WO1995021928A1

WO1995021928A1 - Process for the production of a heterologous protein from a methylotrophic yeast

Info

Publication number: WO1995021928A1
Application number: PCT/EP1995/000320
Authority: WO
Inventors: Martin Comberbach
Original assignee: Smithkline Beecham Biologicals (S.A.)
Priority date: 1994-02-15
Filing date: 1995-01-30
Publication date: 1995-08-17
Also published as: JPH09508528A; AU1536995A; ZA951131B; GB9402832D0; EP0741789A1

Abstract

A process for enhancing expression of a heterologous protein comprising culturing a methylotrophic yeast organism transformed with a vector containing DNA encoding the heterologous protein in a fermenter, characterised in that rate of MeOH addition is controlled during the induction phase of the cultivation, preferably using an automated feedback control loop system. Preferably, the concentration of MeOH in the fermenter broth, or the specific MeOH addition rate are maintained at substantially-constant levels during the induction phase.

Description

PROCESS FOR THE PRODUCTION OF A HETEROLOGOUS PROTEIN FROM A METHYLOTROPHIC YEAST.

The present invention relates to the expression of heterologous protein in methylotrophic yeasts, more particularly to a process for optimising the recovery of 5 protein expressed during methylotrophic yeast fermentations.

The development of recombinant DNA technology in recent years has led to the controlled production of a wide range of useful polypeptides by host microorganisms. The basic techniques of recombinant DNA technology, culminating in expression of a 0 desired polypeptide by a host microorganism, are now well known to those skilled in the art. Such techniques are based on the isolation of a gene encoding one or more desired polypeptides, optimally-provided with appropriate control sequences required for expression in the host microorganism. Insertion of the gene into a vector, usually a plasmid, ensures transfer of the gene into the host cell and, where autonomous 5 replication sequences are present, can give rise to multiple copies of the gene in the cell and enhancement in the expression level of the gene. Finally a suitable host microorganism is selected for transformation of the vector and expression of the information encoded by the gene.

The bacterium Escherichia coli is widely used as a host microorganism for the manufacture of heterologous proteins, for example antigens for incorporation in vaccine products. E. coli is not however an entirely satisfactory host for all such manufacture, eg. for high yield expression of protein. In addition, toxic pyrogenic factors contained in E. coli must be eliminated from any polypeptide for incorporation in a pharmaceutical product.

Attention has accordingly turned to the use of eukaryotic organisms such as yeasts as alternative hosts for the production of polypeptide products. Yeasts offer certain advantages over E. coli as hosts for the production of eukaryotic hererologous proteins. Yeasts are eukaryotic and as such their intracellular environment is more suitable for the correct folding of eukaryotic proteins. Yeasts also have the ability to glycosylate proteins, which can be important for their structural integrity, solubility and biological activity.

Bakers' yeast, Saccharomyces cerevisiae was the initial choice of yeast host. Problems have however been encountered in adapting laboratory- scale processes to an industrial scale. Second generation yeast expression systems have therefore been developed. Yeast expression systems based on methylotrophic yeasts such as Pichia pasto^ήs and Hansenula polymorpha, are particularly suitable. Methylotrophic yeasts can be cultured rapidly to high cell densities on simple, defined media using either glycerol or methanol as the sole carbon source, are easy to scale-up and are readily adaptable to continuous culture processes. Protein expression in methlyotrophic yeasts such as Pichia pastoήs is linked to the alcohol oxidase 1 gene (AOX1) which is tightly regulated by the addition of methanol (MeOH). After attaining a high cell density by fed-batch cultivation on glycerol, protein expression is induced by the addition of MeOH.

The use of methylotrophic yeasts for the production of vaccine antigens is well documented. See for example EP-A-0 173 378 (Unilever); WO 91/15571 (Wellcome Foundation); and EP-A-0 180 899, EP-A-0 226 752 and EP-A-0 341 746 (Phillips Petroleum), the subject matter of which is hereby incorporated by reference.

It is an object of the present invention to enhance the recovery of protein, in particular the recovery of soluble protein, obtained by expression in a methylotrophic yeast host organism when induced in a fermenter provided with means for controlling the host cell environment.

It is known that for optimal expression and control of the cellular environment it is preferable to use a fermenter equipped for monitoring and controlling one or more of the parameters pH, dissolved oxygen concentration, stirrer speed, temperature and aeration. In laboratory-, and production-scale fermentations, non-volatile gaseous components present in the fermenter headspace gas (for example, oxygen and carbon dioxide) are routinely measured on-line using commercially-available analysers (for example, the paramagnetic oxygen, and the near infra-red carbon dioxide analysers).

A more specific, rapid and precise analyser is the capillary inlet mass spectrometer, which in addition to measuring non-volatile components, can also be used to measure volatile components such as ethanol (EtOH) and methanol (MeOH). For fermentation products, such as the products of methylotrophic yeast expression requiring the addition of MeOH, a means of monitoring, on-line, the concentration of MeOH in the fermenter is of use, particularly if there is a fine balance between the optimal induction of product expression and toxicity to the cell from excess MeOH accumulating in the fermentation broth.

A capillary inlet mass spectrometry system allowing on-line fermenter headspace gas analysis of volatile alcohols is described by J-P. Camelbeeck et al. in Biotechnology Techniques, Vol 2(3), 183-188 (1988) and Vol 5(6), 443-448, (1991).

A method has now been developed for controlling the concentration of alcohol in the fermenter broth containing a methylotrophic yeast (Camelbeeck, J-P et al.; ICCAFT-5 & LFAC-BIO-2, Keystone, Colorado, USA, March 29- April 2, 1992). This method is based on the above-mentioned on-line analytical technique for monitoring the concentration of volatile gas in the fermenter headspace using a magnetic sector mass spectrometer. Since fermenters used to cultivate methylotrophic yeasts are usually controlled at fixed temperatures and pressures, the alcohol concentration in the off-gas is directly proportional to the concentration in the fermenter broth. It is therefore possible to feed alcohol, for example MeOH, to the fermenter in a controlled manner in order to maintain its concentration in the fermenter broth at a pre-determined setpoint. Controlled delivery is effected by means of a peristaltic feed pump which supplies MeOH from a reservoir to the fermenter via flexible translucent silicone rubber tubing. The aforementioned pump may be controlled manually with repeated visual reference to the instantaneous concentration of MeOH in the fermenter broth, or for more precise control of the MeOH concentration over longer time periods (several days), the pump may be controlled automatically, using a feedback control loop linked to the mass spectrometer through a computer running commercially-available software, designed specifically for process control (Factory Link, N3 from US Data Corp.).

Experiments with manually-, and automatically-controlled MeOH delivery during the induction phase of protein expression from methylotrophic yeasts such as P. pastoris and H. polymorpha have shown that protein yield is maximised by maintenance of the MeOH concentration in the fermenter broth at a substantially-constant level. In contrast, when the MeOH concentration in the broth is not maintained at a substantially-constant level, protein yield and recovery is significantly reduced.

Furthermore, it has been discovered that there is an optimum MeOH concentration in the fermenter broth which elicits maximum expression, recovery and solubility of a given protein.

In addition, it has been discovered that different methylotrophic yeast strains consume MeOH at different rates depending upon the MeOH concentration in the fermenter broth during induction.

Also, it has been discovered that there is an optimum specific MeOH addition rate which elicits maximum expression, recovery and solubility of a given protein.

Accordingly, the present invention provides a process for enhancing expression of a heterologous protein comprising culturing a methylotrophic yeast organism transformed with a vector containing DNA encoding the heterologous protein in a fermenter, characterised in that rate of MeOH addition is controlled during the induction phase of the cultivation, preferably using an automated feedback control loop system.

In a preferred aspect, the present invention provides a process for the production and isolation of a heterologous protein comprising culturing a methylotrophic yeast organism transformed with a vector containing DNA encoding the heterologous protein in a fermenter, characterised in that either the concentration of MeOH in the fermenter broth, or the specific MeOH addition rate are maintained at substantially- constant levels during the induction phase of the cultivation.

It will be understood that the level of constancy within which MeOH concentration or rate of addition can be controlled will be dependent on the method used for controlling MeOH concentration or addition. Thus, for manual control, a variation of up to about 8 g/L would fall within the definition of substantially contstant, whereas for automated control, for example using the above-described feedback control loop system, variation may be controlled to within tighter limits to ensure optimum protein expression and recovery. For automated control, a variation of up to 5 g/L or less, for example within 3 or even 2 g/L may be achievable.

Methylotrophic yeasts particularly suitable for use in the process of the invention include strains of Pichia, for example Pichia pastoris, and strains of Hansenula, for example Hansenula polymorpha. Suitable MeOH concentrations lie in the range from 0 to about 40 g/L in the fermenter broth.

In a preferred aspect of the invention, the MeOH concentration in the broth is maintained at a level which is substantially equal to its optimum concentration for maximum expression of the heterologous protein. Suitably, the MeOH concentration is maintained within 5 g/L of the optimum concentration and preferably within 2 g/L. For a given protein and yeast strain, the optimum MeOH concentration in the broth is readily determined by experimentation over a range of MeOH concentrations. See, for example, figures. 8, 12 and!7 of the accompanying drawings. In another preferred aspect of the invention, the specific MeOH addition rate is maintained at a level which is substantially equal to the optimum specific MeOH addition rate for maximum expression of the heterologous protein. Suitably, the MeOH addition rate is maintained within 0.04 mL/g DCW h of the optimum addition rate for mamimum expression and preferably within 0.02 mL/g DCW h. For a given protein and yeast strain, the optimum specific MeOH addition rate is readily determined by experimentation over a range of specific MeOH addition rates. See for example, figure 4 in the accompanying drawings.

The invention is illustrated by reference to the expression, extraction and quantification of the following proteins expressed in recombinant strains of Pichia pastoris :- Bordetella pertactin (69kD) and Outer Surface Protein A (OSP A) of Borrelia species. The invention is also illustrated by reference to the expression, extraction and quantification of Hepatitis B surface antigen (HBsAg) in a strain of Hansenula polymorpha.

The production of Bordetella pertactin antigens is described in International Patent Publication WO 91/15571 (Wellcome Foundation). The production of Outer Surface Protein A (OSP A) is described in International Patent Publications WO 92/00055 (Yale University) and WO 93/08299 (Connaught).

The subject matter of publication WO 91/15571, indicated above, incorporated herein by reference, describes expression vectors containing DNA sequences encoding pertactin antigens and Pichia strains transformed therewith.

Bordetella pertactin antigens include the antigens referred to as the 69kD protein from B. pertussis, the 68kD protein from B. bronchiseptica and the 70kD protein from B. parapertussis.

WO 91/15571 describes vector constructs for transformation of Pichia cells containing the promotor from the methanol-inducible AOX1 gene to drive expression of the DNA encoding pertactin antigen, in particular the vector pPIC3-60.5K. This vector digested with Bglll integrates into the host chromosomal AOX1 locus. This vector was used for intracellular expression of the 69kD protein of Bordetella pertussis in Pichia pastoris.

The resultant aoxl" transformants were screened for the presence of multiple integrated copies of pPIC3-60.5K Four multiple-copy transformants (SL3, 4, 18 and 22) were selected for further screening SL3, 18 and 22 were identified as Mut^s phenotypes or slow MeOH utilisers whilst SL4 was identified as a Mut⁺ phenotype or fast MeOH utiliser. In shake-flask inductions, expression levels (% of 69kD protein in total soluble cell protein) for the transformants SL22 (copy no. 21) and SL4 (copy no 30) of 2% and 5% respectively were obtained. Improved protein yields to 10%, were observed for the SL22 transformant after optimal fermenter inductions. No significant change in expression level was observed for the SL4 transformant. The maximum level of expression of 69kD protein from either strain was observed after only 30 hours of induction. Low-expressing transformants gave rise to protein product that was considerably more soluble (approx. 55%), under the conditions used for cell breakage, than that expressed at high levels by the multiple copy transformants SL22 and SL4. (<10% solubility).

WO 91/15571 also describes an optimised method of induction of the

69kD protein in fermenters, a slightly modified version of which was reported previously in an academic publication (Clare, J.J.,Rayment, F.B.,Ballantine, S.P., Sreekrishna, K., and Romanos, M.A. High-level expression of tetanus toxin fragment C in Pichia pastoris strains containing multiple tandem integrations of the gene, Bio/Technology, 9, 455-460, May 1991), and referred to in a subsequent academic publication (Romanos, et al, Vaccine, 9, 901-906, December 1991). This method is used for both Mut^s and Mut⁺ strains of Pichia pastoris. The MeOH is added in a pre¬ determined manner over a period of almost 100 hours following an addition profile given in figure 1. Using the data given in the above publications, and assuming a mean yeast concentration of 115 grammes of Dry Cell Weight per litre of fermenter broth, and a working volume of 1.5 litres in a 2 litre total volume fermenter, it is possible to calculate the specific MeOH addition rate (mL MeOH added/g DCW h) for the duration of the induction phase.

As described in WO 91/15571 and Romanos, et al, samples of the fermenter broth were taken during the MeOH induction phase for analysis of the intracellular concentration of 69kD protein (pertactin) by SDS-gel electrophoresis. Somewhat surprisingly, the concentration of MeOH in the fermenter broth during the induction phase was not followed, or at least not reported. No reported attempt was made, in any of the above publications, to monitor, or control the concentration of MeOH in the fermenter broth during the induction phase.

It has been shown that the pre-determined MeOH feeding profile used by Wellcome for induction of both Mut^s and Mut⁺ strains of Pichia pastoris is not optimal for either strain, being too high for SL22 (Mut^s), and too low for SL4 (Mut⁺). See for example figures 2 and 4 in the accompanying examples. In figure 2, the specific MeOH consumption rate (mL MeOH/g DCW h) is substantially equal to the specific MeOH addition rate (mL MeOH/g DCW h) as shown in figure 4, when the MeOH concentration is maintained at a substantially-constant value in the fermenter broth.

It has now been found that an optimised MeOH feed rate, controlling either the concentration of methanol or the specific methanol addition rate, during the induction phase of cultivation, permits heterologous protein expression to continue to a higher level, long after 30 hours (the duration of cultivation reported by Wellcome giving maximum expression of 69kD).

We have now carried out comparative experiments using strains of P. pastoris expressing the 69kD protein of B. pertussis (P. pastoris strains SL22 and SL4, supplied by Wellcome Laboratories, Beckenham, Kent), and using a strain of H. polymorpha (HP05) expressing HBsAg, both with and without MeOH regulation in the induction phase. We have also carried out comparative experiments using strains of P. pastoris expressing Outer Surface Protein A (OSP A) of Borrelia species (P. pastoris strain PP07, generated in-house) with MeOH regulation in the induction phase.

These experiments clearly illustrate the enhanced expression levels of protein obtained when MeOH concentrations in the fermenter broth are controlled and maintained at substantially-constant levels over extended periods of induction (up to and including 120 hours).

Furthermore, it has been found that protein induction with control of the MeOH concentration in the fermenter broth at a substantially-constant level, over extended periods of induction (up to and including 120 hours), according to the present invention, provides protein which is significantly more soluble under the same conditions for yeast cell breakage. For example, 69kD protein is obtainable which is more than 80% soluble in the buffer used for yeast cell disruption. Enhanced solubility is of particular practical advantage with respect to isolation of the protein from the disruption buffer in that the step of renaturing protein isolated as part of the pellet of insoluble matter obtained after centrifugation of the broken yeast cell suspension can be omitted. It has also been shown for slow MeOH utilising yeast strains, such as the SL22 and PP07 (all Mut^s) transformants of P. pastoris, and the HP05 (Mut^s) transformant of H. polymorpha, that there is an optimum MeOH concentration in the fermenter broth for maximum expression of heterologous protein, lying in the range 0 to 20 g/L, for example, 0.5 to 14 g/L during the induction phase of cultivation. The MeOH concentrations found to give maximum heterologous protein expression levels are :-12 g/L (P. pastoris strain SL22 expressing 69kD); 10 g/L (Hansenula polymorpha strain HP05 expressing HBsAg); 5 g/L (Pichia pastoris strain PP07 expressing OSP A).

Differences in the MeOH concentration optima for maximum heterologous protein expression differ from strain to strain and from protein to protein. This is not surprising given the phenotypic differences between strains and the different chemical nature of the heterologous proteins expressed in these strains. Thus the the optimum MeOH concentration has been shown to be dependent upon either or both the properties of the protein being expressed and the strain of methylotrophic yeast.

In the case of expression of 69kD protein in P. pastoris (SL22), the maximum expression level from a given MeOH concentration is directly proportional to the duration of MeOH induction up to and including a duration of 120 hours, whereas for expression of HBsAg in H. polymorpha (HP05), the maximum expression level from a given MeOH concentration is directly proportional to the duration of MeOH concentration only up to and including a duration of 60 hours, for reasons unknown at this time. For expression of OSP A in P. pastoris, the maximum expression level from a given MeOH concentration is directly proportional to the duration of MeOH concentration only up to and including a duration of 70-80 hours, again for reasons unknown at this time. In all cases, the optimum MeOH concentration in the broth is substantially independent of the duration of the induction phase of cultivation.

It has also been found that the specific MeOH consumption rates of Mut^s strains of Pichia pastoris expressing 69kD, OSP A and Hansenula polymorpha expressing HBsAg increase with an increasing MeOH concentration up to 20 g/L in the fermenter broth during the induction phase. However, the specific MeOH consumption rates of each strain at the MeOH concentration optimal for maximum heterologous protein expression are not the same. For example, the specific MeOH consumption rate for H. polymorpha is 3.5 fold higher than that for P. pastoris at a controlled MeOH concentration of 12 g/L. See for example figures 2 and 3 in the accompanying drawings. This discovery re-enforces the necessity for on-line measurement of the MeOH concentration and control of the specific MeOH addition rate during induction of heterologous protein expression in methylotrophic yeasts.

Enhancement of protein yield, recovery and solubility from Mut⁺ strains of methylotrophic yeasts is also a feature of the present invention. For these strains, the rate of MeOH consumption in the fermenter is so rapid, that MeOH cannot accumulate in the broth and therefore its concentration during the induction phase of cultivation is effectively zero. For these strains, optimum yield, recovery and protein solubility is achieved according to the present invention by controlling, at a constant level, the specific rate of addition of MeOH to the fermenter (mL MeOH per gramme of Dry Cell Weight per hour, abbreviated to mL MeOH/g DCW h). Suitable specific MeOH addition rates for Mut⁺ methylotrophic strains expressing 69kD protein lie in the range from 0.04 to about 0.12 mL/g DCW h.

In a preferred aspect of the invention, the specific MeOH addition rate is maintained at a substantially-constant level to maximise the expression of the heterologous protein. For a given protein and yeast strain, the optimum specific MeOH addition rate is readily determined by experimentation. See for example, figure 4 in the accompanying drawings.

It has also been found that protein induction with control of the specific MeOH addition rate between 0.04 and 0.12 mL/g DCW h over extended periods (up to and including 120 hours) provides protein which is significantly more soluble under the same conditions for yeast cell breakage. For example, 69kD protein is obtainable which is almost 70 % soluble in the buffer used for yeast cell disruption. The benefit of enhanced protein solubility is described above.

The invention is further illustrated with reference to the accompanying examples and figures.

Figure Legends

Figure 1 - A graphical representation depicting the MeOH addition rate and the specific methanol addition rate profiles for induction of 69kD protein expression in Mut^s and Mut⁺ strains of Pichia pastoris used by Wellcome (as described in WO 91/15571 and Romanos et al).

Figure 2 - A graphical representation depicting the variation of the specific MeOH consumption rate of Pichia pastoris (SL22, Mut^s) at different MeOH concentrations in the fermenter broth during induction of 69kD protein.

Figure 3 - A graphical representation depicting the variation of the specific MeOH consumption rate of Hansenula polymorpha (HP05, Mut^s) at different MeOH concentrations in the fermenter broth during induction of HBsAg.

Figure 4 - A graphical representation depicting the variation of the intracellular expression level of 69kD protein in Pichia pastoris (SL4, Mut⁺) at different mean specific methanol addition rates.

Figure 5 - A graphical representation depicting fed-batch cultivation and methanol induction of 69kD protein in Pichia pastoris (SL22) with manual control of the methanol concentration in the fermenter broth at a mean of 5 g/L (fermentation I.D. PPC11).

Figure 6 - A graphical representation depicting the on-line methanol concentration in the fermenter off-gas, as measured directly by the mass spectrometer, during the induction of 69kD protein in Pichia pastoris (SL22, Mut⁺) with automatic control of the MeOH concentration at

18 g/L in the fermenter broth (fermentation I.D. PPC34).

Figure 7 - A graphical representation depicting fed-batch cultivation and methanol induction of 69kD protein in Pichia pastoris (SL22, Mut⁺) with automatic control of the MeOH concentration at 18 g/L in the fermenter broth (fermentation I.D. PPC34).

Figure 8 - A graphical representation depicting intracellular concentrations of 69kD protein in Pichia pastoris (SL22, Mut⁺) during induction at constant MeOH concentrations.

Figure 9 - A graphical representation depicting fed-batch cultivation and MeOH induction of 69kD protein in Pichia pastoris (SL22, Mut^s) without control of the MeOH concentration in the fermenter broth (fermentation I.D. PPC13).

Figure 10 - A graphical representation depicting the on-line MeOH concentration in the fermenter off-gas, as measured directly by the mass spectrometer, during the induction of HBsAg in Hansenula polymorpha (HP05, Mut^s) with automatic control of the MeOH concentration at 16.5 g/L in the fermenter broth (fermentation I.D. C1397).

Figure 1 1 - A graphical representation depicting fed-batch cultivation and methanol induction of HBsAg in Hansenula polymorpha (HP05, Mut^s) with automatic control of the MeOH concentration at 16.5 g/L in the fermenter broth (fermentation I.D. C1397).

Figure 12 - A graphical representation depicting intracellular concentrations of HBsAg in Hansenula polymorpha (HP05, Mut^s) during induction at constant methanol concentrations.

Figure 13 - A graphical representation depicting the fed-batch cultivation and MeOH induction of HBsAg in Hansenula polymorpha (HP05, Mut^s) without control of the MeOH concentration in the fermenter broth (fermentation I.D. C1094).

Figure 14 - A graphical representation depicting the fed-batch cultivation and MeOH induction of 69kD protein in Pichia pastoris (SL4, Mut⁺) during induction with manual control of the MeOH addition at a substantially-constant specific MeOH addition rate (fermentation I.D. PPC10).

Figure 15 - A graphical representation depicting the on-line methanol concentration in the fermenter off-gas, as measured directly by the mass spectrometer, during the induction of OSP A protein in Pichia pastoris (PP07, Mut^s) with automatic control of the MeOH concentration at 5 g/L in the fermenter broth (fermentation I.D. PPL 16).

Figure 16 - A graphical representation depicting fed-batch cultivation and methanol induction of OSP A protein in Pichia pastoris (PP07, Mut^s) with automatic control of the MeOH concentration at 5 g/L in the fermenter broth (fermentation I.D. PPL 16).

Figure 17 - A graphical representation depicting intracellular concentrations of OSP A in Pichia pastoris (PP07) during induction at constant methanol concentrations.

Example 1 On-line measurement of the methanol concentration in the fermenter broth during heterologous protein induction by capillary inlet mass spectrometry with subsequent continuous feedback control.

A magnetic sector mass spectrometer, fitted with a heated capillary inlet, was used to follow the concentration of alcohols in the off-gas of a fermenter containing weighed solutions of EtOH and MeOH in water (Camelbeeck, J-P, et al; Biotechnol Techniques, 2(3), 183-188, 1988)

Pulse-response experiments were performed (Orval, M., ISIL Thesis, Liege, Belgium, 1989) to confirm the utility of this analytical technique for on-line analysis of solvents maintained under various physico-chemical conditions (pH, temp., pressure, etc.).

The total system response time (about 6 minutes) was found to be largely dependent on the response time of the mass spectrometer and the molecular equilibrium within the tube transporting the off-gas to the mass spectrometer (Camelbeeck, J-P, et al.; Biotechnol. Techniques, 5(6), 443-448, 1991). The short response time coupled with the precision of the technique met the requirements for continuous feedback control of alcohol concentrations in fermenters. Continuous feedback control of the MeOH concentration in the fermenter broth was performed either manually, by frequent (at least once every 2 hours) manual adjustment of the MeOH feed pump, with repeated visual reference to the MeOH concentration given by the mass spectrometer, or automatically, by linking the mass spectrometer signal output to a desktop 386 PC. A commercially-available software package (Factorylink) running simple process control algorithms was used to regulate the MeOH feed pump depending upon the setpoint concentration of MeOH desired in the fermenter (Camelbeeck, J-P et al.; ICCAFT-5 & LFAC-BIO-2, Keystone, Colorado, USA, March 29-April 2, 1992).

Example 2

Cultivation method for the expression of 69RD, OSP A and HBsAg with continuous feedback control of the methanol concentration in the fermenter broth.

The SL22 strain of Pichia pastoris (supplied by Wellcome Laboratories) having 21 copies of the 69kD gene incorporated into the genome, the PP07 strain of Pichia pastoris, each possessing phenotype Mut^s and the HP05 strain of Hansenula polymorpha, also possessing phenotype Mut^s, were cultured using the following growth/induction method:

1) Growth to high cell density (60-150 g (Dry Cell Weight)/L) on a glycerol/salts/vitamin medium in fed-batch mode at 30°C, pH 5.0 for ± 40 h.

2) Induction with a solution of micro-elements dissolved in MeOH, using a controlled-feeding rate to maintain the MeOH concentration at a fixed concentration (from 0 to 40 g/L in the fermenter (using the on-line technique described in Example 1). The period of induction was prolonged to maximise the production of :-69kD protein (up to 120 hours' duration); OSP A (up to 80 hours' duration); and HBsAg (up to 100 hours' duration).

3) Samples of the yeast taken from the fermenter at regular intervals during the induction phase were broken with a French press to extract and solubilise the 69kD, OSP A and HBsAg. The intracellular concentrations of 69kD and OSP A were determined by ELISA and expressed as either the fraction of protein present in the total soluble protein released into suspension (69kD), or the fraction of protein present in the dry cell matter (OSP A). The intracellular concentration of HBsAg was determined by a commercially-obtainable AUSRIA test kit (Abbott Laboratories), and expressed as the fraction of HBsAg present in the dry cell matter.

Plots of the MeOH concentration against time showing the controlled concentrations of MeOH during the inductions of 69kD protein, HBsAg and OSP A are given in figures 6, 10 and 15 respectively. The expressions of 69kD protein, HBsAg and OSP A from these cultivations are given in figures 7, 1 1 and 16 respectively. Example 3 Determination of the optimum methanol concentration for maximum expression of 69kD protein from Pichia pastoris (SL22).

Cultivations using the methodology described in Example 2 were carried out at 6 different methanol concentrations for induction periods of 48, 72, 96 and 120 hours' duration. The results (figure 8) show that for P. pastoris SL22 expressing 69kD protein there is an optimum methanol concentration for maximum expression at approximately 12 g/L.

Example 4

Determination of the optimum methanol concentration for maximum expression of OSP A from Pichia pastoris (PP07).

Cultivations using the methodology described in Example 2 were carried out at 10 different methanol concentrations for an induction period of 70-80 hours' duration. The results (figure 17) show that for P. pastoris PP07 expressing OSPA there is an optimum methanol concentration for maximum expression at approximately 5 g/L.

Example 5

Determination of the optimum methanol concentration for maximum expression of HBsAg from Hansenula polymorpha (HP05). Cultivations using the methodology described in Example 2 were carried out at 7 different methanol concentrations for induction periods of 20, 40, 60, 80 and 100 hours' duration. The results (figure 12) show that for H. polymorpha HP05 expressing HBsAg protein there is an optimum methanol concentration for maximum expression at approximately 10 g/L.

Example 6

Cultivation method for the expression of 69kD protein and HBsAg without continuous feedback control of the methanol concentration. The SL22 strain of P. pastoris and the HP05 strain of H. polymorpha were cultured using the method of Example 2 except that the MeOH feed pumps were adjusted manually only once per day with visual reference to the on-line MeOH concentration given by the mass spectrometer. This daily adjustment was clearly insufficient to maintain the methanol concentration in the fermenter broth at a substantially-constant level, due to the rapidly-changing metabolic conditions. Wide oscillations of the methanol concentration in the fermenter broth were therefore observed, which were substantially different from the optimum methanol concentrations given in Examples 3 and 4. These oscillations and the substantially-reduced expression levels of 69kD protein and HBsAg obtained as a result of the absence of continuous feedback control of the methanol concentration in the fermenter broth are given in examples shown in Figures 9 and 13.

Example 7

Cultivation method for the expression of 69kD protein with manual control of the specific methanol addition rate.

The SL4 strain of Pichia pastoris (supplied by Wellcome Laboratories) having 30 copies of the 69kD gene incorporated into the genome and possessing phenotype Mut⁺, was cultured using the following growth/induction method:

1) Growth to high cell density (60-150g (dry weight)/L) on a glycerol/salts/vitamin medium in fed-batch mode at 30°C, pH 5.0 for ± 40 h.

2) Induction with a solution of micro-elements dissolved in MeOH, using a manually-controlled feeding rate to maintain the specific rate of addition of MeOH to the fermenter (mL MeOH per gramme of Dry Cell Weight per hour, abbreviated to mL MeOH/g DCW h) at a substantially-constant level. The value to which the pump was adjusted, was calculated off-line, by repeated reference to a calibrated reservoir containing the MeOH feed solution, and the yeast concentration (gramme of Dry Cell Weight per litre of fermenter broth, abbreviated to g DCW/L), obtained by sampling of the fermenter broth, on average, each 10 hours during the induction phase The period of induction was prolonged to maximise the production of 69kD protein (up to 120 hours'duration).

3) Samples of the yeast taken as above were broken with a French press to extract and solubilise the 69kD protein. The intracellular concentration of 69kD protein was determined by ELISA and expressed as the fraction of 69kD present in the total soluble protein released into suspension.

An example of a cultivation in which this method was used is shown in figure 14.

Example 8

Determination of the optimum specific methanol addition rate for maximum expression of 69kD protein from Pichia pastoris (SL4).

Cultivations using the methodology descibed in Example 7 were carried out at 8 different mean specific MeOH addition rates for a mean induction period of 70 hours' duration. The results (figure 4) show that for P. pastoris (SL4) expressing 69kD protein, there is an optimum mean specific MeOH addition rate for maximum expression at approximately 0.09 mL/g DCW h.

Example 9 Determination of the solubility of 69kD protein

Cultivations using the methodologies described in Examples 7 (Method A), and 2 (Method B) were carried out using P. pastoris SL4 and SL22 respectively. Examples of these cultivations are PPC10 (figure 14), and PPC11 (figure 5). For induction of 69kD protein in P. pastoris SL4, the mean specific MeOH addition rate was controlled manually at 0.04 mL/g DCW h (PPC10), and for induction of 69kD protein in P. pastoris SL22, the methanol concentration was controlled manually at a mean of 5 g/L (PPC11).

Samples of yeast taken from the fermenters at the end of the induction period were centrifuged and stored at -30 °C for one week before being resuspended to an optical density of 40 OD₆₂o in the following buffer-- TRIS-HC1 (50 mM), pH 8.0, NaCl (0.1 M), EDTA(lmM), isopropanol (5%)

A frozen sample of P. pastoris (SL4) received from Wellcome pic, taken from the end of cultivation and methanol induction, performed using the protocol described by Clare, J.J. et al. in Bio/Technology, 9, 455, 1991 (Method C) was treated in exactly the same manner.

The yeast cells were disrupted by 4 passes through the French press at 20,000 psig. The resulting solution was centrifuged at 40,000 x g for 15 minutes, and the supernatant was decanted off and stored at 4 °C before being analysed for the presence of soluble 69kD (by ELISA) and the total soluble protein concentration (Lowry).

The pellet was resuspended in the above buffer containing guanidine-HCl (6M) using a vortex mixer for several minutes at room temperature. After a further centrifugation step, the supernatant (containing the remaining solubilised protein) was decanted off and stored at 4 °C before being analysed as above.

The % solubility of 69kD was calculated by assuming that suspension of the pellet in the presence of guanidine-HCl extracted 100 % of total recoverable protein, whereas disruption with the French press in the presence of TRIS buffer extracted only the soluble proteins.

The results of the analyses can be seen in the following table (Table 1) :

Table 1

The results show that the 69kD protein expressed under the cultivation conditions established above (Methods A and B) is more soluble than the same protein expressed in the same organism cultivated under the prior art conditions (Method C). Since the only difference between the samples was the method of induction, it is clear that the conditions established in the present invention increased the % of recoverable 69kD protein by a significant amount (2.5 - 3.0 times).

Claims

1. A process for enhancing expression of a heterologous protein comprising culturing a methylotrophic yeast organism transformed with a vector containing DNA encoding the heterologous protein in a fermenter, characterised in that rate of MeOH addition is controlled during the induction phase of the cultivation using an automated feedback control loop system.

2. A process for the production and isolation of a heterologous protein comprising culturing a methylotrophic yeast organism transformed with a vector containing DNA encoding the heterologous protein in a fermenter, characterised in that the concentration of MeOH in the fermenter broth, or the specific MeOH addition rate are maintained at substantially-constant levels during the induction phase of the cultivation.

3. A process as claimed in claim 1 or 2 wherein the transformed methylotrophic yeast organism is a strain of Pichia pastoris or Hansenula polymorpha.

4. A process as claimed in claim 3 wherein the transformed methylotrophic yeast organism is a strain of Pichia pastoris expressing the 69kD protein of Bordetella pertussis or Outer Membrane Protein A (OSP A) of Borrelia species or a strain of Hansenula polymorpha expressing Hepatitis B surface antigen (HBsAg).

5. A process as claimed in any one of claims 1 to 4 wherein the concentration of MeOH in the fermenter broth is in the range from 0 to 40 g/L.

6. A process as claimed in any one of claims 1 to 5 wherein the MeOH concentration in the fermenter broth is substantially equal to the optimum concentration for maximum expression of the heterologous protein.

7. A process as claimed in any one of claims 1 to 5 wherein the MeOH addition rate is substantially equal to the optimum addition rate for maximum expression of the heterologous protein.

8. 69kD protein from Bordetella pertussis obtainable by the process of claim 1 which protein is from 65 to 85% soluble in the buffer used to disrupt the yeast cell.