WO1992015688A1

WO1992015688A1 - Production of recombinant proteins

Info

Publication number: WO1992015688A1
Application number: PCT/GB1992/000386
Authority: WO
Inventors: Ian George Charles; Andrew Joseph Makoff; Marcus Dayman Oxer; Timothy Charles Peakman; Jeffrey Alan Cole
Original assignee: Wellcome Foundation Ltd
Current assignee: Wellcome Foundation Ltd
Priority date: 1991-03-05
Filing date: 1992-03-05
Publication date: 1992-09-17
Anticipated expiration: 1993-09-05
Also published as: GB9104596D0; RU2126447C1

Abstract

A process for the expression of a heterologous protein, which process comprises maintaining under anaerobic conditions bacteria in which the expression of the said protein is under the control of a promoter whose activity is induced by anaerobic conditions. Preferably the promoter is the nirB promoter.

Description

PRODUCTION OF RECOMBINANT PROTEINS

This invention relates to the production of recombinant proteins and in particular the use of a promoter suitable for the expression of foreign proteins in bacteria, particularly Escherichia coli.

The large scale production of foreign proteins in E. coli requires the use of a strong promoter that is well regulated, allowing the growth phase and the induction phase to be separated. Without this regulation a highly expressed gene can place a constraint on cell growth and plasmid stability, even if its product is not actually toxic. Of the many E. coli or coliphage promoters, only a few satisfy these requirements: e.g. p_L from lambda, lac and trp from E. coli, the hybrid trp-lac (tac) promoter and the T7 RNΑ polymerase promoter. All of these rely on either a temperature shift or the addition of a chemical to induce their activity. A temperature shift may cause the recombinant protein to form inclusion bodies, and furthermore by activating the heat-shock response may lead to increased proteolysis. Chemical agents are often expensive and require subsequent removal from the recombinant protein at the purification stage.

The nirB promoter has been isolated from E. coli. where it directs expression of an operon which includes the nitrite reductase gene nirB (1), and nirD, nirC and cysG (2). It is regulated both by nitrite and by changes in the oxygen tension of the environment, becoming active when deprived of oxygen (3). Response to anaerobiosis is mediated through the protein FNR, acting as a transcriptional activator, in a mechanism common to many anaerobic respiratory genes (1). By deletion and mutational analysis the part of the promoter which responds solely to anaerobiosis has been isolated and by comparison with other anaerobically-regulated promoters a consensus FNR-binding sice was identified (4,5). It was also shown that the distance between the putative FNR-binding site and the -10 homology region is critical (6).

Some other bacterial oxygen-regulated promoters have been shown to require very low levels of oxygen for full transcriptional activation. The nifA promoter of Rhizobium meliloti (7) and the Vitreoscilla sp. haemoglobin (VHb) promoter (8) are examples of promoters with such a microaerobic optimum. Indeed these promote relatively little transcription under fully anaerobic conditions.

WO 89/03883 relates to the use of changes in the level of oxygen available in the culture medium to control the expression of foreign DNA in a host cell. In particular, this specification is concerned with the use of the VHb promoter referred to above. The fact that this promoter requires the presence of small amounts of oxygen for optimum expression makes expression difficult to regulate since it is difficult to control the amount of oxygen in the culture with sufficient precision.

The present invention provides a process for the expression of a heterologous protein, which process comprises maintaining under anaerobic conditions bacteria in which expression of the said protein is under the control of a promoter whose activity is induced by anaerobic conditions. Expression of the protein can thus occur in a bacterial culture under anaerobic conditions.

The use of a promoter induced by anaerobic conditions has considerable advantages over promoters such as the VHb promoter in terms of the ease with which such conditions can be maintained. Thus, particularly in the case of large scale fermentation involving high cell densities, it is much easier to maintain completely anaerobic conditions rather than the microaerobic conditions required by the VHb promoter.

According to the invention, DNA encoding the heterologouis protein of interest will be placed under control of the promoter, for example in a replicable expression sector such as a plasmid vector. Bacteria, for example an appropriate strain of E. Coli. transformed with the expression vector are then used for expression of the heterologous protein. The process involves a growth phase in which the bacteria are allowed to grow whilst a high oxygen tension is maintained in order to prevent premature activation of the promoter. This is followed by an induction phase in which the bacteria are maintained under anaerobic conditions in order to activate the promoter.

The aerobic conditions required during the growth phase can be maintained, for example, by sparging with an oxygen containing gas such as pure oxygen, oxygen enriched air or air. The anaerobic conditions required to activate the promoter can be established and maintained by sparging with an inert gas such as nitrogen. The anaerobic conditions may be maintained for up to 19 hours, for example for from 5 to 19 hours.

The process according to the invention can be applied to the production of any heterologous protein that can be produced in bacteria. The DNA encoding the protein can be incorporated into an expression vector under control of the promoter by use of standard recombinant DNA techniques. The expression vector will also generally include other elements required for efficient expression. Following expression the heterologous protein can be recovered and purified by conventional methods.

The heterologous protein typically is a physiologically active polypeptide such as an enzyme. The polypeptide may be a polypeptide drug. Polypeptide drugs which may be expressed by the present process therefore include tissue plasminogen activator, luteinizing hormone releasing hormone, human growth hormone, insulin, erythropoietin, an interferon such as α-interferon and calcitonin.

Alternatively, the heterologous protein may be a polypeptide immunogen capable of inducing an immune reponse in a human or animal. A polypeptide immunogen may comprise an antigenic determinant of a pathogenic organism. The polypeptide immunogen therefore typically comprises an antigenic sequence derived from a virus, bacterium, fungus, yeast or parasite. More especially, the immunogen may be an antigenic sequence derived from a type of human immunodeficiency virus (HIV) such as HIV-1 or HIV-2, hepatitis A or B virus, herpes simplex virus, poliovirus type 2 or 3, foot-and-mouth disease virus, influenza virus, coxsackie virus, the cell surface antigen CD4, Chlamydia trachomatis and Plasmodium falciparum. The immunogen may comprise the CD4 receptor binding site from HIV, for example from HIV-1 or -2. Other useful immunogens include E. coli heat labile toxin B subunit (LT-B), E. coli K88 antigens, P.69 protein from B. pertussis, tetanus toxin fragment C and antigens of flukes, mycoplasma, roundworms, tapeworms, rabies virus and rotavirus. The immunogen may be a tumour-specific antigen.

It has been found that the nirB promoter referred to above can be activated by the use of anaerobic conditions and does not show a microaerobic optimum. The nirB promoter also shows a strength and an induction ratio

(the ratio of expression levels before and after induction) which are comparable to known bacterial promoters such as the tac promoter. For this reason the nirB promoter is particularly suitable for use according to the invention.

The nirB promoter has been isolated from E. coli and is particularly suitable for the control of expression in E. coli.

As noted above the nirB promoter is regulated both by nitrite and by changes in oxygen tension and the promoter can be used according to the invetion in this form. However, it is preferred to use only that part of the nirB promoter which responds solely to anaerobiosis. As used herein references to the nirB promoter refer to the promoter itself or a part or derivative thereof which is capable of promoting expression of a coding sequence under anaerobic conditions.

Examples of proteins which have been expressed in E. coli under control of a form of the nirB promoter which lacks the nitrite response regions are tetanus toxin fragment C and Bordetella pertussis pertactin. These proteins are both potential components of sub-unit vaccines against tetanus and whooping cough respectively.

The present invention also proivdes a DNA molecule comprising the nirB promoter sequence operably linked to a DNA sequence encoding a protein heterologous to the host bacteria. The invention further provides a replicable expression vector, suitable for use in bacteria, in which a DNA sequence encoding a heterologous protein is under the control of a promoter whose activity is induced by anaerobic conditions. A preferred vector comprises the nirB promoter sequence operably linked to a DNA sequence encoding a protein heterologous to the host bacteria. The invention also provides bacteria, particularly an E. coli strain, transformed with such an expression vector. A bacterial culture containing the expression vector can thus be provided.

The invention also provides the use of a promoter activated by anaerobic conditions for the control of expression of a heterologous protein in bacteria. More particularly the invention provides the use of the nirB promoter for the control of expression of a heterologous protein in bacteria, preferably an E. coli strain.

The following Examples illustrate the invention. In the accompanying drawings:

Figure 1 shows a Coomassie blue stained sodium dodecyl sulphate (SDS)-polyacrylamide gel of E. coli extracts containing pTETnir215 or pTETnir36; and

Figure 2 shows the accumulation of tetanus toxin fragment C, as % total cell protein (top), as a function of time. Dissolved oxygen tension (DOT) is also shown as % saturation. Example 1 - Construction of plasmids

The Examples used E. coli strain MM294 (9) and plasmids pTETnir15, pPERnir36, pTETnir215 and, for purposes of comparison pTETtac315. Expression plasmids pTETnir15 and pPERnir36 were constructed from pTETtacll5 (10) and pPERtac36 (11) by replacing in each the EcoRI-ApaI region (1354/bp and 1294/bp respectively) containing the lacI gene and tac promoter with the pair of oligos 1 and 2 which contain the nirB promoter:

Oligo-1 5'-AATTCAGGTAAATTTGATGTACATCAAATGGTACCCCTTGCTGAAT Oligo-2 3'-GTCCATTTAAACTACATGTAGTTTACCATGGGGAACGACTTA

CGTTAAGGTAGGCGGTAGGGCC-3 '

GCAATTCCATCCGCCATC-5'

The oligonucleotides were synthesized on a Pharmacia Gene Assembler and the resulting plasmids confirmed by sequencing (12). Plasmid pTETnir215 was constructed by ligating the 1530 bp AatII-BamHI fragment of pTETnir15, containing the nirB promoter and fragment C gene, with the 1983 bp BamHI-AatII fragment from pTETlac5 (13), containing the pUC-based origin of replication and β- lactamase gene. The plasmid pTETtac315 contains the tetanus toxin fragment C sequence under control of the tac promoter (13).

Example 2 -Capped-tube Inductions

Overnight cultures of E. coli (A₆₅₀ approximately

2.0) containing the plasmids described above were diluted 30-fold in fresh L-Broth containing 4 mgml^-1 glucose and

100 μgml^-1 ampicillin and used to completely fill 7ml Bijou bottles. These were then capped tightly and incubated at

37°C for 4-6h before being harvested and analysed for expression.

The plasmid pTETnir15 contains the gene for tetanus toxin fragment C under nirB promoter control. The version of the nirB promoter used was based on the F2 DR25X variant described by Bell and co-workers (6), the BamHI (GGATCC) site of which was replaced by a Kpnl (GGTACC) site, involving only a minimal sequence change. An equivalent plasmid, pPERnir36, had the gene for B. pertussis pertactin [also referred to as P69 (11)] in place of the fragment C gene.

Using these expression vectors in capped-tube cultures showed that in the absence of air the nirB promoter sequence induced synthesis of either fragment C or pertactin, both readily visible on stained SDS-polyacrylamide gels. It has previously been shown that the production of fragment C is mRNA-limited, over a wide range of mRNA levels, and consequently a good indicator of promoter strength (13). In contrast, the production of pertactin becomes mRNA-saturated at relatively low concentrations. Fragment C synthesis was therefore chosen as a model system for studying promoter strength.

Fragment C levels were increased about two-fold when the pAT153 replicon of pTETnir15 was replaced by the higher copy-number pUC19 replicon, a similar improvement to that observed in lac (unpublished observations) and tac expression vectors (10,13). The improved vector, pTETnir215, was used for later experiments as the higher fragment C levels could be more accurately estimated.

Example 3 - Anaerobic Induction in a Fermenter

The above capped-tube experiments could only be relatively crude, with little control over, or facility to measure, the dissolved oxygen tension (DOT) during the induction. Subsequent experiments were therefore carried out in a bench-top stirred-tank fermenter.

All fermentations were carried out in a Braun

Biolab 21 working-volume fermenter equipped with control systems for pH, temperature, and dissolved oxygen. The polarographic probe for measuring DOT was zero-calibrated by sparging the fermenter with oxygen free nitrogen. 1.5 1 of L-Broth containing 10 mgml^-1 glucose and 100 μgml^-1 ampicillin was inoculated with 45 ml of overnight culture of E. coli containing either pTETnir215 or pPERnir36. This was grown aerobically for a period of 4-5h at pH7.2, stirrer rate of 750rpm and air input at up to 1.3vvm to maintain a DOT value in excess of 75% of saturation after which time the A₆₅₀ was typically 1.5-3.0. For inductions the air supply was switched off, the stirrer rate was lowered to 400 rpm and the culture sparged with oxygen-free nitrogen at 1.3 wm until the DOT reading had reached a plateau, and then for a further 2 min before being reduced to 0.25 wm for the remainder of the induction. Samples of the culture were taken at 1 h intervals, their A₆₅₀ determined and 1 ml aliquots pelleted in a microfuge for 2 min, to be analysed for expression.

Cell pellets from the above inductions were resuspended in Laemmli loading buffer and the equivalent of 0.1 A₆₅₀ units of each sample was run on a 7.5% or 6% (w/v) SDS-polyacrylamide gel (22). The gels were either stained with Coomassie blue and expression levels quantitated by densitometer scanning using a Joyce-Loebl Chromoscan 3, or blotted. Western blots were detected as described previously (10,12).

Inductions were carried out on cultures of E.coli containing pTETnir215 or pPERnir36. Analysis of samples showed that recombinant protein accumulated following anaerobiosis, demonstrating that the nirB promoter is active under completely anaerobic conditions.

Figure 1 shows Coomassie blue stained SDS-polyacrylamide gel of E. coli extracts containing pTETnir215 or pPERnir36. Extracts from pTETnir215: preinduction (lane 1) and 19 hours post-induction (lane 2); from pPERnir36: pre-induction (lane 3), and 20 hours post-induction (lane 4). Molecular weight markers are in kDa.

Fig. 2 shows accumulation of fragment C as a function of time in an experiment as described above. The fragment C levels produced under these conditions were very high, approximately 20% total cell protein (tcp), comparable to those obtained by pTETtac315 which has a shortened form of the tac promoter. As was found with all previous production of fragment C in E. coli (10,12), the recombinant protein was fully soluble.

Example 4 - Establishing DOT Conditions for Optimum Expression

As already noted, some bacterial oxygen-regulated promoters, such as the Vitreoscilla haemoglobin promoter, have a microaerobic optimum. Inspection of the Vitreoscilla promoter sequence (8), upstream of the -10 region, revealed a run of nucleotides in which 15 out of 22 matched the consensus FNR-binding site (4). Furthermore, the distance between this sequence and the -10 region is consistent with the critical distance from the putative FNR-binding site to the -10 homology sequence in the nirB promoter (6). This finding might well lead to the assumption that the nirB promoter shows a microaerobic optimum.

A comparison of part of the nirB promoter sequences in pTETnir15, F2 DR25X (6), the FNR consensus sequence (4), wild-type nirB promoter sequence (1) and the Vitreoscilla haemoglobin (Hb) promoter sequence (8) is shown below. Restriction enzyme sites and -10 homology regions and Shine-Dalgarno sequences (SD) are underlined. Conserved nucleotides of the FNR consensus sequence are shown in bold italics. EcoRI FNR KpnI

pTETnir15. . . . GAATTCAGGTAAATTTGATGTACATCAAATGGTACCCCTTGCTGAA

BamHI

F2 DR25X. . . . . . . . . . . AGGTAAATTTGATGTACATCAAATGGATCCCCTTGCTGAA

E. coli nirB . . . . . . . . . . GAATTTGATTTACATCAATAAGCGGGGTTGCTGAAT

Vitreoscilla Hb . . . . . .. .AGTTTTGATGTGGATTAAGTTTTAAGAGGCAATAAA

FNR consensus AAATTTGATATATATCAAATTT

-10 ApaI BglII SD Met.

TCGTTAAGGTAGGCGGTAGGGCCCAGATCTTAATCATCCACAGGAGACTTTCTGATG.

SD Met. . . .

TCGTTAAGGTAGGCGGTAATAGAAAAGAAATCGAGGCAAAAATG . . . .

CGTTAAGGT. .. .

GATTATAAT. . . .

Further inductions were carried out involving two different modifications to the above experimental design to investigate this possibility. In the first of these, the oxygen tension of aerobic cultures was allowed to fall to a low level by reducing the air supply to an average rate of 0.2 wm, stirrer rate reduced to 300-400 rpm and the DOT reading maintained at 2-5% above the background probe reading. This led to a much lower accumulation of fragment C, giving a maximum yield of only 6-7% tcp were obtained. In the second modification, the air supply was turned off during the logarithmic growth phase but stirring maintained, allowing the growing culture to use up the remaining dissolved oxygen. In this way, the oxygen tension was rapidly reduced to undetectable levels. Because this induction was carried out in a fermenter which was open to the atmosphere, it is possible that an immeasurably small amount of oxygen was present in the culture. Under these conditions fragment C levels of 11-12% tcp were obtained. Clearly expression levels achieved during nitrogen-sparging were significantly better than those in all other experiments in which oxygen was not rigorously excluded. It is therefore very unlikely that the nirB promoter has a microaerobic optimum but rather shows an optimum under anaerobic conditions. Example 5 - Regulation of Induction

In order to determine the degree of regulation of the nirB promoter, the background fragment C content of a pre-induction sample on a Western blot was investigated using the method described previously (15). For comparison, the uninduced level resulting from a tac-driven plasmid, pTETtac315, which expresses the fragment C gene at a similar level when induced (13) was also investigated. Using the culture conditions described above, the induction ratio of the nirB promoter closely matches that of the tightly regulated tac promoter.

Example 6 - Induction of Pertactin Synthesis bv the nirB

Promoter

Production of B. pertussis pertactin in E. coli at low to moderate levels has previously resulted in a severe reduction in growth of the culture; paradoxically this does not occur at very high expression levels (11). Because of this, a high background promoter activity would be expected to compromise growth of a pertactin producing culture prior to induction, possibly leading to plasmid loss. The expression of the pertactin gene under nirB promoter control was investigated as a further check on the regulation of the promoter. For this experiment a culture containing the plasmid pPERnir36 and nitrogen-sparged induction conditions were used. No evidence was found for any growth problems before induction and high levels of pertactin were successfully produced. Densitometer scanning of the gel showed that expression was greater than 30% tcp - the same level as has been achieved using the tac promoter (11) and probably represents the maximum achievable in this host. A culture of pPERnir36 was also induced by reducing the DOT to 2-5%. As with pPETnir215, it was found that the level of expression was reduced when compared to the level after nitrogen-sparging, and in this case led to pertactin at 17% tcp. The results of the experiments referred to above are summarised in the following Table in which IPTG denotes isopropyl-β-D-thiogalactopyranoside:

Discussion

The above Examples demonstrate that under anaerobic conditions the nirB promoter is able to express very efficiently genes for both tetanus toxin fragment C and B. pertussis pertactin. The nirB promoter produced similar levels of fragment C to a shortened version of the tac promoter. It has previously been shown that this version produces approximately 50% of the steady-state mRNA level of the full-length tac promoter (10,13). Since it has been possible to use fragment C synthesis to discriminate between promoters with activities up to that of the full-length tac promoter (13), it is likely that under the conditions described in the above Examples the nirB promoter has approximately half the activity of the full tac promoter. However, this conclusion assumes that in changing from aerobic to anaerobic culture conditions the effect on expression is only by activation of the nirB promoter. It is possible that pleiotropic changes brought about by removal of oxygen could affect other parameters of gene expression, such as plasmid copy number, mRNA or protein turnover, or efficiency of translation.

As already noted, the usefulness of a promoter for production systems depends on its regulation as well as its strength. In order to investigate the regulation of nirB, its induction ratio was investigated by measuring fragment C levels before and after induction. The ratio was found to be very similar to the value obtained for the tac promoter. These values would be expected to approximate to the corresponding mRNA ratios because, over most of this range, fragment C levels are directly proportional to mRNA levels (13). As a further check on the regulation of the nirB promoter expression of the B. pertussis pertactin gene was examined, which would be expected to be particularly sensitive to a poorly regulated promoter. Before induction, the growth rate of a culture containing pPERnir36 was comparable to growth rates obtained using pTETnir215. After induction, production of pertactin resulted in levels similar to those obtained using the tac promoter (11). Both these observations are consistent with good regulation of nirB promoter activity.

Because nirB inductions are carried out under different physiological conditions from those used in more conventional expression systems, it is possible that the nature of the product might be affected. Protein solubility in particular can sometimes be sensitive to physiological changes such as elevated temperature. It is of interest that expression of the fragment C gene under anaerobic conditions still results in a product which is fully soluble. Another possible consequence of the use of anaerobic conditions during induction might be a difference in the repertoire of proteases active in vivo. Similar patterns of lower molecular weight bands on Western blots of both C fragment and pertactin, produced under aerobic or anaerobic conditions using the nirB promoter were observed which suggests a similar range of proteolytic activity.

One other oxygen-regulated promoter that has been used for heterologous expression in E. coli is the VHb promoter referred to above (16 and WO 89/03883). This is optimally active under microaerobic conditions. In contrast, no evidence has been found for the nirB promoter requiring low levels of oxygen for maximum activity; its optimum appears to be genuinely anaerobic. This presumably relates to the fact that whereas E. coli is a facultative anaerobe, Vitreoscilla sp. is an obligate aerobe. However both promoters appear to have very similar FNR-binding sites, suggesting a common mechanism at least involving FNR. It is likely that some other feature of the Vitreoscilla promoter prevents it from functioning optimally under anaerobic conditions.

In anaerobic induction optimum is clearly advantageous since, in an actively growing fermenter culture, such conditions are much more easily maintained than microaerobic conditions. In all induction systems, separation of the growth and induction phases allows high biomass to be achieved prior to induction. It is particularly important in the case of the nirB promoter that a high oxygen tension be maintained throughout this initial growth phase, to prevent premature activation of the promoter. This may be achieved by controlling the rate of feed of carbon source to limit the growth rate, and therefore the oxygen consumption rate, of the culture. Additionally, improvements to the oxygen transfer to the culture could be made by increasing sparge and stirrer rates, and by using oxygen-enriched air or pure oxygen as the sparge gas.

REFERENCES

1. Javaraman et al, (1987), J. Mol. Biol., 196, 781-788.

2. Peakmen et al, (1990), Eur. J. Biochem., 191, 315-323.

3. Cole, (1968), Biochim. Biophys. Acta, 162, 356-368.

4. Bell et al, (1989), Nucl. Acids Res., 17, 3865-3874.

5. Javaraman et al, (1989), Nucl. Acids Res., 17, 135- 145.

6. Bell et al, (1990), Molec. Microbiol., 4, 1753-1763.

7. Ditta et al, (1987), J. Bacteriol., 169, 3217-3223.

8. Khosla & Bailey, (1989), J. Bacteriol., 171, 5995- 6004.

9. Meselson & Yuan, (1968), Nature, 217, 1110-1114.

10. Makoff et al, (1989), Nucl. Acids Res., 17, 10191- 10202.

11. Makoff et al, (1990), Bio/Technology, 8, 1030-1033.

12. Makoff et al, (1989), Bio/Technology, 7, 1043-1046.

13. Makoff & Oxer, (1991), Nucl. Acids Res., 19, 11, 2889- 14. Laemmli, (1970), Nature, 227, 680-685, ²⁸⁹²

15. Makoff & Smallwood, (1990), Nucl. Acids Res., 18,

1711-1718.

16. Khosla et al, (1990), Bio/Technology, 8, 554-558.

Claims

1. A process for the expression of a heterologous protein, which process comprises maintaining under anaerobic conditions bacteria in which the expression of the said protein is under the control of a promoter whose activity is induced by anaerobic conditions.

2. A process according to claim 1, in which the promoter is the nirB promoter.

3. A process according to claim 2, in which the promoter is that part of the nirB promoter which responds solely to anaerobiosis.

4. A process according to any one of the preceding claims, in which the bacteria are a strain of E. coli.

5. A process according to any one of the preceding claims, in which the heterologous protein is a polypeptide drug.

6. A process according to any one of claims 1 to 4, in which the heterologous protein is a polypeptide immunogen.

7. A replicable expression vector, suitable for use in bacteria, in which a DNA sequence encoding a heterologous protein is under the control of a promoter whose activity is induced by anaerobic conditions.

8. Bacteria transformed with an expression vector as defined in claim 7.

9. A DNA molecule comprising the nirB promoter sequence operably linked to a DNA sequence encoding a heterologous protein.

10. Use of a promoter activated by anaerobic conditions for the control of expression of a heterologous protein in bacteria.