Pseudomonas aeruginosa and their small diffusible extracellular molecules inhibit Aspergillus fumigatus biofilm formation


  • Editor: Jan Dijksterhuis

Correspondence: Gordon Ramage, Infection and Immunity Research Group, Glasgow Dental School and Hospital, School of Medicine, College of Medicine, Veterinary and Life Sciences, Faculty of Medicine, University of Glasgow, 378 Sauchiehall Street, Glasgow G2 3JZ, UK. Tel.: +44 0 141 211 9752; fax: +44 0 141 331 2798; e-mail:


Aspergillus fumigatus is often isolated from the lungs of cystic fibrosis (CF) patients, but unlike in severely immunocompromised individuals, the mortality rates are low. This suggests that competition from bacteria within the CF lung may be inhibitory. The purpose of this study was to investigate how Pseudomonas aeruginosa influences A. fumigatus conidial germination and biofilm formation. Aspergillus fumigatus biofilm formation was inhibited by direct contact with P. aeruginosa, but had no effect on preformed biofilm. A secreted heat-stable soluble factor was also shown to exhibit biofilm inhibition. Coculture of P. aeruginosa quorum-sensing mutants (PAO1:ΔLasI, PAO1:ΔLasR) did not significantly inhibit A. fumigatus biofilms (52.6–58.8%) to the same extent as that of the PA01 wild type (22.9–30.1%), both by direct and by indirect interaction (P<0.001). Planktonic and sessile inhibition assays with a series of short carbon chain molecules (decanol, decanoic acid and dodecanol) demonstrated that these molecules could both inhibit and disrupt biofilms in a concentration-dependent manner. Overall, this suggests that small diffusible and heat-stable molecules may be responsible for the competitive inhibition of filamentous fungal growth in polymicrobial environments such as the CF lung.


The ubiquitous mould Aspergillus fumigatus is responsible for the majority of human infections caused by Aspergillus spp. The conidia produced by these saprophytic fungi disseminate by aerosolization and are inhaled, finally dwelling in the alveoli of human lungs (Askew, 2008). Aspergillus fumigatus can cause a range of opportunistic infections, ranging from allergic reactions (allergic bronchopulomary aspergillosis) to invasive disease, particularly in immunocompromised individuals, including cystic fibrosis (CF) patients (Skov et al., 2005). Persistent A. fumigatus infection has been shown to be an independent risk factor for hospital admissions in CF patients with reduced respiratory function (Amin et al., 2010).

The pulmonary cavity of CF patients with its thick mucous deposits predisposes to a wide range of opportunistic infections. A diverse microbial ecosystem has been described within the confined space of the lungs of CF patients, which is influenced by both the clinical status and the current antibiotic treatment regimens of the patient (Gilligan, 1991; Valenza et al., 2008). Indeed, a complex mixture of bacterial and fungal pathogens may coexist within the lungs, including Pseudomonas aeruginosa, Staphylococcus aureus, Burkholderia cenocepacia, Candida albicans and A. fumigatus (Valenza et al., 2008). Within this environment, both bacteria and fungi possess the ability to form multicellular biofilm consortia, making it inherently difficult to eradicate infection. In addition, direct physical contact between organisms or indirect molecular signalling interactions may influence microbial pathogenicity, which in turn may influence the disease outcome (Duan et al., 2003). Various studies have confirmed the presence of bacterial quorum-sensing molecules in the sputum of CF patients (Singh et al., 2000; Chambers et al., 2005). As these molecules are known to modulate the pathogenicity of key CF-related pathogens, an investigation of the interactions between these microbial pathogens may provide novel treatment strategies. Our study reports on how direct and indirect interactions of the major prokaryotic CF pathogen P. aeruginosa, and associated molecules, with the eukaryotic pathogen A. fumigatus impact on filamentous growth, leading to biofilm formation.

Materials and methods

Organisms and growth conditions

Aspergillus fumigatus Af293 and four clinical isolates (YHCF1-4) obtained from the Royal Hospital for Sick Children (Yorkhill Cystic Fibrosis Unit, Glasgow) were used throughout this study. Pseudomonas aeruginosa reference strains PAO1, PA14, ATCC 27835, six clinical nonclonal isolates [PA103, PA4384, PA14955, PA15861, PA16190 and PA16191 (gifted by Professor Douglas Storey, University of Calgary Foothills Hospital)] and two mutant strains [PAO1:ΔLasI (unable to synthesize N-acyl homoserine lactones (HSL)) and PAO1:ΔLasR (synthesizes HSL, but cannot respond), gifted by Professor Paul Williams, University of Nottingham] were used in this study. PAO1 mutants were maintained on Luria–Bertani (LB) broth agar plates containing 100 μg mL−1 ampicillin (Sigma-Aldrich, Gillingham, UK) and 20 μg mL−1 gentamicin (Sigma-Aldrich). All working stocks of fungal and bacterial strains were maintained at 4 °C on Sabouraud (Oxoid, Cambridge, UK) agar or LB agar slopes (Oxoid), respectively, and stored in Microbank® vials (Pro-Lab Diagnostics, Cheshire, UK) at −80 °C. For each assay, A. fumigatus was grown on Sabouraud agar and conidia standardized to 1 × 105 mL−1 in 3-(N-morpholino)propanesulphonic acid (MOPS)-buffered RPMI 1640 [pH 7.2 (Sigma-Aldrich)], as described previously by our group (Mowat et al., 2007). Pseudomonas aeruginosa strains were grown overnight in an orbital shaker at 200 r.p.m., standardized to a density equivalent of approximately 1 × 108 CFU mL−1, and diluted to a working concentration of 1 × 106 CFU mL−1.

Direct interaction between P. aeruginosa and A. fumigatus

To examine the direct effect of live P. aeruginosa on A. fumigatus biofilm formation, standardized suspensions of conidia and bacterial cells were combined in equal volumes in a 96-well microtitre plate (Corning, NY) in MOPS-buffered RPMI (Sigma) and incubated overnight at 37 °C. The effect of killed bacterial cells on A. fumigatus biofilm formation was also investigated. Pseudomonas aeruginosa was centrifuged, washed twice in phosphate-buffered saline (PBS) and resuspended in 100% methanol for 2 h. The dead cells were then centrifuged and washed three times in PBS to remove any remaining trace of methanol and finally resuspended to 1 × 106 CFU mL−1 in RPMI. To confirm bacterial killing, aliquots of the bacterial cells were spread onto LB agar plates and incubated overnight at 37 °C. Equal volumes of standardized conidia and methanol-killed bacterial cells were combined in a 96-well microtitre plate and incubated overnight at 37 °C.

Aspergillus fumigatus biofilms were also prepared, as described previously (Mowat et al., 2007), and challenged with P. aeruginosa. The resultant A. fumigatus biomass after exposure of mature biofilms and conidia undergoing morphological differentiation, to both live and dead bacterial cells, were quantified as described previously by our group (Mowat et al., 2007). In addition, scanning electron microscopy (SEM) of A. fumigatus biofilms grown on Thermanox coverslips (Nalge Nunc Inc., Rochester, NY) and challenged with P. aeruginosa (PAO1) for 24 h was examined microscopically, as described previously (Mowat et al., 2007). These were viewed using a Zeiss Evo SEM in high-vacuum mode at 10 kV.

Effect of diffusible endogenous bacterial products on A. fumigatus biofilm formation

A standardized overnight culture of all bacterial strains was centrifuged for 5 min at 3000 g to pellet the cells. The harvested supernatant was then filter sterilized through a 0.22-μM filter (Millipore UK Limited). An aliquot of the supernatant was also heat treated at 80 °C for 10 min. The supernatants were then combined (9 : 1) with 10 × concentrated MOPS-buffered RPMI containing 1 × 105 conidia mL−1, aliquoted into a 96-well microtitre plate and incubated overnight at 37 °C.

To assess the role of an indirect interaction between A. fumigatus and P. aeruginosa, a 12 mm Transwell® (Corning, NY) permeable support system was utilized. The Transwell® system enables the coculturing of the two pathogens in two separate compartments connected via a microporous membrane (0.4 μm). Aspergillus fumigatus conidia were inoculated into the lower compartment and P. aeruginosa were inoculated into the upper chamber of the insert, which was then incubated overnight at 37 °C. The following P. aeruginosa strains were tested using the Transwell® system: PAO1, PAO1:ΔLasI and PAO1:ΔLasR. Wells containing only A. fumigatus or P. aeruginosa were included as controls. Aspergillus fumigatus biofilm growth was assessed using light microscopy and the fungal biomass was quantified as described previously by our group (Mowat et al., 2007).

Effect of small exogenous short-chain carbon molecules on A. fumigatus biofilms

A range of compounds structurally similar to the quorum-sensing molecules produced by P. aeruginosa were tested for their inhibitory properties. These were decanol, decanoic acid, octanoic acid, tetradecanol and dodecanol (Sigma-Aldrich). These were solubilized in ethyl acetate containing 0.01% (v/v) glacial acetic acid (Fisher Scientific, UK) to a 1 M stock concentration. All solutions were stored at −20 °C for a maximum of 1 month. Each compound was diluted to 100 mM in MOPS-buffered RPMI and, using the CLSI broth microdilution M28-A assay (CLSI, 2008), their effect on conidia, biofilm formation and the resultant biomass was evaluated (Mowat et al., 2007). Eight replicates were tested for each compound concentration on three separate occasions with all A. fumigatus strains.

Statistical analysis

For biomass data, an angular transformation was performed and the transformed data were analysed using one-way anova with Bonferroni's multiple comparisons post-test. P<0.05 was considered significant. The analyses were performed using graphpad prism version 4.0 for Windows (GraphPad Software, CA).


Pseudomonas aeruginosa cells and supernatant inhibit A. fumigatus biofilm formation, but have a minimal impact on preformed biofilms

When A. fumigatus conidia were exposed to live P. aeruginosa cells overnight, the resultant fungal biomass was significantly reduced to 14.5% (P<0.001) of the untreated controls. Methanol-treated P. aeruginosa cells also showed this effect (Fig. 1a). Exposure to the P. aeruginosa supernatant resulted in the inhibition of hyphal growth, restricting the biomass to 19.1%. The heat-treated supernatant did not significantly reduce this effect, restricting the biomass to 23.0%. When mature A. fumigatus biofilms were exposed to live P. aeruginosa cells, the fungal biomass was minimally affected (84.8%). SEM analysis revealed individual P. aeruginosa (PA01) cells and microcolonies distributed throughout the intertwined filamentous networks of the mature A. fumigatus biofilms (Fig. 1b). All nine P. aeruginosa isolates examined showed similar effects.

Figure 1.

Pseudomonas aeruginosa inhibition of Aspergillus fumigatus biofilm. (a) Aspergillus fumigatus conidia were exposed to live or dead P. aeruginosa cells, fresh spent bacterial supernatant (SPNT) and heat-treated supernatant (HI SPNT), with the resulting fungal biomass expressed as a proportion of the control. Aspergillus fumigatus preformed biofilms were also exposed to live bacterial cells for 24 h and the resulting biomass was determined. Results are the mean of five A. fumigatus strains exposed to nine P. aeruginosa strains (*P< 0.001). Error bars representing the SEM. (b) SEM evaluation of interaction between P. aeruginosa PAO1 and A. fumigatus biofilms. Bacterial cells adhered throughout the dense intertwined filamentous structure of A. fumigatus. Scale bar represents 2 μm. Arrows indicate examples of P. aeruginosa cells adhering to the hyphae.

Pseudomonas aeruginosa LasIR quorum-sensing network plays a role in the inhibition of A. fumigatus biofilm formation

Aspergillus fumigatus conidia were exposed to live cells from two P. aeruginosa quorum-sensing knockout strains: PAO1:ΔLasI (unable to synthesize HSL) and PAO1:ΔLasR (synthesizes HSL, but cannot respond) (Fig. 2). Aspergillus fumigatus growth was significantly greater (P<0.001) during direct coculture with PAO1:ΔLasI (58.3%) and PAO1:ΔLasR (52.6%) in comparison with the wild-type PAO1 (22.9%). When the Transwell® system was used to determine an indirect effect on A. fumigatus biofilm development, the biomass was restricted to 30.1% of the unchallenged control by the wild-type PAO1. In comparison, the levels of inhibition were significantly less than the wild type (P<0.001) during an indirect coculture with PAO1:ΔLasI (58.8%) and PAO1:ΔLasR (56.8%).

Figure 2.

Pseudomonas aeruginosa LasIR inhibits Aspergillus fumigatus biofilm. Aspergillus fumigatus conidia were cocultured overnight with P. aeruginosa [PAO1 (WT), PAO1:ΔLasI and PAO1:ΔLasR] ± Transwell® system, and the resulting fungal biomass assessed. Results are presented as the mean of two separate experiments with five strains of A. fumigatus exposed to each P. aeruginosa strain (*P < 0.001). Both PAO1:ΔLasI and PAO1:ΔLasR were significantly affected in their capacity to inhibit A. fumigatus biofilms during direct and indirect coculture in comparison with the PA01 WT (P<0.001). Note that the mutant unable to synthesize HSL (LasI) displayed an inhibition profile similar to the mutant able to synthesize it, but not respond (LasR). Interestingly, the levels of inhibition were marginally, but not significantly, reduced.

Aspergillus fumigatus biofilms are inhibited and disrupted by decanol, decanoic acid and dodecanol exposure

All the compounds tested reduced the cellular viability of A. fumigatus conidia in a concentration-dependent manner, with the exception of tetradecanol, which had no notable inhibitory effect after 24 h at any concentration tested compared with untreated controls (Fig. 3a). All four of these inhibitory compounds reduced the biomass by over 80% at the highest concentration (25 mM), with decanol, dodecanol and decanoic acid showing no significant differences between their concentration-dependent inhibitory profiles across the range tested. Biomass inhibition by octanoic acid was not observed until ≥1.6 mM.

Figure 3.

 Small carbon chain molecules inhibit and disperse Aspergillus fumigatus biofilms. Standardized A. fumigatus conidia were exposed to serial doubling dilutions of decanol, decanoic acid, octanoic acid, tetradecanol and dodecanol (0.01–25 mM) and the resulting cellular biomass quantified from challenged (a) conidia and (b) biofilm after 24 h. Results are presented as the mean of eight replicates from five different strains of A. fumigatus performed on two separate occasions. Error bars represent the SE of the mean.

The three most effective exogenous inhibitory compounds were tested against preformed mature A. fumigatus biofilms. The biomass of A. fumigatus biofilms was shown to be reduced by all three compounds in a concentration-dependent manner, with decanol showing a reduction across the entire concentration range tested, whereas both decanoic acid and dodecanol did not reduce the biomass significantly until concentrations of 1.6 mM were applied. All three agents reduced the biomass by ≥85% at 25 mM (Fig. 3b).


The pulmonary cavity of CF patients is a unique environment impacted by a complex microbial ecology. However, to date, relatively little is known about bacterial–fungal cross kingdom interactions within the CF lung. Cell-to-cell signalling is thought to play an important role in determining the ability of particular pathogens to compete with each other for space and nutrients and may contribute to the ability of microorganisms to persist within the CF pulmonary cavity. The data presented herein are suggestive that an antagonistic relationship exists between A. fumigatus and P. aeruginosa, which is influenced through the release of small diffusible extracellular molecules.

Pseudomonas aeruginosa and A. fumigatus are frequently isolated from CF patients. Typically by the age of 18, up to 80% of CF patients are infected with P. aeruginosa, whereas the incidence of A. fumigatus is somewhat variable in CF patients (Bakare et al., 2003; Valenza et al., 2008). This study demonstrated that P. aeruginosa significantly impedes A. fumigatus growth. This is in agreement with reports from elsewhere describing antagonistic properties for bacteria isolated from clinical pulmonary samples (Kerr et al., 1999; Yadav et al., 2005). However, investigation of the antifungal properties of bacterial CF lung pathogens against a panel of fungi, including A. fumigatus, showed that P. aeruginosa clinical isolates were shown to be unable to completely inhibit A. fumigatus (Kerr, 1994a, b). In agreement, our data showed that once filamentous biofilms had been produced, the inhibitory capacity of P. aeruginosa was significantly restricted, with coaggregation upon hyphae observed throughout A. fumigatus biofilms. Recent studies report a similar phenomenon, where P. aeruginosa and C. albicans were shown to exhibit a degree of mutual inhibition within the biofilm (Bandara et al., 2010b), suggesting that these mixed species consortia play a role in the pathobiology of the CF lung.

We have shown that the antagonistic interaction is both mediated by direct contact and secreted extracellular molecules. This is in accordance with previous investigations that have examined supernatants from bacterial strains found in the respiratory and gastrointestinal tracts, which identified P. aeruginosa supernatant to have inhibitory properties against A. fumigatus (Yadav et al., 2005). The main antimycotic agent was shown to be pyocyanin and 1-hydroxyphenazine, which are controlled by multiple quorum-sensing systems (Kerr et al., 1999). These networks of genes may play an important role in controlling the interactions between P. aeruginosa with A. fumigatus. It was reported that both HSL molecules and lipopolysaccharides influence C. albicans morphology and biofilm formation, and that signalling between these two CF pathogens is bidirectional, with farnesol inhibiting the swarming ability of P. aeruginosa (McAlester et al., 2008; Bandara et al., 2010a). Further work is required to determine whether bidirectional chemical interactions exist between P. aeruginosa and A. fumigatus, as no quorum-sensing molecule has been identified as yet for A. fumigatus. This is indeed likely as autoregulatory molecules have been identified from a range of fungal pathogens, including C. albicans (farnesol and tyrosol), Saccharomyces cerevisiae (tryptophol and phenylethylalcohol), Cryptococcus neoformans (11-mer) and Penicillium paneum (octen-3-ol) (Hornby et al., 2001; Chen et al., 2004; Chitarra et al., 2004; Alem et al., 2006; Lee et al., 2007).

The interaction between P. aeruginosa with fungi has been reported, with C. albicans exposure to P. aeruginosa quorum-sensing molecules inhibiting filamentation (Hogan & Kolter, 2002; Hogan et al., 2004; Shirtliff et al., 2009; Holcombe et al., 2010). Our study reported that the deletion of the principal quorum-sensing networks of P. aeruginosa (LasIR) significantly reduced the capacity for A. fumigatus to form hyphae and undergo biofilm development. Given that a similar inhibitory effect was observed both through direct and through indirect interaction suggested that the release of small heat-stable molecule was responsible for the inhibition, which was confirmed as both filtered and heat-killed supernatants also elicited a biological effect. However, similar inhibition profiles were observed for both LasI and LasR, the former of which is unable to synthesize HSL. These data indicate that molecules, other than or in addition to, HSLs may play a role in modulating A. fumigatus filamentation.

Hogan et al. (2004) demonstrated that 3OC12-HSL inhibited the dimorphic switching of C. albicans at a range of concentrations, whereas the smaller molecule C4-HSL had no effect on C. albicans (Hogan et al., 2004). The authors tested 10 different structurally related compounds to assess their ability to inhibit the filamentation of C. albicans, of which four (3OC12-HSL, C12-HSL, dodecanol and farnesol) inhibited the dimorphic switching of C. albicans. We examined a range of similar-sized short-chain carbon molecules in our biofilm system, demonstrating that the 10-carbon-based molecules had the ability to both inhibit filamentation and disrupt biofilm formation. These short-chain carbon molecules have also been reported to have inhibitory properties against S. cerevisiae and C. albicans (Bergsson et al., 2001; Kubo et al., 2003). The size of the chain length is clearly an important factor, which was confirmed in studies of the activity of 40 isomers of farnesol, which concluded that subtle changes in the structure of farnesol can have dramatic effects on the activity against C. albicans (Shchepin et al., 2003). At the molecular level, it is likely that these molecules act to influence key transcription factors, leading to hyphal repression. Both farnesol and dodecanol were shown to affect the cAMP-controlled Ras1-Cdc35 pathway, which is integral to filamentation (Davis-Hanna et al., 2008). Genome analysis of Aspergillus species indicates that cAMP signalling is conserved, thus indicating that these small 10 carbon molecules may play a pivotal role in fungal population control (Lafon et al., 2006). Moreover, recent transcriptional studies to examine the effects of P. aeruginosa supernatant on C. albicans biofilm formation demonstrated that 236 genes were differentially regulated, and interestingly, genes involved in adhesion and biofilm formation were downregulated, in particular YHP1, which encodes a protein known to inhibit biofilm formation (Holcombe et al., 2010).

The suppression of other microbial pathogens via the secretion of small molecules may play a pivotal role in microbial competition. Within the environment of the CF lung, bacteria and fungi exist within close proximity, and given that bacterial quorum-sensing molecules have been identified directly from sputum samples of CF patients, it is plausible that complex microbial interactions are modulated through small defined chemical messengers to allow different bacteria and cross-kingdom interactions to take place that impact microbial pathogenicity (Singh et al., 2000; Shirtliff et al., 2009). Investigation of P. aeruginosa clinical isolates from CF patients has shown that the genetic diversity of quorum-sensing networks is common, with 19 out of 30 CF patients reported to contain lasR mutants (Smith et al., 2006). This indicates that P. aeruginosa may evolve within the complex microbial environment to allow its coexistence with eukaryotes, which is supported by data from a recent study describing mutual inhibition (Bandara et al., 2010b). Interesting observations from the same group showed that exogenous lipopolysaccharide was able to inhibit and disrupt Candida spp. biofilms in a time- and concentration-dependent manner (Bandara et al., 2010a). Collectively, the data demonstrate that P. aeruginosa has the ability to modulate C. albicans behaviour in a number of ways, and under certain circumstances, it can mutually coexist.

Overall, this study demonstrated that filamentation and subsequent biofilm formation of A. fumigatus is inhibited by P. aeruginosa and its associated secreted heat-stable molecules. The analysis of defined mutant isolates revealed that the ability of P. aeruginosa to interfere with the morphological differentiation is dependent on the quorum-sensing networks that regulate an array of virulence factors. However, given that the LasI mutant cannot synthesize HSL, it is likely that this and other undefined small heat-stable molecules influence A. fumigatus and other filamentous fungi, such as those molecules reported herein. These findings could be harnessed to produce novel therapeutics as a means of managing aspergillosis more effectively.


We would like to thank Helen Kennedy (Royal Hospital for Sick Children, Yorkhill Division, Glasgow) for providing all the clinical A. fumigatus isolates used throughout this study. We thank Dr Douglas Storey (University of Calgary, Canada) for provision of the P. aeruginosa isolates and Professor Paul Williams (University of Nottingham) for kindly donating the P. aeruginosa LasIR mutant strains.