MINI-REVIEW
Overview of bacterial expression systems for heterologous protein production:from molecular and biochemical fundamentals to commercial systems
Kay Terpe
Received:30January2006/Revid:18April2006/Accepted:19April2006/Published online:22June2006
索马里人
#Springer-Verlag2006
Abstract During the proteomics period,the growth in the u of recombinant proteins has incread greatly in the recent years.Bacterial systems remain most attractive due to low cost,high productivity,and rapid u.However,the rational choice of the adequate promoter system and host for a specific protein of interest remains difficult.This review gives an overview of the most commonly ud systems:As hosts,Bacillus brevis,Bacillus megaterium,Bacillus sub-tilis,Caulobacter crescentus,other strains,and,most importantly,Escherichia coli li K12and their derivatives are prented.On the promoter side,the main features of the L-arabino inducible araBAD pro-moter(P BAD),the lac promoter,the L-rhamno inducible rhaP BAD promoter,the T7RNA polymera promoter,
the trc and tac promoter,the lambda phage promoter p L,and the anhydrotetracycline-inducible tetA promoter/operator are summarized.
Introduction
The production of recombinant proteins in a highly purified and well-characterized form has become a major task for the protein chemist within the pharmaceutical industry (Schmidt2004).Bacterial expression systems for heterolo-gous protein production are attractive becau of their ability to grow rapidly and at high density on inexpensive substrates,their often well-characterized genetics and the availability of an increasingly large number of cloning vectors and mutant host strains.To produce high levels of protein,it is often uful to clone the gene downstream of a well-characterized,regulated promoter.
In general it is difficult to decide which host and promoter system is the best for heterologous protein production.It depends often on the target protein itlf. This review describes a variety of bacterial host and promoter systems widely ud for heterologous protein production.Nevertheless many bacterial systems are not able to modify proteins posttranslationally such as glyco-sylation.If the posttranslational modification is esntial for bioactivity,bacterial expression systems should not be u
d for heterologous protein production.Alternative hosts such as yeasts,filamentous fungi,or inct and mammalian cell cultures are available for this application(reviewed in Schmidt2004).
Escherichia coli as host
The gram-negative li is the most commonly ud organism for heterologous protein production.One of the reasons ems to be that this organism is very well-known and established in each laboratory.So it is no surpri li systems are also most commonly ud for industrial and pharmaceutical protein production.Large-scale production systems are established.A disadvantage for therapeutic u of produced recombinant proteins in E. coli is the accumulation of lipopolysaccharide(LPS), generally referred as endotoxins,which are pyrogenic in humans and other mammals.Proteins for this application must be purified in a cond step to become endotoxin-free (Petsch and Anspach2000).
In general,overexpresd recombinant proteins accumu-late either in the cytoplasm or periplasmic space.Most
Appl Microbiol Biotechnol(2006)72:211–222 DOI10.1007/s00253-006-0465-8
K.Terpe(*)
IBA GmbH,
37079Göttingen,Germany
e-mail:
frequently,the cytoplasm is the first choice for heterologous protein production becau the higher yield ems to be more attractive.Also,remarkable yields of creted proteins are well-documented(reviewed in Georgiou and Segatori 2005).For routine protein li BL21and K12and their derivatives are most frequently ud (Table1).In contrast to K12strains,BL derivates are lon (Phillips et al.1984)and ompT protea deficient.Over-expression of a gene in a foreign host resulted often in lots of unwanted problems(Table2a).Most problems are the result of the difference between the codon usage of li and the overexpresd protein, e.g.,eukaryotic
li strains most frequently ud for heterologous protein production and their key features
AD494K-12trxB mutant;facilitates cytoplasmic disulfide bond formation
BL21B834Deficient in lon and ompT proteas
BL21trxB BL21trxB mutant;facilitates cytoplasmic disulfide bond formation;deficient in lon and ompT proteas BL21CodonPlus-RIL BL21Enhances the expression of eukaryotic proteins that contain codons rarely ud li:AGG,AGA,
AUA,CUA;deficient in lon and ompT proteas.
BL21CodonPlus-RP BL21Enhances the expression of eukaryotic proteins that contain codons rarely ud li:AGG,AGA,
CCC;deficient in lon and ompT proteas.
BLR BL21recA mutant;stabilizes tandem repeats;deficient in lon and ompT proteas
B834B strain Met auxotroph;35S-met labeling
C41BL21Mutant designed for expression of membrane proteins
C43BL21Double mutant designed for expression of membrane proteins
HMS174K-12recA mutant;Rif resistance
JM83K-12Usable for cretion of recombinant proteins into the periplasm
Origami K-12trxB/gor mutant;greatly facilitates cytoplasmic disulfide bond formation
Origami B BL21trxB/gor mutant;greatly facilitates cytoplasmic disulfide bond formation;deficient in Ion and ompT
proteas
Rotta BL21Enhances the expression of eukaryotic proteins that contain codons rarely ud li:AUA,AGG,
AGA,CGG,CUA,CCC,and GGA;deficient in lon and ompT proteas
Rotta-gami BL21Enhances the expression of eukaryotic proteins that contain codons rarely ud li:AUA,AGG,
AGA,CGG,CUA,CCC,and GGA;deficient in Ion and ompT proteas;trxB/gor mutant;greatly
facilitates cytoplasmic disulfide bond formation
Most strains are also available as DE3and DE3pLysS strains.Strains are commercially available from different manufacturers
家长会板报Table2a Some problems of heterologous protein production li and possible solutions
Symptom Possible problem A collection of solutions
Cell death or no colonies Toxic protein,high basal
expression More stringent control over basal expression Tightly controlled promoter system Weaker promoter
Lowering temperature
Lowering inducer concentration
Insoluble disulfide protein (inclusion bodies)Reduction of disulfide bonds Minimize reduction in cytoplasm
Accumulation in the periplasm
Insoluble protein (inclusion bodies)Too much expression Attenuate expression by:weaker promoter,lowering temperature,lowering
inducer concentration,decrea plasmid copy number,fusion of a hydrophilic
affinity tag
No activity Misfolded protein,affinity
tag can decrea activity Minimize reduction in cytoplasm Accumulation in the periplasm Attenuate expression
Change affinity tag
No protein,truncated li codon usage
(codon bias)
Supply rare tRNAs
Stronger promoter
Increa plasmid copy number
Lower temperature
Tightly controlled promoter system
Nevertheless,another bacterial host li could also solve the problem
proteins.Rare codons could especially be a problem (Table2b).It is well-known that amino acids are encoded by more than one codon,and each organism carries its own bias in the usage of the61available amino acid codons.In each cell,the tRNA population cloly reflects the codon bias of the mRNA population(Dong et al.1996).If the mRNA of heterologous target genes is intended to be overexpresd li,differences in the codon usage can impede translation due to the demand for one or more tRNAs that may be rare or lacking in the expression host (Kane1995;Goldman et al.1995).Insufficient tRNA pools can lead to translational stalling,premature translation termination,translation frameshift,and amino acid mis-incorporation(Kurland and Gallant1996).
Theoretically,modification of culture , lowering the temperature(Table3)or changing me
dia composition might shift the codon usage bias enough to alleviate some codon-usage bad expression problems. However,it was reported that the levels of most tRNA isoacceptors corresponding to rare codons remain un-changed at different growth rates(Dong et al.1996). Translation problems similar to tho caud by codon usage bias can also be created by high-level expression of proteins having an abundant amino acid.In this ca, expression may be improved by supplying the limiting amino acid in the culture medium(Kane1995).Reduction of the inducer concentration ems to be possible also,but mostly unsuccessful.To enhance expression of eukaryotic proteins or proteins that contain codons rarely ud in E. coli,li strains were engineered to meet this problem(Table1).The strains supply additional tRNAs under control of their native promoters.
Independent on host and promoter system,low,middle, and high copy plasmids could be ud to reduce expression problems due to an inadequate expression level.For example,the most frequently ud vectors bad on the plasmid ColE1,which aids in the isolation of large amounts
Table2b Comparison of rare codons li
Organism AGG arginine AGA arginine CUA leucine AUA isoleucine CCC proline GGA glycine Bacterial hosts
Escherichia coli B 2.1 2.4 3.4 5.0 2.48.2
Some organisms
Arabidopsis thaliana10.918.99.912.6 5.324.2 Caenorhabditis elegans 4.015.47.99.4 4.431.6 Clostridium tetani E88 5.025.511.267.4 1.834.6 Drosophila melanogaster 6.3 5.28.29.518.017.7
Homo sapiens11.912.07.27.419.916.5
H.sapiens Mitochondriom0.40.470.444.533.918.9 Plasmodium falciparum 3.317.0 5.340.8 3.020.0
Pichia pastoris 6.620.210.911.7 6.719.1 Picrophilus torridus DSM979022.916.27.830.3 2.618.1 Saccharomyces cerevisiae9.321.313.417.8 6.810.9情归处
The codon usages of bacteria ud for heterologous protein expression and some organisms are list
ed(for Bacilus brevis not available).Codon frequencies are expresd as codons ud per1,000codons encountered.The arginine codons AGG and AGA are recognized by the same tRNA and should therefore be combined.Codon frequency of more than15codons/1,000codons may cau problems for high-level expression li.A complete summary of codon usages can be found at jp/codon/
Table3Incubation temperature and time after induction to change
codon bias li enhancing protein concentration or increasing
soluble protein
Incubation temperature(°C)Incubation time(h)
824–72
1516–24
2012–16
256–12
305–6
373–4
Yield of soluble protein could be also incread by lowering the
temperature of the ,30or20°C
of recombinant protein,may be replaced by low copy number(l cn)vectors when high copy number vectors cannot be ud becau large amounts of the recombinant protein are toxic.The vectors pWKS29,pWKS30, pWKS129,pWKS130and its derivates carry the pSC101 replicon that produces six to eight plasmid copies per cell (Wang and Kushner1991)and could be a uful tool to express proteins li like membrane proteins and kinas.Many different vectors are described but not further discusd in this review.
Nevertheless,lots of proteins could not be expresd in E. coli.For this ca,bacterial hosts like Bacillus brevis,Bacillus megaterium,Bacillus subtilis,or Caulobacter crescentus could be an interesting alternative described in this review.
鸡爪的做法大全家常
Another disadvantage of overexpression in the cyto-plasm is that a lot of proteins form inclusion bodi
es. Strategies for resolution could be lowering temperature, amino acids substitution,coexpression of chaperones, hydrophilic large fusion partner,adding of sorbitol,glycyl betain,sucro,raffino in the growth medium,changing culture ,pH,or changing the bacterial strain (reviewed in Hockney1994;Makrides1996).Alternatively, inclusion bodies can be solubilized and refolded to get functional and active products(reviewed in Singh and Panda2005).
BL21(DE3)derivatives were especially designed for the overexpression of membrane proteins(Miroux and Walker1996).The mutant C41(DE3)and double mutant C43(DE3)could improve expression level of membrane ,b subunit of F1F o ATP syntha(Arechaga et al.2000).
Periplasmic ,proteins with disulfide bonds)
Export of proteins from bacterial cytoplasm is widely employed(Georgiou and Segatori2005).It could be ud
li promoter systems that are in u for heterologous protein production and their characteristics
Expression system bad on Induction(range of inductor)Level of expression Key features Original
reference
lac promoter Addition of IPTG0.2mM
(0.05–2.0mM)Low level up to
middle
Weak,regulated suitable for gene
products at very low intracellular level
Gronenborn
(1976)
Comparatively expensive induction
trc and tac promoter Addition of IPTG0.2mM
(0.05–2.0mM)Moderately high High-level,but lower than T7system Brosius et al.
(1985)
Regulated expression still possible
Comparatively expensive induction
High basal level
T7RNA polymera Addition of IPTG0.2mM
(0.05–2.0mM)Very high Utilizes T7RNA polymera Studier and
Moffatt
(1986)
High-level inducible over expression
T7lac system for tight control of
induction needed for more toxic clones
Relative expensive induction
Basal level depends on ud strain(pLys)
Phage promoter p L Shifting the temperature
from30to42°C(45°C)Moderately high Temperature-nsitive host required Elvin et al.
(1990)
Less likelihood of“leaky”uninduced
expression
Basal level,high basal level by
temperatures below30°C职中可以考大学吗
意识形态学习No inducer
tetA promoter/operator Anhydrotetracycline200μg/l Variable from
middle to high
level Tight regulation Skerra(1994) Independent on metabolic state
Independent li strain
Relative inexpensive inducer
Low basal level
araBAD promoter (P BAD)Addition of L-arabino
0.2%(0.001–1.0%)
Variable from low
to high level
Can fine-tune expression levels in a
do-dependent manner
Guzman et al.
(1995)
Tight regulation possible
Low basal level
Inexpensive inducer
rhaP BAD promoter L-rhamno0.2%Variable from low
to high level Tight regulation Haldimann et
al.(1998) Low basal activity
Relative expensive inducer
for simplifying downstream and N-terminal processing.It generally facilitates correct folding.Proteins with disulfide bonds should be especially accumulated in the periplasm becau the cytoplasm is too reducing.This oxidation process is catalyzed by disulfide binding proteins DsbA, DsbB,DsbC,DsbD,a
nd peptidyl-prolyl isomeras SurA, RotA,Fk1B,and FkpA(Joly and Swartz1994;Shokri et al. 2003).To crete recombinant proteins into the periplasmic space,a fusion with a leader peptide at the N terminus is necessary.Generally,types I and II mechanisms are commonly ud for protein cretion li K-12or B strains.The following signal peptides are in u for cretion of heterologously expresd proteins:Lpp,LamB, LTB,MalE,OmpA,OmpC,OmpF,OmpT,PelB,PhoA, PhoE,SpA,and Tat signal peptides(reviewed in Choi and Lee2004;Mergulhao et al.2005).
In contrast to the signal peptides of the c-system,the Tat signal peptides of the twin-arginine translocation pathway transported folded proteins across the inner membrane.This could be an advantage becau chaperons, who could be responsible of correct folding,are mainly located in the cytoplasm(reviewed in Choi and Lee2004).
Proteins located in the periplasm can be creted into the culture medium with an osmotic shock or cell wall permeabilization.Cell lysis is not necessary,so that cytoplasmic proteins cannot contaminate the purification process(Shokri et al.2003).Most problems for heterolo-gous protein production in the periplasmic space are incomplete translocation across the inner membrane (Baneyx1999),proteolytic degradation(Huang et al. 2001),and insufficient capacity of the export machinery (Mergulhao and Monteiro2004;Ronberg1998).When this capacity is overwhelmed,the e
xcess of expresd recombinant protein is likely to accumulate in inclusion bodies.For optimization of the expression level,a careful balance of the promoter strength and gene copy number is necessary.
Alternatively,proteins with disulfide bonds can be overexpresd in the cytoplasm by thioreducta-deficient (trxB)and glutathione reducta(gor)deficient strains (Bestte et al.1999;Ritz et al.2001;e Table1).
Many promoter systems li are described as tools for protein expression,but only a few of them are commonly ud(Table4).A uful promoter must be strong,has a low basal expression ,it is tightly regulated),must be easily transferable to other E.coli strains,and the induction must be simple and cost-effective, and should be independent on the commonly ud ingredients of culturing media.lac,tac,and trc promoter systems不该放弃
li lacto utilization is one of the well-known regulation mechanism.Many promoters were constructed from lac-derived regulatory element(Polisky et al.1976). The lac promoter is rather weak and rarely ud for high-level production of recombinant proteins.But the leakiness may be an advan
tage for the production of membrane proteins or other gene products that are toxic to the cells. The synthetic tac(De Boer et al.1983)and trc promoter (Brosius et al.1985),which consists of the−35region of the trp promoter and the−10region of the lac promoter, only differ by1bp in the length of spacer domain.The tac promoter is at least five times more efficient than the lac UV5promoter(Amann et al.1983).tac and trc promoters are strong and allow the accumulation of up to 15–30%of total cell protein.Induction of all the promoters including lac promoter could be achieved by adding non-hydrolysable lacto analog isopropyl-β-D-1-thiolgalactopyranoside(IPTG)(Table4).The strength of the tac and trc promoter can be a problem to express successfully recombinant proteins,which are toxic to the cells.Especially,over expression of membrane proteins with the strong trc promoter can result into degradation (Quick and Wright2002).All three promoter systems are regulated by catabolite repression and the metabolic state, which is reprented by the cyclic AMP level.To reduce the problems,the T7RNA polymera system was developed.
T7RNA polymera system
One of the most widely ud expression systems is the T7 RNA polymera system(Studier and Moffatt1986).The T7RNA polymera elongates chains about five times faster li RNA polymera.The two polymeras recognize completely different promoters and can com-pletely be
ud lectively.Thousands of homologous and heterologous proteins were successfully expresd to high levels in E.coli BL21(DE3).The gene of the T7 polymera is in strain BL21(DE3)chromosomal located and under the control of a lac promoter derivate L8-UV5 lac(Grossman et al.1998;Pan and Malcolm2000).The L8-UV5lac promoter contains point mutations that distinguish it from the wild-type lac promoter.Two point mutations are in the−10region,which increa promoter strength and decrea its dependence on cyclic AMP,and its receptor protein called CAP.A third point mutation creates a stronger promoter that is less nsitive to gluco (Grossman et al.1998).This allows strong induction of T7 RNA polymera with IPTG(Table4)even in the prence of gluco.Nevertheless,there is basal expression of T7
RNA polymera for induction(Dubendorff and Studier 1991;Pan and Malcolm2000),which can lead to problems if the produced proteins are toxic for the host cells(Studier 1991).One way to reduce basal level is to work with host strains containing pLysS or pLysE vectors.The vectors express the T7lysozyme,a natural inhibitor of T7RNA polymera(Moffatt and Studier1987).The basal level of BL21compared to BL21(pLysS)is nearly ten times lower but a residual background is still prent which can also result in the described problems.A further problem is that T7lysozyme is a bifunctional protein.It cuts a specific bond in the petidoglycan layer of li cell wall.This ems
to be the reason why growth rates with strains containing pLysS or pLysE decread.Another aspect is that T7lysozyme can reduce expression following induc-tion,resulting in markedly lower yields(Studier1991).The reduction of basal T7RNA polymera level can also be achieved by adding0.5–1.0%gluco into the medium (Grossman et al.1998).This effect of catabolite repression is much stronger for BL21(DE3)than for BL21(DE3) pLys(Pan and Malcolm2000).
Phage promoter p L
Another approach that is widely ud for protein over-expression is to place a gene under the control of a regulated phage promoter p L,which has moderately high expression level.The genes must be cloned downstream of a tightly regulated phage promoter p L that is regulated by the cI repressor.The temperature-nsitive cI857repressor allows control of gene expression by changing the growth temperature instead of induction by a chemical inducer.At 30°C,the cI857repressor is functional and it turns off expression,but at42°C,the repressor is inactivated so that gene expression is induced(Elvin et al.1990;Love et al. 1996).Functional protein could be also purified by a shift up to45°C(Armarego et al.1989).It is interesting to note that the phage promoter p L is constitutive at low temperatures becau cI857repressor becomes fully active at29°C and higher(Lowman and Bina1990).This constitutive system can be ud for the production of proteolytically susceptible prote
ins at low temperatures(Menart et al.2003).
tetA promoter/operator system
The tetA promoter/operator is uful for the tight regulation, high-level synthesis of foreign gene product li (Skerra1994).It is regulated by the tetR repressor,which is not coded by li gene.The system is independent on the li ,B or K12derivates.Cells are induced by low concentration of anhydrotetracycline (Table4).Anhydrotetracycline binds the promoter nearly
35-fold higher than tetracycline and its antibiotic activity is
100-fold lower(Degenkolb et al.1991).Concentrations
starting from50ng/ml cau full induction and have no
effect on the growth li(Lutz and Bujard1997).Lots
of proteins were successfully expresd with this system,
<,F ab fragments or toxins(Skerra1994).The maximum
of induction is ca.100-fold over uninduced level(Korpela
et al.1998).In contrast to other systems,the basal level is
very low and independent on li strain and the
metabolic state.
L-arabino inducible P BAD promoter
The promoter araBAD(P BAD)of the arabino operon is a
uful alternative for heterologous protein production in
promoter,its expression is controlled by the AraC activator.
Expression is induced to high levels on media containing
arabino.Moreover,expression is tightly shut off on media
containing gluco but lacking arabino.In general,genes
cloned under the control of the araC-P BAD promoter system
are efficiently represd.However,the levels of expression
of P BAD controlled genes may not be zero in the represd
state(Guzman et al.1995).However,araC-P BAD promoter
system allows high-level expression,tightly regulated
protein expression,and very inexpensive induction with L-arabino.In bacterial strains,which are deleted for ara genes,expression of the cloned gene reaches maximal
induction upon adding0.001%L-arabino.In strains
that can ferment,1%is necessary for full induction
(Mayer1995).L-arabino acts as inducer with the activator
AraC in the positive control of the arabino regulon
(Haldimann et al.1998).It was proven to function under
high cell density fermentation,but the protein quality was
shown to be lower than in low densities(DeLisa et al.
1999).
L-rhamno inducible rhaP BAD promoter
The rhamno-inducible rhaP BAD promoter is also an
interesting tool for tightly regulated heterologous protein
production in E.coli.The regulon of the operon is
孕妇能吃什么described in detail(Egan and Schleif1993).In principle, L-rhamno acts as an inducer with the activator RhaR for synthesis of RhaS,which in turn acts as an activator in the positive control of the rhamno regulon(Haldimann et al. 1998).The L-rhamno regulons are also regulated by catabolite repression.High cell density fermentation for production L-N-carbamoyla using the rhaBAD promoter is reported(Wilms et al.2001).
