Fusion proteins, their preparation and use.

Proteins which in addition to the desired protein also have an undesirable constituent or ballast constituent in the final product are referred to as fusion proteins. When proteins are prepared by genetic engineering, the intermediate phase of a fusion protein is used particularly if, in direct expression, the fusion protein is relatively rapidly decomposed by endogenous host proteases, resulting in low or totally inadequate yields of the fusion protein. desired protein.

The amount of the fusion protein balast constituent is usually chosen so as to obtain an insoluble fusion protein. This insolubility not only provides the desired protection against endogenous host proteases but also allows for easy seizure. comparison of insoluble cell components. It is generally considered acceptable that the proportion of protein desired in the fusion protein is relatively small, that is, that the cell produces a relatively large amount of balast,

The preparation of fusion proteins with a small constituent of balasto was attempted. For example, a fusion gene encoding a fusion protein was prepared from the first ten amino acids of β-galactosidase and somatosine. However, it was observed that this short chain of amino acids did not adequately protect the fusion protein against decomposition by endogenous host proteases (US-A 4 336 246, Column 15, Paragraph 2).

From EP-A 0 290 005 and 0 292 763, fusion proteins are known whose balasto constituent consists of a β-galactosidase fragment having more than 250 amino acids. These fusion proteins are insoluble, but they can be easily solubilized with urea (EP-A 0 290 005).

The invention relates to a process for the preparation of fusion proteins wherein a mixed oligonucleotide encoding the fusion protein balast constituent is constructed by introducing this oligonucleotide into a vector so that it is functionally linked to a region. regulatory gene and structural gene for the desired protein, transforming suitable host cells with the plasmid population thus obtained and clones producing a high yield of fusion protein are chosen. The best embodiments of this invention are explained below:

The plasmid advantageously encodes at the 3 'end an amino acid or group of amino acids that allows for easy, and preferably enzymatic, cleavage of the desired protein balastate constituent. According to another embodiment, an oligonucle2 * 1 is constructed.

<img file="PT95111B_D0001.tif" />

It produces an insoluble fusion protein that can be easily solubilized. In particular, an oligonucleotide encoding a ballast constituent that does not bend the folding of the desired protein is preferably constructed.

Due to practical reasons, the construction according to the invention of the oligonucleotide to the balasto constituent makes the latter a very small one.

It is surprising to note that even when they have an extremely small balast constituent, fusion proteins not only meet the requirements for protease resistance, but are also introduced at a high expression rate and, if desired, if the protein If the melting agent is insoluble, it can easily be solubilized. In the dissolved or soluble state, the small balasto constituent according to the invention then permits a sterically favorable conformation of the desired protein so that it can be properly folded and easily separated from the balasto constituent.

If the desired protein is formed into a proform, the balasto constituent may be constituted such that its cleavage may occur simultaneously with the transformation of the proprotein into mature protein. In the preparation of insulin, for example, the balasto constituent and the C chain may be removed simultaneously, giving rise to a mature insulin derivative that can be transformed into insulin without any side reactions involving large losses.

The small constituent of balasto according to the invention is actually smaller than the usual protein signal sequences and does not disturb the desired protein folding. It thus does not need to be disposed of before the final processing phase that gives rise to the mature protein.

The oligonucleotide encoding the balasto constituent preferably contains the DNA (coding chain) sequence (dcd)_x wherein D represents A, G or T ex is 4-12, preferably 4-8

In particular, the plasmid is characterized by a DNA sequence (coding strand)

ATG (DCD)_y (NNN)_Z where N in the triplet NNN represents the same or different nucleotides, excluding codon stops, Z is 1-4 and y + z is 6-12, preferably 6-10, where y is at least 4. It has been found advantageous for the oligonucleotide has the DNA sequence (coding strand)

ATG (DCD) 5_g (NNN) especially if DNA sequence (coding cad)

GCW ATG (DCD) 4__B CGW or, advantageously

ATG GCA (DCD) 4 7 CGW where W represents A or T.

The above DNA template sequences meet all of these requirements. The CDC codes encode the amino acids serine, threonine and alanine and thus a relatively hydrophilic protein chain. Stop coding is excluded and the choice of amino acids remains easy. The following is a particularly preferred embodiment of the DNA sequence for the balasto constituent, especially if the desired protein is proinsulin:

ATG GCW (DCD) y, CGW or

ATG GCD (DCD) y¹ CGT

<img file="PT95111B_D0002.tif" />

where y 'means 3 to 6, especially 4 to 6.

second codon, GCD, codes for alanine and completes the recognition sequence for restriction enzyme NcoI, provided that the above regulatory sequence ends with CC. The triplet near the last one codes for threonine and, together with the CGT codon for arginine, represents the recognition sequence for the restriction enzyme M1I. Accordingly, this oligonucleotide can be easily and unambiguously incorporated into gene constructs.

group (NNN)_Z it encodes at the 3 'position for an amino acid or group of amino acids which allows for simple and preferably enzymatic / separation of the balasto constituent from the desired subsequent protein. It is advisable to choose the nucleotides in this group so that at the 3 'end they will decode the cleavage site of a restriction enzyme that allows the structural gene to bind to the desired protein. It is also advantageous for the ATG start codon and if necessary the first DCD triplet to be incorporated into the restriction enzyme recognition sequence such that the gene for the balasto constituent according to the invention can be easily inserted into the usual vectors. .

The upper limit of z is obtained on the one hand at the desired cleavage site for the (enzymatic) cleavage of the fusion protein obtained, that is, it encompasses codons, for example for the amino acid sequence Ile-Glu-Gly-Arg, at if cleavage is performed with factor Xa. In general, the upper limit for the sum of y and z is 12, since the balasto constituent should obviously be as small as possible and, above all, not interfere with the folding of the intended protein.

For practical reasons, bac5 are preferred

<img file="PT95111B_D0003.tif" />

lower eukaryotic cells or cells such as fungi as host organisms in genetic engineering processes, provided that no higher organisms are required. In these processes, expression of the heterologous gene by a homologous regulatory region, that is, one that is most intrinsic to the host or compatible with the host cell. If a pre-peptide is expressed, it also often happens that the pre-sequence is also heterologous to the host cell. In practice, this lack of sequence harmony often results in variable and unpredictable protein yields. Since the ballast sequence according to the invention is adapted to the environment, the selection process according to the invention produces a DNA construct that is characterized by this sequence harmony.

beginning and end of the baastal constituent are set forth in this construct: ethionine is at the beginning, and an amino acid or group of amino acids that allows the desired breakdown of the desired protein balast constituent is at the end. If, for example, the desired protein is proinsulin, it is advantageously chosen from an NNN triplet coding for arginine as the last codon, as this allows for simultaneous cleavage separation. particularly favorable area of ​​the balasto constituent with the removal of the C chain. Of course, the end of the balasto constituent may also be an amino acid or group of amino acids that allows chemical cleavage , for example methionine, so that cleavage is possible with bromide or cyanogen chloride.

The intermediate amino acid sequence should be as small as possible so that the folding of the desired protein is not affected. In addition, this chain should be relatively hydrophilic so that solubilization is facilitated with undissolved fusion proteins and the fusion protein remains soluble. Cysteine ​​residues are undesirable as they may interfere with the formation of disulfide bridges.

<img file="PT95111B_D0004.tif" />

DNA encoding the balasto constituent is synthesized as a mixed oligonucleotide; It is incorporated into a suitable expression plasmid immediately in front of the structural gene for the desired protein and E. coli is transformed with the set of genes obtained in this manner. Suitable gene structures can be obtained in this way. Suitable gene structures can be obtained by selecting the bacterial clones that produce the corresponding fusion proteins.

It has previously been stated that cleavage sites for restriction enzymes at the beginning and end of nucleotide coding for the balasto constituent should be regarded as examples only. Recognition sequences encompassing the ATG start codon and in which any nucleotides following may include the codon for suitable amino acids are, for example, also for the user for restriction enzymes AflIII, Ndel,

Nlalll, NspHI or Styl. Since in the preferred embodiment arginine should be at the end of the ballast sequence and since there are six different codons for arginine, additional restriction enzymes can also be found here to be used in place of Mlul, e.g. Nurl, Avrll, AflIII Clal or Haell.

However, it is also advantageous to use a polymerase chain reaction (PCR) according to Saiki, RK et al., Science 239: 487 491, which may dispense with the construction of restriction enzyme-specific recognition sites.

It has been previously indicated that DNA sequence limitation (DCD) x is for ease of reasons and this does not eliminate other codons such as glycine, proline, lysine, methionine or asparagine.

The most effective embodiment of this DNA sequence is obtained by selecting good fusion protein producers, that is, for example, from the proinsulin-containing fusion protein. This yields the most favorable combination of regulatory sequence, ballast sequence and desired protein, and as a result of which unfavorable combinations of promoter, ballast sequence and structural gene are avoided and good results are obtained with minimal expense in terms of the harmony of above sequence.

It has been surprisingly found that the genes optimized for the balasto constituent according to the invention did not always contain the preferred triplets by E.coli. For Thr, the ACA codon, which is used less frequently by E. coli, was found to actually and often occur in the chosen sequences. If, for example, the following amino acid sequences are optimized

<td colspan="5">according to preferred codon</td><td colspan="3">used by E. coli (pcu; cf.</td>

<td>Aota,</td><td>S. and</td><td>lime:</td><td>Nucleic</td><td colspan="2">Acids Research 16</td><td colspan="2">(supplement) r315,</td>

<td>r316,</td><td>r391,</td><td>r402</td><td> [1988])</td><td colspan="3">, you get a structure</td><td>gene</td>

<td colspan="5">quite different from that obtained</td><td>according</td><td colspan="2">the invention.</td>

<td>(Cf.</td><td>Table</td><td> 1):</td><td></td><td></td><td></td><td></td><td></td>

<td>Allah</td><td>Thr</td><td>Thr</td><td>To be</td><td>Thr</td><td>Thr wing</td><td>Thr</td><td></td>

<td>GCG</td><td>ACC</td><td>ACC</td><td>AGC</td><td>ACC</td><td>GCG ACC</td><td>ACC</td><td>pcu</td>

<td>GCA</td><td>A CA</td><td>A CA</td><td>TCA</td><td>A CA</td><td>GCA ACT</td><td>ACG</td><td>invention</td>

<td></td><td></td><td></td><td colspan="2">In case</td><td>of proteins</td><td colspan="2">merger with a</td>

proinsulin constituent, the starting point was a 10 amino acid balastate constituent. The DNA sequence; from the best producer then served as the basis for variations in this sequence, having been noted that up to 3 amino acids can be deleted without a noticeable loss in relative expression rate. This conclusion is not only surprising, as it is unexpected that such a small balasto protein is adequate, but also very advantageous since obviously the relative proportion of proinsulin in the fusion protein increases as the balasto constituent decreases.

The importance of the balasto constituent in protein is apparent from the following conclusion: human proinsulin contains 86 amino acids. If, for a fusion protein according to EP-A 0 290 005, the limit of 250 amino acids for the balasto constituent is taken, the fusion protein has 336 amino acids, of which only about a quarter occur in the desired protein. In comparison, a fusion protein according to the invention with only 7 non-constituent amino acids of balasto has 93 amino acids, that is, the proinsulin constituent constitutes 92.5%. If the desired protein has many more amino acids than proinsulin, the relationship between balasto and the desired protein becomes even more favorable.

It has been previously noted that a desired proinsulin protein represents only a preferred embodiment of the invention. However, the invention also works with much larger fusion proteins of which an example fusion protein with the active domain of 3-hydroxy-3-methylglutaryl coenzyme A reductase is mentioned. This protein contains 461 amino acids. A gene coding for the latter is known from EP-A 292 803.

Brief Explanation of Drawing:

Figure 1 and its continuation in Figure 1a and Figure 1b shows the construction of a population of the known plamid plasmid (gene set) plNT4x pH154 / 25 * through plasmid pINT40.

The other constructs were not shown graphically as they are readily apparent to one skilled in the art from the Figures.

The invention will be explained in more detail in the following examples. Percentage data refers to weight if not stated otherwise.

Γ

Unless otherwise indicated, all media are prepared according to Maniatis, T; Fritsch, EF and Sambrook, J .; Molecular Cloning, Cold Spring Harbor Laboratory (1982). TP medium consists of M9CA medium but with a glucose and casamino acid content of 0.4% each. Unless otherwise indicated, all media contain 50 µg / ml ampicillin. Bacterial growth during fermentation is determined by measuring the optical density of the cultures at 600 nm (0D).

Example 1: Construction of high expression clone assembly.

of genes and selection of a starting material is plasmid pH154 / 25 * (1), which is known from EP-A 211 299. This plasmid contains a fusion protein gene (D'-Proin) linked to a promoter-trp and a resistance gene for resistance against the ampicillin antibiotic (Amp). The fusion protein gene that contains an E. coli pr-trpD fragment (D ') and monkey proinsulin (Proin). The structure of the plasmid gene results in polycistronic mRNA, which encodes the fusion protein and the resistance gene product. To suppress the formation of the excess resistance gene product, the (commercial) transcriptional terminator sequence was initially introduced between the two structural genes. For this, the plasmid is opened with EcoRI and the protruding terminals are filled with Klenow polymerase. 0 The resulting blunt-ended DNA fragment is ligated with the terminator sequence (2).

5 'AGCCCGCCTAATGAGCGGGCTTTTTTTT3'

3'TCGGGCGGATTACTCGCCCGAAAAAAA5¹ (2) which results in plasmid pINT12 (3).

pH154 / 25 * starting plasmid contains a cleavage site for the Pvul enzyme in the Amp gene, as well as a Hindlll cleavage site in the carboxy terminal area of ​​the trp-d fragment. Both cleavage sites are thus also contained in pINT12 (3). By cutting plasmid (3) with Pvul and HindIII, it is divided into two fragments in which the proinsulin gene (4) is isolated. 0 plasmid pGATTP 5, which is analogously structured to (3) but which instead of the D'-Proin gene contains an interferon gamma (Ifn) gene containing NcoI and HindIII restriction cleavage sites, is also cut with Pvul and HindIII. and the fragment (6) containing the promoter region is isolated. By ligation of this fragment (6) with fragment (4) obtained from (3), plasmid pINT40 (7) is obtained. The small fragment with the remaining part of the interferon gamma gene is cut from the latter with NcoI and M1ul. 0 fragment (8) is ligated with mixed plasmid (9).

5'CATGGCDDCDDCDDCDDCDDCDDCDA3 '

3'CGHHGHHGHHGHHGHHGHHGHTGCGC 5 '(9) wherein D represents A, G or T and H means the complementary nucleotide. This results in a population of plasmid (gene set) pINT4x (10).

mixed oligonucleotide (9) is obtained from synthetic mixed oligonucleotide (9a)

TTCGGGTACCGHHGHHGHHGHHGHHGHHGHTGCGCAG5 '

TTGCCCATGGC3 '(9a) which is filled with Klenow polymerase and cut with Mui and Ncol.

The E. coli strain WS3110 is transformed with the plasmid population (10) and the bacteria are placed on LB agar disc. Six of the resulting bacterial clones are tested to determine their ability to produce a fusion protein with an insulin constituent. For this purpose overnight cultures are prepared in LB medium 100 ml aliquots of the cultures are mixed with 10.5 ml TP medium and shaken at 37 ° C.^sC. At OD600 = 1 the cultures are

<img file="PT95111B_D0005.tif" />

adjusted to 20 pg / ml 3-p-indolacrylic acid (IAA), a solution of 40 mg glucose in 100 ml water is added and the preparation is stirred for a further three hours at 37 ° C.²C. After 6 OD equivalent of culture are removed, the bacteria contained therein are collected by centrifugation and resuspended in 300 µl assay buffer (37.5 mM tris pH 8.5, 7 M urea, 1% (w / v) SDS and 4% (V / V) 2-mercaptoethanol). The suspension is heated for five minutes, ultrasonically treated for two seconds to reduce viscosity, and then aliquoted with SDE gel electrophoresis. With the bacteria producing the fusion protein, one would expect a protein band with a molecular weight of 10 350 D. It is evident that one of the clones, pINT411 (11), produces a suitable protein in relatively large quantities while it is not. this protein formation is observed with the remaining clones. An experiment with immune disclosure with insulin specific antibodies confirms that the protein encoded by (11) contains an insulin constituent.

Table 1 illustrates the DNA and amino acid sequence of the balasto constituent in this fusion protein (as well as other variants of these sequences).

Example 2: Selecting Additional Clones

To detect additional suitable clones, a method according to Helfman, DM ecol is used. (Proc. Natl, Acad. Sci. USA 80: 31-35, 1983). TP-agar discs and medium containing an additional 40 µg / ml IAA are used for this purpose. Fifteen minutes before use, the agar surface of the discs is covered with a top layer of 2mm thick TP agar, a last nitrocellulose filter is placed and freshly transformed cell is placed in the filter. Copies are made of filters on which bacterial colonies grew after incubation at 37 ° C.^sC, and the bacteria of the original filter are lysed. To achieve this, the filters are exposed to a chloroform atmosphere in a desiccator for 15 minutes, then slowly moved for six hours at room temperature in an immune buffer (50 mM Tris pH 7.5.150 mM NaCl, 5 mM MgCl₂. and 3% (w / v) BSA), which contain an additional 1 µg / ml DNase I and 40 µg / ml lysozyme, and then washed twice for five minutes with wash buffer (50 mM tris pH 7.5 and 150 mM NaCl) The filters are then incubated overnight at 3 ° C.^sC in immune buffer with insulin-specific antibodies, washed four times for five minutes with wash buffer, incubated for one hour in immune buffer with a horseradish protein A-peroxidase conjugate, washed again four times for five minutes with buffer wash and colonies that have antibodies linked with a color reaction are visualized. Clones pINT42 and pINT43, which also produce relatively large amounts of fusion protein, are thus discovered in 500 colonies. Sequencing DNA and amino acid sequences derived therefrom were also reproduced in Table 1.

Example 3 ·. Preparation of Plasmid pINT41d

Between the origin of replication and the trp promoter, plasmid, pINT41 contains a nonessential DNA region that is flanked by cleavage sites for the Nsp (7524) enzyme 1: To remove this region of plasmid, pINT41 is cut with NSP (7524) 1, and the largest of the resulting fragments is isolated and reconnected. This gives rise to plasmid pINT41d, whose DNA sequence is reproduced in Table 2.

Example 4; Fermentation and processing of pINT41d fusion protein (i) Fermentation: An LB agitation culture is prepared from E. coli W3110 transformed with pINT41d. Fifteen microliters of this culture, which has an OD value of 2, are then placed in 15.7 l of TP medium and the suspension is fermented for 16 hours at 37 ° C.^sÇ.

The crop, which at this point has a value of OD = 13, is

V

Ι^χ·

<img file="PT95111B_D0006.tif" />

then adjusted to 20 æg / ml IAA, and until the end of the fermentation another five hours, a 50 (w / v) maltose solution is continuously pumped at a rate of 100 ml / hour. A value of 0D = 17.5 is reached in this process. In the end, the bacteria are harvested by centrifugation.

(ii) cell disruption; The cells are resuspended in 400 ml disintegration buffer (10 mM tris pH 8.0, 5 mM EDTA) and disrupted in a French press. The insulin-containing fusion protein is further concentrated during 30 minutes of centrifugation at 23,500 g and washed with disintegration buffer, 134 g of pellet (wet substance) is obtained.

(iii) Sulpholysis: 12.5 g of sediment (wet substance) of (ii) is stirred in 125 ml of an 8M urea solution at 35 ° C.^andç. After stirring for 30 minutes, the solution is adjusted to pH 9.5 with sodium hydroxide solution and reacted with 1 g of sodium sulfite. After another thirty minutes of stirring at 35^sC, 0.25 g of sodium tetrathionate are added and the mixture is stirred again for thirty minutes at 35 ° C.^andÇ.

(iv) DEAE. anion exchange chromatography: The entire batch of (iii) is diluted with 250 ml buffer A (50 mM glycine, pH 9.0) and placed on a chromatography column containing Fractogel (R) TSK DEAE-650 (volume 130 ml column diameter 26 mm) equilibrated with buffer A. After washing with buffer A, the S-sulfonate fusion protein is eluted with a salt gradient consisting of 250 ml each of buffer A and buffer B (50 mM glycine pH 9.0, 3 M urea and 1 M NaCl) at a flow rate of 3 ml / min. The fractions containing the S-sulfonate fusion protein are then combined.

(v) Enzymatic folding and cleavage; The combined fractions (iv) are diluted to 4^sC in a volume ratio of 1 + 9 with folding buffer (50 mM glycine, pH 10.7) and for each liter of the resulting dilution 410 mg of scoric acid and 165 ml of 2-mercapoethanol at 42c are added with stirring. moderate. After correcting the pH to 10.5, stirring is continued for a further 4 hours at 4 ° C.^Qç. Subsequently, N- (2-hydroxyethyl) piperazine-N'-2-ethanesulfonic acid (HEPES) is added to a final concentration of 24 g per liter batch. The mixture now having a pH 8 is digested with trypsin at 25 ° C. During processing the enzyme concentration in the digestion mixture is 80 pg / l. The course of cleavage is analytically followed by RP-HPLC. After two hours, digestion may be stopped by the addition of 130 æg soybean trypsin inhibitor. 0 HPLC assay reveals the formation of 19.8 mg di-Arg insulin from a mixture according to (iii). The identity of the cleavage product is confirmed by sequencing and comparative HPLC with reference substances.

Di-Arg insulin may be purified chromatographically according to known procedures and transformed into insulin with carboxypeptidase B.

Example 5: Construction of Plasmid pINT60 Plasmid pINT60 results in an insulin precursor whose balasto sequence consists of only nine amino acids. For construction of this plasmid, plasmid pINT40 is cut with NcoI and M1ul and the resulting vector fragment is isolated. 0 Insul5 plasmid

TTCGGGTACCGTTGTTGTAGTTTGAGTTGCGCAG 5 '

TTGCCCATGGC 3 'is then synthesized, filled with Klenow polymerase and also cut with these two enzymes. The resulting DNA fragment is then ligated with the vector fragment to obtain plasmid pINT60.

Table 1 shows the DNA and amino acid sequence of the balasto constituent in this fusion protein.

Example 6: Construction of Plasmid pINT67d Plasmid pINT67d is a derivative of pINT41D wherein the amino acid codon at position 9 of the balasto sequence is deleted. This is because, like pINT60, it results in an insulin precursor with a nine amino acid ballast sequence. A process according to Ho, SN et al. (Gene 77: 51-59,1989) is used for its construction. To this end, two separate PCRs with plasmid pINT41d and the two oligonucleotide pairs are performed first.

<td rowspan="2">TIR: DTR8:</td><td rowspan="2"> 5' 5'</td><td rowspan="2">-CTG AAA TGA -CAC AAA TCG</td><td rowspan="2">Gct AGT</td><td rowspan="2">GTT TGC</td><td colspan="2">GAC-3 '</td><td colspan="2">and</td>

<td>TGT</td><td>TGA</td><td>TGT</td><td>TGT-3 'or</td>

<td>DTR9:</td><td> 5'</td><td>-ACA GCA ACT</td><td>CGA</td><td>TTT</td><td>GTG</td><td>AAC</td><td>CAG</td><td>CAC-3 'and</td>

<td>Insull:</td><td> 5'</td><td>-TCA TGT TTG</td><td>A CA</td><td>Gct</td><td>TAT</td><td>CAT-</td><td> 3'</td><td></td>

<td>This produces</td><td>two</td><td colspan="5">fragments that are partially</td><td colspan="2">complemented</td>

summarize each other and when stabilized with each other encode a similar insulin precursor such as pINT41d in which, however, the amino acid at position nine is absent. To complete, the two fragments are combined and subjected to another PCR together with the TIR and Insull oligonucleotides. From the DNA fragment thus obtained, the insulin precursor structural gene is released with NcoI and SalI and purified. 0 plasmid pINT41d is then also cut with these two enzymes, the vector fragment is purified and then ligated with the PCR structural gene fragment to obtain plasmid pINT67d.

The nucleotide and amino acid sequences for the balasto region were reproduced in Table 1.

Example 7: Construction of Plasmid pINT68d

Like plasmid pINT67d, plasmid pINT68d is a reduced derivative of plasmid pINT41d in which the codons of the two amino acids at positions eight and nine of the balasto sequences are deleted. This results in an insulin precursor with a ballast sequence of only eight amino acids. The procedure described above in Example 6 is used for its construction but with the oligonucleotide pairs.

<td>TIR:</td><td> 5'</td><td>-CTG</td><td>AAA</td><td>TGA</td><td>Gct</td><td>GTT</td><td>GAC-3 '</td><td>and</td><td></td>

<td>DTR10:</td><td> 5'</td><td>-CAC</td><td>AAA</td><td>TCG</td><td>TGC</td><td>TGT</td><td>TGA TGT TGT</td><td>TGC-3 '</td><td>or</td>

<td>DTR11:</td><td> 5'</td><td>-TCA</td><td>A CA</td><td>GCA</td><td>CGA</td><td>TTT</td><td>GTG AAC CAG</td><td>CAC-3 '</td><td>and</td>

<td>Insull:</td><td> 5'</td><td>'-TCA</td><td>TGT</td><td>TTG</td><td colspan="3">ACA GCT TAT CAT-3 '</td><td></td><td></td>

The nucleotide and amino acid sequences for the balasto region were reproduced in Table 1.

Example 8: Construction of Plasmid pINT69d Plasmid pINT69d is also a reduced derivative of plasmid pINT41d wherein, containing the canons of the three amino acids at positions seven, eight and nine of the ballast sequence were deleted. This results in an insulin precursor with a ballast sequence of only seven amino acids. The procedure described in Example 6 is also used for its construction but with the two oligonucleotide pairs.

<td>TIR:</td><td> 5'</td><td>-CTG</td><td>AAA</td><td>TGA</td><td>Gct</td><td>GTT</td><td>GAC-</td><td> -3'</td><td></td><td>and</td><td></td>

<td>DTR12:</td><td> 5'</td><td>-CAC</td><td>AAA</td><td>TCG</td><td>TGT</td><td>TGA</td><td>TGT</td><td>TGT</td><td>TGC</td><td>CAT-3 '</td><td>or</td>

<td>DTR13:</td><td> 5'</td><td>'-A CA</td><td>TCA</td><td>A CA</td><td>CGA</td><td>TTT</td><td>GTG</td><td>AAC</td><td>CAG</td><td>CAC-3 '</td><td>and</td>

<td>Insull:</td><td> 5’</td><td>'-TCA</td><td>TGT</td><td>TTG</td><td>A CA</td><td>Gct</td><td>TAT</td><td>CAT-</td><td> -3'</td><td></td><td></td>

The amino acid nucleotide sequences for the balasto region were reproduced in Table 1.

Example 9: Construction of Plasmid pINT72d

<img file="PT95111B_D0007.tif" />

Plasmid pINT72d is a derivative of plasmid pINT69d wherein the entire C-peptide region of the gene with the exception of the first codon of the amino acid arginine is deleted. This results in a miniproinsulin derivative with an arginine residue instead of a C-chain. With plasmid pINT69d as the starting point, the procedure described in Example 6 is also used for its construction, but with the two oligonucleotide pairs.

<td>TIR:</td><td>5'-CTG</td><td>AAA</td><td>TGA</td><td>Gct</td><td>GTT</td><td>GAC-</td><td> -3'</td>

<td>Insu28:</td><td>5'-GAT</td><td>Gcc</td><td>GCG</td><td>GGT</td><td>CTT</td><td>GGG</td><td>TGT-3</td>

<td>Insu27:</td><td>5'-AAG</td><td>ACC</td><td>CGC</td><td>GGC</td><td>ATC</td><td>GTG</td><td>GAG-3</td>

<td>Insull:</td><td>5'-TCA</td><td>TGT</td><td>TTG</td><td>A CA</td><td>Gct</td><td>TAT</td><td>CAT-3</td>

Example 10: Construction of Plasmids pINT73d, pINT88d epINT89d Plasmid pINT73d is also a derivative of plasmid pINT69d (Example 8), wherein the insulin precursor is arranged twice in succession. The plasmid thus results in the formation of a polycistronic mRNA, which can double the yield. For its construction, a PCR reaction is performed with plasmid pINT69d and the two oligonucleotides.

Insu29: .5'-CTA GTA CTC GAG TTC AC-3 'and

Insull: 5'-TCA TGT TTG ACA GCT TAT CAT-3 '

This gives rise to a fragment with the insulin precursor gene and the relevant ribosome binding site which at its 5'-terminal region has a cleavage site for the enzyme XhoI and at its 3 'terminal end a cleavage site for Sall. The fragment is cut with the above two enzymes and purified. 0 plasmid pINT69d is then linearized with SalI, the two terminal DNAs produced are phosphatase (calf intestine) dephorphylated and ligated with the PCR reaction fragment to obtain plasmid pINT73d.

<img file="PT95111B_D0008.tif" />

Plasmids pINT88d and pINT89d are obtained analogously when plasmid p.INT72d (Example 9) is modified analogously by arrangement of the miniproinsulin gene two or three times in sequence.

Example 11: Construction of plasmid pINL41d starting plasmid pRUD3 has a structure similar to that of plasmid pGATTP. However, instead of the trp-promoter region, it contains a tac-promoter region that is flanked by cleavage sites for the EcoRI and NcoI enzymes. The plasmid is cut with EcoRI, after which protruding ends of the cleavage site are filled with Klenow polymerase. Cutting is then performed with NcoI and the resulting promoter fragment is isolated.

plasmid pINT41d trp promoter is flanked by cleavage sites for the enzymes PvuII and NcoI.

Since the plasmid has an additional cleavage site for PvuII, it is completely cut with NcoI, but only partially, with PvuII. The vector fragment, lacking only the promoter region, is then isolated from the resulting fragments. It is then ligated with the tac-promoter fragment to yield plasmid pINL41d.

Example 12: Construction of Plasmid pL41c Plasmid pPL-lambda (obtainable from Pharmacia) has a lambda-pl-promoter region. The latter is flanked by nucleotide sequence:

5'GATCTCTCACCTACCAAACAAT3 'and

5'AGCTAACTGACAGGAGAATCC3 '.

5'ATGAATTCGATCTCTCACCTACCAAACAAT 3 'and 5'TTGCCATGGGGATTCTCCTGTCAGTTAGCT 3'

LPL3 Oligonucleotides:

LPL4:

are prepared by further flanking the promoter region with EcoRI and NcoI enzyme cleavage sites. PCR is performed with these plasmid and pPL-lambda and the resulting promoter fragment is cut with EcoRI and NcoI and isolated. Plasmid pINL41d is then also cut with these two enzymes and the resultant vector fragment, which has no promoter, is then ligated with the lambda-pl-promoter fragment to obtain plasmid pL41c.

Example 13: Construction of the trp-transcriptional terminator plasmid pL41d located between the resistance gene and the fusion protein gene on plasmid pL41c is not effective in E. coli strains that are suitable for fermentation (e.g. E. coli N4830- 1). For this reason, a polycistronic mRNA with a large amount of resistance gene product is formed in the fermentation. To avoid this side reaction, the trp-terminator sequence is replaced by an effective E. coli-rnnB terminator sequence. Plasmid pANGMA has a similar structure to plasmid pINT41d, but has an angiogenin gene instead of the fusion protein and an rrnB sequence (from the commercial plasmid pkk223-3, which can be obtained from Pharmacia) instead of the trp-terminator sequence. . The plasmid is cut with Pvul and SalI and the fragment containing the RnnB terminator is isolated. 0 Plasmid pL41c is then also cut with these two enzymes and the fragment containing the insulin gene is isolated. The two isolated fragments are then ligated to give plasmid pL41d.

Example 14: Construction of pINTI Plasmid

To prepare a plasmid for general use in expression of fusion proteins, the proinsulin gene of plasmid pINT41d is replaced by a polylinker sequence. This gene is flanked by cleavage sites for the enzymes Mlul and Sall. The plasmid is thus cut with the aid of the two above enzymes and the fragment is isolated. It is then ligated to give pINTLI plasmid with the following two synthetic plasmids.

BstEII Accl EcoRI Kpnl BamHI

5 'CGCGCCTGGTTACCTCGAGGTATACTACGAATTCGAGCTCGGTACCCGGGGATCC

GGACCAATGGAGCTCCATATGATGCTTAAGCTCGAGCCATGGGCCCCTAGG

Xhol Saci Xmal sphl Xbal

CTGCAGGCATGCAAGCTTGTCTAGAC -3 '

GACGTCCGTACGTTCGAACAGATCTGAGCT-5 '

PstI HindIII (SalI).

Example 15: Insertion of a gene encoding HMG CoA reductase (active domain) into pINTLI and expression of fusion protein.

EP-A known HMG coA reductase synthetic gene 290803 contains a BstEII cleavage site in the amino acid region Leu and Val at positions 3 and 4 (Table 5 of EP 0 292 803 A2). A salient sequence corresponding to the Xbal enzyme occurs at the end of the gene (in the non-coding area). The corresponding cleavage sites on the pINTLI plasmid polylinker are in the same reading frame. Both cleavage sites are in each case unique.

If pINTLI is cut with BstEII and Xbal and the large fragment is isolated, and if, on the other hand, piasmium pUHIO (EP-A 0 292 803) is digested with the same enzymes and the fragment containing most of the DNA sequence of this plasmid is isolated after ligation of the two fragments, a plasmid encoding a fusion protein is obtained in which arginine follows the first eight amino acids in the pINT41d ballast sequence (Table 1), which is then followed by Leu ^, by the HMG CoA reductase active domain structural gene.

L_

<img file="PT95111B_D0009.tif" />

For comparison purposes, the two initial plasmids are cut with the enzymes NcoI and XbaI and the corresponding fragments are ligated together to obtain a coding plasmid immediately after the splitting dog. the active domain of HMG CoA reductase (according to EP-A DNA sequence 0 292 803)

Expression of the encoded proteins occurs according to Example 4. After cell disruption, centrifugation is performed after which the expected protein of approximately 55 kDa is determined in the supernatant by gel electrophoresis. The band for fusion protein is much more intense here than for directly expressed protein. Individual 100 µl parts of the supernatant are assayed in undiluted form at a 1:10 dilution and a 1: 100 dilution for mevalonate formation. As a further comparison tested to the fusion protein according to Example 4 (fusion protein with insulin constituent); No activity is observed at any of the three concentrations. The fusion protein with the HMG CoA reductase constituent shows maximum activity at all three dilutions, while the direct expression product shows graded activity dictated by concentration. This indicates better axpression than the fusion protein of a factor of at least 100.

Example 16: Construction of Plasmid pB70 Plasmid pINT41d is divided with M1I and SalI and the larger fragment is isolated. Plasmid pIK4 (EP-A 0 347 781, published December 27, 1989, or AU-A 3671/89 or ZA-A 89/4742 or German Patent Application P 38 21 159.9 or US Application 07/369 686, filed June 21, 1989; in any of these documents Figure 1a) contains a mini-proinsular gene whose C-chain consists only of arginine.

The construction of this plasmid is described in Example 2, Parts a) and b) of EP-A 0 347 781 (or a corresponding document). They were reproduced in the appendix together with both^belas in gene fragments Ik I and Ik II, as well as in Figs 22 and 1a. In these tables, the B and A chains of the insulin molecule are in each case indicated by the first and last amino acids. Following the coding region in the Ik II gene fragment, there is a cleavage site for SalI that will be used in the following construct.

Plk4 plasmid is cut with Hpal and SalI and the gene encoding mini-proinsulin is isolated. This gene is ligated with the above large fragment of pINT41d and the following synthetic DNA sequence.

B'

<td colspan="2"></td><td>Arg</td><td>Met</td><td>Gly</td><td>Arg</td><td>Phe</td>

<td>CG</td><td>CGT</td><td>ATG</td><td>GGC</td><td>CGT</td><td>TTC</td><td>GTT</td>

<td>THE</td><td>TAC</td><td>CCG</td><td>GCA</td><td>THE AG</td><td>CAA</td><td></td>

(Mlul) (Hpal)

This gives rise to plasmid pB70, which encodes a fusion protein where the balasto sequence (Table 1, line 1) is followed by the amino acid sequence Met-Gly-Arg which is followed by the amino acid sequence of mini-proinsulin. .

Example 17

Using the oligonucleotides below, plasmids pINT90d through pINT96d are obtained in analogy to the previous examples. An asterisk indicates the same amino acid encoded in the balasto constituent as in pINT41d.

pINT92 encodes a double mutation in the pINT plasmid encoded insulin derivative pINT72d provided that the Arg codon at the end of the balasto constituent and the mini-C-chain is replaced by the Met codon. Thus the expressed preproduct can be cleaved with cyanogen bromide.

pINT90d

TIR:

Insu50: Insu49:

Insull:

pINT91d TIR: InsuSO: Insu49: Insull:

pINT92d Insu56: Insu58: Insu57: Insull:

pINT93d Insu53: Insull:

pINT94d Insu54: Insull:

pINT95d Insu55: Insull:

pINT96d Insu71: Insull:

orders placed with ****** GNSA * (variant of pINT69d)

5'-CTGAAATGAGCTGTTGAC-3 and

5'TGCCGAATTTCCTGTTGATGTTGTTGC-3 'or 5'-GGAAATTCGGCACGATTTGTGAACCAG-3' and 5'-TCATGTTTGACAGCTTATCAT-3 '****** GNSA *' (variant of pINT72d)

5'-CTGAAATGAGCTGTTGAC-3 and

5'-TGCCGAATTTCCTGTTGATGTTGTTGC-3 'or 5'-GGAAATTCGGCACGATTTGTGAACCAG-3' and 5'-TCATGTTTGACAGCTTATCAT-3 '(double mutant pINT72d)

5'-TCGACCATGGCAACAACATCAACAATGTTTGTG-3 and 5'-GATGCCCATGGTCTT-3 'or' -AAGACCATGGGCATC-.3 'and

5'-TCATGTTTGACAGCTTATCAT-3 '****** (variant of pINT68d)

5'-ACCATGGCAACAACATCAACAAAACGATTTGTG-3 'and 5'-TCATGTTTGACAGCTTATCAT-3' ****** (variant of pINT68d)

5'-ACCATGGCAACAACATCAACACCACGATTTGTG-3 'and 5'-TCATGTTTGACAGCTTATCAT-3' ****** (variant of pINT68d)

5'-TCGACCATGGCAACAACATCAACAATGCGATTTGTG-3 'and 5'-TCATGTTTGGACACGTTATCAT-3' ****** (variant of pINT68d)

5'-ACCATGGCAACAACATCAACAGGACGATTTGTG-3 'and 5'-TCATGTTTGACAGCTTATCAT-3'

Some prior published European patent (EP-A) were mentioned in the previous text. Corresponding South African Patents (ZA-A), Australian Published Patent Applications (AU-A) and US Patent Applications (USSN) are filed:

<td>-THE</td><td></td><td>ZA-A</td><td>AU-A</td><td>USSN</td>

<td> 211</td><td> 299</td><td> 86/5556</td><td></td><td> 889,176</td>

<td> 290</td><td> 005</td><td> 88/3208</td><td> 15631/88</td><td> 190,082</td>

<td> 292</td><td> 763</td><td> 88/3517</td><td> 16379/88</td><td> 194,914</td>

<td> 292</td><td> 803</td><td></td><td></td><td> 196,599</td>

<img file="PT95111B_D0010.tif" />

<img file="PT95111B_D0011.tif" />

<img file="PT95111B_D0012.tif" />

Table 1

123456789 10 11 ρΙΝΤ

Met Thr Thr Wing Be Thr Thr Thr Wing

ATG GCA ACA ACA TCA ACA GCA ACT ACG --- CGT 41d

Thr Ser Thr - *** *** *** ** g A * TT * GA * G ** G *** ___ *** 42

Thr Thr Wing Thr Ser - *** ** φ g ** *** A * TT * TA * TT * A *** ___ *** 43

--- Asn Ser - *** *** *** *** ___ AAC T * A *** ___ *** g, Q *** *** *** *** *** ** * *** *** ___ --- ** At 67d *** *** *** *** *** *** --- ** At 68d *** *** * ** *** *** *** --- ** a, 69d, 72d

Gly Asn Ser Ala *** *** *** *** *** * a * T ** GCA ** A 90d, 91d

Lys -------- *** *** *** *** *** AA * ___ ___ ___ ** A 93d

Pro ------— *** *** *** *** *** ç; ** --------- ** A 94d

Met -------- *** *** *** *** *** ATG --- --- --- ** At 95d

Gly --- - *** * 7 * *** *** *** *** * g * ___ ___ ___ ** A 96d

Table 2: Plasmid pINT41c DNA Sequence

30 50

GTGTCATGGTCGGTGATCGCCAGGGTGCCGACGCGCATCTCGACTTGCACGGTGCACCAA

90 110

TGCTTCTGGCGTCAGGCAGCCATCGGAAGCTGTGGTATGGCTGTGCAGGTCGTAAATCAC

130 150 170

TGCATAATTCGTGTCGCTCAAGGCGCACTCCCGTTCTGGATAATGTTTTTTGCGCCGACA

190 210 230

TCATAACGGTTCTGGCAAATATTCTGAAATGAGCTGTTGACAATTAATCATCGAACTAGT

250 270 290

TAACTAGTACGCAAGTTCACGTAAAAAGGGTATCGACCATGGCAACAACATCAACAGCAA

310 330 350

CTACGCGTTTCGTGAACCAGCACCTGTGCGGCTCCCACCTAGTGGAAGCTCTCTACCTGG

370 390 410

TGTGCGGGGAGCGAGGCTTCTTCTACACACCCAAGACCCGCCGGGAGGCAGAGGACCCTC

430 450 470

AGGTGGGGCAGGTGGAGCTGGGCGGGGGCCCTGGCGCAGGCAGCCTGCAGCCCTTGGCGC

490 510 530

TGGAGGGGTCCCTGCAGAAGCGCGGCATCGTGGAGCAGTGCTGCACCAGCATCTGCTCCC

550 570 590

TCTACCAGCTGGAGAACTACTGCAACTAATAGTCGACCTTTGCTTTCATTGTCGATGATA

610 630 650

AGCTGTCAAACATGAGAATTAGCCCGCCTAATGAGCGGGCTCTTTTTTTTAATTCTTGAAGA

670 690 710

CGAAAGGGCCTCGTGATACGCCTATTTTTATAGGTTAATGTCATGATAATAATGGTTTCT

730 750 770

TAGACGTCAGGTGGCACTTTTCGGGGAAATGTGCGCGGAACCCCTATTTGTTTATTTTC

790 810 830

TAAATACATTCAAATATGTATCCGCTCATGAGACAATAACCCTGATAAATGCTTCAATAA

850 870 890

TATTGAAAAAGGAAGAGTATGAGTATTCAACATTTCCGTGTCGCCCTTATTCCCTTTTTT

910 930 950

GCGGCATTTTGCCTTCCTGTTTTTGCTCACCCAGAAACGCTGGTGAAAGTAAAAGATGCT

970 990 1010

GAAGATCAGTTGGGTGCACGAGTGGGTTACATCGAACTGGATCTCAACAGCGGTAAGATC

1030 1050 1070

CTTGAGAGTTTTCGCCCCGAAGAACGTTTTCCAATGATGAGCACTTTTAAAGTTCTGCTA

1090 1110 1130

TGTGGCGCGGTATTATCCCGTGTTGACGCCGGGCAAGAGCAACTCGGTCGCCGCATACAC

1150 1170 1190

TATTCTCAGAATGACTTGGTTGAGTACTCACCAGTCACAGAAAAGCATCTTACGGATGGC

1210 1230 1250

ATGACAGTAAGAGAATTATGCAGTGCTGCCATAACCATGAGTGATAACACTGCGGCCAAC

1270 1290 1310

TTACTTCTGACAACGATCGGAGGACCGAAGGAGCTAACCGCTTTTTTGCACAACATGGGG

1330 1350 1370

GATCATGTAACTCGCCTTGATCGTTGGGAACCGGAGCTGAATGAAGCCATACCAAACGAC

1390 1410 1430 *

GAGCGTGACACCACGATGCCTGCAGCAATGGCAACAACGTTGCGCAAACTATTAACTGGC

1450 1470 1490

GAACTACTTACTCTAGCTTCCCGGCAACAATTAATAGACTGGATGGAGGCGGATAAAGTT

1510 1530 1550

GCAGGACCACTTCTGCGCTCGGCCCTTCCGGCTGGCTGGTTTATTGCTGATAAATCTGGA

1570 1590 1610

GCCGGTGAGCGTGGGTCTCGCGGTATCATTGCAGCACTGGGGCCAGATGGTAACCCCTCC

<img file="PT95111B_D0013.tif" />

1630 1650 1670

CGTATCGTAGTTATCTACACGACGGGGAGTCAGGCAACTATGGATGAACGAAATAGACAG

1690 1710 1730

ATCGCTGAGATAGGTGCCTCACTGATTAAGCATTGGTAACTGTCAGACCAAGTTTACTCA

1750 1770 1790

TATATACTTTAGATTGATTTAAAACTTCATTTTTAATTTAAAAGGATCTAGGTGAAGATC

1810 1830 1850

CTTTTTGATAATCTCATGACCAAAATCCCTTAACGTGAGTTTTCGTTCCACTGAGCGTCA

1870 1890 1910

GAC CCC GTAGAAAAGATCAAAGGATC TTCTTGAGATC CTTTTTTTCTGCGCGTAATCTGC

1930 1950 1970

TGCTTGCAAACAAAAAAACCACCGCTACCAGCGGTGGTTTGTTTGCCGGATCAAGAGCTA

1990 2010 2030

CCAACTCTTTTTCCGAAGGTAACTGGCTTCAGCAGAGCGCAGATACCAAATACTGTCCTT

2050 2070 2090

CTAGTGTAGCCGTAGTTAGGCCACCACTTCAAGAACTCTGTAGCACCGCCTACATACCTC

2110 2130 2150

GCTCTGCTAATCCTGTTACCAGTGGCTGCTGCCAGTGGCGATAAGTCGTGTCTTACCGGG

2170 2190 2210

TTGGACTCAAGACGATAGTTACCGGTAAGGCGCAGCGGTCGGGCTGAACGGGGGGTTCGT

2230 2250 2270

GCACACAGCCCAGCTTGGAGCGAACGACCTACACCGAACTGAGATACCTACAGCGTGAGC

2290 2310 2330

ATTGAGAAAGCGCCACGCTTCCCGAAGGGAGAAAGGCGGACAGGTATCCGGTAAGCGGCA

2350 2370 2390

GGGTCGGAACAGGAGAGCGCACGAGGGAGCTTCCAGGGGGAAACGCCTGGTATCTTTATA

2410 2430 2450

GTCCTGTCGGGTTTCGCCACCTCTGACTTGAGCGTCGATTTTTGTGATGCTCGTCAGGGG

2470 2490 2510

GGCGGAGCCTATGGAAAAACGCCAGCAACGCGGCCTTTTTACGGTTCCTGGCCTTTTGCT

2530 2550 2570

GGCCTTTTGCTCACATGTGTCAGAGGTTTTCACCGTCATCACCGAAACGCGCGAGGCAGC