The Glycosylphosphatidylinositol-Anchored Superoxide Dismutase of Scedosporium apiospermum Protects the Conidia from Oxidative Stress

Scedosporium species are common fungal pathogens in patients with cystic fibrosis (CF). To colonize the CF lungs, fungi must cope with the host immune response, especially the reactive oxygen species (ROS) released by phagocytic cells. To this aim, pathogens have developed various antioxidant systems, including superoxide dismutases (SODs) which constitute the first-line protection against oxidative stress. Interestingly, one of the S. apiospermum SOD-encoding genes (SODD gene) exhibits a glycosylphosphatidylinositol (GPI) anchor-binding site and encodes a conidial-specific surface SOD. In this study, a SODDΔ mutant was engineered from a non-homologous end joining-deficient strain (KU70Δ) of S. apiospermum. Compared to its parent strain, the double mutant KU70Δ/SODDΔ exhibited increased susceptibility to various oxidizing agents and triazole antifungals. In addition, the loss of SodD resulted in an increased intracellular killing of the conidia by M1 macrophages derived from human blood monocytes, suggesting the involvement of this superoxide dismutase in the evasion to the host defenses. Nevertheless, one cannot disregard an indirect role of the enzyme in the synthesis or assembly of the cell wall components since transmission electron microscopic analysis revealed a thickening of the inner cell wall layer of the conidia. Further studies are needed to confirm the role of this enzyme in the pathogenesis of Scedosporium infections, including the production of a recombinant protein and study of its protective effect against the infection in a mouse model of scedosporiosis.


Introduction
Scedosporium species (phylum Ascomycota, order Microascales) are environmental molds usually living as saprophytes, mainly in highly polluted soils and contaminated water [1]. Nevertheless, these fungi may also cause a wide variety of infections in humans. Besides localized infections such as subcutaneous mycetomas or bone and joint infections resulting from traumatic inoculation of some soil fungal elements, pulmonary infections are also described, which are thought to be due to the inhalation of some airborne conidia [2].

Genomic DNA Extraction
Mycelium from 9-day-old cultures on PDA (supplemented or not with phleomycin or hygromycin) was ground in liquid nitrogen before resuspending the crushed material in lysis buffer (10 mM Tris-HCl; 1 mM EDTA; 2% Triton X-100; 1% SDS; 0.1 M NaCl, pH 8). Genomic DNA was extracted using the phenol-chloroform protocol (all reagents from Sigma-Aldrich, Saint-Louis, MI, USA) and precipitated with ethanol. DNA was then treated with RNase A 0.2 mg/mL and kept at 4 • C in Tris-EDTA buffer. Qubit 2.0 Fluorometer (Invitrogen, Carlsbad, CA, USA) was used to quantify DNA, and its integrity was verified by 1% agarose gel electrophoresis.

Disruption of the SODD Gene
For disruption of the gene encoding the GPI-anchored Cu,Zn-SOD, the cassette was prepared using the double-joint PCR procedure as described by Pateau et al. [29] with a few modifications. The selection marker used conferred hygromycin B resistance (HPH). The 5 and 3 flanking regions of the SODD gene were obtained from DNA from the wild-type strain by polymerase chain reaction (PCR) amplification using primers P1-SODD (forward) and P2-SODD (reverse), and primers P5-SODD (forward) and P6-SODD (reverse), respectively (Table 1, Figure 1A). As for the HPH gene, it was amplified from pAN7.1 plasmid using the forward and reverse primers P3-SODD and P4-SODD (Table 1), so that the amplified product overlapped the 5 and 3 flanking regions of the SODD gene. The three amplicons (i.e., 5 flanking region, HPH gene, and 3 flanking region) were then fused with a molar ratio of 1:3:1. A double-joint final PCR was carried out with primers P7-SODD (forward) and P8-SODD (reverse) ( Table 1) to obtain the final disruption cassette. PCR amplifications were performed with Q5 ® high-fidelity DNA polymerase (New England Biolabs, Beverly, MA, USA). Finally, the PCR construct was gel-purified using the Nucleospin ® gel and PCR clean-up kit (Macherey-Nagel, Düren, Germany).   Sequences that are indicated in bold font correspond to homologous sequences overlapping with the resistance marker. and SODD transformants A1, A2, A3, or A4 were digested by XbaI. The 5.5-kbp band corresponding to the SODD wild-type locus was seen for the wild-type strain and SODD transformants A2, A3, and A4, but not for the A1 transformant, which showed a 3-kbp band demonstrating the disruption of the SODD gene by the HPH resistance gene.
The transformation was performed on protoplasts obtained from 24-h-old germ tubes as described by Turgeon et al. [30] and Liu and Friesen [31] with 5 µ g of DNA. Germ tubes were collected by filtration on 20-µm-pore size Miracloth membranes and incubated at 37 °C for 3 h 30 under constant shaking (120 rpm) in OM/glucanex solution (1.2 M MgSO4, 10 mM Na2HPO4, 12.5 g/L glucanex, pH 5.8). Protoplasts were recovered by centrifugation in a Tris-HCl buffer (10 mM) containing 1.2 M sorbitol to maintain the osmotic pressure and then stored at 4 °C in the same buffer supplemented with 10 mM CaCl2. The cassette was integrated into protoplasts by heat shock in the presence of polyethylene glycol (PEG). Afterward, protoplasts were poured onto a soft agar medium (1 M sucrose, 0.2% yeast extract, 0.2% casaminoacids, 1.28% molten agar) in Petri dishes and incubated for 16 h at 37 °C before covering the plates with a layer of the abovementioned medium supplemented with 50 µ g/mL hygromycin B. Cultures were incubated for 5 to 7 days at XbaI restriction sites are indicated, and the black box indicates the hybridization site of the probe (upstream region of the SODD gene). The size of the expected fragments is indicated by arrows. P1, P2, P3, P4, P5, P6, P7, and P8 correspond to the primers used for the construction of the disruption cassette. (B): Genomic DNA of the wild-type strain IHEM 14,462 (WT) and SODD transformants A1, A2, A3, or A4 were digested by XbaI. The 5.5-kbp band corresponding to the SODD wild-type locus was seen for the wild-type strain and SODD transformants A2, A3, and A4, but not for the A1 transformant, which showed a 3-kbp band demonstrating the disruption of the SODD gene by the HPH resistance gene.
The transformation was performed on protoplasts obtained from 24-h-old germ tubes as described by Turgeon et al. [30] and Liu and Friesen [31] with 5 µg of DNA. Germ tubes were collected by filtration on 20-µm-pore size Miracloth membranes and incubated at 37 • C for 3 h 30 under constant shaking (120 rpm) in OM/glucanex solution (1.2 M MgSO 4 , 10 mM Na 2 HPO 4 , 12.5 g/L glucanex, pH 5.8). Protoplasts were recovered by centrifugation in a Tris-HCl buffer (10 mM) containing 1.2 M sorbitol to maintain the osmotic pressure and then stored at 4 • C in the same buffer supplemented with 10 mM CaCl 2 . The cassette was integrated into protoplasts by heat shock in the presence of polyethylene glycol (PEG). Afterward, protoplasts were poured onto a soft agar medium (1 M sucrose, 0.2% yeast extract, 0.2% casaminoacids, 1.28% molten agar) in Petri dishes and incubated for 16 h at 37 • C before covering the plates with a layer of the abovementioned medium supplemented with 50 µg/mL hygromycin B. Cultures were incubated for 5 to 7 days at 37 • C and checked daily for transformants growing upward in the selection marker layer, which were then subcultured on PDA supplemented with 50 µg/mL hygromycin B.

Southern Blot Analysis
Transformants growing in the presence of hygromycin B were subcultured a couple of time to obtain monospore isolates, the genotype of which was then verified by Southern blot as previously detailed [29]. Genomic DNA was extracted and digested overnight with XbaI. After separation of the DNA fragments on agarose gel electrophoresis, gels were incubated successively in 0.25 N HCl, 1.5 M NaCl/0.5 M NaOH, and finally 0.5 M Tris-HCl pH 7.5/1.5 M NaCl. DNA fragments were then transferred on nylon membranes (Amersham Hybond TM -N + , GE Healthcare). After UV-crosslinking for 3 min, the gels were incubated overnight at 55 • C in the presence of the probe [a PCR product obtained from genomic DNA of the wild-type strain using P7-SODD as forward primer and P2-SODD as the reverse primer (Table 1, Figure 1A)] labeled with Illustra™ Shrimp alkaline phosphatase (GE Healthcare life sciences) according to the manufacturer recommendations for validation of the KU70∆/SODD∆ double mutants. Finally, alkaline phosphatase was revealed by the addition of its substrate, and the membrane was imaged by chemiluminescence (LAS4000 GE Healthcare).

Susceptibility Studies
Multiple studies were conducted on the disruptant KU70∆/SODD∆ and its parent strain KU70∆.
Sensitivity to oxidative stress. Unless otherwise specified, the sensitivity to oxidative stress was determined by monitoring the fungal growth in liquid medium by laser nephelometry as previously described [29,32]. Conidia were inoculated into YPD broth to give a final concentration of 10 5 conidia/mL, and the obtained suspension was distributed into the wells of 96-flat bottom wells microtiter plates (300 µL per well). Assay wells were supplemented with one of the following chemicals: 25 µM menadione (Sigma-Aldrich), 0.5 mM diamide (Sigma-Aldrich), 1 mM cumene hydroperoxide (Sigma-Aldrich), or 8 µg/mL honokiol (Ak Scientific, Union City, CA, USA). Menadione was prepared as stock solution in dimethylsulfoxide (DMSO) so that solvent concentration in assay wells did not exceed 1% (v/v), whereas distilled water was used to prepare the stock solutions for the other chemicals. Microplates were incubated at 37 • C for 6 days in a laser-based microplate nephelometer (BMG Labtech, Ortenberg, Germany). During incubation, plates were subjected to shaking at 200 rpm for 5 min every 10 min, and growth was recorded after each shaking. Presented data were obtained from at least three independent experiments, each with triplicate wells. Growth curves expressing the relative nephelometric unit (RNU) as a function of time were generated using GraphPad Prism 6.0 (data not shown), and areas under curves (AUC) were calculated.
Because of its limited stability in liquid medium, we were unable to measure the effect of hydrogen peroxide using laser nephelometry. Sensitivity to H 2 O 2 was therefore investigated by a disk diffusion method. Conidial suspensions were adjusted to 1 Mc Farland and spread onto YPDA plates using a spiral plater. Plates were left drying for 15 min in a biosafety cabinet. Sterile filter paper disks (10 mm in diameter) soaked with 49 or 98 mM H 2 O 2 (100 µL per disk) were placed in the middle of the plates. Plates were incubated at room temperature for 24 h, then transferred to 37 • C. The diameter of the growth inhibition zones was assessed after a total of 72 h of incubation. At least 5 plates were made for each H 2 O 2 concentration during each of three independent repetitions. Susceptibility to antifungal drugs. Antifungal susceptibility testing was performed using the Etest method previously reported to deliver comparable results with the standard broth microdilution method regarding Scedosporium species [33]. Conidia prepared in saline were adjusted to 1 Mc Farland and streaked across the surface of an RPMI agar plate (bioMérieux, Marcy l'Etoile, France) using a spiral plater. Dishes were left drying for 15 min at 37 • C before applying the Etest strips (bioMérieux). The minimal inhibitory concentrations (MIC) were determined after 72 h of incubation at 37 • C; the MIC was defined as the lowest concentration at which the elliptical inhibition zone intercepts the strip. For triazole antifungals, microcolonies within the growth inhibition zone were overlooked. At least 5 Etest strips were used for each antifungal during each repetition.

Phagocytosis Assays
Cell preparation and co-culture. Monocytes were isolated from healthy human donors and differentiated to M1 phenotype-macrophages in complete RPMI 1640 (Lonza, Basel, Switzerland) containing 50 ng/mL of granulocyte-macrophage colony-stimulating factor (GM-CSF) as previously described [34]. Conidia were stained with 40 µg/mL of fluorescein isothiocyanate (FITC; Sigma-Aldrich) in phosphate-buffered saline 0.15 M pH 7.2 (PBS) for 30 min in darkness with constant shaking (120 rpm), washed twice with PBS, and enumerated. FITC-labeled conidia (5 × 10 5 cells) were co-cultured with adherent M1 macrophages (cell ratio of 1:1) in a 6-well plate in a total volume of 500 µL of complete RPMI 1640. Macrophages were allowed to ingest conidia at 37 • C in 5% CO 2 for 6 h. Later, plates were put on ice for 20 min to stop phagocytosis and processed into either the ingestion or the killing experiments. At least three independent replicates with technical triplicates in each were carried out for each strain.
Ingestion assays. Cells were lifted from wells by gentle mixing and transferred to Eppendorf tubes. The suspensions were washed twice with PBS and centrifuged at 500× g for 5 min to eliminate RPMI and non-adherent conidia. Macrophages were then marked with anti CD14 MAb (Miltenyi Biotec, Bergisch Gladbach, Germany) for 15 min at 4 • C, rinsed twice with PBS containing 1% bovine serum albumin and 0.1% sodium azide (PBS-BSA), and stained for viability with 7-aminoactinomycin D (7-AAD; Invitrogen). Cells were analyzed on FACSCanto II cytometer (Becton-Dickinson, Franklin Lakes, NJ, USA) to quantify the percentage of conidia ingested by macrophages. Data were gated to remove dead macrophages. A double-positive population (CD14+ FITC+) was considered to reflect macrophages with adherent and/or ingested conidia.
Killing assays. Non-adherent conidia were removed as described above. Afterward, adherent conidia were removed by discarding the supernatant after additional steps of washing with warm PBS and centrifugation at 500× g for 5 min. After checking the removal of adherent conidia under the microscope, the cell pellets were resuspended with 500 µL of water and incubated at 4 • C for 30 min to allow lysis of macrophages and release of intracellular conidia. After microscopic control of macrophage lysis, the obtained suspensions were centrifuged at 4600 rpm for 5 min to sediment released conidia. Macrophage debris was removed by treating the suspensions with DNase, RNase, and proteinase K (all from Invitrogen) as applied previously [35]. Each replicate included a negative control, reflecting the live conidia (macrophages-unexposed conidia), and a positive control referring to heat-killed conidia (85 • C, 30 min). If necessary, recovered conidia were kept at 4 • C until the next day. Conidia were then labeled with 25 µg/mL propidium iodide (PI; Sigma-Aldrich) for 30 min at room temperature in darkness before flow cytometry analysis. Both controls were used to set gates for the positivity of each marker. Conidia were considered live when they were only FITC-positive and dead in case of double-positivity (FITC+ PI+).

Transmission Electron Microscopy
Conidia collected from cultures grown on PDA were incubated at room temperature for 2 h with or without menadione and processed for transmission electron microscopy (TEM) as described by Ghamrawi et al. [28]. Sections were examined on JEM-1400 transmission electron microscope (Jeol, Paris, France). Only longitudinal sections of conidia (n = 20) were considered to evaluate the width, length, or thickness of the conidial wall.

Statistical Analysis
Statistical analysis was performed using GraphPad Prism 6.0. An unpaired t-test was used to check significant differences between the double mutant and its parent strain regarding the area under the growth curves in the presence of oxidative agents. Otherwise, the one-way analysis of variance Kruskal-Wallis test was run to evaluate differences between the two strains. Results were considered significantly different when p < 0.05.

Generation of a SODD Deficient Mutant
Because of the high frequency of non-homologous recombination events in S. apiospermum, we used a mutant strain deficient for the non-homologous end joining (NHEJ) by disruption of the KU70 gene, and the SODD gene was disrupted in this mutant by introducing the HPH resistance gene at the SODD locus. Culture on hygromycin-containing agar yielded only four colonies, corresponding to potential SODD disruptants (named A1, A2, A3, and A4) based on the hygromycin B resistance phenotype. Monospore isolates prepared from each colony were confirmed by PCR amplification of the HPH gene and finally analyzed by Southern blot after XbaI digestion of genomic DNA ( Figure 1B). As expected for a correct gene disruption event, Southern blot analysis revealed a single 3-kb band for the transformant KU70∆/SODD∆-A1, whereas a 5.5-kb band was observed for the parent strain and the other transformants ( Figure 1B).

Sensitivity to Temperature and Cell Wall Stressing Chemicals
Regarding the sensitivity to heat and chemical stress, no difference was observed between the parent strain (KU70∆ mutant) and the KU70∆/SODD∆ double mutant in size or pigmentation of the colonies grown on PDA whatever was the temperature (28, 37, or 42 • C). Likewise, the disruption of the SODD gene did not affect the tolerance to osmotic agents (NaCl or sorbitol), SDS, or Calcofluor White. Conversely, the addition of Congo red to the culture medium resulted in a one-log growth inhibition compared to the parent strain (data not shown).
The disruption of SODD gene did not affect the growth under normal conditions (YPD), as revealed by laser nephelometry which showed similar AUCs for the double mutant and its parent strain ( Figure 2A). These results demonstrate that SodD is not essential for growth under normal conditions, osmotic stress, or varying temperatures, but it may be involved in the protection against some cell wall stress-inducing agents.

Growth under Chemically-Induced Oxidative Stress
As the main known function of SODs is to protect cells against ROS by breaking down O 2 •− into H 2 O 2 , the growth of the double mutant was evaluated by laser nephelometry in YPD broth containing oxidant agents. Disrupting the SODD gene resulted in a significant increase in susceptibility to all the tested agents, which is reflected in decreased AUC values in comparison to its parent strain ( Figure 2). In the presence of 25 µM menadione, the growth of the double mutant reached only 50% of that of its parent strain ( Figure 2C), whereas it was not affected in the presence of the solvent alone (DMSO; Figure 2B). Likewise, the AUC obtained for the double mutant grown in the presence of 1 mM cumene or 8 µg/mL honokiol was about one-third lower ( Figure 2E,F). The least significant difference in the AUC values between the strains was observed in the presence of 0.5 mM diamide ( Figure 2D).
Evaluation of the sensitivity to H 2 O 2 by the disk diffusion method also suggested the involvement of SodD in the protection of the conidia against ROS. Diameter of the growth-inhibition zone was significantly higher for the double mutant compared to its parent strain ( Figure 3A,B): using disks soaked with 49 mM or 98 mM H 2 O 2 , the diameters of the inhibition zones for the double mutant were 2.278 ± 0.019 cm and 2.913 ± 0.211 cm, respectively, whereas they were at least 20% lower for its parent strain (1.733 ± 0.057 cm and 2.340 ± 0.072 cm, respectively).

Growth Under Chemically-Induced Oxidative Stress
As the main known function of SODs is to protect cells against ROS by breaking down O2 •─ into H2O2, the growth of the double mutant was evaluated by laser nephelometry in YPD broth containing oxidant agents. Disrupting the SODD gene resulted in a significant increase in susceptibility to all the tested agents, which is reflected in decreased AUC values in comparison to its parent strain (Figure 2). In the presence of 25 µ M menadione, the growth of the double mutant reached only 50% of that of its parent strain (Figure 2C), whereas it was not affected in the presence of the solvent alone (DMSO; Figure  2B). Likewise, the AUC obtained for the double mutant grown in the presence of 1 mM cumene or 8 μg/mL honokiol was about one-third lower ( Figure 2E,F). The least significant difference in the AUC values between the strains was observed in the presence of 0.5 mM diamide ( Figure 2D).
Evaluation of the sensitivity to H2O2 by the disk diffusion method also suggested the involvement of SodD in the protection of the conidia against ROS. Diameter of the growthinhibition zone was significantly higher for the double mutant compared to its parent strain ( Figure 3A,B): using disks soaked with 49 mM or 98 mM H2O2, the diameters of the inhibition zones for the double mutant were 2.278 ± 0.019 cm and 2.913 ± 0.211 cm, respectively, whereas they were at least 20% lower for its parent strain (1.733 ± 0.057 cm and 2.340 ± 0.072 cm, respectively).

Susceptibility to Antifungals
The deficiency in the conidial-specific GPI-anchored SOD also affected the in vitro susceptibility to antifungals. Conversely to the susceptibility to fluconazole, amphotericin B, and the echinocandins, which were not impacted, the disruption of the SODD gene led to a significant decrease in the MIC values of isavuconazole, itraconazole, posaconazole, and voriconazole (Table 2).

Susceptibility to Antifungals
The deficiency in the conidial-specific GPI-anchored SOD also affected the in vitro susceptibility to antifungals. Conversely to the susceptibility to fluconazole, amphotericin B, and the echinocandins, which were not impacted, the disruption of the SODD gene led to a significant decrease in the MIC values of isavuconazole, itraconazole, posaconazole, and voriconazole (Table 2).

Interactions with Phagocytes
Since SodD is specifically expressed at the conidial surface and potentially involved in their protection against ROS, we investigated whether the protein interferes with the phagocytosis process. Two parameters in the interaction between macrophages and the conidia, ingestion and killing, were evaluated using a flow cytometry-based technique that enabled robust measurement of each parameter. The ingestion percentage corresponds to the ratio between the number of live macrophages with adherent and/or ingested conidia to the total number of live macrophages multiplied by 100, while the killing was determined from the number of killed conidia among the total number of macrophage-released conidia.
Concerning the ingestion analysis (Figure 4), because there were still non-adherent conidia in the suspensions and since macrophages are subjected to death after several hours of co-culture, we first selected the live macrophage population in the light scattered plot. Flow cytometry effectively differentiated between macrophages alone (CD14+) and macrophages with adherent/ingested conidia (CD14+ FITC+), which served in measuring the ingestion rate. The adherence and ingestion steps were not affected by disruption of the SODD gene since we found almost the same ingestion rate for the double mutant ( Figure 4B) and its parent strain KU70∆ ( Figure 4A).
For the killing assay ( Figure 5), all events should theoretically correspond to intracellular conidia that have been released from macrophages. However, despite treatment of the cell lysate with DNase, RNase, and proteinase K, it remained necessary to eliminate all small events that probably refer to cell debris. Thus, events with low forward scatter (FSC) and side scatter (SSC) values were excluded. Fresh live conidia (FITC+ PI-) and heat-killed conidia (FITC+ PI+) were run as controls to define the area of the correspondent populations on the FITC vs. PI histogram plot ( Figure 5D). However, although still within the positivity zone, the conidia that have been killed intracellularly by macrophages exhibited lower signals for FITC and PI fluorescence ( Figure 5A,B). It is likely that heatkilling and macrophage-mediated killing of the conidia do not have similar impacts on the conidial surface fluorescence and membrane permeabilization; heat-killing results in a rapid membrane permeabilization allowing an important PI uptake, while during the macrophage-mediated killing, the slow membrane damaging leads to a gradual loss of FITC-labelled surface proteins and a lesser PI uptake. Nevertheless, the double-positive population (FITC+ PI+) of macrophage-killed conidia was denser and broader for the double mutant KU70∆/SODD∆ ( Figure 5B) than for its parent strain KU70∆ ( Figure 5A); a significant difference was seen between the two strains in the killing rate with values of 15.25 ± 5.94% for the double mutant vs. 5.2 ± 1.314% for its parent strain ( Figure 5C). These results, which were reproducible even when different donor-derived macrophages were used, support the involvement of the GPI-anchored cell wall SodD in protecting the conidia against macrophage-mediated killing. For the killing assay ( Figure 5), all events should theoretically correspond to intracellular conidia that have been released from macrophages. However, despite treatment of the cell lysate with DNase, RNase, and proteinase K, it remained necessary to eliminate all small events that probably refer to cell debris. Thus, events with low forward scatter (FSC) and side scatter (SSC) values were excluded. Fresh live conidia (FITC+ PI-) and heatkilled conidia (FITC+ PI+) were run as controls to define the area of the correspondent populations on the FITC vs. PI histogram plot ( Figure 5D). However, although still within the positivity zone, the conidia that have been killed intracellularly by macrophages exhibited lower signals for FITC and PI fluorescence (Figures 5A,B). It is likely that heatkilling and macrophage-mediated killing of the conidia do not have similar impacts on the conidial surface fluorescence and membrane permeabilization; heat-killing results in a rapid membrane permeabilization allowing an important PI uptake, while during the macrophage-mediated killing, the slow membrane damaging leads to a gradual loss of FITC-labelled surface proteins and a lesser PI uptake. Nevertheless, the double-positive population (FITC+ PI+) of macrophage-killed conidia was denser and broader for the double mutant KU70Δ/SODDΔ ( Figure 5B) than for its parent strain KU70Δ ( Figure 5A); a significant difference was seen between the two strains in the killing rate with values of 15.25 ± 5.94% for the double mutant vs. 5.2 ± 1.314% for its parent strain ( Figure 5C). These results, which were reproducible even when different donor-derived macrophages were

Morphological Features of the Conidia and Ultrastructure of Their Cell Wall
The higher susceptibility of the double mutant to Congo red known to bind to chitin prompted us to investigate the ultrastructure of the conidial wall ( Figure 6). No differences were found in the width or length of the conidia. Conversely, whereas the conidial wall of the wild-type strain and the KU70∆ mutant exhibited a similar ultrastructure with a very thin electron-dense outer cell wall layer and a thicker inner layer, a thickening of the inner cell wall layer was observed in conidia of the KU70∆/SODD∆-A1 double mutant. used, support the involvement of the GPI-anchored cell wall SodD in protecting the conidia against macrophage-mediated killing.  The higher susceptibility of the double mutant to Congo red known to bind to chitin prompted us to investigate the ultrastructure of the conidial wall ( Figure 6). No differences were found in the width or length of the conidia. Conversely, whereas the conidial wall of the wild-type strain and the KU70Δ mutant exhibited a similar ultrastructure with a very thin electron-dense outer cell wall layer and a thicker inner layer, a thickening of the inner cell wall layer was observed in conidia of the KU70Δ/SODDΔ-A1 double mutant. Figure 6. Length, width and thickness of the cell wall of the conidia of the SodD defective mutant (KU70Δ/SODDΔ) and its parent strain (KU70Δ) as determined by TEM examination. Significant differences were seen in the thickness of the conidial wall between the SodD defective mutant KU70Δ/SODDΔ and its parent strain KU70Δ (*** p = 0.0002), whereas length and width of the conidia were unchanged.

Discussion
Although ranking second among the filamentous fungi colonizing the CF airways, the pathogenic mechanisms that allow Scedosporium species to establish within the respiratory tract and to cause infections, remain poorly understood. Colonization of the CF bronchial mucus is thought to result from inhalation of some airborne conidia and their entrapment into the viscous mucus. Therefore, the mechanisms allowing the conidia to evade the host immune response, mainly represented by the oxidative burst response of phagocytes and the production of ROS, are essential for a successful colonization process. Recently, we showed that Scedosporium species were able to germinate in oxidative stress conditions [36]. In addition, analysis of the S. apiospermum genome revealed the presence of 33 genes encoding putative enzymes possibly involved in the degradation of ROS or RNS [26]. Some of these genes were found to be overexpressed in response to the exposure of hyphae to oxidative stress or after co-culture with lipopolysaccharide-activated macrophages or neutrophils [27]. Nevertheless, the design of these experiments did not allow to Figure 6. Length, width and thickness of the cell wall of the conidia of the SodD defective mutant (KU70∆/SODD∆) and its parent strain (KU70∆) as determined by TEM examination. Significant differences were seen in the thickness of the conidial wall between the SodD defective mutant KU70∆/SODD∆ and its parent strain KU70∆ (*** p = 0.0002), whereas length and width of the conidia were unchanged.

Discussion
Although ranking second among the filamentous fungi colonizing the CF airways, the pathogenic mechanisms that allow Scedosporium species to establish within the respiratory tract and to cause infections, remain poorly understood. Colonization of the CF bronchial mucus is thought to result from inhalation of some airborne conidia and their entrapment into the viscous mucus. Therefore, the mechanisms allowing the conidia to evade the host immune response, mainly represented by the oxidative burst response of phagocytes and the production of ROS, are essential for a successful colonization process. Recently, we showed that Scedosporium species were able to germinate in oxidative stress conditions [36]. In addition, analysis of the S. apiospermum genome revealed the presence of 33 genes encoding putative enzymes possibly involved in the degradation of ROS or RNS [26]. Some of these genes were found to be overexpressed in response to the exposure of hyphae to oxidative stress or after co-culture with lipopolysaccharide-activated macrophages or neutrophils [27]. Nevertheless, the design of these experiments did not allow to investigate the role of genes encoding conidial-specific enzymes, such as SodD, which is a GPI-anchored cell wall protein specifically found on the conidia [26,28]. The present study, therefore, aimed to determine the role of this enzyme. While it is not essential for the growth of S. apiospermum IHEM 14,462 KU70∆, SodD was shown to protect the conidia from damages caused by oxidizing agents.
Previously Pateau et al. [29] produced from a wild-type strain of S. aurantiacum defective mutants for the cytosolic Cu,Zn-SOD SodC and showed a surprising frequency of homologous recombination as high as 30%. Other results from our laboratory sug-gested a markedly lower frequency in S. apiospermum, consistently to those reported in the literature for other filamentous fungi (i.e., <3%) [37]. A mutant strain deficient for the non-homologous end joining system by deletion of the KU70 gene, therefore, was used as parent strain as previously done for the SidD gene [23]. This strategy which is the gold standard to increase the frequency of homologous recombination in filamentous fungi [38], allowed us to disrupt the SODD gene in S. apiospermum. In our study, we have been limited by the very few resistance markers that can be applied to Scedosporium species since they are naturally resistant to many drugs. Currently, only two are effective, phleomycin that was used for engineering of the KU70∆ strain and hygromycin B, the one used for the selection of the defective strain KU70∆/SODD∆. No other selection markers were available, and further genetic tools should be developed to allow the production of complemented strains in Scedosporium species.
The role of SODs in the evasion of fungi to the oxidative stress response of phagocytic cells is well established for numerous human pathogens [29,[39][40][41][42][43][44][45][46][47][48][49][50][51]. The degradation of ROS by microorganisms is an essential process for successful infection, and SODs are the first-line defense against oxidative stress. SODs are mainly intracellular (cytosolic or mitochondrial), and few extracellular SODs have been described to date. SOD defective mutants were obtained in A. fumigatus [46], C. albicans [41,47], C. neoformans [42], Histoplasma capsulatum [48,50], Paracoccidioides brasiliensis [51], and recently in S. aurantiacum [29]. Here, we have successfully constructed a defective mutant for the SODD gene encoding a GPIanchored SOD in S. apiospermum. This enzyme does not only belong to a particular class of extracellular SODs containing a signal peptide at the N-terminus and a GPI-anchor binding site [26], but it is also specifically expressed at the conidial surface [28]. GPI-anchored proteins are known to participate in other fungi to morphogenesis, stress resistance, and virulence [52][53][54][55].
While the role of cytosolic Cu,Zn-SODs in pathogenesis has been extensively investigated in C. albicans [41,45], C. neoformans [42,56,57], and A. fumigatus [46,58,59], little information is available regarding GPI-anchored SODs. Recently, it was shown that the GPI-anchored Sod2 from the phytopathogenic fungus Puccinia striiformis suppressed ROS accumulation during infection and was essential for its virulence [60]. Regarding human pathogenic fungi, three GPI-anchored SODs, i.e., Sod4p, Sod5p, and Sod6p, were described in C. albicans [45], and one of them, Sod5p, was shown to be involved in the protection of blastoconidia against oxidative stress and in virulence [43,45,47,49,61]. Likewise, P. brasiliensis and H. capsulatum SOD3 genes, two orthologs of S. apiospermum SODD gene [26], were shown to be involved in virulence [48,51]. For example, the disruption of the SOD3 gene in P. brasiliensis resulted in attenuated virulence in mice conversely to the disruption of the SOD1 gene encoding a cytosolic Cu,Zn-SOD [51]. Our results suggest that SodD plays an important role in protecting S. apiospermum conidia against superoxide radicals. As with P. brasiliensis Sod3, SodD proved not essential for fungal growth; still, its defect results in increased sensitivity to chemically-induced oxidative stress as attested by the reduced growth of the double mutant exposed to various ROS-induced agents or to H 2 O 2 compared to its parent strain.
Since SodD is localized to the cell envelope, a question arises on how the enzyme protects the fungus from endogenous O 2 •− -inducing agents such as menadione and honokiol. Menadione is essentially an intracellular O 2 •− -inducer; however, the addition of Sod1 to the incubation buffer was shown to protect Saccharomyces cerevisiae cells from the cytotoxic effect of menadione [62], which is in line with our findings of SodD playing a protective role against exogenous oxidative stress imposed by menadione. Honokiol is a natural biphenolic compound that disrupts mitochondrial respiration leading to intracellular ROS accumulation, mainly O 2 •− . Interestingly, all SOD genes harbored by C. albicans were upregulated in response to honokiol, including three GPI-anchored ones [63]. One may hypothesize a defect in membrane trafficking of SodD following such stress. Primary findings on Fusarium oxysporum Sod5 (a GPI-anchored Cu,Zn-SOD) support this hypothesis, showing that the Sod5 subcellular localization is altered under unfavorable environmental conditions [64]. Finally, the peroxidase activities exhibited by all SOD types can explain the increased sensitivity of our mutant defective in SODD gene to H 2 O 2 [65]. While the glutathione system is required for tolerance to cumene hydroperoxide and diamide, it is still unknown how SodD contributes to such resistance. Further studies are needed to unravel the mechanisms by which this enzyme fights various oxidizing agents.
In C. albicans, A. fumigatus, and Cr. neoformans, SODs have been suggested to contribute to tolerance to antifungal drugs, including imidazoles and triazoles [66][67][68][69][70]. Here we showed the contribution of S. apiospermum SodD in tolerance of the fungus to isavuconazole, itraconazole, posaconazole, and voriconazole, and to a lesser extent to amphotericin B and micafungin. There is limited information linking the antifungal activity of azole drugs and oxidative stress. Nevertheless, there is now accumulating evidence for an increased production of endogenous ROS induced by the antifungal drugs [66][67][68][69][70]. Likewise, treating C. albicans or Candida dubliniensis blastoconidia with fluconazole elicited increased SOD and catalase production [71]. As suggested above, one may speculate a defect in membrane trafficking of SodD following exposure to antifungal drugs and the contribution of this enzyme to the degradation of ROS accumulating in mitochondria.
In order to prove the involvement of S. apiospermum SodD in the interaction of the conidia with macrophages, a flow cytometry based-protocol was developed that enabled the differentiation between the different cell populations. To the best of our knowledge, this is the first time such a protocol has ever been used on Scedosporium species or Lomentospora prolificans to investigate the phagocytosis process. The involvement of GPI-anchored SODs in recognition of fungal pathogens by phagocytic cells has never been documented, and as expected, the adherence and ingestion steps of S. apiospermum conidia were not affected by disruption of the SODD gene. Indeed, the ingestion rates were obviously similar for the mutant defective in the SODD gene and its parent strain.
By contrast, some fungal GPI-anchored SODs were shown to contribute to surviving the ROS-dependent host killing. For instance, it was shown that C. albicans Sod4 and Sod5 were essential for protecting the blastoconidia against ROS-dependent macrophage killing [45]. Likewise, in H. capsulatum, the disruption of SOD3 gene also led to a decreased viability of yeast cells exposed to neutrophils and macrophages, and in vivo experiments demonstrated that Sod3 promotes Histoplasma virulence in a murine model of disseminated infection [48]. Consistent with these findings, we showed S. apiospermum GPI-anchored SodD to be relevant in protecting the conidia against the intracellular macrophages-mediated killing. Disruption of SODD gene led to an increased killing of ingested conidia by macrophages, although their germination was not affected. Indeed, no differences were seen in the latency period between the double mutant and its parent strain in laser nephelometry study of the growth under control conditions, e.g., without any oxidative chemical ( Figure S1).
An indirect role of the enzyme in susceptibility of the conidia to killing by macrophages should also be considered because of possible biochemical changes of the cell wall that would have been caused by disruption of the SODD gene. TEM analysis of the conidia revealed for the KU70∆/SODD∆ double mutant a thickening of the cell wall, associated with an increased susceptibility to Congo red, whereas tolerance to Calcofluor White was not affected. Discrepancies in the susceptibility to these cell wall disturbing agents have already been reported. For example, the SKN7 gene contributes to cell wall integrity in A. fumigatus, and its disruption does not modify the susceptibility to Congo red, whereas the SKN7∆ mutant is more resistant to Calcofluor White compared to its parent strain [72]. Calcofluor White is known to bind to chitin, whereas Congo red may also interfere with glucans as described in S. cerevisiae [73]. Contrasting with other fungi, glucans have been detected in the cell wall of Scedosporium species, but mainly alpha-glucans and only a low level of beta-glucans [74]. Nevertheless, the increased susceptibility to Congo red resulting from disruption of the SODD gene suggests a defect in cell wall integrity. Studies in S. cerevisiae and A. fumigatus have demonstrated the importance of some GPI-anchored proteins for cell wall integrity [52,75] and it has been demonstrated that, when the cell wall integrity is compromised, fungal cells respond by increasing cell wall biosynthesis [76]. Therefore, further studies are needed to define the impact of SODD gene disruption on the composition of the cell wall.

Conclusions
In conclusion, an S. apiospermum NHEJ deficient strain was used here to generate a defective mutant for the SODD gene. This allowed us to point out the role of this GPI-anchored SOD in protecting the conidia against various chemically-induced oxidative stresses. The loss of SodD also increased the susceptibility to triazoles, highlighting another function of this enzyme in contributing to tolerance to azole drugs. In addition, a flow cytometry-based protocol was developed that enabled the assessment of conidia-macrophage interactions. SodD does not affect the recognition of conidia, nor their adherence to and ingestion by macrophages, but plays a relevant role in escaping macrophage-mediated killing. Nevertheless, one cannot disregard an indirect role of this enzyme since TEM analysis of the conidia of the double mutant revealed ultrastructural changes in the cell wall with a marked thickening of the inner cell wall layer. Complementary studies will be conducted on the SODD defective mutant and its parent strain to definitively demonstrate the surface location of SodD, and to the analyze the impact of gene disruption on the biochemical composition of the cell wall and on virulence in animal models of Scedosporium infections.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/jof7070575/s1, Figure S1: Growth kinetics of the double mutant KU70∆/SODD∆ and its parent strain KU70∆ under control conditions. Growth was monitored by laser nephelometry in yeast extract-peptone-dextrose broth without any oxidative chemical. Funding: For the realization of her PhD thesis, Cindy Staerck was a recipient of a grant from the French parent association against cystic fibrosis, Vaincre la Mucoviscidose, which is gratefully acknowledged (RF20140501104).
Institutional Review Board Statement: Monocytes were isolated from blood from healthy donors which was obtained from the blood collection centre (EFS Pays de la Loire, Angers, France; agreement ANG-2017-01).

Informed Consent Statement: Not applicable.
Data Availability Statement: All data generated or analyzed during this study are included in this published article. The raw data supporting the conclusions of this article will be made available by the authors to any qualified researcher.