The invention relates to a method for preparing non-glycosylated brazzein by means of expression in Pichia pastoris. The codon usage of the genomic coding sequence for the brazzein protein was optimised as a function of the expression system.

Background of the invention

Sucrose consumption through diet is more or less directly implicated in the onset of various conditions such as for example obesity, dental caries, cardiovascular diseases, hypertension, certain types of cancer and diabetes mellitus. At present, approximately 300 million people worldwide suffer from diabetes and the top ten countries with the highest number of cases of diabetes are: India, China, USA, Indonesia, Japan, Pakistan, Russia, Brazil, Italy, and Bangladesh. The global obesity and overweight epidemic together with the other consequences of the high consumption of sugars, is rapidly becoming one of the main problems of public health. In this context, the demand for new sweetening ingredients, possibly of natural origin, is extremely high.

In the search for "sweetening" compounds to replace sugars, brazzein seems to be of particular interest. It is a protein first isolated from the fruit of the African plant Pentadiplandra brazzeana Bailon in 1994 (Ming D. et al., 1994). Of the six "sweet" proteins so far identified, brazzein seems to be the most promising. With respect to other proteins of this class, it is characterized by high resistance to heat and to extreme pH values, in its native form is not glycosylated and, referring to the data reported in the literature, it has a sweetening power 500-2000 times greater than sucrose in terms of SE (Sucrose equivalents) (Berlec A. et al., 2006).

The limitations connected with the use of brazzein in the food field are linked in a first instance to obtain the plant and to the complications associated with the large- scale production and cultivation, which are the main reasons making the production of brazzein from plants economically disadvantageous. An alternative and more economical solution for the industrial production of this protein is offered by recombinant DNA technology. Pichia pastoris represents an exemplary expression system for the production of proteins for three fundamental reasons: i) it can be easily manipulated at the genetic level; ii) is capable of expressing high levels of recombinant proteins; iii) it makes post-transcriptional changes on the protein sequence (glycosylations, formation of disulphide bridges, proteolytic processing, etc.) in a manner similar to what is observed in higher cellular eukaryotic organisms.

Indeed, Pichia pastoris is a micro-organism that has been commonly used for many years in the production of heterologous proteins for both research and industrial applications (Cregg J.M. et al., 1993; Cereghino J.L and Cregg J.M., 2000).

Miles L. et al. (Miles L. et al. 2010) describe a method for preparing brazzein by expression thereof in yeasts and in Saccharomyces cerevisiae and Pichia pastoris in particular. The protein, or part thereof, produced according to the method described in the above-mentioned patent, is however modified by the expression system with the addition of glycosylations.

It is known that yeasts such as Pichia pastoris and Saccharomyces cerevisiae may modify the protein by adding oligosaccharides that differ from those observed in mammals in the composition and distribution pattern of the saccharide units. Such modifications can cause allergic reactions in humans since:

- the protein with a different glycosylation pattern can be recognised by the immune system as a non-self substance;

oligosaccharide chains of larger dimensions or glycosylations in sites other than those of the native protein can prevent access to the cutting sites on the peptide sequence by proteases of the digestive tract, such as for example pepsin.

The addition of non-native glycosylations can also mask catalytic sites or protein interaction domains, resulting in the partial or complete loss of biological activity and consequently limiting end product efficacy. This is the case, for example, of the Antifreeze ISP Type III HPLC 12 protein, used in the formulation of ice-creams and recombinantly produced with the use of Saccharomyces cerevisiae. At the end of the yeast fermentation process a product containing the protein in active form and a mixture of inactive glycoconjugate isoforms is obtained (EFSA, 2008). With regard to possible glycosylation sites, it is known that Pichia pastoris is able to add oligosaccharides to Asparagine residues within Asn-X-Thr/Ser sequences, where X is any amino acid with the exclusion of Proline. Another glycosylation consensus site in yeast has been identified in the Asn-X-Cys sequence (Knauer et al., 1999); this site is present in the brazzein amino acid sequence. The glycosylation of Serine and Threonine residues is, on the other hand, more unpredictable in that there is no data on the specificity and on the frequency of occurrence of this type of modification. It is generally possible to predict if and how a given protein will be glycosylated by Pichia, on the basis of whether or not it is glycosylated in the organism of origin (Cregg J. M. et al., 2000).

Currently, the strategies that can be adopted to prevent Pichia from conducting unwanted glycosylation of proteins are:

replacing the Asparagines in the potential glycosylation target sites;

avoiding Endoplastic Reticulum targeting

However, in the first case possible complications can arise if the Asparagine residue to be replaced is critical in maintaining the structure/function of the protein. In the second case the passage of the protein through the compartments of the Endoplasmic Reticulum and Golgi, in which the enzymes responsible for the glycosylation process are contained, is avoided, while the advantages in terms of easier purification process, that would be obtained by exploiting the secretion pathway, are loosed. Cytosolic expression is also an alternative hardly viable in the case in which the protein necessitates other post-translational modifications, such as for example the formation of disulphide bridges.

Therefore, there remains a need for a method for preparing brazzein by means of expression technologies applicable at industrial level, in which the protein is expressed in non-glycosylated form and with an appropriate structural conformation.

The obtaining of the brazzein protein in non-glycosylated form in a cell expression system applicable to the food industry is thus an aim of this invention.

Summary The inventors have now found that the non-glycosylated brazzein protein can be obtained by inducible or constitutive expression in a specific strain of P. pastoris: the X33 strain.

This strain of P. pastoris is transformed by integrating into the genome thereof an expression cassette comprising a promoter sequence that is operably linked to a nucleotide sequence encoding a brazzein and a terminator sequence and optionally a signal sequence for the secretion of the expressed protein.

In a first aspect the object of the present invention is therefore a method for preparing a brazzein protein comprising at least the steps of:

- transforming one or more cells of Pichia pastoris with an expression vector comprising a promoter sequence, a coding nucleotide sequence for a brazzein sequence and a terminator sequence and optionally a secretion signal sequence;

growing the transformed cell or cells in a fermentation medium under aerobic conditions,

characterized in that:

the cell or cells of Pichia pastoris are of the X33 strain and the brazzein produced by these transformed cells is a non-glycosylated protein.

In further aspects the object of the present invention are the expressions vectors and the recombinant cells of Pichia pastoris X33 strain comprising at least one copy of a gene construct suitable to express non-glycosilated brazzein.

Different promoter sequences can be used for the expression of brazzein in Pichia pastoris, including the promoter sequence of formaldehyde dehydrogenase (FLD1 , GenBank: AF066054.1 ), the promoter sequence of isocitrate lyase (ILC1 , GenBank: AJ272040.1 ), the promoter sequence of the transcription elongation factor 1 (TEF1 , GenBank: EF014948.1 .), the promoter sequence of phosphoglycerate kinase 1 (PGK1 , GenBank: AY288296.1 ), the promoter sequence of peroxisomal biogenesis factor 8 (PER3, GenBank: L40485.1 ), and other known promoter sequences.

For the purposes of the present invention the preferred promoter sequences are the promoter sequence of alcohol oxidase 1 (pAOX1 , SEQ ID NO:8) for the inducible expression of brazzein and the promoter sequence of glyceraldehyde-3- phosphate dehydrogenase (pGAP, SEQ ID NO:12) for the constitutive expression thereof, while the coding nucleotide brazzein sequence is one of the known sequences of native protein (SEQ ID NO: 5 or SEQ ID NO:6) or a modified form for the first amino acid (SEQ ID NO: 4), optimised for the expression system.

The terminator sequence can be selected from the terminator sequence of alcohol oxidase 1 (AOX1 ) and the terminator sequence of glyceraldehyde-3-phosphate dehydrogenase (GAP, GenBank: U62648.1 ). Preferably, the terminator sequence is AOX1 terminator (SEQ ID NO:10).

In addition, the vector preferably further comprises a secretion signal: the preferred secretion signal is the a-mating factor of S. cerevisiae (a-MF, SEQ ID NO:9), however other secretion signals that are known and usable in yeasts can also be used, such as for example the signal sequence of Pichia pastoris acid phosphatase (PH01 , GenBank: U28658.1 ) or of Saccharomyces cerevisiae invertase (SUC2, GenBank: M13627.1 ). Therefore, in one preferred embodiment, the genome of the Pichia pastoris X33 strain stably comprises a vector comprising a promoter sequence selected from the pAOX1 (SEQ ID NO:8) or pGAP (SEQ ID NO:12) promoter sequences, a secretion signal consisting of the a-factor of Saccharomyces cerevisiae a-MF (SEQ ID NO: 9), a nucleotide coding sequence for a brazzein sequence, preferably optimised for the expression system, the termination sequence AOX1 (SEQ ID NO:10) and a selection marker for the antibiotic Zeocin™ (SEQ ID NO:1 1 ).

At least one and preferably two copies of the gene construct comprising the pAOX1 (SEQ ID NO:8) or pGAP (SEQ ID NO:12) promoter sequences, the secretion signal sequence, the reading frame sequence of the brazzein protein and the AOX1 terminator sequence (SEQ ID NO:10) are integrated into the genome of the Pichia pastoris X33 strain.

In a preferred aspect, the nucleotide coding sequences for the expression of brazzein are selected from the DNA sequences of SEQ ID NO:1 , SEQ ID NO:2 and SEQ ID NO:3.

The expressed protein is then purified from the fermentation medium of the yeasts, according to a method comprising the steps of: a) separating the cells from the surnatant by means of filtration or centrifugation; b) diluting the surnatant obtained in step a) and adjusting the pH; c) ion-exchange chromatography; d) concentrating and diafiltering the protein eluted by step c) to reduce the salt content; e) lyophilising the isolated protein.

At the end of the process described, the protein obtained can be used as a substitute for common sugar for food use in foods and beverages, to give such foods and beverages a taste perceived as sweet. Of the various possible applications that are envisaged, one possible application is the addition of brazzein to bakery products, being that the amount of sugar required to achieve the necessary sweetness often hinders the dough's ability to rise.

Moreover, given the pH resistance, brazzein could also be used to sweeten acid- based beverages, such as fruit juices and fizzy beverages. Another possible application could be to use the protein as an excipient for pharmaceutical products as an alternative to other caloric or synthetic sweeteners in use in the field.

Brief description of the drawings

Figure 1 . The drawing shows the map of the plasmid pPICZa vector for the inducible expression of brazzein in Pichia pastoris with the pAOX1 promoter sequence.

Figure 2. The drawing shows the map of the pGAPa plasmid vector used for the constitutive expression of brazzein in Pichia pastoris with the pGAP promoter sequence.

Figure 3. The drawing shows a schematic diagram of the process for obtaining the plasmid vector with the pGAP promoter sequence from the pPICZaA vector.

Figure 4. The drawing shows: the flow through output from the cation-exchange column (Sepharose CM), after loading of the sample and washing (A). The elution profile of the proteins bound to the CM Sepharose resin, following application of an increasing gradient of NaCI (B). The chromatographic profile of the Fraction III loaded onto a reversed-phase (RP) chromatography column (C). Electrophoresis in Precast polyacrylamide gradient gel (4-20 %). Line 1 : Molecular weights; Line 2: Fermentation medium obtained after 72h hours of culture, diluted 4.6x and brought to pH 4.0; Line 3: Fraction I (Unbound proteins); Line 4: Fraction II ; Line 5: Fraction III (brazzein) (D). Figure 5: The drawing shows: TOCSY spectra of a sample 6.6 mg/ml of brazzein at pH 3.5 reported in "Studies on solution NMR structure of brazzein" (Gao et al. 1999, Science in China) (A). TOCSY spectrum of a sample 3.3 mg/ml of brazzein at pH 3.5 (B). TOCSY spectrum of a sample 3.3 mg/ml of brazzein at pH 3.5 following treatment at 98 °C for 2 h (C).

Figure 6. The drawing shows electrophoresis in Tris-Tricine gel of a sample of brazzein digested with Pepsin. Line 1 ) pepsin 1 .28 mg/ml; Line 2) surnatant collected from the culture medium diluted to a final concentration of brazzein 0.5 mg/ml; Line 3) Ultra Low Molecular Weights; Line 4) Pepsin 1 .28 mg/ml incubated with 0.5 mg/ml of brazzein NaCI in 30 mM, pH 1 .5 37°C at 1 minute; Line 5) the same conditions as line 4 at 5 minutes; Line 6) the same conditions as line 4 at 10 minutes, Line 7) the same conditions as line 4 at 20 minutes, Line 8) the same conditions as line 4 at 40 minutes, Line 9) 80 minutes.

Detailed description of the invention

The present invention provides a new, simple, reliable and convenient method for producing non-glycosylated brazzein using the Pichia pastoris yeast and the purification process thereof. P. pastoris is a methylotrophic yeast, recently reassigned to the Komagataella genus, which can be genetically engineered to express proteins for both research purposes and industrial use.

In addition, P. pastoris can be considered a safe source of food ingredients in that: (a) the American Type Culture Collection (ATCC) classifies Pichia as biosafety level 1 , i.e. as a micro-organism that does not cause diseases in healthy individuals; (b) is compliant with OECD (Organization for Economic Co-operation and Development) criteria for Good Industrial Large Scale Practices; (c) was used for the production of many proteins, even for pharmaceutical use and (d) is used in the feeding of animals.

There are standard procedures for genetic manipulation, similar to those described for S. cerevisiae in Molecular cloning: a laboratory manual, 3 ^rd ed., Sambrook, Joseph; Russell, David W., New York: Cold Spring Harbor Laboratory, 2001 .

The preparation method of brazzein object of the present invention is based on the use of a well-defined strain of Pichia pastoris that, surprisingly and contrary to what is known for these micro-organisms, has proven to express the protein in non-glycosylated form.

For this method all the known promoter sequences for the expression of proteins and/or polypeptides in yeasts and the known gene sequences of brazzein and the variants thereof optimised for expression in yeast can be used.

Hereinafter a detailed description of the preparation method of the non- glycosylated brazzein in preferred embodiments thereof, which provide for the use of a commercially available pPICZa vector for inducible expression or a commercially available pGAPa vector for constitutive expression, is reported.

Definitions

Brazzein is a protein isolated from the fruit of the African plant, Pentadiplandra brazzeana Bailon, the known sequence of which consists of 54 aa and having data bank access number P56552 (UniProtKB: locus DEF_PENBA).

For the purposes of the present invention, the preferred brazzein sequences used for inducible or constitutive expression in P. pastoris, strain X33 are:

(SEQ ID NO:4) (aa 54)

MDKCKKVYEN YPVSKCQLAN QCNYDCKLDK HARSGECFYD EKRNLQCICD YCEY;

(SEQ ID NO:5) (aa 54)

QDKCKKVYEN YPVSKCQLAN QCNYDCKLDK HARSGECFYD EKRNLQCICD YCEY; (SEQ ID NO:6) (aa 53)

DKCKKVYEN YPVSKCQLAN QCNYDCKLDK HARSGECFYD EKRNLQCICD YCEY.

The sequences ID NO:5 and 6 respectively correspond to the wild type 2 and 3 forms of brazzein; the sequence ID NO:4 presents a methionine at the N-terminal end for the optional expression in E. coli.

These brazzein sequences are encoded respectively by the following nucleotide sequences:

SEQ ID NO:1 (165 nt)

ATGGATAAAT GCAAGAAAGT ATATGAGAAT TACCCAGTTA GCAAGTGTCA ATTGGCAAAC 60 CAGTGCAACT ATGACTGTAA ACTAGACAAG CATGCTAGAA GTGGAGAATG TTTTTATGAT 120 GAAAAAAGGA ATTTACAATG CATATGCGAT TATTGTGAGT ATTAA 165

SEQ ID NO:2 (165 nt)

CAGGATAAAT GCAAGAAAGT ATATGAGAAT TACCCAGTTA GCAAGTGTCA ATTGGCAAAC 60

CAGTGCAACT ATGACTGTAA ACTAGACAAG CATGCTAGAA GTGGAGAATG TTTTTATGAT 120 GAAAAAAGGA ATTTACAATG CATATGCGAT TATTGTGAGT ATTAA 165

SEQ ID NO:3 (162 nt)

GATAAATGCA AGAAAGTATA TGAGAATTAC CCAGTTAGCA AGTGTCAATT GGCAAACCAG 60

TGCAACTATG ACTGTAAACT AGACAAGCAT GCTAGAAGTG GAGAATGTTT TTATGATGAA 120 AAAAGGAATT TACAATGCAT ATGCGATTAT TGTGAGTATT AA 162.

Said sequences are preferable in that without altering the amino acid sequence of the protein they have a codon usage that is optimised for expression in P. pastoris. For the inducible expression of brazzein, in a preferred embodiment of the method object of the invention is used the pPICZaA vector containing the AOX1 promoter sequence (SEQ ID NO:8) activated in presence of methanol.

For the constitutive expression of the brazzein in another preferred embodiment of the method object of the invention, the pGAPaA vector obtained from the replacement of the AOX1 promoter sequence (SEQ ID NO:8) with the pGAP promoter sequence (SEQ ID NO:12) is used.

The secretion sequence of the preferred protein is the a-mating factor of S. cerevisiae a-MF of SEQ ID NO: 9.

The preferred terminator sequence is AOX1 of SEQ ID NO:10.

Strain X33 has a wild-type (has no auxotrophy) genotype and Mut ⁺ phenotype; this phenotype (Methanol utilization plus) refers to the strain's ability to grow more quickly using methanol as a source of carbon, with respect to the Mut ^s and Mut ^" strains.

Description

Both the vectors for the inducible pPICZa or constitutive pGAPa production of protein comprise a promoter sequence, a secretion signal, a brazzein nucleotide coding sequence, a terminator sequence and a selection marker for the antibiotic Zeocina™.

Construction of expression vectors

The plasmid comprising the pAOX1 (pPICZaA) promoter sequence for the inducible expression of the protein and a dominant selection marker that confers resistance to the antibiotic Zeocina™, both in P. pastoris and in Escherichia coli, can be commercial and can be, for example, the one sold by Invitrogen having Catalogue No. V195-20. The cDNA of the protein was obtained by chemical synthesis (Sigma genosys) on the basis of a known or new amino acid sequences of brazzein and preferably selected from brazzein sequences SEQ ID NO:4, SEQ ID NO:5 or SEQ ID NO:6. In the case of the brazzein nucleotide coding sequence of SEQ ID NO:1 , preferred cDNA is obtained by replacing the nucleotide transcription codon of glutamine at the N-terminal of brazzein nucleotide coding sequence of SEQ ID NO:2 with a nucleotide transcription codon for a methionine for the optional expression in E. coli, and by optimising the codon usage of the sequence for expression in Pichia pastoris and E. coli.

Codon usage is one of the most important factors that influence the expression of heterologous genes; in general, more rare codons are contained in a sequence, the more difficult it will be to obtain satisfactory levels of protein expression. As a result, the codon optimisation without alteration of the amino acid sequence of the encoded protein, can significantly increase the levels of protein expression. For the method according to the invention, the optimisation of codon usage relates to Pichia pastoris and E. coli.

In addition, all possible restriction sites are to be eliminated from the cDNA that codes for the brazzein, so as to make possible any subsequent gene manipulation steps such as the exchange between different vectors or the removal of fusion proteins and tags.

The cDNA sequence selected from SEQ ID NO:1 , SEQ ID NO:2 and SEQ ID NO:3 is thus cloned in the Xhol/Notl restriction site of pPICZaA to generate the pPICZaA-bra vector (FIG. 1 ).

The expression plasmid containing the pGAP (FIG. 2) is, on the other hand, obtainable by modifying the pPICZaA-bra vector; the pGAP promoter sequence is amplified by means of PCR by the Pichia pastoris genome (chromosome 4) and subsequently inserted into the pPICZaA-bra plasmid in place of the pAOX1 to generate the new pGAPZaA-bra plasmid (FIG.3).

The constitutive production of brazzein is preferable for industrial applications as it eliminates the danger and the costs relating to the use of methanol (storage and transport). In the context of the use for the protein in question, i.e. human consumption, the removal of the methanol from the production process is a mandatory choice as it is toxic to humans.

The latter expression system is also preferable, as it does not require an accurate optimisation of the culture conditions as in the case of induction with methanol. In addition, unlike the plasmids for inducible expression, the vectors that use the GAP sequence promoter allow the recombinant protein to be continuously produced in a simple manner, making this system more suitable for large-scale production.

The pGAPZaA-bra plasmid is therefore the most preferable for the industrial application purposes.

Both plasmids also comprise the secretion signal sequence a-mating factor of

Saccharomyces cerevisiae a-MF after the promoter sequence, thus simplifying the recovery of the heterologous protein secreted. Indeed, the biggest advantage of expressing a heterologous protein as secreted protein is that P. pastoris secretes very low levels of native proteins. Other secretion signals can be used in place of the Saccharomyces cerevisiae a-mating factor as previously mentioned.

Both plasmids for the constitutive and inducible expression are transformed into E. coli strain DH5a (Catalogue No. 18263-012, Invitrogen) in order to amplify the

DNA for the subsequent transformation of the yeasts.

The cDNA preferred sequence is SEQ ID NO:1 and, then, the preferred vectors are respectively the pPICZaA plasmid of SEQ ID NO:7 and the pGAPZaA-bra plasmid of SEQ ID NO:15.

The pPICocA-bra and pGAPZocA-bra plasmids are identified by means of DNA sequencing and used for expression of the brazzein gene in P. pastoris.

Transformation of Pichia pastoris with the pPICZccA-bra or pGAPZocA-bra vectors Following linearisation with an appropriate restriction enzyme, the plasmid is used to transform P. pastoris X33 strain by electroporation. Other methods can be used to transform the yeasts, such as for example chemical methods.

The transformed clones presenting a high level of expression are selected on plates with increasing concentrations of Zeocina™: the clones that are able to grow on the plates with higher concentrations of Zeocina™ are putatively those containing multiple copies of the brazzein gene. The selected clones can be stored at -80 °C in a medium containing 15% glycerol. Selection of the more productive clones through flask experiments

The individual colonies selected are analysed for their level of protein expression in order to identify the most efficient. The cultures are incubated at 30 °C for 3 days, under stirring at 230 rpm; for the inducible expression the methanol is added every 24 hours to maintain the induction. The culture surnatant is obtained every 24 hours by centrifugation and is used to analyse protein expression by means of Tris-Tricine SDS-PAGE electrophoresis.

Brazzein expression in a bioreactor

Fermentation is performed in a bioreactor at 30 °C, DO (dissolved oxygen concentration; this is the percentage of oxygen dissolved in the medium, where 100% represents the medium saturated with O ₂ , corresponding to about 220 μΜ) 20% (maintained by controlling the stirring and the influx of air) and pH 5.0. The cells of Pichia pastoris are maintained in a fed-batch medium for 4 days. For purification of the protein, the medium is recovered by centrifugation and treated as described below.

The medium used for the inducible expression of the brazzein is a low-salt medium as described in the manual Pichia Fermentation Process Guidelines (Invitrogen). After 24 hours of growth, protein expression is induced by adding methanol to the culture, initially at a low rate to allow the adaptation of the culture to the methanol and then progressively at greater rates in order to maintain the required level of dissolved oxygen.

For the constitutive expression of the brazzein a known low-salt medium (Hohenblum H. et al., 2004) is used and recorded below in the detail of example 4. After 24 hours of growth, the feeding step is initiated using a dextrose solution. The surnatant is collected by centrifugation every 24 hours and the level of protein expression thereof is analysed through Tris-Tricine SDS-PAGE electrophoresis. At the end of the fermentation process, the medium is recovered and subjected to purification steps.

Purification process

As reported in the manual EasySelect Pichia Expression Kit (Invitrogen), Pichia pastoris is able to grow at high biomasses in a simple defined medium and also secretes a low level of native proteins. As a consequence, the heterologous protein secreted represents the great majority of the total proteins in the medium, thus making the purification process relatively easy.

The first step of the purification process is the removal of cells from the growth medium by centrifugation. Other methods, such as for example clarification and microfiltration can be used.

The culture surnatant is then diluted to lower the ionic strength value in order to allow the proteins to bind to chromatographic material; the final pH of the solution is subsequently modified on the basis of the type of ion-exchange chromatography that is to be used.

With a cation exchange resin such as Sepharose CM FF at pH values of about 4.0, only brazzein and few other contaminants bind to the resin. It is thus possible to elute the brazzein separately from the other proteins, by applying a linear NaCI gradient to the column.

On the contrary, when using an anion exchange resin, such as for example Sepharose DEAE FF at pH values of about 5.0, only the contaminants present in the medium bind to the resin, thus allowing the recovery of the pure brazzein in the dead volume of the column following the loading of the sample.

Depending on the type of ion-exchange chromatography selected for the purification, various concentration and filtration steps are required, in accordance with the desired characteristics of the end product.

The concentrated brazzein solution can be stored as solution following sterilisation by filtration (filter with 0.22 μιη pores), or as a powder directly lyophilising the purified protein.

This purification process does not envisage hazardous or expensive chemical steps and is thus suitable for application in the food industry.

Non-glycosylated brazzein is obtained by the described method, thus fulfilling the aim of the present invention. The protein obtained has the characteristics as set out below.

Characterisation of the product

At the end of the fermentation, the protein in the surnatant is separated by gel electrophoresis and the brazzein concentration is estimated after staining with Comassie and densitometric analysis. The purity of the protein following the ion- exchange chromatography step is established by means of RP-HPLC analysis (as reported in detail in example 5 below).

Mass spectrometry of the purified protein has given a molecular mass of 6500.89 ±0.67 Da as result, a result comparable to the expected one of 6501 Da. It can be inferred that the brazzein is not degraded by proteases present in the surnatant and that the only post-transcriptional changes present are four disulphide bridges typical of the native structure of the protein; other modifications, such as, for example, glycosylations, which can alter the structure of the protein or mask any proteolytic cutting sites thereof, are not present.

The structural similarity of the recombinant protein thus obtained with the wild-type brazzein directly extracted from its natural source is confirmed by TOCSY NMR experiments, performed as described in the literature (Gao et al., 1999) (Figures 6A and 6B). In addition, as shown in Figure 6C, the protein maintains its structure (and consequently the sweet taste) even following heat treatment at 98 °C for 2 hours. It is thus thermostable and therefore suitable for industrial use.

A first step towards determining the possible allergenicity of the recombinant protein was carried out by in vitro responsiveness test to pepsin, in accordance with the general guidelines published by FAO/WHO (WHO, 2001 ). As shown in Figure 7, the protein is readily digested by the pepsin after 10 minutes of incubation under conditions that simulate the gastric fluid, thus suggesting that the recombinant brazzein has a low probability of being an allergen and has no structural changes such as for example glycosylations preventing the proteolysis thereof.

The present invention thus provides a sweet, properly structured, non- glycosylated, thermostable and pepsin-sensitive protein that is produced through the above-described method.

The protein thus produced is subsequently characterized using both standard biochemical assays and structural characterization techniques, such as for example UV quantification at 280 nm,Tris-Tricine gel electrophoresis, HPLC, mass spectrometry and nuclear magnetic resonance spectroscopy. The method for preparing brazzein object of the invention will further illustrated by the following non-limiting examples.

EXAMPLES

Example 1 . Construction of an expression vector for the inducible or constitutive expression of brazzein

The cDNA of the brazzein (SEQ ID NO: 1 ) was chemically synthesised (Sigma genosys) with optimisation of the codons for expression in P. pastoris and then cloned into the pPICZccA vector.

The pPICZccA vector for the inducible expression of brazzein was obtained from Invitrogen (Catalogue No. V190-20, San Diego, USA). This vector was digested with Xhol (Promega, Catalogue No. R6161 ) and Notl (Promega, Catalogue No. R6431 ). The brazzein cDNA was bound with the pPICZaA vector and then transformed into the DH5a strain of E. coli (Catalogue No. 18263-012, Invitrogen). The transformed clones (pPICZccA-bra, SEQ ID NO:7) were selected on low salt LB plates (1 Og/I tryptone, 5g/l NaCI, 5g/l yeast extract, 15 g/l agar) containing 25 μ9/ιηΙ of Zeocina™ (Invitrogen, Catalogue No. R250-01 ). The plasmid pPICZccA- bra was then sequenced to verify the accuracy of the sequence and then used for the expression of the brazzein gene in P. pastoris X33 strain (Invitrogen, Catalogue No. K1740-01 ). The pPICZccA vector comprises the inducible pAOX1 promoter sequence (SEQ ID NO:8), the secretion signal sequence of the a mating factor of Saccharomyces cerevisiae (SEQ ID NO:9), the brazzein nucleotide coding sequence (SEQ ID NO:2), the AOX1 terminator sequence (SEQ ID NO:10) and the Sh ble gene for the resistance to the antibiotic Zeocina™ (SEQ ID NO:1 1 ), as shown in Fig.1 .

The pGAP promoter sequence (SEQ ID NO:12) was, on the other hand, directly amplified by PCR by the P. pastoris genome (chr 4). The forward and reverse primers used for amplification of the GAP gene (GenBank No GenBank: U62648.1 ) are 5'-CACTTGACAGGATCCTTTTTTGTAG-3' (SEQ ID NO:13) and 3'-CATCGTTTCGAAATAGTTGTTCAATTG-5' (SEQ ID NO:14). The PCR product was then inserted into the pPICZaA-bra plasmid in place of the pAOX1 to generate the pGAPZocA-bra plasmid (SEQ ID NO:15). The pGAP-bra plasmid construction process is shown in Fig. 3. As in the case for the pPICZccA-bra vector, the pGAPZocA-bra vector was transformed into the DH5a strain of E. coli (Catalogue No. 18263-012, Invitrogen) and then sequenced to verify sequence correctness. Example 2. Transformation of the cells of P. pastoris with the vectors for the expression of brazzein by electroporation

Electrocompetent cells of the P. pastoris X33 strain were purchased from Invitrogen (Catalogue No. K1740-01 ), prepared using the protocol suggested in the Easyselect™ Pichia expression kit (Invitrogen) manual: an aliquot of electrocompetent cells of P. pastoris re-suspended in 1 M sorbitol at 0°C (80 μΙ) was mixed with 10 μg of linearised DNA (re-suspended in 10 μΙ of sterile water). The pPICZaA-bra plasmid (SEQ ID NO:7) was linearised with the Sacl restriction enzyme (Promega, Catalogue No. R6061 ). On the other hand, the pGAPZccA-bra plasmid (SEQ ID NO:15) was linearised with Avrll (New England BioLabs; Catalogue No. R0174S).

The mixture was transferred to a 0.2 cm electroporation cuvette pre-cooled to 0°C and incubated on ice for 5 minutes. The cells were then electroporated using a Biorad Gene Pulser II electroporator at 1500V, 25 μΡ and 400Ω.

The transformed cells were incubated with 1 ml of sorbitol 1 M (0°C) at 30 °C for 2 hours.

The cells were then seeded on YPDS plates (10 g/l yeast extract, 20 g/l peptone, 20g/l dextrose, 1 M sorbitol, and 20 g/l agar) containing 100 μg/ml of Zeocina™ as selection marker. The plates were incubated for 3-5 days at 30 °C until formation of the colonies.

Example 3. Identification of transformed cells of P. pastoris X33 strain and small- scale expression test

The transformed clones having the highest level of expression were selected on YPDS plates containing 100, 500, 1000 and 2000 μg/ml Zeocina™ and incubated at 30 °C for 5 days. Clones able to grow on the plates with 2000 μg/ml Zeocina™ are the putative multi-copy clones (i.e. that contain multiple copies of the brazzein coding gene).

The selected colonies can be stored at -80 °C in YPD medium (10 g/l yeast extract, 20 g/l peptone, 20 g/l dextrose) containing 15% glycerol v/v. The yeasts transformed with the plasmid for the inducible expression were re- suspended in 5 ml of BMG medium (100 ml/l of potassium phosphate buffer 1 M at pH 6.0, 100 ml/l of 13.4% Yeast Nitrogen Base (10X YNB) , 2 ml/l of 0.02 % biotin, 100ml/l of 10% glycerol) and incubated in 25 ml Erlenmeyer flasks overnight.

These cultures were used to inoculate 10 ml of BMM medium (100 ml/l of potassium phosphate buffer 1 M at pH 6.0, 100 ml/l of 10X YNB, 2 ml/l of 0.02 % biotin, 100 ml/l of 5% methanol) in 50 ml Erlenmeyer flasks with breakwaters at an initial OD equal to 1 . These cultures were then incubated at 28 °C for 3 days under stirring at 230 rpm; every 24 hours, 100% methanol was added to a final concentration of 0.5 % in order to maintain the induction.

The individual colonies selected for the constitutive expression were transferred into 5 ml of YPD medium (10 g/l yeast extract, 20 g/l peptone, 20 g/l dextrose) and incubated in 25 ml Erlenmeyer flasks 25 ml overnight. These cultures were used to inoculate 10 ml of YPD medium in 50 ml Erlenmeyer flasks with breakwaters at an initial OD equal to 1 . These cultures were then incubated at 28 °C for 3 days under stirring at 230 rpm.

The surnatant of these cultures was obtained by centrifugation every 24 hours and used to analyse the protein expression by means of Tris-Tricine SDS-PAGE electrophoresis.

Through this step it was thus isolated the clone having the greatest efficiency of expression, which will then be used for the subsequent scale-up of the in fermenter process.

Example 4. Brazzein expression in fermenter

The most efficient clone in terms of brazzein expression was selected and grown in 500 ml of medium containing 100 ml/l of phosphate buffer, 100 ml/l of 10X YNB, 2.5 ml/l of 0.02% biotin (500X), 5 ml/l of 100% glycerol, and 792.5 ml/l of distilled water. The culture was used to inoculate the culture for the fermenter (1 .5 I) at an initial OD equal to 1 ; the fermentation medium contains 18.2 g/l K ₂ SO ₄ , 14.9 g/l MgSO ₄ .7H ₂ O, 26.7 ml/l H ₃ PO ₄ (85%), 0.93 g/l CaSO ₄ -2H ₂ O _, 4.13 g/l KOH, 40 g _/ l glycerol, 4.35 ml/l PTM-, . The solution of trace salts (PTM-,) consists of (per litre): 6.0g CuSO ₄ -5H ₂ O, 0.08g Nal, 3.0g MnSO ₄ -H ₂ O, 0.2g Na ₂ MnO ₄ -2H ₂ O, 0.02g H ₃ BO ₃ , 0.5g C0CI ₂ , 20. Og ZnCI ₂ , 65.0g FeS0 ₄ .7H ₂ 0, 0.2g Biotin and 5.0 ml H ₂ S0 ₄ .

The culture was grown for 18-24 hours until complete consumption of the glycerol (this is indicated by a rapid increase in the level of dissolved oxygen). When all the glycerol was consumed, a fed-batch phase with methanol was initiated to induce the expression of the protein; the feeding medium consists of methanol 100% containing 12 ml/l of PTM-i . For the first 2-3 hours, until the culture adapts to the methanol, the feeding rate was set to 3.6 ml/h per litre of initial fermentation volume. When the culture is adapted to the use of methanol the feeding rate was doubled to 7.3 ml/h per litre of initial fermentation volume. After two hours at 7.3 ml /h/litre, the feeding rate of methanol increased to 10.9 ml/h per litre of initial fermentation volume. This rate was maintained until the end of the fermentation. Fermentation for the constitutive expression of the brazzein was carried out in 1 .5 litres of medium containing 9.5 g/l K ₂ S0 ₄ , 7.8 g/l MgS0 ₄ -7H ₂ 0, 23.7 ml/l H ₃ P0 ₄ (85%), 0.6 g/l CaS0 ₄ -2H ₂ 0, 2.6 g/l KOH, 40 g/l glycerol, 4.4 ml/l PTM^ The single colony of yeast selected for the expression was inoculated into 500 ml of medium containing 100 ml/l phosphate buffer, 100 ml/l YNB 10X, 2.5 ml/l biotin 500X, 5 ml/l 100% glycerol, and 792.5 ml/l distilled water, and used as a culture to inoculate the fermenter medium at an initial OD equal to 1 . The culture was grown for 24 hours and then a phase fed-batch phase was initiated with glucose (feeding solution: 550 g/l glucose, 12 ml/l PTM-i) at a rate of 6 ml/h per litre of initial fermentation volume and maintained until the end of fermentation.

At the end of both fermentation processes (72 hours of fed-batch phase), the medium was recovered and centrifuged at 1500g in order to separate the cells from the surnatant. For fermentations on a larger scale other methods such as for example tangential filtration, can be used.

Example 5. Purification process

Cation exchange chromatography

The surnatant obtained following removal of the cells was diluted 4.6 times to lower conductivity from 65 mS/cm to 14 mS/cm and the pH was brought to 4.0. An aliquot of the solution was loaded into Polyacrylamide gel (see Figure 4D, Lane 2). The entire surnatant was then loaded into a Pharmacia Biotech XK 26 column filled with 1 10 ml Sepharose CM FF, pre-equilibrated with Buffer A (20mM Na- acetate buffer, pH 4.0). The flow through output containing the fraction that did not bind to the resin following loading of the sample and washing with Buffer A (Fraction I, Figure 4A) was then collected and an aliquot loaded into gel (Figure 4D, Line 3). Following the first washing step of the column, the concentration of Buffer B (NaCI 1 M in 20mM Na-acetate buffer, pH 4.0) was linearly increased from 0% to 100% in 3.5 column volumes. Two fractions were collected and loaded into gel (see Fraction II and III in Figure 4D). The brazzein can be separated from other protein bound to the column in that it elutes at higher concentrations of Buffer B (approximately 0.6M NaCI). Alternatively, the chromatographic process can be performed using a 2-step elution; the first step at 40% of Buffer B to remove the undesired contaminants and a second step at 60-70% of Buffer B to elute the brazzein. The subsequent diafiltration steps can be carried out using membranes with pore diameter from 1 to 3.5 kDa.

No less that 200 mg of protein are obtained at the end of the described purification process.

Anion-exchange chromatography

The surnatant obtained following removal of the cells, was diluted so as to achieve a final conductivity equal to 5 mS/cm and the pH brought to 5.0 (in this case, the value can be set in the 4-6 range). The solution thus obtained was loaded into a column containing DEAE Sepharose FF, pre-equilibrated with Buffer C (20 mM Piperazine-HCI, pH 5.0). The pure brazzein elutes in the flow through output after the loading of the sample and washes with Buffer C, while the other contaminants are eluted by washing the column with NaCI 1 M. In comparison with the purification method that exploits cation exchange, this method allows the purified protein to be simply recovered in Buffered C in the absence of NaCI.

However in this case the starting surnatant must be diluted by approximately a factor 10 and the protein elutes in larger volumes and requires further concentration steps.

Purity of the protein

The absorbance profile of a reversed-phase chromatography HPLC was monitored at 280 nm and the purity of the brazzein was reported as the relative percentage of peak elution area of the brazzein with respect to the total area subtended by the chromatogram. An aliquot of Fraction III (see Figure 4B) was loaded into a Phenomenex Jupiter 5u C4 column; the eluent A used consists of

0.1 % TFA in H ₂ 0 milliQ, while the eluent B is formed by 0.085% TFA in Acetonitrile. The chromatographic run exploits a non-linear increasing gradient of

Eluent B with a stream at 0.6 ml/min. The purity of the protein at 280 nm was estimated to be 99.9% using a diode-array Agilent 8453 UV-visible spectrophotometer interfaced with Agilent Chemostation software.

References

Ming, D. & Hellekant, G. 1994, "Brazzein, a new high-potency thermostable sweet protein from Pentadiplandra brazzeana B", FEBS letters, vol. 355, No. 1 , pp. 106-108.

Berlec, A., Jevnikar, Z., Majhenic, A.C., Rogelj, I. & Strukelj, B. 2006, "Expression of the sweet-tasting plant protein brazzein in Escherichia coli and Lactococcus lactis: a path toward sweet lactic acid bacteria", Applied Microbiology and Biotechnology, vol. 73, no. 1 , pp. 158-165.

Cregg, J.M., Vedvick, T.S. & Raschke, W.C. 1993, "Recent advances in the expression of foreign genes in Pichia pastoris", Bio/technology (Nature Publishing Company), vol. 1 1 , no. 8, pp. 905-910.

Cereghino, J.L. & Cregg, J.M. 2000, "Heterologous protein expression in the methylotrophic yeast Pichia pastoris", FEMS microbiology reviews, vol. 24, no.

1 , pp. 45-66.

Miles L, Louie M. Subramanian M. WO 2010/030999.

EFSA, 2008, Safety of Ice Structuring Protein (ISP) - Scientific Opinion of the Panel on Dietetic Products, Nutrition and Allergies and of the Panel on Genetically Modified Organisms EFSA.

Knauer, R. & Lehle, L. 1999, "The oligosaccharyltransferase complex from yeast", Biochimica et Biophysica Acta (BBA) - General Subjects, vol. 1426, no. 2, pp. 259-273.

Cregg, J.M., Cereghino, J.L, Shi, J. & Higgins, D.R. 2000, "Recombinant protein expression in Pichia pastoris", Molecular biotechnology, vol. 16, no. 1 , pp. 23-52. Hohenblum, H., Gasser, B., Maurer, M., Borth, N. & Mattanovich, D. 2004, "Effects of gene dosage, promoters, and substrates on unfolded protein stress of recombinant Pichia pastoris", Biotechnology and bioengineering, vol. 85, no. 4, pp. 367-375.

Gao, G., Dai, J., Ding, M., Hellekant, G., Wang, J. & Wang, D. 1999,

"Studies on solution NMR structure of brazzein: Secondary structure and molecular scaffold", Science in China.Series C, Life sciences / Chinese Academy of Sciences, vol. 42, no. 4, pp. 409-419.

WHO, 2001 . Evaluation of Allergenicity of Genetically modified Foods. Report of a Joint FAO/WHO Expert consultation. World Health Organization, Geneva.