Fed batch fermentation a practical guide to scalable recombinant protein production 2015

coli , fermentation, recombinant DNA, yeast, nucleic acids, bacteria, RNA, phosphate plasmid DNA, recombinant protein, media, fed- batch, inclusion body, acetate, glucose, IPTG, cell fa

Therapeutic risk management of medicines

amsterdam • boston • cambridge • heidelberg • london

new york • oxford • paris • san diego san francisco • singapore • sydney • tokyo

Woodhead Publishing is an imprint of Elsevier

Fed-batch fermentation

A practical guide

to scalable recombinant protein production in

Escherichia coli

G arner G M oulton

Langford Lane, Kidlington, OX5 1GB, UK

Copyright © 2014 Woodhead Publishing All rights reserved

No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865

853333; e-mail: permissions@elsevier.com Alternatively, you can submit your request online by visiting the Elsevier website at http://elsevier.com/locate/

permissions , and selecting Obtaining permission to use Elsevier material

British Library Cataloguing-in-Publication Data

A catalogue record for this book is available from the British Library

Library of Congress Control Number: 2014938060

ISBN 978-1-907568-92-3 (print)

ISBN 978-1-908818-33-1 (online)

For information on all Woodhead Publishing publications

visit our website: http://store.elsevier.com/

Typeset by Refi neCatch Limited, Bungay, Suffolk

Printed and bound in the United Kingdom

Cover illustration: From the U.S Department of Energy Genomic program website, http://genomicscience.energy.gov

Figures

1.1 Nucleotide bases made up of pyrimidines and

purines as well as the addition of the sugar

ribose (RNA) or deoxyribose (DNA) and a

1.3 Base pairing in DNA is complementary 7

1.4 Conversion of simple sugars to ethanol and

1.6 Glycolytic pathway and acetyl-CoA

formation 16 1.7 TCA cycle and the formation of acetyl CoA

2.3 Isoproply- β -D-thio-galactoside (IPTG) shown

with the arrow pointing to the sulfur–carbon

bond that is not hydrolysable 39

2.4 Transcription of DNA 40

2.5 Micrograph of many transcription events

taking place on a DNA molecule 41

2.6 E coli micrograph 42

2.7 E coli cell wall structure and components 45

2.8 Transformation of a bacterial cell culture with

2.9 Draw a “T” on the bottom of your Petri dish,

2.10 Touch the inoculating loop to the upper

left-hand corner and then move it across the

agar from left to right, as shown 59

2.11 Touch the loop to the area previously

streaked and then move the loop across

2.12 Touch the loop on the previously streaked

area and then move the loop across the agar

onto the third area, as shown 60

2.13 Incubate the streak plate until you can see

3.1 Exponential growth curve for bacterial growth 64

3.2 Oxygen transport within the cell 76

3.4 ln (C* − C L ) versus Δ time (s) 79

3.5 Oxygen transfer rate and K L a determination 80

3.6 10-liter bioreactor for E coli fermentation 84

3.7 Dissolved oxygen electrode: polarographic

sensor 88 3.8 The pH electrode: Calomel electrode 91

3.9 Typical fed-batch fermentation growth curve 105

3.10 Analysis of residual acetate, glucose and

phosphate during the growth of the

3.11 Typical induction gel at prior to induction and

at 3 hours post-induction 108

4.1 Prokaryotic ribosomal composition 112

4.2 Translation of protein in prokaryotes 114

4.3 A tetrapeptide (V-G-S-A) with the amino

terminus of the peptide on the left and the

carboxyl terminus on the right 118

4.4 Amino acid names, structures and one letter

symbol associated with each 120

4.5 Primary, secondary, tertiary and quaternary

4.6 Bacterial GroES/GroEL complex 123

4.7 Aggregation pathways in vivo 133

Tables

4.1 Codons for amino acids and start and stop

sequences 113 4.2 Protein complexes within prokaryotic and

Gus G Moulton is Chief Scientifi c Offi cer of BioBench LLC,

a contracting facility for purifi cation and fermentation development in Seattle, USA Gus started the company in

2011 and is now pursuing this full time BioBench’s primary focus is initial development for product screening and vaccine Phase I clinical trials

Moulton has more than 20 years of process development experience in the biotechnology community During the last

13 years he has been responsible for setting up and running fermentation labs to generate medium to high cell density fermentations He performed these services for both Corixa Corporation, a former cancer vaccine company bought by GlaxoSmithKline plc, and the Infectious Disease Research Institute (IDRI), a nonprofi t organization which develops diagnostic tests and vaccines to diagnose and treat diseases

in third-world countries, such as India, Brazil and in Africa During Moulton’s career at Corixa he was initially responsible for purifi cation development of the most critical antigens, and subsequently for setting up and developing

recombinant E coli fermentation processes at the 30 liter

scale for Phase I clinical vaccine trials for HER2/neu He also developed an upstream and downstream process for the purifi cation of the recombinant antigen TcF to be used in the diagnostic test for Chagas disease The upstream process was designed per GLP standards for in-house use, while the downstream process was designed for and successfully transferred to Viral Antigens, Inc

During Moulton’s tenure at IDRI he again set up a

fermentation lab for development of recombinant E coli

production of foreign antigens Most fermentation development work Moulton performed at IDRI was for vaccine development against leishmaniasis – a disease caused

by protozoan parasites of the genus Leishmania and

transmitted by the bite of certain species of the sand fl y (subfamily Phlebotominae) – and tuberculosis caused by

Mycobacterium tuberculosis While at IDRI, Moulton

developed a unique feed recipe in which he supplemented

phosphate for a recombinant E coli fermentation using rich

media that tripled the fi nal cell density without any signifi cant increase in process cost or time Moulton also developed an

M smegmatis recombinant system that should easily be

scalable using a wave reactor This project can be used to produce Mtb antigens for both diagnostics and vaccine development

Over the last 13 years Moulton has successfully developed over 30 fermentation processes

Introduction to fermentation

DOI: 10.1533/9781908818331.1

Abstract: The use of yeast or microbial cells for the

production of a foreign protein has changed the approach

of medical research to fi nding healthcare solutions The

application of recombinant systems has become

mainstream in treatment of disease One of the most

important aspects of this new scientifi c discipline is the

ability to design a cell line or strain, in the case of bacterial

or yeast recombinant systems that can be grown under

controlled conditions, to produce signifi cant quantities of

a recombinant protein Recently, E coli has been the

predominant bacteria in research and production

laboratories and plays a key role in the development of

modern biological engineering and industrial microbiology,

enabling foreign proteins to be produced in a prodigious

and cost- effective way This type of cell growth and

production is called fermentation and its history and use

will be discussed along with current developments and

applications of recombinant technology

Key words: E coli , fermentation, recombinant DNA,

yeast, nucleic acids, bacteria, RNA, phosphate plasmid

DNA, recombinant protein, media, fed- batch, inclusion

body, acetate, glucose, IPTG, cell factory

1.1 A brief history of early

fermentation and the discovery

of DNA

It has been known for thousands of years that the fermentation

of carbon sources from grain and/or honey (for beer or mead) and grapes or other fruit (for wine) will yield a beverage, which when fermented correctly is quaffable as well as entertaining to the senses (a feeling of well- being or intoxication) In fact, scientists have shown through chemical analysis that jars found in northern China contained a mixed fermented beverage made from rice, honey and fruit made

9000 years ago [1] Throughout human history, cultures from Greece, Egypt, China and the Americas have produced fermented concoctions for many reasons, including religious, celebratory or personal consumption In Egypt, the god Osiris was believed to have invented beer Because of this, beer was thought of as an important part of society and family and brewed on a daily basis [2]

In Greece, by the 16th century bc , the fermentation of grapes into wine was common By the 3rd century bc , the moderate use of wine was thought of by many, including Plato and Hippocrates, as both benefi cial to health and happiness and of therapeutic or medicinal value [3] During this time, the poet Eubulus stated that three bowls (glasses) of wine were the ideal amount to consume, which roughly equals one

750 ml bottle of wine The cult of Dionysus believed strongly that wine or intoxication from wine would bring the consumers closer to their deities Along these lines, Eubulus, who wrote the play “Dionysus”, has Dionysus saying to his patrons:

Three bowls do I mix for the temperate: one to health, which they empty fi rst; the second to love and pleasure; the third to sleep When this bowl is drunk up, wise

guests go home The fourth bowl is ours no longer, but belongs to violence; the fi fth to uproar; the sixth to drunken revel; the seventh to black eyes; the eighth is the policeman’s; the ninth belongs to biliousness; and the tenth to madness and the hurling of furniture [4]

Interestingly, these words of wisdom and warning have held

up through the thousands of years since they were fi rst penned

In China, one of the fi rst alcoholic drinks made from rice, honey and fruit was thought of as a spiritual sustenance rather than a physical one It was also believed that the moderate use of fermented alcoholic substances was a mandate from heaven and important for inspiration, hospitality and medicinal uses

Needless to say, the fermentation of a few carbon sources by different yeast strains has had a profound effect on the world’s societies, culturally and, albeit much later, scientifi cally The historical aspect of fermentation will be commented on in this introduction but fi rst we need to look at one of the most important scientifi c discoveries in modern times, the discovery

of the cellular molecule, deoxyribonucleic acid (DNA) The DNA molecule was fi rst identifi ed and isolated by the Swiss physician and biologist Friedrich Miescher in 1869, with his work being published in 1871 [5] He had isolated “phosphate rich” molecules from white blood cells, but did not understand the molecules’ signifi cance This came later when Ludwig Karl Martin Leonhard Albrecht Kossel, a German biochemist and pioneer in the study of genetics, worked out the chemical composition of the DNA molecule He was awarded the Nobel Prize for Physiology or Medicine in 1910 for this work Kossel had isolated and described the fi ve organic compounds that are present in nucleic acid: adenine (A), cytosine (C), guanine (G), thymine (T) and uracil (U) Eventually these compounds were to become known as nucleobases, the

foundation for the formation and structure of DNA and ribonucleic acid (RNA) in all living cells

During this same time other scientists were working on determining the structures and chemical nature of these compounds One of these was the Russian biochemist Phoebus Levene, who published many papers on cellular molecules and

is credited with the discovery of the order of the three major components of a nucleotide, the phosphate, the sugar and the base [6] He also identifi ed the sugar components of both the RNA and DNA molecules as ribose and deoxyribose, respectively Levene worked extensively with yeast nucleic acids to identify the components and ultimately (in 1919) proposed that the nucleic acids were made up of one distinct base, a sugar and a phosphate molecule (Figure 1.1) [7] After nearly 30 years of nucleic acid research, the scientifi c community received three important contributions In the 1920s, Frederick Griffi th was studying the differences between two Pneumococcal strains (R (non- virulent) and S (virulent)), and while doing so came upon an interesting fi nding When he heat-killed the virulent S strain and mixed it with a live non- virulent R strain and then injected this mixture into mice, the mice died of pneumonia Griffi th did not realize it at the time, but he had discovered bacterial transformation through the transfer of DNA to a host bacterium In 1944, Oswald Avery and his Rockefeller University colleagues published work along these same lines but with a more defi nitive result They demonstrated a link between DNA and virulence of these same two strains, by transferring DNA from a heat-killed S strain that was treated with proteases (destroys protein), RNAses (destroys RNA) or DNAses (destroys DNA) to a living non- virulent strain (R strain) What they found was that only R cells, transformed with protease or RNAse treated DNA from the S strain, were shown to be virulent The DNAse treated mixture did not convert the R cells to a virulent strain

Soon after this work was presented, the Austrian biochemist, Erwin Chargaff, made a startling discovery when

he analyzed DNA from different species He noticed that the nucleotide composition was not the same from one species to the next He also discovered that within the structure of a DNA molecule, the purines (A and G) and the pyrimidines (C and T) are in equal amounts to the other ([A] = [G] and [C] = [T]) This fi nding of equality between base pairs is known as “Chargaff’s” rule (Figure 1.2) [8]

In 1956, James Watson and Francis Crick co- discovered the structure of the DNA and RNA molecules (Figure 1.3) Along with Maurice Wilkins, they were awarded the Nobel Prize in Physiology or Medicine in 1962 They leaned heavily

on the previous fi ndings of Chargaff and his colleagues at the

Nucleotide bases made up of pyrimidines and purines as well as the addition of the sugar ribose (RNA) or deoxyribose (DNA) and a phosphate group

Figure 1.1

time, as well as having the benefi t of the X-ray crystallography work on the DNA molecule done by Rosalind Franklin and Maurice Wilkins This crucial work led them to defi ne the DNA molecular structure as a double helix [9]

1.2 The rise of biotechnology I

1.2.1 The gene

Approximately 20 years after the determination of the structure of the DNA molecule, the term “biotechnology” was established Wikipedia states that biotechnology is “the application of scientifi c and engineering principles to the processing of materials by biological agents to provide goods and services” Biotechnology has its beginnings in what we call zymotechnology, which are the processes/techniques used for the production of beer Soon after World War I, with the advent of industrial fermentation taking a fi rm

Chargaff’s rule In DNA, the total abundance

of purines is equal to the total abundance of pyrimidines

Figure 1.2

hold on the current larger industrial issues, the path was paved for the increase in scientifi c research in the area of product formation from the single cell

By the 1970s, the term “genetic engineering” was becoming commonplace, ironically being used for the fi rst time in Jack

Williamson’s science fi ction novel Dragon’s Island [10], prior

to the connection of DNA as a hereditary molecule and the

Base pairing in DNA is complementary [4]

The purines (A and G) pair with the pyrimidines (T and C, respectively) to form equal- sized base pairs resembling rungs on a ladder (the sugar- phosphate backbones) The ladder twists into a double- helical structure

Figure 1.3

confi rmation of its structure as a double helix In 1972, the

fi rst recombinant DNA molecule was made by combining the DNA from the lamda virus and the SV40 virus This initial work was done by Paul Berg, which was followed by Herbert Boyer and Stanley Cohen creating the fi rst transgenic organism by inserting antibiotic resistance genes into a

plasmid of E coli [11]

Before 1983, the name Kary Mullis was little known at best Dr Mullis was a writer of fi ction, a baker, but not a candlestick maker He was, in fact, a very good biochemist who worked for the Cetus Corporation in California for seven years after his initial wanderings In this time he worked as a DNA chemist and eventually improved on the already existing polymerase chain reaction (PCR), although improvement is not a strong enough word for the contribution Mullis made to the PCR reaction [11]

A concept similar to that of PCR had been described before Mullis’ work Nobel Prize laureate H Gobind Khorana and Kjell Kleppe, a Norwegian scientist, authored a paper 17 years earlier describing a process they termed “repair replication” [12] Using repair replication, Kleppe duplicated and then quadrupled a small synthetic molecule with the help of two primers and DNA-polymerase The difference between Khorana and Kleppe’s work and Mullis’s is the fact that Mullis used the heat stable taq DNA polymerase instead

of the heat labile DNA polymerase (had to be added anew to each heat cycle) This new PCR method relies on thermal cycling, consisting of cycles of repeated heating and cooling

of the reaction for DNA melting and enzymatic replication

of the DNA Primers (short DNA fragments) containing sequences complementary to the target region along with a DNA polymerase (after which the method is named) are key components to enable selective and repeated amplifi cation

As PCR progresses, the DNA generated is itself used as a

template for replication, setting in motion a chain reaction in which the DNA template is exponentially amplifi ed PCR can be extensively modifi ed to perform a wide array of genetic manipulations [11]

As can be appreciated, within a short time, fermentation and the concept of recombinant protein production has matured and evolved to the point where protein products are produced for modern- day medicine The work of Boyer and Cohen, as discussed above, using plasmids and restriction enzymes to manipulate DNA (recombined with foreign genes) laid the groundwork for what is now known as biotechnology [13,14]

1.2.2 Controlled fermentation

Man has harvested the energy produced by fermentation to generate new and exciting products, used not only in medicine but also in bioremediation and agriculture (Figure 1.4) Surprisingly, even therapeutic antibodies are now being produced using recombinant expression hosts, other than Chinese Hamster Overy cells (CHO), such as yeast

and E coli Signifi cant progress has been made in antibody

engineering, with a particular focus on Fc engineering and

Conversion of simple sugars to ethanol and carbon dioxide

Figure 1.4

glycol- engineering for improved functions, as well as cellular engineering for enhanced production of antibodies in yeast

and bacterial hosts such as E coli [14] Bacteria, yeast and

some mammalian cell systems have been used to produce essential therapeutics such as insulin, as well as recombinant antigens for vaccines, diagnostic and therapeutic purposes [15,16]

E coli is a Gram- negative, rod- shaped bacterium that is

commonly found in the lower intestine of warm- blooded

organisms (endotherms) Most E coli strains are harmless,

but some serotypes can cause serious food poisoning in humans, and are occasionally responsible for product recalls The harmless strains are part of the normal fl ora of the gut, and can benefi t their hosts by producing vitamin K2, and by preventing the establishment of pathogenic bacteria within the intestine

E coli was one of the fi rst organisms to have its genome

sequenced with the complete genome of E coli K12

(MG1655) [17] It was 4.6 million base pairs in length, encoding 4288 protein genes, organized into 2584 operons

It was circular in structure with a large amount of DNA coded for genes (high genetic density) with only 118 base pairs distance between the genes Along with ribosomal RNA and transfer RNA genes, the genome was also shown

to contain a large number of repeat elements, transposable elements, and prophage and bacteriophage sequences

In microbiology studies, E coli has been used to study

metabolic pathways, cell division and mechanisms of cell death In 1946, Lederberg and Tatum discovered bacterial

conjugation using E coli as a model bacterium [18] Phage

genetics studies by early researchers such as Seymour Benzer were used to understand the topography of the gene structure;

to date, Escherichia and Shigella species comprise over 60 complete genomic sequences that are available [15] Only

about 20% of each genome is present in each genomic species, representing a fantastic amount of diversity within the genre The genes present in each individual genome number between 4000 and 5500 genes, while the number

of different genes found among all the E coli strains that

have been sequenced is greater than 16 000! This is called a pan- genome and is thought to have gained its diversity through the process of horizontal gene transfer from other species [19]

Prior to the discovery of restriction enzymes in the 1970s, researchers used ineffi cient ways to modify genetic material, such as what happens when a bacterium is infected by a bacteriophage or a foreign plasmid With the discovery and isolation of the restriction enzyme Hind III in 1970 [20,21] and the subsequent discovery and characterization of numerous restriction endonucleases [22], the 1978 Nobel Prize for Physiology or Medicine was awarded to Daniel Nathans, Werner Arber and Hamilton O Smith [23] During the 1970s, recombinant DNA technology and its use exploded onto the scientifi c scene One of the fi rst important products made with this new technology was the large- scale

production of human insulin for diabetes, using E coli as the

recombinant host

Since the 1920s, animal insulin was used to treat Type II diabetes, along with forms of insulin such as zinc insulin and the lente insulins for Type I diabetes In the 1960s, insulin was chemically synthesized in China, Germany and the United States By the mid-1970s, the separation technology advanced enough to be able to isolate animal (porcine, bovine) insulin to a single component by Novo and Eli Lilly By 1978, scientists from one of the fi rst biotechnology companies, Genentech (San Francisco, CA),

used a genetically engineered plasmid of the E coli bacterium

carrying the foreign gene for human insulin They were able

to produce the recombinant insulin with the same genetic

sequence as human insulin, meaning the E coli transcribed

and translated the foreign gene as it was in humans [15] By

1980, the fi rst recombinant DNA insulin product was injected into a healthy control group in England In 1982, the FDA awarded Eli Lilly the fi rst approved genetically engineered insulin (Humulin R and Humulin N) to be sold

on the US market [24]

After this initial success with recombinant technology, the sky was the limit, or so many scientists thought This new technology was to supply the world with any relevant recombinant protein, which was deemed necessary to address

a medical need Recombinant enzymes, hormones and immunogens (for vaccines) were going to be produced easily and cost- effectively But these expectations were quickly realized as much too grand and efforts to produce such proteins were constantly being stymied The scientifi c community started to realize that most of the proteins made

in the E coli recombinant system were not comparable

to the same protein made from natural source and thus these recombinant proteins were not safe for human use The recombinant proteins had two major obstacles to overcome:

1 proteolysis by host cell proteases [25]; and

2 the formation of inclusion bodies [26]

Ironically, human insulin was one of the fi rst recombinant proteins produced to show formation of inclusion bodies [27] Both of these issues, either combined or separate, interfere with the ability of the process development scientist

to produce recombinant protein products in their native state or at least produced with a consistent product character Part of the problem has been the lack of understanding from

a cell physiology standpoint, how the recombinant E coli

cell is affected by the conditions of a standard fermentation process

As mentioned earlier, recombinant E coli can be used to

develop antigens in vaccine development and proteins for therapeutic uses However, E coli has limitations, and

cannot be used to produce large multimeric heterologous proteins or proteins that require complex disulfi de bond formation or unpaired thiols or proteins that natively contain post- translational modifi cations There is a caveat to these stated limitations, in that in the presentation of a vaccine antigen, the secondary and tertiary structures are important but not essential for an immune response

The epitopes of protein antigens are divided into two categories, conformational epitopes and linear epitopes, based on their structure and interaction with the antibody A conformational epitope is composed of separated sections of the antigen’s amino acid sequence Although these epitopes are separated within the linear sequence of the protein, they are closely oriented spatially in the secondary or tertiary structure of the antigenic protein They interact with the antibody and the surface of a certain type of cell based on their 3-D surface features and shape or tertiary structure

By contrast, linear epitopes interact with the antibody based on their primary structure (amino acid sequence) A linear epitope, usually 8 to 11 amino acids in length, is formed by a continuous sequence of amino acids from the antigen Even though it is thought that most epitopes recognized by the immune system are conformational, it has been shown that non- conformational or aggregated proteins can elicit an immune response Among the critical factors in inducing antibody responses are molecular weight and the insoluble nature of the aggregate [28]

An epitope, also known as an antigenic determinant, is the part of an antigen that is recognized by the immune system,

specifi cally by antibodies, B cells or T cells Although epitopes are usually thought to be derived from non- self proteins, sequences derived from the host can be recognized and are also classifi ed as epitopes This happens in autoimmune diseases, of which there are many, such as Lupus, Crohns Disease and Diabetes mellitus Type I

T cell epitopes are present on the surface of an antigen- presenting cell (APC), where they are bound to MHC molecules T-cell epitopes presented by MHC class I molecules are typically the shorter peptides between 8 and

11 amino acids in length, whereas MHC class II molecules present longer peptides, and non- classical MHC molecules also present non- peptidic epitopes such as glycolipids [28]

1.3 The rise of biotechnology II

1.3.1 Recombinant technology and

E coli

Current research and development using E coli as a

recombinant system has focused on two different strains, E

coli B (BL21) and E coli K (JM109) The K-12A strain was

isolated from a stool sample of a patient and was labeled K-12 in 1922 at Stanford University [29] In the 1940s, the mechanisms of this strain were studied by Charles Clifton for the metabolism of nitrogen Edward Tatum also studied tryptophan biosynthesis using this K-12 strain Today, K-12 strains are used successfully in recombinant protein production, in both research and manufacturing settings

Another common laboratory E coli strain is the B strain,

named by Delbruck and Luria in 1942 (Figure 1.5) This bacteriophage was originally discovered at the Institute Pasteur by Felix d’Herelle in 1918 This strain changed

hands many times prior to coming to Delbruck and Luria, and eventually gave rise to more familiar strains such as BL21 and REL606 Both B and K strains have been studied extensively and have been found to respond differently to glucose concentrations in their growth media, especially when the glucose concentration is 10 g per liter or more [30] These differences in the metabolism of glucose are expressed within the glycolytic pathway and the tri- carboxylic acid (TCA) cycle (Figure 1.6) The B and K strains will process glucose and the subsequent glycolytic product pyruvate at different rates and thus create different anapluerotic stresses that can generate more or less acetate production and consumption This ratio of acetate production/consumption

is crucial to the effi ciency of carbon metabolism and ultimately the growth and recombinant protein production

in the fed- batch culture [31]

The E coli cell

Figure 1.5

Glycolytic pathway and acetyl-CoA formation

Figure 1.6

Under aerobic respiration conditions, glucose is commonly used as the main carbon source and is fed in a non- limiting fashion in order to reach high cell densities Complications can arise when the culture maintains a high growth rate during the exponential phase of growth with the secretion of acetate into the surrounding media This biosynthetic process

is exposing the bacterial culture to metabolic stress through their Central Carbon Metabolism (CCM) and is associated with higher acetate production [30,31] At high enough concentrations, the acetate can inhibit cell growth and/or recombinant protein production [32] Acetate can also decouple trans- membrane pH gradients, affecting amino acid synthesis, osmotic pressure and intracellular pH It has been shown that by adding yeast extract to the fermentation, either initially or during the feed, the acetate formation and its negative effects on the culture can be lessened [32] Other nutrient additions, such as phosphate, to the growing culture may also have positive effects on growth and product formation [33]

The effects of acetate accumulation have been reviewed extensively in the past fi ve years or so [30–35] and will only

be addressed here as it pertains to glucose feed rates and acetogenesis when the culture is grown on excess glucose During this period of aerobic growth on excess glucose, the respiration effi ciency can decrease due to metabolic overfl ow (Figure 1.6) This is called the bacterial Crabtree effect, in which as much as 15% of the glucose is excreted as acetate [36–39] The mechanism that causes the Crabtree effect is not fully understood but most likely involves repression of many TCA promoters and genes that encode enzymes that phosphorylate and transfer glucose to the intracellular matrix for processing Another indication of overfl ow

metabolism is the acetogenesis of the E coli , caused by the

excretion of acetate (Figure 1.7) This is the result of an

imbalance between the fast carbon fl ux into the central metabolism and the limited capacity of the TCA cycle or respiration [40–42]

The feeding method developed for medium- high density

recombinant E coli cell cultures must take the above metabolic

effects into account, in order to maximize culture density and recombinant protein production There are many different feeding strategies to be considered when optimizing fed- batch fermentation for the production of a recombinant heterologous protein product The control of the feed rate and when to

TCA cycle and the formation of acetyl CoA from acetate

Figure 1.7

increase the feed rate is paramount to this end The fed- batch fermentation can be controlled without a feedback loop (using a constant feed, increased feeding or exponential feeding), with a feedback loop with indirect control (DO-stat, pH-stat), or with a feedback loop with direct control (substrate concentration control) The indirect controlled feeding, such

as DO-stat or pH-stat, will be dependent on the rise of the dissolved oxygen (DO) content or pH, which is an indication that more carbon substrate is needed

Choice of media is another critical parameter for the optimal

growth of a recombinant E coli high density culture There are generally three different media that are used in recombinant E

coli fermentation: defi ned media, semi- defi ned media and complex media Defi ned media is a basic recipe of essential micronutrients, salts and a carbon source such as glycerol or glucose This media is generally used when the recombinant product is secreted into the medium and thus purifi cation of the recombinant product is straightforward Recombinant yeast cultures can be grown on defi ned media with glycerol and

methanol, in the case of the Pichia pastoris or P methanolica yeast strains [43] C elegans has also been cultured using a chemically defi ned media [44] Recombinant E coli strains are

also grown on this type of media, optical densities upwards of

400 (A600), but the ability of the culture to grow to high cell densities and maintain acceptable recombinant protein production seems to be host strain dependent [30] A semi- defi ned medium is usually made up of minimal salts and a rich nutrient source such as fetal calf serum (FCS) or yeast extract The yeast extract provides precursors to meet the demand for the high level of synthesis of the expressed protein and thus helps maintain specifi c cellular yield of the expressed protein

As well as providing a nitrogenous source for the expressed protein, it also promotes growth because it is a source for carbon [45]

Yeast supplemented media has been used to grow E coli

[34] and mycoplasmas [46] FCS supplemented media has been used to culture trypanosomes [46,47] and fungi [48] Complex media usually contains yeast extract, salt and a protein digest of tryptone or soytone With an optimized feeding of a concentrated carbon source such as glucose or glycerol, the semi- defi ned and complex medias can grow the recombinant culture to high cell densities, 100 to 190 g dry cell weight/L [30,34], while maximizing product formation The fermentation process takes place in what is known as a cellular bioreactor or cell factory, controlling critical growth parameters such as aeration (DO concentration), pH, temperature and mixing (agitation) These high- cell density cultures have been scaled for the production of recombinant proteins with high yield and high productivities [49]

Recombinant protein production using E coli host strains

are induced by the allolactose mimic isopropyl β thiogalactopyranoside (IPTG) Allolactose is a lactose

-D-1-metabolite that turns on transcription of the lac operon

IPTG does the same thing as allolactose, but is not metabolized because of the sulfur- carbon bond that makes a chemical bond that is not hydrolysable by the cellular metabolic machinery (Figure 1.8)

An inducer, such as IPTG, is used at 0.1 to 1.0 mM concentrations Most proteins, when expressed in this type

of recombinant system, are expressed as inclusion bodies An inclusion body is basically a precipitated protein The cellular machinery forms the recombinant protein products into inclusion bodies for a number of reasons and will be addressed later in this chapter

During expression of the recombinant protein, certain culture conditions can be changed to enhance the probability

of some form of soluble expression, whether it is a soluble non- specifi c multimer or actually a soluble monomeric

protein Conditions that may affect solubility during expression are pH, temperature and amount of IPTG, used

as well as media components and stage of growth of the

recombinant E coli culture [50] The actual sequence and

size of the recombinant protein will also affect solubility characteristics and whether it can be folded correctly Smaller, more hydrophilic sequences tend to have a better chance of soluble expression

Another important factor in soluble expression is the host cell used There are a number of B and K strains to choose from and all can and should be screened for soluble protein production under varying growth conditions Everything else being equal, the most important aspect of the generation

of a correctly folded, soluble protein is the process of folding itself This fi nal structure of the protein is encoded in the primary sequence of the molecule [51] and during the process

of translation the protein goes through numerous transitions

or fl uctuations in the conformation to fi nd the structure with the lowest amount of energy and is thus most stable Protein folding is not a continuous event but is punctuated and predicated on the formation of a “folding nucleus” that is made up of amino acid residues that associate strongly in a

Isopropyl β -D-1-thiogalactopyranoside (IPTG)

Figure 1.8

native structure The rest of the protein sequence then quickly associates around this nucleus to form the completed native protein structure [52]

Whether a recombinant protein is correctly folded and soluble is certainly dependent on its size and if it contains multi- domain regions that require extra time and space (creation of intermediates) to fold properly When a protein

is miss- folded, it has gone through a series of non- native conformational changes that expose different hydrophobic areas, which are normally buried within the protein structure, are now associated with each other and this tends to form a hydrophobic seed for the protein to aggregate around To prevent mis- folding, the cell will use specifi c molecular chaperones that help with the correct folding and other factors such as folding catalysts SurA, FkpA and Skp/OmpH

found in the extra- cytoplasmic space in E coli [53]

It was found by Strandberg and Enfors [54] that at higher temperatures (42 °C) during induction of a heterologous protein, an increase in soluble recombinant protein was observed Lower temperatures of 39 °C showed a signifi cant lack of soluble protein formation At 42 °C this increase in soluble protein production was thought to be attributed in some way to the heat shock protein family called chaperonins There are a handful of known heat shock “family” genomes that are important in protein folding and processing Hsp70

is one such family and is one of the most conserved heat

shock genomes being encoded in all living organisms In E

coli there are three Hsp70 proteins (DnaK, HscC and HscA),

with DnaK being the most characterized of them [55,56]

DnaK has three main roles in E coli :

1 prevents aggregation;

2 controls ATP-dependent unfolding; and

3 helps GroEL (S) in refolding of the translated protein

Interestingly, heat shock proteins GroEL and GroES were

found to reconstitute inclusion body proteins, in vitro , after

they had been denatured using a chaotrope such as urea or guanidine Conversely, Song et al [57,58], using a recombinant system in Rosetta- gami2(DE3), were able to reduce the production of inclusion body formation of the recombinant proteins CaMan and CaCel and increase the soluble form of this protein by drastically reducing the induction temperature from 37 °C to a range of 10 to 6 °C Even at a temperature of 15 °C, the inclusion body formation far exceeded the production of the soluble form Soluble production seemed to level out after 8 °C Not surprisingly, not all host cell strains tested were able to produce the recombinant in a soluble form One test strain produced little to no recombinant protein The only drawback from using this method is that since the temperature is so low the induction time must be increased dramatically They report

an induction time of eight days for maximum production of the recombinant proteins, with the induction fi delity falling signifi cantly after that period of time They did not give a reason for this rapid decrease in product formation

Overall, the recombinant protein product formation at low temperatures depends on a number of bacterial and molecular parameters:

■ host (growth, protease activity at low temperatures);

■ recombinant protein size;

■ hydrophobic/hydrophilic nature of the primary structure

of the recombinant protein; and

■ promoter used

When pursuing a low temperature induction for production purposes, the process development scientist must consider the extended time that it will take for production of the

recombinant product, the cost of extended use of

fermenta-tion suite, and the fi nal fermentafermenta-tion yield of the soluble

form of this product Generally, the product yield in this type

of low temperature process is signifi cantly lower than a

process developed at 37 °C There will be a trade- off to be

considered when developing downstream processes for the

soluble protein product compared with the insoluble

inclusion body protein product

The pH of the media, as mentioned above, can have a

profound effect on the internal pH of the cell Hickey and

Hirshfi eld found that by changing the pH of the media from

7 to 5, the internal pH of the E coli cell was changed by

1 unit [59] This observation leads to a reaffi rmation of the

importance of monitoring and controlling the acetate

concentration within the growing culture In holding with

the basic principles of acid- base kinetics, the acetate molecule

will be in equilibrium between two forms, the dissociated

form and the non- dissociated form:

CH 3 -COOH ⰆⰇ H + + CH 3 -COO − [1.1] The non- dissociated form of acetate, due to its non- ionic

nature, is able to penetrate the cell and then dissociate causing

a lowering of pH within the cell and alter cell growth [59]

The study of the formation of inclusion bodies within the

E coli cell during induction of the recombinant protein has

been a topic of interest for decades The hydrophobic regions

of the protein that would normally be located internally are

now exposed and act to aggregate these miss- folded proteins

Furthermore, other modifi cations that normally could be

made to a eukaryotic protein, such as glycosylation or

lipidation, are not present For all the trouble that inclusion

bodies have caused process development scientists, ironically

the formation of inclusion bodies is a signifi cant fi rst step

towards purifi cation of the recombinant protein The

challenging part of the process is taking this aggregated mess

of a recombinant protein (cellular protein, nucleic acids, lipids and other cell debris) and purifying the recombinant protein to homogeneity Of course, the degree of contaminants will vary depending on the protein being produced and the host strain, as well as methods used for processing of the cells

Inclusion bodies are formed in the cytosol or the periplasmic

space of the E coli cell During this process, the cell machinery

will attempt to fold this foreign protein properly Most of the time the transcription and translation from the recombinant plasmid are so strong that the protein will ultimately not be folded correctly and due to the crowded nature of the cell’s cytoplasm, will precipitate or form aggregates as the protein concentration rises to an amazing 200 mg/ml [60] Furthermore, during this process the cell machinery, seeing that this overly expressed protein is not being folded properly, will try and assimilate it through proteolytic digestion [61,62] Manipulation of these genes as well as chaperone genes through mutation will partially or fully negate their respective activities This results in more manageable recombinant protein products, in either soluble or inclusion body form Put simply, the full length recombinant protein product will have a signifi cant reduction in truncated versions associated with it, making the initial purifi cation development much easier to perform

Another question asked by recombinant bacterial researchers over the years has focused on the purity of the inclusion body as it applies to the total protein content It seemed that within the inclusion body, the majority of the total protein was one protein species and not a heterogeneous

mixture Indeed, it has been shown through in vitro studies

that when an inclusion body of one protein was mixed with

a soluble form of the same protein and another unrelated

soluble protein, only the same protein that was in the inclusion body formed an inclusion body The other unrelated protein did not aggregate [63]

Historically, inclusion bodies have been thought to be a haphazard aggregation of protein that is structurally heterologous in nature, with no internal molecular scaffolding More recent fi ndings have shed increasing understanding on the architecture of the inclusion body, which seems to be much more highly ordered than fi rst believed It takes on the organized forms of β -sheet structures and small α -helical structures The β -sheet may not have the exact structure of a soluble β -sheet, but is most likely a super dense β -sheet matrix [63] As discussed above, lowering of the induction temperature can infl uence the degree to which

a recombinant protein is produced in a soluble form Along those same lines, lowering the temperature during induction can also have a signifi cant impact on the structural formation

of the inclusion bodies Amey and Rosenberg (2006) [28] showed that when the recombinant human granulocyte-

colony stimulating factor (hG-CSF) was induced in E coli at

low temperatures, the inclusion body that was formed was a heterogeneous mixture of what can be termed a soft inclusion body and a hard inclusion body The soft inclusion body can

be solubilized easily, without strong chaotropic agents such

as urea or guanidine The solubilized soft inclusion body was also shown to retain its native structure

Following the idea of the soft inclusion body, it has been observed by a number of process development scientists (including the author) that even though a recombinant protein can be found in the cell lysate, this does not mean it

is soluble It is more likely to have been deposited as “soluble

aggregates” within the cell cytoplasm or periplasm of E coli

Due to these fi ndings, it is diffi cult to correlate solubility with correctly folded structures of E coli recombinant

proteins More likely, there is a heterogeneous mixture of protein conformations, both soluble and insoluble, that is produced within the cell Purifi cation of these different fractions will undoubtedly require different approaches and it will be up to the process development scientist to decide the pros and cons of each, moving forward towards an economical and robust recombinant process [64–66]

Typically, when constructing the fi rst iteration of the recombinant protein plasmid, an affi nity tag is included on the N or C terminus of the protein of interest An affi nity tag that is often used is the hexahistidine tag (6His tag) that can bind to solid bead matrices, usually made of agarose beads, with divalent cations strongly associated with the bead material Cation metals such as nickel, cobalt and zinc have been use with great success in purifying recombinant proteins One of the more compelling reasons to initially use the 6His- tag is that it can make the initial purifi cation process development more straightforward In the case of inclusion body purifi cation, the Ni resin can be exposed to strong solubilizing agents, including strong ionic detergents such as sodium dodecyl sulfate (SDS), deoxycholate or non- ionic detergents Triton 100 or Triton 114 Highly denaturing chaotropes such as 8 M urea or 6 M guanidine-HCl can also

be used These two different types of denaturants both help

in the purifi cation, not just the solubilization of the recombinant protein Once bound to the Ni column, the protein can be washed extensively in this denaturing

background This wash allows E coli cell contaminants,

such as the antigenic lipopolysacharide (LPS) DNA, RNA and host proteins, to be removed from the protein of interest [65]

If the protein was solubilized in the chaotrope urea (8 M), the wash usually consists of urea, salt, varying small

amounts of the imidazole (helps remove nonspecifi c 6His bound molecules) and a moderately strong detergent such as the zwiterionic detergent 3-[(3-cholamidopropyl)dimethylammonio]-1-propane- sulfonate (CHAPS) Typically, strong detergents such as SDS are not miscible with 8 M urea

or 6 M guanidine-HCl Once the wash is concluded, the protein of interest can be eluted by either a pH shift to pH 6.0 or with a step or linear gradient of at least 0.5 M imidazole The imidazole elution is usually the preferred method, since a good number of proteins have isoelectric points (pI) that are in the range of 5 to 6 Using the pH elution method would expose the recombinant protein, unnecessarily, to a pH close to their pI and thus risk precipitation of the Ni eluted product This precipitation would not necessarily happen in the urea, but may happen upon dilution into a non- denaturing buffer system

Although this is protein dependent, with larger recombinant proteins, especially polyproteins (the combination of two or more heterologous primary sequences attached end to end) have a greater tendency to be less stable once they have been exposed to their pI Other affi nity tags that are available and effective to an equal or lesser degree are the FLAG-tag, which

is an octapeptide with the sequence N-DYKDDDDK-C (1012 daltons) It is considered antigenic and when synthesized on the N or C terminal end of a recombinant protein, can be used in an affi nity purifi cation matrix or if there is no antibody to the protein of interest, it can be used

to follow along with additional purifi cation or monitor expression within the cell Other affi nity tags of mention are the HA-tag (human infl uenza hemagglutinin), which corresponds to the sequence YPYDVPDYA (1202 daltons) within the glycoprotein, and myc- tag, which is from the c- myc gene Both can be used for purifi cation of recombinant proteins, applying the protein to an antibody column to the

respective epitope, while the antibody itself can monitor recombinant expression or cellular localization studies using immunofl uorescence as well

The scope of this book will take the reader from development of a working cell bank and inoculant through

to the purifi cation of a recombinant histidine tagged protein The fed- batch fermentation will be of medium cell density at the 2 liter scale, should be straightforward and able to be accomplished in one day The recommended host cell will be HMS 174 (DE3) The focus will be on the parameters that

affect the recombinant E coli fermentation and thus infl uence

the quality and quantity of recombinant protein produced during downstream processing As stated earlier in this chapter, the methods chosen toward processing of the

recombinant E coli inclusion body also play a critical role in

evaluating downstream process parameters, fi nal purity, stability and functionality of the recombinant protein This book will act as a guide for students and researchers alike, who intend on furthering their understanding of recombinant

protein production in the E coli recombinant system

Abstract: When generating a recombinant cell line that

expresses a foreign gene of interest, the work starts with

an E coli host strain such as HMS-174 (DE3) or BL21

(DE3), a plasmid of choice such as pET-28a or

pET-29a, and a gene of interest (usually a protein that is

needed for diagnostic or therapeutic use) A suffi cient

expression screening process of different host cells should

be performed, which allows expression analysis of the

protein of interest in terms of expression fi delity, protein

quality and stability This process matches the host cell

machinery (protease profi le, protein traffi cking, etc.) with

the specifi c properties of the recombinant protein product

(stability, hydrophobic character, toxicity to host cell,

etc.) This chapter presents a short review of the

transcription and translation processes that occur within

the recombinant cell

Key words: plasmid, host cell, bacteria, recombinant,

E coli , cloning, transcription, translation, expression,

gene, protein, competent, media, 2XYS, antibiotic

selection, working cell bank, transformation

2.1 Plasmids

As was stated in the introduction, E coli was the fi rst

microorganism to be thoroughly analyzed by both genetic and molecular biological means This led to it also being the fi rst

to be used for genetic engineering and recombinant protein production Even though much is known about the bacterium

E coli , it is certainly not a trivial task to set up a recombinant

system within this host, or any host for that matter Some of

the main issues to address in a recombinant E coli system are

the instability of the plasmid vectors, initiation and translation problems, along with mRNA stability

Plasmids are generally small circular DNA structures found

in many different strains of bacteria They can also be found in large numbers, over 100 copies per cell, and are self- replicating

In recombinant bacteria such as E coli , this allows for the

production of signifi cant amounts of plasmid DNA or recombinant proteins (potentially) from a small volume of cells For the purposes of recombinant DNA or protein production, plasmids have been developed with a number of advantages for cloning Almost all plasmids carry a gene or genes that encode for antibiotic resistance This antibiotic selection allows the host cell to grow in the media in the presence of an antibiotic such as kanamycin If a cell looses the plasmid or the antibiotic resistant gene, the cell is killed This helps ensure the continued presence of the plasmid as the cell culture grows and is ultimately harvested for the plasmid DNA or an induced foreign protein product

This is important in terms of recombinant E coli high cell

density fermentations when the cells go through many divisions Without the antibiotic selection, the culture would eventually contain a heterogeneous population of cells, some with plasmid and some without The plasmid- free cells would not have the metabolic pressure of maintaining the