Production and Purification of Recombinant Proteins

from Escherichia coli

Nagesh K. Tripathi[1]

Abstract

Development of successful production strategies for ity of the recombinant proteins are not amenable
recombinant proteins is a challenging task due to directly to large scale production processes due to
problems associated with the scale-up of bio- various factors. To meet the growing demand of
processes. Recent developments in fermentation potential bioproducts, there is a need to produce
technology have provided a suitable platform to these products in large quantity employing fermen-
produce bioproducts for the development of diag- tation processes. Development of a suitable purifica-
nostics, therapeutics and vaccines for bacterial, viral tion strategy is yet another important step to
and other diseases. A bacterial system is the com- recover biologically active products. The present pa-
monly used expression system for production of per reviews the current developments in production
recombinant products such as proteins, enzymes, and purification of recombinant products from
and antibodies. Escherichia coli is one of the most E. coli. The bioproducts obtained after fermentation
commonly used bacterial hosts for the production and purification processes may fulfill the require-
of these products. A variety of recombinant proteins ments for the development of various diagnostics,
has been produced and cultivation studies have therapeutics, or prophylactic agents against different
been carried out at small scale. However, the major- kinds of diseases.
Keywords: Escherichia coli, Fermentation, Purification, Recombinant protein, Refolding, Scale-up
Received: January 23, 2016; revised: March 29, 2016; accepted: March 31, 2016
DOI: 10.1002/cben.201600002

1 Introduction duce useful proteins as well as to investigate protein function

[7–10].
Bioprocessing is an essential part of the biotechnology industry, Recombinant proteins are generally expressed using bacterial
which is used for the production of industrial enzymes, vac- systems, yeasts, transgenic plants, mammalian cells, and insect
cines, and therapeutics. Recombinant proteins gain enormous cells. Yeast systems are used when extracellular expression of
importance in detection, protection, and treatment of diseases. recombinant proteins in large amounts is needed [11] and
The demand of proteins for these applications is set to see a when the product contains post-translational modifications
major increase over the coming years. Thus, the process [12, 13]. However, the yeast systems have disadvantages due to
employed to produce these important products must meet the their inability to modify proteins with human glycosylation
requirements of the global market [1]. Approximately 50 % of structures for the expression of recombinant proteins [11]. The
all new medicines are classified as biopharmaceuticals [2]. application of transgenic plants and animals as expression sys-
There are more than 300 biopharmaceutical products in the tems still remains to be addressed due to regulatory issues.
market with sales exceeding US$ 100 billion which includes Even though the choice of expression systems is progressively
therapeutic monoclonal antibodies (>US$ 18 billion), hor- increasing, the bacterial expression system with E. coli as host
mones (>US$ 11 billion) and growth factors (>US$ 10 billion). is still the choice for production of recombinant products
Biopharmaceuticals approved from 2004 to 2013 are pro- [14–16].
duced from mammalian cells (56 %); Escherichia coli (24 %),
Saccharomyces cerevisiae (13 %), transgenic animals, plants and
insect cells [3–6]. Recombinant DNA technology platform has —————
been pivotal in allowing proteins to be produced in sufficient [1]
Dr. Nagesh K. Tripathi
quantities. Using this technique, chimeric proteins can be gen- Bioprocess Scale-up Facility, Defence R&D Establishment, Jhansi
erated with novel functions by gene fusion. In molecular bio- Road, Gwalior 474002, India.
logy, protein engineering has become an efficient tool to pro- E-Mail: [email protected]

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 116
 E. coli is used in many fields for the production Table 1. Examples of industrial products produced in E. coli and their manufac-
of pharmaceuticals, neutraceuticals, and high value turers [17, 33–35].
intermediates. Significant advances have also been
made to express therapeutic proteins, diagnostic Product Manufacturer
intermediates, vaccine molecules, antibodies, en- Insulin Eli Lilly, Aventis
zymes, and complex proteins [9, 17–20]. Foreign
proteins can be produced in E. coli in large PEG interferon alpha-2b Schering-Plough
amounts about 5 to 50 % of the total cellular pro- Interferon beta-1b Novartis
tein [21]. E. coli has been successfully utilized to
produce many products such as proinsulin, growth Pegloticase Savient pharma
hormone, interleukin, and antibody fragments Calcitonin Unigene
[10, 17, 22–24]. Recombinant human insulin was
first produced in E. coli in 1978 and approved in Tissue plasminogen activator Roche
1982 for the treatment of diabetic patients [6]. Asparaginase Merck
Protein production in E. coli can be significantly
Cholera toxin B subunit SBL Vaccine
enhanced by the use of fermentation processes.
Batch, fed-batch and continuous modes of fermen- Human growth hormone Genentech, Eli Lilly, Pfizer, Novo Nordisk
tation are employed for production of bioproducts.
Interleukin-2 Chiron
Researchers obtained more than 100 g biomass per
liter of broth for the recovery of recombinant pro- Human TNF alpha Boehringer Ingelheim
teins using fed-batch processes [25–27]. High yield
Outer surface protein A Smithkline
production of recombinant proteins in E. coli often
accumulates as insoluble aggregates such as inclu- Parathyroid hormone Eli Lilly, NPS pharmaceutical
sion bodies (IBs) [21, 28]. IBs must be solubilized Tumor necrosis factor Boehringer
and then refolded to recover active proteins.
For application as a therapeutic candidate or Urate oxidase-PEGylated Savient
vaccine molecule, the recombinant protein must be Thrombopoietin peptibody, Fc Fusion Amgen
in a highly pure form. To obtain highly pure pro-
tein, chromatographic purification techniques such Glucarpidase BTG International
as affinity, ion exchange, hydrophobic interaction Lucentis Genentech, Novartis (Ranibizumab)
and gel filtration chromatography are commonly
used [20, 26, 29–32]. In the present review, we have Nivestim (filgrastim, rhGCSF) Hospira
focused on the recombinant protein production in
E. coli. The main focus has been on why E. coli is a
working horse for biotechnology and on various steps involved The protein of interest can be expressed as inclusion bodies or
in production and purification of recombinant proteins from in soluble and secretory forms. The main drawback of this
E. coli. system is the production of recombinant products in a non-
functional state since post-translational modification does not
occur in E. coli. The other disadvantage includes loss of plas-
2 Escherichia coli – A Working Horse mid and antibiotic property, unsolicited inducers for gene
for Protein Production expression, improper protein refolding, protein-mediated
metabolic burden and stress, endotoxin issues, and poor secre-
Recombinant proteins have gained enormous importance for tion [6, 9, 15, 38, 40]. The advantages and disadvantages of the
pharmaceutical uses. The E. coli expression system continues E. coli system from an industrial point of view compared to
to dominate the bacterial expression systems. The first yeast and mammalian expression systems are given in Tab. 2.
recombinant therapeutic products have been produced in Among yeast expression systems, S. cerevisiae, and Pichia
E. coli 30 years ago. Recombinant human insulin or recombi- pastoris are very commonly used for production of recombi-
nant human growth hormone entered the market in the early nant proteins [11]. Advantages of these systems include rapid
1980s. To date, a large number of recombinant proteins have growth, easy handling, and ease of various genetic manipula-
been produced using E. coli and many more are currently tions. Yeast-derived proteins are properly folded and glycosy-
under development. The list of some industrial products pro- lated. Recombinant proteins expressed in mammalian cells are
duced using E. coli is given in Tab. 1 [33–36]. properly folded, glycosylated and generally yield a functionally
The availability of alternative expression systems like yeast active protein. However, disadvantages of mammalian expres-
and mammalian systems has experienced a major boost. The sion systems include expensive media, difficult cultivation, and
main advantages, such as well-understood cell biology, easy viral contamination. The comparison of yields of recombinant
handling, use of simple growth medium, rapid cell growth, proteins produced in E. coli and other expression systems is
simple fermentation process, virus-free product, high product given in Tab. 3 [6, 11, 41].
yields, cost effective production, and ease of manipulation, E. coli has been mostly used for the expression of low molec-
make E. coli a popular choice of host for proteins [15, 36–39]. ular weight proteins to considerably high molecular weight

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 117
Table 2. Advantages and disadvantages of the E. coli system for the industrial produc- 3 Recombinant Protein
tion of bioproducts compared to other systems [15, 36, 38].
Production in E. coli
Expression system Advantages Disadvantages

E. coli - Inexpensive culture media, - Inability to form disulfide bonds, 3.1 Host
- Easy to cultivate, - Non-glycosylated proteins,
- Rapid growth rate, - Endotoxin For any recombinant protein production,
- High cell density cultivation, selection of the strain is an important con-
- Ease of genome modifications,
sideration. A variety of hosts is available
- High productivity and product yield,
for the expression of recombinant pro-
- Easy process scale-up,
- Cost effective, teins. Generally, E. coli B and K12 strains
- Virus free are used for the expression of recombinant
proteins [42–45]. Selection of host strains
Yeast (P. pastoris - Inexpensive culture media, - Hyperglycosylation depends on the nature of the heterologous
and S. cerevisiae) - Easy to cultivate,
protein. Rosetta (BLR (DE3)) and BL21
- Rapid growth rate,
(DE3) are used for proteins that contain
- High cell density cultivation,
- Ease of genome modifications, large numbers of rare E. coli codons [45].
- Easy process scale-up The solubility of disulfide-bond-containing
protein can be enhanced to produce prop-
Mammalian cells - Humanized glycosylation pattern, - Difficult to cultivate, erly folded proteins by using AD494
- Properly folded - Low yield,
strains (thioredoxin reductase mutants of
- Expensive culture media,
the K12 strain) and the Origami strain
- Viral contamination
(the thioredoxin reductase and glutathione
reductase mutants) with a more oxidizing
Table 3. Comparison of the yield of some recombinant products produced using E. coli cytoplasmic environment which provide
and other expression systems [6, 11, 41]. the potential to produce properly folded
active proteins [44, 45]. The expression
Recombinant product Expression system Maximum product concentration
level could be improved for toxic proteins
Insulin E. coli 4.34 g L–1 using a strain containing the pLysS plas-
mid (BL21 (DE3)-pLyS) and for globular
P. pastoris 3.07 g L–1
as well as membrane proteins using a
S. cerevisiae 0.075 g L–1 strain containing C41 (DE3) and C43
(DE3) [46]. In some cases, improved
B. subtilis 1 g L–1
expression of toxic proteins may be
Interleukin 6 E. coli 7.5 mg mL–1 achieved in the C41 (DE3) and C43 (DE3)
P. pastoris 0.28 mg mL–1
strains. C41 (DE3) and C43 (DE3) strains
were derived from BL21 (DE3) [46]. These
Spodopterafrugiperda cells 0.001 mg mL–1 strains could tolerate the expression of
Glutamic acid decarboxy- E. coli 12.5 mg mL–1 toxic proteins. Prokaryotes and eukaryotes
lase differ in their codon usage, which may
P. pastoris 0.42 mg mL–1 result in lower yields of nonbacterial pro-
S. cerevisiae 0.46 mg mL–1 teins expressed in E. coli. The Rosetta
strains may help to overcome that [46].
Spodopterafrugiperda cells 0.02 mg mL–1

proteins with desired yields. However, this is not true when a 3.2 Expression Vector
complicated, difficult-to-express protein is produced [19, 35].
The difficulty of expressing complex, multimeric proteins with The expression vectors used today have multiple combinations
a high number of disulfide bonds is associated with E. coli. The of replicons, promoters, selection markers, multiple cloning
formation of correct disulfide bonds is necessary for attaining sites, and fusion protein/fusion protein removal strategies. Dif-
biologically active three dimensional conformations of many ferent E. coli expression vectors (pET, pQE, pBAD, and pUC)
recombinant proteins. The formation of erroneous disulfide with fusion tags such as hexahistidine (6x His-tag), maltose
bonds can lead to protein misfolding, protein inactivation, and binding protein (MBP) and glutathione S-transferase (GST)
aggregation into inclusion bodies [19, 35]. However, apart from are widely used for recombinant proteins [19, 39, 43]. A variety
these disadvantages, E. coli is chosen as the working horse for of promoters (lac, tac, trc, T5, T7, pL, rhaBAD, araBAD, PhyA,
recombinant protein production. and PhoA) have been successfully used for the production of
recombinant products with relatively low to very high expres-
sion levels [10, 23, 47]. Lactose, a sugar, can be used to provide

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 118
induction for the protein production. However, induction is removing the tags, solubility of the desired protein is unpre-
difficult in the presence of glucose as carbon sources. The dictable [61].
vectors that use lac or tac promoters are the pUC series
(lacUV5 promoter) and the pMAL series of vectors (tac pro-
moter). The tac and trc promoters are stronger than the lac 3.3 Small-Scale Expression of Recombinant
promoter and are induced by IPTG (isopropyl b-D-1-thio- Proteins
galactopyranoside). A thermosensitive lac repressor mutant is
available to induce protein expression by shifting temperature Generally small scale cultures are carried out in microtiter
instead of using IPTG. The T7 promoter system present in the plates, test tubes, or shake flasks for expression check and opti-
pET vectors is the most popular one for production of re- mization purposes. Sometimes, shake flask cultures are also
combinant proteins with high level expression of target protein used to produce recombinant proteins for initial characteriza-
[33]. In the majority of the cases, IPTG is used for the expres- tion of the biological activity.
sion of recombinant proteins under the T7 promoter. The
araPBAD promoter is present in the pBAD vectors and arabi-
nose is required for induction. The AraC protein has the dual 3.3.1 Influences on Protein Expression
role of repressor/activator [48]. The pL promoter is associated
with the systems that respond to pH or temperature. E. coli Process optimization at small scale is essential before proceeding
host strains containing the lcI857 protein can be induced by a to large scale production. Various criteria must be considered for
temperature shift from cultivation temperature of 28–30 C to optimization of conditions for the high-level expression of a
40–42 C. Recombinant proteins such as interferon, insulin, recombinant protein in E. coli. This optimization may help in
tumor necrosis factor and granulocyte colony stimulating fac- achieving high product yields and cost effective production of
tor have successfully been produced using this system [49]. recombinant proteins. The first step to enhance the expression of
Recently an auto-induction expression strategy (pET101/ recombinant proteins is to optimize the media composition.
D-TOPO expression vector) under the control of the T7 pro- Generally, complex media such as Luria broth (LB) is used
moter was utilized for the expression of human consensus for the expression of recombinant proteins at small scale cul-
interferon-alpha in E. coli. The protein yield was more in com- tures [62, 63]. However, various researchers used SOB, SOC,
parison to that of IPTG-induced systems [50]. The gene caps 2YT, GYT, MBL, Terrific broth (TB), Super broth (SB), M9 salt
promoter is induced by a downshift in temperature to 15 C. medium, and Enbase Flo media for the enhancement of protein
The pCold series of plasmids have a pUC118 backbone expressions [29, 58, 62, 64–70]. Defined and semi-defined
(a pUC18 derivative) with a cspA promoter [19, 51]. The media were also used by various researchers for the expression
Rhamex vectors, which include the regulatory genes RhaR and of recombinant proteins in E. coli. Generally, defined medium
RhaS have the ability to tune the level of expression of the contains glucose as a carbon source, ammonium sulphate,
target mRNA which may result in increased protein accumula- and/or ammonium chloride as nitrogen source along with
tion and protein solubility [46]. Antibiotics such as ampicillin, magnesium sulphate and phosphate salts [26, 64, 71–74]. Semi-
kanamycin, and tetracycline are used to prevent the growth of defined medium, however, contains glucose or glycerol as car-
plasmid-free cells. The cost of antibiotics and the dissemination bon source, yeast extract, and/or tryptone as complex nitrogen
of antibiotic resistance are major concerns at large-scale culti- source along with the magnesium sulphate and phosphate salts
vations. Efforts have been made to develop antibiotics-free [29, 31]. The medium’s components such as glycerol can
plasmid systems [52], but only limited ones are in use. increase the cell growth and recombinant protein yield.
Various tags are used in the expression of recombinant pro- Complex media such as TB and SB contain glycerol resulting
teins to enhance the solubility of proteins, help in detection by in higher protein concentrations in comparison with other
Western blot using commercial available antibodies, and one media used for recombinant protein expression in E. coli
step affinity purifications [53–55]. Tags are less likely to inter- [62, 65, 75–77]. However, in some cases, the LB medium also
fere for small peptides, but in some cases they may affect the resulted in higher protein concentrations in comparison with
biological activity of the fused chimeric protein [56]. Poly-Arg-, TB medium [30].
FLAG-, poly-His-, c-Myc-, S-, and Strep-II-tags are small pep- Cultivation using semi-defined medium resulted in better
tide tags [57]. Immobilized metal affinity chromatography is product yields in comparison with defined medium for the
used to purify poly-histidine-tagged proteins using nickel- expression of merozoite surface protein-1 [29]. Some nutrients,
charged nitrilotriacetic acid sepharose resins [30, 58]. Bound when present in excess concentration, can inhibit cell growth.
proteins are eluted by decreasing the buffer pH or increasing High concentration of glucose causes acetate accumulation and
imidazole concentrations. Expressed proteins in soluble and leads to a decrease in product concentration [74, 78, 79]. Diva-
refolded form tagged with GST [59] are purified using im- lent cation supplementation (magnesium sulphate) results in
mobilized glutathione. MBP fusion proteins were purified in higher cell growth and protein production. Using yeast extract
a single step using maltose under denaturing conditions [45]. in the cultivation medium, researchers reported significant
The GST showed the poorest solubility enhancement capabil- enhancement in protein yield. It helps to reduce secretion of
ities [59] and MBP resulted in more solubility-enhancing acetic acid during growth of E. coli, enhances the specific cellu-
properties. However, a large size of tags may lead to the lar yield, relieves cellular stresses and also helps in the utiliza-
erroneous assessment of protein solubility [60]. Smaller tags tion of acetic acid during carbon limitation [17, 23, 25, 80]. The
with enhanced solubility capability are essential because after phosphate salts in the media provided a buffering capacity to

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 119
prevent pH fluctuations that could adversely affect normal the use of statistical methods for optimization [90]. Tradition-
metabolic activity. Auto-induction media were also developed ally, the ‘‘one factor at a time’’ approach was used for the opti-
that contain glucose, lactose, and glycerol to avoid the biomass mization of expression conditions by changing one parameter
monitoring for timely addition of the inducer [81, 82]. and keeping the rest of the parameters constant. However, this
To initiate expression of recombinant proteins as IBs in approach takes long time and it does not provide information
E. coli, cultures are induced with inducers. Generally, induction about the interactions between various components or para-
is done during the mid or late log phase and even during the meters. Nowadays, various studies have been conducted to op-
stationary phase [83]. It is necessary to optimize the inducer timize media composition using full factorial design, fractional
concentration as well as the induction time for each new clone factorial design, taguchi orthogonal array, and response surface
so that both cell growth before induction and specific yield are methodology [45]. Design of experiment is used for these pur-
maximized, resulting in high volumetric yield of the protein. poses using Design Expert or Mini Tab software’s. The full fac-
By lowering post induction temperature, expression of some torial approach initiates experiments with little idea of the fac-
proteins resulted in the soluble form. The same has been tested tors affecting the expression of proteins. It provides a complete
for a number of proteins [50, 84]. High temperature can pro- set of experiments and utilizes the results of the same to give
mote cell growth and favor the aggregation reaction due to information about main and interaction effects. Alternative to
hydrophobic interactions. The optimum combination of post- the full factorial design, fractional factorial design is used when
induction temperature and duration of induction is still a trial- a large number of variables is available. Both designs have been
and-error process. The concentration of inducer and duration successfully used in order to enhance the recombinant protein
of induction also affect the product yield. The inducer concen- production using various factors such as media and its compo-
tration should be maintained slightly higher than the critical nents, pH, temperature, agitation, aeration, induction time,
concentration. IPTG (up to 1 mM) concentration did not affect inducer concentration, and biomass before induction [91–95].
the E. coli-specific growth rate or maximum cell concentration Response surface methodology (RSM) has emerged as a
[45]. The auto induction approach was also used successfully powerful technique of statistical optimization and is used by
for various recombinant proteins [50, 81]. many researchers ot improve recombinant protein expression
Aeration and agitation to provide oxygen may also play an [96]. RSM is generally used after conducting screening
important role in protein production. The simple way to increase experiments using factorial designs. Process parameters using
the aeration effect in shake flasks is to increase the agitation speed these methods were successfully optimized for recombinant
of the shaker. Most of the protein expression is carried out at a lipase, recombinant bromelain, human TNF (tumor necrosis
shaking speed range from 150 rpm to 220 rpm. Baffled flask may factor) alpha, human interferon beta and other proteins
also be used to improve the aeration and protein yield [85]. The [45, 92, 97, 98]. Taguchi orthogonal array was also applied to
culture volume to flask capacity ratio of one to four and more optimize process parameters for successful production of
could significantly improve the oxygen availability in the culture. recombinant botulinum neurotoxin using E. coli [94, 99]. The
At higher agitation rates, foaming problems occurred, which advantages of these methods include time saving by conducting
could decrease the availability of oxygen to the cells. Antifoam only a minimum number of experiments as well as to know the
can be added to suppress the foam. However, antifoam can affect interaction effect between media components and better exper-
the cell growth and final product yield. In bioprocesses effects of imental design. The detailed review of statistical optimization
antifoams on the oxygen transfer have been observed. The mass techniques used for recombinant protein expressions was
transfer coefficient and gas hold up within the culture media described elsewhere [45].
were negatively affected by silicone-based antifoam. However,
antifoams without silicone oil did not greatly affect the oxygen
transfer rate [86]. High throughput (HTP) cultivation technol- 3.4 Large-Scale Processes Using Bioreactors
ogy is widely employed today in the process development phase for Protein Production
or process optimization at small scale. The examples of HTP sys-
tems are miniature shaken vessel/wells or microtiter plate After successful expression of recombinant proteins at small
(0.2–4.0 mL), bubble column, or microplate-based mini-bio- scale, development of large scale cultivation processes is neces-
reactors (1.0–10 mL) and stirred mini-tank bioreactors (more sary. Today’s recombinant proteins are used for detection, pro-
than 10 mL) [87–89]. The detailed description of high through- tection, and treatment of various types of bacterial and viral
put cultivation technology was reviewed elsewhere [88]. The diseases as well as other diseases such as diabetes, Crohn’s dis-
small-scale optimization of process parameters may be per- ease, multiple sclerosis, osteoporosis, and cardiovascular dis-
formed using HTP cultivation technologies or shake flasks to ease [6, 34, 35]. Thus, large amounts of biomass are needed to
minimize cost as well as time before being scaled up for any bio- recover recombinant proteins for these purposes.
pharmaceutical products [87–89].

3.4.1 Fermentation Processes

3.3.2 Statistical Methods for Optimization
of Process Parameters The fermentation process is used to produce large amounts of
biomass for purification of the desired bioproduct. Before start-
Recent developments in the optimization of cultivation condi- ing the fermentation process, process parameters could be opti-
tions for the expression of recombinant proteins also include mized to achieve high yield. Generally, shake flask cultivations

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 120
are used to optimize process parameters [66, 100, 101]. The bon source and nitrogen source, respectively. In pH stat feeding
effect of medium composition, pH, temperature, agitation, cul- [108], feed is initiated when pH is increasing whereas in DO
ture to flask volume ratio, inducer concentration, and induc- stat feeding [109], the feed is started when the DO value is
tion time (log phase or stationary phase) have to be optimized increasing. Increases in the pH or DO value during cultivation
as described above. After successful optimization of process indicate nutrient depletion in the culture medium. Recombi-
parameters using shake flasks or microtiter cultures, these con- nant proteins such as Fab fragments, recombinant malaria pro-
ditions could be used for bioreactor cultivations. Bioreactors tein, L-isoleucine, Chikungunya envelope 2 protein, and anti-
provide the facility to control various parameters which affect cancer drugs have been produced using pH stat and DO stat
the cells growth as well as protein expression. These parameters feeding strategies [29, 77, 108, 109]. Exponential feeding using a
include pH, temperature, agitation speed, dissolved oxygen specific growth rate (m) is commonly used for the production
concentration, and most importantly, nutrient addition. Batch, of recombinant proteins with high yields [110]. To keep the
fed-batch, and continuous cultures [102, 103] are the modes of specific growth rate at a preset value, a feed-forward exponen-
fermentation which are employed for the cultivation of bacte- tial feeding is generally used. Nutrients needed to reach the set
ria. Batch are commonly used for E. coli cultures [17, 23]. value of m are calculated and added accordingly during the cul-
Using batch cultivation of E. coli, large amount of biomass can tivation process [111–113]. Fed-batch processes with exponen-
be obtained in short duration. However, batch processes are tial feeding resulted in more than 100 g L–1 of dry cell biomass
only used when small quantities of recombinant proteins are as well as more than 1 g L–1 of protein productivity [28].
needed for detection purposes or laboratory trial studies of vac- Recombinant proteins such as interferon [114], recombinant
cines or therapeutics. Fed-batch processes are used when large alcohol dehydrogenase [67], swine erysipelas [81], and re-
amounts of biomass are needed for purification purposes. combinant nitrilase [105] have been successfully produced
using exponential feeding strategies. Various other factors such
as oxygen supply, metabolic load, substrate inhibition, poor
3.4.2 Optimization of Fermentation Process Parameters heat transfer capacity, cultivation pH, and cultivation tempera-
ture also influence the cell growth and recombinant protein
Before proceeding to pilot scale or industrial scale productions, expressions during the fermentation processes [17]. The most
bioreactor key parameters, apart from shake flask-optimized commonly used pH for E. coli cell cultivation and protein
parameters like effect of dissolved oxygen concentration and expression is around 7.0, ranging from 6.5 to 7.5 [74]. How-
air flow rate (aeration), could be tested for further improve- ever, both pH 6.5 and pH 7.5 are suitable for cell growth and
ment of cell growth and protein production [62, 93]. The tradi- protein expression [74]. Oxygen plays an important role during
tional way or the recently used design of experiment (DOE) cultivations to maintain cell physiology and quality of recombi-
approach are utilized for optimizing these parameters. Using nant protein from E. coli. The effect of acetate, plasmid stabil-
the DOE approach, interaction effects between various para- ity, metabolic load, and oxygen during the fermentation proc-
meters could also be determined [104]. Media components can ess are described elsewhere [17].
reduce the cell growth when they are present in excess
amounts. Thus, simply increasing the concentrations of nu-
trients in batch culture media may not result high cell density. 3.4.3 Disposable Bioreactor Cultivation Techniques
A variety of proteins is produced in E. coli using batch fermen-
tation processes for different applications [22, 64, 73, 76, 105]. Development of disposable cultivation technologies as an alter-
However, the fed-batch process is most commonly used for the native to traditional bioreactor systems were also employed for
production of therapeutic or prophylactic proteins [17, 26, 81, cultivation of E. coli. They offer advantages such as manufac-
106, 107]. Fed batch processes are initiated with low concentra- turing flexibility, simplicity of operation, decreased incidence of
tions of media components and subsequently fed with media contamination, and lower costs for cleaning, sterilization, and
components when cells need to achieve large amounts of bio- validation. Disposable bioreactors are classified by mode of
mass and product concentrations. The simple indication of the power input as mechanically driven (wave-mixed, stirred, orbi-
time to start feeding is that the pH and DO (dissolved oxygen) tally shaken or vertically oscillating), pneumatically driven, or
value start to increase. Controlled feeding of medium compo- hybrid systems. During the last few years, disposable systems
nents should be used to avoid by-product formation like ace- like WAVE bioreactors were utilized for the production of high
tate formation. Acetate formation affects the cell growth as well value biopharmaceuticals in the area of mammalian cell cul-
as production performance [79]. ture. WAVE bioreactor were used to generate inoculum cul-
Various feeding strategies such as constant feed, pH stat, DO tures for E. coli production [115]. Using the BIOSTAT Culti-
stat, combination of pH and DO stat as well as exponential feed Bag rocking motion system, alcohol dehydrogenase was
were employed for enhancement of cell growth and protein successfully produced. Using this technique, it is possible to at-
productivity. High cell density (HCDC) fed-batch processes tain cell densities which are far above those of shake flasks and
have many advantages such as lower production cost, reduced typical for stirred tank reactors with the improved oxygen
culture volume and reduced investment in equipment. Gener- transfer rate [116]. One dimensional and two-dimensional
ally, concentrated feed medium containing a carbon source, a (CELL-tainer) disposable rocking bioreactors were also used
nitrogen source, magnesium sulphate, and trace metals are for the production of recombinant glutathione S-transferase-
used for the majority of recombinant proteins. Glucose or glyc- hCD83ext fusion protein in E. coli and performance was com-
erol and yeast extract or ammonium sulphate are used as car- pared with a stirred tank bioreactor. The culture performance

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 121
was similar in terms of soluble protein yield and inclusion body [121–124]. The important parameters which should be consid-
formation for all bioreactor systems. [117]. ered for scale-up of fermentation processes are shown in Tab. 4.
Constant impeller tip velocity is generally used for scale-up of a
process which deals with shear sensitive cells. Studies reported
3.4.4 Scale-Up of Laboratory-Scale Fermentation that the constant oxygen mass transfer coefficient criteria are
Processes appropriate for E. coli fermentation [125]. Oxygen mass trans-
fer coefficients could be changed with impeller design, agitation
The commercial use of recombinant protein-based therapeu- speed, air flow rate, and oxygen supply as well as bioreactor
tics increased rapidly in recent years. The key component of geometry [126].
the commercial success of any biopharmaceutical product is
the ability to achieve large-scale production. The scale-up pro- Table 4. Important parameters considered for scale-up of the
cess for recombinant proteins in E. coli should be targeted fermentation process [121–123, 125].
with consistency at each scale to achieve high productivity. As
the scale changes from laboratory scale to pilot scale or indus- Parameter Mathematical formula
trial scale, the parameters affecting expression of recombinant
–1
proteins will change, affecting the product yield [31, 106]. Impeller tip velocity [m s ] Vtip = 2pNDi
Thus, the major challenge of successful process scale-up is to Mixing time [s] Tm = (1.54V)/(NDi3)
maintain productivity similar to that of laboratory processes.
Different aspects like effect of multiple seed steps, inoculums Mass transfer coefficient [s–1] kLa =2 · 10–3(Pg/V)0.7(mG)0.2
transfer criteria, stability of strain with greater number of gen- Oxygen transfer rate [kgO2m–3h–1] OTR = kLa (CG–CL)
erations, tank configuration, aeration, dissolved oxygen, anti-
Power input [W] P = NPrN3Di5
foam addition, pH, agitation, temperature control/heat trans-
fer and removal, sterilization strategies across scales, Impeller Reynolds number Re = NDi2r/m
hydrostatic head differences across scales, vessel aspect ratio
Heat transfer coefficient [W m–2K–1] a = (dQ/dt)/AdT
differences, mixing, shear force variation, mass transfer limita-
tions on process intensity, and chiller water capacity must be
carefully considered when scale-up is carried out [17, 23].
During fermentative production of recombinant proteins by Developments in sensor technology resulted in a major
E. coli, cells are typically supplied with air or oxygen-enriched breakthrough in the fermentation process parameters. With the
air. Dissolved oxygen levels should be maintained low enough use of inbuilt sensors, online monitoring of pH, temperature,
to allow satisfactory growth and prevent fermentative metabo- agitation, dissolved oxygen concentration, air flow rate, and
lism, but not high enough to cause oxidative damage to the even cell density as well as nutrient concentrations are possible.
recombinant product. In case of expression of monoamine Constant DO concentration as a scale-up parameter for fer-
oxidase (MAO), the oxygen-enriched culture resulted in sig- mentation processes to produce organisms was also studied.
nificantly higher biomass as compared to aerated culture. This is a simple and easy way to scale-up a process from small
Oxygen seems to enhance the biomass levels, and thus, in- scale to large scale. Constant power per unit volume is gener-
crease the volumetric activity of the protein. However, total ally applied for scale-up of fungal and mammalian cell fermen-
MAO activity is higher for the oxygen-enriched culture while tation. At large scale productions, poor mixing cause’s improp-
the aerated culture resulted in higher specific activity [118]. er distribution of oxygen, nutrients, pH, heat, and metabolites
During the fermentation process, the addition of antifoam inside the fermentor which will ultimately reduce cell growth
could lower the kLa value and adversely affect the biomass and have negative impacts on both productivity and quality
yield and associated products. At large scale fermentation, dif- [122]. Thus, all these factors should be considered for proper
ferences in pH were also observed between the region close to scale-up of protein productions. Researchers have successfully
the addition point of the pH controlling agent and the bulk scaled up the production of allophycocyanin, multiepitope
environment where the pH is often measured [119]. For large malaria vaccine, staphylokinase, ricin A chain vaccine, Dengue
scale production, the process starts with shake flask cultures and Japanese encephalitis domain III protein, human granulo-
(50 mL to 500 mL) and laboratory fermentors (1 to 5 L) to cyte-macrophage colony-stimulating factor, and recombinant
optimize the process parameters which affect the biomass and aldolase in E. coli [31, 87, 106, 112, 127–130]. However,
protein yield [31, 67, 120]. The optimized conditions are then detailed process development studies must be done for each of
transfered to pilot scale (30 to 500 L) to check the consistency the recombinant proteins before starting large scale production.
and then to industrial scale (more than 1000 L) to start pro- The comparative performance of scale-up processes for E. coli-
duction to fulfill the product demand [23]. expressed FFL protein [125] and intracellular protein [123] are
The common criteria for scale-up from laboratory to pilot shown in Tab. 5. The flow chart for large scale fermentation
scale and from pilot scale to plant scale fermentation is to keep processes is shown in Fig. 1. The list of recently produced
one or more parameters constant between scales. These para- recombinant products from E. coli using bioreactors is given in
meters are impeller tip velocity (Vtip), oxygen transfer rate Tab. 6.
(OTR), oxygen mass transfer coefficient (kLa), power input per
unit volume, mixing time, and impeller Reynold’s number

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 122
Table 5. Comparative performance of scale-up of some E. coli- Frozen cell stock of microorganisms
expressed recombinant products.

Cultivation scales Scale-up parameters Yield Ref.

FFL protein Primary culture (Test tube/Shake flask culture, 10-50 ml)

Microwell plate (3 mL) kLa = 247 h–1 5.4 g L–1 [125]

–1 –1
75 L kLa = 247 h 5.3 g L

Intracellular protein Secondary culture (Shake flask culture, 100-1,000 ml)

–1
20 L Vtip = 4.36 m s 100 % [123]
–1
200 L Vtip = 4.39 m s 50 %
Lab scale fermentation, 1-20 l
Table 6. Some recently produced recombinant products in
E. coli using fermentation processes.

Product Cultivation conditions Ref. Pilot scale fermentation, 30-500 l

Aldolase Fed-batch process (50 L) [120]

Antibody fragment (Fab) Fed-batch process (10 L) [131]

TB proteins Batch process (3 L) [132]

Industrial scale fermentation, >1,000 l
Superoxide dismutase Fed-batch process (20 L) [42]

6-O-sulfotransferase-1 Fed-batch process (15 L) [133]

Pullulanase Fed-batch process (3 L) [134]

Glutathione S-transferase Batch process (4 L, 30 L) [74]

Culture broth for downstream purifications
(GST)
Figure 1. Flow chart showing large scale fermentation process-
Plasmid DNA (GALG20) Fed-batch process (14 L) [135]
es for production of recombinant products.
LZ-TRAIL Batch process (30 L) [136]

Human growth hormone Fed-batch (1 L) [137] contained large amounts of biomass (10–30 % w/v). This offers
enhanced productivity in the fermentor, but also puts a high
Antibody fragment (Fab) Fed-batch (20 L) [138] demand on the downstream purifications [141]. The various
Lysostaphin Batch process (5 L) [139] steps involved in the purification of recombinant proteins
expressed as inclusion bodies (IBs) are shown in Fig. 2.
Chikungunya E2 protein Fed-Batch process (6.6 L) [77]

Maltose-binding protein- Batch process (10 L) [70]

NAP 4.1 Cell Harvesting
Dengue virus 3 DIII protein Batch process (100 L) [130]
4.1.1 Centrifugation
Helicobacter multiepitope Fed-Batch process (50 L) [140]
vaccine Centrifugation is used to separate the cell biomass from the fer-
mentation broth [142–144]. Batch centrifuge is used for bio-
mass recovery from culture broth obtained in shake flasks or
laboratory fermentors. In case of large scale (pilot or plant
4 Recovery and Purification scale) production, continuous centrifuges are used for cell
of Recombinant Proteins removal. Mainly three types of centrifuges exist: tubular bowl,
disc stack, and basket centrifuges are generally employed for
After successful expression of recombinant proteins from biomass recovery. Tubular bowl centrifuges are capable of high
E. coli, there will be a need for a purification process to obtain speed centrifugation up to 20 000 g forces and provision of
highly purified product for therapeutic or vaccine uses. Down- cooling is very useful for protein purification. Disc type is the
stream processing of recombinant proteins plays an important most commonly used type and often used continuously [141].
role and contributed about 50–80 % to the costs of the total It may handle feeds with up to 30 % (v/v) solids. Basket centri-
production [16]. Thus, the aim of purification process develop- fuges are also known as filtering centrifuges and are used for
ment is to achieve higher levels of purity with minimum time low g forces. They are not effective for particles below 5 mm
and less cost. The feedstock resulting from HCDC processes and loadings of less than 5 % (v/v) [102, 144].

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 123
 Fermentation/Shake flask culture and viruses from the fermentation broth. In case
of extracellular expressed proteins, ultrafiltrations
are also used for concentration and purification.
Ultrafiltration membranes are usually 1–20 nm in
Centrifugation/Microfiltration to recover cell biomass size and their protein retention is very high
[143]. Buffer exchange and salts removals are car-
ried out using cross-flow ultrafiltration under dia-
filtration mode.

Cells disruption using sonicator or bead mill or high pressure homogenizer

4.2 Cell disruption/Cell Lysis

Recombinant proteins expressed in E. coli resulted

Centrifugation to recover IBs; IBs wash and repeated centrifugation in the either intracellular (cytoplasmic space or
periplasmic space) or extracellular localization of
the proteins as analyzed by sodium dodecyle sul-
phate poly acrylamide gel electrophoresis. Cell dis-
ruption is a process used for disrupting or lysing
IBs solubilization and centrifugation, microfiltration
the cells in order to recover inclusion bodies in case
of intracellular expression of proteins. The most
commonly used techniques for cells disruptions are
high pressure homogenization, bead milling, ultra-
sonication, osmotic shock, freeze thaw, enzymatic
Refolding using dilution or on-column or dialysis and diafiltration
lysis, and chemical lysis [102, 144]. For large scale
processes, the high pressure homogenizer and the
bead mill are commonly used for cells disruption.
The high-pressure homogenizer is generally used
Chromatographic purifications for bacteria, animal or plant cells including spore.
Shear force generated in this treatment are suffi-
cient to completely disrupt many types of cells. In
bead mills, the cell suspension is pumped in a hori-
zontal grinding chamber filled with about 80 %
beads. High shearing and impact forces from beads
Pure and biologically active protein break the cell walls. The bead mill is mostly used
for the disruption of bacterial cells, yeast, algae,
Figure 2. Process steps for purification of recombinant products expressed as and filamentous fungi [144]. Ultrasonicators are
IBs in E. coli. used to disrupt the bacterial cells. An electronic
generator is used to generate ultrasonic waves. Heat
dissipation is an important problem in this process.
4.1.2 Filtration Ultrasonication is generally used at the laboratory scale for cells
disruption [102].
Filtration is the cost effective technique for the separation of
cells from the fermentation broth. Various filtration methods
used for the removal of cells are surface filtration, depth fil- 4.3 Solubilization of Proteins Expressed
tration, centrifugal filtration, cross flow filtration, and rotary in Inclusion Bodies
drum filtration. Rotary drum vacuum filtrations are most
commonly used at large scale for filamentous fungi and yeast High-yield expression of recombinant proteins in E. coli often
cells [144]. Cell separation from the fermentation broth is accumulates large amounts of unfolded and misfolded proteins
also achieved by tangential flow or cross-flow microfiltration. in the cytoplasm as insoluble aggregates and inclusion bodies
Cross-flow microfiltration is used to separate particles rang- [21, 145, 146]. To recover biologically active protein for vaccine
ing from 0.1 to 10 mm. Three types of microporous (poly- or therapeutic applications, inclusion bodies must be solubi-
meric or ceramic) membranes exist: hollow fibers, flat sheets, lized and refolded [147, 148]. Many commercial recombinant
and spiral wound are used for these purposes. This technique products such as insulin, interferons, interleukins, strepto-
replaced centrifugation as well as vacuum filtration for var- kinase, and Fc fusion proteins are produced in IBs due to high
ious applications. Tangential flow or cross-flow microfiltra- yield. The challenge for IB proteins is to achieve biological
tion is mostly used for E. coli cells [102, 142, 143]. Cells con- activity using refolding processes [16]. Refolding of recombi-
centration is carried out using cross-flow microfiltration nant proteins requires in-depth knowledge of the refolding
under diafiltration mode. Cross flow ultrafiltration is used processes. Generally, the recombinant proteins expressed as IBs
for separation of proteins, enzymes, pyrogens, cell debris, are solubilized by the use of high concentrations (6–8 M) of de-

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 124
naturing or chaotropic agents such as urea, guanidine hydro- refolding methodologies using affinity chromatography
chloride (Gd-HCl), and SDS (sodium dodecyl sulfate) along [31, 148, 151, 156], ion-exchange chromatography [32, 152, 157],
with dithiothreitol (DTT), b-mercaptoethanol, or cysteineas as hydrophobic interaction chromatography [158] and size exclu-
reducing agents [149]. Researchers also tried solubilization of sion or gel filtration chromatography [159, 160] have been suc-
IBs using buffer containing low concentrations of a chaotropic cessfully done for various recombinant proteins. Protein aggre-
agent with extreme pH or temperature [21, 150]. Thus, the gation by intermolecular interaction is minimized using
selection of the solubilizing reagent greatly affects the yield of chromatographic methods of refolding because the folding
the refolding process and the overall process costs. The detailed molecules bind to the support matrix and isolated later. Use of
review of IBs solubilization for E. coli expressed proteins were the proper buffer solutions (without denaturing agent) and
described elsewhere [21, 146]. optimization of elution conditions were simultaneously carried
out to obtain pure and biologically active recombinant protein.
Simultaneous refolding and purification of IB proteins using
4.4 Refolding Processes to Recover Biologically affinity chromatography have been reported for interferon,
Active Proteins fibroblast growth factor, interleukin, and Japanese encephalitis
domain III protein [31, 148, 151, 156]. Reduced lysozyme at
To recover biologically active protein, it is necessary to remove very high concentration (9 mg ml–1) has been successfully
the chaotropic agent to attain its native state. For this purpose, refolded into the bioactive form with almost 100 % recovery
the solubilized proteins are refolded to their original state by using immobilized liposome chromatography [25]. Since this
removing the denaturing agents. Initial chromatographic puri- technique is suitable for reducing aggregation of the protein,
fication of the recombinant proteins is generally performed the maximum recovery of the refolded protein may be easily
before refolding (under denaturing conditions) or after refold- achieved. Similarly, recombinant interferon-gamma using
ing steps. The refolding step with chromatographic purification hydrophobic interaction chromatography [158], recombinant
is the best choice because some of the high molecular weight human tumor necrosis factor-alpha and recombinant human
aggregates as well as contaminants can be co-purified in a sin- (Pro) renin using ion exchange chromatography [32, 157], and
gle step. Various refolding techniques such as rapid dilution, recombinant human granulocyte colony stimulating factor
dialysis, diafiltration using ultrafiltration membrane, on-col- using size exclusion chromatography [159, 161] have been suc-
umn refolding using affinity, ion exchange, hydrophobic inter- cessfully refolded for therapeutic studies. Protein refolding can
action or size exclusion chromatography, high hydrostatic pres- also be achieved by the use of diafiltration [149] and dialysis
sure technology, and light scattering [2, 16] are commonly used [162] using ultrafiltration membranes. Large volumes of solubi-
for renaturation of denatured proteins. Rapid dilution of the lized protein solution can be used in refolding using the diafil-
solubilized protein directly into the refolding buffer is a com- tration method. It was found that the use of chromatography
monly used method for renaturation of recombinant proteins. techniques for protein refolding resulted in higher yields
The limitation of the rapid dilution method includes scale-up compared to the dilution method [148]. As on-column chro-
issue and low protein yield. Rapid dilution of large amounts of matography refolding offers many advantages such as buffer
protein requires a large-size refolding vessel, large quantities of exchange, protein refolding and recovery of pure protein, large
refolding buffer as well as additional ultrafiltration steps for scale refolding of IB proteins can be carried out using these
concentration after protein refolding. All of these issues add methods [148, 156].
high costs to the recombinant protein production
[25, 148, 151]. Pulse renaturation, where small amounts of
solubilized protein are added to the refolding buffer at certain 4.5 Purification of Recombinant Proteins Using
time intervals, reduces the volume of refolding buffers. Dilu- Chromatographic Processes
tion-type renaturation processes have been successfully used to
refold recombinant products such as gamma interferon, human Purification is an important step in the bioprocessing of
growth hormone, human granulocyte colony stimulating fac- recombinant proteins to achieve the desired level of purity.
tor, L-asparaginase, and lysozyme [28, 152, 153]. Refolding of Protein purification using chromatographic processes to
biologically active recombinant proteins with little aggregation achieve high purity plays an important role in the production
has also been achieved by using additives in the refolding buf- costs. The final application of recombinant proteins will be
fers [16, 154]. These additives are glycerol, mannintol, CHES, used to decide the required purity levels. It is very easy to use
L-arginine/HCl, low concentration of urea (1–2 M) or guani- tags to separate the desired protein from a mixture of contami-
dine hydrochloride, oxidized and reduced glutathione, and de- nant proteins. To purify recombinant proteins expressed in
tergents [146]. Researchers also reported the positive effect of E. coli cytoplasm, the harvested cells are lysed, which naturally
glycerol and L-Arginine/HCl in reducing aggregation for hu- contributes to the complex nature of protein mixtures. The
man gamma interferon [148, 155], recombinant human bone advantage of IB-expressed (intracellular) proteins as compared
morphogenic protein-2 and fibroblast growth factor [151]. to secretory (extracellular) proteins is the volume to be proc-
Recently, a light scattering approach has been reported for the essed for further purifications as well as high productivity. Gen-
refolding process, but the true potential in protein refolding is erally, extracellular expressed proteins exist in large volumes in
yet to be firmly established [2]. the culture medium. Based on the need of purity levels, various
Refolding of recombinant proteins is also carried out using chromatographic techniques such as affinity chromatography,
on-column chromatographic refolding techniques. On column ion exchange chromatography, hydrophobic interaction chro-

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 125
matography, and size exclusion chromatography are used to achieved [54, 142, 143]. The packing material (affinity matrix)
obtaine the finished product for vaccine and therapeutic uses such as agarose or sepharose must be inert and easily modified.
[143]. The list of some E. coli-expressed recombinant products In the affinity chromatography process, the desired protein
purified using different chromatographic processes are given in binds to the ligands on the affinity matrix until elution is car-
Tab. 7. ried out. The elution of bound protein is carried out using high
salt concentration (0.25 to 0.5 M imidazole), strong chelating
agents and/or low pH (<4.5). Before elution, washing is carried
4.5.1 Affinity Chromatography out to remove unbound proteins. Large numbers of different
fusion partners capable of selective interaction with the ligand
Affinity chromatography is a simple technique used to purify immobilized onto the affinity chromatography matrix have
recombinant proteins with high purity in a single step. Fusion been described in detail [7, 54]. The use of polyhistidine tags
can be provided to the sides of the target gene depending on (6x) has been demonstrated in a wide range of host cells
the specific application. For the majority of proteins, affinity including E. coli, S. cerevisiae, P. pastoris, insect cells as well as
chromatography is used as laboratory-scale purification tech- in mammalian cells [148]. Process optimization parameters for
nique. However, its use at large scale can contribute the costs of the affinity chromatography include pH of buffers, dynamic
the end product. Affinity chromatography needs a biospecific binding capacity, salt (imidazole) concentration, flow rates,
ligand covalently attached to the affinity matrix. Using this bead height, sample load, gradients, and residence, time.
technique, high purity of a protein sample can be easily Parameter testing could be used to enhance protein yield as

Table 7. Some E. coli expressed recombinant products purified using chromatographic processes and their various applications.

Product Chromatographic techniques used Applications Ref.

Protein G Affinity, gel filtration and ion exchange Antibody separation [27]

Dengue and West Nile virus DIII protein Affinity membrane Diagnostics [162]

JE virus NS1 protein Affinity Diagnostics [55]

Chikungunya virus E2 protein Affinity Diagnostics [77]

Anthrax protective antigen Ion exchange and hydrophobic interaction Vaccines [163]

Anthrax protective antigen Affinity Vaccines [164]

ALT-2 filarial protein Hydrophobic interaction Vaccines [30]

Anthrax protective antigen Ion exchange membrane Vaccines [165]

Malaria vaccine (PfMSP-1) Affinity and ion exchange Vaccines [29]

Ricin A-chain protein Affinity Vaccines [166]

Dengue virus 3 DIII protein Affinity and ion exchange Vaccines [130]

Interleukin-24 Affinity, gel filtration and ion exchange Therapeutics [167]

Human interferon alpha Ion exchange and hydrophobic interaction Therapeutics [168]

Ebola virus VP35 protein Affinity Therapeutics [169]

Human Fab Affinity Therapeutics [170]

Human LAT1 transporter Affinity Therapeutics [171]

Streptokinase Affinity Therapeutics [172]

Fibroblast growth factor 19 Affinity and ion exchange Therapeutics [173]

Human growth hormone Affinity and ion exchange Therapeutics [137]

Amide hydrolase Ion exchange, hydrophobic interaction and gel filtration Therapeutics [174]

Human (Pro) renin receptor Ion exchange Therapeutics [157]

Human scFv anti-CEACAM1 Ion exchange and affinity Therapeutics [175]

Human consensus interferon-alpha Ion exchange Therapeutics [50]

IgE-hyporeactive variant Affinity and size exclusion Therapeutics [20]

HECT domain Affinity, ion exchange and gel filtration Therapeutics [176]

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 126
well as purity. Recombinant proteins such as recombinant West Recombinant proteins such as Anthrax protective antigen,
Nile virus domain III protein, human LAT1 transporter, fibro- human interferon alpha, Brugia malayi abundant larval tran-
blast growth factor 19, recombinant protein G, and ricin A script, and amide hydrolase expressed in E. coli have been puri-
chain protein expressed using E. coli have been purified by fied using hydrophobic interaction chromatography
affinity chromatography [27, 128, 162, 171, 173]. [30, 163, 168, 174].

4.5.2 Ion Exchange Chromatography 4.5.4 Size Exclusion or Gel Filtration Chromatography

Ion exchange chromatography (cation and anion exchange) is The size exclusion or gel filtration chromatography (SEC) is
used for the purification of proteins expressed in E. coli. It has based on the molecular mass of the molecules [159]. Large size
the capacity to recover almost any type of charged molecules proteins are excluded from the matrix, whereas intermediate
such as large sized proteins, small size nucleotides, and amino size protein can partly enter and only small size protein can
acids [177]. Ion exchange resins (solid matrices containing freely enter the matrix resin. It is also used to determine the
fixed ionogenic groups) are used to capture the target protein. approximate molecular weights of the protein. The selection of
The two types of ion exchangers such as cation and anion the suitable SEC column depends on the molecular weight and
exchangers having negatively and positively charged functional physical properties of the recombinant proteins to be purified
groups, respectively, were used in ion exchange chromatogra- or separated. This technique is used for purification of small
phy [143]. The commonly used commercially available ion amounts of recombinant proteins. Process optimization
exchange resins are sulpho propyl (SP), Q sepharose, Uno- parameters include size of beads, sample load volume, pH, flow
sphere S, Unosphere Q or diethyl aminoethyl (DEAE) sephar- rate, gradients and bead height that could be tested to enhance
ose. The net surface charge of the protein is responsible for the the protein yield as well as purity. The IgE-hyporeactive mole-
adsorption on the ion exchange resin. The separation of pro- cule and single chain variable fragment against type 1 insulin-
teins in this chromatography is highly dependent on pH and like growth factor receptor expressed in E. coli have been puri-
salt concentration of the buffers. The optimum pH range is fied using size exclusion or gel filtration chromatography
within 1 pH units of the isoelectric point for many proteins. [20, 181]. A single method or a combination of two or three
Elution with increasing salt concentration or increasing/ methods may be used to purify recombinant proteins depend-
decreasing pH of the buffer is the most commonly used strat- ing on the final applications of end products (Tab. 7).
egy in the ion exchange chromatography [26, 29, 142, 178]. The
higher the net charge of protein, the higher ionic strength is
required for elution of the protein. Thus, for process optimiza- 4.5.5 Membrane Adsorption Chromatography
tion, the pH of the buffer is selected so that large differences in
net charge between the sample components are created. Proc- Membrane adsorption chromatography has emerged as a
ess optimization parameters for ion exchange chromatography, promising alternative to conventional column chromatography
that could be tested to enhance the protein yield as well as for the purification of antibodies. It offers volumetric flow rate
purity, include effect of media (resins), pH, salt concentration, independent dynamic capacities, higher separation speed, and
bed height, flow rate, sample load, gradients, and residence easier scale-up. Membrane chromatography has been applied
time. Variety of proteins such as human growth hormone, successfully at lab scale and large scale to remove trace impur-
recombinant human (Pro) renin, recombinant Japanese ence- ities (DNA, virus, host cell proteins). Ion-exchange and affinity
phalitis (JE) virus domain III protein, malaria vaccine, human membrane chromatography were used by researchers to purify
consensus interferon-alpha, and rhGM-CSF expressed in E. coli biological molecules [182]. PEGylated lysozyme [183], human
have been purified using ion exchange chromatography serum albumin, immunoglobulin G, immunoglobulin M,
[29, 31, 50, 129, 137, 157, 179] monoclonal antibodies, Anthrax protective antigen protein
[165], recombinant dengue, and West Nile virus domain III
proteins [162] were successfully purified using membrane
4.5.3 Hydrophobic Interaction Chromatography chromatography. However, membrane adsorbers have had low-
er per volume protein binding capacities than resin columns.
Hydrophobic interaction chromatography (HIC) is used to
separate the recombinant proteins on the basis of relative
hydrophobicity. The recombinant proteins may be captured at 4.6 Scale-up of Chromatographic Processes
an ionic strengths of 1 to 1.5 M from the protein solution onto
matrix and elution is carried out by decreasing the ionic The next challenge is the scale-up of small scale chromato-
strength. HIC is used at small scale as well as in large scale graphic purifications to achieve large amounts of products.
purification of recombinant proteins, hormones, and enzymes Large-scale chromatography operations continue to occupy a
[30, 180]. Process optimization parameters for hydrophobic key position in the overall strategy for the downstream process-
interaction chromatography, that could be tested to enhance ing and purification of protein products for therapeutic use. In
protein yield as well as purity, include hydrophobicity of the commercial manufacturing, a requirement exists to increase
ligand, gradient volume, temperature, effect of media, salt con- the scale of the chromatography operations, which are typically
centration, sample load, bead height, and residence time. developed and optimized in small-scale experiments [184].

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 127
Chromatography scale-up is based on increasing column dia- lowed by centrifugation to recover the supernatant. Then the
meter to obtain a greater column volume. The column volume supernatant solution was further processed chromatographi-
(CV) of load, wash and elution remains unchanged. Superficial cally to achieve high levels of purity. The protein expressed in
velocity or linear flow rate (cm h–1) should remain constant the periplasmic space is less contaminated with host cell
between the scales. Factors like chromatography bead chemis- proteins compared to protein expressed in the cytoplasmic
try and size, load pH, load conductivity, target protein concen- space. Researcher were successfully produced many proteins
tration (mg mL–1), load volume loaded (CV), total protein such as Fab fragments, full length glycosylated antibodies,
loaded (mg mL–1), linear flow rate, wash buffer volume (CV), human growth hormone, scFvs, granulocyte colony-stimulat-
elution buffer volume (CV) and gradient volume (CV) should ing factor (G-CSF), insulin-like growth factor, human consen-
be kept constant across scales. The scaled up process methodol- sus interferon-alpha and interleukin-24 in periplasm of E. coli
ogy should follow the small scale procedure including any pre- [10, 50, 137, 186–188]. However, expression optimization is
column sanitization and equilibration buffer column volume required for most of the protein to lead to periplasmic protein
used. For proper scale-up of chromatography processes, other secretion. Secretion of recombinant proteins into the culture
key parameters like HETP (height equivalent of a theoretical medium offers various advantages such as less endotoxin
plate) may be considered to ensure that the small scale separa- content, low host cell protein content, ease in purification
tion efficiency is retained upon scale-up. Successful scale-up is (minimum steps), no need of additional cell disruption, and
done if the HETP of the total full scale system equals or is refolding steps. Different recombinant products such as inter-
betters than the HETP of the small scale system. Evaluation of leukin-6, Fab, scFv, cholera toxin, OsmY-Leptin, and domain
HETP of a process scale chromatography column is also an antibody were secreted into the culture medium [10, 23, 189].
index of its continued effective performance [185]. Various However, protein secretion into the culture medium remains a
studies have been successfully carried out to scale-up chroma- challenge in E. coli and results in relatively low protein yields
tographic purification processes for E. coli-expressed proteins compared to IB proteins.
such as recombinant Bacillus anthracis protective antigen,
recombinant allophycocyanin, staphylokinase, recombinant
ricin vaccine, human scFv antibody as well as dengue virus type
2 and 3 DIII proteins [31, 106, 127, 128, 142, 163, 175]. By con-
sidering these key issues, bioactive products
with high yield and purity may be possible Fermentation/Shake flask culture
at large scale using E. coli for biopharma-
ceutical and medical applications.

Centrifugation/Microfiltration
4.7 Recovery and Purification
of Solubleand Extracellular
Expressed Proteins

Expression of recombinant proteins in

E. coli at reduced growth temperature and
improved folding environment could en-
hance protein solubility for many products
Extracellular Periplasmic
such as human growth hormone, interfer-
on, and Fab fragments. Soluble expression
of recombinant proteins in E. coli cyto-
plasm can still lead to protein misfolding,
relatively low protein yield, susceptibility to
Product capture Cells disruption and centrifugation
proteolytic activities and labor intensive
downstream purifications [33]. Due to
these reasons, soluble periplasmic expres-
sion is the choice for protein production in
E. coli. An oxidizing environment for the Chromatographic purifications Chromatographic purifications
proper folding of the proteins is created by
the Dsb protein family which introduces
disulfide bonds in the periplasm of E. coli
[10]. The various steps involved in purifica-
tion of periplasmic and extracellular ex- Pure and biologically active protein Pure and biologically active protein
pressed proteins are shown in Fig. 3. The
purification of periplasmic proteins was Figure 3. Process steps for purification of recombinant products expressed in the extra-
simply performed by cell disruption fol- cellular and periplasmic space in E. coli.

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 128
5 Conclusion
Nagesh K. Tripathi studied
Chemical Engineering and
Nearly 20 % of the currently approved recombinant therapeutic
graduated from The Institution
proteins are produced in E. coli. We have reviewed the
of Engineers, Kolkata, India.
recombinant protein production in E. coli. Although significant
He received his doctorate in
improvements and a wide range of expression systems are
Chemical Engineering from
available for recombinant proteins, E. coli is still the preferred
the National Institute of Tech-
microbial cell factory. E. coli is a suitable host for the expression
nology, Rourkela, in 2013. He
of therapeutics. Various host vector combinations along with
did research for his Ph.D.
process parameters have been used to obtain the desired
thesis on the production, puri-
expression of recombinant proteins. Batch and fed-batch culti-
fication and characterization of
vations are used for high yield production of recombinant pro-
recombinant viral proteins. As
teins. Various factors affecting the performance of fermentation
a member of the working
processes have been described to obtain large amounts of bio-
group for the bioprocess scale-
mass for downstream purifications. An increasing number of
up facility, he develops processes for the production of
recombinant protein-based diagnostics, vaccines, and thera-
recombinant proteins. Since 2010, he is scientist at the
peutics will have to be produced in the coming years. E. coli
Defence Research and Development Establishment,
will aid in the more efficient production of these biopharma-
Gwalior. His current research focuses on process develop-
ceuticals. Major constraints of recombinant protein production
ment and scale-up of recombinant protein production.
in E. coli are the secretion of inhibitory products, effects of the
medium, effects of the dissolved oxygen concentration, feeding
strategies, and lowering of the specific cellular protein yield.
Strategies to overcome these constraints have been discussed so Symbols used
that sufficient amounts of protein will be produced to fulfill the
product demand. The use of suitable refolding processes could m [Ns m–2] dynamic viscosity
enhance the product quality for therapeutic and vaccine appli- mG [m s–1] superficial gas velocity per cross
cations. Based on the end use of the final product, it can be sectional area of the bioreactor
easily purified by using a single one or a combination of chro- A [m2] surface area
matographic techniques. By careful selection of process para- CG [kgO2 m–3] oxygen saturation concentration in the
meters for production and purification of recombinant pro- gas phase
teins, considerable improvements can be achieved with respect CL [kgO2 m–3] measured oxygen saturation
to productivity and product quality. Furthermore, the use of concentration in the liquid phase
statistical techniques can also reduce the time for process opti- Di [m] impeller diameter
mization. The successful development of efficient production dQ/dt [W] heat transfer
and purification strategies will strengthen the protein biopro- kLa [s–1] mass transfer coefficient
cessing by significantly reducing time and production costs N [s–1] agitation speed
and expedite the delivery of recombinant proteins for a wide NP [–] impeller power number
range of applications. OTR [kgO2m–3h–1] oxygen transfer rate
P [W] power input
Pg/V [W kg–1] power per working volume under
Acknowledgment aeration condition
Re [–] Reynolds number
The author thanks the Director of the Defence Research and T [K] temperature difference
Development Establishment (DRDE), Gwalior, for his keen Tm [s] mixing time
interest and support of this study. V [m3] liquid volume
Vtip [m s–1] Impeller tip velocity
The author has declared no conflict of interest.
Greek letters
a [W m–2K–1] heat transfer coefficient
r [kg m–3] density of the medium

Abbreviations
DEAE diethyl aminoethyl
DO dissolved oxygen
DOE design of experiment
DTT dithiothreitol
G-CSF granulocyte colony-stimulating factor

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 129
GST glutathione S-transferase [21] S. M. Singh, A. K. Panda, J. Biosci. Bioeng. 2005, 99,
HCDC high cell density 303–310.
HETP height equivalent of a theoretical plate [22] J. Sereikaite, A. Statkute, M. Morkunas, K. Radzevicius,
HIC hydrophobic interaction chromatography V. Borromeo, C. Secchi, V. A. Bumelis, Appl. Microbiol.
HTP high throughput Biotechnol. 2007, 74, 316–323.
IB inclusion body [23] C. J. Huang, H. Lin, X. Yang, J. Ind. Microbiol. Biotechnol.
LB Luria broth 2012, 39, 383–393.
MAO monoamine oxidase [24] H. Nausch, J. Huckauf, R. Koslowski, U. Meyer, I. Broer,
MBP maltose binding protein PLoS One 2013, 8, e54933.
OTR oxygen transfer rate [25] A. K. Panda, Adv. Biochem. Eng. Biotechnol. 2003, 85, 43–93.
RSM response surface methodology [26] R. Khalilzadeh, S. A. Shojaosadati, N. Maghsoudi, J. Moham-
SB super broth madian Mosaabadi, M. R. Mohammadi, A. Bahrami,
SDS sodium dodecyl sulfate N. Maleksabet, M. A. Nassiri-Khalilli, M. Ebrahimi, E. H. Na-
SEC size exclusion or gel filtration chromatography derimanesh, J. Ind. Microbiol. Biotechnol. 2004, 31, 63–69.
SP sulpho propyl [27] H. C. Zhang, J. Yang, G. W. Yang, X. J. Wang, H. T. Fan,
TB terrific broth Mol. Med. Rep. 2015, 12, 3132–3138.
TNF tumor necrosis factor [28] S. M. Singh, A. Sharma, A. K. Upadhyay, A. Singh, L. C.
Garg, A. K. Panda, Protein Expression Purif. 2012, 81, 75–82.
[29] S. Mazumdar, S. Sachdeva, V. S. Chauhan, S. S. Yazdani, Bio-
References process Biosys. Eng. 2010, 33, 719–730.
[30] S. Bhuvanesh, A. Chakkaravarthi, K. Perumal, R. Subrama-
[1] C. J. Huang, A. J. Lowe, C. A. Batt, Appl. Microbiol. Biotech- nian, J. Ind. Microbiol. Biotechnol. 2010, 37, 1053–1059.
nol. 2010, 87, 401–410. [31] N. K. Tripathi, A. Shrivastava, K. C. Biswal, P. V. Rao, Appl.
[2] A. Basu, X. Li, S. S. J. Leon, Appl. Microbiol. Biotechnol. Microbiol. Biotechnol. 2012, 95, 1179–1189.
2011, 92, 241–251. [32] Y. Wang, W. Ren, D. Gao, L. Wang, Y. Yang, Q. Bai, Biomed.
[3] S. Aggarwal, Nat. Biotechnol. 2011, 29, 1083–1089. Chromatogr. 2015, 29, 305–311.
[4] J. Nielsen, Bioengineered 2013, 4, 207–211. [33] K. Graumann, A. Premstaller, Biotechnol. J. 2006, 1,
[5] G. Walsh, Biopharm. Int. 2013, 26, 54–56. 164–186.
[6] N. A. Baeshen, M. N. Baeshen, A. Sheikh, R. S. Bora, M. M. [34] M. N. Baeshen, A. Al-Hejin, R. S. Bora, M. M. Ahmed, H. A.
M. Ahmed, H. A. Ramadan, K. S. Saini, E. M. Redwan, Ramadan, K. S. Saini, N. A. Baeshen, E. M. Redwan, J. Mi-
Microb. Cell. Fact. 2014, 13, 141. crobiol. Biotechnol. 2015, 25, 953–962.
[7] B. Nilsson, G. Forsberg, T. Moks, M. Hartmanis, M. Uhlén, [35] S. K. Gupta, P. Shukla, Crit. Rev. Biotechnol. 2015. DOI:
Curr. Opin. Struct. Biol. 1992, 2, 569–575. 10.3109/07388551.2015.1084264
[8] L. F. Vallejo, U. Rinas, Microb. Cell. Fact. 2004, 3, 11. [36] A. Berlec, B. Strukelj, J. Ind. Microbiol. Biotechnol. 2013, 40,
[9] S. Sahdev, S. K. Kattar, K. S. Saini, Mol. Cell. Biochem. 2008, 257–274.
307, 249–264. [37] J. H. Choi, K. C. Keum, S. Y. Lee, Chem. Eng. Sci. 2006, 61,
[10] K. J. Jeong, S. H. Jang, N. Velmurugen, Biotechnol. J. 2011, 6, 876–885.
16–27. [38] O. Spadiut, C. Herwig, Pharm. Bioprocess. 2013, 1, 283–295.
[11] D. Porro, B. Gasser, T. Fossati, M. Maurer, P. Branduardi, [39] G. J. Gopal, A. Kumar, Protein J. 2013, 32, 419–425.
M. Sauer, D. Mattanovich, Appl. Microbiol. Biotechnol. 2011, [40] M. Berkmen, Protein Expression Purif. 2012, 82, 240–251.
89, 939–948. [41] M. Merlin, E. Gecchele, S. Capaldi, M. Pezzotti, L. Avesani,
[12] D. Mattanovich, P. Branduardi, L. Dato, B. Gasser, M. Sauer, Biomed Res. Int. 2014, 136419.
D. Porro, Methods Mol. Biol. 2012, 824, 329–358. [42] K. Marisch, K. Bayer, M. Cserjan-Puschmann, M. Luchner,
[13] M. Ahmad, M. Hirz, H. Pichler, H. Schwab, Appl. Microbiol. G. Striedner, Microb. Cell Fact. 2013, 12, 58.
Biotechnol. 2014, 98, 5301–5317. [43] C. Wang, J. Zhang, H. Wu, Z. Li, Q. Ye, J. Biotechnol. 2015,
[14] S. Jana, J. K. Deb, Appl. Microbiol. Biotechnol. 2005, 67, 214, 63–68.
289–298. [44] K. Lehmann, S. Hoffmann, P. Neudecker, M. Suhr, W. M.
[15] A. L. Demain, P. Vaishnav, Biotechnol. Adv. 2009, 27, Becker, P. Rosch, Protein Expression Purif. 2003, 31,
297–306. 250–259.
[16] A. S. Rathore, P. Bade, V. Joshi, M. Pathak, S. K. Pattanayek, [45] P. Papaneophytou, G. Kontopidia, Protein Expression Purif.
J. Chem. Technol. Biotechnol. 2013, 88, 1794–1806. 2014, 94, 22–32.
[17] N. K. Tripathi, K. Sathyaseelan, A. M. Jana, P. V. L. Rao, [46] S. Bashiri, D. Vikström, N. Ismail, Biopharm. Int. 2014, 28,
Def. Sci. J. 2009, 59, 137–146. 42–44.
[18] X. R. Jiang, A. Song, S. Bergelson, T. Arroll, B. Parekh, [47] Q. Wang, L. Liu, J. Shi, J. Sun, Y. Xue, Electron. J. Biotechnol.
K. May, S. Chung, R. Strouse, A. Mire-Sluis, M. Schenerman, 2015, 18, 138–142.
Nat. Rev. Drug. Discov. 2011, 10, 101–110. [48] R. Schleif, FEMS Microbiol. Rev. 2010, 34, 779–796.
[19] G. L. Rosano, E. A. Ceccarelli, Front. Microbiol. 2014, 5, 172. [49] N. A. Valdez-Cruz, L. Caspeta, N. O. Perez, O. T. Ramirez,
[20] M. Levin, H. Otten, C. von Wachenfeldt, M. Ohlin, BMC M. A. Trujillo-Roldan, Microb. Cell Fact. 2010, 9, 18.
Biotechnol. 2015, 15, 52.

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 130
[50] N. A. EL-Baky, M. H. Linjawi, E. M. Redwan, BMC Biotech- [80] X. Y. Fu, D. Z. Wei, W. Y. Tong, J. Chem. Technol. Biotechnol.
nol. 2015, 15, 14. 2006, 81, 1866–1871.
[51] J. Yin, G. Li, X. Ren, G. Herrler, J. Biotechnol. 2007, 127, [81] A. J. da Silva, A. C. L. Horta, A. M. Velez, M. R. C. Iemma,
335–347. C. R. Sargo, R. L. C. Giordano, M. T. M. Novo, R. C. Gior-
[52] I. Peubez, N. Chaudet, C. Mignon, G. Hild, S. Husson, dano, T. C. Zangirolami, SpringerPlus 2013, 2, 322.
V. Courtois, K. De Luca, D. Speck, R. Sodoyer, Microb. Cell [82] F. W. Studier, Methods Mol. Biol. 2014, 1091, 17–32.
Fact. 2010, 9, 65. [83] N. S. Berrow, K. Büssow, B. Coutard, J. Diprose, M. Ekberg,
[53] S. Nallamsetty, D. S. Waugh, Nat. Protoc. 2007, 2, 383–391. G. E. Folkers, N. Levy, V. Lieu, R. J. Owens, Y. Peleg,
[54] C. L. Young, Z. T. Britton, A. S. Robinson, Biotechnol. J. C. Pinaglia, S. Quevillon-Cheruel, L. Salim, C. Scheich,
2012, 7, 620–634. R. Vincentellic, D. Busso, Acta Crystallogr. Sect., D: Biol.
[55] N. K. Tripathi, J. Shukla, K. C. Biswal, P. V. L. Rao, Microb. Crystallogr. 2006, 62, 1218–1226.
Biotechnol. 2012, 5, 599–606. [84] A. K. Upadhyay, A. Murmu, A. Singh, A. K. Panda, PLoS
[56] F. Khan, P. M. Legler, R. M. Mease, E. H. Duncan, E. S. Berg- One 2012, 7, e33951.
mann Leitner, E. Angov, Biotechnol. J. 2010, 7, 133–147. [85] K. Ukkonen, A. Vasala, H. Ojamo, P. Neubauer, Microbial.
[57] K. Terpe, Appl. Microbiol. Biotechnol. 2003, 60, 523–533. Cell. Fact. 2011, 10, 107.
[58] N. K. Tripathi, J. Shukla, K. C. Biswal, P. V. L. Rao, Appl. [86] S. Routledge, Comput. Struct. Biotechnol. J. 2012, 3, 1–7.
Microbiol. Biotechnol. 2010, 86, 1795–1803. [87] J. Siurkus, J. Panula-Perälä, U. Horn, M. Kraft, R. Rimše-
[59] L. E. Bird, Methods 2011, 55, 29–37. liene, P. Neubauer, Microbial Cell Fact. 2010, 9, 35.
[60] S. J. Costa, A. Almeida, A. Castro, L. Domingues, H. Besir, [88] Q. Long, X. Liu, Y. Yang, L. Li, L. Harvey, B. McNeil, Z. Bai,
Appl. Microbiol. Biotechnol. 2013, 97, 6779–6791. J. Biotechnol. 2014, 192, 323–338.
[61] D. Esposito, D. K. Chatterjee, Curr. Opin. Biotechnol. 2006, [89] S. J. Wewetzer, M. Kunze, T. Ladner, B. Luchterhand,
17, 353–358. S. Roth, N. Rahmen, R. Klob, A. Costa e Silva, L. Regestein,
[62] D. Manderson, R. Dempster, Y. Chisti, J. Ind. Microb. J. Büchs, J. Biol. Eng. 2015, 9, 9.
Biotechnol. 2006, 33, 173–182. [90] R. S. Islam, D. Tisi, M. S. Levy, G. J. Lye, Biotechnol. Prog.
[63] N. K. Tripathi, A. Shrivastava, K. C. Biswal, P. V. L. Rao, 2007, 23, 785–793.
Biotechnol. J. 2011, 6, 604–608. [91] D. Manderson, R. Dempster, Y. Chisti, Enzyme Microb.
[64] V. Babaeipour, S. A. Shojaosadati, R. Khalilzadeh, N. Magh- Technol. 2006, 38, 591–598.
soudi, A. M. Farnoud, Appl. Biochem. Biotechnol. 2010, 160, [92] L. Beigi, H. R. Karbalaei-Heidari, M. Kharrati-Kopaei,
2366–2376. Protein Expression Purif. 2012, 84, 161–166.
[65] B. A. Fong, D. W. Wood, Microb. Cell Fact. 2010, 9, 77. [93] C. Castro-Martı́nez, S. Luna-Suárez, O. Paredes-López,
[66] F. Volonte, L. Piubelli, L. Pollegioni, Microbial. Cell Fact. J. Biotechnol. 2012, 158, 59–67.
2011, 10, 53. [94] K. Yari, S. S. A. Fatemi, M. Tavallaei, Biotechnol. Appl.
[67] J. Glazyrina, M. Krause, S. Junne, F. Glauche, D. Strom, Biochem. 2010, 56, 35–42.
P. Neubauer, Nat. Biotechnol. 2012, 29, 235–242. [95] G. Marini, M. D. Luchese, A. P. Argondizzo, A. C. de Góes,
[68] G. A. Goncalves, K. L. Prather, G. A. Monteiro, D. M. Pra- R. Galler, T. L. Alves, M. A. Medeiros, A. L. Larentis, BMC
zeres, Vaccine 2014, 32, 2847–2450. Biotechnol. 2014, 14, 1.
[69] N. K. Tripathi, R. Priya, A. Shrivastava, Bioengineered 2014, [96] A. L. Larentis, A. P. Argondizzo, S. Esteves Gdos, E. Jessour-
5, 198–203. on, R. Galler, M. A. Medeiros, Protein Expression Purif.
[70] J. Lu, Q. Song, Z. Ji, X. Liu, T. Wang, Q. Kang, Electron. 2011, 78, 38–47.
J. Biotechnol. 2015, 18, 281–285. [97] L. M. Maldonado, V. E. Hernandez, E. M. Rivero, A. P. Barba
[71] S. S. Yazdani, A. R. Shakri, C. E. Chitnis, Biotechnol. Lett. de la Rosa, J. L. Flores, L. G. Acevedo, A. De Leon Rodriguez,
2004, 26, 1891–1895. Biomol. Eng. 2007, 24, 217–222.
[72] S. Mohan Raj, C. Rathnasingh, W. C. Jung, S. Park, Appl. [98] B. Muntari, A. Amid, M. Mel, M. S. Jami, H. M. Salleh, AMB
Microbiol. Biotechnol. 2009, 84, 649–657. Express 2012, 2, 12.
[73] M. Khamduang, K. Packdibamrung, J. Chutmanop, Y. Chisti, [99] K. Yari, S. S. Fatemi, M. Tavallaei, Bioprocess Biosyst. Eng.
P. Srinophakun, J. Ind. Microbiol. Biotechnol. 2009, 36, 2012, 35, 407–414.
1267–1274. [100] M. Losen, B. Frölich, M. Pohl, J. Büchs, Biotechnol. Prog.
[74] H. Wang, F. Wang, W. Wang, X. Yao, D. Wei, H. Cheng, 2004, 20, 1062–1068.
Z. Deng, PLoS One 2014, 9, e112777. [101] K. Ukkonen, J. Veijola, A. Vasala, P. Neubauer, Microb. Cell
[75] R. D. Madurawe, T. E. Chase, E. L. Tsao, W. E. Bentley, Fact. 2013, 12, 73.
Biotechnol. Progress 2000, 16, 571–576. [102] M. L. Shuler, F. Kargi, Bioprocess Engineering: Basic Concepts,
[76] F. Volontè, L. Pollegioni, G. Molla, L. Frattini, F. Marinelli, 2nd ed., Pearson, London 2003.
L. Piubelli, BMC Biotechnol. 2010, 10, 33. [103] S. Y. Lee, Trends Biotechnol. 1996, 14, 98–105.
[77] N. K. Tripathi, R. Priya, A. Shrivastava, Appl. Microbiol. [104] D. C. Montgomery, Design and Analysis of Experiments,
Biotechnol. 2014, 98, 2461–2471. John Wiley & Sons, New York, 1997.
[78] M. A. Eiteman, E. Altman, Trends Biotechnol. 2006, 24, [105] S. V. Sohoni, D. Nelapati, S. Sathe, V. Javadekar-Subhedar,
530–536. R. P. Gaikaiwari, P. P. Wangikar, Bioresour. Technol. 2015,
[79] S. Leone, F. Sannino, M. L. Tutino, E. Parrilli, D. Picone, 188, 202–208.
Microb. Cell Fact. 2015, 14, 106.

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 131
[106] M. M. Meagher, J. G. Seravalli, S. T. Swanson, R. G. Ladd, [133] O. F. Restaino, U. Bhaskar, P. Paul, L. Lin, M. De Rosa, J. S.
Y. P. Khasa, M. Inan, J. C. Harner, S. K. Johnson, K. Van Dordick, R. J. Linhardt, Appl. Microbiol. Biotechnol. 2013,
Cott, C. Lindsey, R. Wannemacher, L. A. Smith, Biotechnol. 97, 3893–3900.
Prog. 2011, 27, 1036–1047. [134] X. Duan, J. Chen, J. Wu, Bioresour. Technol. 2013, 146,
[107] S. Vemula, R. Thunuguntla, A. Dedaniya, S. Kokkiligadda, 379–385.
C. Palle, S. R. Ronda, Protein Expression Purif. 2015, 108, [135] G. A. Goncalves, K. L. Prather, G. A. Monteiro, D. M. Pra-
62–72. zeres, Vaccine 2014, 32, 2847–2450.
[108] R. Garcı́a-Arrazola, S. C. Siu, G. Chan, I. Buchanan, [136] J. Jiang, X. Liu, L. Deng, P. Zhang, G. Wang, S. Wang, H. Liu,
B. Doyle, N. Titchener Hooker, F. Baganz, Biochemical Eng. Y. Su, Eur. J. Pharmacol. 2014, 740, 722–732.
J. 2005, 23, 221–230. [137] Z. Levarski, A. Soltysova, J. Krahulec, S. Stuchlik, J. Turna,
[109] J. Wang, B. Wen, Q. Xu, X. Xie, N. Chen, Biotechnol. Protein Expression Purif. 2014, 100, 40–47.
Biotechnol. Equip. 2015, 29, 374–380. [138] J. P. Aucamp, R. Davies, D. Hallet, A. Weiss, N. J. Titchener-
[110] L. C. Cheng, L. Y. Hor, J. Y. Wu, T. L. Chen, Biochemical Eng. Hooker, Biotechnol. Bioeng. 2014, 111, 1971–1981.
J. 2003, 14, 101–107. [139] P. Szweda, G. Gorczyca, P. Filipkowski, M. Zalewska, S. Mile-
[111] D. J. Korz, U. Rinas, K. Hellmuth, E. A. Sanders, W. D. wski, Prep. Biochem. Biotechnol. 2014, 44, 370–381.
Deckwer, J. Biotechnol. 1995, 39, 59–65. [140] J. Yang, X. Pan, H. Wang, L. Gao, J. Zhu, Y. Zhou, Y. Li,
[112] J. M. Puertas, J. Ruiz, M. R. De la Vega, J. Lorenzo, G. Cami- M. Li, B. Wang, Int. J. Clin. Exp. Med. 2015, 8, 173–180.
nal, G. González, Process Biochem. 2010, 45, 1334–1341. [141] A. Jungbauer, Trends Biotechnol. 2013, 31, 479–492.
[113] Q. Ren, B. Henes, M. Fairhead, L. Thöny Meyer, BMC [142] N. K. Tripathi, Ph.D. Thesis, National Institute of Technol-
Biotechnol. 2013, 13, 18. ogy, Rourkela, 2012.
[114] V. Babaeipour, S. A. Shojaosadati, R. Khalilzadeh, N. Magh- [143] M. Saraswat, L. Musante, A. Ravidá, B. Shortt, B. Byrne,
soudi, F. Abandeh, Biotechnol. Appl. Biochem. 2008, 49, H. Holthofer, Biomed. Res. Int. 2013, 312709.
141–147. [144] N. K. Prasad, Downstream Process Technology: A New Hori-
[115] E. Mahajan, T. Matthews, R. Hamilton, M. W. Laird, zon in Biotechnology, 1st ed., PHI, Delhi 2010.
Biotechnol. Prog. 2010, 26, 1200–1203. [145] E. Ledung, P. O. Eriksson, S. Oscarsson, J. Biotechnol. 2009,
[116] J. Glazyrina, E. M. Materne, T. Dreher, D. Storm, S. Junne, 141, 64–72.
T. Adams, G. Greller, P. Neubauer, Microb. Cell Fact. 2010, [146] A. Singh, V. Upadhyay, A. K. Upadhyay, S. M. Singh, A. K.
9, 42. Panda, Microb. Cell Fact. 2015, 14, 41.
[117] A. Westbroo, J. Scharer, M. Moo-Young, N. Oosterhuis, [147] B. Fahnert, H. Lilie, P. Neubauer, Adv. Biochem. Eng.
C. Perry Chou, Biochem. Eng. J. 2014, 88, 154–161. Biotechnol. 2004, 89, 93–142.
[118] I. Voulgaris, S. A. Arnold, R. Peight, L. M. Harvey, [148] C. Wang, L. Wang, X. Geng, Biochem. Eng. J. 2009, 43,
B. McNeil, Proc. Biochem. 2011, 46, 721–729. 197–202.
[119] C. J. Hewitt, A. W. Nienow, Adv. Appl. Microbiol. 2007, 62, [149] V. K. R. Dasari, D. Are, V. R. Joginapally, L. N. Mangamoori,
105–135. K. S. B. Adibhatla, Process Biochem. 2008, 43, 566–575.
[120] J. Ruiz, A. Fernández Castané, C. de Mas, G. González, [150] M. Li, H. Fan, J. Liu, M. Wang, L. Wang, C. Wang, Appl.
J. López-Santı́n, J. Ind. Microbiol. Biotechnol. 2013, 40, Biochem. Biotechnol. 2012, 166, 1264–1274.
335–343. [151] M. Alibolandi, H. Mirzahoseini, Biotechnol. Res. Int. 2011,
[121] B. H. Junker, J. Biosci. Bioeng. 2004, 97, 347–64. 973741.
[122] F. R. Schmidt, Appl. Microbiol. Biotechnol. 2005, 68, 425–35. [152] C. Wang, L. Wang, X. Geng, Biomed. Chromatogr. 2007, 21,
[123] T. S. Lee, Vaccine 2009, 27, 6439–6443. 1291–1296.
[124] B. Junker, Biotechnol. Bioeng. 2010, 105, 1011–1020. [153] C. K. Kim, C. H. Lee, S. B. Lee, J. W. Oh, PLoS One 2013, 8,
[125] R. S. Islam, D. Tisi, M. S. Levy, G. J. Lye, Biotechnol. Bioeng. e80109.
2008, 99, 1128–1139. [154] O. Kolaj, S. Spada, S. Robin, J. G. Wall, Microb. Cell Fact.
[126] F. Garcia Ochoa, E. Gomez, Biotechnol. Adv. 2009, 27, 2009, 8, 9.
153–176. [155] K. R. Babu, S. Swaminathan, S. Marten, N. Khanna, U. Ri-
[127] B. Gea, Z. Tanga, F. Zhaoa, Y. Rena, Y. Yanga, S. Qina, Process nas, Appl. Microbiol. Biotechnol. 2000, 53, 655–660.
Biochem. 2005, 40, 3190–3195. [156] S. Esfandiar, S. Hashemi-Najafabadi, S. A. Shojaosadati, S. A.
[128] G. Zhong, A. Yu, B. Shi, Y. Liu, C. Wu, Front. Biol. China. Sarrafzadeh, Z. Pourpak, Biotechnol. Appl. Biochem. 2010,
2009, 4, 75–81. 55, 209–214.
[129] J. F. Zawada, G. Yin, A. R. Steiner, J. Yang, A. Naresh, S. M. [157] F. Wang, J. Guo, Q. Bai, L. Wang, Biotechnol. Prog. 2014, 30,
Roy, D. S. Gold, H. G. Heinsohn, C. J. Murray, Biotechnol. 864–871.
Bioeng. 2011, 108, 1570–1578. [158] X. Geng, Q. Baia, Y. Zhang, X. Lia, D. Wub, J. Biotechnol.
[130] N. K. Tripathi, K. C. Biswal, P. V. Rao, Ind. Biotechnol. 2015, 2004, 113, 137–149.
11, 331–337. [159] C. Wang, L. Wang, X. Geng, Biotechnol. Prog. 2008, 24,
[131] R. Jalalirad, Electron. J. Biotechnol. 2013, 16, 3. 209–213.
[132] L. Piubelli, M. Campa, C. Temporini, E. Binda, F. Mangione, [160] Y. Chen, S. S. J. Leong, Process Biochem. 2010, 45,
M. Amicosante, M. Terenni, F. Marinelli, L. Pollegioni, 1570–1576.
Microbial Cell Fact. 2013, 12, 115. [161] C. Wang, Y. Cheng, Curr. Pharm. Biotechnol. 2010, 11,
289–292.

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 132
[162] L. C. M. Tan, A. J. S. Chua, L. S. L. Goh, S. M. Pua, Y. K. [175] D. Moricoli, M. E. Laguardia, D. C. Carbonella, M. C. Bal-
Cheong, M. L. Ng, Protein Expression Purif. 2010, 74, ducci, S. Dominici, V. Fiori, G. Serafini, M. Flego, M. Cian-
129–139. friglia, M. Magnani, Protein Expression Purif. 2014, 93,
[163] W. Gwinn, M. Zhang, S. Mon, D. Sampey, D. Zukauskas, 38–45.
C. Kassebaum, J. F. Zmuda, A. Tsai, M. W. Laird, Protein [176] J. Jiang, J. Zheng, Y. She, Z. Jia, Protein Expression Purif.
Expression Purif. 2006, 45, 30–36. 2015, 110, 95–101.
[164] G. Wu, C. Feng, Y. Hong, A. Guo, S. Cao, J. Dong, L. Lin, [177] T. Ahamed, B. K. Nfor, P. D. Verhaert, G. W. Van Dedem,
Z. Liu, Appl. Microbiol. Biotechnol. 2010, 87, 609–616. L. A. Van der Wielen, M. H. Eppink, E. J. Van de Sandt,
[165] B. V. Bhut, K. A. Christensen, S. M. Husson, J. Chromatogr. M. Ottens, J. Chromatgr. A. 2007, 1164, 181–188.
A. 2010, 1217, 4946–4957. [178] B. A. Bell, J. F. Wood, R. Bansal, H. Ragab, J. Cargo III, M. A.
[166] T. Zhang, H. Yang, L. Kang, S. Gao, W. Xin, W. Yao, Washington, C. L. Wood, L. A. Ware, C. F. Ockenhouse,
X. Zhuang, B. Ji, J. Wang, Hum. Vaccines Immunother. 2015, A. Yadava, Vaccine 2009, 27, 1448–1453.
11, 1779–1787. [179] S. M. Singh, A. Sharma, A. K. Panda, Protein Expression
[167] J. Yang, W. Zhang, K. Liu, S. Jing, G. Guo, P. Luo, Q. Zou, Purif. 2009, 68, 54–59.
Protein Expression Purif. 2007, 53, 339–345. [180] B. F. Roettger, M. R. Ladisch, Biotechnol. Adv. 1989, 7,
[168] S. S. Salunkhe, B. Prasad, K. Sabnis-Prasad, A. Apte-Desh- 15–29.
pande, S. Padmanabhan, J. Microbial. Biochem. Technol. [181] Y. H. Song, X. W. Sun, B. Jiang, J. E. Liu, X. H. Su, Protein
2009, 1, 005–010. Expression Purif. 2015, 116, 98–104.
[169] L. Zinzula, F. Esposito, E. Mühlberger, M. Trunschke, [182] C. Boi, J. Chromatogr. A. 2007, 848, 19–27.
D. Conrad, D. Piano, E. Tramontano, Protein Expression [183] D. Yu, R. Ghosh, J. Pharm. Sci. 2010, 99, 3326–3333.
Purif. 2009, 66, 113–119. [184] J. J. Milne, Methods Mol. Biol. 2011, 681, 73–85.
[170] R. Rouet, D. Lowe, K. Dudgeon, B. Roome, P. Schofield, [185] Y. Chisti, M. Moo-Young, Biotechnol. Adv. 1990, 8, 699–708.
D. Langley, J. Andrews, P. Whitfeld, L. Jermutus, D. Christ, [186] S. Yim, K. Jeong, H. Chang, S. Lee, Bioprocess Biosyst. Eng.
Nat. Protoc. 2012, 7, 364–373. 2001, 24, 249–254.
[171] M. Galluccio, P. Pingitore, M. Scalise, C. Indiveri, Protein J. [187] A. L. Nelson, J. M. Reichert, Nat. Biotechnol. 2009, 27,
2013, 32, 442–448. 331–337.
[172] S. T. Nguyen, D. T. Quyen, H. D. Vu, Biomed Res. Int. 2014, [188] J. Zhang, X. Lv, R. Xu, X. Tao, Y. Dong, A. Sun, D. Wei, Appl.
324705. Microbiol. Biotechnol. 2015, 99, 6705–6713.
[173] B. Kong, G. L. Guo, PLoS One 2014, 9, e85890. [189] I. Voulgaris, G. Finka, M. Uden, M. Hoare, Appl. Microbiol.
[174] F. Wang, Y. Hou, J. Zhou, Z. Li, Y. Huang, Z. Cui, Appl. Mi- Biotechnol. 2015, 99, 8441–8453.
crobiol. Biotechnol. 2014, 98, 7491–7499.

www.ChemBioEngRev.de ª 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ChemBioEng Rev 2016, 3, No. 3, 116–133 133