HOS2 and HDA1 Encode Histone Deacetylases with Opposing Roles in Candida albicans Morphogenesis

Epigenetic mechanisms regulate the expression of virulence traits in diverse pathogens, including protozoan and fungi. In the human fungal pathogen Candida albicans, virulence traits such as antifungal resistance, white-opaque switching, and adhesion to lung cells are regulated by histone deacetylases (HDACs). However, the role of HDACs in the regulation of the yeast-hyphal morphogenetic transitions, a critical virulence attribute of C. albicans, remains poorly explored. In this study, we wished to determine the relevance of other HDACs on C. albicans morphogenesis. We generated mutants in the HDACs HOS1, HOS2, RPD31, and HDA1 and determined their ability to filament in response to different environmental stimuli. We found that while HOS1 and RPD31 have no or a more limited role in morphogenesis, the HDACs HOS2 and HDA1 have opposite roles in the regulation of hyphal formation. Our results demonstrate an important role for HDACs on the regulation of yeast-hyphal transitions in the human pathogen C. albicans.


Introduction
Candida albicans is the most common fungal pathogen of humans and is the fourth most common cause of nosocomial bloodstream infections [1]. C. albicans pathogenesis depends on its ability to transition between the yeast, pseudophyphal, and hyphal cellular morphologies [2], and these transitions are triggered by diverse environmental cues, including temperature, serum, pH, and starvation [3]. Both the yeast and hyphal morphologies are required for pathogenesis in animal models of infection [4][5][6], and are required for the formation of normal biofilms [7,8], a structure that increases antifungal drug resistance and constitutes a source of inoculum for disseminated and recurrent infections [9]. The different cellular morphologies can also trigger immune tolerance or activation against C. albicans [10][11][12]. Therefore, the ability to switch between morphologies has pleiotropic effects on C. albicans interaction with the host and on its ability to cause infection.
As epigenetic regulators of gene expression, chromatin modifying enzymes regulate diverse aspects of C. albicans biology. For example, histone modifying enzymes are required for the regulation of virulence traits and for pathogenesis in C. albicans [13][14][15][16][17][18][19][20][21][22][23]. Since the yeast-hyphal switch is critical for pathogenesis, we investigated the role of histone deacetylases (HDACs) in the regulation of this virulence trait. Here, we screened mutants in HOS1, HOS2, RPD31, and HDA1 for a role in C. albicans morphogenesis. We found that HOS1 and RPD31 have little to no role in morphogenesis, and that HOS2 and HDA1 encode proteins with opposing roles in morphogenesis: Hos2 functions as a repressor, while Hda1 functions as an inducer of filamentation.

Results
Chromatin remodeling proteins effect diverse aspects of C. albicans biology. Several histone modifying enzymes in C. albicans, including the histone methyltransferase Set1 and the histone acetyl transferase complex NuA4, are required for the expression of virulence factors and for pathogenesis in vivo [16,21]. The yeast-tohyphal transition is one biological property of C. albicans required for pathogenesis, and it is governed at least in part by epigenetic processes [16,22]. To further address the role of chromatin remodeling proteins and epigenetic regulation on pathogenesis, we investigated the role of HDACs in the yeast-to-hyphal transition.
We identified Tn7::UAU1 insertion clones located close to the START codon of HOS1 (orf19.4411), HOS2 (orf19.5377), and RPD31 (orf19.6801) ( Table 1). When available, two clones were used to disrupt the same gene to enhance the robustness of the approach. (Tn7::UAU1 insertions were identified within additional HDACs, but these plasmids had complex or incomplete inserts (data not shown)). We generated hos1/hos1, hos2/hos2, and rpd31/ rpd31 mutants using the Tn7::UAU1 insertional mutagenesis system [24]. The hda1D/D mutant was generated by sequential gene deletion using auxotrophic markers (Table 1). All mutants were tested for filamentation in solid and liquid media (Figures 1  and 2 and Table 2). Since HOS1 and RPD31 had little effect on filamentation (data not shown), we only describe the results for the hos2/hos2 and hda1D/D mutants.
Several different environmental conditions induce the hyphal morphology in C. albicans. Incubation at body temperature (37uC), alkaline pH, starvation, and serum are some of the signals that trigger hyphal morphogenesis in this fungus [3]. Further, incubation on solid surfaces, liquid media, or embedment in a matrix also impact C. albicans morphogenetic responses [25,26]. Thus, we tested the ability of the HDACs mutants to filament in several different environmental conditions, including solid and liquid M199 pH 8, serum, and Spider media, solid SLAD medium, embedded agar, and liquid media supplemented with GlcNAc. The hos2/hos2 mutants consistently showed enhanced filamentation compared to the wild-type strain on most solid media tested ( Figure 1). On M199 pH 8, the hos2/hos2 mutants filamented robustly, and showed a homogeneous peripheral halo of filamentation after 48 hrs of incubation, ,24 hrs earlier than the wild-type strain ( Figure 1 and data not shown). Similar results were observed on Spider medium, in embedded agar, and on serum ( Figure 1). On SLAD, however, the hos2/hos2 mutants showed either no filamentation or irregular filamentation around some colonies (Figure 1 and data not shown). Complementation of the hos2/hos2 mutation restored filamentation to wild-type levels in all media except SLAD. Lack of complementation on SLAD medium may indicate haploinsufficiency of HOS2, as reported previously for other mutants grown on SLAD, such as gap1D/D and gpr1D/D [27,28]. An independent hos2D/D start-to-stop deletion mutant also showed enhanced filamentation, corroborating the results of the insertional mutations (data not shown). Thus,  Hos2 functions as an inhibitor of filamentation, except in conditions of nitrogen starvation (SLAD) in which Hos2 function is required for morphogenesis. The hda1D/D mutant showed poor filamentation compared to the wild-type strain on most solid media tested ( Figure 1). On M199 pH 8 and SLAD, the hda1D/D mutant did not filament. On Spider medium, the hda1D/D mutant showed a slight but reproducible smoother surface than the wild-type strain. In embedded agar, the hda1D/D mutant showed poor filamentation. On serum, the hda1D/D mutant showed a slight defect in hyphal formation. Complementation of the hda1D/D mutation restored filamentation to wild-type on M199 pH 8, Spider, embedded, and serum media, and partially rescued the defects on SLAD. Thus, Hda1 functions as an inducer of filamentation.
In liquid media, the hos2/hos2 strain filamented similarly to wild-type in all media tested ( Figure 2 and Table 2). The hda1D/ D mutant also filamented in all media tested, but the filaments of the hda1D/D mutant appeared shorter than wild-type. Accordingly, we detected a delay in hda1D/D mutant germ tube formation in M199 pH 8 and Spider media compared to the wild-type, hos2/hos2, and hda1D/D+HDA1 strains (Table 2). We noted that the results obtained in liquid media were more variable compared to solid media. Since changes in gene silencing occurs over several generations [29,30], the rapid induction of filamentation in liquid medium may be more susceptible to variations than in solid media because of the differences in incubation time (,1 hr vs .24 hrs, respectively). This difference between liquid and solid medium filamentation may also be due to the fact that liquid filamentation is assessed at the single cell level while solid filamentation is assessed at the population  (colony) level [29]. While the requirement for several generations in order for silencing to be altered may explain the disparate results for the hda1D/D mutant in solid vs. liquid media, it is also possible that Hda1 might be associated with regulators of filamentation that play a more prominent role in solid compared to liquid media. Differences in the function of regulators of hyphal formation in C. albicans when cells are incubated in solid, semi-solid, or liquid media have been previously described [25,26,31,32]. Overall, our results demonstrate that the HDACs HOS2 and HDA1 have opposing roles in the regulation of hyphal formation in C. albicans.

Discussion
Epigenetic mechanisms regulate virulence traits of diverse microbes, including Trypanosoma brucei and Candida glabrata [33,34]. Epigenetic mechanisms also regulate aspects of C. albicans pathogenesis. Set1, a histone methyltransferase, the chromatin remodeling complex Swi/Snf, the histone acetyltransferase NuA4 complex, and the HDAC Sin3 regulate morphogenesis, adherence to epithelial cells, and/or are required for pathogenesis in animal models [16,21,35]. Furthermore, histone acetylation, regulated by the SAGA/ADA coactivator complex is required for the proper response to oxidative stress and antifungals [23]. White-opaque switching is regulated by transcriptional feedback loops and HDACs [13,15,19,36]. HDACs function is also required for antifungal resistance and adhesion to human pneumocytes [14,17,18,20]. Therefore, epigenetic mechanisms play an important role in the pathogenesis of C. albicans.
Here, we show that Hos2 and Hda1 regulate the yeast-tohyphal transition in opposing ways. Previously, Hos2 and Hda1 were reported to have opposing effects on white-opaque switching [15,19]. This suggests that Hos2 and Hda1 may inversely govern a common set of genes. Histone deacetylation is usually associated with transcriptional repression [37,38]. However, HDACs are also required for gene expression, and it has been proposed that acetylation and deacetylation cycles are responsible for maintaining promoter activity [39][40][41]. HDACs can deacetylate histones globally (non-targeted deacetylation) or at specific promoters to which they are tethered in complex with specific transcription factor and other DNA binding proteins (targeted deacetylation) [42,43]. Thus, one possible mechanisms of Hos2 and Hda1 function on filamentation in C. albicans is through the association with transcriptional regulators of hyphal formation, including the positive regulators Cph1, Cph2, Efg1, Tec1, Bcr1, Czf1, and/or Rim101, and the negative regulators Nrg1, Tup1, Rfg1, and/or Sfl1 [3,25,[44][45][46][47]. For example, Hos2 and Hda1 have been associated with Tup1 and Efg1 function in S. cerevisiae and C. albicans, respectively. [48,49,50]. HDACs could also impact filamentation by affecting the expression of the regulators themselves [19,22] or by deacetylating transcription factors and other non-histone proteins that have a direct or indirect role in morphogenesis [40,43,[51][52][53][54][55]. Thus, Hos2 and Hda1 might impact hyphal formation through a diverse array of mechanisms.
Why is HOS2 required for filamentation in SLAD but acts as an inhibitor of hyphal formation in all other conditions tested? In C. albicans, hyphal formation on SLAD is modulated by transcription factors, some of which function specifically during nitrogen starvation, such as Gln3. It is possible that Hos2 is required for the function of these specific transcription factors. Alternatively, loss of Hos2 may promote expression of genes that inhibit morphogenesis during nitrogen starvation. Thus, the hos2D/D effect on morphogenesis in C. albicans varies with the environmental conditions, a phenomenon that has also been observed for the histone deacetylase Set3 [50]. HDAC inhibitors have been proposed as antifungal adjuvants, due to their effect on preventing antifungal resistance in vitro [14,17,20]. However, no studies have shown the efficacy of HDAC inhibitors as antifungals in vivo. These types of experiments become even more critical in lieu of our and others findings that HDACs have differential effects on hyphal formation. Previous reports show conflicting in vitro results on the effect of different HDAC inhibitors on germ tube formation in liquid serum [14,18]. However, inhibiting HDAC function could enhance filamentation in semi-solid surfaces (Figure 1) (such as mucosas), possibly leading to enhanced tissue invasion and biofilm formation, with the potential to cause more damage and increase antifungal resistance [56]. On the contrary, the use of specific HDAC inhibitors might enhance antifungal effectiveness by limiting hyphal development (e.g. against Hda1 (Figure 1)), or by limiting yeast development (e.g. against Hos2 ( Figure 1) and [55]). The critical role of HDACs in C. albicans pathogenesis and survival to antifungal treatment underscores the necessity to study HDAC function in this organism. A combination of in vitro and in vivo studies that assess the role of HDACs in biofilm development, genomic instability, colonization, survival, and pathogenesis could determine the potential of HDAC inhibitors as antifungal drugs. Overall, our results contribute to demonstrate the importance of epigenetic regulators in governing virulence traits in C. albicans, and support the potential of HDAC inhibitors to prevent and/or treat candidal infections.
The HOS2 and HDA1 complementation vectors pDDB503 and pDDB504 were constructed as follows. Wild-type HOS2 and HDA1 open reading frames (ORF), together with ,1kb upstream and 0.5kb downstream of the HOS2 and HDA1 ORF, were amplified in high fidelity PCRs (Pfu Turbo DNA polymerase, Stratagene) from BWP17 DNA using primers HOS2 DDB78  (Table 4). The resulting PCR products were in vivo recombined in S. cerevisiae strain L40 into a NotI/EcoRI-digested pDDB78 to generate plasmids pDDB503 and pDDB504.
All media except that for selection of Ura + transformants were supplemented with 80 mg/ml uridine. For solid media, 2% Bactoagar was added, except for Spider medium and embedded agar which required 1.35% and 1% Bacto-agar, respectively.

Microscopy
Pictures of colonies were taken using a Canon Powershot A560 digital camera on a Zeiss Opton microscope. Images of liquid cultures were captured using a Zeiss Axio camera, Axiovision 4.6.3 software (Zeiss), and a Zeiss AxioImager fluorescence microscope. All images were processed with Adobe Photoshop 7.0 software.