More specifically it relates to improvements in the area of accumulation of proteins which are sweet tasting to man (Homo sapiens) such as the protein Brazzein.

To most mammals sugar is a popular sweetening agent however because of the high calorific value of sugar alternatives have been sought for some time.. This has led to the identification and production of artificial sweetening agents. These alternatives, however, can be problematical when used under certain conditions such as prolonged heat exposure as this can lead to a loss in the desired sweet tasting effect. The above mentioned factors contributed towards the need to identify other alternative low calorie, thermostable sweet tasting proteins, Brazzein being one such protein.

Brazzein was initially identified as a sweet protein present in the berries produced by Pentadiplandra brazzeana Baillon, a tropical African climbing shrub. The protein and its nucleotide and amino acid sequences are described in PCT Applications W094/19467 and W095/31547 and related US Patents US5326580, US5346998 and US5527555. All these documents disclose the use of bacterial systems for the production of quantities of the protein. These documents also suggest that if the proteins were expressed in plant systems a sweet plant would be produced.

The art is familiar with the general techniques for isolating a gene from one organism and transferring it to another. Specifically in plant systems, genetic modification ultimately involves a plant transformation process. This transformation process can be performed via a number of methods, for example: the Agrobacterium-mediated transformation method.

In the microparticle bombardment method, microparticles of dense material, usually gold or tungsten, are fired at high velocity at the target cells where they penetrate the cells, opening an aperture in the cell wall through which DNA may enter. The DNA may be coated on to the microparticles or may be added to the culture medium.

In microinjection, the DNA is inserted by injection into individual cells via an ultrafine hollow needle.

Another method, viz. fibre-mediated transformation, applicable to both monocots and dicots, involves creating a suspension of the target cells in a liquid, adding microscopic

needle-like material, such as silicon carbide or silicon nitride"whiskers", and agitating so that the cells and whiskers collide and DNA present in the liquid enters the cell.

Certain transformation methods may be particularly useful for specific species for example banana may be transformed using the method described in Sagi et al. (1995) Biotechnology. Vol. 13 pp 481-485 or Sagi et al. (1994) Plant Cell Reports. Vol. 13. pp 262- 266. or May et al. (1995) Biotechnology. Vol. 13 pp 486-492.

An object of the present invention is to provide materials and methods for improving the accumulation of proteins, particularly sweet-tasting proteins, in recombinant organisms.

According to the present invention there is provided an isolated polynucleotide encoding a fusion protein comprising fused sequences encoding a membrane targeting peptide and a sweet tasting protein.

It is preferred that the sequence encoding the membrane targeting peptide is not naturally associated with the sweet tasting protein.

Preferably the membrane targeting peptide is heterologous with respect to the sweet tasting protein.

The membrane targeting peptide may be obtained from bacteria, fungi or plants.

In one preferred embodiment of the invention the sequence encoding the sweet tasting protein is obtainable from the plant Pentadiplandra brazzeana Baillon and it is further preferred that the sequence encoding said sweet-tasting protein is SEQ ID NO 4 or a variant thereof which encodes said sweet-tasting protein.

In preferred embodiments the sequence encoding a fusion protein has a sequence selected from SEQ ID NO 1 or SEQ ID NO 2 or SEQ ID NO 3 and such variants which produce the said protein. A sequence which has similarity to one of said sequence may be polynucleotide which is complementary to a said sequence which when incubated at a temperature of between 60°C and 65°C in 0.3 strength citrate buffered saline containing 0.1% SDS followed by rinsing at the same temperature with 0.3 strength citrate buffered saline containing 0.1% SDS still hybridises with the sequences depicted in SEQ ID Nos. 1- 3 wherein said polynucleotide encodes a membrane targeting peptide and a sweet tasting protein.

The invention further provides a DNA construct comprising sequentially a promoter region which is operably linked to at least one of the polynucleotides defined above and a

transcription termination sequence. The promoter may be constitutive or inducible, tissue specific or organ specific.

The invention further provides a method of increasing the accumulation of a sweet tasting protein in an organism comprising inserting into an organism by genetic transformation a DNA construct defined above and selecting transformants having an increased accumulation of sweet tasting protein.

The organism may be transformed using Agrobacterium, or by direct insertion, fibre mediated or particle mediated transformation methods.

Preferably the organism is a plant and, most preferably, genus Lycopersicon or Musa.

Still further the invention provides plants, plant material their fruit and seed produced by the said method.

The present invention also provides a polynucleotide encoding a fusion protein comprising a membrane targeting peptide and a further protein wherein the membrane targeting peptide is derived from a plant defensin and said further protein is not naturally associated with the said membrane targeting peptide.

Preferred polynucleotides encoding are those wherein the membrane targeting peptide is coded for by a polynucleotide having a nucleotide sequence selected from SEQ ID NO 6 to SEQ ID NO 13.

Alternatives may be selected from polynucleotides which are is complementary to one which when incubated at a temperature of between 60°C and 65°C in 0.3 strength citrate buffered saline containing 0.1% SDS followed by rinsing at the same temperature with 0.3 strength citrate buffered saline containing 0.1% SDS still hybridises with any one of the sequences depicted SEQ ID No 6 to 13 wherein said polynucleotides still encode a membrane targeting peptide.

Still further to the invention there is provided a DNA construct comprising sequentially a promoter region operably linked to at least one of the polynucleotides encoding a membrane targeting peptide aforesaid fused to a polynucleotide encoding a protein which is not naturally associated with the said membrane targeting peptide, and a transcription termination sequence.

Also the invention provides a method of directing a protein to and/or through a cellular membrane comprising inserting into an organism by genetic transformation a DNA

construct aforesaid and selecting from the transformants those in which the protein is directed to and/or through a cellular membrane.

The organism may be a microorganism a yeast or, more preferably,. a plant and, most preferably a plant of the genus Lycopersicon or Musa..

In the most preferred embodiment of the invention there is provided an edible fruit of a transgenic plant, said fruit containing the protein Brazzein.

The said sweet tasting protein may be obtainable from Pentadiplandra brazzeana Baillon and the said membrane targeting peptide may be obtainable from bacteria, fungi or plants.

The polynucleotides according to the present invention may comprise at least one of the sequences depicted as SEQ ID NO 1 to 3. Moreover the said polynucleotide The organism may be a microorganism or may be a plant and in particular may be a tomato plant or a banana plant.

The present invention also provides plants, plant material their fruit and seed wherein said plants, plant material their fruit and seed are produced according to any of the preceding methods.

The construct may be inserted into the organism using the Agrobacterium, direct insertion, fibre mediated or particle mediated insertion methods.

It may also be preferred to use the membrane targeting sequences of the present invention to direct antibodies or fragments thereof such as: Fab, Fab', F (ab') 2, Fv, scFv, Fd and Vh, to a cellular membrane.

In its principal embodiment, then, this invention is concerned with improving the expression of a transgene which specifies a proteinaceous sweet compound which may find application as a sugar substitute. The compound may be expressed in an expression vector, and a plant or a yeast is convenient for this application, from which it may subsequently be isolated and purified. Alternatively the sweet compound may be expressed in a plant all or part of which may be part of the human or animal food chain so as to provide sweetening, or additional sweetening, of the edible parts of the plant such as fruit, leaves or seed.

Simple expression of an inserted coding region which specifies a protein under the control of a promoter does not guarantee the best expression of the sweet-tasting protein, nor, indeed that the protein which is thus produced will actually manifest its potential sweetness.

While the explanation of these detrimental phenomena is not fully understood, it may be

theorised that the protein may require post-translational processing, perhaps to ensure formation of its correct tertiary structure. It is believed that transport of the expressed protein to or through a membrane of an organelle, such as transport into the extracellular space, allows that post-translational processing to occur.

It is believed that certain proteins require such post-translational transport to particular sub-cellular components but it is by no means possible to predict which of the thousands of proteins have such requirement. The state of research into the subject as applied to plant cells is reviewed in"Mechanisms of Intracellular Transport and Targeting in Plant Cells", Kermode. R., in Critical Reviews in Plant Science; Volume 15, Issue No. 4 (1996) Conger B. V. Ed.; Published by CRC Press Inc.. From that review, which is incorporated herein by reference, it is apparent that many sequences for targeting of proteins to membranes are known.

Protein transport is also the subject of published International Patent Application No.

WO 98/21346, published 22 May 1998 (Queen's University at Kingston, Ontario, Canada) which is concerned principally with targeting to the plastids using a plastid transport protein named Bce44B. That document is incorporated herein by reference.

Essentially any sequence which would have the effect of directing a protein to and/or through a membrane would suffice. It is preferred that the polynucleotide encoding the fusion protein comprising the membrane targeting peptide and sweet tasting protein, has its codons optimised to correspond with the codon usage of the organism in which it is to be expressed.

It is further preferred to select a membrane targeting peptide for use in the present invention, that naturally directs an endogenous protein to and/or through a cell membrane, with a further preference that the polynucleotide encoding the endogenous protein sequence is substantially similar to that of the polynucleotide encoding the sweet tasting protein.

Typically similarity is observed when 60% or more of the nucleotides are common between the sequences. Even more typically 65%, preferably 70%, more preferably 75% or 80% and especially preferred are 90%, 95%, 96%, 97%, 98%, 99% or more of the nucleotides are common between the sequences.

It is even further preferred that the membrane targeting peptide is naturally associated with the group of antifungal proteins known as the plant defensins. It is also preferred that the membrane targeting peptide is naturally associated with plants exhibiting antifungal

activity. Further preferred is that the membrane targeting peptide is obtainable from Dahlia sp or Raphanus sp. and even more preferred is that the membrane targeting peptide is coded for by the polynucleotides depicted as SEQ ID Nos 6 to 13.

Further according to the present invention there is provided a polynucleotide encoding a fusion protein comprising a membrane targeting peptide and a non naturally associated protein wherein the membrane targeting peptide is derived from a plant defensin.

The membrane targeting protein may be coded for by a polynucleotide as depicted in SEQ ID NO 6 to 13 or may be a polynucleotide which is complementary to one which when incubated at a temperature of between 60°C and 65°C in 0.3 strength citrate buffered saline containing 0.1 % SDS followed by rinsing at the same temperature with 0.3 strength citrate buffered saline containing 0.1 % SDS still hybridises with any one of the sequences depicted in SEQ ID NO 6 to 13 wherein said polynucleotides encode membrane targeting peptide.

Also provided is a DNA construct comprising sequentially a promoter region operably linked to at least one of the polynucleotides encoding a membrane targeting peptide which are described above, fused with a polynucleotide encoding a protein which is not naturally associated with said membrane targeting peptide, and a transcription termination sequence.

Another group of membrane targeting sequences which may be used in the present invention can be isolated using the pathogen related peptide sequence data described in EP440304 and EP460753.

It may also be preferred to select a membrane targeting sequence that directs the protein to a specific intracellular membrane or a specific intracellular organelle or compartment. It may also be preferred to select a membrane targeting sequence which will facilitate the secretion of the protein from the cell. For example, in plants the use of an endoplasmic reticulum target sequence may allow the sweet protein to be secreted via the cell secretory system. This would ensure that the sweet protein was exported to the extracellular space where the protein would not be susceptible or available to attack by intracellular proteases. Furthermore, in microbial systems the use of a membrane targeting peptide which results in the extracellular accumulation of the sweet tasting protein is highly advantageous since this facilitates the subsequent recovery of the protein by obviating the need for cell lysis.

Once a polynucleotide sequence encoding a suitable membrane targeting peptide has been selected, it can be fused to the desired polynucleotide sequence encoding the desired sweet tasting protein, using for example DNA ligase. The resulting polynucleotide may encode the membrane targeting peptide either 5 prime or 3 prime of the polynucleotide sequence encoding the desired sweet tasting protein. It is preferred that the polynucleotide sequence encoding the membrane targeting peptide is 5 prime of the polynucleotide sequence encoding the desired sweet tasting protein.

The polynucleotide encoding the membrane targeting peptide and the desired sweet tasting protein may then be used in transformation of a preferred organism using the preferred regulatory elements.

Host cells which may be transformed with the polynucleotides according to this invention include: mammalian cells for example CHO (Chinese Hamster Ovary) cells, yeast cells for example Saccharomyces cerevisiae; microbial cells such as bacterial cells for example Escherichia coli. Examples of genetically modified plants which may be produced by transformation with the polynucleotide sequences of the present invention include field crops, cereals, fruit and vegetables such as oilseed rape, canola, sunflower, tobacco, sugarbeet, cotton, soya, maize, wheat, barley, rice, sorghum, tomatoes, mangoes, peaches, apples, pears, strawberries, bananas, melons, potatoes, carrot, lettuce, cabbage and onion.

In this specification"membrane targeting peptide"includes any signal peptide and means a sequence which directs the protein to and/or through an intracellular and/or extracellular membrane and organelles including the endoplasmic reticulum, more specifically the rough endoplasmic reticulum or smooth endoplasmic reticulum, the golgi apparatus and various other intracellular organelles including any of the plastids, the mitochondria and chloroplasts and intracellular compartments for example the vacuole.

"Sweet tasting protein"means any isolatable protein that has a sweet taste to man (Homo sapiens). It is preferred that the protein is isolatable from plants.

It is further preferred that the protein is isolatable from the plant Pentadiplandra brazzeana Baillon. It is further preferred that the protein is coded for by the polynucleotide sequence as depicted in SEQ ID No. 4 or a variant thereof. Variant includes any substitution of, variation of, modification of, replacement of, deletion of or addition to the sequence as depicted in SEQ ID No 4 providing the resultant sequence is capable of encoding a protein which is sweet tasting to man. The variant may comprise a polynucleotide which is

complementary to one which when incubated at a temperature of between 60°C and 65°C in 0.3 strength citrate buffered saline containing 0.1 % SDS followed by rinsing at the same temperature with 0.3 strength citrate buffered saline containing 0.1% SDS still hybridises with the sequence depicted SEQ ID No 4.

"Plant defensins"is a term given to a class of antifungal proteins (Terras et al 1995.

Plant Cell, 7. Pp573-583). Further examples of the plant defensins include Rs-AFP 1, Rs- AFP 2, Bn-AFP 1, Bn-AFP 2, Sq-AFP 1, Sq-AFP 2, Br-AFP 1, Br-AFP 2, Dm-AMP 1, Dm-AMP 2, Cb-AMP 1, Cb-AMP 2, Lc-AMP, Ct-AMP 1, Ct-AMP 2, Hs-AFP 1 and Ah AMP 1. Details of these may be found in published International Patent Application Numbers W095/18229 and W093/05153.

"Extracellular space"includes the cell wall, cell surface, middle lamella and any other compartment outside of the cell wall.

In the sequence listings herewith: SEQ ID NO 1 shows a 222 base pair fusion of (1) the MFa signal sequence and (2) the coding region specifying the brazzein protein, the ATG translation start codon of the brazzein coding region having been replaced by CAA and the codon usage optimised for expression in yeast; SEQ ID NO 2 shows a 252 base pair fusion of (1) a signal sequence from radish and (2) the coding region specifying the brazzein protein, the ATG translation start codon of the brazzein coding region having been replaced by CAA and the codon usage optimised for expression in tomato; SEQ ID NO 3 shows a 249 base pair fusion of (1) a signal sequence from Dahlia and (2) the coding region specifying brazzein, the ATG translation start codon of the brazzein coding region having been replaced by CAA and the codon usage optimised for expression in tomato; SEQ ID NO 4 shows a 165 base pair DNA encoding the brazzein protein, codon usage having been optimised for tomato expression; SEQ ID NO 5 shows a 57 base pair DNA of the yeast MFa signal sequence; SEQ ID NO 6 shows an 87 base pair DNA of the radish signal sequence; SEQ ID NO 7 shows an 84 base pair DNA of the Dahlia signal sequence, codon usage optimised for use in tomato and, thus, corresponds to the first 84 nucleotides of SEQ ID NO. 3.

SEQ ID NO 8 to SEQ ID NO 11 show variants of the 84 base pair Dahlia signal sequence: they are possibly members of a gene family; SEQ ID NO 12 shows the sequence encoding the mature signal peptide from Dahlia.

SEQ ID NO 13 shows a 99 base pair signal sequence from Heuchera (HS clone 8).

Sequences depicted in the present application may either be synthesised ab initio using the sequence data provided or may be isolated from the corresponding organism using polynucleotide probes based on the sequence data provided. Polynucleotides encoding membrane targeting peptides may also be isolated from other organisms using the sequence data provided herewith in the construction of probes to screen DNA libraries.

The invention will now be described by way of example with reference to the following Figures of which: Figure 1 is a plasmid map of the vector pPIC3.5K-Bz which was used to transform P pastoris. The construct contains the Saccharomyces cerevisiae Mating Factor Alpha pre signal peptide fused to the Brazzein encoding DNA with codons optimised for use in yeast (SEQ ID No 1). The expression of the construct is under the control of the AOX1 promoter.

Figure 2 is a plasmid map of the vector Yiplacl28-BZ which was used to transform S cerevisiae. The construct contains the Saccharomyces cerevisiae Mating Factor Alpha pre signal peptide fused to Brazzein with codons optimised for use in yeast (SEQ ID No 1). The expression of the construct is under the control of the MF alpha promoter.

Figure 3 is a plasmid map of the vector pZPS37. This construct contains the polynucleotide encoding Brazzein without the use of any signal peptide. The expression of the construct is under the control of the polygalacturonase promoter (PG).

Figure 4 is a plasmid map of the vector pZPS38. This construct contains a <BR> <BR> polynucleotide encoding the Dahlia signal peptide and Brazzein, depicted in SEQ ID NO 3.

The expression is under the control of the polygalacturonase (PG) promoter.

Figure 5 is a map of the vector pZPS39. This construct contains the polynucleotide encoding the Radish signal peptide and Brazzein, depicted as SEQ ID NO 2. The expression is under the control of the polygalacturonase (PG) promoter.

Figure 6 plasmid map of the vector pZPS34. This construct contains the polynucleotide encoding Brazzein without the use of any signal peptide. The expression of the construct is under the control of the Arabidopsis polyubiquitin extension protein promoter (UBQ).

Figure 7 is a plasmid map of the vector pZPS35. This construct contains the<BR> polynucleotide encoding the Dahlia signal peptide and Brazzein, depicted as SEQ ID NO 3.

The expression is under the control of the Arabidopsis polyubiquitin extension protein promoter (UBQ).

Figure 8 is a plasmid map of the vector pZPS 36. This construct contains the polynucleotide encoding the Radish signal peptide and Brazzein, depicted as SEQ ID NO 2.

The expression is under the control of the Arabidopsis polyubiquitin extension protein promoter (UBQ).

Figure 9 Is a graph illustrating the average Brazzein concentration in tomato transformants containing the constructs pZPS34, pZPS35, pZPS36, pZPS37, pZPS38 and pZPS39. Data used in generating Figure 9 are given in Table 1 hereinafter.

The invention will now be described, by way of illustration, in the following Examples. Three promoters are mentioned in these Examples; details of these promoters may be found in the following references which are incorporated herein by reference: Arabidopsis polyubiquitin extension protein promoter: see US Patents 5,510,474 and 5,614,399.

Polygalacturonase promoter: US Patent 5,442,052 The AOX1 promoter is the promoter of the alcohol oxidase gene of the yeast Pichia pastoris.

EXAMPLE 1 Expression of the sweet tasting protein Brazzein in tomato Production of transgenic tomato plants with increased accumulation of sweet tasting protein Brazzein.

Constructs were prepared containing the Dahlia (Dahlia merckii) antimicrobial protein signal peptide fused with the Brazzein under the transcriptional control of the Arabidopsis polyubiquitin extension protein promoter (UBQ) (Construct pZPS35 a map of which is shown in Figure 7) or the polygalacturonase promoter (PG) (Construct pZPS38, a map of which is shown in Figure 4). In addition to this constructs were also prepared containing the Radish (Raphanus sativus) signal peptide fused to Brazzein under the expressional control of either the UBQ promoter (Construct pZPS 36, a map of which is shown in Figure 8) or the PG promoter (Construct pZPS39, a map of which is shown in Figure 5). Constructs were also prepared which encoded Brazzein without a signal peptide

but with an N-terminal methionine by the insertion of ATG nucleotides upstream of the Brazzein gene under the expressional control of either the UBQ promoter (Construct pZPS34, a map of which is shown in Figure 6) or the PG promoter (Construct pZPS37, a map of which is shown in Figure 3). These were prepared as follows: Construction of the transformation vector for expression in tomato with Radish signal peptide fused to Brazzein under the expressional control of either the UBQ promoter or the PG promoter: A synthetic DNA was produced which coded for the radish signal peptide fused to Brazzein. The codons were optimised for expression in tomato. Using appropriate restriction sites the coding sequence was cloned into a plasmid vector. The coding region was excised from the plasmid and cloned between the promoter in question and the terminator in the correct orientation for expression.

Construction of the transformation vector for expression in tomato with Dahlia signal peptide fused to Brazzein under the expressional control of either the UBQ promoter or the PG promoter: A synthetic DNA was produced which coded for the with the Dahlia signal peptide fused to Brazzein. The codons were optimised for expression in tomato. Using appropriate restriction sites the coding sequence was cloned into a plasmid vector. The coding region was excised from the plasmid and cloned between the promoter in question and the terminator in the correct orientation for expression.

EXAMPLE 2 Generation and analysis of plants transformed with the transformation vector.

The vector was transferred to Agrobacterium tumefaciens LBA4404 (a microorganism widely available to plant biotechnologists) and used to transform tomato plants. Transformation of tomato stem segments followed standard protocols (e. g. Bird et al Plant Molecular Biology 11,651-662,1988).

Transformed plants were identified by their ability to grow on media containing the antibiotic kanamycin. Up to 30 individual plants were regenerated with each construct and grown to maturity. The presence of the construct in all of the plants was confirmed by polymerase chain reaction analysis. DNA Southern blot analysis on all plants indicated that the insert copy number was between 1 and 10. Northern blot analysis on fruit from one plant indicated that the Brazzein gene was expressed.

Brazzein production in the fruit of all plants was measured by ELISA (enzyme linked imunoabsorption assay) using a polyclonal and a monoclonal antibody raised against native Brazzein protein isolated from the fruit of the plant Pentadiplandra brazzeana Baillon Two fruit were collected from each transgenic plant at 7 days post breaker (the term"breaker"is used to indicate when the tomato fruit first show signs of the orange colouration characteristic of most mature tomato fruit). Total fruit protein was extracted from a sample of the pericarp of each of the fruit. The amount of Brazzein protein in the total protein extract was measured by ELISA and calculated as the amount of Brazzein per gram fresh weight of the fruit. For each plant the average Brazzein content of the two fruits was calculated. In some plants Brazzein could not be detected in the fruit using the ELISA technique. Western blot analysis of the total protein extract from some of the fruit revealed a 6.5kD protein band. which matches the predicted size of the mature Brazzein protein. This confirmed that the fruit contained Brazzein and that the signal peptide had been cleaved as if the signal peptide had not been cleaved, one would expect the protein to be larger. The Brazzein in fruit from plants which had been transformed with a construct lacking a signal peptide was not detected by Western blot. This is because the Brazzein content in these fruit is below the level of detection by western blot. ELISA is a more sensitive technique than western blot and protein was detected in these fruit by this method.

The results are summarised in Table 1 below and illustrated in Figure, in which the constructs pZPS34 to pZPS39 are as described in Example 1.

What the results show, on a simplified interpretation of the data, is: Comparing pZPS35 (Row 2 in Table 1) with its control pZPS34 (Row 1) the maximum concentration of brazzein obtained increased from about 25ng/g to about 226ng/g when the Dahlia signal peptide was present. Similarly, comparing pZPS36 (Row 3) with Row 1 shows an increase from about 25 ng/g to about 798ng/g when the radish signal peptide is present. The three constructs involved here had the UBQ promoter.

Comparing pZPS38 (Row 5 in Table 1) with its control pZPS37 (Row 4) the maximum concentration of brazzein obtained increased from about 13ng/g to about 51745ng/g when the Dahlia signal peptide was present. Similarly, comparing pZPS39 (Row 6) with Row 4 shows an increase from about 13ng/g to about 56932ng/g when the radish signal peptide is present. The three constructs involved here had the PG promoter TABLE 1 Construct Promoter Signal Peptide No. of No. of Max Brazzein Min Brazzein Mean Brazzein in Standard ß Name Plants Plants ng/g Fresh wt ng/g Fresh wt plants expressing Deviation Tested expressing the gene Brazzein 1pZPS34 UBQ None 29 18 25. 57 Not Detected 6. 85 6. 96 2pZPS35 UBQ Dahlia AMP 1 25 23 226. 5 Not Detected 43. 8 59.46 3pZPS36 UBQ Radish AMP l 30 29 798. 62 Not Detected 126. 67 179. 2 4pZPS37 PG None 15 12. 77 Not Detected 3. 32 4. 47 5pZPS38 PG Dahlia AMPI 13 11 S 1745. 77 Not Detected 12867. 34 13806.89 6 pZPS39 PG Radish AMP1 S S 56932. 16 1758. 74 22540. 03 24553. 47

EXAMPLE 3 Expression of the sweet tasting protein Brazzein in Yeast.

Expression cassettes were constructed which were integrated by homologous recombination in the genome of the methyltrophic yeast Pichia pastoris (P pastoris).

The Mating Factor Alpha signal peptide (pre-peptide) from Saccharomyces cerevisiae was fused to the synthetic Brazzein gene (with codon usage optimised to that of yeast). The construct was under the expressional control of the AOX1 promoter (see Figure 1). The expression cassette containing the pre-peptide of MFa fused to Brazzein under the AOX1 promoter and terminator was cloned into the P pastoris transformation vector pPIC3.5K and with the resulting plasmid P pastoris was transformed by spheroblasting. This resulted in the integration of the expression cassette to the genome of P pastoris. The P pastoris GS 115 transformant containing multiple copies of the expression cassette, expressed Brazzein in the supernatant up to 200mg/l. A high copy transformant of P pastoris KM71 expressed Brazzein up to 70mg/l in shakeflasks. Brazzein was secreted from the transformed P pastoris into the culture medium and detected by ELISA. The Brazzein produced by the transformed P pastoris was purified and confirmed to be the correct size by mass spectometry. In addition, N terminal sequencing revealed that the signal peptide had been cleaved at the predicted position to release mature Brazzein identical in composition to that isolated from the fruit of P brazzeana.

Expression cassettes were also constructed containing the Mating Factor Alpha signal peptide (pre-peptide) from Saccharomyces cerevisiae ligated to the synthetic Brazzein gene (with codon usage optimised to that of yeast) under the expressional control of the constitutive Mating Factor Alpha promoter (see Figure 2). This expression cassette was cloned into the yeast transformation vector called pYIPlacl28. The resulting plasmid was used to transform the yeast S cerevisiae by integration of the expression cassette into the genome at the leu2 locus.. Yeast transformants containing the Brazzein gene were selected by the ability to grow without leucine. The Brazzein produced by the transformed yeast was secreted to the culture medium and was quantified by ELISA. Brazzein in Saccharomyces cerevisiae PMY1 was expressed at 5mg/l. Western blot analysis confirmed that the Brazzein expressed by S cerevisiae was of the correct size at 6.5kD.

EXAMPLE 4 Generation of transformed Musa plants.

Transformed Musa plants containing the brazzein expression vectors are produced by the method described in Sagi et al. (1995) Biotechnology. Vol. 13 pp 481-485. Regenerated transformed plants are identified by their ability to grow on hygromycin and grown to maturity. Ripening fruit are analysed for a modulation in their ripening related or senescence characteristics.

Other suitable transformation methods for banana are described in Sagi et al. (1994) Plant Cell Reports. Vol. 13. pp 262-266.

EXAMPLE 5 Generation of transformed banana plants Vectors are transferred to Agrobacterium tumefaciens LBA4404 (a micro-organism widely available to plant biotechnologists) and are used to transform banana plants.

Transformation of banana meristems follow the protocols described by May et al (1995, Biotechnology 13: 486-492). Transformed plants are identified by their ability to grow on media containing the antibiotic kanamycin. Plants are regenerated and grown to maturity.

Ripening fruit are analysed for modifications to their fruit ripening characteristics.