Deacetylation of Fungal Exopolysaccharide Mediates Adhesion and Biofilm Formation

Abstract

The mold Aspergillus fumigatus causes invasive infection in immunocompromised patients. Recently, galactosaminogalactan (GAG), an exopolysaccharide composed of galactose and N-acetylgalactosamine (GalNAc), was identified as a virulence factor required for biofilm formation. The molecular mechanisms underlying GAG biosynthesis and GAG-mediated biofilm formation were unknown. We identified a cluster of five coregulated genes that were dysregulated in GAG-deficient mutants and whose gene products share functional similarity with proteins that mediate the synthesis of the bacterial biofilm exopolysaccharide poly-(β1-6)-N-acetyl-D-glucosamine (PNAG). Bioinformatic analyses suggested that the GAG cluster gene agd3 encodes a protein containing a deacetylase domain. Because deacetylation of N-acetylglucosamine residues is critical for the function of PNAG, we investigated the role of GAG deacetylation in fungal biofilm formation. Agd3 was found to mediate deacetylation of GalNAc residues within GAG and render the polysaccharide polycationic. As with PNAG, deacetylation is required for the adherence of GAG to hyphae and for biofilm formation. Growth of the Δagd3 mutant in the presence of culture supernatants of the GAG-deficient Δuge3 mutant rescued the biofilm defect of the Δagd3 mutant and restored the adhesive properties of GAG, suggesting that deacetylation is an extracellular process. The GAG biosynthetic gene cluster is present in the genomes of members of the Pezizomycotina subphylum of the Ascomycota including a number of plant-pathogenic fungi and a single basidiomycete species,Trichosporon asahii, likely a result of recent horizontal gene transfer. The current study demonstrates that the production of cationic, deacetylated exopolysaccharides is a strategy used by both fungi and bacteria for biofilm formation.

Importance: This study sheds light on the biosynthetic pathways governing the synthesis of galactosaminogalactan (GAG), which plays a key role in A. fumigatus virulence and biofilm formation. We find that bacteria and fungi use similar strategies to synthesize adhesive biofilm exopolysaccharides. The presence of orthologs of the GAG biosynthetic gene clusters in multiple fungi suggests that this exopolysaccharide may also be important in the virulence of other fungal pathogens. Further, these studies establish a molecular mechanism of adhesion in which GAG interacts via charge-charge interactions to bind to both fungal hyphae and other substrates. Finally, the importance of deacetylation in the synthesis of functional GAG and the extracellular localization of this process suggest that inhibition of deacetylation may be an attractive target for the development of novel antifungal therapies.

FIG 1

Bacterial polysaccharide biosynthetic operons and the putative GAG biosynthetic gene cluster. (A to C) Schematic of the ica operon responsible for the synthesis of polysaccharide intercellular adhesion (PIA) (A), pga operon responsible for the synthesis of PNAG [poly-(β1-6)-N-acetyl-d-glucosamine] (B), and the putative GAG gene cluster (C). (D) Heatmap showing differential gene expression of the A. fumigatus ΔmedA and ΔstuA regulatory mutants compared to wild-type A. fumigatus, highlighting the coregulation of the genes in the GAG biosynthesis gene cluster. Fold induction is shown in red (upregulation), green (downregulation), black (no change), and gray (missing data point). Locus ID, locus identification.

FIG 2

Comparative models and in silico analysis. (A) Comparative models of exopolysaccharide synthesis in bacteria and fungi. The numbered steps are as follows: (step 1) polymerization of sugar residues by the glycosyltransferases indicated in green (Gtb3, IcaA, and PgaC), (step 2) extrusion of the elongating polysaccharide from the cytosolic side to the extracellular side (or periplasm in Gram-negative bacteria) by the dual action of the glycosyltransferase (Gtb3, IcaA, and PgaC) and associated protein for the bacterial species (IcaD and PgaD), and (step 3) deacetylation of the N-acetylhexosamine unit of the nascent polysaccharide by the de-N-acetylase indicated in dark red (Agd3, IcaB, and PgaB). The extracellular matrix (ECM), cell wall (CW), plasma membrane (M), peptidoglycan (PG), outer membrane (OM), and inner membrane (IM) are shown. CoA, coenzyme A. (B) Predicted domains and conserved regions in the Agd3 protein. From the N terminus, these domains include signal peptide (SS), a serine-rich region, a glutamine amidotransferase domain (reductase), metal-coordinating linear motifs (MT1/MT2), β/α barrel, a carbohydrate esterase-4 like domain (CE 4), and a β-strand-rich region. (C) Multiple-sequence alignment showing conserved DXD/DD motif located within the MT1 and MT2 conserved sites. Sequences include S. epidermidis IcaB, E. coli PgaB, P. aeruginosa PelA, and Sinorhizobium meliloti NodB. Highly conserved and similar residues are highlighted in dark and light gray, respectively. Gaps introduced to maximize alignment are indicated by hyphens.

FIG 3

Deletion of agd3 blocks deacetylation of GAG. (A) Detection of primary amine of purified GAG from the strains indicated in the figure as measured by evolution of colorimetric byproduct of the TNBS reaction. O.D., optical density. (B) ¹H NMR analysis of purified GAG from the indicated strains. Black arrows indicate the detection of the hydrogen resonance peak originating from galactosamine. (C) Total secreted GAG production by the indicated strains. (D) Comparison of relative expression of GAG cluster genes between wild-type Af293 and Δagd3 mutant as measured against reference gene tef1 under various growth conditions (RPMI 1640 supplemented with MOPS [RPMI-MOPS], Brian medium, Aspergillus minimal medium [AspMM]). For anaerobic growth, additional reference genes, actin 1, and β-tubulin were also used. For all graphs, data are represented as means plus standard errors of the means (SEM) (error bars). The values of the wild-type A. fumigatus Af293 and the Δagd3 mutant strain were significantly different (P < 0.05 by analysis of variance [ANOVA] with Tukey’s test for pairwise comparison) as indicated by the asterisk.

FIG 4

Deletion of agd3 is associated with loss of adherence and changes in the cell wall. (A) Formation of adherent biofilms by the strains indicated in the figure on either positively charged (poly-d-lysine-treated [PDL]) or negatively charged (tissue culture-treated polystyrene [TC]) surfaces. Biofilms were washed and visualized by staining with crystal violet (gray). (B) Confocal microscopy images of hyphae stained with FITC-tagged soybean agglutinin lectin (top) and corresponding differential interference contrast (DIC) (bottom). (C) Scanning electron micrographs of hyphae grown for 24 h. The white arrows point to hyphal surface decorations associated with GAG production. (D) Confocal microscopy images of hyphae stained with Fc-dectin-1 detected by FITC-tagged Fc-receptor antibody (top) and corresponding DIC (bottom).

FIG 5

Agd3 activity augments the positive charge on the surfaces of hyphae. The graph shows the percentage of negatively charged Sephadex beads bound by hyphae of the indicated strains. Data are presented as means plus standard errors of the means (SEM) (error bars). The values for the indicated mutant strains were significantly different (P < 0.05 by ANOVA with Tukey’s test for pairwise comparison) from the value for wild-type A. fumigatus Af293 strain as indicated by the asterisk. The value for the Δagd3::agd3 strain was not significantly different (n.s.) from the value for the wild-type A. fumigatus Af293 strain.

FIG 6

Agd3 is required for full virulence in a mouse model of invasive aspergillosis. (A) Survival of BALB/c mice treated with cortisone and cyclophosphamide and then infected with the indicated conidial strains. Graphs are the combined results of two independent experiments with 26 mice per group for all groups of mice infected with fungal strains and 24 mice in the PBS sham infection group. There was a significant difference in the survival of mice infected with wild-type Af293 or Δagd3::agd3 strain compared to those infected with the Δagd3 strain as determined by the Mantel-Cox log rank test with pairwise comparison applying Bonferroni’s correction as indicated by the asterisk. (B) Pulmonary histopathology sections from BALB/c mice infected with indicated strains and stained with PAS for visualization of fungal elements. Black arrows indicate fungal elements found within pulmonary lesions. Bars, 100 µm. (C) Pulmonary fungal burden of mice infected with the indicated strains, as measured by quantitative PCR. There were 8 mice in each group. (D) Pulmonary fungal burden of mice infected with the indicated strains, as measured by determination of pulmonary galactomannan content. There were 8 mice in each group. (E) Pulmonary injury as measured by lactose dehydrogenase activity in the bronchoalveolar lavage fluid of mice infected with the indicated strains. There were 8 mice in each group. Values are medians plus interquartile ranged (error bars). There was a significant difference in either the fungal burden or lung injury in mice infected with wild-type Af293 strain and those infected with the Δagd3 mutant strain as determined by the Mann-Whitney test as indicated by the § symbol.

FIG 7

Culture filtrates from the Δuge3 mutant complement the defects in adherence and cell wall morphology of Δagd3 mutant. (A) Biofilm formation by the indicated strains grown alone or in coculture. After the biofilm was washed, adherent biofilm was visualized by crystal violet staining. (B) Biofilm formation by the indicated strains grown in the presence of culture filtrates (C/F) from the Δuge3 or Δagd3 mutant. After the biofilm was washed, adherent biofilm was visualized by crystal violet staining. (C) Scanning electron microscopy visualization of cell wall morphology of the indicated strains grown in the presence of culture filtrates from the Δuge3 mutant. The white arrows indicate cell wall decorations associated with cell wall-bound GAG.

FIG 8

Agd3 localizes to the surfaces of hyphae. Hyphae expressing an Agd3-RFP fusion protein are visualized by confocal microscopy and indirect immunofluorescence.

FIG 9

Agd3 orthologs are found within a wide range of Ascomyces species. A gene tree is shown categorized by taxonomic class of Agd3 orthologs found in the 28 fungal species that possess the GAG gene cluster. The gene tree was built by aligning and trimming Agd3 ortholog sequences, followed by maximum likelihood phylogenetic analysis with 100 bootstrap replicates. The outgroup was rooted with sequences of Arthrobotrys species. The scale bar is the genetic distance representing amino acid substitutions per site.

FIG 10

Trichosporon asahii produces a GAG-like exopolysaccharide. Hyphae of T. asahii were stained with FITC-tagged soybean agglutinin (SBA) lectin binding for the detection of GalNAc-rich exopolysaccharide.