Nonidentical function of Upc2A binding sites in the Candida glabrata CDR1 promoter

Abstract

Increased expression of the Candida glabrata CDR1 gene, encoding an ATP-binding cassette membrane transporter, is routinely observed in fluconazole-resistant isolates of this pathogenic yeast. CDR1 transcription has been well-documented to be due to activity of the Zn₂Cys₆ zinc cluster-containing transcription factor Pdr1. Gain-of-function mutations in the gene encoding this factor are the most commonly observed cause of fluconazole hyper-resistance in clinical isolates. We have recently found that the sterol-responsive transcription factor Upc2A also acts to control CDR1 transcription, providing a direct link between ergosterol biosynthesis and expression of Pdr1 target genes. While this earlier work implicated Upc2A as an activator of CDR1 transcription, our further analyses revealed the presence of a second Upc2A binding site that negatively regulated CDR1 expression. This Upc2A binding site designated a sterol-responsive element (SRE) was found to have significant lower affinity for Upc2A DNA-binding than the previously described SRE. This new SRE was designated SRE2 while the original, positively acting site was named SRE1. A mutant version of SRE2 prevented in vitro DNA-binding by recombinant Upc2A and, when introduced into the CDR1 promoter, caused decreased fluconazole susceptibility and increased CDR1 expression. This negative effect caused by loss of SRE2 was shown to be Pdr1 independent, consistent with the presence of at least one additional activator of CDR1 transcription. The ability of Upc2A to exert either positive or negative effects on gene expression resembles behavior of mammalian nuclear receptor proteins and reveals an unexpectedly complex nature for SRE effects on gene regulation.

Introduction

Regulation of the lipid content of membranes is essential to ensure appropriate function of both the membrane barrier and the proteins embedded within [see Stieger et al. (2021) for a recent review]. The 3 major lipid constituents of eukaryotic membranes are phospholipids, sphingolipids, and sterols. The levels of these lipids must be coordinated and their distribution between the exofacial and cytosolic membrane leaflets are modulated, frequently by transporter proteins (Pomorski and Menon 2016). One of the best-characterized mechanisms is the inward movement of phospholipids by so-called flippase proteins that are P-type transporter proteins (Montigny et al. 2016). This inward movement of phospholipids is balanced by outward flow (flop) thought to be the purview of ATP-binding cassette (ABC) transporter proteins (Nagao et al. 2010). Appropriate regulation of the resulting asymmetric distribution of phospholipids across membranes is essential for the function of these key cellular structures.

Studies in the yeast Saccharomyces cerevisiae identified an ABC transporter called Pdr5 as an important mediator of the flop of the phospholipids phosphatidylethanolamine and phosphatidylcholine to the outer leaflet of the plasma membrane (Decottignies et al. 1998; Pomorski et al. 2003). A better-known role for Pdr5 is its action as a multidrug transporter [reviewed in Prasad and Goffeau (2012)]. We have found that loss of Pdr5 leads to endocytic defects as measured by increased levels of the high-affinity tryptophan transporter Tat2 in the plasma membrane (Johnson et al. 2010). This was interpreted as loss of normal plasma membrane phospholipid distribution disturbing the normal rates of endocytic retrieval of Tat2 from this membrane location (Beck et al. 1999; Hachiro et al. 2013).

In the pathogenic yeast Candida glabrata, the Pdr5 homolog is called Cdr1 and is a major determinant in resistance to the antifungal drug fluconazole (Sanglard et al. 1999). Transcription of CDR1 is controlled by the Pdr1 transcription factor, a Zn₂Cys₆ zinc cluster-containing protein (Vermitsky and Edlind 2004; Tsai et al. 2006). Nearly all clinically derived fluconazole-resistant strains of C. glabrata contain substitution mutations within the PDR1 gene that generate a hyperactive form of Pdr1 and overproduction of Cdr1 (Ferrari et al. 2009). We have recently shown that CDR1 transcription is also regulated by the action of the Upc2A transcription factor (Vu et al. 2019, 2021) which is better-known as the key factor controlling expression of genes encoding enzymes involved in ergosterol biosynthesis (Nagi et al. 2011). This finding provides a direct link between the ergosterol biosynthetic pathway and CDR1 expression in this pathogenic yeast.

Our earlier characterization of the Upc2A control of CDR1 identified a single positively acting binding site for this transcription factor located at position 619 bp upstream of the transcription start site of CDR1 (Vu et al. 2021). The DNA recognition sequence of Upc2A is referred to as the sterol-responsive element (SRE). Loss of this SRE from the CDR1 promoter reduced but did not eliminate fluconazole induction of the resulting mutant promoter. In addition, a strain containing a CDR1 gene with this mutant SRE in the upstream region was more susceptible to fluconazole inhibition. These findings supported our conclusion that the SRE acted as a positive regulatory site for Upc2A.

Our continued analyses of the CDR1 promoter led to the detection of a second SRE located 426 bp upstream from the first. We now refer to the most upstream SRE as SRE2 and the previously described element as SRE1. Comparison of Upc2A DNA-binding to SRE2 with its binding to SRE1 indicated that the affinity was much lower for SRE2 than SRE1. Mutation of SRE2 led to the unexpected finding that this Upc2A binding motif contributed negatively to CDR1 transcription in contrast to its positive SRE1 counterpart. The respective effect of each SRE was tightly linked to its DNA sequence as the replacement of SRE1 sequence with that of SRE2 led to increased negative regulation of CDR1 while replacing SRE2 sequence with that of SRE1 led to increased CDR1 expression. The differential affinity of Upc2A for these 2 SREs, coupled with their opposing actions on gene expression, illuminates a more nuanced regulatory influence of this transcriptional regulator on gene expression.

Materials and methods

Strains and growth conditions

Candida glabrata was grown in rich YPD medium (1% yeast extract, 2% peptone, 2% glucose). YPD media supplemented with 50 μg/ml nourseothricin (Jena Bioscience, Jena, Germany) and/or 2 mM methionine was used to select strain containing pBV65 and pBV133 vector derivatives (Vu and Moye-Rowley 2018). Synthetic complete media (2% glucose, 1 mM estradiol) without methionine was used to recycle the selection cassette on pBV65. All strains used in this study are listed in Table 1.

Table 1.

Strains used in this study.

Strain name	Parent	Genotype
KKY2001	ATCC2001	his3Δ::FRT leu2Δ::FRT trp1Δ::FRT
BVGC442	KKY2001	cdr1Δ::LEU2 his3Δ::FRT leu2Δ::FRT trp1Δ::FRT
BVGC446	BVGC442	CDR1::TRP1 KKY2001
BVGC449	BVGC442	mSRE1-CDR1::TRP1 KKY2001
BVGC452	BVGC442	mSRE2-CDR1::TRP1 KKY2001
BVGC455	BVGC442	dmSRE-CDR1::TRP1 KKY2001
BVGC458	BVGC446	CDR1::TRP1 pdr1Δ::loxP KKY2001
BVGC460	BVGC449	mSRE1-CDR1::TRP1 pdr1Δ::loxP KKY2001
BVGC462	BVGC452	mSRE2-CDR1::TRP1 pdr1Δ::loxP KKY2001
BVGC464	BVGC455	dmSRE-CDR1::TRP1 pdr1Δ::loxP KKY2001

Plasmid construction and promoter mutagenesis

CDR1 promoter reporters

Candida glabrata autonomous plasmids were derived from pBV133 carrying a nourseothricin resistant marker. The Escherichia coli β-galactosidase gene (lacZ) was amplified from pSK80 (Khakhina et al. 2018). CDR1 promoter DNA was amplified from the KKY2001 background, containing the −1 to −1,076 region relative to the ATG as +1. Mutations in the SRE1 region consisted of modifying the SRE core sequence from TCGTTTA to TaaTTaA. Mutations in the SRE2 region consisted of changing the SRE core sequence TCGTAAA to TaaTAtt. The SRE1 with 20 adjacent nucleotides in each side contained the −597 to −643 region while SRE2 with 20 adjacent nucleotides in each side contained the −999 to −1,045 region. Gibson assembly was used to introduce mutations in which overlapping primers were designed to contain the changes. All constructs were verified by Sanger sequencing.

CDR1 promoter replacements

The CDR1 promoter region [−1,715 to −1] was replaced with the LEU2 gene by homologous recombination to provide a recipient background for introduction of the promoter variants. Promoter replacement constructs were made by Gibson assembling the fragments from the immediate upstream and downstream regions flanking the CDR1 promoter region [−1715 to −1], the TRP1 cassette (selection marker), and wt, mSRE1, mSRE2, or dmSRE version of the entire CDR1 promoter region into the pUC19 backbone.

Candida glabrata transformation

Cell transformations were performed using a lithium acetate method (Gietz and Woods 2002). After being heat shocked, cells were either directly plated onto selective CSM agar plates (for auxotrophic complementation) or grown at 30°C at 200 rpm overnight (for nourseothricin selection). Overnight cultures were then plated on YPD agar plates supplemented with 50 μg/ml of nourseothricin. Plates were incubated at 30°C for 24–48 h. In case of chromosomal integration, individual colonies were isolated and screened by PCR for correct insertion of the targeted construct.

Quantification of transcript levels by RT-qPCR

Total RNA was extracted from cells using TRIzol (Invitrogen) and chloroform (Fisher Scientific, Hampton, NH) followed by purification with RNeasy minicolumns (Qiagen, Redwood City, CA). Five hundred nanograms to 1 μg total RNA was reverse-transcribed using an iScript cDNA synthesis kit (Bio-Rad, Des Plaines, IL). Assay of RNA via quantitative PCR (qPCR) was performed with iTaq universal SYBR green supermix (Bio-Rad). Target gene mRNA levels were normalized to transcript levels of 18S rRNA.

Spot test assay

Cells were grown in YPD medium to mid-log phase. Cultures were then 10-fold serially diluted and spotted onto YPD agar plates containing different concentrations of fluconazole (LKT laboratories, St Paul, MN). All agar plates were incubated at 30°C for 24–48 h before imaging was performed.

Electrophoretic mobility shift assays

DNA probes were amplified by PCR with biotinylated primers (IDT, Coralville, IA). The SRE1 (−532 to −732) and SRE2 regions (−920 to −1,119) of the CDR1 promoter and the HO promoter region (−787 to −957) (as a non-Upc2A-responsive control), relative to their respective ATG as +1, were prepared by PCR amplification. Both purifications of the recombinant Upc2A and the detailed electrophoretic mobility shift assay (EMSA) protocol were previously described (Vu et al. 2019). Hill plots were used to determine the apparent K_D for Upc2A-6X His binding. The apparent K_D represents the molar concentration of Upc2A-6X His [log(Upc2A-6X His), plotted on the x-axis] required to shift 50% (equivalent to 0 on the y-axis) of the specific promoter probes [log(F/(1—F)), plotted on the y-axis, where F is the fraction of each promoter probe shifted by Upc2A-6X His]. The values reported in the text for the apparent K_D are the average of 3 independent experiments.

DNase I protection assay

DNA probes were generated by PCR. To generate a 5′ [γ-³²P] singly end-labeled probe, one of the PCR primers was modified [5 Amino-MC6 (Integrated DNA Technologies, Coralville, IA)] at the 5′ end to prevent phosphorylation by polynucleotide kinase. Probes were end-labeled [1 pmol] using 10 µCi of [γ-³²P]-ATP (PerkinElmer, Waltham, MA) and 10 units of polynucleotide kinase (New England Biolabs, Beverly, MA) as instructed by the manufacturer. Unincorporated [γ-³²P]-ATP was removed using a size exclusion column (Qiagen). Samples were digested with DNase I (NEB [1:20 dilution]) for 30 s at room temperature. Binding reactions with purified protein, DNase I foot-printing, and DNA sequencing reactions were performed as previously described (Vu et al. 2021).

β-galactosidase assay

Harvested cells were lysed with glass beads (Scientific Industries Inc.) in breaking buffer (100 mM Tris pH8, 1 mM Dithiothreitol, and 20% Glycerol) at 4°C for 10 min. Lysate was collected and β-galactosidase enzymatic reaction was carried out in Z-buffer (60 mM Na₂HPO₄, 40 mM NaH₂PO₄, 10 mM KCl, 1 mM MgSO₄, 50 mM 2-Mercaptoethanol) with 650 µg/ml o-nitrophenyl-β-D-galactoside [ONPG]. Enzyme activity was quantitated as Miller units based on the equation: (OD₄₂₀ × 1.7)/(0.0045 × total protein concentration × used extract volume × time). The Bradford assay (Bio-Rad) was used to measure the total protein concentration in the lysate.

Chromatin immunoprecipitation

See Vu et al. (2019) for the detailed protocol. Mid-log cells were grown in YPD or YPD supplemented with 10 μg/ml fluconazole for 2 h before fixing with formaldehyde. Rabbit polyclonal Upc2A antibody was used to immunoprecipitate the fixed protein-DNA complex and real-time PCR was used to quantify the immunoprecipitated DNA. The % input method [recently discussed in Solomon et al. (2021)] was used to calculate the level of enrichment of Upc2A on promoter regions of each gene. Primers specifically targeting the ERG11 promoter (−694 to −561 relative to the ATG as +1), PDR1 promoter (−651 to −551) regions, were used. A region of the HO promoter CAGL0G05423g (Srikantha et al. 2005) was used as a negative control.

Antibodies and western immunoblot

Detailed protocol and antibodies employed have been previously described (Vu et al. 2019; Vu and Moye-Rowley 2022). The detected target band fluorescence intensity was normalized against tubulin fluorescence intensity. Data were compiled from 2 biological replicate experiments and 2 technical replicates in each experiment, giving 4 replicates in total.

Statistics

The Student’s T-test was used to assess the statistical significance of results of comparisons of samples. Paired conditions were used for comparisons of results from the same isolate obtained under different treatment conditions, while unpaired conditions were used for comparisons of results from isolates obtained under the same treatment conditions (*P < 0.5; **P < 0.01; ***P < 0.001).

Results

A second sterol response element in the CDR1 promoter

Work from multiple groups has identified CDR1 as a major transcriptional target gene for the Zn₂Cys₆ zinc finger-containing transcription factor Pdr1 (Vermitsky and Edlind 2004; Tsai et al. 2006; Vermitsky et al. 2006). Based on sequence inspection and some functional analysis (Paul et al. 2011), there are 5 different binding sites for Pdr1, referred to as Pleiotropic Drug Response Elements. More recently, we identified a Upc2A binding motif that acts to stimulate CDR1 expression and designated this as the SRE (Vu et al. 2019, 2021). The addition of the antifungal drug fluconazole that inhibits function of the ergosterol biosynthetic Erg11 enzyme [reviewed in Kelly et al. (2001)], led to an increase in levels of Upc2A bound to the CDR1 SRE. These data were most simply interpreted as Upc2A acted as a positive regulator of CDR1 expression in a manner very similar to that seen for the Pdr1 transcription factor.

However, further analysis of the CDR1 promoter indicated the presence of another Upc2A binding site, located further upstream from the site originally reported (Fig. 1a). To distinguish between the 2 SREs, we have designated the previously described SRE as SRE1, and the upstream SRE as SRE2 (Fig. 1a). SRE2 was found to be more than 1,000 bp upstream from the ATG of the CDR1 open reading frame. SRE2 was initially detected on the basis of sequence comparisons with other SREs from C. glabrata genes (Fig. 1b).

Fig. 1.

Characterization of the SRE2 in the CDR1 promoter. a) The relative arrangement of the sterol response element (SRE) SRE1 and SRE2 to the CDR1 transcription start site on C. glabrata chromosome XIII is shown. Numbering refers to the distance from the transcription start site which is indicated by the rightward-facing arrow. The positions of the predicted Pdr1 Response elements are also indicated. b) Nucleotide sequences from several different SREs are shown. The SRE2-specific nucleotide sequence is shown in black and white and the extent of DNA protected from cleavage by the DNase I mapping experiment is shown in gray. The previously reported SRE sequences in CDR1 (SRE1) and PDR1 promoters were included for comparison. c) DNase I protection of the CDR1 promoter is shown. The position of SRE2 is indicated by the bar at the right hand side and DNase I hypersensitive site is noted by the asterisk (*). G/A refers to the purine-specific reaction of Maxam-Gilbert chemical sequencing and was carried out on the same radioactive DNA fragment used in the DNase I reaction. Recombinant Upc2A-6X His was added (+) to the DNA probe or omitted (−) as indicated.

To confirm Upc2A binding to SRE2, DNase I protection assays with a recombinant form of the Upc2A DNA-binding domain were used to locate SRE2 at the nucleotide resolution. A radioactive probe was generated containing SRE2 from the CDR1 promoter and analyzed by DNase I protection (Fig. 1c). The DNase I ladders were electrophoresed in parallel with chemical sequencing reactions on the same probe in order to locate the SRE. The CDR1 probe showed a protected region of DNA and a strong DNase I hypersensitive site located immediately upstream of the SRE2 (Fig. 1c).

During the characterization of Upc2A binding to SRE2 by DNase I, we observed a stronger concentration dependence for DNA-binding than we had found in the case of SRE1 earlier (data not shown). To directly compare the ability of Upc2A to bind to SRE1 or SRE2, we used an EMSA. We confirmed that we could detect Upc2A: SRE2 complexes by EMSA by comparing protein: DNA complex formation using either a wild-type or mutant SRE2 probe (Fig. 2a). Only in the presence of the wild-type probe was a protein: DNA complex observed. These data support our assignment of the region from −1,025 to −1,016 as a binding site for Upc2A.

Fig. 2.

Upc2A binding to SRE1 and SRE2. a) EMSA analysis of Upc2A-6X His binding to wild-type and mutant SRE2 along with the nonspecific HO promoter control. Biotinylated probes were prepared from the indicated C. glabrata promoter regions containing either wild-type (wt) or mutant (mut) versions of SRE2. HO is from the promoter for this gene which is not a Upc2A target. Sequences of these different SREs are shown at the bottom of the panel with the altered residues in lower case. The presence or absence of Upc2A protein is indicated by the (+) or (−), respectively. Representative EMSA used to calculate the apparent K_D for Upc2A-6X His binding to the CDR1 promoter containing SRE1 (b) and SRE2 (c). Specific probes were incubated in the absence of Upc2A-6X His (0 nM) or with increasing concentrations of Upc2A-6X His (1.9–190 nM) for 20 min. d) Hill plots used to determine the apparent K_D for Upc2A-6X His binding. The apparent K_D represents the molar concentration of Upc2A-6X His [log(Upc2A-6X His), plotted on the x-axis] required to shift 50% (equivalent to 0 on the y-axis) of the specific promoter probes [log(F/1-F), plotted on the y-axis, where F is the fraction of each promoter probe shifted by Upc2A-6X His].

We then used this EMSA assay to determine the K_D (equilibrium dissociation constant) for Upc2A binding to both SREs to assess the protein: DNA interaction. We first added differing amounts of Upc2A to biotinylated oligonucleotides corresponding to SRE1. After the binding reaction was complete, free probe and DNA: protein complexes were separated via EMSA (Fig. 2b). This was also carried out using a similar probe corresponding to SRE2 (Fig. 2c). We determined the fraction of DNA bound vs the input protein and plotted these data as described (Fried 1989) to determine the K_D (Fig. 2d).

This analysis estimated the K_D for SRE1 to be 48 nM ± 10 nM while this same parameter was at least 420 nM ± 50 nM for SRE2. This indicates that Upc2A binding to SRE1 is much higher affinity than binding to SRE2. Given these differences in biochemical characteristics, we wanted to compare their roles in CDR1 regulation in vivo.

Phenotypes caused by loss of SRE2 function

To assess the roles of SRE1 and SRE2 in control of CDR1 transcription, we introduced mutant versions of each element into the CDR1 promoter of a CDR1-lacZ low-copy-number plasmid that we have described earlier (Vu et al. 2019, 2021). These mutant SREs were introduced singly or together and analyzed for their effect on expression of the CDR1-lacZ fusion. Transformants were either grown untreated or after 2 and 4 h of fluconazole challenge, followed by measurement of the β-galactosidase activity produced under these conditions (Fig. 3a).

Fig. 3.

Phenotypes of SRE2 mutation. a) Effects of SRE1 and SRE2 on the expression of lacZ gene fusions. Low copy-number plasmids containing translational fusions of CDR1 5′ regulatory region and the E. coli lacZ gene were generated between the wild-type (wt), mutant SRE1 (mSRE1), mutant SRE2 (mSRE2), or double mutant SREs (dmSRE) CDR1 promoter versions. These plasmids were introduced into wild-type cells, grown in the absence or presence of fluconazole (20 μg/ml) and then β-galactosidase activity determined. b) RT-qPCR analysis of CDR1, PDR1, and ERG11 genes transcript levels in cells having the isogenic wild-type (wt) mutant SRE1 (mSRE1), mutant SRE2 (mSRE2), or double mutant SREs (dmSRE) CDR1 promoter integrated into the normal CDR1 chromosomal location. Mid-log growth cells were treated with or without fluconazole (10 μg/ml) for 2 h before total mRNA was extracted for transcript analysis. c) Western immunoblot analysis of Cdr1, Pdr1, and Erg11 protein levels in cells treated with or without fluconazole (10 μg/ml) for 2 h. Tubulin and Ponceau S staining were included as loading controls. The presence of either the wild-type (wt) or mutant (m) SRE is indicated at the bottom of the panel. d) Quantification of band intensities detected in Figure 3C. The error bars and statistics were done based on data from 4 replicates. e) The strains described above were tested by serial dilution for their growth on rich medium (YPD) or the same medium containing fluconazole (4 and 8 μg/ml).

Each single mutant exhibited unique behavior in terms of its impact on expression. Compared to the wild-type promoter, loss of the SRE1 (mSRE1) caused a reduction in CDR1-dependent β-galactosidase activity, irrespective of the presence or absence of fluconazole. Conversely, loss of SRE2 (mSRE2) caused an increase in CDR1-lacZ expression under these conditions. Removing both SREs (dmSRE) produced an expression profile that was indistinguishable from the single mSRE2 mutant CDR1 promoter. We interpret these data to suggest that SRE1 provides a positive transcriptional effect on CDR1 while SRE2 is a negative regulatory input. Quantitatively, it appears that loss of SRE2 has a greater effect on CDR1 expression, even when SRE1 is simultaneously removed.

We moved these SRE mutant promoters back into the native locus in order to evaluate their effects on a more native form of CDR1. CDR1 promoter variants and the wild-type control were all integrated at the natural CDR1 locus with a TRP1 marker placed between the upstream SPT7 locus and 1,700 bp of DNA upstream from the CDR1 ATG at the start of the open reading frame (see Fig. 1a). Promoter designations are the same as were used for the CDR1-lacZ analysis.

These strains were first analyzed for the level of mRNA produced in the absence or presence of fluconazole (Fig. 3b). We compared expression of CDR1 with PDR1 and the ERG11 gene to determine if the changes in CDR1 expression would impact transcription of other genes involved in fluconazole resistance. When fluconazole was absent, levels of CDR1 mRNA were elevated upon removal of SRE2, either as a single mSRE2 mutant or the double dmSRE mutant promoter. These data confirm our findings using the CDR1-lacZ reporter gene above and argue that the effects of Upc2A on CDR1 transcription occur at the level of the CDR1 promoter. There was no observable effect of these CDR1 promoter mutants on CDR1 transcription when subjected to fluconazole challenge. In addition, no changes were caused in either PDR1 or ERG11 transcription, strongly suggesting that the alterations in CDR1 expression did not affect other fluconazole resistance genes.

These strains were also evaluated by western blotting using polyclonal antisera detecting Cdr1, Pdr1, and Erg11 (Fig. 3, c and d). Total protein extracts were prepared from strains grown with and without fluconazole, followed by analysis of the levels of these 3 proteins by western blot along with quantitation of these blots. As seen above for the mRNA, loss of SRE2 elevated Cdr1 protein levels in the absence of fluconazole challenge, irrespective of the presence or absence of SRE1. Fluconazole-induced protein levels were high in the presence of all CDR1 promoter versions tested except when single mutant mSRE1 promoter-containing allele of CDR1 was assessed. This reduction was suppressed in the dmSRE double mutant promoter context. Again, no changes were seen for either Pdr1 or Erg11.

Finally, the fluconazole susceptibility of these strains was assessed by plating serial dilutions of log phase cultures on rich media containing different levels of fluconazole (Fig. 3e). Loss of SRE1 caused a modest increase in fluconazole susceptibility while removal of SRE2 led to a clear decrease in susceptibility, even when combined with the mSRE1 mutation in the dmSRE double mutant.

These data support the view that Upc2A binding to SRE1 acts to stimulate CDR1 expression while binding of Upc2A to SRE2 is a negative transcriptional input. In order to correlate functional changes caused by SRE mutation with Upc2A binding, we carried out chromatin immunoprecipitation analysis.

Loss of normal SRE sequence blocks Upc2A binding in vivo

To confirm that the mSRE versions of each SRE prevented Upc2A binding as we previously determined in vitro (Figs. 1 and 2), we used the strains containing wild-type, mSRE1 or mSRE2 versions of the CDR1 promoter already described above. Chromatin was prepared and immunoprecipitated using anti-Upc2A antiserum. Immunoprecipitated chromatin was then assessed for the presence of DNA corresponding either to the SRE1 or SRE2 regions of the CDR1 promoter (Fig. 4a).

Fig. 4.

Upc2A binding to mutant SRE in vivo. Strains were treated without (Untreated) or with (FLC) fluconazole (10 μg/ml) for 2 h. Total sheared chromatin was prepared from all the strains and immunoprecipitated, after shearing, with anti-Upc2A polyclonal antibody. Immunoprecipitates were analyzed by qPCR using primers specific for the SRE1 and SRE2 regions in CDR1 promoter (a) as well as the SRE regions in PDR1 and ERG11 promoters (b). Fractional occupancies were calculated by comparing the ratios of each PCR product produced using immuno-purified chromatin to that produced from total chromatin (Product from total chromatin was set to a value of 1.0). Candida glabrata strains carrying the isogenic wild-type (wt), mutant SRE1 (mSRE1), or mutant SRE2 (mSRE2) CDR1 promoter in the chromosome were tested. An isogenic upc2AΔ strain was also included as negative control for Upc2A binding.

Binding of Upc2A to SRE1 was readily detectable in the wild-type CDR1 promoter and this was reduced in the presence of the mSRE1 mutation but also when the mSRE2 promoter mutation was present. In this latter CDR1 promoter version, SRE1 is intact but Upc2A binding to this element is reduced when SRE2 is mutated. Treatment with fluconazole induced Upc2A binding to SRE1 in the wild-type CDR1 promoter and the mSRE2 version. The mSRE1 version of CDR1 significantly lowered Upc2A binding to this mutant SRE1 sequence in the presence of drug.

Upc2A binding to SRE2 was diminished in the presence of the mSRE2 mutation but unaffected by loss of SRE1. SRE2 binding was also induced in both the wild-type and mSRE1 version of the CDR1 promoter by growth in fluconazole. The presence of the mSRE2 lesion blocked any induction of Upc2A binding to this mutant promoter. Upc2A binding to all versions of the 2 SREs was approximately 80% reduced in a upc2AΔ strain, indicating that even in the mSRE variants, some residual Upc2A interaction remained. These data argue that the mSRE mutant forms of the 2 SREs can be safely considered as significantly reducing Upc2A binding in vivo.

Upc2A DNA-binding to SREs located at either ERG11 or PDR1 was not affected by SRE mutations linked to CDR1 (Fig. 4b). Coupled with the lack of an effect on gene expression of ERG11 and PDR1, these data argue that the impact of the mSRE alterations in the CDR1 promoter is restricted to effects on CDR1 expression only.

The role of Pdr1 in the transcriptional effects of mutant SREs

As mentioned above, CDR1 gene expression is also strongly dependent on function of the Pdr1 transcription factor as well as Upc2A (Vu et al. 2019, 2021). To determine if the observed effects of the dmSRE on CDR1 transcriptional regulation occurred via changes in Pdr1, we removed the PDR1 gene from the strains containing different versions of the CDR1 SREs and analyzed the behavior of these mutant strains.

In the presence of a wild-type CDR1 promoter, loss of Pdr1 caused a large increase in fluconazole susceptibility (Fig. 5a) as has been seen previously (Vermitsky and Edlind 2004; Tsai et al. 2006; Vermitsky et al. 2006). Importantly, a pdr1Δ strain containing either the mSRE2 or the double mSRE1 mSRE2 CDR1 promoter, still exhibited a decrease in fluconazole susceptibility. Removal of SRE1 from the CDR1 promoter increased fluconazole susceptibility in the pdr1Δ background.

Fig. 5.

SRE mutant phenotypes in the absence of Pdr1. a) Isogenic wild-type and pdr1Δ strains carrying the wild-type (wt), mutant SRE1 (mSRE1), mutant SRE2 (mSRE2), or double mutant SREs (dmSRE) CDR1 promoter in the chromosome were tested by serial dilution for their growth on rich medium (YPD) or the same medium containing fluconazole (0.5 and 2 μg/ml). A strain wild-type for both the CDR1 and PDR1 genes was also included for comparison. b) RT-qPCR analysis of CDR1 and ERG11 gene transcripts during normal growth in pdr1Δ strains described above having the isogenic wild-type (wt), mutant SRE1 (mSRE1), mutant SRE2 (mSRE2), or double mutant SREs (dmSRE) CDR1 promoter integrated in the chromosome. c) Western immunoblot analysis of Cdr1, Pdr1, and Erg11 protein levels in strains described above. Tubulin and Ponceau S staining were included as loading controls. The presence of either the wild-type (wt) or mutant (m) SRE is indicated at the bottom of the panel. d) Quantification of band intensities detected in Fig. 5c. The error bars and statistics are done based on data from 4 replicates.

The retention of the behavior seen when Pdr1 was removed from the cell suggested that Pdr1 was not solely responsible for CDR1 expression alterations caused by the mSRE mutations. To directly confirm this suggestion, we measured the mRNA levels of CDR1 produced in these pdr1Δ derivatives of the SRE mutant promoters (Fig. 5b).

Even in these strains that lack the major regulator Pdr1, loss of SRE2 caused a significant increase in CDR1 expression, irrespective of the presence of SRE1. The presence of the mSRE1 CDR1 promoter still led to a significant decrease in CDR1 expression in the pdr1Δ background. Levels of ERG11 mRNA were not affected by any of these manipulations.

We confirmed these expression changes at the level of protein production via western blots with appropriate antibodies (Fig. 5c) and quantitated these changes (Fig. 5d). The expression of Cdr1 was strongly reduced upon deletion of PDR1 but this decrease was significantly suppressed upon removal of SRE2. This suppression occurred whether only SRE2 was mutagenized or if the mSRE2 was combined with the mSRE1 lesion in the double mutant. These strains all lacked Pdr1 as demonstrated by the absence of any detectable immunoreactive Pdr1 and produced levels of Erg11 enzyme that were unaffected by these manipulations. Together, these data argue for the presence of a positive regulator of CDR1 transcription in addition to Pdr1 and Upc2A.

Lack of positional dependence for SRE1 and SRE2 in CDR1

These analyses suggested that the individual effects of SRE1 or SRE2 on CDR1 expression were quite different, with SRE1 serving primarily a positive role and SRE2 acting negatively. To probe the basis of this differential function, we changed each SRE to replace its cognate partner. For example, we constructed a CDR1 promoter in which both SREs corresponded to either 2 copies of SRE1 or 2 copies of SRE2. Finally, we also switched the relative position of each SRE. In this derivative, we moved SRE2 to the position originally occupied by SRE1 and moved SRE1 to the location that originally corresponded to SRE2. All SREs were either duplicated or moved as a 47 bp element consisting of the 7 bp core SRE and 20 bp upstream and downstream of this site. These 3 new versions of the CDR1 promoter were placed in the context of a CDR1-lacZ fusion plasmid, introduced into wild-type cells and β-galactosidase activities determined as above.

Replacement of SRE2 with SRE1 (SRE1–SRE1) produced a form of the CDR1 promoter that was hyperactive compared to the wild-type version (Fig. 6a). The activity level of this SRE1-SRE1 CDR1 promoter closely resembled that of an mSRE2 CDR1 promoter. In striking contrast to the duplication of SRE1, the presence of a CDR1 promoter with 2 copies of SRE2 (SRE2–SRE2) was strongly repressed compared to the wild-type CDR1-lacZ fusion plasmid. Finally, exchanging the locations of SRE1 and SRE2 yielded a CDR1 derivative with expression indistinguishable from the wild-type version. Together, these demonstrate that each SRE has distinctive, nonidentical properties in the CDR1 promoter and these appear to be tightly linked to a <50 bp fragment containing the core Upc2A binding element at its center. Each SRE can be moved to at least one different position in the promoter while still maintaining its normal influence on gene expression.

Fig. 6.

Position independence of the CDR1 SREs. a) The effects of SRE1 and SRE2 alterations in the CDR1 promoter were tested using a CDR1-lacZ fusion gene carried on a low copy-number plasmid. These plasmids containing versions of the CDR1 promoter corresponding to the wild-type (wt), mutant SRE1 (mSRE1), mutant SRE2 (mSRE2), SRE2 converted into SRE1 (SRE1–SRE1), SRE1 converted into SRE2 (SRE2–SRE2), or rearranged SRE1 and SRE2 (SRE1-2 switched). Plasmids were transformed into wild-type cells, grown to mid-log phase and β-galactosidase levels determined as above. b) Proposed model of Upc2A regulation at each SRE in the CDR1 promoter. The top 2 lines represent the wild-type CDR1 promoter under conditions of ergosterol excess (erg+) or ergosterol limitation (erg−). The width of the bars indicating negative interaction has different thicknesses to denote either less (lighter) or more (darker) of a negative effect on expression. The bottom 3 lines represent the expression of CDR1 with either 2 copies of SRE1 (SRE1–SRE1), 2 copies of SRE 2(SRE2–SRE2) or with the positions of the 2 SREs exchanged (SRE1-2 Switched).

Discussion

Upc2A has been previously described as a homolog of S. cerevisiae and Candida albicans Upc2 and thought to be primarily a positive regulator of the expression of genes encoding enzymes involved in ergosterol biosynthesis as do its Upc2 counterparts in both S. cerevisiae and C. albicans (Crowley et al. 1998; Vik and Rine 2001; Silver et al. 2004). We provide evidence here that the role of Upc2A in C. glabrata is more complex than previously thought as this factor does not only activate but also represses gene expression. This bifunctional behavior of Upc2A resembles that seen in mammalian nuclear hormone receptors which also exhibit differential effects on transcription while binding to similar recognition elements in target genes (Lonard and O'Malley 2007; Meijsing et al. 2009).

We speculate that the differential behavior of Upc2A bound either to SRE1 (where it activates CDR1 transcription) with Upc2A bound to the negatively acting SRE2 may be due to changes in the conformation of the transcription factor when bound to these very different recognition elements. These 2 sites exhibit dramatically different affinities for Upc2A binding with SRE2 possessing a 10-fold higher dissociation constant for Upc2A compared to SRE1. Although our analysis is restricted to the 2 SREs in the CDR1 promoter, their distinct functional behavior suggests the possibility that other SREs in the C. glabrata genome may provide either a positive or negative influence on gene expression.

In our previous analysis of the UPC2A dependence of C. glabrata gene expression, we found 22 genes that were significantly induced at least 2-fold in the absence of UPC2A relative to the wild-type strain yet were bound by Upc2A as measured by ChIP-seq (Vu et al. 2021). The most highly induced gene corresponded to the C. glabrata homolog (CAGL0D04708g) of the Saccharomyces cerevisiae CTR1 gene that encodes the high-affinity copper transporter (Dancis et al. 1994). Previous work in C. glabrata has demonstrated the induction of this gene in response to different cell morphologies and suggest the transporter may be involved in the infection process (Srikantha et al. 2005). Interestingly, upc2Δ/upc2Δ homozygotes in C. albicans have been shown to have increased colonizing potential using a tail vein injection model (Vandeputte et al. 2011). Irrespective of the role of The C. glabrata Ctr1 homolog, its expression is clearly induced upon loss of Upc2A, consistent with the presence of other Upc2A-responsive genes being present in C. glabrata that are negatively regulated by this factor.

The impact of Upc2A on CDR1 is more complex than in the case of the C. glabrata CTR1 gene. Loss of Upc2A strongly elevated expression of CTR1 while CDR1 transcription is only modestly affected (Vu et al. 2021). However, fluconazole induction of CDR1 is nearly completely blocked in the absence of Upc2A. Our experiments demonstrate that there are 2 different Upc2A-dependent inputs to CDR1 transcription: 1 positive and 1 negative. Based on previous ChIP-seq data (Vu et al. 2021), we believe that SRE1 is occupied by Upc2A when sterol levels are relatively high. This occupancy contributes positively to expression of CDR1. When sterol concentrations are lowered (as during fluconazole exposure), nuclear Upc2A levels rise and the low-affinity SRE2 will fill. The data reported here indicate that the presence of SRE2-bound Upc2A acts to reduce the expression of CDR1. These differential behaviors of the different SREs are illustrated in Fig. 6b. In addition, when SRE2 is replaced with SRE1 sequence, the resulting SRE1–SRE1 CDR1 promoter was highly active. Similarly, when SRE1 is replaced with SRE2 sequence (SRE2–SRE2 CDR1 promoter), expression of the resulting gene was held at a level lower than the wild type. Switching the positions of the 2 SREs yielded a promoter that had similar expression to that of the normal CDR1 promoter. We interpret these data to support the view that action from either SRE is position and distance independent relative to the rest of the promoter and that the specific DNA sequence for each SRE is sufficient to confer the characteristic activity of each element.

These data suggest the need for a Upc2A-regulated biphasic induction of CDR1 during differing demands for sterols. When demand is relatively low, primarily SRE1 is occupied by Upc2A, which positively contributes to CDR1 expression. Upc2A occupancy at SRE2 is relatively low, owing to the reduced affinity for this element, but this site still contributes to maintaining CDR1 expression at a relatively lower level. Loss of SRE2 by mutagenesis elevated CDR1 expression under sterol replete conditions by more than 2-fold (see Fig. 3, a–d). This elevated CDR1 expression in the absence of SRE2 is likely responsible for the increased resistance shown by strains containing an mSRE2 mutant version of the CDR1 promoter (see Fig. 3e). This preinduction of CDR1, caused by loss of Upc2A binding to SRE2, is very likely a major driver of the elevated fluconazole resistance. This elevated constitutive CDR1 expression is routinely seen in the most commonly isolated fluconazole hyper-resistant mutants that map to a second major factor controlling CDR1 transporter called PDR1 (Vermitsky and Edlind 2004; Tsai et al. 2006; Vermitsky et al. 2006). Gain-of-function substitution mutant versions of PDR1 drive elevated CDR1 expression and this enhanced expression is required for the observed decrease in fluconazole susceptibility seen in these mutant strains [e.g. see Vermitsky and Edlind (2004), Tsai et al. (2006), Vermitsky et al. (2006), and Ferrari et al. (2009)].

Given the importance of Pdr1 in control of CDR1 transcription, we tested the dependence of the differential function of the 2 SREs for Upc2A on the presence of Pdr1 by removing this gene. It was possible that any effect of a given SRE might be solely through modulation of Pdr1 function. However, while expression of CDR1 was strongly diminished in the presence of the pdr1Δ allele, similar effects of the SRE mutations were observed. Most clearly, the presence of the mSRE2 allele led to induction of CDR1 expression and fluconazole resistance, even in a pdr1Δ background. We suggest that these findings argue for the presence of a transcriptional regulatory protein in addition to Pdr1 and Upc2A, although cannot eliminate that the loss of SRE2 leads to an increase in basal transcription from the CDR1 promoter.

The differential roles for SRE1 and SRE2 in CDR1 transcription are reminiscent of the behavior of binding sites for mammalian transcription factors like the glucocorticoid receptor and the NF-kB transcriptional regulators. While these factors all have characteristic binding sites, small differences in the precise recognition element can lead to large differences in the precise output from these elements. Coactivator requirements were found to change for NF-κB with single nucleotide changes (Leung et al. 2004). The different (but closely related) elements that are recognized by the glucocorticoid receptor have been argued to behave as ligands and alter the structure of receptor protein, depending on the precise sequence contacts of each site (Meijsing et al. 2009). Together, observations like these and our current findings for the differential behavior of Upc2A bound to either SRE on the CDR1 promoter, demonstrate that consideration of the sequence of a binding site alone should not be taken as predictive for the type of influence that site will have on a given promoter.

Finally, these data suggest a complex relationship between the expression of CDR1 and ergosterol biosynthesis. Expression of CDR1 is kept relatively high via Upc2A action at SRE1 when cells are replete for ergosterol. This is also a condition in which most Upc2A is excluded from the nucleus, assuming this factor is regulated like its homologues from S. cerevisiae (Marie et al. 2008; Yang et al. 2015) and C. albicans (Silver et al. 2004). Upon limitation for ergosterol, such as challenge with fluconazole, nuclear levels of Upc2A rise and binding to SRE2 increases. The inhibitory effect of Upc2A bound to SRE2 acts to restrict the level of CDR1 induction that can be seen when this binding is eliminated (see Fig. 3a).

We interpret these findings as indicating that Upc2A has a bifunctional role in control of CDR1 transcription in response to ergosterol levels. In high ergosterol conditions, Upc2A serves a primarily negative role in controlling CDR1 transcription. This can be evidenced by the derepression of CDR1, in the absence of ergosterol limitation, that occurs when the mSRE2 form of the promoter is present. Upon ergosterol limitation, Upc2A acts positively to drive elevated expression of CDR1 mRNA. It is important to note that Upc2A binding to both SRE1 and SRE2 is enhanced by ergosterol limitation but presently we do not have a clear picture of the link between binding and function of Upc2A. The balance between SRE1-mediated activation and SRE2-driven repression is important in ensuring the proper level of Cdr1 in the plasma membrane. Understanding the link between CDR1 expression and ergosterol levels in the cell will provide valuable insight into coordination of membrane protein expression and lipid homeostasis of these membranes.

Data Availability

Strains and plasmids are available upon request. The authors affirm that all data necessary for confirming the conclusions of the article are present within the article, figures, and tables.