The histone acetyltransferase Hat1 facilitates DNA damage repair and morphogenesis in Candida albicans

Authors

  • Michael Tscherner,

    1. Medical University of Vienna, Christian Doppler Laboratory for Infection Biology, Max F. Perutz Laboratories, Campus Vienna Biocenter, Vienna, Austria
    Search for more papers by this author
  • Eva Stappler,

    1. Medical University of Vienna, Christian Doppler Laboratory for Infection Biology, Max F. Perutz Laboratories, Campus Vienna Biocenter, Vienna, Austria
    Search for more papers by this author
  • Denes Hnisz,

    1. Medical University of Vienna, Christian Doppler Laboratory for Infection Biology, Max F. Perutz Laboratories, Campus Vienna Biocenter, Vienna, Austria
    Search for more papers by this author
  • Karl Kuchler

    Corresponding author
    • Medical University of Vienna, Christian Doppler Laboratory for Infection Biology, Max F. Perutz Laboratories, Campus Vienna Biocenter, Vienna, Austria
    Search for more papers by this author

For correspondence. E-mail karl.kuchler@meduniwien.ac.at; Tel. (+43) 1 4277 61807; Fax (+43) 1 4277 9618.

Summary

Chromatin assembly and remodelling is an important process during the repair of DNA damage in eukaryotic cells. Although newly synthesized histone H4 is acetylated prior to nuclear import and incorporation into chromatin during DNA damage repair, the precise role of acetylation in this process is poorly understood. Here, we identify the histone acetyltransferase 1 (Hat1) catalysing the conserved acetylation pattern of histone H4 preceding its chromatin deposition in the fungal pathogen Candida albicans. Surprisingly, Hat1 is required for efficient repair of not just exogenous but also endogenous DNA damage. Cells lacking Hat1 rapidly accumulate DNA damages and switch from yeast-like to pseudohyphal growth. In addition, reduction of histone H4 mimics lack of Hat1, suggesting that inefficient H4 supply for deposition into chromatin is the key functional consequence of Hat1 deficiency. Thus, remarkably, we demonstrate that C. albicans is the first organism known to require histone H4 processing for endogenous DNA damage repair and morphogenesis. Strikingly, we also discover that hat1Δ/Δ cells are hypersusceptible to caspofungin due to intracellular reactive oxygen species induced by this drug. Hence, we propose that targeting this class of histone acetyltransferases in fungal pathogens may have potential in antifungal therapy.

Introduction

The nucleosome, which is composed of 146 base pairs of DNA wrapped around an octamer of the four core histones H2A, H2B, H3 and H4, is the basic unit of eukaryotic chromatin (Luger et al., 1997). Assembly of histones into nucleosomes is an essential step in many biological processes, including regulation of transcription, replication and repair of DNA damage (Kaplan et al., 2003; Saunders et al., 2003; Schwabish and Struhl, 2004; Franco et al., 2005; Tsukuda et al., 2005; Chen et al., 2008). Newly synthesized histones are transiently acetylated immediately after their synthesis in the cytoplasm, which is essential for their subsequent incorporation into chromatin (Ruiz-Carrillo et al., 1975; Sobel et al., 1995; Chen et al., 2008; Yang et al., 2011). During chromatin maturation the acetylation marks are removed (Annunziato and Seale, 1983). Acetylation of histones requires two types of histone acetyltransferases (HAT), referred to as type A and type B (Brownell and Allis, 1996). Nuclear type A HATs utilize only nucleosomal histones as substrates, whereas the type B HATs are defined by their specificity for free non-nucleosomal histones and their partial cytoplasmic localization (Brownell and Allis, 1996).

The founding member of type B HATs is the histone acetyltransferase Hat1 (Kleff et al., 1995). In Saccharomyces cerevisiae, Hat1 forms a complex with a subunit called Hat2 to bind and acetylate newly synthesized histone H4 at lysine 5 and 12 (H4K5ac and H4K12ac), which is a highly conserved acetylation pattern at the N-terminus of H4 (Sobel et al., 1995). Histone H3 then joins and the entire complex is shuttled into the nucleus (Kleff et al., 1995; Parthun et al., 1996; Ai and Parthun, 2004; Campos et al., 2010). In addition, the type B HAT Rtt109 acetylates histone H3 lysine 56 to yield H3K56ac (Driscoll et al., 2007). In S. cerevisiae, the histone chaperone Hif1 recognizes Hat1/Hat2 to form the NuB4 complex (nuclear type B histone acetyltransferase specific for H4) (Ai and Parthun, 2004; Poveda et al., 2004; Parthun, 2007). NuB4 is involved in the incorporation of histone H4-H3 into chromatin at sites of DNA damage, as well as in heterochromatic regions, although the exact mechanism remains elusive (Kelly et al., 2000; Ai and Parthun, 2004; Qin and Parthun, 2006). Interestingly, H3K56ac is required for efficient chromatin reassembly following DNA damage repair (Chen et al., 2008). Acetylation of histone H4 by Hat1 presumably functions in a similar way to facilitate restoration of proper chromatin structure during the final steps of DNA repair (Qin and Parthun, 2002; 2006; Ge et al., 2011). Importantly, newly synthesized histone H4 acetylated by Hat1 represents only a minor fraction of the total cellular histone H4 pool (Poveda et al., 2004; Barman et al., 2008; Poveda and Sendra, 2008). The vast majority of histones are embedded in chromatin.

Candida albicans is an obligatory diploid fungal pathogen able to undergo reversible morphological transitions between three different growth forms known as yeast, pseudohyphae and true hyphae (Gow et al., 2002; Saville et al., 2003). Notably, the ability to undergo morphogenetic transitions is considered a key virulence factor of C. albicans, since a variety of mutants locked in one of these forms showed reduced virulence in different infection models (Lo et al., 1997; Braun et al., 2000; Hwang et al., 2003; Zheng and Wang, 2004; Banerjee et al., 2008). In addition, C. albicans can switch between the so-called white and opaque form (Slutsky et al., 1987), two yeast-like morphologies displaying distinct phenotypes related to host niche colonization (Lachke et al., 2003), interaction with immune cells (Lohse and Johnson, 2008), virulence (Kvaal et al., 1999), as well as mating that occurs exclusively in opaque cells (for a recent review see Lohse and Johnson, 2009).

C. albicans encounters genotoxic stress during infection due to host-derived reactive oxygen species (ROS) produced by immune cells, which damage DNA (Vazquez-Torres and Balish, 1997; Salmon et al., 2004) and kill invading pathogens such as Candida spp. (Frohner et al., 2009), as well as Histoplasma capsulatum (Youseff et al., 2012). Notably, a loss of the DNA damage repair machinery favours a constitutively pseudohyphal morphology and reduces virulence in C. albicans (Leng et al., 2000; Andaluz et al., 2006; Legrand et al., 2007; Shi et al., 2007; Hao et al., 2009). Moreover, inhibition of the DNA damage repair machinery in C. albicans, as well as treatment with genotoxic drugs promotes white-opaque switching (Bennett and Johnson, 2005; Alby and Bennett, 2009). Interestingly, the C. albicans Rtt109 HAT, which generates H3K56ac (Lopes da Rosa et al., 2010), is also involved in DNA repair (Wurtele et al., 2010), white-opaque switching (Stevenson and Liu, 2011) and pathogenesis (Lopes da Rosa et al., 2010).

Hat1 homologues have been implicated in DNA damage repair in different organisms (Qin and Parthun, 2002; Barman et al., 2006; Benson et al., 2007). Here, we aimed to answer the question if Hat1 and other components of the NuB4 complex affect susceptibility to genotoxic stress in this important human pathogen. We demonstrate an essential role for the C. albicans Hat1/Hat2 complex in DNA damage repair and morphogenesis through the acetylation of histone H4. Interestingly, the lack of Hat1 debilitates the repair of both exogenous and endogenous DNA lesions, leads to accumulation of DNA damages, and triggers the switch to pseudohyphal growth. Furthermore, we show that genetic inactivation of the Hat1/Hat2 complex causes hypersensitivity to the widely used antifungal caspofungin (CASP) and triggers white-to-opaque switching. C. albicans is to our knowledge the first organism known to require histone H4 processing by a type B HAT for efficient repair of endogenous DNA damage and morphogenesis. Our data also strengthen emerging notions that the specific inhibition of chromatin remodelling by drugs (Simonetti et al., 2007; Agbor-Enoh et al., 2009; Hnisz et al., 2010; Wurtele et al., 2010) may prove of therapeutic relevance in infectious diseases, including those caused by fungal pathogens.

Results

Deletion of C. albicans NuB4 components leads to constitutive pseudohyphal growth

The C. albicans orf19.779 encodes the orthologue of S. cerevisiae Hat1. Furthermore, a blast search (http://www.ncbi.nlm.nih.gov/blast) identified two additional genes, orf19.2146 and orf19.7185, sharing homology with ScHAT2. In addition, orf19.6843 might represent the third component of the NuB4 complex (Dunleavy et al., 2007). Thus, we employed the SAT flipper technique (Reuss et al., 2004) to generate homozygous deletion strains for all candidate components of the putative NuB4 complex. Interestingly, homozygous hat1Δ/Δ deletion cells displayed a slow growth phenotype with constitutively wrinkled colony morphologies, indicating an active filamentation programme even under yeast-promoting conditions (Fig. 1A), when compared with the wild-type control or the heterozygous HAT1/hat1Δ strain. Liquid cultures of the hat1Δ/Δ strain contained a mixture of yeast cells and elongated cells, which formed pseudohyphae with typical constrictions at the septa (Fig. 1B and C). Importantly, reintegration of HAT1 at its endogenous locus fully restored smooth colony morphology, as well as normal yeast growth. Out of the two putative Hat2 orthologues only deletion of orf19.2146 resulted in altered colony and cell morphology similar to the lack of HAT1 indicating that this ORF is indeed the orthologue of ScHat2 (Fig. 1A and B). Thus, we refer to it as HAT2 in this report. As for orf19.7185, no morphological changes were observed after deleting orf19.6843. Thus, this ORF may not encode the orthologue of ScHif1. Alternatively, CaHif1 is not required for the function of Hat1 in maintaining the yeast morphology (data not shown). Interestingly, deletion of both Hat1 and Hat2 mimicked the phenotype of the respective single deletions, suggesting that Hat1 and Hat2 function in a complex to regulate cell morphology in C. albicans (Fig. 1A and B). Furthermore, to characterize the pseudohyphae produced in the absence of Hat1, we determined transcription levels of hyphae-specific genes under yeast-promoting conditions. We detected a strong upregulation of the typical hyphae-induced genes such as ECE1, HWP1 and SAP5, demonstrating a fully active hyphal transcriptional programme (Fig. 1D). Thus, C. albicans cells lacking the NuB4 components Hat1 and Hat2 are defective in maintaining the yeast morphology and constitutively grow in a pseudohyphal morphology. Furthermore, the constitutive filamentation of hat1Δ/Δ cells is accompanied by a hyphae-specific transcription programme.

Figure 1.

Loss of C. albicans HAT1 or HAT2 induces pseudohyphal growth and hyphae-specific genes.

A. hat1Δ/Δ and hat2Δ/Δ strains form wrinkled colonies on YPD plates at 30°C indicating constitutive filamentous growth. Images were taken after 3 days of incubation. Scale bar corresponds to 1 mm.

B. Cells lacking Hat1 show elongated cell morphology under yeast-promoting conditions (YPD 30°C). Deletion of HAT2 or HAT1 and HAT2 mirrors lack of HAT1. Scale bar corresponds to 20 μm.

C. Elongated cells of the hat1Δ/Δ strain form pseudohyphae with constrictions at the cell junctions. Cells were grown as in (B). Cell wall was stained with Calcofluor White (CW) and arrowheads indicate constrictions. Scale bar corresponds to 20 μm.

D. Filamentation of hat1Δ/Δ cells is accompanied by induction of hyphae-specific genes. RNA was isolated from four independent cultures of cells grown in YPD at 30°C. Transcript levels of ECE1, HWP1 and SAP5 were analysed by RT-qPCR and normalized to RIP1 expression. Data are shown as mean + SD. *P <0.05 and **P <0.01 relative to wild-type cells (Student's t-test).

Lack of HAT1 leads to genotoxin sensitivity and increase in DNA damage followed by filamentation

Hat1 is involved in the repair of DNA damages in different organisms. In C. albicans, a defective DNA damage repair machinery can lead to accumulation of DNA damage and pseudohyphal growth (Andaluz et al., 2006; Shi et al., 2007; Lopes da Rosa et al., 2010). Thus, we asked if the morphogenesis defect upon deletion of HAT1 is due to impaired DNA damage repair. First, we determined the consequence of deleting HAT1 on the sensitivity to various DNA-damaging agents. Interestingly, hat1Δ/Δ and hat2Δ/Δ cells showed a pronounced hypersensitivity to methylmethane sulphonate (MMS), ethylmethane sulphonate (EMS) and 4-nitroquinoline N-oxide (4-NQO) (Fig. 2A). Determination of cfu after transient MMS treatment confirmed that the lack of HAT1 and HAT2 or deletion of both genes had a similar effect on MMS sensitivity (Fig. 2B). Furthermore, the genetic removal of HAT1 also caused hypersensitivity to UV irradiation (Fig. 2C). Importantly, reintegration of HAT1 or HAT2 fully restored the wild-type phenotype, indicating that the sensitivities observed are specifically caused by the absence of the respective proteins. Since these stress conditions cause distinct kinds of DNA lesions, we reasoned that Hat1 and Hat2 play a general role in the repair of different types of DNA damage.

Figure 2.

Lack of Hat1/Hat2 causes genotoxin sensitivity and accumulation of DNA damage followed by filamentation.

A. Lack of Hat1 leads to hypersensitivity to agents causing DNA damage. The removal of HAT2 or HAT1 and HAT2 mimics lack of HAT1. To investigate sensitivity to genotoxic drugs, plates contained 0.025% methylmethane sulphonate (MMS), 0.5% ethylmethane sulphonate (EMS) or 1.25 μM 4-nitroquinoline N-oxide (4-NQO). Fivefold serial dilutions of indicated strains were spotted on agar plates and incubated at 30°C for 3 days.

B. Exponentially growing cells were cultured in MMS-containing YPD for the indicated times at 30°C. Cultures were diluted and plated onto YPD plates for cfu counting. Colonies were counted after 3 days of incubation at 30°C to determine viability. Data are shown as mean ± SD from three independent experiments.

C. The UV sensitivity of indicated strains was tested by plating on YPD plates and irradiating with 4 mJ. Colonies were counted after 3 days of incubation at 30°C. The survival was calculated relative to un-irradiated plates. Data are shown as mean + SD from three independent experiments.

D. Constitutively activated DNA damage response in hat1Δ/Δ cells was detected by measuring expression levels of DNA damage-inducible genes using RT-qPCR. RNA was isolated from three to four independent cultures of cells grown in YPD at 30°C. Transcript levels were normalized to the expression level of RIP1. Data are shown as mean + SD.

E. Increased phosphorylation of histone H2A at serine 129 upon loss of Hat1 was detected by immunoblot analysis. Treatment with MMS (0.02% for 2 h) further increased H2A-S129 phosphorylation. Detection of total histone H2A indicates equal loading.

F. Rad52 was C-terminally tagged with GFP in wild-type and hat1Δ/Δ backgrounds. Rad52-GFP foci were counted in logarithmically growing cells in SC medium at 30°C. Data are shown as mean + SD from three independent experiments. At least 250 cells were counted in each experiment.

G. Rad52-GFP foci formation is followed by cell elongation in hat1Δ/Δ cells. CellASIC Y04C plates were used to perform live cell imaging with cells carrying a RAD52-GFP allele in a wild-type (Rad52-GFP) or hat1Δ/Δ background (Rad52-GFP hat1Δ/Δ). Cells were grown in SC medium at 30°C and pictures were taken at the indicated time points. Scale bar corresponds to 5 μm.

C–F. *P <0.05 and **P <0.01 relative to wild type (Student's t-test).

Next, we assessed transcription levels of several orthologues of DNA damage-induced genes and verified DNA damages in the absence of genotoxic stress by determining H2A serine 129 phosphorylation (H2A-S129), a hallmark modification present in the proximity of double-strand breaks (Shroff et al., 2004). Interestingly, hat1Δ/Δ cells overexpressed several DNA damage-induced marker genes such as RAD6, RAD17 or MEC1 (Fig. 2D), and showed markedly increased levels of H2A-S129 phosphorylation, indicating the accumulation of DNA damages in the absence of Hat1 (Fig. 2E). Because of the heterogeneous cell morphology phenotype of hat1Δ/Δ cells, we investigated the occurrence of DNA damage at a single cell level using fluorescence microscopy. Therefore, we constructed strains expressing fully functional (Fig. S1A) GFP-tagged Rad52 alleles in both wild-type and hat1Δ/Δ backgrounds. Rad52 forms foci at DNA lesions repaired by homologous recombination, which can be visualized by fluorescence microscopy in tagged strains (Lisby et al., 2001). Indeed, loss of HAT1 sharply increased the number of cells containing Rad52-GFP foci even in the absence of any genotoxic stress (Fig. 2F), confirming the H2A-S129 phosphorylation data (Fig. 2E). Furthermore, we aimed to determine whether accumulating DNA lesions are triggering the constitutive filamentation of hat1Δ/Δ cells. Therefore, we subjected Rad52-GFP strains grown in SC medium to live cell imaging. Interestingly, the vast majority of elongating cells (92%, n = 50) that developed a pseudohyphal morphology also contained Rad52-GFP foci and thus must carry accumulated DNA damages before cell elongation (Fig. 2G). As expected, wild-type cells neither elongated nor formed Rad52-GFP foci under these conditions (Fig. 2G). These results clearly demonstrate that accumulation of DNA damages in the absence of Hat1 triggers pseudohyphal growth.

C. albicans Hat1 is required for efficient repair of DNA damage

The observed accumulation of DNA lesions in the absence of Hat1 could have two different explanations. First, DNA damages occurring during normal growth cannot be repaired efficiently and thus accumulate in the cell. This is consistent with what was shown for different DNA damage repair mutants in C. albicans (Shi et al., 2007; Lopes da Rosa et al., 2010). Another possible explanation could be that damages occur more frequently in the absence of Hat1 when compared with wild-type cells. Thus, to investigate, if Hat1 is indeed required for the efficient execution of DNA damage repair via homologous recombination, we measured the dynamics of Rad52-GFP foci in cells after providing a pulse exposure to MMS. This approach has been successfully used in S. cerevisiae to demonstrate the importance of H3K56ac in the completion of DNA damage repair (Wurtele et al., 2011). Thus, we performed live cell imaging using the CellASIC microfluidic plate system, which allows for quick changes of growth medium at specific time points (Lee et al., 2008). Cells grown in SC medium were treated with 0.02% MMS for 90 min. Importantly, hat1Δ/Δ cells do not loose viability in the presence of MMS under these experimental conditions (Fig. S1B). After treatment, the medium was changed back to SC medium lacking MMS, and the fraction of cells containing Rad52-GFP foci, as well as their lifetime was determined. Importantly, without MMS treatment, the fraction of hat1Δ/Δ cells containing Rad52-GFP foci never exceeded 25% (Fig. 3A). The reduced number of cells with Rad52-GFP foci at the beginning of the experiment can be explained by the fact that large elongated cells containing Rad52-GFP foci can hardly enter the culture chamber of the microfluidic plate. In contrast, 100 min after MMS removal, some 78% of wild-type cells and 77% of hat1Δ/Δ cells contained foci (Fig. 3B and C). Interestingly, after 300 min, the fraction of wild-type cells containing Rad52-GFP foci had declined to 10%, whereas foci remained in 52% of the hat1Δ/Δ cells (Fig. 3B and C). Strikingly, lack of Hat1 also dramatically increased the lifetime of individual Rad52-GFP foci when compared with wild-type cells. The majority of foci in wild-type cells disappeared within 2 h. In striking contrast, 78% of the foci formed in hat1Δ/Δ cells persisted for longer than 4 h (Fig. 3D). Taken together, these results demonstrate that Hat1 is indeed required for efficient repair of DNA damage, at least for DNA lesions whose repair requires homologous recombination via Rad52.

Figure 3.

MMS treatment of hat1Δ/Δ cells results in persistent Rad52 foci.

A. The fraction of hat1Δ/Δ cells with Rad52-GFP foci remains constant over time. Live cell imaging was performed using cells harbouring a RAD52-GFP allele in a wild-type (Rad52-GFP) or hat1Δ/Δ background (Rad52-GFP hat1Δ/Δ). Cells were grown in SC medium at 30°C. Pictures were taken at the indicated time points and the fraction of cells carrying foci was quantified. At least 50 cells were inspected for each time point. Data are shown as mean + SD from two independent experiments.

B. Removal of Rad52-GFP foci is delayed in hat1Δ/Δ cells. Live cell imaging was performed as described in (A). After 2 h in SC medium a pulse MMS treatment (0.02% MMS for 90 min) was done. Afterwards, the medium was changed back to SC lacking MMS and the fraction of cells containing Rad52-GFP foci was determined at the indicated time points (+ MMS: start MMS treatment). At least 50 cells were examined for each time point. Data are shown as mean + SD from two independent experiments.

C. Microscopy of cells from the experiment described in (B).

D. The lifetime of Rad52-GFP foci is prolonged upon deletion of HAT1. Live cell imaging was performed as described in (B). After a transient MMS treatment (0.02% MMS for 90 min), cells were grown in medium lacking MMS and the lifetime of Rad52-GFP foci was determined. Fifty cells containing Rad52-GFP foci were examined for each strain. The graph shows the fraction of cells in which Rad52-GFP foci lasted for the indicated period of time. Foci that appeared within 120 min after MMS removal were examined. Data are shown as mean + SD from two independent experiments.

A–D. *P <0.05 relative to the wild-type background (Student's t-test).

Deletion of NuB4 components leads to CASP sensitivity

Echinocandins are antifungal drugs frequently used to treat C. albicans infections (Pfaller and Diekema, 2007; Perlin, 2011). They inhibit the Fks1 subunit of the β-(1,3)-d-glucan synthase and therefore block normal fungal cell wall biosynthesis (Douglas et al., 1997). Interestingly, CASP, a prototype echinocandin, also triggers an oxidative stress response in C. albicans (Kelly et al., 2009). Furthermore, the lack of the Rtt109 histone acetyltransferase leads to CASP sensitivity, which can be rescued by addition of antioxidants. The authors propose that this phenotype is caused by CASP-induced ROS that induce DNA damages (Wurtele et al., 2010). Therefore, we asked if CASP treatment indeed triggers ROS accumulation and if Hat1 mediates tolerance to this drug. To determine if CASP causes ROS accumulation, we measured intracellular ROS levels by FACS analysis using dihydroethidium (DHE) and dihydrorhodamine 123 (DHR). Both DHE and DHR get oxidized by superoxide, yielding in a fluorescent product emitting at 567 nm and 540 nm respectively (Emmendorffer et al., 1990; Zhao et al., 2003). Interestingly, with both dyes we detected ROS accumulation upon CASP treatment (Figs 4A and S2A). Furthermore, addition of the antioxidant vitamin C reduced ROS levels triggered by CASP (Fig. 4A). To determine if CASP-induced ROS accumulation leads to DNA damage, we tested mutants lacking DNA damage repair proteins for CASP sensitivity. Interestingly, deletion of RAD51, RAD52 and RAD53 caused a marked CASP hypersensitivity, which was completely rescued by adding vitamin C, implying that CASP causes DNA damage through accumulating ROS (Figs 4B and S2B). Therefore, we also tested strains lacking HAT1 and HAT2 for their CASP susceptibilities. As expected, both single mutants and the double mutant were clearly CASP-hypersensitive, and growth was fully restored by vitamin C (Fig. 4C). Importantly, constitutive ROS levels in hat1Δ/Δ cells are only slightly increased. Moreover, adding vitamin C to the culture medium did not rescue the slow growth phenotype (Fig. S3). Therefore, increased ROS production is not responsible for the growth defect of the hat1 mutant. Furthermore, we confirmed the CASP sensitivity phenotype of hat1Δ/Δ cells in a liquid assay (Fig. S2C). Finally, because Hat1 and Rtt109 may function in parallel pathways by acetylating histone H4 and H3, respectively, we constructed double-deletion strains and investigated their CASP sensitivity. Interestingly, deletion of both HAT1 and RTT109 clearly exacerbated CASP hypersensitivity of the corresponding single mutants, suggesting that these two HATs function independently to mediate echinocandin tolerance (Fig. 4D). Furthermore, our data strongly suggest that interfering with the function of the NuB4 complex renders C. albicans hypersensitive to the antifungal drug CASP.

Figure 4.

Lack of DNA damage repair proteins or NuB4 components causes ROS-mediated CASP hypersensitivity.

A. Intracellular ROS production in response to CASP treatment was measured by FACS using dihydroethidium (DHE). Logarithmically growing cells were loaded with 20 μM DHE for 1 h, treated with 250 ng ml−1 CASP for 2.5 h and analysed by FACS. Twenty-five mM vitamin C (VitC) was added together with DHE to quench ROS as indicated (a). To quantify the ROS production, the mean fluorescent intensity of each sample was determined. Data are shown as mean + SD from at least two independent experiments (b). *P <0.05 and ***P <0.001.

B. Deletion of DNA damage repair genes encoding Rad52, Rad53 and Rad51 renders cells sensitive to CASP, which can be rescued by addition of vitamin C as antioxidant. Fivefold serial dilutions of indicated strains were spotted onto YPD plates containing 250 ng ml−1 CASP and 25 mM VitC as indicated. Plates were incubated at 30°C for 3 days.

C. Lack of Hat1 or Hat2 leads to CASP hypersensitivity, which can be rescued by addition of vitamin C. Experiment was performed as described in (B). Plates contained 150 ng ml−1 CASP and 25 mM VitC as indicated.

D. Lack of Rtt109 increases CASP sensitivity of hat1Δ/Δ cells. Experiment was performed as described in (B). Plates contained 100 or 150 ng ml−1 CASP as indicated.

Deletion of HAT1 induces white to opaque switching

The accumulation of DNA damages in C. albicans triggers phenotypic switching from the white phase to the mating-competent opaque phase (Alby and Bennett, 2009). Therefore, we investigated a possible role of Hat1 in this process in more detail. White-opaque (W/O) switching can only occur in a mating-type homozygous background (Miller and Johnson, 2002). C. albicans has a diploid genome harbouring a mating type-like locus (MTL) with two alleles, ‘a’ and ‘α’. Thus, we generated mating-competent MTLa/a homozygous hat1Δ/Δ deletion strains and performed quantitative switching assays. Interestingly, lack of Hat1 led to a dramatic increase in the W/O switching frequency towards the opaque from, which is consistent with the fact that DNA damages induce W/O switching in C. albicans (Fig. 5A). Furthermore, the switching frequency in the reverse direction was clearly reduced in hat1Δ/Δ cells (Fig. 5B). Thus, loss of Hat1 must stabilize the opaque state. Moreover, we performed quantitative mating assays with a MTLα/α tester strain to verify the mating-competent opaque form and the mating-incompetent white form. As expected, only opaque cells were able to mate with considerably higher efficiencies than mating-incompetent white cells (Fig. 5C). The slightly increased mating efficiency of white hat1Δ/Δ cells can be explained by increased switching to the opaque form that occurs during the assay. In summary, these data clearly indicate that loss of Hat1 induces W/O switching towards the opaque phase and therefore triggers mating in C. albicans.

Figure 5.

Deletion of HAT1 enhances phenotypic W/O switching towards the opaque form.

A. Quantitative white to opaque switching assays were performed with the indicated strains. The percentages represent the fraction of colonies that showed alterations of the original phenotype. All strains were MTLa/a strains. Data are shown as mean + SD from three independent experiments.

B. Quantitative opaque to white switching assays were performed as described in (A). Data are shown as mean + SD from three independent experiments.

A and B. *P <0.05.

C. To verify the mating-competent opaque and the mating-incompetent white phases quantitative mating assays were performed with an opaque phase MTLα/α tester strain. At least two independent experiments per genotype were performed yielding qualitatively similar results. Values are shown of one representative experiment.

The C. albicans NuB4 complex binds and acetylates free histone H4 at lysine 5 and 12

The efficient repair of DNA damages requires extensive chromatin remodelling and histone deposition, which has been linked to lysine acetylation (Costelloe et al., 2006; Chen et al., 2008; Ejlassi-Lassallette et al., 2010). Thus, to investigate if Hat1 is involved in this process, we characterized its subcellular localization and if it catalyses the acetylation of histone H4 in a deposition-related pattern. Therefore, we constructed strains expressing fully functional GFP-tagged alleles of Hat1 and Hat2 (Fig. S4A). Interestingly, both proteins mainly localized to the nucleus (Fig. 6A). Furthermore, nuclear localization of Hat2 partially required a functional Hat1, since Hat2 lost its predominant nuclear localization in the absence of Hat1 but not vice versa. Importantly, genomic reintegration of HAT1 fully restored the normal nuclear targeting of Hat2 (Fig. 6A). Furthermore, to confirm the physical interaction of Hat1 with Hat2 in C. albicans, we constructed strains expressing a functional myc-tagged allele of Hat1 (Fig. S4A), and performed co-immunoprecipitation experiments. Interestingly, we were able to co-immunoprecipitate a 43 kDa protein with Hat1-myc, which was identified as Hat2 by immunoblotting using Hat2-specific rabbit polyclonal antibodies (Fig. 6B). Furthermore, we were also able to co-immunoprecipitate histone H4 acetylated at lysine 5 and lysine 12, both of which are diagnostic hallmarks for the deposition-related acetylation pattern of newly synthesized histone H4 (Fig. 6B and C) (Sobel et al., 1995; Benson et al., 2006). Interestingly, loss of Hat2 abrogated H4 binding by Hat1 (Fig. 6B). As expected, we failed to detect any changes in the acetylation status of total histone H4 in hat1Δ/Δ cells, because free histones acetylated by Hat1 represent only a minor fraction of the total histone H4 pool (Poveda et al., 2004; Poveda and Sendra, 2008) (Fig. S4B). To confirm that Hat1 is indeed acetylating histone H4, we performed immunoprecipitation coupled with in vitro acetylation assays. Hat1 indeed acetyl-modified free histone H4 to yield H4K5ac and H4K12ac, and the presence of Hat2 was required for full enzymatic activity of Hat1 (Fig. 6D). Importantly, no signal was obtained in the absence of acetyl-CoA, excluding the possibility that the H4K5ac and H4K12ac signals in lane 1 are due to endogenous histone H4 in the IP. These data unequivocally demonstrate that Hat1 and Hat2 are part of the NuB4 complex in C. albicans, which binds and acetylates free histone H4 in a pattern characteristic for newly synthesized histones to be used for chromatin assembly.

Figure 6.

The C. albicans Hat1/Hat2 complex localizes to the nucleus and acetylates free histone H4.

A. GFP-tagged Hat1 and YFP-tagged Hat2 show primarily nuclear localization (Hat1-GFP; Hat2-YFP). Deletion of HAT2 does not change the localization of Hat1 (Hat1-GFP hat2Δ/Δ), but nuclear localization of Hat2 is partially Hat1-dependent (Hat2-YFP hat1Δ/Δ). Genomic reintegration of Hat1 restored the wild-type localization of Hat2 (Hat2-YFP hat1Δ/Δ::HAT1). Cells grown to the exponential growth phase in SC medium at 30°C were inspected. Nuclear localization was confirmed by Hoechst 33342 staining. A wild-type strain (untagged) was used as control. Scale bar corresponds to 10 μm.

B. Immunoprecipitation (IP) of myc-tagged Hat1 in a wild-type (Hat1-myc) and a hat2Δ/Δ background (Hat1-myc hat2) was performed with an anti-myc antibody. The IPs were subjected to SDS-PAGE analysis, followed by immunoblotting with specific antibodies for the myc-tag, Hat2 and the C-terminus of histone H4 respectively. An untagged wild-type strain was used as control. The amount of IP loaded corresponds to 40 OD600 units.

C. NuB4 incorporates H4K5ac and H4K12ac. Immunoprecipitation was performed as described in (B). Histone H4 acetylation was detected with acetylation-specific antibodies for K5 (H4K5ac) and K12 (H4K12ac). An untagged wild-type strain was used as control. The amount of IP loaded corresponds to 40 OD600 units.

D. Immunoprecipitation reactions were used for in vitro acetylation assays using recombinant unacetylated histone H4 and acetyl-CoA (AcCoA). H4K5ac and H4K12ac production by Hat1/Hat2 was detected by immunoblotting with antibodies specific for histone H4 acetyl-lysine 5 (H4K5ac) and acetyl-lysine 12 (H4K12ac). Detection of total histone H4 with an antibody recognizing the C-terminus (H4 C-term) served as loading control. Enzymatic activity of Hat1 was diminished in the absence of Hat2 (Hat1-myc hat2). No acetylation was detected with the IP from an untagged wild-type strain (control) or in the absence of acetyl-CoA. Recombinant histone H4 (rH4) and calf thymus histones (calf hist.) were controls.

Reduction of histone H4 levels mimics inactivation of NuB4

The NuB4 complex is involved in the incorporation of histone H4-H3 dimers into nascent chromatin (Ai and Parthun, 2004). Interfering with this process by deleting NuB4 components debilitates DNA damage repair, which might be caused by defective histone H4 deposition during the final steps of the repair process. To determine if the observed phenotypes upon loss of Hat1 or Hat2 result from insufficient histone H4 incorporation, we aimed to mimic these phenotypes by reducing the gene dosage of histone H4 in C. albicans, since reduction to a single histone H4 gene impairs growth and leads to a constitutively pseudohyphal morphology (Zacchi et al., 2010). Therefore, we sequentially deleted histone H4 genes in C. albicans, revealing a sharply increased MMS sensitivity of the single copy strain, which phenocopied the sensitivities of hat1Δ/Δ and hat2Δ/Δ cells (Fig. 7A). Strikingly, cells containing only one copy of histone H4 also showed markedly increased CASP sensitivity, which was again rescued by vitamin C (Fig. 7B). These data clearly show that interfering with the histone supply or deposition in C. albicans impairs DNA repair and leads to ROS-mediated CASP sensitivity, reminiscent of the phenotype observed for a lack of NuB4 components.

Figure 7.

Reduction of histone H4 gene dosage causes sensitivity to DNA damage and CASP.

A. Reduction to a single copy of histone H4 (1× H4) leads to MMS hypersensitivity. Strains carrying different copy numbers of histone H4 genes were spotted in fivefold serial dilutions onto YPD plates containing methylmethane sulphonate (MMS). Plates were incubated for 3 days at 30°C.

B. Single copy histone H4 cells (1× H4) show elevated sensitivity to CASP, which can be rescued by addition of vitamin C as an antioxidant. Experiment was performed as described in (A). Plates contained 100 or 150 ng ml−1 CASP and 25 mM VitC as indicated.

Discussion

In this study, we analyse the functions of the C. albicans NuB4 histone acetyltransferase complex in DNA damage repair and morphogenesis, as well as antifungal drug resistance. We show that C. albicans requires a fully functional NuB4 complex for efficient repair of DNA damages, normal cell proliferation and maintenance of the yeast morphology. Importantly, our data support the notion that interfering or blocking the function or assembly of the NuB4 complex may have therapeutic relevance for fungal diseases.

C. albicans NuB4 complex is required for repair of various types of DNA damages

The genetic removal of HAT1 or HAT2 impairs the repair of various types of DNA lesions (Fig. 2A and C). MMS and EMS lead to base alkylation, which is primarily repaired by base excision repair (Drablos et al., 2004; Gocke et al., 2009). Furthermore, when the DNA polymerase encounters MMS-induced DNA lesions, it converts them to double-strand breaks (Drablos et al., 2004), whose repair requires distinct but overlapping pathways (Chapman et al., 2012). In contrast, 4-NQO produces a kind of DNA damage, which is mainly repaired by nucleotide excision repair (Ikenaga et al., 1975; Williams et al., 2010). In addition, hat1Δ/Δ cells are also hypersensitive to UV irradiation, resulting in pyrimidine dimers, which are removed by nucleotide excision repair (Franklin et al., 1985; Hoeijmakers, 2001). Thus, we demonstrate here that the NuB4 complex is involved in various DNA damage repair processes. This is consistent with a putative role in the restoration of a functional chromatin structure in the final steps of the repair process, because chromatin remodelling must occur during the repair of different types of DNA damages (Costelloe et al., 2006).

Accumulating spontaneous DNA damage in the absence of Hat1 is followed by filamentation

Treatment of C. albicans with DNA-damaging agents triggers cellular elongation and the development of pseudohyphae. Interestingly, deletion of different components involved in the repair of DNA damage also leads to constitutive pseudohyphal growth (Andaluz et al., 2006; Shi et al., 2007). Obviously, the repair of spontaneous DNA damages appearing during the cell cycle in C. albicans requires a functional repair machinery. If this is not the case, activation of the DNA damage checkpoint and cell cycle arrest ultimately trigger cell elongation (Shi et al., 2007). We show here that inactivation of NuB4 complex components causes accumulation of endogenous DNA damages in the absence of any genotoxic stress (Fig. 2E and F). Furthermore, lack of Hat1 leads to inefficient repair of MMS-induced double-strand breaks (Fig. 3B and D). Thus, spontaneous DNA damages that arise during cell cycle cannot be repaired efficiently and accumulate in the absence of Hat1. In addition, cell elongation in the absence of Hat1 is clearly preceded by the accumulation of DNA lesions, indicating that DNA damage is the trigger for filamentation in the absence of Hat1 or Hat2 (Fig. 2G). We would like to emphasize that our results cannot rule out the possibility that DNA damage arises also more frequently in the absence of Hat1. In fact, there is evidence from S. cerevisiae indicating that the Hat1 orthologue could play a role in histone deposition during DNA replication, because it physically interacts with the origin recognition complex (ORC) (Suter et al., 2007). Thus, an inactive NuB4 complex could destabilize the replication fork, which would promote a fork collapse and subsequent accumulation of DNA damages. The clarification of this possibility will require further experiments. Interestingly, in our experimental set-up, we observed that elongated wild-type cells appearing after transient MMS treatment did not revert back to the yeast morphology within 300 min after removing MMS, even though Rad52-GFP foci had declined to about 10% (Fig. 3B and C). Obviously, the trigger for elongation must last longer than the presence of the foci. The completion of DNA repair, including the restoration of a functional chromatin structure after Rad52 removal might be required for a reversion to the yeast form. In fact, restoration of functional chromatin may signal the completion of DNA repair and enable checkpoint release (Chen et al., 2008). Therefore, completion of the repair process might be required to neutralize the elongation trigger, which may take longer than 300 min in our experimental set-up. However, 24 h after MMS removal, wild-type cells almost fully reverted to the yeast morphology (data not shown).

Lack of Hat1 increases CASP sensitivity and white-to-opaque switching

We also demonstrate that the genetic removal of HAT1 or HAT2 renders cells hypersusceptible to the commonly used antifungal drug CASP (Fig. 4C). Treatment with this drug induces oxidative stress, and antioxidants can suppress CASP sensitivity (Kelly et al., 2009; Wurtele et al., 2010). Interestingly, we show that CASP treatment induces ROS in C. albicans (Figs 4A and S2A), and cells lacking NuB4 components or DNA damage repair mutants are also hypersensitive to CASP in a ROS-dependent manner. This links DNA damage repair to antifungal susceptibility in this important fungal pathogen. Notably, a defective NuB4 complex does not affect susceptibility to other cell wall-damaging drugs or agents known to cause cell wall stress, including SDS or Calcofluor White, excluding the possibility that a cell wall defect might be responsible for the observed CASP sensitivity (data not shown). Thus, ROS-mediated DNA damage might be a secondary and indirect antifungal effect of CASP, which is most likely independent of the β-(1,3)-d-glucan synthase inhibition. Hence, interfering with the function of the NuB4 complex activity or other components implicated in the histone H4-H3 deposition pathway could be of interest in order to increase efficiency of CASP treatment, especially in resistant strains with mutations in Fks1, which stands out as the main mechanism of CASP resistance discovered to date (Balashov et al., 2006). Notably, efflux-based resistance mediated by ectopic overexpression of the Cdr2 ABC transporter in laboratory strains as well as in clinical isolates can also increase CASP tolerance (Schuetzer-Muehlbauer et al., 2003).

The inactivation of the NuB4 complex has also profound consequences on the phenotypic switching process. Genotoxic agents as well as several other environmental stimuli, which cause growth inhibition, increase W/O switching frequencies (Alby and Bennett, 2009). Interestingly, even slowing down the normal cell cycle is sufficient to induce this phenotype due to accumulation of opaque-specific factors that elicit the switch to the opaque state (Alby and Bennett, 2009). Hence, the slow growth phenotype of hat1Δ/Δ cells is most likely the cause for the increased switching frequency.

The C. albicans NuB4 complex acetylates histone H4 in a deposition-related pattern

We demonstrate that the C. albicans NuB4 complex binds and acetylates free histone H4 at K5 and K12 (Fig. 6B and D). Importantly, acetylation of H4K8 was undetectable under these conditions, suggesting a strict specificity of Hat1 for H4K5 and H4K12 (Fig. S4C). This is a highly conserved hallmark acetylation pattern for newly synthesized histone H4 (Sobel et al., 1995; Benson et al., 2006; Campos et al., 2010). Furthermore, the observation that a reduction of histone H4 gene dosage in C. albicans phenocopies the genetic removal of Hat1 strongly suggests that the H4 pool available for chromatin deposition is affected upon inactivation of the NuB4 complex (Fig. 7). Thus, we propose a model whereby Hat1 primarily modifies newly synthesized histones, prior to their incorporation into chromatin at sites of DNA lesions (Fig. 8). Although acetylation of H4K5 and H4K12 is closely correlated with histone deposition, its importance in chromatin assembly varies among different organisms and processes analysed. For example, mutation of H4K5 and H4K12 in S. cerevisiae has no effect on histone deposition onto a 2 μ plasmid or in an in vitro nucleosome assembly assay (Ma et al., 1998; Zhang et al., 1998). In contrast, it was recently shown in Physarum polycephalum that histone H4K5ac and H4K12ac promote chromatin assembly (Ejlassi-Lassallette et al., 2010). Furthermore, deletion of HAT1 as well as mutation of its target lysine residues impairs histone turnover in S. cerevisiae (Verzijlbergen et al., 2011). Therefore, whether H4K5 and H4K12 acetylation is required directly for efficient histone H4 incorporation or whether the essential function of C. albicans NuB4 in this process is restricted to binding and delivery of histones to histone chaperones for subsequent chromatin incorporation remains to be determined. Hat2 seems to be required for both scenarios, since it facilitates binding of Hat1 to histone H4 (Fig. 6B) and it is therefore also necessary for H4 acetylation (Fig. 6D). This role is obviously conserved, because S. cerevisiae Hat1 also requires Hat2 for binding of histone H4 (Parthun et al., 1996). Accordingly, deletion of Hat2 phenocopies the lack of Hat1 in S. cerevisiae (Kelly et al., 2000; Ge et al., 2011) as well as in C. albicans (Figs 1 and 2A).

Figure 8.

Model for NuB4 function in DNA damage repair in Candida albicans. Hat1/Hat2 acetylate newly synthesized histone H4 in a deposition-related pattern to produce cytoplasmic H4K5ac and H4K12ac. The complex is transferred into the nucleus and histone H3, which is acetylated by Rtt109 to H3K56ac (Lopes da Rosa et al., 2010), becomes associated. Subsequently, modified histone H4-H3 are incorporated into chromatin at sites of DNA damage to restore chromatin structure. In the absence of Hat1, newly synthesized histone H4 remains unacetylated. Furthermore, incorporation of histone H4-H3 into chromatin at sites of DNA damage is impaired leading to accumulation of endogenous DNA lesions. This in turn causes a cell cycle delay, which results in cell elongation, pseudohyphal growth and increased white-to-opaque switching.

Although, Hat1 was originally purified from cytoplasmic extracts (Parthun et al., 1996), subsequent work indicates that Hat1 is a predominantly nuclear enzyme (Ai and Parthun, 2004; Poveda et al., 2004). This is consistent with our data, although NuB4 components required for nuclear localization differ between S. cerevisiae and C. albicans. Interestingly, nuclear localization of Hat2 in C. albicans partially requires Hat1 but not vice versa (Fig. 6A). In contrast, S. cerevisiae Hat1 requires Hat2 for proper localization (Poveda et al., 2004), indicating substantial differences in HAT nuclear targeting between these two species. Obviously, Hat1 only transiently localizes to the cytoplasm to bind and acetylate histone H4, which is followed by nuclear translocation. Interestingly, Hat1 is indeed found in complexes with proteins involved in nuclear import of histone H4 (Mosammaparast et al., 2002; Campos et al., 2010). In addition to chromatin assembly, acetylation of H4K5 and H4K12 is also involved in the nuclear targeting of histone H4 in S. cerevisiae, suggesting an additional function of Hat1 in nuclear histone H4 translocation (Glowczewski et al., 2004; Blackwell et al., 2007). If this is also the case in C. albicans remains to be determined.

In conclusion, we show here that the NuB4 complex plays important roles in facilitating DNA damage repair and modulating fungal morphogenesis and antifungal drug tolerance in C. albicans. This has come as a surprise, since Hat1 seems dispensable for normal growth in other organisms as demonstrated by deletion of HAT1 in Schizosaccaromyces pombe or chicken cells (Barman et al., 2006; Benson et al., 2007). In S. cerevisiae, deletion of HAT1 or HAT2 alone does not show any sensitivity or growth phenotypes, due to a functional redundancy with histone H3 acetylation (Parthun et al., 1996; Qin and Parthun, 2002). Obviously, expressing a functional NuB4 complex is more important for C. albicans when compared with other organisms. Thus, C. albicans is to the best of our knowledge the first eukaryotic organism, which depends on histone H4 processing by a type B HAT for repair of endogenous DNA damage, normal cell growth and morphogenesis. Interestingly, slower growth rates and morphological defects due to deletion of DNA damage repair proteins have also been associated with a reduction in virulence (Hao et al., 2009; Lopes da Rosa et al., 2010). While we are certainly interested to identify possible virulence phenotypes using in vivo models, the slow growth rate of hat1Δ/Δ mutants makes it difficult to reliably assess a possible virulence defect due to reduced fitness when compared with wild-type cells (Chauhan et al., 2005). Nevertheless, preliminary virulence experiments in mouse models in vivo indicate attenuated virulence of cells lacking Hat1 (M. Tscherner et al., unpubl. data). The growth defect observed in vitro may lead to fitness defects in vivo, explaining reduced virulence. In contrast, in vitro assays with phagocytes indicate increased resistance to killing by mouse neutrophils, which could be a consequence of altered recognition due to the pseudohyphal morphology of cells lacking Hat1 (M. Tscherner et al., unpubl. data).

Nevertheless, our data suggest that the pharmacological inhibition of Hat1 might be a promising strategy to reduce the pathogenicity of C. albicans or render this pathogen more susceptible to echinocandins. Indeed, there is rising evidence that histone-modification inhibitory drugs (HiMoIDs) could provide a promising approach to treat fungal infections. For example, treatment with HDAC inhibitors can influence the interaction of C. albicans with immune cells (Simonetti et al., 2007) and alter antifungal drug sensitivity (Smith and Edlind, 2002). Furthermore, blocking the Hos2 histone deacetylase (HDAC) complex leads to reduced virulence (Hnisz et al., 2010). Interestingly, a specific Hos2 inhibitor termed MGCD290 (MethylGene, Montreal, Canada) has entered clinical trials as potential antifungal drug (Pfaller et al., 2009). We would like to point out that Hos2 shows even higher conservation between C. albicans and humans, with a Protein blast E-value of 7 × 10−174, while Hat1 yields a lower albeit still significant E-value of 3 × 10−36. Therefore, Hat1 can be considered as a potential antifungal target, despite the fact that the primary sequence is conserved and the target lysine residues are identical in higher eukaryotes. In addition, genetic or chemical inhibition of the HAT Rtt109 and the HDAC Hst3, which target mainly H3K56, reduces C. albicans virulence in mouse infection models (Lopes da Rosa et al., 2010; Wurtele et al., 2011). Since Rtt109 acts in a parallel pathway with Hat1 by acetylating newly synthesized histone H3, we propose that the targeted pharmacological inhibition of the NuB4 complex or other components of the same pathway could be a promising future strategy to combat fungal infections.

Experimental procedures

Media, growth conditions and growth inhibition assays

Rich medium (YPD) and synthetic complete medium (SC) were prepared as previously described (Kaiser et al., 1994). Cultures were grown at 30°C overnight, diluted to an OD600 of 0.1 the next morning, grown until cells reached the logarithmic growth phase and used for the experiments unless otherwise indicated. To assay sensitivity phenotypes cultures were diluted to 1 × 106 cells per millilitre and fivefold serial dilutions were prepared in distilled water. Identical volumes were spotted on agar plates containing the indicated substances. Plates were incubated at 30°C for 3 days and pictures were taken using a SPImager (S&P Robotics, Toronto, ON). MMS, EMS and 4-NQO were obtained from Sigma-Aldrich (Vienna, Austria). Stock solutions of 4-NQO were prepared in acetone. Caspofungin (Merck, Whitehouse Station, NJ) and ascorbic acid (Sigma-Aldrich, Vienna, Austria) were prepared as stock solutions in sterile water. To determine sensitivity to UV irradiation, 200 cells were plated on YPD plates and irradiated with 4 mJ in a Stratalinker 2400 (Stratagene, La Jolla, CA). Colonies were counted after incubation at 30°C for 3 days and survival rate was determined relative to plates without irradiation. For cfu determination after transient MMS treatment, logarithmically growing cells were incubated at 30°C in medium containing the indicated MMS concentration. Before MMS addition (0 h), and at the indicated time points after MMS addition, cultures were diluted and plated onto YPD plates. Colonies were counted after incubation at 30°C for 3 days and viability was determined relative to the zero time point.

Plasmid and strain construction

The complete list of C. albicans strains, plasmids and primers used in this study are listed in Tables S1, S2 and S3 respectively. All strains heterozygous for the mating type were derived from the clinical isolate SC5314 (Gillum et al., 1984). Gene deletion mutants in this background were generated using the recyclable SAT1-flipping cassette (Reuss et al., 2004). For deletion of RAD51, RAD52, RAD53, RTT109 and histone genes, we used a modified plasmid with an additional BglII site in which the SAT1 marker was replaced by the NAT1 marker derived from pJK863 (Shen et al., 2005). The RAD53 deletion cassette was constructed using a fusion-PCR strategy (Noble and Johnson, 2005) with a fragment containing the NAT1 marker and the FLP recombinase fused to the upstream and downstream region. Gene complementation constructs were created by cloning corresponding genes including upstream and downstream regions required for the genomic targeting into pSFS2A. All MTLa/a strains were derived from DHCA202, which is a MTLa/a derivative of SN152 (Hnisz et al., 2009). Gene deletion mutants in this background were done using the fusion-PCR strategy with the Candida maltosa LEU2 and C. dubliniensis HIS1 markers (Noble and Johnson, 2005). The MTLα/α mating tester strain has been previously described (Hnisz et al., 2009). For C-terminal fluorescent tagging, pFA6a-GFP-NAT1 and pFA6a-YFP-NAT1 were created by cloning GFP and YFP derived from pGFP-HIS1 and pYFP-HIS1 (Gerami-Nejad et al., 2001), respectively, into pFA6a-3HA-NAT1. The pFA6a-3HA-NAT1 vector containing the dominant NAT1 marker was created by exchanging the marker in pFA6a-3HA-kanMX6 (Longtine et al., 1998) by the NAT1 marker, including TEF1 promoter and terminator derived from pJK863 (Shen et al., 2005). For construction of the HAT1 and HAT2 fluorescent protein tagging cassettes, the 3′ part of the coding sequence and the terminator region were fused with a fragment containing the fluorescent protein and the NAT1 marker using fusion-PCR (Noble and Johnson, 2005). For tagging of HAT2 with a recyclable marker, YFP derived from pYFP-HIS1 (Gerami-Nejad et al., 2001) was cloned into pSFS3b to yield pSFS3b-YFP and the fusion-PCR strategy was applied. For C-terminal myc-tagging pFA6a-9myc-NAT1 was constructed by replacing the 3HA tag in pFA6a-3HA-NAT1 with a 9-myc tag. The tagging cassette was constructed as described for fluorescent protein tagging. Transformation of C. albicans was done via electroporation as described previously (Reuss et al., 2004).

Colony morphology analysis, microscopy and live cell imaging

Colony morphology and microscopic analysis were performed as described previously (Hnisz et al., 2010). For fluorescence microscopy cells were grown in SC medium, washed once in distilled water and used for microscopy. Nuclear staining was done by adding 2 μg ml−1 Hoechst 33342 dye (bis-Benzimide H 33342 trihydrochloride, Sigma, Vienna, Austria) directly to the medium for 15 min at 30°C. For live cell imaging the ONIX Microfluidic Perfusion Platform (CellASIC, Hayward, CA) with Y04C Microfluidic plates was used. Cells were grown in SC medium to exponential phase and loaded into the culture chambers of the Y04C plate. The media flow rate was set to 1 psi and cells were grown at 30°C for 2 h. Then, the medium was switched to SC containing 0.02% MMS for 90 min. After that, cells were grown again in SC medium till the end of the experiment. Pictures were taken every 10 min during the whole experiment. For Rad52-GFP foci determination five pictures were obtained for each field of cells at 1 μm intervals along the z-axis and analysed.

ROS measurements

To determine intracellular ROS levels, overnight cultures were diluted to an OD600 of 0.25 in SC medium with 10% glucose to avoid CASP-induced flocculation (Gregori et al., 2011), and cultured for 1 h at 30°C. After loading with 20 μM DHE (Invitrogen GmbH, Vienna, Austria) or 20 μM DHR (Invitrogen) for 1 h, cells were treated with 250 ng ml−1 CASP for 2.5 h. Cells were washed once with distilled water followed by FACS analysis with FL1-H (DHR) or FL3-H (DHE) on a FACSCalibur flow cytometer (BD Biosciences, San Jose, CA).

Microbroth dilution assays

To determine the drug sensitivity to CASP in liquid assays, a modified protocol of the microbroth dilution assay was used (Sanglard et al., 1995). Briefly, overnight cultures were diluted to an OD600 of 0.1, grown at 30°C in YPD until cells reached an OD600 of about 1, and an inoculum of 2.5 × 104 cells per millilitre was prepared. Aliquots of 100 μl of twofold concentrated CASP solutions were distributed in duplicates in a flat bottom microtitre plate. After adding a 100 μl inoculum, plates were incubated at 30°C for 48 h in a humid environment to avoid evaporation. OD600 was determined in a Victor2 plate reader (PerkinElmer, Waltham, MA).

RNA isolation and RT-qPCR

RNA isolation was done as previously described with some modifications (Hnisz et al., 2010). Briefly, logarithmically growing cells were harvested at 1500 g for 3 min at 4°C and washed once in ice-cold distilled water. Cells were resuspended in 1 ml of TRI Reagent (Sigma-Aldrich, Vienna, Austria). Afterwards, 200 μl glass beads (425–600 μm, Sigma-Aldrich) were added and cells were broken at 6 m s−1 for 45 s on a FastPrep instrument (MP Biomedicals, Illkirch, France). After centrifugation at 14 000 g for 15 min at 4°C, the supernatant was transferred into a fresh tube and 200 μl of chloroform was added. After another centrifugation step, the supernatant was transferred to a fresh tube and RNA was precipitated by addition of 500 μl of isopropanol for 20 min on ice, washed once with 70% ethanol and resuspended in distilled water. Afterwards, 5 μg of total RNA was treated with DNase I (Fermentas, St. Leon-Rot, Germany). After PCI extraction and ethanol precipitation, 1 μg of total RNA was used for reverse transcription using the RevertAid Reverse Transcriptase (Fermentas, St. Leon-Rot, Germany). cDNA amplification was monitored quantitatively by SYBR Green incorporation in a Realplex Mastercycler (Eppendorf, Vienna, Austria) using the KAPA SYBR FAST Master Mix Universal (Peqlab, Erlangen, Germany). Amplification curves were analysed using the Realplex Software (Eppendorf). Relative quantification of mRNAs was performed using the efficiency corrected ΔΔCt method (Pfaffl, 2001). RIP1 was used as housekeeping gene (Hnisz et al., 2010). Quantification and statistics analysis (Student's t-test) was performed in Excel (Microsoft).

Preparation of whole-cell extracts and immunoblotting

Whole-cell extracts for immunoblotting were prepared by the TCA (trichloroacetic acid) method exactly as described previously (Mamnun et al., 2004). For immunoblot analysis cell lysates equivalent to 0.5 OD600 units were separated by SDS-PAGE and transferred to nitrocellulose membranes (Protran BA79, Millipore, Billerica, MA). For separation of histones, 20% SDS-PAGE gels were used while all other proteins were separated in 12% gels. For detection of total histone H4 irrespective of the acetylation status we used an antibody against its C-terminus (ab10158, Abcam, Cambridge, UK). For detection of acetylated histone H4, we used antibodies recognizing H4K5ac (ab51997, Abcam), H4K8ac (39172, Active Motif, La Hulpe, Belgium) or H4K12ac (07-959, Millipore). Histone H2A phosphorylation was detected with an antibody specific for phosphorylated serine 129 (39271, Acitve Motif, La Hulpe, Belgium) and as loading control an antibody against total histone H2A was used (39236, Active Motif). Tagged proteins were visualized using antibodies against GFP (11814460001, Roche Diagnostics GmbH, Vienna, Austria) and c-Myc (ab32, Abcam) respectively. Calf histones were obtained from Sigma-Aldrich (H9250, Vienna, Austria) and chicken core histones from Millipore (13-107, Billerica, MA).

Rabbit polyclonal anti-Hat2 antibodies and purification

Polyclonal anti-Hat2 antibodies were raised in rabbits against a peptide corresponding to amino acid 363–382 of C. albicans Hat2. The peptide was expressed in E. coli as a glutathione S-transferase fusion protein and purified via glutathione sepharose beads (GE Healthcare, Vienna, Austria). About 1 mg of purified antigen was used for immunization of one New Zealand White Rabbit (Charles River, Sulzfeld, Germany). The antiserum was used at a dilution of 1:1000 and specificity was tested using appropriate cell extracts of wild-type C. albicans and mutants lacking Hat2 as a control.

Immunoprecipitation

Cells were grown in 100 ml of YPD to an OD600 of 2.0, harvested, washed and resuspended in 1 ml of lysis buffer [50 mM HEPES pH 7.5, 140 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% Na-deoxycholate, complete protease inhibitor cocktail (Roche Diagnostics GmbH, Vienna, Austria)]. Afterwards, an equal volume of glass beads (425–600 μm, Sigma-Aldrich, Vienna, Austria) was added and cells were broken by shaking eight times at 6 m s−1 for 30 s on a FastPrep instrument (MP Biomedicals, Illkirch, France). Extracts were cleared by centrifugation twice at 20 000 g. For immunoprecipitation of myc-tagged protein 30 μl of EZview Red anti-c-Myc Affinity Gel (Sigma-Aldrich) was washed six times in 700 μl of lysis buffer, added to the lysates and incubated overnight. Beads were pelleted at 1000 g, washed three times with lysis buffer and resuspended in Lämmli loading buffer for SDS-PAGE analysis. Aliquots of IPs corresponding to 40 OD600 units were loaded.

In vitro histone acetylation assay

Immunoprecipitation reactions used for in vitro acetylation assays were prepared as described above with modifications. Buffer A [50 mM HEPES pH 8.0, 400 mM (NH4)2SO4, 5% glycerol, 0.5 mM EDTA, complete protease inhibitor cocktail (Roche Diagnostics GmbH, Vienna, Austria)] was used as lysis buffer and beads were resuspended in the end in 20 μl of PBS containing a protease inhibitor cocktail. Aliquots of 3 μl of the immunoprecipitation reactions were used for acetylation assays. Reactions contained 1 μg of recombinant histone H4 (Sigma-Aldrich, Vienna, Austria), 40 mM acetyl coenzyme A, 50 mM NaH2PO4 pH 7.4, 15 mM β-mercaptoethanol, 10% glycerol and complete protease inhibitor cocktail (Roche Diagnostics GmbH) in a total volume of 20 μl and were incubated for 15 min at 37°C. To stop the reaction, 20 μl of 2× Lämmli buffer was added and samples were incubated at 70°C for 10 min. Aliquots of 5 μl were used for SDS-PAGE and immunoblot analysis.

White-opaque switching and mating assays

White-opaque switching assays and quantitative mating assays were performed precisely as described previously (Hnisz et al., 2009).

Acknowledgements

We thank all laboratory members for helpful discussions. We are grateful to J. Morschhäuser for providing the pSFS2A plasmid, Julia Köhler for the gift of the plasmid pJK863 and Judith Berman for providing the pGFP-HIS1 and pYFP-HIS1 plasmids. C. albicans sequence data were obtained from the Stanford Genome Center (http://www.candidagenome.org). This work was supported by a grant from the Christian Doppler Society, and in part by a grant from the Austrian Science Foundation (Project FWF-P-25333), to K.K. M.T. was supported through the Vienna Biocenter PhD Programme WK001.

Ancillary