Background of the invention A wide range of yeasts and moulds, among them Candida cloacae and a number of other industrial yeasts, have the ability to use alkanes or fatty acids as the sole carbon source (Watkinson & Morgan 1990 Biodegradation 1,79-92; Kemp et al. 1994 Appl. Microbiol. Biotechnol. 40,873-875) by a diterminal pathway (i. e. these organisms metabolise the substrate by oxidising both ends of the molecule, unlike most bacteria which ß-oxidise one end of the substrate molecule). This is achieved using metabolic pathways in three separate subcellular compartments: endoplasmic reticulum, peroxisomes and mitochondria (see Nauersberger et al, 1987 J. Basic Microbiol. 27,565-582). <BR> <BR> <P>Following import of the fatty acid or hydrocarbon (typically an alkane or an alkene), a oxidation is catalyzed by three sequential enzymes: (a) a P450 linked hydrocarbon/fatty acid oxidase to yield a fatty alcohol, a, ctv-fatty diol or an co-hydroxy fatty acid; (b) a fatty alcohol/fatty diol oxidase, which in the presence of molecular oxygen yields the corresponding fatty aldehyde and H202, and (c) an aldehyde reductase. The immediate product of step (c) is either a monocarboxylic acid or a dicarboxylic acid. Subsequent metabolism of dicarboxylic acids occurs following activation to acyl CoA via 0-oxidation in the peroxisomes. Finally, acetyl CoA is oxidised to CO2 in the mitochondria.

There has been much interest in peroxisomal proliferation and its regulation following alkane induction. Comparatively less attention has been paid to the a, w-oxidation pathway though there has been commercial interest in generating activated fatty acids by preventing further oxidation of dicarboxylic acids (Casey et al., 1990 J. Gen. Microbiol. 136,1197-<BR> <BR> <BR> <BR> <BR> <BR> 1202; Picataggio et al., 1992 Bio/Technology 10,894-898).

Microorganisms possessing this determinal pathway are used commercially in the industrial production of dicarboxylic acids, which are important building blocks for the chemical industry and ingredients for personal care products. Another potential use for these organisms is to clean up environmental pollutants that are alkane-based. Specific strains of both C. cloacae and C. tropicalis which are disrupted in 0-oxidation have been developed for the commercial production of dicarboxylic acids: these strains accumulate activated fatty acids at high levels in the medium (Casey et al. 1990, Picataggio et al.

1992, cited above).

There has been a previous report of the purification of the alcohol oxidase enzyme from <BR> <BR> <BR> <BR> C. cloacae (West et al., 1995 in Plant Lipid Metabolism, Eds. Kader & Mazliak, Kluwer Academic Publishers). However, those authors reported that they were unable to obtain amino acid sequence data on the protein, even from 90 % pure protein preparations. There is a similar report of the purification of a fatty alcohol oxidase enzyme from C. tropicalis (Dickinson & Wadworth 1992 Biochem. J. 282,325-331). Again, however, those authors did not provide any amino acid sequence data. Moreover, C. tropicalis is an organism which is often refractory to transformation and therefore not an organism which is easy to manipulate genetically.

Summary of the invention In a first aspect the invention provides an isolated nucleic acid sequence encoding a polypeptide having fatty alcohol oxidase (FAO) activity. Fatty alcohol oxidase activity is the catalysis of the oxidation of the terminal hydroxy group of a fatty alcohol, or an hydroxy fatty acid, in the presence of molecular oxygen, to form a fatty aldehyde or a fatty acid aldehyde, and H202. FAO activity may conveniently be assayed by the method of Kemp et al, (1988 Appl. Microbiol. Biotechnol. 29,370-374).

It will be understood that a"fatty alcohol oxidase"enzyme can be distinguished from an "alcohol oxidase"by virtue of the respective substrate specificities. Thus, for present purposes, a fatty alcohol oxidase is an enzyme which, inter alia, exhibits greater specific activity for substrates having a chain length of 5 or more C atoms. In addition, fatty alcohol oxidases may often oxidise secondary alcohols (e. g. dodecan-2-ol). Long chain (i. e. 5 or more, preferably 10 or more, C atoms) w-hydroxy fatty acids may be particularly preferred substrates of fatty alcohol oxidases.

Such a nucleic acid sequence has never before been obtained. The present inventors successfully isolated two types of FAO coding sequence from the yeast C. cloacae. Using this information, the inventors were able successfully to isolate a similar further FAO coding sequence from the yeast C. tropicalis, by screening a cDNA library (prepared from cells grown under conditions in which expression of FAO is induced) with sequence- specific probes based on the C. cloacae FAO coding sequence. Those skilled in the art will appreciate that essentially identical techniques should enable the isolation of FAO coding sequences from other organisms, especially from other yeasts (particularly of the genus Candida) or fungi.

The nucleic acid sequence of the invention is preferably a sequence obtainable from a yeast, especially from yeasts of the genus Candida. Examples include C. cloacae and C. tropicalis, the FAO coding sequences of which are disclosed herein.

In a particular embodiment, the invention provides a nucleic acid sequence encoding a polypeptide having fatty alcohol oxidase activity and comprising substantially the amino acid sequence of one of the sequences shown in Figure 7, or a functional equivalent thereof. Functionally equivalent amino acid sequences are those which possess FAO activity yet which are not identical with one of the sequences shown in Figure 7-one or more amino acid substitutions may be present, without substantially affecting the FAO activity of the polypeptide, especially where the substitutions are conservative (e. g. leucine for isoleucine, threonine for serine etc). In particular, analysis of the amino acid sequences shown in Figure 7 reveals portions of the FAO molecule which appear to be less highly conserved, and therefore are more likely to be tolerant of amino acid substitutions. In general, a functionally equivalent amino acid sequence will possess at least 70 % similarity (i. e. identical amino acid residues), preferably at least 75 %, and more preferably at least 80% similarity, with one of the sequences shown in Figure 7.

Further referring to Figure 7, it will be apparent that certain amino acid sequences tend to be conserved among the FAO polypeptides of different species. This information can be employed in order to isolate nucleotide sequences encoding other FAO molecules. For example, an organism whose FAO gene or coding sequence it is desired to clone, can be grown under circumstances such that the FAO polypeptide is highly expressed, and a cDNA library prepared from the organism. The resulting cDNA library can then be screened using degenerate oligonucleotide probe (s) designed to hybridise (preferably under stringent hybridisation conditions) to any nucleic acid sequence encoding one of the conserved amino acid motifs shown in Figure 7. More conveniently, the cDNA library will be subjected to PCR amplification, using a pair of oligonucleotide primers, each primer hybridising to a sequence encoding a conserved amino acid motif. These methods are now routine and within the ambit of those of normal skill in the art. Suitable guidance, if required, is given by Sambrook et al, (Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Press).

Particular amino acid motifs which appear from Figure 7 to be conserved include: IIGSG (X) GAGWA; AGSTFGGG; NWSACLKTP; CG (X) CHLGC; IG (X) NL (X) LHPVS; SAHQMS (X) CRMSG; and PTASG (X) NPM (Seq. ID Nos. 1-7 respectively). It is likely therefore, that other FAO-coding nucleotide sequences will possess nucleotide sequences which encode one or more of the above-specified motifs, and that other FAO polypeptides comprise one or more of the motifs, and possibly other regions wich are conserved in Figure 7.

Specific embodiments of the invention include nucleotides 1127-3220 of the nucleic acid sequence shown in Figure 6, and nucleotides 1702-3814 of the nucleic acid sequence shown in Figure 8, and sequences which hybridise under stringent hybridisation conditions (e. g. as described by Sambrook et al, Molecular Cloning. A Laboratory Manual, CSH i. e. washing with O. lx SSC, 0.5% SDS at 68°C) with the complement of nucleotides 1127-3220 in Figure 6 and/or the complement of nucleotides 1702-3814 in Figure 8. The degeneracy of the genetic code is such that nucleic acid sequences very different to those shown in Figures 6 and 8 may still encode polypeptides which are functional equivalents of those amino acid sequences shown in Figure 7 and accordingly are encompassed within the scope of the present invention.

The nucleic acid sequences of the invention will preferably comprise one or more regulatory elements in addition to the coding sequences. These may comprise, for example, 5'and/or 3'untranslated regions, promoters (for prokaryotic or eukaryotic expression systems), terminators, polyadenylation signals, enhancers and the like. Specific examples of regulatory elements are shown in Figures 6 and 8. In particular, the invention provides replicable nucleic acid constructs comprising the nucleic acid sequence of the first aspect. Desirably the constructs will allow for expression of a fatty alcohol oxidase polypeptide in a suitable host cell, which may be prokaryotic or eukaryotic. The construct may also advantageously comprise a selectable marker (e. g. a gene coding for resistance to an antibiotic such as hygromycin, neomycin, ampicillin, tetracycline and the like).

In a further aspect, the invention provides a host cell into which has been introduced a nucleic acid sequence in accordance with the first aspect of the invention. The host cell may be eukaryotic (e. g. mammalian cell, plant cell, yeast cell such as Candida or Pichia <BR> <BR> <BR> <BR> sp or a fungal cell) or prokaryotic (a bacterial cell such as E. coli). Methods of introducing nucleic acid sequences into host cells are well-known. Such methods may generally be referred to as"transformation"and include classical Ca2+-mediated transformation of bacterial cells, transfection, transduction, protoplast fusion,"biolistic" methods, electroporation and the like. The host cell may be one which does not naturally possess a FAO coding sequence, such that introduction of the nucleic acid sequence may allow the host cell to express a new protein (i. e. FAO). Alternatively, the host cell may be one which already possesses a FAO coding sequence. In this instance, introduction of the sequence of the invention may allow the host cell to express a different FAO to that which is naturally produced by the host cell (e. g. an enzyme with a different pH optimum, or a different substrate specificity). Optionally, several copies of the sequence of the invention may be introduced into the host cell, so as to increase the copy number of the FAO coding sequence, thereby increasing the amount of FAO enzyme produced by the host cell.

Thus the invention also provides a method of transforming a host cell by introducing into the host cell a nucleic acid sequence directing the expression of a FAO polypeptide, preferably such that the transformed host cell produces FAO enzyme. Further, the invention provides a method of altering a substrate, the method comprising contacting the substrate with a host cell transformed with the nucleic acid sequence of the invention, whereby the substrate is metabolised by the host cell via a pathway involving FAO. In some instances, the products of the substrate may not be of interest (for example, where the host cell is added to an alkane-containing pollutant, such as an oil spill or slick, for the purpose of breaking down the alkane). In other circumstances, the metabolic products of the substrate may be of commercial value. In such a case, the method will conveniently comprise the additional step of recovering the product (s), from the host cell and/or from the extra-cellular environment.

In another aspect, portions of at least 200 nucleotides (preferably 300-600 nucleotides or more) of sequences in accordance with the first aspect of the invention may be used to cause specific antisense inhibition of FAO expression in organisms which possess an endogenous FAO gene which exhibits homology with the sequence of the invention. The invention thus provides a method of causing antisense inhibition of an FAO gene.

Of more particular interest in the present invention is the use of a nucleic acid sequence comprising portions at opposed end regions of the FAO gene to delete FAO coding sequences from host cells. Methods, disclosed in the prior art in relation to other genes, are known whereby a nucleic acid sequence comprising opposed end regions of a gene (but from which the majority of the intervening portion has been removed), is introduced into a host cell, wherein the host cell comprises a functional gene whose end regions share a high level of nucleotide sequence identity with the introduced sequence. A homologous recombination event occurs, in which the introduced (inactive) gene is integrated on the host cell chromosome, and the wild type functional gene is excised from the chromosome and subsequently degraded. In this way, deletion mutants can be obtained in which a functional gene in a host cell is replaced with a partial fragment of the gene, which is wholly inactive. Such deletion mutants can possess extremely useful properties.

Thus, in the present instance, the invention provides a method of deleting a functional FAO coding sequence from the chromosome of a host cell, the method comprising: preparing a non-functional fragment of an FAO gene (typically comprising opposed end regions thereof but substantially free of FAO coding sequence); introducing the non- functional FAO gene fragment into a host cell which comprises a functional FAO gene, the opposed end regions of which exhibit a high level of nucleotide sequence identity with the introduced non-functional FAO gene fragment, so as to cause replacement of the functional FAO gene by the introduced non-functional FAO gene fragment. The method may be repeated a number of times if desired, as several industrially significant organisms have polyploid genomes or may, for other reasons, comprise a number of copies of the FAO gene, all of which should preferably be deleted in order to obtain maximum phenotypic change in the deletion mutant. The method may also comprise one or more screening steps (typically after each transformation) to select for further processing those cells which have desirable qualities (e. g. lowest levels of FAO activity, which could be assayed, for example, using the methods described in detail herein).

By way of explanation, for present purposes the term"FAO coding sequence"refers to that portion of the FAO gene which is translated into amino acid. The FAO gene comprises the FAO coding sequence together with regulatory sequences such as the promoter, 5'and 3'untranslated regions, and the like.

The invention also provides a nucleic acid construct for deleting an FAO gene from a cell, the construct comprising a non-functional FAO gene fragment (i. e. a portion of an FAO gene, which portion is insufficient to code for a polypeptide having FAO activity), which fragment comprises one or more portions possessing sequence identity with the FAO gene to be deleted.

The non-functional FAO gene fragment typically comprises at least part of the 5'and 3' untranslated portions of the FAO gene, although it may also include a small amount (for example, up to about 200 nucleotides) of the 5'and/or 3'end portions of the FAO coding sequence. The majority (typically 80 % or more) of the FAO coding sequence is advantageously omitted from the non-functional FAO gene fragment. Typically the omitted portion of the FAO gene in the non-functional FAO gene fragment is replaced by an approximately equivalent length of irrelevant DNA, such that the spacing between the 5'and 3'opposed end portions of the non-functional FAO gene fragment is substantially similar to the spacing between the corresponding 5'and 3'end portions of the functional FAO gene on the host cell chromosome to be deleted.

A high level of nucleotide identity is desirable between the opposed end regions of the functional FAO gene to be deleted, and the corresponding 5'and 3'opposed end regions of the introduced non-functional FAO gene fragment. Preferably this nucleotide sequence identity is 90% or more, over a length of at least 50 nucleotides at each opposed end region.

Those skilled in the art will appreciate that the method defined above can be employed to convert a host cell with a FAO+ phenotype to a FAO-phenotype.

The invention therefore also provides FAO-mutant cells and organisms, from which the FAO coding sequence has been specifically substantially deleted. Such cells may be referred to as FAO deletion mutants. The deletion mutant cell of the invention is a cell in which sufficient of the FAO gene has been specifically deleted so as substantially to abolish FAO activity (which typically will be less than 5 %, preferably less than 2 %, of the activity associated with an otherwise identical cell without the FAO-deletion). Such deletion mutants, with specific, known mutations are greatly preferred to FAO-mutants which might conceivably be obtainable by random mutagenesis, which process is unpredictable and can cause undesirable mutations in other genes. Generally, the deletion mutant of the invention is prepared using a deletion construct as defined above.

The deletion mutant will normally be an organism which, prior to deletion of the FAO coding sequence, is capable of utilising an alkane and/or a fatty acid as a substrate.

Preferred deletion mutant organisms are unicellular micro-organisms, such as yeasts, or other organisms amenable to genetic manipulation (e. g. fungi), especially organisms which are commonly used in industry, such as Candida sp. (especially C. tropicalis), Pichia sp. and Torulopsis sp. Such deletion mutants are expected to offer the possibility of synthesising certain organic compounds of commercial interest on a large scale and in a very cost-effective manner. For example, other yeast mutants (with other metabolic deficiencies) have already proved useful (see, for example, WO 91/14781, WO 94/07837 and EP 0,341,796), and it may be convenient (as described below) to make mutants in which other metabolic pathways (i. e. in addition to those using FAO) are also blocked.

In some embodiments, the FAO-deletion mutant may additionally be of a mutant phenotype with respect to one or more other defined characteristics, especially in relation to enzymes involved in metabolism of fatty alcohols, fatty acids, or dicarboxylic acids.

In particular, it may be advantageous for the FAO-deletion mutant to have one or more blocks in the a-oxidation pathway (i. e. one or more genes encoding an enzyme essential to the (3-oxidation pathway have been deleted or encode mutant, substantially inactive, forms of the enzyme). An FAO-mutant having a disrupted 0-oxidation pathway would be entirely unable to metabolise-hydroxyfatty acids, so that these compounds will accumulate if the organism is provided with a medium comprising fatty acids or compounds which can be metabolised to produce fatty acids. Mutant organisms having disrupted 0-oxidation pathways, and methods of obtaining such organisms, are already available to those skilled in the art (see, for example, EP 0341796).

Thus, for example, an existing (3-oxidation deletion mutant may be transformed with a fatty alcohol oxidase deletion or disruption construct in accordance with the invention to provide a ß-oxidation~, FAO-deletion mutant. Alternatively, the FAO genes may be deleted first, followed by deletion of one or more ß-oxidation genes. Yet another possibility would be to delete both FAO and ß-oxidation genes substantially simultaneously, either by using a single deletion construct comprising sequences to disrupt both sets of genes, or by using two or more deletion constructs. Methods of performing such work will be apparent to those skilled in the art with the benefit of the present specification. In general, deletion mutants accumulate large quantities of those compounds which would normally act as substrates for the missing enzyme. Often, these accumulated intermediate metabolites are produced in such large amounts that they also enter the extra-cellular environment, from where they are readily recovered. In the case of FAO deletion mutants, the substrates for FAO are typically fatty alcohols (HO-CH2-R-CH3), fatty diols (HO-CH2-R-CH2-OH), and-hydroxy fatty acids (HO-CH2-R-COOH), (where R is a saturated or unsaturated hydrocarbon chain, which may be substituted or unsubstituted, branched or [preferably] straight chain, normally comprising 6-22 carbon atoms), so these compounds could be expected to be produced in large amounts by FAO deletion mutants growing on substrates rich in alkanes and/or fatty acids, or substrates which give rise to such compounds upon metabolism by the deletion mutant.

Thus, in a further aspect the invention provides a method of producing fatty alcohols and/or a, w-fatty diols, and/or-hydroxy fatty acids, from a substrate comprising a hydrocarbon (particularly alkenes or alkanes) and/or a fatty acid and/or a fatty alcohol, the method comprising contacting the substrate with a plurality of FAO deletion mutant cells under conditions suitable for metabolism of the substrate by the deletion mutant cells, and recovering from the resulting mixture fatty alcohols, a, w-fatty diols and/or-hydroxy fatty acids. The preferred products are a,-fatty diols and all-hydroxy, saturated or unsaturated, substituted or unsubstituted C8-C22 fatty acids, which compounds are commercially significant (fatty alcohols can generally be prepared by other means).

Where the substrate is a hydrocarbon (especially an alkane), the product may be a fatty alcohol and/or an X-fatty diol; where the substrate is a fatty acid, the product may be an-hydroxy fatty acid; and where the substrate is a fatty alcohol, the product may be an fatty diol.

The conditions generally suitable for metabolism of the substrate are known to those skilled in the art (e. g. typically 20-35°C, preferably with forced aeration etc). The precise conditions will depend on the nature of the substrate and the desired product, the identity of the FAO deletion mutant etc.

Suitable fatty acid substrates particularly include C8-C22 saturated or unsaturated fatty acids, such as erucic, oleic, linoleic, behenic, arachidic, stearic, palmitic, myristic and lauric acids and the like. Suitable alkane substrates particularly include C8-C22 saturated or unsaturated hydrocarbons. Suitable fatty alcohol substrates comprise C8-C22 saturated or unsaturated fatty alcohols. Those skilled in the art will appreciate that the starting substrate upon which a microorganism is grown might not contain a hydrocarbon or fatty acid or fatty alcohol, but may instead comprise a substance which may first be metabolised by the deletion mutant to produce a hydrocarbon, fatty acid or fatty alcohol. For example, a yeast FAO deletion mutant could be provided with an ester, which is readily converted into fatty acids by the deletion mutant. Convenient sources of esters are cheap vegetable oils, such as sunflower oil, rapeseed oil, soya bean oil, palm oils, and the like. Other possible substrates include hydroxy-or epoxy-substituted fatty acids.

The FAO deletion mutant cells may be grown in suspension in liquid batch culture in a fermenter. Alternatively, the cells may be cultured immobilised as a solid matrix, with substrate passed substantially continuously across the matrix. Conveniently, where batch culture techniques are used, the cells are cultured until the majority of the alkane/fatty acid containing or producing substrate has been exhausted.

It should be noted that the inventors have found that C. tropicalis organisms in which all copies of the FAO gene have been deleted/disrupted will grow only poorly (if at all) on media which contain, as the sole carbon source, alkanes or fatty acids. Accordingly, in order to produce fatty alcohols/hydroxy fatty acids etc. in accordance with the method of the invention, it may be desired first to grow the FAO deletion mutant in a medium which has a carbon source readily metabolized by the organism, in order to produce a large biomass of cells of the mutant organism. The organisms are then contacted with the relevant substrate which, although it cannot support any great growth of the cells, may be converted to the desired product by a bioconversion process. A similar approach is taught and described by Picataggio et al, (1992 Bio/Technology 10,894-898). <BR> <BR> <BR> <BR> <BR> <BR> <BR> <P> Preferably the fatty alcohols, a. M-fatty diols and/or-hydroxy fatty acid products of interest are secreted into the extracellular medium, from which they can readily be separated from the cells (e. g. by filtration, centrifugation etc). Alternatively, the cells can be lysed or treated in other ways so as to release their intracellular contents, which may include the products of interest.

The products of interest may be extracted from broth culture using one or more known techniques (e. g. solvent extraction, or acid precipitation). Useful information on these techniques is given, for example, by Hatton in Comprehensive Biotechnology Vol. 2 (1985, Eds. Cooney & Humphrey, published by Pergammon Press, Oxford, UK) and in US 4,339,536. In this latter publication is disclosed an acid precipitation method for the extraction of dioic acids: at the end of fermentation the pH of the broth is increased to 11- 12 to dissolve the dioic acid product, and the cells removed by filtration or centrifugation.

The dioic acid is then precipitated by acidification of the filtrate to pH4 or lower, and the product collected by filtration. In principle, a generally similar approach should also be useful for extraction of-hydroxy fatty acids.

If desired the products of interest may be subjected to one or more conventional purification techniques (e. g. distillation, crystallization, precipitation or chromatography).

All prior art publications mentioned in this specification are considered incorporated herein by reference. The invention will now be further described by way of illustrative example and with reference to the accompanying drawings, in which: Figure 1 is a graph of fatty alcohol oxidase activity against time, showing induction of expression of the enzyme in C. cloacae cells grown on oleic acid substrates; Figure 2 is a photograph of SDS-PAGE analysis of C. cloacae polypeptides produced in cells grown on oleic acid substrates; Figure 3A is an elution profile for elution of alcohol oxidase from a phenyl superose column used to purify the enzyme; Figure 3B is a photograph showing gel electrophoresis analysis of the purified enzyme; Figure 4 is a graph showing the pH profile of the purified enzyme; Figure 5 shows sample Northern blot results of fatty alcohol oxidase mRNA; Figure 6 shows the DNA and deduced amnio acid sequence of a fatty alcohol oxidase gene; Figure 7 is a comparison of the amino acid sequence of three fatty alcohol oxidases; Figure 8 shows the DNA and deduced amino acid sequence of an alcohol oxidase gene; and Figures 9A, B and 10A, B are schematic representations of the preparation of various nucleic acid constructs in accordance with the invention.

Example 1-Purification characterisation and amino acid sequencing of C. cloacae fatty alcohol oxidase Candida cloacae 3152 strain FERM P-736 (used throughout these experiments) originates from the Fermentation Research Institute, the Agency of Industrial Science and Technology, the Ministry of the Industrial Trade and Industry, Japan and is described in GB 1,300,455. It was maintained at 4°C on agar slopes containing 1.5g agar, 0.5g yeast extract, 0.5g peptone and lg sucrose per 100ml. Starter cultures, used to inoculate shake flasks and fermenters, were prepared by aseptically transferring a loop of agar slope culture into 50ml medium containing 0.25g yeast extract, 0.25g peptone and 0.5g sucrose and incubating 24hr in a baffled 250ml flask at 30°C and 90rpm shaking. All biochemical reagents, including the detergent CHAPS, were obtained from Sigma Chemical Co. Poole, Dorset, UK and were of the highest purity available. Reagents for electrophoresis were from BioRad (Hemel Hempstead, Herts, UK).

Initial experiments were performed to investigate the timecourse of enzyme induction, using small scale shake flask cultures: C. cloacae was grown overnight in 1% yeast extract, 1% peptone, 2.5% sucrose medium at 30°C. Baffled flasks containing minimal medium (250 or 500ml) with sucrose as the carbon source, were inoculated with 2.0 and 4.0ml respectively of the starter culture. The minimal medium used at shake flask scale contained (per litre) 25g sucrose, 7.6g NH4Cl, 1.5g Na2SO4,300ml ImM pH 7.0 potassium phosphate buffer, 20mg ZnSO4.7H20, 20mg MnS04. 4H20, 20mg FeS04. 7H20, 2g MgCl2.6H20,100pg biotin, 20mg nicotinic acid, 20mg pyridoxine, 8mg thiamine and 6mg pantothenate. The minimal medium cultures were grown for 48h at 30°C at 90 rpm on a shaker. They were then centrifuged at 2400g for 10min, resuspended in the same volume of minimal medium but with 1 % oleic acid replacing sucrose as the carbon source, and incubation was continued at 30°C and 90 rpm for a further 125h. Samples of 30ml were removed from the flasks immediately before addition of oleic acid, and at 6,24,50 and 125h after addition. They were centrifuged at 10,000g for 10 min and the cell pellet was frozen in liquid nitrogen and stored at-80°C. Cell-free extracts were made from the pellets by grinding each pellet in liquid nitrogen three times then rapidly resuspending the powder in 10 ml of 50mM potassium phosphate pH 7.5 containing 5% glycerol, 0.5% CHAPS and centrifuging it at 10,000g for 10 min for clarification. The alcohol oxidase activity in these extracts was then measured. <BR> <BR> <BR> <BR> <BR> <BR> <P>Fatty alcohol was assayed spectrophotometrically (Kemp et al., 1988 Appl. Microbiol.

Biotechnol. 29,370-374). The assay mixture contained 50 mM-Tris/HCl pH 8.5,0.7 mg/ml ABTS (2,2'-azino-bis- [3-ethylbenzthiazoline-6-sulphonic acid]), 7U of horseradish peroxidase and 50ttM dodecanol previously dissolved in DMSO, in a final volume of 1.0ml unless otherwise specified. Reactions were initiated by addition of sample and the <BR> <BR> <BR> <BR> increase in absorbance at 405nm measured. The value of e for the radical cation of ABTS is 18.4mM-' cmi', and one mole of substrate gives rise to 2 mole of radical cation (Werner et al., 1970 Z. Anal. Chem. 252,224-228).

The induction of alcohol oxidase activity is clearly seen in Figure 1, which is a graph of alcohol oxidase activity (in units per gram cells wet weight) against time (hours) after addition of oleic acid; it rose 4-fold in six hours and reached a maximum of 7-fold at 24h.

In addition, cell extracts from 0,6 and 125h post-induction on oleic acid were run on SDS-PAGE and stained for protein with Coomassie Blue (Figure 2). SDS-PAGE gels consisted of a 5% stacking gel with a 10% running gel and were run on a mini BioRad Protean gel kit. The buffers used were as described by Laemmli (1970 Nature 227,680- 685). Figure 2 shows the results of SDS/PAGE analysis of protein profiles of C. cloacae cells prior to and after induction on oleic acid. Samples are 0,6,125 hours post induction. Arrows on the right show proteins which are elevated upon induction. The arrow on the left marked (a) indicates a protein of Mr 64 kDa which is reduced in quantity upon induction.

As can be seen in the figure, growth on oleic acid caused the induction of several proteins of approximately Mr 102,000,73,000; 61,000,54,000 and 46,000 and the disappearance of another protein. As described below, it was found that the approximately 73,000 Mr protein is alcohol oxidase. Similar results were obtained using hexadecane as inducing agent (data not shown).

Native gel electrophoresis was performed to establish which of the bands identified by SDS-PAGE was associated with alcohol oxidase activity. This was carried out at 200V for 2.5h in a 10% resolving/5% stacking gel as described above for SDS-PAGE except that the SDS was replaced by 1 % and 0.5 % sodium cholate in the gel and running buffer respectively. The sample buffer consisted of lOmM-Tris/HCl pH 6.8,2% glycerol, 1 % sodium cholate and sufficient bromophenol blue to make it visible. The running buffer and apparatus were chilled on ice before and during the run. The sample was 501il of 0.2U of hydroxyapatite-purified C. cloacae material prepared by the method of Dickinson & Wadworth (1992 Biochem. J. 282,325-331). To stain the gel for enzyme activity, buffer, ABTS and peroxidase at lOx the concentration used in the standard alcohol oxidase assay (described above) and dodecanol at 2.7 mM were applied to the surface of the gel, and it was incubated at room temperature for 5 mins. Alcohol oxidase activity was revealed as a region of green stain on the surface of the gel. The region containing the biological activity was cut out from the native gel and then subjected to SDS-PAGE.

Having identified the apparent molecular weight of the alcohol oxidase protein by SDS- PAGE, it was decided to grow up large scale yeast cultures to attempt to purify the enzyme.

Bulk preparation of induced cells for enzyme preparation was performed in 6L fermenters (Biolafitte) using a minimal medium containing (per litre) lOg hexadecane (in place of sucrose as carbon source, and in preference to oleic acid-see below), 4.5g (NH4) 2HPO4, 1.5g yeast extract, 0. lg CaCl2,1.5g Na2SO4. The remaining vitamins and minerals were as detailed above. The minimal medium (5L) was inoculated with 2% (v/v) of starter culture and then grown at 30°C, 0.45L/min aeration and 700 rpm stirring with pH control at 6.4 by auto-addition of 40% NaOH. When the hexadecane substrate was fully utilised (15-20hrs), the cells were harvested (3500g/lOmin), washed with 3L of 50mM HEPES/NaOH pH 7.5 and resuspended in 140ml of 50mM HEPES/NaOH pH 7.5,1mM EDTA, 1mM DTT prior to cell disruption using a French pressure cell. The wet weight of cells was typically 110-120g.

To obtain a preparation of microsomes (within which the enzyme was presumed to be concentrated), 132g (wet weight) of cells were passed through a French pressure cell 3 times at 20,000 psi, the disrupted cell extract centrifuged at 20,000g for 30min and the precipitate discarded. At this and other stages the preparation was snap-frozen in liquid <BR> <BR> <BR> <BR> N2 and stored at-80°C prior to further processing, for the sake of convenience. The supernatant was thawed and 100mM-Tris/HCl pH 7.5 added to give a final volume of 230ml. The microsomal fraction was pelleted by ultracentrifugation at 140,000g for 1.5h at 4°C.

The pellet was washed by suspending in 100mM-HEPES/NaOH, pH 8.0, containing 0.15M-KC1 (final volume 115ml) and re-pelleted by further ultracentrifugation at 140,000g for 1.5h at 4°C. The washed microsomes were resuspended in 50mM-HEPES/NaOH, pH 8.0 (final vol. 62ml). The resuspended pellet was made up to 500ml with 50mM- HEPES/NaOH, pH 8.0. Sodium cholate was added to 1.0% and PMSF in isopropanol to lmM. (NH4) 2SO4 was added to 35% w/v, the solution stirred for 20 min and centrifuged at 20,000g for 5min. (NH4) was added to 55 % w/v to the supernatant, the solution stirred for 20 min and centrifuged at 20,000g for 5min. The pellet was resuspended in 50mM-HEPES/NaOH, pH 8.0, containing 1 % Na cholate to a final volume of 52 ml, dialysed against two changes of 2 litres 50mM-HEPES/NaOH pH 8.0 for 2 x lh and a further 500ml of 50mM-HEPES/NaOH, pH 8.0 containing 1% CHAPS for lh and then centrifuged at 20,000g for 5min to give a clear supernatant.

The inventors initially adopted a similar procedure to that used for the purification of alcohol oxidase from C. tropicalis (Dickinson & Wadworth 1992 Biochem. J. 282,325- 331). It was found essential to prepare the microsomes from alkane induced fermentation as the oleic acid induced cells accumulated fatty acids which interfere with the membrane pelleting during the ultracentrifugation step. Use of this prior art procedure, which involved preparation of microsomes, detergent solubilization, (NH4) precipitation, QAE-cellulose chromatography and hydroxyapatite absorption, did not produce a homogeneous preparation as judged by SDS-PAGE when applied to C. cloacae : it was found that an additional gel filtration step on a Superose 12 column was required to obtain homogeneity.

With the C. cloacae preparation, as described previously with C. tropicalis, proteolysis and degradation was a major problem when using the previous purification procedure, particularly at later stages of purification. Accordingly, the inventors sought to develop a faster technique based on the known hydrophobic character of the protein. By directly applying redissolved (NH4) 2SO4-precipitated material to a Phenyl Superose column and eluting it with a reverse gradient of (NH4) 2SO4, it was possible to obtain a homogeneous preparation in fewer steps.

A Phenyl Superose HR 5/5 column (Pharmacia) was equilibrated in lOmM-Tris/HCI, pH 8.5, containing 1.7M (NH4) and 0.5% CHAPS. A flow rate of 0.6 ml/min was used throughout. For each chromatographic run, 5.0ml of (NH4) 2SO4-precipated resuspended dialysed material was diluted to 15ml with the column equilibration buffer, filtered through a 0.2A4m filter and loaded onto the column. After washing the column with 15ml of buffer a linear reverse gradient was run from 1.7M to 0.85M (NH4) 2SO4 over l0ml followed by a linear gradient from 0.85M to 0.5M (NH4) 2SO4 in 15ml.

The fractions with highest activity eluted at approximately 0.7M (NH4) as seen in Figure 3a, which shows (solid, fluctuating curve) the elution profile of alcohol oxidase from the phenyl superose column when eluted with a decreasing gradient (dashed line) of ammonium sulphate, as described above. In Figure 3a the vertical lines indicate fractions 1-6 of the eluate, which were retained for further study. Biological activity (dotted line, alcohol oxidase units/ml) was only detected in fractions 1-6, these were further analyzed by SDS/PAGE. When active fractions were run on SDS-PAGE, the density of a band at approximately 73 kDa and the activity of alcohol oxidase in each fraction were correlated, leading the inventors to conclude that the alcohol oxidase is a protein of Mr 73,000.

Figure 3b shows the results of this SDS-PAGE analysis of fractions 3-6 from the phenyl superose column. Lanes 1,2,3 and 4 are fractions 3,4,5 and 6 respectively.

Molecular weight markers are indicated on the left hand side. The arrow on the right hand side corresponds to a Mr of 73 kDa.

Table 1 shows the overall purification procedure, giving an enzyme preparation purified 230-fold with a 10.7 % recovery of biological activity. The new procedure was faster than that disclosed in the prior art and gave rise to essentially homogeneous enzyme. Further evidence that the 73 kDa protein was indeed alcohol oxidase was obtained from native gels. The enzyme activity was visualized directly in the gel as a green band when the gel was incubated with enzyme assay reagents. This band was excised and subject to SDS/PAGE, and a major band at 73 kDa was seen, confirming assignment of this band as alcohol oxidase. Volume Units mg Specific Fold Recovery Protein activity Purification (%) Cellextract 230 2012 5750 0. 35 1. 0 100 Membranes 115 1510 1783 0 85 2 4 75 Washed 62 X30 1178 1. 21 3 5 71 membranes Thawedwashed 500 824 1175 0. 7 2 0 41 membranes 55 % (NH4) 2SO4 52 754 156 48 13.7 37 pellet 243 Dialyzed 69 centrifuged (NH4)2504 Pellet PhenylSuperose 23 214 2. 7 80 230 10. 7 pooled fractions Table 1 Purification Table for Fatty Alcohol Oxidase The purified enzyme was assayed for biological activity at several pHs using dodecanol as the substrate. Assays were performed at least in duplicate and usually the variability was less than 5%. The enzyme preparation was homogenous, as judged by SDS-PAGE, and was diluted one in ten in 0. lmg/ml BSA, 0.5% CHAPS, 20% glycerol and 100mM- Tris pH8.5. Each assay contained 10y1 of enzyme. The low pH buffer was Bis/Tris/Propane 50mM and the high pH buffer glycine 50mM. The substrate was 10tel of 5mM dodecanol in DMSO (final concentration 50yM). The results are illustrated in Figure 4, which is a graph of enzyme units against pH. The experiment involved the use of two different batches of enzyme preparation (shown by circle or square symbols respectively). The enzyme exhibits a broad pH optimum between 8.0 and 9.5 but shows much lower activity below pH 7.5. This is similar to the properties of the enzyme from Candida tropicalis, which has maximal activity at pH 9.0 and has no activity at pH 5.5.

The purified enzyme was also the subject of investigations into reaction kinetics. In order to determine the Km for dodecanol three separate experiments were performed using 10 or 20y1 of 1 in 10 diluted enzyme as described above for the pH profile experiment, using 50mM Tris/HCl pH7.5 as buffer and dodecanol stock at 5 and 0.5mM in DMSO. This was to prevent any misinterpretation due to dilution effects. It was found that the apparent Krr, for dodecanol is between 4.0 and 5. OAM, and the V", Determination of K,, and Vx for decanol did not give reliable results: the Km determined from three consecutive experiments on the same day increased from 35pM in the first determination to 65yM in the third. The reason for this variation is unknown.

Finally, the protein preparation was used for the generation and sequencing of alcohol oxidase peptides. This was performed using the Promega procedure with Chromaphor green to detect the protein in the first gel and Endoproteinase Glu C for"in gel"digestion (Elborough et al., 1994 Plant Mol. Biol. 24,21-34).

Material from the post-ammonium sulphate precipitation stage of purification was run on SDS-PAGE, blotted to PVDF membrane and the major band at 73 kDa excised and applied to the amino acid sequencer. Based on specific activity of the enzyme, approximately 600pmole of alcohol oxidase was present in the gel. No amino acid sequence data could be obtained, implying N-terminal blocking. Treatment of the membrane transferred protein with CNBr, to cleave at methionine residues, gave amino acid sequence data at the 50 to 100pmol level, indicating that protein was present on the sequencing disk at approximately the expected levels given the typical recoveries for these processes. Further attempts at N-terminal sequencing protein purified by phenyl superose chromatography were also unsuccessful, confirming that it had a blocked N-terminus.

To obtain amino acid sequence data the partially pure protein (1300pmol based on a specific activity of 80 U/mg of pure material) from the (NH4) 2SO4 dialyzed pellet was separated using SDS-PAGE, stained with dye Chromophor Green, and the major band running at 73 kDa excised from the gel. Only one band in the region of 73 kDa was evident at this stage. The specific activity of this protein was 8.5 U/mg, approximately one tenth of the purified protein. However, most of the impurities were of low molecular weight, and there was only one dominant band on the gel which was at approximately 73 kDa. The excised band was loaded onto a second gel where in-gel digestion with glu-C and separation of the resulting peptides was carried out. Undigested excised 73 kDa band run at the same time as this gel gave a single high molecular weight band, verifying its purity. The resultant peptides were transferred to Problott for sequencing. Three bands were separately applied to the sequencer, but only one gave any sequence data.

An unambiguous sequence of 30 amino acids was obtained with the initial yield of approximately 100pmol (Table 2). This was the first amino acid sequence data for any fatty alcohol oxidase involved in X-oxidation. Searching the international protein data bases and the newly released yeast data base from the yeast total genome sequencing project failed to identify any significant homologies. No indication of any heterogeneity was seen, indicating that, at least in this region of the protein, there are no isoforms.

Taking the expected recoveries from these procedures into account, the sequencing yield was consistent with the amount of protein applied. The availability of amino acid sequence data was an important step towards cloning the cDNA and determination of the full amino acid sequence of the protein in the future.

1 5 10 15 ser gly gly thr ile pro ser thr asn gln gln leu phe met ile 46 94 90 84 91 76 30 60 68 58 78 53 47 41 41 16 20 25 30 ala gly ser thr phe gly gly gly ser thr val asn trp asp ala 43 29 6 19 25 17 25 30 4 8 10 10 3 1 6 Table 2 (Seq. ID No. 8) Amino acid sequence data of fatty alcohol oxidase peptide fragment. The pmole yield is given under each residue.

Example 2-Molecular cloning and characterization of two fatty alcohol oxidases from Candida cloacae Example 1 describes the purification of the fatty alcohol oxidase (FAO) protein and determination of its internal amino acid sequence. Based on that work, the inventors attempted to clone and characterize FAO coding sequences from C. cloacae.

As a first step in the cloning of fatty alcohol oxidase (FAO) coding sequences, genomic and cDNA libraries were prepared. To make a genomic library, C. cloacae DNA was isolated from cells grown on YPD (10 g/1 yeast extract, 20 g/1 peptone, 20g/l glucose) according to Philippsen et al. (1991 Methods in Enzymology 194,169-182). Partially Sau 3A-digested DNA was size fractionated to 14-23 kb via a 10-40% sucrose gradient and ligated into XBluestar BamHI arms (Novagen). Gigapack II Gold packaged particles were transfected into E. coli ER1647 (Novagen).

In order to make the cDNA library, approximately 1.6 x 109 cells of starter culture (0.5% yeast extract, 0.5% peptone, 1% sucrose) were used to inoculate a litre of minimal medium [1.5g/l Na2S04, 20mg/l ZnS04. H20, 20mg/l FeS04.7 H20,100/g/l biotin, 20mg/l pyridoxine, 6mg/l pantothenate, 7.6g/1 NH4Cl, 40.4g/1 KH2PO4,20mg/1 MnSO4 4 H20,2g/1 MgCl2 H2O, 30mg/1 nicotinic acid, 8mg/1 thiamine, 2.5 % sucrose as a carbon source and 4 ml/I of antifoam A emulsion (Sigma)] and grown at 30°C 90 rpm for 45 h after which oleic acid was added to 1 %, to induce expression of FAO. After 24h growth on oleic acid, cells were harvested, washed in 50 mM potassium phosphate buffer pH 7.5, frozen with liquid nitrogen and stored at-80°C.

Poly (A) + mRNA from the frozen cell samples was used for the construction of a random-primed non-directional cDNA library in EcoRI-digested alkaline phosphatase treated XzapII vector (Stratagene). Approximately 5.0, ug of mRNA was used in the reverse transcriptase reaction according to instructions of the TimeSaver cDNA synthesis kit (Pharmacia). EcoRIlNotI adapters were added to the ends and cDNA ligated with EcoRI digested XzapII. Gigapack II Gold packaging extract (Stratagene) was used to form phage particles followed by transfection of E. coli XLl-Blue cells (Stratagene).

There were no known fatty alcohol oxidase (FAO) sequences in the databases that could be used in heterologous screening. Accordingly the inventors had to rely on the amino acid sequence data, obtained as described in example 1. To allow the screening for FAO cDNAs the inventors produced a specific probe by using a cDNA library as a target DNA and a degenerate FAO primer based on amino acid residues 8-16 of Seq. ID No. 8 described above, in combination with cDNA library vector based primers M13 forward, and T3 (Seq. ID No. 9,5'AATTAACCCTCACTAAAGGG 3'), in PCR (see below).

11 (1.5 x 103 pfu) of the 24hr oleic acid induced cDNA library was amplified by using 10 pmol of M13 forward primer (5'TTG TAA AAC GAC GGC CAG T 3'Seq. ID No.

10) or M13 reverse primer (5'CAC ACA GGA AAC AGC TAT GAC C 3'Seq. ID No.

11), and 20 pmol of internal degenerate FAO primer (5'ACN AAY CAR CAR CTN TTY ATG ATH GC 3'Seq. ID No. 12, where N=ACGT, Y=CT, R=AG, H=ACT, which corresponds to the amino acid sequence of residues 8-16 of Seq. ID No. 8) with 2U Taq polymerase and PCR buffer containing 1.5 mM MgCl2 and 1.25 mM dNTPs (Boehringer Mannheim) in the following conditions for 35 cycles: denaturation at 94°C 30 seconds, annealing at 60°C 30 seconds, and extension at 72°C 2 minutes. 5y1 of the PCR products were subjected to gel electrophoresis and the major bands cut from the gel and eluted into MQ water (Quiagen gel isolation kit). The fragments were cloned directly into pGEM-T vectors (Promega) using standard techniques and the length of all clones was analyzed by PCR using internal FAO and T7 (5'GTA ATA CGA CTC ACT ATA GGG CG 3'Seq.

ID No. 13) or SP6 (5'GCT ATT TAG GTG ACA CTA TAG 3'Seq. ID No. 14) primers essentially as described previously, except that annealing was at 58°C for 30 seconds.

This PCR screen yielded products of 1100 bp, 800 bp, 600 bp, 550 bp and 300 bp which, as described above, were directly subcloned into pGEM-T vector. Sequencing of these clones, using an Applied Biosystems Inc. 373 DNA sequencer, showed that they all carried the same FAO derived partial amino acid sequence (data not shown). The recombinant plasmid with the largest insert, pAX17 (1. lkb), was fully sequenced in both directions. The sequence data were entirely consistent with those derived by amino acid sequencing of the alcohol oxidase enzyme from Candida. The insert pAX17 was used to screen the library to obtain full length clones, as described below.

Plasmid AX 17 sequence was digested with PvuII and HindIII to remove polylinker sites and the resulting 1 kb fragment was used as a probe to screen the 24 h oleic acid induced C. cloacae cDNA library. Three independent clones containing inserts of 2. 4 kb (pFA04), 1.0 kb (pFA02), and 0.6 kb (pFA06) respectively, hybridized strongly with the probe.

Partial sequencing of these clones showed that they represented two different classes: class 1 (0.6 kb), with translated amino acid sequence identical to the purified FAO, and class 2 (2.4 kb and 1.0 kb), which contained the variant amino acid sequence NGGALSSTNQQIFIIAGSTFGGGSTVNW (Seq. ID No. 15). Sequencing of the clones also showed restriction enzyme polymorphism between the two classes: Class 1 carried an XbaI site that is not present in the class 2 clones, and class 2 possessed a HindIII site not present in class 1. This polymorphism was ideal for use in the rapid screening of genomic clones.

In order to determine the size of FAO corresponding mRNA and to study if it was induced after growth on oleic acid, Northern blotting experiments were performed. C. cloacae cells were grown for 24 hours in minimal medium containing either 2 % glucose or 1 % oleic acid as a carbon source. The cells were harvested, ground in mortar and pestle under liquid N2 and RNA was extracted using the hot SDS method (Hall et al., 1978 Proc.

Natl. Acad. Sci. USA 75,3196-3200). mRNA was prepared by using oligo (dT) spun columns (Pharmacia mRNA Purification kit). Northern analysis was conducted by running 2 gg of mRNAs into a gel containing 0.66M formaldehyde, blotting the RNA onto a Hybond-Nw membrane (Amersham, UK) and performing a hybridization reaction according to the standard published method (Sambrook et al. 1989 Molecular Cloning: A Laboratory Manual 2nd edition, Cold Spring Harbor Laboratory, NY) with Rediprime random primer labelled (Amersham, UK) probe, or (in subsequent experiments) with the same 1 kb PvuII-HindIII fragment from pAX17 that was used in the original cDNA library screen. Typical results are shown in Figure 5.

Figure 5 shows the induction of FAO mRNA upon 24 h growth on oleic acid probed with PvuII-HindIII fragment from pAX17. Lane 1 is mRNA from cells grown on 2% glucose, lane 2 is mRNA from cells grown on 1 % oleic acid. Molecular size markers are indicated at the side. The signal of 2.4 kb, the size of which corresponds to a protein of approximately 73 kDa known to represent FAO protein was induced 5-7 times on oleic acid compared to glucose grown C. cloacae. A low level of constitutive expression on glucose was also observed, although this is not apparent from the Figure.

Using the insert of pAX17 a genomic bacteriophage library was screened and several hybridising bacteriophage were isolated after three rounds of screening. In order to check if they were FAO sequences each was subjected to PCR analysis using two FAO specific primers (FAO AX1: 5'CAG GTT CGA CTT TTG GTG 3' ; and FAO AX5: 5'GTA CCC AAG CTT GTG GTG 3'Seq. ID Nos. 16 and 17 respectively). All were positively identified using this method and further analysed for classification of gene (type 1 or 2).

When digested with XbaI the PCR product from class 1 genes was digested into two indicative fragments. Similar analysis using HindIII, specific for class 2 clones showed that only class 1 clones were represented in the positive phage isolated. One representative phage was subjected to plasmid rescue using standard techniques. The resulting plasmid, termed pgFAO 14, contained an 18kb fragment insert which was further characterised. Sequencing of 4.3 kb of the genomic clone showed that the whole class 1 type gene was represented.

Computer analysis of the genomic and cDNA sequences was carried out by using DNA Strider (Marck 1988 Nucl. Acids Res. 16,1829-1836) and GCG package from the BBSRC facility at Daresbury, UK. Both the chromosomal clone in pgFA014, and the cDNA clone in pFA04, were sequenced and shown to contain open reading frames of 2094 nt and 2091 nt, encoding polypeptides of 698 and 697 amino acids respectively.

Corresponding genes were named as FAO1 and FA02. The 4.3 kb FAO1 sequence from pgFA014 is shown in Figure 6, together with the FAO1 flanking regions (nucleic acid sequence is Seq. ID No. 18; deduced amino acid sequence is Seq. ID No. 19). Various promoter and terminator elements (discussed below) are shown underlined in the figure.

The FAO1 gene did not carry any apparent intron sequences in the coding region. The deduced amino acid sequence of FAO2 (see Figure 7) contains a carboxy-terminal peroxisome targeting sequence SKL (Gould et al., 1989 J. Cell. Biol. 108,1657-1664).

In FAO1 the corresponding carboxy-terminal sequence is TKL, which fits to the more general consensus (uncharged (neutral)-basic-hydrophobic residue).

Both FAO1 and FA02 contain the consensus sequence of Cys-X-X-Cys-His for the cytochrome c family heme-binding site (Mathews 1985 Prog. Biophys. Mol. Biol. 45,1- 56). Identity between FAO1 and FAO2 was 80.0% at nucleotide level, and deduced amino acid similarity was 89.4%. The aligment of deduced FA01 and FA02 amino acid sequences is shown in Figure 7 (which also includes the amino acid sequence of another alcohol oxidase, FAOT, cloned from C. tropicalis described in detail in Example 3 below). An exhaustive search of EMBL/Genbank databases did not reveal any related sequences, confirming the novelty of the two FAO genes. The deduced amino acid sequence of FA02 and FAOT are shown as Seq. ID Nos. 20 and 21 respectively in the attached sequence listing.

There is a conserved purine at the position-3 (Hamilton et al., 1987 Nucl. Acids Res. 15, 3581-3593) from the putative ATG codon in both FAO1 and FA02 genes. In-frame stop codons are also present upstream of the initiating ATG in the FAO1 gene. The 1126 nt long promoter sequence found in FAO1 does not contain a typical TATA-box, though there is a TATATAC sequence at position-20 from ATG. Computer analysis of the promoter sequence shows two CCAAT sequences (underlined) at positions 183-187 and 924-928, a cAMP-responsive element (CRE) (sequence T (G/T) ACGTCA), see Roesler et al., 1988 J. Biol. Chem. 263,9063-9066) at position-24 (underlined) from the ATG, and a recognition site (AGATAAA) for GATA-binding transcription factors.

FAO1 also contains a typical eukaryotic polyadenylation signal AATAAA (Proudfoot & Brownlee 1976 Nature 263,211-214) as well as a consensus sequence for transcription termination TAG... TA T/A GT... TTT in S. cerevisiae (Zaret & Sherman 1982 Cell 28, 563-573) (see Figure 6).

Codon usage of the two genes is not particularly biased as 89% and 90% of all codons available are used by FAO1 and FAOZ respectively, though preferred codons are slightly different. The most prominent difference relates to codons specifying: tyrosine, where FAO1 uses TAC (72.0%), whilst FA02 favours TAT (52.2%); and aspartic acid, where FAO1 uses GAC (60.5%) whereas FA02 uses GAT (64.3%). Several Candida species use the universal leucine codon CUG as a serine codon (Ohama et al. 1993 Nucl. Acids Res. 21,4039-4045). From the present results, it is not possible to deduce if this non- universal codon usage occurs in C. cloacae, since the partial amino acid sequence of FAO determined from C. cloacae contains one leucine residue encoded by TTG and five serine residues encoded by AGT, TCC, TCG, TCT and TCT, respectively. For both FAO1 and FA02 the usage of the CUG codon is very rare (1.4% and 2.9% respectively of all leucine codons available).

Both FA01 and FA02 were isolatable from a cDNA library, excluding the possibility that one of them might be a pseudogene. Isolation of the FA01 and FA02 genes allows the execution of knock-out experiments to determine their exact functions and enables construction of new industrially significant strains in which the m oxidation pathway is blocked at the alcohol oxidase stage.

Example 3-Isolation of FAO coding sequences from C. tropicalis (FAOT) C. tropicalis NCYC 470, the strain used in this example, originates from the National Collection of Yeast Cultures, Brewing Industry Research Foundation, Great Britain.

Construction and screening of C. tropicalis cDNA library C. tropicalis cells were grown for 24 hours in minimal medium [1.5g/l Na2SO4,20mg/l ZnSO4. H20,20mg/l FeS04.7H20,100pg/1 biotin, 20mg/1 pyridoxine, 6mg/1 pantothenate, 7.6g/l NH4Cl, 40.4g/l KH2PO4,20mg/l MnS04.4H20, 2g/l MgC12. H20, 30mg/l nicotinic acid, 8mg/l thiamine] containing 1 % oleic acid as a carbon source, harvested and ground in mortar and pestle under liquid N2. Total RNA was extracted using the hot SDS method (Hall et al., 1978, cited previously) and mRNA was prepared by using oligo (dT) spun columns (Pharmacia mRNA Purification kit). Poly (A) + mRNA was then used for the construction of a random-primed non-directional cDNA library in EcoRI-digested alkaline phosphatase treated XzapII vector (Stratagene).

Approximately 5. zig of mRNA was used in the the reverse transcriptase reaction according to instructions of the TimeSaver cDNA synthesis kit (Pharmacia), in order to prepare a cDNA library. EcoRIlNotI adapters were added to the ends and cDNA was ligated with EcoRI digested XzapII. Gigapack II Gold packaging extract (Stratagene) was used to form phage particles followed by transfection of E. coli XL1-Blue cells (Stratagene).

Approximately 80,000 pfu of an unamplified cDNA library was transferred onto a Hybond «-N membrane (Amersham) and screened using a 2.3 kb EcoRI fragment from C. cloacae FA04 cDNA clone (obtained by screening a C. cloacae library with the insert from pAX17) as a probe under the following low stringency hybridization conditions: prehybridization 2h at 55°C in 6xSSC-lxDenhardt's, 0.5% SDS, 0.05% sodium pyrophosphate with 0.05mg/ml herring sperm DNA; hybridization overnight at 55*C in essentially the same mix except that instead of herring sperm DNA, 1 mM EDTA was used. Filters were washed twice with 2xSSC-0.1 % SDS at 55 ° C (30 minutes per wash).

Three clones giving positive signals were purified and in vivo excised as Bluescript SK (-) phagemids using ExAssist helper phage in E. coli SOLR cells according to Stratagene's instructions. Plasmid DNA was isolated using Wizard Minipreps (Promega) and sequenced with M13 reverse and M13 forward primers. Partial sequencing of one of these clones FAOT1, with M13 reverse primer showed 64.4% nucleotide sequence identity with the C. cloacae FA04 cDNA clone. Subsequently, complete sequencing of FAOT1 cDNA using sequence specific primers revealed an open reading frame of 617 amino acids.

Comparison with the 697 amino acids of the corresponding C. cloacae clone suggested that the 3'end of the C. tropicalis cDNA clone was missing.

Construction and screening of C. tropicalis genomic library C. tropicalis DNA was isolated from cells grown overnight in YPD according to Philippsen et al. (1991, cited above). Partially Sau3A-digested DNA, size fractionated through a 10-40% sucrose gradient to 10-12 kb, was ligated into ZapExpress BamHI arms (Stratagene). PhageMaker pakaging extract (Novagen) was used to form phage particles, which were then transfected into XL1-Blue MRF'cells (Stratagene).

Approximately 42,000 pfu (6 fold haploid genome) of an unamplified Xlibrary was transferred onto Hybond-N membrane (Amersham) and screened using a 968 bp Asp718 fragment from the FAOT1 cDNA clone as a probe in the following hybridization conditions: prehybridization for 2h at 42 C in 6xSSC, SxDenhardt's, 0.5% SDS, 100pg/ml herring sperm DNA, 50% formamide; hybridization overnight at 42'C in the same hybridization mix. Filters were washed twice (15 min per wash) in 2xSSC, 0.1% SDS, and twice in lxSSC, 0.1 % SDS (15 min per wash), all washes at 42 C.

12 clones giving positive signals were purified and in vivo excised as the pBK-CMV phagemids using ExAssist helper phage in E. coli XLOLR cells according to Stratagene's instructions, except that excision reaction was grown overnight at 37°C and 200, ul phage supernatant was used with 200 yl XLOLR cells which were incubated at 37°C for 3-4 hours before plating. Isolated plasmid DNA (Wizard'Minipreps, Promega) was sequenced with FAOT specific primer FT1 (5'CTT CCC CTC CAT GGT AAC 3', Seq.

ID No. 22) which confirmed that several clones contained the C. tropicalis FAO gene.

Clone gFAOT4 (4.6 kb) was sequenced using sequence specific primers. Sequence comparison showed that the 3'part of the coding sequence and the whole 3'untranslated region (UTR) were missing from the clone.

Cloning of the 3'-end of FAOT 300 pmol of primers T3 and T7, and 10 ftl of X19 phage stock which contained the C. tropicalis FAOT gene isolated as described above, were amplified using Expand'Long Template PCR System (Boehringer Mannheim) with 2.6 U DNA polymerase mixture, 500 zM dNTPs in 1.25 mM MgCl2 and detergents containing PCR buffer (Boehringer Mannheim) in the following conditions: 1 cycle denaturation at 94°C 5 min; 10 cycles denaturation at 94°C 30 seconds, annealing at 56°C 30 seconds, elongation at 68°C 10 minutes; 20 cycles denaturation at 94°C 30 seconds, annealing at 56°C 30 seconds, elongation at 68°C 10 minutes, plus cycle elongation of 20 seconds for each cycle; and 1 cycle prolonged elongation at 68°C for 7 minutes.

The resulting product of approximately 4.5 kb was diluted 1: 5 in buffer and 1 lil amplified with 2.5 U BIOTAQw DNA Polymerase (Bioline), 250 zM dNTPs, and PCR buffer (Bioline) containing 1.5 mM MgCl2, using 30 pmol of FT5 (5'GAT GGT AAA GGA CAT GGC 3'Seq. ID No. 23) and T7 primers in following conditions: 1 cycle denaturation at 94°C 5 minutes; 30 cycles denaturation at 94°C 1 minute, annealing at 56°C 1 minute, elongation at 72°C 2 minutes; 1 cycle prolonged elongation at 72°C for 5 minutes. The resulting 1 kb product was reamplified with gFT7 (5'CCA AGG GAT GAA CGA TCC 3'Seq. ID No. 24) and T7 primers in PCR under essentially the same conditions.

This PCR product of approximately 750 bp was purified using Wizard PCR Preps DNA Purification System for Rapid Purification of DNA Fragments (Promega), ligated into pT7Blue (R) vector (Novagen) and transformed into Epicurian coli XL2-Blue MRF' ultracompetent cells (Stratagene). 733 bp of the resulting plasmid pFAOT3 was sequenced using sequence specific primers. The total FAOT sequence of 4233 bp (Seq. ID No. 25), which contains an open reading frame of 2112 nt (704 amino acid residues) is shown in Figure 8.

In Figure 8, the sequences present in primers F1 or F1', F2, F3, F4 or F4', F5 or F5', F6, F7, F8 or F8' (described below) are underlined (note that primers F2, F4, F6, F8 are from the opposite DNA strand to that shown in the figure). The deduced amino acid sequence is also shown beneath the DNA sequence.

Example 4-Preparation of FAOT Gene Disruption Constructs To disrupt the FAOT gene in the C. tropicalis genome, a plasmid for gene disruption using the FAOT 5'and 3'untranslated regions (UTR) was constructed in two steps: (1) amplication of the 5'-UTR sequence of FAOT and ligation with C. tropicalis transformation vector pDS148 (a generous gift from Dr Dominique Sanglard, CHW, Lausanne, Switzerland), yielding plasmid plSV; and (2) amplificiation of the 3'-UTR sequence of FAOT and ligation with plSV, yielding plasmid pSVUl which was used as template for PCR. Further subcloning resulted in the final FAOT disruption plasmid pSVU2. The process is illustrated schematically in Figures 9A and 9B. By way of explanation, plasmid pDS148 mentioned above was constructed previously by replacing the URA3 HindIII fragment of S. cerevisiae present in pNKY51 (see Alani et al, 1897 Genetics 116,541-545) with an equivalent URA3 fragment from C. tropicalis.

Another construct for gene disruption was prepared, again in a two-step process, using FAOT 5'and 3'coding sequences (CDS): (1) amplification of the 5'CDS sequence of FAOT and ligation with C. tropicalis vector pDS148, yielding plasmid p2SV; and (2) amplification of the 3'CDS seuqence of FAOT and ligation with p2SV, yielding plasmid pSVCl which was also used as template for PCR. Further subcloning resulted the final FAOT disruption plasmid pSVC2. Preparation of pSVC2 is illustrated schematically in Figures 10A and 10B.

4.1 Construction of pSVU2 (1) Amplification of FAOT 5'-UTR sequence and construction of plasmid plSV (Fig.

9A).

A 400 bp 5'-UTR FAOT BGIII fragment was amplified from 10 ng of gFAOT4, described above, using 50 pmol of primers F1 (5'GAA AAG ATC TGT TAT TAG AAG AGT TAC 3'Seq. ID No. 26) and F2 (5'AAA CAA TTA GAT CTC CGA AAC ACA GGC 3'Seq. ID No. 27) (see Fig. 8) with 2U Vent DNA polymerase (New England Biolabs), and 250/in dNTPs in PCR buffer in the following conditions: 1 cycle denaturation at 94°C 5 minutes; 35 cycles denaturation at 94°C 1 minute, annealing at 60°C 1 minute, elongation at 72°C 2 minutes; 1 cycle prolonged elongation at 72°C for 5 minutes.

The resulting fragment was purified using the Wizard PCR Preps DNA Purification System for Rapid Purification of DNA Fragments (Promega). Subsequently the PCR product was ligated into pT7Blue (R) vector and transformed in E. coli XL2-Blue MRF ultracompetent cells to ascertain that BglII restriction enzyme digestion released the required fragment. After BgIII digestion, the fragment was ligated to BgllI-digested, Shrimp Alkaline Phosphatase (Boehringer Mannheim)-treated pDS 148 and transformed into XL2-Blue MRF ultracompetent cells.

The orientation of the 400 bp BglII insert in the resulting plasmid, termed plSV, was checked by PCR of 1: 5 dilution of a transformed bacterial colony resuspended in TE with 2.5 U Biotaq DNA Polymerase (Bioline), 250/au dNTPs, PCR buffer containing 1.5 mM MgCl2 using 30 pmol of primers F1 and URA1 (5'CTG GTT GTT CTT CTG GTG 3'Seq. ID No. 28, from the C. tropicalis URA3 part of the plasmid) under the following conditions: 1 cycle denaturation at 94°C 5 minutes; 35 cycles denaturation at 94°C 1 minute, annealing at 54°C 1 minute, elongation at 72°C 2 minutes; 1 cycle prolonged elongation at 72°C for 5 minutes. Only the correct orientation results in the expected product of 1.4 kb. This was also confirmed by sequencing plasmid plSV using the F1 oligonucleotide as primer, which showed that the 5'-UTR FAOT sequence was followed by the hisG sequence from the plasmid pDS148.

(2) Amplification of FAOT 3'-UTR sequence and construction of plasmid pSVU2 (Fig.

9B).

A 225 bp 3-UTR FAOT BamHI fragment was amplified from 10 ng of gFAOT4, described above, using 50 pmol of primers F3 (5'CCG GAT CCA GCT TGT TGA TTG AAT CC 3'Seq. ID No. 29) and F4 (5'GAA GGA TCC ACA ATT GAT TGC ACA GC 3'Seq. ID No. 30) (see Fig. 8), with 2U Vent DNA polymerase (New England Biolabs), 250 yM dNTPs in PCT buffer in the same conditions as in the step (1) above.

The resulting fragment was purified using the Promega Wizard System. Subsequently the PCR product was ligated into pT7Blue (R) vector and transformed into E. coli XL2-Blue MRF ultracompetent cells to ascertain that BamHl restriction enzyme digestion released the required fragment. After BamHI digestion, the fragment was ligated to BamHI- digested, Shrimp Alkaline Phosphatase (Boehringer Mannheim)-treated plSV and transformed into XL2-Blue MRF ultracompetent cells. The orientation of the 225 bp BamHI insert in the resulting plasmid, termed pSVUl, was checked by PCR of 1: 5 dilution of a transformed bacterial colony resuspended in TE with 2.5 U Biotaq DNA Polymerase (Bioline), 2501tu dNTPs, PCR buffer containing 1.5 mM MgCl2 using 30 pmol of primers F4 and URA2 (5'GGT TGG AAC GCC TAC TTG 3'Seq. ID No. 31, from the C. tropicalis URA3 part of the plasmid) in the following conditions: 1 cycle denaturation at 94°C 5 minutes; 35 cycles denaturation at 94°C 1 minute, annealing at 54°C 1 minute, elongation at 72°C 2 minutes; 1 cycle prolonged elongation at 72°C for 5 minutes. Only the correct orientation results in the expected product of 1.7 kb. This was also confirmed by sequencing plasmid pSVUl using the F4 oligonucleotide as primer, which showed that the 3'-UTR FAOT sequence is followed by the hisG sequence from the plasmid pDS148.

An additional PCR and subcloning step were carried out in order to obtain an optimised cassette which, in essence, contains only the selection marker (URA3) flanked by the homologous recombination sequences (5'and 3'UTR). A 4.6 kb fragment, to which EcoRI restriction sites on both ends were added, was PCR amplified from 10 ng of pSVUl using 50 pmol of primers F1' (5'GAA AAG ATC TGA ATT CAG AAG AGT TAC AAA ACT CCG 3'Seq. ID No. 32) and F4' (5'GAA GGA TCC AGA ATT CAT TGC ACA GCA AAC AAA TAA 3'Seq. ID No. 33) (see Fig. 8). PCR was performed using the Boehringer Expand Long Template PCR System in the following conditions: 1 cycle denaturation at 94°C 5 minutes; 35 cycles denaturation at 94°C 1 minute, annealing at 64.5°C 1 minute, elongation at 72°C 5 minutes; 1 cycle prolonged elongation at 72°C for 10 minutes. The resulting fragment was purified using the Promega Wizard system.

After EcoRI digestion, the PCR product was ligated into EcoRI-digested and Shrimp Alkaline Phosphatase-treated pBuescript SK+ vector to give plasmid pSVU2. pSVU2 was transformed into E. coli XLl-Blue competent cells to ensure that EcoRl restriction enzyme digestion released the required fragment.

4.2 Construction of pSVC2 (1) Amplification of FAOT 5'-CDS sequence and construction of plasmid p2SV (Fig.

10A).

A 437 bp 5-UTR FAOT BglII fragment was amplified from 10 ng of gFAOT4, using 50 pmol of primers F5 (5'GAA AGA GAT CTA CGT GAA AAG GCA TA 3'Seq. ID No.

34) and F6 (5'AAC AAG ATC TTG CGC CAT GTC ATA TG 3'Seq. ID No. 35) (see Fig. 8), with 2U Vent DNA polymerase (New England Biolabs), 250 zM dNTPs in PCR buffer in the following conditions: 1 cycle denaturation at 94°C 5 minutes; 35 cycles denaturation at 94°C 1 minute, annealing at 60°C 1 minute, elongation at 72°C 2 minutes; 1 cycle prolonged elongation at 72°C for 5 minutes.

The resulting fragment was purified using the Promega Wizard System. Subsequently the PCR product was ligated into pT7Blue (R) vector and transformed into E. coli XL2-Blue MRF ultracompetent cells to ensure that BglII restriction enzyme digestion released the required fragment. After BgIII digestion, the fragment was ligated to BglII-digested, Shrimp Alkaline Phosphatase (Boehringer Mannheim)-treated pDS 148 and transformed into XL2-Blue MRF ultracompetent cells.

The orientation of the 437 bp BglII insert in the resulting plasmid p2SV was checked by PCR of 1.5 dilution of a transformed bacterial colony resuspended in TE with 2.5U Biotaq DNA Polymerase (Bioline), 250, uM dNTPs, PCR buffer containing 1.5 mM MgCl2 using 30 pmol of primers F5 and URA1 in the following conditions: 1 cycle denaturation at 94°C 1 minute, annealing at 54°C 1 minute, elongation at 72 ° C 2 minutes; 1 cycle prolonged elongation at 72°C for 5 minutes. Only the correct orientation results in the expected product of 1.4 kb. This was also confirmed by sequencing plasmid p2SV using the F5 oligonucleotide as primer, which showed that the 5'-CDS FAOT sequence is followed by the hisG sequence from the plasmid pDS148.

(2) Amplification of FAOT 3'-CDS sequence and construction of plasmid pSVC2 (Fig.

10B).

A 525 bp 3'-CDS FAOT BamHI fragment was amplified from 10 ng of gFAOT4, using 50 pmol of primers F7 (5'TCC ATG GAT CCA AGC TTC GTT GTT AC 3) and F8 (5 CAT TGG ATC CGC ACC ACT TGC AGT T 3') (Seq. ID Nos. 36 and 37 respectively), with 2U Vent DNA polymerase (New England Biolabs), 250M dNTPs in PCR buffer in the same conditions as in the step (1) above.

The resulting fragment was purified using the Promega Wizard System. Subsequently the PCR product was ligated into pT7Blue (R) vector and transformed into E. coli XL2-Blue MRF ultracompetent cells to ascertain that BamHI restriction enzyme digestion released the required fragment. After BamHI digestion, the fragment was ligated to BamHI- digested, Shrimp Alkaline Phosphatase (Boehringer Mannheim)-treated p2SV and transformed into XL2-Blue MRF ultracompetent cells. The orientation of the 525 bp BamHI insert in the resulting plasmid (termed pSVCl), was checked by PCR of 1: 5 dilution of a transformed bacterial colony resuspended in TE with 2.5 U Biotaq DNA Polymerase (Bioline), 250 zM dNTPs, PCR buffer containing 1.5 mM MgCl2 using 30 pmol of primers F8 and URA2 (5'GGT TGG AAC GCC TAC TTG 3', Seq. ID No. 38 from the C. tropicalis URA3 part of the plasmid) in the following conditions: 1 cycle denaturation at 94°C 5 minutes; 35 cycles denaturation at 94°C 1 minute, annealing at 54°C 1 minute, elongation at 72°C 2 minutes; 1 cycle prolonged elongation at 72°C for 5 minutes. Only the correct orientation results in the expected product of 1.7 kb. This was also confirmed by sequencing plasmid pSVC2, using the F8 oligonucleotide as primer, which showed that the 3'-CDS FAOT sequence is followed by the hisG sequence from the plasmid pDS 148.

An additional PCR and subcloning step were carried out in order to obtain an optimised cassette which, in essence, contains only the selection marker (URA3) flanked by the homologous recombination sequences (5'-and 3-CDS). A 4.6 kb fragment to which EcoRI restriction sites were added, was PCR amplified from 10 ng of pSVCl using 50 pmol of primers F5' (5'GAA AGA GAT CTA CGT GAA TTC GCA TAT GAA TCC CAG 3') and F8' (5'CAT TGG ATC CGC ACG AAT TCC AGT TGG CAA AAC ACT 3') (Seq. ID Nos. 39 and 40, respectively) PCR was performed using the Boehringer Expand Long Template PCR system in the following conditions: 1 cycle denaturation at 94°C 5 minutes; 35 cycles denaturation at 94°C 1 minute, annealing at 60°C 1 minute, elongation at 72°C 5 minutes; 1 cycle prolonged elongation at 72°C for 10 minutes. The resulting fragment was purified using the Promega Wizard System. Subsequently the PCR product was ligated into pGEM-T vector to give pSVC2 which was transformed into E. coli XLl-Blue competent cells to ensure that EcoRI restriction enzyme digestion will release the required fragment.

The entire URA3 coding region of both pSVU2 and pSVC2 was sequenced and confirmed to be correct at amino acid levels.

Example 5: Transformation of C. tropicalis with FAO disruption constructs The EcoRI fragment generated from pSVU2 has the 5'and 3'UTR available for homologous recombination with the FAO gene (as it is specific for it) and the URA3 gene inside the UTRs to use as a positive auxotrophic marker. The transformation cassette also has between the 5'UTR and 3'UTR two identical regions (hisG) which can recombine, so removing the selectable marker URA3 but having completely disrupted the FAO gene in the host cell.

The EcoRI fragment generated from pSVC2 has the 5'and 3'CDS available for homologous recombination with the FAO gene or with a second FAO-like gene (which normally would have nearly 80% homology as is the case in C. cloacae), where second or further alleles of the FAO gene exist (e. g. in diploid or polyploid organisms).

Linear DNA fragments were used to transform with a lithium acetate procedure (descibed by Sanglard et al, 1996 Antimicrobial Agents and Chemotherapy 40,2300-2305) the C. tropicalis strain DSY140 (ade2/ade2, Aacp : : X/Aacp : : X, gall/gall, Aura3 : : CAT/ Aura3 : : CA7). The strain DSY140 is readily available from Dr Dominique Sanglard (Institute of Microbiology, University Hospital Lausanne (CHUV), Rue de Bugnon 44, CH-1011 Lausanne, Switzerland) and is derived from strain Ha900 (Sanglard & Fiechter, 1992 Yeast 8,1065-1075). DSY140 is auxotrophic for adenine and uracil. Transformants were selected for their ability to grow in the absence of uracil in a yeast minimal medium supplemented with adenine.

The genotypes of Ura+ transformants were analysed by PCR to verify the deletion of one of the FAO alleles. The following primers were used in this PCR assay: one specific for the FAO gene (5'GAA GGA TCC ACA ATT GAT TGC ACA GC 3', Seq. ID No. 41) and one specific for the hisG DNA fragment present in the disruption cassette (5'CGC GCG ATA CAG ACC GGT TC 3'Seq. ID No. 42). The PCR reaction comprised yeast DNA mixed with primers and conventional PCR buffers with 30 cycles of: annealing at 54°C for 1 min; elongation at 72°C for 2 min; followed by denaturation at 94°C for 30 seconds.

Following this PCR, specific products with a length of 2 kb were observed in Ura+ transformants obtained from the pSVC2 linear DNA and with a length of 1.5 kb in Ura+ transformants obtained from the pSVU2 linear DNA. The positive transformants were named DSY1670 (from pSVC2) and DSY1671 (from pSVU2). Regeneration of the ura3 genetic marker was obtained by incubating 106 cells from DSY1670 and DSY1671 respectively in a medium containing 5-fluoroorotic acid (5-FOA), which counter-selects for the absence of the URA3 gene contained in the disruption cassette (yeast cells with a functional URA3 gene convert 5-FOA into a toxic compound which kills the cells). Single Ura-derivatives were obtained from DSY1670 and DSY1671 and one transformant from each strain (DSY1688 and DSY1692 respectively) were retained for further work.

These Ura-strains were re-transformed with linear EcoRI fragments as described above. Ura+ transformants were obtained and their genotypes analysed by PCR. Two primers were used which allowed for distinguishing of the FAO wild type allele and the disrupted FAO alleles. These primers had the following sequences 5'TTG CAT TTG AAA GCC ATC CAT TAC CCT 3' (Seq. ID No. 43) and 5'CAA ACA ATC CGA AAC ACC AAC AAC AGT 3' (Seq. ID No. 44). The PCR products obtained from the amplification of the FAO wild type allele and the FAO disrupted allele were expected to be 1.9 kb and 2.1 kb, respectively. Positive transformants, in which the FAO alleles were disrupted, were obtained from the transformation of DSY1688 with linear DNA fragment from pSVC2. In these strains (DSY1700-1, DSY1700-2 and DSY1700-3), no wild type FAO allele could be amplified, only the disrupted FAO allele. The genotype of DSY1700 derivatives is: (Afao : : hisG/Afao ::hisG-URA3-hisG).

No double FAO deletion mutants of C. tropicalis could be obtained using the pSVU construct, presumably because the 5'and 3'UTRs of the respective FAO alleles differed too much to allow homologous recombination to occur between the pSVU construct and the second FAO allele.

The representative strain DSY 1700-2 has been made the subject of a deposit under the Budapest Treaty at the National Collection of Yeast Cultures (NCYC, AFRC Institute of Food Research, Norwich Laboratory, Colney Lane, Norwich NR4 7UA, United Kingdom). The date of deposit is March 1999, and the accession number is