WO2000061763A2 - Proteins related to gaba metabolism - Google Patents

Abstract

Novel polynucleotides encoding enzymes involved in the metabolism of gamma aminobutyric acid (GABA) in plants, specifically, gamma aminobutyric acid transaminase (GABA-T) and gamma hydroxybutyrate dehydrogenase (GHBDH), are described, as well as the GABA-T and GHBDH proteins they encode, and derived antibodies. The polynucleotides are useful for the screening and isolation of other related genes, the antibodies are useful for identifying and isolating related proteins, and the proteins provide a basis for determining how carbon and nitrogen is metabolized in plants.

Description

NOVEL PROTEINS RELATED TO GABA METABOLISM

Field of the Invention

The present invention relates to plant proteins involved in the metabolism of gamma-aminobutyric acid (GABA) and the genes encoding them. In particular, the present invention relates to plant GABA transaminase (GABA-T) and gamma hydroxy butyrate dehydrogenase (GHBDH). Background of the Invention

Gamma-aminobutyric acid (GABA) is a four carbon, non-protein amino acid found in virtually all prokaryotic and eukaryotic organisms as a significant component of the free amino acid pool. GABA is produced via the irreversible decarboxylationof L-glutamate in a reaction catalyzed by the enzyme, glutamate decarboxylase (GAD). Transaminationof GABA to form succinic semialdehyde (SSA) is catalyzed by GABA-transaminase (GABA-T), while SSA is subsequently converted to succinate via succinic semialdehyde dehydrogenase. These reactions constitute the GABA shunt, which is illustrated in Fig. 1, and provide glutamate carbon to the Krebs cycle. In an alternate reaction, SSA may be converted to gamma hydroxybutyrate (GHB) via gamma hydroxybutyrate dehydrogenase (GHBDH). The role of GHB in plants is unknown at present.

GABA-T has been purified from brain and liver tissues of several mammals, including mouse, rat, rabbit, pig and human (Schousboe et al., 1973, Biochemistry 12: 2868-2873; Bloch-Tardy et al., 1974, Biochemie (Paris) 56: 823-832; John and Fowler, 1976, Biochem J 155: 645-651; Buzenet et al., 1978, Biochim Biophys Acta 522: 400-411). GABA-T from a number of other organisms including Pseudomonas sp. F-126, Candida guilliermondii, Streptomyces grisens and E. coli has also been purified to homogeneity. (Yonaha and Toyama, 1980, Arch Biochem Biophys 200: 156-164; Yonaha et al. , 1985, Eur J Biochem 146: 100-106; DerGarabedian et al., 1986, Eur J Biochem 156:589-596; Park et al., 1993, J Biol Chem 268:7636-7639).

The genes encoding GABA-T from a number of different organisms, including Aspergillus nidulans, E. coli, yeast, pig brain, rat brain and Ustilago madis, are highly homologous (Richardson et al., 1989, Mol Gen Genet 217:118- 125; Bartsch et al. , 1990, J Bact 172:7035-7042; Andre and Jauneaux, 1990, Nucl Acid Res 18:3049; Kwon et al., 1992, J Biol Chem 267:7215-7216; Medina- Kauwe et al. , 1994, J Nuerochem 62: 1267-1275 ; Straffon et al. , 1996, Curr Genet 29:360-369). Using partial amino acid sequence data for the pig brain enzyme, Kwon et al. (supra) isolated several overlapping cDN A clones from a pig brain cDNA library and determined that the mature enzyme consists of 473 amino acids and an amino terminal segment, thought to be a mitochondrial targeting peptide, of 27 amino acids. Expression of the recombinant pig GABA-T in E. coli produced a functional protein (Park et al. , 1993 , supra). The amino acid sequence of pig GABA-T bears significant homology (42 %) to a yeast GABA-T which was identified via complementation of a yeast mutant deficient in ugal, the gene which encodes GABA-T (Andre and Jauneaux, 1990, supra). The deduced amino acid sequence for yeast GABA-T shares approximately 45 % homology with yeast, E. coli, rat and human ornithine aminotransferase (Andre and Jauneaux, supra; Bartsch et al. , supra). The pig brain GABA-T cDNA was used to isolate a full length cDNA for human brain GABA-T; the pig and human sequences are 95 % homologous (Osei and Churchich, 1995, Gene 155:185-187).

GHBDH cDNA was first isolated from rat brain and expressed inE. coli (Andriamampandry et al., 1998, Biochem J., 334:43-50). At the amino acid level, it has about 30% similarity to members of the oxidoreductase family, and has a protein molecular mass of 45 kDa. The GHBDH (designated as succinic semialdehyde reductase in some publications) nucleic acid sequence has also been isolated from a number of other sources, including mouse, human, Clostήdium aminobutyricum, Clostήdium kluyveri and Ralstonia eutropha (GenBank Accession numbers; AA403499, AA456318, AJ250267, L21902 and AAC41425, respectively). There is considerable identity at the protein level among the mammalian sources, and among the bacterial sources, but not between the mammalian and bacterial sources.

While the characterization and isolation of some mammalian, fungal and bacterial GABA-T and GHBDH have been conducted, such enzymes have not previously been isolated from any plant sources. The information available regarding plant GABA-T has been obtained from crude plant preparations only, and also arises from speculation based on the isolation of these enzymes from other sources as set out above. It would be desirable, thus, to provide plant GABA-T and GHBDH in isolated form in order to elucidate the specific properties and functions thereof. Summary of the Invention

Accordingly, in one aspect of the present invention, an isolated polynucleotide, consisting of either DNA or RNA, that encodes plant GABA-T is provided, as well as the novel GABA-T protein itself.

In another aspect of the present invention, there is provided an isolated polynucleotide, consisting of either DNA or RNA, that encodes plant GHBDH, as well as the novel GHBDH protein itself.

In other aspects of the present invention, there are provided cells that have been genetically engineered to encode these plant proteins and methods for producing these proteins from such cells. In related aspects of the present invention, recombinant DNA constructs are provided, as well as antibodies to GABA-T and GHBDH.

These and other aspects of the present invention will be described in more detail by reference to the following drawings in which: Brief Description of the Drawings:

Figure 1 illustrates the GABA shunt and related metabolic routes;

Figure 2 is the nucleotide sequence (SEQ ID NO: 1) of a GABA-T- encoding polynucleotide in accordance with the present invention;

Figure 3 is the amino acid sequence (SEQ ID NO: 2) of the GABA-T protein encoded by the polynucleotide of Fig. 2;

Figure 4 is the nucleotide sequence (SEQ ID NO: 3) of a GHBDH- encoding polynucleotide in accordance with the present invention;

Figure 5 is the amino acid sequence (SEQ ID NO: 4) of the GHBDH protein encoded by the polynucleotide of Fig. 4;

Figure 6 illustrates the growth of wild-type yeast and yeast expressing the empty transformation vector pFL61 or the 4-GHBDH on media supplemented with various nitrogen sources; Figure 7 illustrates the effectiveness of chicken antibodies against recombinant GABA-T. Panel A is an SDS-PAGE gel showing total crude protein from ToplO E. coli cells containing either no insert:pTrcHisB or GABAT :pTrcHisB (with insert), and induced with IPTG for 4 hours. Panels B and C are the corresponding Western blots probed respectively, with the AntiExpress antibody specific to the Express protein incorporated into the chimeric GABA-T, and the anti-GABA-T antibody specific to the GABA-T protein; and

Figure 8 illustrates the effectiveness of chicken antibodies against recombinant GHBDH. Panel A is an SDS-PAGE gel showing total crude protein from ToplO E. coli cells containing either no insert:pTrcHisB or GHBDH :pTrcHisB (with insert), and induced with IPTG for 4 hours. Panels B and C are the corresponding Western blots probed respectively, with the AntiExpress antibody specific to the Express protein incorporated into the chimeric GHBDH, and the anti-GHBDH antibody specific to the GHBDH protein. Detailed Description of the Invention

Polynucleotides encoding plant proteins involved in the metabolism of GABA, i.e. proteins of or related to the GABA shunt, have been isolated. In particular, polynucleotides encoding plant GABA-T and GHBDH have been isolated. Not only do such polynucleotides provide a means to prepare plant GABA-T and GHBDH in isolated form, i.e. free from other proteins of plant origin, they are also useful for the screening and isolation of homologous genes from other organisms.

As used herein, the term "GABA-T" is meant to refer to the enzyme "gamma aminobutyric acid transaminase", and specifically to plant GABA-T. Further, the terms "GHBDH" and "4-HBDH" are used interchangeably herein and are meant to refer to the enzyme "gamma hydroxybutyrate dehydrogenase", and specifically to plant GHBDH.

The particular polynucleotides isolated in accordance with the present invention have the nucleotide sequences set out in Figs. 2 and 4. The GABA-T cDNA containing both 5 " and 3 untranslated regions (UTR) comprises 1737 nucleotides as shown in Fig. 2. The GAB A-T-encodingpolynucleotide comprises 1515 nucleotides and encodes a GABA-T enzyme comprising 505 amino acids as set out in Fig.3. The GABA-T enzyme is further characterized as ahomodimer with a 55 kDa subunit, an isoelectric point of 4.8, pyruvate-dependent activity and Michaelis-Menten constants of 1.5 mM and 300 μM for GABA and pyruvate, respectively. The GHBDH-encoding polynucleotide comprises 870 nucleotides and encodes a GHBDH enzyme comprising 289 amino acids as set out in Fig. 5. The GHBDH enzyme is further characterized by a preference for NADPH over NADH in an essentially irreversible reaction, and an estimated molecular mass of 31.8 kDa for the subunit.

DNA coding for the GABA-T and GHBDH enzymes can be obtained by applying selected techniques of gene isolation or gene synthesis. As described in more detail in the examples herein, GABA-T and GHBDH polynucleotides can be obtained by careful application of conventional gene isolation and cloning techniques. This typically will entail extraction of total messenger RNA from a fresh source of plant tissue, followed by conversion of message to cDNA and formation of a cDNA library in plasmidic vectors. The cDNA library is then used to transform competent cells, for example, competent yeast cells. Transformants may then be isolated by growth in a selectable medium. Vectors containing the DNA of interest, e.g. GABA-T or GHBDH DNA, are then isolated for sequencing.

The plasmidic vector harbouring the expression construct typically incorporates a marker to enable selection of stably transformed recombinant cells. The marker generally comprises a gene conferring some survival advantage on the transformants allowing for the selective growth of successful transformants in a chosen medium. For example, common gene markers include genes which code for resistance to specific drugs, such as tetracycline, ampicillin and neomycin. Thus, transformants which have successfully taken up the plasmid DNA will incorporate both the gene of interest, i.e. the GABA-T or GHBDH gene, and the marker gene, e.g. gene for drug resistance such as tetracycline, and will survive culturing in media containing the drug which they could otherwise not tolerate.

Having herein provided the nucleotide sequence of genes encoding plant GABA-T and GHBDH, it will be appreciated that automated techniques of gene synthesis and/or amplification can be performed to generate GABA-T- and GHBDH-encoding DNA. In this case, because of the length of the GABA-T- and GHBDH-encoding DNA, application of automated synthesis may require staged gene construction in which regions of the gene up to about 300 nucleotides in length are synthesized individually and then ligated in correct succession via designed overlaps. Individually synthesized gene regions can then be amplified by PCR.

With appropriate template DNA in hand, the technique of PCR amplification may be used to directly generate all or part of the final gene. In this case, primers are synthesized which will prime the PCR amplification of the final product, either in one piece, or in several pieces that may subsequently be ligated together via step- wise ligation of blunt ended, amplified DNA fragments, or preferentially via step-wise ligation of fragments containing naturally occurring restriction endonuclease sites. Both cDNA or genomic DNA are suitable as templates for PCR amplification. The former may be obtained from a number of sources including commercially available cDN A libraries, single- or double- stranded cDNA, or cDNA constructed from isolated messenger RNA from a suitable tissue sample. Genomic DNA may also be used as a template for the PCR-based amplification of the gene; however, the gene sequence of such genomic DNA may contain unwanted intervening sequences.

Once obtained, the GABA-T- and GHBDH-encoding DNA is incorporated for expression into any suitable expression vector, and host cells are transfected therewith using conventional procedures, such as DNA-mediated transformation including calcium phosphate precipitation, protoplast fusion, microinjection, lipofection and electroporation. Expression vectors may be selected to provide transformed cell lines that express the GABA-T- and GHBDH-encoding DNA in a stable manner. Suitable expression vectors will typically harbour a gene coding for a product that confers on the transformants a survival advantage to enable their subsequent selection. Genes coding for such selectable markers include theE. coli gpt gene which confers resistance to mycophenolic acid, the neo^R gene from transposonTn5 which confers resistance to neomycin and to the neomycin analog G418, the dhfr sequence from murine cells or E. coli which changes the phenotype of DHFR- cells into DHFR+ cells, and the tk gene of herpes simplex virus, which makes TK- cells phenotypically TK+ cells. Other methods of selecting for transformants may of course be used, if desired, including selection by morphological parameters, or detection of surface antigen or receptor expression. The latter can be monitored using specifically labelled antibodies and a cell-sorter, e.g. fluorescent activated.

As one of skill in the art will appreciate, GABA-T- and GHBDH-encoding DNA may be modified prior to its incorporation into an expression vector to enhance protein expression. Specifically, modifications may be made to the 5 ' and 3' non-coding regions of GABA-T- and GHBDH-encoding DNA in order to increase the level of protein expression. For example, the 5 ' non-coding end of GABA-T- and GHBDH-encoding DNA may be modified to provide a 5' " translation-enhancing sequence " (TES) . Such modifications include truncating the 5' end of the GABA-T- or GHBDH- encoding DNA preceding the native translation-enhancing sequence. Alternatively, the DNA is truncated and the native translation-enhancing sequence is replaced with a heterologous translation- enhancing sequence using conventional methods of restriction enzyme digestion followed by ligation techniques. By "heterologous" is meant a sequence that is not native to. Further, by "translation-enhancing sequence" is meant the 5' sequence which is required for translation to occur, and includes the translation initiation codon, i.e. ATG. Modifications may also be made to the 5' and 3' coding regions of GABA-T- and GHBDH-encoding DNA in order to facilitate the isolation of pure protein. For example, the 5' coding end of GABA-T and GHBDH may be modified to provide a hexameric histidine tag.

In order to obtain GABA-T or GHBDH, techniques of genetic engineering are further applied to prepare a plant cell line that incorporates GABA-T- or GHBDH-encoding DNA and is adapted to express GABA-T or GHBDH in functional form as a heterologous product. The construction of such cell lines is achieved by introducing into a selected host cell a recombinant DNA construct in which DNA coding for GABA-T or GHBDH is associated with expression controlling elements that are functional in the selected host to drive expression of GABA-T or GHBDH-encoding DNA, thus elaborating the desired protein. The particular cell type selected to serve as host can be any of several cell types currently available in the art, including both prokaryotic and eukaryotic cell types. Yeast cells, such as Saccharomyces cerevisiae, bacterial cells such as E. coli and insect cells represent suitable host cells for expression and production of plant GABA-T and GHBDH.

A variety of gene expression systems have been adapted for use with these hosts and are now commercially available. Any one of these systems can be selected to drive expression of the GABA-T or GHBDH-encoding DNA. These systems, available typically in the form of plasmidic vectors, incorporate expression cassettes the functional components of which include DNA constituting expression controlling sequences, which are host-recognized and enable expression of GABA-T or GHBDH-encoding DNA when linked 5 ' thereof. GABA-T or GHBDH-encoding DNA is herein referred to as being incorporated "expressibly" into the system, and incorporated "expressibly" in a cell once successful expression from a cell is achieved. These systems further incorporate DNA sequences which terminate expression when linked 3 ' of the receptor- encoding region. Thus, for expression in the selected cell host, there is generated a recombinant DNA expression construct in which the GABA-T or GHBDH- encoding DNA is linked with expression controlling DNA sequences recognized by the host, and which include a region 5' of the GABA-T or GHBDH-encoding DNA to drive expression, and a 3' region to terminate expression.

Included among the various recombinant DNA expression systems that can be used to achieve plant cell expression of the GABA-T or GHBDH-encoding DNA are those that exploit viral or plant promoters that infect plant cells; examples of such promoters include those that are constitutive (e.g. CaMV 35S), and those that are tissue-specific or inducible (e.g. ▲ 0.3 TobRB7).

The present invention also provides, in another of its aspects, antibody to plant GABA-T or GHBDH. To raise such antibodies, there may be used as immunogen either full-length GABA-T or GHBDH, or an immunogenic fragment thereof, produced in a microbial or plant cell host as described above or by standard peptide synthesis techniques. Regions of GABA-T or GHBDH particularly suitable for use as immunogenic fragments include regions which are determined to have a high degree of antigenicity based on a number of factors, as would be appreciated by those of skill in the art, including for example, amino acid residue content, hydrophobicity/hydrophilicityand secondary structure.

The raising of polyclonal antibodies to GABA-T or GHBDH or to desired immunogenic fragments can be achieved using protocols of conventional design, and any of a variety of animal hosts including chickens. Alternatively, for monoclonal antibody production, immunocytes such as splenocytes can be recovered from the immunized animal and fused, using hybridoma technology, to myeloma cells. The fusion cell products, i.e. hybridoma cells, are then screened by culturing in a selection medium, and cells producing the desired antibody are recovered for continuous growth, and antibody recovery. Recovered antibody can then be coupled covalently to a reporter molecule, i.e. a detectable label, such as a radiolabel, enzyme label, luminescent label or the like, using linker technology established for this purpose, to form a specific probe for GABA-T or GHBDH.

According to a further aspect of the present invention, DNA or RNA encoding plant GABA-T or GHBDH, and selected regions thereof, may also be used in detectably labeled form, e.g. radiolabeled form, as hybridization probes to identify sequence-related genes existing in plants (or cDNA libraries) or to locate GABA-T or GHBDH-encoding DNA in particular specimens. This can be done using the intact coding region, due to a high level of conservation expected between related genes, or by using a highly conserved fragment thereof, having radiolabeled nucleotides, for example, P nucleotides, incorporated therein.

Embodiments and aspects of the present invention will now be described by reference to the following specific examples which are not to be construed as limiting.

Example 1 Identification and Characterization of GABA-T Purification of GABA-T

Pyruvate-dependent γ-aminobutyrate transaminase (GABA:pyruvate-T; EC 2.6.1.19) from tobacco (Nicotiana tabacum [L.] cv. Samsun) leaf was partially purified 1530-fold by FPLC anion-exchange chromatography of proteins from isolated mitochondria as follows. Unless indicated, all procedures were performed at 4 °C. One kg of tobacco leaf tissue was harvested from 8- to 10- week-old plants, deveined, sliced into small pieces (4 cm²), and homogenized for 20-40 s (20000 rpm) in 5 volumes of 50 mM Tris-HCl buffer (pH 8.2) containing 3 mM DTT, 1.25 mM EDTA, 2.7 % (w/v) PVP (MW 40 000), 2.5 mM MgCk, 350 mM mannitol, 30 mM ascorbate, 0.2 % (w/v) BSA, 1 mM PMSF, 2.5 μg ml^"1 leupeptinand 2.5 μg ml^"1 pepstatin A (buffer A), using an Ultra Turrax T25 homogenizer (Janke and Kunkel, IKA-Labortechnik, Germany). For examination of total leaf GABA-T activity, mannitol and BSA were omitted and 6 mM CHAPS (3-[(3-cholamidopropyl)dimethylammonio]-l-propanesulfonate)added to the buffer (buffer B). Homogenate from either buffer system was filtered through a single layer of 80-μm Miracloth. A sample of the filtrate, referred to as the crude extract, was incubated for 20 min in 6 mM CHAPS detergent to solubilize the protein and desalted using a Sephadex G-25M PD-10 column (Pharmacia). At each step of the purification procedure, protein concentration and pyruvate- and 2- oxoglutarate-dependentGABA-T activities were determined using the assay described below, in the desalted sample. The crude extract prepared with buffer A was subjected to a series of five spins ranging from 50 to 2 500 xg. The supernatant was then centrifuged at 11 000 xg for 20 min to pellet mitochondria and some contaminating chloroplasts. The pellet was thenresuspended in 25 ml of buffer B . The solubilized, lysed mitochondiral fraction was centrifuged for 75 min at 180000 xg and the supernatant filtered through a 0.45-μm nylon syringe filter.

The filtered mitochondrial fraction was supplemented with both leupeptin and pepstatin A, mixed with 7.5 ml of Fractogel EMD DEAE 650 (S) (EM Separations, Gibbstown, New Jersey), pre-equilibrated in buffer B, incubated for 15 min with constant, gentle mixing and pelleted via a 3-min spin at 2 500 xg. The protein-loaded Fractogel was suspended in 7.5 ml of buffer B and poured into a Waters Protein-Pak anion exchange cylinder (10 mm x 100 mm). The column was packed and washed with 50 mM Tris-HCl buffer (pH 8.2) containing 3 mM DTT, 6 mM CHAPS and 20 % (v/v) glycerol (buffer C) using a flow rate of 1 ml min ^_1 that was generated by a Waters 625 LC System. The proteins were eluted using a 70-min linear gradient of 0-1 M NaCl in buffer C. For all chromatography steps, fractions (1-2 ml) were collected into a solution containing a final concentration of 0.2 mM PLP and 2.5 μg ml^"1 of both leupeptin and pepstatin A . Fractions containing GABA : pyruvate-T activity were pooled , desalted, loaded onto a second Fractogel anion exchange column equilibrated with buffer C, and the protein eluted using a 70-min linear gradient of 0-400 mM NaCl in buffer C . Fractions containing GABA: pyruvate-T activity were pooled, desalted, concentrated using a Centricon-30 concentrator (Amicon, Beverly, MA), and supplemented with leupeptin and pepstatin A each to a final concentration of 2.5 μg ml^"1. After analysis, the remaining sample was used immediately or frozen in liquid nitrogen and stored at -80 °C.

For analysis of total leaf- or mitochondrial-pyruvate- and/or 2- oxoglutarate-dependent GABA-T activities, either crude extract or mitochondrial protein was loaded onto a Fractogel column and eluted with a 70-min linear gradient of 0-400 mM NaCl in buffer C . Where further purification was required, the fractions with GABA:pyruvate-T activity were pooled, desalted and concentrated using a Centricon-30 concentrator, and subjected to low-pressure GABA affinity chromatography and HPLC gel filtration. For GABA affinity chromatography, the sample was loaded onto a 10-ml Pharmacia CIO column packed with EAH Sepharose 4B resin, with the free amino group of GABA acting as the ligand, pre-equilibrated with buffer C at a flow rate of 0.5 ml min¹. Proteins were eluted using an 80-min linear concentration gradient from 0-40 mM GABA in buffer C. Fractions displaying GAB A: pyruvate-T activity were pooled, concentrated and loaded at a flow rate of 0.5 ml min^"1 onto a FPLC BioSil-SEC 250-5 gel filtration column (Pharmacia) pre-equilibrated in buffer C . Fractions containing GABA:pyruvate-T activity were pooled, concentrated and analyzed as above. The remaining sample was divided into aliquots, frozen in liquid nitrogen and stored at -80 °C. Molecular mass standards (Sigma) were chromatographed under the buffer conditions described above and the relationship between logio molecular mass and retention time was used to estimate the molecular mass of native GABA: pyruvate-T.

Pyruvate-dependent activity was further purified by a combination of affinity- and gel filtration-chromatographyand PAGE, including non-denaturing, IEF and denaturing SDS-PAGE. The Bio-Rad Mini Protean II mini-gel apparatus (Bio-Rad Laboratories (Canada) Ltd., Mississauga, Ontario) was employed for all polyacrylamide gel electrophoresis (PAGE) . Non-denaturing PAGE was performed using a discontinuous system similar to that described by Laemmli (1970, Nature 227: 680-685). The 0.5-mm-thick slab gel had a final monomer acrylamide concentration of 7.5 % and 4 % (w/v), respectively, for the separating and stacking gels. Samples containing 5-25 mg of protein were diluted in an equal volume of sample buffer (62.5 mM Tris-HCl (pH 6) containing 10 % (v/v) glycerol, 0.001 % (w/v) bromophenol blue), loaded into each well and electrophoresed at constant voltages of 75 and 155 V through the stacking and separating gels, respectively. The electrode buffer consisted of 25 mM Tris-HCl (pH 8.3) containing 192 mM glycine. Protein bands associated with GABA:pyruvate-T activity were cut from the native PAGE and loaded into single wells of a 0.75-mm-thick native isoelectric focussing (IEF) slab gel (5 % final monomer acrylamide concentration) containing ampholytes (Bio Rad) of pH ranges 3-10 and 5-7 (final concentration of 2.4 %). Proteins were separated at a constant 200 V for approximately 1.6 h. The electrode buffer in the upper chamber (catholyte) was 20 mM NaOH, whereas the lower chamber buffer (anolyte) was 10 mM phosphoric acid. Broad range pi standards (Pharmacia) were run with each gel. Samples cut from the IEF gel were boiled in an equal volume of sample buffer containing 1 mM EDTA, 0.1 % (w/v) SDS and 0.05 % (w/v) β-mercaptoethanol, loaded onto a 12% (w/v) denaturing SDS-PAGE (4% (w/v) stacking gel) and electrophoresed at constant voltages of 75 and 155 V through the stacking and separating gels, respectively. High molecular weight size standards (BioRad) were included in one lane of each gel. Gels were stained with Bio-Rad Silver Stain or with Coomassie brilliant blue R250. IEF gels to be silver stained were incubated for 10 min in 10 % trichloroacetic acid (TCA) followed by overnight incubation in 1 % TCA. Stained gels were stored in a solution of water containing 25 % (w/v) glycerol until photographed.

To assay pyruvate- and 2-oxoglutarate-dependent GABA-T activity an aliquot of sample (5-300 :1) was incubated for 5-180 min at 37 °C in a reaction mixture containing in a final volume of 500 μl: 50 mM Tris-HCl buffer (pH 8.2), 1.5 mM DTT, 0.75 mM EDTA, 0.1 mM PLP, 10% (v/v) glycerol, 2 mM GABA and 2 mM pyruvate or 1 mM 2-oxoglutarate to initiate the reaction. Assays were cunducted with or without substrate. The reactions were terminated with 6 mM (final concentration) cold (4°) sulphosalicylic acid. Following centrifugation of 20 000 xg for 5 min, the supernatant was neutralized with 1 N NaOH, filtered through a 0.45-μm filter and the amino acid products, alanine or glutamate, were separated by reverse-phase HPLC on an Ultrasphere silica-based HPLC column using buffers described by Oaks et al. (Fundamental ecological and agricultural aspects of nitrogen metabolism in higher plants, eds. H. Lambers, J.J. Neetson and R. Stulen. Martinus Niijhoff, Dordrecht, 1986, pg. 197.) following automatic derivatization with ortho-pthalaldehyde (OP A).

To identify GABA:pyruvate-T on a native gel, a filter paper soaked in 100 mM Tris-HCl (pH 8.2) containing 1.5 mM DTT, 0.15 mM PLP, 200 mM GABA, 12 mM pyruvate, 1 mM NADP and 1.5 units of SSADH, was overlaid onto the gel, and the two incubated in an oven at 37°C for 20 min. The NADPH produced in the linked reaction could be viewed under short wave UN light, thereby identifying the location of GABA:pyruvate-T.

Isolation of GABA-T Gene

Protein bands stained with Coomassie brilliant blue R250 were cut from the denaturing SDS-PAGE gel and washed with 2-5ml volumes of 50 % acetonitrile and frozen at -80 °C. The gel pieces were sent for sequencing to the Harvard Microchemistry Facility (Harvard University, The Biological Laboratories, Cambridge, MA) where the protein sample was enzymatically digested (trypsin protease) and the resultant peptides separated via reverse phase high performance liquid chromatography (HPLC). The family of sequence- defining fragments from a polypeptide, determined using Edman degradation in the presence of a terminating agent, was read with matrix-assisted laser desorption ionizationtime-of-flightmass spectrometry (MALDI-TOF MS) performed on a Finnigan Lasermat 2000 (Hemel, UK), thereby generating a complete data set which acts as a protein sequencing ladder (Chait and Kent, 1992, Science 257: 1885-1894). The mass difference between consecutive peaks defines the identity of a particular amino acid, and is distinctive for all amino acids except Leu/Ile.

The putative identity of a small peptide sequence (Y H L P G E T EE E F S T R (SEQ ID NO: 5) or Tyr His Leu Pro Gly Glu Thr Glu Glu Glu Phe Ser Thr Arg), generated via the methods described above, was determined by comparison of its homology to known amino- and nucleic-acid data using BLAST (Basic-Local-Alignment-Search-Tool;Altschul et al, 1990, J Mol Biol 215:403- 410) computer analysis against GenBankJ, EMBL, DDBJ, PDB and SwissProt databases. The peptide had 94 % homology to anArabidopsis dbEST, GenBank accession N97120, which in turn had more than 30 % homology to many known non-plant gaba-t sequences.

Using the Arabidopsis dbEST N97120 as a template, primer sequences for polymerase chain reaction (PCR) were generated using the Primer 3 software (version 3) on the Primer Picking World-Wide Web (www) server (http://www.genome.wi.mit.edU//cgi-bin/primer/primer3.cgi/). The primers were synthesized at the Molecular Supercentre (University of Guelph). The left- (L, 5' GGG TGT GAT ACC TCC ACC TG 3') (SEQ ID NO: 6) and the right-primer (RI, 5' TCT GGG CTC ATA AGA ATG GC 3') (SEQ ID NO: 7) bound to sites within the dbEST coding sequence. An Arabidopsis thaliana [L.] Heynh (Landsberg erecta ecotype) cDNA library within the yeast expression vector pFL61 , which possesses URA3 as a selectable marker and a phosphoglycerate kinase (PGK) promoter and a PGK terminator (Minet et al. , 1992, The Plant Journal 2:417-422), was used as template to amplify and extend dbEST N97120. Vector primers used for the PCRs were; (1) Left primer (PGK-Lv) 5' TAC AGA TCA TCA AGG AAG TAA TTA T 3' (SEQ ID NO: 8) and (2) Right primer (PGK-Rv) 5' TAT TTT AGC GTA AAG GAT GGG GAA A 3' (SEQ ID NO: 9). Conditions used for the PCR were as follows:

4.25 ml sterile water

2.5 ml VENT® polymerase buffer (10 x) (10 mM KCl, 10 mM (NH₄)₂SO₄, 20 mM Tris-HCl (pH 8), 2 mM MgSO₄ and 0.1 % Triton X-100; New England

Biolabs, Inc., Beverly, MA, USA)

2 ml dNTP (2.5 mM)

1 ml primer L (10 pmol ml^"1)

5 ml primer Fv (2 pmol ml^"1)

0.5 units VENT® polymerase in 9.25 ml

1 ml cDNA library (1 mg ml^"1)

Hot start (94 °C)

30 cycles of 94, 55 and 72 °C for 30, 30 and 50 sec, respectively

10 min at 72 °C followed by a 4 °C hold.

A 1-kb PCR product was obtained using L/PGK-Fv primers on the Landsberg library. The product was "Taq-tailed" , using Taq DNA polymerase from Boehringer-Mannheim, according to the procedure outlined in the Original TA Cloning® Kit manual (Invitrogen, U.S. Patent No. 5,487,993). The Taq- tailed product was gel purified using a 1 % agarose gel in Tris-acetate EDTA (TAE) buffer. The product was cut from the gel and isolated using the Bio 101 Geneclean®II Kit (Bio 101 Inc., Vista CA., product #1001-400). Subsequently, the DNA was ligated and transformed using protocols and materials supplied with the Original TA Cloning® Kit.

Six bacterial colonies were isolated from the transformation (plated on solid Luria-Bertani Medium (LB Medium, pH 7.0) containing 50 mg ml^"1 ampicillin, 10 g L¹ tryptone, 5 g L^"1 yeast extract and 10 g L¹ NaCl (Sambrook et al., 1989, Molecular Cloning: A Laboratory Manual. Second Edition. Cold Spring Harbor Laboratory Press, New York.) and grown overnight in 5 ml liquid LB medium containing 50 mg ml^"1 ampicillin. The plasmids from 1.5 ml of the bacterial culture were isolated using alkaline rapid preparation (ARP; Sambrook et al. , supra) for plasmid DNA. The plasmids were digested for 1 h withEcoRl (20,000 units ml^"1; New England Biolabs) restriction enzyme as described in Sambrook et al. (supra) to ensure that an insert was in the vector. One vector contained an insert approximately 1 kb in length. The remaining bacterial culture housing the plasmid with the correct insert was carried through a plasmid isolation procedure using buffers and protocols included in the QIAprep Spin Miniprep Kit (QIAGEN Inc. Canada, Mississauga, product #27104). The Qiagen-isolated plasmid was gel quantified and sent to Geneologics (Agricultural and Food Laboratory Service Branch, Guelph, ON., Can.) for sequencing (automated sequencer model ABI Prism 377).

Sequencing of the isolated plasmid, using M13 forward and M13 reverse sequencing primers (bacteriophage sequence within the vector), generated partial sequence for the ~ 1 kb insert. Additional sequencing, with primers 3Α-2 forward (5' GCT GGA TGG CAG AGT GCT GAT 3') (SEQ ID NO: 10)and 3Α-2 reverse (5^* GAG GGC AAT CTG TGT GCA ACA C 3') (SEQ ID NO: 11) generated using the Primer 3 program and sequence information from the M13 sequencing, gave complete sequence for the insert, identifying a 342 amino acid polypeptide containing the MET (start) codon. When this sequence was analysed using the Basic Local Alignment Search Tool (BLAST at the National Center for Biotechnology Information; http://www.ncbi. nlm.nih.gov/) against the SwissProt data base it identified multiple non-plant GABA-Ts. The sequence was lacking the 3' termini and had to be extended. Using the new sequence data and dbEST N97120, two new PCR primers were identified using the Primer 3 program. PCR, with GABX-1 forward (5' CCG GAT CCG ATG GTC GTT ATC AAC AGT CTC CG 3') (SEQ ID NO: 12) and GABX-1 reverse (5' CGG AAT TCT TCA CTT CTT GTG CTG AGC C 3') (SEQ ID NO: 13) primers and with the Arabidopsis library as template, identified an open reading frame of 1515 bp within the 1737 bp cDNA as set out in Fig. 2. The open reading frame codes for a polypeptide of 505 amino acids to the first termination codon, as shown in Fig. 3, which had more than 30% identity to many known non-plant gaba-t sequences. A submitted nucleotide sequence (GenBank Accession Number AF085149) identified as a probable aminotransferase(M. Aluru, J. Curry, M. O'Connell 1998 Plant Physiology 118:1102) from Capsicum chinense strain Habanero, has 76% identity to Arabidopsis th liana GABA-T at the amino acid level (over 450 a. a.). The Arabidopsis GABA-T has 42 N-terminala.a. more than the Habanero sequence. Expression of GABA-T in E. coli

The resulting PCR product ( ~ 1.7 kb), and pTrcHisB (pTrcHis Xpress™ Kit; Invitrogen, product #K860-01), were digested withEcøR/ and BamHI restriction enzymes. The digested PCR product and the vector were gel purified (1 % agarose gel in TAΕ buffer) and eluted using the Bio 101 Geneclean^® II Kit (Bio 101 Inc., Vista CA., product #1001-400). Ligation and transformation was carried out according to standard protocols. To ensure that the sequence was cloned in frame for proper expression, pTrcHisB-GABA-T was amplified in dH5V cells and sequenced. Subsequently, pTrcHisB-GABA-T was transformed into ToplO cells and selected on SOB medium containing ampicillin (50 μg/ml). Pilot expression experiments to determine kinetics of induction of GABA-T were performed according to Xpress™ System Protein Expression TrcHis Instruction Manual.

To isolate recombinant protein fractions, a 450-ml culture of cells grown in SOB containing ampicillin (50 μg/ml) was treated with 1 mM IPTG when the ODeoo reached 0.6. After 4 h, 37 ml of ToplOE. coli cells containing GABAT:pTrcHisB or 37 ml of ToplOE. coli cells containing no insert:pTrcHisB were pelleted at 3,000 xg for 10 min. The resulting pellet was frozen and thawed and then resuspended in 5 ml of: 50 mM Tris (pH 8.2), 1.5 mM DTT, 0.625 mM EDTA, 0.1 mM PLP, 10% glycerol, 20 mg lysozyme, 12 mM MgCk, 45 μg DNase, 0.5 mM PMSF, 12.5 μg Pepstatin A and 12.5 μg Leupeptin. The mixture was incubated for 30 min at 4 °C to lyse the cells. The mixture was spun at 13,000 xg for 10 min and the pellet was resuspended in 2.5 ml of buffer (50 mM Tris (pH 8.2), 1.5 mM DTT, 0.625 mM EDTA, 0.1 mM PLP, 10% glycerol, 0.5 mM PMSF, 12.5 μg pepstatin A and 12.5 μg leupeptin) containing 6 mM CHAPS detergent. The mixture was incubated at 4 °C for 45 min with gentle rocking. The debris was pelleted at 13,000 xg for 10 min and the supernatant was desalted using a Sephadex G25, PD-10 column (Pharmacia) equilibrated with 50 mM Tris (pH 8.2), 1.5 mM DTT, 0.625 mM EDTA, 0.1 mM PLP and 10% glycerol (v/v).

Expression of GABA-T was determined as pyruvate-dependent GABA-T activity using the methods described above. Activity of the unpurified enzyme from the G4J_L4_T:pTrcHisB line was 2.6 nmol/ mg protein/ min. Pyruvate- dependent GABA-T activity from the no insert: pTrcHisB line was not detectable. Example 2 - Identification and Characterization of GHBDH Competent Yeast Cells and Transformation

Competent yeast cells were made according to Dohmen et al. (1991). A glycerol stock of the S. cerevisiae strain 22.641c (MATa, ura3-l, uga.2-1), deficient in succinate semialdehyde dehydrogenase activity (provided by B. Andr — , University of Brussels), was streaked onto solid YPD medium and incubated at 28 °C. After 5 days, a single colony was selected and used to inoculate 1 ml of liquid YPD medium (2% glucose, 1 % Bacto-y east extract, 2% Bacto-peptone). Cells were incubated overnight at 28 °C. The following day the overnight culture was added to 200 ml of YPD and grown to an ODόooof 0.6. Cells were centrifuged (3000 xg for 5 min) and the pellet washed with 30 ml of Solution A (1 M sorbitol, 2% ethylene glycol, 10 mM Bicine; pH 8.35). The cells were centrifuged again (3000 xg for 5 min) and the pellet resuspended in 2 ml of Solution A. The cells were aliquoted into 100-ml volumes and stored at -70 °C.

Competent cells were transformed with an Arabidopsis cDN A expression library (as described above in Example 1) according to Dohlmen et al. (Yeast 7: 691-692,1991). For each transformation (12 in total), the following was added to 100 ml of frozen yeast cells; 5 ml of sonicated carrier DNA (1 mg ml^"1), 5 ml of 1 M histamine and 1 ml of Arabidopsis cDNA library (1 mg ml^"1). Cells were thawed in a thermomixer with shaking (37 °C for 5 min) before the addition of 1 ml SolutionB (40% polyethylene glycol 1000, 200 mM Bicine; pH 8.35) and cells incubated in a roller drum for 1 hour at 30°C. Subsequently, the cell suspensions were centrifuged (3000 xg for 2 min) and the pellet resuspended and washed in 0.5 ml of Solution C (150 mM NaCl, 10 mM Bicine; pH 8.35). The cells were centrifuged again (3000 xg for 2 min) and the pellet resuspended in 100 ml of solution C and plated on selective medium. In control transformations, competent cells were transformed with 0.1 mg of the empty pFL61 vector. Selection of Transformants

Primary transformants were identified by screening on solid SD medium (2% glucose, 0.5% (NH»)₂SO , 0.17% yeast nitrogen base, without amino acids or ammonium sulfate) for complementation of the nutritional deficiency inuracil biosynthesis. Ura⁺ transformants (approximately 1 x 10⁶ colonies in total from 12 independent transformations) were independently washed from each plate with nitrogen-free liquid SD medium and re-selected on nitrogen-free solid SD medium supplemented with 20 mM GABA as the sole nitrogen source. No background growth of the control transformants was evident, thereby allowing for effective screening of Arabidopsis cDN A candidates.

Recombinant plasmids from candidates capable of growth on GABA (4 clones) were isolated as follows. Single colonies were used to inoculate 2-ml aliquots of liquid SD medium and were grown overnight at 28 °C in roller drum. The following day, 1.5 ml of the overnight culture was transferred to a microfuge tube and the cell pellet (3000 xg for 3 min) was resuspended in 100 ml of Solution 1 (1.2 M sorbitol, 0.1 M sodium citrate, 10 mM EDTA, 0.8% 2-mercaptoethanol and 0.1 % lyticase; pH 7.0) and incubated at 37 °C for 40 min with gentle agitation. Subsequently, 100 ml of Solution 2 (50 mM Tris, 10 mM EDTA and 2% SDS; pH 7.0) was added and mixed gently before the addition of 50 ml of Solution 3 (3 M potassium acetate; pH 7.0). The lysate was vortexed and incubated at 4 °C for 30 min. Subsequently, 500 ml of CIP (chloroform: isoamyl alcohol: phenol; 24: 1:25, respectively) was added, and the suspension vortexed and incubated for 5 min at room temperature. The suspension was microfuged (20 500 xg at 4 °C for 10 min) and the aqueous layer transferred to a clean microcentifuge tube. Two and half volumes of 95 % ethanol was added to the aqueous layer, and the mixture inverted several times. After centrifugation (20 500 xg at 4 °C for 10 min) the DNA pellet was washed twice with 1-ml aliquots of 70% ethanol, dried and resuspended in TE buffer (10 mM Tris, 1 mM EDTA, ; pH 7.0).

Isolated plasmids were amplified in E. coli dH5α strains and re-introduced into the yeast mutant for re-selection on 20 mM GABA. This confirmed that no reversion or second-site mutations had occurred. Further characterization of these plasmids was performed by restriction digest analysis and full or partial sequencing of up to 500 bp at both 3'- and 5 '-ends. This revealed one class of plasmid. DNA Sequencing

The plasmids bearing cDNAs isolated by complementation were sent to Genologics (Agricultural and Food Laboratory Service Branch, Guelph, ON., Canada) for sequencing (ABI PRISM Dye Terminator Cycle Sequencing Ready Reaction Kit on the ABI PRISM Sequencer Model 377; Perkin Elmer). Vector primers used for sequencing were PGK5 ' (5 '-TCA AGA TCA TCA AGG AAG TAA TTA T-3') (SEQ ID NO: 14) and PGK 3' (5'-TAT TTT AGC GTA AAG GAT GAG GAS A-3') (SEQ ID NO: 15).

Sequence analysis of an insert isolated from pFL61 identified a 867-bp open reading frame within the 1032-bp cDNA. The open reading frame codes for a protein containing 289 amino acids with a calculated molecular mass of 31.4 kDa. The protein possesses several potential structural motifs including: 5 casein kinase II and 3 protein kinase C phosphorylation sites, 4 N-myristoylation sites, 2 glycosylation sites and 1 amidation site (PredictProtein, EMBL-Heidelberg). We have designated this gene as 4-HBDH for 4-hydroxybutyrate dehydrogenase (E.C.1.1.1.64). Northern blot analysis of Arabidopsis leaf RNA using 4-HBDH as the probe detected a message of ~ 1.3 kb (data not shown), demonstrating that this gene identified by complementation was derived from plantcDNA. A search of the GenBank™ data base did not identify any homology to known 4-hydroxybutyratedehydrogenases or succinic semialdehyde reductases from plant, fungal, bacterial or animal sources, which suggests that 4-HBDH belongs to a family of oxido-reductases that have not been previously identified. However, the encoded protein sequence did exhibit some sequence similarity (from 33% to 20% identity) to several oxido-reductases, including 3- hydroxyisobutyrate dehydrogenase and 6-phosphogluconate dehydrogenase from numerous bacterial sources. Yeast Growth Assay

The transformed yeast strains and the clones found by complementation were then grown in 50 ml of liquid SD medium to an ODfcoo of 0.5. Cells were washed twice in nitrogen-free SD medium (3000 g for 5 min), and approximate amounts (full inoculating loop) streaked onto nitrogen-free SD plates supplemented with either GABA, proline or (NH )₂SO₄ (20 mM N) The plates were grown for 4 days at 28 °C (Fig. 6). The wild type strain and mutant cells expressing 4-HBDH grew well on each of these nitrogen sources, whereas the mutant cells expressing pFL61 did not grow on GABA. Determination of GABA and 4-hydroxybutyrate (HB) levels in yeast extracts

Single colonies from wild type (S1278b), mutant (22.641c) transformed with the empty yeast vector (pFL61) or Arabidopsis complement (2A), were used to inoculate 1-ml aliquots of liquid SD media and grown overnight in a rotary shaker (150 rpm at 28E C). The following day, each of the three overnight cultures were divided into 3 aliquots and used to inoculate 50 ml of liquid nitrogen-free SD media supplemented with either GABA, proline or (NH )₂SO (20 mM N). The cells were grown to mid-log phase (ODsoo of 1.0 +/- 0.1) at 28 °C for 12-96 hours (depending on the strain). Cells were harvested by centrifugation (5000 xg at 4 °C for 5 min), washed with 50 ml of cold 0.3 M sorbitol, and centrifuged again (5000 xg at 4 °C for 5 min). The pellet was suspended in 50 ml of cold 0.3 M sorbitol, and the OD«χ> re-measured, thereby allowing the cell density of each suspension to be determined. The cells were concentrated by centrifugation (5000 xg at 4 °C for 5 min) to a calculated cell density of 40 absorbance units ml^"1 (~ 1.2 x 10⁹ cells). To break open the cells, 1 g of washed silica and 1 ml of 95 % methanol were added to each yeast pellet and the samples vigorously vortexed (3 x 20 seconds). The broken cells were then microfuged (20 500 g for 5 min), and the supernatant transferred to a clean microfuge tube before the pellet was re- extracted with 1 ml of 70 % methanol by vigorous vortexing (3 x 20 seconds). After microcentrifugation(20 500 g for 5 min), the supernatant from the second extraction was combined with that of the first. The methanol-derived extracts were then analyzed for 4-HB and GABA levels by mass spectrometry according to Gibson et al. (Biomed Environ Mass Spectrom 9: 89-93, 1990) and Kok et al. (J Inter Metab Dis 16: 508-512, 1993).

Yeast cells expressing 4-HBDH had considerably higher levels of 4- hydroxybutyrate (90 μmol (g DW )^_1) than either mutant cells expressing pFL61 (10.9 μmol (g DW) ^]) or wild type cells (0.2 μmol (g DW)!¹), when grown on 20 mM GABA (Table 1). Negligible 4-hydroxybutyrate was synthesized by all strains when grown on proline or NH ⁺as the sole nitrogen source.

Table 1. Comparison of GABA and 4-Hydroxybutyrate (4-HB) pool sizes in yeast strains grown on various nitrogen sources (20 mM N). N.D. indicates not detected. Data represent mean ± SE of four replicates.

Strain Nitrogen GABA pool 4-HB pool source (mol g^"1 DW) (:mol g^"1 DW)

Wild type (NH₄)₂SO₄ 0.081 ± 0.007 0.012 ± 0.012

GABA 59.8 ± 2.0 0.199 ± 0.042 proline 0.666 ± 0.079 0.082 ± 0.051

pFL61 (NH₄)₂SO₄ 0.096 ± 0.013 N.D.

GABA 604 ± 26 10.9 ± 0.3 proline 0.445 ± 0.026 0.117 ± 0.042

4-HBDH (NH₄)₂SO₄ 0.091 ± 0.004 0.023 ± 0.012

GABA 561 ± 29 90.0 ± 5.9 proline 0.467 ± 0.020 0.059 ± 0.031

Expression of 4-HBDH in E. coli

The recombinant yeast expression vector, pFL61 bearing 2A, was used as template to generate a cDNA with EcoRI and ItamHI ends using the primers P2A- for (5' GCC GGA TCC AAT GGA AGT AGG GTTTCT 3')(SEQ ID NO: 16) and P2A-rev ( 5' CCG GAA TTC AAA TGT GTG TTT GGC C 3')(SEQ ID NO: 17). The resulting PCR product ( ~ 1 kb), and pTrcHisB (pTrcHis Xpress Kit; Invitrogen, product #K860-01), were digested withEcoRI and BamHl restriction enzymes. The digested PCR product and the vector was gel purified (1 % agarose gel in TAΕ buffer) and eluted using the Bio 101 Geneclean®II Kit (Bio 101 Inc., Vista C A., product #1001-400). Ligation and transformation was carried out according to standard protocols. To ensure that the sequence was cloned in frame for proper expression, pTrcHisB-GHBDH was amplified in dH5α cells and sequenced. Subsequently , pTrcHisB-GHBDH was transformed into ToplO cells and selected on SOB medium containing ampicillin (50 mg ml^"1). Pilot expression experiments to determine kinetics of induction of GHBDH was performed according to Xpress™ System Protein Expression TrcHis Instruction Manual.

To isolate recombinant protein fractions, a 450-ml culture of cells grown in SOB (2% Bacto-tryptone, 0.5% Bacto-y east extract, 0.05% NaCl) containing ampicillin (50 mg ml^"1) was treated with 1 mM IPTG when the ODωo reached 0.6. After 4 h the cells were pelleted by centrifugation (13 000 xg at 4 °C for 15 min) and resuspended in 10 ml of extraction buffer (100 mg mT¹ lysozyme, 2 M MgCk, 3 ml ml^"1 DNAse, 120 mM PMSF, 2.5 mg ml^"1 leupeptin and 2.5 mg ml^"1 pepstatin A). Cells were incubated at 4 °C for 30 min, sonicated 5 x 20 s, pelleted (13 000 xg at 4 °C for 15 min). The pellet was resuspended in 50 mM Hepes (pH 7.2) containing 10% glycerol (v/v), 6 mM CHAPS and incubated at 4 °C for 30 min. Cellular debris was pelleted (13 000 xg at 4 °C for 15 min) and the supernatant desalted using a Sephadex G25, PD-10 column (Pharmacia) equilibrated with 50 mM Hepes buffer (pH 7.2) containing 10% glycerol (v/v).

Expression of 4-HBDH was determined by measuring the oxidation of NADPH to NADP by a decrease in absorbance at 340 nm using a method adapted from Hearl and Churchich (J Biol Chem 260: 16361-16366), 1985). Standard assay mixtures contained 100 mM potassium phosphate buffer (pH 7.2) containing 10% glycerol (v/v), 1.5 mM succinic semialdehyde, 0.5 mM NADPH, and 200 ml of sample. Assays were intiated by the addition of succinic semialdehyde after a 5 min incubation and were performed at 25 °C.

The GHBDH activity of one recombinant clone (4-HBDH) was measured. Only the membrane soluble fraction possessed activity, which was linear for 5 min of incubation. The enzyme preferred NADPH as a cofactor, over NADH and possessed negligible activity for the conversion of 4-HB to SSA in the reverse direction. The addition of 1 mM pyridoxal 5 '-phosphate fully inhibited enzyme activity, presumably by interfering with a lysine residue at the active site. Activity of the unpurified GHBDH from the GHBDH :pTrcHisB line was 13.3 nmol/ mg protein/ min. GHBDH activity from the no inser pTrcHisB line was not detectable. Example 3 - Production of Antibodies from Synthetic Peptides of GABA-T and GHBDH

Identification and Synthesis of Peptides

Using a number of World Wide Web (WWW) tools and guidelines, we established a protocol for prediction and synthesis of antigenic peptides from GABA-T and GHBDH protein sequences. The primary web site is located at the following address: http://www.sander.emble- heidelberg.de/future/overview/antigen.html. The criteria for prediction were: choose peptides which are surface exposed and hydrophilic; do not choose peptides which are helical; N- and C-terminal peptides often work better; avoid peptides with internal sequence repeats; avoid low complexity sequences; try to avoid proline and cysteine residues; and choose peptides with high calculated antigenicity.

Initially, the "ScanProsite " program, within http://www.expaxy.ch, was used to identify sequence abnormalities. For GABA-T these included a mitochondrial signal peptide (a.a. 1-35), a membrane spanning domain (a.a. 322- 342) and a pyridoxal-1 -phosphate binding domain (a.a. 295-331). These regions were avoided during peptide selection. The GABA-T amino acid sequence was also scanned for accessible residues using the "ProtScale" program within http://www.expaxy.ch. The target regions identified included a.a. 170-200, 220- 260 and 480-504. Using a Hopp and Woods hydrophobicity plot, three hydrophilic regions were identified and coincided with the accessible regions above. The sequence was further examined for antigenic regions via an "Antigenicity Plot" program located at http://www.virus.kyoto- u.ac.jp/JaMBW/3/1/7/index.html. The algorithm for this prediction program was developed by Hopp and Woods (Hopp,T.P. and Woods,K.R. 1981 Prediction of protein antigenic determinants from amino acid sequences. Proc Natl Acad Sci USA 86: 152-15). The three regions identified using the previous two criterion were also identified as being antigenic. Together, this information indicates that there are three regions from which a peptide can be chosen (a.a 170-200, 220-260 and 480-504). The size of the peptide can vary. The two peptides chosen for GABA-T were: (1) C YYNNALGRPEKKK 14 a.a. (171-183) (SEQ ID NO: 18)

(2) TEEKVKELKAQHKK C 15 a.a. (491-504) (SEQ ID NO: 19)

The N- or C-terminal cysteine was added to link the peptide to the keyhole limpet hemocyanin (KLH) immunogen carrier as described in the Imject^®Maleimide Activated mcKLH Kit from Pierce (Rockford, IL.). The entire procedure was carried out according to the manufacturer's instructions.

The identical procedure to that described above was followed for the GHBDH amino acid sequence. No abnormal regions were identified; therefore, no regions were avoided in the evaluation for peptide location. Four regions of surface accessibility were identified (a.a. 115-130, 180-205, 210-230 and 260- 280). Of these, the sequence within the 115-130 a.a. region also had hydrophillic properties slightly greater than the remaining a.a. stretches within the sequence. Determination of antigenicity revealed that two regions (a.a. 115-130 and 230- 240) were antigenic. The first region coincides with those chosen by the other criteria and was, therefore, the region from which the peptide was derived:

(1) C EGPVSGSKKPA 12 a.a. (117-127) (SEQ ID NO: 20)

Again, the N-terminal cysteine was added to link the peptide to the KLH immunogen carrier.

The three sequences were used to generate 3 peptides:

MW Amount Synthesized

OVCpepl (GABA-T 1) 1724.64 118 mg

OVCpep2 (GABA-T 2) 1797.44 137 mg

OVCpep3 (GHBDH 1) 1200.85 105 mg

The synthesis of these peptides was carried out by the Biotechnology Service Centre, Department of Clinical Biochemistry, University of Toronto, Toronto, ON. Production and Evaluation of Antibodies

The synthesized peptides were bound to KLH, the antigen/immunogen complex was mixed with an equal volume of Freund's Incomplete Adjuvent (Sigma), and the mixture was injected into chickens according to the following protocol and procedure. Two pairs of leghorn chickens were used for this study. Eggs were collected from each chicken for 10 days prior to the first injections. These eggs were used to harvest pre-immune chicken IgY. One pair of chickens was inoculated with the GABA-T peptide antigens. Each of the two chickens received 100 μg of each of OVCpepl and OVCpep2 every 10 days for 30 days beginning on day zero (4 injections). For GHBDH, one pair of chickens were each injected with 100 μg of OVCpep3 under the same schedule described for GABA-T. Eggs were collected daily from each chicken and stored at 4 °C until total IgY isolation could be carried out.

The isolation protocol for total IgY was provided by Dr. Yoshinori Mine (Department of Food Science, University of Guelph). One egg yolk was mixed with 40 ml of dH O and stirred. The mixture was stored at -20 °C for 2 days. After thawing at room temperature, the samples were centrifuged at 20,000 rpm for 30 min at 10 °C. The pellet was discarded and an equal volume of saturated ammonium sulfate was slowly added to the supernatant under constant stirring. The mixture was incubated at 4 °C overnight followed by centrifugation at 10,000 rpm for 30 min at 10 °C. The resulting pellet was resuspended in 5 ml of phosphate buffered saline (PBS, pH 7.4) and dialyzed against PBS for 2 days at 4 °C with frequent buffer changes. The solution was removed from the dialysis tubing and stored at -20 °C overnight. These samples, representing total crude IgY, were freeze-dried for storage at 4 °C.

From total crude IgY, the antibodies specific to each peptide were isolated using affinity columns (with specific peptide as ligand) prepared using the SulfoLink^® Kit (Peirce, Rockford, IL.) according to the manufacturer's instructions. PAGE and Western Blot Analysis of GABA-T and GHBDH

A denaturing gel of total crude protein from induced ToplOE. coli cells containing either no insert:pTrcHisB or GAR4_T:pTrcHisB was examined for a prominent band representing recombinant GABA-T . Although no protein band seemed to be preferentially induced (note that an induction time-course revealed no difference in accumulation between 4 and 22 hrs) (Figure 7A), a Western blot, probed with the AntiExpress antibody did recognize a single band in the 'with insert' lane at approximately 53 kDa (Figure 7B). This band corresponds to the predicted size of recombinant GABA-T. The presence of the band suggests that although GABA-T was expressed, its level was too low to detect withCoomassie staining. With the anti-GABA-T specific antibody, no bands were identified in the 'no insert' control (Figure 7C). However, in the 'with insert' lane the recombinant GABA-T band at 53 kDa, and a prominent smaller band at 48 kDa were observed. Since the smaller band was not present in the 'no insert' lane, it is likely a proteolytic product of recombinant GABA-T. The likely reason that the 48-kDa band was not apparent in the AntiExpress blot was that the degradation occurred from the N-terminus where the Express portion is located. In contrast, the GABA-T specific antibody targets the C-terminal portion of the protein.

Examination of denatured total crude protein from induced ToplOE. coli cells containing either no insert:pTrcHisB or GHBDH :pTrcHisB revealed a prominent band representing GHBDH at approximately 35 kDa, that was present in the 'with insert' lane only (Figure 8A). The band was at a relatively higher level of expression than other proteins. The AntiExpress antibody did not detect any protein in the 'no insert' lane, whereas it did recognize a single band corresponding to the recombinant GHBDH at approximately 35 kDa in the 'with insert' lane (Figure 8B). Similarly, the anti-GHBDH specific antibody did not reveal any bands in the 'no insert' control (Figure 8C). In contrast, the same antibody recognized the recombinant GHBDH at 35 kDa, and two smaller bands in the 'with insert' lane. Since these smaller bands were not present in the 'no insert' lane, they were apparently proteolytic products of recombinant GHBDH. They were probably not apparent in the AntiExpress blot because degradation occurred from the N-terminus where the Express portion was located. In contrast, the GHBDH-specific antibody targets the middle of the protein.

Claims

We Claim:

1. An isolated polynucleotide comprising a nucleotide sequence encoding gamma aminobutyric acid transaminase of plant origin.

2. A polynucleotide as defined in claim 1 , wherein the nucleotide sequence encodes the amino acid sequence set out in SEQ ID NO: 2.

3. A polynucleotide as defined in claim 1 , wherein the nucleotide sequence corresponds to that of SEQ ID NO: 1.

4. A polynucleotide comprising a nucleotide sequence encoding gamma hydroxybutyrate dehydrogenase of plant origin.

5. A polynucleotide as defined in claim 4, wherein the nucleotide sequence encodes the amino acid sequence set out in SEQ ID NO: 4.

6. A polynucleotide as defined in claim 4, wherein the nucleotide sequence corresponds to that of SEQ ID NO: 3.

7. A recombinant DNA construct having incorporated therein a polynucleotide as defined in claim 1.

8. A recombinant DNA construct having incorporated therein a polynucleotide as defined in claim 4.

9. A cell that has been engineered genetically to produce plant gamma aminobutyric acid transaminase, said cell having incorporated expressibly therein a heterologous polynucleotide as defined in claim 1.

10. A cell that has been engineered genetically to produce plant gamma hydroxybutyrate dehydrogenase, said cell having incorporated expressibly therein a heterologous polynucleotide as defined in claim 4.

11. A process for obtaining a substantially homogeneous source of plant gamma aminobutyric acid transaminase, comprising the steps of culturing cells having incorporated expressibly therein a polynucleotide as defined in claim 1 , and then recovering said cultured cells.

12. A process for obtaining a substantially homogeneous source of plant gamma hydroxybutyrate dehydrogenase , comprising the steps of culturing cells having incorporated expressibly therein a polynucleotide as defined in claim 4, and then recovering said cultured cells.

13. Plant gamma aminobutyric acid transaminase, in a form essentially free from other proteins of plant origin.

14. Plant gamma aminobutyric acid transaminase as defined in claim 13, encoded by a polynucleotide having the nucleotide sequence set out in SEQ ID NO: 1.

15. Plant gamma aminobutyric acid transaminase as defined in claim 13, having the amino acid sequence set out in SEQ ID NO: 2.

16. Plant gamma hydroxybutyrate dehydrogenase, in a form essentially free from other proteins of plant origin.

17. Plant gamma hydroxybutyrate dehydrogenase as defined in claim 16, encoded by a polynucleotide having the nucleotide sequence set out in SEQ ID NO: 3.

18. Plant gamma hydroxybutyrate dehydrogenase as defined in claim 16, having the amino acid sequence set out in SEQ ID NO: 4.

19. Antibody to plant GABA-T.

20. Antibody to plant GHBDH.