The Yeast Anaerobic Response Element AR1b Regulates Aerobic Antifungal Drug-dependent Sterol Gene Expression*

Background: Saccharomyces cerevisiae sterol gene expression is regulated by a consensus sterol-response promoter element (SRE/AR1c). Results: The anaerobic AR1b promoter element regulates global antifungal-dependent sterol gene expression. Conclusion: Yeast sterol gene expression is regulated by multiple SRE-like elements. Significance: Understanding sterol gene expression will yield valuable information concerning antifungal drug resistance. Saccharomyces cerevisiae ergosterol biosynthesis, like cholesterol biosynthesis in mammals, is regulated at the transcriptional level by a sterol feedback mechanism. Yeast studies defined a 7-bp consensus sterol-response element (SRE) common to genes involved in sterol biosynthesis and two transcription factors, Upc2 and Ecm22, which direct transcription of sterol biosynthetic genes. The 7-bp consensus SRE is identical to the anaerobic response element, AR1c. Data indicate that Upc2 and Ecm22 function through binding to this SRE site. We now show that it is two novel anaerobic AR1b elements in the UPC2 promoter that direct global ERG gene expression in response to a block in de novo ergosterol biosynthesis, brought about by antifungal drug treatment. The AR1b elements are absolutely required for auto-induction of UPC2 gene expression and protein and require Upc2 and Ecm22 for function. We further demonstrate the direct binding of recombinant expressed S. cerevisiae ScUpc2 and pathogenic Candida albicans CaUpc2 and Candida glabrata CgUpc2 to AR1b and SRE/AR1c elements. Recombinant endogenous promoter studies show that the UPC2 anaerobic AR1b elements act in trans to regulate ergosterol gene expression. Our results indicate that Upc2 must occupy UPC2 AR1b elements in order for ERG gene expression induction to take place. Thus, the two UPC2-AR1b elements drive expression of all ERG genes necessary for maintaining normal antifungal susceptibility, as wild type cells lacking these elements have increased susceptibility to azole antifungal drugs. Therefore, targeting these specific sites for antifungal therapy represents a novel approach to treat systemic fungal infections.

CaUpc2 gain-of-function mutations have been associated with azole resistance in Candida albicans clinical isolates (26 -31). Mutations in CaUpc2 result in the constitutive induction of ERG11 gene expression, which is the target of the common azole drugs. Increased ERG11 gene expression has also been seen in azole-resistant isolates of Candida glabrata (32)(33)(34)(35). C. glabrata resistance is on the rise, and it is the second most common cause of disseminated candidiasis (36 -38). As C. glabrata also contains Upc2 orthologs (39), it is only a matter of time before gain-of-function mutations are reported in clinical resistance isolates. Thus, a greater understanding of how Upc2 functions to regulate gene expression is required to fully comprehend how pathogenic yeast gain resistance.
In this study, we determined the in vivo promoter activities of SRE/AR1 elements in several ERG genes. The data show for the first time that two anaerobic AR1 b elements in the UPC2 promoter are required for inducing global sterol gene expression in response to antifungal-induced blocks in sterol biosynthesis. These two elements are the nucleus for global ERG gene expression in response to antifungal assault, as strains lacking these promoter elements show increased susceptibility to antifungal drug treatment. The data have further increased the knowledge of how multiple pathogenic yeast gain resistance to antifungal drug therapies.

EXPERIMENTAL PROCEDURES
Strains, Media, and Miscellaneous Methods-The yeast strains used in this study are isogenic to W303 (MATa leu2-3, 112 trp1-1 ura3-1 his3-11, 15 can1-100) or BY4741 (MATa his3 leu2 met15 ura3). Strains were grown in YEPD (1% yeast extract, 2% bacto-peptone, 2% glucose) or synthetic minimal media containing 0.67% yeast nitrogen base (Difco), supplemented with the appropriate amino acids, adenine, and uracil. Lovastatin (50 g/ml) was added directly to liquid YEPD or synthetic media and was incubated with cells for 16 h. Lovastatin was solubilized as follows. 25 mg of lovastatin (Sigma) was dissolved in 250 l of 0.2 M NaOH/EtOH (0.8 g of NaOH per 100 ml of 100% ethanol) and incubated in a 65°C water bath for 40 min. 600 l of 0.2 M Tris-HCl, pH 8.0, and 150 l of 1 M Tris-HCl were then added. Yeast transformation was performed as described by Ito (40). Escherichia coli XL1Blue cells were used for plasmid propagation and grown in LB medium supplemented with ampicillin (150 g/ml).
Plasmid Construction-The yeast YIp353 integrating plasmid was used to construct promoter-lacZ constructs. Plasmids were integrated at the endogenous URA3 locus. The ERG25 promoter contained 1500 bp, ERG3 contained 1000 bp, ERG1 contained 750 bp, and UPC2 contained 750 bp. The URA3 centromeric plasmid, pRS416, was used to construct low copy ERG vectors. pRS416 constructs contained the promoter sequence, the entire coding sequence, and 500 bp of the 3Ј-untranslated region. pRS416-ERG25 was transformed into an ERG25/erg25:: perg25::HIS3 diploid strain. This strain contains an endogenous HIS3-generated ERG25 promoter disruption allele. ERG25:: perg25::HIS3 pRS416-ERG25 haploids were obtained by sporulation and selecting for His ϩ Ura ϩ . Haploids were FOA s (41), demonstrating the need for plasmid-driven ERG25 activity. pRS416-ERG3 was transformed into an erg3::kan r null strain. This strain was generated in W303 by PCR amplification using template DNA from an erg3::kan r haploid strain (Open Biosystems, Huntsville, AL). All SRE mutant constructs were generated using the QuikChange multisitedirected mutagenesis kit (Stratagene, La Jolla, CA) and were verified by DNA sequencing.
IC 50 Microdilution Assay-IC 50 assays were performed as follows. Yeast cultures were diluted to 5 ϫ 10 3 cells/ml in YEPD. Cell aliquots of 100 l were distributed to wells of a 96-well flat-bottom plate, except for row A, which received 200 l. Drug was added to row A at the desired concentration and then serially 2-fold diluted to rows B-G; row H served as a drug-free control. Plates were incubated at 30°C for 24 and 48 h. Absorbance at 620 nm was read with a microplate reader (Beckman Coulter, Inc., Fullerton, CA); background due to medium was subtracted from all readings. IC 50 values were defined as the lowest concentration inhibiting growth at least 50% relative to the drug-free control.
Total RNA Isolation and Northern Analysis-Cells were grown to exponential phase at 30°C in synthetic medium. Total qRT-PCR Analysis-Cells were grown to exponential phase at 30°C. RNA was resuspended in diethyl pyrocarbonatetreated water. 50 ng of RNA, 11.5 l of the SYBR Green Master Mix (Quanta), and 5 M primer sets were loaded in triplicate onto a 96-well plate. qRT-PCR amplification and analysis were completed using the Stratagene Max-Pro (Mx3000P) software version 4.0.
Phosphatidic Acid Phosphatase Enzymatic Assay-PA phosphatase activity was measured by following the release of watersoluble 32 P i from chloroform-soluble [ 32 P]PA for 20 min at 30°C in a total reaction volume of 0.1 ml. For measurement of PA phosphatase activity, the reaction mixture contained 50 mM Tris-HCl buffer, pH 7.5, 0.2 mM [ 32 P]PA (10,000 -12,000 cpm/ nmol), 2 mM Triton X-100, and 5 mM MgCl 2 (43)(44)(45)(46)(47)(48)(49), and 10 g of cell extract. A unit of PA phosphatase activity was defined as the amount of enzyme that catalyzed the dephosphorylation of 1 nmol of PA/min. Specific activity was defined as units/mg of protein.
To induce expression of fusion proteins, transformed BL21 cells were grown at 37°C in Luria Broth containing 100 g/ml ampicillin and 150 M ZnSO 4 to mid-log phase (A 600 ϳ0.5). Isopropyl ␤-D-thiogalactopyranoside was added to a final concentration of 1 mM, and cultures were incubated with shaking at 37°C for a further 4 h. Soluble cytoplasmic proteins were extracted from cells using BugBuster Master Mix (EMD Chemicals, Gibbstown, NJ) according to the manufacturer's instructions, except that the BugBuster was supplemented with 10 M ZnSO 4 . GST fusion proteins were purified using a GSTrap column (GE Healthcare), following the manufacturer's instructions, except that all buffers contained 10 M ZnSO 4 .
AlphaScreen Assay for Upc2 DNA Binding-AlphaScreen is a bead-based technology. Donor beads contain a photosensitizer that converts ambient oxygen to an excited form of singlet oxygen upon illumination at 680 nm. Singlet oxygen diffuses up to 200 nm in solution before it decays. Interaction between GST-CaUpc2 and SRE brings the donor beads into close proximity with the acceptor beads; singlet oxygen will activate thioxene derivative in the acceptor beads, leading to the emission of light between 520 and 620 nm. In the absence of acceptor beads, the singlet oxygen falls to the ground state with no light emission. Donor beads can release up to 60,000 singlet oxygen molecules/s, resulting in signal amplification. Because signal detection is performed in a time-resolved manner and at lower wavelength than excitation, interference is low (adapted from "Exclusive AlphaScreen and AlphaLISA Assay Technology, PerkinElmer Life Sciences).
Binding of the Upc2 DNA binding domain to DNA was detected using streptavidin-coated AlphaScreen donor beads (PerkinElmer Life Sciences), which capture biotinylated DNA, and anti-GST-coated AlphaScreen acceptor beads, which bind to the GST-Upc2 fusion protein. DNA binding by the fusion protein brings the two types of beads into close proximity, allowing the generation of an AlphaScreen signal. The DNA sequences of probes used for the screening assay are as follows: S. cerevisiae novel AR1 b , 5Ј-biotin-TEG-CTGTATTGTCGTT-TAAAAGTGG-3Ј and 5Ј-CCACTTTTAAACGACAATACA-G-3Ј; C. albicans ERG2 consensus SRE/AR1 c , 5Ј-biotin-TEG-CTGTATTGTCGTATAAAAGTGG-3Ј and 5Ј-CCACTTTT-ATACGACAATACAG-3Ј; C. albicans ERG2 mutant SRE/ AR1 c , 5Ј-biotin-TEG-CTGTATTGTCAGATAAAAGTGG-3Ј and 5Ј-CCACTTTTATCTGACAATACAG-3Ј. TEG is a triethylene glycol spacer. The two probes were hybridized to form a biotinylated, double-stranded oligonucleotide by combining them at a concentration of 50 M each in a buffer consisting of 100 mM KoAc, 30 mM HEPES, pH 7.4. The mixture was heated to 91-95°C for 2 min and allowed to slowly cool.
To perform the assay, fusion protein was diluted to 8 nM (0.34 g/ml) in an assay buffer consisting of 25 mM HEPES, 200 nM NaCl, 0.1% Tween 20, and 3 M ZnSO 4 . Protein was then combined with an equal volume of acceptor beads, which had been diluted to 80 g/ml in assay buffer, and 4 l/well of the mixture was transferred to a white 1536-well assay plate. The plate was incubated for 30 min at room temperature. Biotinylated and double-stranded DNA was diluted to 40 nM in assay buffer and combined with an equal volume of 80 g/ml donor beads in assay buffer. 4 l of this mixture was added to each well of the assay plate, and the plate was incubated for a further 1 h at room temperature in the dark. The results were then read using an Envision microplate reader (PerkinElmer Life Sciences).
Chromatin Immunoprecipitation Assay-Cells were treated with 1% formaldehyde for 15 min at room temperature. Crosslinking was stopped by addition of 125 mM glycine. Cells were pelleted, washed with PBS, and again pelleted. Cells were then lysed using lysis buffer (50 mM HEPES, pH 7.5, containing 140 mM NaCl, 0.1% Triton X-100, 0.1% sodium deoxycholate, and a protease mixture) and spun down to remove cellular debris. 1 mg of the resulting supernatant was used for immunoprecipitating ScUpc2-bound DNA for 1 day at 4°C, using 2 g of anti-Myc monoclonal antibody. Immunoprecipitated complexes were isolated after 1 h at 4°C using protein A-Sepharose/ agarose beads. The beads were washed with lysis buffer containing 500 mM NaCl, and finally with 10 mM Tris-HCl, pH 8.0, containing 0.25 M LiCl, 0.5% Nonidet P-40, 0.5% sodium deoxycholate. ScUpc2-DNA complexes were eluted from beads using 50 mM Tris-HCl, pH 8.0, containing 10 mM EDTA, and 1% SDS. Cross-linking was reversed using 5 M NaCl at 65°C for 6 h. Samples were then diluted using 500 mM Tris-HCl, pH 8.0, containing 10 mM EDTA and 0.67% SDS. Samples were treated with 250 g of proteinase K for 1 h at 37°C. DNA was eluted from samples using the DNA Easy Kit (Qiagen). The degree of binding was determined using qRT-PCR.

ERG25 Promoter-lacZ Fusion Lacking Consensus SRE/AR1 c Elements Has Lovastatin-induced
Activity-We previously demonstrated that ERG25 gene expression was induced in response to blocking several steps in ergosterol biosynthesis and that induction required Upc2 and Ecm22 (9). The ERG25 promoter contains three consensus SRE/AR1 c sites (Table 2) (11, 14 -16). A ␤-galactosidase promoter-lacZ fusion assay was used to determine whether these elements harbored activity. All combinations of ERG25-lacZ SRE/AR1 c promoter deletion/ mutation constructs were tested in response to treatment with the HMG-CoA reductase inhibitor lovastatin. The upc2⌬ ecm22⌬ null cells harboring all constructs were also tested to determine whether ScUpc2/ScEcm22 functions were required for activity.
Of all the combinations tested, only an ERG25-lacZ fusion lacking SRE/AR1 c sites 1 and 3 together had reduced lovastatininduced activity (Fig. 1, A and B, SRE/AR1 c 2 versus SRE/ AR1 c 1,2,3). In some cases, the loss of a single or double SRE/ AR1 c site increased basal activity over the wild type promoter fusion (SRE/AR1 c 1,2,3 versus SRE/AR1 c 1 or SRE/AR1 c 2,3). Importantly, an ERG25-lacZ construct lacking all three SRE/ AR1 c sites retained a wild type activity (Fig. 1B, sre Ϫ /ar1 c Ϫ versus SRE/AR1 c 1,2,3) that required ScUpc2/ScEcm22 (Fig. 1C). Identical results were obtained when SRE/AR1 c sites were mutated by site-directed mutagenesis (data not shown). These results suggested that novel promoter elements regulated ERG25 promoter activity in response to a block in ergosterol biosynthesis.
To determine whether the sre Ϫ /ar1 c Ϫ active phenotype was common to other ERG promoters, ERG1 and ERG3 AR1 c promoter sites were tested using ERG1-lacZ and ERG3-lacZ fusions. ERG1 and ERG3 contain 1 and 2 SRE/AR1 c sites, respectively ( Table 2). Sites were deleted/mutated, and cells were assayed for promoter-lacZ activity in the absence or presence of lovastatin.
The deletion of the single SRE/AR1 c in the ERG1-lacZ promoter resulted in a total loss of lovastatin-induced activity (Fig.  1D, sre Ϫ /ar1 c Ϫ versus SRE/AR1 c 1), whereas loss of each ERG3 SRE/AR1 c site alone or together reduced, but did not abolish, lovastatin-induced ERG3 promoter activity (Fig. 1E, The SRE/AR1 c /Ar1 b locations for ERG25, ERG1, ERG3, and UPC2 are given.

SRE Location
Ϫ359 to Ϫ353 AR1b a Site has been removed due to insertion of HIS3. when compared with the wild type ERG3 promoter-lacZ fusion. Importantly, promoter activities required the presence of Upc2/Ecm22, as induction was lost in upc2⌬ ecm22⌬ cells expressing various promoter-lacZ constructs (Fig. 1F). So, some, but not all, SRE/AR1 c sites are functional.

Regulation of Ergosterol Gene Expression
To mimic the DNA topology of the chromosomal locus in the context of in vivo SRE/AR1 c function, plasmids pRS416-ERG25 and pRS416-ERG3 were constructed, which contained a wild type or sre Ϫ /arl1 c Ϫ promoter driving expression of its coding sequence. Plasmids were transformed into an ERG25 promoterdisrupted haploid strain (ERG25:: perg25::HIS3 ) or an erg3::kan r null strain, respectively. The loss of ERG25 is lethal (50). The erg25:: perg25::HIS3 haploid strain lacks endogenous ERG25 promoter activity but is viable due to plasmid-driven ERG25 expression. Lovastatin-induced mRNA expression was quantitated using Northern analysis.
Lovastatin-induced ERG25 promoter activity was seen in cells containing or lacking all three SRE/AR1 c sites ( Fig. 2A,  lanes 1 and 2 versus 5 and 6). Promoter activity required ScUpc2/ScEcm22 ( Fig. 2A, lanes 1 and 2 versus 3 and 4; lanes 5  and 6 versus 7 and 8). Similar results were obtained for ERG3 promoter activity, although there was a reduction in gene induction in cells expressing the double mutant ERG3 promoter (Fig. 2B, lanes 3 and 4 versus 7 and 8). These results further strengthen the idea that not all consensus SRE/AR1 c sites are functional, as a plasmid-driven promoter lacking these elements still drives lovastatin-induced gene expression. The results also indicate that novel promoter elements function to regulate ERG gene expression.
ERG3 and ERG25 Promoters Contain AR1 b Sites-ERG3 and ERG25 promoter searches revealed the presence of novel variant SRE sites, all containing an identical single nucleotide change; the variant SRE sequence is TAAACGA rather than the consensus SRE/AR1 c site sequence, TATACGA, and is identical to the anaerobic response element AR1 b (15,16). We refer to this element as AR1 b for the remainder of this study.
To determine whether these AR1 b sites were functional, promoter lacZ-fusion assays were used. AR1 b sites were assayed in the absence or presence of SRE/AR1 c sites.
Deleting the three ERG25 promoters, AR1 b or SRE/AR1 c sites alone did not reduce lovastatin-induced activity (Fig. 3A, However, deleting all six resulted in an almost total loss of activity (Fig. 3A, sre Ϫ /ar1 c Ϫ ar1 b Ϫ versus SRE/AR1 c AR1 b ). In the case of ERG3, deletion of the two SRE/AR1 c or three AR1 b elements alone significantly reduced promoter activity (Fig. 3B, . Residual activity was abolished when both elements were deleted together (Fig. 3B, sre Ϫ /ar1 c Ϫ ar1 b Ϫ versus SRE/AR1 c AR1 b ). Importantly, promoter activities required ScUpc2/ ScEcm22 function (data not shown). Thus, AR1 b sites in the ERG3/25 promoters harbor lovastatin-induced activity.
Not All SRE/AR1 Elements Drive Lovastatin-induced Expression-Pah1 catalyzes the conversion of phosphatidic acid to diacylglycerol (48,49). The PAH1 gene promoter contains a single AR1 b site that was tested for lovastatin-induced activity. PAH1 mRNA expression was determined using qRT-PCR, and protein level was determined by Western analysis, and phosphatidic acid phosphatase activity was determined using radiolabeled enzymatic assays. Studies were performed in wild type cells grown in the absence or presence of lovastatin.
No change in PAH1 mRNA expression was observed in the presence of lovastatin when compared with untreated cells (Fig.  4A, black bars versus white bars). This was in contrast to an increase in ERG25 expression upon drug treatment. Erg25 protein level also increased (data not shown), whereas the Pah1 protein level and phosphatidic acid phosphatase enzymatic activity remained constant (Fig. 4, B and C). Under our assay conditions, we cannot tell if the App1, Dpp1, and Lpp1 phosphatases contribute to the phosphatidic acid phosphatase activity seen (43,45,51). However, their contribution would be minimal, as no statistical differences were seen in mRNA expression, protein level, or activity, in the absence or presence of lovastatin.
The LCB1 promoter contains a single consensus SRE/AR1 c , although the LCB2 promoter contains two. These elements were found to be nonfunctional based on LCB1/LCB2-lacZ FIGURE 3. lacZ promoter fusion assays determining ERG3/ERG25/UPC2 SRE/AR1 c and AR1 b activities. A, activities of ERG25 promoter SRE/AR1 c and AR1 b deletion combinations were determined using ␤-galactosidase assays. Uppercase lettering indicates the presence of an element in the lacZ promoter fusion construct. Lowercase lettering indicates that an element is absent. Activity was determined in the absence (black bars) and presence of lovastatin (white bars). B, activities of various ERG3 promoter SRE/AR1 c and AR1 b deletion combinations were determined using ␤-galactosidase assays. C, activities of various UPC2 promoter SRE/AR1 c and AR1 b deletion combinations were determined using ␤-galactosidase assays. promoter fusion assays and Northern analysis (data not shown). Thus, not all SRE/AR1 c /AR1 b elements drive lovastatin-induced promoter activity.
Two UPC2 AR1 b Promoter Sites Drive Lovastatin-induced UPC2 Expression-UPC2 gene expression is induced when sterol biosynthesis is blocked (25). 3 Both ScUpc2 and ScEcm22 are required for lovastatin-induced expression. The UPC2 promoter contains two previously uncharacterized AR1 b sites ( Table 2). These sites were assayed for promoter activity using UPC2 promoter-lacZ assays. Cells were treated with lovastatin.
Lovastatin-induced activity was severely reduced by deletion of either AR1 b site alone when compared with the wild type UPC2 promoter (Fig. 3C, ar1 b Ϫ585 to Ϫ579, ar1 b Ϫ359 to Ϫ353 versus AR1 b ). All residual activity was lost when both AR1 b sites were deleted together (Fig. 3C, ar1 b Ϫ585 to Ϫ579  and ar1 b Ϫ359 to Ϫ353 versus AR1 b ).

Deletion of Endogenous SRE/AR1 c or AR1 b Sites Causes Defects in in Vivo Lovastatin-induced Aerobic Gene Expression-
To definitively show that AR1 b sites drive in vivo gene expression, strains were constructed harboring endogenous ERG promoters lacking either SRE/AR1 c /AR1 b sites alone or in combination. qRT-PCR analysis was used to quantify gene expression in the absence and presence of lovastatin.
The single consensus SRE/AR1 c site in the ERG1 promoter was first tested. When this site was endogenously deleted, Ͼ75% of the lovastatin-induced activity was lost, when compared with the endogenous wild type ERG1 promoter (Fig. 5A, sre Ϫ /ar1 c Ϫ versus AR1 c ). The endogenous ERG3 promoter SRE/AR1 c and AR1 b sites were next assayed. Promoter activity was induced in the presence of lovastatin when both SRE/AR1 c and AR1 b sites were present (Fig. 5B, SRE/AR1 c /AR1 b ). Loss of the AR1 b sites resulted in the near total loss of induction, although loss of SRE/AR1 c sites reduced activity by ϳ50% (SRE/AR1 c /AR1 b versus sre Ϫ /ar1 c /AR1 b , SRE/AR1 c /ar1 b ). This result correlates well with the Northern analysis data (Fig. 2). The loss of both endogenous elements mimicked loss of AR1 b alone (SRE/AR1 c /AR1 b versus sre Ϫ /ar1 c /ar1 b ).
Finally, the activities of the three ERG25 SRE/AR1 c and AR1 b promoter sites were determined. Deleting the three SRE/AR1 c or AR1 b sites alone almost completely abolished lovastatin-induced activity (Fig. 5C, SRE/AR1 c /AR1 b versus sre Ϫ /ar1 c /AR1 b , SRE/AR1 c /ar1 b ). Induced gene expression due to drug treatment required ScUpc2/ScEcm22 (data not shown). Thus, SRE/ AR1 c and AR1 b sites function within the context of their endogenous promoters.
Deletion of the AR1 b Sites in the UPC2 Promoter Results in Defects in Lovastatin-induced Aerobic Gene Expression-If the AR1 b sites in the UPC2 promoter harbor endogenous activity, then they should drive lovastatin-induced expression of ERG genes in trans, through cis up-regulation of UPC2 expression and protein. To test this hypothesis, lovastatin-induced UPC2 and ERG1/3/25 expressions were determined in ecm22 null cells harboring a UPC2 locus-specific integrated wild type or mutant UPC2 promoter lacking AR1 b sites. These strains lack Ecm22 protein and contain only the level of Upc2 expressed due to endogenous UPC2 promoter activity.
The wild type UPC2 promoter induced UPC2 expression in the presence of lovastatin, although the mutant promoter could not (Fig. 6A, AR1 b ,1/AR1 b ,2 versus ar1 b ,1/ar1 b ,2). The mutant promoter lacked the ability to induce Upc2 protein levels in the presence of the drug, when compared with the UPC2 wild type promoter, which is a critical event required for maintaining normal antifungal drug sensitivity (Fig. 6B, AR1 b ,1/AR1 b ,2 versus ar1 b ,1/ar1 b ,2). Moreover, lovastatin-induced ERG1/3/25 expression remained at basal levels in cells harboring a mutant promoter, although UPC2-dependent activity was observed in cells carrying a wild type promoter (Fig. 6C, WT versus mutant). Moreover, the loss of Upc2-driven ERG25 expression resulted in the loss of Erg25. Thus, the two AR1 b sites in the UPC2 promoter were absolutely required for cis induction of UPC2 3 C. Gallo-Ebert and J. T. Nickels, unpublished data. A, wild type-(SRE/AR1 c ) or sre/ar1 c (sre Ϫ /ar1 c )-deleted ERG1 promoter fragment was transformed into wild type cells at the endogenous ERG1 promoter locus. Relative ERG1 expression was determined by qRT-PCR in the absence (black bars) or presence (white bars) of lovastatin. B, wild type (SRE/AR1 c /AR1 b ) or various SRE/AR1 c -or AR1 b -deleted ERG3 promoter fragments were transformed into wild type cells at the endogenous ERG3 promoter locus. Relative ERG3 expression was determined by qRT-PCR in the absence (black bars) or presence (white bars) of lovastatin. C, wild type (SRE/ AR1 c /AR1 b ) or various SRE/AR1 c -or AR1 b -deleted ERG25 promoter fragments were transformed into wild type cells at the endogenous ERG25 promoter locus. Relative ERG25 expression was determined by qRT-PCR in the absence (black bars) or presence (white bars) of lovastatin (Lov). Uppercase lettering indicates the presence of an element. Lowercase lettering indicates the absence of an element. expression and protein and trans activation of ERG gene expression in response to antifungal treatment.
UPC2 AR1 b Sites Are Required for Maintaining Normal Susceptibility to Antifungal Drugs-If the UPC2 AR1 b elements are critical for responding normally to antifungal assault, their loss should increase drug susceptibility. To test this hypothesis, a upc2⌬ ecm22⌬ strain harboring an endogenous mutated UPC2 promoter (upc2 ar1bϪ ) was tested for susceptibility to several antifungal drugs; the mutated strain contains a basal level of Upc2, but it cannot induce Upc2 upon drug treatment. Wild type, upc2 ecm22, and upc2 ar1bϪ strains were tested for growth against amphotericin B (ergosterol-binding agent), terbinafine (targets the Erg24, C-14 sterol reductase), itraconazole (targets the Erg11, lanosterol 14␣-demethylase), and lovastatin (targets the Hmg1, HMG-CoA reductase). IC 50 values were determined after 24 h of treatment.
The upc2 ar1bϪ ecm22 cells had increased susceptibility to all antifungals tested, when compared with values obtained for wild type cells ( Table 3). The strain was 1.6-fold more sensitive to amphotericin B, 8-fold more sensitive to terbinafine, Ͼ45fold more sensitive to itraconazole, and ϳ5.0-fold more sensitive to lovastatin. These values were somewhere in between those seen for wild type and upc2 ecm22 strains. This makes sense, as the upc2 ar1b ecm22 strain does have a basal level of Upc2. The results indicate that the two AR1 b elements have a critical role in mounting a response to antifungal drug treatment, as strains lacking these elements show a higher susceptibility to several antifungals.
Pathogenic and Nonpathogenic Upc2 Directly Bind to SRE/ AR1 c and AR1 b Sites-To determine whether there was a direct physical interaction between Upc2 and SRE/AR1 c /AR1 b sites, an AlphaScreen assay (see "Experimental Procedures") was developed to examine the direct interaction between recombinant expressed Upc2 and SRE/AR1 c or AR1 b sites. Upc2s from pathogenic fungi (C. albicans and C. glabrata) as well as S. cerevisiae Upc2 were tested for binding. All Upc2 orthologs were highly conserved, and their binding to these elements strengthened the argument that Upc2 from multiple common pathogenic fungal species can bind to SRE/AR1 c /AR1 b sites (23).
Upc2 protein binding to SRE/AR1 c and AR1 b sites was first tested. A set concentration of 50 nM of each element was used, and Upc2 protein concentration was varied. A dose-dependent increase in binding to each element was observed with increasing concentrations of each Upc2 tested (Fig. 7, A-C, open and filled circles). Relative K m values were determined to be 1, 0.5, and 25 nM for ScUpc2, CaUpc2, and CgUpc2, respectively. In contrast, no binding was observed to a mutated promoter (TCAGATAA) (Fig. 7, A-C, filled diamonds).
Competition studies using "cold" SRE/AR1 c or AR1 b sites were next performed (no beads attached). In all cases, the respective cold element competed with its identically labeled SRE (Fig. 7, D-F, open and filled circles). Relative IC 50 values were determined for ScUPC2, CaUpc2, and CgUpc2, and all were in the low nanomolar range. In all cases, the mutated SRE was incapable of competing with the SRE/AR1 c or AR1 b sites for Upc2 binding (Fig. 7, D-F, filled diamonds). Thus, pathogenic and nonpathogenic Upc2 have similar binding affinities for both elements. These results indicate that fungal Upc2 proteins can directly bind to either element. The binding is highly specific for specific nucleotide sequences, as no binding was observed to a mutated SRE.
Endogenous Upc2 Binds AR1 b Elements-Finally, the extent of ScUpc2 binding to endogenous UPC2/ERG3/ERG25 Ar1 b elements was determined using ChIP (Fig. 8). A wild type strain expressing an endogenous C-terminal Myc-tagged Upc2 was used to perform the ChIP analysis. Upc2-Myc binding was determined in the absence and presence of lovastatin.
UPC2 AR1 b elements 1 and 2 acted as endogenous binding sites for Upc2 in the presence of lovastatin (Fig, 9, UPC2, AR1 b 1, AR1 b 2; black bar versus white bar). Upc2 bound to both Ar1 b elements in the ERG3 promoter (Fig. 9, ERG3, AR1 b 1 and 2; black bar versus white bar) as well as the three Ar1 b elements in  the ERG25 promoter (Fig. 9, ERG25, AR1 b 1, 2, and 3; black bar versus white bar). The highest degree of binding was seen for the ERG25 AR1 b ,1 element. Based on the results as a whole, the direct binding of Upc2 to AR1 b elements in the UPC2 promoter results in the cis up-regulation of UPC2 and the trans induction of multiple ERG genes, events that are required for inducing sterol biosynthesis in response to blocks in sterol metabolism caused by antifungal treatment.

DISCUSSION
The work presented establishes a novel function for the anaerobic AR1 b promoter element (11,15,16,25,53); it regulates ERG gene expression in response to blocks in ergosterol biosynthesis caused by antifungals, through regulating UPC2 gene expression. Evidence is presented for the first time showing that SRE/AR1 c elements function in vivo to regulate ERG gene expression. Moreover, this is the first report thoroughly examining SRE/AR1 c /AR1 b elements in the context of their endogenous promoters. We demonstrated that not all SRE/ AR1 c /AR1 b elements function in vivo, so only a subset of these promoter sites are critical for mounting an antifungal resistance response. Finally, we have shown that pathogenic CaUpc2 and CgUpc2 bind to AR1 b elements in vitro and that ScUpc2 binds to Ar1 b elements in vivo. Our results indicate that the UPC2 AR1 b elements are critical for initiating a normal response to antifungal assault.
In mammals, SREBPs bind promoters containing sequences other than the consensus SRE (54 -57). Alternative binding sites have been found in the rat farnesyl-diphosphate synthase promoter (58), the 3␤-hydroxysterol ⌬(14)-reductase promoter (59), and the HMG-CoA reductase promoter (60), suggesting the existence of additional SRE-like elements. SREBPs also bind some E-box elements (61). We now show that the AR1 b element is a functional binding site for Upc2 in response to antifungus-dependent blocks in ergosterol biosynthesis.
There are cases where promoter heterogeneity differentially regulates gene expression (62-64). Our data indicates AR1 b elements work in concert with certain SRE/AR1 c sequences to  The UPC2, ERG1, ERG3, and ERG25 promoters are shown that contain the corresponding numbers of promoter elements. AR1 c elements are represented by black circles. AR1 b elements are represented by black squares. Squares with an X through them are those deleted by the HIS3 allele and were not tested in the wild type strain. FIGURE 9. Determination of the degree of Upc2 binding to endogenous AR1 b elements. A wild type UPC2::UPC2-Myc strain was grown to exponential phase at 30°C in the absence (black bars) or presence (white bars) of lovastatin (Lov). Cells were treated with 1% formaldehyde and cross-linked. 1 mg of protein extract was used for immunoprecipitating ScUpc2-bound DNA. ScUpc2-DNA complexes were isolated, and the initial cross-linking was reversed. DNA was eluted from samples using the DNeasy kit (Qiagen). Purified DNA was used for PCR amplification of bound DNA. PCR products were resolved by 2% agarose gel electrophoresis. Binding was assayed in the absence (Ϫ) or presence (ϩ) of lovastatin. The figure is representative of an experiment performed in triplicate.
regulate ERG gene expression, all through directing Upc2/ Ecm22 binding. Our AlphaScreen and ChIP data substantiate this hypothesis and have demonstrated similar ScUpc2 binding affinities for either element, which extends to pathogenic CaUpc2 and CgUpc2. This sets up the opportunity to use the AlphaScreen assay as a high throughput screening tool aimed at developing novel antifungals, as Upc2 is a fungus-specific transcription factor (23).
The 7-bp SRE/AR1 c element found in many ERG genes has been shown to regulate sterol gene expression (11, 14 -16, 25), and it was later found to be required for regulating DAN/TIR gene expression under anaerobic conditions (15,16). In vitro studies have demonstrated a role for this element in lovastatininduced transcriptional induction, through its acting as a binding site for Upc2 in the ERG2/3/10 promoters (11,25). Our ERG3-lacZ and pRS416-ERG3 studies demonstrated that ERG3 SRE/AR1 c sites were functional. A reduction in endogenous promoter activity was seen upon loss of these sites. However, most endogenous ERG3 promoter activity came from AR1 b elements.
Leber et al. (13) identified an ERG1 promoter element consisting of two 6-bp direct repeats separated by 4 bp (AGCTCG-GCCGAGCTCG). Its deletion eliminated sterol-dependent ERG1-lacZ activity (13). Our in vivo studies demonstrated that a SRE/AR1 c promoter site upstream of this element was required for most but not all activity (ϳ25% remaining). Thus, residual activity most likely comes from the repeat sequence, as it was present in the endogenous ERG1 promoter designed.
Kennedy et al. (10) showed that the heme activator protein transcription factors Hap1/2/3/4 (66, 67), the yeast activator protein transcription factor Yap1, and the phospholipid transcription factor complex Ino2/4 regulate ERG9 gene expression. Hap2/3/4 is a heterotrimeric complex that functions as a transcriptional activator (68). A single putative E-box (CANNTG) was also identified (10), which is the binding element for the INO2/4 heterodimeric trans-activator complex (69). Thus, ERG9 expression is regulated by multiple diverse factors, consistent with the idea that yeast sterol biosynthesis as a whole is highly regulated by multiple promoter elements that respond to different environmental signals.
Microarray studies have shown that lovastatin treatment caused an increase in the level of UPC2 transcription but not ECM22 (52). Davies et al. (25) demonstrated that ScUpc2 binding to trans SRE/AR1 c promoter elements significantly increased following lovastatin treatment. Higher levels of ScUpc2 binding were seen at the ERG3 promoter upon a block in sterol biosynthesis, although ScEcm22 levels decreased with a concomitant decrease in ScEcm22 at the promoter. The binding of ScUpc2 to ERG3 AR1 b sites was not tested. In addition, determining whether this had any effect on expression was not determined.
Here, we show that increased UPC2 gene expression is exclusively controlled by two AR1 b elements. These elements drive the critical trans activation of ERG gene expression in response to antifungal treatment, where overexpression gives rise to antifungal resistance. The C. albicans UPC2 promoter contains two elements that are similar but not identical to the AR1 b element (65). Whether these elements function in vivo to reg-ulate ERG gene expression in response to antifungal assault still requires investigation. Studies investigating their function in the context of the endogenous promoter are needed before any significance can be attributed to these elements. Our results using S. cerevisiae suggest this will be the case, further validating the use of S. cerevisiae as an excellent model to study mechanisms leading to fungal pathogenicity (15,16).
We have discovered that S. cerevisiae anaerobic AR1 b elements in the UPC2 promoter act as binding sites for Upc2. We have demonstrated that an endogenous UPC2 promoter strain lacking these sites is unable to induce ERG gene expression in response to blocks in sterol biosynthesis, leading to hypersusceptibility to antifungal drug treatment. Moreover, we have shown that pathogenic Upc2 proteins can bind to the ScAR1 b element. Based on our results and those of others, we conclude that the AR1 b sites in the UPC2 promoter regulate global ERG gene expression. These elements regulate in vivo ERG gene expression in trans, an event that is critical for fungi to respond to blocks in sterol biosynthesis caused by antifungal agents.