Methods in molecular biology vol 1603 heterologous protein production in CHO cells methods and protocols

Methods in
Molecular Biology 1603

Paula Meleady Editor

Heterologous
Protein
Production
in CHO Cells
Methods and Protocols


Methods

in

Molecular Biology

Series Editor
John M. Walker
School of Life and Medical Sciences
University of Hertfordshire
Hatfield, Hertfordshire, AL10 9AB, UK

For further volumes:
/>

Heterologous Protein
Production in CHO Cells
Methods and Protocols

Edited by

Paula Meleady
National Institute for Cellular Biotechnology, Dublin City University, Dublin, Ireland


Editor
Paula Meleady
National Institute for Cellular Biotechnology
Dublin City University
Dublin, Ireland

ISSN 1064-3745    ISSN 1940-6029 (electronic)
Methods in Molecular Biology
ISBN 978-1-4939-6971-5    ISBN 978-1-4939-6972-2 (eBook)
DOI 10.1007/978-1-4939-6972-2
Library of Congress Control Number: 2017935545
© Springer Science+Business Media LLC 2017
This work is subject to copyright. All rights are reserved 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,
express 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.
Cover Illustration: The front cover image, kindly provided by Alan Costello (National Institute for Cellular Biotechnology,

Dublin City University), shows Chinese hamster ovary (CHO) cells with inducible green fluorescent protein (GFP)
expression (from Chapter 6).
Printed on acid-free paper
This Humana Press imprint is published by Springer Nature
The registered company is Springer Science+Business Media LLC
The registered company address is: 233 Spring Street, New York, NY 10013, U.S.A.


Preface
Since their introduction into the market over 20 years ago, biotherapeutics have constituted a large and growing percentage of the total pharmaceutical market, as well as approximately 25% of the R&D pipeline in industry. These biotherapeutics are having a huge
global impact on the treatment of challenging and previously untreatable chronic disease.
Currently biopharmaceuticals generate global revenues of $163 billion, making up about
20% of the pharma market, and predicted to grow to over $320 billion by 2020. The number of approved products in Europe and the USA has steadily increased to 2016 in 2014,
of which 37 have “blockbuster” status, i.e., sales over $1 billion per year, with monoclonal
antibodies (Mabs) representing the most lucrative single product class [1]. Most significantly, nearly 50% of these biopharmaceutical products are produced in a single production
host, i.e., Chinese hamster ovary (CHO) cells. Improving the efficiency of production of
these biologics will be critical in controlling costs to healthcare systems as more of these
drugs come to market.
There has been considerable success in developing high-producing CHO cell culture
processes using approaches such as optimization of media formulation, improvements in
expression vector design, and also improvements in the design of bioreactors. The next
generation of improvements is expected to be made via genetic engineering of the host
(CHO) cell itself to increase or decrease the expression of endogenous genes depending on
the desired outcome, in order to improve the efficiency of the production of therapeutic
protein product. In order to enhance the production capabilities and efficiency of the host
cell line, an increased understanding of cellular physiology of CHO cells is of critical importance. There are substantial research efforts in progress focusing on the ‘omic analysis and
systems biology of CHO cells to understand CHO cell physiology. The publication of the
draft CHO-K1 genome in 2011 represented a major milestone in CHO systems biology.
This information has been supplemented further with the publication of draft genomes for
Chinese hamster and the CHO-S, CHO DG44 and CHO DXB11 cell lines. Availability of

the genome sequence will facilitate the interpretation and analysis of transcriptomic and
proteomic data to assess the physiological state of the cells under different growth and production systems. Combining all levels of regulation through systems biology models will
unveil the underlying complexity inherent in CHO cell biology and will ultimately enhance
and accelerate CHO productive capabilities in the coming decades.
This book includes reviews and protocols for genetic manipulation of CHO cells for
recombinant protein production, including “difficult-to-express” therapeutics. A method is
also included on the use of the recently described genome editing tool, CRISPR/Cas9, and
how this can be applied to CHO cells. The book also includes a review and protocols for
characterization of CHO cells using ‘omic approaches and how these methods can be used
to improve efficiency of recombinant protein production during cell line development.
Analytical methods for characterization of recombinant protein product, such as glycosylation and host cell protein analysis, are also described in this book.

v


vi

Preface

I am deeply grateful to all authors for giving up their valuable time and for contributing
to the book. I would also like to thank the series editor, Prof. John Walker, for help and
guidance during the process of getting the book to publication.
Dublin, Ireland

Paula Meleady

Reference
1. Walsh G (2014) Biopharmaceutical benchmarks 2014. Nat Biotechnol 32(10):992–1000

Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
v
Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
  1 Strategies and Considerations for Improving Expression of “Difficult
to Express” Proteins in CHO Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Christina S. Alves and Terrence M. Dobrowsky
  2 Glycoengineering of CHO Cells to Improve Product Quality . . . . . . . . . . . . . .
Qiong Wang, Bojiao Yin, Cheng-Yu Chung, and Michael J. Betenbaugh
  3 Large-Scale Transient Transfection of Chinese Hamster Ovary Cells
in Suspension . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Yashas Rajendra, Sowmya Balasubramanian, and David L. Hacker
  4 Cloning of Single-Chain Antibody Variants by Overlap-­Extension PCR
for Evaluation of Antibody Expression in Transient Gene Expression . . . . . . . .
Patrick Mayrhofer and Renate Kunert
  5 Anti-Apoptosis Engineering for Improved Protein Production
from CHO Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Eric Baek, Soo Min Noh, and Gyun Min Lee
  6 Conditional Knockdown of Endogenous MicroRNAs in CHO Cells
Using TET-ON-SanDI Sponge Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Alan Costello, Nga Lao, Martin Clynes, and Niall Barron
  7 Application of CRISPR/Cas9 Genome Editing to Improve
Recombinant Protein Production in CHO Cells . . . . . . . . . . . . . . . . . . . . . . . .
Lise Marie Grav, Karen Julie la Cour Karottki, Jae Seong Lee,
and Helene Faustrup Kildegaard
  8 Improved CHO Cell Line Stability and Recombinant Protein Expression
During Long-Term Culture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Zeynep Betts and Alan J. Dickson
  9 Selection of High-Producing Clones Using FACS for CHO
Cell Line Development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Clair Gallagher and Paul S. Kelly
10 The ‘Omics Revolution in CHO Biology: Roadmap to Improved
CHO Productivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Hussain Dahodwala and Susan T. Sharfstein
11 A Bioinformatics Pipeline for the Identification of CHO Cell Differential
Gene Expression from RNA-Seq Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Craig Monger, Krishna Motheramgari, John McSharry, Niall Barron,
and Colin Clarke

vii

1
25

45

57

71

87

101

119

143

153

169


viii

Contents

12 Filter-Aided Sample Preparation (FASP) for Improved Proteome
Analysis of Recombinant Chinese Hamster Ovary Cells . . . . . . . . . . . . . . . . . .
Orla Coleman, Michael Henry, Martin Clynes, and Paula Meleady
13 Phosphopeptide Enrichment and LC-MS/MS Analysis to Study the
Phosphoproteome of Recombinant Chinese Hamster Ovary Cells . . . . . . . . . .
Michael Henry, Orla Coleman, Prashant, Martin Clynes,
and Paula Meleady
14 Engineer Medium and Feed for Modulating N-Glycosylation
of Recombinant Protein Production in CHO Cell Culture . . . . . . . . . . . . . . . .
Yuzhou Fan, Helene Faustrup Kildegaard, and Mikael Rørdam Andersen
15 Glycosylation Analysis of Therapeutic Glycoproteins Produced
in CHO Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Sara Carillo, Stefan Mittermayr, Amy Farrell, Simone Albrecht,
and Jonathan Bones
16 Characterization of Host Cell Proteins (HCPs) in CHO Cell Bioprocesses . . . .
Catherine E.M. Hogwood, Lesley M. Chiverton, and C. Mark Smales

187

195

209

227

243

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 251


Contributors
Simone Albrecht  •  National Institute for Bioprocessing Research and Training (NIBRT),
Dublin, Ireland
Christina S. Alves  •  Biogen Inc., Cambridge, MA, USA
Mikael Rørdam Andersen  •  Department of Systems Biology, Technical University
of Denmark, Kgs. Lyngby, Denmark
Eric Baek  •  Department of Biological Sciences, KAIST, Daejeon, Republic of Korea
Sowmya Balasubramanian  •  Laboratory of Cellular Biotechnology (LBTC), École
Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland
Niall Barron  •  National Institute for Cellular Biotechnology, Dublin City University,
Dublin, Ireland
Michael J. Betenbaugh  •  Department of Chemical and Biomolecular Engineering,
Johns Hopkins University, Baltimore, MD, USA
Zeynep Betts  •  Faculty of Science and Literature, Department of Biology, Kocaeli
University, Izmit, Kocaeli, Turkey
Jonathan Bones  •  National Institute for Bioprocessing Research and Training (NIBRT),
Dublin, Ireland
Sara Carillo  •  National Institute for Bioprocessing Research and Training (NIBRT),
Dublin, Ireland
Lesley M. Chiverton  •  Industrial Biotechnology Centre and School of Biosciences,
University of Kent, Canterbury, Kent, UK
Cheng-Yu Chung  •  Department of Chemical and Biomolecular Engineering,
Johns Hopkins University, Baltimore, MD, USA

Colin Clarke  •  National Institute for Bioprocessing Research and Training (NIBRT),
Dublin, Ireland
Martin Clynes  •  National Institute for Cellular Biotechnology, Dublin City University,
Dublin, Ireland
Orla Coleman  •  National Institute for Cellular Biotechnology, Dublin City University,
Dublin, Ireland
Alan Costello  •  National Institute for Cellular Biotechnology, Dublin City University,
Dublin, Ireland
Hussain Dahodwala  •  Vaccine production program (VPP), VRC/NIAID/NIH,
Gaithersburg, MD, USA; SUNY Polytechnic Institute, Albany, NY, USA
Alan J. Dickson  •  Faculty of Life Sciences, The University of Manchester, Manchester, UK
Terrence M. Dobrowsky  •  Biogen Inc., Cambridge, MA, USA
Yuzhou Fan  •  Department of Systems Biology, Technical University of Denmark, Kgs.
Lyngby, Denmark; The Novo Nordisk Foundation Center for Biosustainability, Technical
University of Denmark, Hørsholm, Denmark
Amy Farrell  •  National Institute for Bioprocessing Research and Training (NIBRT),
Dublin, Ireland
Clair Gallagher  •  National Institute for Cellular Biotechnology, Dublin City University,
Dublin, Ireland

ix


x

Contributors

Lise Marie Grav  •  The Novo Nordisk Foundation Center for Biosustainability, Technical
University of Denmark, Hørsholm, Denmark
David L. Hacker  •  Laboratory of Cellular Biotechnology (LBTC), École Polytechnique

Fédérale de Lausanne (EPFL), Lausanne, Switzerland; Protein Expression Core Facility
(PECF), École Polytechnique Fédérale de Lausanne (EPFL), Lausanne, Switzerland
Michael Henry  •  National Institute for Cellular Biotechnology, Dublin City University,
Dublin, Ireland
Catherine E.M. Hogwood  •  Industrial Biotechnology Centre and School of Biosciences,
University of Kent, Canterbury, Kent, UK
Paul S. Kelly  •  National Institute for Cellular Biotechnology, Dublin City University,
Dublin, Ireland
Karen Julie la Cour Karottki  •  The Novo Nordisk Foundation Center for
Biosustainability, Technical University of Denmark, Lyngby, Denmark
Helene Faustrup Kildegaard  •  The Novo Nordisk Foundation Center for
Biosustainability, Technical University of Denmark, Lyngby, Denmark
Renate Kunert  •  Department of Biotechnology, Vienna Institute of BioTechnology,
University of Natural Resources and Life Sciences-Vienna, Vienna, Austria
Nga Lao  •  National Institute for Cellular Biotechnology, Dublin City University, Dublin,
Ireland
Gyun Min Lee  •  Department of Biological Sciences, KAIST, Daejeon, Republic of Korea
Jae Seong Lee  •  The Novo Nordisk Foundation Center for Biosustainability, Technical
University of Denmark, Lyngby, Denmark
Patrick Mayrhofer  •  Department of Biotechnology, Vienna Institute of BioTechnology,
University of Natural Resources and Life Sciences-Vienna, Vienna, Austria
John McSharry  •  National Institute for Bioprocessing Research and Training (NIBRT),
Dublin, Ireland
Paula Meleady  •  National Institute for Cellular Biotechnology, Dublin City University,
Dublin, Ireland
Stefan Mittermayr  •  National Institute for Bioprocessing Research and Training
(NIBRT), Dublin, Ireland
Craig Monger  •  National Institute for Bioprocessing Research and Training (NIBRT),
Dublin, Ireland; National Institute for Cellular Biotechnology, Dublin City University,
Dublin, Ireland

Krishna Motheramgari  •  National Institute for Bioprocessing Research and Training
(NIBRT), Dublin, Ireland; National Institute for Cellular Biotechnology, Dublin City
University, Dublin, Ireland
Soo Min Noh  •  Department of Biological Sciences, KAIST, Daejeon, Republic of Korea
Prashant  •  National Institute for Cellular Biotechnology, Dublin City University, Dublin,
Ireland
Yashas Rajendra  •  Laboratory of Cellular Biotechnology (LBTC), École Polytechnique Fédérale
de Lausanne (EPFL), Lausanne, Switzerland; Biotechnology Discovery Research, Lilly
Research Laboratories, Eli Lilly and Company, Lilly Corporate Center, Indianapolis, IN, USA
Susan T. Sharfstein  •  SUNY Polytechnic Institute, Albany, NY, USA
C. Mark Smales  •  Industrial Biotechnology Centre and School of Biosciences, University of
Kent, Canterbury, Kent, UK
Qiong Wang  •  Department of Chemical and Biomolecular Engineering, Johns Hopkins
University, Baltimore, MD, USA
Bojiao Yin  •  Department of Chemical and Biomolecular Engineering, Johns Hopkins
University, Baltimore, MD, USA


Chapter 1
Strategies and Considerations for Improving Expression
of “Difficult to Express” Proteins in CHO Cells
Christina S. Alves and Terrence M. Dobrowsky
Abstract
Despite substantial advances in the field of mammalian expression, there are still proteins that are characterized
as difficult to express. Determining the expression bottleneck requires troubleshooting techniques specific
for the given molecule and host. The complex array of intracellular processes involved in protein expression includes transcription, protein folding, post-translation processing, and secretion. Challenges in any
of these steps could result in low protein expression, while the inherent properties of the molecule itself
may limit its production via mechanisms such as cytotoxicity or inherent instability. Strategies to identify
the rate-limiting step and subsequently improve expression and production are discussed here.
Key words Productivity, Difficult to express, Vector design, Cell engineering, Process optimization

1  Introduction
CHO cells have been utilized extensively for recombinant protein
expression; however, not all proteins are expressed at high levels in
this host system. There are many reasons why a protein may be
“difficult to express” and require an alternative strategy to standard
platform workflows for CHO cell production. Although there are
clearly monoclonal antibodies (mAb) that can be challenging to
express at industry standard levels of 5 g/L or more, productivity
improvements for non-mAb therapeutic proteins have lagged
behind [1, 2]. It is more difficult to define productivity levels that
constitute low expression for non-mAb products and a molecule
may be difficult to express not only because of intracellular challenges but also due to its biophysical properties. Determining the
expression bottleneck requires troubleshooting techniques specific
for the given molecule and has historically focused on transcription, protein folding, post-translational processing, and secretion
[3–6]. Alternatively, challenges in producing unique or difficult to
express proteins may have solutions in the bioprocessing space
such as operating parameters or media and feed o
­ ptimization.
Paula Meleady (ed.), Heterologous Protein Production in CHO Cells: Methods and Protocols, Methods in Molecular Biology,
vol. 1603, DOI 10.1007/978-1-4939-6972-2_1, © Springer Science+Business Media LLC 2017

1


2

Christina S. Alves and Terrence M. Dobrowsky

Often, business, regulatory, or biological limitations may require

the introduction of additional process steps or modifications to
reach yield demands with an existing cell line. Previous bioprocess
strategies for improved protein production include chemically
induced specific productivity increases, affecting secreted protein
stability or toxicity in culture, continuous removal of protein from
culture, and general increases to culture biomass. The diversity
that exists among the different CHO host lineages can also be leveraged to improve expression of problematic proteins and antibodies [7, 8].
Engineering CHO cells and their production bioprocess to
better express difficult proteins requires a two-step approach by
which you first determine what is the rate-limiting function and
then develop a strategy to alleviate it. This chapter will outline various strategies that can be used to determine the expression bottleneck and consequently to improve protein expression.

2  Strategy and Methods
2.1  Determining
the Bottleneck

In order to design a comprehensive set of experiments to increase
the production of a difficult to express protein, a fundamental
understanding of the biophysical and biochemical properties of the
protein is essential. The process by which a DNA sequence is converted to a fully folded protein product is complex with steps that
include transcription, translation, post-translational modification,
protein folding, and ultimately secretion (Fig. 1). Any of these individual steps could limit protein expression and may be attributed to
either a poorly designed molecule or suboptimal DNA coding
sequence. Inherent properties of the molecule can also result in the
protein being prone to degradation, aggregation, and other unfavorable inter-protein interactions that can lead to cytotoxicity. Due
to these factors, knowledge of the biology of the protein will aid in
narrowing the scope and focus of troubleshooting efforts.

2.2  Integration
of the Gene of Interest

Transfection of a gene into a cell is followed by the integration of
that gene into the host cell’s genome. This has historically been a
random event in standard cell line development protocols. Efficient
expression of the transgene is highly dependent on both the number of gene copies integrated into the genome as well as the sites of
integration. The latter is greatly influenced by positional effect
variation, which is affected by the local permissiveness of the site as
well as the proximity and interaction with local and distal enhancers.
Although it has been shown that high copy numbers of transgenes
do not always correlate with high cellular productivity [9, 10], the
number of integrated transgenes is an important parameter to
measure as it has been shown to affect expression in some cell lines
[4, 11]. Several methods can be used to measure transgene copy


Optimizing ‘Difficult to Express’ Proteins

3

DNA
TranscripƟon
NUCLEUS

RNA
CYTOPLASM

TranslaƟon

SecreƟon

Protein
Protein Folding
PostTranslaƟonal
ModificaƟon

Fig. 1 Summary of protein synthesis. RNA is transcribed in the nucleus and then transported to the cytoplasm
and translated by the ribosomes. The proteins become bound to the rough ER, where they undergo folding and
processing before moving to the golgi. Soluble proteins undergo post-translational modifications and are subsequently processed through the secretory pathway

number including analysis by southern blot [4], qPCR [11], and
digital droplet PCR [12]. In the situations where gene copy number does not correlate with product expression it is possibly a result
of the transgene being integrated in a suboptimal location in the
genome. Random genome insertion could result in the transgene
benefiting from a location in a highly transcribed genetic region of
euchromatin (referred to as a “hot spot”) or possibly suffering
from the effects of epigenetic gene silencing. By using a targeted
integration approach whereby the gene of interest is inserted into
a predetermined loci in the genome, one can circumvent such
issues by integrating the gene into a known hot spot where a region
of euchromatin and high gene-expression have already been established. Determination of the desired hot spot is often the most
challenging part of developing a targeted integration system. One
approach is to utilize a screen of high expressing cell lines to determine whether expression is driven by a single copy of the transgene. The location of the single integration site may be in an area
that naturally drives high expression and which can be utilized for
other genes of interest. Elaborate systems to determine permissive
loci for integration have been used such as transfecting CHO cells
with a plasmid containing a FRT-tag to specifically screen for single
integration loci with high transcriptional activity [13]. In this work,


Christina S. Alves and Terrence M. Dobrowsky

4

fluorescence in situ hybridization (FISH) was used to locate the
integration site of the FRT sequence or the antibody genes in the
chromosomes. Given the substantial advances in the field of CHO
‘omics, it is now feasible to use next generation sequencing (NGS)
to determine hot spots for integration. Advancements in this technology have increased the speed and throughput of whole genome
(DNA-seq) and transcriptome (RNA-Seq) sequencing such that it
is now feasible to screen clones for the location of genes that have
a high level of expression. A more refined method is targeted
sequencing where the genome is fragmented, incubated with
probes specific for the transgene, and then enriched via a wash
step. This enables sequencing of just the genes with high expression to elucidate their location in the euchromatin. Although it has
not yet been demonstrated, it may be possible to screen early in the
cell line development process and identify clones that display a predefined ‘omics profile that is predictive of productivity using RNAseq [14].
Once a desired site has been elucidated, several methods exist
to insert a gene of interest into a specific location. These methods
which include site-specific recombinases, integrases, or transposases
for the integration of the expression cassettes are summarized in
Table 1. Integrases and transposases allow for multiple integrations
with higher copy numbers at various recognition sites within the
genome [15]. Phage integrases such as PhiC31 integrase rely on
unmodified, native acceptor (attP) and donor (attB) sites, but the
quantity of these sites in the CHO genome may be limiting. On
the other hand, site-specific recombinases have a higher specificity
of integration into a predetermined single site [16]. Flp recombinases have been used in combination with Flp recognition target
Table 1
Methods for targeted integration
Method

Benefits

Disadvantages

References

Integrases and
transposases

Multiple integrations with higher
copy numbers using native
donor and receptor sites

Integrate randomly, limited
number of sites in CHO
genome

[15]

Site-specific
Higher specificity of integration
recombinases (Flp,
into a predetermined single site
Cre/LoxP)

Only can support one
insertion

[18–20]

TALENs

Easy to design for knock in/out,
target DNA sequences using
proteins

High frequency of insertion-­ [21]
deletion mutations,
expensive, and time
consuming to develop

CRISPR/Cas9
system

Target-specific DNA sequences
Can have off target effects,
using RNA, inexpensive, and
IP landscape is undefined
able to screen many sites quickly

[22, 23]


Optimizing ‘Difficult to Express’ Proteins

5

sites (FRT) for targeted integration of transgenes into mammalian
cells with a high specificity of integration and low off target effects

[17]. This is accomplished either by using Flp-in or Flp recombinase-­
mediated cassette exchange (RMCE) strategies. RMCE uses a set
of hetero-specific FRT sites to direct a gene of interest to a predetermined and tagged locus that has been characterized to yield
high protein expression [18]. A binary RMCE expression system
has been used to co-express multiple proteins with different combinations of expression levels [17]. The Cre/loxP system for site-­
specific DNA recombination has also been used as a tool for
transgene integration in CHO cells [19]. Recent work has demonstrated the ability to insert multiple transgenes into a targeted site
of the CHO cell genome using Cre recombinase-incorporating
integrase-defective retroviral vectors [20]. More recently, site-­
specific gene insertion in CHO cells has been performed using
transcription activator-like effector nucleases (TALENs) [21] or
CRISPR/Cas9 RNA-guided nucleases [22]. Precise insertion of a
gene expression cassette at a defined loci in CHO cells has been
accomplished using the CRISPR/Cas9 system following a simple
drug-selection methodology that resulted in homogeneous transgene expression [23].
2.3  Transcription

Traditionally, transcription has been considered the dominant factor in controlling protein production.
In a particular study to elucidate the mechanisms and processes-­
limiting gene expression in CHO cells, transcription appeared to
be the primary limitation for low- and medium-producing cell
lines, whereas in high-producing cell lines post-translational limitations tended to dominate [6]. Within the process of transcription
the rate-limiting step is likely to be initiation. Due to the highly
condensed nature of DNA into chromatin structures, transcription
complexes often have trouble accessing certain regions of
DNA. Chromatin remodeling, the rearrangement of chromatin
structure by various remodeling complexes, is therefore required
for activation of transcription. Additionally, chromatin as well as
other proteins involved in transcriptional control can be altered by
methylation, acetylation, phosphorylation, and other modifications to affect whether a gene is active or inactive [24]. The synergistic effect of these modifications, known as the “histone code,”

adds complexity in the form of epigenetic regulation of genes.
Specific methods to affect these features are detailed here.
Unstable protein expression has been observed in CHO cells
where mRNA decreases despite constant transgene copy numbers
[11], which suggests that either the mRNA is degrading or that the
promoter is being silenced. The expression level of mRNA transcript for the gene of interest can be determined by quantitative
real-time reverse transcription-PCR (qRT-PCR). There have been
mixed results on the correlation between productivity of a cell line


6

Christina S. Alves and Terrence M. Dobrowsky

and mRNA levels in CHO cells. Some studies showing that high
mRNA levels and gene copy numbers in methotrexate amplified
cells correspond to high specific productivities [4], whereas others
have seen no correlation between mRNA levels and expression
[25]. Despite these conflicting reports, it may still be valuable to
assess mRNA levels of the protein of interest to ensure that the
sequence of interest is being adequately transcribed. Additionally, in
the case of antibodies, using qPCR to determine mRNA levels can
be a useful diagnostic tool for determining the ratio of heavy to
light chain which may be important to ensure assembly of the mAb.
Studies have indicated that it is advantageous to have an excess of
light chain in relation to the heavy chain for optimal antibody production [26, 27]. The ratio of heavy to light chain can be influenced by optimizing the quantities of DNA transfected or by the
vector design. If a two-plasmid system is utilized, where the heavy
and light chains are located on different vectors, the mixture of
DNA used at the point of transfection can be used to modulate the
ratio. Alternatively, one can use a single-­plasmid system that employs

IRES-mediated bi- or tri-cistronic vectors that enable control of
heavy to light chain expression at different ratios.
2.3.1  Methylation

DNA methylation has been reported to repress gene expression,
whereas hypomethylation of DNA in the promoter region can elevate gene transcription activity [28, 29]. Enzymatic methylation of
cytosine at carbon 5 is well known as a fundamental epigenetic
mechanism that results in gene silencing [30]. DNA methylation
often occurs at CpG dinucleotides sites within promoter regions
which subsequently renders the promoter transcriptionally inactive.
Bisulfite treatment of DNA can be used to differentiate between
methylated and unmethylated CpG sites. In this method, sodium
bisulfite converts cytosine residues to uracil residues via deamination
at C4, while 5-methylcytosine remains unaffected [31]. Subsequent
amplification of the region by PCR allows for further analysis via
DNA sequencing [31] or microarray analysis [32]. Methylationspecific real-time qPCR is a highly sensitive measurement of promoter methylation and has been utilized to correlate hCMV-IE
methylation with unstable protein expression in recombinant CHO
cell lines [28]. Chemical compounds exist that can affect the degree
of DNA methylation, specifically a class of molecules known as DNA
methyltransferase inhibitors (iDNMTs). These compounds, which
include azacytidine, RG-108, and hydralazine, have been tested in
CHO cells for their capacity to increase cellular productivity in transient gene expression systems with some success [33].

2.3.2  Acetylation

Acetylation of histones typically plays a role in transcriptional control
of active genes [34]. Histone acetyltransferases (HATs) and histone
deacetylases (HDACs) control the enhancement of transcription by
modifying histone acetylation. The most commonly used mechanism

Optimizing ‘Difficult to Express’ Proteins

7

to control acetylation in CHO cell cultures is the use of HDAC
inhibitors to prevent deacetylation. Several studies have demonstrated
that sodium butyrate [35, 36] and/or valproic acid [37] can be used
to enhance mRNA transcription and increase specific productivities.
However, these compounds can also have adverse negative effects on
cell growth due to cytotoxicity and induction of apoptosis [38].
Their appropriateness must be evaluated to determine the optimal
concentration for enhanced productivity, but they are commonly
used over short production durations.
2.3.3  Vector Design
Elements

The design of vectors to promote active transcription by creating a
favorable chromatin environment around the transgene has been
extensively reviewed [39]. The available methods either alter the epigenetic environment of the DNA surrounding the transgene or prevent the surrounding environment from affecting transcription of the
gene of interest [39]. A list of vector design elements for CHO cells
is shown in Table 2. In order to enhance gene transcription and
reduce transgene expression dependence on the surrounding chromatin, strong cellular enhancers such as the Locus Control Region
(LCR) have been utilized [40]. The LCR is a cis-­acting DNA element that controls the expression of human β-globin locus genes.
Unfortunately, these enhancers only function in certain cell lines and
cannot be used as general regulatory elements in all mammalian cells.

Table 2
Vector design elements to enhance transcription
Category

Specific Elements

Description

Reference

Locus control
regions

human β-globin,
HGH

[40]
Can lead to stable high expression of
transgene in a copy number dependent
manner. Limited usefulness in mammalian
cell lines

Insulators

cHS4

Block the positive action of enhancers, can
[41]
modestly increase transgene expression in
CHO, but may not be universally effective

Matrix Attachment
Regions (MARs)

Chicken lysozyme
MAR, human
β-globin MAR, X
MAR

Bind to the nuclear matrix and affect the
arrangement of chromatin into loops.
Have shown some positive effects on
transgene expression in CHO

[42, 43]

HNRPA2B1, CBX3,
Ubiquitous
TBP and PSNB1
Chromatin
Opening Elements
(UCOEs)

Derived from promoters of housekeeping
genes that are transcriptionally active.
Large increases in gene expression were
observed but elements are typically large
(~16 kb), promoter dependent

[41, 44]

Antirepressor or
STAR7, STAR44,

STAR (stabilizing
STAR67
and antirepressor)

Small elements (< 2 kb) that block
[45, 46]
chromatin-associated repressors. Convey
copy number-dependent stable expression


8

Christina S. Alves and Terrence M. Dobrowsky

Regulatory elements that block interactions between the enhancer
and promoter while not directly affecting their individual activity are
referred to as insulators. Insulators such as the chicken β-globin 5′
hypersensitive site 4 (cHS4) have been used to control the effects of
the surrounding chromatin environment on the transgene [41].
Matrix-attachment regions (MARs) are DNA elements that
bind to the nuclear matrix and are believed to influence gene
expression by affecting the arrangement of chromatin into loops.
MARs, such as chicken lysozyme MAR, human β-globin MAR,
and X MAR, can associate with euchromatin and act as boundary
or insulator elements, and hence create an independent chromatin
structure from the surroundings [42]. Although the specific mechanisms by which MARs function in the cell are not entirely understood, they have been effective in enhancing the expression of
target proteins in mammalian cell cultures. MARs can be integrated
into expression vectors that may increase the percentage of high-­
producer cells in a population to reduce the number of clones that
need to be screened. Protocols are available that describe how to

incorporate MARs into vectors that can then be transfected into
CHO cells for increased transgene expression [43].
Other elements that have been shown to protect a transgene from
silencing and convey higher transgene expression are ubiquitous chromatin opening elements (UCOEs), which are derived from the promoters of housekeeping genes that are typically transcriptionally active
[44]. Some well-characterized UCOE pairs include HNRPA2B1 and
CBX3 or TBP and PSNB1, which are DNA regions that contain a pair
of divergent gene promoters that are transcriptionally active in all cells
of an organism. Large UCOEs of up to 16 kb have been used to generate high-level and stable transgene expression for cells in extended
culture by increasing the efficiency of the CMV promoter. Because
UCOEs directly affect transcriptional regulation that is dependent on
the promoter and its activity, these elements have variable effects on
expression of a target protein in CHO cells and need to be tested for
specific host and vector strategies [41].
Antirepressor or STAR (stabilizing and antirepressor) are DNA
elements that block chromatin-associated repressors and have been
used to flank transgenes in mammalian expression vectors. These
elements affect the spread of methylation and histone deacetylation
from the adjacent chromatin environment into the transgene
region. They can enhance protein expression as well as overcome
genetic instability caused by positional effects, epigenetic silencing,
or loss of gene copy number [45]. The positive effects of STAR
elements are most pronounced when high selection stringency is
used to develop stable clones in CHO cells [46].
2.4  Translation

The process of translation consists of initiation, elongation, termination, and recycling. The initiation of mRNA translation is an
essential precursory step that influences cell growth and protein


Optimizing ‘Difficult to Express’ Proteins

9

synthesis via the coordination of numerous initiation factors [47].
The secondary structure of the mRNA can affect translational efficiency. Formation of a closed loop structure consisting of mRNA,
a number of eukaryotic initiation factors (eIFs), and ribosomal
proteins can potentially increase global translation efficiency by
promoting re-initiation of translation. High-producing cell lines
have been shown to maintain appropriate levels of these translation
initiation factors [48]. Use of cell engineering approaches to maintain the levels of these initiation factors may allow for generation of
new host cell lines with high growth and recombinant protein productivity. Another target to improve translation is the global metabolic sensor and processing protein mammalian target of rapamycin
(mTOR). The treatment of CHO cell cultures with adenosine
results in growth arrest but also increases productivity. The adenosine contributes to high ATP levels which increase mTOR activity, inhibiting the key translation initiation repressor 4E-BP1 [49].
mTOR has also been shown to influence ribosomal protein synthesis, translation initiation, and translation elongation in addition to
other cellular functions. Its overexpression in CHO cells has
resulted in increased specific antibody productivity [50] making it
an attractive engineering target for difficult to express proteins.
2.4.1  Codon Optimization

Because it is often the case that human proteins are being expressing in CHO cells and synonymous codons are used with different
frequencies in different organisms (known as codon bias) [51], it is
important to ensure that the transgene sequence is optimized. By
optimizing the DNA sequences for expression in CHO cells, one
can ensure that certain preferred codons are translated more accurately and/or efficiently. Poorly optimized sequences can adversely
affect protein translation, and subsequently protein expression, by
preventing the host from efficiently translating the rare codons.
Codon optimization has been used to increase protein expression
in multiple studies [52, 53] and there are several websites and services that will perform codon optimization for expression in CHO
cells of a given amino acid sequence. A list of codon usage for
CHO cells is shown in Table 3.

Another important consideration is the translation initiation
sequence located upstream of the start codon (AUG). The efficient
consensus sequence GCCACC(AUG)G, known as the Kozak
sequence [54], yields high fidelity and efficiency of initiation and is
typically used at the start of the coding sequence.

2.4.2  Splice Sites

Splice sites are located between an exon and an intron. The splice
site upstream of an intron is referred to as the donor splice site
(5′–3′ direction), while the one downstream of an intron is
the acceptor splice site (3′–5′ direction). The acceptor splice site
corresponds to the end of an intron (AG) and the donor splice site
corresponds to the beginning of an intron (GT). Splice sites can


10

Christina S. Alves and Terrence M. Dobrowsky

Table 3
Codon usage in Chinese hamster genes
Amino
acid.

Relative
Codon frequency

Amino
acid

Relative
Codon frequency

Amino
acid

Relative
Codon frequency

Ala

GCT

22.4

His

CAT

10.2

Ser

TCA

10.3

GCA

16.3

CAC

12.9

AGT

11.4

GCC

25.9

TTG

14.1

TCC

16.5

GCG

5.0

CTC

18.4

AGC

16.4

AGA

10.1

CTG

38.8

TCT

16.0

CGA

7.2

CTA

7.6

TCG

3.4

CGG

10.1

CTT

13.2

ACT

14.1

AGG

10.2

TTA

6.4

ACA

15.7

CGC

9.3

ATT

17.4

ACC

20.3

CGT

5.6

ATC

24.8

ACG

4.5

AAT

17.4

ATA

6.9

Trp

TGG

13.1

AAC

21.2

AAG

38.4

Tyr

TAT

13.1

GAT

24.6

AAA

24.6

TAC

16.4

GAC

28.1

Met

ATG

23.0

GTA

7.8

TGT

9.1

Phe

TTC

22.0

GTT

11.6

TGC

10.3

TTT

19.6

GTG

30.1

CAA

10.3

CCA

15.7

GTC

15.7

CAG

33.4

CCC

17.0

TGA

1.2

GAA

28.4

CCT

16.7

TAA

0.6

GAG

41.1

CCG

4.3

TAG

0.5

GGA

15.8

GGG

13.4

GGT

12.8

GGC

21.3

Arg

Asn

Asp

Cys

Gln

Glu

Gly

Leu

Ile

Lys

Pro

Thr

Val

Stop

The amino acid abbreviation is shown adjacent to the codon and the relative frequency in identified genes of the
Chinese hamster (Cricetulus griseus). The source of these data is These records were
a snapshot of usage as of March 2016. A total of 331 genes and 153,527 codons contributed to this data set.

also unintentionally exist in a coding sequence. As possible acceptor and donor splice sites, every AG and GT in a DNA sequence
needs to be evaluated as either a real splice site or a pseudo splice
site to ensure that the sequence is not compromised during translation. In addition to the sequences immediately adjacent to the


Optimizing ‘Difficult to Express’ Proteins

11

Table 4
Programs available to identify potential splice sites in a DNA sequence
Program

Website

Reference

Gene splicer

/>
[107]

NetGene2

/>
[108]

HSPL

/>
[109]

NNSplice

/>
[110]

GENIO splice site and
exon predictor

/>
SpliceView

/>
[111]

splice event, distal sequences also contribute to the probability of
splicing. Several programs that are summarized in Table 4 exist

online to help evaluate a sequence and the probability that donor
and acceptor splice sites are present.
2.5  Protein Folding
and Processing

Proper protein folding is essential for adequate expression of a
molecule. The ER is responsible for ensuring that proteins are
properly processed and folded and as such there are specific quality
control systems to aid in the efficiency of folding and eliminate
misfolded proteins. When a protein is misfolded in the ER it is
proteolytically destroyed via the ER-associated degradation
(ERAD) pathway. Similarly, the unfolded protein response (UPR),
a signal cascade that protects cells from aggregated protein by
restoring ER function, can be triggered by intracellular accumulation of misfolded protein. Several chaperones and cofactors are
involved in the process of protein folding and assembly and can be
modulated to enhance protein expression.

2.5.1  Chaperones

Molecular chaperones are proteins that assist the folding and assembly of intracellular proteins which may be good targets for cellular
engineering to improve protein expression. Heat shock proteins
(HSPs) function as molecular chaperones and are ­primarily responsible for protein folding, assembly, translocation, and degradation
under cellular stress. Chaperones also prevent newly synthesized
polypeptide chains from aggregating into defective proteins. BiP is
a HSP70 molecular chaperone that binds newly synthesized proteins as they are translocated into the ER, and preserves them in a
state suitable for subsequent folding. Protein disulfide isomerase
(PDI) is an enzyme in the ER that catalyzes the formation and
breakage of disulfide bonds to assist in protein folding [55].
Cyclophilin B (CypB) interacts with other proteins in the ER
including BIP and PDI to form chaperone complexes that facilitate

protein folding. Some work has been done on expressing molecular


12

Christina S. Alves and Terrence M. Dobrowsky

chaperones in CHO cells to improve productivity of difficult to
express proteins. Specifically, co-expression of CypB with a difficult
to express antibody improved cell growth but had no effect on cell
specific productivity whereas, co-expression with BIP increased the
productivity but reduced cell growth [56]. In another CHO cell
line expressing a fusion protein, co-transfection of CypB followed
by addition of chemical chaperones at the start of stationary phase
increased cell-specific production and eliminated protein aggregation [57]. The disparity in these findings suggests that the engineering of molecular chaperones for increased protein expression may
be product and cell line dependent.
Chemical chaperones are a group of compounds that can improve
the folding capacity of the ER, facilitate protein folding in the ER,
and enhance the secretion of protein. Chaperones can be added to
cell culture to potentially improve expression of a difficult to express
protein, especially if misfolding or aggregation is occurring intracellularly. This approach provides a simpler alternative to overexpression
of molecular chaperones given their uncertain effect on productivity.
PBA (Sodium 4-phenylbutyrate) has been used to promote the secretion of a mutated protein C from CHO cells by utilizing an unconventional GRASP55-dependent pathway that restores normal
intracellular trafficking through the ER and golgi [58]. Treatment of
mammalian cells with PBA has been shown to suppress ER stress by
chemically enhancing the ER capacity to cope with the expression of
misfolded protein, ­preventing intracellular aggregates by facilitating
protein degradation [59]. Osmotically active chaperones such as
DMSO, glycerol, and proline have been used to increase specific productivity, but also can have a negative impact on cell growth.
Additionally, DMSO and proline can reduce protein aggregate formation in culture supernatants by an undefined mechanism [60]. To

counteract the negative effect on cell growth and viability, the addition of DMSO [61] and glycerol [62] in a two-staged approach has
been utilized to increase the specific productivity of CHO cell lines.
Similarly, a combination of PBA and glycerol has been added at the
start of stationary phase alongside expression of the molecular chaperone CypB to maximize cell specific production and eliminate protein aggregation [57]. Analogous to the overexpression of molecular
chaperones, the effect of these chemicals may be cell line and protein
specific and their suppression of cell growth requires that their concentration and dosing strategy be carefully considered.
2.5.2  Bioprocess
Modifications

In addition to enabling or improving chaperone protein function,
aggregation can be prevented through bioprocess modifications.
Altering the cellular redox potential by supplementing media with
the antioxidant glutathione can reduce aggregation [63], while
media optimization of components such as cysteine or glycerol can
reduce aggregation [64, 65]. Additionally, a reduction in temperature has been shown to reduce aggregation and positively affect


Optimizing ‘Difficult to Express’ Proteins

13

protein processing by reducing the mis-/unfolded protein destined for degradation via the ERAD pathway [64–66].
While protein folding is typically a primary driver for proper
function, other post-translational modifications in CHO cultures
that could affect protein structure and function include glycosylation, oxidation of methionine, deamidation of asparagine and glutamine, hydroxylation, and sulfation [67]. It is possible that in
trying to achieve the specific ranges or combinations of these modifications, the expression of the protein may be compromised.
The addition of sodium butyrate has been used regularly
used to improve specific protein production, but it has also been
shown to affect post-translational modifications of the protein of
interest by altering histone modification, chaperones, lipid

metabolism, and protein processing [68]. These changes in posttranslational modifications can result in increased microheterogeneity and reduced sialylation which may decrease in vivo activity
[69]. However, a combination of sodium butyrate treatment with
lower production culture temperatures has been shown to mitigate these risks [70].
2.6  Secretion

Translocation of a nascent protein from ribosomes through the
cytosol into the endoplasmic reticulum is mediated by its signal
peptide and is an essential stage in protein secretion. The efficient
secretion of recombinant proteins from CHO cells is strongly
dependent on the signal peptide used, which makes identifying the
optimal signal sequence for each target protein an important step
in maximizing the efficiency of protein secretion [71]. In mammalian cells, a signal peptide that ranges from 5 to 30 amino acids
at the N-terminal end of nascent proteins is recognized by the signal recognition particle (SRP) in the cytosol as the protein is being
synthesized on the ribosome. The SRP then transfers the complex
consisting of the SRP and ribosome-nascent chain to a receptor on
the endoplasmic reticulum (ER) membrane, where it is eventually
translocated to the lumen of the ER and the signal peptide is
cleaved by a signal peptide peptidase [72]. The translocation of
proteins into the ER lumen is considered a bottleneck of the secretory pathway and has motivated further investigation into enhancing the capacity of signal peptides for recombinant protein
expression. Several studies have shown positive effects with native
signal peptides, natural signal peptides derived from human albumin and human azurocidin, as well as optimized signal sequences
[71, 73, 74] indicating the importance of carefully evaluating signal sequences for the expression of a given protein. There are several resources online (Table 5) that assess the probability that a
peptide is a suitable signal sequence as well as how efficiently the
sequence will be cleaved from the protein [75, 76].
Secretion of antibodies has been affected by improper cleavage
of the light chain from the signal peptide due to a dysfunctional SRP


14

Christina S. Alves and Terrence M. Dobrowsky

Table 5
Websites that offer free signal peptide prediction algorithms
SignalP 4.1 Server ( />PrediSi: Prediction of Signal peptides ( />Signal-BLAST Signal Peptide Prediction (e.
sbg.ac.at/signalblast.html)

complex, which results in its precipitation in an insoluble cellular
fraction [3]. Western blotting of intracellular fractions can be used to
determine whether light chain is precipitating in the cells. This can
be achieved by using standard lysis techniques to create protein
extracts, followed by blotting with light chain-specific antibodies. If
this inadequate cleavage of the light chain is affecting expression,
proper processing and secretion can be restored by over-expressing
SRP proteins such as the signal recognition protein, SRP14 [3].
2.6.1  Russell Bodies

Russell bodies are intracellular aggregates of immunoglobulins
stored in the endoplasmic reticulum that can form during protein
biosynthesis. The formation of Russell bodies depends on the physiochemical properties of the protein coded by the variable regions of
the heavy and light chains as well as extrinsic factors such as stressful
cell culture conditions [77]. Immunofluorescent microscopy can be
used to determine whether Russell bodies are forming intracellularly
for a given protein. Cells must be fixed, permeabilized, and stained
with fluorescently conjugated antibodies to the IgG of interest followed by fluorescent imaging and quantification [78]. A frequency
of Russell body phenotype can be calculated by determining the
number of Russell bodies observed and normalizing to the overall
number of cells in the image. This value can be compared to an
alternative cell line that produces an easy to express molecule to
determine whether intracellular aggregation is resulting in reduced

protein expression. Recent studies suggest that there are IgG antibody sequences with intrinsically high condensation/aggregation
propensities that are more prone to form Russell bodies in the ER
lumen [78]. This implies that if a protein is not being expressed due
to the formation of these intracellular aggregates the sequence may
need to be altered to enable better expression.

2.7  Protein Toxicity

Another mechanism that leads to insufficient protein production is
the inherent toxicity of the protein being expressed. Limiting cellular exposure to high concentrations of the protein or adapting
cell lines specifically to be resistant to the toxic protein can improve
growth and subsequently yield. Toxicity of the protein of interest
may diminish a cell’s ability to recover after transfection and selection. Ultimately, this toxicity will result in the unintentional selection of low-producing cells from the population. A dose-response


Optimizing ‘Difficult to Express’ Proteins

15

study with purified protein and host cells is often the most direct
determination of cytotoxicity, wherein the toxicity of the purified
protein and buffer to the naive culture is assessed. If the quantity
of purified protein is limiting, similar results can be confirmed by a
less ideal but easy-to-execute experiment utilizing spent media
from a transient transfection. Clarified culture supernatant from a
transiently transfected culture will likely contain a toxic level of
protein but sufficient unprocessed metabolites for continued
growth. Performing a dose-response study using this supernatant
incrementally blended with fresh media may yield similar results to
dosing purified protein. However, toxicity will be confounded

with other supernatant components such as metabolic waste products and the highest concentration available for testing will be
limited.
Protein toxicity can be mitigated by multiple methods. One
approach, likely the most extreme, is to alter the protein itself to
reduce its cytotoxicity. Modifying the protein to produce a more
stable, less toxic form while retaining its intended biological function can be difficult but possible with extensive knowledge of the
structure function relationship [79]. Introducing stabilizing agents
that can be degraded later can be effective as long as a more complicated downstream purification process is acceptable. N-terminal
tags such as a small ubiquitin-like modifier (SUMO) can be used to
create “dormant fusion” proteins with decreased toxicity that are
capable of being cleaved downstream [80]. An alternative option is
to modify the promoter in the transfected vector rather than the
protein of interest. When the optimum environment for cell
growth varies significantly from that of protein production, an
inducible expression system may be appropriate. An inducible
expression system can enable high cell densities to be achieved
prior to protein production and subsequently alleviate the effects
of toxicity or degradation [81]. Inducible promoter systems are
commercially available [82] for direct implementation. In general,
it can be difficult to ensure that transgene expression is entirely
inhibited prior to the addition of the inducing agent [83].
Therefore, most industrially relevant systems utilize promotortransactivator combinations. In these systems, the activity of a constitutively expressed transactivator is controlled via supplementation
of some complementary ligand [84]. The tetracycline (Tet) inducible system, often referred to as Tet-on, allows for the expression of
the protein of interest in the presence of tetracycline. Alternatively,
protein expression could be repressed in the presence of tetracycline, the Tet-off system, and activated by complete media
exchange. The Tet-on system is often preferred for recombinant
protein production as it is relatively straightforward to supplement
culture with tetracycline while removing it would require significant liquid handling at scale [85, 86]. Other applications for inducible systems include increasing specific productivity by arresting