Candida albicans Spt23p Controls the Expression of the Ole1p Δ9 Fatty Acid Desaturase and Regulates Unsaturated Fatty Acid Biosynthesis*

In Saccharomyces cerevisiae the endoplasmic reticulum membrane proteins scSpt23p and scMga2p control the formation of unsaturated fatty acids by a mechanism that involves their release from the membrane by ubiquitin-mediated proteolysis. The resulting soluble polypeptides act as transcription activators that specifically control the expression of scOLE1, a gene that encodes scOle1p, a Δ9 fatty acid desaturase that forms cis-monounsaturated fatty acids (9Z-16:1 and 9Z-18:1) from saturated fatty acyl-CoA precursors. ScOle1p is the only long chain fatty acid desaturase in Saccharomyces and its membrane and storage lipids contain only saturated fatty acids and the monounsaturated products of that enzyme. Most other fungi, however, express multiple endoplasmic reticulum desaturases, including enzymes that form both mono- and polyunsaturated fatty acids. These typically include Δ12 and Δ15 enzymes that form the polyunsaturated species, 9Z,12Z-18:2, and 9Z,12Z,15Z-18:3, which are the most abundant fatty acids in membrane and storage lipids. An analysis of genomic DNA sequences shows that Candida albicans has a single homologue of the Saccharomyces scSPT23 and scMGA2 genes that we designate here as caSPT23. This study describes the characterization of the caSPT23 gene product and shows that it can repair the unsaturated fatty acid auxotrophy when it is expressed in a Saccharomyces scspt23Δ;scmga2Δ strain. In addition we show caSPT23 is essential for the expression of one of the two Δ9 desaturase homologues in Candida and potentially other functions associated with fatty acid metabolism.

In fungi, unsaturated fatty acids are synthesized by endoplasmic reticulum membrane-bound fatty acid desaturases that form double bonds in the acyl chains of either fatty acyl-CoAs (1) or glycerolipids (2,3). In the yeast Saccharomyces cerevisiae, which only forms 9Z monounsaturated fatty acids, the only fatty acid desaturase is encoded by the scOLE1 2 gene that is under the regulatory control of a pair of homologous endoplasmic reticulum proteins, scSpt23p and scMga2p (4 -6). In this system, the N-terminal fragments of scSpt23p and scMga2p control the expression of scOLE1 as transcription co-activators following their release from the endoplasmic reticulum by a mechanism that involves the ubiquitination of the membrane-bound precursor followed by a selective proteolytic cleavage by the 26 S proteosome (7). The activation and regulation of scOLE1 expression appears to be the only essential function of scSpt23p and scMga2p and disruption of both genes produces a synthetic fatty acid auxotrophy that can be repaired by including cis-unsaturated acids in the growth medium or by expression of the scOle1p desaturase gene under the control of a heterologous promoter (4).
An analysis of the Candida albicans genome identified only one homologue of the Saccharomyces scSPT23/scMGA2 genes. Given the highly specific and essential regulatory role of the scSpt23p/scMga2p proteins in regulating unsaturated fatty acid formation, we examined the hypothesis that caSpt23p exerts a similar control over the expression of the ⌬9 enzyme and possibly other desaturases. This study shows that when caSpt23p is expressed in Saccharomyces it can repair the loss of scOLE1 transcription in an scspt23⌬;scmga2⌬ strain. We further show that repression of caSPT23 in Candida albicans blocks the expression of the caOLE1 desaturase, producing growth defects and loss of viability that can be repaired by a mixture of unsaturated fatty acids. Metabolic labeling studies also indicate that, under conditions that repress caSpt23, the synthesis of both monoenoic and dienoic acids are repressed, suggesting that it plays a fundamental role in the regulation of unsaturated fatty acid metabolism.

EXPERIMENTAL PROCEDURES
Strains and Growth Media-C. albicans and S. cerevisiae strains and plasmids used in this study are shown in Tables 1 and 2. Saccharomyces strains were grown at 30°C on synthetic dextrose) or SG (synthetic galactose) drop out medium as previously described (12). Candida yeast phase cells were also grown at 30°C in synthetic dextrose or synthetic maltose drop-out medium with the addition of 80 M uridine (13). In experiments employing fatty acid supplements, the growth medium used in liquid cultures contained 0.1% Tergitol/Nonidet P-40 (Sigma) to disperse the fatty acids. Tergitol/Nonidet P-40 is not derived from fatty acids or fatty alcohols and is apparently not metabolized by Candida. Fatty acids were obtained from NuChek Prep (Elysian, MN). Escherichia coli NovaBlue Singles competent cells were obtained from Novagen. Growth tests were performed by monitoring A 600 of the cultures or by hemocytometer counting. For viability assays, cells were subjected to brief sonication (2 ϫ 3 s) to disperse clumped cells before plating on the appropriate agar medium.
Cloning, Disruption, and Modifications of caSPT23-The construction of Candida BWP17 strains in which the caSPT23 was placed under the control of the caMAL2 promoter was done in a two-step procedure as shown in Fig. 1. First, one of the caSPT23 alleles was disrupted using the linear DNA fragments containing the caHIS1 gene flanked by caSPT23 upstream and downstream gene sequences. A linear DNA cassette containing the C. albicans caARG4 gene and sequences derived from the promoter of the caMAL2 gene was then inserted before the start codon of the caSPT23 gene on the homologous chromosome. All DNA fragments used for the gene disruption and modification procedures were constructed by PCR methods from the native caSPT23 gene sequences using the oligonucleotide primers shown in Table 3. The native caSPT23 gene used in these manipulations was isolated from C. albicans strain BWP17 by PCR cloning of genomic DNA. Strain BWP17 was transformed with the disrupted or modified forms of DNA fragments encoding the caSPT23 gene using the BD Biosciences Yeastmaker Transformation System 2 method.
Heterologous Expression of the C. albicans caSPT23 Gene in Saccharomyces-The protein coding sequences of the caSPT23 gene were subcloned into a BamHI-XhoI site of vector pESC-TRP (Stratagene) to make vector pEGS (pESC-GAL1-caSPT23-TRP1). This plasmid, which contains the caSPT23 gene under control of the scGAL1 promoter, was transformed into S. cerevisiae strain m⌬s⌬ (mga2::kanMAX4;spt23::kanMAX4), which contains disrupted alleles of the scSPT23 and scMGA2 genes. Expression assays of this strain were performed on tryptophan dropout synthetic glucose agar medium with 1% Tergitol/Nonidet P-40 with no fatty acids or 0.5 mM 16:1 ϩ 0.5 mM 18:1.
Preparation of Radiolabeled Probes-The entire open reading frames of the caOLE1, caOLE2, and caACT1 genes were isolated using PCR primer pairs (Table 3) and cloned into plasmid pSTBlue-1 to make plasmids pSTBlue-OLE1, pSTBlue-OLE2, and pSTBlue-ACT1 ( Table  2). The resulting plasmids were opened with restriction enzymes, NcoI (for pSTBlue-OLE1), ScaI (for pSTBlue-OLE2), and NdeI (for pSTBlue-ACT1), and the linearized plasmid DNAs were separated by agarose gel electrophoresis in 1ϫ Tris acetate-EDTA buffer and purified by the GENECLEAN Kit (Q-Biogene). Radiolabeled RNA probes from the linearized plasmids were synthesized in vitro with [␣-32 P]dUTP (PerkinElmer Life Sciences) by SP6 polymerase using the Strip-EZ RNA kit from Ambion. Unincorporated nucleotides were removed  from the samples using a MicroSpin G-25 column (GE Healthcare). The specific activities of the labeled probes were determined by liquid scintillation counting. ␤-Galactosidase Assays-␤-Galactosidase reporter gene assays were performed as previously described (16) with the modification that cells were disrupted prior to the assay by three cycles of freezing and thawing at Ϫ80 and 30°C.
RNA Isolation and Northern Blot Analysis-The Candida strain, s⌬H-M2pS, was cultured in His/Arg dropout synthetic dextrose medium at 30°C overnight and the cells were washed with and transferred to His/Arg dropout synthetic maltose medium for induction of the caSPT23 gene. The cultures were incubated at 30°C with shaking and harvested by centrifugation. The cell pellet was washed with cold water twice and 2 volumes of RNAlater (Ambion) was added to the pellet prior to freezing at Ϫ80°C. Total RNA was isolated using the RiboPure YeastKit from Ambion. The amount of RNA was measured in a 96-well microplate with the Quanti-iT RNA Assay Kit (Molecular Probes) and the fluorescence signal was quantified by fluorescence imaging with a STORM 860 (GE Healthcare) using the ImageQuant program (Amersham Biosciences). For Northern blot analysis equal amounts of heat-denatured total RNA were fractionated on 1% formaldehyde gels (17) and the separated RNA was transferred to Hybond-N ϩ membrane (GE Healthcare) in 10ϫ SSC for 90 min at 7 inches of mercury using a Vacuum Blotter (model 785), according to instructions from the manufacturer. The membrane was pre-hybridized at 65°C for 2 h in 1 mM EDTA, 0.25 M Na 2 HPO 4 , pH 7.2, and 7% SDS and hybridized at 65°C overnight with 1 ϫ 10 6 cpm/ml of radiolabeled RNA probe. After incubation the membrane was washed twice at 65°C for 30 min with 1 mM EDTA, 40 mM Na 2 HPO 4 , pH 7.2, and 5% SDS. This was followed by two 30-min wash cycles at 65°C with 1 mM EDTA, 40 mM Na 2 HPO 4 , pH 7.2, and 1% SDS. Northern blot radioactivity was quantified by phosphorimaging.
Protein Isolation and Western Blot Analysis-Candida strain s⌬H-M2pS was grown at 30°C in supplemented minimal media to a 1 ϫ 10 7 cells/ml and lysis of cells for immunoblotting was performed as described (18). Cells were resuspended in 300 l of lysis buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 2 mM phenylmethylsulfonyl fluoride) ϩ protease inhibitor mixture (Sigma)) and disrupted with 300 l of acid-washed glass beads (Sigma) in a BioSpec bead beater at full speed 3 times for 1 min. Cells were allowed to cool on ice for 1 min between cycles to avoid overheating the sample. Glass beads were removed from the total lysate by centrifugation for 10 s and the supernatant was mixed with 2ϫ urea sample buffer (8 M urea, 4% SDS, 10% ␤-mercaptoethanol, 0.125 M Tris, pH 6.8) with incubation at 65°C for 10 min. The liquid lysate was then clarified by centrifugation for 5 min. The amount of protein in the final lysate was quantified with the Quanti-iT Protein Assay Kit (Molecular Probes).
Proteins were fractionated by SDS-PAGE using a Mini-PROTEAN 3 electrophoresis cell (Bio-Rad) as described (19). Equal amounts of heat-denatured total lysate in urea sample buffer were applied onto 8% SDS-PAGE gels and separated by electrophoresis at a constant 100 V for 1.2 h or until the bromphenol blue dye was just passed through the gel. Proteins were then electroblotted to a Hybond-P polyvinylidene difluoride membrane. The blotted membrane was placed in TBST (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% Tween 20) with 5% ECL Blocking Agent (GE Healthcare) for 1 h at room temperature. Blots were incubated with either a 1/10,000 dilution of monoclonal anti-HA antibody (Sigma) or a 1/500 dilution of monoclonal anti-␣-tubulin antibody (Sigma) in TBST (with 5% ECL Blocking Agent) at 4°C overnight or for 2 h at room temperature. The membrane was then washed with TBST and reacted with a 1/10,000 dilution of anti-mouse IgG horseradish peroxidase conjugate (Promega) in TBST (with 5% ECL Blocking Agent) for 1 h at room temperature. The secondary antibody-treated membrane was washed with TBST and developed with ECL Plus Western blotting detection reagents (GE Healthcare). The chemifluorescence signal was detected by a STORM 860 imager (GE Healthcare) and quantified with the ImageQuant program (Amersham Biosciences).
Fatty Acid Analysis and Labeling Studies-Fatty acid composition of C. albicans cultures were determined by HCl methanolysis of cells harvested by centrifugation and washed two times with ice-cold distilled water as previously described for Saccharomyces cultures (20). The resulting fatty acid methyl esters were analyzed by gas chromatography as previously described (21). For fatty acid radiolabeling studies, overnight cultures of strain s⌬H-M2pS were harvested by centrifugation, washed with an equal volume of synthetic medium with no carbon source, and inoculated at a density of 2.5 ϫ 10 6 in 15 ml of synthetic maltose medium or 5 ϫ 10 6 /ml in 15 ml of SG medium. After incubation at 30°C for 4 h, 10 Ci of 1-14 C-labeled 16:0 or 18:1 were added to each culture flask. After a 4-h incubation, cells were harvested by centrifugation and washed two times in cold 0.7% NaCl. Washed cell pellets were then frozen at Ϫ80°C overnight. Cell pellets were transferred to Tefloncapped glass tubes in 2 ϫ 1.0-ml aliquots of methanol and the lipids were extracted by the Bligh-Dyer procedure as previously described (22). The lipid fraction was washed two times according to the Folch procedure (23), evaporated under nitrogen, and transesterified by HCl methanolysis as previously described (20). Fatty acid methyl ester fractions were concentrated under nitrogen prior to thin layer chromatography analysis. Argentation TLC analysis was performed using a 70:30 (v/v) hexane:ethyl ether solvent system on Silica Gel G plates presoaked in a solution of methanolic 4% AgNO 3 in methanol:H 2 O (9:1, v/v) and activated for 30 min at 110°C in the dark.

RESULTS
A comparison of the Candida caSPT23 gene with its Saccharomyces scSPT23 and scMGA2 homologues reveals significant conservation within their protein domains despite the divergence of these two species over 700 million years ( Fig. 2A). Although the length of the Candida protein is about 75% of the Saccharomyces homologues, all three proteins show regions of high sequence identity in the N-terminal transcription activation domains, a highly homologous protein-protein interaction domain and near-perfect identities in regions that contain the predicted nuclear export signal and the transmembrane domains. These regions of high homology are located at identical positions on all three proteins (Fig. 2B). Unlike the Saccharomyces homologues, however, the Candida protein lacks the ankyrin repeat sequences located at a C-terminal position to the protein-protein interaction domain.
To test whether the Candida protein can activate scOLE1 transcription in Saccharomyces, a plasmid containing the caSpt23p protein coding sequence under the control of the yeast scGAL1 promoter was introduced into the scspt23⌬;scmga2⌬ strain. Fig. 3A shows that the expression of the Candida gene in medium containing galactose relieves the fatty acid auxotrophy of the strain, indicating that it can function in activating the expression of the scOLE1 gene. Fig. 3B shows that a ␤-galactosidase reporter gene fused to the full-length Saccharomyces OLE1 promoter is activated in the Saccharomyces spt23⌬;mga2⌬ cells by caSpt23 and can be repressed ϳ4.5-fold by exogenous unsaturated fatty acids. This attenuation of reporter gene activity requires a longer time of exposure to unsaturated fatty acids than that observed in wild type cells containing functional copies of the native scSPT23 and scMGA2 genes. Genetic and Biochemical Characterization of Strains Containing Glucose Repressible Forms of caSPT23-Because C. albicans is diploid, tests of the function of the gene requires the inactivation of both alleles. To examine the functions of the caSpt23 gene, we constructed integrating plasmids that placed the native and an N-terminal HA epitope-tagged form of its protein coding sequence under control of the maltose-inducible and glucose-repressible caMAL2 promoter. These were integrated into the remaining caSPT23 locus of a strain that contains a deletion of one of the caSPT23 alleles. The caMAL2 gene encodes a maltase that is the homologue of the Saccharomyces MAL62 gene. Previous studies have shown that the caMAL2 mRNA is undetectable by Northern blotting in C. albicans cells grown on glucose or glycerol, but is induced when cells are grown on maltose (24). Other studies have shown that a green fluorescent protein (25) and the caURA3 gene (26) are expressed when cells are grown on maltose but not on glucose when they are placed under the control of the caMAL2 promoter. Although it cannot be ruled out that a very low level of caMAL2-driven expression occurs in glucose-grown cells, studies of the caURA3 gene under control of caMAL2 promoter in cells grown under glucose is not sufficient to confer 5-fluoroorotic acid sensitivity on the strain, which is a sensitive indicator of weak basal gene activity (24).
Repression of caSPT23 Blocks Growth and Viability of Strains That Express Native and Epitope-tagged Forms of the Protein-Tests of strains that only express the caMAL2-controlled forms of either the wild type or the HA epitope-tagged caSPT23 genes show that after transfer from maltose to glucose medium there is a brief growth period and then cell division is strongly repressed. Fig. 4A shows the results of a representative set of experiments in which cells grown on maltose medium were washed and transferred to glucose containing medium at an inoculum of 2 ϫ 10 6 cells/ml. The cells that were exposed to glucose grew at the same rate as those transferred to fresh maltose carbon source for ϳ4 -6 h (corresponding to about 3.5 generations). At that time, growth of the glucose-repressed cells slows dramatically and progresses into a stationary phase. By 6 h, the glucose-treated cell population begins to exhibit significant clumping and the formation of elongated cells. By 10 h, very few cells with buds are observed and by 24 h a significant fraction of the clumped cells have formed pseudohyphal structures (Fig.  5, A-C). The timing of the initial growth phase, growth arrest, and pseudohyphal formation is consistent over a wide range of inoculum densities ranging from 2 ϫ 10 4 to 5 ϫ 10 6 cells/ml. Viability is also dramatically reduced in the glucose-repressed cultures, resulting in ϳ12% viable cells at 6 h, 5% at 10 h, and 3% after 24 h of growth (Fig. 4C). By contrast, cells containing either the native or HA-tagged genes that were maintained on maltose retained 60 -90% viability during the same period and remained as yeast phase cells as they progressed into stationary phase during the 24-h test period.

Glucose-repressed Growth, Loss of Viability, and Morphogenic
Changes in the caMAL2::caSPT23 Strain Can Be Repaired by Combinations of Unsaturated Fatty Acids-To test the specificity of the caSPT23 on the expression of fatty acid biosynthesis and desaturase genes, we monitored the growth of cells expressing either the MAL2-regulated native or HA-tagged forms of caSpt23p subjected to glucose repression in the presence fatty acids. The effects of fatty acids on the growth of the strain that expresses the native form of caSpt23p are shown in Fig. 4B and the effects of fatty acids on the viability of that strain are shown in Fig. 4C. A mixture of saturated fatty acids did not have any effect on the inhibition of growth resulting from the glucose repression of caSPT23 (data not shown). Growth was not repressed in cells incubated in glucose media that contained either a mixture of saturated and unsaturated fatty acids or a mixture of unsaturated fatty acids. Furthermore, the addition of the unsaturated fatty acid mixture to the glucose medium prevented the loss of viability as well as the formation of pseudohyphae (Fig. 5C) that is characteristic of caMAL2::caSPT23 cells grown under glucose repressed conditions. An identical pattern of unsaturated fatty acid repair of growth, viability, and maintenance of the yeast phase was observed with cells that expressed the caMAL2-regulated HA-tagged form of caSPT23p (data not shown).
Glucose Inhibition of caMAL2::caSPT23 Represses the Synthesis of caSpt23p-Quantitative PCR measurements showed that glucose repression of the caMAL2-regulated transgene resulted in a 2.6-fold (standard deviation (S.D.) ϭ 0.05, n ϭ 3) reduction of the transcript by 6 h. The effects of this repression are also shown by a Western blot analysis of the HA-tagged protein under control of the caMAL2 promoter (Fig. 6). Like the Saccharomyces homologues, a high molecular mass form of the protein (corresponding to the 83-kDa membranebound precursor) and a lower molecular mass species (the proteolytically processed, soluble form) can be found in normally growing cells. Both forms are abundant in cells grown under derepressing conditions (in maltose medium) and semi-quantitative comparisons of protein levels using the tubulin gene as a control showed that the highest total levels of the caSpt23p protein were present immediately following the transfer of cells to fresh maltose medium. Under those conditions ϳ20% of caSpt23p was present as the 85-kDa membrane-bound precursor and 80% as the soluble, 55-kDa proteolytically cleaved form. During the 6 h of logarithmic phase growth, however, total caSpt23p levels declined to about 25% of their initial value and the levels of the membrane-bound and soluble forms of the protein were approximately equal.
By contrast, we were unable to detect the 85-kDa membrane-bound form of the protein in cells that were transferred to glucose containing medium during the course of the experiment. The absence of the membrane-bound caSpt23p in the glucose-repressed cells suggests that synthesis of the protein is strongly repressed and that the remaining membrane-bound precursor is rapidly converted to the soluble, 55-kDa form whose levels are also markedly reduced. We used the ratio of the 83-and 55-kDa forms of caSpt23p to ␣-tubulin to monitor the differences between the total caSpt23 protein levels between glucose-and maltosederepressed cells at each time point. Two hours after the transfer to fresh medium the glucose-repressed cells had ϳ10 -20% of the total caSpt23p found in the corresponding 2-h maltose-grown control cells and by 6 h the levels of the remaining caSpt23p had declined to 5-10% of the corresponding maltose grown control values. These data show that repression of caMAL2-caSPT23 causes a rapid reduction in the active form of the protein, which results in the cessation in growth and reduced viability.
After a shift to either fresh maltose or glucose medium there is a rapid increase in saturated fatty acid levels. This apparently reflects a general increase in fatty acid synthesis relative to desaturation as the cells progress into logarithmic growth phase with either carbon source. Following glucose repression of caSPT23, however, there were highly significant changes in fatty acid levels involving the monounsaturated and polyunsaturated acids and the saturated species 18:0. Cells containing the MAL2-regulated native and HA-tagged forms of caSpt23p showed dramatic decreases in their monoenoic and dienoic fatty acid levels in the first 6 h of growth ( Table 4). The largest decline involved the mono- enoic acids, changing from 35 to 24.7% (native caSpt23p) and 40 to 27.9% (HA-tagged caSpt23p), starting 4 h after the shift. These were accompanied by large increases in the saturated fatty acids (19 -34%) in cells containing the native caSpt23p, whereas, there were increases in both saturated and polyunsaturated species in the cells that contained the HAtagged form of the protein. By contrast, in the BWP17 wild type cells that were exposed to glucose for 6 h there was a significant increase in monounsaturated fatty acid levels from 32 to 40% and the polyenoic acids decreased from 49 to 36% over cells maintained on maltose (data not shown).
Glucose Repression of MAL2::caSPT23 Reduces mRNA Levels of caOLE1; caOLE2 Expression Is Not Affected-A Northern blot analysis was performed to determine the effects of caSPT23 repression on ⌬9 fatty acid desaturase mRNA levels (Fig. 7). C. albicans has two genes with homology to ⌬9 desaturases. The caOLE1 gene is highly homologous to a number of fungal ⌬9 enzymes and disruption of both alleles in Candida produces an unsaturated fatty acid requirement. Furthermore, it has been shown by heterologous expression in a S. cerevisiae ole1⌬ strain to encode an authentic ⌬9 fatty acid desaturase (27). The caOLE2 gene has less homology to fungal desaturases and does not produce an unsaturated fatty acid auxotrophy when disrupted in Candida or function as a desaturase when it is expressed in yeast (27). Quantitative Northern blot analysis using the Candida actin gene as a loading control showed that caOLE1 mRNA levels at 2, 4, and 6 h were Ͻ20% of the levels found in control cells grown in maltose at the same times. By contrast, the caOLE2 levels in the glucose-grown cells did not change significantly from the controls over the same period. These results were further confirmed by QPCR analysis of the two genes at the 6-h point using the Candida tubulin gene (TUB2) as an internal control. Those results showed that the caOLE1 mRNA levels had undergone a 5.6-fold decrease (S.D. ϭ 0.03, n ϭ 3), whereas the caOLE2 transcript levels did not show a statistically significant difference from the maltose-grown cells.
Glucose Repression of the caMAL2::caSPT23 Gene Results in the Repression of ⌬9 and ⌬12 Fatty Acid Desaturase Activity-To test the effects of repressed caSPT23 expression on fatty acid desaturase activities, cells grown in either maltose-or glucose-containing medium for 4 h were exposed to trace amounts (2.75 mol/1.5 ϫ 10 8 cells) of radiolabeled 16:0 and 18:1 and allowed to grow for an additional 4 h. Quantification of the radiolabel in the Bligh-Dyer lipid extracts showed that there were marked differences in the amounts of label that were incorporated from the two fatty acids under the different growth conditions. In the maltose-grown cultures, ϳ14 and 3% of the total radioactivity derived from 16:0 and 18:1, respectively, were incorporated into chloroform-soluble lipids, whereas in the glucose-grown cultures ϳ0.8% of both the 16:0 and 18:1 dpm were found in the lipid fraction.
To determine the relative amounts of conversion of the incorporated fatty acids by endogenous desaturase activities, the radiolabeled lipids were subjected to acid methanolysis and the resulting fatty acid methyl esters were analyzed by chromatography on AgNO 3 -impregnated thin layer plates (Fig. 8). In cells that were grown in maltose medium, ϳ63% of the incorporated radiolabeled 16:0 was desaturated by the ⌬9 enzyme and ϳ2/3 of its products remained as monoenes (16:1 ϩ 18:1, or 40.7% total incorporated dpm). The remaining labeled monenes were converted by a ⌬12 activity to 18:2 (20% of the total incorporated dpm) and a fraction of that species was then desaturated by a ⌬15 activity to 18:3 (2% total incorporated dpm). In comparison, sharply reduced levels of the ⌬9 desaturase activity were observed with the glucose-repressed cells. Only 28% of the total incorporated 16:0 was desaturated. Twenty percent of the incorporated radiolabel was found in the monene fraction, 4.5% was found as 18:2 and 3.5% as 18:3. These data indicate that repression of caSpt23p results in a reduction in both ⌬9 and ⌬12 fatty acid desaturase activities.
The effects of Spt23p repression on the ⌬12 desaturase activity were also supported by the labeling patterns produced from the 14 C-labeled 18:1 species. In the maltose-grown cells ϳ70% of the incorporated 18:1 was converted by ⌬12 desaturase activity to 18:2. Fifty-one percent of the total incorporated label remained as 18:2 and the rest of that species was converted by a ⌬15 enzyme to 18:3, representing 19% of the total incorporated dpm. By contrast, in the glucose-repressed cells only 14% of the incorporated 18:1 was converted to 18:2 and Ͻ1% of the incorporated radiolabel was found in the 18:3 fraction.

DISCUSSION
An analysis of the current genome data bases suggests that homologues of the Spt23p/Mga2p family are found in all fungi, but not in other organisms. This indicates that these genes emerged early in the history of the fungal kingdom, well before the divergence of Candida and Saccharomyces some 730 million years ago (28). When viewed with respect to this length of evolutionary time, the high regional sequence identities found in the Spt23p/Mga2p homologues of these two widely diverged species and the highly conserved order of domains within those proteins suggest that the SPT23/MGA2 family controls basic functions that are necessary for fungal growth and survival.
The data presented here shows that the Candida caSpt23p protein has retained a similar function to that of its Saccharomyces homologues. The loss of viability when caSPT23 is repressed indicates that it is an essential gene and the ability of unsaturated fatty acids to repair the resulting growth and viability phenotypes seems to indicate that its primary function in yeast phase cells is to control the formation of unsaturated fatty acids. Furthermore, the ability of the Candida protein to substitute for its Saccharomyces homologues in yeast indicates that there is a remarkable conservation in its mechanism of function with respect to its ability to activate scOLE1 transcription. We should note that although we have recently demonstrated that the Saccharomyces scMga2p homologue has a secondary function associated with the regulation of OLE1 mRNA stability (6), we currently do not have the reagents to determine whether the Candida protein has similar dual functions in its native host.
S. cerevisiae does not form polyunsaturated fatty acids and the scOle1p desaturase is the only enzyme responsible for the formation of  long chain unsaturated fatty acids: its 9Z-(14 -18):1 products comprise ϳ80% of the fatty acid chains found in membrane lipids. In Candida the ⌬9 enzyme also plays a central role in unsaturated fatty acid formation because it catalyzes the initial step in the formation of unsaturated fatty acids. Unlike Saccharomyces, however, the majority of its monounsaturated products are converted to dienoic and trienoic acids by endogenous ⌬12 and ⌬15 desaturase activities to 18:2 and 18:3.
We had previously observed that the deletion of the Saccharomyces scOLE1 gene causes a fatty acid auxotrophy. When those cells are deprived of unsaturated fatty acids growth is arrested after 3.5 generations (6 h) and is accompanied by rapid cell death (29). During that period, the unsaturated fatty acyl content of membrane lipids is reduced from Ͼ80% to about 10% (29). In this study, we observed a similar pattern after the Candida caSPT23 gene is repressed. But, whereas growth arrest and loss of viability occurs in Candida over a similar time scale of 2.0 -2.2 generations (6 h), there is a much smaller reduction (from 75 to about 73%) in the levels of total unsaturated fatty acids. During that period, however, there are significant rearrangements within the total cellular fatty acyl composition. Most notably, repression of the native form of caSpt23p for 6 h results in a 25% reduction of monoenoic (16:1 and 18:1) acids, a 20% reduction in dienoic (18:2) species, and a 50% increase in the trienoic (18:3) species.
Whereas statistically significant differences in total saturated (c14 -18) fatty acid levels occur only at the 4-h point, within those species there are also highly significant (Ͼ50%) increases in 18:0 at 4 and 6 h after repression of caSpt23p. Taken together with the observed changes in unsaturated fatty acids, these results are consistent with the expected effects if caSpt23p exerts direct control over caOLE1 transcription.
Repression of caSPT23 causes a rapid reduction in the caOle1p ⌬9 desaturase activity, blocking further synthesis of 16:1 and 18:1 and resulting in an accumulation of 18:0 through continued activity of fatty acid synthase and the ELO1-specific (21) 16:0 3 18:0 elongation pathway. In addition, the continuing conversion of a significant fraction of the pre-existing 18:1 to 18:2 and 18:3 by residual ⌬12 and ⌬15 enzyme activities results in the depletion of 18:1 and 18:2 and an accumulation of Comparison of fatty acid composition of strains that express caMAL2-regulated native and N-terminal HA epitope-tagged forms of caSPT23 following transfer to fresh maltose-or glucose-containing medium Data expressed as wild type % total C14 -C18 fatty acid methyl ester mass derived from total cellular lipids. Values are mean Ϯ S.D. for three independent experiments for each strain (values in bold face type show significant deviations).

Time
Carbon source 16:0 18:0 C (n14 -18) :1 a 18:2 18:3 Total saturated b Total unsaturated c 18:3. Whereas these data offer strong support to the hypothesis that caSPT23 exerts direct control over the expression of the ⌬9 desaturase, the radiolabeling studies presented here also suggest that it may also be involved in the regulation of the ⌬12 enzyme. It is not clear whether the growth and viability defects produced by the repression of caSpt23p is caused by the inhibition of functions that are associated with the reduction of monoenoic or dienenoic acids, the increased levels of 18:3 and/or 18:0 or changes in critical molecular species of membrane lipids that result from altered ratios of all or some of these fatty acyl species. The relatively small reduction in total unsaturated fatty acid content associated with the inhibition of growth in Candida compared with that required to block Saccharomyces growth is striking, however, and illustrates how these two organisms have diverged in their ability to adapt to changes in the fatty acyl content of their membrane lipids. These differences in the adaptive abilities of the two organisms undoubtedly reflect the different physiological and nutritional conditions they encounter.
Although C. albicans can grow over a wide range of temperature conditions that are known to affect unsaturated fatty acid biosynthesis and mobilization it is a commensal organism whose primary habitat consists of mammalian tissues that are maintained over a narrow range of temperatures. (An extensive study of the literature has failed to identify a C. albicans isolate from an environment other than a warmblooded host.) Within the domain of host tissues, ranging from the intestine, to various mucosal surfaces and deep organ tissues, however, there are significant differences in the availability of lipid nutrients that might be exploited as a source of fatty acids. Candida also exists in a number of different morphogenic states (30,31), including the budding yeast and pseudohyphal forms described here as well as the filamentous hyphal form associated with the promotion of pathogenesis. All of these conditions may require differential regulation of fatty acid synthetic and desaturation activities. Although this study suggests that the primary function of caSPT23 in the budding yeast form under strictly defined growth conditions is to regulate ⌬9 desaturase gene expression, we cannot rule out that caSPT23 may exert control over other genes, including the ⌬12 and ⌬15 enzymes when it grows under different environmental or developmental conditions.