Background of the Invention In order to mass produce proteins used in medicine and industry, recombinant gene techniques using various microorganisms have been widely used, and to isolate target proteins expressed in such a microorganism, various purification methods have been employed by way exploiting particular properties of the proteins such as charge, solubility, size, hydrophobicity and affinity. However, these methods have the common problem of low efficiency stemming from their multi-steps purification procedures.

Further, one of the serious problems associated with the methods using recombinant microorganisms is that the expressed protein dose not have the correct tertiary structure and often forms an inactive aggregate which precipitates in the form of an inclusion body. Accordingly, a refolding process is required to recover a protein with biological activity from the inclusion body. However, refolding processes in general suffer from the problem of low efficiency primarily due to the occurrence of re- precipitation caused by hydrophobic bonding among intermediates, and this problem becomes particularly severe when refolding is conducted at a high protein concentration.

There have also been reported various methods for preparing an immobilized enzyme. The immobilization method which requires least cost is to electrostatically adsorb a charged protein to a carrier having the opposite charge. However, this method has a problem in that the adsorbed enzyme easily desorbs and remains in solution.

Accordingly, in order to solve the above problems, various efforts have been made to enhance the efficiencies of the purification, immobilization and refolding steps of foreign protein by modifying the solubility, charge or chemical property of the protein by way of introducing an added'tag'or a charged amino acid thereto. For example, used for the purification and identification of various proteins are such tags as polyhistidine, FLAG peptide, strep-tag, polyaspartic acid, polyarginine, polyphenylalanine, polycystein, calmodulin-binding peptide and green fluorescent peptide, some of which are commercially available. The polyhistidine tag (Novagen, USA), in particular, is widely used in laboratories, however it requires the combined use of an expensive carrier and immobilized metal affinity chromatography (IMAC).

Further, Glatz et al. have reported a purification method using a fusion protein prepared by fusing charged anionic amino acid residues to a target protein. They have also been reported: a method using a cationic amino acid, arginine, in the purification step (Brewer, et al., Trends Biotechnol., 3,119-122 (1985)); a method using a cationic amino acid, aspartic acid, in enzyme immobilization (Heng, et al., Biotechnology and Bioengineering, 44, 745-752 (1994); Suominen, et al., Biotec1çnol. Prog, 10, 237-245 (1994); and Enzyme Microb. Technol., 15,593-600 (1993)); and a method using arginine in enzyme immobilization and refolding (Stempfer, et al., Nature Biotechnology, 14, 451-484 (1996)); Nature Biotechnology, 14 329-334 (1996)). However, these methods are currently not practiced in industry or laboratories because of several problems; 1) the purification steps are complicated and give a low purity product; 2) some of the methods require the use of a resin supported heparin which is expensive; and 3) their applicability to various bio-processes is limited.

Accordingly, the present inventors have endeavored to develop an improved amino acid fusion system which can be effectively applied to all of the purification, immobilization and refolding processes, and have found that a fusion protein, which comprises more than two consecutive cationic amino acid residues derived from lysine or arginine fused to a desired target protein, can be employed to enhance the efficiencies of the purification, immobilization and refolding processes of the target protein.

Summary of the Invention Accordingly, it is an object of the present invention to provide a fusion gene which can be advantageously employed in the purification, immobilization and refolding processes of a target protein, a fusion protein encoded therein, an expression vector containing said fusion gene, and a microorganism transformed with said expression vector.

It is another object of the present invention to provide processes for purifying, immobilizing and refolding the fusion protein.

In accordance with one aspect of the present invention, there is provided a fusion protein comprising a target protein and more than two consecutive cationic amino acid residues fused to the target protein, wherein the amino acid is lysine or arginine.

In accordance with another aspect of the present invention, there is provided a fusion gene encoding said fusion protein and an expression vector containing said fusion gene, pTCGTK6, pRCGTR6, pTCGTK10, pLPKlO, <BR> <BR> <BR> pLPK12, pTDSBAK10, pTDSBR10, pTDSBCK10, pTDSBCR10, pTHPPIK10 or pTHIPPIR10.

In accordance with another aspect of the present invention, there is provided a microorganism transformed with said vector, E. coli BL21 (DE3): pLysE: pTCGTK10 (KCTC 0842BP), E. coli BL21 (DE3) : pLysE : pTCGTK6, E. coli BL21 (DE3): pLysE: pRCGTR6, E. coli BL21 (DE3): pLysE: pLPK10, E. coli BL21 (DE3): pLysE: pLPK12, E. coli BL21 (DE3): pLysE: pTDSBAK10, E. coli BL21 (DE3): pLysE: pTDSBR10, E.

coli BL21 (DE3): pLysE: pTDSBCK10, E. coli BL21 (DE3): pLysE: pTDSBCR10, E. coli BL21 (DE3) : pLysE: pTHPPIK10 or E. coli BL21 (DE3): pLysE: pTHIPPIR10.

In accordance with another aspect of the present invention, there is provided a process for preparing a fusion protein comprising culturing said microorganism to induce expression of the fusion protein; obtaining a cell extract from the culture solution; and purifying the fusion protein using an ionic matrix.

In accordance with a further aspect of the present invention, there is provided a process for preparing a fusion protein comprising culturing said microorganism to induce expression of the fusion protein in the form of an inclusion body ; solubilizing the inclusion body; and obtaining the fusion protein from the solubilized inclusion body using an ionic matrix.

In accordance with a further aspect of the present invention, there is provided a process for preparing an immobilized enzyme comprising binding an ion exchange resin with a fusion enzyme containing a target protein and more than two consecutive cationic amino acid residues fused to the target protein, wherein the amino acid is lysine or arginine.

In accordance with a further aspect of the present invention, there is provided a process for refolding an unfolded protein comprising adsorbing the unfolded fusion protein on an ionic matrix and conducting a refolding reaction.

Brief Description of the Drawings The above and other objects and features of the present invention will become apparent from the following description of the invention taken in conjunction with the following accompanying drawings, which respectively show: Fig. 1 : The procedures for constructing vectors pTCGTK6, pTCGTK10 and pTCGTR6 ;

Fig. 2 : Cation exchange resin chromatographic analysis results of recombinant E. coli cell extracts; Fig. 3 : Chromatographic and SDS-PAGE analysis results showing the pre-washing effect of recombinant E. coli cell extract; Fig. 4: SDS-PAGE analysis result showing the effect of the salt concentration in the cell extract on selective adsorption of CGTKlOase to an exchange resin in the adsorbing step; Fig. 5: Effects of salt concentration on the partition coefficient of CGTKlOase protein: Fig. 6: Effects of salt concentration on the enzyme activity adsorbed on cation exchange column; Fig. 7: Effects of amino acid fusion and immobilization on the pH dependence of the enzyme; Fig. 8: Effects of amino acid fusion and immobilization on the stability of the enzyme toward pH changes; Fig. 9: Effect of pH on the inactivation rate constants of amino acid- fused enzyme and the immobilized form thereof; Figs. lOA to lOC : Thermostabilities of fusion of amino acid-fused enzyme and the immobilized form thereof ; Figs. 11A to 11D : Continuous production of cyclodextrins using a fixed-bed reactor containing the immobilized enzyme; Fig. 12: The salt concentration-dependent change in the partition coefficient of the enzyme unfolded by urea; Fig. 13: The Ca+F concentration-dependent variations of the refolding efficiencies of free enzyme in solution and immobilized enzyme; Fig. 14: Salt concentration-dependent variations of the refolding efficiencies of enzymes in solution; Fig. 15: pH-dependent variations of the refolding efficiencies of enzymes in solution; Fig. 16: The salt concentration-dependent change in the refolding efficiency of immobilized enzyme; Fig. 17: The pH-dependent change in the refolding efficiency of

immobilized enzyme; Fig. 18: The protein concentration-dependent variations of the refolding efficiencies of enzymes in solution; Fig. 19 : The protein concentration-dependent change in the refolding efficiency of immobilized enzyme; Fig. 20: Chromatograms of the inclusion bodies of CGTKlOase and LPKlOase purified by selectively binding with cation exchange resin; Fig. 21: Chromatograms of hPPIKlOase andhPPIRlOasepurifiedby using cationic matrix; Fig. 22: The protein concentration-dependent change in the refolding efficiency of lipase ; Fig. 23 : The protein concentration-dependent change in the refolding efficiency of disulfide-bond-promoting enzyme; Fig. 24: The protein concentration-dependent change in the refolding efficiency of disulfide bond isomerase; and Fig. 25: The protein concentration-dependent change in the refolding efficiency of proline isomerase.

Detailed Description of the Invention The inventive fusion protein can be prepared by adding more than 2, preferably 6 to 15, and more preferably 8 to 12 consecutive cationic amino acid residues of lysine or arginine to a desired target protein, usually an enzyme. The cationic amino acid residues can be inserted at the N-terminal or C-terminal of the protein, or anywhere within the protein provided that such insertion does not greatly affect the intrinsic biological function of the protein or enzyme. The term"does not greatly affect the intrinsic biological function of the protein"as used herein refers to a situation that more than 80%, preferably more than 95% of the biological activity of the target protein is maintained after such cationic amino acid fusion.

The target protein which may be used in the present invention is inclusive of such enzymes as cyclodextrin glycosyltransferase (CGTase)

isolated from Bacillus macerans, lipase from Archaeoglobus fulgidus, disulfide-bond-promoting enzyme, disulfide bond isomerase, proline isomerase and the like, but these do not limit the scope of the invention.

Examples of the inventive fusion protein include CGTKlOase which has 10 consecutive lysine residues fused to the C-terminal of wild-type (WT) CGTase, and LPK10 and LPK12 having 10 and 12 consecutive lysine residues, respectively, fused to the C-terminal of wild-type lipase.

Further, the present invention also provides a fusion gene that encodes the inventive fusion protein. However, it is known that several different genes encoding the mutants of the present invention may exist due to the codon degeneracy, and, specifically, a gene modified by introducing preferred codons of microorganism without any change of amino acid sequence can be used for promoting the expression rate of said genes.

Accordingly, the present invention also includes polynucleotides having substantially homologous base sequences with said gene.

The fusion protein of the present invention can be obtained by a process which comprises the steps of synthesizing a gene encoding said fusion protein, inserting to a suitable vector to obtain an expression vector, transforming a suitable host, for example, yeast, E. coli. and the like with said expression vector, and culturing the transformed microorganism under an appropriate culture condition. For example, CGTase fusion protein can be prepared by producing an expression vector pTCGTK10 containing a fusion gene which carries 10 lysine residues inserted in the 3'-end of WT CGTase gene of vector pTCGT1 (Lee, K. C. P and B. Y. Tao., Starch, 46, 67-74 (1994)), transforming E. coli., for example, E. coli. BL21 (DE3) using said expression vector and culturing the same.

WT CGTase does not adsorb to a cation exchange resin (pH 7.4) because its theoretical pI value is about 5.0. However, CGTKlOase which carries fused cationic amino acid residues readily adsorbs to a cationic exchange resin, and thus can be purified thereby.

Accordingly, the present invention provides a process for purifying the inventive fusion protein which comprises the steps of culturing a

transformed microorganism to induce the expression of the fusion protein, obtaining a cell extract from the culture solution and purifying the fusion protein using a cationic matrix; or culturing the transformed microorganism to induce the expression of the fusion protein and recovering the fusion protein from an inclusion body using a cationic matrix. The cationic matrix used in the present invention may be one of those known in the art, e. g., SP sepharose, S sepharose and CM sepharose, preferably SP sepharose, but these do not limit the scope of the invention.

Defining E", as the optimum NaCl concentration at which the maximum activity of a protein is observed during purification by cation exchange chromatograph, E", aX of CGTKIOase is 580 mM, while EmaXS of CGTK6ase, and CGTR6ase, which have 6 fused lysine-and arginine residues, respectively, are 345 mM and 430 mM, respectively. Further, Emax of WT CGTase, which does not adsorb to a cation exchange resin, is 0.

EmaX gives a measure of the adsorbing strength of a protein, and a high EmaX value means that the protein tends to adsorb strongly to a cation exchange resin. Accordingly, a fusion protein having a high Em. value remains adsorbed at a high salt concentration, and thus, such a protein can be separated from other intracellular proteins which readily desorb at a relatively low salt concentration. EmaX of the inventive fusion protein depends on the kind and length of the amino acid residues fused to a target protein. At the same length, EmaX of the fusion protein becomes higher with arginine residues as compared with lysine residues, and if one kind of amino acid is used, EmaX of the fusion protein becomes higher as the multiplicity increases. Moreover, a fusion protein having a high EmaX dissociates in solution more sluggishly. Most of the intracellular proteins readily dissociate at a low salt concentration but some other proteins remain adsorbed to a cation exchange resin even at a high concentration of salt, e. g., 100-300 mM, and act as impurities in the final purifying solution. As described above, as Emax becomes higher, the partition coefficient becomes larger. The partition coefficients of CGTK6ase, CGTR6ase and CGTKlOase, for example, are 2.38,4.97 and 9.88, respectively.

Further, the purity of the desired protein can be raised by pre- washing the column loaded with the adsorbed fusion protein with a buffer solution containing 10 to 1000 mM salt before initiating cation exchange chromatography. This procedure is intended to desorb weakly adsorbed intracellular protein impurities from the column. For example, a purity of more than 95% can be achieved by pre-washing adsorbed crude CGTKlOase with a 400 mM salt solution, according to the results of SDS-PAGE and densitometer analyses, or about 98% when the specific enzyme activity is compared with a pure protein standard obtained by affinity chromatography.

This method can thus be used to obtain the desired protein having an improved purity. In view of the fact that the purity level required for industrial use is generally about 90%, the present invention provides a simple means to produce an enzyme which can be directly used without further purification.

The purification efficiency of the fusion protein can be enhanced by controlling not only the dissociation step as described above but also the adsorption step. This mode of procedure may be represented by a term "selective binding", which provides an additional means for enhancing the recovery and purity of the desired protein. Intracellular proteins compete with the desired protein in the process of binding to ion exchange resin, and thus, when the adsorption step is carried out in the presence of a salt at a concentration in the range of 100-500 mM, the binding of some of the intracellular proteins can be inhibited, thereby enhancing the adsorption efficiency and purity of the desired protein.

Further, when the desired protein is an enzyme, the fusion protein of the present invention can be used in the preparation of an immobilized enzyme by way of electrostatic adsorption to a cation exchange resin. For example, in case of CGTase, more than 90% of the original enzyme activity is maintained after CGTase is fused with cationic amino acid residues and the fused protein retains its full activity even when it is immobilized. The immobilization method of the present invention is much simpler than those described in the prior arts. The inventive immobilized enzyme, which is

fully active and does not leach out, can be used in an immobilized enzyme fixed-bed reactor.

The inventive fusion protein and an immobilized form thereof have similar pH dependencies as WT enzyme. Further, although the fusion enzyme in solution has comparable stability toward pH changes as WT enzyme, the immobilized enzyme has much higher pH stability.

In terms of thermostability, the inventive fusion enzyme is lower than WT enzyme but the immobilized fusion enzyme is much higher.

Accordingly, a desired product can be efficiently produced from a suitable substrate by operating an enzyme reactor containing the inventive immobilized enzyme.

The use of the inventive fusion protein also provides a means to enhance the refolding and other process step efficiencies. This can be accomplished by removing the protein unfolded by the action of a denaturing agent, immobilizing on a cation exchange resin, and inducing refolding.

Further, the present invention provides an efficient means to refold a protein at a high protein concentration without the risk of protein aggregation.

The following Examples are included to further illustrate the present invention without limiting its scope.

Example 1 : Construction of a expression vector containing the fusion protein having cationic amino acid (1) Construction of an expression vector containing CGTase fusion protein PCR (polymerase chain reaction) was carried out using vector pTCGTl (Lee, K. C. P. and B. Y. Tao., Starch, 46,67-74 (1994)) containing WT (wild type) CGTase gene as a template and 5'-primer of SEQ ID NO: 1 and 3'-primer of SEQ ID NO : 2. As a control, PCR was carried out using the fragment of SEQ ID NO : 3 or 4 as the 3'-primer. The PCR program consisted of 1 minutes of denaturation using lmM MgC12 at 95 °C, 2 minutes of annealing at 48 °C, 3 minutes of polymerization at 72 °C for 30 cycles, and 5 minutes of final polymerization at 72 °C. The initial denaturation

reaction was carried out at 95 °C for 5 minutes. The DNA fragment thus obtained was electrophoresed through an agarose gel to confirm its size, treated with BamHI (NEB, England) and SalI (NEB, England), and then, ligated with the fragment obtained by cleaving vector pET21a (Novagen, USA) with the same restriction enzymes, to construct plasmid pTCGTK10 (SEQ ID NO: 5) which contained nucleotides encoding 10 lysines. Plasmids pTCGTK6 (SEQ ID NO: 7) and pTCGTR6 (SEQ ID NO: 9) containing nucleotides encoding 6 consecutive lysines and 6 consecutive arginines, respectively, were prepared as controls. Fig. 1 shows the procedure for constructing expression vectors pTCGTK6, pTCGTR6 and pTCGTK10.

(2) Construction of an expression vector containing lipase fusion protein PCR was carried out using vector pELP-1 as a template and 5'- primer of SEQ ID NO: 11 and 3'-primer of SEQ ID NO: 12 or 13 in accordance with the same method described in (1), wherein pELP-1 is a vector derived by inserting the lipase gene of Archaeoglobusfulgidus (ATCC 49558D) at the NdeI/BamHI section of pET3a vector (Novagen, U. S. A).

The DNA fragment thus obtained was treated with NdeI (NEB, England) and BamHI (NEB, England), and ligated with the fragment obtained by cleaving Vector pET3a (Novagen, USA) with the same restriction enzymes, to construct plasmids pLPKlO and pLPK12 which contained nucleotides encoding 10 and 12 consecutive lysines, respectively.

(3) Construction of vectors for expressing fusion proteins containing disulfide-bond-promoting enzyme, disulfide bond isomerase, and proline isomerase, respectively 5'-primer of SEQ ID NO: 14 and 3'-primer of SEQ ID NO: 15 were mixed and annealed at room temperature, and then, treated with polynucleotide kinase (NEB, USA) at 37°C for 1 hour. pET29-b (Novagen, USA) was cleaved with XhoI and BamHI and ligated with said annealed DNA fragment to construct plasmid pETKlO. In accordance with the same

method, pETRlO was prepared by using 5'-primer of SEQ ID NO: 16 and 3'-primer of SEQ ID NO: 17. PCR was carried out using human proline isomerase cDNA (ATCC 78809) as a template and primers of SEQ ID : 18 and 19 and the product DNA fragment was treated with NdeI (NEB, England) and BamHI (NEB, England). Further, in order to obtain disulfide- bond-promoting enzyme, PCR was carried out using the chromosome of E. coli. as a template and primers of SEQ ID NOS: 20 and 21 and the product DNA fragment was treated with the same restriction enzymes. Moreover, in order to obtain disulfide bond isomerase, PCR was carried out using the chromosome of E. coli as a template and primers of SEQ ID NOS: 22 and 23 and the product DNA fragment was treated with the same restriction enzymes.

Then, vectors pETKlO and pETRlO obtained above were each treated with the same restriction enzymes, and ligated, to construct plasmids pDSBAK10, pDSBCK10, pHPPIK10, pDSBAR10, pDSBCR10 and pHPPIR10, each containing nucleotides encoding 10 consecutive lysines or arginines.

Example 2: Purification of CGTase fusion protein using ion exchange chromatography E. coli BL21 (DE3): pLysE (Novagen, U. S. A.) was transformed with expression vector pTCGTK10 prepared in Example 1 and transformed colonies were selected in a plate containing ampicillin, to obtain a transformant, E. coli BL21 (DE3): pLysE: pTCGTK10, which was deposited with Korean Collection for Type Cultures (Korea Research Institute of Bioscience and Biotechnology) on July 21, 2000 under accession number KCTC 0842BP.

The above transformant, E. coli BL21 (DE3): pLysE: pTCGTl which produces WT CGTase (a control), E. coli BL21 (DE3): pLysE: pTCGTK6 transformed with expression vector pTCGK6, and E. coli BL21 (DE3): pLysE: pTCGTR6 transformed with expression vector

pTCGTR6 were each incubated in 5 of LB medium (containing 50 mg/ of ampicillin) overnight at 37 °C, and then, 1 of the culture solution was re-incubated in lOOTQ of LB medium containing 2 g/ of glucose at 30°C. After 6 hours of incubation when the glucose was exhausted, 0.5 mM isopropyl 3-D-thiogalacto-pyranoside (IPTG, Sigma) and 5 mM CaCl2 were added thereto to induce protein expression. Then, the resultant solution was centrifuged to isolate cells and the cells were suspended in 50mM phosphate buffer (pH 7.4) and disrupted using a French pressure (Aminco, U. S. A.).

The supernatant composed of a soluble cell extract was recovered using a centrifuge and assayed by ion exchange chromatography (Acta, Pharmacia, Sweden). That is, 5mQ of SP-sepharose (Pharmacia, Sweden) cation exchange resin was charged into an XK16 column, the supernatant was loaded thereon, and then, a phosphate buffer (pH 7.4) was introduced as a moving phase at a rate of 1 ml/min for 20 minutes (4 x volume of the resin).

Then, a gradient was created by adding NaCl to the moving phase to a concentration of 1M for 20 minutes, and then, 1M Nal was allowed to flow continuously for 20 minutes. The optical density at 280 nm, conductivity and enzyme activity of the effluent were monitored (Makela MJ and Korpela T., J. Biochem. & Biophys. Method, 15,307-318 (1988)).

Fig. 2 shows the results of such analyses conducted during cation exchange resin chromatography of recombinant E. coli cell extracts, wherein A, B, C and D are results for WT CGTase, CGTK6ase, CGTR6ase and CGTKlOase, respectively, and the lines, (-), (-), and (-- -) represent W absorption at 280nm, gradient and conductivity, respectively, while the bar denotes enzyme activity. As shown in Fig. 2, most of the proteins did not adsorb on the resin when the buffer solution of pH 7.4 was used and also, WT CGTase having a PI of 5.0 did not adsorb on the resin because it was charged negative at that pH (Fig. 3a). Figs. 3b, 3c and 3d represented similar properties except for the enzyme activity. pKa of lysine is 10.0 and that of arginine, 12.0. It was confirmed that the adsorbing strength of the fusion protein having arginine residues was higher than that having lysine residues when the lengths of the lysine and arginine tags are the same. The binding strength of CGTKlOase was higher than those of CGTR6ase and CGTK6ase. The result is shown in table 1.

Table 1 CGTase CGTK6ase CGTKIOase CGTR6ase Protein applied (mg) 1.04 1.18 1.00 1.00 Enzyme activity applied 17.26 5.02 9.76 7.48 Specific enzyme activity 17.26 5.92 9.76 7.48 of the protein inserted (U/mg) Total activity 0 2. 22 4. 92 1. 45 recovered(U) Total activity efficiency 0 44.2 50.0 19.4 recovered (%) Specific enzyme activity 0 14.1 96.4 37.2 of the protein recovered in Em. (U/f) Purification rate in 0 2.38 9.88 4.91 Example 3: Improvement of purity of a fusion protein through Pre- washing Recombinant E. coli cell extracts containing WT CGTase, CGTK6ase, CGTR6ase and CGTKlOase, respectively, were obtained as in Example 2 and subjected to cation exchange chromatography as follows. The cell extract was loaded into the same column and pre-washed with a phosphate buffer solution containing 0,150,200 or 400 mM NaCl, and then, an eluate-was introduced thereto. The resulting solution was subjected to 12% SDS-PAGE using the eluate and a gel was dyed with coomassie blue.

Fig 3 shows effects of pre-washing recombinant E. coli cell extracts on the chromatogram and SDS-PAGE analysis, wherein A represents CGTKase pre-washed with 150 mM NaCI ; B, WT CGTase pre-washed 200 mM NaCI ; C, CGTR6ase pre-washed with 200 mM NaCI ; D, CGTKlOase

pre-washed with 400 mM NaCI ; and S, standard molecular protein. WT CGTase did not adsorb even when the buffer contained no salt. It was confirmed that intracellular proteins remained mostly unremoved and some of the intracellular proteins still remained in case of 200 mM NaCl (Fig. 3B).

Meanwhile, CGTK6ase remained absorbed when prewashed only with the buffer, but it desorbed when 150 mM NaCl was used, while some intracellular proteins remained strongly adsorbed (Fig. 3A). CGTR6ase which strongly binds with the cation exchange resin, desorbed in 300 mM NaCl but some CGTR6ase remained adsorbed in 200 mM NaCl (Fig. 3C).

CGTKlOase remained adsorbed in even 400 mM NaCl and the intracellular proteins except the fusion protein were removed (Fig. 3D). Some intracellular proteins which were strongly adsorbed to the resin can not be removed through pre-washing with 300 mM NaCl but can be mostly removed at 400 mM. These results are in line with the optical density scans at 280 nm: the higher the salt concentration used in pre-washing, the lower the protein concentration recovered by desorbing with 1 M salt. This is due to the removal of intracellular proteins through pre-washing.

Example 4: Improvements of purity and efficiency through selective adsorption Salt was added to the recombinant E. coli cell extract containing CGTKlOase obtained in Example 2 to a concentration of 0, 100,200,300 or 400 mM, adsorbed to SP-sepharose cation exchange resin-filled Poly-prep column (Biorad, USA), and then, washed with a salt solution of the same concentration. The protein recovered using 1M salt was assayed by SDS- PAGE and the zymograms shown in Fig. 4 were obtained. Fig. 4 thus exhibits the effect of varying the condition of the adsorbing step of the ion exchange reaction in the purification of CGTKlOase, wherein S is a standard molecular protein; A, the protein purified by affinity chromatography (Sundberg, L. and Porath J., J. of Chromatogr., 90,87-98 (1974); B, the protein purified by pre-washing with 400 mM salt followed by cation

exchange chromatography in accordance with the method described in Example 3; and C, D, E, F and G, proteins obtained by adsorbing the cell supernatants containing 0,100,200,300 and 400 mM salt, respectively. As is shown, the recovery and purification efficiencies can be increased, as observed in D, E, F, and G, by inhibiting the adsorption of intracellular proteins by way of raising the salt concentration at the adsorption step from 0 to 300 mM. The adsorption of intracellular proteins which usually adsorb more wealdy than the fusion protein is selectively inhibited by added salt at the adsorption step..

Example 5: Immobilization of enzyme through fusion with cationic amino acid residues (step 1) Partition coefficient (a) s were measured at various salt concentrations as follows. A purified CGTKlOase solution was equilibrated (20 mM phosphate buffer solution, pH 7.0) to an ionic concentration of 0 to 500 mM and mixed with a resin (SP-Sepharose, Pharmacia, Sweden) equilibrated at the same concentration to immobilize the fusion enzyme CGTKlOase. The relative amounts of the protein adsorbed and the protein remaining in solution were measured to determine the partition coefficient, according to Equation 1, wherein q is the concentration of the protein adsorbed and p is the concentration of the protein in solution.

Equation 1 q <BR> <BR> <BR> =<BR> <BR> <BR> <BR> <BR> q+p Fig. 5 illustrates the salt concentration-dependant change in the

partition coefficient of CGTKlOase protein. As the result shows, the partition coefficient is about 0.95 at a salt concentration range of 0 to 300 mM. The fusion enzyme CGTKlOase was immobilized by ionic interaction rather than by covalent bonding.

(step 2) In order to examine the eluting behavior of CGTKI Oase immobilized to a resin in buffer (20 mM phosphate buffer solution, pH 7.0) having a salt concentration in the range of 0-1 M, the activity of enzyme dissolved in solution was measured after equilibration. The fusion enzyme was bound to a cation exchange resin, unbounded fusion enzyme was removed by washing, and then, incubated in a 10 x volume amount of a salt solution for 4 hours while mixing. Incubation for 4 hours was considered long enough to accomplish equilibrium, for it is generally known that the adsorption and desorption reaction of a protein reaches equilibrium within several minutes.

Setting the sum total of enzyme activity formed in the equilibrated solution and the enzyme activity after desorbing with 1M NaCl at 100, the salt concentration-dependant change in the equilibrium fusion protein activity in solution can be shown in Fig. 6, wherein solid circles represent eluted enzyme and open circles, uneluted (immobilized) enzyme. As the graph shows, the enzyme did not elute at a salt concentration of less than 100 mM, and more than 90% of the enzyme remained fixed at a salt concentration of even 300 mM.

Example 6: pH-dependant stability of enzyme In order to measure the pH-dependant behaviors of free and immobilized fusion proteins, 0.5 of WT CGTase, non-immobilized CGTKlOase and immobilized CGTKlOase were each added to 1.45 mE of a phosphate buffer solution having a pH in the range of 4 to 9, the enzyme activity relative to a maximum value of 100 was measured. As shown in Fig. 7, both free WT CGTase and CGTKlOase in solution exhibited similar

behaviors. The immobilized enzyme showed a similar trend at pH 6 or higher, but a low enzyme activity at less than pH 6. This is attributed to desorption of the bound fusion protein due to lowered electrostatic binding thereof to the resin. Accordingly, it was confirmed that the pH- dependencies of the free and immobilized fusion protein CGTases are similar to that of WT enzyme.

Meanwhile, in order to measure the stabilities against pH variation of free and immobilized fusion enzymes, WT CGTase, non-immobilized CGTKlOase and an immobilized form thereby were each added to a phosphate buffer solution having a pH in the range of 4 to 10, kept at 50 °C for 1 hour and the enzyme activity was measured. The result is expressed by a value relative to a maximum value of 100. As shown in Fig. 8, WT CGTase and CGTKlOase all exhibit similar trends and the immobilized enzyme was found to be stable at a wide pH range.

The enzyme inactivation process can be represented by Equation 2, wherein A is remaining enzyme activity ; t, time; and k, inactivation rate constant.

Equation 2 dA <BR> <BR> <BR> kA<BR> --=-kA dt Free WT CGTase and CGTKlOase in solution showed almost the same k values in the pH range examined, but the immobilized CGTKlOase exhibited lower inactivation rate constants. The inactivation rate of an enzyme can be determined by measuring the enzyme activity change with time during enzyme incubation in accordance with the method described above. The observed enzyme inactivation processes were analyzed using Equation (2) and the calculated inactivation rate constants are represented in Fig. 9. These results suggest that the fusion of cationic amino acid residues

do not adversely affect the pH stability of enzyme and that increased stability toward pH can be obtained by immobilizing the fusion enzyme.

Example 8: Measurment of thermostability of immobilized enzyme The thermostability of enzymes in 20 mM phosphate buffer (pH 7.0) was examined at 50 C and 25 °C, and the result is shown in Fig. 10. Figs. lOa and lOb show the time-dependent activities of the fusion protein and immobilized fusion protein at 50°C and 25 °C, respectively, which shows that the thermostability of CGTKlOase was lower than free WT CGTase in solution while the immobilized CGTKlOase exhibited enhanced stability that was equal or higher than that of WT CGTase.

Further, CGTase requires the presence of Ca salt in order to maintain its structural integrity. Thus, the enzyme thermostability was measured at 50°C in presence of 5 mM CaCl2 in 20 mM MOPS (3- [N- Morpholino] propanesulfonic acid, Sigma, U. S. A.) (pH 7.0). The result in Fig lOc shows that the thermostability of the immobilized enzyme was markedly enhanced.

Example 9: Stability of immobilized enzyme in a bio-process A CGTKlOase solution equilibrated with 10 mM phosphate buffer (pH 6.0) was mixed with SP-sepharose equilibrated with the same buffer to obtain an immobilized enzyme. Non-immobilized enzymes were removed by washing and the resulting immobilized enzyme was charged to an XK16 column (Pharmacia, Sweden) to fabricate a fixed-bed reactor. The reactor column was kept at 25 °C using a water jacket, and a reactant solution, which was prepared by adding long/. of soluble starch to the same phosphate buffer containing 5mM CaCl2, was continuously introduced to the column using a gear pump at a rate of 0.5 m/min. A sample was collected once a day during a test period of 15 days. Each sample was allowed to stand at- 20 °C and the cyclodextrin (CD) product was assayed by thin membrane

liquid chromatography as follows. Cyclodextrin is composed of a-CD, P- CD and y-CD which are composed of 6,7 and 8 glucose units, respectively.

A TLC plate (KF5, Whatman, USA) was activated in a 110 °C oven for 1-2 hours before use. Each sample was diluted to the concentration range of control standards, and then, 1-2 au of the resulting solution and standards were applied to the plate. The standards contained 0.005-0.05% glucose, maltose, maltotriose and known amounts of the three forms of CD.

The plate was dried and developed twice in a chamber containing a solvent mixture (nitromethane: water: n-propanol, 2: 1.5: 4, v/v). Then, the plate was kept in a 110 °C oven for 10 minutes until the smell of the solvent was not detectable. Added to the developed plate was a mixture of methanol and H2SO4 (95: 5, v/v, containing 3g/L of a-naptol), dried, and kept in a 110 °C oven for 10 minutes to colorize. The colorized spot was analyzed and converted into a concentration term using a densitometer (GS-700, Biorad, USA) and Molecular analyst (Biorad, USA). Fig. 11 shows the amount of CD produced by the immobilized enzyme, wherein A, B, C and D represent a-CD, ß-CD and Y-CD and total CD, respectively. As shown in Fig.

11, a steady amount of CD can be produced using the immobilized enzyme over a long period of time without any significant changes in the relative amounts of three forms of CD produced.

Thus, as shown in Examples 5 to 9, an immobilized enzyme can be efficiently prepared by fusing cationic amino acid residues to a target enzyme, and a desired product can be produced using said immobilized enzyme.

Example 10: Immobilization of the denatured protein To examine whether a protein denatured (unfolded) by urea can be immobilized to a cation exchange resin, the following experiment was conducted. 0-400 mM NaCl was added to 20 mM MOPS buffer (pH 7.0) containing 9M urea, CGTKlOase was added thereto to be denatured for 4

hours, and immobilized on SP-sepharose equilibrated with the same buffer, wherein the resin to the protein volume ratio was 1: 2 and the immobilization process was conducted at room temperature for 2 hours while stirring. The resin was allowed to settle, and the concentration of the protein remaining in solution (designated P) was measured. The precipitated resin was washed with the same solution, and then, a 17 x volume amount of 20 mM phosphate buffer solution (pH 7.0) containing 1M NaCl and 9M urea was added thereto to release the protein adsorbed on the resin. The concentration of the eluted protein in solution was measured and the concentration of the protein adsorbed on the resin was calculated (designated q). The partition coefficient was then assessed in accordance with Equation 2 described in Example 5. The result shown in Fig. 12 depicts that the partition coefficient was more than 0.9 in 1-100 mM NaCl but it dropped precipitously at higher salt concentrations to nearly 0 at 400mM. Thus, it was confirmed that a denatured fusion enzyme could also be easily immobilized.

Example 11: Refolding of unfolded protein (1) Refolding process CGTKlOase was allowed to stand at room temperature in 20 mM MOPS buffer (pH 7.0) containing 9M urea for 4 hours to be denatured, mixed with SP-sepharose pre-equilibrated under the same condition to be immobilized, and then, unreacted, free enzyme was removed by washing.

Free CGTKlOase in solution and the immobilized enzyme were added to refolding buffer (20 mM MOPS buffer solution, pH 7.0) to concentrations of 50 mg/mQ and 1 mg/lTIQ, respectively. After dilution (the concentration of urea was 0.45 M in both cases), and the diluted solution containing free enzyme was allowed to stand at 15 °C for 16 hours and the enzyme activity was measured. In the case of immobilized enzyme, the protein concentration was measured after desorbing with 1M NaCI. Specific enzyme activity refers to the value obtained by dividing enzyme activity by

the protein concentration, and the refolding efficiency was calculated based on the specific enzyme activity (100%) of WT enzyme.

(2) Variation of the refolding efficiency of CGTase with Ca salt concentration Refolding in a refolding solution was examined at various CaCl2 concentrations, in accordance with the method described in (1). As shown in Fig. 13, the refolding efficiency increased with increasing Ca salt concentration. Ca salt is required in the formation of the tertiary structure of CGTase.

(3) Ionic strength and pH-dependency of refolding efficiency of CGTase Refolding reactions were carried out at various NaCl concentrations in the range of 0 to 400mM using free WT CGTase in solution, free CGTKlOase or a mixture of SP-sepharose resin and WT CGTase, in accordance with the method described in (1). WT CGTase which does not bind to a resin was used as a control for examining the influence of the resin on the refolding efficiency caused by resin's interaction with CGTase's protein segments other than lysine. As shown in Fig. 14, WT CGTase as a whole was not influenced by NaCl concentration and the efficiency decreased slightly when the resin was present. The refolding efficiency of CGTKlOase was higher than that of WT CGTase and greatly influenced by NaCl concentration. The higher NaCl concentration, the higher the refolding efficiency, and the refolding efficiency in 400 mM NaCI was 40%, a 2-fold higher value than 20% for WT CGTase. It seemed that the refolding efficiency of free CGTKlOase in solution was enhanced as the result of reduced precipitation of the enzyme, which was brought about by molecular repulsions between strongly charged fusion protein molecules.

Thus, the refolding efficiency increased when WT enzyme or the fusion enzyme was refolded in solution.

Also examined were the refolding behavior of free WT CGTase, CGTKlOase and a mixture of a resin and WT CGTase, in a refolding buffer

solution having a pH in the range of 6-8.5, in accordance with the method described in (1). Because PI of CGTase is 5.0, its net charge became strongly negative at a high pH. By examining the refolding efficiency of WT CGTase mixed with the resin, one can assess to what degree the interaction between the CGTase charge and the resin influences the refolding efficiency. As shown in Fig. 15, the refolding efficiency of WT CGTase depended on pH but was not greatly influenced by the added resin. The refolding efficiency of CGTKlOase was again higher than that of WT CGTase.

(4) Influence of ionic strength and pH on the refolding efficiency of immobilized CGTase Refolding of the immobilized CGTase was examined using refolding solutions containing various amounts of NaCl in the range of 10-400mM, in accordance with the method described in (1).

Fig. 16 displays the NaCl concentration-dependant variation of the refolding efficiency of the immobilized CGTKlOase. As shown in Fig. 16, the refolding efficiency calculated based on the specific enzyme activity of the refolded enzyme was 100% over the entire NaCl concentration range examined. However, as the solid circles of Fig. 16 show, the recovery rate of the protein was lower at a higher salt concentration, possibly due to the presence of some ionized protein which remains in solution.

Also examined was the refolding of the immobilized CGTKlOase in a refolding solution at a pH in the range of 6 to 8.5, in accordance with method described in (1). Fig. 17 displays the pH-dependent change in the refolding efficiency of the immobilized CGTKlOase. As shown in Fig. 17, the refolding efficiency of the immobilized CGTKlOase was in most cases much higher than WT CGTase observed in (2) or free CGTKlOase in solution. The amount of remaining protein was almost constant at pH 6.0- 8.5 but the refolding efficiency was greatly influence by pH. That is, the refolding efficiency decreased due to the fact that some section of the protein becomes positively charged at pH 7.0 or below. This result suggests the

importance of minimizing the attractive interaction between the negative charge of the resin and the positive charge of the protein when the refolding reaction is carried out in the presence of a cation exchange resin which is negatively charged.

(5) Effect of protein concentration on the refolding efficiency of enzyme Refolding reactions were carried out using free WT CGTase, free CGTKlOase in solution and the immobilized CGTlOase at various protein concentrations in the range of 0.004-8 mg/mQn in accordance with the method described in (1).

Figs. 18 and 19 display the protein concentration-dependent variations of the refolding efficiency of free enzyme in solution and immobilized enzyme, respectively. Fig. 18a shows the activity of recovered enzyme, and Fig. 18b, the recovered specific enzyme activity.

As the figures show, the refolding efficiency of the free enzyme in solution decreased with increasing protein concentration. This result is due to the occurrence of increasing re-aggregation between refolded intermediates or denatured proteins at a high concentration. At a protein concentration of more than 0.1 mg/m, WT CGTase exhibited a refolding efficiency of 4%, and CGTKlOase, an efficiency of only 7%. However, as shown in Fig. 19, the fusion protein immobilized through the fused lysine moiety showed a refolding efficiency of 100% at a high protein concentration of even 8 mg As compared with WT CGTase's refolding efficiency of 4%, the refolding of the fusion protein immobilized through the fused lysine moiety occurs with a 25-fold higher refolding efficiency even at a 80-fold higher protein concentration. This represents a 2,000-fold enhancement of the process productivity.

Thus, the free enzyme in solution gives only a low refolding efficiency at a low protein concentration while the immobilized enzyme undergoes refolding with a high efficiency even at a high protein concentration. This is due to fact that the free enzyme in solution tends to undergo by re-aggregation between refolding intermediates, while such re-

aggregation is inhibited in the case of the immobilized enzyme.

Example 12: Purification of lipase containing lysine and CGTKlOase expressed as inclusion body E. coli BL21 (DE3): pLysE (Novagen, U. S. A.) was transformed with each of the expression vectors pLPKlO and pLPK12 prepared in Example 1 (2) and the transformed colonies were selected from a plate containing ampicillin.

E. coli BL21 (DE3): pLysE: pTCGTK10 which produces GCTKlOase, E. coli BL21 (DE3): pLysE: pELP-1 which produces WT lipase, E. coli BL21 (DE3): pLysE: pLPK10 and E. coli BL21 (DE3): pLysE: pLPK12 were each incubated in 5 of LB medium (containing 50 mgli of ampicillin) overnight at 37 °C, and then, 1 ! of the culture solution was re-incubated in IOOW of LB medium containing 2 g/ of glucose at 37°C. After incubating for 4 hours until glucose was exhausted, 1 mM IPTG was added thereto to induce protein expression. Then, the resultant solution was centrifuged to obtain cell mass, which was suspended in 50mM phosphate buffer solution (pH 7.4) and disrupted using a French pressure (Aminco, U. S. A.). As the expressed lipase was mostly in the form of an inclusion body, the supernatant containing soluble cell extracts was removed by centrifugation and the isolated precipitate was dissolved in 9M urea (20 mM MOPS buffer solution, pH 7.0). The solubilized protein was purified by the selective binding method described in Example 4 and subjected to SDS- PAGE. The result is shown in Fig. 20. In Fig. 20, S represents a standard molecular weight protein, and lanes 1 and 5 are inclusion body samples recovered-from E. coli BL21 (DE3): pLysE: pTCGTK10 and E. coli BL21 (DE3): pLysE: pLPK10, respectively. Lanes 2,3,4, and Lanes 6,7,8 correspond, respectively, to enzyme solutions purified by selective binding the inclusion bodies recovered from E. coli BL21 (DE3): pLysE: pTCGTK10 and E. coli BL21 (DE3): pLysE: pLPK10 in 300,350 and 400 mM NaCI, washing with the same salt solutions, and desorbing with 1M NaCl. These

results suggest that the protein denatured by urea can be purified by way of exploiting the fused lysine moiety to a high purity level.

Example 13: Purification of disulfide-bond-promoting enzyme, disulfide bond isomerase and proline isomerase, each containing 10 lysines and arginines E. coli BL21 (DE3): pLysE (Novagen, U. S. A.) was transformed with each of the expression vectors pDSBAK10, pDSBCK10, pHPPIK10, pDSBSR10, pDSBCR10 and pHPPIR10 prepared in Example 1 (3) and transformed colonies were each selected from a plate containing ampicillin to obtain E. coli BL21 (DE3): pLysE: pDSBAK10, E. coli BL21 (DE3): pLysE: pDSBCK10, E. coli BL21 (DE3): pLysE: pHPPIK10, E. coli BL21 (DE3): pLysE: pDSBAR10, E. coli BL21 (DE3): pLysE: pDSBCR10, E. coli BL21 (DE3): pLysE: pDSBAR10, E. coli BL21 (DE3): pLysE: pDSBCR10 andE. coli BL21 (DE3): pLysE: pHPPIR10.

Disulfide-bond-promoting enzyme, disulfide bond isomerase and proline isomerase, each containing 10 lysines and arginines, were purified in accordance with the method described in Example 3. As the result in Fig.

21 shows, these enzymes can be effectively purified by exploiting the fused cationic amino acid residues, as in the case of CGTase.

Example 14: Refolding of disulfide-bond-promoting enzyme, disulfide bond isomerase and proline isomerase immobilized on a solid through cationic amino acid fusion system Disulfide-bond-promoting enzyme, disulfide bond isomerase, each containing fused cationic amino acid residues, were immobilized on a resin and subjected to refolding as in Example 11 and the results are depicted in Figs. 22,23,24 and 25. As the figures show, the protein refolding proceeded smoothly without the occurrence of protein aggregation even at a protein concentration of more than 1 mg/n-l. This process is advantageous

over refolding in solution in that the process volume can be reduced while enhancing the refolding efficiency.

BUDAPEST YREATY ON THE INTERNATIONAL RECOGNITION OF THE DEPOSE OFMICROORGANISM FOR THE PURPOSE OF PATENT PROCEDURE INTERNATIONAL FORM RECEIPT IN THE CASE OF AN ORIGINAL DEPOSIT issued pursuant to Rule 7. 1 TO : SEO Jin-Ho Limkwang Apt. 311-601, Seohyun-dong, Bundang-ku, SSeongnam-si, Kyunggi-do 463-050, Republic of Korea 1. IDENTIFICATION OF THE MICROORGANISM Ident, i, tion refsrence iven by th GSion nuxtr given by t$e IdenfEeationreferencegivenby ie scessionnunber gvenby ie DEPOSITOR.'INTERNATIONAL DEPOSITARY AUTHOPJTY : Ascharkhia coli KCTC 0842BP ,,,., _. _. U. SCENTSt DUCZPTlON AND/OR PROP05ED TMONOC DESIGNATION The micmorgausmidenSi mderlabove wasaccompatedby : xa scientific description ] a proased taxonomie designation (Mark with a cross where applicable) m RECEIPT AND ACCEPTANCE This International Depositary Authority accepts the microorganism identified under l abve, which was receive by it on Jn ! y 24 2000. IV. RECEIPT OF REQUEST FOR CONVERSION The microorganism identified ender I above was received by this International Depositary Authority on and a request to convert the original deposit to a deposit under the Budapest Treaty was received by it on V. INTERNATIONAL DEPOSITARY AUTHORITY Name : Korean Co)) ect ! on for Type Cultures Signature (s) of person (s) having the power to represent the International Depositary Authority of aut} lorized of : s) : Address : Korea Research Institute fR v z Binscience arid Biotechnology (KRIBB) #, tun-don, Yusong-ku, $AE, Kyung ook, ictor "faejon 3Q-33, Republic of Korea Date: July 27 2000 FormBP/4 (KCTC Form 17) sole page