Disruption of the Phospholipase D Gene Attenuates the Virulence of Aspergillus fumigatus

Abstract

Aspergillus fumigatus is the most prevalent airborne fungal pathogen that induces serious infections in immunocompromised patients. Phospholipases are key enzymes in pathogenic fungi that cleave host phospholipids, resulting in membrane destabilization and host cell penetration. However, knowledge of the impact of phospholipases on A. fumigatus virulence is rather limited. In this study, disruption of the pld gene encoding phospholipase D (PLD), an important member of the phospholipase protein family in A. fumigatus, was confirmed to significantly decrease both intracellular and extracellular PLD activity of A. fumigatus. The pld gene disruption did not alter conidial morphological characteristics, germination, growth, and biofilm formation but significantly suppressed the internalization of A. fumigatus into A549 epithelial cells without affecting conidial adhesion to epithelial cells. Importantly, the suppressed internalization was fully rescued in the presence of 100 μM phosphatidic acid, the PLD product. Indeed, complementation of pld restored the PLD activity and internalization capacity of A. fumigatus. Phagocytosis of A. fumigatus conidia by J774 macrophages was not affected by the absence of the pld gene. Pretreatment of conidia with 1-butanol and a specific PLD inhibitor decreased the internalization of A. fumigatus into A549 epithelial cells but had no effect on phagocytosis by J774 macrophages. Finally, loss of the pld gene attenuated the virulence of A. fumigatus in mice immunosuppressed with hydrocortisone acetate but not with cyclophosphamide. These data suggest that PLD of A. fumigatus regulates its internalization into lung epithelial cells and may represent an important virulence factor for A. fumigatus infection.

INTRODUCTION

Aspergillus fumigatus, an airborne fungal pathogen, causes a wide range of diseases, including allergic bronchopulmonary aspergillosis, aspergilloma, and invasive aspergillosis (54, 58). Inhalation of conidia and their colonization in the alveoli may result in their germination and growth in the lung, and in immunocompromised individuals, the pathogen may spread and result in the development of invasive aspergillosis associated with high mortality. However, the molecular mechanisms underlying the pathogenesis of invasive aspergillosis remain poorly understood.

Currently, it is generally accepted that there is no unique essential virulence factor for A. fumigatus, and its virulence appears to be under polygenetic control (38, 58). At least four groups of molecules and genes of A. fumigatus associated with its virulence have been studied intensively (2, 54), including cell wall components required for massive growth during infection (46), such as beta-1,3-glucan and galactomannan (AFMP1 and AFMP2) (70); stress response genes and molecules which have been implicated in the evasion from the host immune response, such as cyclic AMP (cAMP)-dependent protein kinases (pkaR and pkaC) (54) and mitogen-activated protein (MAP) kinases (mpkA) (67); a number of genes and molecules which allow A. fumigatus to compete in its environmental niche, such as genes involved in iron (9) and phosphorous acquisition (10, 50); and toxins and allergens, as well as enzymatic proteins secreted by A. fumigatus, such as alkaline serine proteases (Alp and Alp2) (53) and phospholipases (59) that damage host cells and facilitate tissue infection. Phospholipases hydrolyze ester linkages in glycerophospholipids, which are one of the major chemical constituents of the host cell envelope and, hence, may destabilize the host cell membrane to mediate microbe entry into host cells and tissues. Phospholipases constitute of a heterogeneous group of enzymes, including phospholipases A (PLA), B (PLB), C (PLC), and D (PLD) (21). It has been shown that phospholipases, especially PLB, are essential virulence factors in the pathogenesis of several important pathogenic fungi, including Candida albicans (41) and Cryptococcus neoformans (12). Although the activities of the A. fumigatus phospholipases and three A. fumigatus plb genes have been characterized (5, 59), our current knowledge of the impact of A. fumigatus phospholipases, in particular, PLD, on the virulence of this pathogen is rather limited (1).

PLD hydrolyzes the phosphodiester bond in the phospholipid backbone through its highly conserved HKD motifs to yield phosphatidic acid (PA) and choline or ethanolamine, depending on the specific phospholipid species involved, i.e., phosphatidylcholine or phosphatidylethanolamine (3, 17, 21). Currently, mammalian PLDs are recognized as key enzymes in intracellular signaling involved in processes such as inflammation, endocytosis, and cell shape changes (27), while bacterial PLDs from Corynebacterium pseudotuberculosis and Acinetobacter baumannii have been shown to be the critical virulence determinants of these organisms (25, 29, 44). In fungi, PLD appears to be closely related to fungal cell shape changes, such as sporulation in Saccharomyces cerevisiae (55) and the dimorphic transition of C. albicans (14). Moreover, C. albicans PLD1-deficient mutants exhibit a substantially reduced ability to be internalized by epithelial cells and low virulence in immunodeficient mice, indicating that PLD may also be an important virulence factor in fungal pathogenesis. To date, three PLD isoforms, PLD, PLD1, and PLDA, have been reported in A. fumigatus, but their extracellular existence remains undetermined, and their role in pathogenesis has yet to be studied. Compared to PLD1 and PLDA, PLD of A. fumigatus, encoded by the pld gene, is rather specific and more distinct from the PLDs in other medically important fungi by phylogenetic analysis (26, 31). Therefore, in this study we chose to explore the function of the pld gene in the development and virulence of A. fumigatus.

MATERIALS AND METHODS

Fungal and bacterial strains, media, and chemical reagents.

The strains used in this study are listed in Table S1 in the supplemental material. A. fumigatus B5233 was used as the wild-type (WT) strain for all in vitro and animal model experiments. All A. fumigatus cultures were grown in Aspergillus minimal medium (AMM) at 37°C unless otherwise specified. Escherichia coli DH-5α was used for routine cloning and was grown in Luria-Bertani broth at 37°C. Agrobacterium tumefaciens strains were grown either in Luria-Bertani broth supplemented with 50 mg/liter kanamycin or in induction medium supplemented with 200 μM acetosyringone (IMAS). Transformants were selected using AMM supplemented with 200 mg/liter hygromycin (Roche, Mannheim, Germany) and 200 mg/liter cefotaxime (62). The mammalian PLD1-specific inhibitor VU0359595 (Avanti product no. 857371) (42), the PLD2-specific inhibitor VU0285655-1 (Avanti product no. 857372) (39), and PA (Avanti product no. 840101P) were purchased from Avanti Polar Lipids (Alabaster, AL). Hydrocortisone acetate (Bio Basic Inc., Markham, Canada) and cyclophosphamide (Sigma-Aldrich, Saint Louis, MO) were used for immunosuppression.

Construction of the Δpld and pldC strains of A. fumigatus.

The Δpld mutant of A. fumigatus B5233 was constructed by A. tumefaciens-mediated transformation (ATMT) as described previously (62, 66). Briefly, the deletion vector was constructed by cloning a 5.8-kb sequence, including a 1,220-bp fragment upstream and a 1,032-bp fragment downstream of the coding region of the A. fumigatus pld gene, into pDHt/SK.2 to produce plasmid A using primers P1 and P2 (see Table S2 in the supplemental material). Subsequently, a 2.9-kb PCR product of the hygromycin resistance gene (hph) was amplified from pDHt-hph-hindIII-sacI (62) using primers P3 and P4 (Table S2) to add the BglII and AsuII restriction sites. The hph fragment, which was digested with BglII and AsuII, was ligated into plasmid A to create plasmid B (43, 62). Plasmid B was transformed into competent A. tumefaciens EHA105 using the freeze-thaw method (52, 62). The resulting strain of A. tumefaciens was designated the EHA105h strain. To obtain the Δpld strain, conidia of strains B5233 and EHA105h were cultivated together at a ratio of 1:10 (conidia to bacteria). A total of 100 μl of the Agrobacterium culture was mixed with 100 μl of B5233 conidia and spread onto a nylon filter placed on an AMM agar plate supplemented with 0.2 mM acetosyringone. Plates containing the filters were incubated at 24°C in the dark for 48 h. To select transformants, filters containing the transformants were transferred to AMM agar plates supplemented with 200 μg/ml hygromycin and 200 μg/ml cefotaxime and incubated at 37°C for 72 h (62). Transformants were scraped from the nylon filter and transferred onto Sabouraud medium plates. The Δpld strain was initially screened by PCR with primers designed to amplify the regions of pld (primers P7 and P8) (see Table S2 in the supplemental material) that should have been deleted in the Δpld strain. An additional PCR screen was performed to amplify the junctions of homologous sequences and hph to indicate replacement and homologous recombination (primers P9 and P10) (Table S2).

To ensure that the mutant phenotype obtained could be attributed to the specific desired deletion, the Δpld strain was reconstituted by integration of the B5233 pld allele to create a complementation strain, the pldC strain. Briefly, the 5.8-kb SpeI-KpnI DNA fragment containing the partial pld promoter, pld open reading frame (ORF), and pld terminator was cloned into pDHt/SK to produce plasmid C. The 2.7-kb PCR product of a phleomycin resistance gene (ble cassette) was then amplified from pAN8.1 using primers P5 and P6 (see Table S2 in the supplemental material) and subcloned into plasmid C (19). The resulting plasmid was linearized at the unique KpnI site. Protoplasts of the pld mutant were transformed with 10 μg of this linear construct. Transformants were selected on 375 μg/ml phleomycin plates. To confirm the constructed strains, Southern blot analysis was performed with XbaI-digested genomic DNA. A 454-bp fragment of the pld gene was used as a probe for Southern hybridization (primers P11 and P12) (Table S2).

PLD activity assay.

The conidia (2.5 × 10⁸/ml) from each strain were inoculated in 50 ml of liquid Sabouraud medium and incubated at 37°C with shaking at 150 rpm for 4 h and 8 h to obtain the swollen conidia and hyphae, respectively. The swollen conidia or hyphae and supernatants were separated by centrifugation at 10,000 × g for 2 min and prepared for PLD activity assays. After being washed three times with phosphate-buffered saline (PBS), the swollen conidia or hyphae were resuspended in a volume of 300 μl of lysis buffer consisting of PBS with 1% Triton X-100, 10 μg/ml leupeptin, 10 μg/ml aprotinin, 10 μg/ml soybean trypsin inhibitor, and 100 μM phenylmethylsulfonyl fluoride. Cell extracts were prepared by addition of an equal volume of 0.5-mm glass beads (Sigma-Aldrich) to the cells, followed by shaking at 1,800 rpm in an IKA-Vibrax shaker at 4°C for 15 min. The cell lysates were collected by centrifugation at 10,000 × g for 2 min (11). PLD activity was determined with the Amplex Red phospholipase D assay kit (Invitrogen, Eugene, OR) (29). Each PLD reaction mixture contained 50 μM Amplex Red reagent, 1 U/ml horseradish peroxidase (HRP), 0.1 U/ml choline oxidase, 0.25 mM lecithin, and 100 μl of the sample. The reaction lasted for 30 min, and fluorescence was excited at 530 nm and measured at 590 nm using a fluorescence microplate reader (SpectraMax M5). In parallel, the extracellular PLD activities in the Sabouraud medium culture supernatants were also determined with the same methods.

Morphological characterization and measurement of mycelial growth rate.

A total of 5 × 10² conidia (5 μl) were inoculated centrally in AMM, Sabouraud medium, and egg yolk lecithin medium agar plates and cultured at 37°C for 72 h. Every 12 h, the colonies of each strain were observed and photographed under an inverted microscope (Olympus Leica DMI3000 B). Additionally, the colony diameter was measured every 12 h, and the mycelial growth rate was determined as the increase in colony diameter per hour (mm/h).

Biofilm formation assay.

Quantification of the biofilm formation for A. fumigatus was performed as described in a previous report (22). Briefly, 100 μl of liquid AMM containing 2 × 10⁴ conidia per well was added to a 96-well plate and incubated for 24 h at 37°C. The medium was removed, and the wells were washed three times with PBS. Subsequently, 150 μl of 0.5% crystal violet solution was added for 5 min to stain the residual material in the well. Excess stain was gently removed under running water, and the biofilm was then destained by addition of 200 μl of 95% ethanol. The biofilm density was measured by determining the absorbance of the destaining solution at 570 nm using a microplate reader (SpectraMax M5).

In vitro internalization and phagocytosis assay.

Internalization of A. fumigatus into lung epithelial cells and phagocytosis of the pathogens by macrophages were analyzed as described in a previous report (68). Briefly, human A549 lung epithelial cells or J774 murine macrophages were grown to confluence in 96-well plates (approximately 2 × 10⁴ cells per well). Subsequently, 2 × 10⁵ conidia were added and incubated at 37°C under 5% CO₂ to induce internalization or phagocytosis. After 4 h for internalization or 1 h for phagocytosis, the cell monolayers were washed three times with PBS, and 100 μl Dulbecco modified Eagle medium (DMEM) supplemented with 10% fetal calf serum and 20 μg/ml nystatin was added to each well and incubated for 3.5 h to kill noninternalized conidia. The cell monolayers were then washed and treated with 100 μl of PBS containing 0.1% Triton X-100 for 15 min at 37°C to induce cell lysis and the release of internalized conidia. The released conidia were diluted onto Sabouraud plates and incubated at 37°C for 20 h. Colonies were counted to determine the total number of bound and intracellular conidia (29). The internalization rate was determined to be the percentage of intracellular conidium colonies compared to the initial inoculum of conidia. To ensure that the nystatin had eliminated extracellular conidia, the supernatants of the cell cultures were plated onto Sabouraud plates as controls.

Adherence assay.

The adherence capacity of A. fumigatus to epithelial cells was determined as described previously (22, 60). The 6-well plates were precoated with 1 ml of 0.01 mg/ml fibronectin in PBS (without Ca²⁺/Mg²⁺) and incubated at 37°C for 4 h. A549 cells were seeded in this fibronectin-coated well and grown for 24 h. The conidia (1.5 × 10²) or germlings (germinated for 8 h) in Hanks' balanced salt solution (HBSS)-0.01% Tween 20 were added into the wells and incubated at 37°C for 30 min, followed by being washed three times with HBSS to remove nonadherent organisms and overlaid with AMM agar at 45°C. The number of adherent organisms was quantified by colony counting. The adhesion rate was determined as the percentage of adherent colonies compared to the total number of conidia.

In vivo virulence assay.

The Institutional Animal Care and Use Committee of the Academy of Military Medical Sciences approved the animal experiments. The wild-type, Δpld, and pldC strains were used for experimental infections in white male BALB/c mice weighing 18 to 22 g. Mice were immunosuppressed by one of two methods. One was subcutaneous injection of 5 mg hydrocortisone acetate (Sigma-Aldrich) in 300 μl PBS-0.1% Tween 20 on days −4, −2, 0, 2, and 4 of infection (64); the other was intraperitoneal injection of 3 mg cyclophosphamide (Sigma-Aldrich) on days −4, −1, and 3 of infection and a single intraperitoneal injection of 4 mg hydrocortisone on day 0 (15). Mice were housed under sterile conditions and provided with sterile drinking water containing 500 μg/ml tetracycline hydrochloride. On day zero, mice were anesthetized by inhalation of diethyl ether and infected intranasally with 5 × 10⁶ conidia in 30 μl of PBS-0.01% Tween 20. Control mice were also immunosuppressed with hydrocortisone or cyclophosphamide and inoculated intranasally with 30 μl PBS-0.01% Tween 20. Eight mice per group were used. Morbidity and mortality were monitored for up to 16 days, and the Kaplan-Meier survival analysis with log rank test was used for comparison among the groups. To determine the lung fungal burden, immunosuppressed mice were infected intranasally with 5 × 10⁵ conidia in 30 μl of PBS-0.01% Tween 20. Twelve mice from each group were used. Mice were sacrificed at 2, 72, and 120 h postinfection. One milliliter of lung lavage fluid was harvested and centrifuged at 10,000 × g for 3 min. The precipitate was resuspended, serially diluted, and plated on Sabouraud agar plates for incubation at 37°C. The number of colonies in lung lavage fluid (number of CFU/ml) was counted and calculated after 20 h. In parallel, the lung tissues were weighed, homogenized, and then serially diluted onto Sabouraud agar plates for incubation at 37°C. After 20 h, the number of colonies in lung tissues (number of CFU/g) was also counted and calculated. For histopathological examination, the lung tissue sections obtained from mice from each group at 72 h postinfection were dissected, fixed in 10% (vol/vol) formaldehyde, and stained with hematoxylin and eosin (H&E) and periodic acid-Schiff (PAS). Images were observed and captured under light microscopy (Olympus BX51). The areas of the sites that contained fungal hyphae with inflammatory cell infiltration were determined by use of Image-Pro Express (version 6.0; Media Cybernetics). Briefly, the software was utilized to outline lesions, and the outlined area was determined for 8 sites of infection for each strain tested (61). In addition, we determined the number of lesions per lung for 4 mice that had been infected with the B5233 wild-type (WT), Δpld, or pldC strain (61).

Pretreatment of A. fumigatus conidia with chemical inhibitors.

Samples of 2 × 10⁵ conidia of different A. fumigatus strains were pretreated in 100 μl of DMEM in the presence of 1 μl dimethyl sulfoxide (DMSO), 1% 1-butanol, 4 nM VU0359595 (mammalian PLD1-specific inhibitor), 400 nM VU0285655-1 (PLD2-specific inhibitor), or both 4 nM PLD1 inhibitor and 400 nM PLD2 inhibitor. After 4 h at 37°C, the conidia were washed three times with PBS for a subsequent internalization and germination assay.

Statistical analysis.

The data shown in the figures are either from a representative experiment or presented as the means ± standard errors (SE) from 3 or 4 independent experiments performed in triplicate. For experiments comparing groups, single-factor analysis of variance (ANOVA) was performed, and Student's paired t test was performed between two groups. Survival curves were analyzed using the log rank test. P values of <0.05 were considered statistically significant.

RESULTS

Construction of the Δpld and pldC strains of A. fumigatus.

To disrupt the pld gene using homologous recombination, a binary plasmid pDHt/Δpld::hph with a hygromycin resistance gene (hph) and 5′- and 3′-end flanking sequences of pld was first constructed for homologous recombination (see Table S1 and Fig. S3 in the supplemental material). As Agrobacterium tumefaciens is able to transfer a fragment of DNA between the left and right borders of its tumor-inducing plasmid to the A. fumigatus genome at a high homologous recombination rate (62), it was used in this study to transfer pDHt/Δpld::hph to A. fumigatus B5233. Overall, approximately 75% of the transformants included the hph locus but no detectable pld loci, as shown in Fig. S3B in the supplemental material. To complement the pld defect in the Δpld strain, a PCR product containing the entire B5233 pld locus with an additional 1.2 kb of 5′- and 3′-end flanking sequences and a 2.7-kb bleomycin resis-tance gene was transformed into the recipient Δpld strain (Fig. S3C). The gene deletion and complementation were confirmed by Southern blotting. As shown in Fig. S3D, the A. fumigatus B5233 wild-type strain presented a 4.0-kb band; meanwhile, the pld-deleted strain exhibited a band at this position of 2.3 kb, and the complemented strain had two blot bands, one at 2.3 kb and the other at 7.6 kb (43). A randomly selected pld-deleted strain and a complemented strain were designated the Δpld and pldC strains, respectively, and used throughout this study (51).

Disruption of the pld gene reduces both intracellular and extracellular PLD activities of A. fumigatus.

As it remains not clearly determined whether A. fumigatus secretes PLD outside the mycelium (14), we tested both the intracellular and extracellular PLD activities of the A. fumigatus B5233 wild-type, Δpld, and pldC strains. As shown in Fig. 1A, the intracellular PLD activity of Δpld swollen conidia that had germinated for 4 h was significantly lower (∼56% decrease) than that of wild-type swollen conidia. Importantly, upon complementation of the pld gene, intracellular PLD activity was fully restored in the pldC strain (Fig. 1A). The intracellular PLD activities in the hyphae of conidia that had germinated for 8 h were obviously higher than that of the swollen conidia that had germinated for 4 h, and they were lowered by disruption of the pld gene as well (Fig. 1A). On the other hand, the extracellular PLD activity of the Δpld strain in the supernatants of the A. fumigatus cultures of conidia germinated for 4 h or 8 h exhibited a striking decrease (∼60%) compared to the activities of the wild-type and pldC strains (Fig. 1B). The extracellular PLD activities in different strains increased along with the germination as well (Fig. 1B). These results indicate that A. fumigatus secretes the PLD enzyme, and disruption of the pld gene in A. fumigatus reduces both intracellular and extracellular PLD activities.

Fig 1.

pld gene disruption inhibits both intra- and extracellular PLD activities of A. fumigatus. Resting conidia from the A. fumigatus B5233 wild-type (WT), Δpld, and pldC strains were inoculated and cultured in liquid Sabouraud medium. After 4 h and 8 h, the swollen conidia and hyphae, respectively, were harvested by centrifugation. (A, B) Intracellular PLD activity in swollen conidia or hyphae (A) and extracellular PLD activity in culture medium (B) were measured by an Amplex Red reagent-based assay. The data shown here are the means ± SE (n = 3 to 4 independent experiments performed in triplicate). Depicted are the differences in fluorescence intensities between the wild-type and pld gene mutant strains. *, P < 0.05.

Disruption of the pld gene does not affect conidial morphological characteristics, hyphal growth, and biofilm formation of A. fumigatus.

As the PLD signal is associated with cell shape changes in mammalian cells and other fungi (14, 27, 55), we studied potential morphological alteration in A. fumigatus induced upon the pld gene disruption. Compared to the wild-type and the pldC strains cultured on AMM plates (Fig. 2A), the Δpld strain did not exhibit any morphological changes, including the pigmentation of colonies. Similar results were also observed on the lecithin medium plates that contain lecithin as the sole carbon source (data not shown). The hyphal growth rate of the Δpld mutant was indistinguishable from those of the wild-type and the pldC strains on AMM, Sabouraud, and lecithin medium plates at 37°C (Fig. 2B) and 28°C (data not shown). No differences in the biofilm density under static culture among the Δpld, the wild-type, and the pldC strains were found (Fig. 2C). Taken together, the results suggest that the pld gene of A. fumigatus plays a minor role in its morphology and growth.

Fig 2.

Disruption of the pld gene does not alter the conidia morphological characteristics, hyphae growth, and biofilm formation. Resting conidia from the wild-type (WT) strain, the Δpld strain, and the pldC strain of A. fumigatus B5233 were cultured on AMM, Sabouraud, and lecithin plates at 37°C, respectively. (A) At 24, 36, and 72 h, the colonies of different strains were compared by microscopy. Pictures shown are characteristic colonies on AMM plates for the independent experiments. (B) Every 12 h, the diameter of each colony of different strains was measured. The hyphal growth rates on the three types of media were calculated. (C) Resting conidia from the wild-type (WT) strain, the Δpld strain, and the pldC strain of A. fumigatus B5233 were inoculated into 96-well plates and cultured for 24 h at 37°C, and biofilm density was determined by measuring absorbance using a microplate reader. The data shown here are the means ± SE (n = 3 or 4 independent experiments performed in triplicate). Depicted are the differences in the hyphal growth rates or biofilm density absorbances between the wild-type and pld gene mutant strains.

Disruption of the pld gene reduces A. fumigatus internalization but has no effect on adhesion on A549 cells and phagocytosis by macrophages.

Internalization into nonphagocytic epithelial cells is an important mechanism for A. fumigatus to avoid host immune attack and spread its infection, while the phagocytosis of A. fumigatus conidia by host macrophages represents a critical innate immune reaction against fungal infection. Therefore, we studied the impact of the pld gene disruption in A. fumigatus on the internalization into lung epithelial cells and phagocytosis by macrophages in vitro. As shown in Fig. 3A, the internalization of the Δpld strain into A549 cells decreased significantly compared to that of the wild-type strain (0.60% ± 0.07% versus 1.24% ± 0.05%; P < 0.05), whereas the pldC strain reconstituted the internalization level to that of the wild type (1.22% ± 0.07%). In contrast, the Δpld mutant was phagocytosed by J774 macrophages as efficiently as the wild-type and pldC strains (Fig. 3B). Next we studied whether the reduced internalization of the Δpld mutant into epithelial cells may be due to a defect of its adhesion on A549 epithelial cells. As shown in Fig. 3C, there was no significant difference in the adhesion of conidia and hyphae to A549 cells between the Δpld and wild-type strains. These results indicate that A. fumigatus PLD may improve internalization into lung epithelial cells without affecting the conidial adherence to epithelial cells.

Fig 3.

pld gene disruption suppresses the internalization of A. fumigatus into A549 cells but does not alter the phagocytosis of A. fumigatus by J774 macrophages and the adhesion of A. fumigatus on A549 cells. (A, B) A549 cells (A) and J774 macrophages (B) were infected with the wild-type (WT) strain, the Δpld strain, and the pldC strain of A. fumigatus B5233 at a multiplicity of infection (MOI) of 10. The internalization of A. fumigatus into A549 cells and phagocytosis by J774 macrophages were analyzed by the nystatin protection assay. (C) A549 cells were incubated with conidia or hyphae (germination for 8 h) from the wild-type (WT) strain, the Δpld strain, and the pldC strain of A. fumigatus B5233 at 37°C for 30 min, and the adhesion on A549 cells was measured. The data shown here are the means ± SE (n = 3 or 4 independent experiments performed in triplicate). Depicted are the differences in the percentages of total inoculated conidia between the wild-type strain and the pld gene mutants. *, P < 0.05.

Pretreatment of A. fumigatus conidia with PLD inhibitors suppresses their internalization into A549 cells.

The significant decreases in PLD activity and internalization of the Δpld mutant prompted us to speculate that pharmacological inhibition of the PLD activity of A. fumigatus may also suppress its internalization into lung epithelial cells. To test this hypothesis, we pretreated the conidia of wild-type A. fumigatus with the nonspecific PLD inhibitor 1-butanol and isoform-specific mammalian PLD inhibitors. As shown in Fig. 4A, the conidia pretreated with either 1% 1-butanol or 4 nM mammalian PLD1-specific inhibitor VU0359595 internalized less efficiently into A549 cells than the control group pretreated with PBS. This result was consistent with the effects of pld gene disruption on A. fumigatus internalization. Pretreatment with the mammalian PLD2-specific inhibitor VU0285655-1 had no effect on A. fumigatus internalization into A549 cells. These data gave a hint that the structure and function of A. fumigatus PLD might be more similar to mammalian PLD1 but not PLD2. Moreover, the phagocytosis of conidia by J774 macrophage cells was not affected by pretreatment with any of the inhibitors (Fig. 4B), indicating again that the pld gene function of A. fumigatus does not interfere with macrophage phagocytosis of A. fumigatus. The inhibitor concentrations used were shown not to compromise the fungal cell viability and conidial germination (Fig. 4C and D). These findings indicate that pharmacological inhibition of PLD function in A. fumigatus may impair its internalization into lung epithelial cells.

Fig 4.

Pharmacological PLD inhibitors reduce the internalization of A. fumigatus into A549 cells. Conidia of the A. fumigatus B5233 wild-type strain were pretreated for 4 h with 1× PBS (control), 1% 1-butanol, 4 nM mammalian PLD1-specific inhibitor VU0359595, 400 nM mammalian PLD2-specific inhibitor VU0285655-1, or 4 nM mammalian PLD1-specific inhibitor VU0359595 and 400 nM mammalian PLD2-specific inhibitor VU0285655-1 (both). (A, B) Thereafter, the internalization of A. fumigatus into A549 cells (A) and the phagocytosis of A. fumigatus into J774 macrophages (B) were analyzed by the nystatin protection assay. (C, D) The conidial survival rate (C) and the germination rate (D) of A. fumigatus after pretreatment with the PLD inhibitors were also analyzed. The data shown here are the means ± SE (n = 3 or 4 independent experiments performed in triplicate). Depicted are the differences in the percentages of total inoculated conidia between the wild-type and the PLD inhibitor-pretreated group. *, P < 0.05.

Impact of phosphatidic acid on the internalization of the Δpld strain into A549 cells.

Next, we asked whether the PLD product, PA, regulates the internalization of A. fumigatus into A549 cells (71). As shown in Fig. 5, the presence of PA significantly promoted the internalization of wild-type and Δpld strains into A549 cells in a dose-dependent fashion, peaking at a concentration of 100 μM. Intriguingly, PA did not cause any significant difference in internalization rates between the wild-type and Δpld strains, with the exception that at a concentration of 100 μM, PA was able to restore the internalization of the Δpld mutant to the level of the wild-type strain. Although an effect of exogenous PA on host cellular signal regulation could not be excluded, our data indicate that extracellular addition of PA can improve the internalization of A. fumigatus into A549 cells.

Fig 5.

Impact of phosphatidic acid on the internalization of the Δpld strain into A549 cells. A549 cells were infected with the resting conidia from the wild-type (WT) strain and the Δpld strain of A. fumigatus B5233 at an MOI of 10 at 37°C for 4 h in the presence of the indicated phosphatidic acid concentrations. The internalization of A. fumigatus into A549 cells was analyzed by the nystatin protection assay. The data shown are the means ± SE (n = 3 or 4 independent experiments performed in triplicate). Depicted are the differences in the percentages of total inoculated conidia between the wild-type strain and Δpld strain. *, P < 0.05.

PLD is required for the virulence of A. fumigatus in mice immunosuppressed with hydrocortisone acetate but not with cyclophosphamide.

To determine the contribution of PLD to the virulence of A. fumigatus, freshly harvested conidia from the wild-type, the Δpld, and the pldC strains of A. fumigatus B5233 were inoculated into two groups of BALB/c mice that were immunosuppressed by either hydrocortisone acetate or cyclophosphamide. Mice mortality was monitored for 16 days postinoculation. Mice immunosuppressed with hydrocortisone acetate possessed a clear difference in the survival rates between the groups infected with the wild-type and Δpld strains (P = 0.009) (Fig. 6A). Nearly all wild-type or pldC strain-infected mice (15/16) died within a week, whereas only 60% of the Δpld strain-infected mice died in the same period. At day 16 postinfection, approximately 25% of the Δpld strain-infected mice were still alive. At 72 h postinfection, the number of A. fumigatus conidia in lung lavage fluid from the Δpld strain-infected mice was significantly lower than that of wild-type or pldC strain-infected mice (Fig. 6B). Similarly, a significant difference in the fungal burdens in lung tissue homogenates between the wild-type strain- or Δpld strain-infected group was also found (Fig. 6C). In contrast, mice immunosuppressed with cyclophosphamide possessed no significant difference between the mortality levels of the wild-type and Δpld strains (P = 0.272) (Fig. 6D). In line with this result, no alterations in fungal burden in either lavage fluid or lung homogenates were found between the wild-type and Δpld strain groups (Fig. 6E and F). These results suggest that the virulence of the Δpld strain compared to that of the wild-type strain was reduced in hydrocortisone acetate-treated mice but not in cyclophosphamide-treated mice. Histological analysis of the lung sections confirmed the reduced virulence of the Δpld strain in hydrocortisone acetate-immunosuppressed mice. As revealed by H&E stainings at 72 h postinfection, hyphae were seen primarily within the bronchioles of mice infected with the wild-type or pldC strain (Fig. 7A; see also Fig. S5A in the supplemental material). The lung sections from mice infected with the wild-type and pldC strains had multifocal necrosis and inflammatory cell infiltration, the typical characteristics of invasive pulmonary aspergillosis (63) (Fig. 7A). Though hyphae were also observed in the lung tissues from mice infected with the Δpld strain conidia, the fungal growth and the size and number of lesions per lung were reduced significantly in Δpld strain-infected mice relative to those in wild-type strain-infected mice (Fig. 7A, C, and E; see also Fig. S5A in the supplemental material). Importantly, the reduced Δpld strain-induced pulmonary lesions in comparison to those of the wild-type strain were corroborated by the survival data (Fig. 6A and 7A). In mice immunosuppressed with cyclophosphamide, hyphae and necrosis with inflammatory cell infiltration were also easily observed in the lung sections prepared 72 h postinfection with the wild-type, Δpld, and pldC strains (Fig. 7B; see also Fig. S5B in the supplemental material). In comparison to the infections with the wild-type and pldC strains, the fungal growth and the size and number of lesions per lung caused by the Δpld strain were not significantly altered (Fig. 7D and F), which was also in line with the survival data (Fig. 6D). Taken together, these results indicate that disruption of the pld gene significantly attenuates the virulence of A. fumigatus in mice immunosuppressed with hydrocortisone acetate but not with cyclophosphamide.

Fig 6.

Disruption of the pld gene attenuates the virulence of A. fumigatus in hydrocortisone acetate-treated mice but not in cyclophosphamide-treated mice. BALB/c mice were immunosuppressed with either hydrocortisone acetate (A, B, C) or cyclophosphamide (D, E, F) and inoculated intranasally with 30 μl PBS-0.01% Tween 20 (control) or 30 μl PBS-0.01% Tween 20 containing 5 × 10⁶ conidia of the A. fumigatus B5233 wild-type (WT) strain, Δpld strain, and pldC strain. The Kaplan-Meier survival analysis with log tank test was used to compare the survival rates between the groups (A and D). The data shown here were obtained from two independent experiments. Mice immunosuppressed with either hydrocortisone acetate (B and C) or cyclophosphamide (E and F) were inoculated intranasally with 30 μl PBS-0.01% Tween 20 (control) or 30 μl PBS-0.01% Tween 20 containing 5 × 10⁵ conidia of the A. fumigatus B5233 wild-type (WT), Δpld, or pldC strain. Mice were sacrificed at 72 h postinfection to measure the fungal burden in lung lavage fluid (B and E) and lung tissue (C and F). The data shown here are the means ± SE (n = 3 or 4 independent experiments performed in triplicate). Depicted are the differences in the percentages of total inoculated conidia between wild-type and pld mutant strains. *, P < 0.05.

Fig 7.

Disruption of the pld gene attenuates the size and number of lesions in H&E-stained sections from hydrocortisone acetate-treated mice but not in cyclophosphamide-treated mice. (A, B) Mice immunosuppressed with either hydrocortisone acetate (A) or cyclophosphamide (B) were inoculated intranasally with 30 μl PBS-0.01% Tween 20 (control) or 30 μl PBS-0.01% Tween 20 containing 5 × 10⁵ conidia of the A. fumigatus B5233 wild-type (WT), Δpld, or pldC strain. At 72 h postinfection, the lung tissue sections from mice were dissected, fixed in 10% (vol/vol) formaldehyde, stained with H&E, and observed under light microscopy (Olympus BX51). Black arrows indicate the hyphae (grayish-white round spots indicated by black arrows are the cross sections of hyphae). The ×1,000-magnified images represent the area in black outline shown in the ×200-magnified pictures. (C and D) The areas of the sites that contained fungal hyphae with inflammatory cell infiltration were determined by use of Image-Pro Express 6.0. (E and F) The number of lesions per lung were observed and counted under light microscopy. The data shown here are the means ± SE (n = 3 or 4 independent experiments performed in triplicate). Depicted are the differences in the sizes and numbers of lesions between the lungs infected by the A. fumigatus wild-type and pld gene mutant strains. *, P < 0.05.

DISCUSSION

Targeted gene disruption mediated by homologous recombination is a powerful tool that has been used to investigate the function of many genes and molecules in A. fumigatus virulence. However, none of the genes encoding extracellular phospholipases from A. fumigatus had previously been disrupted and evaluated for its role in virulence (1). In the present study, we successfully constructed a pld gene-deleted mutant of A. fumigatus B5233 using A. tumefaciens-mediated transformation (ATMT). In our experiments, approximately 75% of the obtained transformants exhibited pld gene disruption with mitotic stability. Although this percentage was not as high as that found in a previous study (62), the efficiency was enough to allow production and isolation of the pld-deficient and -complemented strains of A. fumigatus. Furthermore, the mechanism of PLD secretion is still poorly understood; however, the detection of PLD activity in the A. fumigatus culture medium and the obvious decrease of extracellular PLD activity of the Δpld strain compared to that of the wild-type strain are consistent with previous indications that A. fumigatus is able to secrete PLD (6).

In the present study, disruption of the pld gene in A. fumigatus had no influence on conidial morphology, hyphal growth rate, and biofilm formation, but it obviously attenuated virulence in hydrocortisone-treated mice. These findings are similar to those from the functional analysis of other genes of A. fumigatus, such as laeA (63), sho1 (43), mpkA (67), and hdaA (40). In mammalian cells, PLD interacts closely with the actin cytoskeleton to regulate cell shape changes. In S. cerevisiae, Spo14, a phosphatidylcholine-specific PLD, was found to be essential for sporulation (55), while C. albicans PLD1 was found to be required for dimorphic transition (45). However, further phylogenetic analysis showed that A. fumigatus PLD has low homology with Spo14 and C. albicans PLD1 (31). A. fumigatus PLD lacks two critical regulatory domains, the PX (Phox homology) and PH (pleckstrin homology) domains, at its NH₂ terminus, which is significantly different from A. fumigatus PLD1, C. albicans PLD1, Spo14, and mammalian PLDs. Therefore, it is tempting to speculate that A. fumigatus PLD may not be able to interact with phosphoinositides, especially phosphatidylinositol-4,5-biphosphate (PIP₂) and phosphatidylinositol-3,4,5-trisphosphate (PIP₃), which are known cytoskeleton modulator phosphoinositides (13, 56). From our findings, it could be deduced that A. fumigatus PLD, as one of three PLD isoforms in A. fumigatus, may be not the isoform that is predominantly involved in morphological changes and mycelial growth. Disruption of other isoforms of PLD in A. fumigatus, such as A. fumigatus PLD1, is being pursued in future studies to test this hypothesis.

The internalization of many infectious particles (e.g., Listeria monocytogenes, Candida albicans, and Cryptococcus neoformans) and nanoparticles into nonphagocytic cells is a deliberate interaction process between particles and its host (20). It is known that A. fumigatus conidia must cross the anatomical barriers of respiratory epithelial cells after inhalation to traverse tissues and cause invasive disease (47). However, little is known about the role of A. fumigatus PLD in this process. Compared to the wild-type strain, the Δpld strain exhibited a decrease in uptake by A549 cells, whereas in the pldC strain, this ability was fully restored, indicating that the activity of A. fumigatus PLD may be required to facilitate internalization of the conidia. However, the adherence of A. fumigatus to A549 cells was not affected by pld gene disruption, suggesting that A. fumigatus PLD probably plays only a minor role in this first interaction step. This observation was consistent with previous findings showing that several laminin- or galactofuranose-associated proteins may mediate the adhesion between A. fumigatus and epithelial cells (37). In contrast to internalization into epithelia, the phagocytosis of A. fumigatus by macrophages was not altered by either pld gene disruption or pretreatment with PLD inhibitors in A. fumigatus. In the phagocytosis assay, we choose to quantify the germinated conidia in our experiments at 1 h postinfection, because it takes approximately 2 h for macrophages to kill A. fumigatus conidia via phagolysosome acidification (28). Although PLD activity has been closely associated with the intracellular survival of pathogens (44), the survival of A. fumigatus conidia in macrophages at 4 h postinfection was also not altered by pld gene disruption (data not shown). These results indicate that A. fumigatus PLD does not contribute to virulence by promoting phagocytosis by host macrophages and subsequent survival in the macrophages. As fungal phospholipases may regulate various host immune signals, such as inducing host cell-associated phospholipase A2 activation in C. albicans (65), the involvement of A. fumigatus PLD in other host immune responses of leukocytes, such as interleukin release or reactive oxygen species (ROS) responses, can currently not be excluded.

Interestingly, A. fumigatus PLD was involved in the virulence of A. fumigatus in hydrocortisone-immunosuppressed mice but not cyclophosphamide-immunosuppressed mice. This in vivo effect of pld gene disruption on virulence was quite similar to those of other two important factors of A. fumigatus, gliotoxin (36) and DvrA (16), which had also shown significant interference in the survival of hydrocortisone-immunosuppressed mice but not cyclophosphamide-treated mice. These findings suggest that, like gliotoxin, the effect of PLD on A. fumigatus virulence also may be related to neutrophils, which would render PLD to be less important for A. fumigatus virulence in neutropenic mice that lack this interaction. Further investigations are needed to clarify the exact role of PLD in the interaction of A. fumigatus with lung epithelia or neutrophils in vivo.

In the present study, chemical inhibitors of mammalian PLD decreased the internalization of wild-type strain conidia, while the internalization of Δpld strain conidia was rescued by exogenous PA at the concentration of 100 μM. These results hinted that PLD may be rather evolutionarily conserved in many eukaryotic cells, at least in its HKD motif (18); meanwhile, the PLD enzymatic product PA may be a critical modulator of A. fumigatus internalization into lung epithelia. As an important second messenger in the cell, PA is a fusogenic lipid and can induce negative curvatures to promote membrane fission (33, 57, 69). Although the distribution of PA in the reaction system is still poorly understood, it is usually added in the concentrations of 50, 100, and/or 200 μM to restore the cellular downregulation of PLD activity for in vitro studies (8, 34, 35). It is rather difficult to decipher here the rationale that both the maximum of A. fumigatus internalization and rescue of the Δpld strain internalization occurred with PA at the concentration of 100 μM; however, this phenomenon was similar to our recent findings that too high of a PLD activity (more localized PA) may inhibit the internalization of a typical intracellular pathogen, Listeria monocytogenes, into epithelial cells (23). These results might be explained as follows. The higher concentration of PA could induce host stress fiber formation, so as to hinder the internalization (4, 30, 32, 49). Likewise, localized PA concentration, which rises initially and then descend, correlates deliberately with the phagocytic process during phagocytosis in macrophage. Thus, an increased phosphatidic acid concentration may disturb this relationship and ultimately suppress A. fumigatus internalization (7, 48). Interestingly, our recent data indicated that beta-1,3-glucan on the surface of germinated conidia is able to stimulate host cell PLD activity and that this activation is important for the efficient internalization of A. fumigatus into A549 lung epithelial cells (24). It could be deduced that local PA production driven by PLD from A. fumigatus itself may also contribute to the internalization. Nevertheless, the exact role of PLD and PA in A. fumigatus internalization need to be further explored.

In summary, we demonstrated for the first time that PLD of A. fumigatus modulates the internalization of A. fumigatus into epithelial cells and may be a virulence factor of A. fumigatus in invasive aspergillosis under immunosuppression by corticosteroids.