GB2214508A - Plasmid vector for the efficient expression of foreign genes fused with newly conceived transcriptional and translational signal sequences. - Google Patents

Abstract

There is described an autoregulatable expression vector with trancriptional and translational signal sequences in a newly conceived cassette system. The expression vector can be used for the efficient expression of heterologous genes which are fused in frame with a short efficiently translatable N-terminal gene extension segment, e.g. the first 20-30 codons of either the E. coli atpE or atpA gene or the bacteriophage eambda cll gene. Microorganisms transformed with the expression vector, together with a method for the production and purification of short eukaryotic proteins using this vector system are also described.

Description

Plasmid vector for the efficient expression of foreign genes fused with newly conceived transcriptional and translational signal sequences Summary The present invention describes an autoregulatable expression vector with transcriptional and translational signal sequences in a newly conceived cassette system.

The expression vector can be used for the efficient expression of heterologous genes which are fused in frame with a short efficiently translatable N-terminal gene extension segment, e.g. the first 20-30 codons of either the E. coli atpE or atnA gene or the bacteriophage Lambda cII gene.

This invention further relates to microorganisms transformed with the described expression vector. A method for the production and purification of short eukaryotic proteins using this vector system is also described.

Introduction Satisfactory levels of expression of foreign genes in Escherichia coli are often not straightforwardly achieved. Poor yields of gene products are commonly attributable to limitations at the level of translational initiation andlor mRNA or protein instability.

The use of efficient translational signals is therefore an essential prerequisite for high expression of heterologous proteins.

Taking this into account it was possible in certain cases to directly express foreign genes in E. coli in large amounts - e.g. the eglin C gene, (Rink et awl., 1984, Nucl. Acids Res, 12, 6369-6387) or the bovine growth hormone gene, (George et al., 1985, DNA 4, 273281).

It was also possible to highly express poorly translatable genes such as those encoding interleukin 2 and interferon-ss using the strong promoters of bacteriophage Lambda combined with an efficient translational initiation region (TIR) (Mc Carthy et al., 1986, Gene 41, 201-206).

The expression of the interferon ss gene was of particular interest because the protein itself is normally highly unstable in E.coli. But the highly efficient transcription and translation rates afforded by the combined system described previously allowed such a high rate of synthesis that a "threshold concentration" could be achieved which allowed the accumulation of interferon ss in large amounts in inclusion bodies.

However, eyen the application of such a powerful system to direct gene expression cannot be expected to be successful in every case - e.g. if the N-terminal part or the codon usage of the cloned gene sequence has a negative influence on translational initiation - a problem often observed with synthetic genes. In other cases the synthesized heterologous mRNA or protein may be degraded by the appropriate host enzymes much faster than the protein may accumulate.

A number of expression systems have been applied in attempts to overcome the problems of translational inefficiency, mRNA instability and degradation of foreign proteins.

Apart from the various systems that may be used for the (direct) expression of individual foreign genes, there are three further types of systems for the production of foreign proteins: - the secretion systems - the gene polymerisation systems - the gene fusion systems.

Secretion systems: Based on the observation that even foreign gene products may be secreted by E,coli if accompanied by a bacterial signal peptide (see e.g. Talmadge et al., 1980, Proc. Natl. Acad. Sci. USA 77, 3988-3992) special purpose secretion vector systems have been constructed, The secretion vectors contain strong transcriptional and translational signal sequences which enable the efficient synthesis of signal peptides derived from secretory proteins such as ompA (Ghrayed et al., 1984, EMBO Journal 3, 2437-2442) or phoA (Gray et al., 1985, Gene 39, 247-254).

In frame fusions of the signal peptide-encoding DNA segment with foreign genes allowes the synthesis and secretion of the foreign protein across the cytoplasmic membrane into the periplasmic space with concommitant processing of the signal peptide by a leader peptidase.

Although the secretion systems could be successfully used in some cases for the production of heterologous (especially secretory) eukaryotic proteins, these systems have some fundamental disadvantages.

The absolute yield of the secreted foreign proteins in the periplasma is often very low and there is often rate limitation andlor blockage at the membrane transport level. In addition the processing of foreign proteins by the leader peptidases may be incorrect and! our incomplete, The secretory systems are therefore not generally applicable to the large scale production of foreign proteins.

Gene nolvmerisation systems Following the gene multiplication strategy it could be demonstrated that in-frame fusions of multiple copies of the proinsulin and aprotinin gene - each flanked by codons for methionine - considerably increased the stability of the synthesized "multidomain protein" (S.H, Shen, 1984, Proc. Natl. Acad, Sci. USA 81, 4627-4631; Wilcken-Bergmann et awl., 1986, EMBO Journal 5, 32193225). After purification, the polydomain product of the multicopy genes was cleaved into single units by means of cyanogen bromide treatment.

The main disadvantages arising from this expression strategy are the time consuming construction work of the multidomain fusion genes and the high degree of genetic instability, because the high copy number of tandemly arranged genes increase the probability of recombination events manifold. At the protein level the main disadvantage of the gene multiplication strategy is the gamine acid heterogenity of the protein domains after cyanogen bromide cleavage. The flanking methionine residues of the single protein domains are partially modified to homoserine lactones. Therefore, the use of a gene polymerization system, carries with it the risk of genetic and protein heterogenity.

Gene fusion systems: Gene fusion systems useful for the producticn rf a wide range of heterologous proteins in satisfactory yields make use of the naturally existent high expression rates of certain viral and E.coli genes. Commonly employed gene fusions have included large parts of e,g, the bacteriophage MS2 replicase gene (Remault et al., 1981, Gene 15, 81-93) the CAT gene (Bennett et awl., 1984, Er. Pat. appl. 0131362) or the lacZ gene (Goeddel et al., 1979, Proc. Natl, Acad. Sci. USA 76, 106-110; Sumi et al., 1985, J. of Biotechnology 2, 59-74). Inframe fusions of foreign genes with these genes may help to improve the efficiency of expression especially for small proteins.

All of the prokaryotic genes listed above are suitable for the expression of practically any foreign protein, but the required protein constitutes normally only a relatively small part of the fusion product. This generally limits the effective yield of the required protein.

It is therefore desirable to reduce the size of the prokaryotic gene segment as far as possible. However, the use of small prokaryotic gene fusion segments in several cases is more likely to yield an unstable protein fusion product (Derynck et al., 1984, Cell 38, 287297; Guo et awl., 1984, Gene 29, 251-254).

It is well known that the stability of even small fusion proteins is related to their rate of synthesis and thus their cytoplasmic concentration. What is therefore required is a generally applicable expression system that supports the synthesis of relatively unstable fusion proteins to levels high enough to cause the formation of inclusion bodies within the cell. The formation of inclusion bodies protects unstable proteins from protease attack. Moreover, the high density of packing of the inclusion bodies facilitates the purification of large amounts of protein. In designing such an expression system the following points must be taken into account, 1. Highly efficient transcription can normally be guaranteed by using strong and fully repressible promoters positioned appropriately upstream of the gene(s) to be expressed.

2. Highly efficient translation on the other hand re quires the presence of an efficient translational initiation region (TIR), whereby a part of the TIR comprises at least the first 20 bases of the struc tural gene sequence, Thus highly efficient trans lational initiation can only be guaranteed in every case by using a fixed N-terminal coding sequence, i.e. through the use of a gene fusion system.

3. The size of the N-terminal extension element in the hybrid protein should be reduced to the minimum compatible with efficient expression, so that a maximum yield of foreign gene product can be ob tained after hybrid cleavage.

4. Finally, any highly expressed fusion protein should be readily purificable, Consideration of the points outlined above has led to the development of the expression system described in this invention, Brief descriPtion of the invention The invention describes a novel plasmid vector constructed for the efficient expression of heterologous proteins in E.coli. The plasmid vector is especially suitable for the expression and accumulation of small foreign proteins which are usually poorly translated or are unstable in the cytoplasm of the host cell.

The vector contains powerful transcriptional and translational signal sequences arranged in a novel DNAcassette system. The cassette system allows a high degree of constructive flexibility, The DNA-cassettes may be exchanged with functional equivalents by means of digestion with restriction enzymes and religation, The individual sequences from different natural sources have been organized in a compact system which facilitates efficient transcription and translation and thus the efficient synthesis of foreign proteins.

The origins and arrangements of the functional elements of the cassette system are briefly described below: DNA sequences bearing transcriptional promoters and operators as well as a repressor gene were obtained from bacteriophage Lambda. In order to promote highly efficient transcription the rightward and leftward major promoter sequences (PR, PL) of bacteriophage Lambda were combined in a tandem arrangement, together with their respective operator sequences. The promoter/operator cassette PER'PAL has the DNA sequence given in fig 1.

The PR and PL promoters are repressed naturally by the cI repressor protein of bacteriophage Lambda. In order to render the expression vector independent of a chromosomally encoded repressor protein the bacteriophage Lambda cI gene was inserted upstream of the PR and PL promoters (fig 1). In this case a thermosensitive derivative (cIt5 857) has been used which allows heat inactivation of the repressor protein and thus heat induction of the expression system. Thus a temperature shift of the host strain from 280 C to 420 C induces efficient transcriptional initiation from the PR and promoters as the result of disassociation of the temperature-sensitive repressor from the operator binding region.The promotor/repressor cassette is separable by digestion with the restriction enzymes Pst I and Xho I.

In summary the inclusion of the CIts857 repressor gene on the expression vector has a number of advantages: - The cytoplasmic titre of the vector encoded re pressor protein is higher than that of the chromo somally encoded protein (gene dosage effect).

- The relatively high concentration of repressor molecules reduces unwanted leakage of the PR and PL promoters under non induced conditions to a minimum. High level expression of toxic or unstable protein molecules may therefore be restricted to predetermined phases of growth. The host strain is therefore not exposed to the burden of foreign gene expression (and thus to negative selection forces) during growth under noninduced conditions.

- The broad host range of the expression vector faci litates the production of foreign proteins in different genetic backgrounds, i.e. in bacterial strains with reduced protease levels, tRNA suppres sor mutants or RNAse strains.

High expression rates of cloned genes are not attainable without efficient translation of the encoded mRNA. Therefore in addition to a strong promoter sequence the use of a highly efficient and where possible2 generally acting translational initiation region (TIR) is essential for the efficient expression of a foreign gene.

Specific aspects of the overall TIR structure (a sequence of some 50-60 nucleotides) are known to influence translational initiation efficiency (see e.g.

Gold et al., 1981, Ann. Rev. Microbiol. 35, 365-403; Stanssens et al., 1985, Gene 37, 211-223;). The positioning of the Shine-Dalgarno sequence and the gene start codon relative to each other, as well as the precise sequence of the whole TIR including at least the first 20 bases of the structural gene are important (McCarthy, 1987, J, Bioen., Biomem. in press), Since there is no generally applicable strategy for the design of N-terminal structural gene sequences (within the limitations imposed by codon usage possibilities) that will, together with a suitable upstream sequence, allow efficient translation, the required modification programme may be very time consuming.In order to circumvent these problems and to realize high expression of any foreign protein in a short time the present invention describes the use of a newly conceived translational initiation unit (TIU).

The first part of the TIU comprises the well proven TIR sequence of the E.coli atpE gene from position -50 to the start codon (McCarthy et awl., 1985, EMBO Journal 4, 519-526). The second part of the TIU comprises the N-terminal segment of an efficiently translatable gene (fig. 2). For example the N-terminal region of the bacteriophage Lambda cII gene (cII31, encoding 31 amino acids) may be used for this purpose. The positive effect of the N-terminal part of the Lambda cII-gene upon translational efficiency has been described elsewhere (Nagai, K. and Thogerson, H.C,, 1984, Nature 309, 810812; Voradarajan et awl., 1985, Proc. Natl, Acad. Sci,, USA 82, 5681-5684).

Other examples useful for this purpose are the Nterminal coding regions of the K. coli atoE gene or the atnA gene (from position +1 -' +90, McCarthy, 1987, J, Bioen. and Biomem., in press). Alternatively, custom designed, fully synthetic N-terminal sequences can be used.

Heterologous genes, especially synthetic andlor small genes may be efficiently translated after in-frame fusion to the TIU's described above, The general applicability of the newly conceived TIU's is demonstrated by means of experiments performed with an expression vector containing the atDE TIR-cII31-TIU fused with natural or synthetic genes. The described expression system is suitable for the bacterial synthesis of small eukaryotic proteins, such as protease inhibitors which are poorly translatable using unfused expression systems.

A demonstration of the flexibility of the described vector system is given by the construction of multimers of the translational cassettes in order to enhance the expression rate of the fusion proteins.

Finally a rapid purification method for the accumulated fusion proteins involving cleavage of the fusion proteins and isolation of renaturated protease inhibitors is also presented.

In connection with the present application the transformed microorganisms DH 1-pIU 12.4.B, RR 1 15 - pIU 21.4.B and RR 1 15-pIU 18.7.B. have been deposited with National Collection of Industrial Bacteria (NCIB), Torry Research Station, P.O.Box 31, 135 Abbey Road, Aberdeen AB9 8DG, Schottland, Grossbritannien under the designation numbers NCIB 12613, NCIB 12614 and NCIB 12615.

Materials and methods I. Materials Reagents Reagents needed for agarose gels, polyacrylamide gels, reaction buffers and culture media are commercially available from Boehringer (Mannheim), Biorad, Difco, Merck, Serva, Sigma and Biorad.

Enzvmes Enzymes used in molecular cloning were purchased from Boehringer (Mannheim), Biolabs and BRL.

DNA Most of the plasmid-DNA, bacteriophage DNA, DNA adapters and DNA-linkers used in vector construc tions are available from Boehringer (Mannheim)4 Biolabs, BRL and Pharmacia. In some cases the DNA fragments were isolated from DNA molecules, gene rously provided by outside laboratories. The origin of these DNA molecules are listed hereinafter.

E.coli strains E.coli RRIBM15 (ATCC 35102): U. Rather, 1982, Nucleic Acids Res., 10, 5765-5772 E.coli DHl (ATCC 33849) D. Hanahan, 1983, J. Mol. Bio,, 166, 557-580.

Media Growth media with and without antibiotics were pre pared as described in Maniatis et al.g Molecular Cloning, Cold Spring Harbor, USA (1982).

Radiolabelled comPounds L-(35S)methionine, Adenosin 5'-a-135S3thiotriphos phate and (14C)protein molecular weight markers were purchased from Amersham Buchler.

II Methods Standard techniques and methods for recombinant DNA work are described in T. Maniatis et al., Molecular cloning, Cold Spring Harbor, USA (1982), Chemical synthesis of olioonucleotide fragments The DNA-oligonucleotides necessary for vector con structions were synthesized on a manual support using phosphotriester chemistry (Frank et al., 1983, Nucl. Acids Res., 11, 4365-5377). The method is briefly described in U.K. patent application No. 8 700 204, The DNA-oligonucleotides used for the construction of protease inhibitor genes were synthesized either as described above or by using the automated DNA synthesizer from Applied Biosystems, Model 380.

Purification of the single stranded DNA fragments was performed with HPLC-columns or by using polyacrylamide gel electrophoresis. Routinely 0.5-5 O.D. 260 nm units of each fragment were isolated.

DNA-seauencina DNA fragments were subcloned either into M13 bacteriophage vectors (Messing, J., 1983, Methods Enzymol., 101, 20-78) or into pUC 819 plasmid derivatives.

Sequencing of single stranded and double stranded DNA was performed according to a sequencing method described by M. Hattori and S. Sakaki (1986) Analyt. Biochem., 152, 232-238.

Construction of the exPression vector pIU 12.4.B.

The construction of the expression vector p1U12.4.B was based on parts of the R.coli plasmid pBR322 (construction scheme fig 3). The construction was done in accordance with the standard cloning techniques described by Maniatis et al., Molecular cloning, Cold Spring Harbor (1982).

First of all the size of pBR322 was reduced by elimination of the 641 bp AvaI-Pvu II fragment (14252066 bp BR322), The protruding single stranded =ndc were then rendered blunt using dNTP's and hlew enzyme.

The resulting vector was religated.

The pBR322 derivative was linearized with Nru l (972 bp pBR322) in order to insert the 330 bp Hind III fragment of plasmid pGBU207 (Beck et al., 1978, Nucleic Acids Revs. 5, 4495-4503; Gentz et al., 1981. Proc. Natal.

Acas. Sci USA, 78, 4936-4940), The 330 bp Hind III fragment contains the major transcription terminator sequence of the E.coli phage fd. The Nru I and Hind III restriction sites of vector and insert were blunt end ligated after a fill-in reaction using Klenow enzyme.

The 651 bp EcoRI - Sal I fragment (4363-651 bp pBR322) of the vector pfd3700 was replaced by a 633 bp EcoRI-Hae III fragment of the plasmid vector pPL-Lambda (Pharmacia) containing the XPL promoter. Prior to the replacement the Hae III-restriction site of the PPL Lambda fragment was converted to a Sal I recognition sequence using the synthetic DNA linker: 5'-GGTCGACC-3', The new plasmid vector pJLA101 was further modified by the insertion of the PR promoter of phage Lambda and the cI-repressor gene (mutant form). This was done by replacing the small Bam HI fragment (counter clockwise of the PL promoter) with the 1.1 kb Cla I restriction fragment of the recombinant plasmid CM2443 (a generous gift from Kaspar von Meyenburg, University of Kopenhagen, Denmark).Before ligation the Bam HI and Cla I restriction sites of the plasmid vector, and the CM2443 fragment, respectively were adapted by a fill-in reaction using Klenow enzyme. Correct insertion of the 1,1 kb blunt-ended Cla I fragment resulted in the vector pJLA201.

In oder to insert a DNA-polylinker with several cloning sites downstream of the two Lambda promoters the restriction sites of EcoRI, Xho I, and Nde I in the vector pJLA201 were removed. This was achieved by, first of all, cutting the vector with the appropriate restriction enzymes1 secondly, exonucleolysis of the single stranded DNA ends with mung bean nuclease, and finally religation of the resulting blunt ends (precursor vectors pJLA202, pJLA203 and pJLA204), The vector pJLA 204 was cleaved with Sal I and the single stranded DNA ends were removed using mung bean nuclease. After that the DNA-polylinker.

was inserted into the blunt ended Sal I site. Correct insertion of the DNA polylinker yielded the plasmid pJLA501. In oder to combine the Lambda PR and PL promoter with the translation initiation region (TIR) of the atDE-gen of E.coli, the vector pJLA5O1 was digested with with Xho I and Nde I. Thereafter the synthetic atpE-TIR sequence

was ligated with the vector DNA.

The recombinant expression vector pJLA 503 was used for the construction of a universally functioning translational initiation unit (TIU) consisting of the atpE TIR signal sequence and the highly translatable 5'-terminus of the cII-gene of bacteriophage Lambda.

The N-terminus encoding cII-gene segment of bacteriophage Lambda was prepared by digestion of the 651 bp Bgl II fragment (38.103-38.754 kb Lambda Sam 7 DNA with the restriction enzymes Nde I and Alu I. The 95 bp cII31 fragment (encoding 31 amino acids of the cII protein) was subcloned between the Nde I and Sma I restriction sites of pUC9. The recombinant vector pIU1.4.B was cleaved with Nde I and Bam HI to recover the modified cII31-gene fragment. Finally, the c1131- fragment was ligated with pJLA503 digested with Nde I and Bam HI.

The pJLA503 derivative p1U12.4.B was checked for the expression of heterologous genes fused in frame with the newly conceived TIU (see examples), ExamPle 1 Expression of the cII qene of bacteriophaae Lambda The efficiency and regulation of the transcriptional and translational signals in the expression vector pJLA503 and pIU12.4.B, respectively, was tested in vivo using the Lambda cII gene (construction scheme fig 4).

For this purpose the cII gene, located on a 651bp BglII fragment of the Lambda genome (38,103-38.754 kb Lambda) was subcloned into the BamHI restriction site of pUCl2. After transformation of E.coli PRISM15 positive clones were selected on LB-agar plates containing 100 pg ampicillinlml. The pUCl2 derivative pIU14.2 B containing the 651 bp BglII fragment with the cII gene was selected by restriction analysis of plasmid DNA from individual clones.

For expression purposes the cII-protein encoding DNA region of the cloned BglII fragment was isolated from p1U14.2.B by digestion with NdeI and Sail. The NdeI site contains the ATG start codon of the cII gene, The expression vector pJLA503 was cleaved with the corresponding restriction enzymes and the cII gene was inserted via ligation into the NdeI-SalI restriction sites of pJLA503. Transformation and selection of the recombinant plasmid vectors was performed as described above, Correct insertion of the cII gene into the NdeI Sal restriction sites of pJLA503 was verified by DNAsequencing of the recombinant vector pIU7.4.B.

The rate of synthesis of the cII-protein under controll of the tandem PR and PL promoters and the atPE-TIR was investigated in whole cells by means of radioactive "pulse chase" labelling, For this purpose the E.coli strain RRIAM15-pIU7.4.B was grown at 280C ih M9 salt medium supplemented with 1% (wiz) methionine assay medium (Difco Laboratories).

At a cell density of O.D550 = 1.0 the bacterial culture was divided and 1 ml cultures were shifted to 420 C whilst still shaking. After an induction period of 20 minutes 10 uCi of 35S-methionine was added. Radioactive incorporation was stopped 1 minute later by the addition of 100 mM unlabelled L-methionine. After 30 seconds the cells were harvested by centrifugation (00 C) and washed twice with 1 ml TE-buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.5), In order to determine the radioactive incorporation the cells were resuspended in 60 ul SDS sample buffer (60 mM Tris, 10% Glycerol, 0.5% SDS, 5% ss-mercapto- ethanol, 20 Ciglml Bromphenol blue) and cracked in a 950C water bath for 3 minutes.The cell proteins were separated on a SDS-PAGE (6-12,5% gradient gel), An equivalent of 500 000 cpm was separated in each lane. The labelled proteins were detected by means of fluorography.

Summarizing the results of the "pulse chase" experiments the heat induced E.coli strain RRIBM15- piU7.4.B synthesizes large amounts of a new protein with the molecular weight of the Lambda cII protein (10.5 KD/fig. 5). No protein band of this intensity is seen at a growth temperature of 280C or by using the K.coli strain RRIBM15-pJLA503 without the cii gene. This experimental result was taken as confirmation of the perfect functioning of the constructed expression vectors pJLA503 and pIU12.4.B.

ExamPle 2 Expression of the Val-15-Glu-52-aprotinin gene As described in patent application Le A 24 273 gene variants of the bovine proteinase inhibitor aprotinin (58 amino acids) have been synthesized via recombinant DNA technology. Until now the expression of unfused aprotinin gene variants in bacterial hosts could not be achieved except for the expression of Met-aprotinin gene multimers using a special designed expression vector (Wilcken Bergmann et al., 1986, EMBO Journal 5, 32193225).

The application of the constructed expression vector p1U12.4.B to the expression of small eukaryotic genes (fused in frame with the cII fragment) was tested using the Val-15-Glu-52-aprotinin gene variant (construction scheme fig 6). The aprotinin gene variant (fig 7) was isolated from the pUC8 derivative pRK54,1,1 with the restriction enzymes EcoRI and SalI, As preparation for in-frame fusion with the cII31 gene segment the N-terminus-encoding a-block of the Val-15 Glu-52-aprotinin gene was provided with a BglII restriction site.This was achieved by synthesizing the new ablock: B A G P L BglII--block A 2 Met 1 5'-GATCTATGCGTCCGGACTTCTGCCTCGAGCCGCCGTACACTGGGCC-3' -----------------------±---------±-- ATACGCAGGCCTGAAGACGGAGCTCGGCGGCATGTGAC and exchanging this with the previous one in the aprotinin gene variant, An in-frame fusion between the cII31-segment and the modified Val-15-Glu-52-aprotinin gene was achieved by inserting the latter between the BamH and Sall sites of pIUl2.4.B. Standard procedures were used for the transformation of competent E.coli RRIBM15 cells and selection of the recombinant expression vector.

The correct in-frame fusion of the Val-15-Glu-52aprotinin gene with the gene segment of the Lambda cII gene was confirmed in plasmid pIU2l.4.B by DNA sequencing (fig 8).

In order to demonstrate synthesis of cII31-Val-15 Glu-52-aprotinin the E.coli host strain RRIa1S- pIU2l.4.B was inoculated in 10 ml LB-medium containing 100 ug ampicillinlml. The strain was grown at 280 C to a cell density of 5 x 108 cellslml. The temperature was then shifted to 420 C during continuous shaking of the bacterial culture. After 2 hours the bacteria were harvested by centrifugation.

The detection of the cII31 Val-15-Glu-52-aprotinin fusion protein was performed with SDS-PAGE according to Laemmli, U.K., 1970, Nature 277, 680). Per gel lane 1 x 109 cells were solubilized in SDS sample buffer (60 mM Tris, HCl pH 8.8, 10X Glycerol, 1% SDS, 5% 6-mer- captoethanol, 20 ssglml Bromphenol blue) and heated for 5 min at 950 C. After electrophoresis the gels were stained with Comassie blue. A typical pattern of the E.coli proteins and the produced fusion protein is shown in fig 9. The accumulation of the expressed fusion protein is accompanied by the formation of inclusion bodies. This could be visualized by light microscopy.

Example 3 Expression of the PSTI-4A gene A further test of the atPE-TIR cII31-expression vector pIU12.4.B was performed using other small eukaryotic gene sequences which can normally only be poorly expressed in the unfused state. These were homologous to the human gene encoding pancreatic secretory trypsin inhibitor (h-PSTI, 57 amino acids).

Both a gene encoding the natural sequence (L,J, Greene, 1976, Methods Enzym. 45, 813-825) and a series of derivatives, encoding variants of this basic amino acid sequence, were synthesized. The PSTI gene variants were initially inserted between the HincII and HindIII sites of pUC8, as described in patent application Le A 24 828, Before insertion into the expression vector pIU12,4,B the synthetic gene had to be modified at both ends. First of all, an ATG start codon had to be added at the 5' (HincII) end. Secondly, two translational stop codons were added at the 3' (EcoRI) end (construction scheme fig 10).

The PSTI-4A was selected for the first expression experiments. Subsequently other variants were also used.

In order to allow subsequent CNBr-cleavage at the Nterminus of the PSTI-4A protein the structural gene had to be provided with an initial ATG codon.

The PSTI-4A gene was isolated by cleavage of the cloning vector pUC8 with HincII-EcoRI. The HincII-EcoRI fragment was inserted between the NcoI and EcoRI site of pUC12-NcoI (The HindIII site of pUC12 was replaced before by a NcoI site.) The NcoI site of the pUC12 vector had previously been rendered blunt-ended by means of a fill-in reaction using dNTP's and the Klenow fragment of DNA polymerase I. The insertion of the PSTI-4A gene resulted in the regeneration of the NcoI site at the 5'end.

To prevent read through translation, the 3'terminus of the PSTI-4A gene was supplied with two additional translation stop codons by insertion of the DNA oligomer 5'-A TAA TGA ATT CAT TAT-3' into the EcoRI restriction site of the gene. Correct modification of the PSTI-4A gene was yielded with the construction vector pIU3.7.B.

In-frame fusion of the PSTI-4A gene with the cII31 gene segment of pIU12.4.B was achieved as follows.

pIU12.4.B was cleaved with the restriction enzyme BamHI.

The BamHI restriction site was rendered blunt-ended by means of a fill-in reaction using the Klenow large fragment of DNA PolI and the vector was then cleaved with EcoRI.

In order to isolate the modified PSTI-4A gene the plasmid pIU3.7.B was cut with NcoI. The NcoI restriction site was rendered blunt-ended by means of a fill-in reaction using Klenow-enzyme, The synthetic gene was subsequently excised by a secondary cleavage with KcoRI.

After ligation of the PSTI-4A gene with the expression vector pIU12.4.B the K. coli strain RRl8M15 was transformed with the recombinant DNA. Positive clones were selected on LB-agar plates with 100 ug ampicillinlml and were checked by means of restriction analysis of plasmid DNA derived from single colonies. DNA sequencing was used to check the fusion construct inserted in plasmid piU18.7.B (fig 11), Expression and detection of the cII31-PSTI-4A fusion protein in the transformed host strain was investigated using the standard procedures described previously for the cII31-Val-15-Glu-52-aprotinin construct.

The quantity of the cII31-PSTI-4A fusion protein produced as a percentage of total cell proteins was approximately 6%, i.e. similar to that observed in experiments with the cII31-Val-15-Glu-52-aprotinin construct.

ExamPle 4 Expression of multiple copies of cII31-Val-15-Glu- 52-aprotinin and cII31-PSTI-4A arranged in tandem translation units.

A simple approach to increasing the maximal intracellular concentration of the desired protein is to co- struct multiple translation units in tandem behind the PR/PL promoters. Each unit comprises that atE TIRcII31-TIU combined with the gene to be expressed. Tbe construction of these multiple translation units is readily accomplished by isolating the units as XhoI-Sali fragments and ligating them as multimers into the expression vector (construction scheme fig 12).

Starting with the expression vectors piU2l.4.B (cII31-Val-15-Glu-52-aprotinin) and pIU18. 7.B (cII31 PSTI-4A), the plasmid DNA was digested with the restriction enzymes SalI and XhoI. In the case of pIU21.4.B a partial digestion was performed with XhoI because this plasmid contains a second XhoI site in the a-block of the Val-15-Glu-52-aprotinin gene.

After digestion of the plasmid DNA the translation units containing the atpE TIR together with the appropriate fusion genes were isolated from agarose gels and inserted into the SalI restriction site of pIU2l.4.B and pIU18.7.B, respectively. Transformation and selection of positive clones with the duplicated translation units inserted in the correct orientation was performed as described above.

The constructed expression vector derivatives pIU2O.8.B (dimeric atE TIR-cII31-Val-15-Glu-52aprotinin) and pIU6.10.B (dimeric atDE-TIR-cII31-PSTI- 4A) was tested in the host strain RRIEM15.

After heat induction of the E.coli strains harbouring the tandemly arranged translation units, cell proteins were separated using SDS-PAGE and stained with Commassie blue. The dimeric constructs supported the expression of approximately 50% more fusion protein than was observed after induction of the primary monomeric constructs described above (fig. 13).

Preparation of arotinin and PSTI-derivatives from cII31-fusion proteins Many foreign proteins synthesized in large amounts in E4coli bacteria accumulate in an insoluble form - so called inclusion bodies (D.C. Williams et al., 1982, Science 215, 687). These protein aggregates can only be solubilized using strong denaturants and therefore separation from other cell proteins is facilitated. Taking this into account, the following procedure was applied in the purification of the cII31-aprotinin and cII31- PSTI fusion proteins from inclusion bodies (all purification steps were checked by SDS-PAGE, fig. 14): 1. After microbial synthesis of the protease inhibitor the bacteria were centrifuged, resuspended in breaking buffer and lysed by sonication.

2. The cell-lysate was centrifuged in order to recover the inclusion bodies. The inclusion bodies were washed with 2 M urea.

3. The inclusion bodies could be solubilized using 6 M guanidinium hydrochloride containing 10 mM 5- mercaptoethanol. After separation of an insoluble fraction the fusion protein was chromatographed on a Sephacryl S-300 column in 6 M urea containing 10 mM 6-mercaptoethanol (fig. 15), The fractions were appropriately pooled after examination of SDS gels run with the samples (fig. 16).

4. The fusion protein was precipitated by dialysing against dest. H2 containing 10 mM 5-mercapto- ethanol, The wet fusion protein was dissolved in 70% formic acid and cleaved by cyanogen bromide according to Gross (K. Gross, B, Wittkop, 1961, Amer. Chem. Soc.

83, 1510-1511).

5. The protease inhibitor was separated from the un cleaved fusion protein and the other cyanogen bro mide fragments by gel chromatography on a Sephadex G-50 column in 1% acetic acid (fig. 17). The first 8 AAs of the fusion protein were checked by protein sequencing. A mixture of fusion protein was ob tained, i.e. with and without an N-terminal methio nine at the cIl-terminus.

- Met-Val-Arg-Ala-Asn-Lys-Arg-Asn- (about 75%) - Val-Arg-Ala-Asn-Lys-Arg-Asn-Glu- (about 25%).

6, The protease inhibitors were characterized by N terminal sequencing over 19 steps and by amino acid analysis (table 1). The PSTI-variant was renatura ted using a modified form of the procedure des cribed by Tschesche (H. Tschesche, G. Haenisch, 1970, FEBS Letters 11, 209-212), The aprotinin variant was renaturated according to Creighton (T.E. Creighton, UCLA Symposia on Molecular and Cellular Biology New Series, Vol, 39, ed. Dale L.

Oxender, 249-257, Alan R. Liss, Inc., New York, 1986).

The degree of renaturation depends on the pH-value, the concentration of inhibitor and the type of amino acid substitution of the protease inhibitors. The inhibitor activity was measured using a suitable enzyme inhibition assay.

Example 5 Preparation of Leu-18-Glu-19-Arg-21-Ala-32-PSTI (PSTI4A) 1. Isolation and cyanopen bromide cleavage of the cII31-PST1-4A-fusion protein A 2 1 culture of the E.coli strain RRIBM15-pIU 6.10.B was grown at 280 C to a cell density of 5x108 cells/ml. After a rapid temperature shift to 420C the bacteria were continuously shaking for two hours.Thereafter the bacteria were harvested by centrifugation for 20 min at 15,000 g, In order to isolate the cII31-PSTI-4A fusion protein (accumulated in inclusion bodies)1 4,7 g cell pellet was resuspended in 40 ml breaking buffer (0,1 M TRIS-HCl pH 7,5 containing 10 mM EDTA, 5 mM benzamidine-HCl and 10 mH ss- mercaptoethanol) and treated with lysozyme (0.3 mgEml) for 30 min at 370 C.

After cooling to 80 C the cells were lysed in a french-pressure cell at 18 000 psi. Insoluble ma terial was recovered by centrifugation (as above) and was washed twice in buffer (see above) con taining 2 M urea.

The final pellet (wet weight 0.31 g) was dis solved in 10 ml 50 mM TRIS-HCl pH 8.5 containing 6 M guanidinium hydrochlorid-HCl and 10 mM B-mer- captoethanol. The solution was clarified by centri fugation. The supernatant was chromatographed on a column (5x90 cm) of Sephacryl S-300 equilibrated with 50 mM TRIS-HCl pH 8.5 containing 6 M urea and 10 mM ss-mercaptoethanol (fig, 15), Fractions containing the cII31-PSTI-4A fusion protein were identified using SDS-PAGE according to Laemmli (U,K. Laemmli, 1970, Nature 277, 680685) (fig.16)). Fractions 126-137 were pooled and dialyzed against deionized water until precipitation occurred. The precipitate was centrifuged down (described above) and frozen at -70 C.

50 mg of the solubilized fusion protein was dissolved in 10 ml 70% formic acid. The fusion protein was cleaved by adding 0.5 g cyanogen bromide and incubated for 18 h under nitrogen in the dark.

The reaction was stopped by diluting with 200 ml water. The water and the volatile by-products were removed by freeze drying, The cyanogen bromide peptides were dissolved in 1% acetic acid and were chromatographed on a column (2.5 x 85 cm) of Sephadex G-50 (superfine) equilibrated with 1% acetic acid (fig. 17). Fractions containing the PSTI-4A were identified by SDS-PAGE. The fractions No. 40-45 were pooled and lyophilized.

The structure of PSTI-4A was checked by Nterminal sequencing and amino acid analysis (table 1).

In a series of experiments the final yield of PSTI-4A varied from 2 to 4 mgll bacterial culture.

2. Renaturation of PSTI-4A 200 zg PSTI-4A were dissolved in 2 ml 6 M urea containing 20 mM Hepes buffer1 pH 6.5, The solution was dialyzed against 2 x 300 ml 2 M urea containing 1 mM ss-mercaptoethanol and 50 mM Hepes buffer pH 6.5 (cut off 1000 dalton).Dialysis was performed for 20 h (with one change of buffer after 6h). The degree of renaturation was about 35%. The enzymes inhibitory activity was measured using an elastase inhibitory assay and compared to that of native PSTI-4A, obtained from an K.coli secretion system, described in U.K. patent application 8.700,204, ExamPle 6 Preparation of renatured Val-15-Glu-52-aprotinin The cII31-Val-15-Glu-52-aprotinin fusion protein was prepared essentially as described in example 5.

100 mg of the fusion protein was dissolved in 10 ml 70% formic acid, 0.5 g cyanogen bromide was added, and the solution was incubated under nitrogen for 18 hours at room temperature. The reaction was stopped by the addition of 200 ml deionized water. Excess formic acid and cyanogen bromide were removed under reduced pressure.

The concentrated solution (5 ml) was titrated to pH 6,5 by addition of 5 M NaOH and dialyzed exhaustively against 20 mM Hepes pH 6.5 containing 6 M urea and 10 mM ss-mercaptoethanol. For renaturation the dialyzed sample was applied to a column filled with about 5 ml of CM-Sepharose fast flows equilibrated in 20 mM Hepes pH 6,5 containing 6 M urea and 10 mM ss-mercaptoethanol.

The column was washed with this buffer until a stable baseline formed and eluted using a linear gradient formed between 50 ml of the above buffer and 50 ml of 20 mM Hepes containing 2 mM 6-mercaptoethanol followed by a short wash with 20 mM Hepes pH 6.5. Finally the renatured Val-15-Glu-52-aprotinin was eluted from the column with 20 mM Hepes pH 6.5 together with 0.5 M NaCl (fig. 18).

Peak fractions were collected, dialyzed against 20 mM Hepes pH 6,5 and applied to a Mono S column (1 ml) equilibrated in the same buffer using the Pharmacia-FPLC system. The column was eluted using a linear gradient of zero to 0,3 M NaCl, Peak fractions containing Val-15 Glu-52-aprotinin were identified by the elastase inhibition assay, dialyzed against 0.1 M ammonium bicarbonate and lyophilized.

The product was characterized by SDS-PAGE, N-terminal sequencing and amino acid analysis (table 1), Table 1 Amino acid analysis and N-terminal sequencing of Leu-18-Glu-19-Arg-21-Ala-32- (PSTI-4A) cII31 fusion protein, from PSTI-4A and Val-15-Glu-52-aprotinin Amino CII-PSTI 4A PSTI 4A Val-15-Glu-52-aprotinin acid Asp 10.3 (10) 6.9 (7) 5.04 (5) Thr 5.9 ( 6) 3.9 (4) 2.64 (3) Ser 4,1 ( 4) 3.0 (3) 0.82 (1) Glu 11.7 (12) 7.8 (7) 4,06 (4) Gly 7.7 ( 7) 5.2 (5) 5.79 (6) Ala 7.4 ( 7) 2.4 (2) 6.00 (6) Val 3.0 ( 3) 2.0 (2) 1.90 (2) Met 2.4 ( 3) Ile 3.2 ( 4) 2.1 (2) 1.31 (2) Leu 8.8 ( 9) 4.7 (5) 2.03 (2) Tyr 3.3 ( 3) 2.8 (3) 3.98 (4) Phe 1.5 ( 2) 1.0 (1) 4.13 (4) Lys 6.1 ( 6) 2.8 (3) 2.99 (3) Arg 7.0 ( 7) 4.1 (4) 6.10 (6) The amino acids were determined after post column derivatization with o-phthalaldehyde.

Cys and Pro were not determined, 1. N-terminal sequence of the cII31-PSTI-4A-fusion protein; about 1 mmol of the fusion protein was loaded on the sequencer and sequenced for 8 cycles; two sequences were determined: Met-Val-Arg-Ala-Asn-Lys-Arg-Asn Val-Arg-Ala-Asn-Lys-Arg-Asn-Glu

Claims (17)

1. CLAIMS: 1. A plasmid vector comprising transcriptional and translational signal sequences arranged in a DNA cassette system.
2. 2. The plasmid vector according to claim 1 comprising in cassette the form DNA sequences bearing transcrip tional promoters, operators and a repressor gene obtainable from bacteriophage Lambda,
3. 3. The plasmid vector according to claim 2 charac terized in that the DNA-cassette of the rightward and leftward major promoter sequences (PR, PL) of bacteriophage Lambda are combined in a tandem arrangement, together with their respective opera tor sequences.
4. 4. The plasmid vector according to claims 2 and 3 characterized in that the bacteriophage Lambda cl gene is inserted upstream of the PR and PL promo ters.
5. 5. The plasmid vector according to claim 4 charac terized in that the thermosensitive derivative cIt5 857 of bacteriphage Lambda is used for in sertion upstream of the PR and PL promoters.
6. 6. The plasmid vector according to claims 1 to 5 con taining a translational initiation unit (TIU) com prising the translational initiation region (TIR) sequence of the E,coli atnE gene from position -50 to the start codon as a first part and the N terminal segment of an efficiently translatable gene as a second part.
7. 7. The plasmid vector according to claim 6 wherein the translatable gene of the TIU is the N-terminal region of the bacteriophage Lambda cII gene (cII31, encoding 31 amino acids.
8. 8. The plasmid vector according to claim 6 wherein the translatable gene segment of the TIU is the N terminal coding region of the E.coli atPE gene or the K.coli atpA gene.
9. 9. The plasmid Yector according to claim 6 wherein the translatable gene of the TIU is a custom designed fully synthesized N-terminal DNA sequence.
10. 10. The plasmid vector according to claims 1 to 6 con taining the DNA-cassette of the atpE TIR-cII31-TIU fused with natural or synthetic genes.
11. 11. The plasmid vectors according to claims 1 to 9 characterized in that they contain DNA-polylinkers downstream of the atnE TIR-cII31-TIU cassette,
12. 12. Plasmid vector pIU 12.4.B.
13. 13, Plasmid vectors pIU 7.4.B, pIU 21.4.B, pIU 18.7.B, pIU 20.8.B and pIU 6.10.B.
14. 14. E.coli strains transformed with plasmid vectors of claims 12, 13 and 14.
15. 15. E.coli strains transformed with the plasmid vectors of claims 12, 13 and 14 containing fused in frame eukaryotic genes.
16. 16. E.coli strains according to claim 17 wherein the eukaryotic gene is encoding aprotinin homologues or PSTI homogolues.
17. 17. Purification of short eukaryotic proteins of claims 12, 13 and 14 using microorganisms transformed with the expression vectors.