The Aspergillus fumigatus siderophore biosynthetic gene sidA, encoding L-ornithine N5-oxygenase, is required for virulence

Abstract

Aspergillus fumigatus is the leading cause of invasive mold infection and is a serious problem in immunocompromised populations worldwide. We have previously shown that survival of A. fumigatus in serum may be related to secretion of siderophores. In this study, we identified and characterized the sidA gene of A. fumigatus, which encodes l-ornithine N(5)-oxygenase, the first committed step in hydroxamate siderophore biosynthesis. A. fumigatus sidA codes for a protein of 501 amino acids with significant homology to other fungal l-ornithine N(5)-oxygenases. A stable DeltasidA strain was created by deletion of A. fumigatus sidA. This strain was unable to synthesize the siderophores N',N",N'''-triacetylfusarinine C (TAF) and ferricrocin. Growth of the DeltasidA strain was the same as that of the wild type in rich media; however, the DeltasidA strain was unable to grow in low-iron defined media or media containing 10% human serum unless supplemented with TAF or ferricrocin. No significant differences in ferric reduction activities were observed between the parental strain and the DeltasidA strain, indicating that blocking siderophore secretion did not result in upregulation of this pathway. Unlike the parental strain, the DeltasidA strain was unable to remove iron from human transferrin. A rescued strain (DeltasidA + sidA) was constructed; it produced siderophores and had the same growth as the wild type on iron-limited media. Unlike the wild-type and rescued strains, the DeltasidA strain was avirulent in a mouse model of invasive aspergillosis, indicating that sidA is necessary for A. fumigatus virulence.

FIG. 1.

Chemical structures of the A. fumigatus siderophores N′,N",N‴-triacetylfusarinine C (A) and ferricrocin (B).

FIG. 2.

Alignment of the A. fumigatus SidA, A. nidulans SidA, and A. oryzae DffA amino acid sequences. The amino acid sequence of A. fumigatus SidA was predicted using GlimmerM from The Institute for Genomic Research, trained for A. fumigatus. Multiple pairwise alignment was performed with ClustalW (8) and the output generated with Boxshade 3.21. Black and gray boxes represent identical and similar residues, respectively.

FIG. 3.

Restriction map and Southern blot of wild-type and ΔsidA genomic DNAs. (A) Double-crossover gene disruption, showing binding sites for primers and restriction sites for both A. fumigatus wild-type and ΔsidA strains. (B) The Southern blot confirms the disruption of sidA. A. fumigatus genomic DNAs from the wild-type and ΔsidA strains were completely digested with EcoRV, HindIII, or PstI; separated by gel electrophoresis; and transferred to Hybond N+. Blots were probed using a full-length sidA probe constructed using the AlkPhos direct labeling kit. Detection was with the CDP-star reagent. I, wild-type genomic DNA; II, ΔsidA genomic DNA. Numbers on the left represent molecular size markers in kilodaltons.

FIG. 4.

Thin-layer chromatography of siderophores produced by wild-type and ΔsidA strains of A. fumigatus. Siderophores were extracted from wild-type and ΔsidA cultures grown in 5-ml volumes of modified GA medium at 37°C for 3 days as described in Materials and Methods. Extracts were analyzed on silica gel thin-layer chromatography sheets with a mobile phase of 4:1 dichloromethane-methanol, and the locations of authentic TAF and ferricrocin standards run at the same time are noted. Ferricrocin, TAF, and two unidentified siderophores are visible in the wild-type extract (A), while no siderophores are visible in extract from the ΔsidA strain (B). Faint yellow spots in the ΔsidA extracts were also present at the same R_f value in the uninoculated control (data not shown).

FIG. 5.

Iron saturation of human diferric transferrin incubated with wild-type or ΔsidA A. fumigatus mycelia. A. fumigatus wild-type and ΔsidA mycelia were incubated in MEM containing 0.2 mg/ml human diferric transferrin at 37°C and 150 rpm. Iron saturation of transferrin was assessed by urea-PAGE. Gels were stained with SYPRO orange and scanned on a Typhoon 9410 imager, and the transferrin bands were quantified using ImageQuant 5.2. Data are normalized and reported as percent total transferrin present in diferric form/percent transferrin in diferric form at time zero. Error bars represent standard deviations of triplicate measurements.

FIG. 6.

Survival curve for female BALB/c mice infected with the A. fumigatus wild-type strain, the ΔsidA strain, or the rescued strain. Mice were immunosuppressed by subcutaneous injection of 200 mg/kg cortisone acetate on days −3, 0, 2, and 4. Mice were infected with 5 × 10⁶ conidia of either the wild-type, ΔsidA^R, or ΔsidA strain in 20 μl saline on day 0. Control mice were given saline alone. Mice were monitored daily and sacrificed if they displayed symptoms of infection, as described in Results. Survival curves for the wild-type or ΔsidA^R strain versus the ΔsidA strain are significantly different (P < 0.0001) by log rank analysis.

FIG. 7.

Lung tissue sections from cortisone-treated mice exposed to one of the following treatments: saline (A and B), wild-type A. fumigatus conidia (C and D), conidia from the rescued strain (ΔsidA^R) (E and F), and ΔsidA A. fumigatus conidia (G and H). Each panel represents a section from a different animal. (A and B [saline]) Lungs have normal appearance, with clear airways and no inflammatory infiltrate. (C and D [wild type])Fungal hyphae (white arrow) within an airway accompanied by a neutrophil and monocyte infiltration; erythrocytes are seen within the airways. (E [rescued strain]) Hyphae are seen within the bronchiole. (F [rescue]) Fungal growth is evident within an airway (arrow), accompanied by necrosis. (G [ΔsidA strain]) Normal lung tissue surrounding a focus of inflammation confined to the lumen of a bronchiole. (H [ΔsidA strain]) No fungi were observed, but foci containing large numbers of macrophages were observed (arrows). Magnifications, ×340 (A, B, D, E, G, and H) and ×170 (C and F).