Hidden Allies: Decoding the Core Endohyphal Bacteriome of Aspergillus fumigatus

Abstract

Bacterial-fungal interactions that influence the behaviour of one or both organisms are common in nature. Well-studied systems include endosymbiotic relationships that range from transient to long-term associations. Diverse endohyphal bacteria associate with fungal hosts, emphasising the need to better comprehend the fungal bacteriome. We evaluated the hypothesis that Aspergillus fumigatus harbours an endohyphal community of bacteria that influence the host phenotype. We analysed whether 38 A. fumigatus strains show stable association with diverse endohyphal bacteria; all derived from single-conidium cultures that were subjected to antibiotic and heat treatments. The fungal bacteriome, inferred through analysis of bacterial diversity within the fungal strains (short- and long- read sequencing methods), revealed the presence of core endohyphal bacterial genera. Microscopic analysis further confirmed the presence of endohyphal bacteria. The fungal strains exhibited high genetic diversity and phenotypic heterogeneity in drug susceptibility and in vivo virulence. No correlations were observed between genomic or functional traits and bacteriome diversity, but the abundance of some bacterial genera correlated with fungal virulence or posaconazole susceptibility. The observed endobacteriome may play functional roles, for example, nitrogen fixation. Our study emphasises the existence of complex interactions between fungi and endohyphal bacteria, possibly impacting the phenotype of the fungal host, including virulence.

Keywords: Aspergillus fumigatus; clinical isolates; endohyphal bacteria; endosymbionts; fungal bacteriome; fungal virulence.

FIGURE 1

Bacteriome of three Aspergillus fumigatus strains grown with and without antibiotic stress in cultures derived from multiple conidia or single‐conidium. (A) Maximum likelihood midpoint rooted tree of the 100 most abundant bacterial ASVs across the sample set. The maximum likelihood tree was constructed using the general time reversible model with the rate variation among sites described by a gamma distribution and the proportion of invariable sites (GTR + G + I model). Background colours indicate bacterial ASVs assigned at class level. Tree constriction was based on the hypervariable V4 region of 16S rRNA gene sequences, applying 1000 bootstrap replications to estimate confidence. Bootstrap values are indicated above or below the branches. The scale bar indicates nucleotide substitutions per site. Heatmap shows the relative abundances of bacterial ASVs found in tested fungal clinical isolates without (no) and with (yes) antibiotic treatment. The colour intensity shows the ASV percentage in each sample (note that in the colour key the dark blue corresponds to 5%). (B) Heat‐map diagram bacteriome composition at class level of the 3 single spore colonies (marked as C1, C2 or C3) of the three tested fungal clinical isolates. (C) PCoA plot of beta diversity of the single spore cultures based on weighted and unweighted Unifrac distances.

FIGURE 2

Microsatellite genotyping and phenotypic heterogeneity of 38 Aspergillus fumigatus strains assessed in terms of their drug‐resistance and virulence profiles. Hierarchical cluster dendrogram of microsatellite genotypes of A. fumigatus isolates constructed based on the Gower dissimilarity index, with the laboratory model strain Af293 and the soil isolate (Af_SI.00) for comparison. Dots at the end of the dendrogram indicate isolates obtained from the same patient (colour‐coded accordingly). The source of isolation and whether it originated from a cystic fibrosis patient are indicated below the dendrogram (detailed in Table 1). Antifungal susceptibility profiles, assessed via the EUCAST method, and in vivo infection capacity using Galleria mellonella as the infection model (96 h) are represented in the heatmap (Table S1).

FIGURE 3

The core endohyphal bacteriome of the 38 Aspergillus fumigatus strains profiled through sequencing of the V3‐V4 hypervariable region of the 16S rRNA gene, and the derived functional annotation of prokaryotic taxa. (A) Stacked bar chart showing the relative abundance of ASVs from the bacterial V3‐V4 hypervariable region of 16S rRNA sequences, taxonomically classified at the genus level. Low abundance taxa were removed from the visualisation. The order of bacterial genus in the legend is according to its position in the chart. (B) Heat map of the core bacteria at the genus level across the sample set (n = 38), based upon 75% prevalence with at least a 0.1% detection threshold. The y‐axis represents the detection thresholds (indicated as relative abundance), colour shading indicates the prevalence of each bacterial genus among samples for each abundance threshold. (C) The annotation of prokaryotic taxa (FAPROTAX) predicted from the genetic pool of the core endobacteria (75% prevalence with at least 0.1% detection threshold) using the relative abundance ASVs from the bacterial V3‐V4 hypervariable region of 16S rRNA sequences, showing potential functional roles, mostly in categories such as chemoheterotrophy and nitrogen fixation, followed by human pathogens/associated and animal parasites or symbionts.

FIGURE 4

The core endohyphal bacteriome of 9 Aspergillus fumigatus strains profiled through long‐read sequencing of 16S rRNA amplicons. (A) Maximum likelihood tree of V1‐V9 16S rRNA gene bacterial ASVs identified at genus level, with the respective relative abundance (log₁₀) in each A. fumigatus isolate. The genus order in the legend is the same as in the tree, counterclockwise. (B) Heat map of the core bacteria at genus level across the sample set (n = 9), based upon 75% prevalence with at least 0.1% detection threshold. The y‐axis represents the detection thresholds (indicated as relative abundance), colour shading indicates the prevalence of each bacterial genus among samples for each abundance threshold. (C) Venn diagram showing the number of ASVs at genus level found in common in both long length (V1‐V9) and short length (V3‐V4) analysis.

FIGURE 5

Visualisation of endohyphal bacteria in Aspergillus fumigatus strains by microscopy. (A) Fluorescent microscopy, where mycelia were stained with calcofluor‐white (blue) and bacterial DNA with syto9 (green) (Scale bar = 10 μm). Representatives photographs were selected, from left to right: Af_SI.00; Af_CI.01; Af_CI.02; Af_CI.03; Af_CI.12 and Af_CI.18. (B) Fluorescent microscopy, where mycelia were stained with calcofluor‐white (blue), bacterial DNA with syto9 (green) and fungal nuclei with Hoechst (blue) (Scale bar = 10 μm). A representative example is shown: Af_CI.12. The fungal nuclei labelling (↓) enables a clear distinction between fungal and bacterial DNA. (C) Transmission electron micrograph of fungal mycelium of A. fumigatus strain Af_CI.06 containing endohyphal bacteria (B), nucleus (N) and mitochondria (M). (D) Fluorescent in situ hybridization depicting clusters of spherical bacteria (cyan, marked with an arrow) along the hyphae of A. fumigatus (magenta). Fungal rRNA were labelled with the universal eukaryotic 18S rRNA probe; bacteria were co‐stained with a universal 16S rRNA probe and DAPI was used as a global nuclear staining (yellow) (Scale bar = 10 μm).

FIGURE 6

Identifying endohyphal bacteria potentially impacting the virulence and drug‐resistance profiles of Aspergillus fumigatus strains. (A) Hierarchical clustering heat map of fungal bacteriome using Bray–Curtis distance. Samples were clustered with maximum of 0.15 dissimilarity. (B) Spearman's correlation, asterisks indicate significant correlations *p ≤ 0.05; **p ≤ 0.01; ***p ≤ 0.001. Based on region V3–V4 of the 16S rRNA gene.