HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Rognon, B. Articles by Sanglard, D. Search for Related Content PubMed PubMed Citation Articles by Rognon, B. Articles by Sanglard, D. Agricola Articles by Rognon, B. Articles by Sanglard, D.

Identification of promoter elements responsible for the regulation of MDR1 from Candida albicans, a major facilitator transporter involved in azole resistance

¹ Institute of Microbiology, University Hospital Lausanne, Rue du Bugnon 48, CH-1011 Lausanne, Switzerland
² Comenius University, Faculty of Natural Sciences, Department of Microbiology and Virology, 842 15 Bratislava, Slovak Republic

Correspondence
Dominique Sanglard
Dominique.Sanglard{at}chuv.ch

	ABSTRACT

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION NOTE ADDED IN PROOF REFERENCES

 
Upregulation of the MDR1 (multidrug resistance 1) gene is involved in the development of resistance to antifungal agents in clinical isolates of the pathogen Candida albicans. To better understand the molecular mechanisms underlying the phenomenon, the cis-acting regulatory elements present in the MDR1 promoter were characterized using a -galactosidase reporter system. In an azole-susceptible strain, transcription of this reporter is transiently upregulated in response to either benomyl or H₂O₂, whereas its expression is constitutively high in an azole-resistant strain (FR2). Two cis-acting regulatory elements within the MDR1 promoter were identified that are necessary and sufficient to confer the same transcriptional responses on a heterologous promoter (CDR2). One, a benomyl response element (BRE), is situated at position –296 to –260 with respect to the ATG start codon. It is required for benomyl-dependent MDR1 upregulation and is also necessary for constitutive high expression of MDR1. A second element, termed H₂O₂ response element (HRE), is situated at position –561 to –520. The HRE is required for H₂O₂-dependent MDR1 upregulation, but dispensable for constitutive high expression. Two potential binding sites (TTAG/CTAA) for the bZip transcription factor Cap1p (Candida AP-1 protein) lie within the HRE. Moreover, inactivation of CAP1 abolished the transient response to H₂O₂. Cap1p, which has been previously implicated in cellular responses to oxidative stress, may thus play a trans-acting and positive regulatory role in the H₂O₂-dependent transcription of MDR1. A minimal BRE (–290 to –273) that is sufficient to detect in vitro sequence-specific binding of protein complexes in crude extracts prepared from C. albicans was also defined. Interestingly, the sequence includes a perfect match to the consensus binding sequence of Mcm1p, raising the possibility that MDR1 may be a direct target of this MADS box transcriptional activator. In conclusion, while the identity of the trans-acting factors that bind to the BRE and HRE remains to be confirmed, the tools developed during this characterization of the cis-acting elements of the MDR1 promoter should now serve to elucidate the nature of the components that modulate its activity.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION NOTE ADDED IN PROOF REFERENCES

 
The yeast Candida albicans is an opportunistic pathogen that causes oral, vaginal or systemic infections in immunocompromised individuals and immunocompetent women (Coleman et al., 1993). The azole class of antifungals (for example fluconazole) is widely used to treat C. albicans infections. Fluconazole-resistant strains have emerged, especially in HIV-positive patients with recurrent candidiasis who have undergone prolonged fluconazole therapy (Ghannoum et al., 1996; Law et al., 1994; Marr et al., 2000; Rex et al., 1995). The molecular mechanisms proposed to explain azole resistance are diverse: (i) the cellular content of the azole target Erg11p (ergosterol biosynthesis), a 14-lanosterol demethylase, can be elevated, (ii) the affinity of Erg11p for azole antifungal agents can be decreased, (iii) the ergosterol biosynthetic pathway can be altered by inactivation of the sterol ^5,6-desaturase and (iv) cells can fail to accumulate these agents due to the upregulation of multidrug transporters belonging to the ABC transporters and the major facilitator (MF) families. Some azole-resistant strains upregulate the ABC transporters CDR1 (Candida drug resistance) and CDR2, while others upregulate MF MDR1 (multidrug resistance 1) (Sanglard & Odds, 2002; White et al., 1998).

Overexpression and gene inactivation studies have shown that MDR1 confers specific resistance to fluconazole, benomyl, cycloheximide, benzotriazole, methotrexate, 4-nitroquinoline-N-oxide (4-NQO) and sulfomethuron methyl (Ben-Yaacov et al., 1994; Fling et al., 1991; Goldway et al., 1995; Wirsching et al., 2000b). Disruption of MDR1 causes an increased susceptibility to methotrexate, cycloheximide and 4-NQO (Goldway et al., 1995; Sanglard et al., 1996). However, its inactivation does not increase susceptibility to benomyl, suggesting that other mechanisms are involved in this resistance (Goldway et al., 1995). MDR1 is usually not expressed at detectable levels in fluconazole-susceptible C. albicans isolates but its constitutive upregulation in some fluconazole-resistant isolates reveals a correlation between azole resistance and overexpression of MDR1 (Franz et al., 1998; Sanglard et al., 1995; White, 1997; White et al., 1997).

Little is known about the transcriptional regulation of MDR1 in C. albicans. However, a dramatic transient upregulation of MDR1 can be observed in azole-susceptible strains after treatment with toxic hydrophobic agents (benomyl, methotrexate, 4-NQO, o-phenanthroline), oxidizing agents [diamide, H₂O₂, tert-butyl hydrogen peroxide (T-BHP) and diethylmaleate] and the alkylating agent methyl methanesulfonate (Gupta et al., 1998; Harry et al., 2005). Clinical isolates with increased MDR1 expression have not undergone amplification of the MDR1 gene (Wirsching et al., 2000a). Sequence analysis of MDR1 promoters from clinical isolates compared to wild-type promoters did not reveal cis-acting mutations (Harry et al., 2005; Wirsching et al., 2000a). Thus MDR1 upregulation in clinical strains is most likely due to transcriptional activation resulting from mutations of factors acting in trans (Wirsching et al., 2000a). Deletion analysis of the MDR1 promoter, performed with a luciferase reporter gene, suggested that regulatory functions could be assigned to two separate regions: an element required for induction by benomyl was assigned to the region between –399 and –299 upstream of the start codon and another region located between –601 and –500 was responsible for MDR1 induction by the oxidative agent T-BHP.

In Saccharomyces cerevisiae, drug resistance is also mediated by membrane transport proteins belonging to the ABC transporter and major facilitator superfamilies and the different mechanisms controlling the expression of these multidrug resistant genes are relatively well understood (Sa-Correia & Tenreiro, 2002; Taglicht & Michaelis, 1998). The homologue of MDR1, FLR1, confers resistance to multiple drugs such as fluconazole, 4-NQO, cycloheximide, methotrexate and benomyl (Alarco et al., 1997; Broco et al., 1999; Nguyen et al., 2001; Tenreiro et al., 2001). Induction of FLR1 expression by the oxidizing agents H₂O₂, diamide, T-BHP and diethylmaleate, as well by the alkylating agent methyl methanesulfonate (Broco et al., 1999; Jelinsky & Samson, 1999; Nguyen et al., 2001), is mediated by Yap1p (also called Pdr4p or Snq3p), a basic leucine zipper (bZip) transcription factor (Nguyen et al., 2001; Tenreiro et al., 2001). The direct targets of Yap1p, principally genes involved in oxidative stress, have a common promoter element, the palindromic YRE (Yap1p response element: TTAC/GTAA) (Fernandes et al., 1997; Kuge & Jones, 1994; Toone & Jones, 1999).

In C. albicans, Cap1p, a bZip transcription factor structurally and functionally related to S. cerevisiae Yap1p, has been implicated in multidrug resistance and oxidative stress resistance (Alarco & Raymond, 1999; Wang et al., 2006; Zhang et al., 2000). It was also shown to confer fluconazole resistance in S. cerevisiae by transcriptionally activating FLR1 (Alarco et al., 1997), suggesting it has the same sequence specificity as Yap1p. Conversely, the C. albicans MDR1 promoter harbours the putative Cap1p-binding site (TTAGTAA at –532). Together these data indicate that mechanisms regulating MDR1 and its S. cerevisiae orthologue may be grossly similar. However, while in S. cerevisiae inactivation of Yap1p abolished FLR1 transcription (Tenreiro et al., 2001), Alarco & Raymond (1999) showed that disruption of CAP1 in the azole-resistant strain FR2 did not suppress but rather increased the level of MDR1 expression. In C. albicans Cap1p may act as both a repressor and an activator of MDR1 transcription, since experiments performed by Harry et al. (2005) demonstrating the importance of the YRE in the T-BHP response suggest that Cap1p is an activator of MDR1 expression in response to oxidative stress.

In this study, we used a reporter system to explore the transcriptional regulation of MDR1 in azole-resistant and azole-susceptible strains. The system has identified two different cis-acting regulatory elements and has permitted us to locate them at high resolution. The first, located between –296 and –260, was responsible for dramatic MDR1 upregulation in an azole-susceptible strain exposed to benomyl as well as for high constitutive expression in an azole-resistant strain. The second, which harbours two putative binding sites for Cap1p, is located between –561 and –520. Like the trans-acting Cap1p, this cis-acting element was not required for high constitutive expression of MDR1 in an azole-resistant strain. However, the response of MDR1 to H₂O₂ was dependent upon the presence of both Cap1p and its binding site within the MDR1 promoter. The benomyl-dependent response was also partially dependent upon the presence of Cap1p, raising the possibility that interactions may occur between complexes bound at the two regulatory elements.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION NOTE ADDED IN PROOF REFERENCES

 
Strains and media.
The C. albicans strains used in this study are listed in Tables 1 and 2. Strains were grown either in complete medium YEPD [1 % Bacto peptone (Difco), 0.5 % yeast extract (Difco) and 2 % glucose (Fluka)] or in minimal medium YNB [Yeast Nitrogen Base (Difco) and 2 % glucose (Fluka)]. For solid media, 2 % agar (Difco) was added to both media. Escherichia coli DH5 (Hanahan, 1985) was used as a host for plasmid construction and propagation. DH5 was grown in Luria–Bertani (LB) broth or LB agar plates supplemented with ampicillin (0.1 mg ml^–1) when required.

View this table: [in this window] [in a new window]   Table 2. C. albicans strains resulting from transformation with the constructs listed in Tables 3 and 4

Construction of promoter–lacZ fusions.
To generate MDR1–lacZ and CDR2–lacZ fusions used in drug-induction assays, the 999 bp MDR1 and 903 bp CDR2 promoters were first amplified by PCR using the Taq DNA polymerase (Roche), with primers MDR1-5'-999-KPN and MDR1-3'-PST or primers CDR2-KPN and CDR2-PST (Table 3). The generated fragments were cloned into pAU36 (Uhl & Johnson, 2001) previously digested with KpnI and PstI. The resulting plasmids (pBR1 and pDS295, respectively) were digested with either NruI for the MDR1–lacZ fusion or SnaBI for the CDR2–lacZ fusion prior to transformation into C. albicans, resulting in their integration at the MDR1 and CDR2 loci, respectively.

View larger version (26K): [in this window] [in a new window]   Fig. 1. Insertion of MDR1–lacZ fusions at the C. albicans MDR1 genomic locus. (a) Schematic representation of MDR1–lacZ fusions at the genomic MDR1 locus. The plasmids containing the MDR1 promoters (P) fused to lacZ were first digested with NheI and were integrated by double crossing over at the MDR1 genomic locus using sequences (UTR5' and UTR3') flanking MDR1 inserted at the BamHI restriction site of pAU36. Restriction maps of the MDR1 genomic locus are shown with and without insertion of the MDR1–lacZ fusion constructs. (b) Southern blot analysis of genomic DNA from strains derived from FR2 and transformed with different lengths of promoter. Genomic DNA was digested with NsiI and transferred onto a GeneScreen Plus nylon membrane. Hybridization of the restricted genomic DNA with the MDR1 labelled probe [represented in (a) by a black bar] yielded one fragment of 2.7 kb for the wild type strain and an additional fragment, of between 6.1 and 7.3 kb, for the yeast transformants, depending of the length of the inserted promoter. A double insertion of the plasmid into the C. albicans genome yielded an additional signal at approximately 8.4–9.7 kb.

DNA amplification by PCR.
DNA fragments used in cloning steps or as probes were generated by PCR with the Taq DNA polymerase (Roche) following the instructions of the manufacturer. The reaction mixture (final volume 100 µl) was composed of 10 µl of 10x Taq DNA polymerase buffer (Roche), 10 µl of MgCl₂ (Roche), 1 µl of each 100 µM forward and reverse primer, 4 µl of 10 mM deoxynucleoside triphosphates (Roche), 10 U of Taq DNA polymerase and 1 µl of 0.5–1 µg µl^–1 genomic DNA. PCR was performed under the following conditions: one cycle for 4 min at 94 °C, 1 min at 54–60 °C and 2 min at 72 °C; 30 cycles for 30 at 94 °C, 1 min at 54–60 °C and 2 min at 72 °C followed by one terminal cycle for 10 min at 72 °C. After analysis of 1/10 vol. of PCR products by electrophoresis on 1 % agarose gel, DNA fragments were purified with the QIAquick PCR purification kit (QIAGEN) and resuspended in 50 µl sterile water.

Southern blotting.
Southern blots were carried out as described previously (Sanglard et al., 1999). DNA probes used in this study were generated by PCR using the following pairs of primers: for the verification of genomic insertion of pMM43-derived plasmids at the LEU2 genomic locus, CALEU2-3 (5'-CACCACCACTAAAACCCCAACAAATCA-3') and CALEU2-5 (5'-ATGTCTGTTAAAACCCAAACCATTACT-3'); for insertion of pAU36-derived plasmids at the MDR1 locus, MDR1-5'-501-KPN (5'-GCGCAAAGGTACCCAATTTTCATTTTAGGAAATTTACCGAGTTT-3') and LacZ-rev (5'-ATGTCTGTTAAAACCCAAACCATTACT-3'). Radioactive DNA probes were generated by random priming using the Megaprime DNA labelling kit (Amersham). Radioactive signals were revealed by exposure to Kodak BioMax MR film (Amersham Biosciences).

Electrophoretic mobility shift assays (EMSAs).
Protein extracts from C. albicans strain CAF2-1 were prepared from 2 l cultures in YEPD grown to mid exponential phase until the density reached 1.5x10⁷ cells ml^–1. An equal quantity of cells was exposed to benomyl (35 µg ml^–1) for 30 min or left untreated as a control. The benomyl-treated or control cells were centrifuged for 20 min at 5500 r.p.m. and each cell pellet was washed with 10 ml extraction buffer [400 mM (NH₄)₂SO₄, 10 mM MgCl₂, 1 mM EDTA, 10 % (v/v), glycerol, 7 mM mercaptoethanol, 200 mM Tris/HCl pH 8]. Washed cells were centrifuged for 10 min at 5500 r.p.m. and the pellet obtained was resuspended in 10 ml extraction buffer supplemented with a protease inhibitor mix (Roche). Glass beads (10 g) were added to the cell suspension and the resulting mixture was disrupted for 5 min at 4 °C using a Macro bead-beater (Biospec Products). The homogenized solution was kept on ice for 30 min before centrifugation at 4 °C for 30 min at 13 000 r.p.m. The supernatant was transferred to a new tube and the cell pellet was reextracted with 5 ml fresh extraction buffer. The two supernatants were combined and the resulting solution was centrifuged for 1 h at 13 000 r.p.m. at 4 °C. The supernatant was recovered and ammonium sulfate was added to a final concentration of 0.3 g ml^–1. After dissolution of ammonium sulfate, the protein extract was centrifuged for 10 min at 13 000 r.p.m. and 4 °C. Precipitated proteins were resuspended in 250 µl protein buffer (20 mM HEPES, 5 mM EDTA, 7 mM mercaptoethanol, 1 mM PMSF and 20 % glycerol) supplemented with protease inhibitors (Roche). Protein concentrations of the crude extracts were determined using the Bradford method.

Double-stranded oligonucleotide probes were prepared by annealing 5' end labelled oligonucleotides with their complementary strands. The sequences of the probes are described in Figs 7 and 8. Single-strand oligonucleotides (5 pmol) were incubated at 37 °C for 45 min with 10 U T4 polynucleotide kinase (Roche) and 40 µCi (1.48 MBq) [-³²P]ATP (3000 Ci mmol^–1, 10 µCi µl^–1) in a final volume of 10 µl. The excess of [-³²P] ATP was eliminated using the Nucleotide Removal kit (Qiagen). The last elution step was performed with 100 µl of sterile water. After addition of the unlabelled oligonucleotides, the annealing of the complementary strands was performed by progressive temperature reduction after a 3 min treatment at 95 °C.

View larger version (14K): [in this window] [in a new window]   Fig. 7. Insertion of the BRE and HRE into the CDR2 promoter. BRE- (black rectangle) and HRE- (grey rectangle) containing sequences were first cloned upstream of the first 198 bp of the promoter of CDR2 fused to the Renilla luciferase reporter gene. This minimal promoter is lacking the DRE but still contains elements required for efficient transcription. The effects of addition of HRE and/or BRE to the CDR2 promoter were analysed either in DSY449 exposed to benomyl (a) or to H2O2 (c), or in FR2 (b). The results are given as relative increase of reporter activity after benomyl or H2O2 exposure in strain DSY449 or as specific activities for reporter expression, expressed in RLU, in strain FR2. The values are the means of single measurements obtained with three independent clones. Asterisks indicate the position of values that differ significantly (P0.05) from those obtained with the minimal CDR2 promoter in Mann–Whitney tests.

View larger version (51K): [in this window] [in a new window]   Fig. 8. Specific protein complexes are formed with the BRE. (a) EMSA of the BRE with protein extracts from CAF2-1 cells exposed or not exposed to benomyl. Labelled probes 1 and 2 correspond to the sequence of the BRE (–296 to –259), which is longer than the BRE-containing domain delimited by promoter deletion analysis, and to the promoter sequence of RTA4 (–165 to –123), respectively. Labelled probes were separated as described in Methods. Arrows indicate the position of specific and non-specific DNA–protein complexes and of free probes. Specificity of binding was assayed by competition with varying amounts of unlabelled probes 1 and 2. (b) EMSA of the BRE with transversed or deleted oligonucleotide probes. Cell extracts were obtained from CAF2-1 exposed to benomyl. The sequences of probes are shown at the bottom. Transversions are underlined whereas deleted nucleotides are outlined in grey. The BRE probe (probe 1) served as a positive control for DNA–protein complex formation.

Measurement of -galactosidase reporter activities.
Cells were grown overnight in 3 ml YEPD at 30 °C under constant agitation. Cultures were then diluted in 5 ml fresh medium to obtain a cell density of approximately 5x10⁶ cells ml^–1 and were regrown under agitation until the density reached 1.5x10⁷ cells ml^–1. At this point, 1 ml aliquots from cultures that were not exposed to drugs (azole-resistant strains FR2 or MMY412) were centrifuged at 4 °C for 5 min at 5500 r.p.m. Exposure of other strains (azole-susceptible strains DSY449 and MMY411) to different drugs was carried out for defined time periods at 30 °C with constant agitation. After drug exposure, cell density was measured and cultures were centrifuged at 4 °C and 5500 r.p.m. for 10 min. The cell pellets were washed with 1 vol. Z buffer (60 mM Na₂HPO₄, 40 mM NaH₂PO₄, 10 mM KCl, 1 mM MgSO₄) and centrifuged at 4 °C for a further 10 min. The washed cell pellets were resuspended in 150 µl Z buffer supplemented with 38.6 mM mercaptoethanol. After addition of 50 µl chloroform and 20 µl 0.1 % SDS, cells were disrupted by vortexing. To each aliquot was added 700 µl ONPG (1 mg ml^–1; Sigma) in Z buffer supplemented with mercaptoethanol and preheated at 30 °C. The mixtures were incubated at 37 °C until they developed a yellow colour (20 min to 3 h). The enzyme was inactivated by the addition of 500 µl 1 M Na₂CO₃. The mixtures were centrifuged at 13 000 r.p.m. for 10 min and the OD₄₂₀ was read. The -galactosidase activities were calculated using the following equation: -galactosidase activity (Miller units)=(1000xOD₄₂₀)/(txvxOD₆₀₀), where t is the time of reaction expressed in min and v is the volume of culture used in the assay expressed in ml.

Measurement of Renilla luciferase reporter activities.
Cells were grown overnight in 3 ml YEPD at 30 °C under constant shaking, then diluted in 5 ml fresh medium to obtain a cell density of approximately 1.5x10⁷ cells ml^–1. Benomyl (35 µg ml^–1) was added to the cell suspensions and cultures were incubated with constant shaking at 30 °C for a further 30 min. After drug exposure, cells were centrifuged for 10 min at 5500 r.p.m. The pellet was washed twice with 1 ml cold sterile water followed by 1 ml cold RLUC buffer (0.5 M NaCl, 0.1 M K₂HPO₄, 1 mM EDTA, 0.6 mM sodium azide). The final pellet was resuspended in 200 µl cold RLUC buffer supplemented with protease inhibitors (Roche) and an equal volume of chilled acid-washed glass beads was added. The cells were disrupted at 4 °C for 30 s using a Minibead beater (Biospec Products) and then centrifuged at 4 °C for 10 min at 13 000 r.p.m. The supernatants containing soluble proteins were stored at –20 °C or used directly for measurement of luciferase activities.

The relative luciferase activities of the different clones were measured using the Dual-Luciferase kit (Promega). Five microlitres of the crude protein extracts was added to 25 µl LAR II buffer according to the recommendations of the manufacturer and firefly luciferase activities were measured using a Lumat LB 9507 luminometer (E. G. and G. Berthold). Twenty-five microlitres of freshly prepared STOP and GLO buffer was then added to the reactions and Renilla luciferase activities were measured. Specific luciferase activities, expressed in Renilla luciferase units [RLU; i.e. U (µg protein)^–1] were calculated after determination of the protein concentration of cell extracts by the Bradford method.

Statistical analysis.
For the analysis of promoter deletions, a mathematical model was fitted using non-linear least squares. The choice of the model was based on the observation that, starting from a maximal response (H₁ in the equation), the response decreases to a minimum level (H₀) over a very short interval of length. The centre of this interval (A) represents the critical length for loss of response, while the width of this interval indicates the abruptness of the change (the slope B). The final equation was y=H₀+(H₁–H₀)/{1+exp[(A–x)/B]}, where y is the response and x the promoter length. For the mutant strains (cap1/), the same equation was used with parameters written in lower case (h₀, h₁, a, b). The fit of these equations contained eight parameters. The differences between the parameters determined for the wild type strain and the corresponding mutant were next tested using the generalized F test.

Comparisons between two sets of values (e.g. promoter bearing transversion or deletion versus wild type promoter) were performed using the nonparametric Mann–Whitney test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION NOTE ADDED IN PROOF REFERENCES

 
The upregulation of MDR1 occurring in clinical azole-resistant strains is mimicked by exposure to benomyl, H₂O₂ and diamide
To analyse the transcriptional activity of MDR1, the proximal 999 bp of the MDR1 promoter was fused in frame to the lacZ reporter gene from Streptococcus thermophilus (Uhl & Johnson, 2001). The promoter fusion (MDR1–lacZ) was integrated in the azole-susceptible mutant strain DSY449 (cdr1/) at the MDR1 locus as described in Methods. We chose to use this strain because the absence of CDR1 may reduce drug efflux and thereby enhance the sensitivity of drug-dependent MDR1 activation.

Upregulation of MDR1 can occur in C. albicans azole-resistant strains or in azole-susceptible strains exposed to diverse drugs including benomyl and H₂O₂ (Gupta et al., 1998; Harry et al., 2005). Therefore we tested the response of the reporter to a panel of compounds known as inducers of MDR1. In order to test the MDR1 response specificity, we also tested the response to oestradiol and fluphenazine, compounds known to induce CDR2 (de Micheli et al., 2002). For this purpose, a CDR2–lacZ reporter system was integrated in DSY449 in the same fashion as for the MDR1–lacZ construct. The relative increase of MDR1–lacZ and CDR2–lacZ reporter activities after treatment of C. albicans with these compounds is shown in Table 6. Upon exposure to benomyl, H₂O₂ or diamide, MDR1–lacZ reporter activities increased more than 10-fold as compared to untreated conditions, while no increase was observed for the CDR2–lacZ promoter fusion. In contrast, oestradiol and fluphenazine treatment specifically increased CDR2–lacZ reporter activities (by approximately 100-fold). Thus, the response of the two reporter genes to five different agents phenocopies the effects on the endogenous MDR1 and CDR2 genes. Exposure of C. albicans to methotrexate, sulfomethuron methyl and phenanthroline had no influence on transcription of either reporter. Based on these results, responses to the hydrophobic compound benomyl and to the oxidizing agent H₂O₂ were selected as readouts for the functional dissection of the MDR1 promoter. As the chemical properties of these drugs are different, we reasoned that they might reveal different modes of activation of the MDR1 promoter.

A single transformant containing each deletion construct was chosen for testing of reporter activity in the absence and presence of benomyl. As shown in Fig. 2(a), the reporter under control of the proximal 1459 bp of the MDR1 promoter exhibited a 26.1±4.4-fold increase of activity upon exposure to benomyl. Progressive deletions (99 to 141 bp) from –1459 to –302 (with respect to the ATG start codon) did not significantly alter the relative increase of reporter activity due to benomyl exposure. However, deletion of the region between –302 and –201 resulted in a 10-fold decrease in benomyl-dependent reporter activity, indicating that the minimal domain required for dramatic benomyl-dependent upregulation must lie within this region. This element was named BRE (for benomyl response element). To further delimit the BRE, small deletions of the region between nucleotides –302 and –222 were generated (Fig. 2b). Progressive 4 to 7 bp (–302 to –279) and 18 to 20 bp deletions (–279 to –222) revealed that benomyl-dependent MDR1 transcription was significantly lowered by removal of sequence downstream of –296. The benomyl-dependent increase in reporter activity was 22.9±3.1-fold for the construct containing the MDR1 promoter region terminating at –296. It decreased to 4.7±0.7- and 6.4±1.1-fold for constructs with promoters terminating at –290 and –286, respectively. A further decline of relative reporter activity to 2.1±0.5-fold at –279 was also observed. These data indicate that the 5' extremity of the BRE lies between –296 and –279. A residual twofold increase in benomyl-dependent transcription appears to be independent of the BRE and suggests that the basal MDR1 promoter can still respond to the drug.

View larger version (22K): [in this window] [in a new window]   Fig. 2. Identification of a domain in the MDR1 promoter involved in benomyl response. (a) Deletion analysis of the MDR1 promoter fused to the lacZ reporter system under benomyl exposure. Constructs with 5' deletions in the MDR1 promoter (shown on the left of the graph) were introduced into DSY449 (white bars) and MMY411 (cap1/, black bars). The presence of a region involved in benomyl response (BRE) in both strains is shown by a black box. Results are expressed as relative increase of reporter activity, which corresponds to the ratio between the specific activities measured under exposed and unexposed conditions. The values are the means of single measurements obtained with three independent clones. Values significantly (P0.05) different from those obtained with the 1459 bp full-length promoter in DSY449 are indicated by an asterisk. As outlined in Methods, all reporter activities shown (strain DSY449) were used to model an equation with the following terms: y=2.1+(25.4–2.1)/{1+exp[(292.9–x)/1.7]}. From this equation, it could be determined that the critical length for loss of the benomyl response was at –293, meaning that results obtained with promoter lengths below this critical length are significantly different from those obtained with longer lengths. For strain MMY411, the equation was y=2.1+(13.0–2.1)/{1+exp[(292.9–x)/1.7]}. The generalized F test established that the maximal responses (H1=25.4 and h1=13.0) were significantly different (P0.05) between the two strains. (b) Precise delimitation of the BRE. Successive 5' deletions (4 to 7 bp between –302 and –279 or 18 to 20 bp between –279 and –222) were constructed and analysed in DSY449 (white bars). The minimal domain containing the BRE is represented by a black box within each promoter. Asterisks indicate the position of values that differ significantly (P0.05) from those obtained with the promoter of 302 bp in Mann–Whitney tests. The presence of two cytosines at the 3' end of the KpnI restriction site gave promoter lengths of 1101 and 286 bp, obtained from PCR products of 1100 and 284 bp, respectively.

View larger version (20K): [in this window] [in a new window]   Fig. 3. The promoter domain responsible for the H2O2 response in the MDR1 promoter is distinct from the BRE. (a) Deletion analysis of the MDR1 promoter fused to the lacZ reporter system under H2O2 exposure. The presence of a domain implicated in H2O2 response (HRE) in both strains is indicated by a grey box. Results are expressed as relative increase of reporter activity, which corresponds to the ratio between the specific activities measured under exposed and unexposed conditions. The values are the means of single measurements obtained with three independent clones. Values significantly (P0.05) different from those obtained with the 1459 bp full-length promoter in DSY449 are indicated by an asterisk. Results obtained with strains DSY449 and MMY411 exposed to H2O2 allowed the formulation of the following equations: y=1.3+(10.7–1.3)/{1+exp[(540.4–x)/0.9]} and y=1.3. As explained in the legend of Fig. 2(a), the critical length for the loss of H2O2 response was at –540.4, meaning that results obtained with promoter lengths below this critical length are different from those obtained with longer lengths. Except for the variables H0 and h0 (=1.3) representing the minimal H2O2 responses in both strains, all parameters of both equations were significantly different (P0.05) as revealed by the generalized F test. (b)Precisedelimitation of the HRE. Successive 20 bp deletions from –598 to –501 were analysed in strain DSY449 exposed to H2O2 and used to delimit a domain (–560 to –520) containing the HRE, which is indicated by a grey box. Asterisks indicate the position of values significantly different (P0.05) from those obtained with the promoter of 598 bp in Mann–Whitney tests. The presence of one cytosine at the 3' end of the KpnI restriction site gave a promoter length of 561 bp obtained from a PCR product of 560 bp.

View larger version (17K): [in this window] [in a new window]   Fig. 4. The BRE is responsible for the high level of expression of MDR1 in the azole-resistant strain FR2. (a) Deletion analysis of the MDR1 promoter fused to the lacZ reporter in FR2 (white bars) and MMY412 (cap1/, black bars) strains. The domain (BRE) responsible for high reporter activity in both strains is represented by a black box. Specific -galactosidase activities (expressed in Miller units) are the means of single measurements obtained with three independent clones. Values significantly (P0.05) different from those obtained with the 999 bp full-length promoter in FR2 are indicated by an asterisk. As explained in the legend of Fig. 2(a), two equations could be established for all reporter activities obtained in the azole-resistant strain FR2: y=0.3+(57.2–0.3)/{1+exp[(285.8–x)/3.0]}, and the cap1/-derived mutant MMY412: y=0.3+(52.0–0.3)/{1+exp[(285.8–x)/3.0]}. From these equations, the critical length for loss of reporter activity was determined to be –286, meaning that results obtained with promoter lengths below this critical length are different from those obtained with longer lengths. The generalized F test established that all parameters of the two equations in strains FR2 and MMY412 were not significantly different (P0.05). (b) Precise delimitation of the domain responsible for the high reporter activity in strain FR2. Promoters with 5' deletions (4 to 7 bp between –302 and –279 and 18 to 20 bp between –279 and –222) between –302 and –222 were analysed. The domain containing the BRE is represented by a black box. Asterisks indicate values significantly different (P0.05) from those obtained with the promoter of 302 bp in Mann–Whitney tests.

View larger version (19K): [in this window] [in a new window]   Fig. 5. Effects of nucleotide transversions and deletions within the BRE of the MDR1 promoter. (a) Effects of nucleotide transversions within the BRE. The full-length MDR1 promoters with transversions T1 to T4 within the BRE were fused to the lacZ reporter gene and constructs were inserted into strain DSY449, which was exposed to benomyl, and into FR2 in single copies as outlined in Fig. 1(a). Hatched boxes representing transversions T1 to T4 are shown at the corresponding positions (left). Results are given as relative increase of reporter activity after benomyl exposure in strain DSY449 (middle) or as specific -galactosidase activities (expressed in Miller units) for reporter expression in strain FR2 (right). The values are the means of single measurements obtained with three independent clones. Values significantly different (P0.05) from those obtained with the wild type promoter in Mann–Whitney tests are indicated by an asterisk. (b) Effects of deletions within the BRE. Constructs with deletions D1 to D4 were fused to the lacZ reporter gene and analysed either in strain DSY449 exposed to benomyl (middle) or in strain FR2 (right). Deleted sequences and corresponding positions are represented by the crossed boxes for each promoter (left).

The partial retention of reporter activity in constructs lacking parts of the BRE may suggest that this regulatory region is composed of several discrete functional domains but we cannot rule out the involvement of other compensatory mechanisms. Moreover, while 5' deletions only delimit the 5' extremity of regulatory elements (i.e. –296, see Fig. 2b), the results presented above aid the localization of central elements of the BRE. Deletions D2 (–295 to –287), D3 (–286 to –278) and D4 (–277 to –269) all led to a decrease of benomyl-dependent MDR1 response and of reporter activity in strain FR2, thus indicating that they all contain nucleotides important for the functionality of the BRE. Since these three deleted regions are sequential in the MDR1 promoter, it can be concluded that at least all the nucleotides of the central deleted region D3 (–286 to –278) are part of the BRE.

Cap1p, a bZip transcription factor, acts as a trans-acting factor of MDR1 regulation
The presence of two Cap1p-like binding sites within the cis-acting HRE (TTAG/CTAA; at –549 and –532) suggested that Cap1p, a bZip transcription factor involved in oxidative stress, could play a role in regulating H₂O₂-dependent MDR1 expression by binding to these sequences. To further explore this hypothesis, promoter deletion analysis was undertaken in MMY411 (cap1/), a strain derived from DSY449 that lacks CAP1. For all promoter fragments tested, the reporter activities obtained in the absence or presence of H₂O₂ treatment were not significantly different in this strain. The relative levels of reporter activity after H₂O₂ exposure were between 0.8±0.1- and 1.7±0.6-fold higher than that recorded when the strain was not treated with H₂O₂ (Fig. 3a, black columns). Thus, H₂O₂-dependent expression is essentially completely lost in the cap1/ strain irrespective of whether the HRE is present. Therefore Cap1p is likely to be the main positive regulator involved in the response of the MDR1 promoter to H₂O₂. We also examined the benomyl-dependent response of reporters under control of MDR1 promoter fragments in the cap1/ strain. Inactivation of CAP1 did not prevent the different reporter constructs from responding to benomyl. Reporter constructs carrying at least 302 nucleotides of the proximal MDR1 promoter showed a benomyl-dependent increase in activity (13.0±5.1-fold, black columns Fig. 2a). However, since the level of the responses was generally twofold lower than that recorded in the corresponding wild-type strain (25.9±5.3-fold, white columns Fig. 2a), we cannot exclude the possibility that Cap1p plays a minor role in mediating the transient response to benomyl. Finally, we observed no significant differences between the specific reporter activities measured in the azole-resistant strain FR2 (58.2±5.5 U) and the corresponding cap1/ mutant MMY412 (51.3±7.5 U, Fig. 4a), indicating that Cap1p is not necessary for constitutive high expression of MDR1 in FR2.

Importance of the cis-acting HRE for MDR1 regulation
As the response of MDR1 to benomyl was partially reduced in the strain lacking CAP1, the deletion of a region (–552 to –521) within the HRE (delimited between –561 and –520) was performed in the full-length MDR1 promoter and reporter activities were measured under benomyl exposure in DSY449. As shown in Fig. 6(a), deletion of the region between –552 and –521 within the full-length MDR1 promoter did not alter the response to benomyl: the increase in reporter activity after benomyl exposure (25.7±5.0-fold) was not significantly different from that in the control (30.1±3.1-fold). The simultaneous deletion of the BRE and HRE resulted in a similar effect to the deletion of the BRE alone (2.1±0.5-fold versus 1.3±0.2-fold increase in reporter activity). Consistent with the results shown in Fig. 2(a), these results indicate that the upregulation of MDR1 under benomyl exposure is not dependent on the HRE. The HRE was also not involved in high constitutive expression of MDR1 in strain FR2. In this strain, reporter activities were identical in the presence (66.4±5.2 U) or absence (73.6±3.7 U) of the HRE (Fig. 6b). Moreover, reporter activities after the simultaneous deletion of BRE and HRE (23.6±2.6 U) were also similar to those observed in the absence of the BRE (16.3±0.7 U) (Fig. 6b).

View larger version (12K): [in this window] [in a new window]   Fig. 6. Effects of HRE deletions on MDR1 regulation. Nucleotide deletions within the HRE were generated in the full-length promoter or in the BRE-deleted MDR1 promoter. Nucleotide coordinates are given for each deletion. Each HRE deletion was analysed either in DSY449 exposed to benomyl (a) or to H2O2 (c), or in FR2 (b). Results are given as relative increase of reporter activity after benomyl or H2O2 exposure in strain DSY449 or as specific -galactosidase activities (expressed in Miller units) for reporter expression in strain FR2. The values are the means of single measurements obtained with three independent clones. Asterisks indicate the position of values significantly different (P0.05) from those obtained with the full-length promoter in Mann–Whitney tests.

In summary, our results show that the HRE is not essential for MDR1 upregulation by benomyl or in an azole-resistant isolate but is required for H₂O₂-dependent regulation of this gene. In the presence of H₂O₂, Cap1p is likely to act in trans on the MDR1 promoter through the HRE. While we cannot exclude an indirect role for Cap1p in the benomyl response, it does not act through either of the cis-acting elements that we have identified in MDR1.

The BRE and HRE are functional in a heterologous promoter
To further assess the cis-acting role of the BRE in response to benomyl, the element was placed in a minimal 198 bp CDR2 promoter lacking its own regulatory element (DRE). The CDR2 promoter is not responsive to benomyl and is therefore convenient for addressing the involvement of the BRE in this process. A BRE–CDR2 chimeric promoter was constructed and used with a Renilla luciferase reporter system. Because the sequences immediately flanking the BRE may modulate its function, a 50 bp MDR1 promoter sequence was used which spanned nucleotides –309 to –260. The reporter system was inserted at the LEU2 genomic locus of strains DSY449 and FR2. As shown in Fig. 7(a), the presence of the BRE conferred benomyl responsiveness to the minimal CDR2 promoter. Up to 12.0±2.1-fold increase in reporter activity could be measured in benomyl-treated cells as compared to the control, untreated cells. Activity was reduced however as compared to the native 592 bp MDR1 promoter (72.8±5.6-fold). When two copies of the BRE were fused to the CDR2 promoter, the increase in reporter activity (96.8±46.4-fold) was comparable to that of the MDR1 promoter. To assess the role of the HRE in the benomyl-dependent response, a 41 bp MDR1 promoter sequence spanning nucleotides –556 to –516, and containing the HRE with Cap1p-binding sites, was placed upstream of the BRE–CDR2 promoter. The increase in reporter activity induced by benomyl (17.6±1.6-fold) was not significantly different from that observed with the BRE–CDR2 promoter (12±2.1-fold). When the HRE alone was fused to the CDR2 promoter, almost no increase in reporter activity could be detected (1.6±0.5-fold), indicating that the HRE is not involved in the benomyl response of MDR1. In FR2, the BRE–CDR2 chimeric promoter mediated high reporter activity. As shown in Fig. 7(b), the specific luciferase activity obtained with the BRE–CDR2 fusion (1.9±0.45x10⁶ RLU) was only 2.5-fold lower than that obtained with the MDR1 promoter (4.9±1.0x10⁶ RLU). Insertion of two copies of the BRE upstream of the minimal CDR2 promoter gave reporter activities (5.8±1.6x10⁶ RLU) similar to those measured with the MDR1 promoter. Since the specific reporter activities measured with the HRE–BRE–CDR2 and the BRE–CDR2 fusions (1.7±0.3x10⁶ and 1.9±0.5x10⁶ RLU, respectively) were similar and only background reporter activity was detected with the HRE alone (<10⁴ RLU), the HRE was not considered to be involved in the high levels of expression of MDR1 in FR2. The data presented in Fig. 7 also contribute to a more precise localization of the functional BRE. Placing the 5'extremity at position –296 as previously determined, the 3' extremity can be proposed to be at position –260, since this corresponds to the position identified in the testing of BRE functionality.

We investigated whether the HRE could confer an H₂O₂-dependent response to the minimal CDR2 promoter in strain DSY449 (Fig. 7c). However, when the HRE alone was fused to the minimal CDR2 promoter, the relative increase in reporter activity (1.9±0.7-fold) was not different from the background activity increase measured with the basal CDR2 promoter (2.0±1.0-fold). The relative increase in reporter activities after H₂O₂ exposure obtained with the BRE–CDR2 promoter was similar (2.9±1.8-fold) to that with the HRE–CDR2 promoter. However, when the HRE was placed upstream of the BRE–CDR2 promoter, a 43.0±3.1-fold increase in reporter activity generated by H₂O₂ exposure was measured, which was 1.6-fold higher than with the MDR1 promoter (27.3±3.7-fold). Thus, consistent with our observation that deletion of either element reduced the H₂O₂-dependent response of the MDR1 promoter, both elements are required to convey significant H₂O₂-dependent upregulation to a heterologous promoter.

Trans-acting factors bind to the BRE
To assess the possible interaction between the BRE and putative transcription factors present in C. albicans protein extracts, EMSAs were performed with double-stranded probes corresponding to several regions of the BRE. In this experiment, it was possible to observe the formation of a DNA–protein complex with labelled probe 1 (–296 to –259; Fig. 8a), which contained the BRE delimited by 5' promoter deletion analysis and insertion into another promoter context (between –296 and –260). This complex was present at higher levels, as revealed by the intensity of the band, when a protein extract from benomyl-exposed cells was used in comparison to that observed with proteins from untreated cells. The protein–DNA complex was specific for the BRE, since competition with increasing amounts of unlabelled BRE resulted in the disappearance of the complex (Fig. 8a). Furthermore, in the presence of a molar excess of an unlabelled probe, corresponding to a promoter region of the RTA4 gene, the formation of the protein–DNA complex could be still observed. Finally, using this labelled RTA4 as a probe (probe 2), formation of a protein–DNA complex with a similar migration pattern was not observed.

Labelled probes harbouring transversions identical to those used in promoter analysis (T1 to T4: probes 3–6) were also incubated with C. albicans protein extracts. However, as illustrated in Fig. 8(b), no specific band could be detected, thus corroborating the negative effects of these transversions upon reporter activities. Moreover, EMSAs carried out with labelled probes harbouring internal deletions (deletion D5: probe 7, –282 to –277 and deletion D6: probe 8, –285 to –274) showed no specific binding (Fig. 8b). This result is consistent with the finding that deletions within this region strongly reduced reporter activities (Fig. 5b). These results show that the promoter regulatory region previously delimited (–296 to –260) corresponds to that which can bind sequence specific protein complexes in vitro.

To further delimit the BRE sequence necessary for the binding in tra