, or with gentle agitation on a plate shaker at 4  C, in
the dark as the fluorophores are light sensitive (e.g., wrap plate
in aluminum foil).
5. Remove excess stain by adding 200 μl cold FACS wash directly
into each well, then centrifuge the cells to pellet (3–5 min,
300–500  g). Repeat wash once more.
6. After discarding the wash buffer, gently vortex the plate to
resuspend the cell pellets, and add 100 μl fixative/permeabili-
zation agent (see Note 8). Pipette the cells gently to resuspend
thoroughly then incubate for 20–30 min, at 4  C, in the dark.
7. Centrifuge cells to pellet and discard supernatant appropriately.
Wash cells twice as before, but in 200 μl BD Cytowash (dilute
buffer stock as per the manufacturer’s instructions).
8. Prepare a stock staining solution consisting of BD Cytowash
buffer and your selection of appropriately diluted monoclonal
antibodies to stain intracellular cytokines or markers, and
gently resuspend the cells in 50 μl of the stain. Incubate on
ice for 30 min, or with gentle agitation on a plate shaker at
4  C, in the dark.
9. Wash cells twice with 200 μl BD Cytowash per well as before,
before resuspending cells in FACS wash at desired volume. If
required, resuspend instead with a fixative agent as per the
manufacturer’s instructions.
 Immunological Assessment of Inhalational Lipoprotein Vaccines 313

10. Samples may be stored at 4  C in the dark for up to 48 h

although we recommend performing data acquisition on the
cytometer as soon as possible (see Note 9). Be sure to filter the
sample through a 70 μm mesh into a FACS tube immediately
prior to loading onto the cytometer, to ensure a single cell
suspension and reduce the risk of blockages. Aim to acquire at
least 2  106 events to allow accurate quantitation of small
populations of cytokine producing T-cells.
11. For an example gating strategy to identify cytokine-producing
antigen-specific T-cells, see Fig. 1. See [18, 19] for an example
of the final analysis of cytokine producing T-cell populations,
using additional Boolean gating analysis in FlowJo version
10 (BD). Report the data as the percentage (or proportion)
of CD4+ or CD8+ T-cells producing a particular cytokine, or
combination of cytokines (see Note 10).

3.3 IFNγ ELISpot First described by Czerkinsky in 1983, ELISpot provides a quanti-
tative measure of cells in the sample capable of releasing antibody or
a particular cytokine, specifically in response to a recall antigen [36–
38]. It is widely considered a sensitive and reproducible technique
and is utilized in both research and clinical diagnostic settings
[39]. Similar to plate-based sandwich ELISA, ELISpot utilizes
complementary antibody pairs that recognize different epitopes of
the desired analyte, in this case IFNγ, and combines this with
membrane-based Western blotting techniques. The first of the
antibody pairs is coated onto a polyvinylidene difluoride (PVDF)-
backed microtiter plate, then leukocytes are cultured in the plate
with the vaccine peptide or protein antigen of interest. During the
incubation, if the sample contains antigen-specific cells capable of
releasing IFNγ, this is released and captured in place on the mem-
brane by the coating antibody. Leukocytes are then washed out of
the plate and IFNγ is detected with a biotinylated or directly
enzyme-conjugated complementary antibody. If needed, avidin or
streptavidin conjugated enzyme is bound to the antibody, then
substrate solution is added to form colored spots at the sites of
cytokine secretion on the membrane. These can then be counted to
quantitate the number of spot-forming units per well, and extra-
polated to provide the number of responding cells per million
leukocytes.
While the protocol presented here provides the materials and
method for detection of murine IFNγ, this may be adjusted to
detect other cytokines or antibodies of interest using complemen-
tary monoclonal antibody pairs at optimized concentrations. Mod-
ification of the cell culture conditions and incubation period may
also be required.
Subheadings 3.3.1 and 3.3.2 should be performed under asep-
tic conditions. The subsequent steps need not be sterile.
314 Anneliese S. Ashhurst et al.

3.3.1 Preparation 1. Wet the membrane at the base of the wells of the ELISpot plate
and Coating with 40 μl of 35% ethanol (v/v prepared in milliQ sterile water)
of the ELISpot Plate for less than 1 min (see Note 11). Wash the plate three times by
pipette (see Note 12) with 200 μl sterile PBS per well, adding
the first wash straight on top of the ethanol. Ensure the mem-
branes do not dry out for the remainder of the protocol. Leave
final PBS wash on until ready to commence the next step.
2. Coat each well of the ELISpot plate with the primary antibody,
purified anti-mouse IFNγ antibody (clone AN18) by pipetting
100 μl antibody diluted in sterile PBS into each well. Incubate
overnight at 4  C. Alternatively incubate at room temperature
for a minimum of 2 h.
3. Pipette out the primary antibody solution and wash with 200 μl
sterile PBS three times as before.
4. To ensure proteins do not bind nonspecifically to the mem-
brane, add 200 μl complete cell culture media into each well as
a blocking agent. Incubate at 37  C for 2 h, or at room
temperature (22  C) if for longer.

3.3.2 Plating 1. To restimulate vaccine antigen-specific cells, prepare dilutions

of Leukocytes for Antigen of desired peptide or protein antigen in complete cell culture
Restimulation media. Prepare these, as well as negative and positive control
stimuli, at double the desired final concentration for the cell
culture (see Note 13). If not already sterile, filter-sterilize the
antigen stocks using a nonpyrogenic 0.22 μm syringe filter.
2. Remove the blocking solution from the ELISpot plate by
gentle pipetting, then quickly (do not allow membranes to
dry) pipette 100 μl per well of either negative control media,
diluted antigen, or positive control stimulant.
3. Gently plate 100 μl of a 2  106 per ml leukocyte suspension
prepared in complete cell culture media into each well (final
density of 2  105 leukocytes per well; see Note 14).
4. Incubate (humidified) the plate at 37  C, 5% CO2 for 20 h
(minimum 18 h).

3.3.3 Staining 1. Decant cells and wash plate six times with PBST (0.1% Tween
and Developing 20), 200 μl per well, with a pipette. Leave final wash on while
the ELISpot Plate you make up the secondary antibody solution.
2. Add 100 μl per well biotinylated anti-mouse IFNγ antibody
(clone XMG.12, see Note 15) prepared at an appropriate dilu-
tion in 0.5% BSA (w/v)/PBS and incubate at room tempera-
ture for greater than 2 h, or alternatively leave overnight at
4  C.
3. Decant secondary antibody solution and wash plate six times
with PBST as before. Leave final wash on while you make up
secondary antibody label solution.
 Immunological Assessment of Inhalational Lipoprotein Vaccines 315

4. Add 100 μl of secondary antibody label solution Avidin Alka-

line Phosphatase (Sigma) at 1:1000 (v/v) in 0.5% BSA/PBS
and incubate for 45 min at room temperature.
5. Decant solution and wash plate three times with PBST, then
three times with PBS, as before. Before final PBS wash, remove
the plastic backing of the plate and fill with PBS, rinsing the
base of membranes. Leave final PBS wash on while you prepare
the substrate solution.
6. Add 100 μl of alkaline phosphatase conjugate substrate solu-
tion to every well and allow color development by incubating at
room temperature in the dark (wrap in foil) for ~11 min, or
until spots are just visible. Stop reaction by decanting solution
and immediately washing thoroughly with distilled water.
7. Allow plates to air-dry in the dark at 4  C overnight to allow the
spots to darken. Allow further drying in the dark at room
temperature for a few hours before reading with an ELISpot
plate reader (see Note 16).
8. To report the data, multiply the number of spots per well by
5, to provide the spot forming cells per 106 leukocytes for that
sample.

3.4 Sustained To optimize the sensitivity for detecting antigen-specific T-cell

Leukocyte responses to the vaccine peptide or protein antigen, in particular
Restimulation memory responses, it can be beneficial to perform an extended cell
for Recall of Memory culture incubation period, allowing cytokine released from antigen-
Antigen-Specific T-cell specific T-cells to accumulate in the culture supernatant. This
Responses method, in contrast to ELISpot, allows a quantitative measure of
the amount of cytokine released from the leukocytes in response to
antigen restimulation. During this extended incubation time, it is
expected that antigen-specific T-cells will have proliferated, and
these may also be quantitated to give a measure of the magnitude
of the immune response.
This assay should be performed under aseptic conditions.
1. Prepare single cell leukocyte suspensions in complete cell cul-
ture media at 2  106 cells/ml.
2. In a 96-well round bottom plate, plate in duplicate or triplicate,
100 μl per well of appropriately diluted recall antigen, or nega-
tive (media only) or positive (ConA 3–5 μg/ml) controls, as
per Subheading 3.3.2, step 1. Then add 100 μl per well of the
leukocyte suspension (2  105 cells/well), taking care not to
cross-contaminate between wells. Incubate (humidified) at
37  C, 5% CO2, for 72 h (see Note 17).
3. Centrifuge the plate (300–500  g, 3–5 min) to pellet the cells,
and carefully remove the supernatant by pipette into a new
96-well plate, taking care not to disturb the cell pellet. The
316 Anneliese S. Ashhurst et al.

supernatant can be used immediately, or frozen at 30  C until

required, for detection of cytokine release by ELISA or other
method such as cytokine bead array.
4. Resuspend the cell pellets in complete culture media. Prolifera-
tion of the cells may be determined by quantitating the cells per
well (see Note 18).

3.5 Quantitation The majority of clinically utilized vaccines aim to induce humoral
of Vaccine-Induced immunity against the pathogen of interest, with a level of vaccine-
Antigen-Specific specific antibodies determined that correlates with protective
Antibodies by ELISA immunity. Antibodies generated against the vaccine antigen can
be readily detected in serum samples by straightforward ELISA
methods. As mice may be bled routinely throughout vaccination
experiments, serum may be collected sequentially from individual
animals to monitor changes to the isotype and titer of antibodies
over time. Take care to detect isotypes specific to the experimental
mouse strain, for instance, C57BL/6 mice produce IgG2c instead
of the IgG2a produced by BALB/c. However, vaccine-induced
antibody may also be found in many peripheral tissues and samples
taken at euthanasia, such as BALF as described in Subheading
3.1.1, nasal washes or peritoneal washes, may also be assessed for
the presence of vaccine-induced antibodies. This may be particu-
larly valuable to detect mucosal IgA if vaccination is delivered by
inhalation [19].
Here we present a standard indirect ELISA-based method to
quantitatively determine vaccine antigen-specific antibody titer in
serum or BALF.
1. Coat high-binding 96-well ELISA plates with 100 μl per well
of the vaccine peptide or protein antigen, at 1 μg/ml diluted in
coating buffer. Seal plate and incubate for 2 h at 37  C or
overnight at 4  C.
2. Decant coating solution, then thoroughly wash the plate by
filling wells completely with wash buffer then decanting, 4–6
times. This may be performed using a squirt bottle or an
automated plate washer, to forcefully wash the wells. Tap
plate firmly on absorbent paper toweling to blot.
3. Add 200 μl blocking buffer per well, seal plate and incubate at
37  C for 1 h.
4. While the ELISA plate is blocking, prepare serial dilutions of
samples in blocking buffer, in a separate nonbinding 96-well
plate. For serum, we recommend a starting dilution of 1:100,
performing an 8 to 12-point curve and 2–5 fold dilution series.
For BALF samples, start at 1:20 dilution. All samples should be
run in technical duplicate (see Note 19).
 Immunological Assessment of Inhalational Lipoprotein Vaccines 317

5. Decant blocking buffer and wash plate as described in step 2.

Add 100 μl sample per well, taking care to include the relevant
biological and technical negative controls on every plate.
6. Seal the plate and incubate at 37  C for 1 h. Decant sera/BALF
and wash plate as before.
7. Detect bound antigen-specific antibodies via addition of 100 μl
HRP-conjugated detection antibody, diluted in blocking
buffer.
8. Seal the plate and incubate at 37  C for 30 min. Decant
antibody and wash plate as before.
9. Develop ELISA via addition of 100 μl TMB substrate and
monitor for color development (see Note 20). Neutralize the
reaction by addition of 100 μl stop solution (2 M HCl).
10. Read absorbance at 450 nm with an ELISA plate reader. If
possible, also take a reference reading at 570 nm, and subtract
this from the absorbance reading at 450 nm.
11. Analyze data: reporting antibody titer is more informative than
providing raw absorbance values. Plot dilution curves for each
sample, averaging the technical duplicates or triplicates for each
dilution, and generate an equation to fit the curve. Solve to find
the antigen-specific antibody titer, determined as the dilution
at which the absorbance equals the mean absorbance (+1–3
SD) of 1:100 sera or 1:20 BALF from unvaccinated or
adjuvant-only vaccinated control mice.

4 Notes

1. We recommend performing antigen restimulation and staining

for every test sample separately, rather than pooling samples for
an experimental group, to assess biological variability. The
number of cells required for restimulation, staining, data col-
lection and subsequent analysis, will vary depending on the
experimental system, in particular the magnitude of the
vaccine-induced response. This may require optimization;
however, we would not recommend performing this protocol
with less than 2  106 leukocytes per sample. If the investigator
needs to restimulate more than 4  106 leukocytes per sample,
it will be necessary to do this across multiple wells of a 96-well
plate or utilize a larger plate size, as if the cells are more
concentrated, the risk of cell death increases which is detrimen-
tal to the assay.
2. Supernatant may be discarded by decanting the supernatant
from the plate (inside a biosafety hood) onto absorbent paper
towel, within a biohazard bag or container. However, if the
samples contain a pathogen or chemical that represents an
318 Anneliese S. Ashhurst et al.

aerosol hazard, instead remove the supernatant by aspiration,

for example, manually with a pipette. Take care not to cross-
contaminate between wells; this can be aided by spacing sam-
ples one well apart in the plate.
3. Do not pipette cells vigorously, or introduce air bubbles to the
cell suspension, as this can induce cell death.
4. Negative controls—to confirm that cytokine responses are spe-
cific to the peptide or protein antigen used for restimulation,
and not due to nonspecific activation of the cells in culture,
culture cells from the sample in the same manner but in the
absence of any antigen. This may be performed for each sample
individually; however, to conserve time and reagents, this may
be performed using a pooled sample containing an equal num-
ber of cells from each mouse within the same experimental
group. Positive controls—to ensure cells are viable and respon-
sive, as well as to verify the intracellular cytokine staining steps,
stimulate a sample for each organ type processed in the experi-
ment with a non–antigen-specific T-cell stimulant. Purified
monoclonal antibodies against murine CD3e (clone
145-2C11) and CD28 (clone 37.51), at 5 μg/ml, are effective
at inducing T-cell activation and cytokine release. We also
recommend including a few additional wells of spare cells,
treated similarly, that may be used for single-stain control
samples for acquisition on the flow cytometer.
5. Do not shorten the incubation time, as insufficient cytokine
may accumulate to obtain clear immunostaining. For practical
reasons, however, it may be necessary to extend the incubation
time, for example allowing the cells to incubate with Brefeldin
A overnight. This will still provide acceptable cell stimulation
and cytokine accumulation; however, note that cell death will
increase progressively. Where possible incubation time with
Brefeldin A should be limited to 4–6 h. Other Golgi block
reagents are also commercially available and should be used
according to the manufacturer’s instructions.
6. In the example provided in Fig. 1, surface stain antibodies were
anti-mouse CD8-APC/Cy7 (53-6.7; BD Pharmingen),
CD4-PE/Cy7 (RM4-5; BD Pharmingen) and CD3-PerCP/
Cy5.5 (17A2; BioLegend). Intracellular stain antibodies were
anti-mouse IFNγ-FITC (XMG1.2; BD Pharmingen),
IL-17A-PB (TC11-18H10.1; BioLegend), TNFα-PE
(MP6-XT22; BioLegend), and IL-2-APC (JES6-5H4;
BioLegend).
7. It is expected that a proportion of the cells will have died either
during the extraction from tissue, or during the culture restim-
ulation period. Dead cells should be excluded from flow cyto-
metry data analyses as they can nonspecifically take up antibody
Immunological Assessment of Inhalational Lipoprotein Vaccines 319

providing a false positive signal. We strongly recommend opti-

mizing the concentration of fixable viability dyes before use in
an experiment. Test a dilution series of the fixable dead cell
stain on the same number of cells to be used in each sample in
the experiment to establish optimal concentration, for exam-
ple, dilutions of 1:200, 1:500, 1:1000, 1:5000, 1:10000, on
2–4  106 cells. Prepare in the buffer recommended by the
manufacturer; often this is PBS, not FACS buffer. Do not use
the dye at a greater concentration than is necessary to clearly
distinguish live and dead cell populations, and do not exceed
the recommended staining time. Live cells may take up excess
stain, increasing the fluorescent background signal.
8. We find the BD Cytoperm/Cytofix buffer system works well
for intracellular cytokine staining protocols. It may be neces-
sary to centrifuge cells at an increased speed of 500  g, as
components of the Cytoperm/Cytofix buffers result in a looser
cell pellet.
9. As with all acquisitions on a flow cytometer, you will require
single-color stained control samples for every fluorophore in
your antibody panel, including the cell viability dye. We find
the use of compensation beads to be preferable (with the
exception of the cell viability stain) as in general, they will
provide a stronger fluorescent signal than cells. Regardless of
whether cells or compensation beads are utilized for single-
color controls, ensure that they have been stained and fixed
using exactly the same protocol as your experimental samples
to provide the best controls for optimized compensation of
fluorescence spillover signal. Different staining or fixation pro-
tocols will alter the fluorescent signatures of some
fluorophores.
10. As the cells have been cultured, it is expected that there may
have been proliferation or cell death of certain immune popu-
lations; therefore, it is not reliable to present the data as a
quantitation of the number of T-cells producing a particular
cytokine in the sample.
11. It is essential that the entire membrane of each well of the
ELISpot plate is completely wet. You will notice a color change
from a dry white to a gray-white tone as the membrane is
saturated. However, do not exceed 1 min incubation with the
35% ethanol, as this may damage the membrane and reduce the
capacity for binding to the antibodies in subsequent steps.
12. To reduce the risk of contamination, we recommend not
decanting wash buffer at this stage, but rather use a multichan-
nel pipette to aspirate washes. However, take great care not to
scrape or puncture the membrane at the bottom of the well
with the pipette tip. This may be avoided by angling the plate,
and pipetting at a corner of the well, just above the membrane.
320 Anneliese S. Ashhurst et al.

13. The concentration of antigen may need to be optimized, but in

general a final concentration of 0.1, 1, 3, and 10 μg/ml will
provide a suitable gradient of responses, with maximum
response expected when restimulating with 3 or 10 μg/ml
antigen. As a negative control, you will also need complete
cell culture media with no antigen. As a positive control to
ensure cell viability and responsiveness, we recommend the use
of a compound to nonspecifically activate T-cells in the culture,
such as ConA at a final concentration of 3–5 μg/ml in com-
plete culture media. We recommend performing the assay with
at least technical duplicates, preferably triplicates, for each
condition for every experimental sample.
14. If a very strong vaccine response has been generated, it may be
necessary to reduce this cell concentration, in order to provide
sufficient resolution for counting of the spots of the ELISpot
membranes. This may particularly be the case in nonlymphoid
organs where a strong immune response has occurred, for
example, the lungs of mice receiving a potent pulmonary vac-
cine, or following infection with a respiratory pathogen.
15. It is critical to use a different antibody clone to that used to
coat the ELISpot plate in Subheading 3.3.1, that recognizes a
different portion of the IFNγ protein, for example, XMG1.2. If
the antibodies used at the two different steps recognize a
similar epitope, or bind to a similar site on the protein, then
binding of the secondary antibody will be prevented directly
either from blocking of the binding site or by steric hindrance.
16. If necessary, spots can instead be counted manually under a
microscope, or imaged and counted using computer software.
17. The 72 h incubation period is optimized for accumulation of
murine IFNγ. If other cytokines are of primary interest, this
timepoint may need to be altered to optimize accumulation.
18. Cell quantitation may be performed by a variety of methods.
Most simply, perform a viable cell count by hemocytometer or
automated cell counter. Alternatively, radioisotope methods
such as tritiated thymidine incorporation may be performed.
Prelabeling of the leukocyte suspensions prior to the culture
with a cell proliferation dye, such as carboxyfluorescein succi-
nimidyl ester (CFSE) or similar fluorescent dye, would at this
stage allow acquisition of the cells on a flow cytometer to
calculate the proportion of proliferated cells per well.
19. We recommend running every sample individually rather than
pooling samples for an experimental group, to provide a mea-
sure of biological variability. On every ELISA plate, you will
need to run sera/BALF from the experimental negative con-
trol group, that is mice receiving no vaccination or adjuvant-
 Immunological Assessment of Inhalational Lipoprotein Vaccines 321

only vaccination, at 1:100 for serum or 1:20 for BALF to

provide a measure of the baseline background signal. You will
also need to leave 2–3 wells spare on every ELISA plate as a
technical negative control—do not add any samples to these
wells, simply blocking buffer.
20. The time taken to develop the ELISA will vary—as a guide,
when wells containing the most concentrated samples are sky
blue, or after a maximum of 20 min, neutralize the reaction.
The color should change to yellow. Take care to add substrate
and stop solution to wells in the same order, and with similar
pace, so that development time is consistent between wells.

Acknowledgments

We acknowledge support from the National Health and Medical

Research Council (Project APP1044343, and Centres of Research
Excellence in TB Control, APP1043225, APP1153493) to WJB
and the Australian Research Council and its Centre of Excellence
for Innovations in Peptide and Protein Science to RJP.

References
1. Krammer F (2020) SARS-CoV-2 vaccines in receptor 2 ligand. J Med Chem 55
development. Nature 586(7830):516–527 (15):6968–6974
2. WHO (2019) Global Tuberculosis Report. 8. McDonald DM, Byrne SN, Payne RJ (2015)
World Health Organization, Geneva Synthetic self-adjuvanting glycopeptide cancer
3. Andersen P, Doherty TM (2005) The success vaccines. Front Chem 3:60
and failure of BCG—implications for a novel 9. Bettahi I, Dasgupta G, Renaudet O, Chentoufi
tuberculosis vaccine. Nat Rev Microbiol 3 AA, Zhang X, Carpenter D, Yoon S, Dumy P,
(8):656–662 BenMohamed L (2008) Antitumor activity of a
4. Sable SB, Posey JE, Scriba TJ (2019) Tubercu- self-adjuvanting glyco-lipopeptide vaccine
losis vaccine development: Progress in clinical bearing B cell, CD4+ and CD8+ T cell epi-
evaluation. Clin Microbiol Rev 33(1): topes. Cancer Immunol Immunother 58
e00100–e00119 (2):187
5. Zhu X, Ramos TV, Gras-Masse H, Kaplan BE, 10. McDonald DM, Wilkinson BL, Corcilius L,
BenMohamed L (2004) Lipopeptide epitopes Thaysen-Andersen M, Byrne SN, Payne RJ
extended by an Nϵ-palmitoyl-lysine moiety (2014) Synthesis and immunological evalua-
increase uptake and maturation of dendritic tion of self-adjuvanting MUC1-macrophage
cells through a toll-like receptor-2 pathway activating lipopeptide 2 conjugate vaccine can-
and trigger a Th1-dependent protective immu- didates. Chem Commun 50
nity. Eur J Immunol 34(11):3102–3114 (71):10273–10276
6. McDonald DM, Hanna CC, Ashhurst AS, 11. Belnoue E, Mayol JF, Carboni S, Di Berardino
Corcilius L, Byrne SN, Payne RJ (2018) Syn- BW, Dupuychaffray E, Nelde A, Stevanovic S,
thesis of a self-adjuvanting MUC1 vaccine via Santiago-Raber ML, Walker PR, Derouazi M
diselenide-selenoester ligation-deselenization. (2019) Targeting self and neo-epitopes with a
ACS Chem Biol 13(12):3279–3285 modular self-adjuvanting cancer vaccine. JCI
7. Abdel-Aal A-BM, El-Naggar D, Zaman M, Insight 5(11):e127305
Batzloff M, Toth I (2012) Design of Fully 12. Moyle PM, Toth I (2008) Self-adjuvanting
Synthetic, self-Adjuvanting vaccine incorporat- lipopeptide vaccines. Curr Med Chem 15
ing the tumor-associated carbohydrate Tn anti- (5):506–516
gen and Lipoamino acid-based toll-like
322 Anneliese S. Ashhurst et al.

13. Anderson RJ, C-w T, Daniels NJ, Compton BJ, 22. Satti I, Meyer J, Harris SA, Thomas Z-RM,
Hayman CM, Johnston KA, Knight DA, Griffiths K, Antrobus RD, Rowland R,
Gasser O, Poyntz HC, Ferguson PM, Larsen Ramon RL, Smith M, Sheehan S,
DS, Ronchese F, Painter GF, Hermans IF Bettinson H, McShane H (2014) Safety and
(2014) A self-adjuvanting vaccine induces immunogenicity of a candidate tuberculosis
cytotoxic T lymphocytes that suppress allergy. vaccine MVA85A delivered by aerosol in
Nat Chem Biol 10:943 BCG-vaccinated healthy adults: a phase 1, dou-
14. Azmi F, Ahmad Fuaad AA, Giddam AK, Batzl- ble-blind, randomised controlled trial. Lancet
off MR, Good MF, Skwarczynski M, Toth I Infect Dis 14(10):939–946
(2014) Self-adjuvanting vaccine against group 23. WHO (2012) Measles aerosol vaccine project -
a streptococcus: application of fibrillized pep- report to SAGE. WHO, Geneva, Switzerland
tide and immunostimulatory lipid as adjuvant. 24. Counoupas C, Ferrell KC, Ashhurst A, Bhatta-
Bioorg Med Chem 22(22):6401–6408 charyya ND, Nagalingam G, Stewart EL, Feng
15. Batzloff MR, Hartas J, Zeng W, Jackson DC, CG, Petrovsky N, Britton WJ, Triccas JA
Good MF (2006) Intranasal vaccination with a (2020) Mucosal delivery of a multistage sub-
Lipopeptide containing a Conformationally unit vaccine promotes development of lung-
constrained conserved minimal peptide, a uni- resident memory T cells and affords interleu-
versal T cell epitope, and a self-Adjuvanting kin-17-dependent protection against pulmo-
lipid protects mice from group a streptococcus nary tuberculosis. NPJ Vaccines 5(1):105
challenge and reduces throat colonization. J 25. O’Garra A, Redford PS, McNab FW, Bloom
Infect Dis 194(3):325–330 CI, Wilkinson RJ, Berry MPR (2013) The
16. Rai PK, Chodisetti SB, Maurya SK, Nadeem S, immune response in tuberculosis. Annu Rev
Zeng W, Janmeja AK, Jackson DC, Agrewala Immunol 31(1):475–527
JN (2018) A lipidated bi-epitope vaccine com- 26. Counoupas C, Triccas JA, Britton WJ (2019)
prising of MHC-I and MHC-II binder pep- Deciphering protective immunity against
tides elicits protective CD4 T cell and CD8 T tuberculosis: implications for vaccine develop-
cell immunity against Mycobacterium tuberculo- ment. Expert Rev Vaccines 18(4):354–364
sis. J Transl Med 16(1):279 27. Flórido M, Pillay R, Gillis CM, Xia Y, Turner
17. Xu Z, Rivera-Hernandez T, Chatterjee O, SJ, Triccas JA, Stambas J, Britton WJ (2015)
Walker MJ, Moyle PM (2020) Semisynthetic, Epitope-specific CD4+, but not CD8+, T-cell
self-adjuvanting vaccine development: effi- responses induced by recombinant influenza a
cient, site-specific sortase A-mediated conjuga- viruses protect against Mycobacterium tubercu-
tion of toll-like receptor 2 ligand FSL-1 to losis infection. Eur J Immunol 45(3):780–793
recombinant protein antigens under native 28. Barletta KE, Cagnina RE, Wallace KL, Ramos
conditions and application to a model group SI, Mehrad B, Linden J (2012) Leukocyte
A streptococcal vaccine. J Control Release compartments in the mouse lung: distinguish-
317:96–108 ing between marginated, interstitial, and alveo-
18. Ashhurst A, McDonald D, Hanna C, lar cells in response to injury. J Immunol
Stanojevic V, Britton WJ, Payne RJ (2019) Methods 375(1–2):100–110
Mucosal vaccination with a self-adjuvanted 29. Sakai S, Kauffman KD, Schenkel JM, McBerry
lipopeptide is immunogenic and protective CC, Mayer-Barber KD, Masopust D, Barber
against Mycobacterium tuberculosis. J Med DL (2014) Cutting edge: control of Mycobac-
Chem 62(17):8080–8089 terium tuberculosis infection by a subset of lung
19. Hanna CC, Ashhurst AS, Quan D, Maxwell parenchyma–homing CD4 T cells. J Immunol
JWC, Britton WJ, Payne RJ (2021) Synthetic 192(7):2965–2969
protein conjugate vaccines provide protection 30. Flórido M, Muflihah H, Lin LCW, Xia Y,
against Mycobacterium tuberculosis in mice. Sierro F, Palendira M, Feng CG, Bertolino P,
Proc Natl Acad Sci U S A 118(4): Stambas J, Triccas JA, Britton WJ (2018) Pul-
e2013730118 monary immunization with a recombinant
20. Tyne AS, Chan JGY, Shanahan ER, influenza a virus vaccine induces lung-resident
Atmosukarto I, Chan H-K, Britton WJ, West CD4+ memory T cells that are associated with
NP (2013) TLR2-targeted secreted proteins protection against tuberculosis. Mucosal
from Mycobacterium tuberculosis are protective Immunol 11:1743–1752
as powdered pulmonary vaccines. Vaccine 31 31. Ashhurst AS, Flórido M, Lin LCW, Quan D,
(40):4322–4329 Armitage E, Stifter SA, Stambas J, Britton WJ
21. Sabin AB (1983) Immunization against mea- (2019) CXCR6-deficiency improves the con-
sles by aerosol. Rev Infect Dis 5(3):514–523 trol of pulmonary mycobacterium tuberculosis
 Immunological Assessment of Inhalational Lipoprotein Vaccines 323

and influenza infection independent of secreting cells. J Immunol Methods 65

T-lymphocyte recruitment to the lungs. Front (1):109–121
Immunol 10:339 38. Czerkinsky C, Andersson G, Ekre HP, Nilsson
32. Mackay LK, Kallies A (2017) Transcriptional LA, Klareskog L, Ouchterlony O (1988)
regulation of tissue-resident lymphocytes. Reverse ELISPOT assay for clonal analysis of
Trends Immunol 38(2):94–103 cytokine production. I. Enumeration of
33. Joosten SA, Ottenhoff THM, Lewinsohn DM, gamma-interferon-secreting cells. J Immunol
Hoft DF, Moody DB, Seshadri C (2019) Har- Methods 110(1):29–36
nessing donor unrestricted T-cells for new vac- 39. Zhang W, Caspell R, Karulin AY, Ahmad M,
cines against tuberculosis. Vaccine 37 Haicheur N, Abdelsalam A, Johannesen K,
(23):3022–3030 Vignard V, Dudzik P, Georgakopoulou K,
34. Ashhurst TM, Smith AL, King NJC (2017) Mihaylova A, Silina K, Aptsiauri N, Adams V,
High-dimensional fluorescence cytometry. Lehmann PV, McArdle S (2009) ELISPOT
Curr Protoc Immunol 119:5.8.1–5.8.38 assays provide reproducible results among dif-
35. Maciorowski Z, Chattopadhyay PK, Jain P ferent laboratories for T-cell immune monitor-
(2017) Basic multicolor flow cytometry. Curr ing--even in hands of ELISPOT-inexperienced
Protoc Immunol 117(1):5.4.1–5.4.38 investigators. J Immunotoxicol 6(4):227–234
36. Saletti G, Çuburu N, Yang JS, Dey A, Czer- 40. Boulet S, Ndongala ML, Bernard NF (2010)
kinsky C (2013) Enzyme-linked immunospot Dual-color ELISPOT assay for the simulta-
assays for direct ex vivo measurement of neous detection of IL-2 and/or IFN-γ secret-
vaccine-induced human humoral immune ing T cells. Cold Spring Harbor Protocols
responses in blood. Nat Protoc 8 2010(1):pdb.prot5369
(6):1073–1087 41. Gazagne A, Claret E, Wijdenes J, Yssel H,
37. Czerkinsky CC, Nilsson L-Å, Nygren H, Bousquet F, Levy E, Vielh P, Scotte F, Goupil
Ouchterlony Ö, Tarkowski A (1983) A solid- TL, Fridman WH, Tartour E (2003) A Fluoro-
phase enzyme-linked immunospot (ELISPOT) spot assay to detect single T lymphocytes
assay for enumeration of specific antibody- simultaneously producing multiple cytokines.
J Immunol Methods 283(1–2):91–98
 Chapter 17

Determination of Maternal and Infant Immune Responses

to Pertussis Vaccination in Pregnancy
Thomas Rice and Beth Holder

Abstract
The Bordetella pertussis bacterium is the causative agent of whooping cough (pertussis disease). Following
recent outbreaks of pertussis, disproportionately affecting young infants, several countries have introduced
maternal pertussis vaccination strategies, aimed at boosting transplacental transfer of protective antibodies
during pregnancy. Given historical associations between high maternal antibody and blunted infant
responses to vaccination, subsequent research studies have investigated the impact of maternal pertussis
vaccine on infant humoral responses. However, far less is known about the potential impact of the vaccine
on innate immunity. Here, we describe methods to detect in vitro cellular responses to B. pertussis in
mothers and their infants using a B. pertussis stimulation assay and multiplex cytokine assays to address this
research question.

Key words Maternal vaccination, Pregnancy, Bacterial infection, Innate immunity, Cytokines,
Pediatrics

1 Introduction

Epidemics of pertussis (whooping cough) have been observed in

many countries including the USA [1], Australia [2], and the UK
[3], resulting in the deaths of infants too young to be protected by
immunization. As a result, maternal pertussis vaccination during
pregnancy has become the foremost tool in protecting vulnerable
young infants from pertussis in the first weeks of life, through the
transplacental transfer of maternal antibody to the fetus during
pregnancy [4, 5].
The bacterial agent of pertussis, Bordetella pertussis, elicits an
innate immune response upon infecting the host [6, 7], character-
ized by the release of cytokines from innate immune cells such as
natural killer cells [6, 8, 9]. These innate cytokines help direct the
adaptive immune response as the infection progresses, promoting
cytokine release from T cells.

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_17,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

325
326 Thomas Rice and Beth Holder

In this chapter we describe how whole blood collected from

mothers and infants can be utilized to investigate the impact of
tetanus-diphtheria-acellular pertussis (Tdap) vaccination on innate
immune cell responses. Maternal, cord and infant blood collected
from unvaccinated pregnancies, and from pregnancies where
mothers received Tdap vaccination are stimulated in vitro with
heat-killed B. pertussis. Cytokine responses are detected by multi-
plex assay or ELISA performed on supernatants. In addition, we
describe how this assay can be adapted to investigate the contribu-
tion of the plasma compartment in cord blood cytokine responses
to B. pertussis in a plasma removal assay.

2 Materials

2.1 Patient Samples Samples are required from mother-infant pairs, with equal numbers
of Tdap-vaccinated and unvaccinated pregnancies. The gestational
age at time of vaccination of the mother, and the brand of vaccine
administered should be recorded.
1. Maternal sample(s): at minimum, one sample is required from
the mother; usually at term. For practical reasons, this sample is
often taken near the time of birth, when mothers attend a
healthcare setting. The timing of the sample (before/after
birth) and the mode of delivery should be considered, as
labor is an inflammatory process. If samples are taken after
birth, it is preferable to use samples from caesarean sections.
Likewise, if patients are delivering vaginally, it is preferable to
collect blood prior to commencement of labor (see Note 1).
Ideally, gestation-matched controls from unvaccinated preg-
nancies should also be obtained.
2. Infant samples: at minimum, cord blood taken at birth can
serve as a proxy for neonatal blood (see Note 2). If the aim is
to investigate impact on infant vaccine responses, infant blood
is additionally required at 7 weeks of age (1 week prior to
primary pertussis immunization at 8 weeks) and at 5 months
of age (1 month after completion of the primary pertussis
immunization at 4 months; see Note 3).

2.2 Collection 1. Blood collection tubes: commercially available sodium or lith-

and Processing ium heparin tubes. Tube type should be kept consistent.
of Whole Blood 2. Needles: 21-gauge for maternal and cord blood samples,
23-gauge for infant samples.
3. 20 mL syringes.
4. Gauze.
 Detecting Responses to Maternal Tdap Vaccination 327

2.3 Heat-Killed 1. B. pertussis bacteria.

B. pertussis 2. 500 mL shake flasks.
3. THIJS medium: 994.169 mL ultrapure H2O, 3.32 g NaCl,
0.11 g NH4Cl, 0.5 g KH2PO4, 0.5 g KCl, 0.1 g MgCl2·6H2O,
1.53 g Tris–HCl, 1.87 g Na glutamate, 40% w/v L-lactate
3.76 mL, 2.071 mL 5 M NaOH. Heat-sterilize at 110
C for
20 min and store at 4
C (see Note 4).
4. 100 THIJS supplement: 0.2 g L-cystine, 0.13 g CaCl2·2H2O,
0.5 g L-glutathione reduced, 0.05 g FeSO4·7H2O, 0.02 g
nicotinic acid, 0.1 g L-ascorbic acid. Add 44 mL of ultrapure
water and 6 mL of 1 M HCl. Filter-sterilize with a 0.22-μm
filter and store as 1 mL aliquots at 20
C for up to 1 year.
5. Water bath.
6. 2.0 mL cryovials.

2.4 In Vitro Whole 1. 96-well sterile round-bottom tissue culture plate with lid.
Blood 2. Tissue culture hood.
Stimulation Assay
3. Humidified cell culture incubator set to 5% CO2 and 37
C.
4. S-RPMI: Commercially available Roswell Park Memorial Insti-
tute (RPMI 1640) medium supplemented with 10% decom-
plemented fetal calf serum and 1% penicillin–streptomycin.
5. Heat-killed B. pertussis bacteria.
6. 5 mg commercially available bacterial Lipopolysaccharide
(LPS).
7. 3.5 mL bijou.
8. Wet ice.

2.5 Detection 1. Meso Scale Discovery Proinflammatory Panel 1 Human kit (see
of Cytokines by Notes 5 and 6).
Multiplex 2. MESO QuickPlex SQ 120 plate reader and Discovery Work-
and Singleplex Assays bench 4.0 software (free to download).
3. Commercial IL-8 ELISA.
4. High-binding 96-well ELISA plates.
5. Non–high-binding 96-well plate.
6. Reagent reservoirs.
7. Multichannel pipette—ranges 10 μL–200 μL.
8. Adhesive plate sealers.
9. Phosphate buffered saline (PBS): 137 mM NaCl, 10 mM phos-
phate, 2.7 mM KCl, in H2O. Adjust to a final pH of 7.4.
10. ELISA wash buffer: 1 PBS, 0.05% Tween 20.
11. Wet ice.
328 Thomas Rice and Beth Holder

12. Vortexer.
13. 1.5 mL microcentrifuge tubes.
14. Plate shaker.
15. Rocking platform or tube rotator (optional).
16. 2 N H2SO4.
17. Absorbance microplate reader capable of reading 450 nm and
accompanying analysis software.

2.6 Contribution This assay is an adaption of the in vitro whole blood B. pertussis
of the Plasma stimulation assay. As such, reagents and equipment as described in
Compartment in Cord Subheadings 2.2–2.4 are required, in addition to those listed
Blood Cytokine below.
Responses 1. Serum or plasma blood collection tubes.
to B. pertussis
2. Pools of cord blood serum or plasma from unvaccinated and
Tdap-vaccinated pregnancies—40 μL required per assay (see
Note 7).
3. Centrifuge.
4. Microcentrifuge tubes.

3 Methods

3.1 Preparation 1. Grow liquid cultures of B. pertussis overnight in 200 mL THIJS

of Heat-Killed medium at 37
C and 200 rpm (see Note 8).
B. pertussis and LPS 2. Harvest B. pertussis at mid-log growth phase (OD620 0.5–0.6)
Stocks by centrifugation at 3200  g for 10 min at room temperature.
3. Wash the resulting pellet in PBS, centrifuging at 3200  g for
10 min at room temperature.
4. Resuspend in 10 mL THIJS medium and determine the con-
centration of the culture, either using a spectrophotometer or
the plate count method (see Note 9).
5. Adjust the concentration of the B. pertussis culture to
1  108 CFU/mL and heat-inactivate 1 mL aliquots of B. per-
tussis for 30 min at 56
C in a water bath.
6. Flash freeze the 1 mL aliquots of heat-inactivated B. pertussis,
with addition of 10% v/v glycerol, and store at 80
C.
7. LPS is provided as a lyophilized powder. Reconstitute to 5 mg/
mL by the addition of 1 mL of sterile endotoxin-free water,
leave to dissolve for at least 5 min and vortex thoroughly.
8. Dilute to intermediate stock concentration of 2 μg/mL in
S-RPMI and store at 80
C in 800 μL aliquots for future use.
 Detecting Responses to Maternal Tdap Vaccination 329

3.2 Bulk Preparation Plates are prepared in bulk and frozen to reduce variability and to
of B. pertussis simplify the logistics of working with fresh blood samples. Treat-
Stimulation Plates ments are prepared at 2 the final required concentration, as they
will be diluted by the addition of an equal volume of prepared
blood sample in the stimulation assay (see Subheading 3.4, step 3).
1. To prepare ten plates (each with one set of treatments), thaw
one aliquot of frozen heat-killed B. pertussis stock and four
aliquots of frozen LPS stock on ice (see Note 10).
2. Label ten 96-well round-bottom tissue culture plates with
“B. pertussis stimulation” and the date of preparation.
3. Working in a tissue culture hood, prepare 3.5 mL of bacteria at
1  105 CFU/mL. From a stock of 1  108 CFU/mL (see
Note 11), transfer 35 μL of bacteria to 3465 μL of S-RPMI.
Gently vortex to mix.
4. Pour carefully into a reagent reservoir and, using a multichan-
nel pipette, add 100 μL per well in triplicate (in three adjacent
wells in a row) in each of the ten 96-well plates (see Note 12).
5. For the LPS positive control, pool four aliquots of LPS stock in
a sterile 3.5 mL bijou. Vortex gently to mix (see Note 13).
6. Pour into a reagent reservoir and, using a multichannel pipette,
add 100 μL per well in triplicate, adjacent to the wells contain-
ing the heat-killed B. pertussis, in each of the ten 96-well plates.
7. For the no-treatment (i.e., negative control) wells, add 100 μL
per well of S-RPMI to the plates in triplicate, adjacent to the
wells containing the LPS positive control aliquots.
8. Seal plates with the accompanying lid and freeze at 80
C for
future use (see Note 14).

3.3 Processing 1. Peripheral venous blood samples (minimum 1 mL) from mums
of Maternal, Cord, and babies are collected into blood collection tubes (see Note
and Infant 15).
Whole Blood 2. Immediately invert the blood sample tubes end to end at least
twice to ensure blood is mixed with the anticoagulant.
3. Keep tubes at room temperature and process within 2 h (see
Subheading 3.4, step 2).

3.4 In Vitro 1. Thaw preprepared stimulation plates on ice.

B. pertussis 2. In a tissue culture hood, for each sample dilute blood 1:5 by
Stimulation Assay adding 200 μL of whole blood to 800 μL S-RPMI in a sterile
bijou.
3. Pour into a reagent reservoir and, using a multichannel pipette,
add 100 μL diluted blood to each of the nine wells containing a
treatment (i.e., B. pertussis, LPS control and negative control)
per sample. Mix by gently pipetting up and down twice and
330 Thomas Rice and Beth Holder

change tips with each addition. This results in a final blood

dilution of 1:10 and a B. pertussis concentration of 5  105/
mL.
4. Place the plate in the incubator at 37
C, 5% CO2 for 24 h.
5. After 24 h, carefully remove the supernatant into two new
96-well cell culture plates (see Note 16).
6. Place the tissue culture plate lids onto the plates and freeze at
80
C for use in future assays (see Subheadings 3.5 and 3.6).

3.5 Meso Scale MSD plates can only be used once. Therefore, it is important to
Discovery (MSD) accumulate samples before analyzing cytokine levels in supernatants
Multiplex using this assay.
Cytokine Assay 1. Prepare a plate plan for your experiment (see Note 17).
2. Thaw eight sets of supernatant samples (from Subheading 3.4)
on ice and allow all assay reagents to reach room temperature
before being used (see Note 18).
3. Reconstitute the cytokine calibrator provided in the kit by
adding 1 mL of Diluent 2 to lyophilized MSD Calibrator
Blend and vortexing (see Note 19). Allow the reconstituted
calibrator to sit for a minimum of 5 min before using.
4. In this time, add 300 μL of Diluent 2 to each of eight 1.5 mL
microfuge tubes.
5. Prepare a dilution series of the calibrator by transferring 100 μL
to the first of the tubes containing Diluent 2, prepared above,
to create a fourfold dilution. Mix well by vortexing.
6. Transfer 100 μL to the subsequent tube and repeat this process
to serially dilute 6 more times, leaving the eighth tube with
Diluent 2 alone. This final tube will serve as the calibration
blank.
7. Using a multichannel pipette, aliquot 90 μL Diluent 2 per well
into ten columns of wells in a 96-well plate (not supplied in the
MSD kit; use any non–high-binding plate), according to the
MSD plate layout (see Note 17). This plate will be used for the
dilution of supernatant samples only.
8. Dilute thawed supernatant samples (from step 2 above) 1:10
by transferring 10 μL of each sample, using a multichannel
pipette, into appropriate wells (according to prepared plate
layout) containing 90 μL Diluent 2 and mix by pipetting up
and down at least five times. Change tips between sets of
samples. All samples are tested at a single dilution of 1:10.
9. Add 50 μL of standard and samples to the MSD plate according
to the plate layout (see Note 20).
10. Cover plates with an adhesive plate sealer and incubate for 2 h
at room temperature on a plate shaker at 300 rpm.
 Detecting Responses to Maternal Tdap Vaccination 331

11. Discard samples and standards by flicking the plate into a

laboratory sink.
12. Using a multichannel pipette, wash each well with 150 μL
PBS–0.05% Tween, allowing the well to soak for at least 30 s.
IMPORTANT: Do not use an automatic plate washer. After-
ward, blot-dry by tapping plate well-side down on tissue paper.
13. Prepare a cocktail of detection antibodies for all cytokines
(apart from IL-8) at 1 concentration in Diluent 3. The total
volume required for a full plate is 3 mL, which equates to 60 μL
per antibody in 2.46 mL diluent 3 (see Note 21).
14. Add 25 μL detection antibody cocktail per well, ensuring the
whole surface is covered and there are no bubbles. Incubate on
plate shaker at 300 rpm for 2 h at room temperature.
15. Wash the plate three times, as described in step 12 above, with
PBS–0.05% Tween and blot-dry.
16. Add 150 μL 2 Read Buffer to each well (see Note 22).
17. Read plates within 30 min on a MESO QuickPlex SQ 120 and
analyze data using Discovery Workbench 4.0.

3.6 IL-8 ELISA 1. Prepare a plate plan for your experiment (see Note 23).
2. Prepare 1 IL-8 capture antibody in 1 coating buffer (see
Note 24), by adding 40 μL antibody to 10 mL coating buffer
per 96-well plate (see Note 25).
3. Coat ELISA plate with 100 μL per well of IL-8 capture anti-
body (see Note 26).
4. Seal the plate with an adhesive plate sealer and incubate over-
night at 4
C.
5. Aspirate and wash wells three times with 200 μL ELISA wash
buffer to remove unbound antibody (see Note 27). Allow wells
to soak in wash buffer for at least 30 s. Afterward, blot on tissue
paper to remove and excess wash buffer from wells.
6. To prevent nonspecific binding of antigens and antibodies,
block wells with 200 μL 1 diluent (see Note 28).
7. Seal and incubate the plate at room temperature for a minimum
of 1 h.
8. Whilst the plate is being blocked, thaw supernatants on ice (see
Note 29).
9. Using standard nonsterile 96-well plates (not high-binding),
dilute assay supernatants to an intermediate dilution of 1:50 in
PBS according to the plate plan designed in step 1 (see Note
30). For example, in one plate, dilute 10 μL of sample in 90 μL
of PBS for a 1:10 dilution. In a separate plate, take 20 μL of the
diluted sample from plate one, and add to 80 μL PBS (making
an overall intermediate dilution of 1:50; see Note 31). Ensure
332 Thomas Rice and Beth Holder

during dilution steps that sample and PBS are well mixed by
pipetting up and down at least five times. When used in step 15
below, the final dilution for samples will be 1:500 (see Note
32).
10. Reconstitute the IL-8 standard provided in the kit using the
amount of deionized water stated on the vial, and place on a
rocking platform or tube rotator for at least 15 min to solubi-
lize (see Note 33).
11. Whilst the standard is being reconstituted, wash the blocked
ELISA plate twice as described in step 5, and blot-dry. It is
important not to leave the wells dry for long periods of time
between steps.
12. To columns 1 and 2 of the plate, apart from wells A1 and B1
which will contain the reconstituted standard, add 100 μL of
1 diluent. In addition, 100 μL of 1 diluent should be added
to the two designated blank wells G12 and H12.
13. To all wells that will receive samples, add 90 μL of 1 diluent
(see Note 34).
14. Add 200 μL of the reconstituted IL-8 standard to wells A1 and
B1. Using a multichannel pipette, take 100 μL from A1 and B1
into A2 and B2, pipetting up and down to mix with the diluent
present in the well. Change tips, and repeat the process a
further six times, discarding 100 μL from A8 and B8. This
will create a serially diluted eight-point standard curve.
15. Add 10 μL of prediluted 1:50 sample from step 9 above to
designated sample wells containing 90 μL of 1 diluent,
making a final sample dilution of 1:500.
16. Once all samples and standards are added, seal the plate with an
adhesive plate sealer and incubate at room temperature for 2 h.
17. Prepare biotin-conjugated IL-8 detection antibody by adding
40 μL antibody to 10 mL 1 diluent per 96-well plate, or the
same volumes as for capture antibody if running partial plates
(see Note 25).
18. Wash and aspirate plates four times with wash buffer as
described in step 5.
19. Add 100 μL diluted IL-8 detection antibody per well, seal plate
and incubate at room temperature for 1 h (see Note 35).
20. Prepare avidin-conjugated horseradish peroxidase (HRP)
enzyme by adding 40 μL HRP conjugate to 10 mL 1 diluent
for a 96-well plate, or same volumes as for capture antibody if
running partial plates (see Note 25).
21. Wash and aspirate plates four times with wash buffer as
described in step 5.
 Detecting Responses to Maternal Tdap Vaccination 333

22. Add 100 μL prepared HRP per well, seal plate and incubate at
room temperature for 30 min.
23. Wash and aspirate plates six times with wash buffer as described
in step 5.
24. Add 100 μL of 3,30 ,5,50 -tetramethylbenzidine (TMB) per well
(see Note 36).
25. Monitor the colour change of sample and standard wells from
clear to blue to ensure samples and standards do not become
saturated. When the standard curve is colored blue, with a
visible difference between the middle standards (see Note
37), stop the reaction by the addition of 100 μL 2 N H2SO4
(see Note 38).
26. Immediately read the plates at 450 nm using an absorbance
microplate reader. If wavelength subtraction is available, sub-
tract the values of readings at 570 nm from those at 450 nm
and analyze data (see Note 39).

3.7 Contribution 1. Collect blood in blood collection tubes. The assay requires
of the Plasma 400 μL of cord whole blood.
Compartment in Cord 2. Dilute cord blood 1:5 in S-RPMI. After dilution, remove
Blood Cytokine 700 μL of this into a microcentrifuge tube for use in control
Responses wells.
to B. pertussis 3. Centrifuge the rest of the diluted blood at 500  g for 5 min.
4. Remove the top clear plasma layer, without disturbing the cells
beneath and record the volume taken off. Replace this volume
with S-RPMI (see Note 40).
5. Prepare 96-well cell culture stimulation plate (see Fig. 1 for
layout). To eight wells (1A-1D and 2A-2D), add 80 μL S-
RPMI. To eight wells (3A-3D and 4A-4D), add B. pertussis
prepared in 80 μL S-RPMI. To 2 wells (5A and 6A), add
100 μL of 2 μg/mL LPS.

1 2 3 4 5 6
No treatment A S-RPMI
-Plasma B
B. pertussis
-Tdap serum/plasma C
+Tdap serum/plasma D LPS

Fig. 1 Plate layout showing the stimulation conditions for the plasma removal
assay. Cord blood is stimulated with S-RPMI, B. pertussis and LPS. Stimulations
are done with plasma removed from cord blood and replaced with S-RPMI
(plasma), with plasma removed and replaced with cord plasma/serum from
unvaccinated pregnancies (Tdap serum/plasma) or with plasma removed and
replaced with cord plasma/serum from Tdap-vaccinated pregnancies (+Tdap
serum/plasma)
334 Thomas Rice and Beth Holder

6. Add a further 20 μL S-RPMI to wells (except 5A and 6A) in

rows A and B.
7. Add 20 μL of unvaccinated (to wells in row C) and Tdap-
vaccinated (to wells in row D) cord blood serum or plasma, as
shown in Fig. 1.
8. Add 100 μL of blood with plasma to all wells in row 1. These
are the no-treatment control wells for the assay. Mix up and
down using a multichannel pipette.
9. To the rest of the wells on the plate, add 100 μL of cord blood
without plasma. Mix up and down using a multichannel
pipette.
10. Incubate at 37
C and 5% CO2 for 24 h.
11. Remove the supernatant as described in Subheading 3.4, step
5 (see Note 16), and store at 80
C.
12. Analyse cytokine levels in supernatants by multiplex assay or
ELISA as described in Subheadings 3.5 and 3.6, respectively.

4 Notes

1. If desired, additional samples can be taken. These could com-

prise samples taken before vaccination (which can be the day of
vaccination) and 2 weeks post vaccination. This will enable
comparison of innate cell responses before and after vaccina-
tion in the same patient, as well as the measurement of the
antibody response, if desired.
2. The neonatal immune system changes quite rapidly post-birth,
and this should be recognized when using cord blood [10].
3. Pediatric vaccine schedules differ between countries, so it is
important to check the national vaccine schedule for infants in
your study and record the dates and the doses of vaccines
infants receive.
4. For background on THIJS medium, please see Ref. 11.
5. The Proinflammatory Panel 1 Human kit contains an MSD
plate precoated with capture antibodies for each target cytokine
(i.e., IFN-γ, IL-1β, IL-2, IL-4, IL-6, IL-8, IL-10, IL-12p70,
IL-13, and TNF-α) that are located on independent spots in
each well, allowing individual samples to be tested in a single
well for this set of cytokines using a cocktail of the detection
antibodies provided. The kit also contains the necessary stan-
dards, diluents, and buffers for the assay.
6. The chemokine IL-8 is found at high levels and is regularly out
of range of detection in the MSD assay. Levels of IL-8 can be
measured separately by ELISA.
 Detecting Responses to Maternal Tdap Vaccination 335

7. To prepare serum pools, thaw ten serum samples on ice. Pool

and mix, aliquoting and freezing at 80
C for use in future
experiments.
8. All laboratory work with B. pertussis should be carried out in
Biosafety level 2 laboratories.
9. It is helpful to freeze B. pertussis at a working stock concentra-
tion of 10  108 colony forming units (CFU)/mL. The plate
count method consists of serially diluting a sample with sterile
saline or phosphate buffer diluent until the bacteria are dilute
enough to count accurately. CFU/mL can then be calculated
as (no. of colonies  dilution factor)/volume of culture plate.
10. This protocol describes preparing ten plates to be thawed and
used for one blood sample each. To save on plastic laboratory
consumables, and if laboratories expect to receive more than
one blood sample in a day, plates can be prepared to include
more than one sample per plate.
11. B. pertussis stimulation will be performed in triplicate at a final
concentration of 5  105 CFU/mL. For ten plates, with one
sample per plate, a total of 3 mL of bacteria is required. The
bacteria will be diluted in the plate by the addition of an equal
volume of blood.
12. Any set of nine contiguous wells on the plate can be used for
the triplicate aliquots of the three treatments, as this will allow
use of a multichannel pipette for downstream processing. Sur-
round the stimulant wells with water or PBS to help prevent
evaporation in these wells once blood samples are added and
the plate is placed in an incubator. This can be done prior to
freezing, or once the plate has been thawed.
13. LPS is a component of gram-negative bacteria that elicits an
innate immune response through toll-like receptor 4. This
serves as a positive control in cytokine assays. The final concen-
tration used in our assay is 1 μg/mL.
14. Plates can be prepared in batches and frozen at 80
C to limit
variability between plates. This also fits in with the logistics of
pregnancy studies, where the timing of sampling can be
unpredictable.
15. For the collection of cord blood from the placenta, fetal veins
are usually larger, whereas arteries are smaller and more super-
ficial. Arteries usually cross over veins at some point in their
chorionic plate arrangement. Venous cord blood can be taken
from the umbilical cord of the placenta, or the fetal veins on the
surface of the placenta. Ensure the placenta has been wiped
with gauze to prevent contamination with any maternal blood
that may be present of the surface of the placenta.
336 Thomas Rice and Beth Holder

16. Be careful not to disturb the cell pellet sitting at the bottom of
the well. Approximately 160 μL of supernatant can be removed
using a multichannel pipette without disturbing the cell pellet.
This can be done in two 80 μL draws into one plate, mixing by
pipetting up and down, and then removing 80 μL into the
second plate.
17. MSD plates are precoated with capture antibodies for each
target, located on independent spots within each of the
96 wells on the plate. MSD plates are expensive and can only
be used once; therefore, it is good to make maximal use of the
available wells. For the experiment described in this chapter,
each blood sample requires nine wells in order to test the
effects of three treatments, each tested in triplicate. A sug-
gested MSD plate plan that can accommodate eight sets of
samples is shown in Fig. 2. As columns 1 and 2 will contain
duplicates of standards, the eight sets of samples must be
arranged in in consecutive rows (A to H) of nine columns,
which would leave one column empty. In order not to waste
these wells, other samples, unrelated to the experiment
described in this chapter may be tested in this column. A
different layout will be required for the samples from Subhead-
ing 3.7.
18. Upon the first thaw of MSD assay diluents, aliquot into smaller
volumes for future use to avoid repetitive freeze–thaw cycles.
8 mL aliquots of Diluent 2 and 3 mL aliquots of Diluent 3 can
be made. Aliquots can be kept at 20
C.
19. Each kit contains a vial of lyophilized calibrator per MSD plate.
The reconstituted calibrator is stable for 1 day at 2–8
C.

1 2 3 4 5 6 7 8 9 10 11 12
A Standard

B Blank
C
D B. pertussis

E LPS
F
S-RPMI
G
H Empty well

Fig. 2 MSD plate layout. Shown is the suggested layout for testing of eight sets of blood samples, each
stimulated with the three treatments (S-RPMI, B. pertussis, and LPS) run in triplicate. The assay standards and
blanks are applied to wells in columns 1 and 2, and the eight sets of samples are applied, one per row, in wells
of columns 3 to 11, as shown
 Detecting Responses to Maternal Tdap Vaccination 337

20. It is important that the full surface of the well is covered with
sample. Reverse pipetting can help to prevent bubbles in the
wells. To reverse pipette, set the pipette to the desired volume.
When pressing the pipette plunger down to take up a solution,
go past the first stop. Immerse the tip in the liquid, and slowly
release the plunger to full extension. The takes up more vol-
ume of solution than set on the pipette. When dispensing, press
the plunger down to the first stop. A small volume of liquid will
remain in the tip, which helps to prevent bubbles.
21. For the detection antibody cocktail, use 60 μL per antibody, or
60  9 ¼ 540 μL. For 3 mL total volume, add 2.46 mL Diluent
3. NB: the IL-8 antibody is not included in the cocktail, as
sample levels are regularly out of range in the MSD assay, so are
tested separately by ELISA.
22. Read Buffer is provided at 4 concentration, dilute to 2 in
deionized water. For a 96-well plate, add 7.5 mL of Read
Buffer to 7.5 mL of deionized water.
23. A suggested ELISA plate plan that can accommodate eight sets
of samples, tested at a single dilution, is shown in Fig. 3.
Columns 1 and 2 will contain duplicates of standards, and the
eight sets of samples are tested in consecutive rows (A to H) of
columns 3 to 11. The final column contains two wells used for
blanks and six empty wells. A different layout will be required
for the samples from Subheading 3.7.
24. Coating buffer is provided as a 10 stock, dilute to 1 in
deionized water. For one full plate, dilute 1 mL coating buffer
in 9 mL deionized water.

1 2 3 4 5 6 7 8 9 10 11 12
A Standard

B Blank
C
D B. pertussis

E LPS
F
S-RPMI
G
H Empty well

Fig. 3 IL-8 ELISA plate layout. Shown is the suggested layout for testing of eight sets of blood samples, each
stimulated with the three treatments (S-RPMI, B. pertussis, and LPS) run in triplicate. The assay standards are
applied to wells in columns 1 and 2, and the eight sets of samples are applied, one per row, in wells of
columns 3 to 11, as shown. Two blank wells are included in column 12
338 Thomas Rice and Beth Holder

Table 1
Volumes of reagents for partial ELISA plates

Number of ELISA plate columns 12 11 10 9 8 7 6 5 4 3

Volume of antibody (μL) 40.0 36.7 33.3 30.0 26.7 23.3 20.0 16.7 13.3 10.0
Volume of diluent (mL) 10.0 9.2 8.3 7.5 6.7 5.8 5 4.2 3.3 2.5

25. Partial ELISA plates can be run using the reagent volumes
shown in Table 1.
26. Pipette into wells by placing the tips of the pipettes on the rim
of one side of the well. This will help prevent contamination
between wells in later steps.
27. Wells can be washed and aspirated using an automatic plate
washer, or the contents can be tipped out into a laboratory sink
and washed with a multichannel pipette as described earlier in
the protocol (see Subheading 3.5, step 12).
28. Diluent is provided as a 5 stock to be diluted to 1 in
deionized water. Add 10 mL diluent to 40 mL deionized
water. The excess 1 diluent can be saved for later steps in
the ELISA experiment.
29. The IL-8 ELISA and MSD assays can be performed separately,
as stimulation supernatants are frozen in two separate plates, to
minimize freeze-thawing.
30. Standards are prepared separately, so there is no need to add
PBS to columns 1 and 2 for the sample predilution steps.
31. A 1:50 dilution can be done in 96-well plates in two steps (1:10
followed by 1:5) as described, with a final (1:10) dilution being
done in the ELISA plate, giving an overall sample dilution of
1:500. However, these volumes can be changed, as long as the
desired final 1:500 dilution of sample is achieved. For accuracy,
it is advised to avoid pipetting volumes less than 5 μL.
32. A dilution of 1:500 was determined to be optimal for sample
testing at a single dilution.
33. If no rocking platform or tube rotators are available, mixing
can be done periodically over the 15 min by hand. Invert the
bottle gently, avoiding foaming.
34. This volume may change depending on the order in which
samples are diluted to make a 1:500 final dilution, as long as
the final volume in the ELISA plate is 100 μL (see Note 31).
35. After the sample incubation, all subsequent reagents should be
added at a consistent speed and order across the plate to make
the incubation time in each well as consistent as possible.
36. TMB is a chromogenic substrate that reacts with HRP, result-
ing in a colour change when the assay target is present in each
 Detecting Responses to Maternal Tdap Vaccination 339

sample. When pipetting this into the wells, lean the tips on the
opposite side of the well to previous steps to avoid cross-
contamination.
37. It is important to ensure the reaction is not stopped too early,
before wells with low-level concentrations have developed, or
too late, when the colour change becomes saturated leading to
inaccurate standard curve and/or plateauing of samples. It can
be helpful to use standard rows C and D as a guide; as long as
there is still a difference in colour between these wells, the
reaction can continue. If samples wells turn brightly colored
immediately and before the standard curve can develop, stop
the reaction as the sample concentrations will be too high and
out of range. If needed, repeat experiment with a new dilution.
38. Sulfuric acid is a highly corrosive chemical that is potentially
explosive in concentrated form. It can cause severe skin burns,
can irritate the nose and throat and cause difficulties breathing
if inhaled. Laboratory personal protective equipment should
be worn when using, including gloves and a lab coat.
39. The accompanying analysis software to the microplate reader
will produce a concentration of IL-8 in each sample by extra-
polating from the standard curve values, taking into account
the sample dilution factor.
40. Be careful not to disturb the blood layer below the plasma. It
can help to remove the final volume of plasma with a smaller
volume pipette, such as P200 micropipette.

References
1. Sawyer M, Liang JL, Messonnier N, Clark TA 5. Palmeira P, Quinello C, Silveira-Lessa AL et al
(2012) Updated recommendations for use of (2012) IgG placental transfer in healthy and
tetanus toxoid, reduced diphtheria toxoid, and pathological pregnancies. Clin Dev Immunol
acellular pertussis vaccine (Tdap) in pregnant 2012:985646
women--advisory committee on immunization 6. Byrne P, McGuirk P, Todryk S, Mills KHG
practices (ACIP). Morb Mortal Wkly Rep 62 (2004) Depletion of NK cells results in disse-
(7):131–135 minating lethal infection with Bordetella per-
2. Wiley KE, Massey PD, Cooper SC et al (2013) tussis associated with a reduction of antigen-
Pregnant women’s intention to take up a post- specific Th1 and enhancement of Th2, but
partum pertussis vaccine, and their willingness not Tr1 cells. Eur J Immunol 34
to take up the vaccine while pregnant: a cross (9):2579–2588
sectional survey. Vaccine 31(37):3972–3978 7. Higgins SC, Lavelle EC, McCann C et al
3. Donegan K, King B, Bryan P (2014) Safety of (2003) Toll-like receptor 4-mediated innate
pertussis vaccination in pregnant women in IL-10 activates antigen-specific regulatory T
UK: observational study. BMJ 349(9526): cells and confers resistance to Bordetella pertus-
g4219 sis by inhibiting inflammatory pathology. J
4. Gall SA, Myers J, Pichichero M (2011) Mater- Immunol 171(6):3119–3127
nal immunization with tetanus–diphtheria–- 8. Mahon BP, Sheahan BJ, Griffin F et al (1997)
pertussis vaccine: effect on maternal and Atypical disease after Bordetella pertussis respi-
neonatal serum antibody levels. Am J Obstet ratory infection of mice with targeted disrup-
Gynecol 204(4):334. e1 tions of interferon-γ receptor or
340 Thomas Rice and Beth Holder

immunoglobulin μ chain genes. J Exp Med 186 the first week of human life follow a robust
(11):1843–1851 developmental trajectory. Nat Commun 10
9. Kroes MM, Mariman R, Hijdra D et al (2019) (1):1–14
Activation of human NK cells by Bordetella 11. Thalen M, Van Den Ijssel J, Jiskoot W et al
pertussis requires inflammasome activation in (1999) Rational medium design for Bordetella
macrophages. Front Immunol 10:2030 pertussis: basic metabolism. J Biotechnol 75
10. Lee AH, Shannon CP, Amenyogbe N et al (2–3):147–159
(2019) Dynamic molecular changes during
 Chapter 18

Generation of a Universal Human Complement Source

by Large-Scale Depletion of IgG and IgM from Pooled Human
Plasma
Frances Alexander, Emily Brunt, Holly Humphries, Breeze Cavell,
Stephanie Leung, Lauren Allen, Rachel Halkerston, Elodie Lesne,
Elizabeth Penn, Stephen Thomas, Andrew Gorringe, and Stephen Taylor

Abstract
Complement is a key component of functional immunological assays used to evaluate vaccine-mediated
immunity to a range of bacterial and viral pathogens. However, standardization of these assays is compli-
cated due to the availability of a human complement source that lacks existing antibodies acquired either
through vaccination or natural circulation of the pathogen of interest. We have developed a method for
depleting both IgG and IgM in 200 mL batches from pooled hirudin-derived human plasma by sequential
affinity chromatography using a Protein G Sepharose column followed by POROS™ CaptureSelect™ IgM
Affinity resin. The production of large IgG- and IgM-depleted batches of human plasma that retains total
hemolytic and alternative pathway activities allows for improved assay standardization and comparison of
immune responses in large clinical trials.

Key words Complement system, IgG depletion, IgM depletion, Human plasma, FPLC

1 Introduction

Complement-mediated immunity is an important mechanism in

the host defense against invading pathogens and its measurement
is a valued tool in evaluating vaccine-derived immunity. Serum
bactericidal activity (SBA) has long been established as a correlate
of protection against meningococcal disease [1, 2] and many
immunoassays have been developed to understand the role of
complement in protection against a variety of bacterial and viral
pathogens [3–13]. One of the challenges to such functional immu-
noassays is the availability of a standardized complement source
that lacks intrinsic bactericidal and/or opsonophagocytic antibo-
dies against the pathogen of interest which could interfere with the

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_18,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

341
342 Frances Alexander et al.

assay. Vaccination or natural circulation of the bacteria or virus

within the human population means that many researchers use an
animal source of complement such as baby rabbit or guinea pig
serum that is claimed to lack this intrinsic activity [14, 15]. How-
ever, in the case of meningococcal SBA, the high specificity of
meningococcal factor H binding protein for human factor H is
known to result in elevated bactericidal titers when rabbit comple-
ment is used as the complement source [16]. Poor correlation in
bactericidal titers obtained using rabbit or human complement
with meningococcal serogroup A, W, and Y strains [17] also sug-
gests other species-specific interactions may play a role. Depending
on the pathogen of interest it may be possible to screen individuals
to identify a source of human serum with low cross-reactive anti-
bodies to the test strain. Alternatively, small batches (1–2 mL) of
human serum could be IgG-depleted with a Protein G column
immediately prior to addition to immunoassays [7, 18, 19]. While
more biologically relevant, these methods make it difficult to stan-
dardize assays and compare across different strains. Different
donors may be required for each target strain and day-to-day varia-
tion in small batches of IgG-depleted serum can easily be intro-
duced, if they are not fully characterized and adjusted accordingly
[20]. An alternative option to avoid the impact of pre-existing
antibody may be the reconstitution of a functioning complement
system using purified proteins [21].
We have previously described a method to deplete IgG from
300 mL batches of pooled lepirudin-derived human plasma that
retains key complement cascade components and total hemolytic
and alternative pathway activities [22]. This method allows for
greater availability of large batches of human complement for
improved assay standardization and comparison of immune
responses. Here, we present an updated method whereby pooled
hirudin-derived human plasma is depleted of both IgG and IgM in
200 mL batches by sequential affinity chromatography using Pro-
tein G Sepharose column followed by POROS™ CaptureSelect™
IgM Affinity resin.
Due to the delicate interplay between the complement and
coagulation cascades [23, 24], plasma rather than serum was cho-
sen as the complement source. Commonly used anticoagulants can
adversely affect the complement cascade [25] and in our initial
study lepirudin (Refludan™) derived plasma was chosen over hep-
arinized plasma as it showed the greatest total hemolytic activity
[22]. Unfortunately, due to lack of clinical demand for irreversible
non-heparin anticoagulant, the manufacturer ceased production of
lepirudin. The use of bivalirudin (Desirudin™) was trialed but the
reversible nature of the anticoagulant meant the complement
source was not suitable for assays with incubation times that
extended beyond the binding half-life unless excess anticoagulant
 Large-Scale Depletion of IgG and IgM from Human Plasma 343

was added to the assay. In this updated method recombinant hiru-

din, from which lepirudin is derived, is used as the anticoagulant.
Protein G Sepharose was chosen as it was able to completely
remove IgG below detectable levels from the plasma compared to
either Protein A or Protein L Sepharose [22]. We also found that
there was a significant reduction in IgM following IgG depletion
but no change in IgA levels [22]. While IgA is a poor activator of
complement [26], and so should not have intrinsic bactericidal/
opsonophagocytic activity, IgM is able to interact with C1q to
initiate complement activation [27]. The importance of also deplet-
ing IgM varies depending on the bacterial pathogen of interest.
Our initial studies showed the suitability of IgG-depleted human
plasma as a complement source in complement binding (measuring
deposition of complement proteins C3b/iC3b and C5b) and SBA
assays with serogroup B meningococci. However, unencapsulated
human pathogens such as Bordetella pertussis and nontypeable
Haemophilus influenzae are easily killed in vitro using human com-
plement due to activation by IgG [10, 28]. With the additional
removal of IgM, we have since shown complement-mediated
immunity to be important in protection against B. pertussis and
nontypeable H. influenzae [10, 11, 28, 29]. Removal of IgM also
allows assessment of functional immunity to commonly carried
encapsulated Gram-positive bacteria such as Group B Streptococcus
[9, 30] for which donors have high levels of both IgG and IgM.
The Poros CaptureSelect™ IgM resin has a monoclonal anti-
IgM antibody covalently linked to the resin. Screening of other
resins gave poor recoveries of IgM from plasma whereas Poros
CaptureSelect™ IgM resin gave consistent and reproducible
results, removing IgM below detectable levels. The resin is also
sanitizable with 70% ethanol (minimum of 12 h contact time) and
has a linear flow rate (not just volumetric flow rate) that is compati-
ble with the Protein G resin, allowing the plasma to run through
both columns in a single run and so reducing the time needed for
the double depletion.
As described in Brookes et al. [22], C1q and C5 are also
removed following IgG-depletion with Protein G Sepharose and a
NaCl gradient elution step is required to remove these proteins
from the affinity matrix [31]. While the loss of C5 during this step
was unexpected, the removal of C1q is likely due to its activation to
IgG bound to Protein G Sepharose [32, 33]. Both C1q and C5
were identified in the same eluted fractions, concentrated to
pre-depletion levels and added back into the final IgG- and
IgM-depleted plasma. No further loss of complement components
was identified after passing the IgG-depleted plasma through the
POROS™ CaptureSelect™ IgM Affinity column.
By using large pools of plasma from >20 volunteers it is possi-
ble to achieve consistency between batches of IgG- and IgM-
depleted human plasma, with similar concentrations of all tested
344 Frances Alexander et al.

complement proteins; as well as comparable total hemolytic and

alternative pathway complement activities as assessed by functional
analysis of serum bactericidal activity and antibody-mediated depo-
sition of complement proteins C3b/iC3b and C5b-9. However, it
should be noted that not all the complement components have
been exhaustively tested and it is possible that there are some that
are depleted during processing.

2 Materials

2.1 Plasma 1. Freshly drawn blood from volunteer donors (see Note 1), using
and Blood Plasma butterfly cannulas and 50 mL syringe.
Preparation 2. Recombinant hirudin mature variant from Pichia pastoris
freeze dried by request, >1
104 antithrombin units
(ATU)/mg (Creative BioMart). Reconstitute hirudin to
20 mg/mL using sterile water for irrigation (see Note 2).
Add sterile hirudin to 50 mL polypropylene tubes to give
2 mg per 50 mL whole blood.
3. Benchtop centrifuge: speeds up to 3000
g.

2.2 Sanitization 1. Water: HPLC grade, filtered at 0.4 μm.

of Chromatography 2. Phosphate buffered saline (PBS): 0.1369 M sodium chloride,
System and Columns 0.0081 M disodium hydrogen orthophosphate, 0.0015 M
(See Note 3) potassium dihydrogen phosphate, 0.0027 M potassium chlo-
ride pH 7.4  0.2. Prepared in-house.
3. 0.1 M sodium hydroxide, filtered.
4. 20% ethanol in HPLC grade water, filtered.
5. 70% ethanol in HPLC grade water, filtered.
6. 70% ethanol in HPLC grade water, filtered in spray bottle.

2.3 Chromatography 1. Water: HPLC grade filtered at 0.4 μm.

Buffers (See Note 3) 2. Phosphate buffered saline (PBS): 0.1369 M sodium chloride,
0.0081 M disodium hydrogen orthophosphate, 0.0015 M
potassium dihydrogen phosphate, 0.0027 M potassium chlo-
ride pH 7.4  0.2. Prepared in-house.
3. PBS supplemented with 1.5 M sodium chloride.
4. Elution buffer: 0.2 M glycine-HCl. Adjust pH to 2.7.

2.4 Chromatography 1. IgG depletion resin: 360 mL Protein G Sepharose . Linear flow
Columns rate ¼ 150–250 cm/h.
2. IgG depletion column: 400
50 mm h
internal diameter.
3. IgM depletion Resin: 70 mL POROS™ CaptureSelect™ IgM
Affinity resin. Linear flow rate ¼ 150 cm/h.
 Large-Scale Depletion of IgG and IgM from Human Plasma 345

4. IgM depletion column: 400
26 mm height
internal

diameter.
5. Manual switches
8, SRV-1 (GE/Cytiva).

2.5 Chromatography 1. AKTA FPLC, fraction collector placed within a Class II micro-
Systems biological safety cabinet (MSC II), at 22  C with UV monitor
of 280 nm, conductivity cell and pH monitor.
2. AKTA Purifier, in cold room (or 4  C) with UV monitor of
280, 254, 215 nm, conductivity cell, pH, and temperature
monitor.

2.6 Concentration 1. Sterilized dialysis tubing: Snakeskin 3.5 K molecular weight cut
of Plasma, off (MWCO) tubing, autoclaved in HPLC water.
Complement, and C1q 2. Polyethylene glycol 20,000 Da.

2.7 Ancillary Items 1. Collection vessels: 100 mL to 5 L borosilicate glass bottle,

dry-heat sterilized.
2. Collection vessel lids: Bottle lids individually autoclaved.
3. Aluminum foil: 200
200 mm folded and autoclaved.
4. Blunt ended forceps and scissors: dry-heat sterilized in first
instance, and autoclaved thereafter.
5. Sterile lint-free paper: Autoclaved.
6. Tubes: Polypropylene cryotubes with internal thread.
7. Cryogenic storage box with hollow prong dividers.
8. Dry ice and ethanol bath for snap freezing (see Note 4).

2.8 QC Testing— 1. Columbia agar with 5% horse blood media plates.

Bioburden 2. Chocolate agar with PolyViteX media plates.
3. Trypticase soy agar media plates.

2.9 QC Testing— 1. Standard SDS PAGE equipment.

Presence of Ig and C1q 2. Standard Western blot equipment.
3. Anti-human IgG, goat alkaline phosphatase.
4. Anti-human IgM, goat alkaline phosphatase.
5. C1q protein.

2.10 QC Testing— 1. Optilite IE700, Optimised Protein System.

Complement 2. Optilite C1 inactivator kit.
Component Activity
3. Optilite C3c kit.
4. Optilite C4 kit.
5. Optilite CH50 kit.
6. Optilite prealbumin kit.
346 Frances Alexander et al.

7. Optilite IgG kit.

8. Optilite IgM kit.
9. Radial Immunodiffusion Assays (RID).
(a) Total hemolytic complement.
(b) Alternative pathway hemolytic.
(c) C1 inactivator.
(d) C1q.
(e) C2.
(f) C3.
(g) C4.
(h) C4 binding protein.
(i) C5.
(j) C6.
(k) C7.
(l) C8.
(m) C9.
(n) Factor B.
(o) Factor H (β 1H).
(p) Factor I.
10. RID Reader, or microscope with micrometer and imager.

3 Method

All steps should be carried out using aseptic technique and where
possible within an MSC II to maintain sterility, the exterior of all
equipment should be sprayed with 70% ethanol (see Note 5). To
prevent activation of complement during the processing care must
be taken to avoid certain plastics (see Note 6).

3.1 Plasma 1. Collect venous human blood from multiple donors using but-
Preparation terfly cannulas and 50 mL syringes and immediately decant into
polypropylene centrifuge tubes containing hirudin (final con-
centration 0.04 mg/mL) on ice. Mix the blood gently with the
hirudin by inversion of the tube, without frothing.
2. Centrifuge at 3000
g for 10 min to separate red blood cells.
All particulate matter is removed, and the native plasma is snap
frozen in small batches, see Notes 7 and 8.
3. Defrost the small batches under running water and pool before
snap freezing in 45 mL aliquots.
4. On day of depletion, defrost the plasma in approximately
200 mL batches and centrifuge again at 3000
g for 10 min
to remove any particulates or clots, see Note 9. Take 5
1 mL
 Large-Scale Depletion of IgG and IgM from Human Plasma 347

aliquots of native plasma for QC, transfer remaining plasma to

4  C and use immediately, see Note 9.

3.2 Column Setup 1. Place the columns in a clamp within a laminar flow cabinet or
and Sanitization MSC II and dismantle and check the O rings and adaptors (top
and base) for integrity.
3.2.1 Column
Preparation 2. Wash the columns with sterile HPLC grade water (also sanitize
with 0.1 M sodium hydroxide if previously used) and immerse
all wet areas in sterile 70% ethanol for 12 h minimum. Immerse
the top adaptor separately (not within the column).
3. Discard the 70% ethanol and replace with a sterile 20% ethanol
wash before packing the columns with the appropriate resin, see
Note 10.
4. Place the top adaptor on a sterile surface (foil or paper) and
insert and seal when the column has been packed.

3.2.2 Sanitization 1. Assemble the chromatography systems with blank lines in place
of Chromatography of columns, with dual end connector switches.
Systems with Sodium 2. Set the pressure limit for system to 0.5 Pa.
Hydroxide (See Notes 11
3. Purge the system with water or 20% ethanol followed by water
and 12)
to neutralize the pH and remove all traces of salt from the
system. Monitor the pH readings as the valves and lines change;
a plateaued and neutral trace is required and may take 3–6
system volumes (sv) to achieve. A program can be used to
prime pumps and then purge all lines, valves, ports and fraction
collector in sequence.
4. Prime the system with 0.1 M sodium hydroxide and run a
program to sanitize all pumps, valves, ports, lines fraction
collector, and dead space for 1 h contact time.
5. Using a program (see Note 13), prime and flush with HPLC
grade water to remove sodium hydroxide from all pumps,
valves, lines, and dead space settings until the absorbance,
conductivity, and pH traces indicate a neutral pH or the trace
has stabilized. Ensure the sample line is included.
6. The cold room setting may require use of 20% ethanol to clear
Bernoulli effects and enhance clearance of edge effect of the
sodium hydroxide from the tubing and valves—this should be
run through as previously stated. Gather the collection lines in
a single sterile collection vessel and allow these lines to void
20% ethanol.
7. The system is now sterile and ready to be equilibrated. The
columns may now be connected into separate column ports if
they are already in 20% ethanol.
348 Frances Alexander et al.

3.2.3 Sanitization 1. Ensure columns are primed with 20% ethanol prior to transfer
of Columns at 4  C to 4  C.
with 70% Ethanol (See 2. Set pressure limit for 4  C system to 0.35 Pa.
Note 14)
3. Insert the columns into separate column ports in the system,
with 20% ethanol running at a low flow rate to avoid air in lines.
Reverse flow may be applied to the columns. Spray connection
points externally with 70% ethanol when connecting columns
to maintain sterility.
4. Run columns separately for 3–6 column volumes (cv) in 20%
ethanol or until conductivity has plateaued.
5. Prime the pump and flush bypass lines with 70% ethanol.
6. For the Protein G column: run 70% ethanol through the
column for 1 h contact time, and then flush with 20% ethanol.
Close the column top and bottom manual switch valves.
7. For the POROS™ CaptureSelect™ IgM column: equilibrate
the column for 12 h contact time then flush with 20% ethanol.
Close the column top and bottom manual switch valves.
8. Ensure both columns are equilibrated in 20% ethanol. This will
take a minimum of 6
cv.
9. When both columns have been sanitized and equilibrated back
to 20% ethanol, they are ready for use or storage.

3.3 IgG and IgM 1. Ensure the system and columns have been sanitized and are
Depletion of Plasma ready for use.
3.3.1 Equilibration 2. Set the pressure limits for the system and column to 0.35 Pa for
of the Chromatography the Protein G column and 0.5 Pa for POROS™ CaptureSe-
Systems and Columns lect™ column.
at 4  C 3. Open the manual switches and run 20% ethanol slowly through
both column ports.
4. Each column should be connected separately to the chroma-
tography system, see Note 15. Flush columns and all parts of
the system individually with water, including the sample pump.
Continue to flush each column with water until both the UV
and conductivity traces have stabilized.
5. Repeat step 4 using PBS for a minimum of 6
cv and ensure
pH trace has reached 7.4.

3.3.2 Loading 1. Ensure columns and system have been sanitized and equili-
the Plasma onto brated in PBS at 4  C and plasma is prepared for use (as de-
the Columns scribed in Subheading 3.2).
2. Set the pressure limits to 0.35 Pa for the system and columns.
3. Connect the columns together with the Protein G (for IgG
depletion) in front of the POROS™ CaptureSelect™ (for IgM
depletion) (see Fig. 1). Allow the pressure and monitor traces to
 Large-Scale Depletion of IgG and IgM from Human Plasma 349

Fig. 1 Schematic diagram showing the setup of the chromatography system and columns for IgG and IgM
depletion at 4  C. The system is equilibrated and primed with PBS using pump A or B. Plasma is loaded onto
the columns via pump C and passes first over the Protein G Sepharose column, then the POROS™
CaptureSelect™ IgM Affinity column. UV, conductivity, and pH monitors are used to assess the composition
of column breakthrough, which is then directed to waste or collected accordingly

adjust and stabilize by flushing the columns with PBS for 3
–
6
cv (use cv of both columns combined) and autozero the UV
detector once this is achieved. Use manual switch valves and
spray externally with 70% ethanol to ensure sterility when con-
necting the columns.
4. Aseptically place the collection lines into separate sterile collec-
tion vessels.
5. Set pump C running with the sample feed directed to the
system bypass waste and use pump A or B on a slower flow
rate, to pump PBS over the columns to waste. Once pumps are
running, apply the plasma to the columns using pump
C. Gradually raise the flow rate of pump C to approximately
5 mL/min.
350 Frances Alexander et al.

6. When all plasma has been loaded onto the system, transfer
pump C into PBS, taking care not to allow air into the lines
and clear plasma from the sample line and pump. Once cleared,
pump C can be switched off and pump A or B can be used to
direct PBS over the columns.
7. Monitor the UV 280 nm trace and collect the breakthrough
from 20 mAU until the traces indicate the breakthrough has
cleared and the monitors have plateaued again. Collect the
breakthrough until the 280 nm trace has dropped to
20 mAU. This indicates that large proteins have passed over
the columns and been collected.
8. Once the breakthrough is collected, flush the sample pump
with PBS to void, stop all pumps and disconnect columns
from the system and each other and seal with the manual
switches. Transfer columns to the second chromatography sys-
tem at 22  C and allow to acclimatize—this may take up to 3 h.
Sanitize the 4  C system as described in Subheading 3.2 (this
does not have to be carried out immediately after removal of
columns).
9. Keeping the breakthrough at 4  C, pool fractions before
removing QC samples (see Subheading 3.5) and concentrating
the depleted plasma (see Subheading 3.4.3).

3.4 C1q Elution, 1. Set up the fraction collector within an MSC II (to maintain
Concentration sterility) and sanitize both the system and fraction collector.
of Components Aseptically load 15 mL tubes (minus lids) into the fraction
and Reconstitution collector carousel and only insert this once the system has
been sanitized (avoid sodium hydroxide splashing into tubes).
3.4.1 Elution of C1q
at 22  C with Salt Gradient
2. Ensure columns are acclimatized to 22  C and equilibrated
(See Note 16)
in PBS.
3. With the system acclimatized to 22  C, set the pressure limits to
0.35 Pa for the system and columns (see Note 17), and auto-
zero the UV. Autozero with PBS flowing through the
bypass line.
4. Insert the Protein G column between column selection valves
(see Fig. 2) and start pump A (PBS) at a flow rate of 1 mL/min.
5. Apply a linear salt gradient of 0–1.5 M NaCl, at a flow rate of
1.6 mL/min for 1100 min (total 1760 mL). Collect 12 mL
fractions in an MSC II to maintain sterility.
6. Identify fractions containing C1q using the UV monitor trace
and SDS PAGE (see Note 18). Pool all fractions with C1q and
concentrate (see Subheading 3.4.3).
 Large-Scale Depletion of IgG and IgM from Human Plasma 351

Fig. 2 Schematic diagram showing the setup of the chromatography system and columns for the elution of
C1q at room temperature. Pump A is primed with PBS, and pump B is primed with PBS supplemented with
1.5 M NaCl. C1q is then eluted from the Protein G Sepharose using a salt gradient and collected in 12 mL
fractions. The UV monitor trace can be used in identifying which fractions contain eluted C1q

3.4.2 Elution of IgG 1. Ensure the salt gradient has been completed for the Protein G
and IgM for Reuse column and C1q has been collected (see Subheading 3.4.1).
of Column (See Note 19) Ensure the POROS™ CaptureSelect™ column is acclimatized
to room temperature.
2. Connect columns to the system in parallel (see Fig. 3).
3. Equilibrate the Protein G column separately back into PBS.
4. Apply the elution buffer, using pump C, to the Protein G
column for a maximum of 1 h or until a peak has eluted.
5. Apply PBS to the column until it is equilibrated before flushing
with water and 20% ethanol.
6. Collect the peaks manually and assess the protein profile by
SDS PAGE and western blotting, see Subheading 3.5.1, to
ensure antibodies have been eluted.
7. Repeat steps 3–6 for the Poros™ CaptureSelect™ column to
elute the IgM separately.
352 Frances Alexander et al.

Fig. 3 Schematic diagram showing the setup of the chromatography system and columns for the IgG and IgM
elution at room temperature. Pump C is primed with elution buffer and antibodies are eluted from the
respective columns into multiple collection lines. The UV monitor trace can be used in identifying which
fractions contain eluted antibody

8. Sanitize columns and chromatography system as described in

Subheading 3.2. Columns should be transferred back into 20%
ethanol promptly to prevent growth of contaminants.

3.4.3 Concentration 1. Set up sterile dialysis tubing, forceps, scissors, and lint-free
of Plasma and C1q paper in an MSC II.
and Recovery from Dialysis 2. Work aseptically in an MSC II but do not spray the outside of
Tubing tubing containing depleted plasma or C1q with 70% ethanol.
3. Fill the dialysis tubing with the depleted plasma breakthrough
or C1q fractions, void any residual air and knot the ends,
leaving 5 cm lengths of tubing post knots. Dry down the
outside of the tubing. Weigh the tubing once filled, for use as
a guide to the rate of concentration.
 Large-Scale Depletion of IgG and IgM from Human Plasma 353

4. Lay the sealed dialysis tubing onto a bed of chilled polyethylene

glycol (PEG) 20,000 Da and cover the tubing with PEG. Leave
at 4  C until desired volume is reached.
5. The depleted plasma breakthrough is concentrated to 90% of
the original volume (i.e., 180 mL for a 200 mL batch), while
C1q is concentrated to 10% of original plasma volume (i.e.,
20 mL for a 200 mL batch of native plasma). C1q may be
concentrated overnight at 4  C by placing in a bag/vessel and
enveloping the tubing in a thin film of concentrated PEG.
6. Wash excess PEG off the dialysis tubing with water and recover
contents aseptically within a MSC II, see Note 20. Snap freeze
components in a sterile container after removing 1 mL aliquots
for QC.

3.4.4 Reconstitution 1. Pre-chill the labeled internally threaded cryovials and racks.
of Complement with C1q 2. Defrost the required depleted plasma and the paired C1q.
and Aliquoting for Storage
3. Centrifuge the depleted plasma at 3000
g for 10 min to
remove all particulates or clots.
4. Within an MSC II, pool all depleted plasma and C1q into a
sterile glass Duran bottle on ice, mix gently and the aliquot into
50 mL falcon tubes (for ease of aliquoting) and store at 4  C for
immediate aliquoting.
5. Aseptically pipette into 0.5 mL or 1 mL aliquots as required.
Snap-freeze the vials (see Note 4) and store at 80  C.

3.5 Batch Quality 1. Samples are loaded onto 4–12% SDS-PAGE under reduced
Testing conditions; native plasma and reconstituted complement are
loaded at a 1:20 dilution, while C1q and all other samples are
3.5.1 SDS-PAGE
loaded neat.
and Western Blotting
2. One gel should be stained for protein and the subsequent gels
should be transferred to polyvinylidene difluoride (PVDF)
membrane by western blotting and probed with anti-IgG and
anti-IgM primary antibodies to show complete removal of the
immunoglobulins from the final depleted plasma.

3.5.2 Bioburden Check 1. Apply 20 μL of each sample to each of the three types of agar
plate.
2. Incubate at 37  C for 5 days at 5% CO2.
3. Observe contaminant growth.
4. Complete sterility is required for use of the depleted plasma in
functional assays therefore discard any batch where any micro-
bial contamination is seen.
354 Frances Alexander et al.

3.5.3 Confirmation 1. Measurement of individual complement protein levels and

of Complement Activity CH50 activity (see Fig. 4). All assays on the Optilite system
(The Binding Site) are performed according to the manufac-
turer’s instructions. All radial immundiffusion (RID) assays are
performed to the manufacturer’s instructions using either a
calibrated microscope or the available Digital RID Reader
(The Binding Site), which greatly improves the clarity of the
zones. Critical assays are albumin, IgG, IgM, CH50, AH50,
C1q, C3, C5, and FH, all others as listed in materials are
optional. Batches are acceptable if critical assays show 30%
of native plasma.
2. Functional immunoassays (see Table 1). Functional activity
against bacteria is measured using a serum bactericidal assay

Fig. 4 Comparison of percentage recovery compared to native plasma for three

batches of IgG- and IgM-depleted plasma for CH50 activity, C1q, C3, C5, and
FH. Bars represent the standard deviation of a minimum of three measurements,
dotted lines show 30% above and below full recovery, which represent accep-
tance cut-off levels

Table 1
Serum bactericidal activity reciprocal titers against Neisseria meningitidis strain NZ98/254 and
B. pertussis strain B1917 using three different batches of IgG- and IgM-depleted plasma. Reciprocal
bactericidal titers are accepted within 1 doubling dilution of the mean assigned titer calculated
through repeated testing; NT, not tested

N. meningitidis NZ98/254 B. pertussis B1917

(Serum from 1 year post-Bexsero (International pertussis
Depleted plasma batch vaccination) standard NIBSC 06/140)
Batch #1 32 128
Batch #2 32 NT
Batch #3 NT 128
Batch #4 32 128
 Large-Scale Depletion of IgG and IgM from Human Plasma 355

(SBA). Batches are acceptable if the killing titer of a control test

sera is within 1 dilution of the known mean titer. The SBA
against N. meningitidis strain NZ98/254 is performed as
described by Brookes et al. [22] while the SBA against
B. pertussis strain B1917 is performed as described by Lesne
et al. [11].

4 Notes

1. The volunteer blood and plasma have not been screened for
infectious agents and should always be handled wearing gloves,
safety glasses, and lab coat by staff who have been vaccinated
against Hepatitis B.
2. Add sterile water to the hirudin vial and leave for 30 min before
aliquoting the resuspended hirudin to required polypropylene
tubes. The tubes may be frozen if not used immediately. Safety
note: This is a concentrated and irreversible anticoagulant; care
must be taken when handling and the use of sharps should be
avoided.
3. All buffers must be filtered, prepared from the highest quality
grade components possible, and using HPLC grade water.
When preparing the buffers, a sterile lid is aseptically placed
on the filtered buffer bottle and covered with a sterile foil cover.
All buffers and solutions are prepared and filtered ahead of the
purification, the only one prepared fresh on the day is glycine
buffer for the antibody elution. Buffers are left to equilibrate to
required temperature for 3 h prior to use.
4. Snap freezing is performed using an ethanol bath containing
dry ice (solid carbon dioxide pellets). The solution may be as
cold as 70  C so care must be taken to not freeze fingers or
splash ethanol over the body, face, or clothes. The cryotubes
used should be selected to avoid both activation of the comple-
ment and plasma coagulation; the “O” ring and lid composi-
tion should be considered, and in our hands we found
internally threaded cryovials gave the better results. Tubes are
filled to 80% maximum volume, tightly sealed, and immersed
to the level of the liquid to ensure even and immediate super-
cooling of the contents. After reconstitution of the 200 mL
complement, the solution is subdivided into 4 or 5 prechilled
polypropylene tubes or glass vessels and remains chilled until
dispensed. The complement is dispensed as 0.2–1 mL aliquots
into pre-chilled and pre-labeled cryovials ready for use. The
tubes are tightly sealed and placed in racks sitting in the ethanol
bath sufficient to cover the contents but not the lid or the seal.
The use of cryogenic storage boxes with fixed hollow pronged
dividers allows rapid permeation of the freezing mixture
356 Frances Alexander et al.

between the tubes, and supercooling from beneath and within

the prong. As the depleted plasma freezes it changes color from
orange, becoming more yellow in colour.
5. To maintain the sterility of the final product: aseptic technique
should be practiced at all times to reduce the risk of contami-
nation, which could lead to activation of the complement
components and the interference with assays in which the
final product is used. 70% ethanol spray is used in preference
to 70% isopropanol (IPA) as the ethanol evaporates faster. The
70% ethanol spray is used directly on open chromatography
lines and valves when connecting the columns and lines to the
manual switch valves. Collection lines should be placed to the
base of the fraction vessel (glass borosilicate bottle) to lessen
the complement–air interface formed and so prevent comple-
ment activation.
6. Materials to avoid: Complement activation within the native or
IgG- and IgM-depleted plasma from different surfaces should
be avoided. Polyethylene, polyurethane and polystyrene should
be avoided wherever possible, or the contact time reduced to a
minimum. The use of polypropylene tubes and pipettes/past-
ettes should be used wherever possible, with borosilicate glass-
ware for larger volumes. Dialysis tubing made from
regenerated cellulose with a small MWCO (3.5 kDa) should
be used; the use of dialysis cassettes has resulted in the coagu-
lation and/or activation of the complement.
7. For plasma manipulations, use sterile polypropylene pastettes,
which leads to better product. It is important that the plasma/
complement is not frothed when manipulated; when pipetting
(by pastette), when transferring to vessels for loading onto
column, or when preparing dialysis tubing to recover
contents, etc.
8. Avoid hemolysis of the red blood cells and do not transfer any
red blood cells with the plasma. After centrifugation the plasma
should appear as a translucent/transparent solution layer above
the cell layer, with a particulate interface and top surface. Only
the upper plasma layer should be harvested, and the top sur-
face/skim should first be removed using polypropylene past-
ettes before accessing the plasma layer below, to avoid
contamination. All particulate matter must be gently removed
using a polypropylene pastette prior to freezing. If not, on
defrosting, the particulates will cause the plasma to clot/coa-
lesce. Blood should be processed and snap frozen with 1 h of
donation.
9. When defrosting the plasma, do so under running water and
centrifuge as soon as possible at 4  C to keep chilled. The
centrifugation step is required to clarify the plasma and remove
 Large-Scale Depletion of IgG and IgM from Human Plasma 357

particulates again, as a pellicle may form on the surface and base

of the tube and this must be removed.
10. The choice of chromatography system, columns and resin was
determined by the biocompatibility of all wettened surfaces
with the plasma and complement coagulation and activation
pathways. The scalability and reliability of the systems, col-
umns, and resins were an important consideration for down-
stream processing with regard to sanitization and
chromatography running conditions. The two chromatogra-
phy systems, independently equilibrated at 4  C and 22  C,
facilitated both systems being synchronized to optimize both
the C1q recovery at 22  C, and the column and system saniti-
zation. The transfer of chromatography systems to different
temperature-controlled environments risks electrical or
mechanical failure, while the independent cooling of columns
does not chill the system lines and surfaces, should the ambient
temperature not reflect the required temperature. Although
not recommended, a 1 cm buffer head space may be left
above the resin bed to prevent air entering the resin during
the run. Should this occur, the direction of flow can be reversed
and the air voided. Later chromatography systems have an air
sensor that will shut down the system, preventing air entering
the resin and damaging the resin.
11. System sanitization with sodium hydroxide should never be
performed with the columns in place. If this is not possible, a
manual inline switch should be present to protect the columns
and the columns removed from the inner surface of the
switches. A period of 1 h contact time should be allowed for
sanitization of all lines, valves, ports, and pumps in the system.
All fractionating lines should be sanitized as well as the fraction
collector to allow versatility if problems occur. Each column
port should have a blanking line connected by a manual switch
that is sanitized with the system in situ (i.e., port–blanking
line–switch–blanking line–port). When the 4  C system is sani-
tized and placed in 20% ethanol, the columns (already equili-
brated in 20% ethanol) are inserted in tandem for the base prep
elution.
12. Precipitation of crystals during sanitization of columns and
systems may occur if insufficient water is used between the
use of PBS/sodium chloride and ethanol (and vice versa).
Sanitization of the system with sodium hydroxide requires
not only the use of water but also a 20% ethanol flush within
the water wash to prevent crystal formation in the lines espe-
cially when sanitizing the cold room system. Complete flushing
of the system is required for each step when the columns are in
place as the precipitate will result in the chromatography run
being void due to the blockage requiring aseptic clearing.
358 Frances Alexander et al.

Precipitates are dispersed by sonicating the offending line or

valve in warm water while applying gentle pressure using a
hand-held syringe.
13. Programs to sanitize the system and columns may be used to
run the columns overnight/weekend to allow fast turnaround
of columns and systems while the complement base prep and
C1q are being concentrated, recombined and product snap-
frozen. Both columns may be sanitized independently by using
different flow rates. The pressure setting should reflect the
lower column criteria.
Manual switches should be placed either end of each col-
umn, and these, together with the blanking lines, must be
sanitized with sodium hydroxide, not just 70% ethanol. This
is done by rotation of the manual switch valves as the column
lines are inserted or removed.
Chromatography system programs are written specifically
for the configuration of the system used. Many systems have a
default pump wash and prime program, and this may be
incorporated into a bespoke program that reflects the configu-
ration of the system valves, ports, manual valves, bypass lines,
sample pump lines, and fractionating lines; together with the
tubing length and internal diameter. The program reflects the
time to equilibration by assessing UV monitor at 280, 254, and
215 nm, conductivity, and pH; all of which are allowed to
plateau and stabilize. These are checked after being stationary
for at least 5 min; when residual buffers may leach from the
tubing or connections. The program should be written to
cover lines sequentially, as any spike seen after a stationary
sequence results in lengthening of the time of wash, until the
spike is no longer seen. The only program that is not length-
ened is the sodium hydroxide wash which must be completed
within the hour to allow the water wash program to flow after
1 h contact time.
14. Sanitization of the columns with ethanol is best performed at
the end of a chromatography run. Dependent upon the time
left, the system may be sanitized with 0.1 M sodium hydroxide
ahead of the columns. However, the columns should never be
inline when sodium hydroxide is used. Columns are probably
best sanitized separately after transfer to the cold room system
after a depletion has been performed. This prevents the acci-
dental sanitization of the column with sodium hydroxide and
the degradation of the ligand. If required, the columns may be
run in reverse flow mode during sanitization—this allows the
slow buildup of residual protein to be cleared, which maintains
the flow rate and, in turn, lessens the back pressure thereby
prolonging the life of the column.
 Large-Scale Depletion of IgG and IgM from Human Plasma 359

15. Columns should only be added to the systems when the system
contains a sterile and compatible buffer/solution. The flow
rate should be set to a slow flow rate (e.g., 0.5–1.0 mL/
min). The upper blanking line switch should be opened, and
the blanking line uncoupled downstream of the valve so the
liquid flows through; these are held in one hand and sprayed
with 70% ethanol. The column top adaptor line is uncoupled
downside of the column top switch while the base switch
remains closed and the top line sprayed immediately with
70% ethanol. The top adaptor line is inserted into the down-
stream of the flowing blanking line. The base adaptor switch is
uncoupled to prevent back pressure. This process is repeated
with the base switches. The column switches should be period-
ically rotated into the system to ensure they become regularly
sanitized with sodium hydroxide.
16. C1q is in an equilibrium within the plasma and bound to
circulating antibody [34, 35]. When the base plasma prepara-
tion is collected in the column breakthrough, a proportion of
the C1q remains attached to the antibody which is now bound
to the column. Antibody binds to the column by the Fc region,
and so to avoid possible leaching of the antibody by the sudden
addition of high salt to the column, a slow and gentle salt
gradient is applied, thereby resulting in the C1q ionically dis-
sociating from the antibody, and consequently leaving the
antibody attached to the Protein G matrix. In our hands,
transferring the Protein G column from 4  C and equilibrating
to 22  C gives clear separation of the C1q from other proteins
along the gradient. The C1q peak is very minor (~40 mAU)
and may be easily missed (M Pangburn personal communica-
tion and Ref. 31) without analysis by SDS PAGE. Pure C1q is
observed in the second peak of the elution profile
corresponding to ~40% pump B elution. Alternatively, an iso-
cratic elution may be made to clear all proteins by stepping
directly to 40% high salt buffer concentration, and then 100%
to clear all. Elution of the C1q may be scaled up or down using
the ratio of the gradient over time for different column sizes.
17. Use of newer AKTA systems (or other chromatography sys-
tems) allows the selection of the column delta pressure and so
finely tunes the pressure exerted across the column without
being tripped by the internal pressures of the pump. This is
noticeable when eluting the C1q from the Protein G column at
low flow rates when the piston pump reverse flow creates a
pressure in excess of that set for the pressure limit.
18. Sampling of the fractions is made from every sixth tube and the
fractions collected for all six either side of the last C1q sample
identified by SDS PAGE. The C1q remains in the MSC II until
the SDS PAGE gel profile has been confirmed (using neat
360 Frances Alexander et al.

samples). Once the C1q profile is confirmed and fractions

pooled (samples taken), the concentration step occurs and
may take a few hours. Performing the C1q concentration in
narrow bore tubing increases the surface area–volume ratio
and, thereby, the rate of concentration. Agitation of the tubing
ensures a gradient across the dialysis membrane and
encourages the rate of concentrations. Analysis of the C1q
fractions should start as soon as possible in the morning as
this fraction should elute overnight and be ready for pooling
once the fractions can be identified. The peak profile is distinc-
tive for each column used.
19. IgG and IgM elution: The IgG is released from the column by
use of low pH elution releasing the Fc from the Protein G and
the IgM is recovered from the column by low pH isocratic
elution from the resin ligand monoclonal. Both columns are
cleared of bound antibody using a low pH glycine buffer and
pre-equilibrated back to PBS within 1 h contact time to pre-
serve the life of the column and ligand.
20. Recovery from dialysis tubing: when the concentrated compo-
nents are ready, the dialysis tubing is removed from the PEG
and gently washed with HPLC grade water to remove all
residual PEG. Wash the main length including the knots.
Avoid frothing the contents. Lay out the sterile lint-free
paper on the MSC II floor and place the tubing flat upon
it. Gently massage with a gloved hand to void the PEG film
from the dialysis tubing, working finger tips along the length of
the tubing and massaging the sides to ensure all contents are
released—especially the C1q. Pick up the tubing and allow the
contents of the tubing to fall to one end. Twist the tubing over
the contents allowing the empty tubing to fall away on the
other side of your hand. Spray the empty end with 70% ethanol
and, using a pair of sterile scissors, make an angled cut (not
right angled) and cut away the empty end above the fresh twist.
Pour the contents into a fresh container in one smooth
maneuver.

References

1. Goldschneider I, Gotschlich EC, Artenstein 3. Romero-Steiner S, Libutti D, Pais LB, Dykes J,

MS (1969) Human immunity to the Anderson P, Whitin JC, Keyserling HL, Car-
meningococcus. I. The role of humoral anti- lone GM (1997) Standardization of an opso-
bodies. J Exp Med 129:1307–1326. https:// nophagocytic assay for the measurement of
doi.org/10.1084/jem.129.6.1307 functional antibody activity against Streptococ-
2. Goldschneider I, Gotschlich EC, Artenstein cus pneumoniae using differentiated HL-60
MS (1969) Human immunity to the meningo- cells. Clin Diagn Lab Immunol 4:415–422.
coccus. II. Development of natural immunity. J https://doi.org/10.1128/cdli.4.4.415-422.
Exp Med 129:1327–1348. https://doi.org/ 1997
10.1084/jem.129.6.1327 4. Borrow R, Balmer P, Miller E (2005) Menin-
gococcal surrogates of protection - serum
 Large-Scale Depletion of IgG and IgM from Human Plasma 361

bactericidal antibody activity. In: Vaccine. Else- 13. Ji X, Olinger GG, Aris S, Chen Y, Gewurz H,
vier BV, Amsterdam, pp 2222–2227 Spear GT (2005) Mannose-binding lectin
5. Terajima M, Cruz J, Co MDT, Lee J-H, binds to Ebola and Marburg envelope glyco-
Kaur K, Wrammert J, Wilson PC, Ennis FA proteins, resulting in blocking of virus interac-
(2011) Complement-dependent lysis of influ- tion with DC-SIGN and complement-
enza a virus-infected cells by broadly cross- mediated virus neutralization. J Gen Virol
reactive human monoclonal antibodies. J 86:2535–2542. https://doi.org/10.1099/vir.
Virol 85:13463–13467. https://doi.org/10. 0.81199-0
1128/JVI.05193-11 14. Maslanka SE, Gheesling LL, Libutti DE,
6. Townsend K, Ladhani SN, Findlow H, Borrow Donaldson KBJ, Harakeh HS, Dykes JK,
R (2014) Evaluation and validation of a serum Arhin FF, Devi SJN, Frasch CE, Huang JC,
bactericidal antibody assay for Haemophilus Kriz-Kuzemenska P, Lemmon RD,
influenzae type b and the threshold of protec- Lorange M, Peeters CCAM, Quataert S, Tai
tion. Vaccine 32:5650–5656. https://doi.org/ JY, Carlone GM, Mills EL, Ashton FE,
10.1016/j.vaccine.2014.08.010 Diepevaen J, Bielecki J, Bybel M, Cloutier M,
7. Aase A, Herstad TK, Merino S, Brandsdal KT, Poolman JT (1997) Standardization and a
Berdal BP, Aleksandersen EM, Aaberge IS multilaboratory comparison of Neisseria
(2007) Opsonophagocytic activity and other meningitidis serogroup a and C serum bacteri-
serological indications of Bordetella pertussis cidal assays. Clin Diagn Lab Immunol
infection in military recruits in Norway. Clin 4:156–167. https://doi.org/10.1128/cdli.4.
Vaccine Immunol 14:855–862. https://doi. 2.156-167.1997
org/10.1128/CVI.00081-07 15. Yu J, Tostanosk LH, Peter L, Mercad NB,
8. Guttormsen H, Liu Y, Paoletti L (2008) Func- McMahan K, Mahrokhia SH, Nkolol JP,
tional activity of antisera to group B strepto- Liu J, Li Z, Chandrashekar A, Martine DR,
coccal conjugate vaccines measured with an Loos C, Atyeo C, Fischinger S, Burk JS, Slei
opsonophagocytosis assay and HL-60 effector MD, Chen Y, Zuiani A, Lelis FJN, Travers M,
cells. Hum Vaccin Vaccin 4:370–374 Habibi S, Pessaint L, Van Ry A, Blade K,
Brown R, Cook A, Finneyfrock B, Dodson A,
9. Herbert J, Thomas S, Brookes C, Turner C, Teow E, Velasco J, Zahn R, Wegmann F,
Turner P, Nosten F, LeDoare K, Hudson M, Bondzi EA, Dagotto G, Gebr MS, He X,
Heath PT, Gorringe A, Taylor S (2015) Jacob-Dolan C, Kirilova M, Kordana N,
Antibody-mediated complement C3b/iC3b Lin Z, Maxfiel LF, Nampanya F,
binding to group B streptococcus in paired Nityanandam R, Ventur JD, Wan H, Cai Y,
mother and baby serum samples in a refugee Chen B, Schmid AG, Weseman DR, Bari RS,
population on the Thailand-Myanmar border. Alter G, Andersen H, Lewi MG, Barou DH
Clin Vaccine Immunol 22:319–326. https:// (2020) DNA vaccine protection against
doi.org/10.1128/CVI.00803-14 SARS-CoV-2 in rhesus macaques. Science
10. Thomas SR, Leung S, Knox K, Wilkinson 369:806–811. https://doi.org/10.1126/sci
TMA, Staples KJ, Lestrate P, Wauters D, ence.abc6284
Gorringe A, Taylor SC (2018) Development 16. Granoff DM, Welsch JA, Ram S (2009) Bind-
of flow cytometric opsonophagocytosis and ing of complement factor H (fH) to Neisseria
antibody-mediated complement deposition meningitidis is specific for human fH and inhi-
assays for non-typeable Haemophilus influen- bits complement activation by rat and rabbit
zae. BMC Microbiol 18:167. https://doi. sera. Infect Immun 77:764–769
org/10.1186/s12866-018-1314-5
17. Gill CJ, Ram S, Welsch JA, De Tora L, Ane-
11. Lesne E, Cavell BE, Freire-Martin I, mona A (2011) Correlation between serum
Persaud R, Alexander F, Taylor S, bactericidal activity against Neisseria meningi-
Matheson M, van Els CACM, Gorringe A tidis serogroups a, C, W-135 and Y measured
(2020) Acellular pertussis vaccines induce using human versus rabbit serum as the com-
anti-pertactin bactericidal antibodies which plement source. Vaccine 30:29–34. https://
drives the emergence of Pertactin-negative doi.org/10.1016/j.vaccine.2011.10.068
strains. Front Microbiol 11:2108. https://doi.
org/10.3389/fmicb.2020.02108 18. Ray TD, Lewis LA, Gulati S, Rice PA, Ram S
(2011) Novel blocking human IgG directed
12. Mehlhop E, Whitby K, Oliphant T, Marri A, against the pentapeptide repeat motifs of Neis-
Engle M, Diamond MS (2005) Complement seria meningitidis lip/H.8 and Laz lipopro-
activation is required for induction of a protec- teins. J Immunol 186:4881–4894. https://
tive antibody response against West Nile virus doi.org/10.4049/jimmunol.1003623
infection. J Virol 79:7466–7477. https://doi.
org/10.1128/JVI.79.12.7466-7477.2005
362 Frances Alexander et al.

19. Giuntini S, Reason DC, Granoff DM (2012) of classical complement activation. Mol Immu-
Combined roles of human IgG subclass, alter- nol 47:2154–2160
native complement pathway activation, and 28. Brookes C, Freire-Martin I, Cavell B,
epitope density in the bactericidal activity of Alexander F, Taylor S, Persaud R, Fry N,
antibodies to meningococcal factor H binding Preston A, Diavatopoulos D, Gorringe A
protein. Infect Immun 80:187–194. https:// (2018) Bordetella pertussis isolates vary in
doi.org/10.1128/IAI.05956-11 their interactions with human complement
20. Beernink PT, Shaughnessy J, Braga EM, Liu Q, components. Emerg Microbes Infect 7:81.
Rice PA, Ram S, Granoff DM (2011) A menin- https://doi.org/10.1038/s41426-018-
gococcal factor H binding protein mutant that 0084-3
eliminates factor H binding enhances protec- 29. Staples KJ, Taylor S, Thomas S, Leung S,
tive antibody responses to vaccination. J Cox K, Pascal TG, Ostridge K, Welch L, Tuck
Immunol 186:3606–3614. https://doi.org/ AC, Clarke SC, Gorringe A, Wilkinson TMA
10.4049/jimmunol.1003470 (2016) Relationships between mucosal antibo-
21. Mutti M, Ramoni K, Nagy G, Nagy E, Szijártó dies, non-Typeable Haemophilus influenzae
V (2018) A new tool for complement research: (NTHi) infection and airway inflammation in
in vitro reconstituted human classical comple- COPD. PLoS One 11:e0167250. https://doi.
ment pathway. Front Immunol 9:2770. org/10.1371/journal.pone.0167250
https://doi.org/10.3389/fimmu.2018. 30. Le Doare K, Faal A, Jaiteh M, Sarfo F, Taylor S,
02770 Warburton F, Humphries H, Birt J, Jarju S,
22. Brookes C, Kuisma E, Alexander F, Allen L, Darboe S, Clarke E, Antonio M, Foster-
Tipton T, Ram S, Gorringe A, Taylor S (2013) Nyarko E, Heath PT, Gorringe A, Kampmann
Development of a large scale human comple- B (2017) Association between functional anti-
ment source for use in bacterial immunoassays. body against Group B Streptococcus and
J Immunol Methods 391:39–49. https://doi. maternal and infant colonization in a Gambian
org/10.1016/j.jim.2013.02.007 cohort. Vaccine 35:2970. https://doi.org/10.
23. Berends ETM, Kuipers A, Ravesloot MM, 1016/J.VACCINE.2017.04.013
Urbanus RT, Rooijakkers SHM (2014) Bacte- 31. Kolb WP, Kolb LM, Podack ER (1979) C1q:
ria under stress by complement and coagula- isolation from human serum in high yield by
tion. FEMS Microbiol Rev 38:1146–1171. affinity chromatography and development of a
https://doi.org/10.1111/1574-6976.12080 highly sensitive hemolytic assay. J Immunol
24. Amara U, Rittirsch D, Flierl M, Bruckner U, 122:2103–21011
Klos A, Gebhard F, Lambris JD, Huber-Lang 32. Björck L, Kronvall G (1984) Purification and
M (2008) Interaction between the coagulation some properties of streptococcal protein G, a
and complement system. Adv Exp Med Biol novel IgG-binding reagent. J Immunol
632:71–79 133:969–974
25. Mollnes TE, Brekke O-L, Fung M, Fure H, 33. Tao MH, Canfield SM, Morrison SL (1991)
Christiansen D, Bergseth G, Videm V, Lappe- The differential ability of human IgG1 and
gård KT, Köhl J, Lambris JD (2002) Essential IgG4 to activate complement is determined
role of the C5a receptor in E coli-induced oxi- by the COOH-terminal sequence of the CH2
dative burst and phagocytosis revealed by a domain. J Exp Med 173:1025–1028. https://
novel lepirudin-based human whole blood doi.org/10.1084/jem.173.4.1025
model of inflammation. Blood 34. Wang G, de Jong RN, van den Bremer ETJ,
100:1869–1877 Beurskens FJ, Labrijn AF, Ugurlar D, Gros P,
26. Vidarsson G, van der Pol W-L, van den Elsen Schuurman J, Parren PWHI, Heck AJR (2016)
JMH, Vilé H, Jansen M, Duijs J, Morton HC, Molecular basis of assembly and activation of
Boel E, Daha MR, Corthésy B, van de Winkel complement component C1 in complex with
JGJ (2001) Activity of human IgG and IgA immunoglobulin G1 and antigen. Mol Cell
subclasses in immune defense against Neisseria 63:135–145. https://doi.org/10.1016/j.
meningitidis serogroup B. J Immunol molcel.2016.05.016
166:6250–6256. https://doi.org/10.4049/ 35. Feinstein A, Richardson N, Taussig MI (1986)
jimmunol.166.10.6250 Immunoglobulin flexibility in complement
27. Kojouharova M, Reid K, Gadjeva M (2010) activation. Immunol Today 7:169–174
New insights into the molecular mechanisms
 Chapter 19

Assessment of Serum Bactericidal and Opsonophagocytic

Activity of Antibodies to Gonococcal Vaccine Targets
Evgeny A. Semchenko, Freda E.-C. Jen, Michael P. Jennings, and
Kate L. Seib

Abstract
There is no vaccine available to prevent Neisseria gonorrhoeae infection, however there is currently a high
level of interest in developing gonococcal vaccines due to the increasing number of cases and continuing
emergence of antimicrobial resistance worldwide. A key aspect of vaccine development is the investigation
of the functional immune response raised to the vaccine targets under investigation. Here, we describe two
assays used to assess the functional immune response raised against gonococcal vaccine targets: the serum
bactericidal assay (SBA) and the opsonophagocytic assay (OPA).

Key words N. gonorrhoeae, Vaccine, Antigen, Antibody, Complement, Polymorphonuclear leukocyte

(PMN), Neutrophil, Immune response, Serum bactericidal assay (SBA), Opsonophagocytic assay
(OPA)

1 Introduction

Neisseria gonorrhoeae, the causative agent of the sexually transmit-

ted infection gonorrhoea, is a major public health problem world-
wide for which a vaccine is urgently needed [1]. There is an
estimated incidence of 106 million cases of gonorrhoea worldwide
each year [2], infection rates are rising in many parts of the world
(e.g., 67% increase in cases in the USA [3], 80% increase in Australia
[4] over the past 5 years) and infection is increasingly hard to treat
due to emerging antimicrobial resistance [5]. There are various
challenges to developing a gonococcal vaccine, including the high
level of phase and antigenic variation of N. gonorrhoeae surface
structures, and the fact that there is no protective immunity follow-
ing infection, which means there are no established correlates of
protection to guide preclinical vaccine studies (reviewed in [6, 7]).
Key aspects of the general immune response to bacterial pathogens
include generation of antibodies which kill via complement-

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_19,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

363
364 Evgeny A. Semchenko et al.

Fig. 1 Schematic overview of the serum bactericidal assay (SBA) and the opsonophagocytic assay (OPA). For
both SBA and OPA, serial dilutions of antibody are prepared, then Neisseria gonorrhoeae (Ng; ~103 colony
forming units (CFU)) are added and incubated at 37  C/5% CO2 for 15 min. Then complement is added (and
polymorphonuclear leukocyte (PMNs) are added for OPA only) and incubated for 1 h. Samples (5 μL neat, and
dilutions) are then plated and incubated for 16 h. Finally, CFU counts and the SBA or OPA titer are determined

mediated lysis and opsonophagocytic killing. Assays such as serum

bactericidal assay (SBA) and opsonophagocytic assay (OPA) are
commonly used as a correlate or surrogate vaccine-induced immu-
nity [8]. The functional immune response raised in animals to
potential gonococcal vaccine antigens has been evaluated in several
studies using different variations of SBA and OPA [9–13] (see
Fig. 1).
The SBA measures antibody-mediated, complement-depen-
dent killing of a target bacteria. SBAs are performed by incubating
an appropriate bacterial target strain with serial dilutions of serum
(animal or human serum containing the antigen-specific antibody)
and a complement source [14]. Antibody binds to the bacterial
surface and activates the classical pathway of complement, resulting
in lysis death of the bacteria. The SBA titre for each serum is the
reciprocal of the serum dilution that results in
50% killing of the
bacteria, relative to the number of target cells present before incu-
bation with serum and complement. Although the role that bacte-
ricidal antibodies play in protecting against N. gonorrhoeae is
unknown, SBA titres correlate with immunity to the closely related
bacteria Neisseria meningitidis and have been used extensively in
development and licensing of meningococcal vaccines [15, 16].
The OPA assay measures antibody-mediated, complement-
dependent uptake and killing by phagocytic cells [17]. Similar to
the SBA, OPA are performed by incubating an appropriate bacterial
target strain with serial dilutions of serum and a complement
 Gonococcal Vaccine Evaluation 365

source, with the addition of phagocytic cells such as neutrophils.

The OPA titre for each serum is the reciprocal of the serum dilution
that results in
50% killing of the bacteria. The OPA has been used
to support pneumococcal vaccine licensure [18].

2 Materials

1. Benchtop block heater.

2. Biological Safety Cabinet Class II.
3. Calibrated plastic inoculation loops (10 μL).
4. CO2 incubator (set to 37  C and 5% CO2).
5. Microcentrifuge.
6. Temperature controlled centrifuge with swinging bucket rotor.
7. Dry ice–ethanol bath.
8. Light microscope.
9. Hemocytometer.
10. Sterile, round bottom 96-well plates with lids.
11. Polypropylene 1.5, 15, and 50 mL tubes.
12. EDTA blood collection tubes.
13. Serum clot activator tubes with gel separator.
14. Serological pipette.
15. Spectrometer and spectrometer cuvettes (0.5 mL).
16. GC agar plates supplemented with 1% (v/v) IsoVitaleX.
17. GC broth (GCB) supplemented with 1% (v/v) IsoVitaleX.
18. PolymorphPrep™.
19. Hanks’s Balanced Salt Solution (HBSS).
20. Human serum albumin (HSA).
21. Roswell Park Memorial Institute (RPMI) media.
22. Trypan Blue solution (0.4% w/v).
23. 2.5% (w/v) formaldehyde.
24. Sterile Phosphate Buffered Saline (PBS).
25. Sterile solution of 1.7% NaCl (w/v).
26. Sterile solution of 1 M CaCl2.
27. Sterile solution of 1 M MgCl2.
28. Sterile solution of 10% (w/v) saponin in GCB.
29. Sterile milli-Q H2O.
366 Evgeny A. Semchenko et al.

3 Methods

3.1 Preparation 1. Prepare and autoclave the following: PBS, 1.7% NaCl (w/v),
of Assay Media 1 M CaCl2, 1 M MgCl2, milli-Q H2O.
2. Prepare 10% (w/v) saponin in GCB and sterilize by passing
through a 0.22 μm filter.
3. Supplement RPMI with 0.15 mM CaCl2, 0.5 mM MgCl2 and
0.5% (w/v) HSA. Sterilize media through a 0.22 μm filter and
store in the fridge (see Note 1).
4. Bring media to room temperature (RT) before using in the
assays.

3.2 Preparation 1. N. gonorrhoeae is grown on GC agar plates overnight (16 h) at

of Bacterial Inoculum 37  C/5% CO2.
2. Harvest approximately 10 N. gonorrhoeae colonies from the
plate and restreak onto a fresh GC plate. Grow the bacteria
for 4 h (see Note 2).
3. Harvest bacteria from the plate with a sterile loop into 1 mL of
HBSS in 1.5 mL tube and resuspend by pipetting up/down
gently.
4. Wash cells by centrifuging the tube at 4000  g for 5 min at RT
to pellet bacteria. Discard the supernatant.
5. Resuspend cells in HBSS and measure optical density (OD) (see
Note 3).
6. Prepare bacterial inoculum in assay media to OD600 ¼ 0.001
(~1  105 colony forming units (CFU)/mL) (see Note 4).

3.3 Heat Inactivation 1. If using immune serum (mouse, rabbit, human etc.) as a source
of Immune Sera of antibodies for the assay, inactivate heat labile complement
proteins by incubating serum on a block heater (set to 56  C)
for 1 h, then briefly centrifuge (see Note 5). This step is not
necessary if using purified antibodies.
2. Test bactericidal activity of heat-inactivated serum against
N. gonorrhoeae using the same conditions as in the assay (i.e.,
heat-inactivated serum alone should not kill N. gonorrhoeae
after 1 h at 37  C).

3.4 Preparation 1. Collect blood from healthy volunteers (e.g., using Vacuette® Z
of Serum or equivalent Serum Clot Activator Tubes with Gel Separator)
for the Complement and clot as per the manufacturer’s instructions (e.g., invert
Source several times and incubate at room temperature for 10 min
before centrifuging at 2000  g for 10 min, room tempera-
ture). Transfer serum into tubes on ice before combining,
aliquoting and storing (see Note 6).
 Gonococcal Vaccine Evaluation 367

2. Adsorb serum against formaldehyde-fixed bacteria to remove

cross-reactive antibodies as described below (steps 3–10).
3. Grow the N. gonorrhoeae strain(s) being tested in the assay as in
Subheading 3.2, step 1. Harvest bacteria from a GC agar plate
with a sterile loop into 1 mL of PBS in 1.5 mL tube and
resuspend by pipetting up/down gently.
4. Wash cells by centrifuging at 4000  g for 5 min (4  C) to
pellet bacteria. Discard the supernatant. Resuspend cell pellet
in 1 mL of PBS. Repeat twice, then after final wash discard the
supernatant.
5. Resuspend the cell pellet in 1 mL of 2.5% (w/v) formaldehyde
(diluted in PBS) and incubate at room temperature for 15 min.
Wash cells (repeat step 4 above).
6. Resuspend fixed bacteria in PBS and measure optical density
(OD) as described in Note 3.
7. Prepare 1 mL of OD600 3 fixed bacteria and pellet cells at
4000  g for 10 min (discard supernatant).
8. Resuspend cell pellet in 5 mL of prepared pooled serum (see
Note 7). Incubate on ice for 1 h, then pellet the bacteria
(centrifuge precooled to 4  C). Collect and pass the superna-
tant through a syringe filter (0.22 μM) to remove residual cells.
9. Aliquot prepared serum into 1.5 mL tubes and snap freeze in
an ethanol–dry ice bath (see Note 8).
10. Test the bactericidal activity of the prepared serum using the
SBA assay method in Subheading 4 below, with the absence of
added antibody (see Note 9).

3.5 Preparation 1. Collect blood from healthy volunteers (e.g., using Vacuette®
of Polymorphonuclear- K3EDTA blood collection tubes or equivalent). Invert tubes
Cells several times to ensure adequate mixing with the anticoagulant.
2. Dispense 6 mL of PolymorphPrep™ or equivalent into a
15 mL tube and carefully overlay with an equal volume of
freshly collected blood (see Note 10).
3. Centrifuge the tube in a swinging-bucket rotor centrifuge at
500  g for 30 min (RT, no brakes). After centrifugation
carefully transfer the tube to the work bench.
4. Using a serological pipette, carefully collect the PMN buffy
coat and transfer it into a 50 mL tube (see Note 11).
5. Gently add HBSS to the PMN buffy coat, bringing the final
volume to 50 mL. Invert the tube several times to mix the
contents.
6. Centrifuge the tube at 200  g for 10 min.
7. Carefully remove the supernatant and resuspend the PMN
pellet in 10 mL of sterile milli-Q H2O.
368 Evgeny A. Semchenko et al.

8. Approximately 20 s later, add an equal volume of sterile 1.7%

NaCl and invert the tube several times to mix the contents.
9. Centrifuge the tube at 200  g for 10 min and remove the
supernatant.
10. Wash PMNs three times by resuspending the cell pellet in
HBSS and centrifuging at 200  g for 10 min.
11. After the final wash resuspend PMNs in assay media and mea-
sure cell concentration with haemocytometer (see Note 12).
12. Prepare a suspension with ~3  105 PMNs/mL.

3.6 SBA Assay 1. In a round-bottom 96-well plate prepare two-fold serial dilu-
tions of antibody (triplicate wells) (see Note 13).
2. To all appropriate test wells, add bacterial inoculum
(~103 CFU). Place the lid on the plate and incubate at
37  C/CO2 for 15 min.
3. While incubating the assay plate, plate out the bacterial inocu-
lum to determine the bacterial CFU (see Note 14). In a sepa-
rate 96-well plate, prepare serial dilutions of the bacterial
inoculum (e.g., add 5 μL of inoculum to 45 μL of GCB in
the first set of triplicate wells). Mix the contents by repeated
pipetting and transfer 5 μL into subsequent wells containing
45 μL of GCB (1/10 dilution). After performing several ten-
fold dilutions, plate 5 μL spots from all wells onto a GC agar
plates. Allow the spots on the agar plate to dry and then
incubate in 37  C/CO2 incubator overnight.
4. Remove the assay plate from the incubator and add required
volume of serum as a complement source (see Note 9). With-
out delay, make up the final volume to 100 μL with assay media
and return the plate to the 37  C/CO2 incubator. Incubate the
plate for 1 h.
5. While the assay plate is incubating, prepare a 96-well dilution
plate by adding 45 μL of GC broth into appropriate wells for
serial dilutions and plating of bacteria. Also, prepare two GC
agar plates for each set of triplicate wells (i.e., neat/undiluted
and 1/10 dilution).
6. Remove the assay plate from the incubator and mix the con-
tents of wells by gently, repeated pipetting. Transfer 5 μL from
all the wells in the assay plate into prepared 96-well dilution
plate. Plate out 5 μL spots on prepared GC agar plates. Allow
the spots on the agar plates to dry and then incubate in 37  C/
CO2 incubator overnight.
7. Next day, count CFU and determine the SBA titer, defined as
the lowest concentration/dilution of antibody that caused

50% killing relative to no treatment control.
 Gonococcal Vaccine Evaluation 369

3.7 OPA Assay 1. In a round-bottom 96-well plate prepare serial dilutions of

antibody (triplicate wells) (see Notes 13 and 15).
2. To all appropriate wells add bacterial inoculum (~103 CFU).
Place the lid on the plate and incubate at 37  C/CO2 for
15 min.
3. While incubating the assay plate, plate out the bacterial inocu-
lum (see step 3 of SBA assay).
4. Remove the assay plate from the incubator and to all wells
add appropriate volume of serum as a complement source (see
Note 9).
5. Add PMNs (105 cells/well) and assay media (final volume
100 μL/well). Return the assay plate to the 37  C/CO2 incu-
bator and incubate for 1 h.
6. While incubating, prepare a 96-well dilution plate (see step 5 of
Subheading 4, SBA assay).
7. Remove the assay plate from the incubator and add 10 μL of
saponin (10% solution) to all wells to lyse the PMNs. Vigor-
ously mix the contents of all wells by repeated pipetting
up/down. Transfer 5 μL from all the wells in the assay plate
into a prepared 96-well dilution plate and perform 1/10 dilu-
tions. Plate out 5 μL spots on prepared GC agar plates. Allow
the spots on the agar plates to dry and then incubate in 37  C/
CO2 incubator overnight.
8. Next day, count CFU and determine the OPA titre, defined as
the lowest concentration/dilution of antibody that caused

50% killing relative to no treatment control.

4 Notes

1. Complete assay media is stable for 1 week (stored in the fridge,

4  C).
2. Replating ensures bacteria in the assay are in the early log phase
(i.e., to minimize number of nonviable and dead bacteria in the
assay). Only light growth will be apparent.
3. Prepare a spectrophotometer and a cuvette. To a cuvette add
0.9 mL of PBS and take a “blank” measurement at 600 nm.
Add 0.1 mL of resuspended bacterial cells to the cuvette and
apply paraffin tape to seal the top. Invert the cuvette several
times to mix the contents. Take a reading of the cuvette and
record the absorbance. Calculate the OD by adjusting for the
dilution factor (1:10).
4. Bacteria will remain viable for several hours in assay media if
kept at room temperature or 37  C.
370 Evgeny A. Semchenko et al.

5. Heating the serum inactivates its innate bactericidal activity.

Serum volume should be no less than 50 μL per 1.5 mL tube.
Store heat-inactivated serum short term at 4  C (<2 weeks) or
long term at 80  C. Alternatively, to avoid repeated freeze–
thaw cycles that are damaging to antibodies add equal volume
of sterile glycerol (final concentration of glycerol is 50% (v/v)),
mix thoroughly and store at 20  C (solution will remain
liquid). Using serum with 50% (v/v) glycerol does not affect
bacterial viability or performance of assays.
6. Combine serum from several donors (consider using equal
volume and ratio of male/female donor). Serum can be ali-
quoted into appropriate tubes and stored in the 80  C
freezer.
7. The serum for the complement source can be scaled up if large
quantities are required. Use the same ratio of bacteria to serum.
8. Store serum prepared for the complement source at 80  C.
Complement is stable at 80  C for over 6 months and freeze–
thaw cycles are not recommended. When conducting assay,
thaw complement immediately prior to use in the assay and
keep on ice.
9. Typically, N. gonorrhoeae with nonsialylated lipooligosacchar-
ide (LOS) will tolerate 10–2.5% (v/v) complement (i.e., 100%
survival after 1 h). In SBA/OPA use highest concentration of
complement that does not induce killing of N. gonorrhoeae
after 1 h.
10. Bring both the blood and the PolymorphPrep™ to room
temperature before use.
11. After centrifugation of the blood and PolymorphPrep™, there
will be several distinct layers in the tube. Starting from the top
of the tube these are: blood plasma, mononuclear cell buffy
coat, first PolymorphPrep™ layer, PMN buffy coat, second
PolymorphPrep™ layer and erythrocyte layer.
12. In a 1.5 mL tube mix 20 μL of PMNs with equal volume of
Trypan Blue (0.4%) before applying to hemocytometer. Deter-
mine the viability and cell concentration. Typically, cell viability
will be >95%.
13. For example, pipet 40 μL of prepared serum into the top wells
and transfer 20 μL into subsequent wells containing 20 μL of
assay media. Include one set of wells that contain media only
for the no treatment control (i.e., bacteria and complement
only, no antibody). Additional control wells can be also added
(e.g., media only and serum only).
14. Label GC agar plates and leave with lids open at the back of the
Biological Safety Cabinet to dry (5–10 min) before plating
spots.
 Gonococcal Vaccine Evaluation 371

15. The OPA assay plate set up is identical to SBA, however this
assay may require additional control wells (i.e., PMN and
serum only, PMN only). The presence of complete comple-
ment in the OPA assay means that some killing may also be due
to bacterial lysis via SBA. As such it is recommended that SBA
and OPA assays are performed simultaneously if direct compar-
ison is required.

References
1. Gottlieb SL, Ndowa F, Hook EW 3rd, Deal C, blocking antibodies. Infect Immun 85:
Bachmann L, Abu-Raddad L et al (2020) Gon- e00898–e00816. https://doi.org/10.1128/
ococcal vaccines: public health value and pre- IAI.00898-16
ferred product characteristics; report of a 10. Semchenko EA, Day CJ, Seib KL (2020) The
WHO global stakeholder consultation, January Neisseria gonorrhoeae vaccine candidate NHBA
2019. Vaccine 38:4362–4373. https://doi. elicits antibodies that are bactericidal, opsono-
org/10.1016/j.vaccine.2020.02.073 phagocytic and that reduce gonococcal adher-
2. WHO (2012) Global incidence and prevalence ence to epithelial cells. Vaccines (Basel) 8:219.
of selected curable sexually transmitted infec- https://doi.org/10.3390/vaccines8020219
tions - 2008. World Health Organization, 11. Jen FE, Semchenko EA, Day CJ, Seib KL, Jen-
Geneva nings MP (2019) The Neisseria gonorrhoeae
3. Braxton J, Davis D, Emerson B, Flagg E, methionine sulfoxide reductase (MsrA/B) is a
Grey J, La Zetta et al (2018) Sexually transmit- surface exposed, immunogenic, vaccine candi-
ted disease surveillance 2017. Centers for Dis- date. Front Immunol 10:137. https://doi.
ease Control and Prevention, Atlanta. https:// org/10.3389/fimmu.2019.00137
www.cdc.gov/std/stats17/2017-STD-Surveil 12. Gulati S, Beurskens FJ, de Kreuk BJ, Roza M,
lance-Report_CDC-clearance-9.10.18.pdf. Zheng B, DeOliveira RB et al (2019) Comple-
Accessed 18 Jan 2021 ment alone drives efficacy of a chimeric antigo-
4. Kirby Institute (2018) HIV, viral hepatitis and nococcal monoclonal antibody. PLoS Biol 17:
sexually transmissible infections in Australia: e3000323. https://doi.org/10.1371/journal.
annual surveillance report 2018. The Kirby pbio.3000323
Institute, UNSW Sydney 13. McQuillen DP, Gulati S, Rice PA (1994)
5. Unemo M (2020) Optimizations to keep Complement-mediated bacterial killing assays.
gonorrhoea treatable and reduce antimicrobial Methods Enzymol 236:137–147. https://doi.
resistance selection. Nat Rev Urol org/10.1016/0076-6879(94)36013-8
17:609–610. https://doi.org/10.1038/ 14. Findlow J, Balmer P, Borrow R (2019) A
s41585-020-00378-5 review of complement sources used in serum
6. Edwards JL, Jennings MP, Apicella MA, Seib bactericidal assays for evaluating immune
KL (2016) Is gonococcal disease preventable? responses to meningococcal ACWY conjugate
The importance of understanding immunity vaccines. Hum Vaccin Immunother
and pathogenesis in vaccine development. Crit 15:2491–2500. https://doi.org/10.1080/
Rev Microbiol 42:928–941. https://doi.org/ 21645515.2019.1593082
10.3109/1040841X.2015.1105782 15. Borrow R, Aaberge IS, Santos GF, Eudey TL,
7. Rice PA, Shafer WM, Ram S, Jerse AE (2017) Oster P, Glennie A et al (2005) Interlaboratory
Neisseria gonorrhoeae: drug resistance, mouse standardization of the measurement of serum
models, and vaccine development. Annu Rev bactericidal activity by using human comple-
Microbiol 8(71):665–686. https://doi.org/ ment against meningococcal serogroup b,
10.1146/annurev-micro-090816-093530 strain 44/76-SL, before and after vaccination
8. Plotkin SA (2010) Correlates of protection with the Norwegian MenBvac outer mem-
induced by vaccination. Clin Vaccine Immunol brane vesicle vaccine. Clin Diagn Lab Immunol
17:1055–1065. https://doi.org/10.1128/ 12:970–976. https://doi.org/10.1128/
CVI.00131-10 CDLI.12.8.970-976.2005
9. Semchenko EA, Day CJ, Seib KL (2017) MetQ 16. Lucidarme J, Louth J, Townsend-Payne K,
of Neisseria gonorrhoeae is a surface-expressed Borrow R (2019) Meningococcal
antigen that elicits bactericidal and functional serogroup A, B, C, W, X, and Y serum
372 Evgeny A. Semchenko et al.

bactericidal antibody assays. Methods Mol Biol 18. Jódar L, Butler J, Carlone G, Dagan R,
1969:169–179. https://doi.org/10.1007/ Goldblatt D, K€ayhty H et al (2003) Serological
978-1-4939-9202-7_12 criteria for evaluation and licensure of new
17. Dwyer M, Gadjeva M (2014) Opsonophago- pneumococcal conjugate vaccine formulations
cytic assay. Methods Mol Biol 1100:373–379. for use in infants. Vaccine 21:3265–3272.
https://doi.org/10.1007/978-1-62703-724- https://doi.org/10.1016/s0264-410x(03)
2_32 00230-5
 Chapter 20

Opsonophagocytic Killing Assay to Measure Anti–Group A

Streptococcus Antibody Functionality in Human Serum
Helen Wagstaffe , Scott Jones, Marina Johnson, and David Goldblatt

Abstract
The opsonophagocytic killing assay (OPKA) is designed to measure the functionality of strain-specific
antibodies and, therefore, assess protective immunity or the immunogenicity of Group A Streptococcus
(GAS) (type A Streptococcus pyogenes) vaccines. Opsonization of GAS for phagocytosis is an important
mechanism by which antibodies protect against disease in vivo. The Opsonophagocytic Index or Opsonic
Index (OI) is the estimated dilution of antisera that kills 50% of the target bacteria. Here, we describe the
protocol of the standardized GAS OPKA developed by Jones et al., 2018.

Key words Antibody, Phagocytosis, Opsonophagocytic killing assay, S. pyogenes, HL-60 cells, Baby
rabbit complement, In vitro, Functional assay

1 Introduction

Group A Streptococcus (GAS), a Gram-positive bacterium, can

cause asymptomatic infections, mild and severe disease and is the
leading causative agent of pharyngitis in children and adolescents
worldwide [1, 2]. There is currently no licensed vaccine despite a
large global burden of disease resulting in more than 0.5 million
deaths per year [3]. A lack of standardized immunoassays to mea-
sure postvaccination GAS immunity has hindered vaccine develop-
ment to date; the WHO GAS Vaccine Development Technology
Roadmap highlighted this as a key priority activity for vaccine
development in 2019 [3]. The Lancefield assay, widely used in the
past, measures growth and survival of GAS in fresh human or
animal blood (plus immune or nonimmune serum) [4]. Interdonor
variation in neutrophil and complement activity disqualify this type
of assay from full standardization and therefore use in large, multi-
center vaccine studies. The opsonophagocytic killing assay (OPKA)
described in this chapter, first developed by Jones et al., 2018 [5], is
an important tool for the reliable detection and quantification of

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_20,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

373
374 Helen Wagstaffe et al.

functional anti-GAS antibodies, which is important for the

continued development of GAS vaccines.
The GAS OPKA assay makes use of baby rabbit complement
(BRC) as a standard source of complement and continuous human
promyelocytic leukemia cell line (HL-60) as an exogenous source
of phagocytic cells. This method removes the major sources of
variation present in many current and past functional assays used
to measure GAS vaccine immunogenicity. The assay was optimized
for seven clinically relevant GAS strains of emm-type 1, 12, and
6, but other strains can be tested and optimized to be used in the
assay.

2 Materials

Prepare all reagents at room temperature. Pyrogen-free or deio-

nized water is required for buffer preparations (unless stated other-
wise). Volumes can be adjusted accordingly.
1. HL-60 cell culture medium: 500 ml RPMI 1640 medium,
50 ml fetal calf serum (FCS), 5 ml L-glutamine (200 mM).
2. Freezing medium: 9 ml FCS, 1 ml dimethyl sulfoxide
(DMSO).
3. Todd-Hewitt-yeast extract broth (THY broth): 6 g Todd-
Hewitt broth, 1 g yeast extract, 200 ml water. Mix until all
components are dissolved and filter-sterilize using a 0.22 μm
bottle-top filter into a sterile 200 ml bottle.
4. Bacteria storage buffer: 3 g tryptone soya broth, 0.5 g glucose,
10 ml glycerol, 100 ml water. Mix until all components are
dissolved and autoclave.
5. THY agar (THY plates): 48 g Todd-Hewitt broth, 8 g yeast
extract, 24 g bacteriological agar, 1600 ml water. Autoclave
and bring to 50  C in a water bath. Pour 25 ml agar into
100  100 mm square agar plate. Leave on a flat surface to
dry for ~20 min. Invert stacked plates and store at 4  C for up
to 1 month.
6. THY overlay agar: 48 g Todd-Hewitt broth, 8 g yeast extract,
12 g bacteriological agar, 1600 ml water. Make fresh on day of
assay. Autoclave and store in a 50  C water bath until required.
7. 1% gelatin solution: 4 g gelatin, 400 ml water. Dissolve gelatin
in water and autoclave. Store at room temperature (RT) for
2 months.
8. 2,3,5-Tetraphenyltetrazolium chloride (TTC) stock: 1.25 g
TTC, 50 ml water. Dissolve TTC in 40 ml water and make up
to a final volume of 50 ml with the remaining water.
 Group A Streptococcus OPK Assay 375

Sterile-filter with a 0.22 μm filter. Liquid should have a yellow-

ish tinge. Store at 4  C for 1 month.
9. Opsonization buffer (OPS buffer): 5 ml FCS, 40 ml 1 Hank’s
Buffered Salt Solution (HBSS, +Ca/Mg), 5 ml 1% gelatin
solution. Prepare on day of assay and discard after use.
10. 96-well round-bottom microtiter plates.
11. GAS bacteria isolates.
12. HL-60 Cells, preferred source; American Type Culture Collec-
tion (ATCC) (see Note 1).
13. BRC, preferred source; Pel-Freez (see Note 1).

3 Methods

3.1 HL-60 Master HL-60 cells are promyelocytic leukemia cells which are differen-
Stock Propagation tiated to a neutrophil-like cell with 0.8% dimethylformamide
from Master Cell (DMF) for use in the assay. All cell culture is to be undertaken in
Bank Stock a culture hood under sterile conditions. Warm culture medium to
37  C before use.
1. Add 10 ml culture medium to a 15 ml centrifuge tube. Thaw
the master stock of HL-60 cells rapidly by swirling in a 37  C
water bath. Transfer the cells into the culture medium.
2. Centrifuge the tube at 350  g for 5 min at room temperature
(RT). Remove the supernatant.
3. Resuspend the cells in culture medium to a final concentration
of 2  105 per ml and transfer to a culture flask. Incubate at
37  C, 5% CO2.
4. When the cell density reaches 5  105 cells per ml, add further
fresh culture medium to readjust the concentration to 2  105
per ml (see Note 2). Use multiple flasks when the volume
reaches the capacity of the flask. To avoid risk of contamination,
the medium must not reach the cap when the flask is
horizontal.
5. When the concentration reaches 5  105 cells per ml in
10 flasks, freeze the cells. Transfer the contents of the flasks
into 50 ml centrifuge tubes and spin at 350  g for 5 min.
6. Extract and discard the supernatant, being careful not to dis-
turb the pellet.
7. Add 2.5 ml freezing medium to each 50 ml centrifuge tube and
resuspend each pellet.
8. Combine the contents of all tubes in a single culture flask and
aliquot 1 ml into cryovials (each vial should contain
1  107 cells).
376 Helen Wagstaffe et al.

9. Transfer cryovials into controlled-rate freezing containers and

into a 80  C freezer.
10. After a minimum of 2 h, transfer the cryovials into a liquid
nitrogen tank (maximum 1 week at 80  C).

3.2 HL-60 Working 1. Take a 1 ml aliquot of HL-60 master stock prepared as in

Stock Preparation Subheading 3.1, from the liquid nitrogen tank and transfer
from Master Stock into a controlled-rate freezing container.
2. Defrost cells quickly by swirling in a 37  C water bath.
3. Transfer into a 50 ml centrifuge tube containing 50 ml of fresh
culture medium.
4. Spin at 350  g for 5 min at RT. Pour off supernatant and
resuspend pellet in 10 ml fresh culture medium.
5. Transfer into a tissue culture flask and incubate at 37  C, 5%
CO2, overnight.
6. Monitor the cell density using a hemacytometer and readjust to
2  105 per ml in warm culture medium. Repeat after 2–3 days.
7. Maintenance, every 3–4 days, harvest cells into a 50 ml centri-
fuge tube, spin at 350  g for 5 min at RT, pour off supernatant
and resuspend pellet in warm culture medium. Reseed the
required number of flasks with 2  105 cells per ml. Cell
density must remain 1.2  106 per ml.

3.3 Differentiation This section describes the differentiation method in 200 ml

of HL-60 Cells volumes. To differentiate other volumes, the concentration of
cells and DMF remains constant.
1. Count the cells and calculate the volume required to resuspend
at 4  105 cells per ml.
2. Cell viability must be 90% prior to differentiation.
3. Add 1.6 ml DMF to 175 ml fresh culture medium (this will
give a final concentration of 0.8% in 200 ml).
4. Spin cells at 350  g for 5 min at RT and pour off the
supernatant.
5. Resuspend cells in 25 ml culture medium and add to the
175 ml of culture medium containing DMF.
6. Incubate the cells at 37  C, 5% CO2 for 5 or 6 days. Do not
feed the cells during this time.
7. Harvest cells for use in the assay (see Subheading 3.8).

3.4 Bacterial Master All bacterial culture to be undertaken in a bacterial culture hood
Stock Preparation under sterile conditions.
and Maintenance Frozen vials of GAS strains are stored at 80  C. To maximize
the integrity of the bacterial master stocks, the vials should remain
 Group A Streptococcus OPK Assay 377

frozen at all times by keeping on dry ice. Vials can be stored for
several years.

3.4.1 From Clinical Swabs can be stored at 4  C for up to 4 months depending on the
Swabs strain or for longer time if stored at 80  C.
1. Streak the swab along the edge of a horse blood agar plate.
2. Rotate the plate by 90 , streak using a sterile loop and repeat
once more using the same loop (optional, use a new loop, pass
through the final streak and streak on clean zones toward the
center of the plate to ensure isolated colonies).
3. Incubate overnight at 37  C, 5% CO2 until colonies appear.
4. To create a master stock, harvest enough colonies on a loop and
mix well in 1 ml of storage buffer before storing at 80  C.

3.4.2 From Plated Plated bacteria can be stored at 4  C for up to 4 weeks, depending
Bacteria on the storage condition. To create a master stock, harvest enough
colonies on a loop and mix well in 1 ml of storage buffer before
storing at 80  C.

3.5 Bacterial Day 1:

Expansion
1. Remove master vial from freezer and streak a fleck of the frozen
bacteria onto 2 blood agar plates. Immediately return master
vial to cold storage.
2. Label the plates with strain name and date and incubate at
37  C, 5% CO2 overnight.
Day 2:
3. Add 30 ml THY broth to two 50 ml Falcon tubes, label them A
and B plus the bacterial strain name.
4. Harvest a single colony from one blood agar plate using an
inoculating loop.
5. Add harvested bacteria to tube A, repeat for tube B. Use 50 ml
THY broth alone to act as a blank.
6. Measure OD600 of the broth before and after inoculation, and
of the blank. The ODs should be comparable.
7. Incubate tubes at 37  C, 5% CO2. The caps must be kept loose
to allow gas exchange. Incubate for 2–3 h.
8. Label the desired number of sterile 1.5 ml microcentrifuge
tubes per serotype with the bacterial strain and date.
9. Check the OD of the tubes after 2 h and then every hour/half-
hourly until the OD reaches between 0.5 and 0.6 (bacteria are
in exponential growth phase) (see Note 3). OD of blank must
be 0.02.
378 Helen Wagstaffe et al.

10. On reaching required OD, harvest the top 10 ml of broth

(do not mix) and gently mix 1:1 with storage buffer.
11. Aliquot 0.5 ml into the labelled microcentrifuge tubes and
transfer to 80  C freezer, this is the working bacterial stocks.

3.6 Human Serum Human serum must be handled in accordance with local guidelines
Sample Preparation and stored at 80  C. Prior to testing, the serum must be thawed at
RT and heat-inactivated by incubating at 56  C for a minimum of
30 min. Allow the samples to cool to RT before use in an assay.
Samples can be stored at 4  C for up to 1 month during the testing
process to avoid repeated freeze–thaw cycles (if sample requires
retesting).

3.7 Preparing Caution must be exercised when handling BRC as its components
Working Aliquots are extremely heat-sensitive. When receiving BRC stock from the
of Baby Rabbit supplier, ensure that the contents are entirely frozen and transfer to
Complement (BRC) 80  C storage immediately. There is the potential for high degrees
of variation in performance between BRC lots; therefore, prospec-
tive lots need screening before use in an assay.
1. Defrost each bottle of BRC in cold water with constant agita-
tion, for example, inside an ice box containing a mixture of ice
and cold water, placed on top of an orbital shaker.
2. Label tubes with the lot number and aliquot date. Place tubes
on ice to cool. Place the bottle of BRC on ice as soon as it has
thawed.
3. Quickly aliquot the BRC placing the aliquots in ice until fin-
ished (this can be done in a culture hood to ensure sterility).
4. Store all aliquots at 80  C until required.

3.8 Preparation This procedure describes the process for one 200 ml flask; volumes
of Differentiated HL-60 can be altered accordingly. The concentration of cells required for
Cells (for Use use in the assay is 1  107 per ml. Do not use cells >1.5  106 per
in the Assay) ml at initial count. See Note 4 for more detail on HL-60 cell
acceptance criteria.
1. Resuspend differentiated cells by shaking gently to ensure
equal distribution throughout the flask. Count cells and
check viability by trypan blue exclusion. Cell viability must be
80%.
2. Decant cells into four 50 ml centrifuge tubes under sterile
conditions. Spin tubes for 5 min at 350  g at RT.
3. Remove the supernatant and resuspend each 50 ml of cells in
50 ml HBSS (without Ca/Mg). Centrifuge for 5 min at
350  g at RT.
4. Remove the supernatant and resuspend each 50 ml tube in
50 ml HBSS (+Ca/Mg).
 Group A Streptococcus OPK Assay 379

5. Centrifuge for 5 min at 350  g at RT.

6. Remove the supernatant and resuspend cells at 1  107 cells per
ml in OPS buffer. Count the cells and assess the viability. Cell
viability must be 80%.
7. Store at room temperature until required.

3.9 Preparation 1. Remove one vial of bacteria from the 80  C freezer and
of Bacteria (for Use defrost.
in the Assay) 2. Centrifuge the tube for 2 min at 13,000  g at RT.
3. Remove the supernatant using a pipette, being careful not to
dislodge the pellet of cells.
4. Add 1 ml of OPS buffer to each tube and vortex. Centrifuge
the tube at 13,000  g for 2 min at RT.
5. Carefully remove the supernatant and resuspend the pellet in
0.5 ml OPS buffer.

3.10 Procedure Optimal Dilution Experiment 1 (OD1) will determine a rough

for Determining estimate for the dilution of bacteria required in the assay without
Optimal Dilution the inclusion of serum and controls. Optimal Dilution Experiment
of Bacteria 2 (OD2) includes human serum with associated control A and B to
determine more accurately the optimal dilution of bacteria
required. Procedure is carried out in 96-well round-bottom micro-
titer plates. OD assays are run in the presence of BRC and HL-60
cells as they can influence the bacterial growth.

3.10.1 Optimal Dilution 1. Prepare THY overlay agar.

Experiment 1 2. Remove working aliquot of BRC (Subheading 3.7) from the
80  C freezer and defrost on ice.
3. Dilute bacteria (working stock prepared as Subheading 3.9)
twofold (1:1) by mixing 75 μl bacteria with 75 μl of OPS buffer
in a single well (row A) of a microtiter plate (plate 1). Add
120 μl of OPS buffer to rows B to H. Two columns of the plate
will be required per serotype see Fig. 1.
4. Prepare fivefold serial dilutions (1:9, 1:49, 1:249, 1:1249,
1:6249, 1:31,249, 1:156,249) by diluting 30 μl of diluted
bacteria from row A with 120 μl OPS buffer in row B. Repeat
to column H (Fig. 1).
5. In a second microtiter plate (plate 2), add 20 μl OPS buffer to
two columns for each serotype of bacteria to be tested.
6. Transfer 10 μl of diluted bacteria prepared in the first plate to
the appropriate wells in the second plate (in duplicate). Plate
1 can be discarded.
7. Incubate plate 2 at RT on a mini-orbital shaker for 30 min at
700 rpm.
380 Helen Wagstaffe et al.

Fig. 1 Optimal dilution 1 experiment procedure. Bacteria is serially diluted in a plate 1 then transferred to plate
2 when BRC and HL-60 cells are added. After incubation, the mixture is spotted onto THY agar and incubated
overnight

8. Following the incubation period, add 10 μl of active BRC to

each column in use on plate 2. BRC may be prediluted in OPS
buffer prior to addition to the plate to ascertain a final in-well
predetermined concentration. The optimal BRC concentration
was determined for seven different clinically relevant GAS
strains (see Note 5).
9. Resuspend the HL-60 cells. Add 40 μl to each column in use
on plate 2.
10. Incubate on a mini-orbital shaker at 37  C, 5% CO2 for 90 min
at 700 rpm. Incubate multiple plates in a single layer to main-
tain equal CO2 exposure.
11. During this time, remove THY agar plates from the fridge,
remove lids, and lay on the bench to dry for 30 min to 1 h.
12. Place plate 2 on ice for 20 min to halt the phagocytic process.
13. Vortex the plate at a low speed. Using a manual eight-channel
pipette, spot 10 μl from each column of plate 2 side by side on a
THY plate. Tilt the plate to the left and right so the spots
measure approximately 1–1.5 cm across, being careful not to
 Group A Streptococcus OPK Assay 381

allow the spots to merge. If there are multiple serotypes to be

assessed, use a separate agar plate for each serotype.
14. Repeat this procedure for the next column, spot next to the
previous column on the agar plate. Repeat until there are three
columns of spots on the agar plate.
15. Leave agar plates at RT for ~20 min to allow the spots to dry
(see Note 6).
16. Remove the overlay agar from the water bath. Measure out the
amounts of THY overlay agar and TTC (1 μl/ml) required.
Mix well and add 20 ml to each THY agar plate and allow the
agar to solidify. TTC must be added after the agar has cooled to
below 50  C as it turns red upon heating.
17. Invert the stacked plates and incubate at 37  C, 5% CO2 for
16–18 h. Bacteria will form red-colored colonies.
18. After the incubation period, count the colonies using an auto-
mated colony counter (plates can be stored at 4  C for a
maximum of 72 h before reading).
19. Determine the average colony-forming unit (CFU) count of
the duplicate rows and identify the 2 dilutions between which
the average CFU count yields 50–200 CFU.
20. Choose a number of dilutions (maximum of six) between the
range identified in OD1. For example, if the OD1 result was
between 1:249 and 1:1249, dilutions chosen could be 1:249,
1:499, 1:749, 1:999 1:1499 and 1:1999. Proceed to OD2.

3.10.2 Optimal Dilution 1. Make up each dilution of bacteria working stock determined in
Experiment 2 OD1 required for the OD2 experiment.
2. Assay a known positive quality control (QC) serum, Control A
(heat inactivated BRC, no serum) and Control B (active BRC,
no serum) as detailed in Subheading 3.11 (each dilution of
bacteria will be added to four columns of one plate) (Fig. 2).
3. Select the dilution that yields ~100 CFU in Control B and the
OI of the known positive QC serum within the predetermined
accepted range.

3.11 OPKA Procedure The following procedure details the quantities required for four
96-well plates. Unknown test sera can be used neat or prediluted in
OPS buffer, for example, at 1:4 or 1:30 dilution.
1. Prepare 400 ml overlay agar. Store in a 50  C water bath until
required. Ensure a minimum of 16 THY agar plates are avail-
able for use.
2. Remove the BRC from the 80  C freezer and defrost on ice.
3. Label each round-bottomed 96-well microtiter plate with plate
ID (A, B, C or D) and sample layout. The procedure described
382 Helen Wagstaffe et al.

Fig. 2 Optimal dilution 2 experiment plate layout. Each test dilution of bacteria is assayed with a dilution series
of known positive serum, Control A and Control B

Fig. 3 OPKA plate layout

below utilizes 8 dilutions of serum serially diluted threefold, in

duplicate.
4. Add 20 μl of OPS buffer to columns 1 and 2 of each plate, rows
A–H inclusive.
5. Add 20 μl OPS buffer to rows A–G, columns 3–12 inclusive.
6. Add 30 μl of serum sample 1 into row H, columns 3 and 4 of
plate A. Add 30 μl of serum sample 2 into row H, columns
5 and 6. Continue adding a maximum of 5 samples per plate
(4 test samples and a QC sample).
7. Perform threefold serial dilutions in columns 3–12 by transfer-
ring 10 μl from row H to row G making sure no bubbles form
while mixing. Then transfer 10 μl from row G to row F and
continue up the plate. Once 10 μl is transferred from row B to
row A and mixed, remove 10 μl from row A and discard (see
Fig. 3).
8. Add 10 μl of prediluted bacteria working stock (prepared as
Subheading 3.9 and prediluted to optimal dilution determined
in Subheading 3.10) to each well including all control wells.
When adding bacteria, pipet 10 μl by nonreverse pipetting
directly into the liquid at the bottom of the well, release all
the liquid by pressing the pipette to the second stop.
 Group A Streptococcus OPK Assay 383

9. Incubate the 96-well plates for 30 min at RT on a mini-orbital

shaker at 700 rpm.
10. Incubate an aliquot of BRC in a 56  C degree water bath to
heat inactivate for a minimum of 30 min.
11. Following the incubation period, add 10 μl of heat inactivated
BRC (prediluted in OPS buffer if indicated) to column 1 of all
plates. Add 10 μl of active BRC (also prediluted) to all other
wells (columns 2–12). When adding BRC, reverse-pipet onto
the side of the well, then tap the plate to mix in the well.
12. Resuspend the HL-60 cells and add 40 μl to all wells.
13. Incubate on a mini-orbital shaker at 700 rpm at 37  C, 5% CO2
for 90 min. Incubate the plates in a single layer to maintain
equal CO2 exposure.
14. During this time remove THY agar plates from the fridge,
remove lids and lay on the bench to dry for 30 min to 1 h.
15. Place plates on ice for 20 min to halt the phagocytosis
process.
16. Replace the covers on the dried agar plates and stack in piles of
4 (1 stack per 96-well plate). Label the side of each plate with
the 96-well plate ID, the section of 96-well plate (1, 2, 3, or 4),
technician initials and the date.
17. Vortex the plate at a low speed. Using a multichannel pipette,
remove 10 μl from a single column and spot 10 μl on agar plate
by reverse pipetting to minimize bubbles. Tilt the plate to the
left and right so the spots measure approximately 1–1.5 cm
across, being careful not to allow the spots to merge.
18. Repeat this procedure for the next column, spot next to the
previous column on the agar plate. Repeat until there are three
columns of spots on the agar plate.
19. Leave all plates at RT for ~20 min to allow the reaction mixture
to soak into the agar (or until completely dry).
20. Remove the overlay agar from the water bath. Measure out
amount of overlay and TTC (1 μl/ml) required. Mix well and
add 20 ml to each THY agar plate, allow to solidify.
21. Invert the stacked plates and incubate at 37  C, 5% CO2 for
16–18 h. Bacteria will form red-colored colonies.
22. After the incubation period, count the colonies using an auto-
mated colony counter (plates can be stored at 4  C for a
maximum of 72 h before reading).

3.12 Data Analysis 1. The OI is the estimated dilution of serum that kills 50% of the
target bacteria.
384 Helen Wagstaffe et al.

2. The OI is calculated using the Opsotiter software, which is a

Microsoft Excel-based program developed to analyze data from
OPKAs.
3. Raw colony counts generated from the automated colony
counter are copied and pasted into the program. The program
tabulates the opsonic indices for the individual samples as well
as determining the percentage of nonspecific killing (NSK).
4. Each bacterial strain must be analyzed separately giving one file
for each 96-well plate.

3.12.1 Limits The following applies when samples are run neat.
of Detection
1. The upper limit of detection (ULD) is 8748. This is calculated
by taking the fold dilution of serum [3] to the power of
7 (number of fold dilutions), and multiplying by 4 (total vol-
ume in well (80 μl) divided by volume of serum in well (20 μl)).
2. However, Opsotiter automatically multiplies all 8748 results by
2 for statistical purposes. The “OI final” on the results sheet is
therefore displayed as 17,496.
3. Samples that have an opsonic index higher than the upper limit
of detection may be repeated at a predilution of 1 in 30.
4. The lower limit of detection (LLD) is 4, as this is the initial
dilution of serum once all components have been added to the
plate. However Opsotiter automatically divides all results of
4 by 2 for statistical purposes. The “OI final” on the results
sheet is therefore displayed as 2.
5. The lower limit of quantification (LLOQ) was determined for
seven different clinically relevant GAS strains (see Note 7).
Samples below LLOQ can be reported as half LLOQ for
statistical purposes.

3.12.2 Tentative Plate 1. The level of NSK is determined as the killing induced by BRC
Acceptance Criteria and HL-60 cells alone, in the absence of serum. This is deter-
mined as the difference between Control A and Control B: 1—
(average Control B CFU/average Control A CFU)  100.
This should be kept at 35%; see Note 5.
2. The CFU obtained in Control B should be kept between
50 and 150 to ensure accurate and reproducible counting on
the automated colony counter.
3. A QC sample should be included on all plates run, the OI result
should remain between a predetermined acceptability range.

3.12.3 Tentative Sample 1. Sample replicates must cross the 50% killing limit within one
Acceptance Criteria 3-fold dilution. If they do not cross the 50% killing limit, then
the sample will be retested.
 Group A Streptococcus OPK Assay 385

2. Good killing curves are those that show a sigmoid shape.

Curves may be irregular, and “N” or “U” shaped curves may
be identified. Irregular curves may be treated individually as
they may not conform to acceptance criteria.
3. For a sample to be considered positive, the maximum killing
must be greater than 70%. If the sample falls between 40% and
70% maximum killing, the sample must be repeated.

4 Notes

1. HL-60 cells from ATCC are well characterized and qualified,

this is the preferred source of HL-60 cells for this assay. Simi-
larly, BRC from Pel-Freez is of good quality and is the preferred
supplier, however BRC can be purchased from any supplier.
Significant lot to lot variation exists in BRC reagents, new
batches will need to be tested for suitability in the assay first.
2. HL-60 cell concentrations must not exceed 5  105 per ml at
the master stock propagation stage. During working stock
maintenance, cell density must remain 1.2  106 per ml. Cell
viability must remain 80%. HL-60 cell viability tends to
decline over 25–26 passages when a decrease in viability occurs
and irregular cell shapes are seen under the microscope, a new
vial of master stock should be thawed every 3–4 months. A
small volume of culture should checked for mycoplasma and
other microbial contamination using standard tissue culture
mycoplasma screening techniques.
3. Different strains will grow at different rates. Inoculation from a
single isolated colony will ensure that the broth culture is pure.
To avoid contamination from other bacteria that may be pres-
ent in the swab and increase purity of the strain of interest, it is
recommended to inoculate single colonies from the blood agar
plates and streak them onto a new blood agar plate and incu-
bate overnight at 37  C, 5% CO2 until colonies appear.
4. HL-60 cells must be mycoplasma-free. On the day of the assay,
cell viability must be 80%. Phenotype of HL-60 cells is deter-
mined by flow cytometry using mouse anti-human CD35
FITC conjugated antibody and mouse anti-human CD71
PE-conjugated antibody. Differentiated cells are accepted for
use in the assay if the up-regulation of CD35 was 55% of the
cell population and CD71 expressing was down-regulated by
12% when compared to the working stock preparation.
5. The concentration of BRC can be optimized to reduce NSK to
below an acceptable level. NSK at 35% is commonly used as
acceptance criteria in an optimized assay; however, levels up to
60% can also be accepted. The optimal concentration
386 Helen Wagstaffe et al.

determined for strains emm1 (43, 02-12, and GAS05134),

emm12 (611,020, 611,025, and GAS 09437) was 2.1%, for
emm6 (GASOPA6_02) was 3.1% [5]. Other factors that can be
optimized to ensure maximal killing within acceptable range of
NSK are plate shaking speed and the incubation time.
6. For the spots to dry, plates can be left to air-dry on a bench top.
If excessive contamination occurs, plates can be dried in a
laminar flow hood with a shortened length of time.
7. The LLOQ was determined for strains emm1 43; 9, 02-12;
15, GAS05134; 8, emm12 611,020; 7, 611,025; 5, GAS
09437; 7 and emm6 GASOPA6_02; 6 [5]. The average
LLOQ across all strains was 8. The LLOQ can be determined
by spiking low concentrations of known positive serum into
heat-inactivated nonimmune (antibody-depleted) serum to
produce a sample of OI 4–12. Generate each sample individu-
ally and run four times in duplicate.

References
1. Barth DD, Moloi A, Mayosi BM, Engel ME Organization Research and Development tech-
(2020) Prevalence of group a streptococcal nology roadmap and preferred product charac-
infection in Africa to inform GAS vaccines for teristics. Clin Infect Dis 69(5):877–883
rheumatic heart disease: a systematic review and 4. Lancefield RC (1957) Differentiation of group a
meta-analysis. Int J Cardiol 307:200–208 streptococci with a common R antigen into
2. Sims Sanyahumbi A, Colquhoun S, Wyber R, three serological types, with special reference to
Carapetis JR (2016) Global disease burden of the bactericidal test. J Exp Med 106
group a streptococcus. In: Streptococcus pyogenes: (4):525–544
basic biology to clinical manifestations. Univer- 5. Jones S, Moreland NJ, Zancolli M, Raynes J,
sity of Oklahoma Health Sciences Center, Okla- Loh JMS, Smeesters PR et al (2018) Develop-
homa City (OK) ment of an opsonophagocytic killing assay for
3. Vekemans J, Gouvea-Reis F, Kim JH, Excler JL, group a streptococcus. Vaccine 36
Smeesters PR, O’Brien KL et al (2019) The path (26):3756–3763
to group a streptococcus vaccines: World Health
 Chapter 21

Neisseria lactamica Controlled Human Infection Model

Adam P. Dale, Diane F. Gbesemete, Robert C. Read, and Jay R. Laver

Abstract
Neisseria lactamica is a nonpathogenic commensal of the human upper respiratory tract that has been
associated with protection against N. meningitidis colonization and disease. We have previously utilized the
N. lactamica controlled human infection model to investigate the protective effect of N. lactamica
colonization on N. meningitidis colonization, the nature of cross-reactive immune responses mounted
toward N. meningitidis following N. lactamica colonization, and the microevolution of N. lactamica over a
5-month colonization period. More recently, we have assessed the possibility of utilizing genetically
modified strains of N. lactamica to enable use of the commensal as a vehicle for prolonged exposure of
the nasopharynx of humans to antigens of interest, expressed in carried organisms. A controlled infection
with N. lactamica expressing the meningococcal antigen NadA has been executed and the results demon-
strate that this strategy is effective at generating immune responses to the target antigen. Throughout this
chapter, we outline in a step-by-step manner the methodologies utilized when performing controlled
human infection with N. lactamica including procedures relating to: (1) the dilution of N. lactamica
stock vials to derive intranasal inocula, (2) the delivery of intranasal inocula to human volunteers, (3) the
determination of N. lactamica colonization status following intranasal inoculation using oropharyngeal
swabbing and nasal wash sampling, (4) the microbiological procedures utilized to identify N. lactamica
colonization among study volunteers, and (5) the identification of N. lactamica colonies as strain
Y92-1009 using polymerase chain reaction.

Key words Neisseria lactamica, Neisseria meningitidis, Controlled human infection model, CHIME,
Human challenge

1 Introduction

Oropharyngeal carriage of Neisseria meningitidis is a prerequisite

for meningococcal disease, and there is a complex biological inter-
action between this organism and its host [1]. In most continents
other than Africa, the highest carriage rates of N. meningitidis
occur in young adults, particularly in universities and colleges
[2]. Although glycoconjugate meningococcal vaccines such as the
quadrivalent “ACWY” vaccine have had dramatic effects on disease
incidence, this is mostly due to herd protection conferred by
reduced carriage and transmission [3]. However, even prior to the

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_21,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

387
388 Adam P. Dale et al.

advent of meningococcal vaccines, meningococcal disease was rela-

tively rare other than during epidemics despite the fact that menin-
gococcal carriage was prevalent in a significant minority of
asymptomatic individuals. It is widely considered that natural
suppression of widespread dissemination of carriage of disease-
associated N. meningitidis occurs in nature by carriage of non-
pathogenic Neisseriaceae. N. lactamica, like N. meningitidis, is a
member of the Neisseriaceae but in contrast, N. lactamica is non-
capsulate and lacks pathogenic potential. It is a common commen-
sal of the human nasopharynx, particularly in young children.
Age-specific rates of N. meningitidis carriage and disease have
been shown to be inversely associated with carriage of
N. lactamica [4, 5] and mathematical modeling suggests a period
of 4–5 years of protection from meningococcal carriage following
carriage of N. lactamica [6].
To try to determine the mechanism of this relationship, we did
a controlled intranasal infection study of young adults with
N. lactamica Y92-1009 (sequence type 3493, clonal complex
613). Stocks of this strain had been manufactured at the Good
Manufacturing Practices (GMP) pharmaceutical manufacturing
facilities at Public Health England (PHE), Porton Down, United
Kingdom, to generate the seed banks for an outer membrane
vesicle vaccine. Theoretically, strain Y92-1009 is capable of gener-
ating immunity to a broad spectrum of N. meningitidis serogroups
including serogroup B, which expresses the least immunogenic
meningococcal polysaccharide. This intramuscular vaccine had
been tested in a Phase 1 study, and while well tolerated and capable
of generating antibody to N. lactamica, it did not produce cross-
reacting bactericidal antibodies against a bank of pathogenic
N. meningitidis serogroup B strains [7]. Using a N. lactamica
controlled human infection model experiment (CHIME), we
showed that intranasal inoculation with 104 colony forming units
(CFU) of strain Y92-1009 was well tolerated and safe and caused
carriage in 65% of young adults. This carriage of Y92-1009 per-
sisted in most carriers for 6 months. Note that this original group
included active smokers. Carriage was followed by development of
humoral immunity to N. lactamica in those who carried the inocu-
lated strain, but this did not induce significant cross-reactive bacte-
ricidal antibodies against N. meningitidis [8].
Subsequently, we sought direct evidence for a carriage preven-
tion effect by conducting controlled infection with N. lactamica in
healthy university students, the group that has the highest rates of
acquisition of meningococci. To do this, 310 nonsmoking univer-
sity students were inoculated with N. lactamica or were sham-
inoculated, and carriage was monitored for 26 weeks. At baseline,
natural N. meningitidis carriage in the control group of students
was 22.4%, which increased to 33.6% by week 26. Two weeks after
inoculation of N. lactamica, 33.6% of the challenge group became
colonized with N. lactamica. In this group, meningococcal carriage
 N. lactamica Controlled Human Infection Model 389

reduced significantly from 24.2% at inoculation to 14.7% 2 weeks

after inoculation. We found that the inhibition of meningococcal
carriage was only observed in volunteers who were actively colo-
nized with N. lactamica, was due both to displacement of existing
meningococci and to inhibition of new acquisition, and persisted
over at least 16 weeks. At the end of the 26-week period, we did
crossover inoculation of controls with N. lactamica and this repli-
cated the result in the original control group. The impact that we
observed on carriage reduction of N. meningitidis was just as
potent as that observed after glycoconjugate polysaccharide vacci-
nation. Furthermore, genome sequencing showed that the inhibi-
tion affected multiple meningococcal sequence types [9].
A closed genome sequence of strain Y92-1009 was then
defined [10] which enabled us to study microevolutionary changes
of N. lactamica over the course of 5 months of carriage using the
human cohort undergoing controlled infection. We found that
most mutations are transient indels within repetitive tracts of puta-
tive phase-variable loci associated with host–microbe interactions
(pgl and lgt) and iron acquisition ( fetA promoter and hpuA).
Recurrent polymorphisms occurred in genes associated with energy
metabolism (nuoN, rssA) and the CRISPR-associated cas1. In
volunteers who were naturally cocolonized with meningococci,
recombination altered allelic identity in N. lactamica to resemble
meningococcal alleles, including loci associated with metabolism,
outer membrane proteins and immune response activators, but
there was little evidence of recombination in the opposite
direction [11].
Recently, we have delineated B cell responses to N. lactamica
colonization and their cross-reactivity with N. meningitidis [12]
and have investigated the possibility of genetic modification of
N. lactamica to enable use of the commensal as a vehicle for
prolonged exposure of the nasopharynx of humans to antigens of
interest, expressed in carried organisms. A controlled infection with
N. lactamica expressing the meningococcal antigen NadA has been
planned and executed [13] and the results demonstrate that this
strategy is effective at generating immune responses to the target
antigen [14].
Throughout this chapter, we outline in a step-by-step manner
the methodologies utilized when performing N. lactamica
CHIMEs including procedures relating to (1) the dilution of
N. lactamica stock vials to derive intranasal inocula, (2) the delivery
of intranasal inocula to human volunteers, (3) the determination of
N. lactamica colonization status following intranasal inoculation
using oropharyngeal swab and nasal wash sampling, (4) the micro-
biological procedures utilized to identify N. lactamica colonization
amongst study volunteers, and (5) the identification of cultured
N. lactamica colonies as the inoculum strain (Y92-1009) using
polymerase chain reaction (PCR).
390 Adam P. Dale et al.

2 Materials

2.1 Dilution 1. Stock vial containing 1
108 CFU/ml N. lactamica

of N. lactamica Stock Y92-1009 (ST-3493, clonal complex 613, suspended in 1 ml
Vials to 105 CFU/ml Franz medium) produced in the GMP-accredited facilities at
for Intranasal PHE’s Porton Down facility, Salisbury, UK. Following transfer
Inoculation from PHE, stock vials of N. lactamica are stored in a dedicated,
secure and remotely monitored 80  C freezer.
2. Decontaminated class II microbiological safety cabinet (MSC),
dedicated for N. lactamica inoculum preparation.
3. Decontaminated vortex mixer, dedicated for N. lactamica
inoculum preparation.
4. Decontaminated set of calibrated Gilson-style pipettes, dedi-
cated for N. lactamica inoculum preparation.
5. Incubator, set to 37  C, 5% CO2.
6. Sterile filter pipette tips.
7. 10 μl sterile, disposable plastic microbiological loops.
8. 70% alcohol solution (v/v).
9. Single use waste box.
10. Disposable gloves.
11. Sterile phosphate buffered saline (PBS) (1
: autoclaved and
passed through a sterile 0.2 μm filter).
12. Columbia blood agar (CBA) plates: Columbia agar base, 5%
defibrinated horse blood.
13. Gonococcal (GC) agar plates with 5-bromo-4-chloro-3-indo-
lyl-B-D galactopyranoside (X-gal) (GC-X-gal): 36 g/l GC agar
base, 100 ml/l lysed horse blood, 20 ml/l Vitox supplement,
4 g/l glucose, 40 mg/l X-gal, 2 mg/l vancomycin, 7.5 mg/l
colistin, 3 mg/l trimethoprim, and 1 mg/l amphotericin B.
14. Appropriate laboratory source document (to log all stages of
N. lactamica inoculum preparation in real time).
15. Sterile universal containers.

2.2 Intranasal 1. Intranasal inoculum containing 105 CFU N. lactamica sus-

Inoculation pended in 1 ml PBS (see Subheading 3.2).
with 105 CFU 2. Dedicated room within a clinical environment, for example, a
N. lactamica clinical research facility (CRF), complete with an examination
couch.
3. Disposable aprons and gloves.
4. Decontaminated and dedicated 1000 μl Gilson-style pipette.
5. Sterile pipette tips.
 N. lactamica Controlled Human Infection Model 391

6. Clinical waste bin.

7. Appropriate laboratory source document (to log all stages of
N. lactamica inoculation in real time).

2.3 Taking 1. Sterile throat swab for microbiological culture with an appro-
and Processing priate transport medium (e.g., TS/5-17 Probact Amies Clear
an Oropharyngeal medium in a tube with a polystyrene viscose tip swab).
Throat Swab 2. GC-X-gal plate.
to Determine Volunteer 3. Dedicated room within a clinical environment, for example, a
N. lactamica CRF, complete with an examination couch.
Colonization Status
4. Class II MSC housed within an Advisory Committee on Dan-
gerous Pathogens (ACDP) containment level 2 (CL2)
laboratory.
5. Appropriate laboratory waste container.
6. Incubator set at 37  C, 5% CO2.

2.4 Taking 1. Sterile saline (0.9%) for irrigation or injection (20 ml) at room
and Processing temperature.
a Nasal Wash Sample 2. 2
10 ml sterile syringes.
to Determine Volunteer
3. Sterile plastic Petri dish.
N. lactamica
Colonization Status
4. 2
GC-X-gal plates.
5. Sterile PBS.
6. Sterile 50 ml centrifuge tube.
7. Set of calibrated Gilson-style pipettes and sterile tips.
8. Sterile microbiology plate spreaders.
9. Disposable gloves and aprons.
10. Tissues.
11. Transport box with ice.
12. Dedicated room within a clinical environment, for example, a
CRF, complete with an examination couch.
13. Class II MSC housed within an ACDP CL2 laboratory.
14. Centrifuge.
15. Appropriate laboratory waste container.
16. Incubator set at 37  C, 5% CO2.

2.5 Identification 1. Class II MSC.

of N. lactamica 2. GC-X-gal agar plate.
Colonies
3. 10 μl sterile, disposable plastic microbiological loops.
4. Sterile Bacterial Storage Medium: 50:50 (v/v) mixture of
Tryptone Soya Broth supplemented with 0.2% (w/v) yeast
extract, 60:40 (v/v) mixture of glycerol and PBS. Stored in
0.5 ml aliquots in cryogenic storage vials.
392 Adam P. Dale et al.

5. Oxidase reagent-impregnated strips.

6. Glass microscope slides.
7. 0.5% crystal violet solution.
8. Gram’s (Lugol’s) iodine solution.
9. Carbolfuchsin or safranin solution.
10. Gram’s acetone solution.
11. Distilled water.
12. Heat block.
13. Blotting paper.
14. Mineral oil.
15. Microscope with 100
lens (oil immersion).
16. API® NH kit (Biomerieux).
17. Genomic DNA extraction kit.
18. Set of calibrated Gilson-style pipettes.

2.6 Identification 1. Dedicated, decontaminated PCR preparation area or laminar

of N. lactamica flow cabinet.
Colonies as Strain 2. Set of calibrated Gilson-style pipettes and sterile filter tips.
Y92-1009 Using PCR
3. Disposable gloves and dedicated, pre-PCR laboratory coat.
2.6.1 Preparation of PCR 4. Tissues.
Master Mix
5. 10% (v/v) bleach solution in spray bottle.
6. Distilled water in spray bottle.
7. 70% (v/v) alcohol solution.
8. Q5 Hot Start High Fidelity 2
Master Mix.
9. Primer stocks (10 μM each). For list of primers and sequences,
see Table 1.
10. DNase/RNase-free, molecular biology grade water.
11. Thin-walled, 0.2 ml PCR tubes and 0.2 ml tube racks.
12. Benchtop centrifuge.
13. Appropriate laboratory waste container.

2.6.2 Performance 1. Dedicated, decontaminated PCR preparation area or laminar

of Y92-1009-Specific PCR flow cabinet.
2. Set of calibrated Gilson-style pipettes and sterile filter tips.
3. Disposable gloves and dedicated, pre-PCR laboratory coat.
4. Tissues.
5. 10% (v/v) bleach solution in spray bottle.
6. Distilled water in spray bottle.
7. Number of 24 μl aliquots of [1
Y92-1009-specific PCR Mas-
ter Mix] equal to number of isolates for identification.
 N. lactamica Controlled Human Infection Model 393

Table 1
Nucleotide sequences of primers for Y92-1009-specific, multiplex PCR. (Forward—FOR; Reverse—
REV)

Nucleotide sequence
Name (50 to 30 ) Product size (bp)
SeqA_FOR gtgctgaatttatagacgggc 388
SeqA_REV aagctaggtctacttggtttag
SeqB_FOR caattgtttaccgccctgcc 273
SeqB_REV gtttcctgccaatcaatccc
SeqC_FOR agggaccgacatctttcatac 594
SeqC_REV ttgcaggctctttccaaac
lacZ_FOR cgggcaaacttgcgcgg 745
lacZ_REV gcaaaccgaaacggggcagg

8. Genomic DNA extracted from colonies positively identified as

N. lactamica (see Subheading 3.5).
9. Genomic DNA extracted from N. lactamica strain Y92-1009
(@50 ng/μl).
10. DNase/RNase-free, molecular biology grade water.
11. Appropriate laboratory waste container.
12. Laboratory marker pen.
13. Laboratory coat for conducting PCR.
14. Thermal cycler.

2.6.3 Analysis of PCR 1. Laboratory coat and disposable gloves.

Products Using Agarose Gel 2. Set of calibrated Gilson-style pipettes and sterile tips.
Electrophoresis
3. Certified PCR agarose.
4. Tris–acetate–EDTA (TAE) buffer (1
).
5. Nontoxic DNA intercalating agent.
6. DNA molecular weight ladder (200–1000 bp).
7. Digital gel imaging system.
8. Appropriate laboratory source document.

3 Methods

3.1 Dilution N. lactamica inoculum preparation is performed in tandem by two

of N. lactamica Stock trained members of the laboratory technical team (see Note 1).
Vials to 105 CFU/ml
1. Don a pair of disposable gloves and disinfect the class II MSC
for Intranasal with 70% alcohol.
Inoculation
394 Adam P. Dale et al.

2. Label a series of universal containers in preparation for serial

dilution of the N. lactamica stock inoculum vial contents, as
follows: 101, 102, 103, 104, 105, and 106 dilution.
3. Pipet 900 μl PBS into the universal containers labelled 101,
102, 104, 105, and 106 and 4.5 ml of PBS into the
universal container labelled 103.
4. Label the CBA and GC-X-gal plates and ensure they are at
room temperature prior to use.
5. Remove a stock vial containing 1
108 CFU/ml N. lactamica
Y92-1009 from the 80  C freezer and transport this vial
immediately on ice to the class II MSC. Take the stock vial
off ice and place within the class II MSC and allow it to thaw at
room temperature.
6. Immediately upon thawing, vortex the stock vial for 30 s and
then pipet 100 μl of the stock vial contents into the universal
container labelled 101 before vortexing for 30 s.
7. Having applied a new filter tip, pipet 100 μl of the 101
dilution into the universal tube labelled 102 before vortexing
for 30 s.
8. Having applied a new filter tip, pipet 500 μl of the 102
dilution into universal container labelled 103 before vortexing
for 30 s.
9. Having applied a new filter tip, pipet 1 ml volumes of the 103
dilution (containing 105 CFU N. lactamica/ml) into universal
containers and transport to the CRF for use as intranasal
inocula.
10. To verify the purity of the prepared batch of N. lactamica
inocula, streak 10 μl volumes of the 103 dilution onto CBA
and GC-X-gal agar plates using a microbiological loop and
incubate for 48 h at 37  C, 5% CO2. Following incubation,
assess for the pure growth of colonies morphologically in
keeping with N. lactamica that are gray on CBA and blue on
GC-X-gal plates.
11. To enable N. lactamica CFU/ml of the prepared inocula to be
calculated formally, utilize the 103 dilution to immediately
perform three further serial dilutions in the universal contain-
ers labeled 104, 105, and 106, ensuring that a new pipette
tip is utilized between dilution steps and that each dilution is
vortexed for 30 s.
12. Pipet 3
10 μl drops of the 104, 105, and 106 dilutions
onto a GC-X-gal plate and allow to dry for 5 min before
incubating for 48 h at 37  C, 5% CO2. Following incubation,
identify the dilution with an easily countable number of colo-
nies (10–50 colonies within each 10 μl drop). Use the Mean to
calculate N. lactamica CFU/ml in the prepared inoculum
 N. lactamica Controlled Human Infection Model 395

taking into account the dilution factor (CFU/ml ¼ Mean col-

ony count
100
dilution factor).

3.2 Intranasal Prior to enrolment onto a N. lactamica CHIME study, volunteers

Inoculation undergo extensive screening to ensure they are medically fit and
with 105 CFU that it is safe for them to participate. While the full volunteer
N. lactamica screening protocol is beyond the scope of this chapter, key inclusion
and exclusion criteria used across the N. lactamica CHIME studies
during the screening process are outlined in Note 2. Once a
volunteer has passed the screening process and is enrolled onto
the study, the methodology outlined below is followed to ensure
a standardized approach to intranasal inoculation. It must be noted
that all intranasal inoculation procedures must be performed in an
appropriate clinical environment, for example a CRF, by a member
of the clinical research team.
1. Ask the volunteer to drink only water in the 1-h period prior to
their intranasal inoculation appointment.
2. Explain the procedure to the volunteer and obtain informed
consent.
3. Ask the volunteer to don a disposable apron and lie supine on
the clinical examination couch with their head tilted back (see
Fig. 1).
4. Transport the prepared inoculum (105 CFU N. lactamica sus-
pended in 1 ml PBS) from the laboratory to the CRF. The
inoculum must be utilized within 30 min of preparation (see
Note 3).
5. Wash hands and don a pair of gloves and a disposable apron.
6. Instill 0.5 ml of the prepared inoculum slowly in a dropwise
fashion into each nostril, one at a time, using a Gilson-style
pipette. Following intranasal inoculation, the volunteer is asked
to remain supine for 5 min.
7. Dispose of the used pipette tip, apron, and gloves in the clinical
waste.
8. Monitor the volunteer following intranasal inoculation for
30 min. Take a set of clinical observations including pulse,
blood pressure, respiratory rate, and temperature. Assuming
the volunteer is well and has normal observations following
intranasal inoculation, they can then be discharged from
the CRF.

1. Explain the procedure to the volunteer and obtain informed

consent.
2. Wash hands and apply a clean pair of disposable gloves and an
apron.
396 Adam P. Dale et al.

Fig. 1 Intranasal inoculation procedure. Prior to intranasal inoculation, the

volunteer is asked to lie supine with their head tilted back

3.3 Taking 3. While sitting upright, ask the volunteer to tilt their head back
and Processing and open their mouth wide.
Oropharyngeal Throat 4. When the head of the swab has made contact with the pharyn-
Swabs to Determine geal wall, behind the uvula, stroke it across the pharyngeal wall
Volunteer N. lactamica five times.
Colonization Status 5. Remove the swab from the oral cavity, place it into the Amies
transport medium, seal the swab and transfer it immediately to
the class II MSC for further processing.
6. Working in the class II MSC, apply the swab contents directly
to the GC-X-gal plate by streaking the swab vertically in a
continuous fashion while rotating the plate and swab tip.
Ensure the swab makes contact with all areas of the agar.
7. Dispose of the swab in the appropriate waste container and
incubate the GC-X-gal plate immediately at 37  C, 5% CO2 for
48 h.
8. Following 48 h incubation, assess the GC-X-gal plate for
growth of putative colonies of N. lactamica in line with the
methodology outlined in Subheading 3.5.
 N. lactamica Controlled Human Infection Model 397

3.4 Taking 1. Explain the procedure to the volunteer and obtain informed
and Processing Nasal consent.
Wash Samples 2. Ask the patient to don a disposable apron to protect their
to Determine Volunteer clothing.
N. lactamica 3. Position the patient supine with a pillow placed under their
Colonization Density shoulders to enable extension of the neck backward.
4. Don an apron and a pair of gloves.
5. Draw up 10 ml of sterile 0.9% saline into the syringe and expel
any air.
6. Ask the volunteer to hold the Petri dish on their chest to enable
collection of the nasal wash fluid (leave the lid on the Petri dish
at this point).
7. Ask the volunteer to open their mouth, extend their head
backward, and position their tongue to avoid swallowing the
saline. Remind them to breathe through their mouth and not
to swallow.
8. Once in position and the volunteer has signaled they are ready,
place the tip of the syringe in their nostril and gently instill
10 ml of saline. Ask them to hold the saline in their nasophar-
ynx for 1 min.
9. After 1 min, ask the volunteer to open the lid of the Petri dish
and to lean forward as fast as possible to allow the saline to exit
the nose passively and be caught in the Petri dish.
10. Withdraw the nasal wash specimen from the Petri dish using
the syringe and transfer it to the centrifuge tube.
11. Repeat steps 5–10 using the alternate nostril.
12. Transport the nasal wash specimen on ice immediately to the
microbiology laboratory for processing.
13. Measure and record the total volume of nasal wash fluid
retrieved from the volunteer.
14. Centrifuge the nasal wash fluid at 5000
g for 10 min to pellet
the bacteria.
15. Carefully decant the supernatant and resuspend the pellet in a
total volume of 300 μl PBS before vortexing for 30 s to loosen
the mucous.
16. Inoculate 2
GC-X-gal plates with 250 μl and 25 μl volumes of
the resuspended pellet, respectively, and lawn over the plates
using individual microbiology spreaders.
17. Incubate the GC-X-gal plates immediately at 37  C, 5% CO2,
for 48 h.
18. Following 48 h incubation, assess the GC-X-gal plates for
growth of putative colonies of N. lactamica in line with the
methodology outlined in Subheading 3.5.
398 Adam P. Dale et al.

19. Calculate the N. lactamica colonization density within the

nasal wash where necessary as N. lactamica CFU/ml.

3.5 Identification 1. Working in a class II MSC, visually inspect the GC-X-gal agar
of N. lactamica plates inoculated with throat swab or nasal wash contents for
Colonies from Throat evidence of putative N. lactamica colonies, that is, blue colo-
Swab and Nasal Wash nies (beta-galactosidase positive) with morphology consistent
Cultures with Neisseria spp.
2. Pick a well-defined, accessible blue colony and subculture it by
streaking it onto GC-X-gal agar with a 10 μl bacteriological
loop. Incubate at 37  C, 5% CO2, for 24 h.
3. Assuming a pure subculture is obtained, produce a stock of the
isolate in an aliquot of bacterial storage medium. Macerate
colonies against the side of the cryogenic storage vial to pro-
duce a suspension, then incubate for 10 min and freeze at
80  C.
4. Check the oxidase status of the organism by spreading a small
amount of a single blue colony onto an oxidase strip using a
10 μl bacteriological loop. A positive oxidase test (strip turns
blue/black where contact with colony is made) would be in
keeping with potential N. lactamica and should prompt a
Gram stain to be performed. If the bacteria are oxidase nega-
tive, then the isolate is not N. lactamica and the prepared stock
(step 3, above) can be discarded.
5. To perform a Gram stain, the following steps should be
followed:
(a) Place a drop of distilled water onto a clean glass micro-
scope slide and emulsify a small amount of a single blue
colony into it using a 10 μl bacteriological loop.
(b) Dry and fix the slide by placing on a heat block at 65  C.
(c) Remove the slide from the class II MSC and flood the slide
with 0.5% crystal violet solution and leave for 30 s.
(d) Wash off the crystal violet solution with tap water and
then flood the slide with Gram’s acetone and wash off
with tap water rapidly.
(e) Flood the slide with carbolfuchsin or safranin solution and
leave for 30 s.
(f) Wash off the carbolfuchsin or safranin with tap water and
dry with clean blotting paper.
(g) View under oil immersion at 100
magnification.
6. A Gram stain that reveals Gram-negative diplococci is consis-
tent with N. lactamica and should prompt speciation using a
biochemical test strip, for example, the API® NH (Biomerieux)
(see Note 4).
 N. lactamica Controlled Human Infection Model 399

7. Blue colonies on GC-X-gal agar that are oxidase positive, Gram

negative diplococci with an API NH code of 5041 are
N. lactamica. To confirm the cultured N. lactamica as strain
Y92-1009, that is, the inoculum strain, PCR is used (Subhead-
ing 3.6).
8. To provide template material for amplification by the
Y92-1009-specific PCR: return to the frozen stock of the
isolate, culture the bacteria and then isolate genomic DNA
using a genomic DNA extraction kit according to the manu-
facturer’s instructions.

3.6 Identification 1. Don pre-PCR laboratory coat and gloves.

of N. lactamica 2. If not already active, ensure the laminar flow cabinet is empty,
Colonies as Strain and then turn on the laminar flow cabinet. Expose inside of
Y92-1009 Using PCR cabinet to UV light for 15 min (if available).
3.6.1 Preparation of PCR 3. Decontaminate the working area by liberal application of 10%
Master Mix (v/v) bleach solution.
4. Wait for 10 min for the bleach to inactivate environmental
amplicons and other potential DNA contaminants. Change
gloves.
5. Meanwhile: gather together the Q5 Hot-Start High Fidelity
2
Master Mix (in 500 μl aliquots), the primer stocks, the
DNase/RNase-free molecular grade water and the 0.2 ml,
thin-walled PCR tubes and racks. Note that you will need
20
0.2 ml PCR tubes per 500 μl aliquot of Q5 Hot-Start
High Fidelity 2
Master Mix.
6. Spray the working area with distilled water and then mop up
the resulting moisture with paper toweling. Carefully dispose
of the paper toweling in the waste bin. Change gloves.
7. If not already deactivated, deactivate the UV light and then
open the laminar flow cabinet.
8. Inside the laminar flow cabinet, add the following volumes of
the appropriate primers into the 500 μl aliquot of Q5 Hot-Start
High Fidelity 2
Master Mix:
(a) 25 μl: SeqA FOR
(b) 25 μl: SeqA REV
(c) 25 μl: SeqB FOR
(d) 25 μl: SeqB REV
(e) 25 μl: SeqC FOR
(f) 25 μl: SeqC REV
(g) 25 μl: lacZ_FOR
(h) 25 μl: lacZ_REV.
400 Adam P. Dale et al.

9. Add 30 μl of DNAse/RNAse-free molecular biology grade

water to the 500 μl aliquot of Q5 Hot-Start High Fidelity
2
Master Mix supplemented with primers. This is the
Y92-1009-specific 1
Master Mix.
10. Transfer 24 μl aliquots of the Y92-1009-specific 1
Master
Mix into individual 0.2 ml, thin walled PCR tubes and stand in
0.2 ml tube racks. Make sure the tube rack is labelled as con-
taining aliquots of the Y92-1009-specific Master Mix.
11. Place the aliquots of Y92-1009-specific 1
Master Mix into the
fridge (4  C) for use later that day, or proceed directly to
Subheading 3.6.2.
12. If work in the laminar flow cabinet is completed, seal the
laboratory waste container, remove all items from inside the
laminar flow cabinet, spray the inside of the laminar flow cabi-
net liberally with 70% (v/v) alcohol solution and then close the
sash on the laminar flow cabinet.

3.6.2 Performance 1. If using boiled bacterial lysates, either fresh or frozen, remove
of Y92-1009-Specific PCR bacterial debris from each suspension by centrifugation of the
lysates at 17,000
g for 10 min. Note that the Y92-1009-
specific PCR will also amplify target sequences from purified
genomic DNA. Solutions of extracted genomic DNA are most
easy to use at a concentration of 50 μg/ml.
2. Don pre-PCR laboratory coat, safety spectacles and gloves.
3. Decontaminate working area by liberal application of 10%
(v/v) bleach solution.
4. Wait for 10 min for the bleach to inactivate environmental
amplicons and other potential DNA contaminants. Change
gloves.
5. Meanwhile: gather together the appropriate number of ali-
quots of the ‘Y92-1009-specific 1
Master Mix’, such that
there is one aliquot of 1
Master Mix for each sample to be
analysed. Gather additional aliquots of 1
Master Mix suffi-
cient to include one positive and one negative control reaction
for every two rows of wells that will be filled during agarose gel
electrophoresis. Note that each PCR control tube will provide
sufficient material to load 2 wells of control material. One well
containing positive control material and one well containing
negative control material must be loaded in each row of wells,
to allow for: (1) visual comparison of sample amplicons to
those generated from wild type genomic DNA extracted from
N. lactamica strain Y92-1009, and (2) to show that no con-
taminating DNA was present in the 1
Master Mix,
respectively.
 N. lactamica Controlled Human Infection Model 401

A B C D E F G
800 bp

600 bp

400 bp

200 bp

Fig. 2 Interpreting the banding pattern of reaction products from Y92-1009-

specific PCR. Possible banding patterns generated in the Y92-1009-specific
PCR: (a) PCR failure, empty lane, negative control or DNA present from species
other than N. lactamica. (b–f) DNA present from strains of N. lactamica other
than Y92-1009. (g) N. lactamica strain Y92-1009. Note that preliminary identifi-
cation of N. lactamica colonies (i.e., oxidase test, Gram stain and API NH—see
Subheading 3.5) will in most cases prevent PCR from being performed on
species other than N. lactamica. Note that, in addition to amplification of the
bands shown in this figure, strains of N. lactamica other than Y92-1009 may also
produce additional bands of unknown and unpredictable size. Strains can only be
identified as Y92-1009 if the banding pattern shown in (g) is present

6. Spray working area with distilled water and then mop up the
resulting moisture with paper toweling. Carefully dispose of
the paper toweling in the waste bin. Change gloves.
7. For each sample, carefully transfer 1 μl of the lysate or purified,
genomic DNA (50 ng) into a single 24 μl aliquot of “Y92-
1009-specific 1
Master Mix.” Appropriately label each PCR
tube with the laboratory marker pen.
8. Into one of the remaining 1
Master Mix aliquots, add 1 μl of
N. lactamica gDNA derived from wild type strain Y92-1009
(50 ng), and into another add 1 μl of DNase/RNase-free water.
Appropriately label each tube with the laboratory marker pen.
9. Retain all 1.5 ml microcentrifuge tubes containing the boiled
lysates until a given lysate has been definitively identified.
Boiled lysates can be refrozen at 20  C.
10. Place all PCR tubes containing complete reaction mixtures
(including controls) into the Thermal Cycler. Set the Thermal
Cycler to run according to the parameters shown in Note 5.
11. Analyse PCR amplicons on an agarose gel and analyze the
banding pattern present in each lane of the gel, with reference
to Fig. 2.

4 Notes

1. Preparation of N. lactamica inocula from stock vials is per-

formed by two trained members of the research team who
have completed Good Clinical Laboratory Practice training.
402 Adam P. Dale et al.

Each step of the process is verified by both technicians and

logged in real time using the appropriate study-specific labora-
tory source document. This document then forms part of the
volunteer case report form.
2. Key volunteer inclusion and exclusion criteria utilized in
N. lactamica CHIMEs in the United Kingdom are as follows:
Inclusion criteria:
(a) Healthy adults aged 18–45 years inclusive on the day of
enrolment.
(b) Fully conversant in the English language.
(c) Able and willing (in the investigator’s opinion) to comply
with all study requirements.
(d) Written informed consent to participate in the study.
(e) For females only, willingness to practice continuous effec-
tive contraception during the study and a negative preg-
nancy test at the screening visit.
Exclusion criteria:
(a) Active smokers.
(b) Individuals who have a current infection at the time of
inoculation.
(c) Individuals who have been involved in other clinical stud-
ies/trials involving receipt of an investigational product
over the last 12 weeks or if there is planned use of an
investigational product during the study period.
(d) Any confirmed or suspected immunosuppressive or
immunocompromised state, including HIV infection,
asplenia, history of recurrent severe infections or use
(more than 14 days) of immunosuppressant medication
within the past 6 months (topical/inhaled steroids are
allowed).
(e) Allergy to yeast extract.
(f) Any other significant disease, disorder, or finding which
may significantly increase the risk to the volunteer because
of participation in the study, affect the ability of the vol-
unteer to participate in the study, or impair interpretation
of the study data, for example recent surgery to the
nasopharynx.
(g) Occupational, household or intimate contact with immu-
nosuppressed persons.
(h) Pregnancy or lactation.
3. The viability of N. lactamica reduces following thawing and
dilution in PBS. The dilution strategy referred to within Sub-
heading 3.1 will reliably produce a 1 ml volume of inoculum
 N. lactamica Controlled Human Infection Model 403

containing 105 CFU, with the viability being maintained up to

30 min following preparation. By 1 h following inoculum
preparation, the viability reduces to 5
104 CFU/ml. Thus,
to ensure intranasal inoculation with 105 CFU, the inoculum
should be utilized immediately, that is, within 30 min of
preparation.
4. The API® NH testing kit comes ready to use with all required
reagents and clear step-by-step instructions. N. lactamica can
be identified with the API® NH test code of 5041 with the
whole process taking approximately 3 h [15].
5. Thermal cycling parameters for amplification of Y92-1009-
specific target sequences are as follows:

STAGE A: 1
5 min at 98  C
STAGE B: 35
20 s at 98  C (melting)
15 s at 63  C (annealing)
26 s at 72  C (extension)
STAGE C: 1
5 min at 72  C
(Optional): Infinite at 10  C

References
1. Laver JR, Hughes SE, Read RC (2015) Neis- on Neisseria lactamica outer membrane vesi-
serial molecular adaptations to the nasopharyn- cles. Clin Vaccine Immunol 16:1113–1120
geal niche. Adv Microb Physiol 66:323–355 8. Evans CM, Pratt CB, Matheson M et al (2011)
2. Christensen H, May M, Bowen L et al (2010) Nasopharyngeal colonization by Neisseria lac-
Meningococcal carriage by age: a systematic tamica and induction of protective immunity
review and meta-analysis. Lancet Infect Dis against Neisseria meningitidis. Clin Infect Dis
10:853–861 52:70–77
3. Maiden MCJ, Ibarz-Pavón AB, Urwin R et al 9. Deasy AM, Guccione E, Dale AP et al (2015)
(2008) Impact of meningococcal serogroup C Nasal inoculation of the commensal Neisseria
conjugate vaccines on carriage and herd immu- lactamica inhibits carriage of Neisseria menin-
nity. J Infect Dis 197:737–743 gitidis by young adults: a controlled human
4. Gold R, Goldschneider I, Lepow ML et al infection study. Clin Infect Dis 60:1512–1520
(1978) Carriage of Neisseria meningitidis and 10. Pandey AK, Cleary DW, Laver JR et al (2017)
Neisseria lactamica in infants and children. J Neisseria lactamica Y92-1009 complete
Infect Dis 137:112–121 genome sequence. Stand Genomic Sci 12:41
5. Cartwright KA, Stuart JM, Jones DM et al 11. Pandey AK, Cleary DW, Laver JR et al (2018)
(1987) The stonehouse survey: nasopharyn- Microevolution of Neisseria lactamica during
geal carriage of meningococci and Neisseria nasopharyngeal colonisation induced by con-
lactamica. Epidemiol Infect 99:591–601 trolled human infection. Nat Commun 9:4753
6. Coen PG, Cartwright K, Stuart J (2000) Math- 12. Dale AP, Theodosiou AA, Laver JR et al (2020)
ematical modeling of infection and disease due Controlled human infection with Neisseria lac-
to Neisseria meningitidis and Neisseria lacta- tamica induces B-cell responses that are cross-
mica. Int J Epidemiol 29:180–188 reactive with Neisseria meningitidis. J Immu-
7. Gorringe AR, Taylor S, Brookes C et al (2009) nol 204:231.21
Phase I safety and immunogenicity study of a 13. Gbesemete D, Laver JR, de Graaf H et al
candidate meningococcal disease vaccine based (2019) Protocol for a controlled human
404 Adam P. Dale et al.

infection with genetically modified Neisseria 14. Laver JR, Gbesemete D, Dale AP et al (2021) A
lactamica expressing the meningococcal vac- recombinant commensal bacteria elicits heter-
cine antigen NadA: a potent new technique ologous antigen-specific immune responses
for experimental medicine. BMJ Open 9: during pharyngeal carriage. Sci Transl Med
e026544 13:eabe8573
15. APIWEB™, Biomerieux. https://apiweb.bio
merieux.com/login. Accessed 27 Sep 2020
 Chapter 22

Analyzing Macrophage Infection at the Organ Level

Ryan G. Hames, Zydrune Jasiunaite, Joseph J. Wanford, David Carreno,
Wen Y. Chung, Ashley R. Dennison, and Marco R. Oggioni

Abstract
Classical in vivo infection models are oftentimes associated with speculation due to the many physiological
factors that are unseen or not accounted for when analyzing experimental outputs, especially when solely
utilizing the classic approach of tissue-derived colony-forming unit (CFU) enumeration. To better under-
stand the steps and natural progression of bacterial infection, the pathophysiology of individual organs with
which the bacteria interact in their natural course of infection must be considered. In this case, it is not only
important to isolate organs as much as possible from additional physiological processes, but to also consider
the dynamics of the bacteria at the cellular level within these organs of interest. Here, we describe in detail
two models, ex vivo porcine liver and spleen coperfusion and a murine infection model, and the numerous
associated experimental outputs produced by these models that can be taken and used together to
investigate the pathogen–host interactions within tissues in depth.

Key words Ex vivo perfusion, Murine infection model, Immunohistochemistry, Confocal micros-
copy, Fiji, InForm, Image analysis, Correlates of protection

1 Introduction

Correlates of protection are a key issue in preclinical vaccine

research but are highly specific for different pathogens and may
include protective humoral, innate and cellular responses
[1]. Examples of correlates of protection include the titer of bacte-
ricidal antibodies in Neisseria meningitidis vaccine research [2, 3]
or opsonophagocytic titers in pneumococcal vaccine research
[4]. We have recently found that a subpopulation of splenic
tissue-resident macrophages are not only key players in the clear-
ance of encapsulated bacteria but also represent a location for
pathogen replication and subsequent initiation of invasive disease
[5, 6]. The present provides the protocols for an integrated evalua-
tion of the role of splenic tissue macrophages in current in vivo and
ex vivo models with the scope of allowing the testing of this novel
step in cellular immunity for evaluation in vaccine research.

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_22,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

405
406 Ryan G. Hames et al.

This work describes the methodology of the murine intrave-

nous (IV) infection model and porcine ex vivo liver and spleen
coperfusion model. Mouse infection models, in particular for Strep-
tococcus pneumoniae research, are the preferred choice for their high
reliability and excellent reproducibility in research on the patho-
genesis of infection [7, 8]. The IV infection model has the advan-
tage of inducing an almost instantaneous delivery of bacteria to
blood-filtering organs such as the spleen and liver which both have
prominent roles in the control of bacteremia [5, 9–16]. This allows
for the study of the development of invasive pneumococcal infec-
tion while minimizing the effects of confounding factors such as the
lung innate and cellular immunity following intranasal infection.
Therefore, the murine IV infection model is particularly useful for
the study of the early stages of blood borne pathogenesis
[5]. Despite being widely available, practical and having much
comparability to humans, the mouse model has several drawbacks.
The housing of animals in a sterile animal unit environment is costly
and while mice have a high genetic similarity to humans [17], their
bodies have broad physiological and morphological differences as
far down as the cellular population and structure of the organs
[18, 19]. Therefore, while excellent at providing fundamental
data, a more comparable model is required to validate and build
upon these results, enhance their clinical importance for the human
population, and introduce a closer biological match. These aims can
be addressed with the ex vivo perfusion model of the porcine spleen
and liver [6, 20]. The organs are sourced from a local abattoir, from
animals which are to be used within the food industry, therefore
reducing the use of live animals in laboratory settings. In addition,
this approach does not require specialized animal housing and
importantly the utilization of organs of animals slaughtered for
food production means that these ex vivo organ perfusion experi-
ments are excluded from the rules and regulations associated with
the utilization of live animals for experimental research. Further-
more, the model can be adapted for use with various organs [21],
most of which show increased microanatomic comparability to
humans [22]. The expertise gained by running ex vivo perfusions
of porcine organs is also instrumental for the translation to experi-
mental ex vivo perfusion of human organs. Human ex vivo spleen
perfusion to study the early steps of infection, in particular by S.
pneumoniae and SARS-CoV-2 infection, has been recently granted
(ClinicalTrials.gov identifier NCT04620824).
The samples collected from these infection models are prepared
through a common approach of flash freezing and storage at

80  C for future sectioning which is done using a cryostat—a
cryogenic temperature maintaining cutting device often used in
hematoxylin and eosin (H&E) and tissue sample preparation. The
sectioned samples are then prepared for visual analysis using immu-
nohistochemistry (IHC), a technique based on specific target
 Analyzing Macrophage Infection at the Organ Level 407

binding of primary antibodies [23], which is visualized using fluo-

rochrome conjugated secondary antibodies to detect the bound
primary IHC antibodies. Entire immunofluorescence-stained sam-
ples are then visualized using an automated quantitative imaging
system to acquire an image of the whole slice section as opposed to
the limited field-of-view of classic confocal microscopes.
The scanned images are subsequently analyzed via inForm, an
automated image analysis software, or through a more manual
approach using Fiji, an open-source image analysis tool. This allows
for an unbiased distribution assessment of bacteria within the tis-
sue, leaving less room for unconscious bias by the researcher as is
often encountered using classical confocal microscopy. By using
these image analysis techniques, the colocalization of the pathogen
to the various host tissue cells can be quantified, and the structural
distribution of the immune cells within the tissue can be deter-
mined. When used alongside traditional CFU enumeration, these
results provide a stable ground for more robust conclusions by way
of increased experimental outputs, therefore allowing greater
insight into the host–pathogen interactions within tissue on a
cellular level. This method has allowed our group to assess bacterial
distribution in various host cells in multiple experiments, allowing
for a clearer overview of the steps in systemic infection throughout
an infection time course.

2 Materials

All reagents and solutions should be stored as per the manufac-

turer’s instructions, unless otherwise stated.

2.1 Bacterial Strains 1. Bacteria laboratory stock.

and Culture Conditions 2. Broth medium.
3. Petri dishes.
4. Broth medium with agar.
5. Distilled water.
6. Autoclave.
7. Defibrinated horse blood.
8. Universal tube with tight cap.
9. Spectrophotometer.
10. 50% v/v glycerol in phosphate-buffered saline (PBS).
11. Cryovial.
12. 96-well plates.
408 Ryan G. Hames et al.

2.2 Murine 1. Broth medium.

Infection Model 2. Broth agar base (BAB) plates.
3. Bacterial infection stock (see Subheading 3.1).
4. Defibrinated horse blood.
5. PBS.
6. 1.5 mL centrifuge tubes.
7. 96-well plates.
8. HEPA-filtered cages for mice.
9. 29G insulin syringes.
10. 70% v/v ethanol or equivalent disinfectant (such as industrial
methylated spirit (IMS)).
11. Heating chamber for mice.
12. Mechanical restraining device for mice.
13. Dissection board.
14. Pins.
15. Sterile forceps.
16. Sterile scissors.
17. 40 μm cell strainers.
18. Cell strainer pestle or 5 mL syringe.

2.3 Ex Vivo Porcine 1. Perfusion circuit (see Note 1).

Spleen Perfusion 2. Major reservoir for sourcing of blood.
3. Centrifugal pump and corresponding control console (see
Note 2).
4. Oxygenator.
5. Heat exchange unit.
6. Venous reservoir to stimulate portal vein flow.
7. Infusion set (one per vessel to be perfused).
8. Large sterile container for collection of blood.
9. 25,000 IU/mL heparin solution.
10. Scissors.
11. Forceps.
12. Scalpels.
13. Large durable sterile bag to hold porcine organs (see Note 3).
14. Porcine organs (see Note 4).
15. Urine collection catheters (see Note 5).
16. Soltran preservative solution (one 500 mL bag per organ).
17. Large ice container and ice.
 Analyzing Macrophage Infection at the Organ Level 409

18. Saline infusion bag (one 250 mL bag per organ).

19. Sterile 1 L measuring cylinder.
20. 8.4% w/v sodium bicarbonate solution.
21. 0.5 mg epoprostenol sodium solution.
22. 5 mL or 10 mL Syringes.
23. Suture kit.
24. Surgical glue.
25. 2 mL lithium heparin calcium-balanced tubes.
26. Blood gas analyzer.
27. Bacterial growth medium and agar plates (see Subheading 3.1).
28. 96-well microtiter plates.
29. 40 μm cell strainers.
30. Cell strainer pestle (optional).
31. Tissue embedding molds.

2.4 Sample Freezing 1. Metal freezing container.

and Sectioning 2. Ice container with dry ice.
3. 2-methylbutane.
4. Embedding molds.
5. OCT embedding matrix.
6. Forceps.
7. Cryostat (with anti-roll plate and compatible disposable micro-
tome blades).
8. Polylysine adhesion microscopy slides.
9. Relevant primary and secondary antibodies.
10. Blocking solution.
11. Hydrophobic pen or staining rack.
12. Fixation solution.
13. 0.1% v/v Triton™ X-100.
14. PBS.
15. Distilled water.
16. Antifade slide mountant.
17. Slide coverslips.
18. Clear nail polish or coverslip adhesive.

2.5 Whole Slide 1. Vectra® Polaris™ Automated Quantitative Pathology Imaging

Scanning System.
2. 70% v/v IMS.
410 Ryan G. Hames et al.

2.6 Image Analysis 1. Phenochart whole slide viewer.

2. inForm automated image analysis software.
3. Fiji image processing package.
4. Bio-Formats Fiji Plugin.

3 Methods

3.1 Bacterial Strains Bacteria are cultured to the mid-logarithmic phase and aliquots are
and Culture Conditions stored at
80  C with a cryopreservative such as glycerol. The
media selected will depend on the requirements of your bacterial
strain. For example, S. pneumoniae grows well in Brain Heart
Infusion (BHI) broth and BAB culture plates with 3% v/v defibri-
nated horse blood.

3.1.1 Broth 1. Prepare the relevant bacterial growth medium (e.g., BHI;
14.8 g in 400 mL distilled water) and autoclave at 121  C 15 psi
(103 kPa) for 20 min.
2. Ensure color and media pH is within the parameters recom-
mended by the manufacturer.
3. Store media at room temperature (RT).

3.1.2 Blood Agar Base 1. Prepare BAB (16 g in 400 mL of distilled water) and autoclave
(BAB) Culture Plates + 3% at 121  C 15 psi (103 kPa) for 20 min.
v/v Horse Blood 2. Once autoclaved, allow to cool to around 56  C, then add 3%
v/v sterile defibrinated horse blood (12 mL blood in 400 mL
molten agar) and mix (see Note 6).
3. Pour the agar mix into Petri dishes. 400 mL of medium should
make roughly twenty 100 mm  15 mm plates (around 20 mL
per plate).
4. Leave to set in a sterile environment; once set the plates can be
inverted. For long-term storage, store inverted at 4  C.

3.1.3 Bacterial Infection 1. From laboratory stocks, streak to single colonies on blood agar
Stocks plates. Record strain identifier.
2. Confirm the strain, in case it carries an antibiotic resistant
marker, by placing an antibiotic disc onto the plate at the site
of the initial streak with flamed forceps.
3. Incubate the plates inverted at the ideal temperature and time
conditions for the strain selected.
4. Inoculate a sweep of colonies into 10 mL broth in a universal
tube with a tight cap.
5. After 3–3.5 h, check the OD600 nm of the culture. The OD600
nm of the culture should be between 0.4 and 0.6; this
 Analyzing Macrophage Infection at the Organ Level 411

corresponds to the mid-logarithmic phase of bacterial growth.

The CFU can be predicted accurately during the logarithmic
phase.
6. Add 50% v/v glycerol in PBS to the bacterial culture, such that
the final concentration of glycerol is 10%.
7. Divide into 1 mL aliquots in sterile cryovials and freeze at

80  C.
8. After aliquots have been stored at
80  C for over 24 h,
determine the viable CFU/mL by way of 1:10 serial dilutions
and spot each dilution on a BAB plate to calculate the volume
required for the infection dose (see Note 7).
9. Check for contamination by streaking to single colonies on
BAB plates.

3.2 Murine All procedures performed herein were done in the UK in accor-
Infection Model dance with the UK Home Office license P7B01C07A, and were
approved by the University of Leicester Ethics Committee. Proce-
dures documented here are for example purposes only; the actual
procedures utilized in experimental projects should be in line with
the relevant legislation and regulations applicable to the country
where experiments are undertaken.

3.2.1 Infection Dose 1. Thaw an aliquot of bacteria at RT and add the corresponding
volume of broth or PBS, based on the viability counts of the
stocks, to reach the desirable infection dose (e.g.,
1  107 CFU/mL of S. pneumoniae is required for a dose of
1  106 CFU per mouse with an 100 μL inoculation volume).
2. Make a serial dilution of the infection dose in a BAB plate, and
incubate the plate at the optimal time and temperature for the
strain used, to retrospectively determine the viable CFU in the
inoculum (see Note 7).
3. The prepared infection dose should be used within 30 min and
kept at RT.
4. Prepare the individual doses of 100 μL in insulin syringes
equipped with a 29G needle immediately prior to infection.

3.2.2 Intravenous Route 1. Acclimate 6–8-week-old, sex-matched mice in the animal facil-
of Infection ity environment for at least 1 week prior to infection. Animals
should be acclimated under the standard lighting and temper-
ature conditions in individually HEPA-filtered cages with ster-
ile bedding and nesting and free access to food and water
provided ad libitum.
2. Ear tag or tattoo the mice for identification.
412 Ryan G. Hames et al.

3. On the day of infection, carefully remove the mouse from the

cage using a plastic tube to reduce potential stress for the
animal.
4. Place the animal in a heating chamber at 37  C for between
5 and 10 min (see Note 8).
5. Restrain the animal using the mechanical restraint device of
your choice with the tail protruding (see Note 9).
6. Wipe the tail with antiseptic solution such as 70% v/v ethanol.
7. Immobilize the tail and rotate 90 to access the lateral tail vein.
8. Align the needle parallel to the tail with the beveled edge of the
needle facing up.
9. Insert needle into vein starting at the tip of the tail (distally) at
about a 30 angle (see Note 10).
10. Administer the inoculum slowly, removing the needle after
completing the injection.
11. Apply gentle pressure with gauze until bleeding has stopped.
Ensure the mouse is appropriately marked and return it to
the cage.
12. Monitor animal for 5–10 min to ensure homeostasis and nor-
mal behavior.

3.2.3 Monitor Signs 1. Following IV inoculation, monitor the mice every 6–8 h for
of Disease and Euthanasia determination of signs of disease (see Note 11).
2. At predetermined time points after infection or upon reaching
the severity limit outlined in the project license, euthanize the
mice using an approved humane killing method such as cervical
dislocation. Confirm the mouse is dead by cutting the femoral
artery (see Note 12).

3.2.4 Organ Recovery 1. Before organ collection, prepare and label two sterile tubes for
and Bacterial Enumeration each organ to be collected. Add a small volume of PBS or broth
into each tube.
2. Disinfect the work surfaces using 70% v/v ethanol. On a sterile
surgical field, prepare a cleaned dissection board with all sterile
dissection instruments.
3. Pin the mouse carcass to the dissection board with the abdo-
men facing up. Spray the fur of the chest and abdomen with
disinfectant such as 70% v/v ethanol to reduce contamination
by hair.
4. Using sterilized forceps, grasp the skin above the urethral
opening and cut along the ventral midline from the groin to
the sternum using fine point sharp scissors. Peel the skin back
from the peritoneal wall underneath using forceps.
 Analyzing Macrophage Infection at the Organ Level 413

5. With another pair of sterilized scissors and forceps, carefully cut

the peritoneal wall and open up the abdomen. Collect the
spleen, located behind the stomach, and the whole liver.
6. Divide the organs into two portions and place in the prelabeled
tubes. One portion will be placed in a PBS-containing tube for
subsequent embedding in OCT matrix for later microscopy
analysis (see Subheading 3.4). The second portion will be
placed in a broth-containing tube for subsequent bacterial
enumeration after the sample has been homogenized.
7. Dispose of infected carcasses according to the appropriate
guidelines and regulations in place at your institution.
8. Weigh the tubes containing tissue and calculate the weight of
the sample in each tube.
9. For the tube containing the organ suspended in broth, thor-
oughly homogenize the sample by mechanically mashing into a
40 μm cell strainer with a cell strainer pestle or plunger of a
5 mL syringe before washing through with 1 mL bacterial
growth media.
10. Determine the number of viable bacteria in the tissue by
performing serial 1:10 dilutions of the homogenates in sterile
broth and plating on a BAB plate (see Note 7).
11. After incubation, count the viable number of colonies and
calculate the bacterial load in each organ as CFU per gram of
tissue, to normalize for different sample weights.

3.3 Ex Vivo Porcine To prevent extensive ischemia resulting from delays in perfusion,
Liver and Spleen the circuit should be set up prior to organ retrieval. The circuit
Coperfusion Model described herein consists of a reservoir which supplies blood to the
portal circulation and one which supplies the systemic circulation.
3.3.1 Set up of the Ex A diagrammatic representation of the liver–spleen perfusion circuit
Vivo Perfusion Circuit can be found in Fig. 1.
1. Remove the perfusion circuit from its packaging under aseptic
conditions.
2. Connect the drainage lines which will originate from the infe-
rior vena cava and splenic vein to the systemic reservoir.
3. Connect the systemic reservoir to the centrifugal pump fol-
lowed by the oxygenator.
4. Following the oxygenator, the line is split into three outlets.
These should be connected via a cannula to the hepatic artery,
splenic artery, and the reservoir which supplies blood to the
hepatic portal vein once the organ is retrieved.
414 Ryan G. Hames et al.

Fig. 1 The ex vivo normothermic porcine liver and spleen coperfusion. A schematic of the complete perfusion
circuit. The circuit described here consists of a sterile organ chamber that houses the porcine spleen and liver,
cannulated via the splenic artery, hepatic artery, and portal vein respectively. Autologous porcine blood
supplemented with heparin drains passively into a blood reservoir which is fed by saline infusion bags
supplemented with epoprostenol sodium. Blood subsequently flows through the pump and water bath which
ensures physiological temperature and pressure. Following oxygenation, the blood line splits into three, one
for each vessel: splenic artery, hepatic artery and portal vein (via a portal reservoir). The pressure of each
vessel is relayed to, and can be controlled by, the control console. Once the perfusion is deemed stable, the
circuit can be infected via syringe, and blood samples taken from, the inlet/outlet—commonly a three-way
stopcock. (Figure was generated using BioRender)

3.3.2 Organ Retrieval 1. Allow the local qualified team at the abattoir to euthanize pigs
by exsanguination from the jugular vein following stunning
(UK common practice) and collect around 3 L of autologous
blood in a container supplemented with 25,000 units of
heparin.
2. For spleen retrieval: perform a laparotomy and sternotomy
incision and access the celiac trunk to isolate the main splenic
artery and vein.
3. Cannulate the main splenic artery in situ with a catheter of
appropriate size and secure by tying in place with suture thread.
Carefully ligate and divide other associated vessels.
 Analyzing Macrophage Infection at the Organ Level 415

4. Remove the spleen and allow a member of the team not

involved in surgery to immediately infuse with 500 mL Soltran
preservative solution by gentle squeezing of the infusion bag.
5. For liver retrieval: divide the diaphragm and pleurae, before
ligating the suprahepatic inferior vena cava, thoracic aorta, and
esophagus.
6. Cannulate the portal vein and hepatic artery in situ as above,
before clamping and dividing additional vasculature and con-
joining connective tissue.
7. Remove the liver and gall bladder en bloc, before infusing each
vessel with 500 mL of Soltran preservative solution as above.
8. Put the organs into a sterile bag containing Soltran preservative
solution, and place on ice for transportation back to lab. The
ischemic time should be noted for consideration in analysis of
results.

3.3.3 Ex Vivo Perfusion 1. Flush organs with 250 mL saline by gentle squeezing of the
and Bacterial Infection infusion bag to remove excess preservative solution prior to
plugging into the perfusion circuit.
2. Using a measuring cylinder, pour 3 L of autologous pig blood
into the systemic reservoir, and apply a gentle flow rate to the
circuit such that each organ input line is gently releasing blood.
Start the saline infusion lines.
3. Plug the cannulated vessels into the circuit ensuring that no air
bubbles are introduced.
4. Ensure normothermia by setting the heat exchange unit to
37  C, and set the oxygenator to a physiological level of 2 L/min.
5. Once plugged in, set the perfusion pressure to 80 Hg/mm for
the splenic artery and hepatic artery, and 10 Hg/mm for the
portal vein for the length of the experiment. Monitor the flow
rate through each vessel at hourly intervals.
6. Twenty to thirty milliliters sodium bicarbonate can now be
added to the circuit perfusate to minimize the effect of tissue
acidosis, in addition to epoprostenol sodium which should be
added to the saline infusion bag to a final concentration of 2.5
μg/mL to facilitate vasodilation and thus effective oxygenation
of the tissues (see Note 13).
7. Once a stable perfusion flow rate and blood gas parameters
have been achieved, infect the circuit with bacteria via syringe
into the circuit input, commonly a three-way stopcock (see
Note 14).
8. At each predetermined time point, use a scalpel to excise a
1 cm2 tissue biopsy from the organ, which should then be
halved, and take 10 mL of blood via one of the circuit inputs
(see Note 15).
416 Ryan G. Hames et al.

9. Suture or surgically glue the subsequent wound to prevent

leakage of blood from the biopsy site and a drop in flow
pressure.
10. Transfer the blood sample to a lithium heparin calcium-
balanced tube and analyze using a blood gas analyzer to deter-
mine parameters such as pH, hemoglobin oxygenation, lactate
production and levels of physiological electrolytes to confirm
maintenance of physiological blood gas parameters.
11. Serially dilute the blood sample into relevant bacterial agar
growth media plates using a 96-well plate and plate 5 μL of
each dilution on appropriate agar for blood CFU enumeration.
Incubate under conditions appropriate for your bacterium (see
Note 7).
12. To determine bacterial burden within the organ, weigh one
half of the biopsy and thoroughly homogenize by mechanically
mashing into a 40 μm cell strainer with a cell strainer pestle or
plunger of a 5 mL syringe before washing through with 1 mL
bacterial growth media. Serially dilute and plate the homoge-
nate as above. Biopsy CFU should be calculated as CFU per
gram to normalize for different biopsy weights.
13. For later microscopy analysis, flash-freeze the other biopsy half
in OCT within a mold as outlined in Subheading 3.4.1.

3.4 Sample 1. Place a metal freezing container on dry ice and fill it with
Preparation 2-methylbutane up to 1 cm in depth (see Note 16).
3.4.1 Sample Freezing 2. Fill approximately one third of an embedding mold with OCT.
3. Using forceps, place the sample in the middle of the mold.
Ensure the biopsy is submerged in OCT by adding more if
necessary.
4. Place the mold into the 2-methylbutane, being careful not to
get any liquid onto the sample, and let it freeze.
5. Take the frozen sample out of the 2-methylbutane and store it
at
80  C until use.

3.4.2 Sample Sectioning 1. Switch on and set up your cryostat as per manufacturer’s
instructions, setting the internal temperature to
20  C. Insert
the blade and the glass anti-roll plate (see Note 17).
2. Once the temperature is reached, remove the OCT block from
its mold. Add a small amount of OCT onto the specimen disk.
Once the OCT starts to freeze, mount the sample onto the
specimen disk. Once the OCT fully freezes (~5 min) and the
sample is securely attached to the specimen disk, mount the
specimen disk onto the stage.
3. Set the section thickness (see Note 18).
4. Move the specimen as close as possible to the blade by using the
stage forward button (if applicable).
 Analyzing Macrophage Infection at the Organ Level 417

5. Begin sectioning. Handwheel rotation must be smooth and

controlled in order to keep sections even (see Note 19).
6. Place the section onto a polylysine adhesion slide by suspend-
ing the slide just above the sample.
7. Once the section is dry, place the slide into a slide rack on dry
ice for the remainder of the sectioning session. When all the
samples have been sectioned, either thaw and begin staining or
transfer the slides to
80  C storage until use.

3.4.3 Indirect It is important to keep the antibodies, dyes and section out of light
Immunohistochemistry as much as possible throughout the staining process to reduce the
Staining amount of photobleaching that occurs and thus increasing the
quality and fluorescence intensity of the sample.
1. Select the relevant primary and secondary antibodies to label
individual cell types and bacteria, as well as fluorescent dyes to
label nuclei (commonly DAPI) and cell macromolecules (see
Note 20).
2. Dilute the primary and secondary antibodies and dyes in block-
ing solution, and temporarily store away from light at 4  C (see
Note 21).
3. On the slide, circle around the section with a hydrophobic pen
and let dry. Alternatively, place the slides into a specialized
staining rack.
4. Add the fixation buffer (see Note 22).
5. Wash thoroughly with PBS then permeabilize the sample by
adding 0.1% v/v Triton X-100 for 10 min at RT.
6. Block the sample with blocking solution for 30 min to 1 h.
7. Add the primary antibodies and incubate for 1 h at RT (see
Note 23).
8. Wash thoroughly with PBS.
9. Add the secondary antibodies with any dyes and incubate for
45 min at RT (see Note 24).
10. Wash thoroughly with PBS and add DAPI for 10 min.
11. Wash thoroughly with PBS and one final time with distilled
water.
12. Add a drop of mountant onto the sample followed by a cover-
slip (see Note 25).
13. Wipe away any excess mountant from the periphery of the
coverslip and seal the around the edges of the coverslip with
cover slip adhesive or clear nail polish. Let the adhesive dry.
14. The slide is ready to use or can be stored at either 4  C or

20  C until use (see Note 26).
418 Ryan G. Hames et al.

3.5 Sample 1. Wipe the slides to be scanned with 70% v/v IMS.
Microscopy and Image 2. With a fine permanent marker, draw a complete box on the
Analysis underside of the slides around the sample.
3.5.1 Fluorescence 3. Load the slides into the slide carriers (see Note 27).
Whole-Slide Scanning 4. Load the Vectra Polaris software, and load the carriers into the
slide scanner (see Note 28).
5. Select “Edit Protocol”!“New. . .” to create a new protocol. A
new protocol should be created for each combination of anti-
bodies used on the slides to be scanned. Enter a suitable
protocol name, ensure fluorescence imaging mode is selected,
select “multispectral slide scan” (Opal 4 color) and select
(or “create”) a study in which to save the protocols and result-
ing scan images (see Note 29).
6. Under the slide scan settings, change the pixel resolution to
“0.25 μm (40)” and select “auto update” for the exposure
times when prompted. Select “Select Scan Bands” and check
the desired Opals depending on the fluorophores used in this
particular protocol. The color of signal resulting from each
Opal can also be changed here (see Note 30).
7. Select “Scan Exposures”!“Load Carrier” and select a rack
containing a slide stained with the antibody panel
corresponding to the protocol that is being created. With the
DAPI channel automatically selected, locate the sample by
clicking various locations on the appropriate slide depiction to
the right of the live image, or by clicking on the sides of the live
image itself. Alternatively, use “Take Overview” to scan the
entirety of all slides in the carrier, hence showing exactly
where the samples are located (see Note 31).
8. Move around the sample and use “Autofocus” and “Autofluor-
escence” until there are no more overexposed (red) areas in the
image. Repeat this step for the remaining antibody channels,
then select “Back” (see Note 32).
9. Save the protocol. Repeat steps 5–9 to create a new protocol
for each antibody combination used in the slides to be scanned.
Once all protocols have been created, click “Back” (see
Note 33).
10. Select “Scan Slides.” Select the gray circle next to each slot to
bring up the edit slides window. Change the task for all slides to
be scanned to “Scan Slides,” then enter the appropriate study,
protocol and desired slide ID. Select “Scan.”
11. Once scanning is complete, the scanned images of the sample (.
qptiff file) and slide label photos (.tiff file) will appear in the
study folder which was selected prior to scanning. The scans
can then be opened with Phenochart, or analyzed using Fiji or
inForm.
 Analyzing Macrophage Infection at the Organ Level 419

Fig. 2 inForm image analysis window. A screenshot of the inform image analysis window with tools utilized in
this methodology labeled as appropriate. Options on the horizontal toolbar include zoom, draw training regions
(used in the tissue segmentation step), component/composite view (used in the cell segmentation and
phenotyping steps), and edit the view for adjusting channel brightness and contrast. Options on the vertical
toolbar include buttons for the toggling on/off of the tissue segmentation, nuclear segmentation and cytoplasm
segmentation maps along with the cell phenotypes. The image batch containing image stamps is shown at the
bottom of the window

3.5.2 inForm Image Whole slide scan images cannot be directly opened in the inForm
Analysis software, therefore “stamp” sections of the images to be analyzed
must be prepared beforehand in Phenochart. When analyzing the
images in inForm, refer to Fig. 2 for a view of the processing
window with labeled buttons that are used in the methodology
described here.
1. Open the desired image(s) to be analyzed in the Phenochart
software and login using the appropriate institutional
password.
2. Select “Stamp” ! “Select for: inForm Projects”. Select an
appropriate size in fields for the stamp and select areas on the
image to be analyzed in inform (see Note 34).
3. Open the images to be analyzed in inForm through
“File”!“Open”!“Image.” The stamps of each image will be
opened separately in the batch view.
4. Select “Configure. . .” to configure the analysis workflow.
Select “Trainable Tissue Segmentation,” “Adaptive Cell Seg-
mentation,” and “Phenotyping” while skipping the
“Score” step.
420 Ryan G. Hames et al.

5. Prepare the images by selecting which markers are to be

unmixed under the “Spectra for Unmixing” tab, and use
“Edit Markers and Colors. . .” to assign appropriate names
and colors for each cell marker. Under the standard protocol,
all markers will be included in unmixing. Select “Prepare All”
(see Note 35).
6. Once the image batch has finished unmixing, alter the bright-
ness and contrast of each component if needed by selecting the
“Edit the view” button on the horizontal toolbar. Click
“Advance.”
7. Create new tissue categories under the “Tissue Categories”
tab. A tissue category should be created to represent each
marked cell type of interest, with an additional background
category created to represent the blank spaces between cells
(see Note 36).
8. Under the “Components for Training” tab, select which mar-
kers should be taken into consideration by the machine
learning component when segmenting the tissue into its
respective categories (see Note 37).
9. Select a pattern scale; in most instances a medium pattern scale
will suffice (see Note 38).
10. Using the “Draw training regions” button above the image,
and with a tissue category selected, draw around a region of
that respective tissue category on an image. Repeat this step to
create training regions for each tissue category, including a
region with no cells (or an area of a blank slide) for the
background category (see Note 39).
11. Select “Train Tissue Segmenter” and “Segment Image” once
the training is complete. The tissue segmentation map will now
overlay the image, displaying each tissue category in their
respective color as shown under the “Tissue Categories” tab
(see Note 40).
12. In many cases, the first tissue segmentation map will not be
sufficiently accurate. In this scenario, additional training
regions will have to be drawn and the tissue segmenter
retrained until a segmentation map of sufficient accuracy is
created. Once this is achieved, select “Segment All” and
“Advance” onto the next step (see Note 41).
13. If a cytoplasmic marker is included in your staining panel, check
the “Cytoplasm” option under the “Segment” tab alongside
the preselected “Nuclei” option (see Note 42).
14. On the “Cell Segmentation Settings” tab, add a component
under the “Components” section. The marker selected should
be the nuclear marker (commonly DAPI). Selecting the ellipses
adjacent to the marker will allow the alteration of typical
 Analyzing Macrophage Infection at the Organ Level 421

fluorescence intensity of the nuclei; this should be adjusted

until the entirety of each nucleus within the image is high-
lighted (see Note 43).
15. Similarly, if a cytoplasmic stain or antibody is used, add an
additional component, select the relevant antibody/stain
from the dropdown menu and again select the neighboring
ellipses to check the cytoplasm option.
16. Under the “Nuclear Component Splitting” tab, select the
ellipses to select the representative quality of the nuclear stain-
ing within the image, and alter the splitting sensitivity such that
adjacent nuclei can be separated while whole nuclei are not.
Adjust the “Minimum Nuclear Size” to prevent partial selec-
tion of nuclei with a low fluorescence intensity (see Note 44).
17. For any cytoplasmic markers, alter the cytoplasm thickness
under the “Cytoplasm Settings” tab such that the signal is
overlaid sufficiently by the cytoplasm segmentation map (see
Note 45).
18. Select “Segment Image.” The nuclear and cytoplasmic seg-
mentation map will be overlaid on the displayed image, indi-
cating which areas the program has identified as nuclei in
green, and cytoplasm (if segmented) in multicolor. If the selec-
tion and segmentation are accurate, select “Segment All.”
Reevaluate the segmentation on additional images and click
“Advance” if appropriate. Alternatively, alter the cell segmen-
tation settings and resegment the images until accurate.
19. Under the “Phenotypes” tab, add new cell phenotypes for
bacteria positive and bacteria negative cells (see Note 46).
20. Select the “Edit a cell’s phenotype” button on the horizontal
toolbar above the image. Assign a phenotype to at least five
cells which are bacteria positive and five cells which are bacteria
negative. Train the classifier and select “Phenotype All” (see
Note 47).
21. Under the “Export” tab, select an empty export directory
folder in which to save the data and/or images. Select which
images (if any) from the analysis workflow that you wish to
save, along with the desired format, and select the cell segmen-
tation data table. Select “Export for All.”
22. A separate “cell_seg_data” and “cell_seg_summary_data” file
will be created for each image stamp in the selected export
directory. Opening the summary_data file will provide infor-
mation on how many cells of each tissue category are bacteria
positive or negative. In turn, this allows quick and simple
enumeration of the percentage of each cell type which are
infected, along with the percentage of total infected cells
which belong to each cell type (see Note 48).
422 Ryan G. Hames et al.

3.5.3 Fiji Image Analysis 1. Load the Fiji software and open the image to be analyzed by
selecting “Plugins”!“Bio-Formats”!“Bio-Formats
Importer.”
2. In the “Bio-Formats Import Options,” ensure that the color
mode is set to composite (autoscale checked) and that channels
are split into separate windows.
3. Select the series with highest resolution (Series 1) in the “Bio-
Formats Series Options” screen. Click “OK.”
4. On the nuclear marker (commonly DAPI) channel, use the
“Polygon selections” tool in the toolbar to create a region of
interest (ROI) around the entire sample. Add this ROI to the
ROI manager by pressing Ctrl+T, and rename it as “whole slice
ROI” and save.
5. On a macrophage marker channel, alter the threshold by select-
ing “Image”!“Adjust”!“Threshold.” Ensure that the “Dark
background” box is unchecked, and set the maximum fluores-
cence (bottom slider) to its highest level, and the minimum
fluorescence (upper slider) to an appropriate level as to include
the majority of fluorescent signal while negating background.
Click “Apply” to apply the selected threshold levels (see Note
49).
6. The image should now be binary: black with white signal. If
this is reversed, select “Process”!“Binary”!“Options,” select
“Black background” and reapply the threshold settings.
7. If there should be any artefacts in the image that have been
included in the threshold selection, these can be omitted by
drawing around them with the “Polygon selections” or “Free-
hand selections” tools and pressing the delete key.
8. Omit any signal outside of the sample by selecting the “whole
slice ROI” from the ROI manager, inverting the selection by
selecting “Edit”!“Selection”!“Make Inverse” and pressing
the delete key (see Note 50).
9. Select “Edit”!“Selection”!“Create selection” to create
ROIs around the macrophage marker signal. Ensure that the
binary white signal and not the surroundings are selected and
subsequently add the ROIs to the ROI manager using Ctrl+T.
Rename this ROI as the appropriate macrophage marker ROI
and save.
10. On the bacteria channel, apply a threshold to include all bacte-
ria within the sample, as described in step 5. The image should
have a black background with white signal; alter the Binary
settings as outlined in step 6 to reverse this (see Note 51).
11. On the bacteria channel, select the whole slice ROI. Select
“Analyze”!“Analyze Particles. . .” and check the “Summa-
rize” box. The size can be adjusted depending on the average
 Analyzing Macrophage Infection at the Organ Level 423

size of your bacteria present in the sample, although a degree of

leeway should be incorporated (see Note 52).
12. Repeat step 11 with the macrophage marker ROI selected.
13. The summary table gives two values of importance: Total Area
(indicating the area of bacterial signal within the ROI), and %
Area (indicating the area of the ROI that contains bacteria).
This allows for the determination of bacterial location within
the tissue at different time points by analyzing the colocaliza-
tion of bacteria with different cell types. The % Area value can
also be used to analyze the growth of bacterial foci within
certain cell subtypes at different time points within a tissue,
and can be used in conjunction with the infected cell percent-
age data from the inForm analysis to build a more detailed view
of cell infection dynamics within samples over time (see Note
53).

4 Notes

1. The circuit described herein is similar to those used for extra-

corporeal bypass surgery, and was custom made by Medtronic
(Netherlands). This circuit is optimized for coperfusion of the
liver and spleen, but circuits can be adapted for different organ
systems [6, 20, 24].
2. The pump facilitates the blood flow through the cannulated
vessels while the control console provides pressure regulation
and readings.
3. We use a 13  10 in sterile intestinal bag for this purpose.
4. The specific organs (liver and spleen) used in this protocol were
sourced from Joseph Morris Abattoir (South Kilworth, Leices-
tershire, UK), which is in close proximity to the authors’ labo-
ratory. Organs were removed immediately postmortem
following the UK guidelines for processing of animals for
food production. Organs may be sourced from any abattoir
provided sampling does not interfere with local best practice.
This protocol is also amenable to use with experimental specific
pathogen-free (SPF) reared animals in line with a relevant
home office project license.
5. A catheter is required for each vessel and their size will depend
on the vessel diameter. In this protocol the hepatic artery and
splenic artery are cannulated with a 14F catheter, while the
portal vein is cannulated with a 12F catheter.
6. The bottle should be warm but not hot to the touch. When
mixing the agar, take care not to form bubbles which could set
in the Petri dish. If the blood agar plates set brown, then the
424 Ryan G. Hames et al.

bottle was too warm when the horse blood was added. If solid
chunks of agar are poured, the agar was too cold when the
horse blood was added.
7. Serial dilutions from neat to 1:100,000 can be achieved by
diluting 20 μL into 180 μL media in a 96-well plate five
times. A multichannel pipette can then be used to plate 5 μL
of all dilutions onto the plate. To calculate CFU/mL, count
the number of colonies in the highest dilution where bacteria
can still be accurately counted, multiply by the dilution factor
(e.g., 1  104 for the fourth dilution of the neat sample) and
divide by 0.005.
8. This causes vasodilation and provides better vein visibility. Do
not overheat the animal by leaving the animal in the chamber
for more than 10 min. This will cause dehydration and make
accessing the vein and administering the dose more difficult.
9. Sedation or anesthesia for intravenous inoculation is unneces-
sary for trained and experienced personnel. Placing food in the
opposite end of the restraint tube can encourage the mouse to
enter the tube themselves.
10. If the needle is correctly inserted, blood should flash into the
syringe when the plunger is gently withdrawn and the inocu-
lum will flow easily during administration. Do not force the
inoculum—this means the dose is being improperly adminis-
tered to tissue surrounding the vein and can cause distress to
the animal. If the proper placement cannot be confirmed,
attempt to place the needle in a more proximal position
where the vein is wider in diameter.
11. The frequency of which the mice should be observed and/or
scored will depend on the guidelines specific to your project.
12. The time points utilized in the study will depend on experi-
mental objectives; however, a frequently used time course for
the study of early stages of pathogenesis is up to 8 h
postinfection.
13. The volume of sodium bicarbonate that should be added is
dependent on the pH yielded by blood gas analysis and as such
the actual volume required can vary. In our perfusion circuit,
we utilize 200 mL infusion bags of saline, which requires the
addition of 0.5 mg epoprostenol sodium.
14. In our experience, a stable flow rate is reached after around
30 min of perfusion. Our perfusion circuits are infected with
around 1  107 CFU of S. pneumoniae or Klebsiella pneumo-
niae, although the infection bacteria and dosage should be
specifically tailored to the experimental aims and objectives.
Ensure blood is taken into the syringe and dispelled multiple
times to ensure all the bacteria enter the circuit.
 Analyzing Macrophage Infection at the Organ Level 425

15. Sampling times during ex vivo perfusion will depend on the

experimental set up, and the aims of the research project. In
our experimental model, time points were every hour up to 6 h
postinfection. Ensure a biopsy is taken from a well-perfused
area of the organ by ensuring the area is dark red and is warm to
the touch.
16. Metal containers work best as they are good conductors of heat
and can withstand the cold temperatures. 2-methylbutane is
used due to its high thermal conductivity and sample-freezing
efficiency.
17. For spleen and liver samples, we commonly use
20  C. The
optimal sectioning temperature is dependent on the tissue and
can be found in the literature, although
15  C to
25  C is a
common range applicable to most organ types.
18. We commonly use 10 μm sections for liver and spleen samples,
and this thickness is widely used in frozen IHC staining. Thin-
ner sections tend to stain better and produce clearer images but
tear easily. The thickness of the section is ultimately dependent
on the tissue that is being sectioned, and literature should
therefore be consulted for the optimal section thickness.
19. If the section bunches up around the blade, try adjusting the
distance between the edge of the antiroll plate and the edge of
the blade in small increments. In our experience, the antiroll
plate should overhang the edge of the blade by ~2 mm. If the
section rolls while underneath the antiroll plate, adjust the
angle of the glass antiroll plate such that the antiroll plate
becomes closer to the surface underneath. If the section imme-
diately rolls once the antiroll plate is lifted, try briefly warming
the sample by gently placing your sterile-gloved thumb pad on
the specimen for 1–2 s, or raising the internal cryostat temper-
ature slightly. A small paintbrush can also be used to carefully
unroll the sections in these cases.
20. The antibodies and dyes will be dependent on the experimental
aims. The primary antibodies should be specific for the antigen
and species of interest (anti-pig for perfusion samples and anti-
mouse for murine model samples) and all be raised in different
species. Each secondary antibody should be specific for the
species that the corresponding primary antibody was raised
in. Secondary fluorophore-conjugated antibodies should be
used in line with the excitation and emission profile of the
confocal microscope to be used. It is also important that any
dyes used do not have the same excitation wavelength as any of
the fluorophore-conjugated secondary antibodies or other
dyes used in the staining panel.
426 Ryan G. Hames et al.

21. The blocking solution used here is 5% v/v goat serum in PBS;
however, other type of sera used for blocking (i.e., bovine
serum albumin) in IHC are also common. The antibodies
and dyes should be diluted to a concentration outlined by the
manufacturer, although 1:200 for primary antibodies and
1:500 for secondary antibodies are common dilutions.
22. 4% v/v formaldehyde is commonly used in immunohistochem-
istry; however, different fixation buffers such as methanol can
be used in certain applications. The base can also vary depend-
ing on application and often yields different staining intensity
and clarity. A common fixation buffer is PBS, as used here.
23. Alternatively, primary antibodies can often be incubated over-
night at 4  C or for 1 h at 37  C. In some cases, manufacturers
provide specific incubation times for antibodies which should
be adhered to.
24. From this step onwards it is extremely important to keep the
sample out of direct light as much as possible as the
fluorochrome-conjugated secondary antibodies are light sensi-
tive and can become photobleached when exposed to bright
light for extended periods.
25. A coverslip with thickness that is compatible with the micro-
scope should be used. When placing the coverslip onto the
slide, lower down one end slowly to prevent the formation of
air bubbles. If any bubbles form underneath the coverslip, a
gentle rolling pressure to one end can expel the bubbles
through the end of the coverslip.
26. Storage at
20  C is recommended for long-term storage, to
preserve the brightness of the fluorescent tags and dyes,
although care should be taken to minimize the number of
freeze–thaw cycles. 4  C is suitable for short-term storage
when planning to use the samples within the same week.
27. Take a photo of the slides in the carriers to refer back to during
the scan exposures step, thus ensuring that the slides identity
and the location of the sample on the slides can easily be found.
28. Document which slide carriers are placed at which position in
the rack. You will refer back to these when imaging protocols
are assigned to individual slides in the scanning step.
29. Oftentimes, similar sample types stained with the same panel of
antibodies will yield similar fluorescence intensities, that is,
mouse spleens stained with the same antibody combination in
parallel will appear similar. In these cases, one protocol can be
used for all slides harboring the same antibody panel. However,
if the resulting scans show over/underexposed signal between
slides, a protocol should be created for, and assigned to, each
individual slide.
 Analyzing Macrophage Infection at the Organ Level 427

30. If secondary antibodies have not been conjugated to Opal

fluorophores, the Opal corresponding to the fluorophore
used can easily be found online.
31. Locate the sample on the slide by referring back to the photos
taken of the slides in their carriers.
32. The focus can change when moving around the sample if the
section is uneven on the slide. Ensure that autofocus is used
before selecting autofluorescence, as in-focus images often
display higher fluorescence signal intensity. If autofocus is
unable to focus the image, manually use the stage height slider
to bring the image into near-focus before using the autofocus
function.
33. To load different slide carriers when scanning exposures, select
“Unload Carrier,” then “Load Carrier” with the relevant slide
carrier.
34. To make the analysis of the stamps representative of the entire
sample as a whole, we often utilize the largest 3  3 stamp field
size wherever possible and create multiple random stamps per
image. This will ensure that as many phenotypically different
cells as possible can be included in the inForm analysis, while
reducing any bias relating to which areas of the image are
analyzed.
35. For our analysis workflow, bacteria are colored in green, mac-
rophage marker in red, cell stains/additional macrophage mar-
kers in magenta, and autofluorescence in black.
36. For example, a sample stained with DAPI, an anti-bacteria
antibody, and an anti-macrophage antibody will have tissue
categories consisting of macrophage antibody positive, macro-
phage antibody negative, and background.
37. This usually consists of the cell markers and DAPI and should
not include the bacteria or autofluorescence components as the
presence or absence of these markers do not impact on deter-
mining a cell’s phenotype.
38. The pattern scale relates to the size of the area of the image
which will be selected to indicate the different tissue categories.
If the area of one or more cell markers within the image is
particularly small, the pattern scale can be set to small.
39. We recommend zooming in to the image and accurately draw-
ing around a region of around 3–5 adjacent cells of the same
marker. The program uses machine learning to determine the
average fluorescence intensity of each of the markers selected
under the “Components for Training” tab within the training
region. Therefore, care should be taken to only include the
markers of interest, as any area without the markers would
reduce the average intensity of the markers of interest. When
428 Ryan G. Hames et al.

selecting a training region to represent the background, we

commonly use an area outside of the sample—this minimizes
the chance of accidently including a nucleus with low fluores-
cence in the background training region.
40. We recommend only segmenting a single image after each
training of the tissue segmenter as segmentation of the entire
image batch can very lengthy, especially if multiple training
rounds are required. We have found that the most efficient
way of segmenting tissue is to carry out multiple training
rounds (if required) on a single image until accurate, before
training the rest of the image batch to confirm sufficient accu-
racy. Sometimes the trainer will not reach 100%, in this case
stop the training when the accuracy ceases to increase.
41. The tissue segmentation map can be toggled on/off by select-
ing the appropriate button on the vertical toolbar beside the
image. This will help in identifying areas of the tissue which
have been miscategorized. If fluorescence intensity and/or
staining quality of the images within the batch varies drastically,
it is unlikely that tissue segmentation will be accurate for each
image even after multiple training rounds. In this case, images
with similar staining quality/fluorescence intensity should be
grouped and subsequently analyzed as separate inForm
projects.
42. Including a cytoplasmic marker into the analysis is especially
important when analyzing tissues where nuclei may be dis-
persed—if only nuclei are used to segment cells, only cells
with nuclei colocalizing with bacteria will be counted as posi-
tive in later phenotyping steps.
43. Select the nuclear marker channel option under the “Select a
component Image” button on the horizontal toolbar to visua-
lize the image as an H&E-style image. This makes altering the
relative fluorescent intensity easier and more accurate than
when viewing the color/composite image. The nuclear seg-
mentation map can be toggled on/off with the appropriate
button on the vertical toolbar—this can be used to ensure the
nuclear selection is accurate.
44. Move the training box around different images including areas
of differing fluorescence intensity to ensure that all nuclei
throughout the entirety of each image is selected. This allows
for more accurate nuclear selection.
45. Changing to a component image of the cytoplasmic marker or
to a composite image, or by toggling the cytoplasm segmenta-
tion map, will aid when altering the cytoplasm thickness.
46. Do not use green as an indication of either phenotype as the
nuclear segmentation map also appears green and can result in
difficulty in visualizing cells of this phenotype against the
nuclear segmentation map.
 Analyzing Macrophage Infection at the Organ Level 429

47. Toggle the segmentation maps on/off using the respective

button on the horizontal toolbar to switch between locating
bacteria and visualizing cells. Select the “Select a component
image” button on the horizontal toolbar and select the bacteria
channel to allow easier identification of bacteria. If bacteria
cannot accurately be identified when looking only at the
inForm image, use Phenochart to locate known bacteria within
the stamp and subsequently locate the bacterium in the inForm
image. We recommend selecting as many cells of each pheno-
type as possible across different images, thus allowing the
machine learning classifier to be as accurate as possible.
48. More powerful programs such as “R” can also be used on the
raw “cell_seg_data” files to carry out statistical analysis on
additional parameters such as cell position to determine, for
example, whether infected cells are grouped in close proximity
or spread more randomly about the image. This additional
analysis is beyond the scope of this chapter.
49. Changing the color of fluorescent signal from white to green
(“Image”!“Lookup Tables”!“Green”) can aid when setting
the threshold.
50. To determine whether the inside or outside of an ROI is
selected, the cursor will appear as a standard cursor arrow
within selected areas, and a cross in the nonselected areas.
51. Extensively zoom in to the image for a more precise selection
of bacteria—it is not uncommon for bacteria to have a faint
“glow” around them and it is important that only the bacteria
itself is used for setting the threshold. Utilizing a bacterium
with particularly low fluorescence to set the minimum thresh-
old will ensure that the majority, if not all, of the bacteria within
the sample are included in the threshold selection. This can be
done by finding a bacterium with weak signal in Phenochart
(deselecting the other channels to leave only the bacteria chan-
nel can aid in locating bacteria with low signal), and subse-
quently locating and setting the minimum threshold to that
bacterium in Fiji. Alternatively, repeating steps 10–12 multiple
times with a different bacterium each time can produce more
accurate and representative mean values when the intensity of
bacteria within the image vary significantly, particularly in
thicker sample sections.
52. In our analysis, a lower size limit is set; however, the upper limit
remains at infinity. This is especially important for intracellular
bacteria that replicate to form foci as they may be omitted if the
upper size limit is set too small.
53. The “Count” value should be used with caution—this param-
eter gives the number of particles in the given ROI; however,
bacteria that are in close proximity and whose signal overlap
would be counted as only a single bacterium.
430 Ryan G. Hames et al.

Acknowledgements

We thank John Isherwood and Rohan Kumar for help with the
perfusion of the porcine organs at explant, the staff of Joseph
Morris Butchers, and Sarah Glenn and the staff of the Leicester
Preclinical Research Facility for support with the mouse experi-
ments. The grant was in part supported by a collaboration agree-
ment with the University of Oxford and grants from the MRC
MR/M003078/1 and BBSRC BB/S507052/1 to MRO. ZJ is
funded by BBSRC BB/S507052/1.

References

1. Golding H, Khurana S, Zaitseva M (2018) 8. Chiavolini D, Pozzi G, Ricci S (2008) Animal

What is the predictive value of animal models models of Streptococcus pneumoniae disease.
for vaccine efficacy in humans? The importance Clin Microbiol Rev 21(4):666–685. https://
of bridging studies and species--independent doi.org/10.1128/CMR.00012-08
correlates of protection. Cold Spring Harb Per- 9. Kadioglu A, Cuppone AM, Trappetti C et al
spect Biol 10(4):a028902. https://doi.org/ (2011) Sex--based differences in susceptibility
10.1101/cshperspect.a028902 to respiratory and systemic pneumococcal dis-
2. Pizza M, Scarlato V, Masignani V et al (2000) ease in mice. J Infect Dis 204(12):1971–1979.
Identification of vaccine candidates against ser- https://doi.org/10.1093/infdis/jir657
ogroup B meningococcus by whole--genome 10. Gerlini A, Colomba L, Furi L et al (2014) The
sequencing. Science 287(5459):1816–1820. role of host and microbial factors in the patho-
https://doi.org/10.1126/science.287.5459. genesis of pneumococcal Bacteraemia arising
1816 from a single bacterial cell bottleneck. PLoS
3. Borrow R, Carlone GM, Rosenstein N et al Pathog 10(3):e1004026. https://doi.org/10.
(2006) Neisseria meningitidis group B corre- 1371/journal.ppat.1004026
lates of protection and assay standardization— 11. Iannelli F, Chiavolini D, Ricci S et al (2004)
international meeting report Emory Univer- Pneumococcal surface protein C contributes to
sity, Atlanta, Georgia, United States, 16–17 sepsis caused by Streptococcus pneumoniae in
march 2005. Vaccine 24(24):5093–5107. mice. Infect Immun 72(5):3077–3080.
https://doi.org/10.1016/j.vaccine.2006.03. https://doi.org/10.1128/iai.72.5.3077-
091 3080.2004
4. Romero--Steiner S, Frasch CE, Carlone G et al 12. Oggioni MR, Trappetti C, Kadioglu A et al
(2006) Use of Opsonophagocytosis for sero- (2006) Switch from planktonic to sessile life: a
logical evaluation of pneumococcal vaccines. major event in pneumococcal pathogenesis.
Clin Vaccine Immunol 13(2):165–169. Mol Microbiol 61(5):1196–1210. https://
https://doi.org/10.1128/CVI.13.2.165- doi.org/10.1111/j.1365-2958.2006.
169.2006 05310.x
5. Ercoli G, Fernandes VE, Chung WY et al 13. Kerr AR, Paterson GK, McCluskey J et al
(2018) Intracellular replication of Streptococcus (2006) The contribution of PspC to pneumo-
pneumoniae inside splenic macrophages serves coccal virulence varies between strains and is
as a reservoir for septicaemia. Nat Microbiol 3 accomplished by both complement evasion
(5):600–610. https://doi.org/10.1038/ and complement--independent mechanisms.
s41564-018-0147-1 Infect Immun 74(9):5319–5324. https://doi.
6. Chung WY, Wanford JJ, Kumar R et al (2019) org/10.1128/IAI.00543-06
An ex vivo porcine spleen perfusion as a model 14. Manso AS, Chai MH, Atack JM et al (2014) A
of bacterial sepsis. ALTEX 36(1):29–38. random six--phase switch regulates pneumo-
https://doi.org/10.14573/altex.1805131 coccal virulence via global epigenetic changes.
7. Kadioglu A, Weiser JN, Paton JC et al (2008) Nat Commun 5(1):5055. https://doi.org/10.
The role of Streptococcus pneumoniae virulence 1038/ncomms6055
factors in host respiratory colonization and dis- 15. Ricci S, Janulczyk R, Gerlini A et al (2011) The
ease. Nat Rev Microbiol 6(4):288–301. factor H--binding fragment of PspC as a vac-
https://doi.org/10.1038/nrmicro1871 cine antigen for the induction of protective
 Analyzing Macrophage Infection at the Organ Level 431

humoral immunity against experimental pneu- ex vivo perfused model: results with the por-
mococcal sepsis. Vaccine 29(46):8241–8249. cine liver--kidney circuit over 24 h. Artif
https://doi.org/10.1016/j.vaccine.2011.08. Organs 37(5):457–466. https://doi.org/10.
119 1111/aor.12003
16. Rukke HV, Kalluru RS, Repnik U et al (2014) 21. Daniel CR, Labens R, Argyle D et al (2018)
Protective role of the capsule and impact of Extracorporeal perfusion of isolated organs of
serotype 4 switching on Streptococcus mitis. large animals—bridging the gap between
Infect Immun 82(9):3790–3801. https://doi. in vitro and in vivo studies. ALTEX 35
org/10.1128/IAI.01840-14 (1):77–98. https://doi.org/10.14573/altex.
17. Yue F, Cheng Y, Breschi A et al (2014) A 1611291
comparative encyclopedia of DNA elements in 22. Meurens F, Summerfield A, Nauwynck H et al
the mouse genome. Nature (London) 515 (2012) The pig: a model for human infectious
(7527):355–364. https://doi.org/10.1038/ diseases. Trends Microbiol 20(1):50–57.
nature13992 https://doi.org/10.1016/j.tim.2011.11.002
18. Steiniger BS (2015) Human spleen microanat- 23. Ramos-Vara JA, Miller MA (2013) When tissue
omy: why mice do not suffice. Immunology antigens and antibodies get along: revisiting
145(3):334–346. https://doi.org/10.1111/ the technical aspects of immunohistochemis-
imm.12469 try--the red, brown, and blue technique. Vet
19. Mestas J, Hughes CCW (2004) Of mice and Pathol 51(1):42–87. https://doi.org/10.
not men: differences between mouse and 1177/0300985813505879
human immunology. J Immunol 172 24. Butler AJ, Rees MA, Wight DGD et al (2002)
(5):2731–2738. https://doi.org/10.4049/ Successful extracorporeal porcine liver perfu-
jimmunol.172.5.2731 sion for 72 h. Transplantation 73
20. Chung WY, Gravante G, Al-Leswas D et al (8):1212–1218
(2013) The development of a multiorgan
 Chapter 23

Multicolor Flow Cytometry and High-Dimensional Data

Analysis to Probe Complex Questions in Vaccinology
Megan E. Cole, Yanping Guo, Hannah M. Cheeseman, and
Katrina M. Pollock

Abstract
Vaccines induce a highly complex immune reaction in secondary lymphoid organs to generate immunolog-
ical memory against an antigen or antigens of interest. Measurement of post immunization immune
responses generated by specialized lymphocyte subsets requires time-dependent sampling, usually of the
blood. Several T and B cell subsets are involved in the reaction, including CD4 and CD8 T cells, T follicular
helper cells (Tfh), and germinal center B cells alongside their circulating (c) counterparts; cTfh and
antibody secreting cells. Multicolor flow cytometry of peripheral blood mononuclear cells (PBMC) coupled
with high-dimensional analysis offers an opportunity to study these cells in detail. Here we demonstrate a
method by which such data can be generated and analysed using software that renders multidimensional
data on a two dimensional map to identify rare vaccine-induced T and B cell subsets.

Key words Flow cytometry, High-dimensional data analysis, t-SNE, FlowSOM, T cells, Vaccine,
immunophenotyping, FlowJoTM

1 Introduction

Novel techniques to investigate the cellular responses to immuni-

zation in humans, potentially leading to more profound mechanis-
tic or kinetic insights into the immune responses, are of
considerable interest, particularly when these techniques are cou-
pled with the use of innovative vaccine technologies, such as virally
vectored and RNA designs [1–3]. The response to vaccination is
orchestrated by T and B cell cross talk. Among peripheral blood
mononuclear cells (PBMC), activation and induction of several
subsets, including memory T and B cells, plasmablasts (antibody
secreting cells), and circulating Tfh (cTfh), have been observed
after immunization [4–6].
The immune response to vaccination is studied by sampling
specific timepoints in the days before and after immunization.

Fadil Bidmos et al. (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1_23,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

433
434 Megan E. Cole et al.

Design of the immunogenicity experiments is dependent on the

research question(s) and is linked to the vaccine being studied, with
immunogen, formulation, and prime- boost schedule all factors to
be considered when selecting at which timepoints to take samples.
Phenotypic distinction of rare T and B cell subsets responding
to immunization within PBMC can be achieved using multicolor
flow cytometry. Detection of these cells is through staining with
multiple fluorochrome-conjugated antibodies against the cellular
proteins of interest and with a dead cell discriminator dye. Standard
flow cytometry cell preparation techniques are used. The user is
recommended to conduct this with limited exposure to light. The
maximum number of fluorochromes that can be included in panel
design will be dependent on the flow cytometer available. Often-
times, multiple markers are needed to distinguish a particular sub-
set, for example, cTfh are CD3+CD4+CXCR5+ and variably express
programmed cell death protein-1 (PD-1) and Inducible T-cell
COStimulator (ICOS) in the peripheral circulation (Table 1).
Data generated from these analyses are multidimensional. To
explore this data set, conventional two-dimensional gating can fall
short for two reasons, one is the sequential loss of data and two is
the rigidity of the a priori hypotheses driving the gating strategy.
Rare subsets may be overlooked. An alternative is to use unsuper-
vised machine learning to organize complex multidimensional data.
Clusters of data are constructed through similarity of expressed
proteins and these may correspond to cell subsets of interest. One
such method is t-distributed stochastic neighbor embedding
(t-SNE), a machine learning algorithm that can be used to visualize
high dimensional flow cytometry data in two dimensions on a
scatter plot [7]. This is based on stochastic neighbor embedding
with a Student-t distribution to estimate the similarity between two
data points. Rare subsets with previously unforeseen relationships
can be highlighted. Cross-validation of results on the principal
subsets of interest using alternative methods of data analysis includ-
ing two-dimensional gating strategies, is advised. When combined,
these approaches are powerful and innovative methods to explore
the cellular immune response to immunization. Here we describe
these methods, which were used to determine induction of circu-
lating lymphocyte subsets in a study of influenza vaccine
responses [8].

2 Materials

2.1 Equipment 1. Refrigerator 2–8
C.

2. Freezer 20
C.
3. Biological class II safety cabinet.
4. Timer.
 Multicolor Flow Cytometry and High-Dimensional Data Analysis to Probe. . . 435

Table 1
The fluorochrome for each marker has been selected based on the configuration of a Becton
Dickinson (BD) LSRFortessa flow cytometer instrument, following the general principles of multicolor
panel design BV is Brilliant Violet™. BUV dyes are BD Horizon Brilliant™ Ultraviolet. Fluorochrome-
conjugated antibodies were obtained from BD Biosciences and BioLegend

Laser Filter Mirror Fluorochrome CD No. Alternative

355 nM 379/28 BUV 395 CD8
530/30 505LP BUV496 CD4
405 nM 450/50 BV421 CD27
525/50 475LP BV510 CD19
605/12 595LP BV605 CD279 PD-1
655/8 635LP BV650 CD278 ICOS
710/50 700LP BV711 CD127 IL-7R alpha
780/60 735LP BV785 CD183 CXCR3
488 nM 530/30 505LP FITC CD45RA
710/50 685LP PerCP-Cy5.5 CD38
561 nM 582/15 570LP PE CD32
620/10 600LP PE/Dazzle™ 594 CD185 CXCR5
780/60 750LP PE/Cy7 CD197 CCR7
633 nM 670/14 APC CD28
720/40 710LP AF700 CD3
780/60 750LP Near-IR NA Dead cell stain

5. Bench top centrifuge, up to 24,000  g.

6. Bench top microcentrifuge, up to 21,000  g.
7. Bench top vortex mixer, speed approximately 2800 rpm.
8. Heat block capable of heating to 70
C.
9. One channel air-displacement pipettes and tips in a range of
sizes.
(a) 0.2–2 ml.
(b) 2–20 ml.
(c) 20–200 ml.
(d) 100–1000 ml.
10. Calibrated serological pipettes.
(a) 5 ml.
(b) 10 ml.
(c) 25 ml.
(d) 50 ml.
436 Megan E. Cole et al.

11. 96-well U bottomed plates.

12. Pipette controller.
13. 15 ml centrifuge tubes.
14. 50 ml centrifuge tubes.
15. Automated cell counter or light-field microscope with cell
counting slides/hemocytometer and trypan blue dye.
16. Range of sterile lidded plastic tubes of different sizes from 1 to
5 ml.
17. Foil.
18. Foil-lined sealable plastic box.
19. FACS tubes.
20. Flow cytometer with multiparameter capability.

2.2 Software 1. BD FACSDiva™ and FlowJo (FlowJo™ Software, for Win-

dows or for Mac, Version 10.5 or above).
2. Microsoft Excel or similar.

2.3 Reagents 1. Fetal Bovine Serum (heat inactivated) (HI-FBS).

2. Sterile phosphate-buffered saline (PBS); MgCl2/CaCl2 free.
3. Fluorochrome-conjugated antibodies (see Tables 1 and 2).
4. Fc block buffer (e.g., Human TruStain FcX™, BioLegend®).
5. Cell viability dye (e.g. LIVE/DEAD™ Fixable Dead Cell Stain
Kits, ThermoFisher Scientific).
6. Compensation beads (e.g BD™ CompBeads, BD Biosciences).
7. Sodium azide (optional).
8. Commercial fixation buffer (e.g. BD Cytofix™ Fixation
Buffer, BD Biosciences), or 4% paraformaldehyde.

3 Methods

Principles of Good Clinical Practice (GCP)

For investigations that involve samples from clinical trials, per-
sonnel must be appropriately qualified and have received the correct
training.
Safety considerations
Local safety guidance for working with human samples should
be followed. Consideration should be given to the possible pres-
ence of infectious material, particularly blood borne viruses.
Recommendations for the use of personal protective equipment,
for the safe handling of human samples, and for the safe disposal of
waste material should be in place prior to embarking on sample
handling. Use of liquid nitrogen should only be by those trained in
 Multicolor Flow Cytometry and High-Dimensional Data Analysis to Probe. . . 437

Table 2
Examples of antibody clones that could be used in a multicolor flow cytometry panel. Users should
check with suppliers for the fluorescent antibody clone-conjugates available when designing the
panel

Manufacturer Target Clone

BD Biosciences CD4 SK3
BD Biosciences CD8 RPA-T8
BD Biosciences CD278 (ICOS) DX29
BioLegend CD19 HIB19
BioLegend CD279 (PD-1) EH12.2H7
BioLegend CCR7 G043H7
BioLegend CD38 HB-7
BioLegend CD28 CD28.2
BioLegend CD32 FUN-2
BioLegend CD45RA HI100
BioLegend CD27 M-T271
BioLegend CD127 A019D5
BioLegend CXCR3 G025H7
BioLegend CD3 OKT3
BioLegend CXCR5 J252D4

its handling. In this protocol, live cells are fixed with 4% parafor-
maldehyde before acquiring on the flow cytometer.
Sterile working
Standard laboratory aseptic technique should be used through-
out to avoid contamination of live samples. Surfaces, working area,
and equipment are cleaned with 70% ethanol.
Before embarking on the wet-lab procedures, draw up a
96-well plate plan, including control wells and test wells. Use this
as a reference throughout the experiment.

3.1 Preparation Prepare the following buffers (see Note 1):

of Reagents
1. FACS wash buffer (10% HI-FBS in PBS): 50 ml FBS plus
450 ml PBS kept at 2–8
C or on ice during use.
2. FACS stain buffer (5% HI-FBS in PBS): 25 ml FBS plus 450 ml
PBS kept at 2–8
C or on ice during use.
3. Viability dye: According to the manufacturer’s instructions,
prepare one aliquot of the fluorescent reactive dye at room
temperature (RT). Dissolve 1 μl of live–dead fixable dye in
1 ml PBS (1:1000 working solution).
438 Megan E. Cole et al.

4. Fc block buffer: Commercial FC block preparations are avail-

able (e.g., 5 μl Human TruStain FcX™ [BioLegend] is used per
100 μl of FACS stain buffer). Alternatively, to make Fc block
buffer add 200 μl human serum to 9800 μl of FACS stain
buffer.
5. Fixation buffer: Commercial preparations are available. The
working solution should contain 3–4% paraformaldehyde
(PFA).
6. Fluorescence-conjugated antibody cocktails: These must be
prepared fresh on the day of use (see Tables 1 and 2 for
examples).
(a) Prepare the fluorescence-conjugated antibody full stain
cocktail at 2 the optimized staining dilution in FACS
stain buffer. Allow a sufficient overage for pipetting: 25 μl
is required per well, with a final volume for staining of
50 μl. Store at 2–8
C in the dark until needed (see Notes
2 and 3).
(b) Prepare the fluorescence-conjugated antibody fluores-
cence minus one (FMO) control cocktails at 2 the opti-
mized staining dilution in FACS stain buffer. Allow a
sufficient overage for pipetting: 25 μl is required per
well, with a final volume for staining of 50 μl. Store at
2–8
C in the dark until needed (see Note 4).

3.2 Preparation This method can be used on PBMC prepared from fresh whole
of PBMC blood or raised from viable cryopreservation.
for Multidimensional
1. Suspend the washed PBMC in FACS wash buffer and count the
Flow Cytometry cells by trypan blue exclusion or using a cell counter and record
Analysis viability. Adjust the concentration to 1  106 cells/100 μl.
2. Prepare the live–dead cell control sample. Program the heat
block to 70
C. Take 0.5  106 PBMC in FACS wash buffer
and kill the cells by heating in a sealed plastic tube at 70
C for
10 min. Combine the killed 0.5  106 PBMCs in FACS wash
buffer with the same quantity of live cells and make up the vol-
ume to 200 μl. Transfer to a 96-well U bottomed plate.
3. Transfer 2  106 cells per well of live PBMC to the labeled
96-well U bottomed plate.
4. Centrifuge at 800  g for 3 min and remove the supernatant.
5. Resuspend the cells in all wells in 100 μl diluted viability dye
buffer, except for the unstained control well, where 100 μl PBS
is used.
6. Incubate for 5 min at RT.
7. Stop the staining by adding 100 μl of FACS wash buffer.
8. Centrifuge at 800  g for 3 min and remove the supernatant.
 Multicolor Flow Cytometry and High-Dimensional Data Analysis to Probe. . . 439

9. Resuspend cells in all wells in 25 μl of FC block buffer.

10. Incubate for 7 min at RT.
11. Add 25 μl of 2 concentrated antibody staining mix to full
stained cells or 25 μl of 2 concentrated FMO antibody cock-
tail to FMO cells. Add 25 μl FACS buffer to unstained cells and
live–dead cells control wells.
12. Incubate for 20 min in the dark at 2–8
C.
13. Wash twice in 150 μl FACS buffer per well (final volume), and
centrifuge at 800  g 4
C for 3 min and discard the
supernatants.
14. Fix the cells by resuspending in 100 μl fixation buffer for a
minimum of 20 min. Store the fixed cells at 2–8
C in the dark
for up to 24 h before acquiring the data on a flow cytometer.
(Cells can either be washed before acquiring on the flow cyt-
ometer or run immediately in the fixation buffer).
15. When ready to acquire data on the flow cytometer, stain the
compensation beads in the following steps.
16. Add one drop of each compensation bead set (negative and
positive) into labelled FACS tubes, one for each antibody.
17. Add 1 μl of each antibody into the respective FACS tubes.
18. Incubate for 10 min at RT in the dark.

3.3 Acquiring Events 1. Open the instrument software (e.g., BD FACSDiva™ Software
on the Flow Cytometer BD Biosciences) and set up the experiment layout on the
instrument.
2. Run the compensation controls and adjust the voltages for each
parameter for clear negative and positive signal separation with
minimal compensation values (see Note 5).
3. Acquire and record at least one million events in the live cell
lymphocyte gate for analysis of rare populations.

3.4 Data Analysis In this section, we describe how to use the data analysis software,
FlowJoTM (BD Life Sciences) for basic data analysis and plug-ins to
visualize high-dimensional data on a two dimensional (2D) map, to
enable the identification of cell clusters that share similar expression
patterns but would be otherwise difficult to extract for analysis
using conventional analysis methods (see Note 6).

3.4.1 Preliminary 1. Check data quality with FlowJo™ by choosing “Inspect” from
Analysis the populations band of the workspace task in the workspace or
by double-clicking on the circle badge to the left of a sample.
2. Cross-check the compensation matrix or, if necessary, recalcu-
late a new compensation matrix with the compensation
controls.
440 Megan E. Cole et al.

3. Build a suitable gating strategy to identify the key populations

in the sample based on the expression of known markers (CD4+
T cells in this study) using FMO controls to set the gates.
4. Export CD4+ only T cells as new FSC files for further analysis.

3.4.2 Data Down These steps define the number of events for analysis and combine
Sampling them into one file.
and Concatenation
1. Import all exported CD4+ T cell FSC files into one new
workspace.
2. Use the down sample plugin to define the number of events for
analysis, this step will ensure the analysis is not biased due to an
uneven number of events from a few samples or time points.
3. Concatenate the down-sampled populations into one new FCS
file for further analysis, and gate the individual samples based
on their sample ID with the concatenated file.

3.4.3 Running the t-SNE 1. Run the t-SNE plugin on FlowJo™ with the concatenated
data sample to generate x- and y- axis parameters, which project
similar populations from the high-dimensional space into a 2D
plot. Select markers that are important and meaningful to the
research question, to define the x- and y- axis. The running
time for the algorithm depends on the computer RAM capac-
ity, the number of events in the sample, and the number of
markers selected for analysis. The plugin for t-SNE is fully
integrated, and no R or extra package is needed.
2. The 2D plot created shows all events from the concatenated
samples, or one can select the individual sample ID gate to
show the individual sample on the plot. Choose the heatmap
mode to show the expression level of each parameter.
3. To project the conventional populations onto the t-SNE plot,
select the population with t-SNE x-axis and y-axis and copy the
plot onto the experiment layout, and overlay other populations
for comparison.
4. An example of simple data analysis using this method is shown
below (Figs. 1, 2, and 3).

3.4.4 FlowSOM [9] 1. FlowSOM is an algorithm to visualize the cell subsets with Self-
Organizing Maps (SOMs) and Minimal Spanning Trees, which
can reveal how all markers are behaving on all cells and can
detect subsets that might otherwise be missed. The FlowSOM
Plugin needs R packages to run under the command of Flow-
JoTM software (see Note 7 and 8).
2. Run FlowSOM by clicking on the FlowSOM tab and select the
parameters for clusters generation. The outcome of the algo-
rithm run will include distinct populations that are used as
 Multicolor Flow Cytometry and High-Dimensional Data Analysis to Probe. . . 441

250K 250K 250K

Lymphocytes Single Single
200K 200K cells 200K cells

SSC-W
FSC-W
SSC-A

150K 150K 150K

100K 100K 100K

50K 50K 50K

0 0 0
0 50K 100K 150K 200K 250K 0 50K 100K 150K 200K 250K 0 50K 100K 150K 200K 250K
FSC-A FSC-A SSC-A

105 Live 105 105 CD4+CD8-cells

CD3+CD19-cells
cells
LIVE-DEAD

104 104 104

CD19

CD4
103 103 103

0 0 0

–103 –103 –103

0 50K 100K 150K 200K 250K 0 103 104 –103 0 103 104 105
FSC-A CD3 CD8

Fig. 1 Preliminary gating strategy for targeting populations. For each sample, check the sample quality and
select the populations of interest using known markers. The gating strategy shown here is an example to
enrich CD4+ T cells for further analysis

gated population, Population Heatmap and clusters heatmap,

clusters organized onto Minimal Spanning Trees and CSV files.
3. An example of simple data analysis using this method is shown
below (Figs. 4, 5, and 6).

4 Notes

1. These buffers are optimal for cell viability and must be prepared
fresh on the day. For storage up to 1 month at 2–8
C, 0.25 g
sodium azide (antibacterial agent) can be added.
2. Panel design: Successful flow cytometry experiments require
robust panel design. This process is time consuming but,
done correctly, reaps reward. The first stage is to decide how
many target antigens will be in the panel. This will be depen-
dent on the flow cytometer available so always check this first
before designing the panel. The exact configuration and capa-
bility of the flow cytometer must be understood. Determine
the possible fluorochromes in the panel using a matrix and
include the excitation and emission spectra of these fluoro-
chromes. A spectra analyzer or spectrum viewer can help with
viewing the excitation and emission parameters. Spectral over-
lap can result in data that are difficult to interpret and should be
442 Megan E. Cole et al.

Sample A
Heatmap key
120 120 120

tSNE y
CCR7
tSNE y

tSNE y
90 90 90
expression
range 60 60 60

30 30 30

0 0 0
0 30 60 90 120 0 30 60 90 120 0 30 60 90 120
tSNE x tSNE x tSNE x

Time point 1 Time point 2 Time point 3

Sample B

120 20 120

TSNE y
90 90
tSNE y

tSNE y
90

60 60 60

30 30 30

0 0 0
0 30 60 90 120 0 30 60 90 120 0 30 60 90 120
tSNE x tSNE x tSNE x
Time point 1 Time point 2 Time point 3

Fig. 2 t-SNE plots of individual samples to show the difference in cluster abundancy. The selected plots show
a heatmap of CCR7 expression on the concatenated CD4+ T cell FCS file. A cluster with relatively low
expression of CCR7 (red arrow) is more abundant in sample A, compared with sample B at three different time
points

avoided where possible. Some fluorochromes fluoresce more

brightly than others. One approach to managing this is to put
the lowest expressed target antigen on the brightest fluoro-
chrome and the highest expressed on the weakest fluoro-
chrome. Include a viability dye on a target fluorochrome that
is not useful for other targets.
Spectrum viewer links are included below:
(a) https://www.bdbiosciences.com/en-us/applications/
research-applications/multicolor-flow-cytometry/prod
uct-selection-tools/spectrum-viewer
(b) https://www.biolegend.com/en-us/spectra-analyzer
3. Antibody titration: To determine the optimal concentration for
each fluorochrome-conjugated antibody selected for use in the
panel, it is critical that each antibody is titrated. This should be
performed on the cell/sample type to be used in the experi-
ment. For example, if PBMC are used for the investigation,
then the antibody needs to be titrated on PBMC. Stain as per
the staining protocol using a serial dilution. To determine the
best dilution, there are two calculations that can be used. First
the titration data needs to be analysed by gating on the
 Multicolor Flow Cytometry and High-Dimensional Data Analysis to Probe. . . 443

Key ID
105
Name
CD45RA-CCR7- 104

CCR7
CD27
CD45RA-CCR7+
103
CD45RA+CCR7-
0
CD45RA+CCR7+ PD-1
-103

-103 0 103 104 105

CD45RA
CD127

120
CXCR3
tSNE y

90

60
CD38
30

0
0 30 60 90 120
tSNE x CD28

Fig. 3 Overlay of known cell subsets onto the t-SNE plot to reveal the complexity of the populations. In this
example, the CD45RACCR7+ population (orange) and the CD45RACCR7 population (green) distribute into
different cluster areas on the t-SNE plot. The abundancy of each cell subset within individual samples is
shown on the top of the histogram overlay

positively and negatively stained population. Calculate the

mean fluorescence intensity (MFI) for each population and
the standard deviation of the negative population. Export all
the data into an Excel file and either calculate the stain to noise
ratio or the staining index [10].
Stain to noise ratio ¼ MFIpositive population =MFInegative population

MFIpositive population  MFInegative population

Staining index ¼
2  standard deviation of MFInegative population
Once the optimal concentrations are determined, these can
be used to build the antibody cocktail.
4. Controls in the experiment: Several controls are recommended.
(a) Dead cell control: This includes 50% heat killed PBMC
and is single stained with the cell viability marker for
setting compensation in this channel when acquiring on
the flow cytometer.
(b) Fluorescence minus one (FMO) controls: Optimization
of the flow cytometry panel should include an experiment
where these are run for every fluorochrome. These allow
refinement of the gating strategy, particularly where a
444 Megan E. Cole et al.

Pop12 700
Pop1
600
Pop0
Pop6 500
Pop3
Pop7 400
Pop4
Pop9 300

Pop8
200
Pop10
Pop13
Pop5
Pop2
Pop11
CD38

CXCR3

PD-1

ICOS

CCR7

CD45RA

CD32

CXCR5

CD27

CD127

CD28
B
ID
120
CD27
90
tSNE y

PD-1
60
ICOS
30

0
CD127
0 30 60 90 120
tSNE x CXCR3

Name CD45RA
Pop 10
CD38
Pop 9
Pop 8 CD32
Pop 4
Pop 3 CXCR5
Pop 2
Pop 1 CCR7
Pop 0
CD28

Fig. 4 Heatmap of metapopulations; diverse populations identified with the FlowSOM algorithm (A) and how
those populations present on the t-SNE plot (B)

protein is expressed continuously as a spectrum across a

particular cell subset of interest. Once the panel has been
optimized, only selected FMOs may be required, depend-
ing on the data elicited for analysis.
 Multicolor Flow Cytometry and High-Dimensional Data Analysis to Probe. . . 445

Fig. 5 The Minimum Spanning Tree generated by FlowSOM analysis. Each node is represented as a star chart,
where the color-coded height of each segment shows the expression level of that marker. Every chart has a
background colour indicating its meta-population assignment. On the tree, the distance of nodes clustersed
together with each other within a branch shows their similarity to each other. The abundance of each node is
represented by the size of its symbol

(c) Unstained controls: These are useful to gauge back-

ground fluorescence of the cells as treated during the
preparation procedures.
(d) Compensation controls: These are needed to calibrate the
flow cytometer during set-up. To preserve experimental
samples and for ease of use, commercial compensation
beads can be used. These should react with antibodies of
the same species as the fluorochrome-conjugated
antibody.
5. Instrument calibration: Instruments from different providers
will have their own process for standardization and calibration.
This process ensures that the performance of an instrument is
consistent in alignments, signal accuracy and resolutions over a
certain time-period. When the instruments are standardized,
the data acquired on different dates are considered as standar-
dized and can be objectively and quantitatively compared. The
user may wish to run a few events from the unstained sample
and the full stained sample briefly to check voltages initially
446 Megan E. Cole et al.

105 105 105

CCR7

104 104 104

CD27

CD127
103 103 103
0 0 0
–103 –103 –103
–103 0 103 104 105 –103 0 103 104 105
–103 0 103 104 105
CD45RA CD38
CD28

Name
105 105 Pop10
Pop9
104 104
CXCR3
ICOS

Pop8
103 103
Pop4
0 0 Pop3
–103 Pop2
–103
–103 0 103 104 105 Pop1
–103 0 103 104 105
PD-1 CXCR5 Pop0

Fig. 6 Identification of FlowSOM populations with selected markers. Like the use of the heatmap, overlay of
FlowSOM populations with selected markers can help to reveal the nature of the populations and possible
gating strategy for future experiments. For example, population2 (orange) is high in CD27, CD38, PD-1, and
ICOS expression

before running compensation controls. This will help the user

to know in advance that the staining and signal detection are
acceptable for further steps.
6. Computer RAM requirement: To run the plugins smoothly, the
minimum requirement for the computer is 8G RAM, it is
highly recommended to use a computer with more than 16G
RAM to speed up the calculation time. The time needed for the
calculation is based on the size of the data; a high number of
parameters will increase the calculating time.
7. FlowSOM plugin: FlowSOM Plugin can be download from
https://www.flowjo.com/exchange/#/. R and R packages
should be installed to run FlowSOM. Please read the ReadMe
document for the FlowSOM plugin to ensure the correct
installation of R packages and R tools. Further information
about FlowSOM can be found in Van Gassen, S. et al Cyto-
metry A (2015) [9].
8. YouTube videos—self-tutorials: FlowJo have many tutorials
online on their website with helpful directions and instructions
on how to do T-SNE, FlowSOM and many other things, as
well as a YouTube channel with videos under “FlowJo Media.”
 Multicolor Flow Cytometry and High-Dimensional Data Analysis to Probe. . . 447

Acknowledgements

The FluAGE study was conducted with a grant from the National
Institute for Health Research (NIHR) Imperial Biomedical
Research Centre (BRC). The views expressed in this publication
are those of the author(s) and not necessarily those of the NHS, the
National Institute for Health Research or the Department of
Health. The authors would like to thank the participants in the
FluAGE study and the staff of the Imperial College Clinical Trials
Centre, and the Department of Infectious Disease, Imperial Col-
lege London." If a conflicts of interest statement is required then
please write "K.P. reports grants from the National Institute for
Health Research and the Medical Research Council UK Research
and Innovation. K.P. is chief investigator for the Imperial College
London COVID-19 vaccine development programme and princi-
pal investigator for the University of Oxford COVID-19 vaccine
development trials. K.P. reports personal fees from Sanofi, and
honoraria received from the British Society for Antimicrobial Ther-
apy and ITV plc.

References

1. McKay PF, Hu K, Blakney AK et al (2020) Self- 6. Bentebibel SE, Lopez S, Obermoser G et al

amplifying RNA SARS-CoV-2 lipid nanoparti- (2013) Induction of ICOS+CXCR3+CXCR5+
cle vaccine candidate induces high neutralizing T H cells correlates with antibody responses to
antibody titers in mice. Nat Commun influenza vaccination. Sci Transl Med
11:3523. https://doi.org/10.1038/s41467- 5:176ra32. https://doi.org/10.1126/
020-17409-9 scitranslmed.3005191
2. Walsh EE, Frenck RW, Falsey AR et al (2020) 7. Van Der Maaten L, Hinton G (2008) Visualiz-
Safety and immunogenicity of two RNA-based ing data using t-SNE. J Mach Learn Res
Covid-19 vaccine candidates. N Engl J Med 9:2579–2625
383:2439–2450. https://doi.org/10.1056/ 8. Cole ME, Saeed Z, Shaw AT et al (2019)
nejmoa2027906 Responses to Quadrivalent influenza vaccine
3. Folegatti PM, Ewer KJ, Aley PK et al (2020) reveal distinct circulating CD4+CXCR5+ T
Safety and immunogenicity of the ChAdOx1 cell subsets in men living with HIV. Sci Rep
nCoV-19 vaccine against SARS-CoV-2: a pre- 9:1–16. https://doi.org/10.1038/s41598-
liminary report of a phase 1/2, single-blind, 019-51961-9
randomised controlled trial. Lancet 9. Van Gassen S, Callebaut B, Van Helden MJ
396:467–478. https://doi.org/10.1016/ et al (2015) FlowSOM: using self-organizing
S0140-6736(20)31604-4 maps for visualization and interpretation of
4. Ellebedy AH, Jackson KJL, Kissick HT et al cytometry data. Cytometry A 87:636–645.
(2016) Defining antigen-specific plasmablast https://doi.org/10.1002/cyto.a.22625
and memory B cell subsets in human blood 10. Swieboda D, Guo Y, Sagawe S et al (2019)
after viral infection or vaccination. Nat Immu- OMIP-062: a 14-color, 16-antibody panel for
nol 17:1226–1234. https://doi.org/10. Immunophenotyping human innate lymphoid,
1038/ni.3533 myeloid and T cells in small volumes of whole
5. Bentebibel SE, Khurana S, Schmitt N et al blood and pediatric airway samples. Cytometry
(2016) ICOS + PD-1 + CXCR3 + T follicular A 95:1231–1235. https://doi.org/10.1002/
helper cells contribute to the generation of cyto.a.23907
high-avidity antibodies following influenza
vaccination. Sci Rep 6:26494. https://doi.
org/10.1038/srep26494
 INDEX

A BLAST+ .....................................................................78, 92
Bordetella pertussis (B. pertussis) ......................... 4, 48, 49,
Adjuvants ........................................... 112, 141, 155, 163, 51–54, 56, 57, 59, 60, 325–330, 332–337, 343,
167, 172, 207, 209, 219, 302, 303, 309 354, 355
Affinity purification ............................................ 39, 41, 43
Bronchoalveolar ............................................................ 307
Alginate................................................................. 142–148 Brucella ..................................................... 3, 4, 12, 13, 18,
Amino acids ................................... 25, 27, 33, 60, 86, 90, 76, 83, 84, 94
138, 155–158, 178, 231, 232, 249–252, 275, 282
Antibodies ............................... 37–39, 41–45, 60, 63–72, C
75–95, 99, 107, 111, 117, 118, 123, 132, 133,
139, 140, 155, 162–166, 175, 181, 186, 194, C1q ............................................................. 343, 345, 346,
200, 201, 204, 212, 214, 218, 219, 221, 222, 349–354, 357–360
225, 230, 239, 240, 266–268, 274, 284, 285, Candidates .............................................. 1–3, 6, 8–11, 13,
288, 291, 292, 295, 296, 304, 305, 307, 309, 14, 17–19, 21, 25, 27, 31–33, 56, 118, 166
310, 312, 314–320, 325, 331, 332, 334, Capsule ................................................237, 282, 284, 285
336–338, 341–343, 351–353, 355, 359, 360, Carriage ................................................................ 387–389
363–371, 373–386, 388, 405, 407, 409, 417, Cell-mediated ................................................................ 303
418, 421, 425–427, 433–435, 437–439, 442, Cell wall extracts ..........................................38, 40, 42–45
443, 445 CHARMM ................................................. 153, 159, 160,
Antigenicity ................................... 6, 19, 25–27, 32, 101, 162, 165, 166
103, 105, 106, 109, 110, 118, 152, 158, 159, Chromatography ............................... 39, 40, 42, 47, 127,
161, 162, 165 128, 134, 135, 137, 177, 230, 232–234, 237,
Antigen presenting cells (APCs) ......................... 302, 435 238, 267, 282, 283, 286, 290, 293, 296, 298,
Antigens...................................................... 2–4, 6, 12, 13, 342, 344, 345, 347–352, 356–359
18–20, 26, 31–33, 37–45, 47–60, 63, 64, 75, 77, Circulating Tfh (cTfh) .................................................. 433
85, 86, 93, 94, 111, 115–142, 151–167, 171, Class I ..............................................18–20, 26, 27, 29–33
172, 175, 178, 181, 186, 207, 219, 224, 227, Class II ........................................... 19, 20, 26, 27, 29–31,
228, 239, 266, 267, 302–305, 309, 310, 33, 54, 197, 309, 345, 365, 390, 391, 393, 394,
312–318, 320, 331, 364, 389, 425, 441 396, 398, 434
Clostridium spp. ................................................. 76, 83, 94
B Clusters ................................................434, 439–443, 445
Clusters of orthologous genes (COGs)............ 22, 83, 92
Baby rabbit complement (BRC) ........................ 374, 375, Colorimetric ...................................... 39, 40, 42, 75, 247,
378–380, 383–385
249, 255, 272, 273
Bacterial extracellular vesicles Complement.......................................... vi, 341–360, 364,
(BEVs).....................................172–175, 177, 178, 366–371, 373, 374
180–186, 188
ComplexHeatmap .....................................................78, 93
Bacterin..............................................................................vi Conjugation ............................................... 180, 281–299,
Bacteroides thetaiotaomicron 302, 303
(B. thetaiotaomicron)................................ 171–173 Conjugative ................................ 173, 174, 179–180, 197
B cell ........................................14, 18–20, 25, 26, 28, 31,
Controlled human infection model experiment
33, 155, 156, 166, 228, 389, 433 (CHIME).................................................. 388, 394
Bicinchoninic acid (BCA) protein Co-perfusion ................................................413–416, 423
assay............................. 39, 64, 66, 224, 229, 231,
Coupling........................................................... 39–44, 283
246–249, 294 Crossflow ultrafiltration ................................................ 171
Bioinformatics ............................ 1, 18, 19, 51, 53–55, 77

Fadil Bidmos (eds.), Bacterial Vaccines: Methods and Protocols,

Methods in Molecular Biology, vol. 2414, https://doi.org/10.1007/978-1-0716-1900-1,
© The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2022

449
 BACTERIAL VACCINES: METHODS AND PROTOCOLS
450 Index
Cultures ................................................40, 49, 51, 52, 56, G
57, 107–111, 125, 126, 128, 131, 137, 162, 163,
174, 176–177, 179–181, 194–199, 202, 203, Generalized modules for membrane antigens
208, 210, 211, 215, 217–219, 223, 224, 281, (GMMA)................................................... 227–278
289, 295, 304, 305, 307–310, 312, 314–316, Genes ................................................. 3, 9, 11, 14, 18, 56,
318–320, 327–330, 332, 335, 374–376, 378, 64, 75–77, 83, 88, 92, 94, 122–125, 130, 131,
385, 391, 398, 399, 407, 410–411 134, 135, 138, 152, 154, 158, 162, 172–174,
Cytokines ................................... 212, 213, 217–220, 225, 178–179, 194, 196, 197, 202, 208, 209, 212,
303, 305, 307, 309–313, 315, 316, 318–320, 218, 225, 389
325–328, 330–332, 334, 335 Gene synthesis ............................................................... 178
Glycine .......................................39, 40, 43, 65, 175, 193,
D 208–210, 215, 224, 288, 355, 360
Glycoconjugates ......................................... 282–285, 294,
De novo .......................................................................... 178 295, 298, 387, 389
Differential analysis ......................................................... 92 Glycoprotein...................................................................... 8
Dische ......................................................... 229, 234–236, Glycosylation ..................... 282, 283, 292, 294, 296, 297
255, 257, 258 Gram-negative .......................................... 6, 9, 11, 17, 48,
Dynamic light scattering (DLS) ..................................143, 63–72, 87, 99, 112, 115–140, 172, 191, 227,
144, 146, 147, 204, 229, 230, 240, 241 335, 398, 399
Gram-positive ....................................................4, 6, 9, 11,
E
37–45, 48, 87, 95, 373
Electrons .......................................... vi, 97–112, 143, 146, GROMACS ................................................................... 159
147, 192, 204, 209, 225 Group A Streptococcus (GAS) ............................. 373–376,
Endotoxin...........................................137, 207–225, 276, 380, 384, 386
282, 283, 287, 292, 293, 295, 297
Enzyme-linked immunosorbent assays H
(ELISA)..............................................99–101, 105, Hestrin ........................................229, 238, 239, 272, 273
107–109, 111, 117, 118, 122, 123, 131, 132, High-performance liquid chromatography
139, 140, 155, 163, 212–214, 217–222, 304, (HPLC)..................................... 50, 229–232, 238,
305, 307, 312, 316, 317, 321, 326, 327, 242, 250, 251, 260, 262, 270, 274, 344, 345,
331–334, 337, 338 347, 355, 360
Enzyme-linked immunosorbent spot assay High-performance liquid chromatography mass
(ELISpot)................................................. 304, 305, spectrometry (HPLC-MS)...................... 234–236,
312–315, 319, 320 257, 259, 260, 262–265
Epitopes ................................. 2, 8–11, 14, 18–21, 25–33, HL-60 cells.......................................................... 375, 376,
45, 48, 55, 76, 115, 152–163, 165–167, 302, 378–380, 383–385
312, 320 HLA ..........................................19, 20, 26, 27, 29, 30, 33
Escherichia coli (E. coli) ....................................... 108, 125, Human.....................................2, 5, 9–12, 17, 19, 24, 25,
126, 130, 135–138, 154, 162, 167, 173, 174, 27, 32, 33, 38, 39, 41, 43–45, 54, 63, 75–83,
178–180, 194, 195, 197, 202, 208–210, 215, 85–87, 89, 90, 92–95, 99, 117, 151–153, 166,
223, 281–283, 289, 294 327, 334, 341–360, 364–366, 373–403, 406
Extracellular....................................... 12, 13, 94, 95, 116, Human challenge ......................................................75–95
171–189, 191, 192 Humoral .............................. 18, 303, 304, 316, 388, 405
Ex vivo ....................................... 309–313, 405, 406, 408, Hybrids ................................................................. 115–140
409, 413–416, 425
I
F
IgG.............................................41, 44, 90, 91, 123, 133,
Flow cytometry .......................................... 48–50, 52, 53, 155, 165, 194, 200, 214, 221, 222, 225, 230,
57–59, 303, 305, 309–313, 319, 385, 433–446 341–360
FlowJo® ........................................................ 58, 313, 436, IgM ......................................................221, 225, 341–360
439, 440, 446 Immunization................................................. vi, 117, 129,
FlowSOM ...................................................................... 440 134, 152, 155, 158, 162–165, 167, 171, 173,
Fluorescence whole-slide scanning ..................... 418–419 187, 189, 214, 220, 221, 302, 303, 307, 311,
Fusion tag ............................................................. 135, 178 325, 326, 433, 434
 BACTERIAL VACCINES: METHODS AND PROTOCOLS
Index 451
Immunodetection ....................................... 175, 178, 181 M
Immunogenicity ................................................18, 86, 91,
153–155, 162–166, 171, 172, 191, 374, 434 Machine learning (ML) .................................... 1–15, 420,
Immunohistochemistry 427, 429, 434
(IHC) ..............................406, 407, 417, 425, 426 Matrix-assisted laser desorption/ionization time-of-flight
Immunoinformatics ..................................................17, 19 mass spectrometry (MALDI-TOF MS) ........... 236
Immunophenotyping.................................................... 433 Mass spectrometry (MS)............................ 22, 47, 50, 53,
Immunoprecipitation...........................................vi, 37–45 58, 60, 63, 64, 67, 69, 70, 72, 229, 235, 242, 249,
Inactivation................................................... 98, 103, 106, 259, 260, 266, 276, 277
107, 109, 110, 366 Maternal........................................................... vi, 325–339
Infants .............................................................. vi, 325–339 Membranes .............................. 12–14, 18–20, 25, 27, 48,
Infection models ................................166, 387–403, 406, 49, 55–60, 65, 69, 70, 72, 86, 92, 95, 99, 100,
408, 411–413 115–140, 155, 164, 172–187, 191–193, 199,
Inhalational vaccination ................................................ 303 200, 207–225, 227, 239, 240, 274, 286–288,
Innate......................................................6, 325, 326, 334, 291, 292, 296, 304, 305, 307, 312, 314, 315,
335, 370, 405, 406 319, 320, 353, 360, 388, 389
Inoculation ........................................ 108, 136, 174, 219, Memory ...................................... 303–305, 309, 315, 316
225, 365, 377, 385, 388–391, 393–396, 402, Meso scale discovery (MSD) ............ 327, 330, 331, 334,
403, 411, 412, 424 336–338
Insoluble proteins ............................................... 117, 118, MHC ......................................................... 18, 31, 32, 309
128, 130, 198, 199 Microarrays ........................................... 75–79, 83, 85, 86,
Intestinal pig epithelial cell 89, 90, 92–95
(IPEC) ............................................. 64, 66, 67, 71 Microscopy ................................146, 147, 192, 204, 209,
Intravenous.......................................................... 309, 406, 407, 409, 413, 416, 418
411, 412, 424 Monoclonal ................................................ 175, 305, 310,
Intravenous immunoglobulin G 312, 318, 343, 360
(IVIG) ..................................................... 38, 43, 44 Multi-epitope fusion antigen (MEFA) .......... vi, 151–167
Ionizing ........................................................................... 98 Multivalent ........................................................... 151–167
Isoelectric focusing (IEF) ........................... 65, 67, 70, 71 Murine .......................................302, 303, 312, 318, 320,
Isolation ..............................................134, 148, 171–187, 406, 408, 411–413, 425
224, 253, 303, 305, 307–309
N
In vitro ....................................................18, 64, 155, 158,
165, 166, 209, 211–213, 217, 218, 220, Nanoparticles......................................141, 143, 146–148,
326–328, 343 178, 185, 204, 230, 242
In vitro transcription and translations (IVTT) .............. 75 Neisseria gonorrhoeae (N. gonorrhoeae) .............. 129, 363,
In vivo ................................................ 6, 18, 64, 118, 135, 364, 366, 367, 370
152, 165, 209, 214, 219, 221, 222, 405 Neisseria lactamica (N. lactamica) .................... 388–396,
398–403
L Neisseria meningitidis
Lavage ................................................................... 303, 307 (N. meningitidis).......................... 4, 18, 117, 123,
Lawsonia intracellularis 124, 281, 354, 355, 364, 387–389, 405
(L. intracellularis) .....................64, 66, 67, 70, 71 NERVE.............................................................................. 2
Leukocytes .................................................. 303–310, 312, Neutralizing.................................... 63–72, 152, 154, 162
314–317, 320 Neutrophils........................................................... 365, 373
Limulus amebocyte lysate (LAL) ....................... 137, 211, Nuclear magnetic resonance
216, 223, 259, 295 (NMR) ..............................................143–145, 238
Lipoproteins ...........................27, 88, 116, 228, 301–321
O
Liquid chromatography-tandem mass spectrometry
(LC-MS/MS) ............................ 47–50, 53, 59, 60 OAg extraction....................................237, 267, 269, 278
Liver ............................................................ 406, 413–416, Opsonophagocytic assay (OPA) ......................... 364, 365,
423, 425 369–371
Low-energy electron irradiation Opsonophagocytic killing assay
(LEEI).....................................................vi, 97–112 (OPKA)...........................373, 374, 380, 382, 383
Lymphocytes ........................................................ 434, 439 Orthologous.................................................................... 23
 BACTERIAL VACCINES: METHODS AND PROTOCOLS
452 Index
P PSORTb ...................................31, 51, 55, 78, 86, 92, 94
Pulmonary ..................................302, 303, 307, 309, 320
Pam2Cys ............................................................... 302, 311 Purifications..........................................31, 37, 41, 43, 44,
Pam3Cys ........................................................................ 302 118–121, 127–130, 135–138, 177, 178,
Pangenome ...................................................................... 76 184–185, 189, 212, 218, 224, 283, 286, 287,
Panproteomes.................................. 76–79, 82–90, 92–95 290, 295, 355
Pathogen-reactive antibody pool
(PRAP).................................................... 38, 41, 42 Q
Peptides ............................................1, 27, 28, 37, 48–50,
53–60, 88, 95, 118, 132, 135, 141–148, 171, Quantitative real-time PCR........................ 212, 217, 218
172, 178, 235, 259, 282, 302, 303, 305,
R
309–312, 314–316, 318
Peptidoglycans................................................48, 207, 208 Radiation ................................................................ 98, 102
Periplasmic............................................... 13, 95, 138, 172 Radioimmunoprecipitation assay (RIPA) ................64, 66
PglB .....................................................281, 283, 285, 294 Responses ................................................. 14, 18, 86, 141,
Phagocytosis ......................................................... 302, 383 152–155, 158, 163, 166, 301–321, 325–339,
Plasma .......................................... 38, 326, 328, 332–334, 342, 389, 405, 433, 434
339, 341–360, 370 Reverse vaccinology (RV) ................................. v, vi, 1–14,
Polyclonal ............................................................... 99, 111 17–19, 27, 32
Polyelectrolyte ................................................. vi, 141–148 Rodentibacter pneumotropicus
Polymerase chain reaction (PCR) ...............................124, (R. pneumotropicus) ...........................99, 101, 107
125, 193–197, 202, 212, 217–219, 389,
392–393, 399–401 S
Polymorphonuclear leukocyte...................................... 364
SARS-CoV-2 ........................................................... 3, 6, 8,
Polysaccharides .......................................43, 97, 142, 172,
301, 406
228, 253–255, 267, 269, 270, 272, 294, 388, 389
Scores ...........................................6, 8, 14, 20, 27, 28, 80,
Porcine.................. 5, 117, 406, 408, 409, 413–416, 430
86–89, 92, 94, 155, 156, 158, 159, 165, 419
Predictions ..................................................... 1–3, 6, 8–13,
Secretion ............................................... 57, 172, 174–175,
18–21, 25–29, 31–33, 48, 49, 57, 60, 86, 87, 89,
180–182, 221, 312
92, 94, 134, 156, 166
Self-adjuvanting.................................................... 142, 302
Pregnancies...................................................325–339, 402
Sequons ................................................................ 281–283
Profiling .....................................................................76, 77
Serum bactericidal activity
Propidium iodide (PI) ............................................ 49, 50,
(SBA)..............341–344, 354, 355, 364, 367–371
52, 53, 56, 57, 59
Shaving .................................................................vi, 48–60
Protegenicity ..................................................................... 6
Shewanella vesiculosa (S. vesiculosa) ..................... 191–204
Proteinase K assay ................................................ 178, 186
Shigella flexneri (S. flexneri) .....................................17–34
Protein G ............................................................. 342–344,
Shuttle vector ...............................................174, 178–180
348, 349, 351, 359, 360
SignalP ...............................................................51, 55, 78,
Protein glycan coupling technology
86, 87, 94, 130, 178
(PGCT) .............................................................. 281
Size exclusion chromatography
Proteins............................................. 1–3, 6, 8–14, 18–27,
(SEC) ............................................... 177, 184, 189
31–33, 37, 39–45, 47, 48, 51, 53–57, 60, 63, 64,
Size exclusion high-performance liquid chromatography
66–72, 75–83, 85, 86, 88–90, 92–95, 97, 98, 111,
with multiangle laser light scattering (HPLC-
115–140, 152–167, 171, 172, 174–175, 177,
SEC/MALS) ................................... 241, 242, 244
178, 180–182, 184–186, 191–204, 208–212,
Sodium dodecyl sulfate-polyacrylamide gel
215, 218, 223, 224, 227–229, 231, 239,
electrophoresis (SDS-PAGE)...................... 65, 67,
241–243, 245, 247, 249, 251, 257, 259, 266,
68, 70, 71, 127–130, 136, 155, 162–164, 192,
267, 275, 277, 281–285, 290–292, 294–298,
193, 199–201, 229–231, 242, 245, 246, 274,
302, 303, 305, 307, 309, 310, 312, 314–316,
284, 285, 287, 288, 290, 291, 293, 345, 349,
318, 320, 342–346, 350, 351, 353, 354, 358,
351, 353, 359
359, 366, 389, 434, 444
Soluble protein .............................................127–129, 199
Proteomics.................................................... vi, 18, 47–61,
Spectroscopy.................................................................. 144
75, 152, 202
Spleens ............................................... 309, 406, 408, 409,
Protoplasts .................................................................40, 43
413–416, 423, 425, 426
 BACTERIAL VACCINES: METHODS AND PROTOCOLS
Index 453
Staining ............................................ 68, 71, 72, 245, 246, Transports.............................................47, 100, 102, 115,
285, 288–293, 297, 303, 305, 310, 312, 314, 117, 135, 191, 192, 201, 309, 310, 391, 394, 396
315, 317–319, 409, 417, 420, 421, 425, 426, Treponema pallidum (T. pallidum)...........................5, 76,
428, 434, 438, 439, 442, 443, 446 83, 86, 90
Staphylococcus aureus (S. aureus) ...................4, 43, 44, 63 Trimethyl chitosan (TMC) .................................. 142–148
Streptococcus pneumoniae Triparental mating ........................................................ 179
(S. pneumoniae) ..........................4, 18, 43, 76, 77, Tuberculosis (TB) ............................................... 4, 11, 18,
83, 85, 86, 90, 91, 94, 282, 406, 410, 411, 424 152, 302–304
Streptococcus pyogenes (S. pyogenes)..............................4, 48 2-Dimensional gel electrophoresis (2-DE) .............63, 64
Structural vaccinology .................................................. 152
Subunit ..............................vi, 20, 64, 141–148, 171, 302 V
Support vector machine.................................................... 2
Vaccines.................................... v, vi, 1–3, 6, 8–10, 12–14,
Surfaceome ......................................................... 47, 56, 59 17–32, 34, 37–45, 47–60, 64, 75, 76, 78–83,
Swabs ...........................................76, 377, 385, 389, 391, 85–87, 89–95, 97–112, 116, 118, 141–148,
394, 396, 398, 399
151–167, 171–189, 191, 207, 209, 227, 228,
249, 281–298, 301–321, 326, 334, 363–371,
T
373, 374, 387, 388, 405, 433, 434
T7....................................................... 117, 118, 125, 126, Vaxign ...............................................2, 3, 6, 8–14, 18–22,
131, 135, 137, 138 29, 31, 32, 34
T cells ....................................... 14, 18–20, 25–27, 29–31, Vaxign-ML......................................................2, 3, 6, 8–13
33, 155, 166, 302–305, 309–313, 315, 316, Vaxitop .............................................................. 2, 8–11, 32
318–320, 325, 440, 441 Vesicles ........................................... vi, 171–189, 191–204,
t-distributed stochastic neighbor embedding 207–225, 227, 388
(t-SNE) ..........................................................93, 94 Viability.................................................48, 49, 52, 53, 56,
Testing .......................................101, 103, 106, 107, 109, 58, 107, 109, 111, 223, 312, 319, 320, 370, 376,
111, 224, 225, 302, 304, 310, 336–338, 345, 378, 379, 385, 402, 403, 411, 435, 437, 438,
346, 353, 354, 378, 403, 405 441–443
Titration............................. 142, 143, 155, 163, 165, 442 Virulence.............................................6, 54, 56, 152–155,
TMHMM ....................................... 51, 55, 78, 86, 92, 94 158–161, 163, 165, 166
TonB-dependent ..................................... 13, 14, 115, 116
Transconjugants .......................................... 174, 180, 202 W
Transferrin ...................64, 115–118, 122, 132, 138, 139 Western blotting (WB) .......... 63, 64, 69, 175, 181, 186,
Translational fusions ..................................................... 178 198–201, 204, 312, 351, 353
Transporter .....................................................14, 115, 116
Whole blood stimulation .............................................. 327