Ahr1 and Tup1 Contribute to the Transcriptional Control of Virulence-Associated Genes in Candida albicans

Candida albicans is a major human fungal pathogen and the leading cause of systemic Candida infections. In recent years, Als3 and Ece1 were identified as important factors for fungal virulence. Transcription of both corresponding genes is closely associated with hyphal growth. Here, we describe how Tup1, normally a global repressor of gene expression as well as of filamentation, and the transcription factor Ahr1 contribute to full expression of ALS3 and ECE1 in C. albicans hyphae. Both regulators are required for high mRNA amounts of the two genes to ensure functional relevant protein synthesis and localization. These observations identified a new aspect of regulation in the complex transcriptional control of virulence-associated genes in C. albicans.

T he major human fungal pathogen Candida albicans can grow in different morphologies: unicellular yeast cells, pseudohyphae, and (true) hyphae (1). This morphological plasticity is crucial for fungal pathogenicity as formation of hyphae is involved in adhesion to and invasion of host cells and tissues, while yeast cells are required for dissemination within the bloodstream (2,3). Within host cells, hyphae and their associated proteins are involved in the acquisition of trace elements such as iron and zinc (4,5). The different growth forms of C. albicans exhibit distinct interaction patterns with host immune cells. Yeast cells are recognized and taken up by macrophages. However, some cells have the potential to survive and start to form hyphae. These hyphae are able to destroy macrophages by early induction of pyroptosis, host glucose consumption, and toxin production and later direct escape by physical forces (6)(7)(8)(9). In contrast, neutrophils are activated only by hyphae and are crucial for fungal killing (10). Consequently, neutropenic patients have a significantly impaired outcome for infections with C. albicans (11). Fungal polymorphism is also important for the interaction with dendritic cells but not natural killer cells (12,13). In recent years, genes ALS3 and ECE1 were found to encode key contributors to fungal virulence (14,15). Together with DCK1, HGT2, HWP1, IHD1, RBT1, and orf19.2457, they are also part of the core filamentation response (CFR), since their expression is induced whenever C. albicans forms hyphae, regardless of the environmental stimulus (16). ALS3 encodes a multifunctional protein involved in adhesion, invasion, and iron acquisition (4,14,17). The ECE1 product is normally the transcript with the highest abundance in C. albicans hyphae, independently of the environmental stimulus that triggered filamentation (16). The encoded protein is processed intracellularly by the proteinases Kex1 and Kex2 into eight peptides (15,18,19). One of these peptides, candidalysin, is secreted from hyphae into the environment and mediates host cell cytolysis (15). Candidalysin is involved in the immunopathology of Candida vaginitis (20) and drives protective innate type 17 cell responses during oral candidiasis (15,21).
Due to their importance for fungal virulence, knowledge about the regulation of ALS3 and ECE1 expression can contribute to a better understanding of fungal pathogenicity mechanisms. Both genes are downregulated in mutants lacking key activators of filamentation (16,22). In contrast, they are both upregulated in mutants lacking the hyphal growth repressors Tup1 and Nrg1 (23,24). Despite this, tup1Δ and nrg1Δ mutants exhibit attenuated virulence in several infection models (23,(25)(26)(27).
Here, we show that the transcriptional regulators Ahr1 and Tup1 play crucial roles in the activation of ALS3 and ECE1. While Tup1 contributes to repression and activation of the two genes, the presence of Ahr1 is important for high-level, functionally relevant expression of ALS3 and ECE1 in C. albicans hyphae.

RESULTS
Tup1 is required for activation of ECE1 and ALS3. We used RT-qPCR (reverse transcription followed by qPCR) to quantify the transcription of ECE1 in the wild-type strain and in mutants lacking the hyphal growth repressors Nrg1 and Tup1 in minimal medium with or without 10% human serum. As expected, the gene was barely expressed in wild-type yeast cells but was highly upregulated in hyphae (Fig. 1A). It was expressed at higher levels in the filamentous mutants than in the wild type under yeast growth conditions, as expected (Fig. 1A). Under hyphal growth conditions, however, ECE1 was expressed at significantly lower levels in the tup1Δ mutant than in the wild type, which was not the case for the nrg1Δ mutant (Fig. 1A). In fact, ECE1 transcription in the tup1Δ mutant remained at an intermediate level under yeast and hyphal growth conditions. This unexpected low level of expression was not caused by Nrg1 activity as a tup1Δ/nrg1Δ double mutant showed levels of ECE1 transcription similar to those seen with the tup1Δ single mutant (Fig. 1A). We used an established green fluorescent protein (GFP) reporter system for ECE1 to visualize expression in these strains and observed that the level of upregulation of ECE1 in the mutants under yeast growth conditions was not sufficient for a visible GFP signal (Fig. 1B). In accordance with the RT-qPCR data, nrg1Δ filaments showed bright GFP signals under hyphal growth con-ditions, while only weak fluorescence or no fluorescence was observed in the tup1Δ and tup1Δ/nrg1Δ mutants (Fig. 1B). The decreased ECE1 expression in the tup1Δ mutant was not observed exclusively for serum-induced filamentation, as similar results were observed for hyphae which were induced by either pH or medium change (see Fig. S1 in the supplemental material). These results obtained with the GFP reporter system and RT-qPCR correlated with a significantly lower level of release of candidalysin by the tup1Δ mutant than by wild-type hyphae (Fig. 1C). However, we also noted that the nrg1Δ mutant was not able to release as much candidalysin as the wild type, indicating that some other defects in these mutants prevented the translation of high ECE1 mRNA amounts into released candidalysin (Fig. 1C). ECE1 is not the only core filamentation response gene which requires the presence of Tup1 for its high-level transcription in C. albicans hyphae. Similar expression patterns were observed for ALS3 (Fig. 1D), HWP1, and IHD1 but not for the other CFR genes (Fig. S2). As seen with ECE1 and candidalysin, the intermediate levels of expression of ALS3 in the tup1Δ and tup1Δ/nrg1Δ mutants were not sufficient for surface localization of the Als3 protein in these mutants (Fig. 1E). Both strains were also not able to use ferritin as an iron source; such use is known to be dependent on the presence of the Als3 protein on the cell surface ( Fig. 1F) (3).
Overexpression of TUP1 does not inhibit hyphal morphology and expression of ALS3 and ECE1. Due to the surprising finding that Tup1 was required for high-level ALS3 and ECE1 expression, we studied the effects of induced overexpression of TUP1 FIG 1 Tup1 is required for high-level expression of ECE1 and ALS3. (A) Total RNA was isolated from the indicated strains after 6 h of growth in SDG with or without 10% human serum. ECE1 transcription was normalized against ACT1 and the control RNA (wild type, 5 h YPD, 37°C). Asterisks indicate significant transcription differences in a mutant compared to the wild type after growth in SDG with serum (P Յ 0.05, two-tailed, unpaired Student's t test). (B) The indicated strains with integrated pECE1-GFP cassettes were grown for 6 h at 37°C in SDG with or without serum prior to microscopy. Shown are the overlays of the DIC channel and the GFP channel. Scale bar, 20 m. (C) Candidalysin secretion was measured by LC-MS/MS after 18 h of growth in YNBS (pH 7.2). Candidalysin contents measured for wild-type hyphae were defined as 100%. (D) The total RNA isolated as described for panel A was used to determine the normalized relative expression levels of ALS3. Asterisks indicate significant transcription differences in a mutant compared to the wild type after growth in SDG with serum (P Յ 0.05, two-tailed, unpaired Student's t test). (E) After 6 h of growth in SDG with 10% human serum at 37°C, cells of the indicated strains were stained first with a monoclonal anti-Als3 antibody (pink signal) and then with calcofluor white (blue signal). Overlays of the images taken in the Cy5 and DAPI (4=,6-diamidino-2-phenylindole) channel are shown. Scale bar, 50 m. (F) From an overnight culture, 10 3 cells of the indicated strains were dropped on SDG medium with or without an iron chelator (bathophenanthroline disulfonate [BPS]) and with or without ferritin. The plates were grown for 3 days at 37°C in 5% CO 2 before images were taken.
Regulation of Candida albicans Virulence Factors ® and NRG1 using pTET constructs. Overexpression of NRG1, but not TUP1, prevented filamentation in a wild-type background ( Fig. 2A). Both pTET constructs were also integrated into the tup1Δ/nrg1Δ double mutant. Again, overexpression of NRG1 triggered filament-to-yeast reversion whereas overexpression of TUP1 did not affect filamentous growth ( Fig. 2A). Induced overexpression of NRG1 in the wild type led to decreased ALS3 and ECE1 transcript levels (Fig. 2B). In contrast, TUP1 overexpression enhanced the expression of both genes and even restored high mRNA levels of ECE1 in the tup1Δ/nrg1Δ mutant (Fig. 2B). Nonetheless, the presence of Tup1 was important for the repression of ECE1 by NRG1 overexpression (Fig. 2B), illustrating that it contributes to repression and activation of this gene.
Ahr1 is required for the expression of ECE1 and ALS3. In addition to the influence of putative repressors, we also examined the influence of putative activators on the expression of ECE1. The 5= intergenic region of the gene has a size of 3,197 bp and contains potential binding motifs for a variety of transcriptional activators which could at least in theory bind and regulate gene expression ( Fig. 3A; see also Table S1). To visualize ECE1 expression, GFP was integrated into the ECE1 loci of mutants lacking these transcription factors to monitor ECE1 expression levels. The resulting strains were then examined for GFP signals under hyphal growth conditions. The efg1Δ, cph1Δ/ FIG 2 Overexpression of TUP1 promotes ALS3 and ECE1 transcription. TET promoter-driven constructs of NRG1 and TUP1 were integrated into the ADH1 loci of wild-type strain SC5314 and the tup1Δ/nrg1Δ double mutant. (A) The morphology of the resulting strains was studied by microscopy after 6 h of incubation at 37°C in SDG or SDG with 10% human serum with or without the addition of 50 g/ml doxycycline. Scale bar, 20 m. (B) Total RNA of strains expressing either the pTET-TUP1 or the pTET-NRG1 construct was isolated after 6 h growth in SDG with 10% human serum. The pTET constructs were activated by the addition of 50 g/ml doxycycline to the medium. The isolated RNA was used for RT-qPCR to determine the relative levels of gene expression of ALS3 and ECE1. A control RNA (wild type, 5 h YPD, 37°C) and the housekeeping gene ACT1 were used for normalization. Asterisks mark significant differences after overexpression of either NRG1 or TUP1 compared to the same strain without overexpression (P Յ 0.05, two-tailed, unpaired Student's t test). The wild-type strain and regulatory mutants expressing the pECE1-GFP construct were grown for 6 h at 37°C in SDG with 10% human serum prior to microscopy. Shown are the overlays of the DIC and the GFP channels. Scale bar, 20 m. (C) The wild type and the ahr1Δ mutant were grown for 6 h at 37°C in SDG medium with serum prior to isolation of total RNA, which was used for the determination of relative gene expression levels. Asterisks indicate significant changes (P Յ 0.05, two-tailed, unpaired Student's t test) compared to the wild type. (D) Total RNA of the wild-type and ahr1Δ strains with or without AHR1 overexpression. The strains were grown for 6 h at 37°C in SDG with or without 10% human serum for the determination of relative gene expression levels. Asterisks indicate significant changes (P Յ 0.05, two-tailed, unpaired Student's t test) compared to the strain without AHR1 overexpression. (E) Cells of the indicated strains were grown for 6 h in SDG with 10% human serum at 37°C and were then stained with a monoclonal anti-Als3 antibody (pink signal), followed by a second staining with calcofluor white (blue signal). Shown are the overlays of the images taken in the Cy5 and DAPI channels. Scale bar, 20 m. (F) Levels of candidalysin secretion of the wild type and the ahr1Δ mutant were measured by mass spectrometry after 18 h growth in YNBS (pH 7.2). Candidalysin contents measured for wild-type hyphae were defined as 100%. (G) Cytotoxicity of the indicated strains was determined by LDH release from the infected TR-146 cells after a 24 h coincubation. Asterisks indicate significant changes (P Յ 0.05, two-tailed, unpaired Student's t test) compared to the wild type.
Regulation of Candida albicans Virulence Factors ® efg1Δ, ndt80Δ, and tec1Δ mutants were not able to form hyphae, and in accordance with this phenotype, the GFP signal was mostly absent or weak (Fig. 3B). Strains lacking BCR1 or BRG1 partially formed hyphae with GFP signals present (Fig. 3B). The hyphae of the cph1Δ and fkh2Δ mutants showed bright GFP signals (Fig. 3B). The ume6Δ mutant initially formed hyphae with the expected GFP signal, although the hyphal growth quickly reverted to yeast growth (Fig. 3B). The only mutant which was able to form morphologically normal hyphae without a GFP signal was the ahr1Δ mutant (Fig. 3B). RT-qPCR results confirmed the microscopic observations and showed that the level of ECE1 transcription in ahr1Δ filaments was indeed significantly lower than in wild-type hyphae ( Fig. 3C; see also Fig. S3). As seen with ECE1, the levels of transcription of ALS3, HWP1, and IHD1 were significantly reduced in ahr1Δ hyphae (Fig. 3C). Reintegration of AHR1 into the ADH1 locus, which also led to overexpression of the AHR1 gene, restored the high-level expression of ALS3 and ECE1 in ahr1Δ hyphae (Fig. 3D). Interestingly, pADH1-driven AHR1 induced upregulation of ALS3 and ECE1 in the wild type already under yeast growth conditions (Fig. 3D). As observed for the tup1Δ mutant, the intermediate expression levels of the ALS3 and ECE1 genes in the ahr1Δ mutant correlated with an absence of Als3 protein surface localization (Fig. 3E) and low secretion of candidalysin (Fig. 3F). In agreement with these findings, the level of virulence of the mutant in an oral epithelial cell infection model was found to be attenuated (Fig. 3G).
MCM1 overexpression can induce the expression of ECE1 and ALS3 in the presence of Ahr1 and Tup1. Ahr1 is known for its interaction with Mcm1, a transcription factor which itself also binds to the ECE1 promoter (28,29). As deletion of this essential transcriptional regulator is not feasible, we analyzed the effects of MCM1 overexpression under the control of the ADH1 promoter in the wild-type and ahr1Δ strains, each carrying the ECE1-GFP reporter system. This overexpression resulted in observations of GFP signals under yeast and hyphal growth conditions in the wild-type background (Fig. 4A). In the ahr1Δ mutant, however, the MCM1 overexpression led to a GFP signal in hyphae only (Fig. 4A). RT-qPCR showed that MCM1 overexpression triggered a strong upregulation of ECE1 in the wild-type background under both growth conditions (Fig. 4B). This effect of the overexpression was much lower in the absence of Ahr1 under yeast growth conditions (Fig. 4B). However, under hyphal growth conditions, MCM1 overexpression significantly increased ECE1 transcription in the ahr1Δ mutant, although not up to wild-type hyphal levels (Fig. 4B). Due to these observations, we studied the effects of MCM1 overexpression in the filamentous tup1Δ mutant and the nonfilamentous cph1Δ/efg1Δ double mutant. MCM1 overexpression partially rescued filamentation in the cph1Δ/efg1Δ double mutant (Fig. 4C), which was associated with upregulation of ECE1, although not to the level found in wild-type hyphae (Fig. 4D). In contrast, it was unable to compensate for the absence of Tup1 either morphologically or regarding ECE1 expression, which remained at intermediate levels (Fig. 4D). Similar gene expression patterns were observed for ALS3, HWP1, and IHD1 (Fig. S4).
Hyperactive Ahr1 induces ALS3 and ECE1 expression without environmental stimuli by direct binding to the target promoters. It was previously shown that a hyperactive version of Ahr1 can trigger filamentation in C. albicans (30). This construct, consisting of an AHR1 allele which was fused to a Gal4 activator domain (GAD) and a 3ϫ hemagglutinin (HA 3 ) tag, was integrated into the ADH1 locus of a wild-type derivative already carrying the pECE1-GFP construct. The resulting strain showed a bright GFP signal already under yeast growth conditions (Fig. 5A). Independently of the environment, the expression of ECE1 and ALS3 in the wild type with hyperactive Ahr1 was as high as in normal wild-type hyphae (Fig. 5B). Thus, addition of the Gal4 activator domain further enhanced the effects of AHR1 overexpression (Fig. 3D). We also observed slightly increased Als3 protein signals in the wild type with hyperactive Ahr1 under yeast growth conditions (Fig. 5C).
To test whether induction of ECE1 was related to promoter binding of Ahr1, we used the wild-type derivative with the hyperactive Ahr1 to perform chromatin immunopre-cipitation sequencing (ChIP-Seq) experiments under yeast growth conditions. ChIP-Seq provided clear evidence of physical binding of the hyperactive Ahr1 to the promoters of ECE1 and ALS3 (Fig. 5D) but also to those of other CFR genes (Table S2). The Ahr1 binding motif was found to be identical to one previously described (31). A nontagged version of the hyperactive Ahr1 was used to validate the ChIP-Seq results as shown for the ECE1 neighborhood (Fig. S5).
Hyperactive Ahr1 binds to genes encoding hypha-associated transcription factors and induces their expression. Hyperactive Ahr1 also bound to the promoters of transcriptional regulator genes BCR1, BRG1, EFG1, TEC1, and UME6, and this binding correlated with their upregulation (Fig. 6; see also Table S2). We also detected binding to the promoter of EED1, which was associated with increased expression of the gene ( Fig. 6; see also Table S2). In addition, hyperactive Ahr1 bound to the promoter of TCC1, which encodes a Tup1-interacting protein ( Fig. 6; see also Table S2). However, physical attachment to the promoters of NRG1 and TUP1 was not detected. Binding of hyperactive Ahr1 to the promoter of the AHR1 gene itself might indicate a self-controlling feedback mechanism ( Fig. 6; see also Table S2).
Hyperactive Ahr1 induces ALS3 and ECE1 expression in the absence of Cph1 and Efg1 but depends on Tup1. In the next step, we integrated the allele coding for the hyperactive Ahr1 into the nonfilamentous cph1Δ/efg1Δ double mutant. This par- The wild type and the ahr1Δ mutant with integrated pECE1-GFP and with or without pADH1-MCM1 were grown in SDG medium with or without 10% human serum for 6 h at 37°C prior to microscopy. Scale bar, 20 m. (B and D) Total RNA from the wild-type and ahr1Δ mutant strains (B) and from the wild-type and tup1Δ and cph1Δ/efg1Δ mutant strains was isolated after 6 h growth and used to determine the relative ECE1 expression levels. Scale bars, 20 m. Asterisks indicate significant changes (P Յ 0.05, two-tailed, unpaired Student's t test) compared to the corresponding strain without MCM1 overexpression. (C) The indicated strains were grown for 6 h at 37°C in SDG medium with or without 10% human serum prior to microscopy. Scale bar, 10 m.

Regulation of Candida albicans Virulence Factors
® tially restored filamentation in the double mutant (Fig. 7A) and, more strikingly, induced a significant increase of ECE1 transcription to the levels in wild-type hyphae (Fig. 7B). In accordance with these observations, the cph1Δ/efg1Δ strain with hyperactive Ahr1 showed an increased secretion of candidalysin which was absent from the cph1Δ/efg1Δ supernatants (Fig. 7C). Hyperactive Ahr1 also triggered the upregulation of ALS3, HWP1, and IHD1 in cph1Δ/efg1Δ under both yeast and hyphal growth conditions (Fig. S6). This upregulation was sufficient to ensure a surface localization of Als3 protein on the cph1Δ/efg1Δ mutant cells under hyphal growth conditions (Fig. 7D). Cells of the double mutant with or without hyperactive Ahr1 were also able to invade human oral epithelial cells, although not as efficiently as the wild type (Fig. 7E). Despite a modestly increased level of lactate dehydrogenase (LDH) release of infected epithelial cells, hyperactive Ahr1 could restore virulence capacity only partially in the cph1Δ/ efg1Δ mutant as the level of cell damage was still much lower than for the wild type (B) After 6 h of growth, total RNA of the strains grown as described for panel A was isolated and used for determination of relative gene expression levels. Asterisks indicate significant changes (P Յ 0.05, two-tailed, unpaired Student's t test) in mutants with hyperactive Ahr1 compared to their background strains without the hyperactive allele. (C) Cells of the indicated strains were grown for 6 h in SDG with 10% human serum at 37°C and were then stained with a monoclonal anti-Als3 antibody (pink signal), followed by a second staining with calcofluor white (blue signal). Shown are the overlays of the images taken in the Cy5 and DAPI channels. Scale bar, 20 m. (D) ChIP-Seq shows direct binding of hyperactive Ahr1 to the promoters of ECE1 and ALS3. Genomic DNA used for ChIP-Seq was isolated from the wild-type strain with hyperactive Ahr1 after 6 h growth in SDG medium at 37°C. The binding peaks as shown in the IGB viewer are displayed.
Ruben et al. ® (Fig. 7F). This suggests that secretion of candidalysin alone is not sufficient for full virulence in this mutant.
Interestingly, the hyperactive Ahr1 was able to induce a significant increase of ECE1 transcription in mutants lacking other hyphae regulators, such as ndt80Δ, tec1Δ, and ume6Δ mutants (Fig. S7). However, the hyperactive Ahr1 did not enhance the expression of ECE1 and ALS3 in the tup1Δ mutant, indicating that the presence of Tup1 is essential for the upregulation of both genes by the hyperactive Ahr1 (Fig. 8).

DISCUSSION
The results of our study revealed a complex picture of how the virulence-associated genes ALS3 and ECE1 of C. albicans are transcriptionally regulated. Both genes have long been known for high mRNA abundance in hyphae, and we found that these high transcript levels in hyphae are indeed required for function, as even a drop to a more intermediate level of ECE1 expression led to the absence of candidalysin secretion. A similar correlation was found between ALS3 expression and localization of its encoded protein on the cell surface. This striking disparity between intermediate transcript levels and a lack of functional protein (surface Als3 or secreted candidalysin) may indicate posttranscriptional regulation of mRNA and/or protein levels that warrant further studies. Our study results show that Ahr1 and Tup1 are key contributors in this complex regulation of the two genes in the different C. albicans morphologies. The presence of both regulators is required to reach high transcription levels, revealing new functions FIG 6 Binding of hyperactive Ahr1 increases expression levels of regulatory genes. The wild-type strain with hyperactive Ahr1 was grown for 4 h in SDG medium at 37°C prior to the ChIP-Seq analyses. The binding of hyperactive Ahr1 to regulatory genes is displayed in IGB viewer images. After growth under the same conditions, total RNA of the wild type with or without hyperactive Ahr1 was isolated at the same time point and used for the determination of relative gene expression levels. Asterisks indicate significant changes (P Յ 0.05, two-tailed, unpaired Student's t test) in the presence of hyperactive Ahr1 compared to the wild type alone. for both of them. Here, we showed that Ahr1 binds directly to the promoters of ALS3 and ECE1 and induces their transcription. These findings expand the previous knowledge about Ahr1, which was thus far known for its involvement in white-opaque switching and in fungal modulation of the environmental pH but also in regulation of fungal adhesion and filamentation (28,(30)(31)(32). This novel role of Ahr1 might explain why the deletion mutant not only is attenuated in an oral epithelium infection model but also fails to induce pyroptosis in infected phagocytes (7,33,34), for which candidalysin plays an important role (8,35). According to the Candida Genome Database, orthologs of Ece1 can be found in the genomes of C. dubliniensis and C. tropicalis, which are the closest relatives of C. albicans (36). Ahr1 homologs can similarly be found in both fungi, hinting at the possibility that the regulatory mechanism described here is conserved in other Candida species as well.

Regulation of Candida albicans Virulence Factors
In previous works, native Ahr1 was C-terminally tagged and used for ChIP-based identifications of the binding motifs in the promoters of target genes (28,31). Although the hyperactive Ahr1 which we used displayed enhanced activity compared to a native version, we identified the same binding motifs, indicating that the C-terminal addition of the Gal4 activator domain did not lead to false-positive results.
In addition, a hyperactive version of Ahr1 was found to be bound to the promoters of several transcription factor genes, including BRG1, EFG1, TEC1, and UME6, indicating that it regulates a variety of processes which are linked with hyphal growth and virulence. Interestingly, the hyperactive Ahr1 can restore ALS3 and ECE1 expression in deletion mutants of these transcription factors, indicating that it has a central role in the regulation of the two genes. In this context, we also observed that hyperactive Ahr1 restored high-level ECE1 expression and candidalysin secretion in the nonfilamentous cph1Δ/efg1Δ mutant.
Hyperactive Ahr1 was not able to compensate for the absence of Tup1, in contrast to the aforementioned transcriptional activators. This illustrates the importance of the latter for high-level expression of ALS3 and ECE1. Our results also indicate that Tup1 not only participates in the repression of both genes in yeast cells but also is required for their full activation in hyphae. This activating role represents a new concept for C. albicans Tup1 but is in agreement with previous findings from Saccharomyces cerevisiae.  The wild-type strain and tup1Δ derivates with integrated pECE1-GFP and with or without hyperactive Ahr1 were grown for 4 h in SDG medium with or without 10% human serum at 37°C prior to microscopy. Shown are overlays of the DIC and the GFP channels. Scale bar, 10 m. (B) After the same time of incubation in the indicated media, total RNA of the indicated strains was isolated and used for RT-qPCR to determine the relative levels of gene expression of ALS3 and ECE1. Asterisks indicate significant changes (P Յ 0.05, two-tailed, unpaired Student's t test) in mutants with AHR1-GAD compared to their background strains without the hyperactive allele.
Regulation of Candida albicans Virulence Factors ® There, it was shown that Tup1 and its corepressor Cyc8/Ssn6 can contribute to both the repression and activation of several genes, including GAL1 (37)(38)(39)(40)(41). In C. albicans, Tup1 might act as a backbone for the transcriptional control of ALS3 and ECE1 and change its interaction partners depending on whether repression or activation of the genes were required. We also observed that overexpression of MCM1 induced the transcription of these two genes, an effect which was strongly dependent on Tup1 but less so on Ahr1. Further experiments will be required to determine whether Ahr1 can act alone or requires Mcm1 for its function and how Tup1 is involved in the recruitment of Ahr1 or of both transcription factors to their target promoters.
primers used for RT-qPCR are listed in Table S3 in the format R1-gene name for the forward primer and R2-gene name for the reverse primer.
Fluorescence microscopy. GFP fluorescence microscopy was performed with a Zeiss AxioObserver Z.1 microscope (Zeiss, Germany). The illumination time for the differential interference contrast (DIC) channel was 40 ms, and that for the GFP channel was 1,700 ms. The same illumination times were used for all samples to ensure that the GFP signals were comparable.
Immunofluorescence was performed with the anti-Als3 antibody and a Zeiss Z710 laser scanning microscope (Zeiss, Germany). Prior to microscopy, 1 ϫ 10 6 cells/ml of C. albicans strains were grown in SDG or SDG plus 10% human serum for 6 h in petri dishes with a glass bottom (Dish; Mobitec, Germany) at 37°C. After 6 h, cells were washed with 1ϫ phosphate-buffered saline (PBS) and stained with the anti-Als3 antibody (1:500 diluted in 1ϫ PBS) (11) for 60 min at room temperature. After an additional washing with 1ϫ PBS, cells were incubated with a secondary DyLight649 goat anti-rabbit antibody (Thermo Scientific, Germany) or goat anti-rabbit-488 IgG antibody (Jackson Immuno Research, USA). Afterwards, cells were fixed with Histofix (Carl Roth, Germany) for 5 min and washed three times with 1ϫ PBS. Finally, cells were stained with calcofluor white for 15 min. Z-stack image series of hyphae were taken using Zen software (Zeiss, Germany), and images were merged using the extended depth of focus settings.
Chromatin immunoprecipitation sequencing (ChIP-Seq). Two C. albicans strains with a hyperactive Ahr1 (with and without C-terminal HA 3 -tag) in the SC5314 wild-type background were used for ChIP-Seq analyses, which were performed by Active Motif (USA). Formaldehyde fixation of the strains was performed after 6 h of incubation in SDG medium at 37°C following the company's protocol. The strain without the HA 3 tag was used for antibody validation to determine false positives. ChIP-Seq reads were analyzed by Active Motif, resulting in 325 peaks after peak calling and filtering the results. Peak data were further analyzed by BioControl Jena GmbH. Using the chromosomal position of the peaks, neighboring genes on both strands were identified. A sequence of 500 nucleotides centering at the maximum of each peak region was used as input for the online motif analysis tool Meme-ChIP (v. 5.1.0, 50), resulting in a highly significant, centrally enriched motif. Further confirmation was done with MochiView version 1.46 (51) and the Integrated Genome Browser (IGB) (52).
Cytotoxicity and invasion assays. Monolayers of TR146 buccal epithelial squamous cells, grown in Dulbecco's modified Eagle's medium/F12 (DMEM/F12) (Gibco) with 10% heat-inactivated fetal bovine serum, were infected with 4 ϫ 10 4 fungal cells and incubated in serum-free DMEM/F12 for 24 h at 37°C in 5% CO 2 in 96-well plates. To determine cytotoxicity, lactate dehydrogenase (LDH) release in culture supernatants was measured using a cytotoxicity detection kit (Roche). Supernatant from Triton-X-treated, uninfected TR146 cells served as a full-lysis control, corresponding to 100% cytotoxicity. The experiment was repeated three times with biological triplicates in each round. The invasion assay was performed with the same cell line as previously described (27).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.

ACKNOWLEDGMENTS
We thank Gerald Fink, Alexander Johnson, Susan Noble, Peter Sudbery, Arnold Bito, and Jürgen Wendland for providing strains and plasmids. Scott Filler provided the monoclonal Als3 antibody. We are also thankful to Stefanie Allert and Rita Müller for help with the infection assays.
This work was financed by funds from the Deutsche Forschungsgemeinschaft (DFG) within the Collaborative Research Center CRC 124 FungiNet (project C3 to O. Kurzai and project C2 to J. Morschhäuser and S. Vylkova, project C1 to B. Hube, project C5 to I. D.