WO2003062431A2

WO2003062431A2 - Production of epi-hne-4 comprising reduction of improperly processed form

Info

Publication number: WO2003062431A2
Application number: PCT/EP2003/001212
Authority: WO
Inventors: Christine Gerard; Alain Poncin; François Saudubray; Jean-Pierre Petchot-Bacque
Original assignee: Debiopharm Sa
Priority date: 2002-01-23
Filing date: 2003-01-23
Publication date: 2003-07-31
Also published as: WO2003062431A3; AU2003210220A1

Abstract

The present invention concerns novel nucleotide sequences, gene constructs, expression vectors and micro-organisms for secreting a protein, in particular an EPI-hNE protein such as EPI-hNE-4, with very low levels, or undetectable levels, of an improperly processed form of said protein which contains 9 extra amino-terminal amino acids as compared to the correctly processed form.

Description

NOVEL NUCLEOTIDE SEQUENCES, GENE CONSTRUCTS, EXPRESSION

VECTORS AND MICRO-ORGANISMS FOR SECRETING A PROTEIN SUCH AS

EPI-HNE-4, AND EPI-HNE-4 PROTEIN OBTAINED THEREFROM

The present invention concerns novel nucleotide sequences, gene constructs, expression vectors and microorganisms which have been conceived in such a manner as to allow the secretion of a protein, in particular an EPI-hNE protein such as EPI-hNE-4, with very low levels, or undetectable levels, of an improperly processed form of said protein which contains 9 extra amino-terminal amino acids as compared to the correctly processed form. The improperly processed form of EPI-hNE-4 may hereafter be referred to as ^λΕPI-hNE-4 (n+9) ", or as the ^,x (n+9) form of EPI-hNE-4" or as the "mis-processed form of EPI-hNE-4" . These expressions are all equivalent and are used indifferently herein to refer to the improperly processed form of EPI-hNE-4.

International Patent Application WO 96/20278 to Ley et al . describes a number of genetically engineered novel proteins which inhibit human neutrophil elastase (hNE) . As indicated in the above-cited patent application, human neutrophil elastase (also known as human leukocyte elastase) is one of the major neutral proteases of the azurophil granules of polymorphonuclear leukocytes. This enzyme is involved in the elimination of pathogens, and in connective tissue restructuring.

The principal systemic inhibitor of hNE is the α-1- protease inhibitor-, formerly known as αl antitrypsin. In a certain number of pathological situations (hereditary disorders, emphysema, cystic fibrosis, Acute Respiratory Distress Syndrome (ARDS) , Chronic Obstructive Pulmonary Disease (COPD) ) , this inhibitor is either not present in sufficient amounts in the bloodstream or is inactivated, leading to uncontrolled elastolytic activity of hNE, which causes extensive destruction of lung tissue.

WO 96/20278 thus proposes novel proteins which are stable, non-toxic, highly efficient inhibitors of hNE. These inhibitors, referred to as EPI-hNE proteins, are members of a group of inhibitors derived from a Kunitz-type inhibitory domain found in bovine pancreatic trypsin inhibitor (BPTI or a protein of human origin, namely the light chain of human Inter-α-trypsin inhibitor (ITI) . They are, inter alia, EPI- hNE-1, EPI-hNE-2, EPI-hNE-3 and EPI-hNE-4. The inhibitors of WO 96/20278 are produced by fermentation of transformed strains of the yeast Pichia pastoris into which modified DNA sequences of these proteins have been inserted. The modifications concern the biological Kunitz domains, and render these proteins- highly potent. EPI-hNE-3 is described as being secreted partially as a (n + 9) mis-processed form.

These proteins may be used to treat disease conditions due to an excessive activity of hNE, in particular respiratory disorders, cystic fibrosis, emphysema, acute respiratory distress syndrome or chronic obstructive pulmonary disease.

One of these inhibitors, EPI-hNE-4, is of particular interest.

However, the prior art strain which produces this protein, namely strain PEY-53, also produces an incorrectly processed form of EPI-hNE-4 which contains an extra nine residues at the amino terminal of the protein, apparently due to an incorrect cleavage at the moment of post-translational maturation.

The P. pastoris strain PEY-53, which produces EPI-hNE-4, but also produces EPI-hNE-4 (n+9) , contains an expression vector which contains a fused gene construct wherein the Saccharo yces cerevisiae mating factor α prepropeptide is directly upstream of the 5' terminus of the DNA coding for EPI-hNE-4. The (n+9) form which is produced by PEY-53 is the result of an incorrect cleavage of the Saccharomyces cerevisiae α mating factor prepropeptide from the mature Kunitz domain.

Purification of the correctly processed form of EPI-hNE-4 ϊrom the fermentation medium is rendered more difficult due to the presence of the significant amounts of the mis-processed form. It is difficult and costly to completely separate the two forms by conventional chemical/physical separation methods so as to obtain EPI-hNE-4 substantially free of the (n+9) form.

While WO 96/20278 is of undeniable utility for producing EPI-hNE-4 which can be purified on a small scale substantially free of the (n+9) form, it would be desirable to avoid production of that mis-processed form.

The Applicant company has filed a series of European Patent Applications : EP 00203049, EP 00403029 and EP 01202655, as well as an International Patent Application WO 01/10727 claiming priority from these three European Patent Applications, which describe improved methods for purifying EPI-hNE-4 which is produced by fermentation of transformed strains of P. pastoris such as PEY-53 and which is contaminated with large amounts of the (n+9) form.

Even after purification by the improved methods described in the above-mentioned patent applications, it is still possible to detect about 2% of the mis-processed form in the purified EPI-hNE-4.

It would thus be very desirable to obtain a transformed micro-organism which would produce EPI-hNE-4 without, or at least with undetectable levels of, the EPI-hNE-4 (n+9) form, thus facilitating the further purification steps and hence resulting in economies both on a financial as well as on a time-spent scale.

The Applicant company has dedicated much time and effort to the development of such a new micro-organism. The Applicant company has succeeded in constructing a P. pastoris strain closely related to the P. pastoris strain PEY-53, which produces EPI-hNE-4 but which also produces EPI- hNE-4 (n+9) , in such a manner that the new modified strain according to the present invention produces no, or undetectable levels of, EPI-hNE- (n+9) . This modified strain is called PE76A.

In a surprising and unexpected manner, the new microorganism according to the invention not only produces EPI-hNE- 4 without detectable levels of the mis-processed form, but it also allows better results with respect to the purification of EPI-hNE-4 from its conditioned fermentation medium. The Applicant company believes, without being bound by this theory, that the absence (or at any rate very low levels) of the mis-processed form somehow contributes to better separation of EPI-hNE-4 from other contaminants as well.

The invention thus provides a eukaryotic micro-organism which produces into the conditioned fermentation medium a protein apt to be secreted as a (n + 9) mis-processed form, with no or very low (trace) levels of said mis-processed form. The absence or very low levels of the mis-processed form may facilitate to a considerable extent the purification of the correctly processed protein.

The invention is as defined in the appended claims.

In the new microoganism according to the invention, the nucleotide sequence at the (n+9) cleavage site of the fused gene ( mating factor prepropeptide ligated to the 5' terminal of the coding sequence for EPI-hNE-4) has been modified insofar as the codon for Lysine, the amino acid in position 76, has been modified to code for Alanine.

Briefly, strain PE76A was produced as follows: Total genomic DNA of PEY-53 was extracted and a BstBI- EcoRI fragment was amplified by PCR in such a manner that the amplified fragment contains the S. cerevisiae α mating factor prepropeptide sequence fused with the gene encoding EPI-hNE-4. The fragment was cloned into an intermediate vector, PCR .1. The sequence was verified and then site-directed mutagenesis was performed in order to replace the codon for Lysine at position 76 upstream of the N-terminal cleavage site of EPI- hNE-4, by a codon for Alanine. Lysine is a basic residue which is often contained in the recognition sites of endopeptidases . By replacing this basic amino acid residue by a neutral residue (Alanine) , the Applicant company hoped to reduce the extent of incorrect cleavage.

It will of course be understood that the choice of the host micro-organism is not necessarily a limiting factor as far as the invention is concerned. The host strain into which an expression vector containing the modified EPI-hNE-4 sequence is cloned, can be selected from among suitable strains of any eukaryotic microorganisms transformed by a vector for the expression of said proteins. Preferably, the host strain is yeast such as Saccharomyces cerevisiae, Schizosaccharomyces pombe, Yarrovia lipili tica, Hansenula polymorpha, Kl yveromyces lactis, Pichia methanolica and Pichia pastoris, which are capable of expressing high levels of a recombinant protein. Even more preferably, the host strain is Pichia pastoris. The present invention thus concerns a nucleotide sequence coding for the Saccharomyces cerevisiae α mating factor prepropeptide or a variant thereof, which is modified so as to replace the Lysine residue normally present in position 76 of the amino acid sequence of the naturally occurring prepropeptide by an Alanine residue (Figure 1A indicates the amino acid sequence of the naturally occurring prepropeptide; Figure IB provides the nucleic acid and amino acid sequence of the modified prepropeptide according to the invention. The nucleotide and amino acid sequences of this modified peptide are also provided as SEQ ID n°l and 2, and in Figures 7 and 8 respectively) . In the context of the present invention, a variant is to be understood as a natural or synthetic variant of the naturally occurring prepropeptide by substitution of 1 to 5, preferably 1 or 2 amino acids of the naturally occurring sequence by another amino acid which is equivalent, i.e. does not modify the prepro function of that peptide. The present invention also concerns a nucleotide sequence coding for a modified Saccharomyces cerevisiae α mating factor prepropeptide as represented by SEQ ID n° 1 and an amino acid sequence as in SEQ ID n° 2.

The invention further concerns a fused gene construct comprising the above defined nucleotide sequence ligated to the 5' terminal of a nucleotide sequence coding for a protein apt to produce a (n + 9) mis-processed form. Said protein is preferably an EPI-hNE protein, and even more preferably is EPI-hNE-4. A preferred fused gene construct according to the invention is that represented in SEQ ID n°3 (and in Figure 9) . The corresponding amino acid sequence is defined in SEQ ID n° 4 (and in Figure 10) .

The invention further concerns expression vectors comprising the above-mentioned fused gene constructs. Preferably, the expression vector is selected from the group consisting of a vector derived from Saccharomyces cerevisiae such as pESC (Stratagene) , pYES (Invitrogen) , pRS, pYS, a vector derived from Pichia methanolica such as pMET, and a vector derived from Pichia pastoris such as pPIC and pGAP (Invitrogen) , preferably pPIC.

The invention further concerns a eukaryotic micro-organism comprising a fused gene construct or transformed by an expression vector as described above. Preferably the microorganism is yeast such as Saccharomyces cerevisiae, Schizo saccharomyces po be, Yarrovia lipilitica, Hansenula polymorpha, Kluyveromyces lactis, Pichia methanolica and Pichia pastoris, even more preferably being Pichia pastoris .

The micro-organism according to the invention produces EPI- hNE-4 in a fermentation medium with no detectable levels of EPI-hNE-4 (n+9) using Western Blot analysis.

The content of EPI-hNE-4 (n+9) in the fermentation medium of the micro-organism according to the present invention is less than 2 % compared to total EPI-hNE-4, as determined by RP-HPLC analysis .

The invention further concerns an EPI-hNE-4 protein containing no detectable levels of EPI-hNE-4 (n+9) when analysed using Western Blot techniques.

In a preferred embodiment, the EPI-hNE-4 protein according to the present invention contains less than 1.5 %, preferably less than 1 %, and even more preferably less than 0.5% of EPI-hNE-4 (n+9) compared to total EPI-hNE-4, as determined by RP-HPLC analysis.

The invention will be further illustrated by the following examples which are non-limitative.

EXAMPLE 1 ; Pichia pastoris strain PEY-53 producing EPI-hNE-4, contaminated with the (n+9) form. In a manner analogous to that described in WO 96/20278 to Ley et al . for EPI-hNE2, EPI-hNE3 and EPI-hNE-4, a strain of Pichia pastoris is transformed in order to allow it to specifically produce EPI-hNE-4. This strain is named PEY-53 and is a derivative of Pichia pastoris strain GS115 (also known as GTS115) . P. pastoris GS11Ξ {his 4) is obtained by nitrosoguanidine mutagenesis of P. pastoris wild type strain NRRL Y-11430.

The GS115 strain has a defect in the histidinol dehydrogenase activity coded for by the gene HIS4. The strain can grow on histidine-supplemented media or on complex media such as YEPD, but ,not on minimal media such as MD. The strain is stable (spontaneous reversion of GS115 to His+ phenotype is less than 10^~8) . PEY-53 was created by transforming GS115 with an integrating expression plasmid, pPIC-EPI-hNE-4. The expression plasmid was constructed by ligating synthetic DNA sequences encoding the Saccharomyces cerevisiae α mating factor prepropeptide directly to the 5' terminus of synthetic DNA encoding EPI-hNE-4. This fusion gene sequence is sandwiched between an upstream indueible P. pastoris AOX1 gene promoter and downstream AOX1 gene transcription termination and polyadenylation sequences in a derivative of the plasmid pHIL- D2 which also encodes a S. cerevisiae HIS4 gene. Linearized pPIC-EPI-hNE-4 DNA was incorporated by homologous recombination into the genome of GS115 by spheroplast transformation. Regenerated spheroplasts were selected for growth in the absence of added histidine and individual isolates were screened for methanol utilisation phenotype (iτ!ut+) , secretion levels and gene copy number. PEY-53 is a high level production strain which, in high-density mixed-feed fermentation, secretes EPI-hNE-4 into the medium at final levels of between 100 and 300 mg/1.

Southern Analysis of PEY-53 genomic DNA shows this strain contains 3 to 4 copies of the expression plasmid integrated at the AOX1 gene locus.

RP-HPLC performed using the test described in Example 3 showed that this strain produces about 10 to 20 % of the (n+9) form of EPI-hNE-4.

EXAMPLE 2 : New Pichia pastoris strain PE76A according to the invention, which produces EPI-hNE-4 with undetectable levels of the (n+9) form.

This example demonstrates the construction of a new strain of Pichia pastoris, PE76A, which allows the production of EPI-hNE-4, but does not produce detectable amounts of the (n+9) form of EPI-hNE-4.

An expression vector, pPIC-EPI-hNE-4-K76A was constructed as follows: The prepro-EPI-hNE-4 sequence was amplified by PCR from total DNA extracted from the strain Pichia pastoris PEY-53, using the nucleotide primers below:

5' primer: 5' - TTC GAA ACG ATG AGA TTC CCA TC - 3 ' BstBI

3' primer: 5' - GAA TTC TTA CTA TGG AAC ACC ACA G - 3

EcoRI

A 444 bp amplified fragment was cloned in the intermediate vector pCR2.1 (commercialized by Invitrogen) and the nucleotide sequence of three positive clones was verified by single strand sequencing. A clone having no sequence ambiguity was selected for site-directed mutagenesis.

Mutagenesis of the nucleotide sequence corresponding to the (n+9) cleavage site was performed using a "Quick Change Site-Directed Mutagenesis" kit from Stratagene. The protocol followed was according to the manufacturer's instructions. The oligonucleotides necessary for producing the mutant K76A were selected according to the manufacturer's instructions and were as follows:

Primer K76Aa: 5' - GCT TCT ATC GCT GCT GCT GAG GAA GGT GTT TCC - 3'

Primer K76Ab: 5' - GGA AAC ACC TTC CTC AGC AGC AGC GAT AGA AGC - 3'

Restriction enzyme digestion with BstBI and EcoRI produced a BstBI -EcoRI (see Figure IB) fragment containing the modification of a Lysine-encoding codon to an Alanine-encoding codon which was completely sequenced and then cloned into the expression vector pPIC9 (Invitrogen) between Bstl (934) and EcoRI (1223) (see Figure 2) . Since this vector has two BstBI sites, it was first linearized using EcoRI and then partially digested with BstBI. A positive clone was selected and the final construct, pPIC-EPIhNE-4K76A was verified by single strand sequencing. The P. pastoris strain GS115 was transformed by electroporation with the expression vector pPIC-EPIhNE-4K76A linearized using Sacl . The transformants were selected on Histidine-free MD medium. 20 clones were thus selected. The clone No. 5 hereafter referred to as strain PE76A, showed a high level of expression of EPI-hNE-4, as determined by SDS PAGE and anti-EPI-hNE-4 Western Blot. EXAMPLE 3: COMPARISON OF PRODUCTION OF EPI-hNE-4 and EPI-hNE-

4 (n+9) BY PEY-53 (prior art strain) AND PE76A (strain according to the invention)

These strains were then put into culture in Erlenmeyers and fermentors so as to measure their levels of expression of EPI-hNE-4 and of the (n+9) form of EPI-hNE-4. 1) Culture in erlenmeyers

The cell cultures were performed at 30°C in Erlenmeyer flasks fitted with a retort containing 500 ml of YSG+ medium

(non buffered medium containing 6 g/1 yeast extract, 5g/l soya peptone, 20g/l of glycerol) . After 24 hours of culture, the cells were recovered by centrifugation and then resuspended in 150 ml of YSG- medium (non buffered glycerol-free medium containing ' 6 g/1 yeast extract, 5g/l soya peptone) . The expression was induced by addition of methanol (750 μl, final concentration 0.5%) at 0 hours (tO) , 24 hours and 48 hours after cell resuspension. The pH was adjusted during cell growth by the addition of phosphoric acid when the medium became too alkaline. Samples were taken just before induction of cultures at (tO) , and 24, 48 and 72 hours after induction. The optical density (OD) at 600 nm and the pH of the cultures were measured. The results are shown in Table 1 below:

TABLE 1

Note.- at tO, the OD was measured before centrifugation and resuspension.

The supernatants recovered after 72 hours of culture were :

(i) directly concentrated 100-fold by TCA precipitation; or (ii) purified and concentrated 10-fold on Macroprep highS columns (commercialized by BioRad) ; or (iii) purified and concentrated 10-fold on Macroprep HighS columns and further concentrated 15-fold by TCA precipitation. The various samples were then analyzed by SDS-PAGE (Tris- glycine 18%), anti-EPI-hNE-4 Western Blot and anti-EPI-hNE-4 (n+9) Western Blot. These Western Blots were performed using polyclonal antibodies raised in rabbits against EPI-hNE-4 and the (n + 9) peptide coupled to KLH (Keyhole Limpet Hemocyanin) , respectively.

The results demonstrate that the transformed strain PE76A produces a level of EPI-hNE-4 comparable to PEY-53, but without the presence of the (n+9) form of EPI-hNE-4. The latter was clearly detected on both SDS-PAGE and anti-EPI-hNE-4 (n+9) Western Blot for strain PEY-53 and was not detectable for strain PE76A.

2) Culture in fermentor

The fermentation process including the fermentation mineral medium was performed according to the «Pichia Fermentation Process Guidelines » from Invitrogen. The fermentation was divided in four phases:

• Seed culture in shake-flask = inoculum of the fermentation • Fermentation Batch with glycerol = production of biomass

• Fed-batch with glycerol = production of biomass

• Fed-batch with methanol = induction of expression of the recombinant protein. The temperature was set to 30 °C. The pH was set to 5.0 during the. growth on glycerol . At the beginning of the methanol induction the pH set point was shifted to 3.0.

For the strain PEY-53 (50 1 scale fermentation) the glycerol was added by fed-batch for 4 hours . The OD_60o before induction of the expression of EPI-hNE4 was 110. After 68h of methanol induction, the final optical density was 270.

For the strain PE76A (5 1 scale fermentation) the glycerol was added by fed-batch during 24 hours. The OD_S00 before induction of the expression of EPI-hNE4 was 350. After 66 hours of methanol induction, the final optical density was 450.

The RP-HPLC test consists in passing lOOμl of sample on a silica column Licrosphere 100RP from Merck: gradient of water

+ 1 % TFA and acetonitrile + 1 % TFA, with detection at 280 nm and 220 nm (gradient : 0-20%B in 3 min, 20-70%B in 17 min and 100%B in 3 min) .

The HPLC system was an HP1100 Agilent equiped with a diode array detector.

EPI-hNE-4 is eluted in 3 closed peaks : mis-processed (n+9) form at a retention time of 13.2 min, EPI-hNE-4 (main peak) at 13.5 min and a pyroglutamic form at 13.8 min.

The quantitation of the three forms is obtained by integration of the signal at 220 nm, taking as baseline valley to valley integration.

RP-HPLC with detection at 220 nm showed about 11.2 % EPI- hNE-4 (n ₊ 9) compared to total EPI-hNE-4 (total of the 3 forms) for strain PEY-53 (see Figure 3) , and less than about 2 % EPI-hNE-4 (n + 9) compared to total EPI-hNE-4 for strain PE76A (see Figure 4) .

EXAMPLE 4; Purification of EPI-hNE-4 produced by the prior art strain PEY-53

The procedure described in international application WO 01/10727, Example 1, which gives the lowest percentage of the (n+9) form of EPI-hNE-4 in the purified protein, was reproduced.

100 1 of the PEY-53 culture medium obtained as described in Example 3 (2) above were collected and passed over an expanded bed as follows: 10 1 of chromatographic matrix (Streamline SP from Amersham-Pharmacia) was equilibrated in 50 mM ammonium acetate pH 3.5 and fluidized in the same buffer to 30 1 at 300 cm/h. After loading, the column was washed in the 10 mM ammonium acetate pH 3.5 to obtain an absorption at 280 nm below 0.05. The beads were packed to 10 1 and EPI-hNE-4 was recovered by washing the column in 1 M ammonium acetate pH 4.5 buffer.

Thus was obtained a 10 1 solution containing about 60 g of proteins and pigments (as determined by spectrometric assay at 280 nm and by Coomassie protein assay) . RP-HPLC (performed as described in Example 3 but with integration of the signal at 280 nm) showed that the amount of EPI-HNE-4 was only about

30 g and the mis-processed form was not separated from the main form. Furthermore, contamination by green pigments was detectable.

The solution was sterile-filtered on a 22 μm filter (Millipack 200 from Millipore) before further purification.

Hydrophobic interaction chromatography was conducted by passing the above 10 1 solution on a BioProcess (Pharmacia) system, using a phenyl-sepharose Fast Flow matrix from Pharmacia in a 15 1 BPTG column from Pharmacia. The buffers used were A: sodium acetate 50 mM pH 4.5 + 1M NaCl, and B: sodium acetate 50 mM pH 4.5. The elution was performed by one step at 100 % B with a flow rate of 300 cm/h.

The eluate contained about 15 g of purified EPI-hNE-4 (as determined by spectrometric assay at 280 nm, Coomassie protein assay and biological activity assay) .

RP-HPLC showed that the mis-processed form was not separated from the correctly processed form of EPI-hNE-4.

Cation exchange chromatography was then performed using a Bioprocess chromatographic system from Pharmacia. The matrix used was Macroprep High S matrix from BioRad (rigid matrix based on cross-linked methacrylate carrying sulphonate surface groups) , in a 15 1 BPG200 column from Pharmacia. The buffers used were A: ammonium acetate lOmM pH 3.5, B sodium acetate 50 mM pH 6.2 and C: lOmM ammonium bicarbonate pH 7.8. A first elution in buffer B was used to separate the mis-processed form. Elution was then performed by one step at 100%C with a flow rate of 300 cm/h.

The eluate contained about 12 g of purified EPI-HNE-4 (as determined by spectrometric assay at 280 nm, Coomassie protein assay and biological activity assay) , corresponding to an overall yield of the purification process of about 40 %. RP-HPLC (performed as described in Example 3 but with integration of the signal at 280 nm) showed about 1.5 % of the (n+9) form of EPI-hNE-4 form, compared to total EPI-hNE-4 (see Figure 5) .

EXAMPLE 5; Purification of EPI-hNE-4 produced by the strain according to the invention PE76A 1.8 1 of the supernatant after centrifugation of PEY-53 culture medium (obtained as described in Example 3 above) were collected and passed over a packed bed as follows: 56 ml of chromatographic matrix (Streamline SP from Amersham-Pharmacia) was equilibrated in 50 mM ammonium acetate pH 3.5 (packed bed) . After loading, the column was washed in 10 mM ammonium acetate pH 3.5 to obtain an absorption at 280 nm of below 0.05. EPI-hNE-4 was recovered by washing the column in 1 M ammonium acetate pH 4.5 buffer. Thus was obtained a 240 ml solution containing about 380 mg of proteins and pigments (as determined by spectrometric assay at 280 nm and by Coomassie protein assay) . Compared with Example 4 above, contamination with pigments was far less significant (the solution was only slightly yellow instead of dark-green) . This decrease in pigment contamination was confirmed by RP-HPLC analysis. No (n+9) form of EPI-hNE-4 was detected.

One third of the peak after Streamline (80 ml, corresponding to 0.6 1 of fermentation) was further puri ied on Cation exchange using a FPLC chromatographic system from Pharmacia. The matrix used was Macroprep High S matrix from BioRad (rigid matrix based on cross-linked methacrylate carrying sulphonate surface groups) , in a 1 ml HR5/5 column from Pharmacia. The buffers used were A: ammonium acetate lOmM pH 3.5, B sodium acetate 50 mM pH 6.2 and C: lOmM ammonium bicarbonate pH 7.8. No (n+9) form was detected in buffer B but the peak contained pigments. Elution was then performed by one step at 100%C with a flow rate of 300 cm/h.

The eluate contained about 35.1 mg of purified EPI-HNE-4 (as determined by spectrometric assay at 280 nm, Coomassie protein assay and biological activity assay) , corresponding to an overall yield of the purification process of about 40 %. The purified EPI-hNE-4 was then crystallized according to the procedure already described in WO 01/10727. It appears that EPI-hNE-4 produced by PE76A crystallizes more easily : the purified EPI-hNE-4 from PEY-53 crystallized in such a manner that about 10 mg /ml of protein remained in solution

(as determined by UN spectrometry) while the purified EPI-hΝE-4

-from the new clone PE76A crystallized in such a manner that only about 2.7 mg /ml of protein remained in solution (as determined by ON spectrometry) . These results allow a substantial increase of the yield of the purification procedure of EPI-hΝE-4 (from about 40 % to more than 70 %) with the strain according to the invention. This leads the inventors to the hypothesis that the (n+9) form or the pigments somehow interferes with the crystallization of the purified protein.

RP-HPLC (performed as described in Example 3 but with integration of the signal at 280 nm) showed no traces of the (n+9) form of EPI-hΝE-4 (see Figure 6) .

Claims

1. A fused gene construct comprising a nucleotide sequence coding for the Saccharomyces cerevisiae α mating factor prepropeptide or a variant thereof, which is modified so as to replace the Lysine residue in position 76 of the amino acid sequence of the naturally occurring prepropeptide by an Alanine residue, ligated to the 5' terminal of a nucleotide sequence coding for EPI-hNE-4.

2. A fused gene construct according to claim 1, comprising the nucleotide sequence as represented by SEQ ID N° 1, or coding for an amino acid sequence as represented by SEQ ID n° 2.

3. A fused gene construct according to either of claims 1 and 2, having a nucleotide sequence as represented by SEQ ID n°3.

4. A fused gene construct according to claim 3, coding for an amino acid sequence as represented by SEQ ID n°4.

5. An expression vector comprising a fused gene construct according to any one of claims 1 to 4.

6. An expression vector according to claim 5, wherein the expression vector is selected from the group consisting of a vector derived from Saccharomyces cerevisiae such as pESC (Stratagene) , pYES (Invitrogen) , pRS, pYS, a vector derived from Pichia methanolica such as pMET, and a vector derived from Pichia pastoris such as pPIC and pGAP (Invitrogen) , preferably pPIC.

7. A eukaryotic micro-organism comprising a fused gene construct according to any of claims 1 to 4 , or transformed by an expression vector according to claim 5 or 6.

8. A micro-organism according to claim 7, wherein the microorganism is yeast such as Saccharomyces cerevisiae, Schizosaccharomyces pombe, Yarrovia lipili tica, Hansenula polymorpha, Kluyveromyces lactis, Pichia methanolica and Pichia pastoris, preferably Pichia pastoris .

9. A micro-organism according to any one of claims 7 and 8, which produces EPI-hNE-4 in a fermentation medium with no detectable levels of EPI-hNE- (n+9) using Western Blot analysis .

10. A micro-organism according to any one of claims 7 to 9, wherein the content of EPI-hNE-4 (n+9) in the fermentation medium is less than 2 % compared to total EPI-hNE-4, as determined by RP-HPLC analysis.