Yas3p, an Opi1 Family Transcription Factor, Regulates Cytochrome P450 Expression in Response to n-Alkanes in Yarrowia lipolytica*

In the alkane-assimilating yeast Yarrowia lipolytica, the expression of ALK1, a gene encoding cytochrome P450 that catalyzes the first step of n-alkane oxidation, is induced by n-alkanes. We previously demonstrated that two basic helix-loop-helix proteins, Yas1p and Yas2p, activate the transcription of ALK1 in an alkane-dependent manner by forming a heterocomplex and binding to alkane-responsive element 1 (ARE1), a cis-acting element in the ALK1 promoter. Here we identified an Opi1 family transcription factor, Yas3p, involved in the alkane-dependent transcription regulation of ALK genes. Deletion of YAS3 caused a significant increase in ALK1 mRNA in cells grown on glucose, glycerol, and n-alkanes. The YAS3 deletion also resulted in a marked elevation of reporter gene expression driven by an ARE1-containing promoter on glycerol and n-decane. Bacterially expressed Yas3p bound specifically to Yas2p, but not to Yas1p, in vitro. In addition, although green fluorescent protein-tagged Yas3p was localized in the nucleus in glucose-containing medium, it changed its localization to an endoplasmic reticulum-like compartment upon transfer to medium containing n-decane. These findings suggest that Yas3p functions as a master regulator of transcriptional response, which changes its localization between the nucleus and endoplasmic reticulum membrane in response to different carbon sources. Furthermore, quantitative real time PCR analysis of 12 ALK genes in YAS1, YAS2, and YAS3 deletion mutants suggested that Yas3p is involved in the transcriptional repression of a variety of ALK genes, including ALK1. In contrast, YAS3 deletion did not affect the mRNA level of an INO1 ortholog in Y. lipolytica, indicating functional diversity of Opi1 family transcription factors.

Cytochromes P450s (P450s) 4 are heme-containing monooxygenases that catalyze diverse reactions in the metabolism of various endogenous and exogenous lipophilic compounds. P450s constitute a superfamily and have been isolated from a wide variety of species ranging from bacteria to mammals (1). Many microorganisms, including bacteria, yeasts, and fungi, have developed metabolic systems to assimilate n-alkanes as a carbon source using P450s (2)(3)(4). The cytochrome P450ALKs, which are classified into the CYP52 family, are involved in the terminal mono-oxygenation of n-alkanes. P450ALKs have been found in many alkane-assimilating yeasts, including Candida tropicalis (5)(6)(7), Candida maltosa (8,9), and Yarrowia lipolytica (10,11). The expression of many P450ALKs in these yeasts is induced when they are cultivated in the presence of n-alkanes as a sole carbon source and repressed when grown on glucose or glycerol. However, the molecular mechanisms underlying the transcriptional regulation of these P450ALKs are not fully understood.
Y. lipolytica has a remarkable ability to utilize a wide variety of hydrophobic compounds, and this makes Y. lipolytica potentially important both in fundamental research and in biotechnological applications (12,13). In a previous study, we isolated eight ALK genes (ALK1 to ALK8) encoding P450ALKs in Y. lipolytica (10,11). Furthermore, the existence of four additional ALK genes was suggested from its genome information (13)(14)(15). Gene deletion analysis was performed for ALK1, ALK2, ALK3, ALK4, and ALK6 (11). The deletion of ALK1 caused a defect in growth on medium containing n-decane, a short carbon chain n-alkane, as a sole carbon source, but it did not confer an apparent growth defect on n-hexadecane, a longer chain n-alkane. The double deletion of ALK1 and ALK2 resulted in growth defect on n-hexadecane. These results suggest the vital role of ALK1 and ALK2 in n-alkane assimilation (11). However, the functions of other ALK genes remain to be elucidated.
The expression of ALK1 is highly induced by n-alkanes, including n-decane and n-hexadecane, and repressed strongly by glycerol and weakly by glucose (11). We identified an upstream activating sequence that provides the transcriptional induction in response to n-alkanes in the ALK1 promoter, and we named it alkane-responsive element 1 (ARE1) (16,17). ARE1 contains an E-box-like sequence, and ARE1-like sequences are found in the promoters of several genes involved in n-alkane assimilation, including a subset of ALK genes and PAT1 encoding acetoacetyl-CoA thiolase (18). Recently, we identified YAS1 and YAS2 encoding basic helix-loop-helix (bHLH) transcription factors essential for the transcriptional induction of ALK1 by n-alkanes (17,19). Deletion of YAS1 or YAS2 abolishes the n-alkane-dependent transcriptional induction of ALK1 and impairs the growth on n-alkanes. bHLH proteins are known to form homo-and/or heterodimers and to bind to E-box motifs (20 -22). Yas1p and Yas2p also form a heterocomplex in vivo and in vitro and bind to ARE1 in vitro only when both are present (19).
The bHLH motifs in Yas1p and Yas2p share highest sequence similarity with those of Ino4p and Ino2p, respectively, in the budding yeast Saccharomyces cerevisiae. The heterodimer complex formed by Ino2p and Ino4p binds to a DNA element called UAS INO /ICRE, which is found in the promoters of phospholipid biosynthetic genes, and this induces the expression of the genes in the absence of myo-inositol (23)(24)(25)(26)(27)(28). Transcription of these phospholipid biosynthetic genes is repressed by a negative regulator, Opi1 (29). According to the model proposed by Loewen et al. (30), Opi1p is sequestered on the ER membrane through interactions with the integral membrane protein Scs2 and with phosphatidic acid (PA) in the absence of myo-inositol. When myo-inositol is provided, PA is consumed for synthesis of phosphatidylinositol. Upon the decrease of PA in the ER membrane, Opi1p translocates into the nucleus, binds to Ino2p (31), and represses the Ino2p-Ino4p-dependent transcription.
In this study, we identified a Y. lipolytica ortholog of Opi1p, which we named Yas3p (yeast alkane signaling 3), and we investigated its function in n-alkane-dependent transcriptional regulation. The results suggest that Yas3p regulates the transcription of ALK1 and a subset of other ALK genes in response to n-alkanes through interaction with Yas2p. Importantly, our data provide evidence for the functional diversity of Opi1 family proteins that are widely and specifically encoded in fungal genomes.

EXPERIMENTAL PROCEDURES
Yeast Strains and Growth Condition-Y. lipolytica strain CXAU/A1 (ura3, ade1::ADE1) (17) was used as a wild-type strain. The ⌬yas1 and ⌬yas2 strains were described previously (17,19). The ⌬yas3, ⌬lyas3N, and ⌬lyas3 strains were obtained by replacing the second exon, the first exon, and the whole coding region of the long form of YAS3, respectively, with the ADE1 gene. The deletion cassettes (described below) were introduced into CXAU1 strain (ura3, ade1) (10), and Ade ϩ transformants were selected. To obtain a strain that expresses Yas3p fused to EGFP (Yas3-EGFP) or Sec61p fused to DsRed (Sec61-DsRed) under the native promoter from its own chromosomal locus, pBSUYAS3-EGFP or pBSUSEC61-DsRed was digested with PshAI or BamHI, respectively, and introduced. Transformants were selected, and then the cells, in which the DNA region corresponding to intact YAS3 or SEC61, URA3, and the vector was popped out, were selected on YPD medium containing 1 mg/ml 5-fluoroorotic acid (32). The correct integration or recombination was confirmed by Southern blot analysis and/or DNA sequencing.
An appropriate carbon source was added to YNB (0.17% yeast nitrogen base without amino acids and ammonium sulfate; Difco, 0.5% ammonium sulfate) as follows: 2% (w/v) glucose; 2% (w/v) glycerol; 2% (v/v) n-decane; and 2% (v/v) n-hexadecane. Uracil (24 mg/liter) was added, if necessary. For solid media, 2% agar was added. n-Alkanes were supplied in the vapor phase to YNB solid media; a piece of filter paper was soaked with n-alkanes and placed on the lid of a Petri dish, which was sealed and kept upside down. SD medium with or without inositol was prepared as described previously (19). Yeast cells were grown at 30°C. The growth curve was obtained with an automatically recording incubator TN1506 (Advantec).
Plasmids-The short form of the YAS3 gene with its flanking region from CXAU1 total DNA was PCR-amplified with primers 5Ј-GCGATTGGTGATTAGCG-3Ј and 5Ј-GTGGCCAAT-CAGACGC-3Ј. The amplified fragment was cloned into the pGEM-T vector (Promega) to obtain pT-YAS3. The short form of YAS3 gene with its flanking region was also PCR-amplified from CXAU1 total DNA with primers 5Ј-ACCCGAAGCTTT-TCGAGGTGCCCATTCAGAT-3Ј and 5Ј-AGCGCCGCGGT-CCAGCCGTTGTCTTGTCA-3Ј. The amplified fragment was cloned into the HindIII-SacII site of the pSUT5 vector (18) to obtain pSUT5-YAS3.
The deletion cassette for the second exon of the long form of YAS3 was constructed as described below. The plasmid pT-YAS3 was digested with EcoRI and EcoRV to remove the second exon of YAS3, blunted, and ligated with the ADE1-carrying BamHI fragment (filled) of pSAT4 to obtain pT-dyas3. The YAS3 deletion cassette was liberated by digestion of pT-dyas3 with ApaI and SpeI.
The deletion cassette for the first exon of the long form of YAS3 was constructed as follows. The promoter region of the long form of YAS3 and the 3Ј-adjacent region of the first exon were amplified by PCR using primers 5Ј-CAGAATTCCACG-TGACTTCATGCGGGGTAGAAC-3Ј and 5Ј-GCTGTGGT-ACTCTCGTTCCCGT-3Ј, and 5Ј-TTGGATCCGGAAACG-ACGACCGGTGAGGA-3Ј and 5Ј-GGTCTAGATATGTGT-GGGCAGGGAACTTAG-3Ј, respectively. The amplified fragments were cloned into pBluescript II SK(ϩ) (Stratagene) after digestion with EcoRI or XbaI and BamHI. The ADE1carrying BamHI fragment from pSAT4 was cloned into BamHI site of this plasmid to obtain pBSdlyas3N. The deletion cassette for the first exon of YAS3 was liberated by digestion of pBSdlyas3N with EcoRV and XbaI.
The plasmid, pBSdlyas3, containing the deletion cassette for the whole region of the long form of YAS3 was constructed similarly to pBSdlyas3N, except that the 3Ј-noncoding region of YAS3 amplified by PCR using primers 5Ј-TTGGATCCTGTG-GGTGCTCCAGAAGATGGCT-3Ј and 5Ј-GGTCTAGAGG-CGGTAACAGTGTACTCCTGGT-3Ј was cloned into pBluescript II SK(ϩ), instead of 3Ј-adjacent region of the first exon. To express His 6 -tagged Yas1p and Yas2p, we used pET-YAS1 and pET-YAS2, respectively (19). To obtain pGEX-s-YAS3 for expression of GST-fused s-Yas3p, a DNA fragment was PCRamplified with the primer pair 5Ј-CATGGATCCATGCCCAA-GGCTCTGTCTTC-3Ј and 5Ј-TAAGGATCCTTAGGCATC-TCCCATCTCAA-3Ј using pSUT5-YAS3 as a template, cut with BamHI, and cloned into the corresponding site in pGEX-4T-3 (GE Healthcare). For the lacZ reporter assay, pS3ϫLZ, which contains lacZ under the control of three copies of ARE1 and LEU2 minimal promoter, was used (17).
The plasmid, pSYAS2-EGFP, to express Yas2p fused to EGFP at its C terminus (Yas2-EGFP) from its own promoter was constructed as follows. The YAS2 ORF with its 5Ј-flanking region was amplified from CXAU1 total DNA by PCR using primers 5Ј-TCGACCAAGCTTCGATCTCCGTTATGTCCG-3Ј and 5Ј-TTGTGGGAATTCCTCATCAATCTTGGGAGG-3Ј, and the amplified fragment was digested with HindIII and EcoRI. The YAS2 3Ј-flanking region was also amplified by PCR using primers and 5Ј-AAGATTGAATTCTAAATGCCACAATCA-CCC-3Ј and 5Ј-TGAAGCTCTAGATACTGTGTGGGACAA-CTG-3Ј, and the amplified fragment was digested with EcoRI and XbaI. These fragments were cloned into pSUT5, which had been cut with HindIII and XbaI, resulting in pSYAS2-EcoRI. The GFP coding region was amplified from pEGFP (Clontech) by PCR using primers 5Ј-GGAATTCATGGTGAGCAAGGG-CGAGGAG-3Ј and 5Ј-GGAATTCCTTGTACAGCTCGTCC-ATGCC-3Ј. The amplified fragment was digested with EcoRI and cloned into the corresponding site of the pSYAS2-EcoRI yielding the plasmid pSYAS2-EGFP.
To obtain a strain that expresses Yas3-EGFP under its own promoter, we constructed the plasmid, pBSUYAS3-EGFP, as follows. The short form of YAS3 ORF with its 5Ј-flanking region was amplified from CXAU1 total DNA by PCR using primers, 5Ј-ACCCGAAGCTTTTCGAGGTGCCCATTCAGAT-3Ј and 5Ј-CGAATTCGGCATCTCCCATCTCAACGT-3Ј, and the amplified fragment was digested with HindIII and EcoRI. The YAS3 3Ј-flanking region was also amplified by PCR using primers 5Ј-GGAATTCCGCTCTAGAGCGGCTGGCATGACAC-GACAAG-3Ј and 5Ј-AGCGCCGCGGTCCAGCCGTTGTCT-TGTCA-3Ј, and the amplified fragment was digested with XbaI and SacII. The GFP coding region was amplified from pEGFP by PCR using primers 5Ј-GGAATTCATGGTGAGCAAGGG-CGAGGAG-3Ј and 5Ј-GCTCTAGATTACTTGTACAGCTC-GTCCAT-3Ј, and the amplified fragment was digested with EcoRI and XbaI. These DNA fragments were cloned into the HindIII-SacII sites of the pSUT5 to obtain pSYAS3-EGFP. URA3-carrying fragment was excised from pSUT5 with NcoI, blunted, and cloned into EcoRV site of pBluescript II SK(ϩ) to obtain pB-URA3. YAS3-EGFP-containing fragment was amplified from pSYAS3-EGFP by PCR using primers 5Ј-GAAGGATCCTTCGAGGTGCCCATTCAGAT-3Ј and 5Ј-GGAGGATCCTCCAGCCGTTGTCTTGTCA-3Ј. The amplified fragment was digested with BamHI and cloned into the corresponding sites of pB-URA3 to obtain pBSUYAS3-EGFP.
To obtain a strain that expresses Sec61-DsRed under its own promoter, we constructed the plasmid, pBSUSEC61-DsRed, as follows. The SEC61 ORF with its 5Ј-flanking region was amplified from CXAU1 total DNA by PCR using primers 5Ј-GGAATTCGCTCCACCTACAAGTAGC-CCA-3Ј and 5Ј-AAGGCCTAATAGCAGCATCAACGTA-GCC-3Ј, and the amplified fragment was digested with EcoRI and StuI. The SEC61 3Ј-flanking region was also amplified by PCR using primers and 5Ј-AAGGCCTTAATT-GCGCTAGGGTGTAACTGA-3Ј and 5Ј-GCTCTAGAGC-GAGTCATCTTCAAACCATCCA-3Ј, and the amplified fragment was digested with StuI and XbaI. These DNA fragments were cloned into the HindIII-XbaI site of the pBluescript II SK(ϩ) to obtain pBSSEC61. The coding region of DsRed was amplified from the pDsRed monomer (Clontech) by PCR using primers 5Ј-TCCCCCGGGATGGACAACACCGAGGACGT-CATCAAG-3Ј and 5Ј-TCCCCCGGGCTGGGAGCCGGAG-TGGCGGGCCTCGGC-3Ј, and the amplified fragment was digested with SmaI and cloned into the StuI site of the pBSSEC61 to obtain pBSSEC61-DsRed. The plasmid pBSSEC61-DsRed was digested with ApaI and XbaI, and SEC61-DsRed-containing fragment was cloned into the corresponding site of pSAT4 to obtain pSASEC61-DsRed. SEC61-DsRed-containing fragment was excised from pSASEC61-DsRed with EcoRI and XbaI and cloned into the corresponding sites of pB-URA3, to obtain pBSUSEC61-DsRed.
After PCR amplification, the DNA fragments were checked by sequence analysis. Restriction enzyme sites in the primers were indicated with underlines.
Transformation of Y. lipolytica-Y. lipolytica was transformed by electroporation as described previously (10).
␤-Galactosidase Activity Assay-We performed ␤-galactosidase activity assay as described (16), except that the cells were incubated with various carbon sources for 3 h.
For binding assay of Yas1p and Yas2p to Yas3p immobilized on the beads, cell lysate (ϳ1.8 mg of protein) containing GSTs-Yas3p was incubated with 10 l of glutathione-Sepharose beads (GE Healthcare) for 2 h at 4°C with constant rotation. Cell lysate (ϳ1.0 mg of protein) containing His 6 -tagged proteins was added to the beads and incubated overnight at 4°C. The beads were washed five times with 1 ml of lysis buffer. For binding assay of Yas1p to the preformed Yas2p-Yas3p complex, cell lysate (ϳ700 g of protein) containing GST-s-Yas3p was incubated with 10 l of glutathione-Sepharose beads for 2 h at 4°C. Cell lysate (ϳ400 g of protein) containing His 6 -Yas2p was added to the beads and incubated for an additional 4 h at 4°C. The beads were washed as described above. Then cell lysate (ϳ400 g of protein) containing His 6 -Yas1p was added to the beads and incubated overnight at 4°C. The beads were washed as described above. Bound proteins were eluted by boiling in 50 l of SDS/sample buffer (125 mM Tris-Cl, pH 6.8, 10% glycerol, 2% SDS, 1% 2-mercaptoethanol, 0.0005% bromphenol blue), and 10 l were subjected to SDS-PAGE followed by immunoblotting. His 6 -tagged proteins were detected with antipoly Histidine monoclonal antibody (R & D Systems) at 1:5,000. Horseradish peroxidase-conjugated secondary antibody (Cell Signaling Technology) at 1:5,000 and enhanced chemiluminescence reagents (GE Healthcare) were used to visualize the resolved proteins.
Fluorescence Microscopy-For DAPI staining, cells were collected, washed with phosphate-buffered saline, and fixed in 70% ethanol for 15 min. Cells were washed twice and stained in 1 g/ml DAPI for 10 min at room temperature. Cells were washed twice and subjected to microscopic observation. Microscopic images were acquired a BX52 microscope (Olympus, Tokyo, Japan) equipped with ORCA-ER (Hamamatsu Photonics, Hamamatsu, Japan).
Quantitative Real Time PCR (qRT-PCR)-Total RNA was prepared using ISOGEN (Nippon Gene) and treated with DNase I (Takara) to remove genome DNA. The total RNA was reverse-transcribed with ExScript RT reagent kit (Takara). qRT-PCR was performed using SYBR Premix Ex Taq (Takara) and the gene-specific primers (Table 1) with SmartCycler II (Cepheid). To quantify the mRNA expression levels, standard curves were prepared using CXAU1 total DNA as a template. Specific amplification of each ALK gene was confirmed by sequence analysis of five clones amplified using corresponding primers.

RESULTS
Identification of Yas3p-The bHLH motifs of Yas1p and Yas2p share the highest sequence similarity with those of S. cerevisiae Ino4p and Ino2p, respectively. Opi1p is involved in the transcriptional repression of phospholipid biosynthetic genes controlled positively by Ino2p and Ino4p. These observations suggest the possibility that an Opi1p counterpart is also conserved in Y. lipolytica and that it is involved in the transcriptional regulation of ALK1. We searched the Y. lipolytica genome data base for an Opi1p ortholog using the BLAST pro-gram provided by the Génolevures project (available on line) (14,15,33) using the full-length amino acid sequence of Opi1p as a query. The highest score, 86.0 bits, was calculated for the YALI0C14784g translation product, and we named this gene YAS3. In the Y. lipolytica genome data base, it is predicted that the YAS3 is 2,388 bp in length with a 525-bp intron and that it encodes a 620-amino acid protein. However, our RLM-RACE analysis using RNA isolated from cells incubated on glucose suggested that there exist two types of YAS3 transcripts. The shorter transcripts initiate in the region annotated as the second exon of YAS3 and do not have an intron. YAS3 ORF in these transcripts are predicted to start at 1,119 bp downstream of the annotated start codon and to encode a 422-amino acid protein, which we designated the short form of Yas3p, s-Yas3p (Fig. 1A). The longer transcripts start in the region annotated as the 5Ј-noncoding sequence of YALI0C14762g, a neighboring ORF, have a 1,600-bp intron, and are predicted to encode a 727amino acid protein (Fig. 1A). We designated this translation product as the long form of Yas3p, l-Yas3p. The N-terminal 153 amino acids of the l-Yas3p is derived from YALI0C14762g. In RLM-RACE analysis using RNA isolated from cells incubated on n-decane for 1 h, the DNA fragments derived from the shorter transcripts of YAS3 were predominantly amplified. The deduced amino acid sequence of l-and s-Yas3p have 13.2 and 19.1% identity and 35.5 and 54.3% similarity to that of Opi1p, respectively. A significant similarity between Yas3p and Opi1p was observed in three regions as follows: a leucine zipper domain (amino acids 139 -160 in Opi1p) that partially overlaps with the PA binding domain, a central uncharacterized domain (amino acids 216 -273), and the activator interaction domain (AID) that is involved in the interaction with Ino2p (amino acids 321-380) (Fig. 1A). A PSORT II search also predicted a leucine zipper motif in Yas3p. However, no significant similarity was observed in the N-terminal portion of the PA binding domain of Opi1p. Yas3p does not have a short motif containing two phenylalanines in an acidic tract (FFAT) motif that is

ORF
Primer Sequence important for Opi1p to bind to the ER membrane protein Scs2p. Yas3p also does not have glutamine-rich regions that are characteristic features of Opi1p.
Opi1p orthologs or Opi1p-like proteins are also encoded in the genomes of diverse yeasts and filamentous fungi among the ascomycetes, zygomycetes, and basidiomycetes (Fig. 1B), including Candida albicans and Kluyveromyces lactis in which Opi1p orthologs were reported by Heyken et al. (34). These Opi1 family proteins can be classified into several subfamilies. Yas3p was classified in a subfamily distinct from those including the Opi1 orthologs of S. cerevisiae, C. albicans, and K. lactis.
Expression of ALK1 in the ⌬yas3 Mutant-To elucidate the function of YAS3, we constructed a YAS3 deletion mutant (⌬yas3), in which the second exon of the long form of YAS3 was deleted ( Fig. 2A). ⌬yas3 cells grew more slowly on glucose and n-hexadecane than the wild-type cells and did not grow on n-decane (data not shown). The reason for the growth defect of ⌬yas3 cells on n-decane is currently under investigation. One possibility is that the excessive accumulation of intermediate metabolites of n-decane caused by overproduced P450ALKs (see below) might impair the growth of the ⌬yas3 cells. Alternatively, the transcription of as yet unidentified genes required for n-decane assimilation might be diminished in ⌬yas3 cells.
We performed Northern blot analysis using the wild-type and ⌬yas3 cells to assess whether the YAS3 gene is involved in the transcriptional regulation of ALK1 (Fig. 2B). The yeast cells were cultured in glucose-containing medium and transferred to medium containing glucose, glycerol, n-decane, or n-hexadecane. After 1 h of incubation, the transcript levels of ALK1 were increased on n-decane and n-hexadecane in the wild-type cells but not on glucose or glycerol (Fig. 2B). In contrast, ALK1 mRNA was significantly increased in the ⌬yas3 cells on all carbon sources (Fig. 2B). Analysis of reduced CO difference spectra also revealed that ⌬yas3 cells produced more P450s than the wild-type cells on both glycerol and n-decane (data not shown).
The transcription of ALK1 is activated by the Yas1p and Yas2p heterocomplex via ARE1 (17,19). We analyzed the expression of a reporter gene under the control of an ARE1containing promoter in ⌬yas3 cells (Fig. 2C). ⌬yas3 cells harboring pS3xLZ, which carries three copies of ARE1 upstream of the minimal Y. lipolytica LEU2 promoter directing the expression of the lacZ gene (17), were precultured in glucose-contain- ing medium, transferred to medium containing glycerol or n-decane, and incubated for 3 h. In the wild-type cells, the ARE1-containing promoter was inactive on glycerol but activated in the presence of n-decane. In contrast, the reporter activities were greatly elevated in extracts of the ⌬yas3 cells grown on both glycerol and n-decane, indicating that the AREcontaining promoter was hyperactive in ⌬yas3 cells independent of carbon source (Fig. 2C). These results suggest that Yas3p represses the expression of ALK1 via ARE1.
mRNA Levels of YAS1, YAS2, and YAS3 in Yas Mutants-Next, we analyzed the expression of YAS3 on each carbon source. Transcripts of YAS3 were detected on all carbon sources we tested (Fig. 3A). Consistent with the results of RLM-RACE analysis, the YAS3 transcripts of lower mobility, which probably encodes the l-Yas3p, were observed in the RNA pre-pared from cells incubated on glucose but not in that on glycerol, n-decane, or n-hexadecane. The levels of the smaller YAS3 transcripts, which probably encode the s-Yas3p, were markedly increased on n-decane and n-hexadecane compared with glucose and glycerol in the wild-type cells but not in ⌬yas1 or ⌬yas2 cells (Fig. 3A), suggesting that the transcription of the shorter form of YAS3 is positively regulated by Yas1p and Yas2p. Interestingly, the YAS3 transcripts of lower mobility were detected in the RNA prepared from ⌬yas1 or ⌬yas2 cells incubated on n-alkanes.
We have previously demonstrated that Yas1p also binds to its own promoter and that the expression of YAS1 is induced by n-alkanes (17). In this study, we analyzed the levels of YAS1 and YAS2 mRNA in ⌬yas3 cells. Whereas the YAS1 and YAS2 mRNA levels were low in the wild-type cells on glucose and glycerol, they were elevated in ⌬yas3 cells (Fig. 3, B and C), implying that the transcription of YAS1 and YAS2 is repressed by Yas3p. Taken together, these results suggest that the transcription activator complex Yas1p-Yas2p and the transcription repressor Yas3p mutually regulate their own transcription.
Functions of Two Forms of Yas3p-To illuminate the functional difference of l-Yas3p and s-Yas3p, we constructed a deletion mutant, ⌬lyas3N, in which the first exon of the long form of YAS3 was deleted ( Fig. 2A). This mutant does not express the long form of YAS3, but it has a 1.6-kbp intact 5Ј-noncoding region of the short form of YAS3 and appears to express the short form of YAS3 normally on glucose, glycerol, and n-decane (Fig. 4A). In addition, we constructed another deletion mutant, ⌬lyas3, in which the whole region of the long form of YAS3 was deleted ( Fig. 2A). We examined the expression of ALK1 in these mutants. Any defect in the expression of ALK1 was not observed in ⌬lyas3N on glucose, glycerol, and n-decane (Fig.   FIGURE 2. The expression of ALK1 in ⌬yas3 cells. A, deletion mutants constructed in this study. The second exon, the first exon, and the whole region of the long form of YAS3 were replaced with ADE1 in ⌬yas3, ⌬lyas3N, and ⌬lyas3, respectively. B, mRNA levels of ALK1 in ⌬yas3 cells. Cells were grown in glucose-containing medium, transferred to minimal medium containing glucose (Glc), glycerol (Gly), n-decane (C10), or n-hexadecane (C16), and cultured for 1 h. Total RNA was extracted and subjected to Northern blot analysis using ALK1-specific probe. Ribosomal RNAs stained with ethidium bromide are shown as loading control. C, ARE1-dependent transcriptional regulation by Yas3p. Cells transformed with pS3ϫLZ were grown in glucose-containing medium, transferred to minimal medium containing glycerol or n-decane, and cultured for 3 h. Crude extracts from the cells were assayed for ␤-galactosidase activity. Results represent an average of three independent experiments Ϯ S.E. WT, wild type. 4B). Furthermore, the growth defect of ⌬lyas3N was not observed on these carbon sources (data not shown). In contrast, no difference in the expression of ALK1 and the growth between ⌬lyas3 and ⌬yas3, in which the second exon of the long form of YAS3 was deleted, was found ( Fig. 4B and data not  shown). These results indicate that l-Yas3p is dispensable for the proper transcriptional regulation of ALK1 and normal growth on these carbon sources, and suggest that the expression of the short form of YAS3 is sufficient in these regards.
Interaction of Yas3p with Yas2p-We analyzed the interaction of s-Yas3p with Yas1p and Yas2p by in vitro binding assays using bacterially expressed recombinant proteins. When GSTs-Yas3p was expressed in bacterial cells, it was found exclusively in the pellet fraction in the absence of 1% Triton X-100 but recovered in the soluble fraction in the presence of the detergent, suggesting that Yas3p associated with membrane components in E. coli (data not shown). Therefore, the preparation of crude extracts containing GST-s-Yas3p and in vitro binding assays was performed in the presence of 1% Triton X-100. Bacterially expressed GST-s-Yas3p was immobilized on glutathione-Sepharose beads and then incubated with the extract containing His 6 -Yas1p or His 6 -Yas2p. Bound protein complexes were eluted from the beads, resolved by SDS-PAGE, and immunoblotted with anti-His tag antibody. His 6 -Yas2p, but not His 6 -Yas1p, was co-purified with GST-s-Yas3p (Fig.  5A). Nonspecific interaction of His 6 -Yas2p with GST alone was not observed (data not shown). Furthermore, GST-Yas2 also bound to His 6 -s-Yas3p and His 6 -Yas1p (data not shown). These results suggest the specific interaction of Yas3p with Yas2p, but not with Yas1p, and raise the possibility that Yas3p represses the transcription of target genes by binding to Yas2p, a component of an activator complex. When GST-s-Yas3p was incubated with both His 6 -Yas2p and His 6 -Yas1p, interaction of GST-s-Yas3p with His 6 -Yas2p was somehow decreased (data not shown). However, His 6 -Yas1p was co-purified with the preformed complex of GST-s-Yas3p and His 6 -Yas2p immobilized on glutathione-Sepharose beads (Fig. 5B), suggesting that a ter-nary complex, including Yas1p, Yas2p, and Yas3p, can be formed.
Subcellular Localization of Yas2p and Yas3p-We previously demonstrated that Yas1p is localized in the nucleus independent of the presence of n-decane (17). In this study, we analyzed the localization of Yas2p and Yas3p in cells grown on glucose or n-decane. We investigated the localization of Yas2p tagged with EGFP (Yas2-EGFP) at its C terminus expressed under its own promoter from a low copy YCp plasmid. Expression of Yas2-EGFP complemented growth defects of ⌬yas2 strain on n-decane, indicating that Yas2-EGFP is functional (data not shown). Yas2-EGFP appeared intact in Y. lipolytica cells, because no degradation products were detected in Western blotting analysis of the extracts prepared from cells incubated on glucose or on n-decane using anti-GFP antibody (data not shown). Nuclear accumulation of Yas2-EGFP was observed in the cells grown on both glucose and n-decane, as revealed by simultaneous staining with DAPI, suggesting that Yas2p is localized in the nucleus independently of carbon source (Fig. 6A).
Next, we investigated the localization of Yas3p tagged with EGFP (Yas3-EGFP) at its C terminus. Yas3-EGFP was expressed under the native promoter from its own chromosomal locus in the cells grown on glucose or n-decane. The  strain expressing Yas3-EGFP instead of Yas3p grew normally on glucose and n-decane (data not shown). These data indicate that Yas3-EGFP is similarly functional as Yas3p. Yas3-EGFP appeared mostly intact as assessed by Western blotting analysis of the extracts prepared from cells incubated on glucose or n-decane using anti-GFP antibody, although a faint band was observed at ϳ25 kDa (data not shown).
When cells were incubated on glucose, nuclear localization of Yas3-EGFP was observed, as evidenced by simultaneous staining with DAPI (Fig. 6B). In contrast, when cells were shifted to n-decane, accumulation of Yas3-EGFP was observed in the ER-like compartment that was visualized by the fluorescent signal of DsRed fused to Sec61p (Sec61-DsRed), a component of the translocon complex in the ER membrane in addition to unidentified punctate structures (Fig. 6C) (35). Sec61-DsRed also appeared to be functional, because the strain expressing the fusion protein under its own promoter instead of native Sec61 grew normally. These results suggest that Yas3p changes its subcellular localization between the nucleus and ER in response to different carbon sources.
Expressions of the 12 ALK Genes in ⌬yas1, ⌬yas2, and ⌬yas3 Cells-We have previously isolated eight ALK genes (ALK1 to ALK8) in Y. lipolytica (10,11). We identified an additional four genes showing homology to the ALK genes YALI0B06248g, YALI0B20702g, YALI0C10054g, and YALI0A20130g through a Y. lipolytica genome data base search. We named these genes ALK9, ALK10, ALK11, and ALK12, respectively (Fig. 7). We analyzed the mRNA levels of these 12 ALK genes on glucose, glycerol, n-decane, and n-hexadecane using qRT-PCR with specific primers. The mRNA levels of the ALK genes varied by almost 4 orders of magnitude, and the amounts of many ALK mRNAs were increased on n-alkanes compared with those on glucose or glycerol (Fig. 8). To determine the involvement of Yas1p, Yas2p, and Yas3p in the transcriptional regulation of these ALK genes, we also examined their mRNA levels in ⌬yas1, ⌬yas2, and ⌬yas3 cells (Fig. 8). The transcript levels of ALK2, FIGURE 6. Subcellular localization of Yas3p on various carbon sources. A, localization of Yas2-EGFP to the nucleus in the cells cultured on glucose. ⌬yas2 cells harboring pSYAS2-EGFP were grown in glucose-containing medium from early to mid log phase, transferred to YNB containing n-decane, and cultured for 1 h. Localization of Yas2-EGFP was observed as described under "Experimental Procedures." B, localization of Yas3-EGFP to the nucleus in the cells cultured on glucose. Cells expressing Yas3-EGFP were grown in glucose-containing medium to early to mid log phase. Localization of Yas3-EGFP was observed. C, localization of Yas3-EGFP to the ER-like compartment in the cells cultured on n-decane. Cells co-expressing Yas3-EGFP and Sec61-DsRed were grown in glucose-containing medium, transferred to YNB containing n-decane, and cultured for 1 h. Localization of Yas3-EGFP and Sec61-DsRed was observed. DIC, differential interference contrast.

DISCUSSION
In this study, we identified a novel transcription factor, Yas3p, involved in the transcriptional regulation of ALK1 and probably a subset of other ALK genes in Y. lipolytica. Based on our results, we propose a model in which Yas3p is transported into the nucleus and represses the expression of ALK1 and other target genes by binding to Yas2p in the absence of n-alkanes, whereas Yas3p is trapped on the ER membrane, and consequently the Yas1p-Yas2p complex activates the transcription of target genes in the presence of n-alkanes.
Structural Features of Yas3p-Opi1p and Yas3p share significant sequence similarities in their leucine zipper domain, uncharacterized conserved domain, and AID (Fig. 1A). The AID of Opi1p and the repressor interaction domain of Ino2p FIGURE 8. The expression of 12 ALK genes. Cells were grown in glucose-containing medium, transferred to YNB containing glucose (black bar), glycerol (gray bar), n-decane (white bar), or n-hexadecane (hatched bar), and cultured for 1 h. The copy number of mRNA was determined by qRT-PCR using specific primers for each of the 12 ALK genes. Note that the mRNA levels vary by 4 orders of magnitude among the ALK genes. WT, wild type.
are involved in the interaction of these two proteins (31,37). A short stretch of amino acids with weak similarity to receptor interaction domain, which is widely conserved among Ino2 orthologs in various yeasts (37), was also found in Yas2p (amino acids 165-182) (19), suggesting that Yas3p interacts with Yas2p through these two domains. In contrast, the functional significances of the leucine zipper domain and the uncharacterized conserved domain remain to be determined. Because mutations in these two domains of Opi1p caused the constitutive overexpression of target genes (38), the domains may play important roles in the function or the structural integrity of Opi1p and Yas3p.
Functional Diversity of Opi1 Family Proteins-The Opi1 family proteins are widely encoded in the genomes of a diverse range of fungi for which genome information is currently available. To date, functional characterizations have been performed only for S. cerevisiae and C. albicans; the biological processes in which other Opi1 family members are involved remain unknown. Although it has been reported that OPI1 (CaOPI1) of C. albicans functionally complements the inositol/ choline-dependent gene regulation defect of the opi1 mutant of S. cerevisiae, it is not clear whether CaOpi1p is involved in the transcriptional regulation of phospholipid biosynthetic genes in C. albicans cells (34). Indeed, it has been reported that the orthologs of Ino2p and Ino4p in C. albicans regulate the transcription of ribosomal protein genes (39). In Y. lipolytica, Yas3p does not appear to be involved in the transcriptional regulation of YlINO1 (Fig. 9). Furthermore, deletion of neither YAS1 nor YAS2 confers inositol auxotrophy on Y. lipolytica (19). These data lead to the tempting hypothesis that Opi1 family proteins in a particular subfamily, including S. cerevisiae Opi1p, are engaged in the inositol response, but that those in other subfamilies are involved in the regulation of distinct biological processes that may also be implicated in the biosynthesis or the metabolism of lipophilic compounds in fungi.
Mechanisms for Transcriptional Regulation by Yas3p-Key unresolved questions concerning n-alkane-dependent tran-scriptional regulatory mechanisms in alkane-assimilating yeasts are how the presence of n-alkanes is sensed and how the transcription of target genes is regulated. Our findings raise the simple and intriguing possibility that Yas3p itself senses the nalkanes, their metabolites, or other membrane components and functions as a pivotal regulator of the transcriptional response to n-alkanes. The fact that bacterially expressed Yas3p was recovered from the pellet fraction after disruption of cells, but was solubilized in the presence of Triton X-100, also suggests the interaction of Yas3p with membrane components. However, the Ino2-Ino4-Opi1 system is not completely conserved in the Yas1-Yas2-Yas3 system. The FFAT motif, which is involved in the interaction of Opi1p with the ER membrane protein Scs2p, is not found in Yas3p (Fig. 1A). Consistent with this observation, the mRNA level of ALK1 was not appreciably affected by single or double deletion of YlSCS2 and/or YlSCS22 (40). This is in contrast to the expression of INO1, which was reduced in ⌬scs2 (41) and ⌬scs2⌬scs22 mutants. The latter double mutant exhibited a more severe inositol auxotrophy than the ⌬scs2 mutant in S. cerevisiae (42). Furthermore, localization of Yas3-EGFP in the ER-like compartment was observed in ⌬ylscs2 cells incubated on n-decane (data not shown). Therefore, regulatory mechanisms distinct from those in S. cerevisiae will be involved in the transcriptional repression and derepression of ALK genes by Yas3p in Y. lipolytica. The low similarity between Yas3p and Opi1p in the regions other than leucine zipper motifs, central uncharacterized regions, and C-terminal AID is consistent with this possibility. The mechanism for controlling the subcellular localization of Yas3p in response to different carbon sources is currently under investigation, particularly the interaction of Yas3p with n-alkanes, alkane derivatives, or other membrane components.
Alternatively, it is possible that n-alkanes are sensed by unknown mechanisms and that the generated signals are transduced to regulate the synthesis or the activity of Yas3p. Although the functional difference of the two forms of Yas3p remains obscure, the transcription of YAS3 appears to be regulated by carbon sources. Regulation of Yas3 activity by phosphorylation is also feasible, because protein kinase A, protein kinase C, and casein kinase II have been reported to phosphorylate Opi1p and to regulate its activity in S. cerevisiae (43)(44)(45).
Functions of Two Forms of Yas3p-Our results indicate that l-Yas3p is dispensable for the proper transcriptional regulation of ALK1 and normal growth on glucose, glycerol, and n-decane. Although we have not obtained a mutant that is specifically defective in the expression of the short form of YAS3 but expresses the long form of YAS3 at an equivalent level to the wild-type cells, results in this study suggests that s-Yas3p plays a pivotal roles in the transcriptional regulation of ALK1 and the growth on these carbon sources. Indeed, the shorter transcripts were prominent in glycerol-containing medium, in which the expression of ALK1 is strictly repressed, and in alkane-containing medium, in which ALK1 expression is highly induced (Fig.  3A). However, a significant amount of the longer transcripts was found in cells grown in the medium containing glucose. Furthermore, in our preliminary study, l-Yas3p tagged with GFP at its N terminus, which was expressed under the native promoter from low copy plasmid, was localized in the nucleus FIGURE 9. The mRNA levels of YlINO1 in Y. lipolytica. Cells were grown in SD medium with or without 200 M myo-inositol to early to mid log phase. Total RNA was extracted and subjected to Northern blot analysis using YlINO1specific probe. Ribosomal RNAs stained with ethidium bromide are shown as loading control. on glucose. Thus, l-Yas3p could have unidentified function, which is not essential for the transcriptional regulation of ALK1.
Expression of YAS1, YAS2, and YAS3 in Yas Mutants-The level of the shorter transcript of YAS3 was increased in the wild-type cells on n-alkanes but not in ⌬yas1 or ⌬yas2 cells. Conversely, compared with the wild-type cells, the mRNA levels of YAS1 and YAS2 in ⌬yas3 cells were elevated on glucose and glycerol. These results suggest the mutual regulation of the transcription of their coding genes by the activator proteins, Yas1p and Yas2p, and the repressor Yas3p. Indeed, ARE1-like sequences are found in the 5Ј-noncoding regions of YAS2 and in that of the short form of YAS3, in addition to that of YAS1 (17). In S. cerevisiae the expression of OPI1 is also regulated positively by Ino2p and Ino4p and negatively by Opi1p itself (46), and the expression of INO2 is down-regulated by Opi1p (47). Although the significance of the transcriptional interdependence of these genes also remains to be determined, it may provide the mechanisms that enable yeast cells to rapidly adapt to environmental carbon sources or myo-inositol availability.
Transcriptional Regulation of ALK Genes-With the exception of ALK1 and ALK2, the functions of most of the ALK genes found in the Y. lipolytica genome remain unclear. The qRT-PCR analysis revealed that the transcript levels of many ALK genes were elevated on n-alkanes, suggesting the possibility of their involvement in n-alkane assimilation. In particular, our qRT-PCR results suggested that the transcription of ALK2, ALK4, ALK6, ALK9, and ALK11 was activated on n-alkanes by the Yas1p-Yas2p complex and repressed by Yas3p, as for ALK1. In addition, the transcript levels of these ALK genes on n-alkanes are relatively higher compared with other ALK genes. Like ALK1 and ALK2 that are vital for growth on n-alkanes (11), these results suggest that ALK4, ALK6, ALK9, and ALK11 are also involved in n-alkane assimilation.
The transcript levels of ALK3 and ALK12, both of which were increased on n-alkanes, were decreased in ⌬yas3 cells, whereas no effect of deletion in YAS1 or YAS2 was observed. These results suggest the positive regulatory roles of Yas3p on the transcription of these genes. In this respect, it is interesting that Opi1 also appears to be involved in the positive regulation of a subset of genes that are responsive to inositol and choline, as suggested by microarray analysis (48), although the underlying molecular mechanism is unknown. Functional and structural analyses of Yas3p, in combination with the analysis of ALK gene promoter, will provide novel insights into the regulatory mechanisms mediated by the Opi1 family transcription factors.