Direct Refolding of Inclusion Bodies Using Reversed Micelles

Masafumi Sakono,† Yu-mi Kawashima,‡ Hirofumi Ichinose,‡ Tatsuo Maruyama,†Noriho Kamiya,† and Masahiro Goto*,†,‡

Department of Applied Chemistry, Graduate School of Engineering, Kyushu University, 6-10-1, Hakozaki,Fukuoka 812-8581, Japan, and Japan Science and Technology Corporation, PRESTO, Japan

The protein refolding of inclusion bodies was investigated using reversed micellesformed by aerosol OT (AOT). Ribonuclease A (RNase A) was overexpressed inEscherichia coli and used as native inclusion bodies. The enzymatic activity of RNaseA was completely regained from the inclusion bodies within 14 h by solubilization inreversed micelles. To further enhance the refolding rate, a molecular chaperone, GroEL,was incorporated into the refolding system. The resultant refolding system includingGroEL showed better performance under optimized conditions for the refolding ofRNase A inclusion bodies. The refolding rate was considerably improved by the additionof the molecular chaperone, and the refolding step was completed in 1 h. The proteinrefolding in the GroEL-containing refolding system was strongly dependent on thecoexistence of ATP and Mg2+, suggesting that the GroEL hosted in the reversedmicelles was biologically active and assisted in the renaturation of the inclusion bodies.The addition of cold acetone to the reversed micellar solution allowed over 90% recoveryof the renatured RNase A.

Introduction

The expression of recombinant proteins in Escherichiacoli often results in the formation of insoluble aggregatedproteins that are called inclusion bodies. To regain thebiological activity of the target proteins, many strategieshave been proposed for protein refolding (1-4). Dilutionmethods are one example, in which protein refolding isachieved by diluting the aqueous solution containing thedenatured proteins and high concentrations of denatur-ants with a large amount of aqueous buffer solution.Although this method has the distinctive feature of easyoperation, the yield of active proteins is rather lowbecause protein refolding is a competitive process throughinter- and intramolecular interactions (5, 6). As theformer interaction causes the reaggregation of denaturedproteins, development of an efficient refolding methodhas been anticipated.

To overcome the above problems associated withprotein refolding, we have utilized reversed micelles asthe refolding medium. Reversed micelles are nanostruc-tured molecular assemblies composed of surfactant mol-ecules, an organic solvent, and a small amount of water.They contain nanoscaled aqueous droplets called waterpools. Because the size of the water pools is comparableto the molecular size of proteins, reversed micelles canfacilitate the encapsulation of proteins in the water pools.Figure 1 shows a schematic illustration of protein refold-ing of an inclusion body using reversed micelles. In ourprevious studies, we proposed a total system for proteinrefolding using reversed micelles (abbreviated to “RM-mediated refolding system”). It is composed of three

steps: solubilization, renaturation, and back extraction.When inclusion bodies are solubilized in reversed mi-celles, each protein is separated from the protein ag-gregates into the water pools of the reversed micelles.As a result, reversed micelles can provide a restrictedenvironment for spontaneous refolding by reducing in-termolecular interactions. It was found that the co-solubilization of GroEL, a typical molecular chaperone,enhances the efficacy of protein refolding (7). The rena-

* To whom correspondence should be addressed. Tel: +81-92-642-3575. Fax: +81-92-642-3575. E-mail: mgototcm@mbox.nc.kyushu-u.ac.jp.

† Kyushu University.‡ JST, PRESTO.

Figure 1. Schematic illustration of the refolding of an inclusionbody using reversed micelles.

1783Biotechnol. Prog. 2004, 20, 1783−1787

10.1021/bp049887j CCC: $27.50 © 2004 American Chemical Society and American Institute of Chemical EngineersPublished on Web 09/21/2004

tured proteins in the reversed micelles are finally recov-ered from the micellar solution by stripping the watervia the addition of cold acetone. However, all of thesestudies were carried out using artificially denaturedmodel proteins including RNase A (8, 9) and cytochromec (10).

In the present study, we investigated the refolding ofreal recombinant RNase A inclusion bodies prepared byoverexpression in E. coli. Factors affecting the yield ofactive proteins from the inclusion bodies by the RM-mediated refolding system were compared with factorsobtained with artificially denatured RNase A. The resultsdemonstrate that the reversed micelles are effective forrenaturing inclusion bodies of proteins.

Materials and Methods

Chemicals. Sodium bis-2-ethylhexyl sulfosuccinate(AOT) and urea were purchased from Kishida Chemical(Osaka, Japan). RNase A from bovine pancreas andcytidine 2′:3′-cyclic monophosphate were obtained fromSigma (St. Louis, MO). Oxidized and reduced glutathione(GSSG and GSH, respectively) and 2-amino-2-hydroxy-methyl-1,3-propanediol (Tris) were purchased from WakoPure Chemical Industries (Osaka, Japan). Isopropyl-thio-â-D-galactopyranoside (IPTG) and DNase I were obtainedfrom TaKaRa (Tokyo, Japan). All reagents were of thehighest grade available commercially. Deionized waterwas obtained with a Milli Q System (Millipore, Billerica,MA).

Expression Vector Construction and Recombi-nant Protein Preparation. The gene fragment encod-ing RNase A was amplified by PCR from the genomicDNA of Bos taurus. Amplification of cDNA fragments wasperformed with the primer combination of 5′-CCAAG-GAAACTGCAGCAGCC-3′ as the upstream primer and5′-TTGAATTCAGACCTACACTGAAGCATC-3′ as thedownstream primer. The PCR mixture (100 µL) contained0.1 µg of genomic DNA, 2.5 U of Pyrobest DNA polym-erase (TaKaRa), 1x Pyrobest Buffer II, 100 pmol of eachprimer, and dNTPs at a final concentration of 200 µM.The PCR temperature was programmed as follows: 94°C for 3 min, 55 °C for 2 min, and 72 °C for 1 min for 1cycle followed by 94 °C for 20 s, 60 °C for 20 s and 72 °Cfor 1 min for 29 cycles, and then a final extension at72 °C for 1 min. The resultant cDNA fragment wasdigested with EcoRI, phosphorylated by alkaline phos-phatase (TaKaRa), and ligated into the pET22 plasmid,which had been treated with BalI and EcoRI (11, 12).The obtained recombinant plasmid was transferred intoE. coli BL21(DE3)pLysS (TaKaRa). The transformantswere grown in TB broth containing ampicillin (100 mg/L) with shaking (200 rpm) at 37 °C until the OD600reached 1.2. IPTG was then added to the culture to afinal concentration of 0.5 mM, and the bacterial cells wereincubated with shaking (200 rpm) at 37 °C for 3 h. Therecombinant protein was purified using 50 mM Tris-HClbuffer solutions (pH 8.0) containing Triton X-100 (1% w/v)or 1 mM EDTA. The homogeneity of the recombinantprotein was analyzed by sodium dodecyl sulfate-poly-acrylamide gel electrophoresis (SDS-PAGE) and matrix-assisted laser desorption/ionization time-of-flight massspectrometry (MALDI-TOF MS; Perseptive Biosystems).

Solubilization of RNase A Inclusion Bodies intoReversed Micelles by Solid-Liquid Extraction. TheRNase A inclusion bodies (0.4 mg) were pretreated witha small amount (0.018 mL) of 8 M urea solution, and theresultant wetted inclusion bodies were added to 0.856 mLof 400 mM AOT-isooctane solution. An aqueous solution

(0.126 mL of 0.1 M Tris-HCl buffer, pH 8.5) was addedto the organic solution, and the mixture was sonicatedfor 20 min (80 W). The Wo value () [water]/[AOT]) of theresultant reversed micellar solution became 20 underthese experimental conditions. After the sonication,insoluble impurities present in the inclusion bodies wereremoved from the solution by centrifugation (10 000g ×3 min), resulting in a reversed micellar solution includingthe RNase A inclusion bodies.

Refolding of Artificially Denatured RNase A inReversed Micelles. Oxidized and reduced glutathione(10 mM GSSG and 20 mM GSH, respectively) weresolubilized in 0.1 M Tris-HCl buffer (pH 8.5), and theresultant solution (0.144 mL) was added to the 400 mMAOT-isooctane solution (0.856 mL). Renaturation wasinitiated by mixing 1 mL of the reversed micellar solutionincluding glutathione and 2 mL of the reversed micellarsolution including denatured RNase A. The mixture wasincubated at 298 K.

Measurement of RNase A Activity. (a) In Re-versed Micelles. A reversed micellar solution containinga substrate for RNase A was prepared by injecting 0.6mL of 10 mg/mL cytidine 2′:3′-cyclic monophosphate in0.05 M borate buffer (pH 9.5) into 36 mL of 50 mM AOT-isooctane solution. A 200-µL aliquot of the reversedmicellar solution including RNase A was added to 1 mLof the substrate solution. The RNase A activity wasdetermined by the increase in absorbance at 286 nmusing a UV-Vis spectrophotometer (Vbest-570; JASCO,Tokyo). The refolding yield was defined by the catalyticactivity of native RNase A solubilized in reversed micellesas units under the same experimental conditions withrespect to the RNase A inclusion bodies.

(b) In Aqueous Solution. An aqueous RNase Asolution (0.06 mL) was added to 0.84 mL of the substratesolution (0.1 mg/mL cytidine 2′:3′-cyclic monophosphatedissolved in 0.025 M phosphate buffer, pH 7.5). TheRNase A activity was determined by measuring theabsorbance change at 286 nm as described above.

Quantitative Analysis of the Number of Thiols inRNase A Aggregates. After preparation of 0.3 M 2,2′-dithiodipyridine (2PDS) solution dissolved in DMSO, 10µL of the 2PDS solution was added to 1.5 mL of an 8 Murea solution containing denatured RNase A (0.2 mg/mL).The number of thiols was estimated by measuring theabsorbance at 343 nm from the degradation of 2PDS (13).

Back Extraction of Refolded RNase A from theReversed Micellar Solution. One volume of reversedmicellar solution containing renatured RNase A wasadded to 20 vol of cold acetone at 273 K. The precipitatedRNase A was collected by centrifugation and lyophilizedusing a freeze-dryer (EYELA FD-5N) for 24 h. Thecatalytic activity of the recovered RNase A was measuredafter dissolution in a Tris-HCl buffer solution as de-scribed above for the measurement of RNase A activity.

Incorporation of GroEL into the RM-MediatedRefolding System. A reversed micellar solution con-taining GroEL was prepared by injecting 0.324 mL of0.01 M Tris-HCl buffer (pH 8.5) containing 1 mM ATP,1 mM MgCl2, 1 mM KCl and 2 mg/mL GroEL into the400 mM AOT-isooctane solution (0.64 mL). RNase inclu-sion bodies (0.4 mg) were added to 0.036 mL of 8 M ureasolution, and the resultant wetted inclusion bodies wasadded to the reversed micellar solution, making a totalvolume of 1 mL. The mixture was sonicated for 30 min.After the sonication, the insoluble fraction present inthe inclusion bodies was removed by centrifugation(10 000g × 3 min), resulting in a reversed micellarsolution including both denatured RNase A and GroEL.

1784 Biotechnol. Prog., 2004, Vol. 20, No. 6

The refolding yield in reversed micelles was determinedby the comparison with native activity of RNase A in thereversed micellar solution including GroEL. The catalyticactivity of RNase A was not influenced by the coexistenceof GroEL in reversed micelles.

Results and DiscussionPreparation of RNase Inclusion Bodies in E. coli.

The expression of recombinant RNase A in E. coli wasconfirmed by SDS-PAGE analysis, and the expressionof RNase A was significantly induced after the additionof IPTG (Figure 2). Little RNase A was identified in thesoluble fraction of the cell lysates (data not shown),suggesting that the majority of the expressed RNase Aexisted in the insoluble fraction. SDS-PAGE analysis ofthe insoluble fraction showed that the molecular weightof the major protein band originating from the proteina-ceous substances was in good agreement with that ofnative RNase A (lanes 5 and 6 in Figure 2). MALDI TOF-MS analysis also provided clear evidence for the success-ful expression of full-length RNase A (theoretical MW13.7 kDa; obtained MW 13.7 kDa). On the basis of theseresults, we employed the insoluble fraction as RNaseinclusion bodies. The inclusion bodies were further puri-fied with Tris-HCl buffers including both Triton X-100and EDTA, because renaturation was not enhanced usingunpurified inclusion bodies.

RM-Mediated Refolding of RNase A InclusionBodies. Several factors affecting the recovery of activeproteins using the RM-mediated refolding system havebeen revealed (8-10). In the present study, we investi-gated experimental parameters for the efficient refoldingof RNase A inclusion bodies. The incubation time forrefolding was fixed at 20 h for all of the experiments inthis section.

Effect of Wo Values on Renaturation of RNase AInclusion Bodies. The diameter of AOT reversed mi-celles is in proportion to the Wo values (14), and the sizeof reversed micelles influences the refolding efficiency asa result of the degree of interaction between the surfac-tant and the denatured protein (9). Figure 3 shows therelationship between the Wo value and the refolding yieldof RNase A in reversed micelles. The optimal range ofWo values for sufficient renaturation of RNase A wasaround 10-20. On the other hand, the refolding yieldswere markedly lower with Wo values under 10 or over20. It is considered that the decrease in refolding yieldat Wo ) 5 may be due to an unfavorable interactionbetween the surfactants and the protein, whereas the low

yield at Wo ) 25 might be caused by either the dilutionof the redox reagents participating in the protein refold-ing in the reversed micelles or insufficient prevention ofprotein aggregation by the increase in the size of waterpools. In the RM-mediated refolding system for theartificially denatured RNase A (i.e., model inclusionbodies), the optimal Wo value for the refolding yield wasaround 15 (9), indicating that the optimal size of reversedmicelles for RNase A refolding was basically identical tothat for refolding of model inclusion bodies of the protein.

Effect of Water Pool pH in Reversed Micelles onRenaturation of RNase A Inclusion Bodies. The pHof aqueous refolding buffers affects the refolding yield andthe formation of three-dimensional bridge structures inproteins through disulfide bonds. It is evident fromFigure 4 that the refolding yield was significantly influ-enced by the pH of the water pools of the reversedmicelles, and the optimal pH was around 8-9. It isknown that the pKa value for effective disulfide bondformation is 8.5, and the refolding yield was thusincreased. When the pH was less than 7.0, the refoldingyield was reduced because the dissociation of disulfidebonds was impossible. In a similar manner, RNase A maybe unable to fold correctly at pH 10 as the formation ofdisulfide bonds is disrupted. In the case of the artificiallydenatured RNase A, the optimum pH value for therefolding yield was about 9.0, suggesting that there wasno significant difference between the model and realinclusion bodies concerning the optimal pH for proteinrefolding (9).

Effect of Molar Ratio of Reduced and OxidizedGlutathione on Renaturation of RNase A InclusionBodies. The reoxidation of reduced thiols is essential forobtaining the native conformation and recovering theenzymatic activity of the refolded RNase A (15, 16). In

Figure 2. SDS-PAGE analysis of the expression of recombi-nant RNase A. Lane 1, molecular weight markers; lane 2, beforeIPTG induction; lane 3, 1 h after IPTG induction; lane 4, 3 hafter IPTG induction; lane 5, a purified inclusion body; lane 6,native RNase A.

Figure 3. Relationship between Wo and the refolding yield inreversed micelles.

Figure 4. Effect of pH on the refolding yield in reversedmicelles.

Biotechnol. Prog., 2004, Vol. 20, No. 6 1785

the study reported here, we employed reduced andoxidized glutathione (GSH and GSSG, respectively),which are widely used as reshuffling reagents in thiol/disulfide conversion, and investigated the optimum con-ditions for refolding of RNase A inclusion bodies. Figure5 shows the effect of GSH/GSSG molar ratios at aconstant concentration ([GSH] + [GSSG] ) 30 mM) onthe refolding yield in reversed micelles. As shown inFigure 5, the refolding yield was significantly influencedby the molar ratio of GSH/GSSG and sufficient recoveryof active RNase A was obtained using GSH alone. In ourprevious study with model RNase A inclusion bodies,GSSG alone was effective for the renaturation, and therefolding yield was decreased with increasing concentra-tions of GSH (9). The different observations for the modeland real inclusion bodies were hypothesized to be due toa change in the state of the thiols in RNase A, which hasfour disulfide bonds in the native state. The numbers offree thiols in the model and real inclusion bodies werethus checked using 2PDS. The results obtained indicatedthat all of the disulfide bonds in the model inclusionbodies were dissociated to thiols by reduction. On theother hand, we found that only one of the four disulfidebonds in the real inclusion bodies was reduced to freethiols, implying that the remaining three disulfide bondsexisted as a shuffling structure. The difference in thenumber of disulfide bonds in the model and real inclusionbodies led to the different optimum conditions for theglutathione ratio in the RM-mediated refolding system.

Comparison of RM-Mediated Refolding Systemwith Conventional Refolding Method Using Dilu-tion. To evaluate the performance of the RM-mediatedrefolding system, it was compared with a typical dilutionmethod (17). The RNase A inclusion bodies were com-pletely activated within 14 h using the RM-mediatedrefolding system, whereas the refolding yield using thedilution method was only 40% active RNase A within 20h (Figure 6) and insoluble precipitates were observed.The reason for the greater refolding yield with thereversed micellar method should be due to the suppres-sion of unfavorable intermolecular interactions amongfolding intermediates by the isolation of each protein inreversed micelles (9).

Effective Enhancement of RNase A Renaturationwith the Help of a Molecular Chaperone. Increasingthe performance of the RM-mediated refolding system isof great importance from a practical viewpoint. In aprevious study, we developed an advanced protein refold-ing method by introducing a molecular chaperone, GroEL,into the reversed micelles (abbreviated to “RM/GroEL-mediated refolding system”). The efficacy for refoldingof model RNase A inclusion bodies was enhanced by the

RM/GroEL-mediated refolding system (7). Figure 7 showsthe effect of the addition of folding enhancers (GroEL andglutathione) on the refolding yields as a function of time.When the reversed micelles were prepared with a buffersolution containing glutathione, 20% of the RNase Ainclusion bodies were refolded in a 1-h incubation. Whenthe RM/GroEL-mediated refolding system was employedwith glutathione, the refolding yield was markedlyincreased and over 80% of the RNase A activity wasrecovered in only 1 h. Without the addition of the twofolding enhancers, the refolding yield was only 3% in 1h. When the same conditions in the ordinary dilutionmethod was applied to the reversed micellar system, therefolding yield was not enhanced as a result of thedilution of the folding enhancers (data not shown).

On the sigmoidal tendency in refolding yield in Figure7, we assume that there are two stages in RM/GroEL-mediated protein refolding: one is in the presence andthe other is in the absence of GroEL. Since such a two-phase increase was not observed without GroEL (opensquares in Figure 7), the increase in the refolding yieldafter 30 min is due to the aid of GroEL. One possibleexplanation for this phenomenon is that the GroEL-assisted refolding may be significantly accelerated afterthe accumulation of folding intermediates. This impliesthat the folding intermediates were more effectivelyrefolded by GroEL than the fully denatured proteins inreversed micelles as well as in an aqueous solution (18,19). A similar phenomenon was observed in our studywith an artificially denatured RNase A (7).

Figure 5. Effect of the molar ratio of GSH to GSSG on therefolding yield in reversed micelles.

Figure 6. Time courses of RNase A refolding by the RM-mediated refolding system or a dilution method: (0) RM-mediated refolding system; (9) dilution method.

Figure 7. Effect of the addition of refolding enhancers inreversed micelles: (9) with GroEL and glutathione; (0) withglutathione; (2) no additives.

1786 Biotechnol. Prog., 2004, Vol. 20, No. 6

To better understand the mechanism of RM/GroEL-mediated refolding, the effects of the addition of ATP andMg2+, which are required for GroEL biological function,were investigated. As shown in Figure 8, the RM/GroEL-mediated refolding was strongly inhibited by the lack ofATP or Mg2+. The degree of refolding was similar to thatof spontaneous refolding when the reaction mixturelacked ATP or Mg2+. These results demonstrate that theGroEL hosted in the reversed micelles is biologicallyactive in the same manner as in vivo.

Recovery of Renatured RNase A from ReversedMicelles. The final step of the RM-mediated refoldingsystem is the recovery of renatured proteins from thereversed micellar solution to a fresh aqueous solution bya back extraction procedure. However, this procedureoften gave a low recovery yield and caused the formationof irreversible precipitates at the organic-aqueous inter-face, which could reduce the activity of proteins recoveredduring the procedure. In a previous paper, we reportedthat the addition of cold acetone into a reversed micellarsolution containing active RNase A obtained from themodel inclusion bodies was effective for the recovery yield(9). Similarly, with the refolded RNase A from the realinclusion bodies in the present study, over 90% of theactive RNase A was recovered from the reversed micellarsolution.

ConclusionsThe RM-mediated refolding system was applied to the

recovery of active recombinant RNase A from inclusionbodies produced by overexpression in E. coli. The RNaseA inclusion bodies were completely refolded in 14 h underoptimized conditions with respect to Wo, pH, and themolar ratio of reduced and oxidized glutathione. WhenRM/GroEL-mediated refolding was applied, the refoldingrate was markedly enhanced and the molecular chaper-one was found to be very useful for improving therefolding kinetics. Since all the experimental conditions(except for an appropriate redox reagent) for gaining thehigh refolding yield with the real inclusion bodies weresimilar to those for the model denatured RNase A, itappears that the restricted nanostructured environmentpromotes RNase A refolding. The results also suggestthat the optimum refolding conditions for a real inclusionbody in reversed micelles can be predicted by optimiza-tion with an appropriate artificial model inclusion body.

AcknowledgmentThis research was supported as a research project by

PREST (JST) and a Grant-in-Aid for Scientific Research

(no. 13010359) from the Ministry of Education, Scienceand Culture of Japan.

References and Notes

(1) Machida, S.; Ogawa S.; Shi, X. H.; Takaha, T.; Fujii, K.;Hayashi, K. Cycloamylose as an efficient artificial chaperonefor protein refolding. FEBS Lett. 2000, 486, 131-135.

(2) Yoshimoto, N.; Hashimoto, T.; Felix, M. M.; Umakoshi, H.;Kuboi, R.; Artificial chaperone-assisted refolding of bovinecarbonic anhydrase using molecular assemblies of stimuli-responsive polymers. Biomacromolecules 2003, 4, 1530-1538.

(3) Gu, Z.; Su Z.; Janson J. C. Urea gradient size-exclusionchromatography enhanced the yield of lysozyme refolding. J.Chromatogr. A 2001, 918, 311-318.

(4) Nomura, Y.; Ikeda, M.; Yamaguchi, N.; Aoyama, Y.; Aki-yoshi, K. Protein refolding assisted by self-assembled nano-gels as novel artificial molecular chaperone. FEBS Lett. 2003,553, 271-276.

(5) Zettlmeissl, G.; Rudolph, R.; Jaenicke, R. Reconstitution oflactic dehydrogenase. Noncovalent aggregation vs reactiva-tion. 1. Physical properties and kinetics of aggregation.Biochemistry 1979, 18, 5567-5571.

(6) Clark E. D. B. Protein refolding for industrial processes.Curr. Opin. Biotechnol. 2001, 12, 202-207.

(7) Sakono, M.; Ichinose, H.; Goto, M. Protein refolding innanostructured reversed micelles including a molecularchaperone. J. Biosci. Bioeng. 2003, 96, 275-278.

(8) Hashimoto, Y.; Ono, T.; Goto, M.; Hatton, T. A. Proteinrefolding by reversed micelles utilizing solid-liquid extractiontechnique. Biotechnol. Bioeng. 1998, 57, 620-623.

(9) Goto, M.; Hashimoto, Y.; Fujita, T.; Ono, T.; Furusaki, S.Important parameters affecting efficiency of protein refoldingby reversed micelles. Biotechnol. Prog. 2000, 16, 1079-1085.

(10) Sakono, M.; Goto, M.; Furusaki, S. Refolding of cytochromec using reversed micelles. J. Biosci. Bioeng. 2000, 89, 458-462.

(11) Dodge, R. W.; Scheraga, H. A. Folding and unfoldingkinetics of the proline-to-alanine mutants of bovine pancreaticribonuclease A. Biochemistry 1996, 35, 1548-1559.

(12) Chatani, E.; Tanimizu, N.; Ueno, H.; Hayashi, R. Structuraland functional changes in bovine pancreatic ribonuclease Aby the replacement of Phe120 with other hydrophobic resi-dues. J. Biochem. 2001, 129, 917-922.

(13) Humphrey, R. E.; Hinze, W. Spectrophotometric determi-nation of cyanide with organic disulphides. Talanta 1971, 18,491-497.

(14) Adachi, M.; Harada, M.; Shioi, A.; Sato, Y. Extraction ofamino acids to microemulsion. J. Phys. Chem. 1991, 95,7925-7931.

(15) Welker, E.; Narayan, M.; Volles, M. J.; Scheraga H. A. Twonew structured intermediates in the oxidative folding ofRNase A. FEBS Lett. 1999, 460, 477-479.

(16) Vinci, F.; Ruoppolo, M.; Freedman, R. B.; Marino, G. Earlyintermediates in the PDI-assisted folding of ribonuclease A.Protein Sci. 2000, 9, 525-535.

(17) Maeda, Y.; Ueda, T.; Imoto, T. Effective renaturation ofdenatured and reduced immunoglobulin G in vitro withoutassistance of chaperone. Protein Eng. 1996, 9, 95-100.

(18) Staniforth, R. A.; Cortes, A.; Burston, S. G.; Atkinson, T.;Holbrook, J. J.; Clarke, A. R. The stability and hydrophobicityof cytosolic and mitochondrial malate dehydrogenases andtheir relation to chaperonin-assisted folding. FEBS Lett. 1994,344, 129-135.

(19) Ranson, N. A.; White, H. E.; Saibil, H. R. Chaperonins.Biochem. J. 1998, 333, 233-242.

Accepted for publication August 12, 2004.

BP049887J

Figure 8. Effect of ATP and MgCl2 in the RM/GroEL-mediatedrefolding system on the refolding yield.

Biotechnol. Prog., 2004, Vol. 20, No. 6 1787