GB2167760A - Expression vectors and method for their production - Google Patents

Abstract

An expression vector comprising a hybrid operon which combines parts of an E.coli rRNA operon and a part of a protein coding operon, which may be constructed by removing most of the structural part of an E.coli rRNA operon carried by a suitable plasmid to obtain a shortened rRNA operon, combining said shortened rRNA operon with a part of a protein-coding operon to produce an effective hybrid expression control region and finally inserting a DNA fragment coding for a desired polypeptide into the combined operon to ensure gene expression under the control of said hybrid expression control region.

Description

SPECIFICATION Expression Vectors and Method for their Production The present invention relates to expression vectors. More particularly it relates to expression vectors providing for foreign gene expression under the control of a hybrid operon which combine parts of an E.

coli rRNA operon and a part of a protein coding operon. Moreover the invention relates to the production of said expression vectors. The method for constructing expression vectors according to the invention comprises: removing most of the structural part of an E. coli rRNA operon carried by a suitable plasmid to obtain a shortened rRNA operon, combining said shortened rRNA operon with a part of a protein-coding operon to produce an effective hybrid expression control region and finally inserting a DNA fragment coding for a desired polypeptide into the combined operon to ensure gene expression under the control of said hybrid expression control region.

The recombinant DNA techniques made it possible to introduce DNA from any source into bacterial cells. the stabil maintenance of a foreign gene in bacterial cells and the expression of the encoded genetic information may be ensured by specific carrier molecules, so-called expression vectors. In bacterial cells a gene coding for a protein is expressed by a complex series of reactions involving transcription of the DNA into RNA and the subsequent translation of the RNA into protein. Both of these complex processes are carried out and controlled by specific bacterial proteins which recognize expression signal or control sequences situated ahead of the coding sequence. As the structures of these signal sequences in genes of different organisms are diverse, eukaryotic and foreign prokaryotic expression signal sequences are poorly recognized by bacterial enzymes.So, in order to ensure efficient expression of a foreign gene in a new host the DNA piece coding for the desired protein should be joined to an endogenous expression signal sequence which directs efficient transcription and translation. Moreover, because over-production of a foreign gene product may be harmful to host cells and may lead to decreased stability of the recombinant molecule, it is desirable to use controllable expression signal sequences which allow the modulation gene of expression during bacterial growth.Thus a good expression vector which is suitable to direct efficient protein production in E. coli cells, should be characterized by: 1. having a strong E. coli promoter which provides high level of transcription; 2. bearing a suitable ribosome binding site in proper distance from that promoter to ensure the correct and efficient initiation of translation and; 3. allowing the regulation of gene expression at the level of transcription.

In the case of pER expression vectors which are described in this invention the above mentioned advantageous properties, are provided by a hybrid expression control region in which parts of an E. coli rRNA operon and parts of a protein-coding operon are combined.

All living bacterial cells synthetise a large number of ribosomes. The protein and RNA components of these are encoded by DNA. The final products of the transcription of rRNA operons are three rRNA species.

These rRNAs are not only more stable than the mRNAs but they are synthetised also in much higher amount in a growing cell than any one of the RNAs coding for protein. The rRNA operons contain the most actively transcribed genes of bacteria. Although they make up only about 0.8% of the genome, more than 50% of all transcription takes place on them in exponentially growing cells. The high rate of rRNA synthesis is provided by specific DNA sequences i.e. promoters preceding the rRNA genes which direct RNA polymerase binding and initiation of transcription. The nucleotide sequence of these promoters are known and they are thought to be constitutive, i.e., they are not under the control of repressors, so that in growing cell they continually promote the expression of genes linked to them.

The structures and functions of numerous protein-coding bacterial operons are also well known and some of these frequently are used for foreign gene expression in E. coli (e.g. lac, gal. trp). The control of expression of these operons operates on a similar principle: the presence or absence of a small molecule determines whether transcription is blocked or initiated at the promoter and proceeds through the regulatory region into the structural genes of the operon. Accordingly, transcription of these operons could be regulated by adding to or removing from the media, the appropriate small molecules.

The present invention is based on our recognition that the promoters of rRNA operons are able to direct protein synthesis when they are combined with a part of a protein-coding operon which ensures the initiation of translation to produce a hybrid expression control region and that the promoter of an rRNA operon as a part of said hybrid expression control region preserves its advantageous properties concerning the ability to initiate transcription with a high rate and finally that said hybrid expression control region is controllable as it is in the said protein coding operon from which it has been constructed.

To construct expression vectors having expression control sequences which fulfil the above mentioned criteria it is necessary to clone the strong rRNA promoters in stable form on a multicopy plasmid. It is known that recombinant clones containing strong promoters are unstable, probably because the efficient transcription initiated by these promoters interferes with plasmid replication or drains the transcription machinery of the cell. In order to avoid this possibility we created long deletions in the rRNA operon, removing most of the structural part coding for rRNA and fusing directly the promoter and terminator regions. Plasmids carrying this shortened rRNA operon provded to be stable and allowed the further modification of the promoter region.In accordance with the above considerations the main features of gene expression ensured by the expression vectors constructed according to the present invention may be summarized as follows: - The transcription of the inserted gene is provided by a hybrid expression control region in which parts of an E. coli rRNA operon and a part of a protein-coding operon are combined.

- The termination of transcription takes place at the terminator region of the rRNA operon, consequently the transcription does not interfere with plasmid replication.

- The transcription is controllable and the expression of the inserted gene can be induced or repressed by the same method which characterizes the protein coding gene involved in the operon.

- The result of transcription is a hybrid RNA which comprises three regions of different origin: the first corresponds to the protein coding gene which was used to construct the hybrid regulatory region, the second codes for the desired, protein and the third corresponds to a DNA piece coding for rRNA.

The presence of nucleotides corresponding to the mature rRNA increases the stability of the bybrid RNA.

- The translation results in a fused polypeptide having a C-terminal fragment encoded by the inserted foreign gene and an N-terminal fragment encoded by the piece of the protein coding operon which was used to construct the hybrid control region. This latter may increase the stability of the fused protein and if it is an assayable protein it could be used to follow the expression of the inserted gene.

In accordance with the present invention it is advantageous to produce expression vectors fulfilling the above criteria and also having: - the ribosome binding site at the optimal distance from the transcription start site to provide for high level of translation initiation, - a low molecular weight, - a gene for ampicillin resistance, - unique cleavage sites of restriction endonucleases to allow insertion of a foreign gene in all the three possible reading frames.

For purposes of examplification we will describe the construction of a plasmid series and their use according to the present invention. The examples (described in greater detail below) involve: 1st example: - insertion of rRNA operon containing DNA fragment into multicopy plasmid vector, - in vitro deletion of most of the structural part of that cloned rRNA operon to produce a shortened rRNA operon with intact promoter and terminator region, - in vitro deletion of the unnecessary part of the vector plasmid to obtain as small recombinant plasmid forfurtherwork.

- insertion of the first part of the E. coli lac operon into the shortened rrnB operon, - construction of a maximally efficient expression control region by altering the distance between the transcriptional and translational start sites, - construction of unique restriction sites downstream from the expression control region to obtain plasmids which allow insertion of foreign DNA in all the three possible reading frames.

2nd example: - insertion of a prokaryotic gene into expression vectors produced as described above, 3rd example; - insertion of a eukaryotic gene into expression vectors produced as described above, The plasmids obtained by this process and referred to herein as pER plasmids, may be characterized by having a hybrid expression control region to direct foreign gene expression which comprises the promoter and terminator region of the E. coli rrnB operon and the first part of the E. coli lac operon, also having a region which provides ColEI type replication, a gene coding for ampicillin resistance, unique restriction sites of EcorRI and Clal enzymes allowing insertion of foreign DNA in all the three reading frames, and Hindlil, BamHI, Bglli, Xbal and Pvull sites in one reading frame.

The inocution will be illustrated in greater detail by the following non-limiting Examples.

EXAMPLE 1 For the isolation of a E. coli rRNA operon the DNA source may be total E. coli DNA or any one of those transducing bacteriophage DNAs which carry one of these operons. The insertion of rRNA operon containing DNA fragments into appropriate vector DNA could be done by following the methods described for the isolation of the rrnB operon from Arif bacteriophage DNA (A. Kiss et al., Cloning of an E. coli ribosomal RNA gene and its promoter region from Arid18. Gene 4, 137-152, 1978) orfrom total E. coli DNA (I. Boros et al., Physical map of the seven ribosomal RNA genes of Escherichia coli, Nucl. Acids Res. 6, 1817-1830, 1979). The first step to construct the pER plasmids described in this invention was the construction of plasmid pBB9 (Fig. 1). To obtain this plasmid a 7.5 kb fragment of the BamH digested Arif DNA containing the rrnB operon was inserted into the BamHI site of the well known vector pBR322 (J. G.

Sutcliffe, Complete nucleotide sequence of the Escherichia coli plasmid pBR322. Cold Spring Harbor Symp.

Quant. Biol. 43,77-90, 1978). The inserted fragment carries the complete rrnB operon of E. coli. It contains a portion of approximately 400 bp originating from the bacteriophage DNA, a sequence of 1721 nucleotides corresponding to the region located in front of the 16S rRNA gene the three genes coding for the 16S,23S and 5S rRNA, and the terminator region of the rrnB operon which is followed by a sequence of aDproximately 500 nucleotides from an unidentified operon located beyond the rrnB operon. The region in front of the 16S rRNA gene involves the two promoters of the rRNA operon (designated as P1 and P2) which are located approximately 180 and 300 bp upstream from the 16S rRNA gene. Recombinant plasmids containing a complete rRNA operon are generally unstable.This behavior may be explained by assuming that the very intensive transcription along the long operon directed by strong promoters, drains the transcription machinery of the host cell. As a consequence such plasmids can be stabilised by shortening the transcribed region. Therefore we introduced long deletions into the cloned rRNA operon in order to obtain stable plasmids with intact promoter and terminator regions. We aimed to remove all those parts of the rrnB operon which were thought to be not essential for the construction of the desired expression vectors. Plasmid pBB9 was cleaved with Hpal at the unique site in the middle of the rrnB operon and the resulting linear molecules were treated with BAL31 nuclease. This enzyme degrades the polynucieotide chains from both ends and produces shortened molecules.The BAL31 digested DNA was then treated with DNA polymerase-Klenowfragment in the presence of the four dNTPs in order to produce blunt-ended DNA fragments and it was ligated under conditions allowing circularisation. The ligated DNA was transformed into E. coli HB101 cells and ampicillin resistant colonies were selected. Plasmid DNAs were prepared from a number of colonies that grew on the transformation plates and these were used to determine the extension of the deletions. To do this, purified plasmid DNAs were digested with appropriate restriction endonucleases and the fragment patterns were analysed by agarose gel electrophoresis.

On the basis of the restriction map we selected long deletions beginning in the first region of the 16S RNA gene and ending slightly ahead of the 5S RNA gene. One such deleted plasmid was chosen for further work and it was denoted pBB9A9. Its structure is shown on Fig. 2. Nucleotide sequence analysis of pBB9A9 DNA indicated that the above described BAL31 treatment removed 4696 bp from the rrnB operon and joined the promoter and terminator regions. In addition to the intact promoter and terminator regions the plasmid contains the gene coding for the 5S rRNA and the first 267 nucleotides corresponding to the mature 16S rRNA. The nucleotide sequence around the fusion junction in pBB9A9 is shown on Fig. 3.

In constructing the expression vectors of the invention the next two steps were the removal of further unnecessary parts of the pBB9A9 plasmid DNA in order to obtain a small plasmid which bears only a few restriction sites. First, unnecessary sequences beyond the terminator region were removed. This was done in a similar manner as it was described in connection with the construction of pBB9A9. Purified plasmid DNA was linearised at the unique Sall site of the vector, it was then treated with BAL31 nuclease to remove approximately 800 nucleotides from each end, and it was digested with Pvull restriction endonuclease and then treated with DNA polymerase-Klenow fragment in the presence of the four dNTPs to fill in the BAL31 generated ends. The DNA fragment was recircularised by ligating the Pvull generated blunt-ends to the Klenow repaired blunt-ends.Following transformation into HB101 cells, plasmid DNA was prepared from amplicillin resistant clones, digested with several restriction endonucleases and the resulting fragment pattern was analysed by agarose gel electrophoresis. One suitable deleted plasm id, was denoted pBB9-b28 and chosen for further work. In this plasmid, the above step removed the last 0.5 kb of the insertion (beyond the terminators) and 1.4 kb from the vector. Schematic structure of pBB9-b28 and the nucleotide sequence around the fusion junction are shown on Fig. 4 and Fig. 3, respectively.

The aim of the next step was to remove unnecessary sequences of the insertion in front of the promoters. Plasmid pBB9-b28 was cleaved with EcoRI and Fspll and the two resulting fragments were separated on agarose gel. Both of these enzymes cut the plasmid at unique sites. The Ecorl site is in the vector region while the Fspll site is in the rrnB region approximately 400 bp upstream from the 16S rRNA gene. The longer fragment containing a pBR backbone (from the Pvull site to the EcoRI site) and the shortened rrnB operon with the two promoters was reisolated from the gel, treated with Klenow fragment of DNA polymerase to convert the EcoRI and Fspll generated sticky end into flush ends and ligated with polynucleotide ligase under conditions allowing blunt-end ligation and recircularisation.This procedure regenerated the EcoRI and the Fspll sites with the removal of an approximately 1.7 kb DNA segment, so the resulting plasmid, designated p408-5 (Fig. 5) could be cut at the vector-insert junction. Plasmid p408-5 was further modified by cleaving at the EcoRI site, digesting for various times with BAL31 nuclease and religating. Following transformation into E. coli HB101 cells single colonies were picked, grown in culture and the plasmid DNA was reisolated from each. These were members of the p419-series. The DNAs of p41 9 plasmids were analysed with respect to the number and location of sites for the restriction enzymes Bspl, Mspl and Hhal and some of the plasmids were further characterized by determining their nucleotide sequences inthe promoter region. In the two plasmids p419-13 and p419-10 which we chose for further work, the length of the deletions were 164 and 197 bp, respectively. In p419-13 both of the two promoters of the rrnB operon were intact, while in p419-10 the P, promoter has been removed by the deletion. During the next steps of expression vector cqnstruction we used these two plasmids as starting material. As their structures are basically identical, and they bear the same restriction sites, in order to simplify the description the following steps will be described only for plasmids derived from p419-10. One may follow the same process using as starting material anyone of the plasmids from the p41 9-series. The nucleotide sequence around the pBR322-rrnB fusion junction in p419-10 is shown on Fig. 3.

In order to produce an expression vector according to the present invention the shortened rRNA operon must be combined with a part of a protein coding-operon which provides for the translational signal sequences. For this purpose we chose a DNA piece from the well known E. coli lac operon. A suitable DNA fragment can be obtained by digesting plasmid pLBU3 both with EcoRI and Hindlll. pLBU3 was constructed by R. Gentz et al (Gentz, R., Langer, A., Chang, A.C.Y., Cohen, S.N. and Bujard, H., Cloning and analysis of strong promoters is made possible by the downstream placement of a RNA termination signal. Proc. Natl.

Acad. Sci. USA 78, 4936-40, 1981). Among the resulting three fragments the desired one is about 410 bp in length and it comprises, from the EcoRI end to the Hindlll end, a stretch of about 160 nucleotides of unknown origin, a stretch of nucleotides corresponding to the Pribnowbox of the lac promoter, the lac operator, the ribosome binding site, the ATG translational initiation signal, followed by a piece of DNA coding for the first 68 amino acids of betagalactosidase. It is known that the N-terminal polypeptide of beta-galactosidase coded by this fragment does not have enzyme activity alone but it can complement the C-terminal fragment of the enzyme coded by the chromosome of an appropriately chosen host cell, therefore its presence can be easily detected on X-gal indicator plates.

In order to obtain a suitable plasmid which allows the insertion of the EcoRI-Hindlll generated pLBU3 fragment into the shortened rrnB operon in correct orientation, the p419-10 plasmid was modified as follows: the plasmid was cut with Stul between the P2 promoter and the unique Hindlll site, EcoRI linkers were ligated again to yield plasmid p808-2 (Fig. 6). This plasmid was cleaved with EcoRI and Hindlll and ligated with the 410 bp EcoRI-Hindlll fragment from pLBU3. Following transformation into E. coli ED8800 cells colonies containing the desired plasmid were screened for beta-galactosidase activity on X-gal indicator plates. Pale blue colonies appeared after 2436 hours on incubation, indicating a low level of lac alfa peptide expression.Plasmid DNA was isolated from single colonies and their structure i.e. the presence of the 410 bp EccRI-Hindlll fragment inserted into p808-2 between the appropriate sites as confirmed by restriction enzyme analysis. One representative of these clones was designated as p827-10 (Fig. 6).

For creating expression vectors having maximally efficient expression control region plasmid p827-10 was further modified by varying the distance between the rrn B promoter and the lac ribosome binding site.

Differing number of the approximately 400 base pairs between the start points of translation and transcription were removed by creating in vitro deletions with controlled BAL31 digestion. This was carried out by cleaving plasmid p827-10 with ECoRI, digesting with BAL31 nuclease, filling in the ends and finally ligating the DNA. Following transformation into ED8800 cells, colonies were screened for betagalactosidase activity on X-gal indicator plates. On the basis of their dark blue color on the indicator plates those colonies were selected forfurther analysis which showed a high level of a-peptide synthesis. From this clones plasmid DNA was isolated and sequenced to determine the exact structure of the control region.

These plasmids are the members of pER series. The plasmids were found to be of different types, depending on the junction between the rrnB and lac region. Some of them contained a hybrid promoter consisting of the -35 region from the rrnB P2 promoter and the -10 region from the lac promoter. We designated this hybrid promoter region "rac" and picked one plasmid bearing the rac promoter for further work, designated as PERIl-8. The structure of the hybrid expression control region in pERlIl-8 is shown in Fig. 7.

The above plasmids comply with almost all requirements of an ideal expression vector but they offer only limited possibilities to insert foreign genes. This may be done using either the unique Pvull site in the region coding for the alfa peptide or the Hindlll site which is also unique at the end of the lac region. In order to obtain more versatile expression vectors which allow the insertion of DNA fragments generated by using the most common restriction endonucleases and in all possible reading frames, we inserted a short polylinker fragment into the Hindlll site of pERllI-8. This fragment was a derivative of the polylinker carried by plasmid pVX. It was produced as follows plasmid pHC314 (Boros I., P6sfai Gy. the Venetianer P.High copy-number derivaties of the plasmid cloning vector pBR322, 1984, Gene 30,257-260) containing the pVX polylinker (Seed, B., Purification of genomic sequences from bacteriophage libraries by recombination and selection in vivo. Nucl. Acids. Res., 11, 2427-2446, 1983) between two EcoRI sites was digested with EcoRI and religated to yield a plasmid with two polylinkers in head-to-head orientation. From this latter plasmid a modified form of the original polylinker bearing all the cutting sites but in an other order can be easily isolated after digestion with Hindlll. This was inserted into the Hindlll site of pERlIl-8 to produce pERII1-8pl.

Further modifications of the above plasmid resulted in four plasmids (pERlll-8rl, -8r2, -8r3, -8r0) which bears unique sites of the EcoRI, Clal, BamHI, Gblll and Xbal endonucleases at the end of the lac region. Three out of the four plasmids bear the Clal and EcoRI sites in different position with respect to the phasing of lac translation without in phase stop codon. Therefore these allow the insertion of foreign DNA into these two sites in all possible reading frames. Two of the plasmids offer the possibility to introduce DNA two different reading frames into Bglu sites, while the phasing of the BamHI, Xbal, Hindlll and Pvull sites are the same in all plasmids. The constructions of these pERlll-8 derivatives are shown in detail on Figs & 1. The positions of the restriction sites and their relations to the phasing of the lac translation is summarised in Table 1. (In order to simplify the description of the position of a restriction cleavage site with respect to the translational reading frame the phases are defined and used in this text as follows: pH 1; when the 6 nucleotides of the recognition site of a given restriction endonuclease form two intact amino acid codons, PH 2; when the first 5 nucleotides of the recognition site and the nucleotide preceding them form two complete amino acid codons, and PH 3; when the last 5 nucleotides from the recognition site and the next nucleotide form two codons. According to this convention the Hindlll site at the end of the lac region is in PH 1.

TABLE I Plasmids PH1 PH 2 PH3 pER111-8p1 Hindlll Bglll Xbal EcoRI BamHI Pvull Clal pERlll-8r1 Hindlll Clal Xbal EcoRI Pvull pERlll-8r3 Hindlll - EcoRI BamHI Pvull Clal pER111-8rO Hindlll - Xbal Clal Pvull pERlll-8r2 Hindlll EcoRI Xbal Clal Pvull EXAMPLE 2 In order to illustrate the utility of the expression vectors produced according to the present invention we will describe the pER plasmid-directed expression of chloramphenicol acetyl transferase in E. coli cells.

The described procedure involves the insertion of a promoterless CAT gene into several pER plasmids to obtain recombinant plasmids which produce native CAT under the control of different hybrid regulatory regions and the further modifications of these plasmids which result in the synthesis of CAT as part of a fused protein.

The gene which mediates resistance towards chloramphenicol was identified at first in a transposon, then it was translocated into several recombinant plasmids from which it could be easily isolated (Alton K.N. and Vapnek D.: Nucleotide sequence analysis of the chloramphenicol resistance transposon Tn9.

Nature, 282,864-869, 1979). We used one of these recombinants, plasmid pBR329 (Covarrubias Land Bolivar F.: Construction and characterization of new cloning vehicles VI. Plasmid pBR329, a new derivative of pBR238 lacking the 482-base-pair inverted duplication. Gene, 17, 79--89, 1982) as a source of DNA to isolate the CAT gene without its own promoter region. Plasmid pBR329 was cleaved with Taql endonuclease electrophoresed on 1.2% agarose gel, the band corresponding to the 785 bp fragment was cut out and the DNA recovered from the agarose. This fragment contains the DNA sequences coding for the CAT protein and the sequences signalling for the initiation and termination of translation, (i.e. ribosome binding site, start and stop codons) but it does not contain the promoter of the CAT gene.The schematic structure of the pBR321 Taql 785 bp fragment is shown in Fig. 13. The isolated fragment was ligated with Clal cleaved pERlll-8rO DNA and transformed into E. coli JM107. Transformed cells were plated on complete media and screened for ampicillin and chloramphenicol resistance. Plasmid DNA was isolated from 12 independent ampicillin and chloramphenicol resistant clones, digested with restriction endonucleases and the fragment pattern was analysed by agarose gel electrophoresis. We found that all these plasmids contained the 785 bp Taql fragment in the same orientation, i.e. the direction of transcription of lac and CAT was the same. One representative of these plasmids was designated as pER-CAT.On the basis of the known sequences of pER111-8rO and pBR329, pER-CAT should carry a hybrid operon, producing an RNA containing two regions for translation initiation. Therefore cells containing one of the pER-CAT plasmids will produce two new polypeptides from which one is the native CAT protein. In accordance with this expectation, the SDS-polyacrylamide gel electrophoretic pattern of proteins extracted from pER-CAT plasmid containing cells shows an extra band of a 24 kd protein which can be identified as native CAT enzyme (Fig. 14).

As we wanted to compare the effectiveness of different rrnB-lac hybrid regulatory regions the fragment containing the pERlll-8 type expression control region in pER-CAT was replaced with the corresponding fragments from different pER plasmids. pER-CAT DNA was restricted with Pvul, the resulting two fragments were separated by agarose gel electrophoresis and the longer one was reisolated from the gel. This fragment contains the C-terminal part of the amplicillin resistance gene, the region of replication, the rrnB terminator region and a short region coding for the last amino acids of lac-alfa peptide. The hybrid control region was isolated in a similar way from different pER plasmids. Aliquots of the isolated pER-CAT Pvul fragment were ligated with each isolated pER fragment and the ligated DNAs were transformed into JM107 cells.Plasmid DNAs were isolated from ampicillin and chloramphenicol resistant clones and analysed with restriction endonucleases to verify the presence of the desired control region. The protein content of the several clones containing one of the above-mentioned plasmids was analysed Dy SDS-polyacrylamid gel electrophoresis. We found significant differences in the amount of CAT (some of the ot tained patterns are shown in Fig. 14). On the basis of densitometric scanning of comassiebriliiant-blue-sta" n i gels the amount of CAT synthesised in cells containing one of the best plasmids was estimated to be as high as 50% of total cell protein.

All of the above mentioned plasmids contain the CAT gene with its own ribosome binding site, which separates translation of the alfa-peptide and that of the CAT protein. In order to characterise the hybrid control region with respect to translation initiation, in the next step we fused the two protein coding regions by removing the CAT translational signal sequences. This was achieved by treating the original pER-CAT DNA with Xbal at the unique site between the lac and CAT coding region, digesting with BAL31 and recircularising with ligase. The proper reading frame was ensured by applying chloramphenicol selection for CAT expression. Plasmids obtained from chloramphenicol resistant clones were characterized by restriction analysis and one of them having the largest deletion was chosen forfurtherwork. It is designated as pER-CATf3.It has a fused operon in which the codon of the 60th amino acid of lac alfa peptide is followed by the codon of the 9th amino acid from CAT. As pER-CATf3 contains only one ribosome binding site in the rrnB-lac hybrid regulatory region it has to direct the synthesis of a fused protein with a mol. weight of 30 kd.

This was verified by analysing the patterns of proteins extracted from pER-CATf3 containing cells (Fig. 14).

Several different pER hybrid regulatory regions were used to repiace the signal sequences in pER-CATf3 in a similar manner as described in connection with pER-CAT. The amount of the fused polypeptide synthetised under the control of a given hybrid regulatory region was found to be proportional to the amount of intact CAT synthetised under the control of the same regulatory region.

EXAMPLE 3 The effectiveness of the pER plasmids of expression vectors for inserted eukaryotic DNA may be seen from the results obtained by insertion of a proinsulin gene into pERlll-8. The cDNA coding for human proinsulin was isolated from a recombinant plasmid which was constructed in the Biochemical Institute of BRC by using conventional methods. This plasmid is a pBR322 derivative it carries the poly DC-poly dG tailed proinsulin gene in the Pstl site. In order to obtain a DNA fragment coding for proinsulin in such a form that it could be easily inserted into one of the pER plasmids the above plasmid was restricted with Pstl endonuclease the proinsulin gene fragment was isolated from agarose gel, treated with BAL31 nuclease and then ligated with Clal linkers.Insertion of such fragment into the Clal site of an appropriate vector resulted in a series of recombinant plasmids. These were analysed by DNA sequencing and one of them having the insert as shown in Fig. 13 was selected as a source to isolate the proinsulin gene. In this plasmid the three nucleotides coding for the first amino acid of the proinsulin gene were replaced by the last nucleotides of the Clal linker which formed a Met codon. To isolate the proinsulin gene the above plasmid was restricted with Clal, electrophoresed on 1.5% agarose gel and the 257 bp fragment was cut out and recovered from the gel. This fragment was inserted into the Clal site of pERlll-8rO. Following transformation plasmid DNAs were prepared from a number of colonies and these were used to confirm that some of the plasmids had the desired structure.On the basis of restriction mapping we determined that in plasmid pSz1153 the inserted insulin gene is in the same orientation as the direction of the transcription of the rrnB-lac hybrid region. Moreover as the determination of the nucleotide sequence revealed, the proinsulin gene was in-frame with the lac translation. Therefore we expected pSz1153 to produce a lac-insulin fused polypeptide with a molecular weight of 1 6--18 kd.Although the production of this polypeptide was detected by anaiysing pSz1153 containing cell-extracts on SDS-polyacrylamide gel, and the presence of the peptide fragment corresponding to human proinsulin was confirmed by precipitating the fused protein with antibodies produced against human insulin the amount of the fused protein was found to be very low, even in extracts from IPTG induced cells. Using radioactively-labelled amino acids we found that the fused protein was degraded rapidly in E. coli JM107 cells, so we tried to increase the production of the fused protein by increasing its half-life. It is known that the stability of a foreign protein in E.coli cells could be increased by combining it with native bacterial proteins.In accordance with this, we modified plasmid pSz1153 by elongating the DNA piece coding for beta-galactosidase. A plasmid containing the intact beta-galactosidase gene (see for example: pRLT2, Erdei et al.: A novel type of bacterial transcription unit, specifying rRNA,andtRNA. Mol. Gen. Genet. 191,162-164,1983) was restricted with Pvulland Clal endonucleases and a fragment of 700 bp coding for the 33-263rd amino acids of beta-galactosidase was isolated from agarose gel. This fragment was ligated with the longer fragment of pSzi153 which was obtained by using the same restriction endonucleases.This latter fragment is a linear derivative of pSz1153 from which the last nucleotides of the alpha-peptide-coding region (up to the Pvull site) and the first nucleotides of the polylinker region were missing. Following ligation and transformation, plasmids were selected on the basis of restriction fragment pattern and one ofthem which showed the desired structure was chosen as a modified vector for the insertion of the proinsulin gene. Digestion of pSz1153 resulted in two fragments. These were separated on 1.5% agarose gel and the smaller one containing the proinsulin gene was recovered from the gel. This fragment was ligated with the Clal-cleaved modified vector and the DNA was transformed into E. coli JM107 celis. In order to ensure the correct reading frame before ligation both of the two fragments were treated with Klenow fragment of DNA polymerase in the presence of each of the four dNTPs to fill in the Clal generated protruding 5' ends. Colonies of E. coli each containing one of the resulting plasmids were then screened for ampicillin resistance and plasmid DNA was isolated from a number of individual clones. Detailed restriction analysis revealed that some of these plasmids carried the inserted insulin gene in proper orientation so that the translation of the hybrid operon results in a fused polypeptide consisting of the 263 N-terminal amino acids of beta-galactosidase and the 86 amino acids of human proinsulin at the C-terminal region. This type of plasmid was designated as pSzL23.Protein patterns of pSzL23 containing E. coli cells confirmed the production of an extra polypeptide of the molecular weight calculated for the fused protein (Fig. 14). Replacing the pERIl 1-8 type expression control region of pSzL23 with the corresponding regions of different pER plasmids we found that just as in the case of CAT expression, the amounts of the fused protein synthetised under the control of different hybrid regulatory regions varied considerably. Among the compared plasmids there were some which provided for the synthesis of the fused protein up to 510% of the total protein.

On the basis of these experiments we believe that the expression vectors constructed according to the present invention may be used successfully for programming high level of protein production in E. coli cells.

LEGEND TO FIGURES Fig. 1 Schematic structure of the recombinant plasm id pBB9 The vector DNA (pBR322) is symbolised with an open double line the inserted DNA with a thick black line. The internal circle shows the location of a functional units: amp=DNA coding for resistance to ampicillin; tet=DNA coding for resistance to tetracycline; 16S, 23S, 5S= DNA sequences corresponding to the mature ribosomal RNAs; P=promoter; T=terminator; A=DNA originating from the bacteriophage lambda chromosome. Restriction endonuclease cleavage sites are indicated by arrows: E=EcoRI; B=BamHI; P=Pstl; H=Hindlll; S=Sall, Hp=Hpal; Pv=Pvull, F=Fspil.

Fig. 2 Schematic structure of the recombinant plasmid pBB9a9 Symbols are as in Fig. 1. Deletion is represented as a shaded area.

Fig. 3 The nucleotide sequences of the junctions in plasmids obtained by removing parts of the recombinant molecules step by step.

A, Junction between the first part of 16S rRNA gene and the last part of 23S rRNA gene. i.e. the junction which was created by deleting part of the rrnB operon from pBB9 to construct pBB9A9.

B, Junction between bacterial DNA following the terminator region of rrnB operon and pBR322 DNA i.e., the junction created by the deletion which was made to construct pBB9-b28 from pBB9A9.

C, Junction between the promoter region of rrnB operon and pBR322 DNA i.e. the junction created by the deletion which was made to construct p419-10 from p408-5.

Fig. 4 Schematic structure of the recombinant plasmid pBB9-b28 Symbols are as in Fig. 1 and Fig. 2.

Fig. 5 Schematic structure of plasmid p408-5 Symbols are as in Fig. 1 and Fig. 2.

Fig. 6 Successive steps of the construction of pER plasmids Only those restriction cleavage sites are indicated which were used in construction. Solid black bar: region coding for rRNA. Hatched bar: region coding for beta-galactosidase. P and T; promoters and terminators of rrnB, ApR: region coding for beta-lactamase, RO:origo of replication. In the centre the origins of the fragments constituting the plasmids are shown. a: deletions.

Fig. 7 Nucleotide sequence of the hybrid regulatory region of plasmid pER111-8.

Nucleotides originated from the lac operon are written in italics. Dotted line: AT-rich prepromoter region of rrnB P2; solid line: -35 region of rrnB P2 promoter; double line: -10 region (Pribnow-Box) of the lac promoter; *: start point of transcription initiation; broken line: ribosome binding site; The lac operator and the first two codons of betagalactosidase are also indicated.

Fig. 8 Construction of the polylinker region of plasmid pERlll-8r1.

The polylinker fragment of pERlll-8r1 (bottom) was obtained by cleaving pERlll-8p1 both with Xbal and EcoRI (a), filling in the sticky ends of the DNA (ss) and religating (y). This procedure removed 69 nucleotides (i.e. 23 codons), therefore the phasing remained unchanged (Hindlll, EcoRI: PH 1; Clal: PH 2; Xbal: PH 3.) Fig. 9 Construction of the polylinker region of plasmid pERlll-8r3.

The polylinker fragment of pERlll-8r3 (bottom) was obtained by cleaving pER111-8p1 both with XBAI and BamHI (a), filling in the sticky ends of the DNA (ss) and religating (y). This procedure removed 56 nucleotides (i.e. 18 codons+2 nucleotides), therefore changed the phasing from the BamH1 site. Hindlll, BamHI, Clal: PH 1; EcoRI: PH 3).

Fig. 10 Construction ofthe polylinker region of plasmid pERI11-8rO.

The polylinker fragment of pERllI-8r0 (bottom) was obtained by cleaving pERllI-8r1 (top) with EcoRI (a) filling in the sticky ends of the DNA (ss) and religating (y). This procedure added nucleotides, therefore the phasing from the Hindlll site was changed. (Hindlll: PH1; Xbal, Clal: PH 3).

Fig. 11 Construction of the polylinker region of plasmid pERlll-8r2.

The polylinker fragment of pERlll-8r2 (bottom) was obtained by cleaving pERllI-8p1 (top) both with Bglll and BamHI (a), filling in the sticky ends of the DNA (ss) and religating (y). This procedure removed 50 nucleotides (i.e. 16 codons+2 nucleotides), therefore changed the phasing from the Clal site (Hindlll, Clal: PH 1; EcoRI: PH 2; Xbal: PH3) Fig. 12 Schematic structure of the pER expression vectors and the nucleotide sequences of the polylinker regions.

Only those restriction sites are indicated which are suitable to change the regulatory regions between different pER plasmids. Over the polylinker sequences the reading frame of the translation starting at the lac initiator codon is indicated.

Fig. 13 Schematic structure of the two DNA fragments which were inserted into pER Plasmids to demonstrate protein production A, Structure of the pBR329 Taql fragment coding for CAT.

B, Structure of the DNA fragment coding for human proinsulin.

Fig. 14 Proteins extracted from JM107 cells containing different pER plasmids.

Samples 1 JM107-pER-CAT (pERlIl-8 promoter region) 2 JM107-pER-CAT (pERVI-21 promoter region) 3 JM107-pER-CAT (pERVI-20 promoter region) 4 JM107-pER-CAT (pERVI-23 promoter region) 5 JM107-pER-CATf3 (pERIll-8 promoter region) 6 JM107-pER-CATf3 (pERVI-21 promoter region) 7 JM107 - 8 JM107-pSZIL23 (pERVI-23 promoter region) 9 JM107-pSUIL23 (pERVI-23 promoter region) STRAINS AND PLASMIDS Escherichia coli HB101 (pro~, let~, thi-, lac-, strR, r, m, endow~: Boyer, H. W., Roulland-Dussoix, D. (1969) A complementation analysis of the restriction and modification of DNA in E. coli. J. Mol.Biol. 41, 45472) Escherichia coli ED8800 (supE, supF, hsdS-, met~, IacZM15, recA56; Murray, N. E., Brammer, W. J. and Murray, K. (1977) Lambdoid phages that simplify the recovery of in vitro recombinants. Mol. gen. Genet. 150, 5361 ) E. coli K12 JM107; Yanisch-Perron, C., Vieria, J. and Messing, J.: Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13 mp18 and pUC19 vectors. Gene 33 (1985) 103--119.

arid1 8 (Kirschbaum, J. B. and Konrad, E.B. (1973) Isolation of a specialized lambda transducing bacteriophage carrying the beta subunit gene for Escherichia coli ribonucleic acid polymerase. J. Bacteriol. 116, 517-526) pBR322 (Bolivar, F., Rodriguez, R. L., Green, P. J., Betlach, M., Hyneker, H. L., Boyer, H. W., Crosa, J. and Falkow, S. (1977) Construction and characterization of new cloning vehicles. Gene 2, 9113) pBR329 (Covarrubias L. and Bolivar F.: Construction and characterization of new cloning vehicles VI. Plasmid pBR329, a new derivative of pBR328 lacking the 482-base-pair inverted duplication. Gene, 17, 7989, 1982) pLBU3 (Gentz, R., Langer, A., Chang, A. C. Y., Cohen, S. N. and Bujard, H. (1981) Cloning and analysis of strong promoters is made possible by the downstream placement of a RNA termination signal. Proc. Natl. Acad.

Sci. USA 78, 4936-40) pHC314 (Boros I., Pósfai Gy. and Venetianer P. (1984) High copy-number derivatives of the plasmid cloning vector pBR322, Gene 30,257-260) MATERIALS AND METHODS Restriction endonucleases: BamHI, BspRI, Hpal, Sall, Pvull, EcoRI, Mspl, Hindlll, Pstl, Bglll, Xbal were purified in the Biochemical Institute of BRC according to established protocols Roberts, R. J., Restriction and modification enzymes and their recognition sequences, Nucleic Acids Res. 11, r1 35r1 73, 1983) or were from New England Biolabs, Fspll and T4 DNA ligase were generous gifts of Dr. M. Szekeres and Dr. A.

Udvardy respectively. Clal, Stul and Hhal restriction endonucleases, BAL31 nuclease and E. coli DNA polymerasel Klenow fragment were from BRL, bacterial alkaline phosphatase from Worthington, RNAse and lysisimefrom REANAL.

(y~32p) ATP and (a-32P) dATP were from the Hungarian Isotope Institute, Budapest.

All other reagents were analytical grade commercial products.

Media E. coli strains were grown in complete medium containing 10 g tryptone, 5 g yeast extract and 5g NaCI per liter. Colonies were grown on the same media containing 1.5% Bacto agar.

The in vivo beta-galactosidase alpha-complementation was detected on X-gal-indicator plates containing: 20 g Casamino acids, 6 g Na2HPO4, 3 g KH2PO4, 0.5 g NaCI, 1 g NH4Cl, 15 g Bacto-Agar, 20 mg X-gal -dimetil4ormamide/liter, supplemented with 2 mM MgCl2 0.1 mM CaCI2-1001lg/ml amp.

Enzyme reactions Restriction endonucleases were used according to the recommendations of New England Biolabs.

Cohesive-end ligations were performed at volumes of 30--40 Cil in 66 mM Tris-HCI pH 7.6,5 mM MgCl2, 5 mM dithiothreitol, 1 mM ATP using 1 unit of T4 ligase for 0.5--1.0 g of DNA. Reactions were incubated overnight at 14"C.

Blunt-end ligations were carried out in 25 mM Tris HCI pH 7.4, 5 mM MgCl2, 5 mM dithiothreitol, 0.25 mM spermidine, 1 mM ATP, 10 pg/ml BSA, 30--40 Cig/ml DNA, 50 pimp T4 ligase, at 14"C, for 12-14 hours.

Filling in of 5' extensions was achieved by treatment with DNA polymerase I Klenow fragment in the presence of the four dNTPs. The reactions contained 50 mM Tris-HCI pH 7.4,7 mM MgCl2, 1 mM dithiothreitol, 0.1 mM each of the four deoxynucleotide triphosphates, 20250 ugiml DNA and 100 p/ml DNA polymerase Klenow fragment in 10--20 Cil final volume. The reactions were carried out at 37"C for 15 min.

For constructing deletions appropriately cleaved plasmid DNA was extracted with phenol-chloroform, precipitated with ethanol and dissolved in 600 mM NaCI, 20 mM Tris-HCI pH 8.0,1.0 mM EDTA, 12 mM CaCI2, 12 mM MgCl2to obtain a 200--250 pg/mlfinal concentration of DNA. 0.41.2 U BAL31 nuclease was added for 1 pg DNA and incubated at 30"C. Aliquots were removed after different times of digestion.

Progress of the reaction was monitored by agarose gel electrophoresis and digestion with restriction endonucleases which were known to have cleavage sites in the area to be analysed. After digestion the reaction was stopped by extraction with phenolchloroform, the DNA was precipitated with ethanol and treated with DNA polymerase Klenow fragment to produce blunt ends.

DNA preparation For plasmid DNA purification cultures of E. coli strains containing the desired plasmids were grown in complete media. At an A600 of 0.6 chloramphenicol was added to cultures to a final concentration of 170 pg/ml and incubation continued at 37 C to amplify plasmids. For large-scale plasmid preparation cleared lysates were prepared as described by Clewel, D. B. and Helinski, D. R. (Supercoiled circular DNA-protein complex in Escherichia coli: purification and induced conversion to an open circular DNA form. 1969, Proc.

Natl, Acad. Sci. USA, 62 1159--1166). DNA was isolated from cleared lysates either by using the caesium chloride/ethidium bromide method or by Sephacryl-Slooo chromatography. Small-scale preparations were done by the method of Birnboim and Dolly (A rapid alkaline extraction procedure for screening recombinant plasmid DNA. Nucleic Acids Res. 1979.7, 1513).

Electrophoresis Electrophoresis of DNA fragments of 0.2% horizontal agarose slab gels or 48% vertical polyacrylamide slab gels and protein electrophoresis on 12% SDS-polyacrylamide gels were carried out under conventional conditions. DNA fragments were reisolated from agarose and polyacrylamide gels by electroelution or by using the method described by Hammarskjöld M. and Winberg G. (Isolation of DNA from agarose gels using DEAE paper. Application to restriction site mapping of adenovirus type 16 DNA.

1980. Nucleic Acids Res. 8,253).

All other methods of molecular biology were used as described in Maniatis, T., E. F. Fritsch, Sambrook J.: Molecular cloning, Cold Spring Harbor Lab. NewYork, 1982.

Deposition Date of the Plasmid/Strain Number Deposition pBB9 00300 October31,1984 pBR322 00298 September 1984 p 408-5 00301 October 31, 1984 p827-10 00307 October31,1984 pERlll-8 00302 October31, 1984 pHC314 00280 March 7,1984 pERlli-8r1 00303 October31, 1984 pERIII-8r3 00304 October 31,1984 pERIII-8r0 00305 October 31, 1984 pERIll-8r2 00306 October31, 1984 E. coli HB 101 00290 July 25, 1984 E. coli ED 8800 00291 July 25, 1984 pBR 329 00299 September 1984

Claims (19)

1.An expression vectorfor protein production in E. coli cells, containing a hybrid operon which comprises parts of an E. coli rRNA operon and a part of a protein coding operon.

2. An expression vector according to claim 1 characterized in that said rRNA operon contains a deletion which removes most of the structural part of the operon.

3. An expression vector according to any one of claims 1 and 2 characterized in that said part of a protein coding operon is a part of the E. coli lac operon.

4. An expression vector according to claim 3 characterized in that said part of the E. coli lac operon is the EcoRI-Hindlll fragment of plasmid pLBU3.

5. An expression vector according to any one of claims 1 to 4 characterized in that the vector which carries said hybrid operon is a derivative of pBR322.

6. An expression vector according to any one of claims 1 to 5 characterized in that the combination of said rRNA operon and said protein coding operon results in a hybrid promoter region.

7. An expression vector according to any one of claims 1 to 6 characterized in that it also contains a polylinker beyond said protein coding operon.

8. Plasmid pERlll-8, comprising a portion of the E. coli lac operon (nucleotides from - 14 to +243, where + 1 is the first nucleotide of the lac mRNA) and a portion of the pBR322 (nucleotides from 2068 to 4278, where +1 is the first nucleotide of the EcoRI site of the pBR322) which are flanked by portions of the E. coli rrnB operon,the nucleotide at -14 of the lac operon being joined to the nucieotide at -189 ofthe rrnB operon (where +1 is the first nucleotide of the mature 16S rRNA) creating a hybrid promoter which contains a -35 region from the rrnB P2 promoter and a -10 region from the lac promoter, in which the portion of the lac operon also contains the ribosome binding site and the region coding for the N-terminal 68 amino acids of beta-galactosidase, this portion is attached to a second portion of the rrnB operon through a Hindlll site, following the lac region the portion of rrnB operon (a streach of nucleotides from +78 to +270 joined to a streach of nucleotides from +4965 to 5508) carries the terminator region of the rrnB operon, the portion of pBR322 contains the region which ensures ColEl-type replication and the gene coding for ampicillin resistance, this portion links the two portions of the rrnB operon such that the terminator region from the rrnB operon and the region of replication from pBR322 are neighboring.

9. A derivative of plasmid pERlll-8 referred to as pERI11-8rO having a 44 bp insert in the Hindlll site, in which the insertion bears unique sites of Xbal and Clal endonucleases allowing the insertion foreign DNA into them.

10. A derivative of plasmid pERIl 1-8 referred to as pERlll-8r1 having a 40 pb insert in the Hindlll site, in which the insert bears unique sites of the Xbal, EcoRI and Clal endonucleases allowing the insertion foreign DNA into them.

11. A derivative of plasmid pERlll-8 referred to as pERllI-8r2 having a 58 bp insert in the Hindlll site, in which the insert bears sites of the Xbal, Clal and EcoRI endonucleases allowing the insertion foreign DNA into them.

12. A derivative of plasmid pERlll-8 referred to as pERlll-8r3 having a 54 bp insert in the Hindlll site, in which the insert bears sites of the BamHI and EcoRI endonucleases allowing the insertion foreign DNA into them.

13. A process for the production of expression vectors which comprises removing most of the structural part of an E. coli rRNA operon carried by a suitable vector to obtain a shortened rRNA operon, combining said shortened rRNA operon with a part of a protein-coding operon to produce an effective hybrid expression control region, inserting a DNA fragment coding for a desired polypeptide into the combined operon to place under the control of said hybrid expression control region.

14. A process as claimed in claim 13 in which said shortened rRNA operon is constructed in such a manner that it contains deletions which terminate in the rRNA structural genes.

15. A process as claimed in claims 13 or 14 in which said suitable vector is a derivative of pBR322.

16. A process as claimed in claim 15 characterized in that the unessential parts of said pBR322 are removed by enzyme digestion.

17. A process as claimed in claims 13 to 16 in which said hybrid expression control region is constructed by combining parts of an rRNA operon and a part of the E. coli lac operon and creating deletions between the two regions to obtain a vector series from which those expression vectors are selected which show high lac activity.

18. A process as claimed in claim 17 in which said vector series are produced by digestion with BAL31 nuclease in such a way that hybrid promoters containing a part of the rRNA operon and a part of the lac promoter are resulted in.

19. A process as claimed in claims 13 to 18 characterized in that a polylinker fragment is attached to the end of the lac region to obtain expression vectors which allow to insert foreign DNA in different reading frame.