A Consensus Sequence for Long-chain Fatty-acid Alcohol Oxidases from Candida Identifies a Family of Genes Involved in Lipid ω-Oxidation in Yeast with Homologues in Plants and Bacteria*

The yeast Candida cloacaeis capable of growing on alkanes and fatty acids as sole carbon sources. Transfer of cultures from a glucose medium to one containing oleic acid induced seven proteins of M r102,000, 73,000, 61,000, 54,000, and 46,000 and two in the region ofM r 45,000 and repressed a protein ofM r 64,000. The induction of theM r 73,000 protein reached a 7-fold maximum 24 h after induction. The protein was confirmed by its enzyme activity to be a long-chain fatty-acid alcohol oxidase (LC-FAO) and purified to homogeneity from microsomes by a rapid procedure involving hydrophobic chromatography. An internal peptide of 30 amino acids was sequenced. A 1100-base pair cDNA fragment containing the LC-FAO peptide coding sequence was used to isolate a single exon genomic clone containing the full-length coding sequence of an LC-FAO (fao1). The fao1 gene product was expressed inEscherichia coli and was translated as a functional long-chain alcohol oxidase, which was present in the membrane fraction. In addition, full-length coding sequences for a Candida tropicalis LC-FAO (faoT) and a second C. cloacae LC-FAO (fao2) were isolated. The DNA sequences obtained had open reading frames of 2094 (fao1), 2091 (fao2), and 2112 (faoT) base pairs. The derived amino acid sequences of fao2 and faoTshowed 89.4 and 76.2% similarities to fao1. Thefao1 gene is much more highly induced on alkane than isfao2. Although this study describes the first known DNA sequences encoding LC-FAOs from any source, there are unassignedArabidopsis sequences and an unassignedMycobacterium sequence in the GenBankTM Data Bank that show strong homology to the described LC-FAO sequences. The conservation of sequence between yeast, plants, and bacteria suggests that an as yet undescribed family of long-chain fatty-acid oxidases exists in both eukaryotes and prokaryotes.

Candida cloacae and Candida tropicalis are industrial yeast species capable of utilizing both alkanes and long-chain fatty acids as sole carbon sources for growth. These water-immiscible substrates are metabolized to carbon dioxide by two sequential oxidative pathways: (a) the membrane-bound -oxidation pathway and (b) the ␤-oxidation pathway located in peroxisomes. During -oxidation, the methyl end of the molecule is oxidized successively by a cytochrome P450 alkane/fatty-acid oxidase, a hydrogen peroxide-generating alcohol oxidase, and an aldehyde dehydrogenase, producing -alcohols, -aldehydes, and -fatty acids, respectively. Growth on both alkanes and fatty acids results in dicarboxylic acid formation as a metabolic intermediate. Such long-chain dicarboxylic acids are very versatile raw materials for the oleochemical industry and are used in the production of fragrances, polyamides, polyesters, adhesives, and macrolide antibiotics (1). The dicarboxylic acid products of -oxidation are further oxidized in the peroxisome by the ␤-oxidation pathway following activation to acyl-CoAs.
Two different biological modes of alcohol dehydrogenation have been identified. The first is a nicotinamide-dependent reaction catalyzed by alcohol dehydrogenases, and the second is a flavin-dependent reaction catalyzed by alcohol oxidase. The latter uses molecular oxygen as acceptor and generates hydrogen peroxide (2,3). Flavin-dependent alcohol oxidases have been isolated from a number of different fungal sources (4 -9). In all organisms studied to date, the enzyme is octameric, with the exception of C. tropicalis, where it is dimeric (7). The substrate specificity of the enzyme differs considerably depending on the source of the enzyme. In C. tropicalis, the enzyme is most active on long-chain fatty alcohols (7), whereas secondary alcohols such as dodecan-2-ol and long-chain -hydroxy fatty acids are also good substrates of the enzyme (7). Alcohol oxidases from Kloeckera sp., Hansenula polymorpha, and Candida boidinii are very active toward methanol and ethanol, but are inactive toward substrates of chain length longer than C 5 and will not oxidize secondary alcohols or -hydroxy fatty acids (4,6). Substrate specificity has also been investigated for the alcohol oxidase from the filamentous fungus Aspergillus flavipes grown on hexadecanol. Although this enzyme will oxidize long-chain alcohols, it will not utilize ␣,-diols, -hydroxy fatty acids, or secondary alcohols (9). The enzymes from Candida maltosa (10), C. tropicalis (7), and C. cloacae (11) all oxidize -hydroxy fatty acids, whereas those from other organisms studied lack the ability to use these substrates. Long-chain alcohol oxidase has been purified from C. tropicalis (7), but to date, there is no gene or amino acid sequence available.
Oxidation of long-chain alcohols by alcohol oxidase has been reported in germinating seedlings of the plant Simmondsia chinensis (jojoba) (12), which accumulates long-chain fatty alcoholcontaining waxes as its main storage product. The utilization of these storage products as an energy source requires an appropriate oxidative pathway, presumably having -oxidation as a key component. However, to date, no potential long-chain alcohol oxidase DNA coding sequences have been identified in plants.
There has been much interest in cloning and manipulating genes of the alkane/fatty acid oxidation pathway in industrial yeast species (1,13,14). Strains giving high yields of dicarboxylic acids from alkane-or fatty acid-containing media have been produced by selective mutagenesis of the oxidation pathways. However, comparatively little is known at the molecular level about long-chain fatty-acid alcohol oxidases and aldehyde dehydrogenases from these organisms.
In this study, we report the induction of C. cloacae proteins upon transfer to a lipid-based growth medium, purification and partial amino acid sequence determination of long-chain fattyacid alcohol oxidase, the first isolation of cDNA and genomic clones from C. cloacae and C. tropicalis, and the functional expression of an alcohol oxidase in Escherichia coli. It is expected that disruption of these genes will result in industrial yeast strains capable of producing -hydroxy fatty acids from fatty acid substrates.

Yeast Strains and Growth Conditions
C. cloacae strain 3152 FERM P-736 was used throughout these experiments (15). It was maintained at 4°C on agar slopes containing (per 100 ml) 1.5 g of agar, 0.5 g of yeast extract, 0.5 g of peptone, and 1 g of sucrose. Starter cultures, used to inoculate shake flasks and fermentors, unless otherwise specified, were prepared by transferring a loop of agar slope culture into 50 ml of medium containing 0.25 g of yeast extract, 0.25 g of peptone, and 0.5 g of sucrose and incubating for 24 h in a baffled 250-ml flask at 30°C with shaking at 90 rpm. The minimal medium for shake flask cultures contained (per liter) 25 g of sucrose, 7.6 g of NH 4 Cl, 1. Bulk preparation of induced cells for enzyme preparation was performed in 6-liter fermentors (Biolafitte) using a minimal medium containing (per liter) 10 g of hexadecane (in place of sucrose as carbon source), 4.5 g of (NH 4 ) 2 HPO 4 , 1.5 g of yeast extract, 0.1 g of CaCl 2 , and 1.5 g of Na 2 SO 4 . The remaining vitamins and minerals were as detailed above. The minimal medium (5 liters) was inoculated with 2% (v/v) starter culture and then grown at 30°C and 0.45 liters/min aeration with stirring at 700 rpm and a pH control at 6.4 by auto-addition of 40% NaOH. When the hexadecane substrate was fully utilized (15-20 h), the cells were harvested (3500 g/10 min); washed with 3 liters of 50 mM HEPES/NaOH (pH 7.5); and resuspended in 140 ml of 50 mM HEPES/ NaOH (pH 7.5), 1 mM EDTA, and 1 mM dithiothreitol prior to cell disruption using a French pressure cell. The wet weight of cells was typically 110 -120 g. C. tropicalis strain NCYC 470 was grown in the same way as C. cloacae on an alkane medium.

Induction by Oleic Acid
Starter cultures of C. cloacae were grown overnight in 1% yeast extract, 1% peptone, and 2.5% sucrose at 30°C. Baffled flasks containing 250 ml of minimal medium and sucrose as the carbon source were inoculated with 2.0 ml of starter culture and grown for 48 h at 30°C and 90 rpm on a shaker. The cells were then centrifuged at 2400 ϫ g for 10 min; the resulting cell pellet was resuspended in the same volume of minimal medium but with 1% oleic acid as the carbon source instead of sucrose; and incubation was continued at 30°C and 90 rpm for a further 125 h. Samples of 30 ml were removed from the flasks immediately before addition of oleic acid and at 6, 24, 50, and 125 h after addition. They were centrifuged at 10,000 ϫ g for 10 min, and the cell pellet was frozen in liquid N 2 and stored at Ϫ80°C. Cell-free extracts were made from the pellets by grinding each pellet in liquid N 2 three times and then rapidly resuspending the powder in 10 ml of 50 mM potassium phosphate (pH 7.5) containing 5% glycerol and 0.5% CHAPS 1 and centrifuging at 10,000 ϫ g for 10 min.

Purification of Long-chain Fatty-acid Alcohol
Oxidase from C. cloacae The procedure developed involved microsomal preparation, solubilization, and (NH 4 ) 2 SO 4 precipitation followed by chromatography on phenyl-Superose. All procedures were performed at 4°C unless otherwise specified. Cells (132 g, wet weight) were passed through a French pressure cell three times at 20,000 p.s.i.; the disrupted cell extract was centrifuged at 20,000 ϫ g for 30 min; and the precipitate was discarded. At this and other stages, the preparation was snap-frozen in liquid N 2 and stored at Ϫ80°C prior to further processing. The supernatant was thawed, and 100 mM Tris-HCl (pH 7.5) was added to give a final volume of 230 ml. The microsomal fraction was pelleted by ultracentrifugation at 140,000 ϫ g for 1.5 h. The pellet was washed by suspending in 100 mM HEPES/NaOH (pH 8.0) containing 0.15 M KCl (final volume of 115 ml) and pelleted at 140,000 ϫ g for 1.5 h. The washed microsomes were resuspended in 50 mM HEPES/NaOH (pH 8.0) (final volume of 62 ml). The resuspended pellet was made up to 500 ml with 50 mM HEPES/ NaOH (pH 8.0). Sodium cholate was added to 1.0%, and phenylmethylsulfonyl fluoride (in isopropyl alcohol) was added to 1 mM, (NH 4 ) 2 SO 4 was added to 35% (w/v); the solution was stirred for 20 min and centrifuged at 20,000 ϫ g for 5 min. (NH 4 ) 2 SO 4 was added to 55% (w/v) to the supernatant; the solution was stirred for 20 min and centrifuged at 20,000 ϫ g for 5 min. The resulting pellet was resuspended in 50 mM HEPES/NaOH (pH 8.0) containing 1% sodium cholate to a final volume of 52 ml, dialyzed 2 ϫ 1 h against two changes of 2 liters of 50 mM HEPES/NaOH (pH 8.0) and a further 500 ml of 50 mM HEPES/NaOH (pH 8.0) containing 1% CHAPS for 1 h, and centrifuged at 20,000 ϫ g for 5 min to give a clear supernatant. The (NH 4 ) 2 SO 4 precipitant was resuspended, dialyzed, and diluted to 15 ml with the column equilibration buffer; filtered through a 0.2-m filter; and loaded onto a phenyl-Superose HR 5/5 column (Amersham Pharmacia Biotech) equilibrated in 10 mM Tris-HCl (pH 8.5) containing 1.7 M (NH 4 ) 2 SO 4 and 0.5% CHAPS. After washing the column with 15 ml of buffer, a 10-ml linear reverse gradient was run from 1.7 to 0.85 M (NH 4 ) 2 SO 4 , followed by a 15-ml linear gradient from 0.85 to 0.5 M (NH 4 ) 2 SO 4 . A flow rate of 0.6 ml/min was used throughout.

PCR Isolation of a Probe for fao
To produce a probe for the cDNA library screen, 1 l (1.5 ϫ 10 3 plaque-forming units) of the C. cloacae oleic acid-induced cDNA (nonamplified) library was placed into a reaction mixture containing 10 pmol of M13 forward primer (5Ј-TTG TAA AAC GAC GGC CAG T-3Ј) or M13 reverse primer (5Ј-CAC ACA GGA AAC AGC TAT GAC C-3Ј), 20 pmol of internal degenerate FAO primer (5Ј-ACN AAY CAR CAR CTN TTY ATG ATH GC-3Ј, where N ϭ ACGT, Y ϭ CT, R ϭ AG, and H ϭ ACT; corresponding to the amino acid sequence TNQQLFMIA), 2 units of Taq polymerase, 1.5 mM MgCl 2 -containing PCR buffer, and 1.25 mM dNTPs (Roche Molecular Biochemicals). Amplification was performed under the following conditions for 35 cycles: denaturation at 94°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 2 min. A mixture of products resulted from the reverse and internal primer reactions, and these were subcloned into pGEM ® -T vector (Promega) and analyzed by sequencing. The largest open reading frame clone, named pAX17, was used in subsequent hybridization experiments as described below.

Construction of DNA Libraries
cDNA Libraries-Poly(A) ϩ mRNAs from 24-h oleic acid-induced C. cloacae and C. tropicalis were used for the construction of a randomprimed non-directional cDNA library in EcoRI-digested alkaline phosphatase-treated ZapII vector (Stratagene). Approximately 5 g of mRNA was used in the reverse transcriptase reaction (TimeSaver cDNA synthesis kit, Amersham Pharmacia Biotech). EcoRI/NotI adapters were added to the ends, and cDNA was ligated with EcoRI-digested ZapII. Gigapack II Gold packaging extract (Stratagene) was used to form phage particles, which were used to transfect E. coli XL1-Blue cells (Stratagene).
Genomic Libraries-Genomic C. cloacae and C. tropicalis DNAs were isolated from cells grown on 10 g/liter yeast extract, 20 g/liter peptone, and 20 g/liter glucose according to Philippsen et al. (20). A Sau3A partial digest of C. cloacae DNA was size-fractionated to 14 -23 kb on a 10 -40% sucrose gradient and ligated into Bluestar BamHI arms (Novagen). Gigapack II Gold-packaged particles were used to transfect E. coli ER1647 (Novagen). A Sau3A partial digest of C. tropicalis DNA was size-fractionated on a 10 -40% sucrose gradient to 10 -12 kb and ligated into ZapExpress BamHI arms (Stratagene). PhageMaker packaging extract (Novagen) was used to form phage particles, which were then used to transfect XL1-Blue MRFЈ cells (Stratagene).

Hybridization Screening of DNA Libraries
cDNA Libraries-For the C. cloacae cDNA screen, prehybridization was carried out at 65°C for 2 h in 6ϫ SSC, 1ϫ Denhardt's solution, 0.5% SDS, 0.05% sodium pyrophosphate, and 0.05 mg/ml herring sperm DNA, and hybridization was carried out at 65°C overnight in essentially the same buffer but without herring sperm DNA and with 1 mM EDTA. Filters were washed at 65°C with 2ϫ SSC and 0.1% SDS (2 ϫ 15 min) and with 0.2ϫ SSC and 0.1% SDS (2 ϫ 15 min). For the C. tropicalis screen, prehybridization was carried out at 55°C for 2 h, and hybridization was carried out at 55°C overnight in the same buffers as listed above. Filters were washed 2 ϫ 30 min at 55°C with 2ϫ SSC and 0.1% SDS.
Genomic Libraries-For the C. cloacae screen, prehybridization was carried out for 2 h at 55°C in 6ϫ SSC, 5ϫ Denhardt's solution, 0.5% SDS, 100 g/ml herring sperm DNA, and 50% formamide, and hybridization was carried out with the probe overnight at 55°C in the same hybridization mixture. Filters were washed 3 ϫ 15 min in 2ϫ SSC and 0.1% SDS at 55°C. For the C. tropicalis screen, prehybridization was carried out for 2 h at 42°C, and hybridization was carried out overnight at 42°C in the same hybridization mixture as described above. Filters were washed 2 ϫ 15 min in 2ϫ SSC and 0.1% SDS and 2 ϫ 15 min in 1ϫ SSC and 0.1% SDS at 42°C.

Northern Analysis
C. cloacae cells were grown for 24 h in minimal medium containing either 2% glucose or 1% oleic acid as a carbon source. The cells were harvested and ground in a mortar and pestle under liquid N 2 , and RNA was extracted using hot SDS (21). mRNA was prepared using oligo(dT) spun columns (Pharmacia mRNA Purification kit). Northern analysis was conducted using 2 g of mRNAs on agarose gels containing 0.66 M formaldehyde, blotting onto Hybond-N TM membrane (Amersham Pharmacia Biotech), and hybridizing according to standard methods (22) with a Rediprime random primer-labeled probe (Amersham Pharmacia Biotech).

RT-PCR Quantification of fao1 and fao2 Gene Expression
RNA samples were prepared from cells grown under identical conditions described for Northern analyses, except that samples were grown on hexadecane. They were treated with RNase-free DNase I (Life Technologies, Inc.) at room temperature for 15 min to remove possible DNA contamination. DNase I was subsequently inactivated at 65°C for 10 min in the presence of 2.5 mM EDTA. The resulting RNA (2 g) was used as a template for RT-PCR using the CLONTECH RT-PCR kit following the manufacturer's instructions. The oligonucleotides 5Јatgtcccatcaagttgaagacc-3Ј and 5Ј-atgaatcccgttgttgaagacagcc-3Ј are specific to fao1 and fao2 5Ј-end coding sequences, respectively; the oligonucleotides 5Ј-ctaaagtttagtttgtggttgcaagtc-3Ј and 5Ј-ctaaagtttactttgttgc-3Ј are specific to fao1 and fao2 3Ј-end coding sequences, respectively. PCRs were carried out for 29 cycles with denaturing at 94°C for 1 min, annealing at 60°C for 1 min, and extension at 72°C for 3 min and were performed on a Robocycler (Stratagene). Initial experiments demonstrated that up to 30 cycles of PCR product formation was linear.

Overexpression of C. cloacae Long-chain Fatty-acid Alcohol Oxidase FAO1 in E. coli
Oligonucleotide primers were designed to PCR-amplify the fulllength fao1 coding sequence: 5Ј-CTAGCTAGCATGTCCCATCAAGTT-GAAGACC-3Ј and 5Ј-CGGGATCCCTAAAGTTTAGTTTGTGG TTG-CAAGTC-3Ј. The purified PCR product was digested by NheI and BamHI and subsequently cloned into the overexpression vector pET17b (Novagen); the resulting plasmid was designated pET17b-fao1. The sequence of the fao1 coding region was confirmed by DNA sequencing. The expression construct was transformed into E. coli BL21(DE3) cells according to the Stratagene protocol, and the cells were grown on an LB/ampicillin plate overnight. One colony was transferred into 5 ml of LB medium containing ampicillin (100 mg/ml), and the culture was grown overnight at 37°C with shaking at 200 rpm. The cells were harvested by centrifugation (5000 rpm, 10 min) and resuspended in 250 l of buffer containing 20 mM Tris-HCl (pH 8.0) and 2 mM EDTA. For enzyme assay, microsomal fractions were prepared from 500 ml of overnight culture grown on LB medium. The pellet was finally washed and resuspended in 1 ml of buffer containing 20 mM Tris-HCl (pH 8.0) and 2 mM EDTA.

Miscellaneous Methods
SDS-Polyacrylamide Electrophoresis-SDS-polyacrylamide gels consisted of a 5% stacking gel with a 10% running gel and were run on a Bio-Rad Mini-Protean gel kit. The buffers used were as described by Laemmli (16).
Native Gel Electrophoresis and Biological Activity Staining-Native PAGE was carried out at 200 V for 2.5 h on a 10% resolving and 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 buffers, respectively. The sample buffer consisted of 10 mM Tris-HCl (pH 6.8), 2% glycerol, 1% sodium cholate, and sufficient bromphenol blue to make it visible. The running buffer and apparatus were chilled on ice before and during the run. The sample was 50 l of 0.2 units of hydroxylapatite-purified C. cloacae material prepared by the method of Dickinson and Wadforth (7). To stain the gel for enzyme activity, buffer, ABTS, and peroxidase at 10 times the concentration used in the standard LC-FAO assay and dodecanol at 2.7 mM were applied to the surface of the gel. This was incubated at room temperature for 5 min. Alcohol oxidase activity was identified 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.
Enzyme Assays-LC-FAO was assayed by spectrophotometry (3). The assay mixture contained 50 mM Tris-HCl (pH 8.5) 0.7 mg/ml ABTS, 7 units of horseradish peroxidase, and 50 M dodecanol previously dissolved in Me 2 SO, in a final volume of 1.0 ml unless otherwise specified. Reactions were initiated by addition of enzyme, and the increase in absorbance at 405 nm was measured. The value of ⑀ for the radical cation of ABTS is 18.4 mM Ϫ1 cm Ϫ1 , and 1 mol of substrate gives rise to 2 mol of radical cation (17). One unit of enzyme activity catalyzes the conversion of 1 nmol of substrate to product/min.
Protein Concentration-This was determined using the dye binding method of Bradford (18) with bovine serum albumin as a standard.
Generation and Sequencing of Long-chain Fatty-acid 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 digestion (19). For N-terminal amino acid sequence determination, proteins were transferred onto Problot membrane (Applied Biosystems, Inc.) and visualized by Coomassie Blue staining as directed in the Problot manual. Amino acid sequencing was performed with an ABI Model 477 sequencer.
DNA Sequencing-Sequencing was carried out using an Applied Biosystems Model 373 DNA sequencer. Computer analysis of DNA sequences was carried out using DNA Strider (23) and FASTA alignments.
Chemicals-All biochemical reagents were obtained from Sigma and were highest purity available. Reagents for electrophoresis were from Bio-Rad.

Alterations in the Protein Profile upon Oleic Acid Induction-
Both LC-FAO enzyme activity and SDS-PAGE protein profiles were analyzed in cell extracts of C. cloacae following transfer from a sucrose-containing growth medium to one containing oleic acid as the sole carbon source. The induction of LC-FAO enzyme activity increased 4-fold in 6 h and reached a maximum of 7-fold at 24 h; similar results were obtained using the alkane hexadecane as an inducer (data not shown).
C. cloacae cell extracts from cultures grown for 0, 6, and 125 h after addition of oleic acid were run on SDS-polyacrylamide gel and stained for protein with Coomassie Blue. Growth on oleic acid resulted in the induction of proteins of M r 102,000, 73,000, 61,000, 54,000, and 46,000 and two in the region of M r 45,000 and the disappearance of a protein of M r 64,000 (Fig. 1).
Purification of Long-chain Fatty-acid Alcohol Oxidase from C. cloacae-We initially adopted a procedure similar to that used for the purification of LC-FAO from C. tropicalis (7). The C. tropicalis procedure did not, however, result in a homogeneous preparation with C. cloacae extracts. Accordingly, we developed a procedure with hydrophobic chromatography on phenyl-Superose as a key step in the purification, taking advantage of the hydrophobic nature of the protein. It was essential to prepare the microsomes from alkane-induced cells as oleic acid-induced cells accumulated fatty acids, which interfere with membrane pelleting during the ultracentrifugation step. By directly applying the resuspended (NH 4 ) 2 SO 4 precipitate to a phenyl-Superose column and eluting it with a reverse gradient of (NH 4 ) 2 SO 4 , it was possible to obtain a homogeneous preparation. The fractions with highest activity eluted at ϳ0.7 M (NH 4 ) 2 SO 4 ( Fig. 2A). Analysis of alcohol oxidase activity and protein profiles by SDS-PAGE demonstrated that the activity coincided with the elution of the M r 73,000 protein (Fig. 2B). The enzyme was purified 230-fold with a 10.7% recovery of biological activity (Table I). Further evidence that LC-FAO is an M r 73,000 protein was obtained from native gels. The enzyme activity was visualized directly on the gel as a green band when the gel was incubated with enzyme assay reagents. This band was excised and subjected to SDS-PAGE, and a major band at M r 73,000 was seen, confirming assignment of this band as LC-FAO. The molecular weight of the alcohol oxidase from C. cloacae is the same as that reported for the LC-FAO enzyme from C. tropicalis (7). The pH optimum of the enzyme from C. cloacae was between 8.5 and 9.0, and the apparent K m for dodecanol was between 4.0 and 5.0 M (data not shown).
Amino Acid Sequence of Long-chain Fatty-acid Alcohol Oxidase-N-terminal amino acid sequencing was performed three times on purified LC-FAO following transfer to Problot. No amino acid sequence data were obtained, indicating that the alcohol oxidase from C. cloacae had a blocked N terminus. When protein bound to a polyvinylidene difluoride membrane was treated with CNBr to cleave methionine residues, amino acid sequence data were obtained at the 50 -100-pmol level, indicating that protein was present on the sequencing disc at approximately the expected levels given the typical recoveries for these processes. In-gel digestion of the M r 73,000 protein with Glu-C resulted in three major peptides, which were separated electrophoretically and blotted onto Problot. One of these gave an unambiguous sequence of 30 amino acids (Table  II), which was suitable for degenerate primer design.
Isolation of LC-FAO-encoding cDNAs from C. cloacae-The cDNA library and a degenerate primer based on the internal amino acid sequence SGGTIPSTNQQLFMIAGSTFGGGST-VNW from C. cloacae LC-FAO were used in combination with library vector-based M13 forward and T3 primers to create a probe for screening the cDNA library. PCR yielded products of 1100, 800, 600, 550, and 300 bp, which were directly subcloned into pGEM ® -T vector. Sequencing of these clones showed that they all carried the same translated LC-FAO-derived partial amino acid sequence, TNQQLFMIAGSTFGGGSTVNW (data not shown). Screening of the C. cloacae cDNA library with the insert from the subcloned 1100-bp PCR product (pAX17) resulted in four independent clones containing inserts of 2.4, 1.0, 0.6, and 0.6 kb. It was clear from sequencing analysis that the four clones represented two classes of coding sequence. One cDNA of 600 bp had an internal translated amino acid sequence identical to the sequenced peptide and was designated as fao1 class. Three cDNAs with 2400-, 1000-, and 600-bp inserts with identical overlapping sequences were grouped as fao2 class. The fao2 cDNAs encoded a 2-amino acid variant of the amino acid sequence obtained by direct protein sequencing. The other factor determining fao classification was the polymorphism in restriction endonuclease target sites. The fao1 class of LC-FAO carried an XbaI site not present in fao2; additionally, fao2 carried a HindIII site not present in fao1.
Northern Analysis-To verify the full-length size of LC-FAOencoding cDNA, the mRNA from oleic acid-induced C. cloacae was probed in a Northern analysis with the insert from clone As in the case of LC-FAO, fao mRNA was induced five to seven times on transfer of the organism to an oleic acid medium (Fig.  3A). It should be borne in mind that this probe will hybridize to fao1 and fao2 as described below.
Isolation of the fao1 Class Full-length Coding Sequence-The only representative of the fao1 class was not full-length, and we required the promoter sequence driving the oleic acid induction for further studies. To obtain the full-length coding sequence, a genomic library was screened with the fao1 600-bp fragment of cDNA already isolated. Twenty hybridizing clones were purified, and after PCR isolation of the complementary 600-bp fragment, XbaI/HindIII restriction enzyme analysis was carried out to compare them with fao1 class genes. Plasmids were rescued from clones, and plasmid pgFAO14, which carried an insert of 18 kb containing the full-length fao1 class gene, was further characterized. Sequencing of the 4.3-kb region of the genomic clone containing the LC-FAO gene identified a single exon with an open reading frame (2094 bp) in the same order as that predicted by Northern analysis (2.4 kb). The 3Ј-end of the fao1 gene derived from genomic DNA contained a typical eukaryotic polyadenylation signal, AATAAA (24), as well as a Saccharomyces cerevisiae consensus sequence for transcription termination, TAG . . . TA(T/A)GT . . . TTT (25). The derived protein of M r 77,300 of both fao1 and fao2 sequences (698 and 697 amino acids, respectively) is also consistent with LC-FAO protein analysis by SDS-PAGE.
The deduced amino acid sequence of fao2 contained a carboxyl-terminal peroxisome-targeting sequence, SKL (26). However, the corresponding fao1 carboxyl-terminal sequence was TKL. It is possible that in fao1 the corresponding sequence may not function in targeting the protein into peroxisomes, and thus, the two gene products would be located in different cell compartments. Both fao1 and fao2 contain the consensus sequence Cys-X-X-Cys-His for a cytochrome c family heme-binding signature (27). The fao1 and fao2 sequences showed 80.0% sequence identity at the nucleotide level, and the deduced amino acid similarity was 89.4%. The FASTA alignment of the deduced fao1 and fao2 amino acid sequences is shown in Fig. 4.
RT-PCR Quantification of fao1 and fao2 Gene Expression-Using gene-specific primers for fao1 and fao2, we used RT-PCR to investigate the expression of both the fao1 and fao2 genes following alkane induction (Fig. 3B). fao1 gene expression could be detected in the alkane-free medium, whereas fao2 could not. Following induction, fao1 was elevated ϳ6-fold, and fao2 could be detected, but was the minor species. From a mRNA abundance perspective, fao1 would seem to be the dominant form of alcohol oxidase in C. cloacae and is 10 times more highly expressed than fao2 following alkane induction.
Overexpression of C. cloacae Long-chain Fatty-acid Alcohol Oxidase FAO1 in E. coli-The fao1 gene product was expressed in E. coli using the pET17b expression system. Both membrane   fractions and soluble proteins were analyzed by SDS-PAGE. Membranes isolated from the fao1-containing vector showed an elevated level of a protein of M r ϳ73,000, consistent with expression of the fao1 gene (Fig. 5). Membranes were prepared from BL21 cells transformed with either the empty vector or one containing the fao1 gene and assayed for alcohol oxidase activity. Using 144 g of membrane protein/incubation, the pET17b-fao1 membranes had 2000 units of enzyme activity with juniperic acid (-hydroxy-C16:0) and 10,000 units with 1-dodecanol. No activity was detected with membranes isolated from BL21 cells transformed with the empty vector sequence even after a 2-day incubation. In contrast, the reaction with membranes from BL21 cells transformed with the vector sequence virtually went to completion within 5 min. The specific activity was five times higher with 1-dodecanol (0.069) than with juniperic acid (0.013). This difference in specific activity between the two substrates is very similar to that reported for the purified enzyme from C. tropicalis (7). The fao2 gene would not express in E. coli, and the reason for this is unknown.
Cloning and Sequencing of Fatty-acid Alcohol Oxidase from C. tropicalis-Much of our current knowledge about yeast species that are able to utilize alkanes and long-chain fatty acids comes from C. tropicalis. To expand and to complement this knowledge, we isolated fao1-like clones from a C. tropicalis cDNA library using the C. cloacae fao1 fragment from pg-FAO14 as a probe. One of the cDNA clones isolated contained an open reading frame of 617 amino acids. Comparison with the corresponding derived 679-residue-long C. cloacae fao1 gene sequence suggested that the 3Ј-end of this cDNA clone was missing. To obtain the rest of the sequence, we generated a product of ϳ750 bp with identical overlapping sequence from the C. tropicalis genomic library by PCR. This fragment contained the 3Ј-end of the C. tropicalis faoT gene. The combined faoT sequence of 4233 bp has an open reading frame of 2112 bp, which corresponds to 704 amino acid residues (Fig. 4). faoT shares 60.6 and 61.7% nucleotide identities and 74.8 and 76.2% amino acid sequence similarities with C. cloacae fao1 and fao2, respectively (Fig. 4).
Identification of Unassigned Sequences Homologous to Candida fao Genes-Alignment comparison of the three Candida LC-FAOs revealed large regions of homologous domains throughout the full length of the derived sequences (Fig. 4). Further analysis of the GenBank TM Data Bank with the FASTA-generated Candida fao consensus sequence showed that there were two high-scoring unassigned Arabidopsis sequences (accession numbers AB015474 and AL022580) and an unassigned genomic fragment from Mycobacterium tuberculo-sis (accession number Z77162) in the data base. Both of the plant sequences contained one intron with classical intron-exon border sequences. Removal of these intron sequences revealed five domains (domains I-V) that were analogous in all three Candida and both Arabidopsis sequences (see Fig. 4). Domain I was present in all the Candida sequences described in this paper as well as in the Arabidopsis and Mycobacterium sequences. Domains II and III (which contain a consensus sequence for the cytochrome c family heme-binding site) were unique to the Candida and Arabidopsis sequences, whereas domain IV was present in Arabidopsis, Candida, and a Caenorhabditis elegans cosmid data base entry (accession number AF036689). Partial similarities to domain V were observed in many data base entries, including several known oxidases (a hypothetical Rhizobium glucose-methanol-choline-type oxidoreductase (accession number AE000087), M. tuberculosis cholesterol oxidase (accession number X99343), Hansenula polymorpha methanol oxidase (accession number X02425), Coriolus versicolor pyranose oxidase (accession number D73369), and the Streptomyces cholesterol oxidase choM gene (accession number U13981)).
The first 200 derived amino acids of the Candida sequences, although similar to each other, showed little homology to any other sequence in the data base. This region may represent a region for membrane attachment and/or a site for proteinprotein interaction. Membrane-spanning sequence analysis us- ing a transmembrane ␣-helix prediction model, of the translated gene sequences (28) showed that both C. cloacae fao genes contained three and C. tropicalis faoT contained four putative transmembrane-spanning regions within the first 200 amino acids. In addition, all of them have a putative transmembranespanning region in the carboxyl-terminal region around amino acids ϳ550 -600 (Fig. 6). DISCUSSION Transfer of C. cloacae from a sucrose-containing medium to one containing alkane or fatty acid as the carbon source results in a distinct alteration in the protein profile. Seven new proteins were present in the lipid-containing medium, and one protein disappeared. When C. tropicalis and C. cloacae cells are grown on lipid-containing medium, there is a strong induction of both theand ␤-oxidation pathways together with peroxisomal proliferation. Proteins that are likely to be up-regulated are (a) those involved in import of the substrate into the cells such as alkane/fatty acid transporters; (b) -oxidation pathway enzymes including cytochrome P450 reductase, alkane cytochrome P450, alcohol oxidase, and aldehyde dehydrogenase; (c) components of the ␤-oxidation pathway including acyl-CoA synthetase, acyl-CoA oxidase, acetoacetyl-CoA thiolase, 3-ketoacyl-CoA thiolase, and the trifunctional enzyme containing acyl-CoA hydrolase, 3-hydroxyacyl-CoA dehydrogenase, and 3-hydroxyacyl-CoA epimerase activities; and (d) peroxisomal components such as catalase and appropriate transporters. From molecular weight assignments, the M r 54,000 protein is possibly the alkane-induced cytochrome P450. Assignment of functions of the other proteins presented here will have to await further protein sequencing and protein function studies. Sequencing of these induced proteins in Candida will provide valuable information about the pathway of lipid degradation in this organism. The identity of the M r 73,000 protein has been confirmed in this paper as long-chain fatty-acid alcohol oxidase following cloning and expression in E. coli.
The highest sequence homologies of fao genes determined in this study are to genes in Arabidopsis and Mycobacterium of unknown function/assignment. There is also high homology in domain V to other known oxidoreductases, in particular cholesterol oxidase (29), choline oxidase (30), and cellobiose oxidase (31). These three enzymes are flavoproteins that utilize molecular oxygen in their reactions and that generate hydrogen peroxide, characteristics shared with the alcohol oxidase from Candida species. The purified cellobiose oxidase protein has been studied spectroscopically and has a heme and flavin domain (31). It is of interest that domain III present in all three of the Candida species contains a heme-binding motif.
Alcohol oxidases from both C. cloacae and C. tropicalis have wide substrate specificity, oxidizing both straight-chain alcohols and -hydroxy fatty acids. The kinetic constants and the pH optimum of the enzyme from C. cloacae are similar to those reported for C. tropicalis (7). It is known that C. cloacae can utilizes a wide range of fatty acids as sole carbon sources (11). The two C. cloacae genes identified may code for proteins of slightly different substrate specificity or cellular location. In C. maltosa, direct gene disruption has been used to elucidate the function of the n-alkane-inducible cytochrome P450 genes ALK1, ALK2, ALK3, and ALK5 (14). Each of the four individual genes is sufficient for growth on n-alkanes, but disruption of all four genes prevents growth on alkanes. The ALK5 gene exhibits narrower chain length specificity and cannot support growth on shortchain n-alkanes (14). The next metabolic step in -oxidation involves an LC-FAO enzyme. C. cloacae contains two LC-FAO genes, fao1 and fao2, cloned as part of this study. The fao1 gene is much more highly expressed than the fao2 gene both in the absence of alkane and following induction. Gene disruption studies should permit the generation of a strain of C. cloacae that can utilize fatty acids as substrates and produce -hydroxy fatty acids. This creates the possibility of making a wide range of -hydroxy fatty acids, which are at present unavailable in either natural or synthetic form, by growing C. cloacae on a variety of both saturated and unsaturated fatty acids (11). The limiting factor is the availability of a transformation system. There is one for C. tropicalis based on URA3 selection (32), but at present, none is available for C. cloacae.
Availability of a Candida strain disrupted in the fao genes would also allow complementation studies with the unassigned Arabidopsis and Mycobacterium genes identified in this paper. Their similarity to the data presented in this study suggests they are part of an -oxidation pathway. Complementation could confirm their proposed cellular function as LC-FAOs. Such complementation studies using plant genes in both yeast and bacterial systems have been successful (33,34). Since there are no identified long-chain alcohol oxidase genes in Saccharomyces or bacteria, generation of a null mutant in Candida would provide an ideal test system for the direct isolation of alcohol oxidases from diverse species, as has been used for acyltransferases (34,35).