Engineering nutrient and by-product metabolism of CHO cells · Engineering nutrient and by-product...

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.

https://orbit.dtu.dk/en/publications/83ef725c-686f-4082-b2ee-cb9e812b6883

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

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

https://doi.org/10.1002/biot.201700499

http://www.biotechnology-journal.com

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)

http://www.advancedsciencenews.com


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)


Biotechnol. J. 2018, 1700499 © 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim1700499 (3 of 13)16



Table 1. (Continued)

Metabolite

KEGGcompound


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


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


Cysteine and methionine

metabolism

[57,58]

Phenylalanine C00079 Recommended to keep concentration below 0.5–1mM in fed-batch


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


[57,58]

Threonine C00188 Recommended to keep concentration below 0.5–1mM in fed-batch


Glycine, Serine and Threonine

metabolism

[57,58]

Tryptophan C00078 Recommended to keep concentration below 0.5–1mM in fed-batch




[57,58]

Tyrosine C00082 Recommended to keep concentration below 0.5–1mM in fed-batch




[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)






Metabolite

KEGGcompound


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


Tryptophan metabolism [57,58]

Indolelactate C02043 Metabolic by-product; recommended to keep concentration below 3mM


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


[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


Phenylalanine metabolism [57,58]

2-hydroxybutyric acid C05984 Metabolic by-product; recommended to keep concentration below


Cysteine and methionine

metabolism

[57,58]

3-(4-hydroxyphenyl)lactate C03672 Metabolic by-product; recommended to keep concentration below




[57,58]

4-hydroxyphenylpyruvate C01179 Metabolic by-product; recommended to keep concentration below 1mM




[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)






Metabolite

KEGGcompound


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.


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




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.


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





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


http://www.kegg.jp




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


http://www.CHOgenome.org

https://chomine.boku.ac.at

http://CHO.sf.net

http://CHO.sf.net

http://bigg.ucsd.edu/models/iCHOv1


http://www.CHOgenome.org

http://www.kegg.jp

http://www.kegg.jp

http://www.reactome.org

http://www.hmdb.ca



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]


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)

http://www.chogenome.org

http://www.chogenome.org





http://www.kegg.jp



http://www.hmdb.ca




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,





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.





[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.




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.


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]


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:

[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.

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


preparation

3 MiSeq_Bcat2_NW_003614570.1+1

68308_fwd TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGA

CAGCATCAGACCAAGGGG


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

Engineering nutrient and by-product metabolism of CHO cells · Engineering nutrient and by-product...

Documents

Transcript of Engineering nutrient and by-product metabolism of CHO cells · Engineering nutrient and by-product...