WO2000063399A1

WO2000063399A1 - Glycogen branching enzymes and carbohydrates modified thereby

Info

Publication number: WO2000063399A1
Application number: PCT/GB2000/001371
Authority: WO
Inventors: Mark Andrew Taylor; Howard Vivian Davies
Original assignee: Scottish Crop Research Institute
Priority date: 1999-04-17
Filing date: 2000-04-17
Publication date: 2000-10-26
Also published as: AU4577600A

Abstract

There are provided novel cDNA polynucleotide sequences obtained from Aureobasidium pullulans which encode novel carbohydrate modifying enzymes. Host cells and in particular transgenic plants transfected with recombinant genetic constructs comprising the novel cDNA sequences are described. The enzymes encoded by the cDNA polynucleotide sequences may be used to produce carbohydrates or to modify existing carbohydrates, and this may be achieved in vitro or in vivo (for example by a transgenic plant). The modified carbohydrates so produced form a further apsect of the invention.

Description

GLYCOGEN BRANCHING ENZYMES AND CARBOHYDRATES MODIFIED THEREBY

The present invention relates to novel enzymes involved in carbohydrate modification and synthesis, and also to genetic constructs coding for those enzymes, and to the carbohydrate products of the enzymes.

The glucose polymers glycogen and starch are natural storage carbohydrates. Glycogen is a major storage carbohydrate in a wide range of organisms ranging from bacteria to mammals whilst starch is the major plant storage carbohydrate. One of the major differences between the two polymers is that the mature glycogen molecule is much more highly branched than starch (Goldsmith et al . , 1982, J. Mol . Biol . 156, pages 411- 427) . Starch is a mixture of two glucose polymers, amylose and amylopectin. Amylose is an essentially linear polymer of αl , 4 linked glucose monomers whereas amylopectin contains αl , 6 linked branch points within the αl , 4 linked chain (Banks and Greenwood, 1975, "Starch and its components", Edinburgh, UK, University Press, pages 242-305) . The frequency of the branch points and the size distribution of the sidechains in amylopectin are major determinants of the physio- chemical properties of the starch (reviewed m Smith and Martin, 1993, starch biosynthesis and the potential for its manipulation m "Biosynthesis and Manipulation of Plant Products", Vol. 3, Plant Biotechnology Series (Gπerson ed) London, Blackie A & P, pages 1-54) . Although the mechanism of amylopectin biosynthesis is not fully understood, it is known that the formation of these αl , 6 branch points is catalysed by starch branching enzymes. Although the degree of amylopectin branching is highly dependent on the properties of the branching enzymes (Tolmasky and Krisman, 1987, Eur. J. Biochem. 168 :393-397) , the activity of a debranchmg enzyme may also influence amylopectin structure (reviewed by Nakamura, 1996, Plant Science 121 :1-18) . Based on ammo acid sequence homology, it has been proposed recently that m plants there are two families of starch branching enzyme (Burton et al . , 1995, Plant J. 7:3-15) .

Glycogen branching enzymes from a diverse range of organisms also exhibit significant sequence similarities to the starch branching enzymes. In most plants in which the biosynthesis of starch has been studied m detail (for example, maize, rice and pea) , the two families of starch branching enzyme and their corresponding genes (SBEI and SBEII) have been isolated and characterised. The different families of starch branching enzyme have been shown to be expressed at different developmental stages. In developing pea embryos for example, SBEI is expressed earlier than SBEII. Data from developing pea embryos also suggest that the two branching enzyme families catalyse the transfer of branches of different lengths (Burton et al . , 1995, supra) . In maize it has been demonstrated that SBEII transfers shorter branches than SBEI when branching amylose in vivo (Takeda et al . , 1993, Carbohydrate Res. 24_0: 253-263) . Expression of bacterial glycogen branching enzymes in plants has been demonstrated to have an effect on starch branching. For example when an E. coli glgB gene (encoding glycogen branching enzyme) was expressed in an amylose free mutant of potato, the number of branchpoints in starch isolated from tubers of the transgenic plants was increased by up to 25% (Kortstee et al . , 1996, Plant J. 10:83-90) .

Many micro-organisms produce glucose polymers other than glycogen. For example some strains of Aureobasidium pullulans , a ubiquitous saprophyte, produce, under certain environmental and developmental conditions, in addition to glycogen, the extracelluar glucan, pullulan (Finkelman and Vardanis, 1987, CRC Critical Reviews in Biotechnology 5:185-193). Pullulan is thought to consist of maltotriose and (more rarely) maltotetraose units comprised of three or four α-1,4- linked glucose molecules respectively, that are polymerised in a linear fashion via α-1 , 6-linkages (Catley, 1979, Pullulan Biosynthesis by Aureobasidium Pullulans, In "Microbial polysaccharides and polysaccharases" (Berkley et al . eds . ) , Academic Press, London, pages 62-84) . Pullulan has a wide range of applications including use as packaging film, colourless adhesive, blood plasma substitute and food additives (Tsujisaka and Mitsuhashi, 1993, Pullulan, In "Microbial Gums", 3rd Edition, Academic Press). Despite the commercial importance of pullulan, little is known about its biosynthesis.

The present Application describes the isolation of a cDNA clones from Aureobasidium pullulans . The expression product of the cDNA clones affect glycogen branching and also the branch structure of starch.

The cDNA clones were each expressed in the E. coli mutant KV832 (which lacks the usual branching enzyme activity) and which had been modified with E. coli glgC to ensure high levels of glycogen synthesis (hereinafter "glgC modified KV82"). The glycogen subsequently produced was shown to have modified branching characteristics. The cDNA clones may be used to transform potato plants; the amylose content of the starch may be altered in the transgenic plants. In one particular cDNA clone (herein termed " 2" - see SEQ ID No 4) transgenic plants were produced and the amylose content of the starch was found to be altered.

In one aspect, the present invention provides a polynucleotide comprising a nucleotide sequence as set out in any one of SEQ ID Nos 1 to 4. The amino acid sequences encoded by the nucleotide sequences of SEQ ID Nos 1 to 4 may be determined in accordance with standard techniques. For SEQ ID No 4 the ammo acid sequence is given in SEQ ID No 5. Thus in a further aspect, the present invention provides a polypeptide encoded by a protein encoding portion of any one of the nucleotide sequences set out in SEQ ID Nos 1 to 4 and provides a polypeptide having an amino acid sequence as set out in SEQ ID No 5.

Those skilled in the art will appreciate that the nucleotide sequences of SEQ ID Nos 1 to 4 correspond to one allele of the cDNA genes, and that allelic variation is likely. Allelic variants can be cloned by probing cDNA or genomic libraries according to standard procedures. Allelic variants of the DNA sequence shown in SEQ ID Nos 1 to 4 , including those containing silent mutations and those in which mutations result in amino acid sequence changes, are within the scope of the present invention, as are proteins which are allelic variants of those encoded by SEQ ID Nos 1 to 4.

The present invention also includes modified sequences retaining or improving the biological function of the reference sequence or peptide (ie the ability to modify carbohydrates) and having at least 70% homology (preferably 80% homology, especially preferably 85-90% homology) with the nucleotide sequence in question. Thus, functional equivalents of such polynucleotides are also part of this invention. In particular, we include nucleotide substitutions which do not affect the amino acid expressed. The term "functional equivalent" used herein refers to any derivative in which nucleotide base(s) and/or amino acid(s) have been added, deleted or replaced without a significantly adverse effect on expression of the gene product or on biological function thereof. Thus, for example, amino acid Glu may be encoded by the codon gag or by the codon gaa and each construct may be varied in such a way without affecting the sequence of the expressed peptide.

Modifications also covered include polynucleotides incorporating intron sequences as well as domain rearrangement proteins, fusion or chimeric proteins and the like.

The polynucleotides may be in any form (for example DNA or RNA, double or single stranded) , but generally double stranded DNA is the most convenient. Likewise the polynucleotides according to the present invention may be present as part of a recombmant genetic construct, which itself may be included in a vector (for example an expression vector) and eukaryotic or prokaryotic vectors (as well as plant cell vectors) are of interest. Alternatively, the construct may be incorporated into the genome of a transgenic plant. Any vectors or transgenic plants comprising a polynucleotide as described above form a further aspect of the present invention.

A vector including the 2 coding sequence has been deposited at European Collection of Cell Cultures (ECACC) , Salisbury, Wiltshire, SP4 OJG, United Kingdom on 15 April 1999 under No 99041515. This deposit forms a further aspect of the present invention. In one embodiment the present invention provides a recombmant genetic expression system comprising and able to express a polynucleotide described above.

The term "expression system" is used herein to refer to a genetic sequence which includes a protem-encodmg region and is operably linked to all of the genetic signals necessary to achieve expression of the protein encoding region. Optionally, the expression system may also include a regulatory element, such as a promoter or enhancer, to increase transcription and/or translation of the protem-encodmg region, or to provide control over expression. The regulatory element may be located upstream or downstream of the protem-encodmg region, or may be located at an mtron (non-coding portion) interrupting the protein encoding region. Alternatively the regulating protein may be expressed from a gene encoded on another plasmid (optionally stably integrated m the genome) , acting m trans on a regulatory element situated upstream or downstream of the protein encoding sequence which may be present on another plasmid (optionally also stably integrated elsewhere m the genome) . It is also possible for the sequence of the protem-encodmg region itself to comprise a regulatory ability.

Genetic engineering has been recognised as a powerful technique not only m research but also for commercial purposes. Thus, by using genetic engineering techniques (see Maniatis et al Molecular Cloning, a Laboratory Manual Cold Spring, Harbor Laboratory, Cold Spring Harbor, New York 1982 and "Principle of Genetic Engineering", Old and Primrose, 5th edition, 1994, both incorporated herein by reference) exogenous genetic material can be transferred to a host cell and the protein or polypeptide encoded by the exogenous genetic material may be replicated by and/or expressed withm the host. For the purposes of simplicity genetic engineering is normally carried out with prokaryotic micro-organisms, for example bacteria such as E. coli , as host. However, use has also been made of plant cells and of eukaryotic organisms, in particular yeasts or algae, and m certain applications plant cell or eukaryotic cell cultures may also be used.

Additionally, the present invention comprises a vector containing such a recombmant expression system and host cells transfected with such a recombmant expression system (optionally m the form of a vector) .

Suitable host cells will depend upon the vector and there is now a vast amount of information available to those skilled m the art on the selection of host cells. The invention encompasses microbial, animal and plant host cells. Preferred host cells include those from monocotyledonous plants such as maize, wheat, rice and barley and those from dicotyledonous plants such as potato, tomato, etc. Additionally, yeast and algal cells and prokaryotic cells may be suitable hosts.

In particular one preferred embodiment is a stable cell line able to express one or more of the cDNA clones herein described (for example the W2 clone) under the influence of a plant storage organ specific promoter system, especially the patatm and granule bound starch synthase promoters from potato or constitutive promoters such as the ubiquitm promoter or the 35S Cauliflower Mosaic Virus promoter.

In a further aspect, the present invention provides a transgenic plant, said plant having cells incorporating a recombmant expression system adapted to express one or more of the cDNA clones herein described (for example the W2 clone) . Generally the recombmant expression system will be stably integrated into the genome of the transgenic plant and will thus be heritable so that the offspring of such a transgenic plant may themselves contain the transgene and thus also be covered by the present invention. Suitable transgenic plants may be monocotyledonous or dicotyledonous and include (but are not limited to) maize, wheat, rice, barley, potato, tomato, etc.

In a yet further aspect, the present invention further provides carbohydrate (in particular glucose polymer carbohydrates) produced or modified by the expression product of one of the polynucleotides of the present invention.

We envisage that the carbohydrate could be synthesised m vi tro through use of a combination of enzymes, including any one of the novel polypeptides, for example the novel W2 polypeptide. The enzymes could be simply incubated with suitable simple sugar units (for example maltose) . Thus, the present invention provides a process for producing a modified carbohydrate wherein simple sugar units are incubated with carbohydrate forming enzymes, wherein at least one such enzyme is a polypeptide as described above .

Alternatively, the enzyme could be expressed in transgenic plants and act upon and affect the in vi vo synthesis of carbohydrate.

The branching structure of the carbohydrate will be affected by the enzyme encoded by the polynucleotides of SEQ ID Nos 1 to 4; the W2 clone is of particular interest in this regard.

The present invention will now be further described with reference to the examples and accompanying sequence listing and figures, in which:

Sequence Listing

SEQ ID No 1: Nucleotide sequence of the cDNA clone of the gf8 gene.

SEQ ID No 2 : Nucleotide sequence of the cDNA clone of the gfl gene.

SEQ ID No 3: Nucleotide sequence of the cDNA clone of the b3 gene.

SEQ ID No 4: Nucleotide and amino acid sequences of the cDNA clone of the W2 gene. SEQ ID No 5 : Amino acid sequence of W2.

Figure Legends

Fig.l Iodine vapour stained colonies of E. coli strain KV832 and KV832 transformed with W2 encoding plasmid. Typical examples labelled A and B respectively.

Fig.2 HPAEC analysis of the debranched α-glucans isolated from E. coli strains (A) KV832, (B) KV832 transformed with W2 and (C) DH5α.

EXAMPLE 1 ; Production of 2 clone and effect on E. coli

Materials and Methods

Growth of Aureobasidi um pullulans .

Aureobasidium pullulans strain IBT-89, a colourless mutant strain that produces high yields of pullulan, was kindly provided by L. Tarabasz-Szymanska and E. Galas, Institute of Technical Biochemistry, Technical University of Lodz, Poland (Tarabasz-Szymanska and Galas, 1993, Enzyme Microb. Technol . 15:317-320) . Cultures were grown at 28°C with shaking, m liquid medium (2g/l yeast extract, 0.5g/l (NH₄)₂S0₄, lg/1 NaCl , 0.2g/l MgS0₄, 3g/l K₂HP0₄, O.Olg/l FeS0₄, O.Olg/l MnS0₄, O.Olg/l ZnS0₄, 20g/l sucrose, pH 6.0) . Under these conditions, pullulan was produced withm 24 hours of sub-culturmg with a 5% mnoculum.

Construction of a cDNA library from Aureobasidium pullulans .

RNA was extracted from cells of Aureobasidium pullulans from cultures that were producing pullulan (see above) . Cells were harvested from 500 mis of culture by centπfugation (6000g, 10 mms) and washed twice m sterile distilled water. The pelleted cells were transferred to a pre-cooled pestle and were frozen m liquid nitrogen prior to grinding. RNA was extracted using Qiagen columns (Qiagen Ltd, UK) following the manufacturer's protocol. The yield of total RNA was approximately lmg from 500ml of culture. Polyadenylated RNA was purified from total RNA using oligo dT spin column chromatography (Pharmacia) . A cDNA library was constructed from lOμg of polyadenylated RNA using a cDNA synthesis kit and the Uni-Zap XR vector (Stratagene) . The library contained approximately 1 x 10⁶ primary clones and the average insert size was approximately lkb.

Isolation of cDNA clones from the Aureobasidium pullulans library

The cloned inserts from the A . pullulans lambda library were excised in vivo to form a representative phagemid library following a standard protocol (Stratagene) . The glycogen branching enzyme-deficient E. coli strain KV832 (Kiel et al . , 1987, Mol . Gen. Genet. 207:294-301) was transformed with E. coli gig C so that the strain produced high yield of glycogen. This modified KV832 strain was then transformed with an aliquot of the phagemid library by electroporation (Dower et al . , 1988, Nucl. Acid. Res. 16:6127-6145). Transformants were plated on rich medium (YT supplemented with 1% glucose, 0.5mM IPTG, lOOmg/ml ampicillin, 25mg/ml chloramphenicol and 50mg/ml kanamycin) at a density of 10,000 colonies per 150mm diameter plate. Colonies were stained with iodine vapour and those that stained differently (ranging from pale yellow to brown) to the blue colour, indicative of unbranched α-glucan, were selected for further rounds of screening by preparing plasmid DNA from the colony (Birnboim and Doly, 1979, Nucl. Acid. Res. =1513) and re-transforming the modified KV832 strain. Clones that stained pale yellow consistently through three rounds of screening were selected for further analysis. One clone with these characteristics (W2) was isolated from screening approximately 50,000 colonies. Plasmid DNA isolated from this colony was transformed into E. coli strain DH5α (Life Technologies) , purified from this strain and insert sequence (both strands) was obtained by cycle sequencing (DyeDeoxy Terminator kit, Perkin Elmer) and a 373 automated DNA sequencer (Applied Biosystems) . DNA sequence analysis was carried out using software available on the SEQNET Computational Molecular Biology Facility at SERC Daresbury Laboratory UK. The nucleotide sequence of the W2 clone is given in SEQ ID No 4 and the amino acid sequence of the protein encoded is given in SEQ ID No 5.

Analysis of glycogen branching structure

Glycogen was purified from cultures of E. coli strains (glgC modified KV832, glgC modified KV832 transformed with W2 , and wild type (DH5α) ) . Overnight cultures were grown in LB medium supplemented with 1% glucose, 0.5mM IPTG and the antibiotics as follows: Wild type (DH5α) : no antibiotics; glgC modified KV832 : 25 mg/ml chloramphericol and 50 mg/ml kanamycin; glgC modified KV832 transformed with W2 : 25 mg/ml chloramphericol, 50 mg/ml kanamycin and 100 mg/ml ampicillin. The overnight cultures were then harvested by centπfugation, and resuspensed m 50mM Tπs acetate buffer, pH 7.5 at a density of lg wet weight cells per 2ml buffer. Following sonication, glycogen was extracted as described by Preiss et al . , 1975, J. Biol . Chem. 250 : 7631-7638. Debranched α-glucan was prepared by treatment of α-glucan with isoamylase (Megazyme) as described by Guan et al . , 1995. Chamlength distribution of the debranched α-glucan was determined by high-performance anion-exchange chromatography (HPAEC) using a Dionex DX500 system with a pulsed amperometric detector and an amperometπc flow-through cell with a gold working electrode. The column and gradient programme were as described by Guan et al . , 1995, PNAS USA 92:964-967.

The type and quantity of linkages present m the α- glucans isolated from the different E. coli strains was determined by methylation analysis (Needs and Selvendran, 1993, Photochemical Analysis 4:210-216). The products of methylation were analysed on a Hewlett Packard GC/MS using a SGE BP5 30m x 0.32mm ι.d., 0.25mm capillary column. The temperature programme was as follows: 400°C for 3 rnms rising to 1300°C at 200°C per minute, then rising to 2300°C at 40°C per minute then remaining at 2300°C for ten minutes. The peaks and mass breakdown of the mam chain αl-4 and branch point αl- 4,6-1 linked residues were identified by comparison with commercial potato amylose and amylopectin (Sigma, UK) . Measurement of starch branching enzyme activity

Cells from overnight cultures of E. coli were harvested by centrifugation and resuspended m 0.5 M sodium citrate buffer pH 6.4. Following somcation, the supernatant following centrifugation (12,000g, 5 mms) was passed through a Sephadex G-25 column (PD-10, Pharmacia) and assayed for branching enzyme activity (A. Korstee, PhD Thesis, University of Wagenmgen, Netherlands, 1997) . This assay is based on the rate of decrease m absorbance maxima of the lodme/glucan complex, on branching of the amylose substrate.

RESULTS

Isolation of a glycogen branching enzyme encoding cDNA

Iodine staining of bacterial or yeast colonies is an effective method for the rapid screening of genes that effect glucan metabolism. For example, functional complementation of glycogen branching enzyme deficiency in a Saccharomyces cerevi siae mutant has led to the cloning of glycogen branching enzymes from Saccharomyces and human (Rowen et al . , 1992, Molecular and Cellular Biology 12:22-29; Thon et al . , 1993, J. Biol . Chem. 268 : 7509-7513) . In this study, a variation of the iodine staining technique, using an E . coli branching enzyme-deflcient mutant, was employed to isolate cDNAs encoding enzymes involved m glucan metabolism m the yeast-like fungus, Aureobasidium pullulans . One clone was isolated that gave consistently pale-sta mg colonies on treatment with iodine vapour compared with the dark blue staining normally observed with the modified KV832 strain (Fig. 1) . The plasmid of Aureobasidium pullulans origin contained m this clone was transformed into E. coli strain DH5α and the insert was sequenced (SEQ ID No 4) . An estimate of the size of the transcript from Northern analysis indicates that the clone is near to full length (data not shown) . The longest open reading frame (also indicated m SEQ ID No 4) extends from an ATG codon (A position 56) to a TAG translation termination codon (T position 710) - see SEQ ID No 5. The open reading frame encodes a polypeptide of molecular weight 23,451 and an isoelectric point of 6.4. The closest match m sequence similarity of this deduced sequence is with an open reading frame from chromosome VIII of Saccharomyces cerevisiae (encoded by the gene YHR076w, Johnston et al . , 1994, Science 265:2077-2082) . No function has been ascribed to the polypeptide encoded by the YHR076w gene located m the ERG7-NMD2 intergenic region, although the open reading frame encodes a polypeptide of 41.2 kD . There is 36% identity and 57% similarity between the YHR076w gene product and the sequence encoded by the W2 clone.

Characterisation of the branching enzyme activity encoded by the W2 gene

In order to demonstrate that the W2 gene does encode glycogen branching activity, the W2 cDNA was expressed the modified E. coli mutant, glgC modified KV832. The glycogen produced on induction of W2 gene expression was analysed and compared with glycogen isolated from strains of E . coli that produce branched glycogen and also from the non- transformed strain. Analysis of branching pattern clearly demonstrated that glycogen was branched to a greater extent than in the non- transformed host strain (Fig. 2) in which very little glycogen branching could be detected. However, the degree of branching and distribution of chain lengths on transformation with W2 was considerably different to that observed in glycogen isolated from the branching competent DH5α strain. The chain length (CL) distribution of the glycogen isolated from the DH5α strain of E. coli was similar to that previously detected in E. coli B (Guan et al . , 1995, PNAS USA 92:964-967) with CLs 6-8 being the most abundant and CLs greater than 10 relatively reduced on a molar basis. Glycogen from the W2 transformant, on the other hand, had the highest frequency of branches with CL 3 and 6 and relatively few longer branch points. The longer chains (CL > 25) detected in glycogen from the W2 transformants and in the glgC modified KV832 strain were present in samples that had not been treated by isoamylase and so are probably long chain contaminants of the preparations. The extent of branching was also considerably less in the W2 transformant than in DH5α.

Methylation analysis of the glycogen produced by these strains supported the HPAEC chromatography data and confirmed the presence of branch point residues m the glycogen isolated from the W2 transformed strain (Table 1) . Table 1 Methylation analysis of α-glucans isolated from E . coli strains.

Strain Ratio of peak areas (αl, 4/1-6: αl,4) glgC modified KV832 0.000 glgC modified KV832 transformed with W2 0.111 Wild type (DH5α) 0.314

The frequency of the branch point residue was approximately one third that observed in the glycogen from the DH5α strain. No branch point residues could be detected in the host glgC modified KV832 strain. No account was taken for differences in detector response factors for the main chain and branch point residues. However, due to the structural similarity of the residues it is unlikely that they are very different and have affected the estimation of branch point residues.

A direct assay of starch branching activity in the E. coli strains was also carried out (Table 2) . Starch branching activity measured in the W2 transformed strain was significantly greater than in the non- transformed KV832 strain (no activity detectable) and the specific activity was approximately 30% of the level measured in DH5α . Table 2 Branching enzyme activity measured in E . coli strains .

Specific activity units* glgC modified KV832 0.00 glgC KV832 transformed with W2 0.21 Wild type (DH5α) 0.72 * nm red shift in absorbance maximum of the amylose substrate per minute per mg protein.

Discussion

Several lines of evidence suggest that the W2 gene product is involved in glucan metabolism. Firstly, the basis of selection of the clone, a frequently used methodology for isolating genes involved in glucan metabolism, clearly indicated changed iodine staining in the branching enzyme-deficient strain of E. coli on expression of the W2 gene. In fact the colonies stained very pale yellow on treatment with iodine vapour (Fig. 1) . Expression of the maize branching enzyme II gene in the same strain led to a much darker brown staining (data not shown) . The second line of evidence that demonstrates that the W2 gene encodes glycogen branching activity comes from analysis of the glycogen produced in the glgC modified KV832 strain transformed with the W2 gene. HPAEC chromatography of the isoamylase treated glycogen shows a small (approximately 10% of the wild type level) degree of branching. However, this level of branching is significant because in the untransformed glgC modified KV832 strain the degree of branching is very low (approximately 1% of that observed the wild type) . Polysacchaπde methylation followed by GC-MS analysis of the products is a powerful method for the determination of the type of linkages present in carbohydrates (Needs and Selvendran, 1993, Phytochemical Analysis 4:210-216). Analysis of glycogen preparations confirmed that on induction of the W2 gene m the glgC modified KV832 mutant, αl , 6 linkages could be detected. The frequency of these linkages was approximately one third that observed on analysis of glycogen from the DH5α strain. No αl , 6 linkages could be detected by this analysis m glycogen isolated from the glgC modified KV832 mutant. Hence the methylation analysis is broadly consistent with the HPAEC data. Branching enzyme activity could be measured m extracts of the E. coli mutant following complementation with the W2 gene, further evidence that W2 does encode branching enzyme activity.

The levels of branching enzyme activity detected, and the degree of branching measured m the glycogen isolated from the complemented KV832 strain were much less than m a similar study with maize starch branching enzymes (Guan et al . , 1995, supra) . The specific activity of branching enzyme m the mutant transformed with the maize genes was increased 4-30 fold compared with the wild type. This may indicate that the branching activity encoded by the W2 gene is not the mam function of the W2 gene product. Alternatively, it may be that the W2 gene product does not function efficiently in a prokaryotic background.

The structures of the glycogen and starch branching enzymes are predicted to be related to that of the α- amylases (Jesperson et al . , 1993, J. Protein Chem. 12:791-805). The catalytic domain of the α-amylases contains a characteristic (β/α) 8 barrel domain and a similar secondary structure pattern has been predicted for the glycogen and starch branching enzymes. Alignments of starch and glycogen branching enzyme primary sequences show strong sequence conservation in most of those regions predicted to form the α-helical or β-sheet domains apart from in the carboxyl-terminal region (Burton et al . , 1995, Plant J. 7:3-15). The W2 gene product does not have significant sequence similarity to any previously characterised branching enzyme and the predicted folding of the W2 polypeptide does not fit the (β/α) pattern. This suggests that the mechanism of branching catalysed by the W2 gene product is entirely novel. It might be expected that the biosynthesis of pullulan involved different mechanisms to glycogen biosynthesis, consistent with the novel structure of the polypeptide encoded by the W2 gene.

EXAMPLE 2: Expression of the Aureo-basidiu-m pullulans 2 gene in potato

Gene constructs based on the W2 sequence were built in order to express the W2 gene in the plastid of transgenic potato. The aim was to produce starch with modified branching characteristics, and consequently modified functionality, due to the expression of W2 activity in the plastid where starch is made.

The construct contained the entire W2 open reading frame (see SEQ ID Nos 4 and 5) fused in-frame with the sequence encoding the potato granule bound starch synthase transit peptide which targets proteins to the plastid and the potato granule bound starch synthase promoter. Granule bound starch synthase (GBSS) is the enzyme responsible for amylose production in the potato. A pUC18 plasmid carrying an EcoRI fragment of the genomic clone of potato GBSS including both the promoter and transitpeptide region has been described in the literature (see van der Leij et al . , 1991, Theor. Applied Genetics 8_2: 289-295 and van der Leij et al., 1991, Mol . Gen. Genet. 228:240-248) and could be used to produce the present construct.

Potato (cv. Desiree) was transformed with the constructs and 20 independent transgenic lines (as determined by Southern Analysis) were produced.

Chain length analysis by HPAEC did not reveal significant differences between starch from any of the W2 transformants and "empty-vector-transformed" controls.

Analysis of amylose/amylopectin content of the starch was determined using a Megazyme amylose/amylopectin assay kit. The amylose content was lower in a number of transgenic lines with 4 of 12 lines analysed containing less than 50% of the amylose content of starch from control lines.

The low amylose content is reminiscent of starch produced in plants in which the granule bound starch synthase activity is down-regulated. It is possible that the low amylose effect is due to co-suppression of the potato GBSS as sequences from this gene were used in making the construct. Additional constructs containing a constitutive promoter (35SCaMV) and plastid targeting sequence (RuBISCo small sub-unit) not related to the granule bound starch synthase have been built. These constructs have been introduced into potato by AgroJbacterium-mediated transformation and if the low amylose effect is still observed, it is likely to be due to the effect of W2 expression.

EXAMPLE 3 ; Production of gf8, gfl and b3 clones and effect on E. coli

Materials and Methods

Growth of Aureobasidium pull ulans .

Aureobasidium pullulans strain IBT-89, a colourless mutant strain that produces high yields of pullulan, was kindly provided by L. Tarabasz-Szymanska and E. Galas, Institute of Technical Biochemistry, Technical University of Lodz, Poland (Tarabasz-Szymanska and Galas, 1993) . Cultures were grown at 28°C with shaking, in liquid medium (2g/l yeast extract, 0.5g/l (NH₄)₂S0₄, lg/1 NaCl, 0.2g/l MgS0₄, 3g/l K₂HP0₄ , O.Olg/l FeS0₄, O.Olg/l MnS0₄, O.Olg/l ZnS0₄, 20g/l sucrose, pH 6.0). Under these conditions, pullulan was produced within 24 hours of sub-culturing with a 5% innoculum.

Construction of a cDNA library from Aureobasidium pullulans .

RNA was extracted from cells of Aureobasidium pullulans from cultures that were producing pullulan (see above) . Cells were harvested from 500 mis of culture by centrifugation (6000g, 10 mins) and washed twice in sterile distilled water. The pelleted cells were transferred to a pre-cooled pestle and were frozen in liquid nitrogen prior to grinding. RNA was extracted using Qiagen columns (Qiagen Ltd, UK) following the manufacturer's protocol. The yield of total RNA was approximately lmg from 500ml of culture. Polyadenylated RNA was purified from total RNA using oligo dT spin column chromatography (Pharmacia) . A cDNA library was constructed from lOμg of polyadenylated RNA using a cDNA synthesis kit and the Uni-Zap XR vector (Stratagene) . The library contained approximately 1 x 10⁶ primary clones and the average insert size was approximately lkb.

Isolation of cDNA clones from the Aureobasidium pullulans library

The cloned inserts from the A . pullulans lambda library were excised in vivo to form a representative phagemid library following a standard protocol (Stratagene) . The glycogen branching enzyme-deficient E . coli strain KV832 (Kiel et al . , 1987) was transformed with E. coli gig C so that the strain produced high yield of glycogen. This modified KV832 strain was then transformed with an aliquot of the phagemid library by electroporation (Dower et al . , 1988) . Transformants were plated on rich medium (YT supplemented with 1% glucose, 0.5mM IPTG, lOOmg/ml ampicillin, 25mg/ml chloramphenicol and 50mg/ml kanamycin) at a density of 10,000 colonies per 150mm diameter plate. Colonies were stained with iodine vapour and those that stained differently (ranging from pale yellow to brown) to the blue colour, indicative of unbranched α-glucan, were selected for further rounds of screening by preparing plasmid DNA from the colony (Birnboim and Doly, 1979) and re-transforming the modified KV832 strain. Clones that stained pale yellow consistently through three rounds of screening were selected for further analysis. Three clones with these characteristics (gf8, gfl and b3 ) were isolated from screening approximately 50,000 colonies. Plasmid DNA isolated from this colony was transformed into E. coli strain DH5α (Life Technologies) , purified from this strain and insert sequence (both strands) was obtained by cycle sequencing (DyeDeoxy Terminator kit, Perkin Elmer) and a 373 automated DNA sequencer (Applied Biosystems) . DNA sequence analysis was carried out using software available on the SEQNET Computational Molecular Biology Facility at SERC Daresbury Laboratory UK. The nucleotide sequence of the clones is given in SEQ ID Nos 1 to 3.

Analysis of glycogen branching structure

Glycogen was purified from cultures of E. coli strains (glgC modified KV832, glgC modified KV832 transformed with gfδ, gfl or b3 and wild type (DH5α) ) . Overnight cultures were grown in LB medium supplemented with 1% glucose, 0.5mM IPTG and the antibiotics as follows: Wild type (DH5α) : no antibiotics; glgC modified KV832 : 25 mg/ml chloramphericol and 50 mg/ml kanamycin; glgC modified KV832 transformed with gfδ, gfl or b3 : 25 mg/ml chloramphericol, 50 mg/ml kanamycin and 100 mg/ml ampicillin.

The overnight cultures were then harvested by centrifugation, and resuspended in 50mM Tris acetate buffer, pH 7.5 at a density of lg wet weight cells per 2ml buffer. Following sonication, glycogen was extracted as described by Preiss et al . , 1975, J. Biol . Chem. 250 : 7631-7638. Debranched α-glucan was prepared by treatment of α-glucan with isoamylase (Megazyme) as described by Guan et al . , 1995. Chainlength distribution of the debranched α-glucan was determined by high-performance anion-exchange chromatography (HPAEC) using a Dionex DX500 system with a pulsed amperometric detector and an amperometric flow-through cell with a gold working electrode. The column and gradient programme were as described by Guan et al . , 1995, PNAS USA 92 :964-967.

The type and quantity of linkages present in the α- glucans isolated from the different E. coli strains was determined by methylation analysis (Needs and Selvendran, 1993, Photochemical Analysis 4:210-216). The products of methylation were analysed on a Hewlett Packard GC/MS using a SGE BP5 30m x 0.32mm i.d., 0.25mm capillary column. The temperature programme was as follows: 400°C for 3 mins rising to 1300°C at 200°C per minute, then rising to 2300°C at 40°C per minute then remaining at 2300°C for ten minutes. The peaks and mass breakdown of the main chain αl-4 and branch point αl- 4,6-1 linked residues were identified by comparison with commercial potato amylose and amylopectin (Sigma, UK) .

Measurement of starch branching enzyme activity

Cells from overnight cultures of E. coli were harvested by centrifugation and resuspended in 0.5 M sodium citrate buffer pH 6.4. Following sonication, the supernatant following centrifugation (12,000g, 5 mins) was passed through a Sephadex G-25 column (PD-10, Pharmacia) and assayed for branching enzyme activity (A. Korstee, PhD Thesis, University of Wageningen, Netherlands, 1997) . This assay is based on the rate of decrease in absorbance maxima of the iodine/glucan complex, on branching of the amylose substrate. RESULTS

Isolation of a glycogen branching enzyme encoding cDNA

Iodine staining of bacterial or yeast colonies is an effective method for the rapid screening of genes that effect glucan metabolism. For example, functional complementation of glycogen branching enzyme deficiency in a Saccharomyces cerevisiae mutant has led to the cloning of glycogen branching enzymes from Saccharomyces and human (Rowen et al . , 1992, Molecular and Cellular Biology .12:22-29; Thon et al . , 1993, J. Biol . Chem. 268:7509-7513) . In this study, a variation of the iodine staining technique, using an E . coli branching enzyme-deficient mutant, was employed to isolate cDNAs encoding enzymes involved in glucan metabolism in the yeast-like fungus, Aureobasidium pull ulans . Clones were isolated that gave consistently pale-staining colonies on treatment with iodine vapour compared with the dark blue staining normally observed with the modified KV832 strain. The plasmids of Aureobasidi um pullulans origin contained in these clones were transformed into E . coli strain DH5α and the inserts were sequenced (SEQ ID Nos 1 to 3, respectively) .

Characterisation of branching enzyme activity

In order to demonstrate that the cloned genes encode glycogen branching activity, the cDNA was expressed in the modified E . coli mutant, glgC modified KV832.

Methylation analysis of the glycogen produced by these strains confirmed the presence of branch point residues in the glycogen isolated from the transformed strains (Table 3) .

Table 3 Methylation analysis of α-glucans isolated from E. coli strains .

Strain Ratio of peak areas (αl, 4/1-6: αl , 4 ) glgC modified KV832 0.000 glgC modified KV832 transformed with gf8 0.000 glgC modified KV832 transformed with gfl 0.007 glgC modified KV832 transformed with b3 0.005 DH5α (wild type) 0.314

No branch point residues could be detected in the host glgC modified KV832 strain. No account was taken for differences in detector response factors for the main chain and branch point residues. However, due to the structural similarity of the residues it is unlikely that they are very different and have affected the estimation of branch point residues. Discussion

Several lines of evidence suggest that the gf8, gfl and b3 gene products are involved in glucan metabolism. Firstly, the basis of selection of the clone, a frequently used methodology for isolating genes involved in glucan metabolism, clearly indicated changed iodine staining in the branching enzyme- deficient strain of E. coli on expression. In fact the colonies stained yellow on treatment with iodine vapour. Expression of the maize branching enzyme II gene in the same strain led to a much darker brown staining (data not shown) . The second line of evidence that demonstrates that the gf8, gfl and b3 genes encode glycogen branching activity comes from analysis of the glycogen produced in the glgC modified KV832 strain transformed with the different gene.

EXAMPLE 4; Expression of the Aureo-basidit-un pullulans gf8, gfl and b3 genes in potato

Gene constructs based on the gf8, gfl and b3 sequences may be built in order to express the encoded genes in the plastid of transgenic potato. The aim is to produce starch with modified branching characteristics, and consequently modified functionality, due to the expression of gf8, gfl or b3 activity in the plastid where starch is made.

The constructs contained the entire open reading frame of one of gf8, gfl or b3 fused in-frame with the sequence encoding the potato granule bound starch synthase transit peptide which targets proteins to the plastid and the potato granule bound starch synthase promoter. A pUClδ plasmid containing an EcoRI fragment of the genomic clone of potato GBSS (including the promoter and transitpeptide region) has been described in the literature (see van der Leij et al . , 1991, Theor. Applied Genetics 82:289-295 and van der Leij et al., 1991, Mol . Gen. Genet. 228:240-248) and could be used to produce the present construct .

Potato (cv. Desiree) is transformed with the constructs and independent transgenic lines (as determined by Southern Analysis) produced.

Chain length analysis by HPAEC may be performed to reveal significant differences between starch from any of the transformants and "empty-vector-transformed" controls.

Analysis of amylose/amylopectin content of the starch may be determined using a Megazyme amylose/amylopectin assay kit.

Low amylose content may be reminiscent of starch produced in plants in which the granule bound starch synthase activity if down-regulated. It is possible that any low amylose effect could be due to co- suppression of the potato GBSS as sequences from this gene were used in making the construct. Additional constructs containing a constitutive promoter (35SCaMV) and plastid targeting sequence (RuBISCo small sub-unit) not related to the granule bound starch synthase may be built . These constructs may be introduced into potato by Agrobacterium-mediated transformation and if the low amylose effect is still observed, it is likely to be due to the effect of transgene expression.

Claims

1. A polynucleotide encoding a glycogen branching enzyme obtainable from Aureobasidium pullulans, said polynucleotide comprising a nucleotide sequence as set out in any one of SEQ ID Nos 1 to 4.

2. A polynucleotide as claimed in Claim 1 comprising the nucleotide sequence as set out in SEQ ID No 4.

3. A polynucleotide as claimed in Claim 1 comprising the W2 coding sequence forming part of the vector deposited under No 99041515 of ECACC .

4. A polypeptide encoded by a polynucleotide as claimed in any one of Claims 1 to 3.

5. A polypeptide having an amino acid sequence as set out in SEQ ID No 5.

6. A recombmant genetic construct comprising a polynucleotide as claimed in any one of Claims 1 to 3.

7. A recombmant genetic construct as claimed in Claim 6 in the form of an expression vector.

8. A recombinant genetic construct being the vector deposited under No 99041515 of ECACC.

9. Host cells transfected with a recombinant genetic construct as claimed in any one of Claims 6 to 8.

10. Host cells as claimed in Claim 9 selected from the group consisting of maize, wheat, rice, barley, potato, tomato and the like.

11. A stable cell line able to express the protein encoding region of a polynucleotide as claimed in any one of Claims 1 to 3 wherein the protein encoding region is under the influence of a plant storage organ specific promoter system.

12. A cell line as claimed in Claim 11 wherein said plant storage organ specific promoter system is selected from the group consisting of potato patatin promoter, potato granule bound starch synthase promoter, ubiquitin promoter and the 35S Cauliflower Mosaic Virus promoter.

13. A transgenic plant having cells incorporating a recombinant genetic construct as claimed in any one of Claims 6 to 8.

14. A transgenic plant as claimed in Claim 13 which is a maize, wheat, rice, barley, potato or tomato plant.

15. Use of a host cell as claimed in Claims 9 or 10, of a cell line as claimed in Claims 11 or 12, or of a transgenic plant as claimed in Claims 13 or 14 to produce a carbohydrate modifying enzyme.

16. Carbohydrate produced or modified by an expression product of a polynucleotide as claimed in any one of Claims 1 to 3.

17. Carbohydrate as claimed in Claim 16 which is a glucose polymer carbohydrate.

18. A process for producing a carbohydrate as claimed in either one of Claims 16 and 17 wherein single sugar units are incubated with carbohydrate forming enzymes, characterised in that at least one such enzyme is a polypeptide as claimed in either one of Claims 4 and 5.

19. A process for producing a carbohydrate as claimed in either one of Claims 16 and 17 wherein said carbohydrate is produced in vi vo by a transgenic plant as claimed in either one of Claims 13 and 14.