Refolding of Therapeutic
`Proteins Produced in
`Escherichia coli as
`Inclusion Bodies
`
`Satoru Misawa1
`Izumi Kumagai2
`1 Pharmaceutical and
`Biotechnology Laboratory,
`Japan Energy Corporation,
`3-17-35 Niizo-Minami,
`Toda-shi, Saitama 335-8502,
`Japan
`
`2 Department of Biomolecular
`Engineering,
`Graduate School of
`Engineering,
`Tohoku University,
`Aoba-yama 07,
`Sendai, 980-8579, Japan
`
`Abstract: Overexpression of cloned or synthetic genes in Escherichia coli often results in the
`formation of insoluble protein inclusion bodies. Within the last decade, specific methods and
`strategies have been developed for preparing active recombinant proteins from these inclusion
`bodies. Usually, the inclusion bodies can be separated easily from other cell components by
`centrifugation, solubilized by denaturants such as guanidine hydrochloride (Gdn-HCl) or urea, and
`then renatured through a refolding process such as dilution or dialysis. Recent improvements in
`renaturation procedures have included the inhibition of aggregation during refolding by application
`of low molecular weight additives and matrix-bound renaturation. These methods have made it
`possible to obtain high yields of biologically active proteins by taking into account process
`parameters such as protein concentration, redox conditions,
`temperature, pH, and ionic
`strength. © 1999 John Wiley & Sons, Inc. Biopoly 51: 297–307, 1999
`
`Keywords:
`refolding; inclusion body; renaturation; recombinant protein; high-level expression;
`therapeutic protein
`
`INTRODUCTION
`
`Major advances in genetic engineering have resulted
`in the development of bacterial expression systems,
`particularly those in Escherichia coli, capable of pro-
`ducing large amounts of proteins
`from cloned
`genes.1,2 The supply of many valuable proteins that
`have potential clinical or industrial use, such as hor-
`mones, cytokines, and enzymes, is often limited by
`their low natural availability. Initially, this approach
`
`Correspondence to: Satoru Misawa; email:
`energy.co.jp
`Biopolymers (Peptide Science), Vol. 51, 297–307 (1999)
`© 1999 John Wiley & Sons, Inc.
`
`s.misawa@j-
`
`employing E. coli seemed to guarantee an unlimited
`supply of recombinant proteins. For example, recom-
`binant DNA technology has facilitated the efficient
`production of therapeutic-grade proteins such as in-
`sulin,3 growth hormone (GH),4 and interferon (IFN).5
`However, high-level expression of recombinant pro-
`teins in E. coli often results in the formation of insol-
`uble and inactive aggregates known as inclusion bod-
`ies.6,7 To obtain biologically active recombinant pro-
`teins from inclusion bodies, it is necessary to develop
`
`CCC 0006-3525/99/040297-11
`
`297
`
`APOTEX EX1044
`
`Page 1
`
`

`
`298
`
`Misawa and Kumagai
`
`a simple and efficient procedure for renaturation of
`these proteins.8,9
`The formation of inclusion bodies offers several
`advantages for the production of recombinant pro-
`teins. These proteins may be unstable in the cyto-
`plasm of E. coli due to proteolysis and may be toxic
`to the host cell in the native conformation. Under
`appropriate conditions, the recombinant protein de-
`posited in inclusion bodies amounts to about 50% or
`more of the total cellular protein. Because inclusion
`bodies have a relatively high density,10 they can be
`isolated from the cellular proteins by centrifugation,
`and the purity of the resulting preparation may reach
`90% under optimal conditions. Therefore, the produc-
`tion of many human therapeutic proteins as inclusion
`bodies is a cost-effective downstream process.2,11,12
`Recent advances in procedures for refolding inclusion
`body proteins have made it possible to obtain large
`amounts of authentic human proteins for therapeutic
`use. This review summarizes the improvements that
`have been made in the in vitro refolding of therapeu-
`tically relevant proteins containing disulfide bonds
`
`after production at high yield as inclusion bodies in
`E. coli.
`
`HIGH-LEVEL EXPRESSION OF
`RECOMBINANT PROTEINS IN E. COLI
`
`The expression of cloned genes in E. coli for the
`production of recombinant proteins has provided a
`valuable system for developing therapeutic proteins
`such as human insulin and human GH. Many success-
`ful E. coli expression systems have been described
`and are available from a variety of academic and
`commercial sources. Therefore, E. coli expression
`systems are suitable for the industrial-scale produc-
`tion of recombinant proteins. A number of criteria
`must be considered when optimizing conditions for
`the high-level expression of a recombinant protein.
`These include the stability of the mRNA,13 the effi-
`ciency of transcription directed from a strong pro-
`moter,14 the efficiency of protein synthesis (transla-
`tion),15 the formation of inclusion bodies, and the
`
`Table I High-Level Expression of Recombinant Proteins for Therapeutic Use in E. colia
`
`Recombinant Protein
`
`hEGF
`Human insulin
`hIFN-b
`hIFN-g
`Human prourokinase
`hGH
`hGH
`hIGF-I
`hIGF-I
`ht-PA
`ht-PA
`hTIMP-1
`hTIMP-2
`Human calcitonin
`hG-CSF derivative
`hbFGF derivative
`hIL-2
`hIL-6
`Human glucagon
`Hirudin
`Hirudin
`Arginine deiminase
`Humanized F(ab9)2
`Chimeric Fab L-chain
`
`Mode of
`Expression
`
`Level of
`Expression
`(% of Total Protein)
`
`Level of
`Production
`(mg/L)
`
`Inclusion
`Body
`Formation
`
`Promoter
`
`Reference
`
`Fusion
`Fusion
`Direct
`Direct
`Direct
`Direct
`Secretion
`Fusion
`Secretion
`Direct
`Secretion
`Direct
`Fusion
`Fusion
`Direct
`Direct
`Direct
`Direct
`Fusion
`Fusion
`Secretion
`Direct
`Secretion
`Secretion
`
`NE
`20
`NE
`40
`6
`NE
`14
`20
`30
`10
`NE
`15
`5
`NE
`15
`NE
`20
`20
`34.5
`18
`NE
`20
`NE
`NE
`
`60
`NE
`20
`NE
`NE
`169
`25/A550
`1240
`8500
`460
`0.18
`NE
`NE
`478
`2800
`1700
`700
`NE
`42
`200
`1000
`400
`2000
`2880
`
`trp
`lac
`trp
`trp
`trp
`trp
`phoA
`trp
`phoA
`lPL
`araB
`T7
`T7
`lac
`trp
`T7
`trp
`trp
`trp
`trp
`trp
`tac
`phoA
`tac
`
`1
`1
`2
`1
`1
`1
`2
`1
`1
`1
`2
`1
`1
`1
`1
`2
`1
`1
`1
`1
`2
`1
`2
`1
`
`62
`3
`63
`64
`65
`66
`67
`68
`69
`45, 47
`70
`51
`57
`71
`72
`73
`74
`75
`76
`77
`78
`25
`79
`80
`
`a hEGF, human epidermal growth factor; hIFN, human interferon; hGH, human growth hormone; hIGF-I, human insulin-like growth
`factor-I; ht-PA, human tissue-type plasminogen activator; hTIMP, human tissue inhibitor of metalloproteinases; hG-CSF, human granulocyte
`colony-stimulating factor; hbFGF, human basic fibroblast growth factor; hIL, human interleukin. NE, not estimated.
`
`Page 2
`
`

`
`Refolding of Therapeutic Proteins
`
`299
`
`Because of their refractile character, they can be ob-
`served directly in the living host cell by phase-con-
`trast microscopy. We have shown that porcine muscle
`adenylate kinase is expressed in E. coli as inclusion
`bodies at high levels up to 40% of total cellular
`protein (Figure 1).18 Because inclusion bodies are
`characterized by a relatively high specific density,
`they can be harvested after cell lysis by centrifugation
`at moderate rotor speeds.19 To purify the inclusion
`bodies from their associated impurities, they can be
`washed with detergents such as Triton X-100, deoxy-
`cholate, or a low molar concentration of cha-
`otroph.2,20 However, it should be kept in mind that an
`excessively high concentration of urea or Gdn-HCl
`will lead to solubilization of the inclusion bodies
`themselves. Table II shows several examples of dif-
`ferent washing solutions used for the purification of
`inclusion bodies.11 On average, the purity of the in-
`clusion body preparation may reach 90% under opti-
`mal conditions.
`Next, the purified inclusion bodies must be solu-
`bilized by strong denaturants such as 6M Gdn-HCl or
`8M urea. For this purpose, Gdn-HCl is usually pref-
`erable to urea for two reasons.8 First, Gdn-HCl is a
`rather strong chaotroph, which may allow solubiliza-
`tion of extremely aggregated inclusion bodies that are
`resistant to solubilization by urea. Second, urea solu-
`tions may contain isocyanate, leading to carbamyla-
`tion of free amino groups of the polypeptide, espe-
`cially upon long-term incubation at alkaline pH val-
`ues.21 In the case of cysteine-containing proteins, the
`isolated inclusion bodies usually contain non-native
`intramolecular and intermolecular disulfide bonds,22
`which reduce the solubility of the inclusion bodies in
`the absence of reducing agents such as dithiothreitol
`(DTT), dithioerythritol, glutathione (GSH), cysteine,
`cystamine, or b-mercaptoethanol. Addition of these
`thiol reagents in combination with chaotrophs allows
`reduction of the disulfide bonds by thiol-disulfide
`
`FIGURE 1 Electron micrograph of inclusion bodies con-
`taining recombinant porcine muscle adenylate kinase ex-
`pressed in E. coli. The cells were harvested and washed with
`1% NaCl and stained with 1% uranyl acetate. The dense
`material shown in the elongated E. coli is the inclusion
`bodies.
`
`susceptibility of the product to proteolysis.16,17 All of
`these criteria must be considered for each product
`individually. Representative examples of the high-
`level expression of recombinant proteins for therapeu-
`tic use are presented in Table I. In the majority of
`cases, the expressed proteins are in an insoluble form.
`A number of human proteins expressed in E. coli
`IFN-g,
`directly, e.g., GH,
`interleukin-2 (IL-2),
`prourokinase, and tissue-type plasminogen activator
`(t-PA), or as fusion proteins, e.g., proinsulin, calcito-
`nin, and insulin-like growth factor-I (IGF-I), have
`been shown to exist as aggregates or inclusion bodies
`(see Table I for references).
`
`ISOLATION AND SOLUBILIZATION OF
`INCLUSION BODIES
`
`Inclusion bodies obtained by cytosolic overexpression
`of a recombinant protein are large, spherical particles.
`
`Table II Purification of Inclusion Bodies by Different Washing Solutionsa
`
`Recombinant Protein
`
`Mode of Expression
`
`Washing Solution
`
`Reference
`
`Human prourokinase
`ht-PA
`ht-PA
`hM-CSF
`Arginine deiminase
`hIGF-I
`Bovine GH
`Prochymosin
`HRP
`
`Direct
`Direct
`Direct
`Direct
`Direct
`Fusion
`Direct
`Direct
`Direct
`
`0.1% Triton X-100
`5M urea, 2% Triton X-100
`1% Triton X-100, 1% b-DPG
`2% Triton X-100
`4% Triton X-100
`0.5% Sarcosyl
`2% deoxycholate
`0.5% Triton X-100
`2M urea
`
`65
`45
`46
`81
`25
`68
`82
`83
`84
`
`a ht-PA, human tissue-type plasminogen activator; hM-CSF, human macrophage colony-stimulating factor; hIGF-I, human insulin-like
`growth factor-I; GH, growth hormone; HRP, horseradish peroxidase C. b-DPG, octyl-b-D-thioglucopyranoside.
`
`Page 3
`
`

`
`300
`
`Misawa and Kumagai
`
`FIGURE 2 Effects of temperature and pH on renaturation of recombinant arginine deiminase
`(r-AD). The lyophilized inclusion bodies derived from 10 mL of cultured E. coli cells were
`solubilized in 1 mL 50 mM Tris HCl (pH 8.5) containing 6M Gdn-HCl and 10 mM DTT and
`incubated at 37°C for 1 h. The solubilized proteins were diluted rapidly with 100 mL of 10 mM
`potassium phosphate buffer, and the solutions were stirred at various temperature for 45 h at pH 7.0
`(A) and at various pH values for 45 h at 15°C (B). The extent of r-AD renaturation was monitored
`by measuring the AD activity.
`
`exchange.8,23 Various experimental protocols used for
`the solubilization of inclusion bodies have been com-
`pared by Fischer et al.11 If the purity of the solubilized
`inclusion bodies is low, purification can be achieved
`by reverse-phase high-performance liquid chromatog-
`raphy, gel filtration, or ion-exchange chromatography
`in the presence of a denaturant.
`
`RENATURATION OF RECOMBINANT
`PROTEINS
`
`To obtain the correctly folded proteins after solubili-
`zation of the inclusion bodies, excess denaturants and
`reducing thiol reagents have to be removed, and the
`reduced proteins transferred to oxidizing conditions.
`Renaturation of solubilized inclusion bodies is initi-
`ated by removal of the denaturant by either dilution or
`dialysis. The efficiency of renaturation depends on the
`competition between correct folding and aggrega-
`tion.24 To slow down the aggregation process, refold-
`ing is usually performed at low protein concentra-
`tions, within the range 10 –100 mg/mL. Furthermore,
`the renaturation conditions must be carefully opti-
`mized with regard to external parameters such as
`temperature, pH, and ionic strength for each individ-
`ual protein.9,23
`
`Both folding and association of proteins depend
`strongly on temperature and pH. For example, we
`have shown that recombinant Mycoplasma arginine
`
`FIGURE 3 Time course of r-AD renaturation. The 6M
`Gdn-HCl-solubilized inclusion bodies containing r-AD
`were diluted rapidly 100-fold with 10 mM potassium phos-
`phate buffer (pH 7.0) and the solutions were stirred at 4°C
`(h), 15°C ((cid:130)), and 25°C (E) for 0 –90 h. The extent of
`r-AD renaturation was monitored by measuring the AD
`activity at various time intervals.
`
`Page 4
`
`

`
`Refolding of Therapeutic Proteins
`
`301
`
`Table III Optimal Conditions for Renaturation of Proteins from Inclusion Bodiesa
`
`Recombinant Protein
`
`hIFN-g
`Human prourokinase
`Prochymosin
`Human angiogenin
`Bovine GH
`Arginine deiminase
`Porcine ADK
`hIGF-I
`Salmon GH
`
`Solubilizing
`Reagent
`
`Refolding
`Method
`
`6M Gdn-HCl
`6M Gdn-HCl
`8M urea
`7M Gdn-HCl
`6M Gdn-HCl
`6M Gdn-HCl
`6M Gdn-HCl
`6M Gdn-HCl
`7M urea
`
`Dilution
`Dilution
`Dialysis
`Dilution
`Dialysis
`Dilution
`Dialysis
`Dilution
`Dilution
`
`pH
`
`7
`8.8
`10.5
`8.5
`8.5
`7.5
`7.4
`8
`8
`
`Temperature
`(°C)
`
`4
`15
`Room temp.
`4
`Room temp.
`15
`4
`25
`4
`
`Time
`(h)
`
`Overnight
`24
`6
`24
`24
`90
`Overnight
`72
`One day
`
`Reference
`
`85
`65
`86
`87
`82
`25
`18
`68
`88
`
`a hIFN, human interferon; GH, growth hormone; ADK, adenylate kinase; hIGF-I, human insulin-like growth factor-I.
`
`deiminase, developed as an antitumor agent, is effi-
`ciently renatured at 15°C and at pH 7.5 by 100-fold
`rapid dilution of inclusion bodies solubilized with 6M
`Gdn-HCl (Figure 2).25 The time required for complete
`renaturation may extend over a range of seconds to
`days. Upon renaturation of antibody Fab fragments
`from inclusion bodies, it was shown that the amount
`of functional antibody increased over 100 h.26 Also,
`renaturation of recombinant Mycoplasma arginine de-
`iminase exhibited exceedingly slow kinetics (over
`90 h) even at 15°C by the rapid dilution method
`(Figure 3). Table III shows several of the optimal
`conditions for renaturation of proteins from inclusion
`bodies.11
`Most secretory proteins contain disulfide bonds in
`their native state. If a target protein contains disulfide
`bonds, the renaturation buffer has to be supplemented
`with a redox system. Addition of a mixture of the
`reduced (RS2) and oxidized (RSSR) forms of low
`molecular weight thiol reagents such as glutathione,
`cysteine, and cysteamine (molar ratios of reduced to
`oxidized compounds 5 : 1 to 10 : 1, respectively)
`
`usually provides the appropriate redox potential to
`allow formation and reshuffling of disulfides.9,27,28
`These systems increase both the rate and yield of
`renaturation/reoxidation by facilitating rapid reshuf-
`fling of incorrect disulfide bonds according to.23,29
`
`In order to accelerate thiol– disulfide exchange, the
`pH of the renaturation buffer should be at the upper
`limit that still allows the protein to form its native
`structure. In order to prevent fortuitous oxidation of
`thiols by molecular oxygen, which is catalyzed by
`ions (e.g., Cu21), EDTA
`trace amounts of metal
`should be added to the buffer solutions. Reoxidation
`of protein disulfide bonds is performed by dilution of
`the reduced solubilized inclusion bodies in the
`“oxido-shuffling” system.23,29 Table IV summarizes
`
`Table IV Optimal Conditions for Renaturation and Reoxidation of Proteins from Inclusion Bodies by the
`Glutathion System
`
`Recombinant Protein
`
`Fab-fragment
`ht-PA
`Trancated ht-PA
`Trancated hM-CSF
`hIL-2
`hIL-4
`hIL-6
`hTIMP-1
`Trancated hTIMP-2
`
`Number of
`Disulfide
`Bonds
`
`Reduced
`Glutathione
`(mM)
`
`Oxidized
`Glutathione
`(mM)
`
`5
`17
`9
`9
`1
`3
`2
`6
`3
`
`5
`0.5
`2
`0.5
`10
`2
`0.01
`2
`0.78
`
`0.5
`0.3
`0.2
`0.1
`1
`0.2
`0.002
`0.2
`0.44
`
`pH
`
`8
`8.75
`8.6
`8.5
`8
`8
`8.5
`8
`9.75
`
`Temperature
`(°C)
`
`Time
`(h)
`
`Reference
`
`10
`15
`20
`4
`Room temp.
`Room temp.
`22
`4
`25
`
`150
`24
`24
`48
`16
`4
`16
`16
`2
`
`26
`45
`32
`81
`89
`90
`91
`61
`59
`
`Page 5
`
`

`
`302
`
`Misawa and Kumagai
`
`FIGURE 4 Schematic drawing of ht-PA showing the finger (F), growth factor (G), kringle (K1,
`K2), and serine protease (P) domains. Solid bars indicate potential disulfide bridges based on
`homology with other proteins.
`
`several of the conditions for renaturation of proteins
`from inclusion bodies by the glutathione reoxidation
`system.11
`In addition to the control of parameters such as
`temperature, pH, or redox conditions, the presence of
`low molecular weight compounds in the renaturation
`buffer may have a marked effect on the yield of
`renaturation.8,9,30 A large series of low molecular
`weight additives are, in certain cases, very efficient
`refolding enhancers:
`for examples, nondenaturing
`concentrations of chaotrophs such as urea or Gdn-HCl
`are essential for the renaturation of reduced chymo-
`trypsinogen A.31
`The most popular additive is L-arginine.8,9 In the
`case of human t-PA29 or its truncated form,32 the yield
`of renaturation is markedly increased in the presence
`of 0.5M L-arginine, whereas in its absence almost no
`reactivity is observed. The positive effect of L-argi-
`nine on renaturation efficiency has also been con-
`firmed for various other proteins such as antibody Fab
`fragments,26 single-chain immunotoxins,33 and sin-
`gle-chain Fv fragments.34 The mechanism by which
`L-arginine supports renaturation is still unknown. Al-
`though L-arginine contains a guanidino group, it does
`not destabilize the native folded structure as strongly
`as Gdn-HCl. The beneficial effect of L-arginine on
`protein refolding probably originates from increased
`solubilization of folding intermediates.9
`
`In the case of bovine carbonic anhydrase B, stoi-
`chiometric amounts of polyethylene glycol (PEG) sig-
`nificantly enhanced the recovery of active protein by
`reducing aggregation.35 Furthermore, three recombi-
`nant human proteins—deoxyribonuclease, t-PA, and
`IFN-g—were refolded efficiently in the presence of
`PEG (MW 3350).36 Therefore, PEG has significant
`potential for enhancing the recovery of active proteins
`from inclusion bodies.
`Likewise, increased solubilization of folding inter-
`mediates can explain the positive effect of detergents
`on the refolding yield. Using lauryl-maltoside,
`CHAPS (3-[3-chloramidopropyl]dimethylammonia-
`1-propane sulfonate) or some other detergents during
`renaturation, the yield of renatured protein can be
`improved.37,38 Refolding in the presence of a deter-
`gent followed by addition of cyclodextrin has been
`claimed to be analogous to a molecular chaperone
`system in terms of function.39 To prevent aggregation
`during refolding, other techniques such as renatur-
`ation in reversed micelles40 or in aqueous two-phase
`systems41 have also been explored.
`Another possibility for suppressing unspecific in-
`termolecular interactions is the coupling of the dena-
`tured protein to a matrix. When denatured a-glucosi-
`dase fused to a polyarginine tag was bound to heparin-
`Sepharose, renaturation under conditions allowing the
`protein to remain bound to the matrix resulted in high
`
`Page 6
`
`

`
`yields of active protein even at a high gel load of up
`to 5 mg/mL.42 Another matrix used for this kind of
`renaturation is Ni21-nitrilotriacetic acid (NTA) resin,
`which was originally developed for efficient protein
`purification. After binding the denatured protein to the
`matrix via a His tag, the column is equilibrated with
`renaturation buffer, and the refolded protein can be
`eluted by imidazole using a pH gradient.43
`
`CASE STUDIES OF THE PRODUCTION
`OF THERAPEUTIC PROTEINS
`
`Example 1: t-PA
`
`t-PA is a serine protease that has an important func-
`tion in the fibrinolytic system. It catalyzes the con-
`version of plasminogen to plasmin in the presence of
`a fibrin clot. Human t-PA (ht-PA) is a glycosylated
`single-chain polypeptide of 527 amino acids, and con-
`tains 17 disulfide bonds.44 A schematic drawing of
`ht-PA is shown in Figure 4. The molecule comprises
`five distinct structural domains: a finger domain, an
`epidermal growth factor-like domain, two kringle do-
`mains, and a C-terminal protease domain. The affinity
`of t-PA for fibrin, and the 100 –200-fold increase in its
`activity in the presence of fibrin make it an attractive
`thrombolytic agent because it should generate plasmin
`locally at the fibrin surface and achieve thrombolysis
`without systemic activation of plasminogen.45 Several
`attempts to produce ht-PA or a truncated form in E.
`coli have been reported.29,32,45– 48 Recombinant t-PA
`(rt-PA) accumulated as inclusion bodies in the cyto-
`plasm. t-PA best exemplifies the challenges associ-
`ated with the production of multidisulfide complex
`proteins in E. coli by refolding from inclusion bodies.
`Sarmientos et al. reported that ht-PA was produced
`as an insoluble, aggregated form in E. coli (5–10% of
`total cellular protein), with a yield of 460 mg/L fer-
`mentation broth.45,47 The inclusion bodies obtained
`after centrifugation of the sonicated extract were first
`washed with a solution containing 5M urea and 2%
`Triton X-100, then dissolved in 7M Gdn-HCl and 50
`mM b-mercaptoethanol. The reducing agent was then
`removed by dialysis, and the solubilized inclusion
`bodies were diluted at least 50-fold into a renaturation
`buffer containing 2.5M urea, 10 mM lysine, and a
`redox coupler at appropriate concentrations (0.5 mM
`GSH and 0.3 mM GSSG), under carefully controlled
`incubation conditions (15°C, no air). After renatur-
`ation, Tween 80 was added to a final concentration of
`0.01%, and the t-PA activity was measured. As a
`result, t-PA activity corresponding to a concentration
`of 2–5 mg/mL of fully active t-PA was consistently
`detected in the renaturation solution. Further charac-
`
`Refolding of Therapeutic Proteins
`
`303
`
`FIGURE 5 Schematic diagram for the production of rena-
`tured rt-PA from inclusion bodies expressed in E. coli.10,47
`
`terization and purification of the renatured rht-PA
`strongly suggested that it was a fully active enzyme,
`very similar to natural t-PA, despite the lack of gly-
`cosylation. The purification yield of 2.8% for the
`overall process reflects a 20% step yield for the re-
`folding operation, and a 56% yield for the subsequent
`ultrafiltration step at a refolding concentration of 2.43
`mg rt-PA/L. Details of the scheme of rt-PA produc-
`tion from E. coli are shown in Figure 5.47
`However, Grunfeld et al. reported that a 90% yield
`for the refolding process was achieved within an
`optimal concentration range of 2.6 –3.7 mg rt-PA/L of
`reactivation mixture.46 They also reported that ab-
`sence of arginine in the reactivation mixture resulted
`
`Page 7
`
`

`
`304
`
`Misawa and Kumagai
`
`FIGURE 6 Amino acid sequence and two-dimensional representation of the structure of
`hTIMP-2.
`
`in yields that were only 10% of those achieved when
`arginine was present at a final concentration of 0.25M.
`
`Example 2: Tissue Inhibitors of
`Metalloproteinases (TIMPs)
`
`Tissue inhibitors of metalloproteinases (TIMPs) play
`important roles in regulating the activities of matrix
`metalloproteinases (MMPs),49,50 a family of enzymes
`responsible for the breakdown of connective tissue
`components. Inappropriate matrix breakdown is asso-
`ciated with a number of pathologies including peri-
`odontal disease, rheumatoid arthritis (RA), and tumor
`metastasis. The control of MMP activity is therefore
`an important therapeutic target, and TIMPs or low
`molecular weight MMP inhibitors could lead to new
`therapeutic strategies for the treatment of RA and
`cancer.51
`Three members of the TIMP family have been
`identified to date: TIMP-1, TIMP-2, and TIMP-3.52–54
`The TIMP proteins show a high degree of amino acid
`sequence similarity including the conservation of 12
`cysteinyl residues known to form 6 disulfide bonds.55
`TIMP-1 is a glycoprotein of 184 amino acids in
`length,52 whereas TIMP-2 and TIMP-3 are not glyco-
`sylated.53,54 The amino acid sequence of human
`
`TIMP-2 is shown in Figure 6. The overproduced
`recombinant TIMPs or the N-terminal domain of
`TIMPs accumulate as inclusion bodies in E. coli, and
`can be refolded into active forms.51,56 – 61 Particularly,
`recombinant
`human TIMP-2
`(rhTIMP-2)
`and
`rhTIMP-3 have been prepared by controlled refolding
`in the solid phase using an ion metal affinity col-
`umn.57,60
`Negro et al. reported that rhTIMP-2 was expressed
`in E. coli as a fusion protein with a 34-amino-acid
`NH2-linked tail containing 6 histidine residues, and
`that the TIMP-2 fusion protein immobilized on Ni21-
`NTA resin was refolded at a high concentration (1
`column.57 The
`mg/mL resin)
`in the
`refolded
`rhTIMP-2 was eluted from the resin with 250 mM
`imidazole. It showed specific binding and inhibitory
`activity against 72-kDa gelatinase (MMP-2).
`Furthermore, Negro et al. reported that rhTIMP-3
`was expressed in E. coli as a fusion protein with a
`36-amino-acid NH2-linked tail containing 6 histidine
`residues, and that the rhTIMP-3 was refolded in the
`Ni21-NTA column by very slowly removing the de-
`naturing and reducing agents (urea and b-mercapto-
`ethanol).60 The refolded rhTIMP-3 was treated with
`mild acid (70% formic acid) in order to remove the
`the
`NH2 tail from TIMP-3. After this treatment,
`
`Page 8
`
`

`
`rhTIMP-3 was loaded on the Ni21-NTA column in
`order to separate the cut (eluted) from the uncut
`(column retained) recombinant form. rhTIMP-3 from
`which the NH2 tail had been removed showed inhib-
`itory activities against both MMP-2 and MMP-9, and
`CD, fluorescence and second-derivative UV spectro-
`scopic analyses
`supported correct
`refolding of
`rhTIMP-3. Solid-phase refolding may prove to be
`useful for a variety of other proteins in which correct
`disulfide bridging plays a critical role.
`
`CONCLUSIONS
`
`Although recombinant DNA technology now permits
`burst synthesis of heterologous proteins in E. coli,
`these proteins often accumulate as insoluble inclusion
`bodies, and therefore solubilization and renaturation
`systems are necessary in order to obtain the fully
`active proteins with a native conformation.
`Similar to protein purification, protein refolding
`protocols still have to be developed on a case-by-case
`basis. Various procedures introduced in this review
`for in vitro refolding are available. Choosing the right
`procedure should allow renaturation of most recom-
`binant proteins deposited in inclusion bodies, giving
`high yields.
`Structural and functional analyses of proteins, es-
`pecially those for therapeutic or industrial applica-
`tions, require large amounts of recombinant proteins.
`The E. coli system for production of recombinant
`protein as inclusion bodies, together with a suitable
`renaturation procedure, provides an efficient avenue
`for meeting these requirements. In the near future, in
`vitro refolding of inclusion body proteins will become
`a powerful tool for commercial production of ex-
`tremely complex proteins in which multidisulfide
`bonds play a critical role.
`
`REFERENCES
`
`1. Kane, J. F.; Hartley, D. L. Trends Biotechnol 1988, 6,
`95–101.
`2. Marston, F. A. O. Biochem J 1986, 240, 1–12.
`3. Goeddel, D. V.; Kleid, D. G.; Bolivar, F.; Heyneker,
`H. L.; Yansura, D. G.; Crea, R.; Hirose, T.; Kraszewski,
`A.; Itakura, K.; Riggs, A. D. Proc Natl Acad Sci USA
`1979, 76, 106 –110.
`4. Goeddel, D. V.; Heyneker, H. L.; Hozumi, T.; Arent-
`zon, R.; Itakura, K.; Yansura, D. G.; Ross, M. J.;
`Miozzari, G.; Crea, R.; Seeburg, P. Nature (London)
`1979, 281, 544 –548.
`5. Staehelin, T.; Hobbs, D. S.; Kung, H.-F.; Pestka, S.
`Methods Enzymol 1981, 78, 505–511.
`
`Refolding of Therapeutic Proteins
`
`305
`
`6. Taylor, G.; Hoare, M.; Gray, D. R.; Marston, F. A. O.
`Bio/Technol 1986, 4, 553–557.
`7. Schein, C. Bio/Technol 1990, 8, 308 –317.
`8. Rudolph, R.; Lilie, H. FASEB J 1996, 10, 49 –56.
`9. Lilie, H.; Schwarz, E.; Rudolph, R. Curr Opin Biotech-
`nol 1998, 9, 497–501.
`10. Mukhopadhyay, A. Adv Biochem Eng Biotechnol
`1997, 56, 61–109.
`11. Fischer, B.; Sumner, I.; Goodenough, P. Biotechnol
`Bioeng 1993, 41, 3–13.
`12. Datar, R. V.; Cartwright, T.; Rosen, C.-G. Bio/Technol
`1993, 11, 349 –357.
`13. Gross, G.; Hollatz, I. Gene 1988, 72, 119 –128.
`14. Sawers, G.; Jarsch, M. Appl Microbiol Biotechnol
`1996, 46, 1–9.
`15. Gold, L.; Stormo, G. D. Methods Enzymol 1990, 185,
`89 –93.
`16. Gottesman, S. Methods Enzymol 1990, 185, 119 –
`129.
`17. Nygren, P.-A.; Sta¨hl, S.; Uhlen, M. Trends Biotechnol
`1994, 12, 184 –188.
`18. Hibino, T.; Misawa, S.; Wakiyama, M.; Maeda, S.;
`Yazaki, K.; Kumagai, I.; Ooi, T.; Miura, K. J Biotech-
`nol 1994, 32, 139 –148.
`19. Tayler, G.; Hoare, M.; Gray, D. R.; Marston, F. A. O.
`Bio/Technol 1986, 4, 553–557.
`20. Marston, F. A. O.; Hartley, D. L. Methods Enzymol
`1990, 182, 264 –276.
`21. Hagel, P.; Gerding, J. J. T.; Fieggen, W.; Bloemendal,
`H. Biochem Biophys Acta 1971, 243, 366 –372.
`22. Schoemaker, J. M.; Brasnett, A. H.; Marston, F. A. O.
`EMBO J 1985, 4, 775–780.
`23. Rudlph, R.; Bo¨hm, G.; Lilie, H.; Jaenicke, R. In Protein
`Function: A Practical Approach, 2nd ed.; Oxford: IRL
`Press, 1997; p 57–99.
`24. Zettlmeissl, G.; Rudolph, R.; Jaenicke, R. Biochemistry
`1979, 18, 5567–5571.
`25. Misawa, S.; Aoshima, M.; Takaku, H.; Matsumoto, M.;
`Hayashi, H. J Biotechnol 1994, 36, 145–155.
`26. Buchner, J.; Rudolph, R. Bio/Technol 1991, 9, 157–
`162.
`27. Ahmed, A. K.; Schaffer, S. W.; Wetlaufer, D. B. J Biol
`Chem 1975, 250, 8477– 8482.
`28. Wetlaufer, D. B.; Branca, P. A.; Chen, G.-X. Protein
`Eng 1987, 2, 141–146.
`29. Rudlph, R. In Modern Methods in Protein and Nucleic
`Acid Research; Tschesche, H., Ed.; Walter de Gruyter:
`New York, 1990; p 149 –172.
`30. Hofmann, A.; Tai, M.; Wong, W.; Glabe, C. G. Anal
`Biochem 1995, 230, 8 –15.
`31. Orsini, G.; Goldberg, M. E. J Biol Chem 1978, 253,
`34 –39.
`32. Kohnert, U.; Rudolph, R.; Verheijen, J. H.; Weening-
`Verhoeff, E. J. D.; Stern, A.; Opitz, U.; Martin, U.; Lill,
`H.; Printz, H.; Lechner, M.; Kresse, G.-B.; Buckel, P.;
`Fischer, S. Protein Eng 1992, 5, 93–100.
`33. Brinkmann, U.; Buchner, J.; Pastan, I. Proc Natl Acad
`Sci USA 1992, 89, 3075–3079.
`
`Page 9
`
`

`
`306
`
`Misawa and Kumagai
`
`34. Tsumoto, K.; Shinoki, K.; Kondo, H.; Uchikawa, M.;
`Juji, T.; Kumagai, I. J Immunol Methods 1998, 219,
`119 –129.
`35. Cleland, J. L.; Hedgepeth, C.; Wang, D. I. C. J Biol
`Chem 1992, 267, 13327–13334.
`36. Cleland, J. L.; Builder, S. E.; Swartz, J. R.; Winkler,
`M.; Chang, J. Y.; Wang, D. I. C. Bio/Technol 1992, 10,
`1013–1019.
`37. Tandon, S.; Horowitz, P. M. J Biol Chem 1987, 262,
`4486 – 4491.
`38. Cerletti, N.; McMaster, G.; Cox, D.; Schmitz, A.; Mey-
`hack, B. European Patent Application No. 0 433 225
`A1, 1990.
`39. Rozena, D.; Gellma, S. H. J Am Chem Soc 1995, 117,
`2373–2374.
`40. Hagen, A.; Hatton, T. A.; Wang, D. I. C. Biotechnol
`Bioeng 1990, 35, 955–965.
`41. Forciniti, D. J Chromatography A 1994, 668, 95–100.
`42. Stempfer, G.; Ho¨ll-Neugebauer, B.; Rudolph, R. Nature
`Biotechnol 1996, 14, 329 –334.
`43. Holzinger, A.; Phillips, K. S.; Weaver, T. E. Biotech-
`niques 1996, 20, 804 – 806.
`44. van Zonneveld, A. J.; Veerman, H.; MacDonald, M. E.;
`van Mourik, J. A.; Pannekoek, H. J Cell Biochem 1986,
`32, 169 –178.
`45. Sarmientos, P.; Duchesne, M.; Denefle, P.; Boiziau, J.;
`Fromage, N.; Delporte, N.; Parker, F.; Lelievre, Y.;
`Mayaux, J.-F.; Cartwright, T. Bio/Technol 1989, 7,
`495–501.
`46. Grunfeld, H.; Patel, A.; Shatzman, A.; Nishikawa, A.
`Appl Biochem Biotechnol 1992, 33, 117–138.
`47. Dater, R. V.; Cartwright, T.; Rosen, C.-G. Bio/Technol
`1993, 11, 349 –357.
`48. Fromage, N.; Denefle, P.; Cambou, B.; Duchesne, M.;
`Joyeux, C.; Kovarik, S.; Martin, J.; Imbault, F.; Uzan,
`A.; Cartwright, T. Fibrinolysis 1991, 5, 187–190.
`49. Woessner, J. F., Jr. FASEB J 1991, 5, 2145–2154.
`50. Docherty, A. J. P.; O’Connell, J.; Crabbe, T.; Angal, S.;
`Murphy, G. Trends Biotechnol 1992, 10, 200 –207.
`51. Kleine, T.; Bartsch, S.; Bla¨ser, J.; Schnierer, S.; Triebel,
`S.; Valentin, M.; Gote, T.; Tschesche, H. Biochemistry
`1993, 32, 14125–14131.
`52. Docherty, A. J. P.; Lyons, A.; Smith, B. J.; Wright,
`E. M.; Stephens, P. E.; Harris, T. J. R.; Murphy, G.;
`Reynolds, J. J. Nature (London) 1985, 318, 66 – 69.
`53. Boone, T.; Johnson, M. J.; DeClerck, Y. A.; Langley,
`K. E. Proc Natl Acad Sci USA 1990, 87, 2800 –2804.
`54. Apte, S. S.; Olsen, B. R.; Murphy, G. J Biol Chem
`1995, 270, 14313–14318.
`55. Williamson, R. A.; Marston, F. A. O.; Angal, S.; Kokli-
`tis, P.; Panico, M.; Morris, H. R.; Carne, A. F.; Smith,
`B. J.; Harris, T. J. R.; Freedman, R. B. Biochem J 1990,
`268, 267–274.
`56. Kohno, T.; Carmichael, D. F.; Sommer, A.; Thompson,
`R. C. Methods Enzymol 1990, 185, 187–195.
`57. Negro, A.; Onisto, M.; Masiero, L.; Garbisa, S. FEBS
`Lett 1995, 360, 52–56.
`58. Huang, W.; Suzuki, K.; Nagase, H.; Arumugam, S.; van
`Doren, S. R.; Brew, K. FEBS Lett 1996, 384, 155–161.
`
`59. Williamson, R. A.; Natalia, D.; Gee, C. K.; Murphy, G.;
`Carr, M. D.; Freedman, R. B. Eur J Biochem 1996, 241,
`476 – 483.
`60. Negro, A.; Onisto, M.; Grassato, L.; Caenazzo, C.;
`Garbisa, S. Protein Eng 1997, 10, 593–599.
`61. Hodges, D. J.; Reid, D. G.; Rowan, A. D.; Clark, I. M.;
`Cawston, T. E. Eur J Biochem 1998, 257, 562–569.
`62. Shimizu, N.; Fukuzono, S.; Harada, Y.; Fujimori, K.;
`Gotoh, K.; Yamazaki, Y. Biotechnol Bioeng 1991, 38,
`37– 42.
`63. Mizukami, T.; Oka, T.; Itoh, S. Biotechnol Lett 1986, 8,
`611– 614.
`64. Nishi, T.; Fujita, T.; Nishi-Takaoka, C.; Saito, A.; Ma-
`tsumoto, T.; Sato, M.; Oka, T.;