engineering nutrient and by-product metabolism of cho cells · engineering nutrient and by-product...
TRANSCRIPT
General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
You may not further distribute the material or use it for any profit-making activity or commercial gain
You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.
Downloaded from orbit.dtu.dk on: Sep 29, 2020
Engineering nutrient and by-product metabolism of CHO cells
Domingues Pereira, Sara Isabel
Publication date:2019
Document VersionPublisher's PDF, also known as Version of record
Link back to DTU Orbit
Citation (APA):Domingues Pereira, S. I. (2019). Engineering nutrient and by-product metabolism of CHO cells. TechnicalUniversity of Denmark.
Engineering nutrient and by-product metabolism of CHO cells
Ph.D. Thesis
of
Sara I. D. Pereira
Main supervisor: Mikael Rørdam Andersen, Professor MSO, Technical University of Denmark
Co-supervisor: Helene Faustrup Kildegaard, Senior Scientist, Novo Nordisk
The Novo Nordisk Foundation Center for Biosustainability
Technical University of Denmark, DTU
I
In memory of those who left us
Em memória dos que partiram
Abstract Chinese Hamster Ovary (CHO) cells are the preferred hosts for the production of therapeutic glycoproteins
used for treating severe health conditions. It is of interest to improve the production of such proteins in CHO
cells to cut the production costs without compromising on product quality, which is critical for patient safety.
However, a major challenge is that CHO cells have an inefficient metabolism, characterized by the build-up of
toxic metabolites, such as lactate and ammonia. These impair cell growth and decrease productivity during cell
cultivations. Recent advances in the field, such as sequencing of Chinese hamster (Cricetulus griseus) and CHO
cell lines genomes, the publishing of accurate metabolic models and appearance of new precise genome editing
tools, such as the CRISPR/Cas9 system, create a favorable landscape for the rational engineering of CHO cells
towards optimal nutrient and by-product metabolism. Overall, this thesis aims to identify and study
metabolites that are toxic and inhibit cell growth in a similar way as the well investigated by-products of the
mammalian metabolism, followed by cell line engineering approaches to generate cells with improved
phenotypes. First, a review article describing metabolites that are biomarkers of the metabolic status of the cells
and are linked to cell growth inhibition or cell death is presented. The second part of this work covers
applications of cell line engineering tools to target the cell metabolism. Thus, as we have identified targets
participating in amino acid catabolic pathways, engineering of the nutrient metabolism is described as a
strategy to obtain enhanced cell factories. The single or combinatorial disruption of eleven genes using the
CRISPR/Cas9 system was carried out to decrease by-product formation and increase amino acid availability
for protein biosynthesis and important cellular processes. Moreover, this section includes the study of the
effects of engineering the co-factor metabolism via G6pd overexpression. The findings related to cell
physiology and resistance to induced cellular stress are also described. Finally, to understand how ammonium
is affecting the CHO cells used in-house, the study of dose- dependent effects of ammonia chloride in cell
growth is presented. Altogether, this thesis compiles a set of studies employing state-of-the-art methods for cell
line development and metabolic engineering. Concluding remarks and future perspectives are presented to
close this work.
III
Dansk sammenfatning Kinesisk Hamster Ovarie (CHO)-celler er den foretrukne vært til produktion af terapeutiske glycoproteiner,
der anvendes til behandling af alvorlige sygdomme. Det er stor interesse i at forbedre produktionen af sådanne
proteiner i CHO-celler for at reducere produktionsomkostningerne uden at gå på kompromis med
produktkvaliteten, hvilket er kritisk for patientsikkerheden. En stor udfordring er dog, at CHO-celler har en
ineffektiv metabolisme, der er kendetegnet ved akkumulering af toksiske metabolitter, såsom laktat og
ammonium. Disse svækker cellevækst og mindsker produktiviteten under celledyrkning. Nye fremskridt inden
for feltet, tilladt af sekventeringen af kinesisk hamster- og CHO-cellelinie genomer, publikationen af nøjagtige
metabolske modeller og tilkomsten af nye præcise genomredigeringsværktøjer, såsom CRISPR/Cas9-systemet,
skaber et gunstigt landskab for rationel konstruktion af CHO-celler der kan føre til optimal næringsstof og
biproduktmetabolisme. Samlet set sigter denne afhandling mod at identificere og studere metabolitter, som er
toksiske og som hæmmer cellevæksten på samme måde som de allerede velundersøgte biprodukter af
pattedyrmetabolismen, efterfulgt af cellelinie-konstruktionsmetoder til generering af værtsceller med
forbedrede fænotyper. For det første, præsenteres en review-artikel der beskriver metabolitter der er
biomarkører for cellernes metabolske status og er forbundet med celledød eller hæmmelse af cellevækst. Den
anden del af dette arbejde dækker over anvendelsen af værktøjer rettet mod cellemetabolismen. Således,
eftersom som vi har identificeret mål der deltager i aminosyre katabolske veje, beskrives modificering af
metabolisme som en strategi til at opnå forbedrede cellefabrikker. Enkelt eller kombinatorisk forstyrrelse af
elleve gener ved anvendelse af CRISPR/Cas9-systemet blev udført for at mindske biproduktdannelse og øge
tilgængeligheden af aminosyre, da disse er byggestenene i biosyntese af proteiner og anvendes i vigtige cellulære
processer. Desuden omfatter dette afsnit undersøgelsen af virkningerne af modificering af co-faktor
metabolisme via G6pd over-ekspression. Resultater relateret til cellefysiologi og resistens over for induceret
celle stress er også beskrevet. For at forstå, hvordan ammonium påvirker internt anvendte CHO-celler, bliver
undersøgelsen af dosisafhængige virkninger af ammonium-chlorid i cellevækst præsenteret. Samlet set
udarbejder denne afhandling et sæt undersøgelser, der anvender state-of-the-art metoder til udvikling af
cellelinier, metabolske teknikker og metabolomics. Afsluttende bemærkninger og fremtidige perspektiver
præsenteres for at afrunde dette arbejde.
IV
Acknowledgements JJ
Dedico este trabalho à minha querida Mãe, ao meu querido Pai e à minha querida mana. Obrigada pelo “amor e amizade” e pelo apoio incondicional apesar da distância! Agradeço também aos meus avós, primas e primos, tios e tias por todo o apoio. My sincere thank you to my main supervisor, Mikael, for opening the doors to the world of CHO cells! Thank you for accepting me as your student and for giving me the unique opportunity to join the eCHO Systems ITN. Thank you for always sharing your knowledge and optimistic views during the ups and downs of the project and, nonetheless, for the boldness and good sense. Helene, thank you for the guidance and co-supervision, for being an example of leadership and great management. To Christian Müller, thank you for giving me the possibility of doing my external stay at AGC Biologics A/S. I would like to extend my acknowledgments and thank yous: To Ankita, Nusa, and Thomas for the uncountable foosball sessions and fun times! To the office mates, Julie (thank you for translating the abstract to Danish) and Lise for the fruitful discussions and for always being available to help. To Daniel for the opportunity to learn and work together with you. To the current and former members of CLED Che Lin, Daria, Henning, Hooman, Jae, Johan, Kai, Kim, Manuel, Nachon, Saranya, Thomas K., TK, thank you for creating a friendly atmosphere in the lab. To the core units at CfB a special thank you to the analytics and to Nachon and KK and for assisting my experiments using FACS. To the current and former members of CHO CORE: Johnny, KK, Zulfiya, Karoline, Mikkel, Marianne, Maria, and Kristian – thank your having the analyzer up and running – Stef and Helle – for the precious help in the protein lab – and Sara for always having nice wise words to share. To all the members of CHO management in Denmark and abroad. To all members of the eCHO consortium: PI’s and scientific board, thank you for the feedback in every annual meeting. To all my fellow eCHOs around Europe: You are the best! Thank you for the great memories and great science! To the girls (and guy) from my basketball team at DTU for accepting me in your midst. To my friends, Rachel for sending me off to this side of the Öresund bridge, Vera for turning short visits easily into longer ones and for listening to me, and Ivana for the nice meetups before lab work took over. To Lars, for standing by my side and giving me the strength to move forward when my own was not enough. I am looking forward to the rest of our lives together! <3 SDG
V
List of publications This thesis includes the following articles and manuscripts:
I. Impact of CHO Metabolism on Cell Growth and Protein Production: An Overview of Toxic and Inhibiting Metabolites and Nutrients. Pereira, S., Kildegaard, H. F. and Andersen, M. R. (2018), Biotechnol. J., 13: 1700499. doi:10.1002/biot.201700499
II. Reprogramming amino acid catabolism in CHO cells with CRISPR/Cas9 genome editing improves cell growth and reduces byproduct secretion Daniel Ley, Sara Pereira, Lasse Ebdrup Pedersen, Johnny Arnsdorf, Hooman Hefzi, Anne Mathilde Lund, Tae Kwang Ha, Tune Wulff, Helene Faustrup Kildegaard, Mikael Rørdam Andersen (2018) – manuscript in submission
III. Physiological study of CRISPR/Cas9-mediated disruption of branched-chain amino acid
transaminases in CHO cells Sara Pereira, Daniel Ley, Lise Marie Grav, Helene Faustrup Kildegaard, and Mikael Rørdam Andersen – manuscript revised after peer-review, in submission as “BCAT1 and BCAT2 disruption in CHO cells has cell line-dependent effects”
IV. A targeted study of stable overexpression of Glucose-6-phosphate dehydrogenase (G6pd) in
CHO-S cells: effect on cell growth and protective properties against ROS inducers and cytotoxic agents Sara Pereira, Lise Marie Grav, Tune Wulff, Helene Faustrup Kildegaard, and Mikael Rørdam Andersen – manuscript ready for submission
VI
Table of contents Preface .................................................................................................................................................................................I
Abstract ............................................................................................................................................................................III
Dansk sammenfatning ....................................................................................................................................................IV
Acknowledgements ..........................................................................................................................................................V
List of publications .........................................................................................................................................................VI
Table of contents ............................................................................................................................................................VII
Thesis structure ..................................................................................................................................................................1
1. Introduction ............................................................................................................................................................3
1.1. CHO cells in the biopharmaceuticals market .................................................................................................... 3
1.2. CHO cell factories .................................................................................................................................................... 4 1.2.1. The advantages and disadvantages of using CHO cell factories to produce recombinant therapeutic proteins ......................................................................................................................................................................... 4 1.2.2. Cell line development ..................................................................................................................................... 6
1.3. Nutrient and by-product metabolism of CHO cells ......................................................................................... 6 1.3.1. Glucose metabolism and by-product formation ...................................................................................... 7 1.3.2. Amino acid catabolism and by-product formation ................................................................................. 8
1.4. Engineering of CHO cells using synthetic biology tools .................................................................................. 9 1.4.1. Genome editing tools ..................................................................................................................................... 9 1.4.2. Engineering CHO cells for improved protein production and improved metabolism ................. 10
Chapter 1 - Overview of nutrients and growth inhibitory metabolites affecting CHO cell metabolism .....................12
Paper I – Impact of CHO Metabolism on Cell Growth and Protein Production: An Overview of Toxic and Inhibiting Metabolites and Nutrients......................................................................................................................................... 13
Chapter 2 - Engineering the metabolism of CHO cells .................................................................................................27
Paper II – Reprogramming amino acid catabolism in CHO cells with CRISPR/Cas9 genome editing improves cell growth and reduces byproduct secretion .................................................................................................................................28
Paper III – Physiological study of CRISPR/Cas9-mediated disruption of branched-chain amino acid transaminases in CHO cells .......................................................................................................................................................... 60
Paper IV – A targeted study of stable overexpression of Glucose-6-phosphate dehydrogenase (G6pd) in CHO-S cells: effect on cell growth and protective properties against ROS inducers and cytotoxic agents .................................. 80
Chapter 3 – Study of dose-dependent effects of metabolite additions on cell growth .............................................107
Conclusion and future perspectives ............................................................................................................................111
References ......................................................................................................................................................................115
Appendices.....................................................................................................................................................................122
Appendix 1: Paper II – Supplementary materials ..................................................................................................................123
Appendix 2: Paper III – Supplementary materials ................................................................................................................ 135
Appendix 3: Paper IV – Supplementary materials ................................................................................................................ 154
Thesis structure
This thesis aims to shed some light on the causes of metabolic inefficiency of Chinese hamster ovary (CHO)
cells observed in cultivations, where by-products accumulate leading to poor cultivation performance, and
to generate cells with improved phenotypes. The work consists of the following parts: (i) identification of
metabolites linked to impaired performance of cells in culture based on published reports; (ii)
characterization of cell-line-specific effects of toxic metabolites known to be secreted to media during
cultivation; (iii) employment of a genome engineering-based approach to generate CHO cells with
increased growth and efficient nutrient and by-product metabolism.
The thesis starts with an Introduction to the use of CHO cells used for the production of
biopharmaceuticals and describes some of the challenges relevant to the work presented here.
Chapter 1 introduces Paper I, a bibliographic study that surveys metabolites reported to be cytotoxic,
growth inhibitory, depleting or accumulating, or that are markers of metabolic inefficiency during cell
cultivation. This review article is included in one of the CHO special issues published by Biotechnology
Journal in 2018.
Chapter 2 presents cell line engineering approaches for reprogramming the metabolism of CHO cells by
reducing nutrient utilization and, in consequence, by-product formation. First, nutrient metabolism,
specifically amino acid catabolism, was engineered using the CRISPR/Cas9 system for single and
combinatorial gene disruption in CHO cells followed by respective physiological studies. Paper II addresses
the disruption of nine target genes in these pathways in host cells and Paper III addresses the disruption of
two targets genes involved in branched-chain amino acid (BCAA) catabolism in CHO-S cells and in CHO-
1
S derived producer cells with reduced growth variation. Next, an attempt to increase cell growth and
resistance to induced cellular stress through the overexpression of a gene related to cofactor metabolism in
a recombinase-mediated cassette exchange (RMCE)-ready parental cell line is presented in Paper IV.
Chapter 3 features a study of the effect of metabolite additions in cell growth. Here, CHO host cells were
cultivated in small scale using basal medium supplemented with different concentrations of ammonium
chloride (NH4Cl).
To close this thesis, a summary of findings and contributions to the field are discussed in Conclusion and
future perspectives.
2
1. Introduction
The introduction aims to explain how the performance of Chinese hamster ovary cells (CHO) cells is
affected by intrinsic and extrinsic factors during small scale and industrially-relevant cultivations and why
and how it can be improved. But first, the role of CHO cells in the biopharmaceutical field is presented by
covering the background about biological drugs used as a treatment for severe diseases and how the
demands of the pharmaceutical market call for shorter production timelines without compromising patient
safety. Second, the characteristics of CHO cell factories are explored and compared with other well-studied
cell factories. Next, arriving at the main point of this thesis, a look into the metabolism of CHO cells is
provided, with a focus on nutrient metabolism and its by-products. Finally, a mention of the recent
advances in the technologies used to generate “omics” data and in the development of cell line engineering
tools is made. The stated subparts of this section highlight some of the main challenges encountered
throughout the more than 30 years of CHO cells as the workhorse for biopharmaceutical production and
set the stage for this thesis. Some of the points mentioned above are also described in Paper I.
1.1. CHO cells in the biopharmaceuticals market
Biopharmaceuticals such as recombinant therapeutic proteins are used to treat severe diseases. Examples of
recombinant therapeutic proteins and their indications include: monoclonal antibodies (mAbs) used to
treat some cancer types, autoimmune diseases and in transplantation, erythropoietin (EPO) to treat anemia,
and blood factors such as Factor VIII and Factor IX indicated to treat hemophilia [1]. Monoclonal
antibodies alone represent a large slice of all biopharmaceuticals, with a market value of $140 billion in 2013
[1], and with mAbs sales between 2014 and 2017 reaching $103.4 [2]. Recombinant therapeutic proteins are
in its majority produced in mammalian expression systems, from which CHO) cells have been the preferred
expression system used for the production of biological drugs for over 30 years [2–4]. Yet, the production
3
costs and the lengthy timelines to produce such therapies represent challenges to the field. Therefore,
optimization of the production pipeline is beneficial for both patients and the biopharmaceutical industry.
1.2. CHO cell factories
CHO cells were originally isolated by Puck in the late 1950s from Chinese hamster (Cricetulus griseus) [5]
and were used for the first time to express a biopharmaceutical drug, tissue plasminogen activator (tPA), in
1986 [3]. Meanwhile, several CHO cell lines have been generated and are commonly used as expression
systems in industrial bioprocesses. How the CHO cell lines CHO-K1, CHO-DG44, CHO-DXB11, and
CHO-S differ from each other has been described by F. Wurm [6] and others [7].
1.2.1. The advantages and disadvantages of using CHO cell factories to
produce recombinant therapeutic proteins
Mammalian expression systems have advantages and disadvantages to other cell factories. A small number
of biopharmaceutical drugs are expressed in bacterial or yeast cells [1,8], as well as insect and plant cells [1],
although these last two will not be brought to consideration in this discussion. Bacteria such as Escherichia
coli (E. coli) and yeast such as Saccharomyces cerevisiae or Pichia pastoris are very well studied cell factories
that have been used in industrial production of therapeutic proteins [1]. E. coli cell factories have the
advantages of easy set-up, fast growth and productivity, and available genome sequences as well as synthetic
biology tools [9]. However, E. coli do not secrete proteins and do not have the machinery to perform
complex post-translational modifications (PTMs) [10]. Yeast cell factories have also been studied
extensively [11] and are able to perform PTMs. However, the glycosylation patterns these cells generate
differ from those found in humans [12]. Glycosylation is the most important PTM, as it is essential for the
effector function of therapeutic proteins and for patient safety. Aberrant glycoforms can induce severe
immunogenic reactions and alter the pharmacokinetics of the biopharmaceutical [13]. Therefore, an
expression system that mimics the glycosylation patterns found in humans is required.
4
The mammalian cells used for therapeutic protein secretion are both of human and non-human origin.
Mammalian cells grow slower than microbial cells but are able to secrete most proteins, resulting in fewer
separation and purification steps in downstream processing. In addition, mammalian cells just like CHO
cells have been adapted to grow in suspension using serum-free chemically defined media, are easy to
transfect and can achieve high productivities [14]. However, human cell lines such as human embryonic
kidney 293 (HEK293) have the disadvantage of being susceptible to infection by viruses [14]. CHO cells, on
the other hand, have been shown to be free of genes responsible for viral entry [15,16]. Other examples of
non-human expression systems are baby hamster kidney cells (BHK21) and mouse myeloma cells (NS0 and
Sp2/0) [1,17,18]. However, in the end, the choice of the ideal host falls on the ability to perform human-like
PTMs [19], to be a safe host, and to be cultivated in suspension to allow easy scale-up [20,21]. CHO cells
are therefore the preferred host for recombinant protein production. In sum, the main advantages of using
CHO cells include the ability to be cultivated in suspension using chemically defined media, low
susceptibility to human viral infections, ability to perform PTMs compatible to those found in humans.
In addition to these advantages, there are many resources for engineering CHO cells. These include
established gene amplification methods and, in recent years, genomic sequences of Chinese hamster and
CHO cell lines became available [16,22–25], alongside other omics technologies used to characterize the
cells throughout the cell line development (CLD) process [26–28]. The number of genetic engineering tools,
including the Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)-based methods for
CHO cell engineering is also expanding [29–34]. Moreover, metabolic models are available [35,36] and
these can aid in identifying novel targets and to predict the effects of gene engineering.
A disadvantage of CHO cells genome is that their genome is unstable as they undergo frequent genomic
rearrangements, with a high rate of translocations [3,6,37] and changing copy number of transgenes [27],
which can affect productivity and cell metabolism. Nevertheless, CHO remains a popular expression
5
system. A high percentage of biological drugs are being produced in CHO, proving its safety; because of
this new drugs expressed in CHO cells are likely to be approved by the regulatory authorities.
1.2.2. Cell line development
To obtain a cell line suitable to be used in a bioprocess, several CLD steps are necessary. These include 1)
design and cloning of the vector encoding the recombinant protein of interest, 2) transfection for delivery
of the transgene into the cells 3) selection pressure and expansion of cells 4) generation and high-
throughput screening for high producer cells 5) characterization of high-producer clone in small scale
cultivations. Altogether, these steps can take 6-12 months [3], or even longer if the product is a difficult-to-
express protein. These steps include screening rounds of a large number of clones [38]. These are required
to identify the high producing cells that are robust during upstream development since existing
bioprocesses are required to be adapted to each selected clone [39]. Physiological characterization of these
clones is required for media and feed optimization able to sustain cell growth and productivity during a
fed-batch process. Often, the cells exhibit an inefficient metabolism and perform poorly during cultivation.
Therefore, knowledge of the inner workings of the cell is necessary to rationally design a producer cell line
with enhanced performance during culture.
1.3. Nutrient and by-product metabolism of CHO cells
The inefficient metabolism of CHO cell is characterized by high uptake rates of substrates glucose and
glutamine [40], used as carbon and nitrogen sources. These lead to the formation of lactate and ammonia
that are the main metabolic by-products of the mammalian metabolism known to be toxic and inhibitor of
cell growth [41–43].
Paper I reports on a number of additional metabolites related to cell growth inhibition, apoptotic cell death,
and that accumulate or deplete during cell cultivation [44]. These metabolites represent a waste of carbon
that is diverted from the main metabolic pathways. Amino acid catabolism, in particular, forms toxic
6
intermediates and diverts amino acids from recombinant protein production [45] while the majority of the
remaining compounds reported in Paper I are reporters of or related to glycolysis and the TCA cycle. The
following brief description of CHO cell metabolism focuses on these pathways.
1.3.1. Glucose metabolism and by-product formation
Glucose, the main carbon source for CHO cells, is supplied in media and feeds used in industrial
cultivations. In cells with normal metabolism growing in aerobic conditions, glucose enters glycolysis to
form pyruvate, which is transported into the mitochondrial matrix where it enters the TCA cycle, followed
by oxidative phosphorylation pathway. In CHO cells that have an inefficient metabolism and in aerobic
conditions, glucose is taken up by the cell in high rates to form pyruvate, which is converted into lactate
along with the oxidation of reduced nicotinamide adenine dinucleotide (NADH) to oxidized nicotinamide
adenine dinucleotide (NAD+) by the action of the enzyme lactate dehydrogenase A (LdhA), instead of
continuing towards a more oxidative metabolism via the TCA cycle. This represents a diversion carbon flux
away from the TCA cycle and results in a less efficient form of energy production. This phenomenon, also
observed in cancer cells, is called the Warburg effect [44].
Previous studies reveal that 35% [45] and 70% [46] of the metabolized glucose in CHO cells is used in the
formation of lactate. The effects of lactate to the cell have been reviewed by other research groups [54–57].
The negative effects of lactate accumulation during cultivation are inhibition of cell growth, apoptosis, and
reduced productivity of recombinant therapeutic products, due to changes in pH and osmolality [41,42,45–
49]. It has been observed during mammalian cell cultivations that lactate is produced from glucose in the
initial phases of the cultivation and consumed in the later phases. However, co-consumption of lactate and
glucose has also been reported [52,53].
Glucose 6-phosphate which is formed upon glucose entry into the cell is also used in the pentose phosphate
pathway (PPP). The products of this pathway are reduced nicotinamide adenine dinucleotide phosphate
7
(NADPH), pentoses and precursors of nucleotide synthesis. These are linked to cell growth as the generated
nucleotides are used for DNA synthesis and replication and resistance to oxidative stress as NADPH is a
cofactor for enzymes responsible for reduction of glutathione, the main scavenger of reactive species in the
cell [50]. NADPH is formed in the first irreversible step of the oxidative phase of the PPP and is catalyzed
by glucose-6-phosphate dehydrogenase (G6pd), while pentoses and precursors of nucleotide synthesis form
during the non-oxidative phase. Because of this, G6pd was chosen to be overexpressed in CHO cells (Paper
IV).
1.3.2. Amino acid catabolism and by-product formation
Amino acids contribute to biomass formation and are used for protein synthesis [51,52]. The amino acid
catabolic reactions lead to the formation of TCA cycle intermediates that support an oxidative metabolism.
High availability of amino acids can lead to accumulation of ammonia and other toxic intermediates.
Glutamine is the main cause of ammonium formation, that is produced by enzymatic transamination of
glutamine to form glutamate that is deaminated to α-ketoglutarate. Accumulation of ammonia in the cell
culture media during biopharmaceutical production can result in variation in product quality attributes
[43,53], reduced productivity, decreased cell growth and cell death [42,47,54,55].
The catabolism of certain amino acids leads to the formation of additional toxic intermediates. This is the
case of phenylalanine, tyrosine, tryptophan, methionine, leucine, serine, threonine, and glycine. The
intermediates they form cause growth inhibition [56]. But once the initial concentrations are controlled,
the levels of the toxic metabolites remain low under certain cultivation conditions [56]. Furthermore, there
are reports of CHO cells with bottlenecks in the TCA cycle in response to the addition of feed. For instance,
the buildup of the intermediates citrate, succinate, fumarate, and malate in the cell culture medium after
the addition of a feed containing pyruvate and amino acids (aspartate, asparagine, and glutamate) is linked
to growth limitation [57]. Media and feeds need to be prepared in accordance with the metabolic needs of
the cell, which can be achieved through the study of cell physiology as clones are selected.
8
1.4. Engineering of CHO cells using synthetic biology tools
Synthetic biology tools enable gene insertion, disruption, and altering the gene expression levels to obtain
cells with enhanced properties. These are widely used in cell line development to perform gene
amplification, to obtain cells with extended longevity by engineering apoptotic pathways, to debottleneck
the secretory pathway, to get specific glycosylation patterns and to improve cell metabolism. Here, a
description of the genome editing tools used in the experimental work for this thesis (Chapter 2) is given,
followed by a brief mention of cell engineering efforts carried out by other researchers.
1.4.1. Genome editing tools
Genome editing tools such as Zinc-finger nucleases (ZFN) and Transcription activator-like effector
nucleases (TALENs) were used in the past to perform genome editing [58]. As reports of CRISPR/CRISPR-
associated protein 9 (Cas9)-mediated genome editing in mammalian cells emerged [59–62], these were also
adapted in the field of cell line engineering for biopharmaceutical production. CRISPR are short repeating
DNA sequences that are part of a bacterial immune system. The endonuclease Cas9 is responsible for
cleaving the DNA. The CRISPR/Cas9 system can be used to induce double-strand breaks on a target DNA
sequence complementary to the single guide RNA (sgRNA). The sgRNA is a 20 nucleotide sequence,
complementary to the target and flanked by NGG, where N is any nucleotide (also referred as the
protospacer adjacent motif (PAM) sequence) [63,64]. Both sgRNA and Cas9 can be delivered to the cell via
virus transduction or plasmid transfection. The endonuclease can be delivered in a plasmid as RNA or
protein. The sgRNA guides Cas9 to the target site where it will disrupt the DNA sequence. Then the cell’s
DNA repair machinery repairs the double strand break using either pathway – the error-prone non-
homologous end joining (NHEJ) or the homology-directed repair (HDR) pathway. The NHEJ repair leads
to insertion and deletion (indel) mutations that can vary in size. The inserted sequences used by NHEJ can
be random DNA stretches present inside the cell or use a donor template. NHEJ is mostly used to disrupt
9
or knock out genes, but can also be used for targeted gene insertion. HDR repair has low efficiency in
mammalian cells [61]. HDR can be used for specific integration of heterologous sequences. A donor
template that the cell will use during repair of the DNA break must be delivered into the cell alongside the
sgRNA and Cas9. Although CRISPR-based engineering is highly efficient, off-target effects can occur.
Mismatches placed within 8-12 nucleotides upstream the PAM influence the cleavage by Cas9 [61]. Despite
the chance for off-targeting, CRISPR allows for fast and specific gene editing. Examples of other applications
of CRISPR include screenings using knock-out libraries, gene regulation using a deactivated version of
Cas9. However, the legal situation of CRISPR makes it difficult to adopt in an industrial setting [65].
Recombinase-mediated cassette exchange (RMCE) is a two-step method used for targeted integration [66].
First, a landing pad carrying a target site for recombinase activity, and a reporter gene, used to assess the
expression levels after integration of the landing pad, are inserted in the genome either randomly or at pre-
identified hot-spot suitable for gene expression. Second, a recombinase catalyzes the exchange of the
reporter gene by the gene of interest carried by the donor cassette. An example of this system is the Cre/Lox
system [67].
1.4.2. Engineering CHO cells for improved protein production and
improved metabolism
Two systems for gene amplification commonly in use are dihydrofolate reductase (DHFR) where
methotrexate (MTX) is used to inhibit DHFR, and glutamine synthetase (GS) developed more recently,
where methionine sulfoximine (MSX) is used as selection pressure to inhibit GS [68,69]. The GS system is
advantageous since it allows for the reduction of by-product formation. The endogenous gene encoding GS
first is knocked out and reintroduced along with the vector for recombinant protein expression, after which
ammonia along with glutamate are utilized to form of glutamine.
10
Several reviews [70–72] gather a number of cell line engineering approaches carried out in CHO cells. Some
of the examples of the use of metabolic engineering to reduce the levels of lactate secretion include the
downregulation of LdhA using RNAi technology, where lactate levels were reduced without affecting cell
growth and productivity [73]. Lactate secretion was also reduced through downregulation of LdhA and
pyruvate dehydrogenase kinase (Pdhk) isoenzymes 1, 2, and 3 in a CHO cell line producing antibody [74].
However, the full disruption of LdhA using ZFNs in cells in combination with Pdhk 1, 2, and 3
downregulation was lethal [74,75]. Using a different strategy, the overexpression of Aralar1, part of the
malate–aspartate shuttle (MAS), it was possible to induce a shift in the metabolism from lactate production
to lactate consumption [76]. Other approaches included overexpression of the GLUT5 transporter [77],
yeast pyruvate carboxylase [77–79] and malate dehydrogenase II (MDHII) [80]. Very recently, the amino
acid metabolism was engineered with the aim to reduce the levels of toxic catabolic intermediates [81].
11
Chapter 1 - Overview of nutrients and growth inhibitory metabolites affecting CHO cell metabolism This thesis starts with a literature review of nutrient and by-product metabolism of CHO cells. Paper I
presents literature referent to more than 45 metabolites reported to affect cultivation performance and/or
be markers of the metabolic status of the cell [82]. Besides describing metabolites that affect cell growth
negatively, the review also focuses on the amino acid and glutathione metabolism. These two main pathways
were subjected to further scrutiny due to their relevance for recombinant protein expression in mammalian
host cell factories seen in two ways: amino acids are the building blocks for protein synthesis [51,52], and
high expression of recombinant proteins leads to cellular stresses [83] that can potentially be scavenged by
glutathione. Additionally, this review article covers the majority of the topics included in this thesis work.
These include a description of the central metabolism and the appearance of metabolic by-products, cell
line engineering approaches employed in CHO cells, and perspectives for media and feed design and
optimization are also described. Thus, by reading the review, the reader will understand the framework of
this PhD project.
12
Paper I – Impact of CHO Metabolism on Cell Growth and Protein Production: An Overview of Toxic and Inhibiting Metabolites and Nutrients
13
Amino Acid Metabolism www.biotechnology-journal.com
REVIEW
Impact of CHO Metabolism on Cell Growth and ProteinProduction: An Overview of Toxic and InhibitingMetabolites and Nutrients
Sara Pereira, Helene Faustrup Kildegaard, and Mikael Rørdam Andersen*
For over three decades, Chinese hamster ovary (CHO) cells have been thechosen expression platform for the production of therapeutic proteins withcomplex post-translational modifications. However, the metabolism of thesecells is far from perfect and optimized, and requires substantial know howand process optimization and monitoring to perform efficiently. One of themain reasons for this is the production and accumulation of toxic andgrowth-inhibiting metabolites during culture. Lactate and ammonium are themost known, but many more have been identified. In this review, an overviewof metabolites that deplete and accumulate throughout the course ofcultivations with toxic and growth inhibitory effects to the cells is presented.Further, an overview of the CHO metabolism with emphasis to metabolicpathways of amino acids, glutathione (GSH), and related compounds whichhave growth-inhibiting and/or toxic effect on the cells is provided. Addition-ally, relevant publications which describe the applications of metabolomics asa powerful tool for revealing which reactions occur in the cell under certainconditions are surveyed and growth-inhibiting and toxic metabolites areidentified. Also, a number of resources that describe the cellular mechanismsof CHO and are available on-line are presented. Finally, the application ofthis knowledge for bioprocess and medium development and cell lineengineering is discussed.
1. Introduction
Chinese hamster ovary (CHO) cells are the mammalian host ofchoice for the production of recombinant biological compounds.The market of therapeutic recombinant proteins presentscumulative sales values, ranging between $107 to $140 billionfrom 2010 to 2013.[1] The first drug produced in this expression
S. Pereira, Dr. H. F. KildegaardThe Novo Nordisk Foundation Center for BiosustainabilityTechnical University of Denmark2800 Kgs. Lyngby, Denmark
Prof. M. R. AndersenDepartment of Biotechnology and Biomedicine TechnicalUniversity of Denmark2800 Kgs. Lyngby, DenmarkE-mail: [email protected]
The ORCID identification number(s) for the author(s) of this articlecan be found under https://doi.org/10.1002/biot.201700499.
DOI: 10.1002/biot.201700499
Biotechnol. J. 2018, 1700499 © 21700499 (1 of 13)14
system was tissue plasminogen activator(tPA), which reached the marked in 1987.[1]
Examples of products expressed in CHOcells include erythropoietin (EPO) indi-cated for the treatment of severe anemia,coagulation factors as factor IX used as atherapeutic in hemophilia, interferon usedfor treating multiple sclerosis and mono-clonal antibodies (mAbs) with the indica-tion for treating Crohn’s disease, differentlymphomas, and cancers (e.g., breastand gastric cancer).[1] From the biologicaldrugs approved between 2006 and 2010,about 55% were produced in mammaliancells[2,3]: from those between 2010 to themiddle of 2014, 60% of the recombinanttherapeutic proteins were also produced inmammalian cells. This shows an increas-ing trend which favors the use of expres-sion systems of mammalian origin. Underthe latter-mentioned time period, 33% oftotal approvals were for drugs expressed inCHO cells.[1] In retrospect, CHO cells haveshown to be safe hosts and therefore, morelikely to obtain approval for novel thera-peutic proteins manufactured in this cellplatform by the regulatory agencies.
Themain advantages of using CHO cellscompared to other microbial or mamma-lian cells[4] include the ability of these
cells to perform post-translational modifications similar tothose found in human proteins, such as glycosylation, which isconsidered to be a critical quality attribute. The presence of anaberrant glycan profile will decrease the efficacy,[5] affects theprotein drug pharmacokinetics,[6] and alters biological proper-ties.[7–10] In addition, CHO cells have been demonstrated todisplay reduced susceptibility to human viral infections,[11]
which represents an additional advantage over cell lines ofhuman origin. Genomic and transcriptomic analysis of CHO-K1showed that genes encoding for viral entry receptors, as well asother genes required for a successful viral infection, are absentor not expressed in the cell line.[12] CHO cells can grow inchemically defined medium, which reduces the chances forbatch-to-batch variation and have the ability to be cultured insuspension, to facilitate the scale-up of the bioprocess.[13]
The metabolism of CHO cells, characterized by high uptakerates of substrates used as carbon and nitrogen sources,[14–16] isgenerally inefficient and suboptimal. The nutrients supplied in
018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advancedsciencenews.com www.biotechnology-journal.com
the media and feeds at certain concentrations may lead to theaccumulation of metabolites, intermediates, and by-products.This indicates the existence of metabolic bottlenecks in keypathways and inefficient flux distribution. Furthermore, theseaccumulating compounds may decrease cell growth, productiv-ity,[17,18] and protein quality.[19,20] As monitoring the metabolitesand changing the related pathways has the potential to improverecombinant protein production in CHO cell culture, these 45compounds are presented in a tabulated form (Table 1). Thetable contains the main reports on the effects of the individualmetabolites, and provides helpful primary references on, forexample, the effect of concentrations of individual amino acidsand sugars.
Furthermore, for some of these compounds, the reportedeffects are complex and surprising. To provide additional detailand context of these metabolites, we present overviews of thepathways for generation and consumption of the main toxicand inhibiting metabolites; glycolysis, the tricarboxylic acid(TCA) cycle, and amino acid metabolism, as well as glutathionemetabolism. Moreover, some lipids have been shown to affectgrowth, for which we also discuss the details. In addition, wepresent methods and methodologies, which can be used todecrease or remove the presence of such toxic and inhibitingmetabolites.
2. An Overview of CHO Metabolism
CHO cells have an inefficient metabolism, which is character-ized by high uptake rates of substrates used as carbon andnitrogen sources (e.g., glucose and glutamine[14]). Thesubstrates are not fully used for production of biomass orrecombinant proteins: thus, based on reports, 35%[21] and70%[22] of glucose can be diverted into the formation of wasteproducts, which impact the cell culture performance.[15,21,22]
Examples of toxic or inhibiting metabolites can be foundthroughout metabolism.[17–19] Table 1 summarizes compoundsthat are reported to correlate with cell growth inhibition,apoptosis and/or have additional negative effect in culture dueto accumulation or depletion. These metabolites are reportersof metabolic inefficiency and represent a waste of carbondiverting from the main metabolic pathways. Additionally,some of the listed metabolites function as alternative redoxsinks (sorbitol, threitol, and glycerol), while others (aminoacids) are catabolized instead of contributing directly torecombinant protein production and lead to the formation oftoxic intermediates. In this review, we have chosen to focuson the main pathways where the majority of these compoundshave been reported; glycolysis, the TCA cycle, amino acidmetabolism, and glutathione (GSH) metabolism.
2.1. Central Metabolism: Nutrient Uptake and By-ProductFormation
The main carbon source in CHO cells is glucose, which issupplied in media and feeds that are used in batch and fed-batchbioprocesses. Glucose is taken up by the cell at high rates andphosphorylated to glucose-6-phosphate (G6P), and used in
Biotechnol. J. 2018, 1700499 1700499 (215
glycolysis to form adenosine triphosphate (ATP), reducednicotinamide adenine dinucleotide (NADH) and pyruvate.Instead of proceeding to the full oxidation of glucose in aerobicconditions, pyruvate is converted into lactate along with theoxidation of NADH to oxidized nicotinamide adenine dinucleo-tide (NADþ) by the action of lactate dehydrogenase A (LdhA).This represents a diversion of a flux of carbon away from theTCA cycle, to lessen the energy production and decreaseproduction of important C4-6 precursors which are requiredfor biomass formation (Figure 1). This phenomenon is alsoobserved in cancer cells and called the Warburg effect.[23] For abetter understanding of the relation between NADþ/NADHand glycolysis, we suggest reading the review by J. Locasaleand L. Canteley.[24]
As seen in Figure 1, lactate is one of the main toxicmetabolites found in central metabolism. The consequences ofthe accumulation of lactate inmammalian cell culture have beenwidely mentioned in the literature (Table 1). Reports have shownthat lactate inhibits cell growth, induces apoptosis and reducesproductivity of recombinant therapeutic products, due tochanges in pH and osmolality.[17,18,25–27] Many other studieshave explored the underlying motives for such phenotype.[28–34]
In a bioprocess, two distinct phases of lactate metabolism havebeen described; initially, glucose consumption is accompaniedby lactate production while, in later phases, the consumptionof lactate is observed, although simultaneous consumptionof glucose and lactate has likewise been described.[29,34] It isimportant to note the link between lactate consumptionphenotype and increased productivity,[35] as well as the metabolicshift from lactate production to lactate consumption as a markerof metabolic efficiency.[33] Reports show that initial lactatesupplementation can induce a shift from high to low glycolyticflux, even in the presence of high glucose concentration.[36]
When lactate is added into the medium along with pyruvate,glucose uptake rate was reduced by 50%.[37] For additional detail,the role of lactate has been extensively reviewed. We suggest thereader examines a set of particularly excellent reviews andresearch papers.[15,16,28,38]
Alternative fates for carbon have also been suggested sinceglucose can also be converted into glycerol along with oxidationof NADH to NADþ, as well as sorbitol and threitol – in bothcases, accompanied by the oxidation of reduced nicotinamideadenine dinucleotide phosphate (NADPH) to oxidized nicotin-amide adenine dinucleotide phosphate (NADPþ). These com-pounds are formed from glycolytic intermediates andaccumulate both intracellularly and extracellularly in thetransition of the exponential to stationary phase of culture,after the addition of feed containing depleted nutrients(Table 1).[26,39] Additionally, G6P is shunted away from glycolysisto enter the pentose phosphate pathway (PPP) where the sugarprecursors are required for the synthesis of nucleotides,NADPH, and glycolytic intermediates are produced. Thepresence of nucleosides and nucleotides – adenosine, adenosinemonophosphate (AMP), adenosine diphosphate (ADP), guano-sine diphosphate (GDP), and guanosine monophosphate (GMP)– in the culture medium at low concentrations (1mM) hasshown to arrest cell growth and to contribute to proteinproduction.[40] However, as specified in Table 1, adenosine, ADP,and AMP have been reported to be cytotoxic. In particular, the
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheimof 13)
Table 1. Metabolites with growth inhibitory effects, apoptosis inducers and waste products in CHO cell cultivations.
Metabolite
KEGGcompound
no. Comment/effect Pathways Reference
Central/energy metabolism
Acetate C00033 Formed from acetyl-CoA starts to build up after the onset of the stationary
phase.
Pyruvate metabolism [43]
Citrate C00158 Accumulation indicates TCA cycle truncation; increased along with
alanine, upon supplementation of the medium with growth limiting
nutrients (aspartate, asparagine, glutamate and pyruvate) in glutamine
synthetase (GS) expression system.
TCA cycle; Alanine, aspartate and
glutamate metabolism
[26]
TCA cycle and fatty acid/lipid biosynthesis intermediates; related to
mitchondria/cell redox status. In fed-batch, appeared in the medium
during culture in response to feed addition, representing changes in the
mitochondria and changes in C-fluxes to alternative fates;
[39]
Secreted during exponential phase [43]
Fructose C00085 Increased intracellularly after addition of feed containing glucose; build up
may occur in connection to sorbitol.
Sorbitol pathway [33]
Fumarate C00122 Secreted during exponential phase TCA cycle; Alanine, aspartate and
glutamate metabolism
[43]
Lactate C00186 Inhibits cell growth; Accumulation results in lowered pH and changes in
osmolarity due to the presence of base, added to counter the effects of
decreased pH from lactate formation.
Pyruvate metabolism [17,27]
Reduces cell growth due to acidification; reported to inhibit cell growth of
murine hybridoma cell lines, in cultivations that do not employ pH
control.
[18]
When present in the cell culture medium, reduces growth and induces cell
death in baby hamster kidney cells.
[25]
Accumulation in medium is linked to growth phase of culture. [26]
Malate C00149 Accumulated extracellularly; linked to aspartate supplied in the medium
and to enzymatic bottleneck at malate dehydrogenase II in TCA cycle.
TCA cycle; Pyruvate metabolism [94]
Accumulated in medium during a fed-batch culture in response to feed
addition; represents changes in the mitochondria and changes in C-fluxes
to alternative fates;
[39]
Secreted during exponential phase [43]
Sorbitol C00794 Released into the medium and represents carbon losses to the cell;
alternative redox sink for the cell; linked to the cellular redox state
(NADPH/NADPþ) and inform of cell well-being during culture.
Fructose and mannose metabolism;
Galactose metabolism
[26]
Linked to the cellular redox state (NADPH/NADPþ). [39]
Builds up intracellularly in cells growing media containing high and low
copper, related to metabolic shift from lactate production to lactate
consumption phenotype
[33]
Succinate C00042 Secreted during exponential phase TCA cycle, Oxidative
phosphorylation, Alanine, aspartate
and glutamate metabolism
[43]
Threitol C16884 Linked to the cellular redox state (NADPH/NADPþ) and is an alternative
redox sink for the cell.
[39]
Amino acid metabolism
Alanine C00041 Accumulation in the medium results in, negative effect for cell growth;
Inhibits pyruvate kinase and TCA pathway; potential source of ammonia
Alanine, Aspartate and Glutamate
metabolism
[53]
Produced during culture. Formed by transamination from pyruvate; [39]
Accumulated in the medium along with glycine and citrate in the
transition of exponential phase to stationary
[26]
Produced from pyruvate at late stages of culture. [42]
Accumulated in the medium [43]
(Continued)
www.advancedsciencenews.com www.biotechnology-journal.com
Biotechnol. J. 2018, 1700499 © 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim1700499 (3 of 13)16
Table 1. (Continued)
Metabolite
KEGGcompound
no. Comment/effect Pathways Reference
Ammonia C00014 Decreases specific cell growth rate, increases consumption rates of
glucose and glutamine and decreases antibody product titer in hybridoma
cells.
Amino acid metabolism [18]
Affects intracellular pH, cell growth and recombinant protein productivity,
and product glycosylation.
[100]
Reduces growth rates and maximal cell densities, changes metabolic
rates, affects protein processing in mammalian cells.
[20]
Production of ammonia and alanine is linked to the consumption of
asparagine and glutamine in a GS-CHO cell line.
[53]
Asparagine C00152 Asparagine consumption has been correlated with accumulation of
ammonia and alanine.
Alanine, Aspartate and Glutamate
metabolism
[59]
Highest consumed amino acid in GS-CHO cells treated with butyrate. [43]
Glutamine C00064 Degradation of glutamine generates ammonium and glutamate. Glutamate metabolism [48,100]
Extracellular supply of glutamine and pyruvate are sources of lactate
formation.
[101]
Glycine C00037 Product of serine catabolism Glycine, Serine and Threonine
metabolism
[43]
Accumulation in the medium indicates a positive effect. [53]
Accumulated in the medium along with alanine, in the transition of
exponential to stationary phase.
[26]
Recommended to keep concentration below 0.5–1mM in fed-batch
process due to growth inhibition.
[57,58]
Accumulation of glycine is beneficial for the cells due to its role in GSH
biosynthesis.
[59]
Leucine C00123 Recommended to keep concentration below 0.5–1mM in fed-batch
process due to growth inhibition.
Valine, Leucine and Isoleucine
metabolism
[57,58]
Lysine C00047 Oversupplied nutrient; accumulates in the medium during death phase. Lysine metabolism [56]
Methionine C00073 Recommended to keep concentration below 0.5–1mM in fed-batch
process due to growth inhibition.
Cysteine and methionine
metabolism
[57,58]
Phenylalanine C00079 Recommended to keep concentration below 0.5–1mM in fed-batch
process due to growth inhibition.
Phenylalanine, Tyrosine and
Tryptophan metabolism
[57,58]
Serine C00065 Highly consumed amino acid in GS-CHO cells treated with butyrate Glycine, Serine and Threonine
metabolism
[43]
Recommended to keep concentration below 0.5–1mM in fed-batch
process due to growth inhibition.
[57,58]
Threonine C00188 Recommended to keep concentration below 0.5–1mM in fed-batch
process due to growth inhibition.
Glycine, Serine and Threonine
metabolism
[57,58]
Tryptophan C00078 Recommended to keep concentration below 0.5–1mM in fed-batch
process due to growth inhibition.
Phenylalanine, Tyrosine and
Tryptophan metabolism
[57,58]
Tyrosine C00082 Recommended to keep concentration below 0.5–1mM in fed-batch
process due to growth inhibition.
Phenylalanine, Tyrosine and
Tryptophan metabolism
[57,58]
Amino acid derivatives
Dimethylarginine (DARG) C03626 Accumulates in the media over culture time; linked to excessive supply of
Arginine.
Arginine metabolism [59]
Induces apoptosis in endothelial cells due to intracellular oxidant
production and related to p38 mitogen-activated protein kinase (MAPK)/
caspase-3-dependent signaling pathway.
[102]
Known to induce apoptosis in human endothelial cells, by increasing the
formation of intracellular reactive oxygen species.
[103]
(Continued)
www.advancedsciencenews.com www.biotechnology-journal.com
Biotechnol. J. 2018, 1700499 © 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim1700499 (4 of 13)17
Table 1. (Continued)
Metabolite
KEGGcompound
no. Comment/effect Pathways Reference
Formate C00058 Product of serine catabolism Glycine, serine and threonine
metabolism
[43]
Product of serine catabolism, growth inhibitory at concentrations between
4–10mM.
[104]
Metabolic by-product; recommended to keep concentration below 2mM
in fed-batch process due to growth inhibition.
[57,58]
Homocysteine C00155 Metabolic by-product; recommended to keep concentration below
0.5–1mM in fed-batch process due to growth inhibition.
Cysteine and Methionine
metabolism
[57,58]
Indole-3-carboxylate C19837 Metabolic by-product; recommended to keep concentration below 1mM
in fed-batch process due to growth inhibition.
Tryptophan metabolism [57,58]
Indolelactate C02043 Metabolic by-product; recommended to keep concentration below 3mM
in fed-batch process due to growth inhibition.
Tryptophan metabolism [57,58]
Isobutyrate C02632 Accumulated in culture as result of breakdown of the branched-chain
amino acids.
Valine metabolism [43]
Isovalerate C08262 Accumulated in culture as result of breakdown of the branched-chain
amino acids.
Leucine metabolism [43]
Metabolic by-product; recommended to keep concentration below 1mM
in fed-batch process due to growth inhibition.
[57,58]
Methylglyoxal C00546 Detrimental to cultured cells; D-lactic acid is the end product of
methylglyoxal metabolism in mammalian cells;
Glycine, serine and threonine
metabolism; Pyruvate metabolism
[105]
Inhibits cell growth and induces apoptosis when added to the medium in
hybridoma cell cultures; By-product formed through non-enzymatic
decomposition of dihydroxyacetone phosphate and glyceraldehyde-3-
phophate;
Glycolysis [65,106]
Ornithine C00077 Present in death phase of culture and associated with apoptosis. Arginine and proline metabolism [56,107]
Phenyllactate C05607 Metabolic by-product; recommended to keep concentration below 1mM
in fed-batch process due to growth inhibition.
Phenylalanine metabolism [57,58]
2-hydroxybutyric acid C05984 Metabolic by-product; recommended to keep concentration below
0.5–1mM in fed-batch process due to growth inhibition.
Cysteine and methionine
metabolism
[57,58]
3-(4-hydroxyphenyl)lactate C03672 Metabolic by-product; recommended to keep concentration below
0.5–1mM in fed-batch process due to growth inhibition.
Phenylalanine, Tyrosine and
Tryptophan metabolism
[57,58]
4-hydroxyphenylpyruvate C01179 Metabolic by-product; recommended to keep concentration below 1mM
in fed-batch process due to growth inhibition.
Phenylalanine, Tyrosine and
Tryptophan metabolism
[57,58]
Nucleotide metabolism
Adenosine C00212 Cytotoxic: induces apoptosis in cells of the immune system, nervous
system and endothelium. Results in increased metabolic rates.
Purine metabolism; Signaling
pathways
[40]
ADP C00008 Arrests cell cycle in G1 in CHO cells overexpressing p27, a cyclin-
dependent kinase (CDK) inhibitor, and increased Secreted embryonic
alkaline phosphatase (SEAP) specific productivity.
Purine metabolism; Oxidative
phosphorylation
[108]
Results in increased metabolic rates. [40]
Adenosine
monophosphate (AMP)
C00020 Cytotoxic: induces apoptosis when added to cell culture medium at 2mM,
in at lower concentrations (1mM) arrests cell growth and increases
productivity. Results in increased metabolic rates.
Purine metabolism; Signaling
pathways
[40]
Arrests cell cycle in G1 in CHO cells overexpressing p27, a CDK inhibitor,
and increased SEAP specific productivity.
[108]
Addition to fresh CHO mAb cultures lead to apoptosis. [41]
Adenosine triphosphate
(ATP)
C00002 Cytotoxic: induces apoptosis in cells of the immune system, nervous
system and endothelium. Increased metabolic capacity of the cell.
Purine metabolism; Oxidative
phosphorylation
[40]
Arrests cell cycle in G1 in CHO cells overexpressing p27, a CDK inhibitor,
and increased SEAP specific productivity.
[108]
(Continued)
www.advancedsciencenews.com www.biotechnology-journal.com
Biotechnol. J. 2018, 1700499 © 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim1700499 (5 of 13)18
Table 1. (Continued)
Metabolite
KEGGcompound
no. Comment/effect Pathways Reference
Guanosine diphosphate
(GDP)
C00035 Leads to cell growth arrest. Purine metabolism, Signaling
pathways
[40]
Guanosine
monophosphate (GMP)
C00144 Added to culture medium and decreased cell growth. This effect was
shown not to be cell line dependent. May improve protein production
after arresting cell growth.
Purine metabolism, Signaling
pathways
[40,41]
Addition to fresh CHO mAb cultures lead to apoptosis.
Lipid metabolism
Choline phosphate
(PCHO)
C00588 Depleting over time (144h) in fed-batch cultivation; linked to the build-up
of extracellular G3PC and to cell growth limitation.
Glycerophospholipid metabolism [59]
By-product of choline, builds up after 72 h of cultivation. [43]
Glycerol C00116 Possibly formed from glycerol-3 phosphate; Released into the medium
and represent carbon losses to the cell; Linked to the cellular redox state
(NADH/NADþ) and informs of cell well-being during culture.
Glycerolipid metabolism [26,42]
Glycerol accumulated over culture time, as a result of branching
from glycolysis at dihydroxyacetone-phosphate (DHAP) with NADH
oxidation.
[43]
Alternative redox sink and related to mitochondria/cell redox status [39,42]
Glycerol-3-phosphate C00093 Builds up intracellular concentration and these changes are related to
phospholipid synthesis and cell growth.
Glycerolipid metabolism;
Glycerophospholipid metabolism
[26,42]
Glycero-3-phospho-
choline (G3PC)
C00670 Builds-up over time as intracellular precursors of PE and PC deplete;
linked to cell growth limitation.
Glycerophospholipid metabolism [59]
By-product of choline, builds up after 72 h of cultivation. [43]
Redox metabolites
GSSG C00127 Addition to fresh CHO mAb cultures lead to apoptosis. Linked to
oxidative stress; potential growth-limiting factor.
Glutathione metabolism [41]
Accumulates extracellularly towards the end of the culture. [59]
CDK, cyclin-dependent kinase; GS, glutamine synthetase; G3PC, Glycero-3-phosphocholine; PC, Phosphatidylcholine; PE, Phosphatidylethanolamines; SEAP, Secretedembryonic alkaline phosphatase.
www.advancedsciencenews.com www.biotechnology-journal.com
extracellular concentrations of AMP as high as 2mM becomecytotoxic[40] while both AMP and GMP have been correlated withearly apoptotic events in CHO cells.[41]
An oxidative metabolism is characterized by the channelingof the carbon molecules from glycolysis into the TCA cycle.Intermediates of the TCA cycle (citrate, succinate, fumarate andmalate) accumulate during culture phases, which indicates abottleneck (Table 1) (reviewed by Dickson[42]). These intermedi-ates were observed to build up in the cell culture mediumafter the addition of a feed, which contains pyruvate and aminoacids (aspartate, asparagine, and glutamate), and has been linkedto growth limitation.[26,39,43]
2.2. Amino Acid Metabolism in CHO Cell Culture
In CHO cell culture, amino acids are supplied in the growthmedium and/or produced via biosynthetic pathways. These arerequired to support cellular functions, such as cell growth, andutilized as building blocks for protein synthesis.[44,45] Through
Biotechnol. J. 2018, 1700499 1700499 (619
the catabolism of amino acids, the cell can utilize the carbonbackbones for the formation of TCA cycle intermediates, whichare to be used in the central metabolic pathways (Figure 1).When these are supplied in excess, a wasteful cellularmetabolism will lead to the formation of by-products, inparticular ammonium. Ammonium is mainly formed fromthe breakdown of glutamine. The catabolism of several aminoacids also leads to the formation of ammonium. This occursvia transamination reaction as an amino group is transferredto α-ketoglutarate and forming glutamate – that, in its turn, isdeaminated to release ammonium ion, NH4þ.[46] Amino acids,such as Serine and Threonine, can undergo direct deamina-tion.[46] Ammonia has a negative impact on the product qualityattributes when it accumulates and, similarly, affects productiv-ity and cell growth.[18–20,25,47,48] The mechanism by whichammonia affects growth is still not fully understood. It has beenreported that the increasing concentration of ammonia modifiesthe electrochemical gradient and acidifies the intracellularmilieu, which disrupts enzymatic activity and leads to apopto-sis.[18,49] Cell growth is inhibited in mammalian cell lines by
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheimof 13)
Figure 1. Schematic of main biosynthetic and catabolic pathways of CHO cells linked to production of toxic or inhibiting compounds. See Table 1 fordetails on individual metabolites.
www.advancedsciencenews.com www.biotechnology-journal.com
ammonia concentrations ranging from 1.8 to 33mM.[50] ForCHO cells, cell growth inhibition has been reported for anammonia concentration of 5.1mM[25] and a reduction of 50% ofgrowth was observed for ammonia concentrations were above8mM.[51] In an additional report, apoptotic cell death was notdetected when CHO cells were exposed to 50mM of ammoniumchloride upon engineering of apoptotic genes.[52] Therefore, in achemically-defined medium, the initial amino acid concen-trations should be well controlled and adjusted to the cell’sspecific metabolic requirements, based on prior studies of thecell line.[50]
2.2.1. Amino Acid Catabolism Leads to Formation of ToxicIntermediates
Changes in amino acid concentrations at defined culture phaseshave been correlated to cell growth inhibition and cell death. Forexample, asparagine consumption[26,39,53] has a negative effect incell growth,[53,54] while alanine production[26,39,53] inhibits TCAcycle[53,54] and also represents a source of ammonium.[55] Inanother study, lysine was supplied in excess in mediaformulations (considering a relatively low cell density) andaccumulates during the death phase.[56] A more completedescription is available in Table 1. Interestingly, it has recentlybeen shown that the catabolism of phenylalanine, tyrosine,tryptophan, methionine, leucine, serine, threonine, and glycineleads to the formation of nine intermediates (Table 1), which areidentified to be inhibiting cell growth.[57,58] These metabolitesshould, in principle, be immediately converted to the nextmetabolite in the catabolic pathway and, thus, end up in the TCAcycle. However, the pathways are not optimally regulated and,thus, under conditions of low lactate and ammonia, and high cell
Biotechnol. J. 2018, 1700499 1700499 (720
density observed in the later stages of cultivation, “leak” theinhibiting compounds into the medium, where the inter-mediates accumulate. Furthermore, it was demonstrated thatcontrolling the concentrations of these amino acids resulted ina reduction of the formation of inhibitory intermediates andimproved cell growth and product titers during fed-batchcultivation for antibody production.[57]
2.3. Control of Lipid Metabolism is Required for ProductiveCell Culture
A number of lipids have been shown to deplete or build up overtime in CHO cell culture (Table 1). In particular, glycerolipidshave been seen in multiple studies to affect growth in a variety ofways.[26,39,42,43,59] Choline phosphate (PCHO), glycero-3-phos-phocholine (G3PC), and glycerol-3-phosphate (G3P) have allbeen shown to build up in culture over time, which has beenseen to lead to growth limitation in both fed-batch and batchcultures.[26,42,43,59] This is an interesting phenomenon, as itsuggests a poor regulation of glycerolipids and membranecomposition in CHO cells.
PCHO has also been reported to deplete over time in longercultivation, which also leads to growth limitations, possibly dueto a resulting build up of G3P in the culture.[59] Finally, one canalso see glycerol as linked to glycerolipids, despite the compoundhaving many other functions in the cell, such as an osmoticregulator and a storage compound/redox sink.[39] As such, thismakes the monitoring of glycerol over culture interesting, as ithas been reported to also accumulate in the cells overculture.[26,43] Given the many roles of glycerol, it is difficult toevaluate the reason or effect of this accumulation, but the tieswith the redox potential of the cell and the possible links to
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheimof 13)
www.advancedsciencenews.com www.biotechnology-journal.com
lactate metabolism through NADH metabolism makes it anintriguing prospect.
Overall, themetabolism of glycerol and glycerolipids appear tobe suboptimal in CHO cells and with possible growth-limiting or-arresting effects. It is thus clear that these aspects must betightly controlled for optimal cell culture performance.
2.4. Glutathione Metabolism
Glutathione is a small abundant non-protein thiol, which isassembled from three amino acids found in eukaryoticcells.[60,61] In mammals, its main role is to act as a protectivemolecule against oxidative stress by transitioning between theoxidized (GSSG) and reduced (GSH) form of the molecule(Figure 2), and have as such been linked to cellular stressresponses. In particular, glutathione participates in redoxsignaling, detoxification of xenobiotics, regulation of cellproliferation, apoptosis, and is involved in immune functionevents.[60,62] While GSH and GSSG are important for the overallcellular metabolism, we focused on the key mechanismregarding recombinant protein production in this review.
2.4.1. Biosynthesis of Glutathione
The de novo biosynthesis of GSH takes place in the cytosol, asthe first reaction is catalyzed by the enzyme glutamate-cysteine
Figure 2. Glutathione biosynthesis and cycling based on the genes present(http://www.kegg.jp). Glutathione (GSH) biosynthesis occurs in the cytosolcatalyzed by the enzyme Gcl, which is composed by two subunits: Gcl catalyticatalyzed by Gss. Reactions of GSH with ROS are mediated by enzymesglutamamte cysteine ligase; Gss, glutathione synthetase. Ggct, gamma-gthioredoxin domain containing 12; Gsr, glutathione-disulfide reductase; G6dfamily; Anpep, alanyl aminopeptidase membrane; Lap3, leucine aminopeptidtype enzymes. “?”, reaction (represented with dashed line arrow) required foaccording to the consulted database.
Biotechnol. J. 2018, 1700499 1700499 (821
ligase (Gcl) which assembles the amino acids cysteine andglutamate at the cost of one ATP to form γ-glutamyl-cysteine –this represents the rate-limiting step in this pathway. Gcl is anenzyme composed of two subunits that are coded by twodifferent gene sequences as seen in higher eukaryotes – thesesubunits are the catalytic (Gclc) and the modifier (Gclm). Gclactivity is regulated by the concentration of GSH present inthe cell in a feedback inhibition manner.[63] Furthermore, theavailability of cysteine, as it donates the sulphur from itssidechain, represents another important limiting factor in thebiosynthesis of GSH. The second step in the generation of GSHis catalysed by the enzyme glutathione synthase (Gss) whereglycine is added to γ-glutamyl-cysteine, also at the cost of oneATP to form γ-L-glutamyl-L-cysteinyl-glycine – GSH (Figure 2).
As GSH interacts with reactive oxygen species and other redoxproteins, it is converted into glutathione disulphide (GSSG),where two molecules of GSH are required to form one GSSG.The salvage pathway of glutathione formation occurs whenGSSG is reduced back to GSH in a reaction catalyzed byglutathione-disulfide reductase (Gsr) which requires the avail-ability of NADPH as a co-factor, along with the magnesium ionMg2þ. Post-translational regulation of Gcl involves modifica-tions of Gclc via phosphorylation, caspase-mediated cleavage,which may have a mild impact on overall Gcl activity.[64]
Additionally, NADPþ and NADPH can alsomodulate Gcl activityin vitro.[64] Glutathione, additionally, can detoxify the cell fromthe toxic compound methylglyoxal (Table 1), which is formedspontaneously as its free form reacts with GSH.[65]
in Chinese hamster (C. griseus) genome, derived from KEGG pathwaysusing glutamate, cysteine and glycine as precursors. The first reaction isc subunit (Gclc) and Gcl regulatory subunit (Gclm), followed by a reactionfrom the Gst family. Abbreviations: ATP, adenosine triphosphate; Gcl,lutamylcyclotransferase; Gpx�, glutathione peroxidase family; Txndc12,p, glucose-6-phosphate dehydrogenase; Gstp�, glutathione S-transferasease 3; Cth, cystathionine gamma-lyase; Ggt�, gamma-glutamyltransferaser the endogenous formation of cystine, not present in C. griseus genome
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheimof 13)
www.advancedsciencenews.com www.biotechnology-journal.com
2.4.2. Glutathione in the Context of Recombinant ProteinProduction
During the production of recombinant proteins, the cellmetabolism is characterized by high glycolytic metabolismduring cell growth, while maximum antibody production isassociated with a more oxidative metabolism.[21] Due to theirhigh proliferative nature, CHO cells may experience increasedlevels of oxidative stress and, consequently, require higher levelsof GSH,[66,67] as observed in most types of cancer cells. Chonget al.[41] used a metabolomics approach with LC-MS analysis forthe identification of compounds which induced apoptosis in afed-batch cultivation of a mAb-producing CHO cell line. Theshortlisted extracellular metabolites which correlated withintracellular caspase activity were GSSG, AMP, GMP, andamino acid derivatives, which included dimethylarginine andacetylphenylalanine (Table 1). In particular, the presence ofGSSG in the medium resulted in an increased fold-change incaspase activity, which showed a strong link between GSSGaccumulation and the early signal of apoptotic cell death. Thisobservation suggests that GSSG is an additional cause for celldeath in prolonged cell cultures, other than those linked tolactate and ammonia. In a subsequent study, the group definedthe GSH as a marker of productivity, as high mAb producershave high intracellular GSH content.[68] This same trait wasobserved in CHO cells that produced a different mAb in a studywhich combined different proteomics methods to determinedifferentially expressed proteins.[69] Interestingly, this study hasalso shown that, among other pathways, glutathione biosynthe-sis enzymes were upregulated in the producer cells. Theengineering strategy developed by the same group is furtherdiscussed in section 5.
3. Metabolomics as an Evaluator of Presenceof Growth-Inhibiting and Toxic Metabolites
Metabolomics allows the quantitative analysis ofmetabolites thatare present inside and outside the cell, and provides evidenceregarding which pathways and reactions are active in the cellunder given culture conditions.[70–72] Metabolomics represents acomplement to other ‘omics,[73] since the data gathered fromthese investigations can also be integrated into metabolicmodels. While metabolic profiling is employed when a small setof metabolites that are linked to a phenotype is known,metabolomics is employed to measure and identify all possiblemetabolites and, consequently, explore hitherto undeterminedmetabolic links to a phenotype.[34] Thus the cell metabolome,along with the other ‘omics data, can inform which cellularevents are responsible for a specific phenotype (e.g., high proteinproducer). When reviewing metabolomics in the context ofrecombinant protein expression, Dickson[42] has argued that theinterpretation of these data sets can potentially assist in theidentification or generation of the best producer cells, either viacell engineering and/or the optimization of media and feeds.Moreover, the raw materials can be controlled using metab-olomics approaches and therefore, minimize batch-to-batchvariations, as part of bioprocess development. This methodologyis indeed central for understanding CHO cell metabolism and,
Biotechnol. J. 2018, 1700499 1700499 (922
when employed in combination with other ‘omics data, gives asnapshot of the cell’s metabolic state.
4. Online Resources for Metabolism
In order to continually follow the updates in CHO metabolismand identify new growth-inhibiting and toxic metabolites, anumber of online resources are helpful. www.CHOgenome.orgis the access point for all publicly available genome-wide data ofChinese hamster and CHO cell lines.[74] Similarly, the CHOmine(https://chomine.boku.ac.at) is a data warehouse for CHO datawith analysis tools. Additionally, CHOmine provides links toexternal websites and integrates recently published genomescale models (GEMs).[75] Such models, developed by a consor-tium of researchers, allow for the integration ‘omics data –genomics, transcriptomics, proteomics, and metabolomics – forguiding hypothesis-driven discovery and metabolic engineer-ing.[76] The GEMs are also excellent sources for an overview ofCHO metabolites as models specifically developed for CHOcells exist and can support cell line engineering approaches andCHO cells’ bioprocesses.[77] An earlier model[59] allowed for theidentification of growth limiting factors and is available at http://CHO.sf.net (v1.1). More recently, the constraint-based models ofChinese hamster and CHO cell lines (CHO-S, CHO-K1 andCHO-DG44) were made available to researchers in the CHOfield and can be downloaded from http://bigg.ucsd.edu/models/iCHOv1 and www.CHOgenome.org.[77]
Additional useful databases are the metabolic database KyotoEncyclopedia of Genes and Genome (KEGG) (http://www.kegg.jp), Reactome (http://www.reactome.org) and the HumanMetabolome Database (HMDB) (http://www.hmdb.ca). KEGGprovides information about metabolites and genes coding forenzymes which catalyze reactions participating in biochemicalpathways.[78–80] Reactome is a tool for the visualization of thereactions, networks in the context of cellular compartmentswhere anabolic and catabolic pathways occur.[81] Detailedinformation about small molecule metabolites that are presentin the human body[82–84] can be found in HMDB. This databasecan hint to which metabolites might affect CHO cells in culture,based on the toxic effects of the molecules to human cells,tissues, or organs.
In conclusion, the resources presented in Table 2 can aid andprovide clues for medium development and for finding targetsfor engineering cells with improved phenotypes, based on theavoidance of unwanted metabolites.
5. Cell Line Engineering for Improved NutrientMetabolism
Many strategies for cell line engineering have been employed inattempts to tackle the problematic of metabolic waste productswhich arise during cell culture, most extensively towards lactateproduction. The reviews by Ficher et al.[85] and Kim et al.[2] gathera number of cell line engineering approaches that were carriedout in CHO cells. We briefly discuss cell line engineeringapproaches carried out in CHO cells which resulted in reducedlactate production. Reports have demonstrated the potential of
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheimof 13)
Table 2. Summary of publicly available resources relevant for CHO research.
Name Description/comment URL Reference
CHOgenome Host of all published CHO-related data. Compiles genome-scale information of Chinese hamster
and CHO-K1.
http://www.
chogenome.org
[74]
CHOmine Data warehouse for CHO data and provides links to outside websites containing information on
gene and protein. It integrates the recently published genome scale model for C. griseus. and
CHO cell lines.
https://chomine.
boku.ac.at
[75]
Genome scale model for C. griseus.
and CHO cell lines
Genome scale model of global model of Chinese hamster (C. griseus) metabolism and cell line-
specific models of CHO-S, K1, and DG44.
http://bigg.ucsd.
edu/models/iCHOv1
[77]
Standardized network
reconstruction of CHO cell
metabolism
Genome-scale network reconstruction of CHO cell metabolic network, as based on genome
sequence and literature.
CHO.sf.net (v1.1) [59]
Kyoto Encyclopedia of Genes and
Genome (KEGG)
Metabolic database that provides information about metabolites and genes coding for enzymes
catalyzing reactions part of metabolic pathways. Data accessible for several organisms including
C. griseus.
http://www.kegg.jp [78–80]
Reactome Knowledgebase Archive of biological processes and a tool for discovering functional relationships in data;
Provides visualization of reaction networks and details of single reactions; some N-glycosylation
pathways for C. griseus are available and other metabolic maps are not fully accessible. Therefore,
we recommend using this tool based on well annotated and closely related organisms, such as
H. sapiens or M. musculus, as a guide.
http://www.
reactome.org
[81]
Human metabolome database The Human Metabolome Database (HMDB) contains detailed information about small
molecule metabolites found in the human body. It is intended to be used for applications
in metabolomics among others. Additionally, it provides chemical, clinical, and molecular
biology/biochemistry data.
http://www.hmdb.ca [82–84]
www.advancedsciencenews.com www.biotechnology-journal.com
engineering CHO cells, as seen in the case of the LdhA gene thatwas downregulated using RNAi technology, yielding reducedlactate production rates without impacting cell growth norproductivity of human thrombopoietin.[86] Similar results wereobserved with the downregulation of LdhA and pyruvatedehydrogenase kinase (Pdhk) isoenzymes 1, 2, and 3 in antibodyproducing-CHO cells.[87] However, the knockout of LdhA usingzinc finger nucleases (ZFNs) in cells where Pdhk 1, 2, and 3 wasdown regulated was revealed to be lethal.[88] The overexpressionof Aralar1, part of malate–aspartate shuttle (MAS), in a lactate-producing cell line led to a metabolic shift from lactateproduction to consumption.[89] This way, the authors found alink betweenMAS and this metabolic shift. Other investigationalwork involved the stable expression of fructose transporter(GLUT5).[90] When cells used this sugar as a carbon source, theuptake rate of fructose was such (low) that the overflow of excesscarbon to lactate was avoided. A number of research articlesdescribe the effects of overexpressing the enzyme pyruvatecarboxylase. The overexpression of yeast pyruvate carboxylase(PYC2) resulted in a significant decrease in lactate productionand increase in productivity.[91] An identical outcome wasobserved when human pyruvate carboxylase was engineeredusing a similar approach.[92] In a more recent study, theoverexpression of codon optimized PYC2, reduced lactateproduction, and improved mAb production and glycosylation.[93]
Additionally, improved cell metabolism was observed with theoverexpression of malate dehydrogenase II (MDHII), which leadto an increase in intracellular ATP and NADH, and integralviable cell number.[94] The LC-MS analysis of the extracellularmetabolites revealed the accumulation of malate, which was aresult of an excess supply of aspartate in the medium and thepresence of a bottleneck in MDHII in the TCA cycle.
Biotechnol. J. 2018, 1700499 1700499 (123
More recently, the glutathione biosynthesis pathway wasengineered through the stable overexpression of Gclc, whichyielded increased GSH concentrations but did not improveproductivity.[95] However, when the modifier subunit of Gcl wasstably overexpressed in CHO host cells, an increase in specificproductivity was observed once a mAb was transiently expressedby these cells. Surprisingly, the findings of this work allowed theconclusion that the GSH content does not contribute to theimprovement of productivity of mAb in CHO cells, contrary towhat was previously stated.[68,69]
Furthermore, the development of glutamine synthetase (GS)selection system exemplifies an important advance in recombi-nant protein expression using CHO cells, representing analternative to dihydrofolate reductase (DHFR) expressionsystem. The GS system is based on the knockout of the geneencoding for GS, which is reintroduced into the cell along withthe vector encoding for recombinant protein.[96] The cells growin glutamine free-medium under the selection pressure ofmethionine sulfoximine (MSX). An additional advantage of theGS system is that it allows for the reduction of by-productformation, as once the GS gene is reintroduced, ammonia alongwith glutamate is utilized to form of glutamine. Glutaminebecomes available for the formation of TCA cycle intermediatessuch as α-ketoglutarate.
6. Applications and Future Perspectives
In upstream process development for the production ofrecombinant therapeutic protein, both media and feed design,and cell line engineering can be employed. Cell line- or clone-specific media optimization may be required for each shortlisted
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim0 of 13)
www.advancedsciencenews.com www.biotechnology-journal.com
candidate that is generated in one cell line developmentexperiment, as these may display different growth phenotypesand by-product levels, as well as to account for the effects ofclone-medium interactions.[97]
The quantification of toxic metabolites in cell culture can aidthe media development efforts, by indicating which precursormedia and feed components are required to be supplied incontrolled amounts. However, while media and feed optimiza-tion has enabled achievement of higher cell densities andincreased productivity, it is still far from challenging the cell’smaximized growth and production capacity that is predicted bymetabolic network models. Metabolomics, along with other‘omics, provides an extra layer of knowledge on the cellmetabolism and can lead to breakthroughs that improve theseparameters. For instance, after the identification of toxicmetabolic intermediates, such as the ones presented in Table 1,one can employ cell engineering tools to limit the formation ofthese inhibitory molecules. The metabolic pathways where thesecompounds are involved should be analyzed for identification ofthe target genes. Thereafter, it is essential to select the mostsuitable engineering strategy to perform specific genomicchanges (e.g., downregulation or deletions of genes, or theoverexpression of heterologous pathways that convert the toxicintermediates into “safer” molecules) for targeting genesencoding for enzymes forming such molecules. GEMs can beused to predict the effect of the transformation. The resultingphenotypic changes of the cell may indicate better nutrient usageand reduced the formation of toxic and inhibiting metabolites.Effective tools for genome engineering, such as ZFNs, andtranscription activator-like effector nucleases (TALENs)employed in the past revealed themselves to be rather costly.The less costly and still efficient tool clustered regularlyinterspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 system (CRISPR/Cas9) system (reviewedby Lee et al.[98]) permits faster but yet specific gene targeting inmammalian cells. This genome editing tool offers newcapabilities for streamlining CHO cell line developmentprocesses to obtain improved cell factories. A multiplexing cellengineering approach successfully reduced apoptosis andyielded non-fucosylated secreted proteins, through the simulta-neous triple knockout of apoptotic Bcl-2 antagonist killer protein(BAK), Bcl-2-associated X protein (BAX), and fucosyltransferase8 (FUT8) using CRISPR/Cas9 in CHO-S cells.[99] A similarapproach can be employed to target metabolic genes.
Taken together – themetabolic models generated based on theintegration of ‘omics data, the employment of metabolomics forobtaining a detailed view of all active metabolic reactions in thecell and the recent genome editing tools which offer newcapabilities for engineering and generating cells with idealphysiologic traits – form a well-connected trio which canenhance CHO cells factories as platforms for expression oftherapeutic recombinant proteins.
AbbreviationsADP, adenosine monophosphate; AMP, adenosine monophosphate;Anpep, alanyl aminopeptidase membrane; ATP, Adenosine triphos-phate, BAK; Bcl-2 antagonist killer protein; BAX, Bcl-2-associated X
Biotechnol. J. 2018, 1700499 1700499 (124
protein; Cas, CRISPR-associated; Cas9, Cas protein 9; CHO, Chinesehamster ovary; CRISPR, Clustered regularly interspaced short palin-dromic repeats; EPO, erythropoietin; ER, Endoplasmatic Reticulum;FUT8, fucosyltransferase 8; Gcl, Glutamamte cysteine ligase; Gclc,Glutamamte cysteine ligase catalytic subunit; Gclm, Glutamamtecysteine ligase regulatory subunit; GDP, guanosine diphosphate;GEM(s), genome-scale models; Ggt, gamma-glutamyltransferase; Ggct,gamma-glutamylcyclotransferase; GMP, guanosine monophosphate;Gpx, glutathione peroxidase; GS, glutamine synthetase; Gsr, Glutathi-one-disulfide reductase; GSH, reduced glutathione; Gss, Glutathionesynthetase; Gst, glutathione S-transferase; GSSG, oxidized glutathioneor glutathione disulfide; G3P, glycerol-3-phosphate; G3PC, glycero-3-phosphocholine; G6d, glucose-6-phosphate; G6pd, glucose-6-phos-phate dehydrogenase; HMDB, Human Metabolome Database; KEGG,Kyoto Encyclopedia of Genes and Genome; LdhA, lactate dehydrogenaseA; mAb(s) monoclonal antibody(ies); MAS, malate–aspartate shuttle(MAS); MDHII, malate dehydrogenase II; NaCl, sodium chloride;NADþ, Nicotinamide adenine dinucleotide oxidized; NADH, Nicotin-amide adenine dinucleotide reduced; NADPþ, Nicotinamide adeninedinucleotide phosphate oxidized; NADPH, Nicotinamide adeninedinucleotide phosphate reduced; PCHO, Choline phosphate; Pdhkpyruvate dehydrogenase kinase; PYC2, yeast pyruvate carboxylase;TALENs, transcription activator-like effector nucleases; TCA, Tricarbox-ylic acid; tPA, tissue plasminogen activator; Txndc12, thioredoxindomain containing 12; ZFNs, zinc finger nucleases.
AcknowledgementsThis work is funded by Marie Skłodowska-Curie Actions under the EUFramework Programme for Research and Innovation for eCHO systemsITN (Grant no. 642663). H.F.K. and S.P. additionally thank the NovoNordisk Foundation (Grant no. NNF10CC1016517) for the support.
Conflict of InterestThe authors declare no conflict of interest.
Keywordsamino acid metabolism, Chinese hamster ovary cells, glutathionemetabolism, metabolic by-products
Received: July 24, 2017Revised: December 21, 2017
Published online:
[1] G. Walsh, Nat. Biotech. 2014, 32, 992.[2] J. Y. Kim, Y. G. Kim, G. M. Lee, Appl. Microbiol. Biotechnol. 2012, 93,
917.[3] G. Walsh, Nat. Biotechnol. 2010, 28, 917.[4] E. Wells, A. S. Robinson, Biotechnol. J. 2017, 12.[5] R. Jefferis, Trends Pharmacol. Sci. 2009, 30, 356.[6] A. M. Sinclair, S. Elliott, J. Pharm. Sci. 2005, 94, 1626.[7] A. Helenius, M. Aebi, Science 2001, 291, 2364.[8] C. F. Goochee, M. J. Gramer, D. C. Andersen, J. B. Bahr,
J. R. Rasmussen, Biotechnology. (N. Y). 1991, 9, 1347.[9] C. F. Goochee, Dev. Biol. Stand. 1992, 76, 95.
[10] S. Elliott, T. Lorenzini, S. Asher, K. Aoki, D. Brankow, L. Buck,L. Busse, D. Chang, J. Fuller, J. Grant, N. Hernday, M. Hokum,S. Hu, A. Knudten, N. Levin, R. Komorowski, F. Martin, R. Navarro,
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim1 of 13)
www.advancedsciencenews.com www.biotechnology-journal.com
T. Osslund, G. Rogers, N. Rogers, G. Trail, J. Egrie, Nat. Biotechnol.2003, 21, 414.
[11] A. Berting,M. R. Farcet, T. R. Kreil, Biotechnol. Bioeng. 2010, 106, 598.[12] X. Xu, H. Nagarajan, N. E. Lewis, S. Pan, Z. Cai, X. Liu, W. Chen,
M. Xie, W. Wang, S. Hammond, M. R. Andersen, N. Neff,B. Passarelli, W. Koh, H. C. Fan, J. Wang, Y. Gui, K. H. Lee,M. J. Betenbaugh, S. R. Quake, I. Famili, B. O. Palsson, J. Wang,Nat. Biotechnol. 2011, 29, 735.
[13] L. Chu, D. K. Robinson, Curr. Opin. Biotechnol. 2001, 12, 180.[14] J. Neermann, R. Wagner, J. Cell. Physiol. 1996, 166, 152.[15] J. D. Young, Curr. Opin. Biotechnol. 2013, 24, 1108.[16] B. C. Mulukutla, S. Khan, A. Lange, W.-S. Hu, Trends Biotechnol.
2010, 28, 476.[17] M. S. Lao, D. Toth, Biotechnol. Prog. 1997, 13, 688.[18] S. S. Ozturk, M. R. Riley, B. O. Palsson, Biotechnol. Bioeng. 1992, 39,
418.[19] Y. Fan, I. Jimenez Del Val, C. Müller, J. Wagtberg Sen,
S. K. Rasmussen, C. Kontoravdi, D. Weilguny, M. R. Andersen,Biotechnol. Bioeng. 2015, 112, 521.
[20] M. Schneider, I. W. Marison, U. Von Stockar, J. Biotechnol. 1996, 46,161.
[21] N. Templeton, J. Dean, P. Reddy, J. D. Young, Biotechnol. Bioeng.2013, 110.
[22] A. Woo Suk, M. R. Antoniewicz, Metab. Eng. 2013, 15, 34.[23] O. Warburg, Science (80-.). 1956, 123, 309.[24] J. W. Locasale, L. C. Cantley, Cell Metab. 2011, 14, 443.[25] H. J. Cruz, C. M. Freitas, P. M. Alves, J. L. Moreira, M. J. T. Carrondo,
Enzyme Microb. Technol. 2000, 27, 43.[26] C. A. Sellick, A. S. Croxford, A. R. Maqsood, G. Stephens,
H. V. Westerhoff, R. Goodacre, A. J. Dickson, Biotechnol. Bioeng.2011, 108, 3025.
[27] N. Ma, J. Ellet, C. Okediadi, P. Hermes, E. McCormick, S. Casnocha,Biotechnol. Prog. 2009, 25, 1353.
[28] F. Zagari, M. Jordan, M. Stettler, H. Broly, F. M. Wurm, N.Biotechnol. 2013, 30, 238.
[29] V. S. Martínez, S. Dietmair, L.-E. Quek, M. P. Hodson, P. Gray,L. K. Nielsen, Biotechnol. Bioeng. 2013, 110, 660.
[30] I. H. Yuk, S. Russell, Y. Tang, W. T. Hsu, J. B. Mauger, R. P. S. Aulakh,J. Luo, M. Gawlitzek, J. C. Joly, Biotechnol. Prog. 2015, 31, 226.
[31] N. Templeton, K. D. Smith, A. G. McAtee-Pereira, H. Dorai,M. J. Betenbaugh, S. E. Lang, J. D. Young,Metab. Eng. 2016, 43, 218.
[32] I. H. Yuk, J. D. Zhang, M. Ebeling, M. Berrera, N. Gomez, S. Werz,C. Meiringer, Z. Shao, J. C. Swanberg, K. H. Lee, J. Luo, B. Szperalski,Biotechnol. Prog. 2014, 30, 429.
[33] J. Luo, N. Vijayasankaran, J. Autsen, R. Santuray, T. Hudson,A. Amanullah, F. Li, Biotechnol. Bioeng. 2012, 109, 146.
[34] B. C. Mulukutla, M. Gramer, W. S. Hu, Metab. Eng. 2012, 14, 138.[35] H. Le, S. Kabbur, L. Pollastrini, Z. Sun, K. Mills, K. Johnson,
G. Karypis, W. S. Hu, J. Biotechnol. 2012, 162, 210.[36] B. C. Mulukutla, A. Yongky, S. Grimm, P. Daoutidis, W. S. Hu, PLoS
ONE 2015, 10, 1.[37] J. Li, C. L. Wong, N. Vijayasankaran, T. Hudson, A. Amanullah,
Biotechnol. Bioeng. 2012, 109, 1173.[38] C. Altamirano, J. Berrios, M. Vergara, S. Becerra, Electron. J.
Biotechnol. 2013, 16.[39] C. A. Sellick, A. S. Croxford, A. R. Maqsood, G. M. Stephens,
H. V. Westerhoff, R. Goodacre, A. J. Dickson, Biotechnol. J. 2015, 10,1434.
[40] S. S. Santos, M. Haury, M. J. T. Carrondo, Biotechnol. Prog. 2003,1000.
[41] W. P. K. Chong, F. N. K. Yusufi, D.-Y. Lee, S. G. Reddy,N. S. C. Wong, C. K. Heng, M. G. S. Yap, Y. S. Ho, J. Biotechnol.2011, 151, 218.
[42] A. J. Dickson, Curr. Opin. Biotechnol. 2014, 30, 73.
Biotechnol. J. 2018, 1700499 1700499 (125
[43] N. Carinhas, T. M. Duarte, L. C. Barreiro, M. J. T. Carrondo,P. M. Alves, A. P. Teixeira, Biotechnol. Bioeng. 2013, 110, 3244.
[44] L. M. Carrillo-Cocom, T. Genel-Rey, D. Araíz-Hernández, F. L�opez-Pacheco, J. L�opez-Meza, M. R. Rocha-Piza~na, A. Ramírez-Medrano,M. M. Alvarez, Cytotechnology 2014, 809.
[45] A. Salazar, M. Keusgen, J. Von Hagen, Amino Acids 2016, 48, 1161.[46] S. L. Berg JM, JLTymoczko, Biochemistry, 5th ed. WH Freeman, New
York 2002, Ch. 23.[47] P. Chen, S. W. Harcum, J. Biotechnol. 2005, 117, 277.[48] M. W. Glacken, R. J. Fleischaker, A. J. Sinskey, Biotechnol. Bioeng.
1986, 28, 1376.[49] K. Martinelle, L. Häggström, Cytotechnology 1999, 29, 45.[50] S. Rose, T. Black, D. Ramakrishnan, Handbook of Industrial Cell
Culture�Mammalian, Microbial, and Plant Cells, 1st ed. (Eds: VinciVA, Parekh SR. Humana Press, Totowa, NJ 2003, Ch. 4.
[51] N. Kurano, C. Leist, F. Messi, S. Kurano, A. Fiechter, J. Biotechnol.1990, 15, 113.
[52] A. J. Mastrangelo, J. M. Hardwick, S. Zou, M. J. Betenbaugh,Biotechnol. Bioeng. 2000, 67, 555.
[53] D. Reinhart, L. Damjanovic, C. Kaisermayer, R. Kunert, Appl.Microbiol. Biotechnol. 2015, 99, 4645.
[54] Z. Xing, B. Kenty, I. Koyrakh, M. Borys, S. H. Pan, Z. J. Li, ProcessBiochem. 2011, 46, 1423.
[55] T. M. Duarte, N. Carinhas, L. C. Barreiro, M. J. T. Carrondo,P. M. Alves, A. P. Teixeira, Biotechnol. Bioeng. 2014, 111.
[56] S. E. Mohmad-Saberi, Y. Z. H. Y. Hashim, M. Mel, A. Amid,R. Ahmad-Raus, V. Packeer-Mohamed, Cytotechnology 2013, 65,577.
[57] B. C. Mulukutla, J. Kale, T. Kalomeris, M. Jacobs, G. W. Hiller,Biotechnol. Bioeng. 2017, 114, 1779.
[58] G.W. Hiller, B.C. Mulukutla, Method of cell culture, WO 2015/140708 A1, 2015.
[59] S. Selvarasu, Y. S. Ho,W. P. K. Chong, N. S. C. Wong, F. N. K. Yusufi,Y. Y. Lee, M. G. S. Yap, D.-Y. Lee, Biotechnol. Bioeng. 2012, 109,1415.
[60] A. Meister, M. E. Anderson, Annu. Rev. Biochem. 1983, 52, 711.[61] Y. Li, G. Wei, J. Chen, Appl. Microbiol. Biotechnol. 2004, 66, 233.[62] S. C. Lu, Biochim Biophys Acta 2013, 1830, 3143.[63] G. F. Seelig, R. P. Simondsen, A. Meister, J. Biol. Chem. 1984, 259,
9345.[64] C. C. Franklin, D. S. Backos, I. Mohar, C. C. White, H. J. Forman,
T. J. Kavanagh, Mol. Aspects Med. 2009, 30, 86.[65] B. M. Roy, T. D. Rau, R. R. Balcarcel, Cytotechnology 2004, 46, 97.[66] M. P. Gamcsik, M. S. Kasibhatla, S. D. Teeter, O. M. Colvin,
Biomarkers 2012, 17, 671.[67] Y. Liu, A. S. Hyde, M. A. Simpson, J. J. Barycki, Adv. Cancer Res. 2014,
122, 69.[68] W. P. K. Chong, S. H. Thng, A. P. Hiu, D. Y. Lee, E. C. Y. Chan,
Y. S. Ho, Biotechnol. Bioeng. 2012, 109, 3103.[69] C. A. Orellana, E. Marcellin, B. L. Schulz, A. S. Nouwens, P. P. Gray,
L. K. Nielsen, J. Proteome Res. 2015, 14, 609.[70] R. Goodacre, S. Vaidyanathan, W. B. Dunn, G. G. Harrigan,
D. B. Kell, Trends Biotechnol. 2004, 22, 245.[71] P. Datta, R. J. Linhardt, S. T. Sharfstein, Biotechnol. Bioeng. 2013,
110, 1255.[72] S. Dietmair, N. E. Timmins, P. P. Gray, L. K. Nielsen, J. O. Krömer,
Anal. Biochem. 2010, 404, 155.[73] A. M. Lewis, N. R. Abu-Absi, M. C. Borys, Z. J. Li, Biotechnol. Bioeng.
2016, 113, 26.[74] B. G. Kremkow, J. Y. Baik, M. L. MacDonald, K. H. Lee, Biotechnol. J.
2015, 2011, 931.[75] M. P. Gerstl, M. Hanscho, D. E. Ruckerbauer, Database 2017.[76] M. A. Oberhardt, B. Ø. Palsson, J. A. Papin, Mol. Syst. Biol. 2009,
5, 1.
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim2 of 13)
www.advancedsciencenews.com www.biotechnology-journal.com
[77] H. Hefzi, K. S. Ang, M. Hanscho, A. Bordbar, D. Ruckerbauer,M. Lakshmanan, C. A. Orellana, D. Baycin-Hizal, Y. Huang, D. Ley,V. S. Martinez, S. Kyriakopoulos, N. E. Jim�enez, D. C. Zielinski,L. E. Quek, T. Wulff, J. Arnsdorf, S. Li, J. S. Lee, G. Paglia, N. Loira,P. N. Spahn, L. E. Pedersen, J. M. Gutierrez, Z. A. King, A. M. Lund,H. Nagarajan, A. Thomas, A. M. Abdel-Haleem, J. Zanghellini,H. F. Kildegaard, B. G. Voldborg, Z. P. Gerdtzen, M. J. Betenbaugh,B. O. Palsson, M. R. Andersen, L. K. Nielsen, N. Borth, D. Y. Lee,N. E. Lewis, Cell Syst. 2016, 3, 434.
[78] M. Kanehisa, Nucleic Acids Res. 2006, 34, D354.[79] M. Kanehisa, Y. Sato, M. Kawashima, M. Furumichi, M. Tanabe,
Nucleic Acids Res. 2016, 44, D457.[80] M. Kanehisa, M. Furumichi, M. Tanabe, Y. Sato, K. Morishima,
Nucleic Acids Res. 2017, 45, D353.[81] A. Fabregat, K. Sidiropoulos, P. Garapati, M. Gillespie,
K. Hausmann, R. Haw, B. Jassal, S. Jupe, F. Korninger, S. McKay,L. Matthews, B. May, M. Milacic, K. Rothfels, V. Shamovsky,M. Webber, J. Weiser, M. Williams, G. Wu, L. Stein, H. Hermjakob,P. D’Eustachio, Nucleic Acids Res. 2016, 44, D481.
[82] D. S. Wishart, D. Tzur, C. Knox, R. Eisner, A. C. Guo, N. Young,D. Cheng, K. Jewell, D. Arndt, S. Sawhney, C. Fung, L. Nikolai,M. Lewis, M. A. Coutouly, I. Forsythe, P. Tang, S. Shrivastava,K. Jeroncic, P. Stothard, G. Amegbey, D. Block, D. D. Hau, J. Wagner,J. Miniaci, M. Clements, M. Gebremedhin, N. Guo, Y. Zhang,G. E. Duggan, G. D. MacInnis, A. M. Weljie, R. Dowlatabadi,F. Bamforth, D. Clive, R. Greiner, L. Li, T. Marrie, B. D. Sykes,H. J. Vogel, L. Querengesser, Nucleic Acids Res. 2007, 35, 521.
[83] D. S. Wishart, C. Knox, A. C. Guo, R. Eisner, N. Young, B. Gautam,D. D. Hau, N. Psychogios, E. Dong, S. Bouatra, R. Mandal,I. Sinelnikov, J. Xia, L. Jia, J. A. Cruz, E. Lim, C. A. Sobsey,S. Shrivastava, P. Huang, P. Liu, L. Fang, J. Peng, R. Fradette,D. Cheng, D. Tzur, M. Clements, A. Lewis, A. de souza, A. Zuniga,M. Dawe, Y. Xiong, D. Clive, R. Greiner, A. Nazyrova,R. Shaykhutdinov, L. Li, H. J. Vogel, I. Forsythei, Nucleic AcidsRes. 2009, 37, 603.
[84] D. S. Wishart, T. Jewison, A. C. Guo, M. Wilson, C. Knox, Y. Liu,Y. Djoumbou, R. Mandal, F. Aziat, E. Dong, S. Bouatra, I. Sinelnikov,D. Arndt, J. Xia, P. Liu, F. Yallou, T. Bjorndahl, R. Perez-Pineiro,R. Eisner, F. Allen, V. Neveu, R. Greiner, A. Scalbert, Nucleic AcidsRes. 2013, 41, 801.
[85] S. Fischer, R. Handrick, K. Otte, Biotechnol. Adv. 2015, 33, 1878.
Biotechnol. J. 2018, 1700499 1700499 (126
[86] S. H. Kim, G. M. Lee, Appl. Microbiol. Biotechnol. 2007, 74, 152.[87] M. Zhou, Y. Crawford, D. Ng, J. Tung, A. F. J. Pynn, A. Meier,
I. H. Yuk, N. Vijayasankaran, K. Leach, J. Joly, B. Snedecor, A. Shen, J.Biotechnol. 2011, 153, 27.
[88] S. S. M. Yip, M. Zhou, J. Joly, B. Snedecor, A. Shen, Y. Crawford,Mol.Biotechnol. 2014, 56, 833.
[89] F. Zagari, M. Stettler, L. Baldi, H. Broly, F. M. Wurm, M. Jordan,Pharm. Bioprocess. 2013, 1, 19.
[90] K. F. Wlaschin, W. S. Hu, J. Biotechnol. 2007, 131, 168.[91] M. B. Fogolín, R. Wagner, M. Etcheverrigaray, R. Kratje, J. Biotechnol.
2004, 109, 179.[92] S. H. Kim, G. M. Lee, Appl. Microbiol. Biotechnol. 2007, 76, 659.[93] S. K. Gupta, A. Sharma, H. Kushwaha, P. Shukla, Front. Pharmacol.
2017, 8, 1.[94] W. P. K. Chong, S. G. Reddy, F. N. K. Yusufi, D. Y. Lee, N. S. C. Wong,
C. K. Heng, M. G. S. Yap, Y. S. Ho, J. Biotechnol. 2010, 147, 116.[95] C. A. Orellana, E. Marcellin, P. P. Gray, L. K. Nielsen, Biotechnol.
Bioeng. 2017, 114, 1825.[96] L. Fan, C. C. Frye, A. J. Racher, Pharm. Bioprocess 2013, 1, 487.[97] S. Dietmair, M. P. Hodson, L.-E. Quek, N. E. Timmins,
P. Chrysanthopoulos, S. S. Jacob, P. Gray, L. K. Nielsen, Biotechnol.Bioeng. 2012, 109, 1404.
[98] J. S. Lee, L. M. Grav, N. E. Lewis, H. F. Kildegaard, Biotechnol. J 2015,10, 979.
[99] L. M. Grav, J. S. Lee, S. Gerling, T. B. Kallehauge, A. H. Hansen,S. Kol, G. M. Lee, L. E. Pedersen, H. F. Kildegaard, Biotechnol. J2015, 10, 1446.
[100] P. Chen, S. W. Harcum, Metab. Eng. 2006, 8, 123.[101] J. Dean, P. Reddy, Biotechnol. Bioeng. 2013, 110, 1735.[102] D. J. Jiang, S. J. Jia, Z. Dai, Y. J. Li, J. Mol. Cell. Cardiol. 2006, 40, 529.[103] R. H. Böger, S. M. Bode-Böger, P. S. Tsao, P. S. Lin, J. R. Chan,
J. P. Cooke, J. Am. Coll. Cardiol. 2000, 36, 2287.[104] T. J. Ihrig, M. A. Maulawizada, B. D. Thomas, F. S. Jacobson, Animal
Cell Technology: Developments Towards the 21st Century. (Eds:Beuvery EC, Griffiths JB, Zeijlemaker WP. Dordrecht, SpringerNetherlands 1995.
[105] S. Kingkeohoi, F. W. R. Chaplen, Cytotechnology 2005, 48, 1.[106] F. W. R. Chaplen, Cytotechnology 1998, 26, 173.[107] K. E. Tobias, C. Kahana, Cell Growth Differ. 1995, 6, 1279.[108] A. V. Carvalhal, I. Marcelino, M. J. T. Carrondo, Appl. Microbiol.
Biotechnol. 2003, 63, 164.
© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim3 of 13)
Chapter 2 - Engineering the metabolism of CHO cells This chapter describes cell line engineering approaches to improve cell growth and reduce by-product
formation. Our research group has optimized genome editing tools such as the CRISPR/Cas9 system [84,85]
paired with RMCE in CHO cells [32]. These allow for targeting selected genes of interest for disruption and
overexpression. Here, we focus on targeting genes that directly or indirectly influence the nutrient and by-
product metabolism of CHO cells. First, targeting genes in the amino acid catabolic pathways for single and
combinatorial gene disruption using the CRISPR/Cas9 system is presented in two manuscripts (Paper II
and Paper III). Second, as an alternative strategy to improve cell growth, a gene related to the cofactor
metabolism was overexpressed using RMCE (Paper IV).
27
Paper II – Reprogramming amino acid catabolism in CHO cells with CRISPR/Cas9 genome editing improves cell growth and reduces byproduct secretion
28
Reprogramming AA catabolism in CHO cells with CRISPR/Cas9 genome editing improves cell growth and reduces byproduct secretion Daniel Ley1,2, Sara Pereira2, Lasse Ebdrup Pedersen2, Johnny Arnsdorf2, Hooman Hefzi3,4, Anne Mathilde
Lund1, Tae Kwang Ha2, Tune Wulff2, Helene Faustrup Kildegaard2,*, Mikael Rørdam Andersen1,*.
(1) Department of Biotechnology and Biomedicine, Technical University of Denmark, Kgs. Lyngby, Denmark.
(2) The Novo Nordisk Foundation Center for Biosustainability, Technical University of Denmark, Kgs.
Lyngby, Denmark. (3) Department of Bioengineering, University of California, San Diego, La Jolla, CA 92093,
United States. (4) Novo Nordisk Foundation Center for Biosustainability at the University of California, San
Diego, School of Medicine, La Jolla, CA 92093, United States.
*Corresponding authors: [email protected], [email protected]
Phone: +45 45 25 26 75, Fax: +45 45 88 41 48
Address: Søltofts plads, bygning 223, 2800 Kgs Lyngby, Denmark.
Declaration of interests: The authors declare no competing interests.
Grant numbers: The Novo Nordisk Foundation and eCHO Systems H2020 MSC-ITN (Grant no. 642663)
provided funding for this work.
Keywords: Chinese hamster ovary cells, Lactate, Ammonium, AA catabolism, CRISPR, Metabolic network
reconstruction.
29
Abstract
Chinese hamster ovary (CHO) cells are the preferred host for producing biopharmaceuticals. Amino
acids are biologically important precursors for CHO metabolism; they serve as building blocks for
proteogenesis, including synthesis of biomass and recombinant proteins, and are utilized for growth and
cellular maintenance. In this work, we studied the physiological impact of disrupting amino acid catabolic
pathways in CHO cells. We aimed to reduce secretion of growth inhibiting metabolic by-products derived
from amino acid catabolism including lactate and ammonium. To achieve this, we engineered nine genes in
seven different amino acid catabolic pathways using the CRISPR-Cas9 genome editing system. For
identification of target genes, we used a metabolic network reconstruction of amino acid catabolism to follow
transcriptional changes in response to antibody production, which revealed candidate genes for disruption.
We found that disruption of single amino acid catabolic genes reduced specific lactate and ammonium
secretion while specific growth rate and integral of viable cell density were increased in many cases. Disruption
of multiple amino acid catabolic genes further reduced secretion of lactate and ammonium, but did not
increase growth. This study demonstrates the potential of engineering amino acid catabolism in CHO cells to
achieve improved phenotypes for bioprocessing.
1. Introduction
Chinese hamster ovary (CHO) cells are the predominant cell factory for producing recombinant
therapeutic proteins, a segment of the pharmaceutical industry worth more than 140 billion USD in 2013 alone
(Walsh, 2014). A key element for improving CHO-based production of active pharmaceutical ingredients
(APIs) has traditionally been the development of optimized growth media that provide cells with excess
nutrients to support growth and protein productivity. Today, fed-batch bioprocessing of CHO cells is
complicated by accumulation of toxic metabolic by-products, mainly lactate and ammonium, which inhibit
growth (Lao and Toth, 1997), impair recombinant protein quality (Andersen and Goochee, 1995; Borys et al.,
1994; Thorens and Vassalli, 1986; Yang and Butler, 2000), and productivity (Hansen and Emborg, 1994). To
30
address this problem, bioprocessing- and cell line engineering efforts have mainly targeted glucose and
glutamine metabolism (reviewed by Ahn and Antoniewicz, 2012; Altamirano et al., 2013; Young, 2013).
However, the impact of amino acid (AA) catabolism on lactate and ammonium production has received less
attention, despite that AA catabolism is directly linked to ammonium production, and indirectly, through
redox metabolism, to lactate production. This prompted us to study the physiological effects of reprogramming
AA catabolism in CHO cells. Specifically, the physiological impact of disrupting AA catabolic pathways, and
in particular, the allocation of AAs to catabolism and biomass synthesis.
AA catabolism is in many ways an unwanted process in a CHO cell producing a biopharmaceutical.
For one, the AAs would be more efficiently used directly in protein biosynthesis and biomass production. AA
catabolism generate ammonium by transamination, a chemical reaction that transfers an amino group to α-
ketoglutarate to form glutamate, which is deaminated to yield ammonium (Ahn and Antoniewicz, 2012). AA
catabolism contributes to lactate production as well, either directly by fueling carbon to glycolysis (Templeton
et al., 2014), or in an indirect redox-dependent manner. Many AA catabolic pathways reduce NAD+ to NADH,
which perturb the redox equilibrium, and force the cell to regenerate cytosolic NAD+ pools through lactate
synthesis, to maintain redox homeostasis (Templeton et al., 2014). To our knowledge, no studies have targeted
lactate and ammonium production derived from AA catabolism (excluding glutamine) in CHO cells, despite
these pathways accounting for 25 % of the total carbon pool fueling central carbon metabolism (Nicolae et al.,
2014).
Recent evidence suggests that AA catabolism in CHO cells produces a wide range of growth-inhibiting
compounds besides lactate and ammonium. Mulukutla et al. 2017 identified nine growth-inhibiting
compounds from catabolism of phenylalanine, tyrosine, tryptophan, methionine, leucine, serine, threonine
and glycine. They demonstrated that controlling these AAs at low concentrations reduced inhibitor
accumulation and improved peak cell density and antibody titers in fed-batch culture. Similarly, González-
Leal et al., 2011 found that leucine and threonine inhibit peak cell density and maximum specific growth rate
during exponential growth, and suggested a feeding strategy, in which these AAs remain at sufficiently low
31
concentrations, to avoid growth inhibiting effects while sustaining protein synthesis. Clinical studies provide
several indications that intermediates in AA catabolism are toxic to mammalian cells in general, and thus may
be growth-inhibiting in CHO cells: Sallée et al., 2014 identified toxic AA catabolites in the primary catabolic
pathway of tryptophan, the kynurenine pathway. Beltrán-Valero de Bernabé et al., 1999 and Rodríguez et al.,
2000 found toxic intermediates in the common catabolic pathway of tyrosine and phenylalanine. Additionally,
Hallen et al., 2013 suggest that the main catabolic pathway of lysine, the saccharopine pathway, produces
reactive aldehydes that are potentially toxic, as they form adducts and condensation products with proteins
and DNA. In combination, these studies highlight a broad range of potentially detrimental activities associated
with AA catabolism in CHO bioprocessing.
To address the impact of AA catabolism on CHO physiology in a progressive manner, we devised a
rational genetic engineering strategy to study the physiological response to disruption of individual AA
catabolic pathways before proceeding to disrupt multiple pathways in concert. For selection of target genes,
we utilized a network reconstruction of AA catabolism in CHO cells to follow transcriptomic changes in AA
catabolic pathways in response to protein production, which provided a rational basis for disrupting genes
using the RNA-guided Cas9 nuclease. We monitored physiological changes in terms of maximum specific
growth rate (µmax), integral of viable cell density (IVCD), and major exo-metabolite secretion in parallel shake
flask and bioreactor cultures. Our data contribute to the fundamental understanding of AA catabolism in
relation to CHO-based bioprocessing, and highlight the applicability of rational cell line engineering strategies
to reduce endogenous production of toxins derived from catabolism of AAs.
2. Materials and methods
2.1 RNA seq data generation
Three CHO cell lines; two expressing a recombinant human IgG and one not expressing any recombinant
protein (Lund et al., 2017), were cultivated in batch mode in 125 mL shake flasks (Corning). RNA was extracted
when the cells were in exponential growth phase and in stationary phase. RNA-seq was performed by
32
Multiplexed cDNA library generation using the TruSeq RNA Sample Preparation Kit v2 (Illumina, Inc., San
Diego, CA) and next-generation sequencing were performed by AROS Applied Biotechnology (Aarhus,
Denmark) using eight samples per lane in an Illumina Hiseq 2000 system for paired-end sequencing (SRA
accession: SRP073484) as described by Lund et al., 2017. RNA-Seq data were processed as described in by Lund
et al., 2017 and trimmed reads were mapped to CHO-K1 genome (assembly and annotation) released in 2014
(NCBI Accession: GCF_000223151.1) using TopHat2 version 2.0.9 (using Bowtie 2.2.0) with default settings
(Kim et al., 2013; Langmead et al., 2009). Read counts for each transcript were obtained with HTSeq count
version 0.5.4p3 using the intersection none-empty mode (Anders et al., 2015). The read counts were
normalized using EdgeR (version 3.6.8) (Robinson et al., 2010)in R (Ihaka and Gentleman, 1996). Genes with
detected counts per million in less than two samples were disregarded. Differential analyses were performed
using the GLM likelihood ratio test in EdgeR for the experiments with multiple factors. A p-value of 0.05 and
a false discovery rate < 0.05 as well as ± log2.0 fold changes were used as the default thresholds to identify the
differentially expressed genes.
2.2 Metabolic network reconstruction
For integration of transcriptomic data and selection of target genes for genome editing, we employed a CHO-
specific metabolic network reconstruction of glycolysis and AA catabolism as described previously (Ley &
Kazemi et al., 2015). Briefly, a metabolic network reconstruction of glycolysis and AA catabolism in CHO cells
was generated using mouse genomic and biochemical pathway information from the KEGG database
(Kanehisa and Goto, 2000)as starting point. To identify orthologous metabolic genes in CHO, a protein BLAST
search of AA metabolic genes from mouse was conducted against the CHO-K1 genome (Xu et al., 2011), hosted
at http://www.CHOgenome.org (Hammond et al., 2012). The resulting list of CHO genes was manually
curated for inclusion based on information from literature. The finalized reconstruction features 319 proteins
catalyzing 183 reactions with 188 metabolites.
33
2.3 Single-guide RNA target design, transfection, single cell sorting and genotype verification
Design and selection of single-guide RNA (sgRNA) target sites was performed with the online tool “CRISPy”
(Ronda & Pedersen et al., 2014). The sgRNA expression vectors were constructed as previously described by
Ronda & Pedersen et al., 2014. Prior to transfection, CHO-S suspension cells obtained from Life Technologies
were grown in CD-CHO medium (Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 8 mM
L-glutamine (Life Technologies) and 0.2 % anti-clumping reagent (Life Technologies) in a humidified shaking
incubator operated at 37°C, 5 % CO2and 120 rpm. One day prior to transfection, cells were washed and seeded
in medium without anti-clumping reagent at 5 x 105 cells/mL. Cells were transfected with expression vectors
encoding Cas9 nuclease linked to GFP via a 2A peptide (GFP_2A_Cas9) as described by Grav et al., 2015, and
sgRNAs targeting Aass, Afmid, Ddc, Gad1, Gad2, Hpd, LOC100759874, Prodh and Prodh2 individually (See
Table 1 for details), to generate single-gene knockout transfectants. For each sample, cell cultures with cell
density of 1 x 106 cells/mL in 125 mL shake flasks (Corning), were transfected with 17.7µg DNA using
FreeStyle™ MAX reagent together with OptiPRO SFM medium (Life Technologies), according to the
manufacturers recommendations. For generation of multiple gene knockout transfectants, cells were co-
transfected with equimolar amount of each plasmid encoding GFP_2A_Cas9 and selected sgRNAs, in a total
of 17.7 µg expression vector DNA. Anti-clumping reagent (0.5 %) was added one day after transfection. Two
days subsequent to transfection, clones expressing GFP_2A_Cas9 were enriched from the population of
transfectants, and single cell sorted using fluorescence activated cell sorting (FACS), as described by Grav et
al., 2015. Single cell sorted clones were cultured in 96-well U bottom plates (BD Biosciences) for 14 days and
genotypes were determined using deep sequencing analysis as described previously by Grav et al., 2015 (Miseq
primers described in Supplementary Table 1) .
2.4 Cell cultivation in shake flasks
Cultures were initiated from cryopreserved vials in preheated CD-CHO medium (Thermo Fisher Scientific)
supplemented with 8 mM L-glutamine (Gibco) and 0.2 % anti-clumping reagent (Gibco). Pre-cultures were
passaged three times during a seven-day period prior to inoculation. To avoid cell aggregation, cells were
34
passed through a 40 µm cell strainer (Sigma #CLS431750-50EA) before inoculation. For characterization of
single gene disrupted clones, cell culture was performed in 250 mL vented Erlenmeyer shake flasks (Corning,
NY, USA) with a working volume of 80 mL in a humidified shaking incubator operated at 37°C, 5 % CO2 and
120 rpm. For validation of physiological effects across multiple single gene disrupted clones, cell culture was
performed in 125 mL vented Erlenmeyer shake flasks (Corning, NY, USA) with working volume of 40 mL,
under similar conditions. All cultures were seeded at 3 x 105 cells/mL and grown for 6 days in batch culture.
Samples were drawn daily and analyzed for cell density and viability using two methods: for the
characterization of single gene disrupted clones we used the Nucleocounter NC-200 (Chemometec, Allerød,
Denmark), while the for validation of physiological effects across multiple single gene disrupted clones we used
a high-throughput method described by Hansen et al., 2015 for determining viable cell density and viability.
Briefly, a dye master mix, containing 5 μg/mL Hoechst-33342 (Life Technologies), for staining of total cell
population, and 0.4 μg/mL propidium iodide (Life Technologies), for staining of dead cells, was prepared in
CD-CHO medium supplemented with 8 mM L-Glutamine, and was transferred to a 96-well optical-bottom
microplate (Greiner Bio-One, Frickenhausen, Germany) containing 3 µl of cell suspension in a total volume
of 200 µl per well. After incubation at room temperature, the cells were imaged using the appropriate channels
in Celígo Imaging Cell Cytometer (Nexcelom Bioscience, MA, USA). Culture supernatants were analyzed for
glucose, lactate, glutamine, glutamate and ammonium using BioProfile 400 Plus (Nova Biomedical, Waltham,
MA, USA) and for AAs as described in section 2.6. Cultures were sampled for RNA and intracellular proteins
in mid-exponential growth phase. Genomic DNA was extracted to verify the culture genotype using sanger
sequencing (Eurofins Genomics) at the end of the cultures.
2.5 Cell cultivation in bioreactors
Pre-cultures were handled as described for cultivation in shake flasks. Cell culture was performed in single-use
bioreactors (Eppendorf DASbox Mini Bioreactor, Jülich, Germany) with a working volume of 250 mL.
Cultures were seeded and sampled as described in section 2.4. Temperature was maintained at 37°C, agitation
rate was fixed at 200 rpm, dissolved oxygen was maintained at 50 % of atmospheric air saturation using air, O2
35
and CO2 operated at a constant flow rate of 0.6 L/h. Culture pH was maintained at 7.15 with a deadband of
0.25 using intermittent CO2 addition to the gas mix and 2M sodium carbonate.
2.6 HPLC quantification of AAs
Supernatants were quantified for AAs using the method described by Valgepea et al. 2017, with the following
modifications: AAs were derivatized in a HPLC autosampler (Dionex Ultimate 3000) and samples were
injected into a Gemini C18 column (3 µm, 4,6 x 150 mm, Phenomenex PN: 00F-4439-E0) with a guard column
(Security Guard Gemini C18, Phenomenex PN: AJO-7597). The HPLC gradient was 5-22% B from 0-9,5 min,
kept at 22% B to 11 min, 22-35% B from 11-14 min, kept at 35% to 20 min, 35-60% B from 20-24.5 min, 24.5-
25.5% to 100% B, kept at 100% B to 27 min, decreased to 5% B at 27.1-30 min where chromatography finished.
Buffer A was 40 mM Na2HPO4, 0.02% NaN3(w/v) at pH 7.8. Buffer B was 45% (v/v) acetonitrile, 45% (v/v)
methanol and 10% (v/v) water. Flow rate was 1 mL/min from 0-26 min and 1.5 mL/min from 26.1-29 min
thereafter 1 mL/min to 30 min. Derivatized AAs were monitored using a fluorescence detector. OPA-
derivatized AAs were detected at 340ex and 450em nm and FMOC-derivatised AAs at 266ex and 305em nm.
Quantifications were based on standard curves derived from serial dilutions of an in-house prepared mixed
AA standard. The upper and lower limits of quantification were 75 and 0.5 μg/mL, respectively.
Chromatograms were integrated using Chromeleon version 7.1.3.
2.7 Calculations and statistics
Maximum specific growth rate was calculated using exponential regression of viable cell density from day 0 to
day 3. Average specific production rates of lactate and ammonium were calculated during the time interval
from day 0 to day 3, by dividing the increase in metabolite concentration by the increase in integral of viable
cell density (IVCD). Similarly, specific consumption rates of AAs were determined from day 0 to day 3. For
assessing differences in rates between gene edited and wild type clones, Levene’s test for means was used
initially to test for homogeneity of variances. The statistical test for difference between clones was performed
using Student’s or Welch’s t-test as appropriate, with significance level of a= 0.05.
36
2.8 Preparation of DNA, RNA, cDNA and qPCR experiments
Genomic DNA (gDNA) was isolated from pellets of 0.5 x 106 CHO cells using QuickExtract DNA Extraction
Solution (Epicentre, Madison, WI, USA). For DNA preparation, the cell pellet was homogenized by vortexing
in 200 µL 65°C preheated QuickExtract solution. The mixture was incubated at 65°C for 15 min and 98°C for
5 min and centrifuged at 5000 x g for 3 min. The supernatant containing DNA was used for further
experimentation.
Total RNA was extracted from at least 106 cells using Trizol reagent (Life Technologies #15596-026)
according to the manufacturer’s description. For qPCR experiments, cDNA was prepared from 500 ng
TURBO-DNase (Life Technologies #AM1907)treated RNA using the qScript Flex cDNA kit (Quanta
Bioscience #95049-100) with random priming. qPCR was performed in an Mx3005P (Agilent Technologies)
using Brilliant III Ultra-Fast SYBR® Green master mix (Agilent Technologies #600882). Primers for qPCR are
listed in Supplementary Table 1. Relative gene expression levels were calculated using the ∆∆CT method with
Gapdh has reference gene.
2.9 Sample preparation for proteomic analysis
Preparation of protein extract from CHO cells were done as previously described in Bonde et al., 2016. Liquid
chromatography was performed on an Easy-nLC system (Thermo scientific) coupled to an 75 µm x 15 cm C18
easy spray column (Thermo Scientific). Using a stepped gradient, going from 6% to 60% acetonitrile in water
over 120 minutes, the samples were sprayed into an Orbitrap Fusion mass spectrometer (Thermo Scientific).
MS-level scans were performed with Orbitrap resolution set to 60,000; AGC Target 2.0e5; maximum injection
time 50 ms; intensity threshold 5.0e3; dynamic exclusion 45 sec. Data dependent MS2 selection was performed
in Top Speed mode with HCD collision energy set to 28% and ion trap detection (AGC target 1.0e4, maximum
injection time 35 ms).The resulting data were analysed using MaxQuant with the following settings: Fixed
modifications: Carbamidomethyl(C). Variable modifications: oxidation of methionine residues. First search
mass tolerance 20 ppm and a MS/MS tolerance of 20 ppm. Trypsin was selected as enzyme and allowing one
missed cleavage. FDR was set at 0.1%. For gel-based proteomics, gel bands were excised from the gel and
37
washed first in 50% acetonitrile followed by water. After drying, the spots were in-gel-digested overnight at 37
°C with trypsin as enzyme. Peptide containing samples were analyzed on a Synapt G2 (Waters, Manchester
UK) Q-TOF instrument as previously described in Ørnholt-Johansson et al., 2017. RAW-files were analyzed
using Progenesis QI for Proteomics version 2.0.
3. Results & discussion
3.1 Selection of AA catabolic genes for genome editing
The ideal metabolic engineering strategy should increase availability of AAs for proteogenesis, while
reducing synthesis of toxic metabolites, as well as avoiding adverse effects on cellular maintenance processes.
To achieve this, we carefully selected catabolic pathways in AA metabolism for targeted disruption based on
three criteria: (i) Pathways must not be essential for cell survival. (ii) Pathways should include transaminases
or dehydrogenases, as these biochemical reactions were hypothesized to contribute to production of
ammonium and lactate, respectively. (iii) Pathways should preferably be upregulated in our network
reconstruction of AA metabolism, which integrated differential gene expression data generated from an IgG
producing cell line and a non-producing cell line. Hence, the dataset was assumed to reflect changes in gene
expression levels as a direct response to recombinant protein production, and consequently should reveal
which AA catabolic pathways contribute the most to energy metabolism, when exposed to the metabolic
burden of recombinant protein production. We targeted genes encoding the first catabolic reaction in each
pathway to avoid accumulation of potentially toxic pathway intermediates. In cases where the first catabolic
reaction was performed by isoenzymes encoded by three or more genes, we decided to target the second
catabolic reaction, to reduce the number of gRNAs needed to disrupt pathway activity.
When inspecting the gene expression landscape in AA catabolic pathways, we found that the L-
tryptophan, L-lysine, L-phenylalanine and L-tyrosine catabolic pathways were upregulated in the IgG
producing cell line (Figure 1; the complete map is found in Supplementary Figure 1), indicating that CHO cells
38
increase catabolism of said AAs when producing recombinant proteins. To disrupt these pathways, we selected
a total of four genes encoding catabolic enzymes for knock-down (Table 1). In addition to these genes, we
selected five genes in catabolic pathways of L-glutamate, L-proline and L-threonine (Table 1), as these
pathways contained dehydrogenases and/or transaminases, and thus were expected to reduce specific lactate
and ammonium. All target genes were tested for lethality by simulating their disruption using the consensus
genome-scale reconstruction of CHO cell metabolism (Hefzi et al., 2016), and found to be non-lethal when
deleted.
3.2 Generation of single gene disruptions and characterization of clone genotypes
In order to disrupt the target genes (Table 1) , we selected a Cas9-based strategy to generate indels in
5’ proximal exons in the coding region of each target gene causing out-of-frame mutations leading to
premature termination of translation and/or translation of non-functional peptides. In order to exclude the
potential impact of simultaneous production of a heterologous protein, we disrupted genes in a non-producing
cell line.
We evaluated four target sites and screened each gRNA for indel generation efficiency by PCR
amplification and deep sequencing of the targeted genomic region. The most efficient (i.e. highest ratio of indel
over wild type sequence) gRNA sequence was selected for co-transfection with GFP_2A_Cas9. An off-target
effect prediction was made based on the CHO-K1 genome (Supplementary Table 2), and found to be specific
for the target genes (at least 3 mismatches in off-target sequences for the gRNA). We enriched transfected cell
pools for cells expressing GFP-linked Cas9 nuclease using FACS, which has previously been shown to
dramatically increase the indel frequency and thus reduce downstream clone screening efforts (Grav et al.,
2015). We characterized single cell sorted clones for indels in target loci using PCR amplification and deep
sequencing. The deep sequencing identified a set of clones with indels disrupting the genes of interest (Indel
sizes found in Table 1), which were selected for further analysis (Henceforth named based on the disrupted
genes).
39
3.3 Molecular characterization of gene disruption
Introduction of targeted frame-shift mutations in genes has been shown to effectively disrupt the
activity of encoded proteins in CHO cells (Grav et al., 2015; Ronda & Pedersen et al., 2014), leading to
translation of non-functional and/or truncated target proteins. We chose to examine the molecular effect of
the disruption at both RNA and protein level.
As the mutation is introduced in the coding region, the promoter activity is assumed to be unaffected
by the mutation. Indeed, we detected active transcription of all genes using qPCR, except for Prodh2, which
was not expressed at detectable levels in the potential knockout clone (Figure 2). However, we found that most
genes were transcribed at a lower level in potential knockout clones than the wild type (Figure 2), suggesting
that the gene editing affected mRNA stability. Nonsense-mediated mRNA decay (NMD) is a cellular quality
control system that prevents translation of dysfunctional proteins by degrading mRNAs with premature stop
codons (reviewed by Lykke-Andersen and Jensen, 2015), and thus offer an explanation to the observed
decrease in mRNA abundance. Assembly of the NMD complex on a premature stop codon induces
endonucleolytic cleavage of the mRNA followed by decapping, deadenylation and complete mRNA cleavage
by general 5’-3’ and 3’-5’ exonucleases. To investigate possible impact of NMD activity on our qPCR data, we
decided to quantify mRNA abundance using two primer pairs located towards the 5’ and 3’ end of the coding
region in each gene. For four genes (Afmid, Ddc, Gad2 and Prodh) qPCR results indicated a difference in
determined mRNA levels, suggesting a possible impact of NMD on the quantification of mRNA abundance
(Figure 2 and Supplementary Figure 2), However, no clear trend was observed for the 3' versus 5' end of the
transcripts. We further attempted to characterize target proteins using LC-MS/MS, and while we detected 2900
proteins in total, but target protein levels were below the detection limit (data not shown).
In summary, the deep sequencing shows correct disruption of the genes (Table 1), and a follow-up
qPCR analysis of transcript levels showed that most genes additionally have a knock-down effect at the
transcriptional level (Figure 1), in the non-functional transcript.
40
3.4 Physiological characterization of single gene disrupted clones
To investigate potential improvement in bioprocessing derived from disrupting AA catabolic
pathways in CHO cells, we performed a physiological comparison of nine gene-disrupted clones (genotypes in
Table 1) with unedited wild type cells as control. We monitored changes in growth (i.e. µmax and IVCD) and
specific metabolic by-product secretion (i.e. qLac and qNH3) between the gene-disrupted clones and the control
in batch culture performed in three separate shake flask experiments.
As presented in Figure 3, the distribution of growth curves indicated a strong biological response to
disruption of single AA catabolic genes, seen by increased maximal viable cell density, in comparison to wild
type control, in clonal cell lines where the following genes were disrupted: Aass, Afmid, and Hpd (figure 3A);
Gad2 and LOC100759874 (figure 3B); and Prodh (figure 3C). Additionally, in comparison to the control cell
line, eight of nine gene-disrupted clones displayed increased mean µmax, up to 115 % of the wild type µmax (Figure
4). The mean IVCD was increased in 6 of 9 clones up to 136 % of wild type IVCD. For specific secretion of
metabolic by-products, we found that single gene disruptions decreased mean qLac in 4 of 9 clones (up to 119
% of wild type rate, highlighting the connection between lactate and AA metabolism (Nicolae et al., 2014)).
Similarly, mean qNH3 was decreased in 5 of 9 clones to 91 % of wild type rate, indicating a decrease in
transamination associated with AA catabolism.
When comparing the physiological impact of each gene disruption individually, we found that some
gene disruptions produced statistically significant (t-test, a< 0.05) improvements across multiple investigated
parameters (i.e. increased µmax and IVCD and reduced qLac and qNH3), while other disruptions produced minor,
but statistical insignificant improvements (Figure 4). Of particular interest was the Hpd-disrupted clone, which
improved in all investigated parameters, and the Gad2-disrupted clone, which improved in all parameters
except qNH3 (for qNH3, p-value was 0.08). Of lesser interest and effect were the other gene-disrupted clones, for
instance the Afmid-disrupted clone which featured increased µmax and IVCD, but with no significant change in
specific by-product secretion (although for qNH3, p-value was 0.08), while the Aass-disrupted and Prodh-
41
disrupted clones featured increased µmax and IVCD, respectively. These results highlighted the potential in
engineering AA metabolism towards improved bioprocessing performance. To this end, improvement of
IVCD is especially desirable, as it represents the cell-work-hours available for protein synthesis and is
correlated to protein titer (Altamirano et al., 2004).
Notably, we found that the reductions of specific lactate and ammonium secretion in the gene edited
clones were generally not reflected in absolute concentrations of the by-products in the culture medium (Figure
5). Only the Hpd knock-down clone produced lower lactate concentration of about 2-3 mM on average
throughout the culture compared to the control.
To our knowledge, this is the first study describing engineering of AA metabolism in CHO cells using
genome editing tools, which complicates the comparison of our results to other studies. However, medium
optimization studies offer insight into the physiological response to various AA concentrations and associated
uptake rates. Previous reports have demonstrated that availability of AAs above a certain threshold inhibits
cell growth (Chen and Harcum, 2005; Parampalli et al., 2007), as excess AA availability increase AA catabolism,
which result in accumulation of associated growth inhibiting compounds (Mulukutla et al., 2017). Our
observations are consistent with these reports. For example, I. J. González-Leal et al., 2011 found that threonine
had a negative effect on growth rate. In agreement with this, we found that disruption of the first gene in
threonine catabolism, LOC100759874, increased mean µmax and IVCD indicating that threonine catabolism
may produce growth inhibiting compounds. Furthermore, Mulukutla et al. 2017 found that phenylalanine,
tyrosine and tryptophan catabolism inhibit growth. Our results are consistent with this, as disruption of genes
in phenylalanine/tyrosine and tryptophan catabolism, Hpd and Afmid, respectively, significantly improved
µmax and IVCD. We found that disrupting Aass significantly increased µmax, indicating that lysine catabolism is
associated with growth inhibition. Mammalian Aass encodes a mitochondrial alpha-aminoadipic
semialdehyde synthase, a bifunctional enzyme catalyzing the first two steps in the main catabolic pathway of
lysine, the saccharopine pathway (Pena et al., 2016). The pathway produces reactive aldehydes, which are
potentially toxic, as they form adducts and condensation products with proteins and DNA (Hallen et al., 2013),
42
which may explain the observed growth benefit. Gad1 and Gad2 encode the isoenzymes glutamate
decarboxylase 1 and 2, respectively (Erlander and Tobin, 1991), which catalyze the first reaction in glutamate
degradation through a three-step anaplerotic pathway that leads to succinate (Supplementary Figure 1). To
our surprise, disrupting Gad1 and Gad2 had different effects in all physiological parameters (Figure 4). In rats,
Gad1 and Gad2 are both involved in γ-aminobutyric acid (GABA) synthesis and are expressed in distinct
subcellular locations of the rat neuronal tissues. Immunochemical analysis suggest that Gad1 is expressed in
cell bodies and dendrites, while Gad2 is predominantly expressed in nerve endings (Erlander and Tobin, 1991),
indicating that subcellular location may explain the different physiological effect of disrupting Gad1 and Gad2.
However, the subcellular location of these proteins in CHO cells has not yet been determined. Prodh and
Prodh2 encode isoenzymes that catalyze the first catabolic reaction in degradation of proline to glutamate; a
catabolic pathway consisting of two reactions (Supplementary Figure 1). Both enzymes localize to the
mitochondria in mouse (Cruz et al., 2003; Pagliarini et al., 2008), suggesting that the enzymes localize to the
mitochondria in CHO cells as well. Still, we found that disrupting Prodh and Prodh2 produced different
responses in all physiological parameters (Figure 4), despite the apparent similar catalytic function of the
enzymes, suggesting that they are not isoenzymes in a functional sense, but rather structurally similar enzymes.
3.5 Evaluation of AA Consumption Rates in Gad2- and Hpd-disrupted clones
As the analysis showed above, the two most interesting phenotypes were found in the Gad2- and Hpd-
disrupted clones. To further investigate the impact of these disruptions, we measured AA uptake rates by AA
quantification in culture. We did not find any statistically significant difference (t-test, a< 0.05) in specific AA
uptake rates between gene edited clones and wild type, suggesting that these AAs are primarily utilized for
protein synthesis instead of catabolism in the mutants (Supplementary Tables 3-4). Additionally, the fact that
AA uptake rates remained unchanged enables straight-forward engineering of AA catabolism in existing
production cell lines, since requirements for bioprocess adaptation upon engineering are minimal.
43
3.5 Physiological characterization and validation of physiological effect across multiple Gad2- or
Hpd-disrupted clones
Functional heterogeneity in CHO cell populations, also known as clonal variation, is a general challenge
in CHO cell engineering. To exclude that the improved phenotype observed in gene disrupted CHO clones
was caused by random clonal variation, we characterized five clones with either Gad2- or Hpd-disrupted, as
these gene disruptions had the biggest physiological effect and thus the most interesting leads. The clones were
generated together with the single gene-disrupted clones described above, and had all disrupting indels (Table
2). The five clones and wild type were subjected to a similar physiological characterization as the clones above,
including growth characterization (Figure 6 and Table 2), and AA consumption rates (Supplementary Tables
6 and 7). In summary, these biological replicates all confirm the phenotypes seen above. All gene-edited clones
had increased µmax and IVCD and reduced by-product secretion relative to the wild type (except for △Gad2#1,
which showed marginally increased qLac). Notably, we found a larger variation between Gad2-disrupted clones,
highlighting the importance of characterizing multiple clones when performing metabolic engineering of CHO
cells due to clonal variation.
44
3.6 Combined disruption of multiple AA catabolic pathways
To explore potential synergistic effects of disrupting multiple pathways, we targeted Aass, Afmid, Ddc
and Hpd (genes responsible for catabolism of the three aromatic AAs L-tryptophan, L-phenylalanine and L-
tyrosine as well as L-tyrosine). As these genes were found to be upregulated during protein production, we
assumed the corresponding catabolic pathways were contributing more carbon to central metabolism when
producing proteins, which makes them potential interesting engineering targets for the bio-manufacturing
industry. We transfected parental CHO-S cells with gRNAs targeting all four genes. However, we did not
isolate a clone with full disruption of all genes. Still, we isolated two clones with interesting genotypes
(Supplementary Table 7); clone 1 had indels in four genes with one remaining wild type allele of Afmid and
clone 2 had indels in all genes except Afmid. Furthermore, both clones had partial disruption of Ddc (95 %
reads from deep sequencing of PCR amplified gRNA target-locus indicated frame-shift mutation and 5 % in-
frame insertion of 105 base pairs).
To study the physiological impact of disrupting multiple AA catabolic genes, we compared the gene
edited clones (i.e. clone 1 and clone 2) to a wild type control in duplicate batch cultures in bioreactors, using
the same bioprocess parameters as for single gene disrupted clones. To our surprise, we found that µmax and
IVCD were not increased in the gene edited clones (Table 3 and Figure 7A), however both specific lactate and
ammonium secretion rates were substantially lower in gene edited clones (table 3), which – in contrast to single
gene disrupted clones – led to lower concentrations of lactate and ammonium in the cultures (figure 7B-C).
45
5. Conclusion
In combination, our results suggest that growth improvements can be achieved from disrupting single AA
catabolic pathways. In particular, disruption of Hpd and Gad2 had desirable phenotypes. However, in some
cases, the effect is sensitive to disruption of additional catabolic pathways. Even so, reduction of ammonium
and lactate secretion improves as more pathways are disrupted. Thus, we recommend combinatorial
disruption of multiple AA catabolic pathways, to identify a set of disruptions that sustain growth
enhancements while reducing lactate and ammonium secretion.
Conflicts of Interest
The authors state that D.L, H.F.K, and M.R.A have filed patent no.: WO2017EP70682, addressing some of the
findings of this manuscript. The remaining authors have no conflicts of interest.
Acknowledgements
We acknowledge Karen Katrine Brøndum and Zufiya Sukhova for technical assistance with generation of
genome edited cell lines. Moreover, we thank Sara Bjørn Petersen for cloning plasmids and Thomas Beuchert
Kallehauge for sharing his experience in design of quantitative PCR experiments, Lene Holberg Blicher for
assisting in the proteomics experiment, Mette Kristensen and Lars Boje Petersen for assisting in the HPLC
analysis. The Novo Nordisk Foundation and eCHO Systems H2020 MSC-ITN (Grant no. 642663) provided
funding for this work.
46
Author contributions
D.L, S.P., H. F. K. and M.R.A. designed experiments and wrote the manuscript. D.L., S.P., T.K.H. & T.W
performed the experiments. H.H. validated essentiality of target genes. A.M.L. provided differential gene
expression data. D.L., S.P., L.E.P., J.A., T.W., H.F.K. and M.R.A. analysed the data. All authors reviewed the
manuscript.
References
Ahn, W.S., Antoniewicz, M.R., 2012. Towards dynamic metabolic flux analysis in CHO cell cultures. Biotechnol. J. 7, 61–74. https://doi.org/10.1002/biot.201100052
Altamirano, C., Berrios, J., Vergara, M., Becerra, S., 2013. Advances in improving mammalian cells metabolism for recombinant protein production. Electron. J. Biotechnol. 16, 1–16. https://doi.org/10.2225/vol16-issue3-fulltext-2
Altamirano, C., Paredes, C., Illanes, A., Cairó, J.J., Gòdia, F., 2004. Strategies for fed-batch cultivation of t-PA producing CHO cells: Substitution of glucose and glutamine and rational design of culture medium. J. Biotechnol. 110, 171–179. https://doi.org/10.1016/j.jbiotec.2004.02.004
Anders, S., Pyl, P.T., Huber, W., 2015. HTSeq--a Python framework to work with high-throughput sequencing data. Bioinformatics 31, 166–9. https://doi.org/10.1093/bioinformatics/btu638
Andersen, D.C., Goochee, C.F., 1995. The effect of ammonia on the O-linked glycosylation of granulocyte colony-stimulating factor produced by chinese hamster ovary cells. Biotechnol. Bioeng. 47, 96–105. https://doi.org/10.1002/bit.260470112
Beltrán-Valero de Bernabé, D., Peterson, P., Luopajärvi, K., Matintalo, P., Alho, a, Konttinen, Y., Krohn, K., Rodríguez de Córdoba, S., Ranki, a, 1999. Mutational analysis of the HGO gene in Finnish alkaptonuria patients. J. Med. Genet. 36, 922–3.
Bonde, M.T., Pedersen, M., Klausen, M.S., Jensen, S.I., Wulff, T., Harrison, S., Nielsen, A.T., Herrgård, M.J., Sommer, M.O.A., 2016. Predictable tuning of protein expression in bacteria. Nat. Methods 13, 233.
Borys, M.C., Linzer, D.I., Papoutsakis, E.T., 1994. Ammonia affects the glycosylation patterns of recombinant mouse placental lactogen-I by chinese hamster ovary cells in a pH-dependent manner. Biotechnol. Bioeng. 43, 505–514. https://doi.org/10.1002/bit.260430611
Chen, P., Harcum, S.W., 2005. Effects of amino acid additions on ammonium stressed CHO cells. J. Biotechnol. 117, 277–286. https://doi.org/10.1016/j.jbiotec.2005.02.003
Cruz, S. Da, Xenarios, I., Langridge, J., Vilbois, F., Parone, P.A., Martinou, J., 2003. Proteomic Analysis of the Mouse Liver Mitochondrial Inner Membrane * □ 278, 41566–41571. https://doi.org/10.1074/jbc.M304940200
47
Erlander, M.G., Tobin, A.J., 1991. Two Genes Encode Distinct Glutamate Decarboxylases 7, 91–100. https://doi.org/10.1016/0896-6273(91)90077-D
González-Leal, I.J., Carrillo-Cocom, L.M., Ramírez-Medrano, A., López-Pacheco, F., Bulnes-Abundis, D., Webb-Vargas, Y., Alvarez, M.M., 2011. Use of a Plackett-Burman statistical design to determine the effect of selected amino acids on monoclonal antibody production in CHO cells. Biotechnol. Prog. 27, 1709–17. https://doi.org/10.1002/btpr.674
Grav, L.M., Lee, J.S., Gerling, S., Beuchert Kallehauge, T., Holmgaard Hansen, A., Kol, S., Lee, G.M., Ebdrup Pedersen, L., Faustrup Kildegaard, H., 2015. One-step generation of triple knockout CHO cell lines using CRISPR Cas9 and fluorescent enrichment. Biotechnol. J. n/a-n/a. https://doi.org/10.1002/biot.201500027
Hallen, A., Jamie, J.F., Cooper, A.J.L., 2013. Lysine metabolism in mammalian brain: An update on the importance of recent discoveries. Amino Acids 45, 1249–1272. https://doi.org/10.1007/s00726-013-1590-1
Hammond, S., Kaplarevic, M., Borth, N., Betenbaugh, M., Lee, K., 2012. Chinese Hamster Genome Database: An Online Resource for the CHO Community at www.CHOgenome.org. Biotechnol. Bioeng. 109, 1353–1356.
Hansen, H.A., Emborg, C., 1994. {I}nfluence of ammonium on growth, metabolism, and productivity of a continuous suspension {C}hinese hamster ovary cell culture. Biotechnol Prog 10, 121–124. https://doi.org/10.1021/bp00025a014
Hansen, H.G., Nilsson, C.N., Lund, A.M., Kol, S., Grav, L.M., Lundqvist, M., Rockberg, J., Lee, G.M., Andersen, M.R., Kildegaard, H.F., 2015. Versatile microscale screening platform for improving recombinant protein productivity in Chinese hamster ovary cells. Sci. Rep. 5, 18016. https://doi.org/10.1038/srep18016
Hefzi, H., Ang, K.S., Hanscho, M., Borth, N., Lee, D., Lewis, N.E., 2016. Article A Consensus Genome-scale Reconstruction of Chinese Hamster Ovary Cell Metabolism Article A Consensus Genome-scale Reconstruction of Chinese Hamster Ovary Cell Metabolism 434–443. https://doi.org/10.1016/j.cels.2016.10.020
Hiller, G.W., Mulukutla, B.C., 2015. Method of cell culture. WO 2015/140708 A1. Ihaka, R., Gentleman, R., 1996. R: A Language for Data Analysis and Graphics. J. Comput. Graph. Stat.
5, 299. https://doi.org/10.2307/1390807 Kanehisa, M., Goto, S., 2000. KEGG: kyoto encyclopedia of genes and genomes. Nucleic Acids Res. 28,
27–30. Kim, D., Pertea, G., Trapnell, C., Pimentel, H., Kelley, R., Salzberg, S.L., 2013. TopHat2: accurate
alignment of transcriptomes in the presence of insertions, deletions and gene fusions. Genome Biol. 14, R36. https://doi.org/10.1186/gb-2013-14-4-r36
Langmead, B., Trapnell, C., Pop, M., Salzberg, S.L., 2009. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 10, R25. https://doi.org/10.1186/gb-2009-10-3-r25
Lao, M.S., Toth, D., 1997. Effects of ammonium and lactate on growth and metabolism of a recombinant Chinese hamster ovary cell culture. Biotechnol. Prog. 13, 688–91. https://doi.org/10.1021/bp9602360
Ley, D., Seresht, A.K., Engmark, M., Magdenoska, O., Nielsen, K.F., Kildegaard, H.F., Andersen, M.R., 2015. Multi-omic profiling -of EPO-producing Chinese hamster ovary cell panel reveals metabolic
48
adaptation to heterologous protein production. Biotechnol. Bioeng. 112, 2373–2387. https://doi.org/10.1002/bit.25652
Lund, A.M., Kaas, C.S., Brandl, J., Pedersen, L.E., Kildegaard, H.F., Kristensen, C., Andersen, M.R., 2017. Network reconstruction of the mouse secretory pathway applied on CHO cell transcriptome data. BMC Syst. Biol. 11, 37. https://doi.org/10.1186/s12918-017-0414-4
Lykke-Andersen, S., Jensen, T.H., 2015. Nonsense-mediated mRNA decay: an intricate machinery that shapes transcriptomes. Nat. Rev. Mol. Cell Biol. 16, 665–677. https://doi.org/10.1038/nrm4063
Mulukutla, Bhanu Chandra, Jaitashree Kale, Taylor Kalomeris, Michaela Jacobs, and Gregory W. Hiller. 2017. “Identification and Control of Novel Growth Inhibitors in Fed-Batch Cultures of Chinese Hamster Ovary Cells.” Biotechnology and Bioengineering 114 (8): 1779–90.
Nicolae, A., Wahrheit, J., Bahnemann, J., Zeng, A.-P., Heinzle, E., 2014. Non-stationary 13C metabolic flux analysis of Chinese hamster ovary cells in batch culture using extracellular labeling highlights metabolic reversibility and compartmentation. BMC Syst. Biol. 8, 50. https://doi.org/10.1186/1752-0509-8-50
Pagliarini, D.J., Calvo, S.E., Chang, B., Sheth, S.A., Vafai, S.B., Ong, S., Walford, G.A., Sugiana, C., Boneh, A., Chen, W.K., Hill, D.E., Vidal, M., Evans, J.G., Thorburn, D.R., Carr, S.A., Mootha, V.K., 2008. A Mitochondrial Protein Compendium Elucidates Complex I Disease Biology 112–123. https://doi.org/10.1016/j.cell.2008.06.016
Parampalli, A., Eskridge, K., Smith, L., Meagher, M.M., Mowry, M.C., Subramanian, A., 2007. Developement of serum-free media in CHO-DG44 cells using a central composite statistical design. Cytotechnology 54, 57–68. https://doi.org/10.1007/s10616-007-9074-3
Pena, I.A., Marques, L.A., Laranjeira, A.B.A., Yunes, J.A., Eberlin, M.N., Arruda, P., 2016. Simultaneous detection of lysine metabolites by a single LC – MS / MS method : monitoring lysine degradation in mouse plasma. Springerplus 1–9. https://doi.org/10.1186/s40064-016-1809-1
Robinson, M.D., McCarthy, D.J., Smyth, G.K., 2010. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–40. https://doi.org/10.1093/bioinformatics/btp616
Rodríguez, J.M., Timm, D.E., Titus, G.P., Beltrán-Valero De Bernabé, D., Criado, O., Mueller, H. a, Rodríguez De Córdoba, S., Peñalva, M. a, 2000. Structural and functional analysis of mutations in alkaptonuria. Hum. Mol. Genet. 9, 2341–50.
Ronda, C., Pedersen, L.E., Hansen, H.G., Kallehauge, T.B., Betenbaugh, M.J., Nielsen, A.T., Kildegaard, H.F., 2014. Accelerating genome editing in CHO cells using CRISPR Cas9 and CRISPy, a web-based target finding tool. Biotechnol. Bioeng. 111, 1604–1616. https://doi.org/10.1002/bit.25233
Sallée, M., Dou, L., Cerini, C., Poitevin, S., Brunet, P., Burtey, S., 2014. The Aryl Hydrocarbon Receptor-Activating Effect of Uremic Toxins from Tryptophan Metabolism: A New Concept to Understand Cardiovascular Complications of Chronic Kidney Disease. Toxins (Basel). 6, 934–949. https://doi.org/10.3390/toxins6030934
Templeton, N., Lewis, A., Dorai, H., Qian, E.A., Campbell, M.P., Smith, K.D., Lang, S.E., Betenbaugh, M.J., Young, J.D., 2014. The impact of anti-apoptotic gene Bcl-2Δ expression on CHO central metabolism. Metab. Eng. 25, 92–102. https://doi.org/10.1016/j.ymben.2014.06.010
Thorens, B., Vassalli, P., 1986. Chloroquine and ammonium chloride prevent terminal glycosylation of immunoglobulins in plasma cells without affecting secretion. Nature 321, 618–620. https://doi.org/10.1038/321618a0
49
Walsh, G., 2014. Biopharmaceutical benchmarks 2014. Nat. Biotechnol. 32, 992–1000. https://doi.org/10.1038/nbt.3040
Xu, X., Nagarajan, H., Lewis, N.E., Pan, S., Cai, Z., Liu, X., Chen, W., Xie, M., Wang, W., Hammond, S., Andersen, M.R., Neff, N., Passarelli, B., Koh, W., Fan, H.C., Wang, J., Gui, Y., Lee, K.H., Betenbaugh, M.J., Quake, S.R., Famili, I., Palsson, B.O., Wang, J., 2011. The genomic sequence of the Chinese hamster ovary (CHO)-K1 cell line. Nat. Biotechnol. 29, 735–741. https://doi.org/10.1038/nbt.1932
Yang, M., Butler, M., 2002. Effects of ammonia and glucosamine on the heterogeneity of erythropoietin glycoforms. Biotechnol Prog 18, 129–138. https://doi.org/10.1021/bp0101334
Yang, M., Butler, M., 2000. Effects of ammonia on {CHO} cell growth, erythropoietin production, and glycosylation. Biotechnol Bioeng 68, 370–380.
Young, J.D., 2013. Metabolic flux rewiring in mammalian cell cultures. Curr. Opin. Biotechnol. https://doi.org/10.1016/j.copbio.2013.04.016
50
Tables
Table 1. List of target genes, gene description, associated AA catabolic pathway and corresponding indel sizes.
Gene symbol Gene ID Protein names Associated catabolic pathway
Indel sizes
Aass 100751161 Alpha-aminoadipic semialdehyde synthase
L-lysine -1
Afmid 100773211 Kynurenine formamidase
L-tryptophan +1
Ddc 100761742 Aromatic-L-amino-acid decarboxylase
L-tyrosine/L-phenylalanine
-29/-16
Gad1 100765882 Glutamate decarboxylase 1 L-glutamate +28
Gad2 100757642 Glutamate decarboxylase 2 L-glutamate -13
Hpd 100768220 4-hydroxyphenylpyruvate dioxygenase
L-tyrosine/L-phenylalanine
+1
LOC100759874 100759874 L-threonine 3-dehydrogenase L-threonine +1
Prodh 100750856 Proline dehydrogenase 1 L-proline -20
Prodh2 100773901 Proline dehydrogenase 2 L-proline -22
51
Table 2. Comparison of growth characteristics (µmax and IVCD), by-product secretion rates (qNH3 and qLac) and indel sizes in respective target genes across wild type, Hpd- and Gad2-disrupted clones. Values represent mean ±one standard deviation of three biological replicates.
Wildtype △Gad2 #1 △Gad2 #2 △Gad2 #3 △Hpd #1 △Hpd #2
µmax
(Day-1)
0,89±0,13 0,9±0,1 0,97±0,11 1,01±0,12 0,99±0,21 1,06±0,16
qNH3 (pmol/cell/day)
1,2±0,2 1,16±0,25 0,89±0,17 1,03±0,26 1,12±0,13 1±0,11
qLac (pmol/cell/day)
4,29±1,03 4,37±1,05 3,25±0,62 3,34±0,85 3,68±0,49 3,77±0,56
IVCD (106cell*h/mL)
598,89±9,26 649,65±43,77 835,7±35,01 897,43±31,83 717,01±57,55 759,41±29,94
Indel size - -7 -25 -13 +1 +1
Table 3. Comparison of growth characteristics and by-product secretion rates across wild type and multiple gene disrupted clones. Replicate values are listed.
Wildtype Clone 1 Clone 2
µmax (Day-1) 0,95 / 0,89 0,97 / 1,03 0,84 / 0,8
qNH3 (pmol/cell/day) 0,91 / 1 0,63 / 0,63 0,82 / 0,8
qLac (pmol/cell/day) 4,57 / 4,63 3,47 / 3,51 3,01 / 4,04
IVCD(106cell*h/mL) 632 / 552 733 / 703 580 / 550
52
Figures
Figure 1. Simplified overview of targeted AA catabolic pathways. Multiple reactions have been collapsed for
simplicity. Circles next to reaction arrows indicate genes encoding the corresponding enzymes that catalyze
each reaction. Circle colors indicate differential gene expression levels comparing an IgG producing cell line
to a non-producing cell line (i.e. log fold-change [IgG / WT]). Grey circle color indicates missing gene
expression data. AAs are colored blue, redox active compounds are colored red. We targeted the genes: Aass,
Afmid, Ddc, Gad1, Gad2, Hpd, LOC100759874, Prodh and Prodh2, which are indicated with bold circles.
53
Figure 2. Gene expression levels of target genes in single gene disrupted clones. Transcription rate was
quantified using two primer pairs targeting coding regions upstream and downstream relative to the gRNA
target site. Gene expression levels are normalized to the wild type expression. Error bars indicate standard
deviation of three biological replicates.
54
Figure 3. Profile of cell growth and viability of single gene knockout clones and wild type cells. Growth
curves were generated from three separate experiments where sets of clonal cell lines with disrupted single
genes Aass, Afmid, Ddc and Hpd (A), Gad1, Gad2, Prodh2 and LOC100759874 (B) and Prodh (C) were
cultivated in parallel with CHO-S wild type cells. Error bars indicate SEM calculated for three biological
replicates.
55
Figure 4. Comparison of maximum specific growth rate, integral of viable cell density, specific lactate and
ammonium secretion across nine knockout clones. Values have been normalized to the wild type. Stars
indicate statistically significant difference to the wild type. Error bars indicate standard deviation of three
replicates.
56
Figure 5. Extracellular metabolite profiles of lactate and ammonium in single gene knockout clones and
wild type. Data were generated in three separate experiments in biological triplicates. Error bars indicate SEM.
57
Figure 6. Growth and viability of multiple Gad2 (Left) and Hpd (Right) disrupted clones and wild type
cells. Growth curves were generated from the cultivation of multiple clonal cell lines with disrupted single
genes Gad2 (left) and Hpd (right) were cultivated in parallel with CHO-S wild type cells. Error bars indicate
SEM calculated for three biological replicates.
58
Figure 7. Comparison of wild type and multiple AA catabolic pathway disrupted clones in bioreactors. A
Growth and viability. B Extracellular lactate concentrations. C Extracellular ammonium concentrations. Error
bars indicate standard deviation.
59
Paper III – Physiological study of CRISPR/Cas9-mediated disruption of branched-chain amino acid transaminases in CHO cells
60
Physiological study of CRISPR/Cas9-mediated disruption of branched-chain amino acid
transaminases in CHO cells
Sara Pereira1, Daniel Ley1,2, Mikkel Schubert1, Lise Marie Grav1, Helene Faustrup Kildegaard1,3, Mikael
Rørdam Andersen4
1The Novo Nordisk Foundation, Center for Biosustainability, Technical University of Denmark, Kongens
Lyngby, Denmark, 2Current address: AGC Biologics A/S, Vandtårnsvej 83, 2860 Søborg, Denmark, 3Current
address: Novo Nordisk, Department of mammalian expression, Måløv, Denmark, 4Department of
Biotechnology and Biomedicine, Technical University of Denmark, Kongens Lyngby, Denmark
Correspondence: [email protected] for enquiries on the computational analysis and strategy,
[email protected] for correspondence on the molecular biology.
Author contributions
S. P. performed experiments, performed data analysis and wrote the manuscript, S.P., D.L. and H.E.F.
and M.R.A. designed experiments, and H.F.K and M.R.A supervised the project and edited the
manuscript. L. M. G. generated the parental cell line essential for running the experiments.
Keywords: Chinese hamster ovary cells, branched-chain amino acids, nutrient metabolism, by-product,
CRISPR/Cas9
Abbreviations
BCAA(s) – Branched-chain amino acid(s); Bcat1 – branched-chain amino acid transaminase 1, cytosol;
Bcat2 – branched-chain amino acid transaminase 2, mitochondrial; Cas9 – CRISPR-associated protein
9; CHO – Chinese hamster ovary; CRISPR – Clustered Regularly Interspaced Short Palindromic
Repeats; FACS – Fluorescence-activated cell sorting; gRNA – guide RNA; indel – insertion or deletion;
IVCD – Integral Viable Cell Density; mAbs – monoclonal antibodies; RMCE – Recombinase Mediated
Cassette Exchange; sgRNA – single guide RNA; WT – wild type; VCD – Viable cell density; µmax –
Specific growth rate
61
Abstract
In recombinant protein expression using Chinese hamster ovary (CHO) cells, chemically defined media
contain essential amino acids such as the branched-chain amino acids (BCAAs) leucine, isoleucine, and
valine. Availability of amino acids is critical as building blocks for protein synthesis. However,
breakdown of amino acids can lead to build-up of toxic intermediates and metabolites that decrease cell
growth, productivity, and product quality. BCAA catabolic reactions hamper the usage of BCAAs for
protein synthesis. In this work, we studied the effects of disrupting the genes responsible for the first
step of BCAA catabolism: branched-chain aminotransferase 1 (Bcat1) and branched-chain
aminotransferase 2 (Bcat2). We evaluated the effect of disrupting the genes individually and in
combination and examined the effects in CHO cells stably expressing mCherry and non-producer host
cells. Our results show a cell-line dependent effect as Bcat1 disruption improves cell growth in producer
cells but not in non-producers. Bcat2-disruption has a minor negative effect on growth in producer
cells, and no effect in non-producers. Simultaneous Bcat1 and Bcat2 disruption results in improved cell
growth in producer cells. The changes in by-product metabolism are cell line-, clone- and producer-
dependent. Overall, our results show that the effects of targeting Bcat1 and Bcat2 are cell line-dependent,
and linked to the burden of recombinant protein expression.
Introduction
Chinese hamster ovary (CHO) cells are the preferred host for the production of therapeutic
glycoproteins. Total sales of monoclonal antibodies (mAbs) reached $103.4 billion in 2017. Eighty-four
percent of mAbs that reached the market between 2014 and 2017 was expressed in CHO cells (Walsh,
2018). Despite being a widely used host cell line, CHO cells have an inefficient metabolism caused by
the high uptake rates of glucose and glutamine that lead to the formation toxic by-products such as
lactate and ammonia that affect cell growth (Lao and Toth, 1997). Additionally, amino acid catabolism
leads to the build-up of additional toxic intermediates and metabolites that affect cell growth
productivity and product quality attributes negatively (Ahn and Antoniewicz, 2013; Mulukutla et al.,
2017). Therefore, there is a need to improve the cellular performance towards reduction of the amounts
of toxic and inhibitory metabolites and improvement of metabolic efficiency.
Amino acid availability is critical in recombinant protein expression as these are the building blocks
required for protein synthesis. In our previous studies (Paper II of this thesis), we tested the effects of
62
disrupting genes in the amino acid catabolic pathways of glutamate, lysine, tryptophan, proline,
phenylalanine and tyrosine, and showed that in a number of cases, we could reduce the formation of
by-products lactate and ammonia, and improve cell growth, that are preferred phenotypes in
bioprocessing. This work focuses on the catabolism of branched-chain amino acids (BCAAs), as BCAA
catabolism hampers the usage of the BCAAs for protein synthesis. BCAAs participate in nutrient-
sensitive signaling pathways, such as phosphoinositide 3-kinase-protein kinase B (PI3K-AKT),
mammalian target of rapamycin (mTOR) (Nie et al., 2018). The enzymes participating in the first
reaction of BCAAs catabolism are shared by leucine, isoleucine, and valine. The initial step is catalyzed
by branched-chain amino acid transaminase 1 (Bcat1) present in the cytosol and branched-chain amino
acid transaminase 2 (Bcat2) active in mitochondria followed by branched-chain alpha-keto acid
dehydrogenase a (Bckdha) and branched-chain alpha-keto acid dehydrogenase b (Bckdhb). Defects in
the branched-chain amino acid transaminases are associated with health conditions linked to
accumulation of BCAAs in the urine and serum, such as hypervalinemia and hyperleucine-
isoleucinemia (Wang et al., 2015), while mutations in the branched-chain alpha-keto acid
dehydrogenase (BCKD) enzyme complex are associated with maple syrup urine disease characterized
by the accumulation of toxic metabolic intermediates from the BCAA catabolism (Blackburn et al.,
2017). Based on this knowledge, we selected Bcat1 and Bcat2 as targets for engineering. Recently,
another lab demonstrated that Bcat1 deletion has a beneficial effect in a producer cell line (Mulukutla
et al., 2019).
In this work, we test the hypothesis that disrupting the genes involved in the BCAAs catabolism can
induce phenotypic changes in CHO cell metabolism. In addition, we hypothesize that disrupting Bcat1
and Bcat2 will increase the availability of BCAAs leucine, isoleucine, and valine that can be used for
biomass formation and ultimately improve recombinant protein production. We used the
CRISPR/Cas9 system to target Bcat1 and Bcat2 in two cell lines: CHO-S wild type (WT) cells and T2_6
cells expressing mCherry. Finally, in our experiments, we evaluated the physiological effects of these
disruptions by following cell growth, nutrient uptake, and by-product formation.
63
Materials and methods
Cells
In this study, two background cell lines were used: CHO-S (Thermo Fisher Scientific) and a CHO-S-
derived parental cell line, T2_6, harboring a stably integrated LoxP/Lox2272 landing pad for expression
of genes of interest via recombinase-mediated cassette exchange (RMCE) (Petersen et al., 2018). CHO-
S and T2_6 parental cells were maintained in CD-CHO medium (Life Technologies) supplemented with
8 mM L-Glutamine (Thermo Fisher Scientific) and 0.2% anti-clumping agent (Gibco). The cells were
cultivated in 125 mL Erlenmeyer shake flasks (Corning Inc., Acton, MA), incubated at 37°C, 5% CO2 at
120 rpm and passaged every 2-3 days. Viable cell density (VCD) and viability were monitored using the
NucleoCounter NC-200 Cell Counter (ChemoMetec).
Single guide RNA target design, transfection and generation of knockout cell lines
Single guide RNA target sites were identified using the CRISPy online tool with the genomic sequences
of Bcat1 (NW_003613704.1) and Bcat2 (NW_003614570.1) and the respective expression vectors were
constructed as previously described (Ronda et al., 2014). The purity and concentration of the plasmids
carrying the sgRNA sequences was determined using NanoDrop (Thermo Scientific). The sequences of
sgRNAs are presented in Table S1.
The CHO-S wild type and parental cell line (T2_6) cells were maintained as described above, were
seeded at a VCD of 1x106cells/ml to be transfected, according to the manufacturers recommendations,
using FreeStyle MAX reagent (Gibco) and OptiPRO SFM medium (Gibco), with sgRNA targeting Bcat1
or Bcat2 and GFP_2A_Cas9 plasmid at a 1:1 ratio, to generate single-gene knockout transfectants (Grav
et al., 2017, 2015). To generate cells with simultaneous double gene disruption, cells were co-transfected
with equal mass of each plasmid encoding GFP_2A_Cas9 and two gRNAs plasmids, each encoding a
sequence targeting Bcat1 or Bcat2, by adding the volume corresponding to the equal mass of each
plasmid to the transfection preparation mix. As controls, CHO-S cells were treated similarly but without
adding any plasmid (WT) to the transfection mix, or just with the plasmid carrying Cas9 (referred to as
hereafter as T2_6+Cas9 cells). Transfection efficiency was assessed 48h post-transfection by measuring
the fluorescence of GFP using BD FACSJAzz (BD Biosciences) flow cytometer, and GFP positive clones
were single-cell sorted onto 384-well plates (Corning). After 10-14 days, the growing colonies were
transferred onto 96-well plates (Corning). Genomic DNA was extracted from cells growing in 96-well
plates with QuickExtract™ (Epicentre®) and the genotypes were determined using deep sequencing
64
analysis (Ronda et al., 2014). For validation purposes, the genotypes were reanalyzed using deep
sequencing using samples from the seed train, before the batch cultivation. Primers used for clone
screening are presented in Table S2. Gene off-targets were located following an approach based on
previous work by Ronda et al. (Ronda et al., 2014): All 13 bp k-mers upstream of PAM sites in the CHO
K1 (GCF_000223135.1) genome were indexed, and k-mers with no more than 3 mismatches compared
to those of gRNAs used in this study were collected. Genomic features overlapping the off-target gRNAs
were located in the CHO K1 genomic annotation, excluding those which the overlap was located in
introns (Tables S3 and S4).
Batch cultivation
Clonal cells derived from CHO-S and from T2_6 parental cells were inoculated at 5 x 105 cells/ml and
cultivated in 125 ml or 250 mL Erlenmeyer shake flasks (Corning Inc., Acton, MA) a working volume
of 40 or 60 ml, respectively, of CD CHO medium supplemented with 8 mM L-Glutamine (Thermo
Fisher Scientific) incubated at 37°C, with 5% CO2, shaking at 120 rpm. Viable cell density (VCD) and
viability were monitored using the NucleoCounter NC-200 Cell Counter (ChemoMetec). Cultivations
were stopped when viability was below 70%. Specific growth rates (µmax) and integral viable cell density
(IVCD) were determined for all clones. Samples of 1 ml were collected every day throughout the batch
cultivation, centrifuged at 2000 g, 5 minutes and the pellet and supernatant fractions were separated,
stored at -20°C further downstream analysis.
Metabolite profile and determination of specific rates
Extracellular concentrations of glucose, lactate, glutamine, glutamate and ammonium present in
supernatant samples collected throughout the batch cultivation were monitored using BioProfile 400
Plus (Nova Biomedical, Waltham, MA, USA). Specific consumption or production rates of each
metabolite were determined in exponential phase of culture.
HPLC quantification of amino acids
Supernatants from exponential growth phase were prepared for quantification of amino acids using the
method described by Valgepea et al. (Valgepea et al., 2017), with the following modifications: amino
acids were derivatized in an HPLC autosampler (Dionex Ultimate 3000), and samples were injected into
a Gemini C18 column (3 µm, 4,6 x 150 mm, Phenomenex PN: 00F-4439-E0) with a guard column
(SecurityGuard Gemini C18, Phenomenex PN: AJO-7597). Buffer A was 40 mM Na2HPO4, 0.02%
NaNO3 (w/v) at pH 7.8. Buffer B was 45% (v/v) acetonitrile, 45% (v/v) methanol and 10% (v/v) water.
65
The HPLC gradient was 5-22% B from 0-9.5 min, kept at 22% B to 11 min, 22-35% B from 11-14 min,
kept at 35% to 20 min, 35-60% B from 20-24.5 min, 24.5-25.5% to 100% B, kept at 100% B to 27 min,
decreased to 5% B at 27.1-30 min where chromatography finished. The flow rate was 1 mL/min from 0-
26 min and 1.5 mL/min from 26.1-29 min; thereafter, 1 mL/min until 30 min. Derivatized amino acids
were monitored using a fluorescence detector. OPA-derivatized amino acids were detected at 340ex and
450em nm and FMOC-derivatised amino acids at 266ex and 305em nm. Quantifications were based on
standard curves derived from serial dilutions of an in-house prepared mixed amino acid standard. The
upper and lower limits of quantification were 75 and 0.5 μg/mL, respectively. Chromatograms were
integrated using Chromeleon version 7.1.3.
Specific consumption rates of BCAAs were determined during the exponential phase of the culture.
Data and statistical analysis
One-way ANOVA was used to assess the differences in specific rates between control and edited cells,
with a significance level set to α = 0.05.
Results
Generation of single disruption of Bcat1 and Bcat2 in CHO-S cells
In this work, genes involved in the catabolism of branched-chain amino acid pathways, Bcat1 and Bcat2,
were targeted for engineering using CRISPR/Cas9 system in CHO-S cells. Single-cell clones were
obtained by fluorescence-activated cell sorting (FACS) of transfected cells and were genotyped using
next-generation sequencing of the knock out target locus for identification of frameshift insertion and
deletion (indel) mutations. Disruptions were confirmed, and indel sizes are shown in Table 1.
Table 1 – Size of insertion and deletion mutations obtained by genotyping of Bcat1 and Bcat2 disruption in CHO-S-derived clones performed via deep sequencing of PCR amplified gRNA target loci.
CHO-S-Cas9+Bcat1 CHO-S-Cas9+Bcat2
Clones A2 B2 B10 F5 F10 G7
Bcat1 -1 -4/-2 -17 - - -
Bcat2 - - -32 -154/-10 -154/+1
66
Study of physiological changes in CHO-S engineered clones in batch cultivations: cell growth
In order to assess the influence of disrupting Bcat1 and Bcat2 in cell growth, batch cultivations were
performed. Clonal cell lines with disrupted Bcat1 and Bcat2 were cultivated, and cell growth and
viability (Figure 1) were measured every day followed by determination of specific growth rate (µmax)
and IVCD (Figure 2). The results show that one of the three Bcat1-disrupted clones displays a significant
change in cell growth. Clone Bcat1_B2 reached the highest maximal VCD (above 10 x 106 cells/mL on
day 6) and displays higher viability, compared to the remaining characterized clones that have a
maximal VCD around 7.5 x 106 cells/ml. When the mitochondrial version of Bcat, Bcat2, was disrupted,
the cell growth profile of the engineered clones remained similar to wild type cells. Furthermore, no
significant changes in µmax are observed in neither case of disruption of Bcat1 and Bcat2 genes (Figure
2). IVCD values of Bcat1_B2 clone showed a statistically significant increase compared to WT cells
(Figure 2B), while disrupting the Bcat2 gene in CHO-S cells lead to an overall mean increase of IVCD,
although not statistically significant (Figure 2D).
Figure 1 – Growth profiles of CHO-S WT cells and Bcat1- and Bcat2-disrupted cells. CHO-S cells with disruption of Bcat1: A –
Viable cell density and B – Viability. CHO-S cells with disruption of Bcat2: C – Viable cell density and D – Viability. Viable cell
density (VCD) and viability were determined during simultaneous batch cultivation. Error bars indicate standard deviations
(triplicate cultures).
67
Figure 2 – Specific growth rates (µMax) calculated from day 1 to day 3 of cultivation and terminal integral cell viable cell density
(IVCD) determined for parallel batch cultivations of CHO-S Bcat1- and Bcat2-disrupted clones compared to CHO-S wild type
cell. Error bars indicate standard deviations (triplicate cultures).
Study of physiological changes in CHO-S engineered clones in batch cultivations: nutrient and by-
product profile
We continued the study of cell physiology by assessing whether the effects of disruption of Bcat1 and
Bcat2 also lead to changes in nutrient and by-product metabolism during the batch cultivation. For that,
we examined the exo-metabolite profile of glucose, glutamine, glutamate, lactate and ammonium
(Figure 3) and determined the specific rates of growth and metabolite formation and consumption
(Figures S1). Small changes in the concentration profiles of glucose and glutamine of Bcat1-disrupted
cells compared to WT are observed (Figure 3). In the fast-growing clone (Bcat1_B2), the specific glucose
consumption rate along with specific lactate secretion rate were decreased in the same clone. These rates
remained unchanged in the two remaining Bcat1 disrupted clones. For all the Bcat2-disrupted clones,
both specific glucose and specific lactate production rate remained similar to WT (Figure S1). Further,
clone Bcat2_G7 shows a significant increase in specific glutamine consumption rate while the other two
remained unchanged. The increased glutamine uptake rate does not seem to have significant influence
on the specific production rate of ammonium of Bcat2_G7 since it remains at WT levels (Figure S1).
Only clone Bcat2_F10 shows a significant decrease in specific ammonium production rate. These results
WT
Bcat1
_A2
Bcat1
_B2
Bcat1
_B10
0.00
0.25
0.50
0.75
1.00
1.25
µMax (day-1)
WT
Bcat2
_F5
Bcat2
_F10
Bcat2
_G7
0.00
0.25
0.50
0.75
1.00
1.25
WT
Bcat1
_A2
Bcat1
_B2
Bcat1
_B10
500
600
700
800
900
1000P = 0.0058
**
IVCD (106 cells*h/ml)
WT
Bcat2
_F5
Bcat2
_F10
Bcat2
_G7
500
600
700
800
900
1000ns
p > 0.05
Bcat1
Bcat2
68
show a possible increase in metabolic efficiency in a fast-growing Bcat1-disrupted clone as lower uptake
rates of glucose lead to slower lactate secretion. The single disruption of Bcat1 and Bcat2 in CHO-S cells
does not affect the nutrient and by-product metabolism in the majority of tested host cells.
Figure 3 – Exo-metabolite profile of CHO-S wild type and clonal cell lines with Bcat1- and Bcat2-disrupted cells. Concentrations
of glucose, lactate, glutamine, ammonium, and glutamate determined from the analysis of supernatants samples collected from
batch cultivations from day 0 to day 7. Error bars represent standard deviations (n=3).
0 2 4 6 8
0
10
20
30
40
50
Glu
cose
(mM
)
CHO-S_WTCHO-S-Bcat1_A2CHO-S-Bcat1_B2CHO-S-Bcat1_B10
2 4 6 80
10
20
30
40
Lact
ate
(mM
)
0 2 4 6 80
2
4
6
8
10
Glu
tam
ine
(mM
)
0 2 4 6 80
5
10
15
Am
mon
ium
(mM
)
0 2 4 6 80
1
2
3
4
5
Time (days)
Glu
tam
ate
(mM
)
0 2 4 6 8
0
10
20
30
40
50CHO-S WTCHO-S_Bcat2-F5CHO-S_Bcat2-G7CHO-S_Bcat2-F10
2 4 6 80
10
20
30
40
0 2 4 6 80
2
4
6
8
10
0 2 4 6 80
5
10
15
0 2 4 6 80
1
2
3
4
5
Time (days)
Bcat1 Bcat2
69
Generation of Bcat1- and Bcat2-disrupted clones in T2_6 parental cells
In order to verify the reproducibility of the results obtained after the disruption of Bcat1 and Bcat2 in
CHO-S wild-type cells, we selected a CHO-S cell line stably expressing mCherry, the T2_6 cell line
(Petersen et al., 2018), for disrupting the same genes. The T2_6 cell line is derived from CHO-S and
carries a targeted integrated LoxP/Lox2272 landing pad in T2 site, and cells express mCherry as a model
protein. Additionally, T2_6 cells have reduced variation in cell growth after recombinase-mediated
cassette exchange (RMCE) with donor plasmids carrying recombinant therapeutic proteins and
subsequent sub-cloning (unpublished data). This producer cell line, generated in-house by our
colleagues, was selected as it allows for studying the effects of engineering the selected target genes, in
this case, Bcat1 and Bcat2, in a stable and producer cell line relevant for industrial applications.
Furthermore, the rationale behind the choice of mCherry as reporter protein relates to the experimental
design used to generate the parental cell line performed similarly to as described by Grav et al. (Grav et
al. 2018). Briefly, to generate cells with targeted integration of mCherry and RMCE elements, the
CRISPR/Cas9 system was used to do double-strand break in the pre-selected locus, followed by
homology-directed repair for which a repair template is required. The repair template consisted of a
donor plasmid carrying the sequences of (from 5’ to 3’): 5’ homology arm, promoter EF1α, mCherry
within the sequences needed for RMCE (LoxP and Lox2272), BGH polyA signal sequence followed by
SV40 promoter driving the expression of antibiotic resistance marker for Neomycin (NeoR), the
sequence of respective polyA signal (SV40 pA) and 3’ homology arms sequence. The donor plasmid also
encoded sequences of CMV promoter, a second fluorescence marker (Zsgreen1) and BGH poly A signal
– all these downstream of the 3’ homology arm sequence. This way the cells harboring the repair
template in the genome were selected and bulk sorted as a mCherry positive/ZsGreen1-DR negative cell
population.
Bcat1 and Bcat2 were targeted in T2_6 cell line using the same workflow as for the CHO-S cells.
However, in this case, we also attempted simultaneous targeting of both Bcat1 and Bcat2 via co-
transfection of two plasmids, each encoding sgRNA targeting Bcat1 or Bcat2 with a plasmid encoding
GFP_2A_Cas9. Control T2_6 cells (T2_6+Cas9 cells) were made by transfection only with the plasmid
encoding Cas9. All cells were single-cell sorted using FACS and were expanded in a similarly. Gene
targeting was validated using deep sequencing of the locus targeted by each sgRNA (Table 2). For the
double gene targeting, full disruption of both Bcat1 and Bcat2 was achieved for only clone F-A8, as the
remaining indels are in-frame.
70
Table 2 – Sizes of insertion and deletion mutations obtained by genotyping of Bcat1 and Bcat2 knock out T2_6-derived clones performed via deep sequencing of PCR amplified gRNA target loci. Group of T2_6 derived-clones: (D) disruption of Bcat1, (E) disruption of Bcat2 and (F) disruption of Bcat1&Bcat2.
T2_6-Bcat1 (D) T2_6-Bcat2 (E) T2_6-Bcat1+Bcat2 (F)
Clone D-A2 D-A12 D-B11 D-C2 E-B6 E-B11 E-C11 E-A5 F-A7 F-A8 F-B4 F-B12
Bcat1 -20 -7 -25 -25 - - - - 7/-3* -22 25/-3* 33*/ -6*
Bcat2 - - - - -13 -13 -19 -67 -13 +1 -13 +1
* indicates in-frame indel mutations.
Study of physiological changes in engineered T2_6 clones in batch cultivations: cell growth
To study the physiological impact of disrupting Bcat1 and Bcat2, we characterized the clones in batch
cultivations in shake flask and monitored changes in viable cell density (Figure 4) and viability (Figure
S2). Control T2_6 cells and sub-cloned cells with single or combinatorial disruption of Bcat1 and Bcat2
were cultivated in parallel. Growth rates and IVCD are presented in Table S5. Bcat1-disrupted T2_6
cells showed higher peak VCD for 3 of 4 clones (T2_6+Bcat1_D-A12, T2_6+Bcat1_D-B11, and
T2_6+Bcat1_D-C2). The highest peak VCD was observed for the T2_6+Bcat1_D-B11 that reached 14.4
x 106 cells/mL on day 4, while the highest peak VCD attained by control clones was 10x106 cells/mL
(T2_6+Cas9_C-A8). For Bcat2-disrupted T2_6 cells, only 1 of 4 clones displayed a peak VCD that was
higher than that of control clones (clone T2_6+Bcat2+E-B11 with peak VCD 14x106 cells/mL). For
combinatorial disruptions, the verified double-disruption mutant (T2_6+Bcat1&2_F-A8) showed the
highest of all measured VCDs (16x106 cells/mL) and the highest IVCD (almost 1500x106 cells*h/mL).
We determined the specific growth rates (µMax) and identified the fast grower clones in each group:
T2_6+Bcat1_D-B11 and T2_6+Bcat1_D-C2 of Bcat1 disrupted cells and T2_6+Bcat1&2_F-A7,
T2_6+Bcat1&2_F-A8 and T2_6+Bcat1&2_F-B12 for double disrupted cells, while most Bcat2-disrupted
clones behaved like the wild type clones.
71
Figure 4 – Growth profiles of T2_6 Bcat1 and Bcat2-disrupted cells. T2_6+Cas9 cells were used as control and T2_6 cells with A
– Bcat1 and B – Bcat2 single gene disruption, and C – simultaneous disruption of Bcat1 and Bcat2. Viable cell densities (VCD)
determined during simultaneous batch cultivation (Triplicate cultures). Error bars represent standard deviation (n=3).
Study of physiological changes in Bcat1- and Bcat2-engineered T2_6 cells in batch cultivations:
nutrient and by-product profile
We characterized physiological changes linked to the disruption of Bcat1 and Bcat2 in T2_6 cells by
monitoring the concentration profiles of the by-products lactate and ammonium (Figure 5) and
nutrients glutamate, glutamine, and glucose (Figure S3) obtained from in batch cultivations. Overall,
ammonium and lactate profiles were quite similar between gene-disrupted clones and the controls. One
exception was a higher ammonium concentration in two of the Bcat2-disrupted clones, which was not
seen in the Bcat1&Bcat2-disrupted mutant. Bcat2-disrupted cells converted glutamine a bit faster than
the control (Figure S3). Glutamate concentrations increased for clones T2_6+Bcat1_D-A2,
T2_6+Bcat2_E-A5 and T2_6+Bcat2_E-C11, and T2_6+Bcat1&2Bca1&2_F-B4 but had a decreasing
trend in the remaining clones.
We determined the specific rates of consumption of glucose and glutamine and production of lactate
and ammonium for each clone (Figure S4). The analysis showed reduced rates of lactate production in
T2_6+Bcat1_D-A12, T2_6+Bcat1_D-B11, T2_6+Bcat1&2_F-A8, and Bcat1&2_F-B12. Besides, specific
ammonium production rate decreased in clones T2_6+Bcat1_D-A12, T2_6+Bcat1_D-B11,
T2_6+Bcat1_D-C2 and T2_6+Bcat1&2_F-A8. Then we proceeded to group the clones according to
disrupted gene: a control group of T2_6+Cas9 clones (T2_6+Cas9_C), Bcat1- (T2_6+Bcat1_D), Bcat2-
(T2_6+Bcat2_E) and Bcat1&2-disrupted cells (T2_6+Bcat1&2_F) (Figure S5). This analysis revealed
that the consumption rates of glucose and glutamine were similar to control in groups Bcat1- and
72
Bact1&2, but increased in group Bcat2, while lactate secretion was significantly increased in T2_6+Bcat2
cells.
Figure 5 – Exo-metabolite profile of ammonium (A-C) and lactate (D-F) determined from the analysis of supernatants samples
collected from batch cultivations of T2_6 cells and T2_6 cells-derived clonal cell lines with Bcat1- and Bcat2-disrupted genes, from
day 0 to day 6. Curves represent mean values, and error bars represent standard deviation (n=3).
Branched chain amino acid metabolism in engineered clones
Since Bcat1 and Bcat2 catalyze the first step in the degradation pathways of BCAAs, we were interested
in the changes in concentration of isoleucine, leucine, and valine in the exponential phase of the
cultivation. We measured the metabolism of L-leucine (Figure 6A-C), L-isoleucine (Figure 6D-F), and
L-valine (Figure 6G-I) by performing HPLC analysis of supernatant samples obtained from day 0 to day
3. Our results showed that the control cells presenting higher initial BCAAs concentrations.
Furthermore, in order to investigate whether the engineering strategy had resulted in changes in the
uptake rates of BCAAs, we determined the specific consumption rates for each clone, shown in Table
S6, and each group of clones with the same disruption (Figure S7). Our results show a decrease in the
uptake rates of all three BCAAs in all engineered cells after analyzing the rates for each group.
Specifically, the decrease in isoleucine uptake was significant in T2_6+ Bcat1_D and T2_6+Bcat1&2_F,
and a significant decrease in valine uptake in T2_6+ Bcat2_E was also observed. The specific leucine
consumption rates in each group of engineered cells was also reduced.
73
Figure 6 – Concentrations of extracellular of BCAAs (A-C) leucine, (D-F) Leucine and (G-J) valine measured in supernatant
samples obtained from day 0 to day 3 of cultivation of Bcat1- and/or Bcat2-disrupted clones derived from T2_6 cells. Supernatant
samples were diluted 20 X in ultrapure water and internal standard mix. The samples from day 1 were also analyzed but the data
points are not included in the analysis due to poor separation of the analytes. T2_6+Cas9 clonal cells were used as control. The
curves represent average values, and error bars represent standard deviation (n=2).
Discussion
In this study, we engineered the catabolism of BCAAs by targeting Bcat1 and Bcat2 for disruption in
two cell lines. The genes Bcat1 and Bcat2 that catalyze the first step of BCAA degradation were disrupted
in CHO-S WT host cells and in T2_6 cells expressing mCherry using CRISPR/Cas9 system. We assessed
74
the physiological changes resulting from disruption of Bcat1 and Bcat2, by following cell growth, by-
product formation, and BCAA consumption during batch cultivations.
In order to achieve optimal and reproducible performance, CHO cells are cultivated in chemically
defined media. Often, the formulations supply the cells with excessive amounts of nutrients (e.g., amino
acids) that leads to the accumulation of metabolic by-products. These can have a negative impact on the
cell metabolism of CHO cells used for the production of recombinant therapeutic proteins. Motivating
this approach is that BCAAs, specifically leucine, have a key role in mechanistic target of rapamycin
complex 1 (mTOR1) activation, responsible for several signaling pathways linked to cell growth,
autophagy amongst other activities (Nie et al., 2018). By reducing or eliminating their expression, we
hypothesized that changes in cell growth and availability of these amino acids for protein synthesis
would occur, while preventing the formation of toxic metabolic intermediates.
To assess the impact of the disruption in CHO-S WT cell physiology, we have characterized cell growth
and metabolite profile of Bcat1- and Bcat2-disrupted cells compared to WT in batch cultivations. We
observed that both Bcat1- and Bcat2-disrupted cells have specific growth rates comparable to WT cells
(Figure 2), although a Bcat1-disrupted clone significantly improved IVCD. Based on this, we conclude
that disruption of Bcat1 and Bcat2 has a mild positive effect in a CHO-S background, which is contrary
to what is observed in some cancer cells (Ananieva and Wilkinson, 2018). Furthermore, the recent work
of Mulukutla et al. (Mulukutla et al., 2019) shows an increase in cell growth when Bcat1 was disrupted,
which we only see as a mild effect here. Moreover, we see that Bcat1 disruption in CHO-S cells with
reduced lactate formation resulting from in slightly lower glucose consumption, lactate formation while
the lactate secretion in Bcat2-disrupted cells similar to WT.
Next, we replicated the experiment carried out in CHO-S WT cells by targeting Bcat1 and Bcat2 for
disruption on T2_6 cells expressing mCherry from a strong promoter as a model protein. mCherry
expression was present for all T2_6 cells and not for CHO-S WT cells, based on red fluorescence
measurements performed using a cytometer (data not shown). Here we see an improved cell growth in
the Bcat1-disrupted cells, but decreased growth in most Bcat2-disrupted clones. Moreover, a confirmed
Bcat1&2 disrupted mutant showed a significant increase in cellular growth, compared to the control
clone with highest cell growth profile. There is a significant reduction of specific lactate production rate
of 2 of 4 of Bcat1-disrupted clones and in specific ammonium production rate in one of the clones with
the same disruption, showing that Bcat1 disruption leads to a reduction specific production rate of by-
products in a clone dependent manner.
75
It has been seen that suppression of Bcat1 results in lower secretion of glutamate in glioma cells (Tönjes
et al., 2013) as Bcat1 and Bcat2 catalyze the first step in the degradation pathways of BCAAs and the
transamination reaction uses α-ketoglutarate to form glutamate. We thus expected to find changes in
glutamate levels in cells with disrupted Bcat1/2, especially in Bcat2-disrupted cells as it is expressed in
the mitochondria. However, the glutamate profiles of disrupted clones are within the variation of the
controls, both for disrupted clones derived from CHO-S and mCherry-expressing (T2_6). It does seem
to be a trend however, that glutamate concentrations increase for cultures with lower VCDs.
To complete our physiological study, we followed the changes in concentration of leucine, isoleucine,
and valine in the exponential phase of the cultivation (day 0 to day 3) using HPLC analysis of amino
acids present in supernatant samples and calculated the uptake rates. At a first look, the concentrations
BCAAs present in the supernatant is higher in the control cell line than in engineered cells, contrary to
the reports in the literature (Wang et al., 2015), where increased concentration of BCAAs is seen in
mutated Bcat2. Possibly the different response is specific to a whole organism relative to cell culture.
When evaluating the specific consumption rates of each BCAA, the decrease in leucine uptake rates in
the group of disrupted mutants was not significant, while the uptake rates of isoleucine are significantly
lower for Bcat1 and Bcat1&2 disruptions, and valine uptake rate is also significantly lower in Bcat2
disruption. The results for Bcat1 are in line with the work Mulukutla et al. (Mulukutla et al., 2019),
where Bcat1 disruption in CHO cells showed a decrease in consumption of all three BCAAs. In that
work, a different cell line, expressing a monoclonal antibody at high levels is used, which could explain
some of the differences, as such a cell line would have a higher metabolic load than our cell line.
Cells with increased cell growth, but with low metabolite consumption rate and decreased by-product
secretion have an efficient metabolism. T2_6 clones with Bcat1 and Bcat1&2 disruption displaying
increased cell growth rates also displayed a reduction in glutamine and isoleucine specific consumption
rates, accompanied by reduction the specific production of ammonium and, to a smaller extent, of
lactate is observed in these clones. Surprisingly, targeting Bcat2 in T2_6 cells resulted in detrimental
changes in cell growth, and increased consumption of nutrients and production of by-products. These
effects were not seen when we engineered CHO-S WT cells, showing that Bcat1 and Bcat2 have cell line-
and clone-dependent effects.
76
Conclusions
In summary, Bcat1 disruption does not change cell growth in a CHO-S non-producer background,
while it improves cell growth in T2_6 cells producing mCherry. Bcat2 disruption does not change
growth in CHO-S non-producer cells and causes a minor reduction in cell growth in T2_6 cells. The
single targeting of Bcat1 and Bcat2 result in changes in by-product profile in CHO-S cells. Bcat1
disruption reduces lactate secretion although both effects are small, while Bcat2-disrupted cells behave
like WT. In T2_6 cells, lactate and ammonium production is reduced in Bcat1 disruption but not in
Bcat2. In a Bcat1&2 double-disrupted mutant in the T2_6 background, the phenotype was similar to
that of Bcat1 disruption, as it seems to be much stronger than the phenotype of Bcat2 disruption. Finally,
we showed that BCAA the specific consumption rate of isoleucine is reduced upon Bcat1- and Bcat1&2
disruption in T2_6 cells but not for the other BCAAs. Overall, our results allow us to conclude that
Bcat1 disruption may improve cell growth and the effects of targeting Bcat1 and Bcat2 are cell line and
clone-dependent. The metabolic effect is also dependent on the additional burden of expression of a
recombinant protein.
Acknowledgments
The authors acknowledge Sara Petersen Bjørn for cloning of plasmids, Karen Katrine Brøndum and
Zulfiya Sukhova for technical assistance with generation of part of genome edited cell lines, Nachon
Charanyanonda Petersen for the assistance in FACS analysis and single cell sorting, Mette Kristensen
and Lars Boje Petersen for assisting in the HPLC analysis. The authors S.P., H.F.K. and M.R.A. thank
the Marie Skłodowska-Curie Actions under the EU Framework Programme for Research and
Innovation for eCHO systems ITN (Grant no. 642663) for funding this work. S.P., D.L, L.M.G. and
H.F.K. additionally thank the Novo Nordisk Foundation (Grant no. NNF10CC1016517) for the
support.
References
1. Walsh, G. Biopharmaceutical benchmarks 2018. Nature Biotechnology 36, 1136–1145 (2018).
2. Lao M-S, -S. Lao M, Toth D. Effects of Ammonium and Lactate on Growth and Metabolism of
a Recombinant Chinese Hamster Ovary Cell Culture. Biotechnol Prog. 1997;13: 688–691.
3. Mulukutla, B.C., Mitchell, J., Geoffroy, P., Harrington, C., Krishnan, M., Kalomeris, T., Morris,
C., Zhang, L., Pegman, P., Hiller, G.W., 2019. Metabolic engineering of Chinese hamster ovary cells
77
towards reduced biosynthesis and accumulation of novel growth inhibitors in fed-batch cultures.
Metab. Eng. 54, 54–68.
4. Mulukutla BC, Kale J, Kalomeris T, Jacobs M, Hiller GW. Identification and control of novel
growth inhibitors in fed-batch cultures of Chinese hamster ovary cells. Biotechnol Bioeng. 2017;114:
1779–1790.
5. Ahn WS, Antoniewicz MR. Parallel labeling experiments with [1,2-(13)C]glucose and [U-
(13)C]glutamine provide new insights into CHO cell metabolism. Metab Eng. 2013;15: 34–47.
6. Nie C, He T, Zhang W, Zhang G, Ma X. Branched Chain Amino Acids: Beyond Nutrition
Metabolism. Int J Mol Sci. 2018;19. doi:10.3390/ijms19040954
7. Wang XL, Li CJ, Xing Y, Yang YH, Jia JP. Hypervalinemia and hyperleucine-isoleucinemia
caused by mutations in the branched-chain-amino-acid aminotransferase gene. J Inherit Metab Dis.
2015;38: 855–861.
8. Blackburn P, Gass J, Vairo FP e., Farnham K, Atwal H, Macklin S, et al. Maple syrup urine
disease: mechanisms and management. Appl Clin Genet. 2017;10: 57–66.
9. Petersen SD, Zhang J, Lee JS, Jakociunas T, Grav LM, Kildegaard HF, et al. Modular 5’-UTR
hexamers for context-independent tuning of protein expression in eukaryotes. Nucleic Acids Res. 2018;
doi:10.1093/nar/gky734
10. Ronda C, Pedersen LE, Hansen HG, Kallehauge TB, Betenbaugh MJ, Nielsen AT, et al.
Accelerating genome editing in CHO cells using CRISPR Cas9 and CRISPy, a web-based target finding
tool. Biotechnol Bioeng. 2014;111: 1604–1616.
11. Grav LM, Lee JS, Gerling S, Kallehauge TB, Hansen AH, Kol S, et al. One-step generation of
triple knockout CHO cell lines using CRISPR/Cas9 and fluorescent enrichment. Biotechnol J. 2015;10:
1446–1456.
12. Grav LM, la Cour Karottki KJ, Lee JS, Kildegaard HF. Application of CRISPR/Cas9 Genome
Editing to Improve Recombinant Protein Production in CHO Cells. Methods Mol Biol. 2017;1603: 101–
118.
13. Valgepea K, Loi KQ, Behrendorff JB, Lemgruber R de SP, Plan M, Hodson MP, et al. Arginine
deiminase pathway provides ATP and boosts growth of the gas-fermenting acetogen Clostridium
autoethanogenum. Metab Eng. 2017;41: 202–211.
14 Grav LM, Sergeeva D, Lee JS, Marin de Mas I, Lewis NE, Andersen MR, et al. Minimizing Clonal
Variation during Mammalian Cell Line Engineering for Improved Systems Biology Data Generation.
ACS Synth Biol. 2018;7: 2148–2159.
78
15. Ananieva EA, Wilkinson AC. Branched-chain amino acid metabolism in cancer. Curr Opin
Clin Nutr Metab Care. 2018;21: 64–70.
16. Tönjes M, Barbus S, Park YJ, Wang W, Schlotter M, Lindroth AM, et al. BCAT1 promotes cell
proliferation through amino acid catabolism in gliomas carrying wild-type IDH1. Nat Med. 2013;19:
901–908.
79
Paper IV – A targeted study of stable overexpression of Glucose-6-phosphate dehydrogenase (G6pd) in CHO-S cells: effect on cell growth and protective properties against ROS inducers and cytotoxic agents
80
A targeted study of stable overexpression of Glucose-6-phosphate dehydrogenase (G6pd)
in CHO-S cells: effect on cell growth and protective properties against ROS inducers and
cytotoxic agents
Sara Pereira (1), Lise Marie Grav (1), Tune Wulff (1), Helene Faustrup Kildegaard (1)(2), Mikael
Rørdam Andersen (3)
Affiliations:
(1) The Novo Nordisk Foundation, Center for Biosustainability, Technical University of Denmark, Kongens
Lyngby, Denmark, (2) Current address: Novo Nordisk A/S, Department of mammalian expression, Måløv,
Denmark, (3) Department of Biotechnology and Biomedicine, Technical University of Denmark, Kongens
Lyngby, Denmark
Correspondence: [email protected]
Author contributions
S.P designed and performed the majority of experiments, conducted the data analysis and wrote the
manuscript. L.M.G, H.F.K, M.R.A helped design experiments. H.F.K and M.R.A supervised the project.
T.W performed the proteomics analysis. All authors revised and commented on the manuscript.
Keywords: Chinese Hamster ovary cells, Glucose-6-phosphate, overexpression, RMCE, cell line
engineering, cellular stress
Abbreviations
BiP – Binding of immunoglobulin protein
CHO – Chinese Hamster ovary
CHOP – C/EBP homologous protein
ER – Endoplasmic Reticulum
FACS – Fluorescence-activated cell sorting
GOI – Gene of interest
GSH – reduced glutathione
81
GSSG – oxidized glutathione
G6pd – Glucose-6-phosphate dehydrogenase
H2O2 - Hydrogen peroxide
LFQ - Label Free Quantification
mAb(s) – Monoclonal Antibody(ies)
NaBu – Sodium Butyrate
NaCl – Sodium chloride
PPP – Pentose Phosphate Pathway
RMCE – Recombinase Mediated Cassette Exchange
ROS – Reactive Oxygen Species
UPR – Unfolded Protein Response
USER – uracil-specific excision reagent
VCD – Viable cell density
Abstract
CHO cells are driven to express high amounts of recombinant therapeutic proteins. These result in
different types of cellular stresses that the cell cannot cope with. Examples are unfolded protein response
(UPR) in the ER, linked to expression of proteins classified as “difficult-to-express. This leads to the
formation of reactive oxygen species (ROS). Oxidative stress can be halted when glutathione (GSH)
reacts with ROS, forming glutathione disulfide (GSSG). GSH regeneration is limited by NADPH
availability. This co-factor is mainly generated in the rate limiting step of Pentose Phosphate Pathway
(PPP), catalysed by Glucose-6-phosphate dehydrogenase (G6pd). G6pd supplies the cell with NADPH
required for fighting oxidative stress, miscellaneous biosynthetic pathways, including biosynthesis of
several amino acids, and is related to cell growth. Overexpression of G6pd has been shown to confer
resistance to ROS and to improve growth in many other cell types. In this study, we investigate the effect
of overexpressing G6pd, as this has been shown in many other systems to have a beneficial effect on
protein production and oxidative stress. First, we inserted the G6pd gene sequence into a stably
integrated landing pad located in a pre-selected genomic locus of a CHO-S derived parental cell line.
Thereafter, we tested the effect of this transformation on cellular physiology, by assessing cell growth
and exo-metabolite profile, and specific rates. Furthermore, we cultivated the cells in the presence of
82
100 µM H2O2, a ROS inducer, and 0.5 mM sodium butyrate (NaBu), a cytotoxic agent to test whether
G6pd overexpression improves protection against cellular stress. Contrary to results in many other
organisms, our results show that overexpression of G6pd did not improve cell growth nor changed the
metabolite profile significantly, and no additional protective capacity was observed in the engineered
CHO cells.
1. Introduction
Chinese Hamster Ovary (CHO) cells are the preferred mammalian host for the production of
recombinant therapeutic proteins. Examples of products expressed in CHO cells include erythropoietin
(EPO), blood coagulation factors, such as factor IX, and monoclonal antibodies (mAbs) (Walsh 2014).
The market of therapeutic recombinant proteins presents cumulative sales values, ranging between $107
to $140 billion from 2010 to 2013 (Walsh 2014). CHO cells have several advantages compared to
microbial or other mammalian cells (Wells and Robinson 2017). CHO cells have the ability to perform
complex post-translational modifications similar to those found in human proteins, such as
glycosylation, which is considered to be a critical quality attribute by the regulatory authorities.
The availability of genomic sequences of the Chinese Hamster and CHO cell lines (Lewis et al. 2013;
Rupp et al. 2018; Xu et al. 2011), other ‘omics (reviewed in (Stolfa et al. 2018) and cell engineering tools,
is advancing the CHO cell line engineering field. These tools can guide the study of important genes
that underlie an optimal host cell phenotype (reviewed by Fischer et. al. in (Fischer, Handrick, and Otte
2015)). However, a general issue for host cells is the generation of stress when the cells produce certain
recombinant proteins. In this study, the issue of limited redox precursor availability and generation of
oxidative stress is of particular interest. Thus, a strategy to overcome this issue in commonly used in
other protein-producing cell types, is to overexpress the Pentose Phosphate Pathway (PPP). This is
commonly done by overexpressing glucose-6-phosphate dehydrogenase (G6pd), which is part of the
first irreversible step in the PPP producing the co-factor NADPH (Davy, Kildegaard, and Andersen
83
2017). The function of G6pd is related to cell growth, as the PPP generates sugar precursors required
for the biosynthesis of nucleotides used for DNA synthesis/replication and plays an important role in
one carbon metabolism. Another crucial role of G6pd is connected to supplying the cell with NADPH
required to scavenge reactive oxygen species causing oxidative stress to the cell. This occurs via the
oxidation of 2 molecules of reduced glutathione (GSH) into oxidized form (GSSG), in reactions
catalysed by enzymes from the Glutathione peroxidase (Gpx) and Glutathione-S-Peroxidase (GST)
families (Lu 2009). Simultaneously, an enzymatic redox reaction (catalysed by glutathione reductase)
regenerates glutathione to its reduced form GSH, while NADPH is reduced to NADP+. Hence, an
important role of NADPH is to help the cells fight oxidative stress. The relevance of glutathione
pathways in CHO cell factories has previously been reviewed by our group (Pereira, Kildegaard, and
Andersen 2018). Glutathione may play a role in protein folding by providing a suitable redox
environment for protein folding or by directly participating in the reduction of some proteins (Ellgaard,
Sevier, and Bulleid 2018). The high expression of recombinant proteins, especially those classified as
difficult-to-express, can lead to accumulation of misfolded or unfolded proteins in the ER. Misfolded of
unfolded proteins can trigger the unfolded protein response (UPR) and increase the appearance of
reactive oxygen species (ROS) (Cao and Kaufman 2014; Walter and Ron 2011), which are also generated
the mitochondria.
Addition of antioxidants is a widely used method to scavenge oxidative stress in the cell (Ilnicka et al.
2014; Ercal et al. 1996; Skała et al. 2016; Ha et al. 2017), although cell line engineering approaches have
been employed to address ER stress and increase mAb productivity (Haredy et al. 2013; Ku et al. 2008).
The strategy of overexpressing G6pd has to our knowledge not previously been tested in CHO cells. In
this work, we investigated the effect G6pd overexpression has on cell growth and metabolism, and if it
displays protective properties against ROS inducers and cytotoxic agents. We stably overexpressed the
Chinese Hamster G6pd gene in a CHO-S-derived master cell line harboring a landing pad in a well-
defined integration site (Petersen et al. 2018). We assessed whether increased expression of G6pd
84
induced changes in physiology by measuring viable densities of two G6pd overexpressing cell pools and
their nutrient consumption and production of toxic metabolites. We further tested if the cell pools
displayed resistance to stress by exposing them to stress inducing chemicals such as reactive oxygen
species hydrogen peroxide (H2O2) that leads to oxidative stress and sodium butyrate (NaBu). NaBu is a
highly cytotoxic compound that is widely used for cell cycle arrest and to boost productivity (Field and
Brown 1990; Seong Lee, Lee, and Lee 2012; Kim and Lee 2000; Ganne et al. 1991; Kantardjieff et al. 2010;
Sunley and Butler 2010). Under the tested set-up, stably overexpression of G6pd showed no significant
effect on cell growth, nor any significant protective properties against oxidative stress. Further analyses
are necessary to conclude what effect overexpression of the PPP may have on CHO cells as the current
results indicate that overexpression of G6pd alone does not induce metabolic changes linked to cell
growth and resistance to cellular stress; overexpression of other genes participating in the PPP might be
required.
2. Materials and methods
2.1 Plasmid construction
Promotorless donor plasmids for recombinase-mediated cassette exchange (RMCE) were constructed
via uracil-specific excision reagent (USER) cloning method (Davy, Kildegaard, and Andersen 2017;
Lund et al. 2014). The backbone (pJ204), lox2272 and loxP sequences were amplified from plasmid
loxP-mCherry-lox2272-BGHpA (Grav et al. 2018). The G6pd gene was amplified from cDNA prepared
from RNA extracted from CHO-S wild type cells (named G6pd-1) and additionally synthesized as a
gBlock based on G6pd annotation NM_001246727 from NCBI (https://www.ncbi.nlm.nih.gov) (named
G6pd-2), the sequences are listed in supplementary Table 1. Three donor plasmids were constructed by
assembling the backbone with either the G6pd-1, G6pd-2 or the 3Xstop+SpA sequence (made of 3
subsequent stop codons followed by a small synthetic polyadenylation signal sequence (Levitt et al.
1989)), flanked by 5’ LoxP and 3’ Lox2272. All primers used for PCR amplification and USER cloning
are listed in Table 2. After assembly, the plasmids were transformed into E. coli Mach1 competent cells
85
(Life Technologies), verified by Sanger sequencing and purified using NucleoBond Xtra Midi EF
(Macherey-Nagel) according to manufacturer’s instructions. For Cre recombinase expression, a PSF-
CMV-CRE recombinase expression vector was used (OGS591, Sigma-Aldrich).
2.2 Cells
A CHO-S-derived parental cell line with a defined RMCE landing pad expressing mCherry in T2 site
were used in this study (Petersen et al. 2018). The landing pad consists of the following parts: 5’ and 3’
homology arms, the promoter Elongation factor 1-alpha (EF-1α), mCherry coding sequence with LoxP
sequence at 5’ end and Lox2272 sequence at 3’ end, Bovine Growth Hormone Polyadenylation
(BGHpA) and NeoR expression cassette (pSV40-NeoR-SV40pA). The cells were maintained in CD-
CHO medium (Life Technologies) supplemented with 8 mM L-Glutamine (Thermo Fisher Scientific)
and 0.2 % anti-clumping agent (Gibco), and was cultivated in 125 mL Erlenmeyer shake flasks (Corning
Inc., Acton, MA), incubated at 37°C, 5% CO2at 120 rpm and passaged every 2-3 days. Viable cell density
(VCD) and viability were monitored using the NucleoCounter NC-200 Cell Counter (ChemoMetec).
2.3 Generation of RMCE-based CHO cell pools
The parental cell line (T2_6) was seeded at a viable cell density (VCD) of 1x106 cells/mL and transfected
with promoterless donor plasmid and Cre-recombinase vector in a 3:1 ratio (w:w) in 6-well plate
(Corning) using FreeStyle MAX transfection reagent (Gibco). Control cell lines (CHO-S wild type and
parental mCherry cell line) were treated similarly, but did not receive exogenous DNA. Cell pools were
passaged at least 3 times after transfection and bulk sorted after 7 days using Fluorescence-activated cell
sorting (FACS) FACSjazz (BD Biosciences). The mCherry expressing parental cell line was used as
gating control for mCherry negative cells. The RMCE efficiency ranged between 0.85-2.5% based on the
percentage of mCherry negative cells (Table S1). The bulk sorted cell pools were expanded.
86
2.4 Extraction of genomic DNA and insert PCR
The presence of correct RMCE in the cell pools were verified by insert PCR and Sanger sequencing of
the inserted G6pd-1, G6pd-2GOI or the 3xstop+SpA sequence. Pellets of approximately 1.5 x 106 cells
were sampled from cells growing in mid-exponential growth phase of pre-culture and used for genomic
DNA extraction. The amplification of the insert was done by using primers designed to amplify the
sequentially the inserted a) EF1α promoter region, loxP sequence, GOI, lox 2272 and Bovine growth
hormone (BGH) polyadenylation signal sequences inserted (Out-Out) and b) EF1α promoter region,
loxP sequence, and partial sequence of the insert GOI (Out-In). Primers are presented in Table S2. The
PCR products were purified by electrophoresis prepared in the following conditions: 1% agarose in 1x
TAE buffer, 400mA, 80-100 V, 40 minutes, GeneRuler 1 kb DNA Ladder (Thermo Scientific). The
bands with the expected length were excised and purified using NucleoSpin Gel and PCR Clean-up
(Macherey-Nagel). The purified product was then confirmed to contain the correct sequence by Sanger
sequencing (Eurofins Genomics).
2.5 RNA extraction and cDNA first strand synthesis
Approximately 2.5 x 106 cells were harvested from cells from pre-culture while growing in exponential
phase, while 1x106 cells were sampled from the batch cultivation on day 5. The cells were centrifuged at
2000 g for 5 min and the supernatant was discarded, while the pellet was stored in -80°C. Total RNA
was extracted from the pellets using the RNeasy plus kit (Qiagen) following the manufacturer
instructions. RNA concentrations were measured with NanoDrop 2000 (Thermo Scientific).
Complementary DNA first strand synthesis was performed using Maxima First Strand cDNA Synthesis
Kit for RT-qPCR with dsDNase (Thermo Scientific) using oligo-dT and random hexamers mix as
priming strategy.
2.6 RT-qPCR for analysis of gene expression
The relative expression of the GOIs and mCherry was determined using cDNA template prepared from
pre-culture samples. A master mix was prepared using TaqMan Multiplex Master Mix (Applied
87
Biosystems) and TaqMan Gene Expression Assays (Thermo Scientific), that include target-specific pre-
designed primers and probes that span exon-exon boundaries in order to ensure amplification of cDNA
and not genomic DNA. Primers and probes are presented in Table S2. Amplification was executed with
the following conditions: 50°C for 2 min, 95°C for 10 min; 40X: 95°C for 15s, 60°C for 1 min. The fold
changes in expression were determined using ddCT method, and normalization to two reference genes
(Fkbp1p1a and Gnb1) The RT-qPCR was performed using QuantStudio 5 Real-Time PCR System
(Applied Biosystems). Each experiment included no template controls in every PCR run and had 3
replicates. Additionally, RT-qPCR was used for assessment of changes in ER stress and apoptosis-
related markers after induced cellular stress. The running conditions were the same as described above
with the following changes: custom made assays with probes spanning exon/exon boundaries specific
to CHOP, Caspase 3 and Caspase 7 were mixed in a master mix using TaqMan Multiplex Master Mix
(Applied Biosystems) while BiP probe was mixed with Gene Expression Master Mix (Applied
Biosystems) and TaqMan Gene Expression Assays (Thermo Scientific), according to manufacturer’s
instructions using cDNA synthesized from on samples collected on day 5, as described in the batch
cultivation in the presence of stress inducers. For BiP, the normalization was made to Fkbp1p1a as
reference gene.
2.7 Batch cultivation for characterization of cell pools
Cell pools overexpressing G6pd-1 and G6pd-2, non-producing cell pools (3xstop+SpA) and control cell
pools (CHO-S WT and mCherry parental cells) were inoculated at 3 x 105 cells/ml. They were cultivated
in 125 mL Erlenmeyer shake flasks (Corning Inc., Acton, MA) in a working volume of 40 ml of cell
culture media comprising CD CHO medium supplemented with 8 mM L-Glutamine (Thermo Fisher
Scientific) and incubated at 37°C, 5% CO2, shaking at 120 rpm. Viable cell density and viability were
monitored using the NucleoCounter NC-200 Cell Counter (ChemoMetec) and cultivations were
stopped when viability reached below 60% or for a maximum of 8 cultivation days. Specific growth rate
was determined for all clones. A 1 ml sample of cell suspension was collected every day and centrifuged
88
at 2000 g, 5 minutes. The pellet and supernatant fractions were then separated and stored at -20°C for
further downstream analysis.
2.8 Metabolite profile and determination of specific rates
Extracellular concentrations of glucose, lactate, glutamine, glutamate and ammonium present in
supernatant samples were monitored using BioProfile 400 Plus (Nova Biomedical, Waltham, MA,
USA). Specific consumption or production rates of each metabolite were determined in exponential
phase of culture.
2.9 Batch cultivation of recombined cell pools in the presence of oxidative stress inducer
and cytotoxic chemicals
CHO-S wild type cells, control parental mCherry expressing cell line, 3xstop+SpA cell pool, and G6pd-
1 and G6pd-2 expressing cell pools, were cultivated in batch mode as described above, except for a few
changes. Cultivations were carried out using 250 ml shake flasks (Corning Inc., Acton, MA) and 60 ml
cell growth media working volume. On day 3, the cells were transferred to 6-well plates with media
supplemented with 1 µM H2O2, 5 mM NaBu, 5 mM NaCl and a similar volume of sterile filtered milliQ
H2O was used as control. Pellets of 1x106 cells were collected on days 5 for RNA, stored at -80°C. until
further downstream analysis.
2.10 Sample preparation for proteomic analysis
Preparation of protein extract from CHO cells were done as previously described in (Bonde et al. 2016).
Liquid chromatography was performed on a Cap-LC system (Thermo scientific) coupled to an 75 µm x
15 cm 2µm C18 easy spray column (Thermo Scientific). The flow rate was set to 1.2 µl and using a
stepped gradient, going from 4% to 40% acetonitrile in water over 50 minutes, the samples were sprayed
into an Orbitrap Q Exactive HF-X mass spectrometer (Thermo Scientific). MS-level scans were
performed with Orbitrap resolution set to 60,000; AGC Target 1.0e6; maximum injection time 50 ms;
intensity threshold 5.0e3; dynamic exclusion 25 sec. Data dependent MS2 selection was performed in
89
Top 12 mode with HCD collision energy set to 28% (AGC target 1.0e4, maximum injection time 22
ms). The resulting data were analyzed using MaxQuant with the following settings: Fixed modifications:
Carbamidomethyl (C). Variable modifications: oxidation of methionine residues. First search mass
tolerance 20 ppm and a MS/MS tolerance of 20 ppm. Trypsin was selected as enzyme and allowing one
missed cleavage. FDR was set at 0.1%. And data was searched against the Chinese hamster database
retrieved from Uniprot with proteome Id UP000001075.
2.11 Statistical analysis in RT-qPCR
We used one-way ANOVA to assess the differences in gene expression (determined by RT-qPCR as
mentioned in section 2.6), between edited and parental cells (cultivated as mentioned in Section 2.7)
and between cultivation conditions (control, H2O2, NaBu and NaCl) for cells obtained (as described in
Section 2.9), with a significance level set to α= 0.05.
3. Results
3.1 Insertion of G6pd via recombinase mediated cassette exchange
In order to test whether G6pd overexpression – like in other organisms – has a positive effect on protein
secretion stress, we stably expressed G6pd in a CHO-S derived cell line harboring a landing pad with
the EF-1α promoter upstream LoxP site, the BGHpA signal sequence downstream the Lox2272 site and
pSV40-NeoR-SV40pA used for generating the cell line (Petersen et al. 2018), referred to as the parental
cell line. We generated recombineering-ready constructs of two G6pd sequences, one from the CHO-S
genome (G6pd-1) and one based on ncbi annotation NM_001246727 (G6pd-2), and a stop codon
sequence (3xstop+SpA) that serves as a non-producer control. These sequences were then exchanged
with the mCherry sequence in the landing pad of the parental cell line using Cre/Lox based RMCE (Grav
et al. 2018). In the parental cell line, mCherry is flanked by LoxP and Lox2272 sites. The mCherry serves
as a selection marker for cells where RMCE has taken place, as these should be mCherry negative upon
FACS. The RMCE efficiencies for the G6pd constructs were 0.85% and 0.87%, and for the stop-codon
90
sequence 2.5% (Figure S1 and Table 1). To validate that mCherry has been exchanged with G6pd-1,
G6pd-2 or 3xstop+SpA sequences in the bulk sorted cell pools, the inserted sequences were confirmed
by PCR (Suppl. Material Fig. S2). This was further confirmed by Sanger sequencing of PCR-amplified
bands. Additional bands with lower intensity and of similar size as the mCherry sequence are observed.
These indicate that neither RMCE nor the enrichment step using FACS bulk cell sorting of mCherry
negative cells were 100% efficient. Another band of approximate 1000 bp length is also present,
indicating that recombineering of the lox sites has taken place - leaving some cell without any donor
DNA within the Lox sites. This confirms that part of the bulk sorted cell populations have received the
correct insert, and should be expressing G6pd-1, G6pd-2, or contain the 3xstop+SpA sequence.
3.2 Test of recombinant gene expression using qPCR
Figure 1 – Determination of mRNA expression of G6pd (A) and mCherry (B) using RT-qPCR. A 2.5- and 5-fold
increase in G6pd transcript levels was observed relative to the parental cell line (expressing mCherry), in cells
transfected with promoterless plasmids carrying the Chinese Hamster G6pd sequence from cDNA (G6pd-1) and
synthesized from ncbi annotation NM_001246727.1 (G6pd-2), respectively. In all recombined cells residual
mCherry expression relative to the parental cell line (mCherry) was observed. The fold changes in expression were
determined using ddCT method. The error bars represent standard deviations in technical replicates (n=3).
91
In order to assess the change in G6pd and mCherry expression after the RMCE event, we analyzed the
cell lines using a TaqMan assay (Figure 1). Here, we used a probe spanning an exon/exon boundary of
G6pd (verified against the recently published Chinese Hamster genome (Rupp et al. 2018)). The results
were normalized to the geometric mean of the two reference genes, Gnb1 and Fkbp1a, as these have
been reported to be stable in CHO cells (Brown et al. 2018). We observed an 2.5- and 5-fold increase in
G6pd in G6pd-1 and G6pd-2 respectively (Figure 1 - A). In all recombined cells, residual mCherry
expression relative to the parental cell line (mCherry) was observed (Figure 1 - B), which may be a result
of FACS sorting efficiency of mCherry negative cells. These results confirm an increased transcription
of G6pd in cells post-RMCE, despite the residual mCherry expression.
3.3 Determination of G6pd expression at protein level using LC-MS-based proteomics
In order to clarify whether the overexpressed G6pd is being translated, protein analysis of G6PD was
performed. We started by analyzing whole cell lysates using SDS-Page in reduced and non-reduced
forms to see if a clear change in the migration pattern was observable (Suppl. Material Fig. S2). However,
this analysis revealed itself inconclusive. Proteomics analysis was used instead, as it is a more precise
and powerful tool to detect expressed proteins. Cell pellets from cultivation day 3 were processed as
described in (Bonde et al. 2016) and were analyzed using LC-MS. The results in Figure 2 - A, show an
increase in relative label free quantification (LFQ) intensity of G6pd expression in the sample with
G6pd-2 cells in comparison to the parental cell line, showing that G6pd-2 cells express G6pd at much
higher levels. There is negligible variation in relative LFQ intensity of G6pd among the parental cell line,
non-producer cells, wild type CHO-S cells and even the G6pd-1 cells. The proteomics analysis also
shows residual expression of mCherry in all recombined cell pools (Figure 2 -B).
92
Figure 2 – Determination of protein expression using untargeted LC-MS proteomics. The charts represent relative
Label Free Quantification (LFQ) intensity of G6pd (A) and mCherry (B). LFQ intensity was normalized to parental
cell line expressing mCherry. The charts represent single measurements. An increase in G6PD protein levels were
only observed for the G6pd-2 expressing cell pool (A). All recombined cells showed residual mCherry protein
levels, relative to the parental cell line (mCherry) (B).
3.4 Growth profiles of G6pd overexpressing cells
In order to study the changes in physiology, primarily in cell growth, we cultivated 5 cell pools in parallel
batch cultivations using shake flasks. In Figure 3, growth (A) and viability (B) curves of cell pools
expressing G6pd-1 and G6pd-2, three controls including the parental cell line expressing mCherry, a
nonproducing cell pool with the 3xstop+SpA DNA sequence recombined in the landing pad, and CHO-
S wild type cells are presented. At the end of the exponential phase, the highest maximum VCD (12.8 x
106 cells/ml) was achieved by the parental cell line that expresses mCherry, followed by G6pd-2. The
highest IVCD on day 7 (1358.42 x 106 Cells/h/ml) was also reached by the parental cell line (Table 3).
As for the growth performance in the parallel batch cultivation, the two cell pools expressing G6pd
differ in growth profile: G6pd-2 expressing cells reach higher cell densities at a faster rate than G6pd-1
expressing cells.
93
Figure 3 – Growth profiles obtained from batch cultivations of cell pools. (A) Viable cell density and (B) viability
measured throughout a batch cultivation of nonproducing cells (3xstop+SpA), G6pd expressing cells (G6pd-1 and
G6pd-2), parental cell line (mCherry) and wild type CHO-S cells (WT). Data points represent single
measurements (n=1). The batch cultivations were terminated on day 7.
3.5 Metabolite profile and specific consumption and production rates
We further characterized the metabolite profile of the engineered cells in relation to their growth profile.
The consumption of glutamine and glucose are directly linked to the formation of ammonia and lactate,
respectively. These main by-products of the mammalian metabolism can affect growth and productivity
in recombinant protein producing mammalian cells (Lao and Toth 1997). For studying the metabolic
differences in cells overexpressing G6pd versus and control cells, the concentrations of glucose,
glutamine, lactate and ammonia were determined throughout the batch cultivation (Figure 4).
Figure 4 – Time course measurements for metabolite profiling. Extracellular concentrations of (A) glucose, (B)
lactate, (C) glutamate, (D) glutamine, and (E) ammonium, measured throughout a batch cultivation of
nonproducing cells (3xstop+SpA), G6pd expressing cells (G6pd-1 and G6pd-2), parental cell line (mCherry) and
wild type CHO-S cells (WT). Data points represent single measurements (n=1).
94
In comparison to CHO-S wild type (WT) cells, all the remaining cell pools derived from the parental
cell line (mCherry), including the parental cell line, (herein referred as RMCE cells) present a similar
metabolite profile. WT cells (untransfected) present an overall higher consumption of glucose,
translated by the lower glucose concentrations measured in the spent media compared to the RMCE
cells (Figure 4 - A). Higher lactate concentrations on day 5 are observed in WT cells, while the RMCE
cells reach higher lactate secretion on day 6. WT (Figure 4 - B). The profile of glutamate concentration
differs for WT and RMCE cells. For WT cells, glutamate concentrations have low variability between
days 1 and 6 and increases on day 7 for WT cells, while for RMCE cells glutamate concentration has a
declining trend during that time period, reaching the minimum value on day 7 (Figure 4 - C). Complete
glutamine depletion is observed on day 5 for WT cells and on day 6 for the RMCE cell pools (Figure 4 -
D). Finally, ammonia concentrations are higher in WT cells until day 5, where both groups (WT and
RMCE cells) reach the same concentration from day 5 to day 7 (Figure 4 - E). The metabolite
consumption and production rates are presented in Table 4. It shows that overexpression of G6pd leads
to increased glutamine (qGln) and glucose (qGluc) consumption rates and specific production of lactate
(qLac) compared to parental cell line expressing mCherry.
3.6 Cell cultivation in the presence of H2O2 and sodium butyrate
To test whether the overexpression of G6pd increases the resistance to induced cell stress, we exposed
them to the known stress inducing compounds sodium butyrate (NaBu) and hydrogen peroxide (H2O2).
We added 5 mM NaBu, using 5 mM NaCl as control, and 100 µM H2O2, using H2O as control, to cell
culture media on day 3 of batch cultivations. Furthermore, we assessed the resulting changes in viability
(Figure 5) and cell growth (Suppl. Material Fig. S3). All cells show a drastic decrease in viability when
cultivated in the presence of NaBu and, unexpectedly, a decrease in viability was observed in the
presence of 5 mM NaCl for G6pd-2 cells. The results show that cell viability of G6pd-2 cells and the
parental cell line (mCherry) remains higher in most cultivation conditions than G6pd-1 and WT cells,
and are therefore more robust, as they are less affected by the induced stress. No reproducible effect is
observed between G6pd-1 and G6pd-2 cells.
95
Figure 5 – Changes in viability of cell pools expressing G6pd (G6pd-1 and G6pd-2), parental cell line (mCherry)
and CHO-S wild type (WT) cells transferred to 6-well plates on day 3. The cells were cultivated in cell culture
media supplemented with hydrogen peroxide (H2O2), a reactive oxygen species inducer of oxidative stress, H2O
used for volume control, the cytotoxic agent sodium butyrate (NaBu), and NaCl used as osmolarity control, added
on day 3. Viability was measured from day 4 to day 7. Five media formulations were included: Control – Basal
media made of CD-CHO+8 mM L-glutamine+0,2% anti-clumping agent; H2O2 – Basal media supplemented with
100 µM H2O2; H2O – Basal media with addition of same volume of H2O as in H2O2; NaBu – Basal media
supplemented with 5 mM NaBu; NaCl – Basal media supplemented with 5 mM NaCl.
3.7 Evaluation of effect of induced cellular stress in ER stress and apoptosis
In order to examine in more detail, whether G6pd overexpression plays a protective role against cellular
stress that leads to apoptotic cell death, we examined specific markers related to ER stress and apoptosis.
We used samples obtained on day 5 of batch cultivation to determine the effect the induced cellular
stress has on the ER stress markers C/EBP homologous protein (CHOP) and Binding of
immunoglobulin Protein (BiP) ER chaperone, as well as in apoptosis effectors, Caspase 3 (and Caspase
7). The addition of NaBu to the cell culture media results in a statistically significant (p < 0.05) increased
expression of CHOP, BiP and caspase 3 in G6pd-2 cells. However, the expression of caspase 7 decreases
significantly for WT and G6pd-1 cells cultivated in the presence of this cytotoxic chemical (Figure 6 -
A and B). When considering apoptosis, all cells show a similar profile, suggesting that overexpression
of G6pd does not influence the expression of caspase 3 or 7 in stress conditions in this experiment
(Figure 6 - C and D).
96
Figure 6 – Response of ER stress markers (A) BiP and (B) CHOP , and apoptosis effectors (C) Caspase 3 and (D)
Caspase 7 to induced cellular stress. Gene expression of ER stress markers CHOP and BiP, and apoptosis effectors
Caspase 3 (and Caspase 7) were analysed using RT-qPCR on samples collected on day 5 of batch cultivation. Cells
overexpressing G6pd (G6pd-1 and G6pd-2, respectively), the parental cell line (mCherry) and CHO-S wild type
(WT) cells were cultivated in the presence of CD-CHO (control), media supplemented with 100 µM H2O2, 5 mM
NaBu and NaCl (control). The fold changes were determined using ddCT method. The expression levels are all
relative to samples grown in the control medium (CD-CHO). Error bars represent standard deviation in technical
replicates (n=3). One-way ANOVA was used to determine whether the differences between the means for each
analyzed cell pool were statistically significant, α = 0.05. **** P=0.0001, *** P=0.0004, * P<0.05.
4. Discussion
During cell line development it is desirable to achieve high viable cell densities, as it leads to increased
titers of recombinant therapeutic glycoproteins. The cells are subject to increased cellular stress from
producing recombinant proteins normally linked to high production rates, complex protein structures
and their post translational modifications. Simultaneously, high viable cell densities leads to cellular
stresses imposed by changes in the culture milieu. Therefore, generating cells with increased resistance
to these types of stresses is a desirable feature in cell culture and engineering field. With this knowledge
in mind, we aimed at generating a CHO cell line that is overexpressing G6pd. G6pd plays an important
97
role in cell growth, via formation of nucleotides in PPP and biosynthesis of lipids needed for membrane
formation linked to cofactor requirements, as well as supplying required NADPH for biosynthesis of
amino acids usable for protein production, and resistance to cellular stress via reduction of oxidized
glutathione also requiring NADPH. We hypothesized that by overexpressing G6pd, we would be able
to create a positive effect on cell proliferation and increase the resistance to cell stress. In this work, we
inserted the G6pd gene into a parental cell line harboring a landing pad in a well-defined locus (Petersen
et al. 2018), using Cre/lox-based RMCE. Transfected cells were maintained for 7 days and bulk sorted
for enrichment of cells where the recombination event had successfully occurred (mCherry negative
cells). Using genomic DNA extracted from the cell pools, we amplified the inserts using PCR, using
primers specific to the regions surrounding the lox sites and using an oligo complementary to G6pd,
followed by Sanger sequencing. The PCR results shows inserts with the expected length (Suppl.
Materials Figure S1) and Sanger sequencing allowed for confirmation of the identity of the amplified
fragments, as these were correctly aligned to the reference sequence. We proceeded with further
verification at the transcript level, to understand whether there was an increase in G6pd mRNA. The
gene expression analysis of G6pd shows an increase of 2.5 - fold for G6pd-1 cells and approximately 5-
fold for G6pd-2 cells compared to the expression levels in the parental cell line (Figure 1 - A). G6pd-1
cells were generated from CHO-S wild type cDNA for USER cloning, while synthesized DNA based on
the G6pd coding sequence was used to generate G6pd-2 cells (Table 1), but as the sequences are
identical, the changes are likely to be due to variations in RMCE efficiency. Residual expression of
mCherry is present in all RMCE cells (Figure 1 - B). Next, proteomics analysis using LC-MS proteomics
revealed that there is an increase in G6pd protein levels only in the G6pd-2 cells, corresponding to a 4-
fold increase in G6pd protein expression compared to the parental cell line and even G6pd-1 cells
(Figure 2). This is well aligned with the RT-qPCR showing a higher expression of G6pd in the G6pd-2
cells.
98
G6pd is upregulated during exponential cell growth (Meleady et al. 2011; Orellana et al. 2015) and
highly translated (Courtes et al. 2013) in CHO cells, and is also upregulated in proliferative (tumor) cells
(Kuo, Lin, and Tang 2000). Based on this, we studied the effects of the transformation in cell growth
and physiology of the generated cells. We cultivated the two cell pools overexpressing G6pd, the parental
cell line expressing mCherry, an “empty-sequence” control (3xstop+SpA), and CHO-S WT, in batch
mode using shake flasks (Figure 3). Overall, the parental cell retains high viability for longer time
compared to the WT and the cells that exchanged donor DNA sequences. Surprisingly, the two cell
pools expressing G6pd have a dissimilar growth profile. This can possibly be due to variation at the
seeding density or at sampling VCD determination, or from the biological point of view, it could be an
effect of varying levels of G6pd mRNA expression in each cell pool. We conclude that overexpression
of G6pd did not improve growth significantly in the tested cells.
We continued studying cell physiology with the analysis of the exo-metabolite profile of the cells (Figure
4) and determined specific consumption and production rates of glucose, glutamine, lactate and
ammonium (Table 4). Differences between wild type cells and cells derived from the parental cell line
are observed. A reproducible metabolite profile is observed in cells derived from and including the
parental cell line (mCherry) that differs from the WT cells. WT cells display higher lactate and ammonia
concentrations a cultivation day earlier than RMCE cells (day 4 vs day 5). Additionally, the
overexpression of G6pd leads to higher qGln, qGluc, and qLac as seen in both G6pd-1 and G6pd-2.
Upregulation of G6pd is related to cellular stress responses as reduced glutathione (GSH) is the main
cellular scavenger of reactive oxygen species (Lu 2009). When GSH reacts with ROS, it becomes oxidized
to GSSG and in order for the cell to restore its reduced form, an enzymatic reaction is required, catalyzed
by Gsr or in a reaction requiring NADPH as a cofactor, which is mainly produced by the reaction
catalyzed by G6pd. Having this as premise, we tested the hypothesis that overexpression of G6pd in
CHO cells would confer higher resistance to induced cellular stress. We induced a cellular stress
response in two ways; supplementation of basal media with 100 µM H2O2, as reactive oxygen species
99
that induces oxidative stress, or 5 mM NaBu, that is a highly cytotoxic agent, mainly used in bioprocess
as cell cycle arrest and subsequent increase in recombinant protein expression in mammalian cell hosts
(Field and Brown 1990; Seong Lee, Lee, and Lee 2012; Kim and Lee 2000; Ganne et al. 1991; Kantardjieff
et al. 2010; Sunley and Butler 2010; Sung et al. 2004). We performed batch cultivation in shake flasks
and, on day 3, the cells were transferred to a 6-well plate format and cultivated in fresh cell culture media
supplemented with 100 µM H2O2, 5 mM NaBu and 5 mM NaCl as osmolality control. Figure 5 shows
that G6pd- and mCherry-expressing cells perform in a similar way in the majority of conditions,
regarding viability in response to induced cellular stress. Furthermore, we wanted to understand how
these forms of stress are related to ER stress and apoptosis and if there is a differential response in cells
expressing G6pd compared to control cells. We used samples from day 5 after addition of stress inducers
for conducting a gene expression analysis of CHOP, Bip, Caspase 3 and Caspase 7 using RT-qPCR
(Figure 6). The results show a similar response in all conditions in ER stress marker CHOP in G6pd-2
(with the higher G6pd expression) and mCherry expressing cells. In the presence of NaBu, these cells
show a higher increase in CHOP expression levels. Our results seem to contradict a number of reports
in the literature where the benefits of overexpression and upregulation of G6pd (Lee et al. 2012; Ghosh,
Zhao, and Price 2011) link to cell growth and protection to ROS via regeneration of NADPH required
for reduction of GSSG to GSH (Leopold et al. 2003; Tian et al. 1998, 1999; Kuo, Lin, and Tang 2000;
Courtes et al. 2013). One possibility could be that the observed phenotypes in other studies are
expression level-dependent. This explains some of the variations between the replicates we have in our
study, but does not fit the observation that we don’t see the expected phenotype in any of the two pools.
Another possibility is that the positive effect reported in other studies, is due to improved biosynthesis
of amino acids and/or nucleotides, and this is not an advantage in our experiments, as we use a rich
defined medium, as is common in CHO cell culture.
100
5. Conclusion and future experiments
This study aimed to generate cells with improved cell growth and increased resistance to induced
cellular stress. We overexpressed G6pd in a parental cell line with a defined RMCE integration site and
obtained two cell pools, where one of them with the highest G6pd expression levels was validated at
three levels (genomic, transcriptomic and proteomic), while the other was validated at the genomic and
transcriptomic levels. When studying the physiology of the re-engineered cells in batch cultivation
experiments, the results described subtle changes that might or might not be a characteristic phenotype
of CHO cells overexpressing G6pd. When cellular stress was induced in the cells by addition of 100 µM
H2O2 and 5 mM NaBu, a similar response was shown across cells when considering the ER stress
markers CHOP and Bip and apoptosis effector caspase 3.
The cells generated in this study can be single cell sorted to form clonal cell lines. After due validation
experiments, one may then test the hypothesis studied in this study and help determining the phenotype
of CHO cells overexpressing G6pd. One could additionally use functional studies on the activity of
G6pd, and determine NADPH/NADP+ and GSH/GSSG ratios in order to obtain a better understanding
of redox status of the cell and its organelles.
Acknowledgments
The authors thank Nachon Charanyanonda Petersen for the assistance in FACS analysis and cell sorting
and Saranya Nallapareddy the assistance with DNA cloning. The authors S.P., H.F.K. and M.R.A. thank
the Marie Skłodowska-Curie Actions under the EU Framework Programme for Research and
Innovation for eCHO systems ITN (Grant no. 642663) for funding this work. S.P., L.M.G. and H.F.K.
additionally thank the Novo Nordisk Foundation (Grant no. NNF10CC1016517) for the support.
References
Bonde, Mads T., Margit Pedersen, Michael S. Klausen, Sheila I. Jensen, Tune Wulff, Scott Harrison, Alex T. Nielsen, Markus J. Herrgård, and Morten O. A. Sommer. 2016. “Predictable Tuning of Protein Expression in Bacteria.” Nature Methods 13 (3): 233–36.
101
Brown, Adam J., Suzanne Gibson, Diane Hatton, and David C. James. 2018. “Transcriptome-Based Identification of the Optimal Reference CHO Genes for Normalisation of qPCR Data.” Biotechnology Journal 13 (1). https://doi.org/10.1002/biot.201700259.
Cao, Stewart Siyan, and Randal J. Kaufman. 2014. “Endoplasmic Reticulum Stress and Oxidative Stress in Cell Fate Decision and Human Disease.” Antioxidants & Redox Signaling 21 (3): 396–413.
Courtes, Franck C., Joyce Lin, Hsueh Lee Lim, Sze Wai Ng, Niki S. C. Wong, Geoffrey Koh, Leah Vardy, Miranda G. S. Yap, Bernard Loo, and Dong-Yup Lee. 2013. “Translatome Analysis of CHO Cells to Identify Key Growth Genes.” Journal of Biotechnology 167 (3): 215–24.
Davy, Anne Mathilde, Helene Faustrup Kildegaard, and Mikael Rørdam Andersen. 2017. “Cell Factory Engineering.” Cell Systems 4 (3): 262–75.
Ellgaard, Lars, Carolyn S. Sevier, and Neil J. Bulleid. 2018. “How Are Proteins Reduced in the Endoplasmic Reticulum?” Trends in Biochemical Sciences 43 (1): 32–43.
Ercal, Nuran, Piyanee Treeratphan, Paula Lutz, Terese C. Hammond, and Richard H. Matthews. 1996. “N-Acetylcysteine Protects Chinese Hamster Ovary (CHO) Cells from Lead-Induced Oxidative Stress.” Toxicology 108 (1-2): 57–64.
Field, R., and M. Brown. 1990. “Sodium Butyrate Increases Productivity of Hybridoma Cultures.” Cell Biology International Reports 14: 219.
Fischer, Simon, René Handrick, and Kerstin Otte. 2015. “The Art of CHO Cell Engineering: A Comprehensive Retrospect and Future Perspectives.” Biotechnology Advances 33 (8): 1878–96.
Ganne, V., P. Guerin, T. Faure, and G. Mignot. 1991. “INCREASED EXPRESSION OF FACTOR VIII BY BUTYRATE IN CHINESE HAMSTER OVARY CELLS.” In Production of Biologicals from Animal Cells in Culture, 104–6.
Ghosh, Amit, Huimin Zhao, and Nathan D. Price. 2011. “Genome-Scale Consequences of Cofactor Balancing in Engineered Pentose Utilization Pathways in Saccharomyces Cerevisiae.” PloS One 6 (11): e27316.
Grav, Lise Marie, Daria Sergeeva, Jae Seong Lee, Igor Marin de Mas, Nathan E. Lewis, Mikael Rørdam Andersen, Lars Keld Nielsen, Gyun Min Lee, and Helene Faustrup Kildegaard. 2018. “Minimizing Clonal Variation during Mammalian Cell Line Engineering for Improved Systems Biology Data Generation.” ACS Synthetic Biology, August. https://doi.org/10.1021/acssynbio.8b00140.
Haredy, Ahmad M., Akitoshi Nishizawa, Kohsuke Honda, Tomoshi Ohya, Hisao Ohtake, and Takeshi Omasa. 2013. “Improved Antibody Production in Chinese Hamster Ovary Cells by ATF4 Overexpression.” Cytotechnology 65 (6): 993–1002.
Ha, Tae Kwang, Anders Holmgaard Hansen, Stefan Kol, Helene Faustrup Kildegaard, and Gyun Min Lee. 2017. “Baicalein Reduces Oxidative Stress in CHO Cell Cultures and Improves Recombinant Antibody Productivity.” Biotechnology Journal 13 (3): 1700425.
Ilnicka, Anna, Katarzyna Roszek, Andrzej Olejniczak, Michal Komoszynski, and Jerzy P. Lukaszewicz. 2014. “Biologically Active Constituents from Salix Viminalis Bio-Oil and Their Protective Activity against Hydrogen Peroxide-Induced Oxidative Stress in Chinese Hamster Ovary Cells.” Applied Biochemistry and Biotechnology 174 (6): 2153–61.
Kantardjieff, Anne, Nitya M. Jacob, Joon Chong Yee, Eyal Epstein, Yee-Jiun Kok, Robin Philp, Michael Betenbaugh, and Wei-Shou Hu. 2010. “Transcriptome and Proteome Analysis of Chinese Hamster Ovary Cells under Low Temperature and Butyrate Treatment.” Journal of Biotechnology 145 (2): 143–59.
Kim, No Soo, and Gyun Min Lee. 2000. “Overexpression of Bcl-2 Inhibits Sodium Butyrate-Induced
102
Apoptosis in Chinese Hamster Ovary Cells Resulting in Enhanced Humanized Antibody Production.” Biotechnology and Bioengineering 71 (3): 184–93.
Kuo, W., J. Lin, and T. K. Tang. 2000. “Human Glucose-6-Phosphate Dehydrogenase (G6PD) Gene Transforms NIH 3T3 Cells and Induces Tumors in Nude Mice.” International Journal of Cancer. Journal International Du Cancer 85 (6): 857–64.
Ku, Sebastian C. Y., Daphne T. W. Ng, Miranda G. S. Yap, and Sheng-Hao Chao. 2008. “Effects of Overexpression of X-Box Binding Protein 1 on Recombinant Protein Production in Chinese Hamster Ovary and NS0 Myeloma Cells.” Biotechnology and Bioengineering 99 (1): 155–64.
Lao, M-S, and D. Toth. 1997. “Effects of Ammonium and Lactate on Growth and Metabolism of a Recombinant Chinese Hamster Ovary Cell Culture.” Biotechnology Progress 13 (5): 688–91.
Lee, Jeong Wook, Dokyun Na, Jong Myoung Park, Joungmin Lee, Sol Choi, and Sang Yup Lee. 2012. “Systems Metabolic Engineering of Microorganisms for Natural and Non-Natural Chemicals.” Nature Chemical Biology 8 (6): 536–46.
Leopold, Jane A., Ying-Yi Zhang, Anne W. Scribner, Robert C. Stanton, and Joseph Loscalzo. 2003. “Glucose-6-Phosphate Dehydrogenase Overexpression Decreases Endothelial Cell Oxidant Stress and Increases Bioavailable Nitric Oxide.” Arteriosclerosis, Thrombosis, and Vascular Biology 23 (3): 411–17.
Levitt, N., D. Briggs, A. Gil, and N. J. Proudfoot. 1989. “Definition of an Efficient Synthetic poly(A) Site.” Genes & Development 3 (7): 1019–25.
Lewis, Nathan E., Xin Liu, Yuxiang Li, Harish Nagarajan, George Yerganian, Edward O’Brien, Aarash Bordbar, et al. 2013. “Genomic Landscapes of Chinese Hamster Ovary Cell Lines as Revealed by the Cricetulus Griseus Draft Genome.” Nature Biotechnology 31 (8): 759–65.
Lund, Anne Mathilde, Helene Faustrup Kildegaard, Maja Borup Kjær Petersen, Julie Rank, Bjarne Gram Hansen, Mikael Rørdam Andersen, and Uffe Hasbro Mortensen. 2014. “A Versatile System for USER Cloning-Based Assembly of Expression Vectors for Mammalian Cell Engineering.” PloS One 9 (5): e96693.
Lu, Shelly C. 2009. “Regulation of Glutathione Synthesis.” Molecular Aspects of Medicine 30 (1-2): 42–59.
Meleady, Paula, Padraig Doolan, Michael Henry, Niall Barron, Joanne Keenan, Finbar O’Sullivan, Colin Clarke, et al. 2011. “Sustained Productivity in Recombinant Chinese Hamster Ovary (CHO) Cell Lines: Proteome Analysis of the Molecular Basis for a Process-Related Phenotype.” BMC Biotechnology 11 (July): 78.
Orellana, Camila A., Esteban Marcellin, Benjamin L. Schulz, Amanda S. Nouwens, Peter P. Gray, and Lars K. Nielsen. 2015. “High-Antibody-Producing Chinese Hamster Ovary Cells up-Regulate Intracellular Protein Transport and Glutathione Synthesis.” Journal of Proteome Research 14 (2): 609–18.
Pereira, Sara, Helene Faustrup Kildegaard, and Mikael Rørdam Andersen. 2018. “Impact of CHO Metabolism on Cell Growth and Protein Production: An Overview of Toxic and Inhibiting Metabolites and Nutrients.” Biotechnology Journal 13 (3): e1700499.
Petersen, Søren D., Jie Zhang, Jae S. Lee, Tadas Jakociunas, Lise M. Grav, Helene F. Kildegaard, Jay D. Keasling, and Michael K. Jensen. 2018. “Modular 5’-UTR Hexamers for Context-Independent Tuning of Protein Expression in Eukaryotes.” Nucleic Acids Research, August. https://doi.org/10.1093/nar/gky734.
Rupp, Oliver, Madolyn L. MacDonald, Shangzhong Li, Heena Dhiman, Shawn Polson, Sven Griep, Kelley Heffner, et al. 2018. “A Reference Genome of the Chinese Hamster Based on a Hybrid Assembly Strategy.” Biotechnology and Bioengineering 115 (8): 2087–2100.
103
Seong Lee, Jae, Jae Seong Lee, and Gyun Min Lee. 2012. “Effect of Sodium Butyrate on Autophagy and Apoptosis in Chinese Hamster Ovary Cells.” Biotechnology Progress 28 (2): 349–57.
Skała, Ewa, Przemysław Sitarek, Marek Różalski, Urszula Krajewska, Janusz Szemraj, Halina Wysokińska, and Tomasz Śliwiński. 2016. “Antioxidant and DNA Repair Stimulating Effect of Extracts from Transformed and Normal Roots of Rhaponticum Carthamoides against Induced Oxidative Stress and DNA Damage in CHO Cells.” Oxidative Medicine and Cellular Longevity 2016 (February): 5753139.
Stolfa, Gino, Matthew T. Smonskey, Ryan Boniface, Anna-Barbara Hachmann, Paul Gulde, Atul D. Joshi, Anson P. Pierce, Scott J. Jacobia, and Andrew Campbell. 2018. “CHO-Omics Review: The Impact of Current and Emerging Technologies on Chinese Hamster Ovary Based Bioproduction.” Biotechnology Journal 13 (3): e1700227.
Sung, Yun Hee, Yeon Jung Song, Seung Wook Lim, Joo Young Chung, and Gyun Min Lee. 2004. “Effect of Sodium Butyrate on the Production, Heterogeneity and Biological Activity of Human Thrombopoietin by Recombinant Chinese Hamster Ovary Cells.” Journal of Biotechnology 112 (3): 323–35.
Sunley, Kevin, and Michael Butler. 2010. “Strategies for the Enhancement of Recombinant Protein Production from Mammalian Cells by Growth Arrest.” Biotechnology Advances 28 (3): 385–94.
Tian, W. N., L. D. Braunstein, K. Apse, J. Pang, M. Rose, X. Tian, and R. C. Stanton. 1999. “Importance of Glucose-6-Phosphate Dehydrogenase Activity in Cell Death.” The American Journal of Physiology 276 (5 Pt 1): C1121–31.
Tian, W. N., L. D. Braunstein, J. Pang, K. M. Stuhlmeier, Q. C. Xi, X. Tian, and R. C. Stanton. 1998. “Importance of Glucose-6-Phosphate Dehydrogenase Activity for Cell Growth.” The Journal of Biological Chemistry 273 (17): 10609–17.
Walsh, Gary. 2014. “Biopharmaceutical Benchmarks 2014.” Nature Biotechnology 32 (10): 992–1000. Walter, Peter, and David Ron. 2011. “The Unfolded Protein Response: From Stress Pathway to
Homeostatic Regulation.” Science 334 (6059): 1081–86. Wells, Evan, and Anne Skaja Robinson. 2017. “Cellular Engineering for Therapeutic Protein
Production: Product Quality, Host Modification, and Process Improvement.” Biotechnology Journal 12 (1). https://doi.org/10.1002/biot.201600105.
Xu, Xun, Harish Nagarajan, Nathan E. Lewis, Shengkai Pan, Zhiming Cai, Xin Liu, Wenbin Chen, et al. 2011. “The Genomic Sequence of the Chinese Hamster Ovary (CHO)-K1 Cell Line.” Nature Biotechnology 29 (8): 735–41.
104
Tables
Table 1- Sequences used for construction of promotorless plasmids used for Recombinase Mediated Cassette exchange.
Transcript name Sequence Accession number
G6pd ATGGCAGAGCAGGTGGCCCTGAGCCGGACCCAGGTGTGTGGCATCCTGAGGGAAGAGTTGTACCAGGGTGATGCCTTCCACCAAGCTGATACACATATATTTATCATCATGGGTGCATCGGGTGACCTGGCCAAGAAGAAGATCTATCCCACTATCTGGTGGCTGTTCCGGGATGGCCTTCTACCCGAAGACACCTTCATTGTGGGCTATGCCCGCTCCCGACTCACAGTGGATGACATCCGCAAGCAGAGTGAGCCCTTCTTTAAAGCCACCCCAGAGGAAAGACCCAAGCTGGAGGAGTTCTTTGCCCGTAACTCCTATGTGGCTGGCCAGTATGATGATCCAGCCTCCTACAAGCACCTCAACAGCCACATGAATGCCCTGCATCAGGGGATGCAGGCCAACCGCCTATTCTACCTGGCCTTGCCCCCCACAGTCTATGAAGCTGTCACCAAGAACATTCAAGAGACCTGCATGAGTCAGACAGGCTGGAACCGCATCATAGTGGAGAAGCCCTTCGGGAGAGACCTGCAGAGCTCCAACCAGCTGTCGAACCACATCTCCTCTCTGTTCCGTGAGGACCAGATCTACCGCATTGACCACTACCTGGGCAAGGAGATGGTCCAGAACCTCATGGTGCTGAGATTTGCCAACAGGATCTTTGGCCCCATCTGGAACCGAGACAACATTGCCTGTGTGATCCTCACATTTAAAGAGCCCTTTGGTACTGAGGGTCGTGGGGGCTACTTTGATGAATTTGGGATCATCAGGGATGTTATGCAGAACCACCTCCTGCAGATGTTGTGTCTGGTGGCCATGGAAAAACCTGCCTCCACAGATTCAGATGATGTCCGTGATGAGAAGGTCAAAGTGTTGAAATGTATCTCAGAGGTGGAAACCAGCAATGTGGTCCTTGGCCAGTATGTGGGGAACCCCAATGGAGAAGGAGAAGCTACCAATGGGTACTTGGATGACCCCACAGTGCCCCGTGGGTCCACCACTGCCACCTTTGCAGCAGCTGTCCTCTATGTGGAGAATGAGCGGTGGGATGGGGTACCCTTCATCCTGCGCTGTGGCAAAGCCCTGAATGAACGCAAGGCTGAGGTGAGACTACAGTTCCGAGATGTGGCAGGCGACATCTTCCACCAGCAGTGCAAGCGTAATGAGCTGGTGATTCGTGTGCAGCCCAATGAGGCTGTATACACCAAGATGATGACCAAGAAGCCTGGCATGTTCTTCAACCCTGAGGAGTCAGAGCTGGACTTGACCTATGGCAACAGATACAAGAATGTGAAGCTCCCTGATGCCTATGAACGCCTCATCCTGGATGTCTTCTGTGGGAGCCAGATGCACTTTGTCCGCAGTGATGAACTCAGGGAAGCCTGGCGTATCTTCACACCACTGCTGCACAAGATTGATCAAGAGAAGCCCCAGCCTATCCCCTATGTTTATGGCAGCCGCGGCCCCACAGAGGCAGATGAGCTGATGAAGAGAGTGGGCTTCCAGTATGAGGGCACCTACAAATGGGTGAACCCTCACAAGCTCTGA
NM_001246727
3x stop codon + SpA
TAATAGTGAATAAAATATCTTTATTTTCATTACATCTGTGTGTTGGTTTTTTGTGTG -
Table 2- Primer sequences used for the generation of promoterless plasmids via USER cloning.
# Primer name Sequence Template
1 1stbackboneFwd AATAACUTCGTATAGGATACTTTAT Plasmid backbone
2 1stbackboneRev ATAACTUCGTATAATGTATGCTATA Plasmid backbone
3 kzg6pd2 AAGTTAUCGCCACCATGGCAGAGCA G6pd (cDNA and gblock)
4 kzG6p2GOIRev AGTTATUCAGAGCTTGTGAGGGTTCACCC G6pd (cDNA and gblock)
5 3x STOP_SpA_Backbone_forward
ATTTTCATUACATCTGTGTGTTGGTTTTTTGTGTGATAACTTCGTATAGGATA
Plasmid backbone
6 3x STOP_SpA_Backbone_reverse AATGAAAAUAAAGATATTTTATTCACTATTAATAACTTCGTATAATGTA
Plasmid backbone
105
Table 3 - Integral viable cell density (IVCD) calculated on day 3 and on day 7.
Sample IVCD at Day 7
(106 Cells/h/ml)
G6pd-1 1106,13
G6pd-2 1239,54
3xstop+SpA 1244,42
mCherry 1358,42
WT 1156,09
Table 4 - Specific consumption and specific production rates determined from day 0 to day 3. Data represents single measurements. * indicates biological replicates.
Specific consumption rates
(pmol/cell/day)
Specific production rates
(pmol/cell/day)
Cell qGln qGlu qGluc qLac qNH4+
G6pd-1* -0,47 0,02 -2,05 3,65 0,71
G6pd-2* -0,43 0,03 -1,96 3,30 0,66
3xstop+SpA -0,30 -0,03 -1,74 3,01 0,66
mCherry -0,26 0,02 -1,57 3,00 0,66
WT -0,34 0,04 -1,97 2,82 0,67
106
Chapter 3 – Study of dose-dependent effects of metabolite additions on cell growth
Introduction
Ammonia is known to decrease cell growth and productivity in mammalian cells (Ozturk, Riley, and
Palsson 1992) and to alter glycosylation patterns of expressed therapeutic proteins (Zhou and Kantardjieff
2014; Schneider 1996). Even though ammonia has been well described to be cytotoxic to mammalian cells,
the exact mechanisms behind this toxicity remain poorly understood. Ammonia is mainly a by-product of
glutamine degradation. We therefore studied the effects of supplementing cell culture media with different
doses of ammonium chloride (NH4Cl) on the growth profile of CHO-S wild type host cells used in house.
Based on Lao and Toth’s study (Lao and Toth 1997), we employed an experimental set-up adapted to CHO-
S cells that are cultivated in suspension and maintained in serum-free chemically defined media
supplemented with glutamine.
Materials and methods
CHO-S cells growing exponentially in suspension were seeded at 1 x 106 cells/mL to 6-well plates (Corning)
in CD-CHO medium (Life Technologies) supplemented with 8 mM L-Glutamine (Thermo Fisher
Scientific), 0.2 % anti-clumping agent (Gibco) and with 0, 5, 10, 15, 20, 25, 30, 35 and 40 mM of NH4Cl
(Sigma-Aldrich) in test samples. The same concentrations of sodium chloride (NaCl) (Sigma-Aldrich) were
used in control samples. Cells were maintained in humidified incubators, at 37°C, 5% CO2, shaking at 120
rpm. Viable cell density (VCD) and viability were monitored every day using NucleoCounter NC-200 Cell
Counter (ChemoMetec) over the course of 6 days of cultivation. Each day, after cell counts, the cells were
spun down (200 g, 5 min) and the supernatant was discarded. The cells were resuspended in fresh
107
supplemented cell culture media, transferred again to 6-well plates and maintained under the conditions
described above.
Results and discussion
We studied the effects of NH4Cl on cell growth and viability of CHO-S host cells. For that, CD-CHO
medium supplemented with 8 mM L-Glutamine, 0.2 % anti-clumping agent, from here on referred to as
cell culture media, and with 0, 5, 10, 15, 20, 25, 30, 35 and 40 mM of NH4Cl, in test samples and NaCl at the
same concentrations in control samples. Over the course of 6 days, the cells were cultivated in cell culture
media supplemented with the abovementioned concentrations of NH4Cl and NaCl. Fresh cell culture media
were supplied every day to prevent the accumulation lactate and ammonia produced by the cells.
Cell culture media supplemented with 0 and 5 mM NH4Cl have an identical growth profile while media
supplemented with 5 mM NaCl seem to have an improved cell growth (Figure 1, Top Right). The next level
of growth inhibition by NH4Cl is observed at 20 and 25 mM. Finally, 30 and 40 mM NH4Cl has a very large
effect on cell growth. Furthermore, supplementation of media with 20 and 25 mM NH4Cl decreases cell
counts more than 50%. This concentration is about 2-fold higher than what we normally observe in batch
cultivations of CHO-S cells (data not shown). We observed comparable terminal VCD values across control
conditions of 0 mM and 10-40 mM NaCl (Figure 1).
Based on these results, there is a dose-dependent effect of NH4Cl on cell growth that is not observed when
the cells are cultivated in media supplemented with identical doses of NaCl. We have identified the
conditions that lead to 50% reduction of cell growth (20-25 mM NH4Cl) and to cell growth arrest (40 mM
NH4Cl). A follow up experiment can be performed in combination with screening studies in long term
cultivations, to e.g. determine changes in gene expression or non-coding regulatory sequences that become
up- or down-regulated in the presence of different concentrations of NH4Cl, used as selection pressure.
108
In Paper I, a number of metabolites were reported to affect the performance of cells in culture. These can
be studied in a similar way to the work presented in this chapter to improve understanding of the effects of
their accumulation on cell metabolism.
Figure 1 – Viable cell density of CHO-S host cells cultivated in CD-CHO medium supplemented with 8 mM L-Glutamine, 0,2 % anti-clumping agent, and 5, 10, 15, 20, 25, 30, 35 and 40 mM NH4Cl (Top left) and NaCl (Top right) and terminal VCD at day 6 of cell cultivation in media supplemented with various concentrations of NH4Cl (Bottom left) and NaCl (Bottom right). The cells were cultivated in duplicates (n=2) and error bars represent standard deviation.
109
References
Lao, M. S., and D. Toth. 1997. “Effects of Ammonium and Lactate on Growth and Metabolism of a Recombinant Chinese Hamster Ovary Cell Culture.” Biotechnology Progress 13 (5): 688–91.
Ozturk, Sadettin S., Mark R. Riley, and Bernhard O. Palsson. 1992. “Effects of Ammonia and Lactate on Hybridoma Growth, Metabolism, and Antibody Production.” Biotechnology and Bioengineering 39 (4): 418–31.
Schneider, M. 1996. “The Importance of Ammonia in Mammalian Cell Culture.” Journal of Biotechnology 46 (3): 161–85.
Zhou, Weichang, and Anne Kantardjieff. 2014. Mammalian Cell Cultures for Biologics Manufacturing. Springer.
110
Conclusion and future perspectives
The aim of this project was to study CHO cell metabolism and generate cells with improved metabolism.
In Chapter 1, we identified metabolites that affect cell performance by surveying literature. Many are
related to nutrient oversupply in the media, while others represent limitations inherent to the cell’s
metabolism. Paper I is a useful resource that this project contributed to the CHO community, in which we
provided leads for media and feed optimization, as well as for identifying gene engineering targets. In
Chapter 3, the effect of ammonium on cell growth was tested in CHO-S cells, since often these effects are
cell-line specific. This small-scale study revealed that there is a dose-dependent effect of NH4Cl on cell
growth. Similar studies could be performed with selected metabolites mentioned in Paper I and a step
further could be taken into identifying novel targets for engineering based on the analysis of the
transcriptomic or proteomic signatures in cells cultivated under different selection pressures or through
screening experiments.
We presented cell line engineering approaches used to obtain CHO cells with optimal nutrient and by-
product metabolism. We studied CHO cell physiology and were able to generate cells with improved
metabolism in some of our experiments presented in Chapter 2. We showed that cell growth increased, as
well as lactate and ammonia decreased when genes part of the amino acid catabolism are disrupted in
Papers II and III. We chose to target genes encoding dehydrogenases that indirectly affect lactate secretion
and prevent amino acids conversion in catabolic reactions, which, themselves, lead to the buildup of
intermediates reported to be growth inhibitory [56]. Moreover, we aimed to disrupt genes leading to
ammonia formation. According to our hypothesis, after disrupting catabolic pathways, the amino acids
would become available for cellular processes, such as biomass formation and, in the case of producer cells,
recombinant protein production. Our studies showed that amongst the 9 targeted genes (Paper II), the
individual disruption of Hpd and Gad2 lead to improvements in cell growth, while combinatorial
disruption of genes leads to a reduction of ammonium and lactate secretion without changes in cell growth.
111
Even so, a number of challenges were encountered during this project. Clonal variation, a common topic of
discussion in the CHO field [38,86–88], may be responsible for masking the real effects of the gene targeting
which complicates the phenotype interpretation. Clonal variation can be an inherent property of the cell
pool from where the clones originate and become evident only after expanding engineered clones derived
from a pool of cells. Ideally, parental cells used for metabolic studies should display unvarying cell growth.
Thus, we used the learnings from the experimental work that led to Paper II when preparing the
experiments related to Paper III, where we disrupted genes participating in BCAA catabolism in two cell
lines, a CHO host cell and a producer cell line with reduced growth variation. We were able to reprogram
the metabolism in clones displaying increased cell growth and improved nutrient consumption and reduced
by-product production rates. Nonetheless, these results appear to be cell-line- and clone-specific. After the
completion of the experimental work leading to this thesis, a study employing a strategy contrary to ours in
Paper II and targeting BCAA metabolism was published using a different producer CHO cell line [81]. We
also understand that given the low expression of some genes across cell lines [7], the effects of gene
disruption may be subtle. Furthermore, our approach differs from other attempts to reduce by-product
formation that used RNAi technology to engineer genes related to lactate secretion in transient manner [73,
74].
During industrial cultivations, the cells are constantly monitored by following several parameters, such as
metabolite concentration, temperature, and pH amongst others, along with pre-planned feed and base
addition[89]. However, these can also generate new problems. For instance, in a controlled bioprocess, as
lactate builds up and pH decreases, base is added to maintain the pH at physiological levels. This has
detrimental effects on the cell due to the resulting increase in osmolarity [90]. Using cells engineered with
reduced lactate formation phenotype has the potential to avoid such a situation. Despite providing a
straightforward solution to adapt to the cell’s metabolic inefficiency, this example shows the limitations of
112
bioprocess optimization. Different approaches to engineer CHO cells with reduced secretion of lactate
[73,76,80,91,92] and ammonia [73,76,80,91,92] have been reported.
Media and feeds are routinely optimized in order to achieve higher productivity [39,93]. However, most of
these approaches fail to take into consideration the metabolic characteristics of the cell as several nutrients
are supplied at too high concentrations and lead to increased consumption rates and production of toxic
catabolic intermediates [51]. Most optimal media recipes and industrial cell lines alike are part of the trade
secrets of biopharmaceutical companies and are not publicly available, complicating efforts to assess the
industrial applicability of novel approaches using cell engineering. We also see that engineered cells with
reduced lactate secretion could compete with existing bioprocessing-based approaches.
In Paper IV, we overexpressed G6pd since the PPP has been reported to be upregulated during the
proliferative growth phase of several cell types [94–96]. However, our study showed unexpected opposite
results. These pathways are highly regulated and the overexpression of a combination of genes, rather than
a single gene, might be required to attain the predicted result. Established strategies used to improve cell
growth and viability during cultivations included engineering of apoptosis by overexpression of anti-
apoptotic gene Bcl-2, Bcl-XL or disruption of Bax and Bak, granting longer cultivation and, in some cases,
productivity was also increased [84,97–100].
Overall, based on the results obtained in Chapter 2, we conclude that detailed knowledge of the cell’s
genome, gene expression, and metabolic networks is essential to proceed with rational engineering of CHO
cells. But since sequencing is becoming affordable, future efforts to genetic engineer metabolic genes will
be made easier. Interesting targets for engineering would be related to lipid metabolism since these
accumulate inside the cell beyond the cell’s growth requirements [101]. Moreover, with the tendency to
move from fed-batch cultivation mode into intensified processes such as perfusion, high producer cells able
113
to withstand shear pressures and high cell densities may be required. One could attempt to generate such
cells using cell engineering.
The work of this thesis provides metabolic efficient cell lines and insights into the improvement of
bioprocesses. Further enhancement of the CHO cell metabolism through metabolic engineering may lead
to an improved producer of recombinant therapeutic proteins and efficient industrial cultivation processes.
These efforts cannot rely on one discipline alone. Success at generating CHO host and producer cell lines
requires a combined effort of cell culture specialists and with knowledge of cell line engineering and
bioprocessing, and bioinformaticians.
114
References
1. Walsh G. Biopharmaceutical benchmarks 2014. Nature Biotechnology. 2014;32: 992–1000.
2. Walsh G. Biopharmaceutical benchmarks 2018. Nat Biotechnol. 2018;36: 1136–1145.
3. Wurm FM. Production of recombinant protein therapeutics in cultivated mammalian cells. Nat Biotechnol. 2004;22: 1393–1398.
4. Bandaranayake AD, Almo SC. Recent advances in mammalian protein production. FEBS Letters. 2014;588: 253–260.
5. Puck TT, Cieciura SJ, Robinson A. Genetics of somatic mammalian cells. III. Long-term cultivation of euploid cells from human and animal subjects. J Exp Med. 1958;108: 945–956.
6. Wurm F. CHO Quasispecies—Implications for Manufacturing Processes. Processes. 2013;1: 296–311.
7. Singh A, Kildegaard HF, Andersen MR. An Online Compendium of CHO RNA-Seq Data Allows Identification of CHO Cell Line-Specific Transcriptomic Signatures. Biotechnol J. 2018;13: e1800070.
8. Davy AM, Kildegaard HF, Andersen MR. Cell Factory Engineering. Cell Systems. 2017;4: 262–275.
9. Marisch K, Bayer K, Cserjan-Puschmann M, Luchner M, Striedner G. Evaluation of three industrial Escherichia coli strains in fed-batch cultivations during high-level SOD protein production. Microb Cell Fact. 2013;12: 58.
10. Kleiner-Grote GRM, Risse JM, Friehs K. Secretion of recombinant proteins from E. coli. Engineering in Life Sciences. 2018;18: 532–550.
11. Nielsen J. Yeast Systems Biology: Model Organism and Cell Factory. Biotechnol J. 2019; e1800421.
12. Wildt S, Gerngross TU. The humanization of N-glycosylation pathways in yeast. Nat Rev Microbiol. 2005;3: 119–128.
13. Liu L. Antibody glycosylation and its impact on the pharmacokinetics and pharmacodynamics of monoclonal antibodies and Fc-fusion proteins. J Pharm Sci. 2015;104: 1866–1884.
14. Dumont J, Euwart D, Mei B, Estes S, Kshirsagar R. Human cell lines for biopharmaceutical manufacturing: history, status, and future perspectives. Crit Rev Biotechnol. 2016;36: 1110–1122.
15. Berting A, Farcet MR, Kreil TR. Virus susceptibility of Chinese hamster ovary (CHO) cells and detection of viral contaminations by adventitious agent testing. Biotechnology and Bioengineering. 2010;106: 598–607.
16. Xu X, Nagarajan H, Lewis NE, Pan S, Cai Z, Liu X, et al. The genomic sequence of the Chinese hamster ovary (CHO)-K1 cell line. Nature Biotechnology. 2011;29: 735–741.
17. Estes S, Melville M. Mammalian cell line developments in speed and efficiency. Adv Biochem Eng Biotechnol. 2014;139: 11–33.
115
18. Lalonde M-E, Durocher Y. Therapeutic glycoprotein production in mammalian cells. Journal of Biotechnology. 2017;251: 128–140.
19. Kumar SR. Industrial production of clotting factors: Challenges of expression, and choice of host cells. Biotechnol J. 2015;10: 995–1004.
20. Butler M, Spearman M. The choice of mammalian cell host and possibilities for glycosylation engineering. Curr Opin Biotechnol. 2014;30: 107–112.
21. Chu L, Robinson DK. Industrial choices for protein production by large-scale cell culture. Current Opinion in Biotechnology. 2001;12: 180–187.
22. Lewis NE, Liu X, Li Y, Nagarajan H, Yerganian G, O’Brien E, et al. Genomic landscapes of Chinese hamster ovary cell lines as revealed by the Cricetulus griseus draft genome. Nat Biotechnol. 2013;31: 759–765.
23. Rupp O, MacDonald ML, Li S, Dhiman H, Polson S, Griep S, et al. A reference genome of the Chinese hamster based on a hybrid assembly strategy. Biotechnol Bioeng. 2018;115: 2087–2100.
24. Brinkrolf K, Rupp O, Laux H, Kollin F, Ernst W, Linke B, et al. Chinese hamster genome sequenced from sorted chromosomes. Nat Biotechnol. 2013;31: 694–695.
25. Hammond S, Swanberg JC, Kaplarevic M, Lee KH. Genomic sequencing and analysis of a Chinese hamster ovary cell line using Illumina sequencing technology. BMC Genomics. 2011;12. doi:10.1186/1471-2164-12-67
26. Kildegaard HF, Baycin-Hizal D, Lewis NE, Betenbaugh MJ. The emerging CHO systems biology era: harnessing the ’omics revolution for biotechnology. Curr Opin Biotechnol. 2013;24: 1102–1107.
27. Yusufi FNK, Lakshmanan M, Ho YS, Loo BLW, Ariyaratne P, Yang Y, et al. Mammalian Systems Biotechnology Reveals Global Cellular Adaptations in a Recombinant CHO Cell Line. Cell Systems. 2017;4: 530–542.e6.
28. Stolfa G, Smonskey MT, Boniface R, Hachmann A-B, Gulde P, Joshi AD, et al. CHO-Omics Review: The Impact of Current and Emerging Technologies on Chinese Hamster Ovary Based Bioproduction. Biotechnol J. 2018;13: e1700227.
29. Ronda C, Pedersen LE, Hansen HG, Kallehauge TB, Betenbaugh MJ, Nielsen AT, et al. Accelerating genome editing in CHO cells using CRISPR Cas9 and CRISPy, a web-based target finding tool. Biotechnology and Bioengineering. 2014;111: 1604–1616.
30. Sergeeva D, Camacho-Zaragoza JM, Lee JS, Kildegaard HF. CRISPR/Cas9 as a Genome Editing Tool for Targeted Gene Integration in CHO Cells. Methods Mol Biol. 2019;1961: 213–232.
31. Xiong K, Marquart KF, la Cour Karottki KJ, Li S, Shamie I, Lee JS, et al. Reduced apoptosis in Chinese hamster ovary cells via optimized CRISPR interference. Biotechnol Bioeng. 2019;116: 1813–1819.
32. Lee JS, Kallehauge TB, Pedersen LE, Kildegaard HF. Site-specific integration in CHO cells mediated by CRISPR/Cas9 and homology-directed DNA repair pathway. Scientific Reports. 2015;5. doi:10.1038/srep08572
116
33. Shin J, Lee N, Song Y, Park J, Kang TJ, Kim SC, et al. Efficient CRISPR/Cas9-mediated multiplex genome editing in CHO cells via high-level sgRNA-Cas9 complex. Biotechnology and Bioprocess Engineering. 2015;20: 825–833.
34. Schmieder V, Bydlinski N, Strasser R, Baumann M, Kildegaard HF, Jadhav V, et al. Enhanced Genome Editing Tools For Multi-Gene Deletion Knock-Out Approaches Using Paired CRISPR sgRNAs in CHO Cells. Biotechnol J. 2018;13: e1700211.
35. Hefzi H, Ang KS, Hanscho M, Bordbar A, Ruckerbauer D, Lakshmanan M, et al. A Consensus Genome-scale Reconstruction of Chinese Hamster Ovary Cell Metabolism. Cell Syst. 2016;3: 434–443.e8.
36. Selvarasu S, Ho YS, Chong WPK, Wong NSC, Yusufi FNK, Lee YY, et al. Combined in silico modeling and metabolomics analysis to characterize fed-batch CHO cell culture. Biotechnol Bioeng. 2012;109: 1415–1429.
37. Martinet D, Derouazi M, Besuchet N, Wicht M, Beckmann J, Wurm FM. Karyotype of CHO DG44 cells. Cell Technology for Cell Products. : 363–366.
38. Ko P, Misaghi S, Hu Z, Zhan D, Tsukuda J, Yim M, et al. Probing the importance of clonality: Single cell subcloning of clonally derived CHO cell lines yields widely diverse clones differing in growth, productivity, and product quality. Biotechnol Prog. 2018;34: 624–634.
39. Pan X, Streefland M, Dalm C, Wijffels RH, Martens DE. Selection of chemically defined media for CHO cell fed-batch culture processes. Cytotechnology. 2017;69: 39–56.
40. Neermann J, Wagner R. Comparative analysis of glucose and glutamine metabolism in transformed mammalian cell lines, insect and primary liver cells. Journal of Cellular Physiology. 1996;166: 152–169.
41. Lao MS, Toth D. Effects of ammonium and lactate on growth and metabolism of a recombinant Chinese hamster ovary cell culture. Biotechnol Prog. 1997;13: 688–691.
42. Ozturk SS, Riley MR, Palsson BO. Effects of ammonia and lactate on hybridoma growth, metabolism, and antibody production. Biotechnology and Bioengineering. 1992;39: 418–431.
43. Fan Y, Del Val IJ, Müller C, Sen JW, Rasmussen SK, Kontoravdi C, et al. Amino acid and glucose metabolism in fed-batch CHO cell culture affects antibody production and glycosylation. Biotechnology and Bioengineering. 2015;112: 521–535.
44. Warburg O. On the Origin of Cancer Cells. Science. 1956;123: 309–314.
45. Templeton N, Dean J, Reddy P, Young JD. Peak antibody production is associated with increased oxidative metabolism in an industrially relevant fed-batch CHO cell culture. Biotechnology and Bioengineering. 2013;110: 2013–2024.
46. Ahn WS, Antoniewicz MR. Parallel labeling experiments with [1,2-13C]glucose and [U-13C]glutamine provide new insights into CHO cell metabolism. Metabolic Engineering. 2013;15: 34–47.
47. Cruz HJ, Freitas CM, Alves PM, Moreira JL, Carrondo MJT. Effects of ammonia and lactate on
117
growth, metabolism, and productivity of BHK cells. Enzyme and Microbial Technology. 2000;27: 43–52.
48. Sellick CA, Croxford AS, Maqsood AR, Stephens G, Westerhoff HV, Goodacre R, et al. Metabolite profiling of recombinant CHO cells: designing tailored feeding regimes that enhance recombinant antibody production. Biotechnol Bioeng. 2011;108: 3025–3031.
49. Ma N, Ellet J, Okediadi C, Hermes P, McCormick E, Casnocha S. A single nutrient feed supports both chemically defined NS0 and CHO fed-batch processes: Improved productivity and lactate metabolism. Biotechnology Progress. 2009;25: 1353–1363.
50. Lu SC. Glutathione synthesis. Biochimica et Biophysica Acta (BBA) - General Subjects. 2013;1830: 3143–3153.
51. Carrillo-Cocom LM, Genel-Rey T, Araíz-Hernández D, López-Pacheco F, López-Meza J, Rocha-Pizaña MR, et al. Amino acid consumption in naïve and recombinant CHO cell cultures: producers of a monoclonal antibody. Cytotechnology. 2015;67: 809–820.
52. Salazar A, Keusgen M, von Hagen J. Amino acids in the cultivation of mammalian cells. Amino Acids. 2016;48: 1161–1171.
53. Chen P, Harcum S. Effects of elevated ammonium on glycosylation gene expression in CHO cells. Metabolic Engineering. 2006;8: 123–132.
54. Chen P, Harcum SW. Effects of amino acid additions on ammonium stressed CHO cells. Journal of Biotechnology. 2005;117: 277–286.
55. Glacken MW, Fleischaker RJ, Sinskey AJ. Reduction of waste product excretion via nutrient control: Possible strategies for maximizing product and cell yields on serum in cultures of mammalian cells. Biotechnol Bioeng. 1986;28: 1376–1389.
56. Mulukutla BC, Kale J, Kalomeris T, Jacobs M, Hiller GW. Identification and control of novel growth inhibitors in fed-batch cultures of Chinese hamster ovary cells. Biotechnol Bioeng. 2017;114: 1779–1790.
57. Dickson AJ. Enhancement of production of protein biopharmaceuticals by mammalian cell cultures: the metabolomics perspective. Curr Opin Biotechnol. 2014;30: 73–79.
58. Boettcher M, McManus MT. Choosing the Right Tool for the Job: RNAi, TALEN, or CRISPR. Molecular Cell. 2015. pp. 575–585. doi:10.1016/j.molcel.2015.04.028
59. Jinek M, East A, Cheng A, Lin S, Ma E, Doudna J. RNA-programmed genome editing in human cells. Elife. 2013;2: e00471.
60. Hsu PD, Lander ES, Zhang F. Development and Applications of CRISPR-Cas9 for Genome Engineering. Cell. 2014;157: 1262–1278.
61. Sternberg SH, Doudna JA. Expanding the Biologist’s Toolkit with CRISPR-Cas9. Molecular Cell. 2015;58: 568–574.
62. Mali P, Esvelt KM, Church GM. Cas9 as a versatile tool for engineering biology. Nature Methods.
118
2013;10: 957–963.
63. Jinek M, Chylinski K, Fonfara I, Hauer M, Doudna JA, Charpentier E. A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. Science. 2012;337: 816–821.
64. Mojica FJM, Díez-Villaseñor C, García-Martínez J, Almendros C. Short motif sequences determine the targets of the prokaryotic CRISPR defence system. Microbiology. 2009;155: 733–740.
65. Sherkow JS. The CRISPR Patent Landscape: Past, Present, and Future. The CRISPR Journal. 2018;1: 5–9.
66. Turan S, Zehe C, Kuehle J, Qiao J, Bode J. Recombinase-mediated cassette exchange (RMCE) - a rapidly-expanding toolbox for targeted genomic modifications. Gene. 2013;515: 1–27.
67. McLellan MA, Rosenthal NA, Pinto AR. Cre-loxP-Mediated Recombination: General Principles and Experimental Considerations. Curr Protoc Mouse Biol. 2017;7: 1–12.
68. Barnes LM, Bentley CM, Dickson AJ. Advances in animal cell recombinant protein production: GS-NS0 expression system. Cytotechnology. 2000;32: 109–123.
69. Fan L, Frye CC, Racher AJ. The use of glutamine synthetase as a selection marker: recent advances in Chinese hamster ovary cell line generation processes. Pharmaceutical Bioprocessing. 2013;1: 487–502.
70. Kim JY, Kim Y-G, Lee GM. CHO cells in biotechnology for production of recombinant proteins: current state and further potential. Applied Microbiology and Biotechnology. 2012;93: 917–930.
71. Fischer S, Handrick R, Otte K. The art of CHO cell engineering: A comprehensive retrospect and future perspectives. Biotechnology Advances. 2015;33: 1878–1896.
72. Jadhav V, Hackl M, Druz A, Shridhar S, Chung C-Y, Heffner KM, et al. CHO microRNA engineering is growing up: recent successes and future challenges. Biotechnol Adv. 2013;31: 1501–1513.
73. Kim SH, Lee GM. Down-regulation of lactate dehydrogenase-A by siRNAs for reduced lactic acid formation of Chinese hamster ovary cells producing thrombopoietin. Appl Microbiol Biotechnol. 2007;74: 152–159.
74. Zhou M, Crawford Y, Ng D, Tung J, Pynn AFJ, Meier A, et al. Decreasing lactate level and increasing antibody production in Chinese Hamster Ovary cells (CHO) by reducing the expression of lactate dehydrogenase and pyruvate dehydrogenase kinases. Journal of Biotechnology. 2011;153: 27–34.
75. Yip SSM, Zhou M, Joly J, Snedecor B, Shen A, Crawford Y. Complete Knockout of the Lactate Dehydrogenase A Gene is Lethal in Pyruvate Dehydrogenase Kinase 1, 2, 3 Down-Regulated CHO Cells. Molecular Biotechnology. 2014;56: 833–838.
76. Zagari F, Stettler M, Baldi L, Broly H, Wurm FM, Jordan M. High expression of the aspartate–glutamate carrier Aralar1 favors lactate consumption in CHO cell culture. Pharmaceutical Bioprocessing. 2013;1: 19–27.
77. Wlaschin KF, Hu W-S. Engineering cell metabolism for high-density cell culture via manipulation of
119
sugar transport. Journal of Biotechnology. 2007;131: 168–176.
78. Fogolı́n MB, Wagner R, Etcheverrigaray M, Kratje R. Impact of temperature reduction and expression of yeast pyruvate carboxylase on hGM-CSF-producing CHO cells. Journal of Biotechnology. 2004;109: 179–191.
79. Kim SH, Lee GM. Functional expression of human pyruvate carboxylase for reduced lactic acid formation of Chinese hamster ovary cells (DG44). Applied Microbiology and Biotechnology. 2007;76: 659–665.
80. Chong WPK, Reddy SG, Yusufi FNK, Lee D-Y, Wong NSC, Heng CK, et al. Metabolomics-driven approach for the improvement of Chinese hamster ovary cell growth: overexpression of malate dehydrogenase II. J Biotechnol. 2010;147: 116–121.
81. Mulukutla BC, Mitchell J, Geoffroy P, Harrington C, Krishnan M, Kalomeris T, et al. Metabolic engineering of Chinese hamster ovary cells towards reduced biosynthesis and accumulation of novel growth inhibitors in fed-batch cultures. Metab Eng. 2019;54: 54–68.
82. Pereira S, Kildegaard HF, Andersen MR. Impact of CHO Metabolism on Cell Growth and Protein Production: An Overview of Toxic and Inhibiting Metabolites and Nutrients. Biotechnol J. 2018;13: e1700499.
83. Handlogten MW, Zhu M, Ahuja S. Intracellular response of CHO cells to oxidative stress and its influence on metabolism and antibody production. Biochem Eng J. 2018;133: 12–20.
84. Grav LM, Lee JS, Gerling S, Kallehauge TB, Hansen AH, Kol S, et al. One-step generation of triple knockout CHO cell lines using CRISPR/Cas9 and fluorescent enrichment. Biotechnol J. 2015;10: 1446–1456.
85. Grav LM, la Cour Karottki KJ, Lee JS, Kildegaard HF. Application of CRISPR/Cas9 Genome Editing to Improve Recombinant Protein Production in CHO Cells. Methods in Molecular Biology. 2017. pp. 101–118.
86. Orellana CA, Marcellin E, Palfreyman RW, Munro TP, Gray PP, Nielsen LK. RNA-Seq Highlights High Clonal Variation in Monoclonal Antibody Producing CHO Cells. Biotechnol J. 2018;13: e1700231.
87. Ghorbaniaghdam A, Chen J, Henry O, Jolicoeur M. Analyzing clonal variation of monoclonal antibody-producing CHO cell lines using an in silico metabolomic platform. PLoS One. 2014;9: e90832.
88. Lee JS, Kildegaard HF, Lewis NE, Lee GM. Mitigating Clonal Variation in Recombinant Mammalian Cell Lines. Trends Biotechnol. 2019. doi:10.1016/j.tibtech.2019.02.007
89. Brunner M, Fricke J, Kroll P, Herwig C. Investigation of the interactions of critical scale-up parameters (pH, pO and pCO) on CHO batch performance and critical quality attributes. Bioprocess Biosyst Eng. 2017;40: 251–263.
90. Becker M, Junghans L, Teleki A, Bechmann J, Takors R. The Less the Better: How Suppressed Base Addition Boosts Production of Monoclonal Antibodies With Chinese Hamster Ovary Cells. Front Bioeng Biotechnol. 2019;7: 76.
120
91. Noh SM, Park JH, Lim MS, Kim JW, Lee GM. Reduction of ammonia and lactate through the coupling of glutamine synthetase selection and downregulation of lactate dehydrogenase-A in CHO cells. Applied Microbiology and Biotechnology. 2017;101: 1035–1045.
92. Jiménez N, Martínez VS, Gerdtzen ZP. Engineering CHO cells galactose metabolism to reduce lactate synthesis. Biotechnology Letters. 2019;41: 779–788.
93. Torkashvand F, Vaziri B, Maleknia S, Heydari A, Vossoughi M, Davami F, et al. Designed Amino Acid Feed in Improvement of Production and Quality Targets of a Therapeutic Monoclonal Antibody. PLoS One. 2015;10: e0140597.
94. Ghosh A, Zhao H, Price ND. Genome-scale consequences of cofactor balancing in engineered pentose utilization pathways in Saccharomyces cerevisiae. PLoS One. 2011;6: e27316.
95. Stanton RC. Glucose-6-phosphate dehydrogenase, NADPH, and cell survival. IUBMB Life. 2012;64: 362–369.
96. Zhang C, Zhang Z, Zhu Y, Qin S. Glucose-6-phosphate dehydrogenase: a biomarker and potential therapeutic target for cancer. Anticancer Agents Med Chem. 2014;14: 280–289.
97. O’Connor S, Li E, Majors BS, He L, Placone J, Baycin D, et al. Increased expression of the integral membrane protein ErbB2 in Chinese hamster ovary cells expressing the anti-apoptotic gene Bcl-xL. Protein Expr Purif. 2009;67: 41–47.
98. Kim Y-G, Park B, Lee S, Ahn JO, Jung J-K, Lee HW, et al. Effect of mitochondrial and ER-targeted Bcl-2 overexpression on apoptosis in recombinant Chinese hamster ovary cells treated with sodium butyrate. Process Biochemistry. 2012;47: 2518–2522.
99. Kim NS, Lee GM. Response of recombinant Chinese hamster ovary cells to hyperosmotic pressure: effect of Bcl-2 overexpression. Journal of Biotechnology. 2002;95: 237–248.
100. Misaghi S, Qu Y, Snowden A, Chang J, Snedecor B. Resilient immortals, characterizing and utilizing Bax/Bak deficient Chinese hamster ovary (CHO) cells for high titer antibody production. Biotechnol Prog. 2013;29: 727–737.
101. Pan X, Dalm C, Wijffels RH, Martens DE. Metabolic characterization of a CHO cell size increase phase in fed-batch cultures. Appl Microbiol Biotechnol. 2017;101: 8101–8113.
121
Appendices
122
Appendix 1: Paper II – Supplementary materials Reprogramming amino acid catabolism in CHO cells with CRISPR/Cas9 genome editing improves cell growth and reduces byproduct secretion Authors: Daniel Ley1,2, Sara Pereira2, Lasse Ebdrup Pedersen2, Johnny Arnsdorf2, Hooman Hefzi3,4,
Anne Mathilde Lund1, Tae Kwang Ha2, Tune Wulff2, Helene Faustrup Kildegaard2,*, Mikael Rørdam
Andersen1,*.
(1) Department of Biotechnology and Biomedicine, Technical University of Denmark, Kgs. Lyngby,
Denmark. (2) The Novo Nordisk Foundation Center for Biosustainability, Technical University of
Denmark, Kgs. Lyngby, Denmark. (3) Department of Bioengineering, University of California, San Diego,
La Jolla, CA 92093, United States. (4) Novo Nordisk Foundation Center for Biosustainability at the
University of California, San Diego, School of Medicine, La Jolla, CA 92093, United States.
*Corresponding authors:
Phone: +45 45 25 26 75, Fax: +45 45 88 41 48
Address: Søltofts plads, bygning 223, 2800 Kgs Lyngby, Denmark.
123
p-va
lue
> α
= 0.
05
p-va
lue
≤ α
= 0.
05
Log
fold
cha
nge
7.7
-10.
60
Di
eren
tial g
ene
expr
essio
n in
CH
O-K
1 ce
lls p
rodu
cing
Ig
G re
lativ
e to
a n
on-p
rodu
cer
Met
abol
ite co
lors
Esse
ntia
l am
ino
acid
Ines
sent
ial a
min
o ac
id
Redo
x ac
tivity
Miss
ing
gene
exp
ress
ion
data
LOC1
0075
4689
Tdo2
Ido1
Ido2
Kmo
LOC1
0077
3211
Kynu
Haa
oAc
msd
Aass
Aass
Aada
tO
gdhl
Ogd
h
LOC1
0075
6509
LOC1
0075
8466
LOC1
0075
7019
LOC1
0077
4025
Had
ha
Ehha
dh
Dld
LOC1
0076
3658
LOC1
0075
2344
Hsd
17b1
4
LOC1
0075
3968
LOC1
0077
4234
LOC1
0077
3651
LOC1
0077
4277
Dbt
Dld
Ivd
Mcc
c1
Mcc
c2
LOC1
0077
4025
Oxc
t1
LOC1
0077
1204
LOC1
0077
2647
LOC1
0075
5385
Ehha
dh
Had
hb
Hm
gcs2
Acat
2
Acat
1
LOC1
0076
0294
LOC1
0076
1491
Adsl
LOC1
0075
3207
Il4i1
LOC1
0075
2344
LOC1
0076
3658
Hsd
17b1
4
LOC1
0075
3968
LOC1
0077
4234
LOC1
0077
3651
LOC1
0077
4277
Dbt
Dld
LOC1
0076
3159
Acad
sb
Acad
s
LOC1
0075
4654
Ehha
dh
Had
ha
LOC1
0076
1491
Ehha
dh
Hsd
17b1
0
LOC1
0076
5544
LOC1
0075
3943
LOC1
0075
4813
Acaa
2
LOC1
0076
7356
Pcca
Pccb
Mce
e
Mut
LOC1
0076
3658
LOC1
0075
2344
LOC1
0077
4277
LOC1
0077
4234
Dbt
Dld
LOC1
0076
3159
Acad
sb
Acad
s
Acad
8
LOC1
0075
4654
Ehha
dh
Had
ha
Hib
chLO
C100
7614
91
Ehha
dh
LOC1
0076
0581
LOC1
0075
7658
LOC1
0075
6509
LOC1
0077
2225
LOC1
0076
5484
LOC1
0076
5774
LOC1
0075
9157
Bzw
1
Aldh
6a1
LOC1
0076
4455
Hgd
LOC1
0076
8220
Got
1G
ot2
Il4i1
LOC1
0075
3207
Tat
Pah
LOC1
0076
1742
LOC1
0076
2926
Mao
a
Mao
b
Aoc2
LOC1
0076
3332
Aldh
3a1
Aldh
1a3
LOC1
0076
1742
LOC1
0076
2926
Mao
a
Mao
b
Aoc2
LOC1
0076
3332
Aldh
3a1
Aldh
1a3
Gpt
LOC1
0076
8176
Ldha
Pkm
Pklr
Eno1
Eno2
LOC1
0077
1169
Pgam
2LO
C100
7660
80
LOC1
0075
1347
Pgm
2Pg
m5
Gpi
Fbp2
LOC1
0076
1539
Pfk1
Pfkm
Pfkp
Aldo
aAl
dob
Aldo
c
Tpi1
Gap
dhLO
C100
7591
00G
apdh
s
Pgk1
Pgk2
Pgam
1Bp
gm
Eno3
Pck2
Pck1
Ldhb
Ldha
16b
LOC1
0076
4142
LOC1
0075
1672
Pgm
211
Pdha
1
Dld
Dla
t
LOC1
0077
2790
Pdhb
Bckd
hb
Pdhx
LOC1
0077
4281
LOC1
0076
0225
Asns
Got
1
Got
2
Il4i1
LOC1
0075
3207
LOC1
0076
7083
LOC1
0075
1825
LOC1
0075
0580
LOC1
0077
3325
Arg1
Arg2
Otc
LOC1
0075
0753
Asl
LOC1
0076
4994
LOC1
0075
4097
Oat
Glu
l
LOC1
0077
4486
Gls
Cad
Nfu
1G
fpt2
Amdh
d1
Uro
c1Zx
dc
Hal
Ccdc
38
Cps1
LOC1
0075
7642
Gad
1G
adl1
LOC1
0076
0149
Glu
d1
Abat
Aldh
5a1
Prod
h
Pycr
2
P4ha
1P4
ha2
P4ha
3
Pycr
1Py
cr2
LOC1
0075
8600
Prod
h2
LOC1
0076
4994
LOC1
0076
4994
Got
1G
ot2
Hog
a1
LOC1
0077
1382
Aldh
1b1
Aldh
9a1
LOC1
0075
1303
LOC1
0075
6509
LOC1
0075
7658
LOC1
0075
6588
Odc
Agm
at
Abat
Aldh
5a1
LOC1
0076
5797
LOC1
0076
9680
Gat
m
LOC1
0075
5964
Hm
mr
Mat
1aM
at2a
LOC1
0077
2831
Dnm
t3l
Dnm
t3a
Ahcy
l1
Ahcy
LOC1
0076
2713
LOC1
0075
1727
LOC1
0075
4476
Bhm
tBh
mt2
LOC1
0076
3416
LOC1
0077
4596
Mta
p
Mri1
Apip
Enop
h1
Adi1
Tat
Il4i1
Znf6
92LO
C100
7532
07
LOC1
0075
5964
LOC1
0075
1960
LOC1
0076
2141
Cbs
Cth
Cth
Got
1
Got
2
LOC1
0076
3920
Tst
LOC1
0077
0821
LOC1
0075
1672
LOC1
0077
0347
Ldha
LOC1
0076
0315
Got
1G
ot2
LOC1
0076
2141
Shm
t1Ag
xtSh
mt2
Sds
Sdsl
Agxt
LOC1
0075
2744
Gly
ctk
Bpgm
Pgam
2
LOC1
0076
6080
Phgd
h
Psph
LOC1
0076
9555
Gca
t
SdsM
tmr9
Mao
aM
aob
Aoc2
LOC1
0076
2926
Pgam
1
Gly
ctk
Psat
1
Gat
m
Dao Ag
xt
Gld
cAm
t
Dld
Alas
2Al
as1
LOC1
0076
9585
Pipo
x
Sard
h
LOC1
0075
6588
Aldh
9a1
Aldh
1b1
L-Pr
olin
e (S
)-1-P
yrro
line-
5-ca
rbox
ylat
e
Acce
ptor
Redu
ced
acce
ptor
NAD
+NA
DH
+ H+
L-G
luta
mat
e
NAD
+ +
2 H 20
NAD
H +
H+
L-G
luta
mat
e5-
sem
iald
ehyd
e
Spontaneous
Ort
hini
ne
L-Gl
utam
ate
2-O
xogl
utar
ate
Citr
ulin
e
L-A
rgin
ino
succ
inat
e
L-Ar
gini
ne
NAD
+ +
H 20NA
DH
+ H+
+ NH
3
L-G
luta
min
e ATP
+ NH
3
ADP
+ O
rtho
phos
phat
e
H 2O
Carb
amoy
lPh
osph
ate
2 AT
P +
CO2 +
H20
2 AD
P +
Ort
hoph
osph
ate
Ort
hoph
osph
ate
Ure
a cy
cle
4-H
ydro
xypr
olin
e2-O
xogl
utar
ate
+ O
2
Succ
inat
e +
CO2
L-1
-Pyr
rolin
e-3-
hydr
oxy-
5-ca
rbox
ylat
e
NAD
P+
NAD
PH +
H+
NAD
+
NAD
H +
H+
L-e
ryth
ro-4
-Hyd
roxy
glut
amat
e
NAD
+ +
2 H 20
NAD
H +
H+
L-1
-Pyr
rolin
e-3-
hydr
oxy-
5-ca
rbox
ylat
e
Spon
tane
ous
NAD
+ +
H 20
NAD
H +
H+
(4R)
-4-H
ydro
xy-2
-oxo
glut
arat
e
2-O
xogl
utar
ate
L-Gl
utam
ate
Pyru
vate
Glyo
xyla
te
H 20
Urea
ATPAM
P +
Dip
hosp
hate
Fum
arat
e
Citri
c aci
d cy
cle
Putr
esci
ne
CO2
4-A
min
obut
yral
dehy
de
O2 +
H20
NH3 +
Hyd
roge
n pe
roxi
de
4-A
min
obut
anoa
teNAD
+ +
H 20
NAD
H +
H+
Sper
mid
ine
5'-M
ethy
lthio
aden
osin
e
5'-M
ethy
lthio
aden
osin
eSper
min
e
Beta
-ala
nine
met
abol
ism
Glut
athi
one
met
abol
ism
S-Ad
enos
yl-L
-met
hion
ine
S-Ad
enos
ylm
ethi
onin
amin
e
H+
CO2
Gua
nidi
noac
etat
e
Glyc
ineL-
Orn
ithin
eCr
eatin
e
Agm
atin
eCO
2
Urea
H 20
L-M
ethi
onin
e
Ort
hoph
osph
ate
+ D
ipho
spha
teAT
P +
H 2O
S-Ad
enos
yl-L
-hom
ocys
tein
e
DNA
DNA
5-
met
hylc
ytos
ine
L-H
omoc
yste
ine
H 20Ad
enos
ine
Beta
ine
N,N-
Dim
ethy
lgly
cine
5-M
ethy
ltetra
hydr
ofol
ate
Tetra
hydr
ofol
ate
S-M
ethy
l-5-t
hio-
D-
ribo
se-1
-pho
spha
te
S-M
ethy
l-5-t
hio-
D-
ribu
lose
-1-p
hosp
hate
2,3-
Dik
eto-
5-m
ethy
lthio
pent
yl-1
-pho
spha
te
4-M
ethy
lthio
-2-o
xobu
tano
ic
acid
Ort
hoph
osph
ate
Aden
ine
H 20
1,2-
Dih
ydro
xy-5
-(met
hylth
-io
)pen
t-1-
en-3
-one
Ort
hoph
osph
ate
H 20
Form
ate
O2
L-Gl
utam
ate
2-O
xogl
utar
ate
H 20 +
O2
NH3 +
Hyd
roge
n pe
roxi
de
L-Cy
stat
hion
ine
H 20
L-Cy
stei
ne
H 20
NH3 +
2-O
xobu
tano
ate
H 20
NH3 +
Hyd
roge
nsul
de
Pyru
vate
Mer
capt
opyr
uvat
e
2-O
xogl
utar
ate
L-Gl
utam
ate
Sul
te
Thio
sulfa
te
3-M
erca
ptol
acta
teNAD
H +
H+
NAD
+
O2
3-Su
lno
-L-A
lani
ne
3-Su
lny
lpyr
uvat
e 2-O
xogl
utar
ate
L-Gl
utam
ate
Spon
tane
ous
Sul
te
L-Se
rine
Gly
cine
Gua
nidi
noac
etat
e
Gly
oxyl
ate
Hydr
ogen
sul
de
H 20
5,10
-Met
hyle
nete
trahy
drof
olat
e +
H 20
Tetro
hydr
ofol
ate
Glyo
xyla
te
Hydr
oxyp
yruv
ate
Pyru
vate
Sul
te
Spon
tane
ous
NH3
O-P
hosp
ho-L
-ser
ine
Hyd
roxy
pyru
vate
3-Ph
osph
onoo
xypy
ruva
te
D-G
lyce
rate
3-Ph
osph
o-D
-gly
cera
te
2-Ph
osph
o-D
-gly
cera
te
Pyru
vate
L-Al
anin
e
NAD
+ NA
DH
+ H+
ADP
ATP
Glyc
olys
is
NAD
+
NAD
H +
H+
L-Gl
utam
ate
2-O
xogl
utar
ate
H 20
Ort
hoph
osph
ate
L-Ar
gini
neL-
Ort
hini
ne
S-Ad
enos
yl-L
-met
hion
ine
S-Ad
enos
yl-L
-hom
ocys
tein
e
H 2O +
O2
NH3
+ hy
drog
en p
erox
ide
Glyo
xyla
te
met
abol
ism
Glyo
xyla
te
met
abol
ism
Pyru
vate
L-Al
anin
eLO
C100
7599
06
L-Th
reon
ine
Acet
alde
hyde
L-2-
Am
ino-
3-ox
obut
anoi
c ac
id
Acet
yl-C
oA
CoA
2-O
xobu
tano
ate
NH3
NAD
+
NAD
H +
H+
Spon
tane
ous
Am
inoa
ceto
neM
ethy
lgly
oxal
H 2O +
O2
NH3 +
Hyd
roge
n pe
roxi
de
Pyru
vate
met
abol
ism
S-Am
inom
ethy
ldi
hydr
olip
oylp
rote
in
Lipo
ylpr
otei
n
Dih
ydro
lipoy
lpro
tein
CO2
Tetro
hydr
ofol
ate
5,10
-Met
hyle
nete
trahy
drof
olat
e +
NH3
NAD
+ NA
DH
+ H+
5-A
min
olev
ulin
ate
Succ
inyl
-CoA
CoA
+ CO
2
Porp
hyrin
eBi
osyn
thes
is
Sarc
osin
e
S-Ad
enos
yl-L
-met
hion
ine S-
Aden
osyl
-L-h
omoc
yste
ine
H 2O +
O2
Form
alde
hyde
+ hy
drog
en p
erox
ide
Redu
ced
elec
tron-
trans
ferr
ing
avop
rote
inEl
ectro
n-tra
nsfe
rrin
g av
opro
tein
+ H
2O
Glyc
erop
hosp
ho-
lipid
met
abol
ismO
xalo
acet
ate
Citr
ate
cis-
Acet
onitr
ate
D-Is
ocitr
ate
2-O
xogl
utar
ate
Succ
inyl
-CoA
Succ
inat
e
Fum
arat
e
Mal
ate
Citr
ic a
cid
cycl
e
L-A
spar
tate
Aden
ylos
ucci
nate
4-A
min
obut
anoa
teSu
ccin
ate
sem
iald
ehyd
e
Acet
yl-C
oA
4-Fu
mar
ylac
etoa
ceta
te
L-Ty
rosi
ne
L-Ph
enyl
alan
ine
L-A
spar
agin
e AMP
+ D
ipho
spha
te
ATP
+ NH
3
2-O
xogl
utar
ate
L-Gl
utam
ate
O2 +
H20
NH3 +
Hyd
roge
n pe
roxi
de
GTP
+ IM
PGD
P +
Ort
hoph
osph
ate
AMP
CoA
2 AT
P +
H 20 +
HCO
3-
2 AD
P +
Ort
hoph
osph
ate+
L-Gl
utam
ate
D-G
luco
sam
ine
6-ph
osph
ate
D-F
ruct
ose
6-ph
osph
ate
L-Gl
utam
ate
N-F
orm
imin
o-L-
glut
amat
e
5-F
orm
imin
otet
rahy
drof
olat
e
Tetra
hydr
ofol
ate
4-Im
idaz
olon
e-5-
prop
anoa
te
H 2O
H 2O
Uro
cana
te
L-H
istid
ine
NH3
CO2
L-Gl
utam
ate
Succ
inat
e se
mia
ldeh
yde
Succ
inat
eCi
tric a
cid
cycl
e
NAD
+ +
H 20
NAD
H +
H+
NAD
+ +
H 20NA
DH
+ H+
L-Gl
utam
ate
5-P
hosp
ho-a
lpha
-D-r
ibos
e 1-
diph
osph
ate
5-Ph
osph
orib
osyl
amin
e +
Dip
hosp
hate
L-Gl
utam
ine
+ H 20
Purin
e m
etab
olism
L-Is
oleu
cine
(S)-3
-Met
hyl-2
-oxo
pent
a-no
ic a
cid
H 2O +
Hyd
roge
n pe
roxi
deNH
3 + H
ydro
gen
pero
xide
2-O
xogl
utar
ate
L-Gl
utam
ate
S-(2
-met
hylb
utan
oyl)d
i-hy
drol
ipoy
llysi
ne
Enzy
me
N6-(l
ipoy
l)lys
ine
[Dih
ydro
lipoy
llysin
e-re
sidue
(2-m
ethy
l-pr
opan
oyl)t
rans
fera
se]
(S)-2
-Met
hylb
utan
oyl-C
oA
CoA
Enzy
me
N6-(d
ihyd
rolip
oyl)l
ysin
e
NAD
+ NA
DH
+ H+
2-M
ethy
lbut
-2-e
noyl
-CoA
Acce
ptor
Redu
ced
acce
ptor
(2S,
3S)-3
-Hyd
roxy
-2-
met
hylb
utan
oyl-C
oA
H 2O
2-M
ethy
lace
toac
etyl
-CoA
NAD
H +
H+NA
D+
Prop
anoy
l-CoA
S-M
ethy
lmal
onat
e-se
mi-
alde
hyde
R-M
ethy
lmal
onyl
-CoA
Met
hylm
alon
ate
Spon
tane
ous
S-M
ethy
lmal
onyl
-CoA
CoA
ADP
+ O
rtho
phos
phat
eAT
P +
HCO
3-
(S)-3
-Hyd
roxy
isob
utyr
yl-C
oAL-
Valin
e
2-O
xogl
utar
ate
L-Gl
utam
ate 3-
Met
hyl-2
-oxo
buta
noat
eS-
(2-m
ethy
lpro
pano
yl)d
i-hy
drol
ipoy
llysi
ne
Enzy
me
N6-(l
ipoy
l)lys
ine
[Dih
ydro
lipoy
llysin
e-re
sidue
(2-m
ethy
l-pr
opan
oyl)t
rans
fera
se]
2-M
ethy
lpro
pano
yl-C
oA
CoA
Enzy
me
N6-(d
ihyd
rolip
oyl)l
ysin
e
NAD
+ NA
DH
+ H+
Met
hyla
cryl
yl-C
oA
Acce
ptor
Redu
ced
acce
ptor
H 2O
(S)-3
-Hyd
roxy
isob
utyr
ate
H 2OCo
ANA
DH
+ H+
NAD
+ NA
DH
+ H+
NAD
+
2 Hy
drog
en p
erox
ide
2 H 2O
+ O
2
CoA
+ NA
D+
CO2 +
NAD
H +
H+
L-Le
ucin
e
2-O
xogl
utar
ate
L-Gl
utam
ate 4-
Met
hyl-2
-oxo
pent
anoa
teS-
(3-m
ethy
lbut
anoy
l)di-
hydr
olip
oylly
sine
Enzy
me
N6-(l
ipoy
l)lys
ine
[Dih
ydro
lipoy
llysin
e-re
sidue
(2-m
ethy
l-pr
opan
oyl)t
rans
fera
se]
3-M
ethy
lbut
anoy
l-CoA
CoA
Enzy
me
N6-(d
ihyd
rolip
oyl)l
ysin
e
NAD
+ NA
DH
+ H+
3-M
ethy
lcro
tony
l-CoA
Elec
tron-
trans
ferr
ing
avop
rote
inRe
duce
d el
ectro
n-tra
nsfe
rrin
g av
opro
tein
LOC1
0076
3159
FAD
FAD
H 2
3-M
ethy
lglu
taco
nyl-C
oA
ATP
+ HC
O3-
ADP
+ O
rtho
phos
phat
e
(S)-3
-Hyd
roxy
-3-m
ethy
lglu
tary
l-CoA
H 2O
Acet
oace
tate
Acet
oace
tyl-C
oA
Succ
inyl
-CoA
Succ
inat
e
Acet
yl-C
oA
CoA
Bran
ched
chai
n fa
tty a
cid
synt
hesis
Bran
ched
chai
n fa
tty a
cid
synt
hesis
Bran
ched
chai
n fa
tty a
cid
synt
hesis
L-Ly
sine
Sacc
haro
pine
2-O
xogl
utar
ate
+ NA
DPH
+ H
+NA
DP+ +
H2O
NAD
+ + H
2ONA
DH
+ H+ +
L-Gl
utam
ate
L-2-
Am
inoa
dipa
te
6-se
mia
ldeh
yde
NAD
+ + H
2ONA
DH
+ H+
L-2-
Am
inoa
dipa
te
2-O
xogl
utar
ate
L-Gl
utam
ate 2-
Oxo
adip
ate
S-gl
utar
yldi
hydr
olip
oylly
sine
Enzy
me
N6-(l
ipoy
l)lys
ine
[Dih
ydro
lipoy
llysin
e-re
sidue
succ
inyl
-tra
nsfe
rase
] + C
O2
CoA
Enzy
me
N6-(d
ihyd
rolip
oyl)l
ysin
e
NAD
+ NA
DH
+ H+
NAD
+
NAD
H +
H+
Glu
tary
l-CoAEl
ectro
n-tra
nsfe
rrin
g av
opro
tein
Redu
ced
elec
tron-
trans
ferr
ing
avop
rote
in +
CO
2 Crot
onoy
l-CoA
Crot
onoy
l-CoA
H 2O
L-Tr
ypto
phan
O2
L-Fo
rmyl
kynu
reni
ne
H 2OFo
rmat
e
L-Ky
nure
nine
NAD
PH +
H+ +
O2
NAD
P+ + H
2O
3-H
ydro
xy-L
-kyn
uren
ine
H 2OL-
Alan
ine
3-H
ydro
xyan
thra
nila
te
O2
2-A
min
o-3-
carb
oxym
ucon
ate
sem
iald
ehyd
e
CO2
2-A
min
omuc
onat
e se
mia
ldeh
yde
2-O
xoad
ipat
e2
mis
sing
gen
esLy
sine
cata
bolis
m
Acet
oace
tate
H 2O
4-M
aley
lace
toac
etat
e
O2
Hom
ogen
isat
e3-
(4-H
ydro
xyph
enyl
)pyr
uvat
e
O2
CO2
L-Gl
utam
ate
2-O
xogl
utar
ate
H 2O +
O2
NH3 +
Hyd
roge
n pe
roxi
de
Tetra
hydr
obio
pter
in +
O2
Dih
ydro
biop
terin
+ H
2O
Tyra
min
e
CO2
H 2O +
O2
NH3
+ Hy
drog
en p
erox
ide
s uoenat nopS
4-H
ydro
xyph
enyl
acet
-al
dehy
de4-
Hyd
roxy
phen
ylac
etat
e
H 2O +
NAD
+ NA
DH
+ H+
Phen
ethy
lam
ine
CO2
H 2O +
O2
NH3
+ Hy
drog
en p
erox
ide
Phen
ylac
etal
dehy
de
H 2O +
NAD
+ NA
DH
+ H+
Phen
ylac
etic
aci
d
NH3
Extra
cellu
lar s
ecre
tion
Pyru
vate
L-La
ctat
eL-
Ala
nine
2-O
xogl
utar
ate
L-Gl
utam
ate
Phos
phoe
nolp
yruv
ate
ATP
ADP
2-Ph
osph
o-D
-gly
cera
te
H 2O
3-Ph
osph
o-D
-gly
cera
te
ADP
ATP
Glu
cose
-1-p
hosp
hate
Glu
cose
-6-p
hosp
hate
Fruc
tose
-6-p
hosp
hate
Fruc
tose
-1,6
-bis
phos
phat
e
H 2O
Ort
hoph
osph
ate
Gly
cera
ldeh
yde-
3-ph
osph
ate
Gly
cero
ne p
hosp
hate
3-Ph
osph
o-D
-gly
cero
yl
phos
phat
eNAD
H +
H+
Ort
hoph
osph
ate
+ NA
D+
GDP
+ CO
2GT
P
2-O
xogl
utar
ate
Citri
c aci
d cy
cle
[Dih
ydro
lipoy
llysin
e-re
sidue
ace
tyltr
ansf
eras
e] +
CO
2
Enzy
me
N6-(l
ipoy
l)lys
ine
Enzy
me
N6-(d
ihyd
rolip
oyl)l
ysin
e
Extra
cellu
lar s
ecre
tion
S-ac
etyl
dihy
drol
ipoy
llysi
ne
[Dih
ydro
lipoy
llysin
e-re
sidue
ace
tyltr
ansf
eras
e] +
CoA
Extra
cellu
lar s
ecre
tion
NAD
H +
H+
NAD
+
NAD
+NA
DH
+ H+
Reco
nstr
uctio
n of
the
amin
o ac
id m
etab
olis
m in
Chi
nese
Ham
ster
Ova
ry ce
lls
III
Figure S3. Gene expression levels of target genes in single gene disruption of multiple
clones. Transcription rate was quantified using two primer pairs targeting coding regions
upstream and downstream relative to the gRNA target site. Gene expression levels are
normalized to the wild type expression. Error bars indicate standard deviation of three
biological replicates.
125
S1: P
rimer
sPr
imer
nam
eTa
rget
gen
ePu
rpos
ePr
imer
seq
uenc
e 5'
to 3
' M
iSeq
_AAS
S_12
1891
5_Fw
dAa
ssM
iseq
prim
erTC
GTC
GG
CAG
CGTC
AGAT
GTG
TATA
AGAG
ACAG
TGAG
AGTG
CAG
AGAC
CGAG
AM
iseq
_AAS
S_12
1891
5_Re
vAa
ssM
iseq
prim
erG
TCTC
GTG
GG
CTCG
GAG
ATG
TGTA
TAAG
AGAC
AGCG
TCG
ATTG
GAA
GG
CTG
GAT
MiS
eq_A
FMID
_487
593_
Fwd
Afm
idM
iseq
prim
erTC
GTC
GG
CAG
CGTC
AGAT
GTG
TATA
AGAG
ACAG
AGG
AGAA
GCT
GG
GTC
AGG
ATM
iSeq
_AFM
ID_4
8759
3_Re
vAf
mid
Mis
eq p
rimer
GTC
TCG
TGG
GCT
CGG
AGAT
GTG
TATA
AGAG
ACAG
TTTG
TCCT
GG
AAG
ACCA
GCC
MiS
eq_D
dc_1
0053
45_f
wd
Ddc
Mis
eq p
rimer
TCG
TCG
GCA
GCG
TCAG
ATG
TGTA
TAAG
AGAC
AGG
CCCC
ACAG
TAAC
TGTT
CCA
MiS
eq_D
dc_1
0053
45_r
evDd
cM
iseq
prim
erG
TCTC
GTG
GG
CTCG
GAG
ATG
TGTA
TAAG
AGAC
AGTG
AAG
CCAA
TGCA
GCC
GAT
AM
iSeq
_Gad
1_N
W_0
0361
3606
.1_1
7770
49_f
wd
Gad
1M
iseq
prim
erTC
GTC
GG
CAG
CGTC
AGAT
GTG
TATA
AGAG
ACAG
ACAG
TAG
AGAC
CCCG
AGAC
CM
iSeq
_Gad
1_N
W_0
0361
3606
.1_1
7770
49_r
evG
ad1
Mis
eq p
rimer
GTC
TCG
TGG
GCT
CGG
AGAT
GTG
TATA
AGAG
ACAG
CCCC
AGCT
GCA
GTC
CATT
TAM
iSeq
_Gad
2_N
W_0
0361
5130
.1_1
5548
9_fw
dG
ad2
Mis
eq p
rimer
TCG
TCG
GCA
GCG
TCAG
ATG
TGTA
TAAG
AGAC
AGG
GAG
ACTC
TGAG
AAG
CCAG
CM
iSeq
_Gad
2_N
W_0
0361
5130
.1_1
5548
9_re
vG
ad2
Mis
eq p
rimer
GTC
TCG
TGG
GCT
CGG
AGAT
GTG
TATA
AGAG
ACAG
CGCC
TTTA
CCTG
TTG
CGTT
GM
iSeq
_Hpd
_437
652_
fwd
Hpd
Mis
eq p
rimer
TCG
TCG
GCA
GCG
TCAG
ATG
TGTA
TAAG
AGAC
AGTG
AGAC
TTCT
TCTT
GCC
CGG
MiS
eq_H
pd_4
3765
2_re
vHp
dM
iseq
prim
erG
TCTC
GTG
GG
CTCG
GAG
ATG
TGTA
TAAG
AGAC
AGCA
CTCA
AGG
GG
TGTG
TCCT
CM
iSeq
_LO
C100
7598
74_N
W_0
0361
3673
.1_1
8457
40_f
wd
LOC1
0075
9874
Mis
eq p
rimer
TCG
TCG
GCA
GCG
TCAG
ATG
TGTA
TAAG
AGAC
AGTG
GCT
GG
ATG
GAT
GTA
AGG
CM
iSeq
_LO
C100
7598
74_N
W_0
0361
3673
.1_1
8457
40_r
evLO
C100
7598
74M
iseq
prim
erG
TCTC
GTG
GG
CTCG
GAG
ATG
TGTA
TAAG
AGAC
AGCA
GG
GG
AGCT
GCC
TAG
AAAC
MiS
eq_P
rodh
_NW
_003
6138
98.1
_123
1978
_fw
dPr
odh
Mis
eq p
rimer
TCG
TCG
GCA
GCG
TCAG
ATG
TGTA
TAAG
AGAC
AGG
GTC
TCTC
AACA
GG
GCC
GM
iSeq
_Pro
dh_N
W_0
0361
3898
.1_1
2319
78_r
evPr
odh
Mis
eq p
rimer
GTC
TCG
TGG
GCT
CGG
AGAT
GTG
TATA
AGAG
ACAG
GCG
ATCT
GG
ACCA
CCG
AAAT
MiS
eq_P
rodh
2_N
W_0
0361
4167
.1_6
2300
6_fw
dPr
odh2
Mis
eq p
rimer
TCG
TCG
GCA
GCG
TCAG
ATG
TGTA
TAAG
AGAC
AGCT
GG
AAG
CTG
ACCT
CCAC
TGM
iSeq
_Pro
dh2_
NW
_003
6141
67.1
_623
006_
rev
Prod
h2M
iseq
prim
erG
TCTC
GTG
GG
CTCG
GAG
ATG
TGTA
TAAG
AGAC
AGCC
CCTC
CCCA
GTG
TCAC
TTA
Aass
5' #
2 Fw
dAa
ssqP
CRTG
AGAG
TGCA
GAG
ACCG
AGA
Aass
5' #
2 Re
vAa
ssqP
CRAT
CACT
GG
CTTG
TGG
TGG
AGAa
ss 3' #
4 Fw
dAa
ssqP
CRG
GG
CTTA
CTG
GG
GG
ATG
AAC
Aass
3' #
4 Re
vAa
ssqP
CRCC
AGAA
GG
ATG
TCTG
ATG
CCA
Afm
id 5' #
3 Fw
dAf
mid
qPCR
TCCC
CTAT
GG
AGAT
GG
CGAA
Afm
id 5' #
3 Re
vAf
mid
qPCR
ACCA
TGAA
GG
CCG
AGTC
ATC
Afm
id 3' #
1 Fw
dAf
mid
qPCR
CGTG
ATG
GTG
GTG
ATG
GTA
ATA
Afm
id 3' #
1 Re
vAf
mid
qPCR
TTTG
TGCT
GG
TGG
TCTC
TGDd
c 5'
#3
Fwd
Ddc
qPCR
GAG
GG
ACG
TGCT
GTG
TATC
CDd
c 5'
#3
Rev
Ddc
qPCR
ATAT
GCA
TCCG
GTT
CCTG
GG
Ddc
3' #
3 Fw
dDd
cqP
CRTC
GG
GG
CTCA
TCAC
TGAC
TADd
c 3'
#3
Rev
Ddc
qPCR
TGTA
AGCC
TGCA
GTC
CCTT
GGa
d1 5' #
4 Fw
dG
ad1
qPCR
TAG
CCCA
TGG
ATG
CACC
AGA
Gad1
5' #
4 Re
vG
ad1
qPCR
GAG
GAC
TGCC
TCTC
CCTG
AAGa
d1 3' #
4 Fw
dG
ad1
qPCR
CGAG
CAG
ATCC
TGG
TTG
ACT
Gad1
3' #
4 Re
vG
ad1
qPCR
CAG
CCAC
TCG
CCAG
CTAA
AGa
d2 5' #
4 Fw
dG
ad2
qPCR
CTG
CACC
TGCG
ACCA
AAAA
CGa
d2 5' #
4 Re
vG
ad2
qPCR
AATG
CCAG
TGTG
GG
TCTC
TCGa
d2 3' #
1 Fw
dG
ad2
qPCR
CTTC
TTCC
GCA
TGG
TCAT
CTGa
d2 3' #
1 Re
vG
ad2
qPCR
AGTT
TGAT
GAG
CGAG
GTG
ATTA
Hpd
5' #
1 Fw
dHp
dqP
CRAC
AAG
TTCG
GG
AAG
GTG
AAG
Hpd
5' #
1 Re
vHp
dqP
CRG
GCC
TCAA
ATCC
AGG
TAAG
AAHp
d 3'
#4
Fwd
Hpd
qPCR
GCA
GG
CAAC
TTCA
ACTC
CCT
126
Hpd
3' #
4 Re
vHp
dqP
CRAT
TCCT
GAC
CTCA
CCCC
GTT
LOC1
0075
9874
5' #
3 Fw
dLO
C100
7598
74qP
CRAG
AAG
CCCT
CTG
CTCA
TGTC
LOC1
0075
9874
5' #
3 Re
vLO
C100
7598
74qP
CRCC
AGCT
GAT
ACG
GTG
GTT
CALO
C100
7598
74 3' #
3 Fw
dLO
C100
7598
74qP
CRTG
ATG
ACAG
CAAT
GCC
CGTA
LOC1
0075
9874
3' #
3 Re
vLO
C100
7598
74qP
CRTT
TGCT
TGG
GCA
ATTC
TGG
TGPr
odh
5' #
1 Fw
dPr
odh
qPCR
CGAG
GAC
CAG
GAG
TCTA
TCA
Prod
h 5'
#1
Rev
Prod
hqP
CRG
GG
CTCA
TATC
TTCC
TCCA
TTC
Prod
h 3'
#1
Fwd
Prod
hqP
CRTC
TGCA
GG
ATG
GAG
GAG
TTA
Prod
h 3'
#1
Rev
Prod
hqP
CRCT
GG
CCTA
ATG
GG
AAG
CTAA
TPr
odh2
5' #
4 Fw
dPr
odh2
qPCR
TGG
TGCC
TTCC
ATCT
CAAG
GPr
odh2
5' #
4 Re
vPr
odh2
qPCR
CTG
AAAC
GCT
AGTC
CATG
GG
TPr
odh2
3' #
4 Fw
dPr
odh2
qPCR
TGG
TGG
CTTC
CCAC
AATG
AAPr
odh2
3' #
4 Re
vPr
odh2
qPCR
GTT
GTC
CAAA
ACAG
ACCG
GC
Aass
gDN
A-se
q #1
Fw
dAa
ssIn
del s
eque
ncin
g on
gDN
AAA
ATCT
CAG
GG
GG
AGCG
TTG
Aass
gDN
A-se
q #1
Rev
Aass
Inde
l seq
uenc
ing
on g
DNA
AGAC
CACA
CGAG
AAG
CAAG
GAf
mid
gDN
A-se
q #1
Fw
dAf
mid
Inde
l seq
uenc
ing
on g
DNA
ACAG
AGG
GG
AGG
GAG
GC
Afm
id g
DNA-
seq
#1 R
evAf
mid
Inde
l seq
uenc
ing
on g
DNA
CCCT
CTG
TCTT
TCCA
CAG
TAAA
Ddc
gDN
A-se
q #1
Fw
dDd
cIn
del s
eque
ncin
g on
gDN
AAG
TCG
GTT
CACC
ACAG
TGAC
Ddc
gDN
A-se
q #1
Rev
Ddc
Inde
l seq
uenc
ing
on g
DNA
CTG
ATAG
GCT
GG
GCA
GTA
GC
Gad
1 gD
NA-
seq
#1 F
wd
Gad
1In
del s
eque
ncin
g on
gDN
ACC
TTG
GAA
GCC
CCTA
AGCT
CG
ad1
gDN
A-se
q #1
Rev
Gad
1In
del s
eque
ncin
g on
gDN
ATG
CCCT
CACT
CGTC
AATA
GC
Gad
2 gD
NA-
seq
#1 F
wd
Gad
2In
del s
eque
ncin
g on
gDN
AG
CTCT
ATG
GAG
ACTC
TGAG
AAG
CG
ad2
gDN
A-se
q #1
Rev
Gad
2In
del s
eque
ncin
g on
gDN
AG
TTTG
GG
AAAT
GCC
TTCG
GA
Hpd
gDN
A-se
q #1
Fw
dHp
dIn
del s
eque
ncin
g on
gDN
ACG
GCA
CCCC
CATT
ATAG
TCC
Hpd
gDN
A-se
q #1
Rev
Hpd
Inde
l seq
uenc
ing
on g
DNA
GTG
ACTC
GTA
GCT
GTC
ACCG
LOC1
0075
9874
gDN
A-se
q #1
Fw
dLO
C100
7598
74In
del s
eque
ncin
g on
gDN
AAG
GCT
GTT
CGCA
GCT
TACT
ALO
C100
7598
74 g
DNA-
seq
#1 R
evLO
C100
7598
74In
del s
eque
ncin
g on
gDN
AAG
GG
TTG
CCAT
CGCA
ATG
AAPr
odh
gDN
A-se
q #1
Fw
dPr
odh
Inde
l seq
uenc
ing
on g
DNA
GTC
TCTC
AACA
GG
GCC
GC
Prod
h gD
NA-
seq
#1 R
evPr
odh
Inde
l seq
uenc
ing
on g
DNA
CGG
GCT
CCTT
TTCC
TGTG
TPr
odh2
gDN
A-se
q #1
Fw
dPr
odh2
Inde
l seq
uenc
ing
on g
DNA
GG
TGG
TGCC
TTCC
ATCT
CAA
Prod
h2 g
DNA-
seq
#1 R
evPr
odh2
Inde
l seq
uenc
ing
on g
DNA
CTCA
TGTC
CTG
TCCT
CCG
TGAa
ss c
DNA-
seq
#2 F
wd
Aass
Inde
l seq
uenc
ing
on c
DNA
TCTC
CACC
ACAA
GCC
AGTG
AAa
ss c
DNA-
seq
#2 R
evAa
ssIn
del s
eque
ncin
g on
cDN
AG
ATG
GCC
CGTC
GAT
TGG
AAG
Afm
id c
DNA-
seq
#2 F
wd
Afm
idIn
del s
eque
ncin
g on
cDN
ATG
GCA
GAG
CGG
AAG
TAAA
GA
Afm
id c
DNA-
seq
#2 R
evAf
mid
Inde
l seq
uenc
ing
on c
DNA
TTG
GAT
ACCG
CCTC
TGTA
GG
ADd
c cD
NA-
seq
#2 F
wd
Ddc
Inde
l seq
uenc
ing
on c
DNA
GG
AACC
GG
ATG
CATA
TGAA
GA
Ddc
cDN
A-se
q #2
Rev
Ddc
Inde
l seq
uenc
ing
on c
DNA
CTTC
CCCA
GCC
AATC
CATC
AG
ad1
cDN
A-se
q #2
Fw
dG
ad1
Inde
l seq
uenc
ing
on c
DNA
TCTT
CCAC
TCCT
TCG
TCTG
CAA
Gad
1 cD
NA-
seq
#2 R
evG
ad1
Inde
l seq
uenc
ing
on c
DNA
CATC
CATG
GG
CTAC
GCC
ACG
ad2
cDN
A-se
q #2
Fw
dG
ad2
Inde
l seq
uenc
ing
on c
DNA
TAG
CTCA
AAAG
TTCA
CCG
GC
Gad
2 cD
NA-
seq
#2 R
evG
ad2
Inde
l seq
uenc
ing
on c
DNA
AGTT
GAC
ATCC
GCT
TTG
GG
GHp
d cD
NA-
seq
#2 F
wd
Hpd
Inde
l seq
uenc
ing
on c
DNA
CAAC
CAAC
CCG
ACCA
GG
AAHp
d cD
NA-
seq
#2 R
evHp
dIn
del s
eque
ncin
g on
cDN
ATT
GAT
GG
ACTC
CTCA
TAG
TTG
GC
Prod
h cD
NA-
seq
#2 F
wd
Prod
hIn
del s
eque
ncin
g on
cDN
AG
CGAG
CTCA
GG
GG
CTG
Prod
h cD
NA-
seq
#2 R
evPr
odh
Inde
l seq
uenc
ing
on c
DNA
GAG
TAAC
TGTT
CGTG
GTG
CGPr
odh2
cDN
A-se
q #2
Fw
dPr
odh2
Inde
l seq
uenc
ing
on c
DNA
GAG
GCC
TGG
TATG
AGG
GG
AAC
Prod
h2 c
DNA-
seq
#2 R
evPr
odh2
Inde
l seq
uenc
ing
on c
DNA
GTG
CTG
GTT
AGTG
CTG
TCAT
CT
127
Day
Aver
age
SDAv
erag
eSD
Aver
age
SDAv
erag
eSD
SDAv
erag
eSD
01,
30,
011,
350,
030,
920,
070,
910,
070,
158,
760,
14
11,
270,
041,
250,
040,
780,
060,
770,
10,
397,
450,
05
21,
10,
071,
130,
040,
620,
060,
650,
040,
185,
230,
07
31,
050,
030,
970
0,65
0,05
0,65
0,05
0,17
3,1
0,17
40,
890,
030,
830,
020,
520,
040,
470,
020,
391,
370,
26
KOW
T
2,36
Tabl
e S6
.1: C
ompa
rison
bet
wee
n ge
ne e
dite
d cl
ones
and
wild
type
bas
ed o
n qu
antif
icat
ion
of P
heny
alan
ine
and
Tyro
sine
(for
Hpd
kno
ck-o
ut) a
nd G
luta
min
e (G
ad2
for
knoc
k-ou
t) de
term
ined
via
HPL
C a
naly
sis.
KO
Aver
age
8,62
7,21
5,24
3,52
Hpd
knoc
k-ou
t G
ad2
knoc
k-ou
t
Phen
ylal
anin
e (m
M)
Tyro
sine
(mM
)G
luta
min
e (m
M)
WT
KOW
T
128
Aver
age
SDAv
erag
eSD
Aver
age
SDAv
erag
eSD
Aver
age
SDAv
erag
eSD
-0,0
50,
01-0
,06
0-0
,04
0,02
-0,0
40,
01-1
,07
0,09
-1,0
10,
07
KO
Tabl
e S6
.2: M
ean
spec
ific
cons
umpt
ion
rate
s (p
mol
/cel
l/da
y) o
f phe
nyla
lani
ne (q
Phe)
, tyr
osin
e (q
Tyr)
and
glu
tam
ine
(qG
ln) d
eter
min
ed fr
om d
ay 0
to d
ay 3
. Hp
d kn
ock
out
Gad
2 kn
ock
out
qPhe
qT
yrqG
ln
WT
KOW
TKO
WT
129
Tabl
e S6
.2 H
PLC
am
ino
acid
qua
ntifi
catio
n fo
r wild
type
and
Hpd
kno
ck o
ut c
ells
(KO
)
No.
In
ject
ion
Nam
eU
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)U
ndilu
ted
resu
lts (µ
g/m
l)
Aspa
rtic
acid
Glu
tam
ic a
cid
Cys
tein
eAs
parg
ine
Serin
eG
luta
min
eH
istid
ine
Gly
cine
Thre
onin
eAr
gini
neAl
anin
eTy
rosi
neVa
line
Met
hion
ine
Tryp
toph
anPh
enyl
alan
ine
Isol
euci
neLe
ucin
eLy
sine
Prol
ine
1 B
lank
_M9_
IS0,
050,
02n.
a.0,
020,
080,
03n.
a.0,
110,
020,
030,
070,
020,
02n.
a.2,
33n.
a.0,
120,
041,
97n.
a.
2 B
lank
_M9_
IS0,
050,
02n.
a.0,
020,
070,
020,
020,
110,
020,
030,
070,
020,
03n.
a.1,
93n.
a.0,
080,
031,
95n.
a.
3 A
Amix
_M9_
0,1
0,14
0,11
0,07
0,11
0,17
0,11
0,12
0,27
0,12
0,14
0,13
0,12
0,12
0,1
1,8
0,09
0,09
0,13
2,81
n.a.
4 A
Amix
_M9_
0,25
0,27
0,24
0,17
0,24
0,3
0,26
0,27
0,48
0,26
0,26
0,26
0,27
0,26
0,23
1,67
0,27
0,25
0,27
2,91
n.a.
5 A
Amix
_M9_
0,5
0,48
0,47
0,34
0,48
0,53
0,49
0,49
0,87
0,51
0,51
0,56
0,51
0,51
0,48
1,96
0,49
0,61
0,51
0,9
n.a.
6 A
Amix
_M9_
10,
940,
930,
70,
971,
021
11,
661,
031,
021,
081,
031,
031,
011,
791,
011,
151,
041,
4n.
a.
7 A
Amix
_M9_
2,5
2,23
2,27
1,82
2,34
2,37
2,4
2,44
3,91
2,47
2,43
2,46
2,5
2,47
2,43
3,33
2,47
2,56
2,49
2,94
n.a.
8 A
Amix
_M9_
54,
534,
64,
084,
744,
764,
854,
887,
825,
044,
844,
885
4,93
4,9
5,72
5,01
4,95
4,99
5,53
n.a.
9 A
Amix
_M9_
109,
89,
969,
7710
,21
10,1
610
,31
10,4
216
,87
10,7
610
,34
10,4
110
,64
10,5
210
,52
11,2
610
,68
10,5
910
,63
12,5
9n.
a.
10 A
Amix
_M9_
2524
,12
24,4
925
,82
24,7
724
,78
24,8
925
,05
40,8
426
,01
24,6
424
,93
25,2
825
,18
25,1
425
,34
25,3
725
,09
25,2
26,1
2n.
a.
11 A
Amix
_M9_
5050
,25
50,5
55,5
850
,65
50,6
50,7
650
,47
83,5
753
,12
50,4
450
,76
50,9
850
,84
50,8
550
,02
51,2
250
,45
50,7
950
,86
n.a.
12 A
Amix
_M9_
7575
,09
75,0
883
,84
74,7
474
,83
74,3
372
,190
,75
78,1
674
,75
74,7
774
,92
74,5
774
,82
74,4
275
,13
73,8
874
,47
74n.
a.
13 B
lank
_M9_
IS0,
060,
03n.
a.0,
020,
070,
030,
020,
130,
030,
050,
070,
020,
03n.
a.1,
670,
020,
10,
042,
76n.
a.
14 H
pd _
rep_
1_D
ay0
19,6
932
,03
0,06
78,4
355
,52
113,
7917
,74
1,2
38,8
437
,66
1,72
15,2
337
,84
12,7
521
,75
22,3
438
,97
55,4
440
,26
n.a.
15 H
PD_r
ep_1
_Day
121
,94
32,9
30,
0572
,56
52,0
394
,06
17,2
32,
7937
,96
36,3
84,
3413
,24
36,5
812
,27
21,7
121
,33
37,6
853
,538
,94
n.a.
16 H
PD_r
ep_1
_Day
228
,335
,20,
0455
,12
43,8
660
,716
,36
5,76
35,7
532
,85
11,0
811
,74
33,6
510
,86
21,5
19,1
334
,94
49,1
735
,83
n.a.
17 H
PD_r
ep_1
_Day
335
,81
35,8
20,
0324
,11
30,2
428
,47
14,6
210
,82
31,6
127
,75
22,8
810
,829
,11
8,98
20,2
116
,11
30,7
942
,57
30,7
6n.
a.
18 H
PD_r
ep_1
_Day
437
,55
38,3
1n.
a.5,
117
,26
8,42
14,4
319
,35
30,2
325
,16
38,3
58,
8226
,79
7,95
23,5
414
,129
,06
39,2
129
,52
n.a.
19 W
T_re
p_1_
Day
019
,22
31,0
30,
0575
,23
53,3
610
8,62
16,8
41,
3337
,69
36,3
32,
0915
,21
36,4
12,3
320
,99
21,6
137
,52
53,4
138
,73
n.a.
20 W
T_re
p_1_
Day
122
,48
33,2
24,
3670
,21
51,5
893
,53
17,3
3,26
38,0
936
,46
5,16
13,9
736
,63
12,2
521
,95
21,3
737
,55
53,3
738
,94
n.a.
21 W
T_re
p_1_
Day
227
,48
33,9
22,
3548
,73
42,1
262
,56
15,8
46,
1634
,97
32,1
811
,57
9,92
32,8
10,6
20,9
18,5
334
,148
,01
34,8
4n.
a.
22 W
T_re
p_1_
Day
334
,34
35,6
10,
0322
,45
31,8
237
,88
15,1
611
,74
3329
,21
22,4
11,7
630
,41
9,48
21,3
217
32,1
744
,66
32,3
8n.
a.
23 W
T_re
p_1_
Day
433
,87
36,9
60,
187,
8420
,38
17,2
14,2
218
,81
31,0
526
,65
34,8
38,
9928
,08
8,49
21,2
914
,92
30,3
341
,36
29,8
n.a.
24 H
PD_r
ep_2
_Day
020
,14
32,4
50,
0579
,25
56,3
311
5,11
181,
1839
,55
38,1
81,
7917
,72
38,2
812
,96
22,1
622
,85
39,5
956
,23
40,9
3n.
a.
HPL
C a
naly
sis
of a
min
o ac
ids
in s
uper
nant
ants
ext
ract
ed fr
om b
atch
cul
tivat
ions
of t
he k
nock
out
CH
O c
ell l
ines
and
wild
type
stra
ins.
Pro
line
was
not
an
alys
ed in
all
sam
ples
due
to s
atur
atio
n. C
yste
ine
is a
labi
le a
min
o ac
id a
nd th
eref
ore
the
pres
ente
d co
ncen
tratio
ns a
re s
olel
y in
dica
tive.
130
25 H
PD_r
ep_2
_Day
120
,330
,10,
0565
,97
47,4
785
,57
15,9
12,
6634
,65
33,2
54,
0815
,98
33,4
811
,23
20,1
119
,99
34,3
748
,97
36,0
4n.
a.
26 H
PD_r
ep_2
_Day
228
,89
34,4
52,
1653
,11
43,1
59,5
216
,16,
1234
,84
32,3
811
,15
12,4
832
,95
10,6
620
,718
,87
34,6
948
,63
35,5
4n.
a.
27 H
PD_r
ep_2
_Day
336
,51
35,2
20,
4222
,63
29,8
228
,35
14,3
411
,18
31,0
927
,52
22,8
912
,48
28,6
68,
9119
,75
16,0
630
,71
42,4
30,7
8n.
a.
28 H
PD_r
ep_2
_Day
436
,56
36,4
80,
084,
4216
,29
8,54
13,1
819
,04
28,6
224
,37
36,9
78,
6425
,76
7,7
19,3
713
,57
28,1
37,9
627
,72
n.a.
29 W
T_re
p_2_
Day
018
,72
30,1
50,
0472
,851
,910
5,42
16,6
91,
4236
,52
35,4
2,05
17,5
935
,42
12,0
420
,34
21,2
336
,66
52,1
838
,14
n.a.
30 W
T_re
p_2_
Day
120
,87
30,8
83,
6665
,48
47,9
786
,92
16,0
93,
2135
,333
,97
4,85
15,2
833
,95
11,4
520
,34
20,2
135
,149
,91
36,6
9n.
a.
31 W
T_re
p_2_
Day
233
,92
35,3
50,
0322
,52
31,5
537
,48
15,1
211
,96
32,6
529
,14
22,4
611
,79
30,2
29,
4520
,73
16,8
432
,04
44,4
632
,26
n.a.
32 W
T_re
p_2_
Day
326
,44
32,6
52,
0647
,12
40,6
860
,35
15,1
26,
1333
,64
31,0
411
,210
,76
31,6
210
,320
,01
17,9
632
,96
46,4
233
,8n.
a.
33 W
T_re
p_2_
Day
432
,42
34,1
80,
145,
9918
,715
,613
,15
17,7
628
,63
24,7
132
,64
10,3
25,8
27,
8919
,32
14,0
228
,07
38,3
527
,86
n.a.
34 H
PD_r
ep_3
_Day
019
,29
31,2
24,
1876
,81
54,3
210
8,65
17,2
31,
2737
,96
36,7
31,
6816
,51
36,7
412
,48
21,0
821
,94
38,0
554
,49
39,1
9n.
a.
35 H
PD_r
ep_3
_Day
121
,12
31,6
73,
6370
,26
50,2
288
,65
16,6
82,
8736
,635
,06
4,21
12,8
35,1
311
,86
20,9
220
,55
36,2
451
,937
,68
n.a.
36 H
PD_r
ep_3
_Day
228
,333
,49
0,03
47,4
939
,74
51,1
815
,29
6,51
33,5
630
,712
,82
11,1
731
,49
10,1
320
,16
17,8
832
,76
46,4
333
,53
n.a.
37 H
PD_r
ep_3
_Day
334
,79
34,5
70,
0624
,24
30,1
27,3
214
,96
10,9
731
,18
27,3
722
,43
11,8
228
,67
8,88
22,5
116
,04
30,2
942
,19
31,7
5n.
a.
38 H
PD_r
ep_3
_Day
435
,01
35,4
20,
15,
4416
,95
8,19
13,5
818
,328
,28
23,9
235
,52
8,04
25,3
27,
6122
,21
13,4
427
,39
37,5
128
,05
n.a.
39 W
T_re
p_3_
Day
018
,96
30,5
80,
0474
,27
52,7
910
4,82
16,7
61,
4537
,22
35,8
92,
0917
,05
35,8
512
,220
,79
21,5
537
,08
53,1
338
,45
n.a.
40 W
T_re
p_3_
Day
122
,16
32,9
23,
6470
,18
51,4
390
,89
17,2
63,
3337
,936
,24
5,1
13,2
536
,16
12,1
921
,93
21,1
937
,23
53,3
338
,93
n.a.
41 W
T_re
p_3_
Day
228
,55
34,4
62,
149
,46
43,3
762
,43
16,1
6,45
35,6
732
,89
11,8
111
,75
33,5
610
,87
21,3
519
,18
34,8
849
,44
35,6
6n.
a.
42 W
T_re
p_3_
Day
332
,82
35,1
30,
0324
,21
32,1
536
,42
14,9
211
,53
32,8
529
,13
22,0
612
,61
30,3
39,
4720
,79
17,0
532
,07
44,7
331
,97
n.a.
43 W
T_re
p_3_
Day
435
,637
,14
0,17
7,23
21,5
516
,96
14,3
518
,95
31,2
127
,15
34,7
38,
9928
,38,
6420
,79
15,0
730
,89
42,2
630
n.a.
44 B
lank
_M9_
IS0,
050,
02n.
a.0,
020,
060,
020,
020,
10,
020,
040,
050,
010,
02n.
a.2
n.a.
0,05
0,02
2,35
n.a.
45 A
Amix
_M9_
0,1
0,14
0,12
0,07
0,11
0,17
0,12
0,13
0,27
0,12
0,14
0,13
0,13
0,12
0,09
1,63
0,12
0,16
0,12
2,8
n.a.
46 A
Amix
_M9_
0,25
0,26
0,24
0,13
0,24
0,3
0,26
0,25
0,48
0,26
0,28
0,26
0,26
0,26
0,23
1,56
0,25
0,31
0,27
2,87
n.a.
47 A
Amix
_M9_
0,5
0,49
0,46
0,26
0,47
0,54
0,51
0,5
0,86
0,51
0,52
0,56
0,51
0,5
0,48
1,52
0,48
0,55
0,52
0,73
n.a.
48 A
Amix
_M9_
10,
940,
940,
540,
971,
011,
011,
011,
671,
031,
031,
071,
031,
010,
991,
721,
021,
061,
061,
31n.
a.
49 A
Amix
_M9_
2,5
2,24
2,26
1,38
2,36
2,35
2,4
2,42
3,92
2,48
2,44
2,46
2,47
2,43
2,42
3,44
2,48
2,46
2,5
2,81
n.a.
50 A
Amix
_M9_
54,
514,
63,
084,
744,
714,
824,
857,
874,
994,
874,
874,
974,
94,
875,
84,
955,
015,
035,
21n.
a.
51 A
Amix
_M9_
109,
759,
947,
5510
,17
10,0
810
,34
10,4
616
,95
10,6
710
,38
10,3
410
,56
10,4
710
,43
11,3
310
,53
10,6
510
,61
12,2
3n.
a.
52 A
Amix
_M9_
2524
,05
24,4
320
,76
24,7
824
,624
,87
24,9
441
,28
25,7
24,9
224
,84
25,1
625
,01
24,9
524
,97
2525
,27
25,2
525
,98
n.a.
53 A
Amix
_M9_
5050
,33
50,4
44,9
950
,56
50,6
750
,66
52,0
384
,95
52,9
150
,45
50,4
750
,38
50,5
550
,51
51,3
150
,24
51,0
550
,65
50,3
2n.
a.
54 A
Amix
_M9_
7575
,27
74,7
667
,54
74,6
74,5
574
,74
76,1
50,
5367
,05
74,7
374
,41
73,8
874
,32
74,1
474
,03
73,6
174
,83
74,2
673
,76
n.a.
55 B
lank
_M9_
IS0,
060,
03n.
a.0,
020,
080,
020,
020,
110,
030,
040,
060,
020,
03n.
a.1,
320,
030,
040,
032,
49n.
a.
56 B
lank
_M9_
ISn.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.
131
Tabl
e S6
.3 A
min
o ac
id q
uant
ifica
tion
of w
ild ty
pe a
nd G
ad2
knoc
k ou
t (K
O)
No.
In
ject
ion
Nam
eU
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)U
ndilu
ted
resu
lts(µ
g/m
l)
Aspa
rtic
acid
Glu
tam
ic a
cid
Cys
tein
eAs
parg
ine
Serin
eG
luta
min
eH
istid
ine
Gly
cine
Thre
onin
eAr
gini
neAl
anin
eTy
rosi
neVa
line
Met
hion
ine
Tryp
toph
anPh
enyl
alan
ine
Isol
euci
neLe
ucin
eLy
sine
Prol
ine
1Bl
ank_
M9_
IS0,
050,
02n.
a.0,
020,
090,
020,
030,
10,
020,
040,
080,
020,
02n.
a.1,
14n.
a.0,
310,
041,
69n.
a.
2Bl
ank_
M9_
IS0,
050,
02n.
a.0,
030,
080,
01n.
a.0,
10,
030,
040,
080,
020,
03n.
a.1,
07n.
a.0,
260,
042,
26n.
a.
3AA
mix
_M9_
0,1
0,15
0,13
0,06
0,12
0,17
0,12
0,12
0,24
0,13
0,14
0,12
0,12
0,12
0,09
1,22
0,08
0,29
0,14
3,15
n.a.
4AA
mix
_M9_
0,25
0,28
0,25
0,13
0,25
0,3
0,24
0,26
0,43
0,26
0,26
0,24
0,25
0,25
0,23
1,23
0,2
0,44
0,26
2,7
n.a.
5AA
mix
_M9_
0,5
0,49
0,47
0,3
0,5
0,53
0,49
0,53
0,79
0,51
0,53
0,57
0,52
0,51
0,48
1,21
0,53
0,48
0,54
0,77
n.a.
6AA
mix
_M9_
10,
970,
960,
641,
021,
051,
031,
081,
521,
071,
061,
121,
071,
031,
011,
530,
991,
331,
061,
211,
13
7AA
mix
_M9_
2,5
2,23
2,24
1,71
2,36
2,36
2,39
2,51
3,46
2,48
2,41
2,47
2,46
2,44
2,43
2,99
2,38
2,71
2,48
2,84
1,99
8AA
mix
_M9_
54,
714,
814,
14,
984,
955,
045,
217,
245,
255,
045,
115,
185,
125,
135,
95,
095,
415,
236,
934,
75
9AA
mix
_M9_
109,
799,
989,
5110
,27
10,2
710
,34
10,7
714
,98
10,8
610
,35
10,4
210
,58
10,5
310
,511
,33
10,5
110
,87
10,6
210
,79,
79
10AA
mix
_M9_
2524
,55
24,9
225
,57
25,1
625
,325
,23
25,8
736
,81
26,4
125
,125
,35
25,5
825
,53
25,5
925
,67
25,5
725
,71
25,5
726
,63
24,6
1
11AA
mix
_M9_
5048
,74
4951
,33
49,1
249
,61
49,1
150
,22
7251
,58
48,8
549
,19
49,2
949
,32
49,3
248
,82
49,3
649
,27
49,1
749
,46
50,2
9
12AA
mix
_M9_
7576
,11
75,7
980
,72
75,4
775
,68
75,4
376
,47
1,37
68,1
775
,61
75,4
175
,59
75,5
475
,57
75,6
675
,47
7575
,275
,38
n.a.
13Bl
ank_
M9_
IS0,
060,
02n.
a.0,
030,
070,
020,
030,
10,
020,
040,
060,
020,
03n.
a.1,
81n.
a.0,
280,
022,
12n.
a.
14W
T_re
p_1_
Day
019
,05
30,7
64,
974
,45
54,0
512
8,48
17,3
11,
1637
,66
36,3
92,
0816
,01
36,0
112
,23
20,7
321
,25
37,7
652
,93
38,9
752
,37
15W
T_re
p_1_
Day
121
,29
32,8
74,
2770
,92
51,8
311
0,55
18,0
52,
6837
,49
35,9
35,
1413
,28
35,6
111
,99
22,3
520
,49
37,3
452
,03
39,7
928
,37
16W
T_re
p_1_
Day
226
,33
32,8
72,
1550
,03
42,2
275
,09
15,8
74,
5533
,19
30,8
910
,111
,01
31,1
810
,121
,26
17,5
733
,08
45,8
34,7
327
,8
17W
T_re
p_1_
Day
329
,83
33,4
80,
0230
,05
33,0
449
,77
13,9
26,
9630
,34
27,2
17,2
811
,97
27,9
68,
719
,415
,64
30,1
941
,28
30,1
146
,01
18W
T_re
p_1_
Day
443
,04
45,5
0,18
11,7
328
,19
33,3
217
,46
15,4
636
,87
31,4
138
,13
12,3
132
,83
9,82
26,6
317
,43
36,2
548
,51
35,5
133
,84
19G
AD2_
rep_
1_D
ay0
17,8
429
,52
0,02
73,9
52,6
126,
816
,65
1,02
36,6
135
,36
1,19
14,8
935
,15
11,8
920
,02
20,5
936
,951
,67
37,8
49,9
6
20G
AD2_
rep_
1_D
ay1
19,3
631
,23
3,89
71,1
349
,87
108,
0616
,64
2,77
36,3
34,7
93,
2911
,55
34,4
511
,59
20,4
219
,86
36,1
550
,66
37,5
2n.
a.
21G
AD2_
rep_
1_D
ay2
26,6
534
,43
1,75
56,0
843
,36
77,0
816
,77
5,8
35,1
832
,58
8,38
11,6
932
,97
10,6
522
,13
18,6
634
,84
48,3
436
,56
28,5
22G
AD2_
rep_
1_D
ay3
33,7
337
,33
0,02
34,6
333
,84
47,9
215
,31
9,79
33,3
629
,51
17,3
12,2
830
,59,
3720
,94
16,8
33,3
45,2
332
,63
49,6
7
23G
AD2_
rep_
1_D
ay4
49,2
52,9
3n.
a.7,
3924
,99
22,2
119
,51
22,2
340
,89
33,5
944
,81
11,1
936
,06
10,3
131
,57
18,2
40,0
152
,86
38,6
342
,9
24W
T_re
p_2_
Day
018
,43
29,9
14,
4672
,252
,46
125,
2116
,75
1,19
36,6
535
,42
2,07
15,0
735
,02
11,9
120
,15
20,5
936
,85
51,7
138
,04
51,6
5
25W
T_re
p_2_
Day
118
,82
29,2
83,
663
,646
,399
,27
15,5
32,
4333
,65
32,2
34,
6211
,731
,88
10,7
919
,23
18,4
733
,42
46,7
535
,05
47,4
3
26W
T_re
p_2_
Day
228
,03
34,6
52,
1852
,35
44,5
779
,53
15,9
84,
6235
,05
32,7
210
,57
11,2
232
,82
10,6
320
,94
18,6
534
,77
48,3
235
,27
50,7
1
27W
T_re
p_2_
Day
332
,44
35,9
70,
0232
,435
,81
54,2
515
,16
7,33
32,8
629
,31
18,3
912
,43
30,1
99,
3520
,616
,85
32,5
244
,45
32,2
848
,54
28W
T_re
p_2_
Day
451
,88
54,2
20,
213
,88
34,1
240
,67
19,8
317
,83
44,0
337
,68
44,9
415
,44
39,1
911
,75
28,5
620
,97
43,0
558
,02
40,8
6n.
a.
29G
AD2_
rep_
2_D
ay0
18,2
130
,21
4,53
75,4
53,8
313
0,36
17,0
31,
0537
,34
36,1
61,
0614
,04
35,6
712
,15
20,3
20,9
37,6
452
,73
38,5
7n.
a.
HPL
C a
naly
sis
of a
min
o ac
ids
in s
uper
nant
ants
ext
ract
ed fr
om b
atch
cul
tivat
ions
of t
he k
nock
out
CH
O c
ell l
ines
and
wild
type
stra
ins.
Pro
line
was
not
an
alys
ed in
all
sam
ples
due
to s
atur
atio
n. C
yste
ine
is a
labi
le a
min
o ac
id a
nd th
eref
ore
the
pres
ente
d co
ncen
tratio
ns a
re s
olel
y in
dica
tive.
132
30G
AD2_
rep_
2_D
ay1
20,2
131
,99
0,07
71,5
250
,79
109,
5217
,85
3,02
36,8
335
,33,
4410
,68
35,0
411
,81
21,9
20,0
136
,73
51,4
139
,39
31,5
5
31G
AD2_
rep_
2_D
ay2
26,2
633
,83
1,53
54,1
42,0
475
,26
15,4
65,
7334
,11
31,8
28,
2610
,232
,04
10,3
920
,23
18,0
234
,02
47,2
634
,43
49,9
8
32G
AD2_
rep_
2_D
ay3
31,2
733
,88
0,02
30,3
130
,37
43,1
113
,93
8,83
30,1
826
,51
15,4
79,
327
,81
8,49
19,7
15,0
630
,06
40,8
929
,37
47,1
8
33G
AD2_
rep_
2_D
ay4
35,4
537
,04
n.a.
4,13
17,1
215
,56
12,8
615
,35
28,8
923
,48
30,6
98,
3525
,41
7,23
19,7
312
,89
28,3
837
,57
24,6
445
,74
34W
T_re
p_3_
Day
018
,21
29,4
94,
4772
,07
51,8
912
4,2
16,4
31,
0736
,32
35,0
22,
0314
,06
34,7
11,7
619
,95
20,3
636
,25
50,8
837
,351
,43
35W
T_re
p_3_
Day
120
,12
31,2
93,
7567
,65
49,4
210
6,11
16,2
72,
6135
,75
34,4
24,
9711
,134
,03
11,4
820
,09
19,5
35,7
949
,94
36,9
849
,6
36W
T_re
p_3_
Day
225
,13
32,2
2,01
50,2
941
,69
74,9
414
,74
4,17
32,8
730
,53
9,81
10,4
730
,69
1019
,34
17,3
932
,67
45,3
733
49,4
5
37W
T_re
p_3_
Day
328
,43
32,8
0,02
31,7
533
,22
50,4
313
,84
6,79
3026
,97
16,5
211
,51
27,7
18,
6318
,79
15,4
729
,99
41,0
829
,87
45,8
4
38W
T_re
p_3_
Day
435
36,7
0,15
11,2
524
,429
,39
13,7
612
,08
30,5
126
,05
30,3
311
27,2
48,
1920
,414
,65
30,1
740
,42
28,7
746
,91
39G
AD2_
rep_
3_D
ay0
17,7
329
,40,
0273
,56
52,3
812
6,92
16,5
21,
0436
,43
35,2
21,
0415
,75
34,9
211
,86
2020
,736
,55
51,3
837
,550
,62
40G
AD2_
rep_
3_D
ay1
19,8
631
,83,
5371
,38
50,1
910
8,94
16,5
42,
8436
,67
35,1
93,
4912
,97
34,9
411
,73
20,6
420
,27
36,3
851
,09
37,5
549
,15
41G
AD2_
rep_
3_D
ay2
26,2
434
,57
1,44
56,0
442
,82
7715
,59
5,82
34,9
432
,52
8,5
11,1
732
,94
10,5
720
,91
18,5
834
,76
48,3
135
,14
48,1
4
42G
AD2_
rep_
3_D
ay3
30,4
834
,76
0,03
33,5
131
,36
44,7
414
,22
8,87
31,4
827
,34
15,8
12,0
128
,58
8,77
20,0
615
,81
30,9
142
,22
30,2
448
,56
43G
AD2_
rep_
3_D
ay4
52,5
355
,43
n.a.
6,87
25,3
22,1
218
,93
22,3
42,6
734
,93
45,2
412
,46
37,4
210
,62
29,1
819
,02
41,5
554
,85
38,0
5n.
a.
44Bl
ank_
M9_
IS0,
050,
02n.
a.0,
020,
090,
030,
040,
10,
030,
050,
080,
020,
03n.
a.1,
36n.
a.0,
20,
052,
54n.
a.
45AA
mix
_M9_
0,1
0,14
0,12
0,05
0,12
0,18
0,13
0,12
0,24
0,13
0,15
0,13
0,12
0,13
0,1
1,34
0,1
0,23
0,14
2,57
n.a.
46AA
mix
_M9_
0,25
0,26
0,24
0,12
0,25
0,29
0,26
0,25
0,41
0,26
0,28
0,25
0,25
0,26
0,23
2,11
0,26
0,24
0,29
0,34
n.a.
47AA
mix
_M9_
0,5
0,5
0,48
0,26
0,49
0,54
0,51
0,52
0,78
0,54
0,54
0,5
0,53
0,52
0,49
1,33
0,47
0,71
0,55
0,61
n.a.
48AA
mix
_M9_
10,
970,
970,
541
1,06
1,04
1,05
1,51
1,07
1,06
1,11
1,06
1,04
1,02
2,11
1,03
1,24
1,08
1,15
n.a.
49AA
mix
_M9_
2,5
2,24
2,23
1,4
2,35
2,4
2,39
2,4
3,43
2,53
2,47
2,47
2,47
2,44
2,4
3,43
2,44
2,65
2,51
2,65
n.a.
50AA
mix
_M9_
54,
74,
783,
454,
964,
995,
035,
077,
185,
285,
15,
15,
175,
125,
085,
945,
145,
275,
225,
2n.
a.
51AA
mix
_M9_
109,
779,
927,
9610
,19
10,1
310
,31
10,3
914
,85
10,7
610
,34
10,3
710
,49
10,4
810
,43
11,1
210
,56
10,7
110
,59
12,1
n.a.
52AA
mix
_M9_
2524
,55
24,8
222
,17
25,1
325
,17
25,2
225
,12
36,4
226
,55
25,1
25,2
825
,43
25,3
125
,26
25,3
25,5
325
,51
25,5
524
,57
n.a.
53AA
mix
_M9_
5048
,78
48,8
745
,84
49,0
848
,36
49,0
948
,39
71,7
251
,149
,07
49,1
449
,13
49,1
249
,01
47,8
249
,07
49,3
549
,33
49,1
8n.
a.
54AA
mix
_M9_
7575
,96
75,7
872
,49
75,5
975
,46
75,5
373
,96
110,
0578
,79
75,6
175
,38
74,9
675
,08
75,1
375
,71
75,0
575
,24
75,2
474
,57
n.a.
55Bl
ank_
M9_
IS0,
050,
03n.
a.0,
020,
090,
030,
040,
110,
030,
050,
080,
020,
04n.
a.1,
36n.
a.0,
150,
032,
53n.
a.
56Bl
ank_
M9_
ISn.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.n.
a.
133
Table S5. List of multiplex gene disrupted clones with target genes and indel sizes.Clone Target gene Indel sizesClone 1 Aass 0,1
Afmid -9 / wildtypeDdc +98 / +105 (in frame 5%)Hpd 1
Clone 2 Aass -1 / -10Afmid wildtypeDdc +98 / +105 (in frame 5%)Hpd 1
134
Appendix 2: Paper III – Supplementary materials Physiological study of CRISPR/Cas9-mediated disruption of branched-chain amino acid transaminases in CHO cells
Sara Pereira1, Daniel Ley1,2, Mikkel Schubert1, Lise Marie Grav1, Helene Faustrup Kildegaard1,3, Mikael Rørdam
Andersen4
1The Novo Nordisk Foundation, Center for Biosustainability, Technical University of Denmark, Kongens Lyngby,
Denmark, 2Current address: AGC Biologics A/S, Vandtårnsvej 83, 2860 Søborg, Denmark, 3Current address: Novo
Nordisk, Department of mammalian expression, Måløv, Denmark, 4Department of Biotechnology and
Biomedicine, Technical University of Denmark, Kongens Lyngby, Denmark
Correspondence: [email protected] for enquiries on the computational analysis and strategy,
[email protected] for correspondence on the molecular biology.
135
Supplementary figures
Figure S1 - Specific rates of consumption of glucose, glutamine and glutamate and of production lactate and
ammonium of CHO-S cells and Bcat1 and Bcat2-disrupted clonal cells. Consumption rates are presented as
absolute values.
136
Figure S2 – Viability of T2_6 Bcat1 and Bcat2-disrupted cells. (A) T2_6+Cas9 cells were used as control and
T2_6 cells with (B) Bcat1 and (C) Bcat2 gene disruption. Viable cell densities (VCD) determined during
simultaneous batch cultivation (Triplicate cultures). Error bars were omitted.
137
Figure S3 – Extracelular glutamine, glucose and glutamate profiles for batch cultivation of Bcat1-, Bcat2- and
Bcat1&2- disrupted T2_6 cells producing mCherry compared to control T2_6+ Cas9 cells.
138
Figure S4 – Specific consumption rates glucose and glutamine and specific production rates of lactate and
ammonium calculated from day 0 to day 3 of batch cultivation for Bcat1-, Bcat2- and Bcat1&2- disrupted T2_6
cells producing mCherry compared to control T2_6+Cas9 cells also producing mCherry. Consumption rates
are presented as absolute values.
139
Figure S5 – Specific consumption rates glucose and glutamine and specific production rates of lactate and
ammonium calculated from day 0 to day 3 of batch cultivation for Bcat1-, Bcat2- and Bcat1&2- disrupted T2_6
cells producing mCherry compared to control T2_6+Cas9 cells also producing mCherry. Consumption rates
are presented as absolute values.
140
Figure S6 – Specific consumption rates of BCAAs leucine, isoleucine and valine calculated from day 0 to day
3 of batch cultivation for Bcat1-, Bcat2- and Bcat1&2- disrupted T2_6 cells producing mCherry compared to
control T2_6+Cas9 cells also producing mCherry. Consumption rates are presented as absolute values.
141
Figure S7 – Specific consumption rates of BCAAs leucine, isoleucine and valine calculated from day 0 to day
3 of batch cultivation for Bcat1-, Bcat2- and Bcat1&2- disrupted T2_6 clones producing mCherry compared
to control T2_6+Cas9 clones also producing mCherry. Consumption rates presented as absolute values.
142
Figure S8 – Specific growth rates calculated from day 0 to day 3 of batch cultivation for Bcat1-, Bcat2- and
Bcat1&2- disrupted T2_6 cells producing mCherry compared to control T2_6+Cas9 cells also producing
mCherry, presented by clone and by group of clones according to gene target.
143
Figure S9 – Sum of integral viable cell density (IVCD) of Bcat1-, Bcat2- and Bcat1&2- disrupted T2_6 cells
producing mCherry compared to control T2_6+Cas9 cells also producing mCherry in batch cultivation,
presented by clone and by group of clones according to gene target.
144
Supplementary tables
Table S1- List of gRNAs used in for disrupting Bcat1 and Bcat2 in CHO cells.
# Name Sequence (5’->3’)
1 grna_design_Bcat1_NW_003613704.1-688351 GCTCTGACATATTTCGGAT
2 grna_design_Bcat2_NW_003614570.1+168308 CAGCACAGGCCGCACGTAG
Table S2 – List of primers used in PCR for screening. Primers used in deep sequencing have overhangs (in italic).
# Primer name Sequence Application
1 MiSeq_Bcat1_NW_003613704.1-
688351_fwd TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGAA
AGCCCTGCTCTTTGTGGT
1st PCR for deep sequencing library
preparation
2 MiSeq_Bcat1_NW_003613704.1-
688351_rev GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGC
ACCAGACTGACTTACCCGC
1st PCR for deep sequencing library
preparation
3 MiSeq_Bcat2_NW_003614570.1+1
68308_fwd TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGA
CAGCATCAGACCAAGGGG
1st PCR for deep sequencing library
preparation
4 MiSeq_Bcat2_NW_003614570.1+168
308_rev GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGAGA
GGGTGGCTTAGGGGATC 1st PCR for deep sequencing
library preparation
145
Ta
ble
S3 –
List
of o
ff-ta
rget
s for
sing
le gu
ide
RNA
use
d to
disr
upt B
cat1
.
Offt
arge
t H
ead
Hea
dErr
ors
KM
er
KM
erEr
ror
s Su
mEr
ror
s Re
fSeq
Po
siti
on
Stra
nd
Gen
e G
eneI
D
Feat
ures
Pr
oduc
ts
GG
CTCT
GA
CATA
TTTC
GG
ATT
GG
G
GCT
CTG
0
ACA
TATT
TCG
GA
T 0
0 N
W_0
0361
370
4.1
6883
58
- Bc
at1
1007
6365
8 ex
on;C
DS
bran
ched
chai
n am
ino
acid
tr
ansa
min
ase 1
;bra
nche
d-ch
ain-
amin
o-ac
id
amin
otra
nsfe
rase
cyto
solic
CC
CTCT
GTC
ATA
TGTG
GG
ATT
GG
Cc
CTC
TG
1 tC
ATA
TgTg
GG
AT
3 4
NW
_003
613
630.
1 84
465
- LO
C103
1596
44
1031
5964
4 nc
RNA
unch
arac
teriz
ed
LOC1
0315
9644
TG
CTTT
GA
CATA
TATC
AC
ATA
GG
TG
CTtT
G
1 A
CATA
TaTC
acA
T 3
4 N
W_0
0361
497
9.1
1577
89
+ Ba
nk1
1007
6246
3 tr
ansc
ript
B-ce
ll sc
affo
ld p
rote
in w
ith
anky
rin re
peat
s 1
CAG
CCG
GA
CATA
TTTC
TG
ATG
GG
Ca
gcCg
G
4 A
CATA
TTTC
tG
AT
1 5
NW
_003
613
657.
1 55
572
3 -
LOC1
0797
8317
10
7978
317
ncRN
A un
char
acte
rized
LO
C107
9783
17
AG
CTG
CCA
CACT
TTTC
GG
ATT
GG
A
GCT
gcc
3
ACA
ctTT
TCG
GA
T 2
5 N
W_0
0361
387
8.1
1599
323
+
LOC1
0316
3634
10
3163
634
ncRN
A un
char
acte
rized
LO
C103
1636
34
ATG
ACT
GA
CCTA
TTTC
TG
ATG
GG
A
tgaC
TG
3
ACc
TATT
TCt
GA
T 2
5 N
W_0
0361
362
8.1
2805
532
+
LOC1
0316
0180
10
3160
180
ncRN
A un
char
acte
rized
LO
C103
1601
80
CTTT
ATG
ACT
TATT
TCG
TA
TTG
G
CttT
aTG
3
ACt
TATT
TCG
tAT
2 5
NW
_003
613
919.
1 99
806
6 +
Ypel
1 10
0759
808
exon
yi
ppee
like
1
TTTT
CTA
ACT
TATT
TCA
GA
TAG
G
TttT
CTa
3 A
CtTA
TTTC
aG
AT
2 5
NW
_003
613
727.
1 76
241
0 -
LOC1
0797
7783
10
7977
783
ncRN
A un
char
acte
rized
LO
C107
9777
83
AA
GTC
TGTC
ATA
TTTG
GG
GTA
GG
A
agTC
TG
2
tCA
TATT
TgG
GgT
3
5 N
W_0
0361
486
4.1
1094
49
- LO
C107
9783
88
1079
7838
8 nc
RNA
unch
arac
teriz
ed
LOC1
0797
8388
A
GA
TATG
ACA
TGTT
TTG
GG
TTG
G
AG
aTaT
G
2 A
CATg
TTTt
GG
gT
3 5
NW
_003
613
803.
1 95
508
0 +
LOC1
0797
9455
10
7979
455
ncRN
A un
char
acte
rized
LO
C107
9794
55
AG
ATC
AG
ACA
GA
CTTC
GG
CTA
GG
A
GaT
CaG
2
ACA
gAcT
TCG
GcT
3
5 N
W_0
0361
399
7.1
2004
68
- Sl
c25a
48
1007
5644
1 ex
on;C
DS
solu
te ca
rrie
r fam
ily 2
5 m
embe
r 48
AG
CTCA
TCCA
CATC
TCG
GA
TTG
G
AG
CTC
at
2 cC
AcA
TcTC
GG
AT
3 5
NW
_003
613
700.
1 28
022
9 +
Ars
b 10
0759
794
exon
;CD
S ar
ylsu
lfata
se B
AG
TTG
TGTC
TTA
TTTT
GG
ATT
GG
A
GtT
gTG
2
tCtT
ATT
TtG
GA
T 3
5 N
W_0
0361
420
8.1
5530
20
- Cy
sltr2
10
0752
628
tran
scrip
t cy
stei
nyl l
euko
trie
ne re
cept
or
2 A
GTT
TTG
GCA
TGTT
TGG
GA
TAG
G
AG
tTtT
G
2 gC
ATg
TTTg
GG
AT
3 5
NW
_003
613
904.
1 68
471
4 -
R3hd
m4
1007
6387
4 ex
on
R3H
dom
ain
cont
aini
ng 4
ATC
ACT
GA
AA
TCTT
TGG
GA
TAG
G
AtC
aCT
G
2 A
aATc
TTTg
GG
AT
3 5
NW
_003
617
236.
1 21
97
+ LO
C100
7725
12
1007
7251
2 ex
on
ATC
ACT
GA
CACA
TTTT
GT
ATG
GG
A
tCaC
TG
2
ACA
cATT
TtG
tA
T 3
5 N
W_0
0361
364
2.1
2707
647
-
LOC1
0315
8872
10
3158
872
ncRN
A un
char
acte
rized
LO
C103
1588
72
ATC
TCTC
ACA
AA
TTTG
GG
GTG
GG
A
tCTC
Tc
2 A
CAaA
TTTg
GG
gT
3 5
NW
_003
613
648.
1 50
274
1 -
LOC1
0315
9647
10
3159
647
ncRN
A un
char
acte
rized
LO
C103
1596
47
ATC
TGTG
TGA
TATT
TCA
GA
TTG
G
AtC
TgT
G
2 tg
ATA
TTTC
aGA
T 3
5 N
W_0
0361
387
6.1
1312
51
- Ld
ah
1007
7418
1 tr
ansc
ript
lipid
dro
plet
ass
ocia
ted
hydr
olas
e CG
GTC
AG
ACA
TACT
TCG
TACT
GG
CG
gTC
aG
2 A
CATA
cTTC
GtA
c 3
5 N
W_0
0361
536
5.1
8855
7 -
LOC1
0077
2299
10
0772
299
exon
;CD
S pr
otoc
adhe
rin-3
-like
CTCC
CTG
TCA
TATG
TCG
GA
GA
GG
Ct
CcCT
G
2 tC
ATA
TgTC
GG
Ag
3 5
NW
_003
613
935.
1 10
536
18
- LO
C107
9789
55
1079
7895
5 nc
RNA
unch
arac
teriz
ed
LOC1
0797
8955
146
GCA
TCTG
CCA
GA
TTTC
TG
ATG
GG
G
caTC
TG
2
cCA
gATT
TCtG
AT
3 5
NW
_003
614
654.
1 81
004
+ LO
C103
1626
49
1031
6264
9 nc
RNA
;exo
n un
char
acte
rized
LO
C103
1626
49
GG
AG
CTG
ACA
TATT
GTG
GA
AG
GG
G
Gag
CTG
2
ACA
TATT
gtG
GA
a 3
5 N
W_0
0361
401
3.1
1187
700
-
Inm
t 10
0773
805
exon
in
dolet
hyla
min
e N-
met
hyltr
ansfe
rase
G
GG
TATG
ATA
TATT
TGA
GA
TAG
G
GG
gTa
TG
2 A
tATA
TTTg
aGA
T 3
5 N
W_0
0361
549
2.1
2177
2 -
Cspp
1 10
3164
367
tran
scrip
t ce
ntro
som
e and
spin
dle p
ole
asso
ciate
d pr
otei
n 1
GG
TTCT
CACA
TATT
GG
GG
TTG
GG
G
GtT
CTc
2
ACA
TATT
ggG
GtT
3
5 N
W_0
0361
431
4.1
9849
6 -
Mag
ee1
1007
6525
9 ex
on
MA
GE
fam
ily m
embe
r E1
GG
TTTT
GA
CATG
TTTT
AG
ATG
GG
G
GtT
tTG
2
ACA
TgTT
TtaG
AT
3 5
NW
_003
615
474.
1 64
974
+ LO
C103
1643
62
1031
6436
2 nc
RNA
unch
arac
teriz
ed
LOC1
0316
4362
G
TGTC
TGA
CCTA
TTTG
TG
ATG
GG
G
tgTC
TG
2
ACc
TATT
TgtG
AT
3 5
NW
_003
614
095.
1 40
956
7 +
Abh
d18
1007
7226
9 ex
on;C
DS
abhy
drol
ase d
omai
n co
ntai
ning
18;
prot
ein
ABH
D18
TA
CCCT
GCC
ATA
TACC
GG
ATT
GG
Ta
CcCT
G
2 cC
ATA
TacC
GG
AT
3 5
NW
_003
614
654.
1 36
055
7 -
Mcm
3 10
0774
903
exon
;CD
S D
NA
repl
icatio
n lic
ensin
g fa
ctor
M
CM3;
min
ichro
mos
ome
mai
nten
ance
com
plex
co
mpo
nent
3
TGA
GCT
GA
TTTA
TTTC
GG
ACA
GG
TG
agCT
G
2 A
ttTA
TTTC
GG
Ac
3 5
NW
_003
617
737.
1 34
070
+ LO
C107
9775
05
1079
7750
5 nc
RNA
unch
arac
teriz
ed
LOC1
0797
7505
TG
CTA
GG
ACA
GA
TCCC
GG
ATC
GG
TG
CTag
G
2 A
CAgA
TccC
GG
AT
3 5
NW
_003
614
605.
1 54
542
1 +
LOC1
0315
9877
10
3159
877
exon
;CD
S tr
ansla
tion
initi
atio
n fa
ctor
IF-
2-lik
e TG
CTA
GG
ACA
GA
TCCC
GG
ATC
GG
TG
CTag
G
2 A
CAgA
TccC
GG
AT
3 5
NW
_003
614
605.
1 54
542
1 +
Myo
6 10
0765
167
exon
m
yosin
VI
TGCT
ATA
CCA
TATT
GCT
GA
TGG
G
TGCT
aTa
2
cCA
TATT
gCtG
AT
3 5
NW
_003
613
794.
1 20
003
89
+ LO
C107
9800
89
1079
8008
9 nc
RNA
unch
arac
teriz
ed
LOC1
0798
0089
TG
GG
CTG
GCA
TATT
TCT
TATA
GG
TG
ggCT
G
2 gC
ATA
TTTC
ttA
T 3
5 N
W_0
0361
520
3.1
6169
1 +
LOC1
0075
3753
10
0753
753
tran
scrip
t lip
oma
HM
GIC
fusio
n pa
rtner
TGTC
CTG
ACA
TTTT
TAG
GA
CTG
G
TGtc
CTG
2
ACA
TtTT
TaG
GA
c 3
5 N
W_0
0361
358
0.1
3915
279
-
LOC1
0797
8640
10
7978
640
ncRN
A un
char
acte
rized
LO
C107
9786
40
TTCT
CAG
ATA
TATT
TCTG
AG
AG
G
TtCT
CaG
2
AtA
TATT
TCt
GA
g 3
5 N
W_0
0361
511
9.1
2929
17
+ LO
C103
1591
58
1031
5915
8 nc
RNA
unch
arac
teriz
ed
LOC1
0315
9158
TT
CTCT
AA
TATA
TTTC
CGCT
TGG
Tt
CTCT
a 2
AtA
TATT
TCc
GcT
3
5 N
W_0
0361
461
6.1
2714
24
+ Zn
f644
10
0762
160
tran
scrip
t zi
nc fi
nger
pro
tein
644
147
Tabl
e S4
– L
ist o
f off-
targ
ets f
or si
ngle
guid
e RN
A u
sed
to d
isrup
t Bca
t2.
Offt
arge
t H
ead
Hea
dErr
ors
KM
er
KM
erEr
ror
s Su
mEr
ror
s Re
fSeq
Po
siti
on
Stra
nd
Gen
e G
eneI
D
Feat
ures
Pr
oduc
ts
TCA
GCA
CAG
GCC
GCA
CGT
AG
AG
G
TCA
GC
AC
0 A
GG
CCG
CACG
TAG
0
0 N
W_0
0361
4570
.1
1683
05
+ Bc
at2
1007
523
44
exon
;CD
S br
anch
ed ch
ain
amin
o ac
id tr
ansa
min
ase
2;br
anch
ed-c
hain
-am
ino-
acid
am
inot
rans
fera
se
mito
chon
dria
l A
GA
GTA
TAG
GCC
GA
ACG
TA
GG
GG
A
gAG
tAt
3 A
GG
CCG
aACG
TAG
1
4 N
W_0
0363
6756
.1
1090
+
LOC1
0075
379
3 10
0753
793
ex
on;C
DS
olfa
ctor
y re
cept
or 8
S1
ACA
CCA
GA
GG
ACG
CACT
TA
GA
GG
A
CAcC
Ag
2 A
GG
aCG
CACt
TA
G
2 4
NW
_003
6136
23.1
31
844
74
- Zn
f133
10
0767
427
ex
on;C
DS
zinc
fing
er p
rote
in 1
33
CCA
GG
CCA
GG
CCG
CAG
GG
AG
AG
G
CCA
Ggc
C 2
AG
GCC
GCA
gGgA
G
2 4
NW
_003
6136
10.1
34
395
07
- Lt
br
1007
514
11
exon
;CD
S ly
mph
otox
in b
eta
rece
ptor
;tum
or n
ecro
sis
fact
or re
cept
or
supe
rfam
ily m
embe
r 3
TCTT
CACA
GG
GCG
CAG
GT
AG
GG
G
TCttC
AC
2 A
GG
gCG
CAgG
TAG
2
4 N
W_0
0361
4206
.1
1672
25
- Ta
f8
1007
663
20
exon
;CD
S TA
TA-b
ox b
indi
ng
prot
ein
asso
ciat
ed fa
ctor
8;
trans
crip
tion
initi
atio
n fa
ctor
TFI
ID su
buni
t 8
ACA
CCA
CAG
GCA
GCA
GG
TA
ATG
G
ACA
cCA
C 1
AG
GCa
GCA
gGTA
a 3
4 N
W_0
0361
5850
.1
1517
93
- Er
rfi1
1007
589
04
exon
;CD
S ER
BB re
cept
or fe
edba
ck
inhi
bito
r 1
TGA
GCA
CAG
GCT
GCA
GTT
AG
AG
G
TgA
GC
AC
1 A
GG
CtG
CAgt
TA
G
3 4
NW
_003
6138
67.1
59
651
5 -
Kds
r 10
0774
578
ex
on
3- keto
dihy
dros
phin
gosin
e re
duct
ase
ACA
GG
GA
AG
GCC
TCA
GG
TA
GTG
G
ACA
Gg
ga
3 A
GG
CCtC
AgG
TA
G
2 5
NW
_003
6168
92.1
30
486
+ Xr
cc6
1007
702
14
exon
;CD
S X-
ray
repa
ir co
mpl
emen
ting
defe
ctiv
e rep
air i
n Ch
ines
e ham
ster c
ells
6;X-
ray r
epai
r cro
ss-
com
plem
entin
g pr
otei
n 6
CATG
TACA
GG
CTG
CACT
TA
GG
GG
Ca
tGtA
C 3
AG
GCt
GCA
CtT
AG
2
5 N
W_0
0361
3732
.1
8633
26
+ LO
C107
978
045
1079
780
45
ncRN
A un
char
acte
rized
LO
C107
9780
45
CCA
GA
GG
AG
GCC
GG
GCG
TA
GA
GG
CC
AG
agg
3 A
GG
CCG
ggCG
TAG
2
5 N
W_0
0361
3803
.1
1120
563
-
Zbtb
47
1007
672
62
exon
;CD
S zi
nc fi
nger
and
BTB
dom
ain
cont
aini
ng
47;zi
nc fi
nger
and
BTB
dom
ain-
cont
aini
ng
prot
ein
47
CCCG
TCCA
GG
CTG
CAG
GT
AG
GG
G
CCcG
tcC
3 A
GG
CtG
CAgG
TA
G
2 5
NW
_003
6145
52.1
28
190
2 +
Fhod
1 10
0760
903
ex
on;C
DS
FH1/
FH2
dom
ain-
cont
aini
ng p
rote
in
1;fo
rmin
hom
olog
y 2
dom
ain
cont
aini
ng 1
148
CGA
GG
AA
AG
TCCC
CACG
TA
GTG
G
CgA
Gg
Aa
3 A
GtC
CcCA
CGT
AG
2
5 N
W_0
0361
4101
.1
4025
54
- Fk
bp9
1007
686
35
exon
;CD
S FK
506
bind
ing
prot
ein
9;pe
ptid
yl-p
roly
l cis-
tran
s iso
mer
ase F
KBP
9 A
AA
TCA
CAG
GCC
TCA
CCT
GG
TGG
A
aAtC
AC
2 A
GG
CCtC
ACc
TgG
3
5 N
W_0
0361
3741
.1
9660
49
+ A
dcy7
10
0756
132
ex
on;C
DS
aden
ylat
e cyc
lase
7;
aden
ylat
e cyc
lase
type
7
ACA
CCA
GA
GG
ACC
CACG
GA
GA
GG
A
CAcC
Ag
2 A
GG
aCcC
ACG
gA
G
3 5
NW
_003
6136
09.1
37
716
40
- Zn
f169
10
0758
704
ex
on;C
DS
zinc
fing
er p
rote
in 1
69
ACA
GA
AA
AG
GCT
GCC
AG
TA
GA
GG
A
CAG
aA
a 2
AG
GCt
GCc
aGT
AG
3
5 N
W_0
0361
3595
.1
9229
02
- LO
C107
978
628
1079
786
28
ncRN
A un
char
acte
rized
LO
C107
9786
28
ACA
GA
GCA
GG
CCG
AA
GG
AA
GTG
G
ACA
Gag
C 2
AG
GCC
GaA
gGa
AG
3
5 N
W_0
0361
5800
.1
8059
5 +
Sh3b
gr
1007
551
37
exon
;CD
S SH
3 do
mai
n bi
ndin
g gl
utam
ate r
ich
prot
ein;
SH3
dom
ain-
bind
ing
glut
amic
acid
-ric
h pr
otei
n A
CAG
ATC
AG
GTC
CCA
CGC
AG
GG
G
ACA
Gat
C 2
AG
GtC
cCA
CGc
AG
3
5 N
W_0
0361
3748
.1
2186
689
-
Dus
p4
1007
562
33
exon
du
al sp
ecifi
city
ph
osph
atas
e 4
ACA
GCG
TAG
GCC
ACA
TGA
AG
GG
G
ACA
GC
gt
2 A
GG
CCaC
AtG
aA
G
3 5
NW
_003
6143
00.1
22
035
0 +
Mrp
l38
1007
585
80
exon
;CD
S 39
S rib
osom
al p
rote
in
L38
mito
chon
dria
l;mito
chon
dria
l rib
osom
al p
rote
in
L38
ACA
GTA
TGG
GCC
GG
ACA
TA
GTG
G
ACA
Gt
At
2 gG
GCC
GgA
CaT
AG
3
5 N
W_0
0361
3677
.1
3369
92
+ Ct
dspl
2 10
0752
292
tr
ansc
ript
CTD
smal
l pho
spha
tase
lik
e 2
CAA
ACA
CAG
GG
TGCA
CCT
AG
TGG
Ca
AaC
AC
2 A
GG
gtG
CACc
TA
G
3 5
NW
_003
6139
25.1
63
144
5 -
Dclk
1 10
0768
340
tr
ansc
ript
doub
leco
rtin
like
kin
ase
1 CA
AG
CAG
AG
GCT
GCC
AG
TA
GG
GG
Ca
AG
CA
g 2
AG
GCt
GCc
aGT
AG
3
5 N
W_0
0361
4990
.1
3818
51
+ LO
C107
979
629
1079
796
29
ncRN
A un
char
acte
rized
LO
C107
9796
29
CAA
GCC
CAG
GCC
ATA
AGT
AG
GG
G
CaA
GCc
C 2
AG
GCC
atA
aGT
AG
3
5 N
W_0
0361
3665
.1
2195
085
+
Nca
ph
1007
512
28
tran
scrip
t no
n-SM
C co
nden
sin I
com
plex
subu
nit H
CC
ACC
GCG
GTC
AG
CACG
TA
GCG
G
CCA
cCg
C 2
gGtC
aGCA
CGT
AG
3
5 N
W_0
0361
3593
.1
4562
015
+
Pcdh
10
1007
544
59
exon
;CD
S pr
otoc
adhe
rin-1
0
CCA
GA
CCA
GG
CCG
CGCA
TCG
AG
G
CCA
Gac
C 2
AG
GCC
GCg
CaTc
G
3 5
NW
_003
6143
93.1
36
106
+ Zb
tb39
10
0752
450
ex
on;C
DS
zinc
fing
er an
d BT
B do
mai
n co
ntai
ning
39
;zinc
fing
er an
d BT
B do
mai
n-co
ntai
ning
pr
otei
n 39
CC
AG
AG
CAG
GA
CGCA
CGT
CCTG
G
CCA
Gag
C 2
AG
GaC
GCA
CGTc
c 3
5 N
W_0
0361
7481
.1
4053
2 +
LOC1
0797
748
8 10
7977
488
nc
RNA
unch
arac
teriz
ed
LOC1
0797
7488
CC
AG
CCA
AG
GCT
TCA
AG
TA
GA
GG
CC
AG
Cca
2
AG
GCt
tCA
aGT
AG
3
5 N
W_0
0361
7388
.1
4742
2 +
LOC1
0797
746
5 10
7977
465
nc
RNA
unch
arac
teriz
ed
LOC1
0797
7465
CC
AG
GTC
AG
GCT
GCA
CCT
ATC
GG
CC
AG
gtC
2 A
GG
CtG
CACc
TA
t 3
5 N
W_0
0361
4196
.1
7790
9 -
Tcf1
9 10
0756
450
ex
on;C
DS
tran
scrip
tion
fact
or 1
9
CCA
GTG
CAG
GCA
GCA
CTT
CGA
GG
CC
AG
tgC
2 A
GG
CaG
CACt
TcG
3
5 N
W_0
0361
4582
.1
5769
46
+ LO
C100
765
361
1007
653
61
exon
;CD
S ch
rom
osom
e unk
now
n op
en re
adin
g fra
me
hum
an
C16o
rf46;
unch
arac
teriz
e
149
d pr
otei
n C1
6orf4
6 ho
mol
og
CCA
TCCC
AG
CCG
GCA
TGT
AG
AG
G
CCA
tCc
C 2
AG
cCgG
CAtG
TA
G
3 5
NW
_003
6146
70.1
40
535
6 -
Abc
c8
1007
643
94
exon
;CD
S A
TP b
indi
ng ca
sset
te
subf
amily
C m
embe
r 8;
LOW
QU
ALI
TY
PRO
TEIN
: ATP
-bi
ndin
g ca
sset
te su
b-fa
mily
C m
embe
r 8
CCTC
CACG
GG
CCG
CGCG
TA
AA
GG
CC
tcCA
C 2
gGG
CCG
CgCG
TAa
3 5
NW
_003
6135
97.1
23
879
74
+ LO
C103
163
145
1031
631
45
ncRN
A;e
xon
unch
arac
teriz
ed
LOC1
0316
3145
CG
AG
CCCA
TGCC
GA
ACG
TG
GTG
G
CgA
GCc
C 2
AtG
CCG
aACG
TgG
3
5 N
W_0
0361
3739
.1
2144
751
-
Kia
a092
2 10
0771
764
ex
on;C
DS
KIA
A09
22
orth
olog
;tran
smem
bran
e pro
tein
131
-like
CG
TGCA
CAG
GCA
ACA
TGT
AG
AG
G
CgtG
CAC
2 A
GG
CaaC
AtG
TA
G
3 5
NW
_003
6135
80.1
37
807
03
- A
sb13
10
0766
461
ex
on;C
DS
anky
rin re
peat
and
SOCS
box
cont
aini
ng
13;a
nkyr
in re
peat
and
SOCS
box
pro
tein
13
CTA
GA
ACA
GG
CCG
CCA
TTA
GA
GG
Ct
AG
aAC
2 A
GG
CCG
Ccat
TA
G
3 5
NW
_003
6137
32.1
18
752
37
+ LO
C107
979
034
1079
790
34
ncRN
A un
char
acte
rized
LO
C107
9790
34
CTA
GCT
CAG
GCC
GA
ACT
TTG
AG
G
CtA
GCt
C 2
AG
GCC
GaA
CtT
tG
3 5
NW
_003
6148
46.1
45
758
1 +
LOC1
0076
110
2 10
0761
102
ex
on;C
DS
LOW
QU
ALI
TY
PRO
TEIN
: iso
citra
te
dehy
drog
enas
e [N
AD
P]
cyto
plas
mic;
isocit
rate
de
hydr
ogen
ase [
NA
DP]
cy
topl
asm
ic G
AA
GCC
CAG
GCC
ACA
CCT
GG
TGG
G
aAG
CcC
2 A
GG
CCaC
ACc
TgG
3
5 N
W_0
0361
6908
.1
8501
4 -
LOC1
0316
452
3 10
3164
523
nc
RNA
unch
arac
teriz
ed
LOC1
0316
4523
G
CACC
TCA
GA
CCG
CACG
CTG
AG
G
GCA
cCt
C 2
AG
aCCG
CACG
ctG
3
5 N
W_0
0361
3623
.1
3138
456
-
Dza
nk1
1007
671
39
tran
scrip
t do
uble
zinc
ribb
on an
d an
kyrin
repe
at d
omai
ns
1 G
CAG
ACC
AG
GG
GG
CACG
GA
GTG
G
GCA
Gac
C 2
AG
Ggg
GCA
CGg
AG
3
5 N
W_0
0361
3658
.1
1498
265
+
Zbtb
37
1007
749
39
exon
;CD
S zi
nc fi
nger
and
BTB
dom
ain
cont
aini
ng
37;zi
nc fi
nger
and
BTB
dom
ain-
cont
aini
ng
prot
ein
37
GCA
GCT
TGG
GCA
GCA
CGC
AG
CGG
G
CAG
Ctt
2 gG
GCa
GCA
CGc
AG
3
5 N
W_0
0361
3822
.1
1937
604
+
Trim
56
1007
698
53
exon
;CD
S E3
ubi
quiti
n-pr
otei
n lig
ase T
RIM
56;tr
ipar
tite
mot
if co
ntai
ning
56
GCA
TCA
GA
GG
ACG
CACA
CA
GG
GG
G
CAtC
Ag
2 A
GG
aCG
CACa
cA
G
3 5
NW
_003
6136
18.1
62
702
2 -
LOC1
0075
737
0 10
0757
370
ex
on;C
DS
zinc
fing
er p
rote
in 1
2
GCA
TCTC
AG
ACA
GCA
CGG
AG
TGG
G
CAtC
tC
2 A
GaC
aGCA
CGg
AG
3
5 N
W_0
0361
3646
.1
2427
176
+
Arfg
ap2
1007
555
28
exon
;CD
S A
DP
ribos
ylat
ion
fact
or
GTP
ase a
ctiv
atin
g pr
otei
n 2
GCA
TTA
CTG
GCC
CCA
CCT
AG
TGG
G
CAttA
C 2
tGG
CCcC
ACc
TA
G
3 5
NW
_003
6142
10.1
50
93
- LO
C103
164
099
1031
640
99
exon
br
eakp
oint
clus
ter
regi
on p
rote
in
150
GCT
GCC
CGG
CGCG
CACG
TA
GTG
G
GCt
GCc
C 2
gGcg
CGCA
CGT
AG
3
5 N
W_0
0361
3867
.1
5547
86
+ K
dsr
1007
745
78
exon
;CD
S 3- ke
todi
hydr
osph
ingo
sine
redu
ctas
e G
CTG
GA
CGG
GCA
GCA
GG
TA
GTG
G
GCt
GgA
C 2
gGG
CaG
CAgG
TA
G
3 5
NW
_003
6137
88.1
18
999
4 -
Parp
3 10
0764
843
ex
on;C
DS
poly
[AD
P-rib
ose]
po
lym
eras
e 3;
poly
(AD
P-rib
ose)
po
lym
eras
e fam
ily
mem
ber 3
TA
AG
CAG
AG
GCT
GG
AGG
TA
GG
GG
Ta
AG
CA
g 2
AG
GCt
GgA
gGT
AG
3
5 N
W_0
0361
3745
.1
7847
33
- LO
C103
160
422
1031
604
22
ncRN
A un
char
acte
rized
LO
C103
1604
22
TCA
CGA
CAG
AG
CGCA
CGA
AG
AG
G
TCA
cgA
C 2
AG
agCG
CACG
aA
G
3 5
NW
_003
6136
99.1
15
781
88
+ Pr
pf3
1007
621
16
exon
;CD
S U
4/U
6 sm
all n
ucle
ar
ribon
ucle
opro
tein
Pr
p3;p
re-m
RNA
proc
essin
g fa
ctor
3
TCA
TCA
AA
TGCT
GCA
CCT
AG
TGG
TC
AtC
Aa
2 A
tGCt
GCA
CcT
AG
3
5 N
W_0
0361
4266
.1
6757
34
- N
pc1
1006
894
24
exon
;CD
S N
PC in
trac
ellul
ar
chol
este
rol t
rans
port
er
1;N
iem
ann-
Pick
C1
prot
ein
prec
urso
r TC
ATC
CCA
GG
CCCC
ACA
CA
GA
GG
TC
AtC
cC
2 A
GG
CCcC
ACa
cA
G
3 5
NW
_003
6164
37.1
99
833
+ LO
C100
774
062
1007
740
62
ncRN
A un
char
acte
rized
LO
C100
7740
62
TGA
ACA
CAG
GCA
GCA
GG
CA
GTG
G
TgA
aCA
C 2
AG
GCa
GCA
gGc
AG
3
5 N
W_0
0361
5656
.1
2158
27
- K
cnk7
10
0763
826
ex
on;C
DS
pota
ssiu
m ch
anne
l su
bfam
ily K
mem
ber
7;po
tass
ium
two
pore
do
mai
n ch
anne
l su
bfam
ily K
mem
ber 7
TG
AG
AA
CAG
GCA
GCA
CGC
TGA
GG
Tg
AG
aAC
2 A
GG
CaG
CACG
ctG
3
5 N
W_0
0361
4689
.1
1747
93
- Tr
af3i
p1
1007
714
21
exon
;CD
S TR
AF3
-inte
ract
ing
prot
ein
1 TG
ATC
ACA
GCC
CGCC
CAT
AG
TGG
Tg
AtC
AC
2 A
GcC
CGCc
CaT
AG
3
5 N
W_0
0361
4150
.1
5531
96
+ Er
cc2
1006
892
72
exon
;CD
S ER
CC ex
cisio
n re
pair
2 TF
IIH
core
com
plex
he
licas
e sub
unit;
TFIIH
ba
sal t
rans
crip
tion
fact
or co
mpl
ex h
elica
se
XPD
subu
nit
151
Tabl
e S5
– G
row
th an
d sp
ecifi
c rat
es o
f nut
rient
s and
by-
prod
ucts
of c
ontr
ol a
nd en
gine
ered
T2_
6 ce
lls w
ith si
ngle
and
sim
ulta
neou
s disr
uptio
n of
Bca
t1 a
nd B
cat2
gen
es. I
VC
D –
Inte
gral
vi
able
cel
l den
sity,
µM
ax –
spec
ific g
row
th, b
y-pr
oduc
t rat
es IV
CD
, spe
cific
con
sum
ptio
n ra
te o
f glu
cose
(qG
luc )
, glu
tam
ine (
q Gln
), gl
utam
ate (
q Glu),
spec
ific p
rodu
ctio
n ra
te la
ctat
e (q L
ac) a
nd
amm
oniu
m (q
NH
4+).
Gro
up
C-T
2_6-
Cas
9 D
-T2_
6-C
as9+
Bcat
1 E-
T2_6
-Cas
9+Bc
at2
F-T2
_6-C
as9+
Bcat
1+Bc
at2
Clo
ne
C-A
5 C
-A8
C-B
3 D
-A2
D-A
12
D-B
11
D-C
2 E-
A5
E-B6
E-
B11
E-C
11
F-A
7 F-
A8
F-B4
F-
B12
IVC
D
(106 ce
ll*h
/mL)
735,
1 ±
28,4
2
931,
72
± 74
,43
986,
98
± 70
,87
664,
96
± 96
,19
1204
,99
± 39
,41
1360
,75
± 81
,69
1186
,91
± 98
,32
582,
37
± 30
,51
697,
7 ±
16,0
1
1563
,62
± 91
,28
650,
8 ±
47,9
2
1185
,29
± 15
9,27
1506
,61
± 16
7,7
676,
92
± 33
,96
1204
,22
± 74
,61
µ max
(day
-1)
0,8 ± 0,
09
0,74
± 0,08
0,69
± 0,12
0,66
± 0,03
0,8 ± 0,
06
0,92
± 0,03
0,99
± 0,06
0,65
± 0,09
0,76
± 0,04
0,78
± 0,04
0,71
± 0,13
0,92
± 0,15
1,01
± 0,05
0,61
± 0,07
1 ± 0,02
q Glu
c
(pm
ol/c
ell
/day
)
-1,9
6 ± 0,29
-1,6
9 ± 0,33
-2,1
1 ± 0,36
-1,7
9 ± 0,1
-1,4
9 ± 0,08
-1,5
6 ± 0,07
-1,6
5 ± 0,09
-2,5
4 ± 0,22
-2,6
1 ± 0,38
-1,7
0 ± 0,18
-2,3
4 ± 0,44
-1,9
9 ± 0,33
-1,3
7 ± 0,08
-2,6
0 ± 0,56
-1,6
9 ± 0,08
q Gln
(pm
ol/c
ell
/day
)
-0,6
9
± 0,10
-0,5
5 ± 0,12
-0,7
2 ± 0,12
-0,5
7 ±
0,00
3
-0,4
6 ± 0,04
-0,4
3 ± 0,02
-0,4
3 ± 0,02
-0,7
5 ± 0,12
-0,8
1 ± 0,11
-0,5
5 ± 0,06
-0,6
2 ± 0,15
-0,5
2 ± 0,10
-0,2
9 ± 0,02
-0,5
1 ± 0,09
-0,3
7 ± 0,01
q Glu
(pm
ol/c
ell
/day
)
0,04
± 0,01
-0,0
5 ± 0,01
-0,0
3 ± 0,02
0,07
± 0,01
-0,0
2 ± 0,01
0,00
± 0
-0,0
1 ± 0,01
0,07
± 0,01
0,02
± 0
-0,0
2 ± 0
0,02
± 0
-0,0
2 ± 0,02
-0,0
2 ± 0
0,02
± 0,01
-0,0
2 ± 0,01
q Lac
(pm
ol/c
ell
/day
)
2,88
± 0,31
2,09
± 0,46
4,39
± 0,72
2,47
± 0,1
2,08
± 0,37
1,92
± 0,21
2,55
± 0,18
5,67
± 0,55
2,8 ± 0,
41
2,11
± 0,12
2,94
± 0,75
3,29
± 0,85
2,12
± 0,13
4,62
± 0,41
2,59
± 0,11
q NH
4+
(pm
ol/c
ell
/day
)
0,78
± 0,11
0,65
± 0,14
0,83
± 0,13
0,76
± 0,02
0,59
± 0,05
0,53
± 0,01
0,55
± 0,05
1,12
± 0,14
1,05
± 0,14
0,68
± 0,06
1,04
± 0,1
0,66
± 0,15
0,45
± 0,03
0,9 ± 0,
17
0,56
± 0,01
152
Tabl
e S6
– S
peci
fic co
nsum
ptio
n ra
tes o
f leu
cine
(qLe
u), is
oleu
cine
(qIle
) and
val
ine (
q Val) o
f con
trol
and
engi
neer
ed T
2_6
cells
with
sing
le a
nd si
mul
tane
ous d
isrup
tion
of B
cat1
and
Bca
t2
gene
s. G
roup
C
-T2_
6-C
as9
D-T
2_6-
Cas
9+Bc
at1
E-T2
_6-C
as9+
Bcat
2 F-
T2_6
-Cas
9+Bc
at1+
Bcat
2
Clo
ne
C-A
5 C
-A8
C-B
3 D
-A2
D-A
12
D-B
11
D-C
2 E-
A5
E-B6
E-
B11
E-C
11
F-A
7 F-
A8
F-B4
F-
B12
q Leu
(p
mol
/ce
ll/da
y)
Rep
1 0,
35
0,23
0,
08
0,33
0,
23
0,22
0,
10
0,15
0,
23
0,12
0,
17
0,20
0,
11
0,25
0,
42
Rep
2 0,
31
0,20
0,
44
0,20
0,
11
0,09
0,
14
0,17
0,
13
0,08
0,
21
0,14
0,
11
0,18
0,
10
q Ile
(pm
ol/
cell/
day)
Rep
1 0,
55
0,36
0,
31
0,45
0,
29
0,29
0,
20
0,40
0,
40
0,20
0,
36
0,34
0,
18
0,39
0,
38
Rep
2 0,
44
0,37
0,
61
0,33
0,
22
0,17
0,
21
0,41
0,
26
0,20
0,
33
0,23
0,
17
0,32
0,
18
q Val
(pm
ol/
cell/
day)
Rep
1 0,
21
0,17
0,
04
0,21
0,
16
0,15
0,
08
0,11
0,
15
0,08
0,
07
0,14
0,
07
0,12
0,
27
Rep
2 0,
21
0,16
0,
27
0,10
0,
07
0,05
0,
10
0,11
0,
09
0,03
0,
13
0,10
0,
07
0,09
0,
05
*Rep
- Re
plic
ate
153
Appendix 3: Paper IV – Supplementary materials
A targeted study of stable overexpression of Glucose-6-phosphate dehydrogenase (G6pd)
in CHO-S cells: effect on cell growth and protective properties against ROS inducers and
cytotoxic agents
Sara Pereira (1), Lise Marie Grav (1), Tune Wulff (1), Helene Faustrup Kildegaard (1)(2), Mikael
Rørdam Andersen (3)
(1) The Novo Nordisk Foundation, Center for Biosustainability, Technical University of Denmark, Kongens
Lyngby, Denmark, (2) Current address: Novo Nordisk A/S, Department of mammalian expression, Måløv,
Denmark, (3) Department of Biotechnology and Biomedicine, Technical University of Denmark, Kongens
Lyngby, Denmark
Correspondence: [email protected]
Supplementary Materials 1 - FACS analysis of Recombinase-mediated cassette exchange efficiency
Figure S1- Recombinase-mediated cassette exchange efficiency determined using FACS analysis, based on the percentage of T2_6 cells displaying no mCherry fluorescence after RMCE.
154
Table S1- Recombinase-mediated cassette exchange efficiency determined using FACS analysis, based on the percentage of T2_6 cells displaying no mCherry fluorescence after RMCE.
Sample RMCE Efficiency (%)
G6pd-1 (T3.1) 0,87
G6pd-2 (T3.2) 0,85
3xstop+SpA (T5) 2,5
Non-transfected (NT) -
Supplementary Materials 2 – Target validation and verification
Table S2 - List of primers used in PCR and RT-qPCR (TaqMan assay) experiments.
# Primer name Sequence Application
1 Ef1-a_Fwd_seq CTCGTGCTTGAGTTGAGGC Insert PCR Out/Out and Out/In
2 Bgh_pA_rev_seq AGATGGCTGGCAACTAGAAG Insert PCR Out/Out
3 G6pd_seq2_rev CTTCTCCTTCTCCATTGGGGTTC Insert PCR Out/In
4 Gnb1_fwd CCATATGTTTCTTTCCCAATGGC RT-qPCR
5 Gnb1_rev AAGTCGTCGTACCCAGCAAG RT-qPCR
Gnb1_Probe ACTGGTTCAGACGATGCTACGTGC RT-qPCR
6 Fkbp1a_fwd CTCTCGGGACAGAAACAAGC RT-qPCR
7 Fkbp1a_rev GACCTACACTCATCTGGGCTAC RT-qPCR
8 Fkbp1a_Probe ATGCTAGGCAAGCAGGAGGTGATC RT-qPCR
9 mCherry_fwd GACTACTTGAAGCTGTCCTTCC RT-qPCR
10 mCherry_rev CGCAGCTTCACCTTGTAGAT RT-qPCR
11 mCherry_Probe TTCAAGTGGGAGCGCGTGATGAA RT-qPCR
12 G6pd_Probe TCTATCCCACTATCTGGTGGCTGTT RT-qPCR
155
Figure S2 - Target validation using PCR amplification of the inserted gene of interest into the target locus. Red arrows indicate the bands where G6pd was inserted. The term Out-out refers to PCR product resulting from primer sets hybridizing to Ef1a promoter and Bovine polyA signal sequences, while out-in refers to primers hybridizing to Ef1α promoter and a pre-selected region within the coding sequence of the gene of interest.
Supplementary Materials 3 - Changes in growth profile (VCD) cultivated in cell culture media
supplemented with inducers of cellular stress
Figure S3 - Changes in growth profile (VCD (cells/ml)) of cell pools expressing G6pd (G6pd-1 and G6pd-2), parental cell line (mCherry) and CHO-S wild type (WT) cells transferred to 6-well plates on day 3. The cells were cultivated in cell culture media supplemented with hydrogen peroxide (H2O2), a reactive oxygen species inducer of oxidative stress, H2O used for volume control, the cytotoxic agent sodium butyrate (NaBu), and NaCl used as osmolarity control, added on day 3. Viability was measured from day 4 to day 7. Five media formulations were included: Control - Basal media made of CD-CHO+8 mM L-glutamine+0,2% anti-clumping agent; H2O2- Basal media supplemented with 100 µM H2O2; H2O - Basal media with addition of same volume of H2O as in H2O2 ; NaBu - Basal media supplemented with 5 mM NaBu; NaCl - Basal media supplemented with 5 mM NaCl.
156
Supplementary Materials 4 - Total protein analysis SDS-Page gel
Figure S4 - Total protein analysis by electrophoretic separation of whole cell lysate using SDS-Page gel. Running conditions: MOPS buffer, ladder - PageRuler Plus Prestained (Cat. No. 26619, Thermo Fisher)
157