The following strains were used in this study: S. (Pseudallescheria) boydii CBS 120157, S. (Pseudallescheria) boydii CBS 117410, S. (Pseudallescheria) boydii CBS 117432, P. ellipsoidea CBS 301.79, P. angusta CBS 254.72, S. aurantiacum CBS 136910, S. aurantiacum CBS 136046, S. aurantiacum CBS 136047, and S. aurantiacum CBS 136049. They were generously provided by Sybren de Hoog from the Westerdijk Fungal Biodiversity Institute, Utrecht, The Netherlands. A list of all nine strains and where they were isolated is shown in Table 1. All strains were maintained in modified Sabouraud media (0.5% yeast extract, 1% peptone, and 2% glucose). To obtain conidia, the cells were grown on Sabouraud-agar plates at room temperature, and after 7 days, the surface growth was scraped off using sterile PBS, and the collected conidia were filtered and washed twice with sterile PBS.

Biofilms were grown on the surface of sterile polystyrene microplates (96- or 24-well), on central venous catheter (CVC) sections or on glass-bottom petri dishes, as previously described (Vila et al., 2015). Briefly, a standardized suspension (10⁷ conidia/ml) was added to each well of the microplate, and the plates were then incubated at 36°C for 1.5 h (adherence phase). After adherence, the supernatant containing the non-adherent cells was removed, and fresh RPMI 1600 media (Sigma-Aldrich, United States) supplemented with 2% glucose and 20% fetal bovine serum (FBS, Gibco, United States) was added to each well. The microplates were then incubated for 30 min or 2, 4, 6, 8, 12, 24, 48, or 72 h. At each time point during incubation, biofilm formation was quantified using two different approaches.

Biofilm formation on the surface of polystyrene microplates was quantified using the following two different methods: the overall biomass, which comprised both cells and the ECM, was quantified using crystal violet assays, while the metabolic activity of the cells inside the biofilms, which reflects viability and cell density, was quantified using XTT-reduction assays (Mello et al., 2016).

Crystal violet assay: At each incubation time point, the biofilms were washed with sterile PBS (pH 7.2) to remove any un-adherent cells. The remaining biofilm was fixed in methanol for 15 min and then stained with crystal violet (0.02%) for 20 min. The solution was discarded, and the biofilms were washed two times with PBS. The impregnated crystal violet was dissolved using a 33% acetic acid solution for 5 min, and the colored solution was transferred to a clean microplate and measured using a spectrophotometer at 590 nm (SpectraMAX 340 Tunable; Molecular Devices Ltd., United States).

XTT assay: The biofilms were quantified using a XTT-reduction assay, as previously described (Pierce et al., 2008). First, the biofilms were washed with sterile PBS, and then 100 μl of a XTT: Menadione solution (0.5 mg/ml: 1 μM) was added to the cells. The microplate was incubated for 2 h at 36°C while protected from light. Finally, the colored solution was transferred to a clean microplate, and its absorbance was measured at 490 nm using a spectrophotometer (SpectraMAX 340 Tunable; Molecular Devices Ltd., United States).

A standardized suspension containing 10⁵ conidia was added to each well of a microplate with RPMI 1640, and the plates were then incubated at 36°C for 2, 4, 6, and 12 h. At each time point, a total of 200 cells were counted using an optical microscope, and the percentage of germinated cells was calculated.

The biofilms were grown on the surface of sterile CVC sections (0.5 cm) that were cut longitudinally cut to expose the interior. A catheter section was placed in each well of a 96-well microplate, and biofilms were formed on the inner CVC surface for 24 or 48 h as described above. Then, catheter sections that contained biofilms were processed for scanning electron microscopy (SEM) as previously described (Vila et al., 2013). Briefly, the CVCs were washed in 0.01 M PBS (pH 7.2) and fixed in 2.5% glutaraldehyde and 4% formaldehyde in 0.1 M cacodylate buffer for 1 h at room temperature. Subsequently, the CVCs were washed in the same buffer, post-fixed in 1% osmium tetroxide and 1.25% potassium ferrocyanide for 30 min, and dehydrated in a series of ethanol solutions with increasing concentrations (30, 50, 70, 90, 100% and ‘ultra-dry’ ethanol) for 30 min at each concentration. Then, the samples were critical point-dried in CO₂, coated with gold and observed using a FEI Quanta 250 scanning electron microscope (FEI, Netherlands).

Biofilms were formed on glass-bottom petri dishes (CellView^TM, Greiner Bio-One, Germany) as described above. Then, after the biofilms were gently fixed in a 2% formaldehyde solution, they were incubated with the following fluorescent markers: Concanavalin A conjugated to Alexa Fluor 488, Filmtracer^®Sypro Ruby (both from Molecular Probes, Invitrogen, United States), and Calcofluor White M2R (Sigma-Aldrich, United States). The biofilms were observed using a Leica TCS-SPE confocal scanning microscope (Leica, Germany), and Z-stack reconstructions were analyzed using Fiji software (Schindelin et al., 2012).

Susceptibility to antifungal drugs was evaluated according to the EUCAST protocols, with some modifications (Taj-Aldeen et al., 2016). Briefly, to each well of a microplate, 100 μl of a standardized suspension (2 × 10⁵ conidia/ml) was added to 100 μl of antifungal drugs serially diluted in RPMI 1640 (supplemented with 2% glucose and buffered with 3-(N-morpholino) propanesulfonic acid (MOPS) 0.165 mol/l, pH 7.0). The concentrations of caspofungin, fluconazole, itraconazole and voriconazole ranged from 0.125 to 128 μg/ml. Control cells were grown in the presence of DMSO (for itraconazole and voriconazole) or water (for caspofungin and fluconazole). The plates were incubated at 36°C for 72 h.

To evaluate biofilm susceptibility, biofilms were formed as previously described (Pierce et al., 2008). Antifungal drugs were added to 24 h-old biofilms at the same concentration range described for planktonic cells. The plates were then incubated for an additional 24 h at 36°C. For both susceptibility tests (planktonic cells and biofilms), a visual reading was performed to define the minimal inhibitory concentration (MIC), and the results were confirmed using a XTT-reduction assay as described above.

A survival analysis was performed using the larvae of G. mellonella as described by de Lacorte Singulani et al. (2016). Each group of larvae (10 larvae per group) was infected with 10 μl of 10⁶ conidia via the last left pro-leg using a Hamilton syringe. The S. boydii CBS 120157, S. boydii CBS 117410, S. aurantiacum CBS 136046 and S. aurantiacum CBS 136047 strains were selected for these experiments. A group of uninfected larvae and a group of uninfected larvae that were inoculated with PBS were used as controls. The larvae were incubated in Petri dishes at 37°C, and population death was evaluated daily for a total of 9 days.

All statistical analysis were performed using GraphPad Prism 6.0 software (GraphPad, United States). A variance two-way ANOVA was performed using Tukey’s and Bonferroni’s comparisons tests to evaluate the kinetics of biofilm formation in addition to the germination process and survival experiments.

The growth kinetics of all nine strains that formed biofilms on polystyrene microplates were analyzed by correlating the results of two different analyses: the total biofilm biomass (including cells, both dead and alive, and extracellular matrix), which was evaluated using crystal violet assays (Figures 1A,C,E) and metabolic activity, which was quantified using XTT-assays (Figures 1B,D,F). These nine strains are representative isolates defined on a meeting of the Working Group on Pseudallescheria/Scedosporium Infections, whose researchers received the samples to generate more data about these fungi.

In our analysis of biofilm biomass, among the three S. boydii strains, the clinical isolate obtained from human lungs (S. boydii CBS 120157) grew significantly faster than the other two isolates (CBS 117432 and CBS 117410), especially at 6 and 12 h (Figure 1A). The environmental strain (S. boydii CBS 117410) was the slowest-growing strain, and it reached the same biomass as S. boydii CBS 120157 only after 48 h. These results suggest that the clinical S. boydii strains form biofilms faster than the environmental strain but that all three strains were able to form thick biofilms after 48 h (Figure 1A). Interestingly, the S. boydii clinical strain (CBS 120157) had significantly lower metabolic activity inside the biofilm at 24 h, but at 48 h, all three S. boydii strains had similar metabolic activity (Figure 1B). The typical environmental species P. ellipsoidea and P. angusta produced a similar amount of biomass and P. angusta exhibited a significantly higher metabolic rate at 6, 12, and 48 h (Figures 1C,D).

The S. aurantiacum clinical isolates shared a similar pattern of biomass growth, and both strains produced biomass levels that were the same as that produced by the S. boydii clinical isolates within 24 h (Figure 1E). Remarkably, across all of the S. aurantiacum strains, there was no significant difference in the amount of biomass produced or metabolic activity rates, excepting at 72 h when S. aurantiacum CBS 136047 presented significantly less biomass comparing to the other strains (Figures 1E,F).

For the further analysis, two strains of S. aurantiacum and S. boydii were chosen, since the experiments present a significantly high cost to be performed with all nine isolates. A germination assay was performed using fresh conidia. One clinical and one environmental strain of each species (S. boydii CBS 120157 and CBS 117410 and S. aurantiacum CBS 136046 and CBS 136047) were selected to evaluate growth profiles (Figure 2). For all of the included strains, regardless of the germ tube length, approximately 10% of the cells were germinating after 2 h of incubation, and 100% were germinating after 12 h (Figure 2A). The most important difference between the germination profiles was observed after 4 and 6 h of incubation when the clinical isolates, S. boydii CBS 120157 and S. aurantiacum CBS 136046, were found to have germinated 2- to 3-fold faster than the environmental isolates, S. boydii CBS 117410 and S. aurantiacum CBS 136047 (Figure 2A). Similar results were observed when the germ tube lengths were measured, as follows: although similar germ tube lengths were observed after 2 h of incubation, the tube lengths were different after 4 h of incubation (Figure 2B).

To compare the biofilm structures of clinical and environmental strains of S. boydii and S. aurantiacum, we selected one clinical and one environmental strain of each species (S. boydii 120157 and 117410, respectively, and S. aurantiacum 136046 and 136047, respectively) and imaged the biofilms using CLSM. Biofilm cells were stained with (a) Concanavalin A, which stains glucose and mannose residues, both of which are present in cell walls and biofilm ECM; (b) Calcofluor white, which stains chitin structures in the cell walls of fungi, and (c) Filmtracer^®sypro ruby, which stains glycoproteins known to be abundant in the ECM of other fungi.

A more robust biofilm was formed by the S. boydii clinical isolate (120157) than the corresponding environmental strain (117410) (Figures 3A,B). These results corroborate the data shown in Figures 1, 2. However, there was little to no visible difference between the biofilms formed by the clinical and the environmental strains of S. aurantiacum (Figures 4A,B).

Biofilms were formed on catheters using the clinical isolates of each species (i.e., S. boydii CBS 120157 and S. aurantiacum CBS 136046) with the aim of evaluating their ability to adhere to and develop biofilms on medical devices. The biofilms were visualized using SEM.

Similar to the results previously described for the biofilms grown on polystyrene and glass-bottom surfaces in this study, biofilms formed faster on CVCs in the S. aurantiacum clinical isolate than in the S. boydii clinical isolate. Whereas 24 h was sufficient for S. aurantiacum to colonize the entire catheter surface (Figure 6), the S. boydii isolate required 48 h to reach a similar biomass density (Figure 5). The presence of ECM was observed among hyphae that had adhered to the catheter surface, especially in S. boydii and less in S. aurantiacum, suggesting the formation of a mature biofilm structure (arrows in Figures 5D, 6B).

To confirm that S. boydii and S. aurantiacum grow in a similar manner on all of the different surfaces used in this study (i.e., polystyrene, glass bottom dishes and catheters), biofilm formation was simultaneously measured using crystal violet and XTT-reduction assays in clinical isolates of S. boydii (CBS 120157) and S. aurantiacum (CBS 136046) that were grown on all three surfaces. After 24 h, both isolates had adhered to and grown biofilms on all three surfaces and, as expected, faster growth was observed on all 3 surfaces by S. aurantiacum as compared with S. boydii (Figure 7).

Because increased resistance to drugs is the most clinically relevant phenotypical alteration observed in biofilms, we evaluated the susceptibility of S. boydii and S. aurantiacum strains to different antifungal drugs and compared the results to those obtained for their planktonic counterparts (primary suspension inoculums were used to form the biofilms). As expected, the biofilms were generally less susceptible to all the azole drugs (MIC > 128 μg/ml) than were their planktonic cells. Interestingly, caspofungin was the only drug for which we obtained a lower MIC value in the biofilms (32 μg/ml) than in the planktonic cells (64 μg/ml) (Table 2).

In the planktonic cells, itraconazole (MIC, 1–2 μg/ml) and voriconazole (MIC, 0.5–1 μg/ml) were the most active antifungals. As was expected for filamentous fungi, these isolates were less sensitive to fluconazole (MIC between 16 and 64 μg/ml) (Table 2). No relevant difference was observed in antifungal susceptibility between these distinct genera and or by sample source.

An in vivo infection model was established using larvae of G. mellonella. In this assay, we evaluated the pathogenicity of the strains used in this study to construct survival curves (Figure 8). Interestingly, the S. aurantiacum clinical isolate (136046), which had the fastest germ tube elongation (Figure 2B), was also the most virulent strain in that it had caused 100% mortality in G. mellonella at 6 days after infection (Figure 8). The other three strains (S. aurantiacum 136047, P. boydii 120157 and S. boydii 117410) had similar virulence patterns and had killed the larvae by 8–9 days post-infection (Figure 8).