Improvement of downstream processing of recombinant proteins by means of genetic engineering methods
ABSTRACT The rapid advancement of genetic engineering has allowed to produce an impressive number of proteins on a scale which would not have been achieved by classical biotechnology. At the beginning of this development research was focussed on elucidating the mechanisms of protein overexpression. The appearance of inclusion bodies may illustrate the success. In the meantime, genetic engineering is not only expected to achieve overexpression, but to improve the whole process of protein production. For downstream processing of recombinant proteins, the synthesis of fusion proteins is of primary importance. Fusion with certain proteins or peptides may protect the target protein from proteolytic degradation and may alter its solubility. Intracellular proteins may be translocated by means of fusions with signal peptides. Affinity tags as fusion complements may render protein separation and purification highly selective. These methods as well as similar ones for improving the downstream processing of proteins will be discussed on the basis of recent literature.
- [Show abstract] [Hide abstract]
ABSTRACT: The principle of metal chelate-protein interaction was applied for protein extraction by means of reverse micellar phases. For this purpose, an affinity surfactant was synthesized exposing iminodiacetic acid as the hydrophilic moiety. The application of this substance as a cosurfactant leads to enhanced extraction of proteins exhibiting histidine groups on their surface.Biotechnology Techniques 8(5):307-312.
- [Show abstract] [Hide abstract]
ABSTRACT: A Tn5-based transposon bearing the kil gene (killing protein), mediating controlled export of periplasmic proteins into the culture medium, was constructed (Tn5-KIL3). This transposon contained the kil gene of the ColEl plasmid under the growth-phase-dependent promoter of the fic gene (filamentation induced by cAMP) of Escherichia coli, an interposon located upstream of kil, a kanamycin/neomycin-resistance gene, a multiple cloning site and the mob site. The transposition of Tn5-KIL3 to Acetobacter methanolicus showed a moderate transposition frequency (10(-5) -10(-6). By insertion of a Bacillus hybrid beta-glucanase (bgl) as a model protein into the transposon (Tn5-LF3) it was shown that the secretion function as well as the gene of the target protein had been transferred to and stably integrated into the chromosome of A. methanolicus, and that the transposition of Tn5-LF3 was non-specific. beta-Glucanase was highly overexpressed and secreted into the medium during stationary phase. Total and extra-cellular production of beta-glucanase varied depending on the integration site of the transposon. The viability of the bacterial cells was not affected, and cell lysis did not occur.Applied Microbiology and Biotechnology 06/1997; 47(5):530-6. · 3.69 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: We established a purification system for glutathione-S-transferase (GST) fusion protein using glutathione coupled magnetic particle. Glutathione was coupled covalently to the surface of magnetic particles with isothiocyanate functional groups. Cell lysate, containing the fusion protein, was then incubated with these glutathione coupled magnetic particles at room temperature. Unbound and non-specifically bound proteins were removed by wash steps. Subsequently, the GST-fusion protein was eluted from the magnetic particles by the addition of reduced glutathione. The resulting fusion protein was tested for purity using SDS-PAGE and demonstrated by Western blotting. The concentration of the fusion protein was measured by Bradford method. Both the conditions for incubation and washing were optimized. The results showed that 150 microg glutathione could be bound on 1 mg of particle surface and 10 mg of the glutatione-coupled magnetic particles was suitable for 100 microL lysate, the optimal incubation time for reaction between particles and lysate was 40 min. The magnetic particles could help purify efficiently GST-fusion protein with a yield of around 516 microg fusion protein per 10 mg particles. Magnetic particles can be successfully used in a simple, rapid and reliable method for the purification of GST-fusion proteins.Sheng wu gong cheng xue bao = Chinese journal of biotechnology 08/2009; 25(8):1254-60.
Improvement of downstream processing of recombinant
proteins by means of genetic engineering methods
Flaschel, Erwin; Friehs, KarlFlaschel, Erwin; Friehs, Karl
Flaschel, Erwin ; Friehs, Karl (1993) Improvement of downstream processing of recombinant
proteins by means of genetic engineering methods. Biotechnology Advances, 11(1), pp. 31-77
Posted at BiPrints Repository, Bielefeld University.
Improvement of downstream processing of recombinant
proteins by means of genetic engineering methods
The rapid advancement of genetic engineering has allowed to produce an impressive number
of proteins on a scale which would not have been achieved by classical biotechnology. At the
beginning of this development research was focussed on elucidating the mechanisms of
protein overexpression. The appearance of inclusion bodies may illustrate the success. In the
meantime, genetic engineering is not only expected to achieve overexpression, but to improve
the whole process of protein production. For downstream processing of recombinant proteins,
the synthesis of fusion proteins is of primary importance. Fusion with certain proteins or
peptides may protect the target protein from proteolytic degradation and may alter its solubility.
Intracellular proteins may be translocated by means of fusions with signal peptides. Affinity
tags as fusion complements may render protein separation and purification highly selective.
These methods as well as similar ones for improving the downstream processing of proteins
will be discussed on the basis of recent literature.
Biotech. Adv. Vol. 11, pp. 31-78,1993
Printed in Great Britain. All Rights Reserved.
@ 1993 Pergamon Press Ltd
IMPROVEMENT OF DOWNSTREAM PROCESSING OF
RECOMBINANT PROTEINS BY MEANS OF GENETIC
ERWIN FLASCHEL and KARL FRIEHS
Universittlt Bielefeld, Technische Fakulttlt, Arbeitsgruppe Fermentationstechnik,
Postfach 10 01 31, 4800 Biel~Celd 1, Germany
The rapid advancement of genetic engineering has allowed to produce an impressive
number of proteins on a scale which would not have been achieved by classical
biotechnology. At the beginning of this development research was focussed on elucidating
the mechanisms of protein overexpression. The appearance of inclusion bodies may
illustrate the success. In the meantime, genetic engineering is not only expected to achieve
overexpression, but to improve the whole process of protein production. For downstream
processing of recombinant proteins, the synthesis of fusion proteins is of primary
importance. Fusion with certain proteins or peptides may protect the target protein from
proteolytic degradation and may alter its solubility. Intracellular proteins may be
translocated by means of fusions with signal peptides. Affinity tags as fusion complements
may render protein separation and purification highly selective. These methods as well as
similar ones for improving the downstream processing of proteins will be discussed on the
basis of recent literature.
Downstream processing, recombinant protein, genetic engineering, fusion protein, fusion
complement, purification tag, affinity handle, protein secretion, protein export.
E. FLASCHEL and K. FRIEHS
Proteins are the primary products of the translation of the genetic code. They are therefore
of central interest for life science research as well as applications in a multitude of areas.
Thus, they are a very important class of biotechnology-derived products. Enzymes have
found a large area of applications due to their catalytic properties. Spectacular
advancements in molecular genetics (regulation of transcription and translation, gene
transfer, polymerase chain reaction etc.) and in the cultivation of animal and human cells
(cell fusion, hybridoma-technique, transfection) have given the opportunity that intensive
efforts are undertaken in order to develop processes for the production of species-specific
peptide hormones (insulin, somatostatin, growth hormones, etc.), immunologically active
proteins (interferons, interleukins, etc.), coagulation factors (urokinase, t-PA, factor VIII,
etc.), as well as of vaccines and monoclonal antibodies.
Such proteins may only be obtained economically by means of biotechnological processes,
since they represent polymers composed out of the 21 different proteinogeneous L-s-
amino acids (including selenocysteine). Chemical synthesis of proteins  is only
competitive for the production of short chain peptides. However, chemical synthesis is
important when non-proteinogeneous amino acids have to be incorporated into peptides.
The chemical route suffers from a laborious and expensive protection group chemistry and
an error rate which is still too high. In addition, it has to be taken into account that proteins
are only active in their natural folding state and that proteins of eukaryotes often need
posttranslational modifications (glycosylation, proteolysis). In consequence, protein
synthesis will need the use of microorganisms and of cells of animal or human origin. This
also means that proteins will need to be separated from complex natural media. Cell-free
translation systems are another possible method to produce recombinant proteins. The
usefullness of such systems in bioprocessing is under investigation . The importance
of downstream processing very likely will at least remain or even increase [2, 3] because
the separation and purification of proteins will remain the main source of costs for the
manufacturing of proteins.
A whole spectrum of downstream processes has been developed for the recovery of
proteins [4-7] and these have been documented in recent books and monographs [8-15].
Uncertainty with the use of novel processes, the partiality for problem-specific solutions and
new possibilities for the production of a multitude of pharmacologically active proteins have
considerably promoted the interest in downstream processing. Since the tool box of genetic
engineering is more and more used for improving the production of proteins, it has been
straightforward not only to use these techniques in order to achieve overexpression but to
aim at the improvement of downstream processing as well. Quite some novel approaches
for improving the downstream processing of proteins have already been worked out [16-
19]. These novel methods may lead to the opportunity that very different proteins might be
IMPROVING DOWNSTREAM PROCESSING 33
purified by means of one unique process strategy. On the one hand, the physico-chemical
properties of the proteins may be changed or amplified by means of protein design or site-
directed mutagenesis. Recombination of different genes for the production of fusion
proteins, on the other hand, is of primary importance because unique fusion partners may
be chosen for the complementation of the target protein. Such fusion partners are
adequate which are carriers of a specific affinity useful for the downstream process applied.
The class of fusion proteins englobe the use of signal sequences in order to change the
localization of the target proteins.
Such approaches and similar ones will be reviewed - taken from recent literature. The
following chapter will deal with questions referring to the choice of host organisms and the
requirement for producing biologically active natural proteins. Only a few issues can be
emphasized with respect to host organisms and expression in general for not overloading
the article. A further chapter discusses possibilities of manipulating the localization of
proteins and the properties of cells which have direct impact on downstream processing.
The last chapter is dedicated to strategies of protein modification for the improvement of
protein separation and purification.
PRODUCT CONCENTRATION AND BIOLOGICAL ACTIVITY
The economy of downstream processing of proteins highly depends on the achievable
product concentration. Therefore, recombinant protein production aims at the improvement
of the overall protein concentration. However, aiming at a high product concentration may
fail, unless a high proportion of biologically active proteins is obtained.
The first step for improving the concentration of the target protein is the choice of an
adequate biological system by means of which overexpression may be achieved. Many
such expression systems are available. For the optimisation of these systems a series of
factors have to be taken into account referring to the replication, transcription and
translation of the genes.
The replication of genes coding for recombinant proteins strongly influences the stable
transmission as well as the copy number of these genes. Most expression systems for
bacteria use plasmids as vehicles of recombinant DNA. Plasmid stability depends on both
the structural and the segregative stability. The structural stability refers to the error rate
during plasmid replication, whereas the segregative stability depends on the transmission
of the plasmid on to daughter cells. Besides the correct replication of plasmids the
structural stability is influenced by the rate of mutation and the gene repair facilities of the
host organism. The segregative stability, commonly called plasmid stability, depends on a
34 E. FLASCHEL and K. FRIEHS
series of factors which cannot be discussed here in detail . Means for stabilizing or
enhancing plasmid stability have been discussed by Kumar et al. .
The rate of gene expression may be accelerated by increasing the number of plasmids per
cell. This may be achieved by using so-called multi-copy plasmids, the copy number of
which may reach 100 owing to their particular mode of replication. Since the plasmid copy
number has to be high only at the time when the target protein shall be produced, a
controlled induction of plasmid replication represents an advantageous strategy. This may
be achieved by using so-called runaway replication plasmids the copy number of which
may easily reach 1,000 . An obvious additional strategy consists of integrating several
copies of the coding gene into the plasmid. However, it has to be kept in mind that a high
plasmid copy number may affect the physiological state of the host organism leading to a
limited rate of gene expression.
Both a promoter-operator system and an appropriate termination sequence are required in
order to achieve efficient transcription. In addition, transcription should be inducible. During
the development of expression systems different sequences have been recognized as
strong promoters. A strong promoter is the prerequisite of a high transcription frequency.
The best known inducible bacterial promoter-operator systems are those using operators
derived from the lac- , trp-  and the ~,-  systems. The lac-operator is often used
for the induction of different promoters like lac and tac. Induction is achieved by the
addition of IPTG (isopropyl-J3-D-l-thiogalactopyranoside). Promoters coupled with the trp-
operator may be induced by iAA (3-(3-indolyi)-acrylic acid). An induction by means of a
temperature jump is applied in the case of X-promoters using a temperature sensitive
mutant strain with respect to the ~.-repressor. Chemical inducers may be quite expensive
and may lead to contamination of the product. Temperature induction may cause
undesirable secondary products to appear due to an altered metabolism. Therefore, there
is still a need for improved novel promoters  - like ones inducible by oxygen depletion
 or by means of a pH-shift . When strong promoters are applied, a strong
terminator is required as well. Without an appropriate termination sequence transcription
would proceed beyond the target gene and transcription would lead to useless products.
Appropriate sequences like the trpA-transcription terminator are commercially available
Transcription is the first step during gene expression leading to mRNA. The second step
involves the ribosomes which are responsable for the translation of the mRNA into the
amino acid sequence of the recombinant protein. Translation efficiency depends on the
stability of the mRNA, the ribosomal binding site (RBS), the correct termination of
translation as well as the use of special codons. The stability of mRNA is a function of its
susceptibility for hydrolysis by ribonucleases. It is commonly assumed that the lifetime of
mRNA is influenced mainly by the presence of 3'-ribonucleases, because the secondary
IMPROVING DOWNSTREAM PROCESSING 35
structure of the 3'-terminal sequence of the mRNA is of crucial importance . In
consequence, mRNA stability may be enhanced by changing the 3'-terminal sequence or
shifts in growth rate . The lifetime of the mRNA determines how often it may be used
by the ribosomes for translation . Newest results suggest that RNA degradation may
depend on 5'-terminal base pairing in E. coll. .
Recombinant genes are commonly introduced behind an inducible promoter, but they often
have their own ribosomal binding site (RBS). This may be the reason for a low translation
frequency, although the rate of transcription may be high. Optimized gene expression
systems include an RBS especially designed for the particular host organism . Thus,
AT-rich sequences may be introduced up- and downstream of the RBS in order to increase
the initiation frequency of translation . The distance between RBS and the start codon
AUG may also be changed in order to achieve higher productivities . Synthetic RBS-
sequences have not led to improvements of gene expression . A strong terminator is
required for efficient translation, too. Appropriate sequences are commercially available
If a synthetic gene is used the sequence of which has been deduced from the amino acid
sequence of a protein, the base sequences of the codons for the amino acids should be
chosen with care. Studies with genes of highly expressed proteins have shown that certain
codons out of the pool of the degenerate genetic code are preferred [34, 266].
After translation the stability of the protein itself determines how much product will be
obtained. Protein stability mainly depends on the susceptibility of the protein for proteolytic
decomposition. Especially in the case of recombinant proteins, proteases may cause a
considerable loss of product [35, 222]. It seems as if heterologous proteins were
recognized as belonging to a different species with the consequence of a rapid proteolytic
degradation. Different proteases may be responsible for protein digestion. In the case of E.
colithe protease La coded on the /on-gene is accused of mainly being responsible for the
degradation of recombinant proteins . By using ~on-minus mutants this problem may be
partially overcome. However, the use of such mutants may cause a series of difficulties
during cultivation. The product of the Ion-gene is also controlling cell division and the
development of the polysaccharide capsule. Therefore, Ion-minus mutants tend to
overproduce polysaccharides causing strong production of slime and interference with cell
division and growth . This problem may be circumvented by controlling the Ion-gene by
inducing the Ion-function during the growth phase and switching it off during the production
phase. However, this control is quite delicate since the concentration of the Ion-gene
product has to be kept in a very narrow band. This may not be achieved with commonly
used promoters because of the constitutive gene expression taking place in spite of
36 E. FLASCHEL and K. FRIEHS
repression . By inactivation of structural and regulatory genes of the slime capsule
synthesis it is possible to prevent slime production and the growth problems related with
that . Modifications of the sulA-gene the product of which regulates the Ion-coded
protease may yield a significant improvement of cell division for Ion-negative strains.
However, it has to be taken into account that Ion-negative strains are very sensitive with
respect to the induction of stress proteins causing additional growth problems.
degradation of foreign proteins by phage infection, using T4-promoter  and to consist
the influence of growth rate on proteolysis . An easier to implement strategy for
reducing protein degradation consists in the synthesis of fusion proteins. Thus, small
heterologous peptides which commonly are very labile have been stabilized by fusion with
large proteins like J~-galactosidase or Staphylococcus protein A [39-43, 272]. In the case of
such fusions it may be necessary to split off the fusion complement in order to obtain the
target protein in the desired form.
methods to reduce protein degradation are to inhibit the host-mediated
The export of proteins is another elegant possibility to avoid the intracellular environment
rich of proteases . However, some secretion systems may also export proteases into
the extracellular environment. The problems related to the presence of proteases have to
be considered especially for the operation of cell disintegration . Therefore, the addition
of protease inhibitors may be necessary during cell disruption.
Many proteins, in particular eukaryotic proteins of pharmaceutical interest, require further
modification after protein biosynthesis. Such posttranslational modifications may consist in
phosphorylation, glycosylation, the cleavage of sequences from pre-pro-proteins of
especially secretory proteins as well as the cleavage of the primary protein chain. Without
appropriate modification these proteins are often obtained in a biologically inactive form.
In many cases it is relatively easy to produce large quantities of protein by means of
bacterial systems like E. coil However, strains of E. cofiare not provided with a pathway for
the glycosylation of proteins. If glycosylation is of crucial importance, other less convenient
expression systems have to be used. Yeast cells, simple eukaryotes, have glycosylating
activity, but often they use other glycosidic residues then those required for the particular
case . There is certainly still a long way to go until bacterial systems or yeasts may
have learned to achieve different posttranslational modifications by means of genetic
engineering. Cells of animal and human origin will remain the systems of choice for
proteins needing adequate posttranslational processing . Another way to circumvent this
problem is to look for the biologically active part of the target protein which, by
circumstance, may not need further modification .
IMPROVING DOWNSTREAM PROCESSING
Inclusion Bodies, Renaturation end Protein Foldino
Recombinant proteins are often found as insoluble aggregates in the cytoplasm [17, 48].
Extremely high protein concentrations due to overexpression may be responsible for this
phenomenon . These accumulations of solid insoluble proteins are called inclusion
bodies . The formation of these refractile particles may represent an advantage as well
as a disadvantage for the production of recombinant proteins. Thus, inclusion bodies are
relatively simple to be recovered from cell homogenates [17, 50, 51], and are protected
from proteolytic cleavage. However, a great disadvantage must be seen in the fact that
proteins bound in inclusion bodies are biologically inactive. Therefore, these highly
insoluble protein aggregates have first to be denatured under extreme conditions in order
to be solubilized [51, 52]. Subsequently, these solubilized proteins have to be renatured by
means of adequate methods in order to be transformed into their biologically active form.
This procedure may yield heavy losses which may be acceptable if a high overproduction
goes in common with a high added value. Genetic engineering methods for the reduction of
aggregate formation consist in forcing protein secretion as well as in changing specific
properties of the target protein which may be responsable for the occurence of inclusion
bodies. Wilkinson et a/.  describe factors influencing the solubility of recombinant
proteins. The critical factors englobe the protein concentration, the distribution of surface
charge, the portion of particular secondary structures, the contents of cysteine and proline
as well as the overall hydrophobicity. These properties may be changed deliberately by
site-directed mutagenesis as well as by fusion of the target protein with an appropriate
complement [53, 277]. Some approaches along these lines have been discussed by
Kotewicz . The most simple possibility for reducing or even preventing inclusion body
formation seems to be by decreasing the rate of expression [52, 223], e.g. by lowering the
temperature after induction .
In the case of heterologous proteins it is often inevitable that aggregates are formed and
the whole procedure of solubilization and renaturation has to be applied . The
achievement of the native structure is the most crucial aspect for the production of
recombinant proteins . Achieving correct protein folding is a fundamental problem and
of enormous economic impact . Protein folding depends on a variety of different factors
[56-58]. Certain additives like e.g. PEG  may improve folding .
kinetics of folding reveals a rather complex pathway .
In recent years it has become clear that in vivo protein folding is assisted by catalytic
protein complexes. Such proteins obviously play a crucial role for protein folding and
assembly. They belong to the heterogeneous group of heat-shock proteins (hsp), and are
now divided into two more or less distinct classes . The class of the chaperonins
belonging to the hspl0 and hsp60 families occur both in eubacteria, mitochondria and
plastids, whereas the class of the hsp70 family seems to be present everywhere where
38 E. FLASCHEL and K. FRIEHS
protein folding occurs. Evidence is also emerging for the presence of cytosolic chaperonins
in archaebacteria as well as in eukaryotes [226, 228]. However, the term chaperone is
generally used for proteins which assist in protein folding. Anyway, the division of
chaperones into two classes seems to be a quite simplistic approach, because it has been
shown in in vivo experiments that five chaperones cooperate in a sequential ATP-
dependent folding pathway . Both in vitro and in vivo studies of protein folding have
already revealed quite some details about their action [61, 62, 230, 231, 267] and other
proteins assisting in protein folding like peptidyl-prolyl cis-trans isomerase and protein
disulfide isomerase [232, 234]. Such studies may finally lead to the fine-tuning of the
cellular machinery of protein folding. This would certainly lead to a considerable
improvement for the manufacturing of correctly folded recombinant proteins.
(~leava_oe of Fusion Proteins
Fusion proteins exhibiting the desired biological activity may be used as such, if there are
no special regulations requiring them to be transformed into their natural structure, as it is
still common practice for proteins of therapeutical use. In the latter case the cleavage of the
fusion complement has to be accomplished. For this purpose, the amino acid sequence of
the fusion protein has to be constructed with a specific cleavage side.
Fusion proteins may be cleaved by chemical or enzymatic means. A chemical cleavage
may be relatively unexpensive, but the rather drastic reaction conditions may lead to non-
specific cleavage and denaturation of the target protein. Enzymatic methods are generally
preferred, because specific cleavage sites may be introduced between the target protein
and the fusion complement [63, 235]. Over 200 protein-cleaving enzymes are listed in a
book of B. Keil . Costs for the cleavage procedure may be reduced considerably by
the development of an adequate standard cleavage sequence for a specific proteolytic
enzyme. Table 1 contains an overview of available chemical and enzymatic cleavage
methods and their respective cleavage sequences.
At2breviations used in Table 1:
BNPS-skatol = 3-bromo-3-methyl-2-(2-nitrophenyl mercapto)-3H-indol,
CAT = chloroamphenicol acyltransferase, HIV = human immunodeficiency virus,
ompA = outer membrane protein A, Kex 2 = yeast endoprotease
IMPROVING DOWNSTREAM PROCESSING
Table 1: Cleavage of fusion proteins
Cleavage agent Cleavage-site sequence
Asn ~ Gly
Asp ~ Pro
Asp g Pro
Enzyme Cleavage-site sequence Ref.
Poly His t)
Poly Arg U, Poly Lys ;)
Trp U, Tyr 8, Phe ~
Pro-X 8 Gly-Pro
X-Tyr ~ (X not Pro)
CAT- ~ -HIV-1 protease fusion, selfsplitting
Y-Pro ~t X-Pro; Y = Pro, Ala, Gly, Thr
X = Thr, Ser, Ala
-Lys ~ Arg
Tyr-Ile-His-Pro-Phe-His-Leu ~ Leu
Arg ~, Lys
Ubiquitin 8 Relaxin ¢¢-chain
S. aureus strain V8 protease
E. FLASCHEL and K. FRIEHS
The localization of proteins is obviously of major importance for downstream processing. If
the protein is to be found in the cytoplasm, the concentrated cell mass has at first to be
broken. The target protein, in consequence, has to be separated from a complex mixture of
proteins, nucleic acids; eventually cellular compartments and debris. A more gentle cell
disruption may be applied, if the target protein is secreted into the periplasm of Gram-
negative bacteria. Obviously disruption is superfluous in the case of extracellular proteins.
Separation of Cells and Cell Disruotion
Biomass is most often harvested by centrifugation or cross-flow filtration . For E. coli, a
genetic engineering approach has been developed for changing the properties of the cells
in order to facilitate biomass recovery. A gene has been cloned which codes for a protein
localized on the outer membrane surface and which is responsable for the flocculation
properties. This modification resulted in a higher sedimentation velocity of the cells .
Similar interventions should also yield better properties for centrifugation and filtration.
Mechanical as well as non-mechanical techniques are applied for cell disruption on a large
scale [6,89]. The mechanical methods comprise high pressure homogenization and bead
milling. Both processes may exhibit considerable protein losses due to uncomplete protein
liberation and thermal denaturation. Non-mechanical methods englobe chemical as well as
biochemical processes. Cells may be permeabilized by organic solvents or by enzymatic
lysis of the cell wall . The use of organic solvents implies appropriate safety precautions
, whereas enzymic processes are expensive on a large scale .
By means of genetic modifications it becomes possible to control cell lysis. The product of
the kil-gene of the plasmid ColE1 may yield complete lyses of the cells . A recent patent
describes how the kil-gene, under the control of the lac-promoter may lyse the cells after
induction with IPTG . This strategy yields an expression system for which cell lyses may
be induced under controlled conditions after the recombinant product has accumulated.
Based on the same principle, the lysis gene E of phage ~X174 may be employed - under
control of the ~,-PL-promoter . In the presence of the temperature-sensitive ~,-repressor
cl897, cell lysis may be induced by increasing the temperature to 42 °C.
Protein export may circumvent some problems inherent to the expression of recombinant
proteins . It has already been discussed that the formation of inclusion bodies and
protein degradation due to cytoplasmic proteases may be prevented. In addition, some
IMPROVING DOWNSTREAM PROCESSING 41
proteins may be letal for a host when overproduced. Secretion obviously may reduce this
phenomenon . However, secretion also may yield uncorrectly folded proteins especially
in the case of eukaryotic proteins synthesized by bacteria.
Nevertheless, protein secretion into the medium considerably simplifies downstream
processing. In this case the cells may be separated from the medium by e.g. centrifugation
and the proteins may be isolated from the supernatant. The early separation of unbroken
cells leads to reduced contamination by other proteins or cellular constituents - an obvious
advantage. A minor disadvantage of this strategy is the need for handling large volumes of
The transport of proteins into the periplasm of Gram-negative bacteria represents a special
case of protein secretion. After separation of biomass the product may be liberated from
the periplasm prior to being isolated out of a relatively small volume . Contamination
due to cytoplasmic constituents may be avoided by gently removing the cell capsule and
the outer membrane. This strategy has actually gained much attention. The periplasmic
space representing about 20 to 40% of the cellular volume in the case of E. coil is large
enough for the accumulation of large amounts of proteins. Besides the advantage that
proteins in the periplasm are protected against the attack by cytoplasmic proteases (but not
against outermembrane bound proteases , the environment of the periplasm favours
the correct folding of proteins.
The gene of the target protein has to be coupled with an appropriate signal leader
sequence and the host organism has to be provided with a cellular transport system in
order to allow secretion of proteins. In addition, the target protein should not show
properties preventing secretion.
which enable proteins to use the sorting and transport systems of a particular organism. In
bacteria signal peptides with a length of 15 to 30 amino acids are found [96,97]. These
structures represent positively charged leader sequences allowing proteins to cross
membranes. It may be mentioned that hemolysin of E. co/i carries its signal peptide at the
C-terminus. Between signal peptide and core protein a cleavage site is found at which the
signal peptide is cleaved off by means of a specific signal peptidase after transport of the
core protein has occured.
Signal sequences commonly are short N-terminal peptides
Secreting proteins may show additional domains besides the signal sequence which may
be necessary for a successful transport. In this category inner sequences are found which
facilitate or even stop transport, like in the case of membrane proteins, giving rise to the
development of an export competent conformation. Therefore, fusion with a signal peptide
42 E. FLASCHEL and K. FR[EHS
Table 2: Signal leader sequences for protein secretion
Target protein I M/P
Host : Escherichia coil
Amylase, B. stearothermophi/us
bla Proinsulin, human
IgG of mouse
Epidermal growth factor, rat
alk. Phosphatase, c~-amylase
CD4 receptor of HIV
Gene 5 protein, phage M13
Nuclease A, S. aureus
Colony stimulation factor, human
Superoxide dismutase, human
Antiviral protein, Mirabilis
Nuclease A, S. aureus
Antibody VH-domain, mouse
Trypsin inhibitor, bovine
Epidermal growth factor, human
Fusion : #-galactosidase-alk. phosphatase
Fusion : MBP-~-galactosidase
Growth hormone release factor, human
Parathyroid hormone, human
Insulin-like growth factor, human
Ovalbumin Ovalbum in
IMPROVING DOWNSTREAM PROCESSING
Table 2: Signal leader sequences for protein secretion (continued)
Host : E$cherichia coil (cont.)
Preproineulln (rat) Proinsulin, rat
Enterotoxin LTA Epidermal growth factor, human
Synthetic Interferon e¢2.
Metalloproteese Metalloprotease, with helper protein
BRP Insulin-like growth factor, human
I~-Iactamase & a-amylase
Pseudomonas cholesterol esterase
Host : Bacillus aubtilis
alkaline Phosphatase M 
amy, B. amy/oliquefaciens Amylase, B./icheniformis M 
amy, prepro-peptide Amylase & human growth hormone M 
ble, E. coil Amylase, B./icheniformis M 
Prepro-neutral-protease Growth hormone, human M 
Host : Saccharomyces cerevisiae
Yeast killer toxin a-Amylase, mouse M 
Prepro-a-factor Viral proteins, human papillomavirus M 
Abbreviations used in Table 2:
amy = amylase, bla = ~lactamase, BRP = bacteriocin release protein, cgt = cyclodextrin
glycosyltransferase of B. circulans, hly = hemolysin, M = secretion into the medium, male =
maltose binding protein, MBP = maltose binding protein, OM = secrection into the outer
membrane, ompA = outer membrane protein A, ompF = outer membrane protein F, P =
secretion into the periplasm, phoA =alk. phosphatase, phoS = phosphate binding protein,
spa = Staphylococcus aureus protein A