Aspergillus fumigatus conidial cell wall is covered by a rodlet layer and a melanin pigment layer (Aimanianda et al., 2009; Gastebois et al., 2009; Bayry et al., 2014). For swollen conidia and germlings, it has been known that exposure of polysaccharides in the conidial cell wall, including β-1,3-glucan, α-1,3-glucan, chitin, and galactomannan, is required for adherence to the host cell surface, extracellular matrix, and a variety of other substrates (Dague et al., 2008; Lamarre et al., 2009; Fontaine et al., 2010; Loussert et al., 2010; Sheppard, 2011; Gravelat et al., 2013; Valsecchi et al., 2019; Ball et al., 2020).

In order to assess the influence of GPI anchoring on the adhesive properties of A. fumigatus, the attachment of dormant, swollen, and germinating conidia to different surfaces was investigated. The conidia of the wild-type (WT), Δgpi7, and Regpi7 strains were spotted onto microscopy slides made from polystyrene, Permanox, and glass. As can be seen in Figure 1 , all dormant conidia were not able to adhere to glass, polystyrene, or Permanox. After 2–6 h of incubation, all adhered conidia on hydrophobic surfaces were swollen and germinating. As compared with the WT or Regpi7, more mutant conidia adhered to the hydrophobic surface, indicating an increased adherence of the Δgpi7 mutant.

Usually conidial rodlet disappears after 5–6 h from the start of conidial swelling and germination (Ball et al., 2020). Under our experimental conditions, about 92% of the mutant conidia are germinated, while only 55% of the WT start germinating after incubation at 37°C for 6 h (Ouyang et al., 2019). As summarized in Table 1 , after 6 h of incubation at 37°C, the adhered conidia of the mutant on Permanox and polystyrene were 29- and 15-fold than those of the WT, respectively. These results clearly demonstrate that the significant increase of adherence is due to the fast germination of the Δgpi7 mutant.

Gpi7 is involved in the transport and localization of GPI-anchored proteins required for cell wall organization, such as β-1,3-glucanosyltransferase Gel1, cell wall galactomannoprotein Mp1, and Ecm33 (Ouyang et al., 2019). To elucidate if Gpi7 affects the organization of conidial cell wall, we further determined the cell wall contents of the mutant conidia. As summarized in Table 2 , the Δgpi7 mutant showed 57.6% increase of chitin and 20.5% increase of β-glucan as compared with the WT. Interestingly, the amount of glucosamine released from the cell wall proteins of the mutant increased by 85.4% though the content of the cell wall proteins extracted from the mutant conidia was similar with that from the WT. Also, galactose and mannose residues were detected in the mutant. These results indicate that glucosamine-containing polysaccharide, chitin, and β-glucan is increased in the mutant.

It has been shown that during germination of A. fumigatus cell wall, β-1,3-glucan and mannan are exposed to the conidial surface (Alsteens et al., 2013). When the dormant and germinating conidia were detected with ConA, a lectin specifically recognizes mannose and glucose, both dormant and germinating conidia of the mutant were positively stained by FITC-labeled ConA ( Figures 2A, B ), indicating an increased exposure of cell wall polysaccharides on the surface of dormant and germinating conidia of the mutant. When conidia were detected with FITC-labeled WGA, a lectin binds N-acetylglucosamine (GlcNAc), the dormant conidia of the mutant exhibited an increased positive staining as compared with the WT ( Figure 2C ), while the germinating conidia of the mutant were similar with the WT ( Figure 2D ). When both dormant and germinating conidia were detected with FITC, no difference was observed between the WT and mutant ( Figures 2E, F ). These results suggest that the lack of Gpi7 affects the organization of the conidial cell wall and results in an increased exposure of cell wall polysaccharides in the dormant mutant conidia and fast exposure of mannose- and glucose-containing polysaccharides during germination of A. fumigatus.

Aspergillus fumigatus dormant conidia are covered by a non-covalently attached hydrophobin rodlets (Templeton et al., 1995; Valsecchi et al., 2019). To define the mechanism of the increased exposure of the cell wall polysaccharides in the mutant, we further determined the rodlet of the mutant by using the protocol described by Paris et al. (2003). As shown in Figure 3 , the surface proteins extracted from the mutant were similar with those from the WT, indicating that the content of RodA was not affected in the mutant.

Taken together, it is likely that the increased adherence of the mutant is contributed by fast germination, which led to the fast exposure of increased cell wall mannose- and glucose-containing polysaccharides of the mutant.

As we observed changes in the conidial cell wall components of the mutant, we further checked its cell wall integrity by determining the survival rate of the dormant mutant conidia in water. As a result, 40% of the mutant conidia lost their viability after 8 weeks of incubation in water at room temperature in comparison with 90% viability of the WT ones ( Figure S1A ). Interestingly, the viability of the mutant conidia remained 60% after 4 weeks of incubation in water at 42°C, whereas the viability of the WT conidia was only 40% ( Figure S1B ), which suggest a better temperature tolerance of the mutant conidia.

To evaluate the contribution of the increased adherence on the virulence of the Δgpi7 mutant, freshly harvested conidia from the WT, Δgpi7, and Regpi7 strains were inoculated into immunocompromised mice. The mice were monitored for 30 days after inoculation. Although no significant difference in mortality was documented between the WT and the Δgpi7 mutant, IA was observed in the lung tissues of mice inoculated with the mutant conidia at day 3 post-inoculation ( Figure 4A ). The histological feature of mice infected with the Δgpi7 mutant was observed with necrosis encompassed with numerous neutrophils and macrophages, while neutrophilic infiltration and necrosis were much gentler in the WT, indicating that the Δgpi7 mutant can stimulate stronger inflammatory response than the WT. The lung sections were also stained with periodic acid-Schiff stain, and the invasive hyphae of the mutant could be easily observed in the bronchial tubes and alveoli ( Figure 4B ). It was also noted that the number of conidia of the Δgpi7 mutant was much larger than that of the WT ( Figure 4B ).

After 24 h post-inoculation, the lung tissue was separated from immunosuppressed mice and then ground. The A. fumigatus conidia in ground lung tissue were washed out and counted. As shown in Figure 4C , the conidia in the lung infected by the Δgpi7 mutant were 2-fold of those from the WT or Regpi7. These results indicate an increase of adhesion to the lung cells and a resistance to killing of the Δgpi7 mutant.

To assess the immune response of the dormant conidia of the Δgpi7, phagocytic ratio and ROS production were measured. As shown in Figure 5A , at 2 h post-incubation with THP−1−derived macrophages, the phagocytic ratio of the Δgpi7 strain was significantly higher than that of the WT or Regpi7 strain. Although a previous study has shown that the mutant conidia germinate 2 h earlier than the WT, it should be pointed out that none of the WT or mutant conidia germinated after incubation at 37°C within 2 h (Ouyang et al., 2019). As shown in Figure 1 , only a few conidia of the WT or mutant were swollen after incubation at 37°C for 2 h, and no significant difference was observed between the WT and mutant. Therefore, we postulate that the increased phagocytosis is contributed by the exposure of polysaccharides on the surface of the dormant mutant conidia. In addition, an intracellular ROS production assay was carried out by co-culture of the mutant conidia with human polymorphonuclear neutrophils (PMNs) at 37°C for 1 h. As shown in Figures 5B, C , the neutrophils exposed to the Δgpi7 strain produced higher intracellular ROS than those exposed to the WT or Regpi7 strain. Collectively, these results demonstrate that the Δgpi7 dormant conidia can induce stronger immune response in vitro.

To detect whether Gpi7 influences the survival capacity of A. fumigatus in immune cells in vitro, colony-forming units (CFUs) were determined after 2 h of incubation of A. fumigatus conidia with THP−1−derived macrophages. As was noticed, CFUs recovered from the THP−1−derived macrophages infected by the Δgpi7 strain were significantly higher than other strains, which indicates a resistance to killing of the Δgpi7 strain and is consistent with the higher fungal loads of the Δgpi7 mutant in the lung of infected mice.

Aspergillus fumigatus WT, Δgpi7, and Regpi7 strains used in this study were described previously (Ouyang et al., 2019). Aspergillus fumigatus strain was grown for 2 days at 37°C on complete medium (CM) (Cove, 1966). Conidia were harvested from solid CM medium with 0.1% Tween 20. The concentration of conidia was measured by hemocytometer counting.

Spore solutions of the WT, Δgpi7, and Regpi7 strains were prepared in 0.1% Tween 20 in saline at a concentration of 1 × 10⁸ conidia/ml. Ten microliters of each spore solution was added into 200 μl of CM medium sitting on a microscopy slide and mixed well by pipetting and stirring. Two different hydrophobic surfaces were used: polystyrene cell culture slides and Permanox slides (Thermo Scientific Nunc). Ordinary glass microscopy slides were used as a non-hydrophobic surface. The slides were incubated at 37°C for 0–6 h, the culture medium was removed, the slides were washed shortly in 0.1% Tween 20 in saline, and the adherent spores were removed by running 1 ml 1% Tween 20 in saline over the slide. For each slide, this solution was collected in an E-cup, appropriate dilutions in 0.1% Tween 20 in saline were prepared, and 100 μl of each dilution was spread on the CM plate and incubated at room temperature for 2 days, after which the number of colony-forming units was determined. For statistical significance, each experiment was performed five times for each strain and condition.

Conidia (1 × 10⁸) were washed with deionized water and disrupted by 0.2 g of glass beads (0.5 mm diameter) containing 50 mM NH₄HCO₃ at pH 8.0. The conidia was broken by Disruptor Genie (Scientific Industries) 15 times for 5 min each time. Then, the cell homogenates were centrifuged and washed several times by distilled water. Three independent samples of lyophilized conidia were used for cell wall analysis, and the experiment was repeated three times. After washing, cell walls were treated with 1 M KOH and incubated at 70°C for 30 min to release glycoprotein and α-glucan. The alkali-soluble materials were acidified with acetic acid to pH 5.0, and the precipitated α-glucans were collected by centrifugation and washed with water. The glycoprotein in the supernatant was precipitated with 2 volumes of ethanol, washed twice with 64% ethanol, and dissolved in distilled water. The glycoprotein concentration was determined using the Lowry protein assay (Takaku et al., 2010). Monosaccharides were liberated from glycoproteins by acid hydrolysis (6 M HCl at 100°C for 2 h) and separated on a CarboPac PA1 anion-exchange column, equipped with an Amino Trap guard column. Elution was performed at room temperature at a flow rate of 1 ml/min with 18 mM NaOH. The alkali-insoluble materials were washed with water several times and digested in 6 M HCl at 100°C for 2 h to release monosaccharides from β-glucan and chitin. After digestion, HCl was evaporated and the residues were dissolved in 0.2 ml distilled water (Yamashita et al., 1999; Lamson et al., 2002; Guo et al., 2015). The amounts of α-glucan and β-glucan present were estimated by measuring released glucose using the phenol/sulfuric acid method (Zhang et al., 2016). Chitin content was determined by measuring the N-acetylglucosamine released after digestion (McMillan et al., 1999).

Spores (1 × 10⁵) were inoculated in 200 ml of CM and incubated at 37°C. The resting spores and swollen conidia were separately stained with ConA-FITC or WGA-FITC and then examined under a fluorescence microscope.

The rodlet layer was extracted as described by Paris et al. (2003). Briefly, the conidia were subjected to sonication. After removal of the remaining conidia by low-speed centrifugation, the supernatant was ultracentrifuged for 1 h at 50,000×g. The pellet was boiled in SDS-PAGE loading buffer and then washed twice with SDS-PAGE loading buffer and three times with distilled water. The resulting pellet was lyophilized. The lyophilized material was subsequently treated with 100% trifluoroacetic acid (TFA) for 10 min at room temperature. Then, the acid was removed under a stream of nitrogen. Dried extracts were dissolved in SDS-PAGE loading buffer and incubated in boiling water for 15 min. Proteins were subjected to SDS-PAGE (15% polyacrylamide) and visualized by silver staining.

The conidia (1 × 10⁵) were kept on distilled water for different storage times (1, 2, 4, 8 weeks) at different temperatures (room temperature and 42°C) separately. To calculate the spore survival, the conidia were inoculated on CM and cultured at 37°C. The germination of conidia was counted. For statistical significance, each experiment was performed five times for each strain.

The human monocyte cell line—THP−1 cells—was cultured in RPMI 1640 (Gibco, Thermo Fisher Scientific, Inc., Waltham, MA, USA) containing 10% fetal bovine serum (FBS, Gibco, Thermo Fisher Scientific) and 1% penicillin–streptomycin antibiotic mixture at 37°C with 5% CO₂. To prepare THP−1−derived macrophages, THP−1 cells (1 × 10⁶/ml) were plated onto six-well plates and incubated with phorbol 12-myristate 13-acetate (PMA) at a concentration of 50 ng/ml for 48 h. After incubation, adherent macrophages were maintained in complete medium at 37°C with 5% CO₂ and utilized in phagocytosis and killing experiments (Shabani et al., 2020; Qin et al., 2020).

THP−1−derived macrophages (1 × 10⁶/ml) were exposed to resting conidia of the WT, Δgpi7, and Regpi7 strains (MOI = 10) in six−well plates at 37°C and 5% CO₂ for 2 h, respectively. At 2 h time point, non−adherent cells and non−phagocytosed conidia were removed by washing the cells three times with PBS. To lyse the cells and harvest the conidia, sterile water was added and mixed vigorously with standing for 5 min. Cellular lysis was confirmed by microscopy. The serial dilutions were performed in sterile water and immediately plated on potato dextrose agar (PDA; BD Biosciences, San Jose, CA, USA). Colonies were counted following incubation for 24 h at 37°C (Liu et al., 2017).

To prepare fluorescein isothiocyanate (FITC)−labeled conidia, a total of 1 × 10⁷/ml conidia were suspended in 100 ml 0.05 M sterile carbonate–bicarbonate buffer (cat. no. C3041; Sigma−Aldrich; Merck KGaA) (pH 9.6). After 10 mg FITC powder was dissolved in 1 ml DMSO, the FITC solution was quickly added to the above carbonate–bicarbonate solution containing 1 × 10⁷/ml conidia, and then the solution was stirred with a magnetic stirrer about 1 h at 4°C in the dark. The suspension was then washed three times with PBS. The FITC−labeled conidia were resuspended in PBS and adjusted to the desired concentration (1 × 10⁹/ml). The FITC-labeled conidia were confirmed by a fluorescence microscope.

THP−1−derived macrophages (1 × 10⁶/ml) were co−cultured with FITC−labeled resting conidia of WT, Δgpi7, and Regpi7 strains (MOI = 10) in 2 ml complete RPMI 1640 medium at 37°C and 5% CO₂ for 2 h, respectively. The supernatants were discarded and the wells were washed gently three times with ice-cold PBS. THP−1−derived macrophages were lifted from the wells with gentle pipetting with wash buffer. To define the ingested conidia, THP−1−derived macrophages were further stained with FITC-conjugated anti-human CD11b antibody. Data were acquired on a BD FACSCalibur system and analyzed with the FlowJo 7.6 software (Marr et al., 2001).

Human PMNs were isolated from whole blood specimens (three healthy donors) by density-gradient centrifugation using the Ficoll-Paque Plus (GE Healthcare, USA) as described previously (Voyich et al., 2005). PMN was resuspended in 1 ml HBSS and adjusted to the concentration of 2 × 10⁶/ml. Using a 96-well flatbottom plate, 50 μl PMN (2 × 10⁶/ml), 50 μl 40% FBS, and 50 μl conidia (4 × 10⁷/ml) were put together, with 50 μl luminol solution added finally. Chemiluminescence was measured at 37°C for 60 min in 3 min intervals in an automated LB96V MicroLumat Plus luminometer (EG&G Berthold, Germany) (Guerra et al., 2016).

Virulence of the WT, Δgpi7, and Regpi7 strains was detected with immunosuppressed mice (Latgé, 1999; Li et al., 2007; Sugui et al., 2007; Li et al., 2012). Briefly, four groups (control, WT, Δgpi7, and Regpi7) of strains each containing 20 male BALB/c mice (18–20 g) were used for the virulence experiments. Fresh conidia were washed from CM plates and suspended in 0.01% Tween 20 in saline with inoculum of 3 × 10⁵ CFU/g mouse weight in 30 μl volume. Mice were immunosuppressed by injection with 150 mg/kg mouse weight cyclophosphamide on days −3, −1, +3, +6, and +9 and 200 mg/kg mouse weight hydrocortisone on day −1. Mice were inoculated with conidia by nasal feeding on day 0 and monitored twice each day for 30 days after inoculation and mortality was recorded. Mice surviving in the experiment were humanely terminated on day 30. The lung from each mouse was homogenized by OSE-Y30 (TIANGEN) at day 1 post-infection. The conidia were washed out from the same amount of homogenate, diluted, and counted by flat dilution counting. The right lung from each mouse was dissected at day 3 post-infection and fixed in 4% (v/v) paraformaldehyde in physiological saline. Sections were stained with hematoxylin–eosin (HE), and periodic acid-Schiff (PAS) by standard techniques. For statistical significance, each experiment was performed five times for each strain.