Inducible Cell Fusion Permits Use of Competitive Fitness Profiling in the Human Pathogenic Fungus Aspergillus fumigatus

Abstract

Antifungal agents directed against novel therapeutic targets are required for treating invasive, chronic, and allergic Aspergillus infections. Competitive fitness profiling technologies have been used in a number of bacterial and yeast systems to identify druggable targets; however, the development of similar systems in filamentous fungi is complicated by the fact that they undergo cell fusion and heterokaryosis. Here, we demonstrate that cell fusion in Aspergillus fumigatus under standard culture conditions is not predominately constitutive, as with most ascomycetes, but can be induced by a range of extracellular stressors. Using this knowledge, we have developed a barcode-free genetic profiling system that permits high-throughput parallel determination of strain fitness in a collection of diploid A. fumigatus mutants. We show that heterozygous cyp51A and arf2 null mutants have reduced fitness in the presence of itraconazole and brefeldin A, respectively, and a heterozygous atp17 null mutant is resistant to brefeldin A.

Keywords: Aspergillus fumigatus; anastomosis; antifungal; chemical genomics; functional genomics; hyphal fusion; imaging.

FIG 1

Cell fusion was not observed in strain AF293 on solid culture, or in prototrophic fluorescently tagged strains under standard liquid culture conditions. (A) Low-temperature scanning electron micrograph of germinating AF273 conidia grown on solid oatmeal medium after 12 h of incubation at 37°C showing an absence of cell fusion between neighboring germlings. Bar, 10 µm. (B) Static liquid coculture of TurboFP635- and GFP-expressing strains (MFIGRag29 and MFIGGFP4) in AMMNO₃ liquid medium. Tiled confocal images demonstrate the lack of colocalization of GFP and TurboFP635 fluorescence proteins after 22 h. Panels show the TurboFP635, GFP, and brightfield images plus a merged image of both fluorescence channels showing no colocalization. Bar, 100 µm.

FIG 2

Cell fusion between conidia involving short thin cell protrusions that resemble conidial anastomosis tubes. Cell fusion has occurred in a static liquid coculture of nitrogen assimilation mutants expressing TurboFP635 and GFP (AfTurb635FPniaD and AfGFPcnx, respectively) after 24 h of incubation in AMMNO₃ liquid medium. Cell fusion is very rare (0.2%) under these conditions. (A) TurboFP635, GFP, and brightfield images plus a merged image of both fluorescence channels that shows very little colocalization. Fused cells expressing both GFP and TurboFP635 are white and nonfused cells are either magenta or green; a rare fusion event is highlighted by a yellow box. Bar, 100 µm. (B) Zoomed-in images of a direct fusion event (arrow in brightfield panel) which resembles a fusion between two very short CATs/one very short CAT and a conidium; the putative CAT(s) is much narrower in width than germ tubes. Bar, 10 μm.

FIG 3

Cell fusion frequency is increased under nutritional and antimicrobial stresses under solid culture conditions. Representative images of TurboFP635- and GFP-expressing A. fumigatus cells grown on various solid media for 48 h and visualized with live-cell confocal imaging using the inverted agar technique (26). Colocalized pixels are shown in white in the panels on the right. The cells containing white colocalized pixels are the result of fusion between GFP- and TurboFP635-expressing cells. The white colocalization is commonly observed in conidiophores, from which chains of spores arise, and is as a result of fusion of underlying hyphae from which the conidiophores are derived. The magenta or green cells remain unfused. (A) Nitrogen assimilation mutant strains (AfTurboFP635niaD and AfGFPcnx) that have undergone nitrogen starvation on AMMNO₃ solid medium. (B) Prototrophic strains (MFIGRag29 and MFIGGFP4) grown on solid AMMNO₃ medium containing 25 µg/ml cerulenin. (C) Prototrophic strains (MFIGRag29 and MFIGGFP4) grown on solid AMMNO₃ medium containing 0.5 µg/ml itraconazole. (D) Prototrophic strains (MFIGRag29 and MFIGGFP4) grown on solid AMMNO₃ medium containing 6.25 µg/ml brefeldin A. Bars, 100 μm. (E) Quantitative analyses of cell fusion under various conditions from images exhibiting GFP and TurboFP635 colocalization. Each data point (open circles) represents a biological replicate. Bars, SEMs. P values are displayed above each condition and are all relative to “No stress” treatment, obtained by means of a nonpaired two-tailed Student’s t test.

FIG 4

Schematic representation of the chemical genomics methodology employed in this study. A library of knockout mutants was generated in the A. fumigatus diploid isolate AFMB3. The library was pooled and grown in the presence and absence of an antifungal agent or inhibitory drug. Genomic DNA was extracted from the pooled library and sheared by sonication. Illumina asymmetric linkers were ligated to the sheared DNA. The KO-cassette (pyrG) 3′ target flanks were enriched by PCR amplification. The primers used for amplification (green arrows) incorporate the Illumina sequencing primer site and anneal to the selectable marker and the noncomplementary region of the asymmetric linker. After amplification, sequencing reads were mapped to a reference library that included the 3′ flank regions of all 46 mutants in the library. Relative counts per flank were used to establish the fitness of each strain.

FIG 5

Chemical genomics to determine antifungal mechanism of action. (A) Dot plot comparing the normalized read counts for each strain (n = 44) in the mutant pool. The x and y axes represent data from counts of two independent replicate cultures exposed to brefeldin A. (B) Box plot showing the relative fitness of each strain in the mutant pool when exposed to itraconazole and brefeldin A. Whiskers represent 25th and 75th percentile plus 1.5× interquartile range (IQR). Dots represent strains falling outside this range. (C) Real-time PCR quantitative analysis of the DNA analyzed by en mass sequencing for brefeldin A. All data were normalized to the Δmot1 mutant (100%) (error bars are 95% confidence intervals [CIs]). (D) Real-time PCR quantitative analysis of the DNA analyzed by en mass sequencing for itraconazole. All data were normalized to the Δmot1 mutant (100%) (error bars are 95% CIs). For panels C and D, P values were calculated using pairwise reallocation randomization test (52).

FIG 6

Phenotypic evaluation of strains with altered fitness upon drug exposure. Resistant and sensitive isolates identified from the competitive fitness study were grown on RPMI agar with increasing concentrations of brefeldin A (A) and itraconazole (B) for 72 h. A strain for which fitness correlates with that of the rest of the pool (Δmot1 mutant) is shown for comparison.