Erythorbic acid is a C-5 epimer of ascorbic acid and has essentially identical chemical properties including anti- oxidant activity. However, the vitamin C activity of erythorbic acid is very low compared to that of ascorbic acid and for practical purposes it is not considered to be a vitamin. Erythorbic acid has GRAS status and is used as an anti-oxidant in a number of food and other applications. Currently used chemical methods for the manufacture of erythorbic acid are based on esterification of 2-ketogluconic acid followed by base- catalyzed cyclization of the ester to produce sodium erythorbate.

It has been known since the 1960's that certain wild-type fungi belonging to the genus Penicillium produce small amounts of erythorbic acid when grown on glucose [Yagi J. et al. Agric. Biol. Chem. 31 (3) 340-345 (1967)]. Through an

extensive chemical mutagenisis/selection program Penicillium notatum strains capable of converting glucose to erythorbic acid with yields up to 40% have been isolated [Shimizu K. et al. Agric. Biol. Chem 31 (3) 346-352 (1967)]. However, the fermentation time needed to complete the conversion is very long (1-2 weeks) making the industrial application of this process impractical.

Subsequent studies of the enzymology of the glucose-to- erythorbic acid pathway in Penicillium have established that the pathway is comprised of two reactions [Takahashi T. Biotechnology and Bioengineering 11,1157-1171 (1969)]. The first reaction is the well-known oxidation of glucose to gluconolactone by glucose oxidase. The second reaction in this pathway is the oxidation of D- gluconolactone by molecular oxygen with the formation of erythorbic acid and hydrogen peroxide. This reaction is catalyzed by D-gluconolactone oxidase (D-GLO), an enzyme detected in only several fungal species. Takahashi and co-workers [Takahashi T. et al. Agric. Biol. Chem. 40, 121-129 (1976)] have elucidated the basic enzymological properties of D-GLO from a strain of Penicillium cyaneo- fulvum (subsequently re-classified as Penicilliw griseoroseum ATCC 1043).

The difficulties associated with the direct conversion of glucose to erythorbic acid at an economically acceptable rate remain unresolved. The present invention provides isolated nucleic acids encoding D-GLO which are useful in the biotechnological process for the production of erythorbic acid and its salts.

The present invention is directed to the isolation and identification of nucleic acid molecules which encode the enzyme D-gluconolactone oxidase (D-GLO) useful in the production of erythorbic acid and related salts.

Accordingly, in one embodiment, this invention is directed to newly isolated nucleic acid molecules defined by SEQ ID NO: 1 (cDNA), SEQ ID NO: 2 (coding) and SEQ ID NO: 3 (mature). The present invention further contemplates nucleic acid molecules which hybridize under stringent conditions to any one of SEQ. ID. NOS. 1-3.

In another embodiment of the present invention, vectors containing any one of the nucleic acid molecules identified herein, as well as host cells transformed with such vectors are also provided. Recombinant methods using the identified nucleic acids to produce D-GLO are also conternplated by this invention.

A further embodiment of the present invention is directed to the D-GLO protein encoded by the nucleic acid molecules identified herein including the protein identified by SEQ ID NO: 4 and proteins having at least 70 o sequence identity with SEQ ID NO: 4.

In another embodiment of this invention, methods of producing erythorbic acid and related salts by the conversion of glucose and/or D-gluconolactone using the D-GLO of the present invention are also provided.

Figure 1 is an illustration of the plasmid pGTY (GLO) which codes for the MScrl prepropepetide and the mature portion of the P. griseoroseum GLO.

Figure 2 graphically depicts cell density and GLO activity for untransformed S. cerevisiae and S. cerevisiae transformed with pGTY (GLO).

Figure 3 is a schematic representation of the conversion of glucose to D-erythorbic acid using glucose oxidase and D-GLO.

Figure 4 is a schematic representation of the conversion of glucose to D-erythorbic acid using glucose dehydrogenase and D-GLO.

Figure 5 is an illustration of the plasmid pPIC3.5K (GLO) which contains the complete coding region of the D-GLO gene under the control of the P. pastoris promoter.

The present invention relates to isolated nucleic acid molecules encoding a D-GLO of fungal origin. Preferably, the fungus is of the genus Penicillium, e. g., Penicillium

griseoroseum, Penicillium notatum, Penicillium cyaneum and Penicillium decumbens. The term"D-gluconolactone oxidasell (D-GLO) as used herein refers to and includes any natural or man-made variant of fungal D-GLO. For example, the nucleic acid molecules of the present invention which encode the fungal D-GLO can have the sequence of SEQ ID NO: 1, SEQ ID NO: 2 or SEQ ID NO: 3; or can have a sequence that hybridizes under stringent conditions to an isolated nucleic acid molecule having SEQ ID NO: 1, SEQ ID NO: 2, or SEQ ID NO: 3; or can have a sequence that encodes a protein having a sequence identity of about 70% or greater when compared to the amino acid sequence set forth in SEQ ID NO: 4. Amino acid sequence identities of about 70% or greater can be defined as a percentage of positives identified by the BLASTP algorithm as implemented at the Internet site http://www. ncbi. n/m. nih. gov/egi-bin/BLAST/ in a search using the default parameters of the program (Matrix = Bl. osum 62; Gap existence cost = 11; Gap extension cost = 1).

More specifically, the nucleic acid molecules of the present invention include variations of SEQ ID NO: 1, SEQ ID NO: 2 and SEQ ID NO: 3, such as deletions, insertions, additions and mutations wherein such sequences encode a protein which retains the enzymatic activity of naturally occurring D-GLO, i. e., the ability to convert D- gluconolactone to erythorbic acid.

Vectors and transformed host cells or transgenic organisms containing such nucleic acid molecules of the

present invention are also included within the scope of this invention. By"vectors or expression vectors"is meant any nucleic acid molecule or virus containing regulatory elements or reporter genes for the purpose of, but not limited to, expression in prokaryotic or eukaryotic cells or organisms. By"transformed host cell"is meant a host cell into which (or into an ancestor of which) has been introduced, by means of molecular biological techniques, a nucleic acid encoding a D-GLO, preferably from a fungal source and most preferably from the genus Penicillium. After introduction into the cell, this nucleic acid can exist extrachromosomally or become integrated into the host genome. It is routine for those skilled in the art to construct expression vectors into which a nucleic acid molecule of the present invention is placed in operable linkage to a desired promoter in order to effect the expression and secretion of the D-GLO proteins encoded by such nucleic acid molecules in a wide variety of host cells. Sambrook, et al, (Molecular Cloning, A Laboratory Manual, Sambrook, J., Fritsch, E. F., and Maniatis, T., 2 d ed. (1989) Cold Spring Harbor Laboratory Press).

Host cells useful for the practice of the present invention can be any available host cell which is amenable to transformation, procedures. For example, bacterial cells can be used such as bacteria belonging to genera Eschericha, Erwinia, Pantoea, Bacillus, Lactobacillus or Pseudomonas. Yeasts can also be used as host cells and include, for example, yeast belonging to the genera Saccharomyces, Kluyveromyces, Pichia,

Hansenula, Candida and Schwanniomyces. Unicellular algae may also be used such as, for example, those of the genera Synechosystis, Chlamydomonas, and Euglena. Higher plant cells can also be transformed with the isolated nucleic acids of the present invention. Methods for transforming plant cells with nalced DNA or with vectors comprising promoters which function in plants operably linked to heterologous genes, are widely known in the art. Exemplary plants include soybean, maize, potato, tomato, sugarbeet and the like. Mammalian cells may also be used as host cells. Preferably, the expressed GLO should be targeted to the culture medium or one of the organells, such as vacuole, chloroplast, microsome, peroxisome and the like.

The term"nucleic acid or nucleic acid molecule" encompasses both RNA and DNA, including cDNA, genomic DNA, and synthetic (e. g., chemically synthesized or modified) DNA. The nucleic acid molecules of the present invention can be double-stranded or single-stranded.

Where single stranded, the nucleic acid can be a sense strand or an antisense strand. The term"isolated nucleic acid"refers to a nucleic acid which may be flanked by non-natural sequences, such as those of a plasmid or virus. Thus, the nucleic acid can include none, some, or all of the 5'non-coding (e. g., promoter) sequences which are immediately contiguous to the coding sequence. The term, therefore, includes, for example, a recombinant DNA which is incorporated into a vector including an autonomously replicating plasmid or virus, or into the genomic DNA of a prokaryote or eucaryote

other than Penicillium, or which exists as a separate molecule (e. g., a cDNA or a genomic DNA fragment produced by PCR or restriction endonuclease treatment) independent of other sequences. This term also includes a recombinant DNA or RNA which is part of a hybrid gene encoding an additional polypeptide sequence. Moreover, the term is meant to include nucleic acid fragments which are not naturally occurring as fragments and would not be found in the natural state.

These recombinant nucleic acids may further include any of the varieties of sequences which increase, regulate or modify the transcription of the D-GLO coding sequence in the various recombinant hosts; namely, constitutive or regulated promoters, transcriptional enhancers and terminators and other sequences regulating the expression of D-GLO by various known mechanisms such as transcriptional repressor binding, attenuation or antitermination.

By"hybridizes under stringent conditions"is meant, the conditions in which a nucleic acid forms a stable, sequence-specific, non-covalent bond with the nucleic acid of SEQ ID NO: 1, SEQ ID NO: 2, or SEQ ID NO: 3 in solution or on a solid support under the low salt and high temperature conditions, regarded as stringent and set forth in Sambrook, et al, (Molecular Cloning, A Laboratory Manual, Sambrook, J., Fritsch, E. F., and Maniatis, T., 2nd ed. (1989) Cold Spring Harbor Laboratory Press). For example, reference nucleic acids such as SEQ ID N0: 1 can be immobilized on nitrocellulose filters, and

any other nucleic acids specifically and non-covalently binding to the immobilized reference nucleic acids in the presence of 0.2 X SSC (1.75 g/1 NaCl, 0.88 g/1 sodium citrate dihydrate; pH 7.0) and 0.1% (w//v) sodium dodecylsulfate at 623°C are considered to be hybridized under stringent conditions.

A GLO preparation or D-GLO protein obtained from a host cell transformed with the nucleic acid molecules of this invention is also included within the scope of the present invention. Host cells transformed with the nucleic acid molecules of the present invention can be cultured and the D-GLO recovered from the cultured cells. The D-GLO may be secreted or expressed within the host cells; the whole D-GLO coding region or parts of the coding region can be expressed without any additional modifications or after adding various C-and N-terminal extensions, such as an initiator codon or an oligo- histidine sequence. Fusions of D-GLO with other proteins which retain the enzymatic activity of D-GLO are within the scope of the current invention as well as mutated forms of all of the D-GLO-related proteins identified herein. These mutant forms may be obtained by random or directed mutagenesis. The preparations of such proteins having D-GLO activity may be crude or purified, used in solution or be immobilized on various carriers cnown in the art. Likewise, recombinant host cells expressing said proteins can be used to catalyze the conversion of D-gluconolactone into erythorbic acid during the fermentation of the host, or even in a resting state of the host. Notably, killed recombinant host cells of the

present invention can also be used as a catalyst of D- gluconolactone oxidation in the conversion of glucose and/or D-gluconolactone to erythorbic acid. Specifically, glucose can be converted to D-gluconolactone in the presence of glucose oxidase or glucose dehydrogenase; the D-gluconolactone can then be converted to erythorbic acid by contacting the D-gluconolactone substrate with the D- GLO of the present invention for a time, and under conditions sufficient to produce erythorbic acid. D- gluconolactone, as a substrate for D-GLO of the present invention, can be prepared through the chemical conversion of gluconic acid or by the conversion of glucose by glucose oxidase or glucose dehydrogenase. D- gluconolactone is thus converted to erythorbic acid by contacting the substrate with D-GLO for a time and under conditions sufficient to produce erythorbic acid.

The full-length D-GLO coding sequence may be used without any modifications or may be modified by methods well known in the art. For example, the N-or C-terminal amino acid sequences may be deleted or substituted with various known pre-or prepro-peptides (signal peptides) improving the secretion of D-GLO in suitable hosts.

Various other amino acid sequences regulating the sorting and targeting of the D-GLO polypeptide may be fused with full-length D-GLO or parts of D-GLO coding sequences retaining the GLO activity.

The D-GLO of the present invention can also be expressed as a fusion protein. In particular, it can be fused to protein domains which provide an auxiliary and/or

associated enzymatic activity, such as glucose oxidase, catalase and the like. Likewise, D-GLO can be fused to domains which provide other known useful functions such as high affinity for a specific ligand (e. g. streptavidin, maltose-binding protein, cellulose-and other polysaccharide-binding domains) or improved stability and/or solubility of the fusion protein (e. g. superoxide dismutase or glutathione S-transferase).

Erythorbic acid can be produced using the recombinant D- GLO of the present invention as a key element in the enzymatic process, preferably in an efficient host such as a filamentous fungus (e. g. a fungus belonging to the genera Aspergillus or Penicillium), or a yeast species with high secretory potential (e. g. those belonging to the genera Pichia or fiansenula), to convert glucose to erythorbic acid. In addition to D-GLO, such process incorporates at least two other enzymes, preferably including glucose oxidase (or an enzyme with overlapping specificity, such as glucose dehydrogenase, hexose oxidase or pyranose oxidase) and a catalase. The sequence of chemical reactions underlying the process based on glucose oxidase and GLO is illustrated by Fig.

3. Briefly, glucose is oxidized by molecular oxygen to D- glucono-d-lactone. This reaction is catalyzed by the glucose oxidase (or, hexose oxidase, pyranose oxidase, glucose dehydrogenase and the like) and generates hydrogen peroxide as a by-product. D-GLO subsequently catalyzes the conversion of D-glucono-d-lactone to erythorbic acid. Also in this reaction, molecular oxygen is consumed and hydrogen peroxide is formed. D-glucono-d-

lactone in water solution is known to be in equilibrium with gluconic acid and D-glucono-y-lactone. D-glucono-y- lactone is also a substrate of D-GLO. Both D- gluconolactones are oxidized by GLO to the same product, erythorbic acid. The spcntaneous reaction of gluconolactone hydrolysis is relatively slow and if sufficiently high concentrations of GLO and molecular oxygen are present in the reaction mixture, the hydrolysis reaction can be minimized. Moreover, because of the reversibility of this reaction, gluconic acid eventually forms a lactone that can be oxidized to erythorbic acid. Both the reaction catalyzed by the glucose oxidase and that catalyzed by D-GLO, generate hydrogen peroxide as a by-product. Hydrogen peroxide is a highly reactive substance that can cause inactivation of the enzymes involved in the process. Therefore, the hydrogen peroxide has to be continuously removed from the reaction mixture. The preferred method for the removal of hydrogen peroxide is through the use of a catalase. The use of other known scavengers which remove various activated oxygen species, e. g. superoxide dismutase, can also be advantageous in practicing this invention.

If glucose dehydrogenase is used to catalyze the conversion of glucose to gluconolactone, the removal of hydrogen peroxide can be coupled to the regeneration of the NAD+ which is consumed in the glucose dehydrogenase- catalyzed reaction. Fig. 4 illustrates the overall reaction scheme of this particular implementation of the present invention. The preferred manner of converting glucose to erythorbic acid according to the present

invention is to conduct the whole process in a single reactor wherein the oxidation of glucose and the oxidation of gluconolactone proceed concurrently.

However, implementation, wherein the conversion of glucose to gluconolactone and the conversion of gluconolactone to erythorbic acid are conducted separately, is also acceptable. It is also satisfactory to use gluconolactone, gluconic acid or a mixture of the two as the starting material for production of erythorbic acid using the recombinant D-GLO of the present invention.

The D-GLO nucleic acids of the present invention can also be expressed in hosts already capable of converting glucose to gluconolactone. A number of such microbial hosts are known in the art. Typically, such microorganisms oxidize glucose to gluconolactone using glucose oxidase (e. g. many fungal species belonging to the genera Aspergillus and Penicilium) or glucose dehydrogenase. Many bacterial species belonging to several genera that produce membrane-bound glucose dehydrogenases are suitable. It is advantageous to choose hosts which also express sufficiently high levels of catalase. When such recombinant hosts expressing GLO are fermented on glucose-containing cultural media, erythorbic acid can be obtained directly from glucose.

Additionally, a recombinant D-GLO-expressing host of the present invention can be co-cultured with a different host expressing glucose oxidase, glucose dehydrogenase

and/or a catalase. In this embodiment, erythorbic acid can be produced in a one-step mixed fermentation.

Two types of recombinant hosts would be particularly suitable for direct fermentation of glucose into erythorbic acid-the yeast and the filamentous fungi.

In this regard, Example 9 demonstrates that the D-GLO gene can be expressed efficiently in yeast such as Picllia pastoris. Any other yeast host known to support efficient secretion of heterologous protein, for example, Hansenula polymorpha or Kluyveromyces marxianus can also be used. The expression system used in Example 9 is based on glucose-repressible promoter and therefore is not well suited for construction of recombinant hosts fermenting glucose into erythorbic acid. However, equally efficient expression systems for P. pastoris based on strong promoters that are not repressed by glucose are known for example, one based on Stratagene's expression pGAPZ vector series and should be used for construction of such hosts. Other promoters of glycolytic genes can also be used.

In addition to expressing the GLO gene, a recombinant yeast host used in this fermentation process should also express a glucose oxidase e., g. glucose oxidase from Aspergillus niger is known to be expressed efficiently in yeast (De Baetselier A. et al. Fermentation of a yeast producing A. niaer glucose oxidate... Bio/Technology 9, 559-561 (1991).) Also, over-expression of a preferably secreted catalase gene is a very useful additional

genetic trait of a yeast host fermenting glucose into erythorbic acid.

In contrast to the yeast hosts, many wild-type filamentous fungi do express high levels of glucose oxidase and catalase. Therefore, genetically engineered over-expression of genes coding for these two enzymes is not as essential as in the case of the yeast hosts. For the (over-) expression of D-GLO, promoters of highly expressed glycolytic genes are suitable. Other promoters that retain high activity during cultivation on glucose can also be used, for example, the fungal TEF1 promoter.

Unlilce yeast, many filamentous fungi produce gluconolactonase. For example, in A. niger grown under high aeration conditions, gluconolactonase levels are extremely high (Whtteveen C. F. B., et al. Induction of glucose oxidate, catalase and lactone in Aspergillus niger, Curr. Genetics 24,408-416 (1993)).

Gluconolactonase competes with GLO for a common substrate glucono-b-lactone, and thus interferes with the conversion of glucose to erythorbic acid. Therefore, gluconolactonase should preferably be inactivated. The inactivation may be achieved either by genetic means through mutation of the gluconolactonase gene (s) or by selecting fermentation conditions which selectively inhibit gluconolactonase but not GLO. One preferable way to achieve inactivation is to add a selective inhibitor of gluconolactonase to the fermentation broth.

In another embodiment of this invention, erythorbic acid is produced from gluconolactone, gluconic acid or a mixture of these two substrates by fermenting each with a microbial host expressing sufficiently high levels of recombinant D-GLO.

The present invention is further illustrated, but not limited by, the following examples.

EXAMPLE 1 Glo Activity The activity of GLO was. measured using, as a substrate, an equilibrium mixture of glucono-b-lactone, glucono-y- lactone (D-gluconolactone) and gluconic acid. This mixture was prepared by dissolving crystalline glucono-b- lactone in water at 50% concentration (w/v) and allowing it to stand at 50°C for several days. The reaction was performed in a 50 mM potassium biphthalate buffer, pH 5.6 containing 2mM hydroxyquinoline, 12 uM 2,6- dichlorophenolindophenol and 70 mM substrate. Both 2,6- dichlorophenolindophenol and the substrate were added to the reaction mixture in the form of stock solutions immediately before the measurement. Also, the aliquot of the substrate solution to be used in the experiment was quickly adjusted to pH 5.8 immediately before use.

The enzymatic reaction was followed by recording the time course of the absorption at 600 nm caused by the reduction of 2,6-dicholorophenolindophenol by erythorbic acid. A control, differing from the reaction mixture only in that the enzyme solution was boiled for 2 min before the assay, was included. The amount of the erythorbic acid produced in the reaction was calculated using a calibration curve obtained by adding known amounts of erythorbic acid to the reaction mixture. The activity was expressed as umoles of erythorbic acid produced per minute under the conditions described above and at 30°C.

EXAMPLE 2 Purification of the Homogenous GLO From P. griseoroseum The GLO from P. griseoroseum strain ATCC 10431 was purified to homogeneity using a procedure similar to the method published earlier Takahashi T, et al. Agric. Biol.

Chem. 40,121-129 (1976).

Several 200 ml portions of YEPD medium (2% bacto-peptone, 1% yeast extract, 2% glucose) in 2 liter Erlenmelyer flasks were inoculated with 2 ml of a suspension of spores of P. griseoraseum grown on potato-dextrose agar (Difco) plates for one week. The cultures were grown on a rotary shaker at 30°C, 180 rpm for 2 days, 0.51 of this culture was used to innoculate a 10 1 induction medium (8% glucose, 0.2% KH2PO4, 0. 1% (NH4) 2SO4 0.1% (NH)CO, 0. 1%0.1% (NH)CO, 0. 1% NaNO3 0.1% MgSO4*7H2O, 190 MnSO4, 7H2O, 0.001% ZnSO4*7H2O, 0.5% CaC03, pH 5.5) in a 15 1 fermentor. Chloramphenicol (2.5 mg/1) and tetracycline (3 mg/1) were used in some fermentations to avoid the risk of bacterial contamination. The fermentation was allowed to proceed for 60 h at 30°C (aeration-51/min, stirring-300 rpm).

The mycelium was collected by filtration on a sintered glass filter and washed with water and buffer A (10 mM phosphate buffer, pH 6.5, containing 0.1 mM EDTA). The cells were disintegrated using a glass bead mill in the same buffer containing 1 mM phenylmethylsulphonyl fluoride. The disintegration process was done in cycles with ice-water cooling between and during the cycles.

The length of a cycle was adjusted to maintain the temperature of the suspension within a range of 4°C-22°C and the number of cycles was adjusted to achieve at least 90% cell breakage (evaluated microscopically). The homogenate was centrifuged for 30 min at 19000 x _q and the supernatant was used for the purification of the enzyme.

Approximately 80 ml of DEAE-Sepharose FF (Pharmacia) per liter of the cell extract was added and the suspension was incubated overnight at 4°C with gentle stirring. The resin was removed by filtration and the filtrate was treated with a fresh portion of DEAE-Sepharose under the same conditions. This treatment removed a significant proportion of the ballast protein while only a minute amount of GLO was adsorbed on the resin. GLO was precipitated from the filtrate with ammonium sulfate (60- 100% saturation). The ammonium sulfate precipitate was dialyzed against several changes of buffer A. The dialyzed enzyme solution was diluted with enough of buffer A to bring its conductivity to 1.25 mS and applied to a column of DEAE-Sepharose FF equilibrated with the same buffer (approx. 5ml column bed volume per ml of sample). The column was eluted with buffer A at 0.15 bed volume/h until the absorption at 280 nm dropped to the background value followed by elution with a linear gradient of NaCl in buffer A (0-100 mM NaCl, total gradient volume = 2 column bed volumes). The active peak fractions were pooled, concentrated using an Amicon ultrafiltration device and XM 50 membrane and applied on top of a 200 cm column of Sephacryl S-300 HR (Pharmacia)

equilibrated with buffer A. The column was eluted at a linear rate of 0.5 cm/min. Active fractions from the gel filtration column were collected and applied on top of a hydroxyapatite column (Bio-Gel HT, Bio-Rad) equilibrated with buffer A. The column bed volume was 0.13 of the sample volume and elution rate 0.1 bed volume per min.

GLO was eluted from the column with a linear gradient of ammonium sulfate (0-7.5%) in buffer A. The total volume of gradient was approximately 36 column bed volumes. The fractions with the highest activity of GLO were pooled and the pool was analyzed by polyacrylamide gel electrophoresis. A single strong diffuse band corresponding to apparent molecular weights of 68-80 kDa and a very weak band at approximately 34 kDa were observed. The results of a typical purification experiment are summarized in Table 1.

EXAMPLE 3 Amino Acid Sequence Analysis of The GLO The analysis was done as a commercial service in the Protein Analysis Laboratory of the Institute of Biotechnology, Helsinki. The purified GLO preparation of Example 2 was found to be homogenous in terms of the N- terminal sequence analysis. The following N-terminal sequence was found: Tyr Arg Trp Phe Asn Trp Gln Phe Glu Val Thr Nnn Gln Ser Asp Ala Tyr Ile Ala Pro His Asn Glu His... (SEQ ID No.: 5) ("Nnn"means that no interpretable signal was observed at this position, which is most probably explained by the presence of an underivatized cysteine residue or glycosylated amino acid residue).

The protein was further alkylated with 4-vinylpyridine, digested with Lys-C-protease and several peptides isolated from the digest using reverse phase HPLC. The following peptide sequences were determined with mass- spectrometry: Peptide 1-Glu His Asp Arg Met Thr Val Cys Gly Pro His Phe Asp Tyr Asn Ala Lys (SEQ ID NO: 6); Peptide 2-Glu Tyr Ile Cys Tyr Asp Glu Val Thr Asp Ala Ala Ser Cys Ser Pro Gln Gly Val Val (SEQ ID NO: 7); Peptide 3-Cys Gln Phe Val Asn Glu Phe Leu Val Glu Gln Leu Gly Ile Thr Arg (SEQ ID NO: 8).

EXAMPLE 4 Isolation of P. griseoroseum Chromosomal DNA P. griseoroseum was grown in YEPD medium as described in Example 2. The mycelium was washed with water and buffer A (Example 2) and freeze-dried. About 50 mg of the dry mycelium was ground in a mortar under liquid nitrogen.

The finely ground mycelium was suspended in 500 ul of extraction buffer (250 mM NaCl, 25mM EDTA, 200 mM Tris HCl, pH 8.5,0.5% SDS), 350 pl of phenol was added and the mixture was shaken to form a homogeneous suspension.

150 pi of chloroform was added to the suspension followed by a one hour high speed centrifugation (13 500 rpm in a table-top mini-centrifuge). The water phase was transferred to a new test tube and 10 pi of 10% ribonuclease A solution was added. The mixture was incubated at 37°C for 1 hour. After the incubation 1/10 vol. of 5 M Na-acetate buffer (pH 5.4) was added to the solution followed by 0.6 vol. of isopropanol. The DNA was recovered by centrifugation (10 min. 13500 rpm), washed 2 times with 70% ethanol, vacuum-dried and dissolved in 100 ul of water.

EXAMPLE 5 Cloning of a fragment of chromosomal DNA encoding GLO Based on the partial amino acid sequences of GLO SEQ ID NO: 5-SEQ ID NO: 8, a number of different oligonucleotides were synthesized and tested as primers in a PCR using chromosomal DNA of P. griseoroseum (Example 4) as a template.

The PCR was performed in a PTC-255 DNA Engine apparatus (MJ Research Inc., MA, USA) using the following program: 2 min at 94°C; 10 cycles of (30 sec at 94°C; 45 sec at 50°C; 3 min at 72°C) followed by 30 cycles of (30 sec at 94°C; 45 sec at 60°C; 3 min at 72°C). Each reaction was performed in 15 pi of a solution containing about 25 ng of template DNA, 0.75 unit of Taq DNA polymerase (Boehringer Mannheim), 0.75 uM of each of the oligonucleotide primers, 200 uM of each of the four deoxynucleoside triphosphates (dATP, dTTP, dCTP, dGTP), 1.5 ul of the 10 X buffer concentrate (supplied by the manufacturer of the Taq polymerase). The products of the PCR were analyzed by agarose gel electrophoresis using conventional techniques. [Maniatis, T. et al. (1982) Molecular cloning. Cold Spring Harbor Laboratory].

One pair of oligonucleotide primers produced the best results: a sense oligonucleotide TAYCGITGGTTYAAYTGGCA (SEQ ID NO: 9) and an antisense oligonucleotide CCIARYTGYTCIACIARRAAYTCRTTIACRAAYTGRCA (SEQ ID NO: 10).

In these sequences"I"represents an inosine phosphate

residue, R-a mixture of adenosine and guanosine phosphate residues, and Y a mixture of thymidine and cytosine phosphate residue. The PCR product (approximately 1.2 kb) obtained with this pair of oligonucleotides was purified by agarose gel electrophoresis and cloned into pCR2.1-TOPO vector (Invitrogen) using the TOPO TA Cloning kit supplied by the same manufacturer resulting in plasmid pCR (GLO).

EXAMPLE 6 Construction of a P. griseoroseum cDNA library Total RNA was isolated from P. griseoroseum mycelium grown on a mineral medium under conditions inducing GLO production (Example 2). The mycelium (stored frozen at- 70°C) was ground in a mortar under liquid nitrogen. 3 g of the finely ground mycelium was suspended with vigorous shaking in 10 ml of ice-cold RNA-extraction buffer (4 M guanidine thiocyanate, 0.5% Na laurylsarc osine, 25 mM Na citrate, 100 mM P-mercaptoethanol). The mixture was centrifuged at 4°C and 10000 rpm (SS-34 rotor, Sorvall) for 6 min. 10 ml of the supernatant was transferred to a new tube and 4 g of CsCl was added to it. All solutions used at subsequent steps were prepared using diethylpyrocarbonate-treated water and glassware. The RNA-containing solution was layered on the top of 1.2 ml of 5.7 M CsCl, 0.1M EDTA, pH 7 and centrifuged at 15°C and 33000 rpm for about 20 h. The precipitate was quickly rinsed with a small amount of water and dissolved in 100 ul of water. mRNA was isolated from this preparation using Oligotex Midi Kit (Qiagen) according to the manufacturer's instructions. A cDNA library was prepared from the P. griseoroseum mRNA using Stratagen's cDNA Synthesis Kit and XZAP-cDNA Gigapack III Gold Cloning Kit according to the instructions supplied with these kits.

EXAMPLE 7 Isolation of the full-length GLO cDNA from the P. griseoroseum cDNA library The 1.2 kb DNA fragment containing part of the chromosomal GLO gene was isolated from the plasmid pCR (GLO) (Example 5) by EcoRI restriction and preparative agarose gel electrophoresis. Standard genetic engineering techniques were used for restrictase digestion, isolation of plasmid DNA and DNA fragments etc. [Maniatis, T. et al. (1982) Molecular cloning. Cold Spring Harbor Laboratory], This fragment was radioactively labeled using the Random Primed DNA labeling kit (Boehringer Mannheim) and [aP'2]-dCTP.

The A-phage library of Example 6 was plated and screened by DNA hybridization using the labeled 1.2 kb fragment according to the manual provided by Stratagene with the XZAP-cDNA Gigapack III Gold Cloning Kit. A number of positive plaques were identified and the recombinant phages from twenty of them were purified and converted to the plasmid form according to the protocols from the same manual. These 20 plasmids were analyzed by restriction with EcoRI and XhoI and found to have inserts of different sizes. One small DNA fragment was present in all plasmids suggesting that all of the cDNA clones are derived from the same gene. The largest plasmid (named pGLO 1.8) contained an insert of about 1.8 kb size. The whole insert was sequenced using the commercial service

of Eurogentec Bel. S. A. (Belgium). Sequence analysis revealed that the insert of the plasmid pGLO 1.8 contained the complete coding region of the GLO cDNA (1443 bp, including the stop codon), 70 bp of the 5'- untranslated sequence and 261 bp of the 3'-untranslated sequence (SEQ ID NO: 1). The sequence encoded a protein of 480 amino acid residues (SEQ ID NO: 4).

The deduced amino acid sequence of P. griseoroseum GLO was compared to the known protein amino acid sequences using the BLAST service provided by GenBank over the Internet (http ://www. ncbi. nhn. nih. Kov/ci-hin/BLAST/np) i- newhlast ?. lfrnm=O). A number of homologous protein sequences were identified. If only the proteins with established function are considered, the highest homology was observed with the other lactone oxidases, such as D- arabinonolactone oxidase from Candida albicans and L- gulonolactone oxidase from rat.

The N-terminal amino acid sequence determined using the purified GLO from P. griseoroseum (Example 3, SEQ ID NO: 5) is identical to the deduced sequence of GLO (SEQ ID NO: 4) starting at amino acid residue 21. This observation and the predominance of hydrophobic amino acid residues in the area 1-20 of the translated GLO coding sequence strongly suggest that P. griseoroseum GLO contains an N-terminal signal peptide. The presence of a signal peptide in GLO is unexpected, since other lactone oxidases do not have signal peptides and are not known to be secreted. Furthermore, the deduced sequence of P. griseoroseum GLO contains 8 tentative-linked

glycosylation sites. The isolated GLO appears as a diffuse band on the polyacrylamide gel suggesting that is indeed a glycoprotein. However, we have isolated GLO from the cell extracts of P. griseoroseum. It may be speculated that in its native host GLO is either directed from the Golgi apparatus to one of the intracellular organelles or is secreted but remains associated with the cells.

EXAMPLE 8 Expression of the P. griseoroseum GLO gene in a heterologous host Since the deduced sequence of GLO displayed many features of a secreted protein, a yeast secretory expression seemed to be the most suitable for testing the functionality of the cloned GLO cDNA. The expression vector used for the expression of GLO is based on the well known yeast-E. coli shuttle vector pJDB207 [Beggs.

J. D-, in Wiliamson R., (ed.) Genetic Engineering 2, Academic Press (1981)]. Construction of the expression vector was accomplished in two stages. Firstly, pAC109 was constructed by simultaneously ligating three DNA fragments: (1) a 0.45 kb BamHI-Eco47III fragment from the promoter area of the Saccharomyces cerevisiae PH05 gene; (2) a 0.38 kb HaeIII-HindIII fragment from the S. cerevisiae MFal gene containing 116 bp of the 3'- noncoding area of the MFgl gene and part of coding area corresponding to the sequence of the prepropeptide of the yeast a-factor precursor protein (Mfa1-prepropeptide); (3) an approximately 6.5 kb fragment of pJDB207 obtained by restriction with BamHI and ffindIII. Secondly, a synthetic polylinker was inserted into the HindIII site of pAC109.

The polylinker was composed of two oligonucleotides: the top strand nucleotide AGCTCTCGAGATCTCCCGGGA (SEQ ID NO: 11) and the bottom strand nucleotide AGCTTCCCGGGAGATCTCGAG (SEQ ID NO: 12). The plasmid that has

the polylinker inserted in such an orientation that the HindIII site is located proximally to the Mf a1 prepro- area was selected and named pGTY.

The DNA sequence of the cloned GLO gene was modified into a form suitable for constructing a fusion with the Mf a1- prepropeptide by conducting a PCR with the plasmid pGLO1.8 as template and the two oligonucleotide primers, the"sense"primer: GAAGAAGCTTACCGGTGGTTCAATTGGCAGTTTTTGGT (SEQ ID NO: 13) and the"anti-sense"primer: CACGACGTTGTAAAACGACGGCCAG (SEQ ID NO: 14), annealing in the vector downstream of the GLO gene. At this step, a modification was introduced into the GLO gene. All of the 5'-noncoding sequence and the sequence corresponding to the amino acid residues 1-20 of the deduced GLO coding sequence was deleted and a HindIII site was introduced in a position allowing for the in-frame fusion of the Mfal- prepropeptide and mature GLO. The product of this PCR reaction was digested with HindIII and Xhol and ligated with pGTY d : igested with the same pair of restrictases.

The resulting plasmid pGTY (GLO) (Fig. 1) codes for a fusion protein composed of the Val prepropeptide and the presumptive mature part of the P. griseoroseum GLO (SEQ ID NO: 15).

S. cerevisiae strain GRF18 (ATCC 64667, genotype: MATs, leu2-3, leu2-212, his3-11, his3-15 was transformed to leucine prototrophy using the"lithium"transformation method [Sherman F. et al. Laboratory Course Manual for

Methods in Yeast Genetics pp-121-122, Cold Spring Harbor Laboratory (1986)].

One of the transformed clones as well as the recipient strain used as a control were grown until early stationary phase (rotary shaker, 180 rpm, 30°C) in 0.3 1 SC-his medium (0.67% Yeast Nitrogen Base w/o amino acids, Difco, 2% Glucose, 100 mg/1 histidine; for the non- transformed control strain leucine was also added at 100 mg/1). The yeast cells from these cultures were used to inoculate two identical 15 1 fermentors each containing 10 1 of a low-phosphate (PEP) medium. To prepare the PEP medium an 8% solution of bacto-peptone (Difco) was treated with CaCl2 (added to 0.4M concentration) at pli 11 and 100°C for 5 min. The peptone solution was cooled to room temperature, adjusted to pli 5.5, filtered through paper and 0.4 urn pore-size membrane and used as the stock solution of phosphate-depleted peptone. PEP medium contained 2% phosphate-depleted peptone and 5% glucose.

The. fermentation conditions were: stirring-300 rpm, aeration-5 1/min, pH-5.5 maintained by addition of 4M NaOH. Samples of the cultures were taken at suitable intervals. Cell density was followed by measured the absorption at 600 nm. The GLO activity was measured after removing the cells by centrifugation, concentrating the samples of the medium about 500-fold using Centriplus membranes (Amicon) and removing the low molecular weight components of the fermentation medium by gel filtration using disposable EconoPack 10 DG mini-columns (BioRad).

The peak levels of GLO (about 2.4 mU/ml) were measured after about 60-70 hours of fermentation. No GLO activity

was found in the control fermentation of the untransformed recipient strain.

The results of this experiment show conclusively that the cDNA clone of the GLO gene in the plasmid pGLO 1.8 is indeed functional. S. cerevisiae is known to be a relatively inefficient host for secretion of heterologous proteins. Therefore, much higher expression levels of recombinant GLO may be expected when the GLO gene is introduced into other fungal species-filamentous fungi or other yeast with higher secretory potential.

Example 9 Expression of the GLO gene in methylotrophic yeast The coding region of the GLO gene was amplifie by PCR using the oligonucleotide primers: CAAAGCTTCTAGAGCCTCAGACCACTCATATCACATC (SEQ. ID No: 14) and CCAACAATTGATGCTGAGCCCTAAGCCGGCTTTCCTGC (SEQ. ID No: 16). The resulting DNA fragment was digested with restriction endonucleases Xbal and MfeI and ligated with the plasmid pPIC3 5K (Multi-Copy Pichia Expression Kit, Invitrogen Corp.) digested with AvrII and EcoRI. The resulting plasmid PPIC3.5K (GLO 51-3) contains the complete coding region of the GLO gene under control of the P. pastoris AOXI promoter (Figure 5).

P. pastoris strain GS115 was transmfored with PPIC3.5K (GLO 51-3) using the method recommended by Invitrogen.

Several independently obtained transformed clones were cultivated in a rotary shaker (30°C, 200rpm) in BMGY (yeast extract-1% peptone-2% potassium phosphate buffer, pH 6.0-lOOmM, 1% glycerol, 1.34% Yeast Nitrogen Base (Difco), 0.4 mg/1 biotin). After reaching early stationary phase, the cells were collected by low speed (4000 rpm) centrifugation, re-suspended and cultivated overnight in an equal volume of BMMY is identical to BMGY except that the glycerol is replaced with 0.5% methanol.

The highest GLO expression levels measured in culture supernatant in these experiments were about 0.4-0.5 U/ml.

This value is approximately 200 times higher than the GLO expression levels in S. cerevisiae.

EXAMPLE 10 Purification of recombinant GLO produced P. pastoris For the preparative isolation of recombinant GLO, a recombinant P. pastoris strain GS115:: pPIC3.5 (GLO51-3) was cultivated in a 101 fermentor using a fed-batch mode essentially as described in (K. Sreckrishna, et al.

Biochemistry, After the cultivation, the cells were separated by centrifugation, the pH of the clarified culture medium was adjusted to 6.5 and 500 ml of DEAE Sepharose FF was added. The suspension was stirred overnight at 4°C after which the DEAE Sepharose was collected by sedimentation, packed into a column and eluted with 0-0.2 M gradient of NaC : I in 10Mm sodium phosphate buffer (pH 6.5) containing 1mM EDTA.

The fractions containing the highest activity of GLO were pooled and subjected to hydroxyapatite chromatography under conditions described in Example 2. The fractions containing the highest GLO activity were analyzed by acrylamide gel electrophoresis and found to contain an almost homogeneous GLO preparation (approximately 80-900 purity as judged by the intensity of Coomassie Brilliant Blue G250 staining).

The specific activity of this preparation was about 24 U/mg protein, i. e. approximately 4-times higher than the specific activity of homogeneous GLO purified from P. cyaneo-fulvus.

EXAMPLE 11 Immobilization of the recombinant GLO 1 ml of N-hydroxysuccinimide-activated Sepharose 4 FF (Pharmacia) was incubated with 0.5 ml of 0.35 mg/ml solution of GLO purified according to Example 10 and adjusted to pH 8.0. The coupling reaction and subsequent treatment were carried out according to the instructions of the manufacturer of the resin.

The activity of immobilized GLO was measured by a slight modification of our standard assay (Example 1). 1 ml of the GLO-Sepahrose was gently shaken with 10 ml of the substrate solution (Example 1) for 30 min, the resin was separated by sedimentation and the amount of erythorbic acid formed was measured in a reaction with 2,6- dichlorophenolindophenol.

It was found that GLO, the activity of immobilized GLO (about 2.5 U/ml resin) and the GLO activity remaining unbound approximately (5.5 U) match the amount of enzyme used in the reaction. Thus, GLO retains most of its enzymatic activity after immobilization on N- hydroxysuccinimde-activated Sepharose. Therefore, immobilized GLO can be used in the production of erythorbic acid from gluonolactone.

EXAMPLE 12 Enzymatic conversion of glucose into erythorbic acid 120 tir ouf a reaction m : i} : ture containing 60OU/ml gl. ucose oxidate, 1.2 U/ml of catalase and 1% glucose in 100 mM potassium phosphate buffer, pH 6.0 was incubated for 1 hour at 35°C. At this point, 120 ul of potassium phthalate buffer pH 5.6 and 100 pi of purified recombinant GLO solution containing about 0.8 activity units (Example 10) was added and the reaction was allowed to proceed for an additional 1 hour. The reaction was terminated by freezing and erythorbic acid was analyzed by HPLC (using the conditions described by L. W. Doner, K. Hicks, Anal. Biochem. 115,225-230 (1981)) of by measuring the reduction of 2,6-dichlorophenolindophenol.

Aproximately 1.5 mg/ml of erythorbic acid was found in the reaction mixture corresponding to about 40% yield.

In another experiment, 1 ml of a reaction mixture containing 0.1 mg/ml glucose, 0.06 U/ml glo, 24 U/ML catalase, 800 U/ml glucose oxidate in 0. 1M potassium phosphate buffer was incubated at room temperature in an open test tube. About 0.3 mg/ml of erythorbic acid was formed corresponding to about 30% yield.

Notably, no attempt to optimize the conversion yield was rnade in these experiments. For example, the yield of erythorbic acid could be further improved by introduction of automatic pH control and other protocols well known to those skilled in the art.