INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

The lipid composition of cellular membranes is regulated to maintain membrane fluidity (22). A key enzyme involved in this process is the membrane-bound 9 fatty acid desaturase, which catalyzes the introduction of the initial double bond between the 9th and 10th carbons of palmitoyl-coenzyme A (CoA) and stearoyl-CoA (35). The correct ratio of saturated to monounsaturated fatty acids contributes to membrane fluidity. Alterations of this ratio have been implicated in various diseases including cardiovascular diseases, obesity, non-insulin-dependent diabetes mellitus, hypertension, neurological diseases, immune disorders, and cancer in mammals (35). In yeast, this ratio has been suggested to be related to the heat shock response, ethanol tolerance, and mitochondrial movement and inheritance (1, 8, 45). The regulation of the expression of 9 fatty acid desaturase is, therefore, of considerable physiological importance.

In Saccharomyces cerevisiae, 9 fatty acid desaturase is encoded by OLE1 (46). The steady-state level of OLE1 mRNA is regulated at the level of transcription and by mRNA stability, and both regulatory processes are affected by the presence of unsaturated fatty acids (UFAs) in the growth medium (6, 10, 20). Addition of exogenous UFA represses the transcription of OLE1 and promotes the decay of OLE1 mRNA. A fatty acid-regulated (FAR) element which is essential for UFA repression of OLE1 transcription under aerobic conditions has been identified by promoter deletion analysis (10). Simultaneous disruption of the two long-chain (C₁₄ to C₁₈) fatty acyl-CoA synthetase genes, FAA1 and FAA4, blocks incorporation of long-chain saturated fatty acids and UFAs such as oleic acid into cells and does not cause UFA repression of OLE1 transcription (9, 10). A putative fatty acid transport protein (FATP) with 33% homology to an adipocyte FATP (42) was identified (15). However, more-recent reports indicate that this protein is an acyl-CoA synthetase specific for very-long-chain fatty acids rather than a plasma membrane fatty acid transporter (9, 52). Therefore, a fatty acid transport protein and UFA signal transducers and responding transcriptional regulators have not yet been identified.

Since the desaturation reaction requires molecular oxygen (O₂) as an electron acceptor (35), it has been proposed that cells derepress expression of OLE1 under hypoxic conditions to allow more-efficient use of limiting O₂ (53). Well-established O₂-responsive repressor Rox1p is believed to repress OLE1 transcription under aerobic conditions because putative Rox1p-binding sequences are present in the OLE1 promoter (53). However, there had been no conclusive evidence that OLE1 transcription is indeed repressed by O₂ until the recent report by Kwast et al. (26), who showed that OLE1 transcription was induced when cells were shifted to anaerobic conditions.

In this report, we show that O₂ repression of OLE1 is mediated by a Rox1p-independent novel mechanism. Furthermore, we newly identified a 50-bp region essential for O₂ repression, designated the O2R element, and this element was also responsible for UFA repression of anaerobically derepressed OLE1 transcription despite not being included within the FAR element. Therefore, there appears to be a close connection between O₂ and UFA signals that act on the short 50-bp element in the OLE1 promoter.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Microorganisms and media. The S. cerevisiae strains used in this work are listed in Table 1. Escherichia coli strains TG1 (40) and DH5 (40) were used as hosts for the propagation and manipulation of plasmid DNA. Yeast cells were grown in nutrient medium (YPDA; i.e., nutrient high-P_i medium) (38) or YPDA supplemented with 1 mM oleic acid (Wako Chemicals, Osaka, Japan) and 1% (vol/vol) Triton X-100 (Wako Chemicals) (16). The Luria-Bertani medium for E. coli was as described previously (40).

Disruption of ROX1. A 1,109-bp fragment of ROX1 (nucleotides [nt] 2 to +1107) was amplified by PCR using chromosomal DNA of S. cerevisiae strain S288C (33) as a template and oligonucleotides 5'-GCGGATCCCAATGAATCCTAAATCCTCTACAC-3' and 5'-CGCTCGAGTCATTTCGGAGAAACTAGGC-3', corresponding to the nt 2 to +22 and +1107 to +1088 of ROX1, as forward and reverse primers, respectively. The PCR product was doubly digested with BamHI and XhoI and cloned into the BamHI-XhoI gap of pRS305 (44) to obtain plasmid p1513. A 1,148-bp fragment of ROX1 (nt 1121 to +27) was amplified by PCR using chromosomal DNA of S288C as a template and oligonucleotides 5'-CCAAGCTTCCATTGAGAAGGACAACATT-3' and 5'-CTGGATCCTTAGGTGTAGAGGATTTAGG-3', corresponding to the nt 1121 to 1102 and +27 to +7 of ROX1, as forward and reverse primers, respectively. The PCR product was doubly digested with HindIII and BamHI and cloned into the HindIII-BamHI gap of pUC18 (40) to obtain plasmid p1514. A 2.9-kb BamHI-ScaI fragment from p1513 and a 2.1-kb BamHI-ScaI fragment from p1514 were ligated to obtain plasmid p1515. A 525-bp BamHI-BglII fragment of p1515 containing the ROX1 open reading frame (ORF), corresponding to nt +28 to +523, was replaced with a 1.7-kb BamHI fragment containing LEU2 from YDp-L (4) to obtain plasmid p1516. The rox1::LEU2 disruptant, SH5128, was constructed by transformation of SH4041 with HindIII- and XhoI-digested plasmid p1516. The rox1::LEU2 disruption was verified by Southern analysis.

OLE1p-PHO5 fusions. Construction of plasmid p1166, which contains a 935-bp fragment upstream of OLE1 fused with the structural region of PHO5 encoding repressible acid phosphatase (rAPase; EC 3.1.3.2), was described previously (16). The activation and repression assay vector pRAV, containing the UAS_PHO84E-PHO84p-PHO5 reporter gene, was constructed previously (31). The UAS_PHO84E-PHO84p-PHO5 reporter gene consists of a 272-bp fragment upstream of PHO84 containing Pho4p binding site E fused with the PHO5 ORF. Plasmid p1785 (see Fig. 2), which has the 934 to 586 region of OLE1 (taking base A of the ATG start codon as +1) upstream of the UAS_PHO84E-PHO84p-PHO5 reporter gene was constructed as follows. A 349-bp region of OLE1 was amplified by PCR using p1166 as a template and oligonucleotides 5'-CTCGAATTCAGCTTTTCGTTTGCAGGTTT-3' and 5'-CTCAAGCTTAGTTAGTTTTTGGGCCACCG-3', corresponding to the nt 934 to 915 and 586 to 605 of OLE1, as the forward and reverse primers, respectively. The PCR product was doubly digested with EcoRI and HindIII and cloned into the EcoRI-HindIII gap of pRAV to obtain plasmid p1785. Plasmids p1787, p1781, p1783, p1860, p1862, and p1864, which each contain nt 585 to 456 (the FAR element), 455 to 307, 306 to 159, 455 to 406, 405 to 357, and 356 to 307 of OLE1, respectively, were constructed in a similar way to plasmid p1785 by inserting EcoRI-HindIII fragments containing each region amplified by PCR into the EcoRI-HindIII gap of pRAV. Plasmid p1872, which carries OLE1p lacking the O2R element (OLE1pO2R) fused with the structural region of PHO5 was constructed as follows. The 306 to 1 region of OLE1 was amplified by PCR using p1166 as a template and oligonucleotides 5'-CTCAAGCTTTTCTACGAGTCTTGCTCACT-3' and 5'-GCGGATCCTTTGTTGTAATGTTTTAG-3', corresponding to the nt 306 to 287 and 1 to 19 of OLE1, as the forward and reverse primers, respectively. The PCR product was doubly digested with HindIII and BamHI and cloned into the HindIII-BamHI gap of pSH39 (34) to obtain plasmid p1870. The 934 to 357 region of OLE1 was amplified by PCR using p1166 as a template and oligonucleotides 5'-CTCGAATTCAGCTTTTCGTTTGCAGGTTT-3' and 5'-CTCAAGCTTAAAGAAAGCTGCCGACTATG-3', corresponding to nt 934 to 915 and 357 to 376 of OLE1, as the forward and reverse primers, respectively. The PCR product was doubly digested with EcoRI and HindIII and cloned into the EcoRI-HindIII gap of p1870 to obtain plasmid p1872. Plasmid p1904, in which the 5'-GATAA-3' sequence in the O2R element is changed to 5'-ACGCC-3' by site-directed mutagenesis, was constructed as follows. Oligonucleotides 5'-AATTCCGGACGTTGAAACACTCAACAAACCGGCGTTAGTGCCCAACCAGGTGTGCA-3' and 5'-AGCTT GCACACCTGGTTGGGCACTAACGCCGGTTTGTTGAGTGTTTCAACG TCCGG-3', corresponding to nt 356 to 307 and 307 to 356 of OLE1, respectively, in which the 5'-GATAA-3' sequence (331 to 327) is changed to 5'-ACGCC-3' (underlined) were annealed and cloned into the EcoRI-HindIII gap of pRAV to obtain plasmid p1904. Plasmid p1858, which contains a 567-bp fragment upstream of ANB1 (27) fused with the PHO5 ORF, was constructed as follows. A 567-bp fragment of ANB1 (nt 567 to 1) was amplified by PCR using chromosomal DNA of S288C as a template and oligonucleotides 5'-CTCAAGCTTCCGGGAATTTTAGATTCAGG-3' and 5'-CTCGGATCCGTTTTAGTGTGTGAATGAAA-3', corresponding to nt 567 to 548 and 1 to 20 of ANB1, as the forward and reverse primers, respectively. The PCR product was doubly digested with HindIII and BamHI and cloned into the HindIII-BamHI gap of pSH39 to obtain plasmid p1858. All constructs were analyzed by sequencing the respective promoter regions. The resulting plasmids were digested with StuI and integrated into the URA3 locus of SH5143, SH4041, or SH5128 by transformation. Single-copy integration was confirmed by Southern analysis of genomic DNA digested with HindIII from the respective transformants.

Northern blot analysis. The preparation of RNA and Northern blot hybridization were performed as described previously (38). Cells were cultivated in 10 ml of YPDA medium to stationary phase at 30°C with vigorous shaking. The stationary-phase culture was inoculated into 30 ml of YPDA or YPDA containing oleic acid in a 100-ml Erlenmeyer flask to give an optical density at 660 nm (OD₆₆₀) of 0.1. The cultures were shaken at 30°C for aerobic growth. For anaerobic growth, the cultures were sealed with rubber stoppers and bubbled with pure nitrogen gas for 2 min to purge O₂ after inoculation or sampling, and shaken at 30°C. Anaerobic conditions were verified based on confirming the transcription of ANB1 (28), a well-known hypoxic gene. Total RNAs were prepared from the cells harvested in the logarithmic growth phase (OD₆₆₀ = 0.7 to 1.6). DNA fragments containing the OLE1 ORF (nt 9 to +1905 relative to ATG), the PHO5 ORF (nt 18 to +2116), and the ANB1 ORF (nt +1 to +455), amplified by PCR, and the 1.0-kb HindIII-XhoI fragment carrying ACT1 prepared from pYA301 (19) were labeled with ³²P as described previously (39) to generate DNA probes.

rAPase assay. Cells were cultivated in 10 ml of YPDA medium to stationary phase at 30°C with vigorous shaking. The stationary-phase culture (0.2 ml) was inoculated into 10 ml of YPDA medium or YPDA medium containing oleic acid. The cultures were shaken for 5 h at 30°C for aerobic growth or left to stand for 24 h at 30°C in a sealed Anaero Pack (Mitsubishi Gas Chemical Co., Tokyo, Japan) for anaerobic growth, and then rAPase activities were measured as described previously (48). The rAPase activities presented are the averages of at least three independent experiments.

Genetic and biochemical methods. S. cerevisiae and E. coli cells were transformed as described by Ito et al. (24) and Sambrook et al. (40), respectively. Yeast chromosomal DNA was prepared as described previously (23). Southern blot analysis and other DNA manipulations were performed using standard methods (40). Bacterial plasmid DNA was isolated by the alkaline lysis method (40). Nucleotide sequences were determined by the dideoxy chain termination method (41).

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

Transcription of OLE1 is repressed by O₂ in a Rox1p-independent manner. To clarify whether OLE1 transcription is indeed repressed by O₂, we investigated OLE1 transcription in cells cultivated aerobically or anaerobically. Northern blot analysis (Fig. 1A) showed that the levels of transcripts of OLE1 and OLE1p-PHO5 fusion gene are significantly increased under anaerobic conditions, compared to the levels under aerobic conditions, in the wild-type strain (Fig. 1A, lanes 1 and 3), indicating that OLE1 transcription is repressed by O₂. The same observation that OLE1 transcription is induced when cells are shifted to anaerobic conditions has been reported recently (26). In S. cerevisiae, a Rox1p-dependent mechanism is known to mediate transcriptional repression of various hypoxic genes by O₂ (53). Since the biosynthesis of heme requires O₂, heme accumulates under aerobic conditions and binds to Hap1p. Hap1p with bound heme acts as a transcriptional activator to activate ROX1 transcription. Rox1p binds to its recognition sites, 5'-YYYATTGTTCTC-3' (where Y represents pyrimidine) (3), upstream of anaerobically expressed genes such as ANB1 and forms a complex with the general repressors Tup1p and Ssn6p to repress target genes under aerobic conditions (53). As three putative binding sites for Rox1p are present in the OLE1 promoter (Fig. 1B), we disrupted ROX1 to determine whether OLE1 repression requires Rox1p. In the rox1 disruptant, ANB1 transcription was, as reported previously (53), derepressed under aerobic conditions (Fig. 1A, lane 2). However, OLE1 transcription was unexpectedly not derepressed (Fig. 1A, lane 2), indicating that Rox1p is not involved in the O₂ repression of OLE1. Although there are many hypoxic genes which have consensus binding sequences for Rox1p in their promoters (53), for some of them the function of these sequences has not been experimentally determined by, for example, deletion or mutation analysis of the consensus sequence or disruption of ROX1. Our results for OLE1 demonstrate that one must not decide on the operating mechanism only by the presence of a consensus sequence. What is the mechanism of O₂ repression of OLE1? Recently, another group has shown that the respiratory chain is involved in the anaerobic induction of OLE1 transcription and that cytochrome c oxidase is likely the hemoprotein sensor for O₂ (26). However, nothing is known about the signal transduction pathway downstream of the O₂ sensor. Identification of O₂-responsive transcriptional factors involved in Rox1p-independent O₂ repression should make it possible to elucidate the repression mechanism.

View larger version (33K): [in this window] [in a new window]   FIG. 1.   Transcription of OLE1 is repressed by O2 in a Rox1p-independent manner. (A) Northern analysis of the OLE1, OLE1p-PHO5, and ANB1 transcripts in rox1 disruptants. Total RNA samples were prepared from cells of the wild-type strain (SH5420) and rox1 disruptant (SH5421) cultivated aerobically (lanes 1 and 2) or anaerobically (lanes 3 and 4) in YPDA medium. Equal amounts of RNA (5 µg) were electrophoresed in a 1.5% agarose gel in the presence of formaldehyde, transferred to a nylon filter, blotted, and hybridized with probes consisting of 32P-labeled DNA fragments containing OLE1, ANB1, and PHO5 for detection of the transcripts of OLE1p-PHO5 or ACT1 as an internal control. The ANB1 probe also hybridized to the TIF51A transcript, which is not repressed by Rox1p (53). (B) Putative Rox1p binding sites in the OLE1 promoter. Numbers indicate positions relative to the first nucleotide of the initiation codon (+1). Lowercase letters, nucleotides that differ from the consensus Rox1p-binding sequence (3).

The 50-bp O2R element is sufficient and essential for O₂ repression of OLE1 transcription. To elucidate the Rox1p-independent O₂ repression mechanism of OLE1, we first identified the O₂ signal-responsive region in the OLE1 promoter. The OLE1 promoter was divided into three subregions, i.e., the region harboring the FAR element and the regions upstream and downstream of this region, and these regions were inserted into the region upstream of the PHO84p-PHO5 reporter gene in the activation and repression assay vector pRAV (31). The transcription of the PHO84p-PHO5 reporter gene is repressed when cells are cultivated under conditions whereby a sufficient amount of P_i is used, such as YPDA medium (31). The levels of expression of these reporter genes were measured as rAPase activities under aerobic and anaerobic conditions (Fig. 2). Control experiments with the ANB1p-PHO5 reporter gene (p1858) validated this reporter assay system to monitor O₂-mediated transcriptional repression. The rAPase activity from p1166 was derepressed 4.4-fold under anaerobic conditions compared with aerobic conditions. Although p1787, which contains a FAR element, was not as responsive to anaerobic conditions (rAPase activity was derepressed 1.4-fold), p1781, containing the downstream region of the FAR element, responded significantly (rAPase activity was derepressed 9.7-fold). In order to delineate the O₂-regulated element in the 149-bp OLE1p region of p1781, we further divided this region into three subregions and determined which region responds to anaerobic conditions. Only p1864 showed significant derepression of rAPase activity (7.3-fold) under anaerobic conditions, indicating that the 50-bp OLE1p region (nt 356 to 307) is sufficient for anaerobic derepression. Deletion of the 50-bp region from OLE1p (p1872) decreased the level of anaerobic derepression from 4.4- to 1.4-fold, indicating that the 50-bp region is essential for anaerobic derepression of OLE1 transcription. We therefore designated the 50-bp region as an O₂-regulated element (O2R). That the O2R element does not contain consensus Rox1p-binding sites is consistent with a Rox1p-independent mechanism for O₂ repression of OLE1 transcription (Fig. 2). These results indicate that the O2R element is an anaerobic upstream activation site and further imply that some anaerobically responsive activators acting on the element exist. Thus, the system for O₂ repression of OLE1 transcription appears to be similar to the mammalian HIF-1 hypoxia-sensing system, whose regulation is mediated by transcriptional activator HIF-1 (51). On the other hand, the Rox1p-dependent O₂ repression system is not parallel to the HIF-1 hypoxia-sensing system because it is mediated by transcriptional repressor Rox1p, which binds to the upstream repression site of O₂-repressed genes (53). It is noted that a homologue of HIF-1 is not found in S. cerevisiae (21) and that the O2R element does not contain the consensus core sequence, 5'-RCGTG-3', of HIF-1 binding sites (43).

View larger version (28K): [in this window] [in a new window]   FIG. 2.   The O2R element is present in the 356 to 307 region of the OLE1 promoter. rAPase activity in cells of the wild-type strain (SH5143) harboring the respective reporter genes cultivated aerobically (+O2) or anaerobically (O2) in YPDA medium was measured as described previously (48). Numbers indicate the position relative to the first nucleotide of the initiation codon (+1) of OLE1. PHO84 promoter sequences (nt 272 to 1) contain binding site E (nt 262 to 257) for the transactivator Pho4p and two putative TATA boxes (nt 122 and 99). Shaded boxes, structural regions of PHO5. The ANB1p-PHO5 reporter gene was used to verify anaerobic conditions. Error bars, standard deviations determined from a minimum of three independent measurements. The actual values of rAPase activity with their standard deviations are indicated. Other symbols are the same as those described for Fig. 1B.

5'-GATAA-3' sequence in the O2R element is important for O₂ repression. What is the transcriptional activator acting on the O2R element? We searched for binding motifs of already-known transcriptional regulators on the O2R element and found the 5'-GATAA-3' sequence, which is the binding site for GATA family transcriptional activation factors Gln3p and probably Gat1p (Fig. 3A) (5, 13). Although the 5'-GATAA-3' sequence is on the complementary strand of the O2R element, it is known that the 5'-GATAA-3' sequence functions in an orientation-independent manner (12). The GATA family of DNA-binding proteins are present in organisms from S. cerevisiae to humans (29, 30). These binding proteins contain one or more characteristic C₄ zinc finger motifs and bind to DNA sequences with sequence GATA at their cores (11). In vertebrates, six GATA factors, GATA-1, GATA-2, GATA-3, GATA-GT1, GATA-GT2, and GATA-5, have been identified and are involved in regulation of various genes such as those encoding globins, erythropoietin receptors, preproendothelin-1, T-cell receptors, and the gastric proton pump (29). To determine whether the 5'-GATAA-3' sequence on the O2R element functions in O₂ repression, we mutated the 5'-GATAA-3' sequence to 5'-ACGCC-3'. As shown in Fig. 3B, this mutation resulted in a decrease in the level of anaerobic derepression from 7.3- to 3.3-fold, indicating that the 5'-GATAA-3' sequence is important in O₂ repression. Since it is known that there are only four GATA factors, i.e., transcriptional activators Gln3p and Gat1p and transcriptional repressors Dal80p and Deh1p, in S. cerevisiae (11), Gln3p and Gat1p could be candidates for the transcriptional activator acting on the O2R element. These four proteins respond to the nitrogen source signal to regulate nitrogen catabolite-repressed genes (11) but have not been reported to be involved in transcriptional regulation by O₂ or UFAs. However, in vertebrates, GATA-1 has been demonstrated to play a major role in the regulation of various specific erythroid genes, for example, genes encoding globins and heme biosynthetic enzymes involved in terminal erythroid differentiation in vivo (14, 36, 37, 47, 49, 50). This finding appears to imply a relationship between GATA factors and O₂ signals.

View larger version (33K): [in this window] [in a new window]   FIG. 3.   (A) The O2R element contains a 5'-GATAA-3' sequence (arrow) on the complementary strand. (B) The 5'-GATAA-3' sequence plays an important role in O2 repression. Dotted box, O2R element (nt 356 to 307); solid box, O2R in which the 5'-GATAA-3' sequence is changed to 5'-ACGCC-3' by site-directed mutagenesis. The conditions used for measurement of rAPase activity in cells of the wild-type strain (SH5143) harboring the respective reporter genes and the symbols employed are as described in the legend to Fig. 2.

O₂ and UFA signals act on the O2R element. To further clarify the relationship between O₂ and UFA repression mechanisms, we investigated the effect of UFAs on anaerobic derepression of OLE1 transcription. Northern blot analysis showed that anaerobic derepression of OLE1 transcription (Fig. 4A, lanes 1 and 2) did not occur in the presence of UFA (Fig. 4A, lanes 3 and 4). This suggests that UFA repression is epistatic to anaerobic derepression. From a physiological viewpoint, this result is reasonable because, if there is a sufficient amount of UFA in the medium, cells do not have to produce UFA even under anaerobic conditions.

View larger version (43K): [in this window] [in a new window]   FIG. 4.   (A) Anaerobic derepression of OLE1 transcription is repressed by oleic acid. Cells of the wild-type strain (SH5141) were inoculated into YPDA medium (lanes 1 and 2) or YPDA medium containing 1 mM oleic acid (lanes 3 and 4) and cultivated at 30°C aerobically (lanes 1 and 3) or anaerobically (lanes 2 and 4). Total RNA was prepared from cells in the logarithmic growth phase. The RNA samples (each 10 µg of RNA) were subjected to Northern blot hybridization as described in the legend to Fig. 1A. (B) The O2R element is responsible for repression by UFA. Cells of SH5143 (wild type) harboring one copy of the respective reporter genes integrated at the ura3 locus were cultivated aerobically without oleic acid (dotted bars), anaerobically without oleic acid (solid bars), aerobically with oleic acid (open bars), or anaerobically with oleic acid (hatched bars) and subjected to an rAPase assay as described in the legend to Fig. 2.

View larger version (28K): [in this window] [in a new window]   FIG. 5.   A model for signal transduction pathways of O2 and UFA regulating OLE1 transcription. X, unknown transcriptional activators acting on the FAR element; M and C, mitochondrial respiratory chain and cytochrome c oxidase, respectively; T, unknown fatty acid transporter located at the cellular membrane. Arrows and blunt arrows, positive and negative interactions, respectively. UFA repression is epistatic to anaerobic derepression (thick line). See Discussion for details.