The Involvement of the Mid1/Cch1/Yvc1 Calcium Channels in Aspergillus fumigatus Virulence

Aspergillus fumigatus is a major opportunistic pathogen and allergen of mammals. Calcium homeostasis and signaling is essential for numerous biological processes and also influences A. fumigatus pathogenicity. The presented study characterized the function of the A. fumigatus homologues of three Saccharomyces cerevisiae calcium channels, voltage-gated Cch1, stretch-activated Mid1 and vacuolar Yvc1. The A. fumigatus calcium channels cchA, midA and yvcA were regulated at transcriptional level by increased calcium levels. The YvcA::GFP fusion protein localized to the vacuoles. Both ΔcchA and ΔmidA mutant strains showed reduced radial growth rate in nutrient-poor minimal media. Interestingly, this growth defect in the ΔcchA strain was rescued by the exogenous addition of CaCl2. The ΔcchA, ΔmidA, and ΔcchA ΔmidA strains were also sensitive to the oxidative stress inducer, paraquat. Restriction of external Ca2+ through the addition of the Ca2+-chelator EGTA impacted upon the growth of the ΔcchA and ΔmidA strains. All the A. fumigatus ΔcchA, ΔmidA, and ΔyvcA strains demonstrated attenuated virulence in a neutropenic murine model of invasive pulmonary aspergillosis. Infection with the parental strain resulted in a 100% mortality rate at 15 days post-infection, while the mortality rate of the ΔcchA, ΔmidA, and ΔyvcA strains after 15 days post-infection was only 25%. Collectively, this investigation strongly indicates that CchA, MidA, and YvcA play a role in A. fumigatus calcium homeostasis and virulence.


Introduction
Calcium ions are universally important secondary messengers involved in numerous signaling pathways in all eukaryotic cells. Calcium-mediated signaling regulates a diverse array of biological processes including gene transcription, protein conformation, energy metabolism, endo-and exo-cytosis, cytoskeletal arrangement, and cell physiology [1,2]. Additionally, in yeasts and filamentous fungi, Ca 2+ is important for the regulation of cell cycle, sporulation, hyphal morphogenesis, hyphal orientation and pathogenesis [3][4][5][6][7][8]. Calcium homeostasis and signaling are based on the equilibrium between the accumulation and release of calcium from intracellular stores, such as the vacuole, endoplasmic reticulum (ER), mitochondria and golgi apparatus. In Saccharomyces cerevisiae free cytosolic Ca 2+ concentrations are very low, but in response to external stimuli or stress, Ca 2+ channels are opened and Ca 2+ is rapidly mobilized from intracellular stores. Cytosolic Ca 2+ binds to calmodulin that activates the phosphatase calcineurin, which in turn dephosphorylates the Crz1 transcription factor, resulting in its nuclear translocation, the modulation of gene transcription including calcium transporters, and the restoration of calcium homeostasis [9][10][11][12].
Calcium homeostasis is maintained by calcium channels, pumps and transporters. Calcium channels allow the passive flow of Ca 2+ across cell membranes into the cytosol. Two major calcium uptake pathways including the high-affinity (HACS) and low-affinity (LACS) calcium uptake system have been identified in fungi [7], [8], [13][14][15][16]. HACS consists of the voltage-gated Ca 2+ channel Cch1 [17], [18], the stretch-activated calcium channel/regulatory protein Mid1 [19] and the PMP22_Claudin superfamily member Ecm7 [14,20]. Voltage-gated calcium channels sense the membrane potential and upon its depolarization open a gate allowing Ca 2+ influx [1]. Stretch-activated channels are activated by cell deformations associated with contact sensing and thigmotropism [1]. HACS is the major route of calcium entry into the cell when availability is low [7], [8], [14]. Alternatively, LACS is active when calcium availability is high and consists of the Fig. 1, a pheromone-inducible plasma membrane protein [7,21].
In S. cerevisiae, endoplasmic reticulum and environmental stresses as well as exposure to mating pheromone trigger a response of the HACS [7,19], [22][23][24]. In filamentous fungi, HACS mutants demonstrate impaired calcium metabolism and reduced vegetative growth. In A. nidulans, CchA and MidA are important for the regulation of conidiation, hyphal polarity and cell wall components in low-calcium environments [8]. In phytopathogenic fungi, such as Magnaporthe oryzae and Gibberella zeae, Mid1 and Cch1 mutants have minor defects in growth and development, but they are still virulent [25][26][27]. However, the deletion of mid1 in the Claviceps purpurea resulted in a complete loss of plant virulence [28]. Recent studies have indicated that in the human fungal pathogens Candida albicans and Cryptococcus neoformans, HACS was essential for sensing, and adapting to, the human host [8], [29][30][31]. These studies showed that HACS was required for the response to various stresses and also the tolerance of antifungal compounds, suggesting that Cch1 and Mid1 calcium channels are important in virulence and could represent targets for novel antifungal compounds.
The fungal vacuole is a major store for calcium. Fungal vacuolar calcium transporters, such as Ca 2+ ATPases and Ca 2+ /H + exchangers, are involved in removing Ca 2+ ions from the cytosol and loading them into internal stores to avoid calcium toxicity [32]. In S. cerevisiae, PMC1 is responsible for this process preventing growth inhibition by the activation of calcineurin in the presence of elevated calcium concentrations [33]. Conversely, a vacuolar membrane localized calcium channel, YVC1, which is a member of the transient receptor potential (TRP) family [34], [35] mediates calcium release from the vacuole in response to hypertonic shock [36], [37]. The null mutant of the C. albicans YVC1 homologue showed a decreased stress response, altered morphogenesis and attenuated virulence [38], [39], suggesting that ycv1 homologues may also play a role in pathogenicity. To date no functional analyses of YVC1 homologues in filamentous fungi have been reported.
Aspergillus fumigatus is a major opportunistic pathogen and allergen of mammals [40], [41]. Multiple components of the calcium signaling pathway impact upon A. fumigatus virulence [6], [42][43][44][45]. In addition, three PMC1 calcium transporter homologues in A. fumigatus, pmcA-C, were shown to be involved in calcium and manganese metabolism, but only pmcA influenced virulence [45]. Subsequently, the presented study characterized the A. fumigatus homologues of the CCH1, MID1 and YVC1 calcium channels. The phenotypes of the constructed cchA, midA and yvcA null mutants demonstrated their involvement in calcium metabolism and homeostasis. The A. fumigatus cchA, midA, and yvcA null mutants were avirulent in a murine model of invasive pulmonary aspergillosis. Therefore, HACS and the release of vacuolar calcium stores are essential for full virulence.

Identification of A. fumigatus Mid1, Cch1 and Yvc1 homologues
In the A. fumigatus genome a single gene showed significant identity to each of the S. cerevisiae calcium channel encoding genes, CCH1, MID1 and YVC1. These A. fumigatus gene were subsequently named cchA (Afu1g11110), midA (Afu5g05840) and yvcA (Afu3g13490). The predicted 2,125 amino acid sequence of cchA encoded four hydrophobic transmembrane domains (Figure 1A). The 642 amino acid protein of midA contained an Nterminal signal peptide and six putative transmembrane domains Figure 2. mRNA accumulation of the A. fumigatus cchA, midA, and yvcA genes. The wild-type and mutant strains were grown for 16 hours in YG medium at 37uC and the mycelia was either transferred to YG without any calcium for 30 minutes (Control:C) or for YG supplemented with 200 mM of calcium. RT-qPCR was used to quantify calcium channel mRNA abundance, which was normalized using btubulin (Afu1g10910). The relative quantitation of cchA, midA, yvcA, and b-tubulin was determined by comparison with a standard curve (C Tvalues plotted against logarithm of the DNA copy number). The results presented are the means (6 standard deviation) of four sets of experiments. doi:10.1371/journal.pone.0103957.g002  ( Figure 1B). The 670 amino acid protein encoded by yvcA had seven putative transmembrane domains ( Figure 1C). The overall topology of the three calcium channels in A. fumigatus was similar to that of Ca 2+ voltage-gated channels in higher eukaryotes. The A. fumigatus midA, cchA, and yvcA are single copy genes and do not have paralogues ( Figure S1 and Table S1). Moreover, they have very close homologues in other ascomycetes but more distantly related ones in S. pombe, C. albicans, and basidiomycetes ( Figure S1).
To determine if the predicted calcium channels in A. fumigatus were influenced at the transcriptional level by calcium, cchA, midA and yvcA mRNA accumulation was assessed in the parental strain when exposed to 200 mM CaCl 2 . Calcium exposure for 10 and 30 minutes increased yvcA, cchA, and midA mRNA abundance by 3-and 9-fold, none and 22-fold, and 12-and 5-fold, respectively ( Figure 2). These results suggest that cchA, midA and yvcA are regulated at transcriptional level by calcium. YvcA localization was assessed using a C-terminal GFP fusion protein. Growth of the YvcA::GFP strain resembled the parental strain under nutrientrich and nutrient-poor conditions (data not shown), suggesting that the replacement of the wild-type gene with the GFP fusion construct in the transformed strain did not impair YvcA function. Within the hyphae the YvcA fusion protein localised to structures that resemble vacuoles during growth on minimal or complete media ( Figure 3). The vacuolar identity of these structures was confirmed by staining the germlings with the vacuolar specific fluorophore CMCA ( Figure 3, see insets). We were not able to see any increase or relocalization of the YvcA::GFP upon exposure to 200 mM of CaCl 2 for 10 to 60 minutes (data not shown). These results strongly indicate YvcA is located on the vacuoles.
Construction of the A. fumigatus cchA, midA and yvcA null mutants The function of cchA, midA and yvcA was investigated by generating the DcchA, DmidA, and DyvcA strains using a gene replacement strategy ( Figure S2). In addition, a DcchA DmidA double mutant was created by inserting a deletion midA::pyrG + cassette into DcchA pyrG 2 mutant strain aiming to assess possible unique functions of each gene. The individual gene deletions were also complemented with the corresponding wild-type genes aiming to confirm the occurrence of possible secondary mutations during the construction of the deletion strains. The DcchA, DmidA and DcchA DmidA strains demonstrated reduced radial growth on solid MM, when compared to the parental strain ( Figure 4A). Interestingly, the addition of different concentrations of CaCl 2 rescued the growth defect of the DcchA (25 to 200 mM), DmidA  Figure 4A). This positive CaCl 2 effect reflects a compensation of lacking the calcium channels for suitable levels of intracellular calcium. All single, or double, gene deletion mutants and the respective complemented strains displayed parental-like responses to temperature, cell wall (congo red and calcofluor white), membrane (SDS), oxidative (t-butyl) and osmotic stress (NaCl and Sorbitol). However, the DcchA, DmidA and DcchA DmidA strains demonstrated increased sensitivity to the oxidative stress inducer, paraquat ( Figure 4A). The DyvcA mutant showed no differential sensitivity to any of agents tested (data not shown). Curiously, no growth reduction was observed in liquid MM and YG media ( Figure 4B). None of the three single gene deletions had an impact upon conidiation or germination rate (data not shown). The complemented DchhA::cchA + , DmidA::midA + and DyvcA::yvcA + showed parental phenotypes, indicating that the observed phenotypes were due to the gene deletion(s) in the respective strains ( Figure 4A).
Cyclosporine A (CsA) is an immunosuppressive drug that inhibits calcineurin signaling by forming a complex with the immunophilin, cyclophilin, which then inhibits calcineurin. The DcchA, DmidA and DcchA DmidA strains showed an increased resistance to cyclosporine in both MM and YAG; the DyvcA mutant was as sensitive as the wild-type strain ( Figure 5). Growth in the presence of the Ca 2+ -chelating agent EGTA was subsequently utilized to assess the impact of the gene deletions during calcium limitation. Growth of the parental strain was not dramatically affected by the presence of 10 or 20 mM EGTA, while DcchA growth was restricted at 10 or 20 mM and DmidA growth was restricted at only 20 mM ( Figure 6). As previous, the DcchA DmidA demonstrated a phenotype reminiscent of the DmidA. Interestingly, the presence of EGTA resulted in the DcchA, DmidA, and DcchA DmidA strains resembling the A. fumigatus calcineurin catalytic subunit deletion [6], i.e., thicker swollen hyphae with dichotomous hyphal tips ( Figure 6). Similar to the parental strain, EGTA had limited influence on the growth of the DyvcA mutant ( Figure 6).
Calcium is an important cellular signal for fungal drug resistance [46]. The DcchA and DmidA strains were more susceptible to various fungicides including voriconazole, ketoconazole, and itraconazole, while DcchA was also more susceptible to posaconazole ( Figure 7). In contrast, both the DcchA and DmidA strains were more resistant to caspofungin ( Figure 7). The DyvcA mutant showed no difference in susceptibility, from the parental strain, to any of the antifungal agents tested (data not shown).
The concentration of free calcium within the cell was evaluated in the A. fumigatus parental, DcchA, DmidA, and DyvcA mutant strains using Fluo-3 (Invitrogen), which is a highly sensitive fluorescent dye for the rapid measurement of calcium. Fluo-3 passively diffuses across cell membranes. Within the cell, the esters are cleaved by intracellular esterases yielding a fluorescent indicator that can no longer diffuse across membrane. Upon Ca 2+ -binding, Fluo-3 exhibits an absorption shift from 506 to 526 nm. Thus, the relative intracellular Ca 2+ concentration was evaluated based upon the fluorescence ratio after dual-wavelength excitation. During the exposure to extracellular calcium concentrations from 0, 100, and 200 mM, in the parental strain there was an increase in intracellular calcium ( Figure 8). The DyvcA, DcchA and DmidA strains showed a constant low level of intracellular calcium. Alternatively, the DcchA strain demonstrated higher intracellular calcium than the parental strain when in the presence of 0, 20, and 100 mM extracellular calcium ( Figure 8). Collectively, these results suggest that CchA and MidA play an important role in the calcium uptake while YvcA is important for removing high calcium concentrations from the cytoplasm.
Recently, C. albicans ability to adhere to a polystyrene surface and buccal epithelial cells was inhibited by exposure to verapamil, a calcium channel blocker [47]. Biofilm formation on polystyrene was significantly inhibited about 25 to 30% in the DmidA and DyvcA mutant strains when measured by crystal violet staining ( Figure S3). These results were further confirmed by a reduction of about 20% biomass in both strains when grown in static growth on polystyrene Petri dishes ( Figure S3). As previously shown, all the mutant strains have growth comparable to the wild-type strain ( Figures 4B and S3). Histopathological examinations of the parental, DcchA, DmidA and DyvcA infected tissues were performed to identify the differences in growth rate, tissue invasion and the inflammatory response. After 72 hours post-infection the lungs of mice infected with the parental strain contained multiple sites of invasive hyphal growth that penetrated the pulmonary epithelium in major airways ( Figures 9B, 10B, and 11B) and formed pockets of branched invading hyphae originating from the alveoli ( Figure 9B, 10B, and 11B). In contrast, DcchA, DmidA, and DyvcA infections typically contained inflammatory infiltrates in bronchioles, some of which contained poorly germinated or ungerminated conidia ( Figures 9B, 10B, and 11B). Fungal burden measured by real-time PCR showed that the growth of the DcchA, DmidA and DyvcA strains within the lungs was less than the parental and complemented strains ( Figures 9C, 10C, and 11C, p,0.0001). This data strongly indicates that CchA, MidA, and YvcA play a role in A. fumigatus virulence.

Discussion
Calcium homeostasis and signaling is essential for numerous biological processes and importantly impacts upon the growth, stress tolerance and virulence of fungal pathogens [1], [48], [49]. Several elements of the calcium signal transduction machinery sense and transport Ca 2+ , which in turn translate the spatiotemporal fluctuation in Ca 2+ levels into a cellular response [48]. The filamentous ascomycete A. fumigatus is an opportunistic pathogen and major allergen of mammals [40,41]. Multiple components of the calcium signaling cascade have previously been shown to impact upon A. fumigatus calcium homeostasis and virulence  [6,33,42], [49][50][51][52][53]. For example, the calcineurin catalytic subunit calA deficient strain had severe defects in hyphal extension, branching and conidial architecture [6], while the downstream calcineurin-activated transcription factor CrzA directly controlled the transcription of calcium transporters pmcA-C [45]. Significantly, all the A. fumigatus DcalA, DcrzA, and DpmcA mutants were avirulent in the murine model of invasive pulmonary aspergillosis [6,42,45]. Generally, cytosolic Ca 2+ concentrations are very low (50 to 100 nM). Calcium channels, pumps, and transporters are essential to control intracellular calcium levels. However, how A. fumigatus uptakes and stores calcium is unclear. Ca 2+ influx and efflux channels such as Mid1, Cch1, and Yvc1 from S. cerevisiae have been identified in several filamentous fungi and linked to pathogenesis [54], [55]. Therefore, the function of three calcium channels in maintaining Ca 2+ homeostasis during A. fumigatus infection was investigated. Growth in the presence of calcium-chelator EGTA represents calcium limiting conditions that would require HACS. The DcchA and DmidA strains demonstrated growth defects reminiscent of DcalA during growth under such conditions, strongly indicating that A. fumigatus MidA and CchA are important for calcium uptake in low-calcium environments. Moreover, the accumulation of intracellular calcium was lower in the DmidA and DcchA strains, while the fact that DcchA and DmidA were also more resistant to cyclosporine, suggests a low level of calcineurin activity and implies that intracellular calcium level were low. The exogenous supply of high extracellular calcium concentrations restored the growth defect of DcchA and DmidA strain.
In S. cerevisiae, Mid1 is located in the plasma membrane as well as in the ER membrane [56], while in Botrytis cinerea the Mid1 homologue is located to the ER membrane and nuclear envelope [16]. The interaction between Mid1 and Cch1 has been demonstrated in S. cerevisiae by co-immunoprecipitation [57], [58] and in A. nidulans using the yeast two-hybrid system [8]. However, in B. cinerea a split ubiquitin-based yeast two-hybrid approach did not reveal any interaction between Mid1 and Cch1 [16]. Although Cch1 and Mid1 have been shown to physically interact in various fungal systems, it remains to be addressed if A. fumigatus MidA and CchA physically interact. The corresponding individual A. fumigatus midA and cchA null mutants showed complementing and contrasting phenotypes, resulting in the hypothesis of unique gene functions for either Cch1 or Mid1. This was observed in several species such as S. cerevisiae [17,18,59], C. albicans [60], A. nidulans [8], and B. cinerea [16].
Both the A. fumigatus DcchA and DmidA mutants were more sensitive to oxidative stress. Peiter et al. [23] demonstrated that the S. cerevisiae Dcch1 Dmid1 double mutant was hypersensitive to iron, that iron stress led to increased oxidative stress and that Cch1p-Mid1p was essential to tolerate this condition. The C. albicans homologue Ecm7 is an integral protein membrane with a role in HACS, which is also involved in the oxidative stress response and hyphal development [20]. During the aging of mammalian cells, Ca 2+ signaling is altered and excess ROS can be produced due to an ineffective mitochondrial respiration [61]. It could be speculated that the impairment of calcium homeostasis, caused by deficiencies in the calcium uptake system, affect mitochondrial function and in turn cause oxidative stress [62].
The A. fumigatus DcchA and DmidA strains demonstrated attenuated virulence in a murine model of pulmonary aspergillosis. In other human pathogenic fungi, such as C. albicans and C. neoformans, the virulence of CCH1 and MID1 null mutants was attenuated in the respective mouse infection models [29,60]. In phytopathogenic fungi, such as C. purpurea and G. zeae, the calcium channel mutants respectively were non-pathogenic or showed slower disease progression [26][27][28]. In contrast, in B. cinerea deletion of Cch1 and Mid1 did not impact on virulence [16]. The observed difference in the impact of Cch1 and Mid1 deletion in the aforementioned phytopathogens could reflect differences in infection strategy. As the biotroph C. purpurea and hemibiotroph G. zeae both colonize live host tissues to establish infection, possibly experiencing calcium limiting conditions, while the necrotroph B. cinerea rapidly induces host cell death ahead of infection and may not rely on the HACS. Recently, Jiang et al. [63] characterized the A. fumigatus Mid1 homologue. The phenotypes shown for the afmid1 null mutant mostly agreed with what was observed in the presented study. However, Jiang and colleagues [63] showed that the deletion of afmid1, referred to as DmidA in the current investigation, led to hypervirulence in the immunosuppressed mice model. The contrasting results observed by these authors concerning afmid1 virulence could be due to the   different genetic backgrounds of the strains used. However, the reason for this discrepancy in virulence remains to be investigated.
The S. cerevisiae YVC1 vacuolar calcium channel mediates the release of Ca 2+ from the vacuole in response to hyperosmotic stress [34,36]. The investigation of the A. fumigatus yvcA represents the first YVC1 homologue to be functionally characterized in filamentous fungi. Krishnan et al. [64] recently showed that a novel mRNA isoform of yvcA was induced in the A. fumigatus ER translatome during ER stress, in an unfolded protein responsedependent manner. As predicted due to its function in S. cerevisiae, the A. fumigatus YvcA located to the vacuolar membrane. Surprisingly, no phenotype except for the decrease in the adhesion/biofilm formation and an attenuated virulence, was observed for the A. fumigatus yvcA null mutant. It is possible we have not observed any phenotype in the DyvcA mutant because there is a functional redundancy between A. fumigatus YvcA and PmcA. Actually, we have observed PmcA is located at the cell and vacuolar membranes (de Castro, unpublished results).
In conclusion, the A. fumigatus calcium channels MidA, CchA, and YvcA are all required for full virulence in an animal infection model. This investigation implies that CchA and MidA participate in the A. fumigatus Ca 2+ -calcineurin signaling pathway, influencing intracellular calcium homeostasis. Further studies are necessary to address the role of YvcA and how calcium homeostasis and oxidative stress are regulated by CchA and MidA.

Ethics statement
The principles that guide our studies are based on the Declaration of Animal Rights ratified by the UNESCO in January 27, 1978 in its articles 8th and 14th. All protocols used in this study were approved by the local ethics committee for animal experiments from the Campus of Ribeirão Preto from Universidade de São Paulo (Permit Number: 08.1.1277.53.6; studies on the interaction of Aspergillus fumigatus with animals). All animals were housed in groups of five within individually ventilated cages and were cared for in strict accordance with the principles outlined by the Brazilian College of Animal Experimentation (Princípios É ticos na Experimentação Animal -Colégio Brasileiro de Experimentação Animal, COBEA) and Guiding Principles for Research Involving Animals and Human Beings, American Physiological Society. All efforts were made to minimize suffering. Animals were clinically monitored at least twice daily and humanely sacrificed if moribund (defined by lethargy, dyspnoea, hypothermia and weight loss). All stressed animals were sacrificed by cervical dislocation. File S1 reports all the ARRIVE checklist for reporting animal studies.

Strains, media and culture methods
The A. fumigatus strains used in this study were CEA17 (pyrG 2 ) and CEA17-80 (as the parental in all the experiments). The media used were: complete medium composed of 2% w/v glucose, 0.5% w/v yeast extract, trace elements (YAG) or YUU (YAG supplemented with 1.2 g of uracil and uridine) and minimal medium (MM) composed of 1% glucose or 2% glycerol, trace elements and nitrate salts [65], pH 6.5. Solid versions of the aforementioned media were prepared via the addition of 2% w/v agar. Strains were grown at 37uC.

Construction of the A. fumigatus mutants
The cassettes for gene replacement were constructed by in vivo recombination in S. cerevisiae as previously described by Colot et al. [66]. Approximately 1.5 kb from the 59-untranslated region (UTR) and 39-UTR flanking region of the targeted genes were selected for primer design. The primers 5F and 3R contained a short sequence homologous to the multiple cloning site (MCS) of the pRS426 plasmid. Both the 5-and 3-UTR fragments were PCR-amplified from A. fumigatus genomic DNA (gDNA). The pyrG inserted into the gene replacement cassettes was amplified from pCDA21 plasmid and was used to generate a marker for prototrophy in the mutant strains. Each fragment along with the BamHI/EcoRI cut pRS426 plasmid were transformed into the S. cerevisiae strain SC94721 using the lithium acetate method [67]. The transformant DNA was extracted according to Goldman, et al. [68]. The cassette was PCR-amplified from the plasmids utilizing TaKaRa Ex Taq DNA Polymerase (Clontech Takara Bio) and used for A. fumigatus transformation. Southern blot analyses were used to confirm a single homologous integration event at the targeted A. fumigatus locus. The deleted mutant strains were complemented by co-transforming the respective gene plus a 1 kb flanking regions together with the pHATa vector [69] and selecting for hygromycin resistance on MM containing 150 mg/ml of hygromycin B.
The C-terminal Yvc::GFP fusion protein was created by cloning the yvc ORF (minus the stop codon) in frame with the green fluorescent protein (GFP) gene. The gfp and yvc sequences were separated by four additional codons that after translation produce a four amino acid linker (glycine-threonine-arginine-glycine) [70]. First, the yvc ORF plus 500 bp of the 59-UTR were amplified from gDNA of the wild-type strain using the primers yvc pRS426 5Fw and yvc SPACER GFP Rv. The GFP ORF was amplified from the pMCB17apx plasmid (provided by Vladimir P. Efimov) using primers Spacer GFP Fw and GFP VE39 AF. The selective marker pyrG was PCR amplified from the pCDA21 plasmid using the primers GFP pyrG Fw and pyrG Rv. Approximately 600 bp of the 39-UTR were amplified using the primers Afu yvc 3Fw and Afu yvc 3Rv primers. As describe previously, the S. cerevisiae in vivo recombination system was used for production of the transformation cassette. The PCR-amplified cassette was transformed into the A. fumigatus wild-type strain. The primers used above are described in Table S2.

Phenotypic assays
The E-tests (Biomérieux) were performed according to the manufacturer's instructions as follows, 1610 7 conidia were inoculated into 20 ml YAG medium containing half the concentration of agar. E-test strips were placed on the surface. Plates were incubated at 37uC for 24 h. The minimum inhibitory concentration was determined by analyzing the growth inhibition formed around the strip of voriconazole, amphotericin B, caspofungin, ketoconazole, itraconazole, posaconazole, anidulafungin.
YvcA::GFP microscopy YVC::GFP conidia were grown on coverslips in 4 ml of MM for 16 h at 30uC. After incubation, the coverslips with adherent germlings were left untreated or treated with CaCl 2 (200 mM) for 10, 30 and 60 min. CMAC (CellTracker Blue CMAC -Molecular Probes) staining of vacuoles was performed as described previously [71]. After growth, 10 mM CMAC dye was added to the cultures for 15 min, at 30uC. Subsequently, the coverslips were rinsed with PBS and mounted for examination. Slides were visualized on a Carl Zeiss Observer Z1 fluorescence microscope using the excitation wavelength of 450 to 490 nm, and emission wavelength of 500 to 550 nm. DIC (differential interference contrast) images and fluorescent images were captured with an AxioCam camera (Carl Zeiss) and processed using AxioVision software (version 4.8).

Crystal violet assay and biofilm formation
Crystal violet assays were used to assess biofilm formation as described previously by Mowat et al [72]. A. fumigatus biofilms were formed in 96-well microtitre plates (CorningH CostarH). Biofilms were formed by inoculating 1610 4 conidia/ml in MM containing 0.1% glucose. The plates were incubated at 37uC for 16 hours. Subsequently, the media was discarded and the biofilms were carefully washed three times with sterile PBS remove nonadherent cells. The biofilms were then dried at 60uC for 10 min and 200 ml of 0.5% crystal violet (Sigma) solution added. The plate was then incubated at room temperature for 5 min and then extensively washed with distilled water until all unbound crystal violet was removed. The plate was dried again and then 200 ml ethanol was added to each well to elute the bound crystal violet. The ethanol was then transferred to a clean 96-well microtitre plate and the absorbance read at 570 nm using a SpectraMax i3 spectrometer (Molecular Devices). A minimum of 6 replicates was performed for each strain.
In addition, A. fumigatus biofilm biomass was determined as described by Shopova et al. [73]. Briefly, conidial suspensions of 1610 3 to 1610 6 conidia/per well were inoculated in HEPES-buffered RPMI 1640 with L-glutamine. This suspension was inoculated into pre-sterilized polystyrene petri dish (static growth) or glass flasks (planktonic growth) and grown for 96 hours at 37uC. Biomass was determined by dry weight.

RNA extraction and real-time PCR reactions
Mycelia were harvested by filtration, washed twice with H 2 O and immediately frozen in liquid nitrogen. For total RNA extractions, the mycelia were ground in liquid nitrogen with pestle and mortar. Total RNA was extracted with Trizol (Invitrogen, USA). The integrity of the RNA from each treatment was assessed using an Agilent 2100 Bioanalyzer. RNase-free DNase I treatment was carried out as previously described by Semighini et al. [74]. Twenty micrograms of total RNA was treated with DNase I, purified using the RNAeasy kit (Qiagen) and cDNA synthesized using the SuperScript III First Strand Synthesis system (Invitrogen) with oligo(dT) primers, according to the manufacturer's instructions. All the qPCR reactions were performed using an ABI 7500 Fast Real-Time PCR System (Applied Biosystems, USA) and SYBR Green PCR Master Mix (Applied Biosystems, USA). The reactions and calculations were performed according to Semighini et al. [74]. The primers used are described in Table S2.

Murine model of pulmonary aspergillosis, lung histopathology and fungal burden
The murine model of pulmonary aspergillosis was performed according to Dinamarco, et al. [45]. Outbreed female mice (BALB/c strain; body weight, 20 to 22 g) were housed in vented cages containing 5 animals. Mice were immunosuppressed with cyclophosphamide, which was administered intraperitoneally (150 mg/kg of body weight) on days -4, -1 and 2 prior to and post infection. Hydrocortisonacetate (200 mg/kg) was injected subcutaneously on day -3. Fresh A. fumigatus conidia were grown on YAG for 2 days prior to infection, then harvested in PBS and filtered through Miracloth (Calbiochem). Conidial suspensions were spun for 5 min at 3,0006 g, washed three times with PBS, counted using a hemocytometer, and resuspended at a concentration of 2.5610 6 conidia/ml. The viability of the administered inocula were confirmed by incubating a serial dilution of the conidial suspension, on YAG media, at 37uC. Mice were anesthetized by halothane inhalation and infected by intranasal instillation of 5.0610 4 conidia in 20 ml of PBS. As a negative control, a group of 5 mice received PBS only. Mice were weighed every 24 h from the day of infection and visually inspected twice daily. In the majority of cases, the endpoint for survival experimentation was identified when a 20% reduction in body weight was recorded, at which time the mice were sacrificed. The statistical significance of comparative survival values was calculated using log rank analysis and the Prism statistical package. Additionally, at 3 days post-infection, 2 mice per strain were sacrificed and the lungs were removed, fixed, and processed for histological analysis.

Lung histopathology and fungal burden
After sacrifice, the lungs were removed and fixed for 24 hours in 3.7% formaldehyde-PBS. Samples were washed several times in 70% alcohol before dehydration in a series of alcohol solutions of increasing concentrations. Finally, the samples were diafanized in xylol and embedded in paraffin. For each sample, sequential 5mm-thick sections were collected on glass slides and stained with Gomori methenamine silver (GMS) or hematoxylin and eosin (HE) stain following standard protocols [40]. Briefly, sections were deparaffinized, oxidized with 4% chromic acid, stained with methenamine silver solution, and counterstained with picric acid. For HE staining, sections were deparaffinized and stained first with hematoxylin and then with eosin. All stained slides were immediately washed, preserved with mounting medium, and sealed with a coverslip. Microscopic analyses were done using an Axioplan 2 imaging microscope (Carl Zeiss) at the stated magnifications under bright-field conditions.
To investigate fungal burden in murine lungs, mice were infected as described previously, but with a higher inoculum of 1610 6 conidia/20 ml. The higher inoculum, in comparison to the survival experiments, was used to increase fungal DNA detection. Animals were sacrificed 72 h post-infection, and both lungs were harvested and immediately frozen in liquid nitrogen. Samples were homogenized by vortexing with glass beads for 10 min, and DNA was extracted and the subsequent use of the phenolchloroform method. DNA quantity and quality were assessed using a NanoDrop 2000 spectrophotometer (Thermo Scientific). At least 500 mg of total DNA from each sample was used for quantitative real-time PCRs. A primer and a Lux probe (Invitrogen) were used to amplify the 18S rRNA region of A. fumigatus (primer, 59-CTTAAATAGCCCGGTCCGCATT-39; probe, 59-CATCACAGACCTGTTATTGCCG-39) and an intronic region of mouse GAPDH (glyceraldehyde-3-phosphate dehydrogenase) (primer, 59-CGAGGGACTTGGAGGACA-CAG-39; probe, 59-GGGCAAGGCTAAAGGTCAGCG-39). Six-point standard curves were calculated using serial dilutions of gDNA from all the A. fumigatus strains used and the uninfected mouse lung. Fungal and mouse DNA quantities were obtained from the threshold cycle (CT) values from an appropriate standard curve. Fungal burden was determined as the ratio between picograms of fungal and micrograms of mouse DNA.