Pseudomonas-Candida Interactions: An Ecological Role for Virulence Factors

See allHide authors and affiliations

Science  21 Jun 2002:
Vol. 296, Issue 5576, pp. 2229-2232
DOI: 10.1126/science.1070784


Bacterial-fungal interactions have great environmental, medical, and economic importance, yet few have been well characterized at the molecular level. Here, we describe a pathogenic interaction betweenPseudomonas aeruginosa and Candida albicans, two opportunistic pathogens. P. aeruginosa forms a dense biofilm on C. albicans filaments and kills the fungus. In contrast,P. aeruginosa neither binds to nor kills yeast-form C. albicans. Several P. aeruginosa virulence factors that are important in disease are involved in the killing of C. albicans filaments. We propose that many virulence factors studied in the context of human infection may also have a role in bacterial-fungal interactions.

Interactions between prokaryotes and eukaryotes are ubiquitous. Although the pathogenic and symbiotic relationships bacteria have with plants and animals have garnered the most attention, the prokaryote-eukaryote encounters that occur among microbes are likely far more common. Many of the virulence factors that we study in the context of human disease may also have an ecological role within microbial communities.

Bacteria and unicellular eukaryotes, such as yeasts and filamentous fungi, are found together in a myriad of environments and exhibit both synergistic and antagonistic interactions (1,2). Here, we describe a pathogenic relationship between a fungus, Candida albicans, and a bacterium, Pseudomonas aeruginosa, that involves genes important for bacterial virulence in mammals. P. aeruginosa is prevalent in soils and is often found on the skin and mucosa of healthy individuals (3). In compromised hosts, however, P. aeruginosa uses an arsenal of virulence factors to cause serious infections associated with burns, catheters, and implants. C. albicans is also a benign member of the skin and mucosal flora. When host defenses falter, however,C. albicans initiates invasive growth that can lead to severe disease (4). In the host, C. albicans exists as both yeast-form and filamentous cells, and the ability to induce filamentation is important for its virulence (5). Several studies suggest that P. aeruginosaand C. albicans interact with each other in the human body (6–11). A molecular understanding of bacterial-fungal interactions, such as those between P. aeruginosa and C. albicans, should allow us to more effectively explore the interface between bacterial pathogenesis and microbial ecology.

Upon mixing cultures of P. aeruginosa andC. albicans, we observed that P. aeruginosareadily attached to C. albicans filaments (Fig. 1A), but almost never adhered to yeast-form C. albicans cells even after prolonged incubation (Fig. 1B) (12). Differences in the cell walls of yeast-form and filamentous C. albicans likely explain the selective attachment of P. aeruginosa to fungal filaments (13). Initial contact with the filament was usually made by one pole of the bacterial cell (Fig. 1A), and at least some cells were attached to the filament by the pole opposite the flagellum (Fig. 1A, inset) (14). Attachment to the fungus was affected by the physiological state of P. aeruginosa. For example, many more bacteria attached to filaments after a short period of coincubation when P. aeruginosa were taken from stationary-phase (WT) rather than from exponential-phase cultures (WT-EP) (Table 1) (12). Over the course of 24 to 48 hours, bacteria attached along the filamentous portion of C. albicans cells (Fig. 1C) and ultimately formed biofilms containing bacterial cells at high density surrounded by phase-bright material suggestive of an extracellular matrix (Fig. 1D). In the conditioned medium used in these experiments, biofilm formation occurred predominantly on the fungal filaments and not on the underlying glass coverslip (Fig. 1D). Given the scarcity of nutrients under these conditions, it is likely that forming a biofilm on fungal filaments enables P. aeruginosato obtain nutrients from the fungus.

Figure 1

Microscopic examination of the physical interactions between P. aeruginosa and C. albicans. (A) Phase-contrast image shows a single filamentous C. albicans cell after incubation with P. aeruginosa for 1 min. (Inset) Transmission electron micrograph shows polar attachment to the fungal filament. The flagellum is indicated by a black arrow. (B) Phase-contrast image shows P. aeruginosaincubated with yeast-form cells for 24 hours. (C) Scanning electron micrograph shows a fungal filament that was incubated in the presence of P. aeruginosa for 48 hours. Scale bar, 2 μm. (D) A phase-contrast image shows a biofilm surrounding a fungal filament after incubation for 72 hours. See ScienceOnline for experimental details (12).

Table 1

Initial attachment of P. aeruginosa toC. albicans filaments. See ScienceOnline for experimental details (12). P. aeruginosa wild type (WT), pilB, andpilC were taken from early stationary-phase cultures (OD600 of 1.2). WT-EP cells were taken from exponential phase cultures (OD600 of 0.4). The percentage of filaments colonized was determined by microscopic examination of 150 filaments per sample after a 20-min incubation. Each value represents the average of triplicate samples. In the last column, number represents the average number of bacteria attached to 300 filaments.

View this table:

Plate count assays showed that C. albicansfilaments were killed by P. aeruginosa, whereas yeast-formC. albicans retained full viability (Fig. 2). The viability of C. albicans filaments was measured using the constitutively filamentous tup1 mutant (15) to circumvent the difficulties associated with measuring the survival ofC. albicans filaments in the presence of a population of yeast cells. The number of surviving C. albicans tup1decreased after 24 hours in the presence of P. aeruginosa(Fig. 2B), but not after incubation in conditioned medium from stationary phase P. aeruginosa cultures without bacteria, showing that the presence of P. aeruginosa was necessary for fungal killing. Furthermore, the death of the fungal cell occurred after the onset of biofilm formation, suggesting a causal relationship between biofilm formation and fungal killing.

Figure 2

Survival of C. albicans in the presence of P. aeruginosa. The viability of (A) yeast-form and (B) the constitutively filamentous C. albicans tup1mutant was measured in the presence (squares) or absence (circles) of P. aeruginosa cells. The total volume was 2 ml in a 13-mm by 100-mm tube, and cells were incubated with mild shaking at 30°C. C. albicans titers were determined by plating on YPD medium supplemented with tetracycline (60 μg/ml), gentamicin (30 μg/ml), and chloramphenicol (30 μg/ml) to suppress the growth of P. aeruginosa.

To investigate the role of P. aeruginosavirulence genes in P. aeruginosaC. albicansinteractions, we analyzed three classes of P. aeruginosamutants (12) for their ability to form biofilms on C. albicans filaments and to kill the fungus (Fig. 3). In these experiments, we included P. aeruginosa mutants defective in (i) surface structures, (ii) secreted factors, or (iii) regulatory molecules. Although most of the mutants formed biofilms that were indistinguishable from those of the wild type, the biofilms produced by the P. aeruginosa flgK and lasRmutants were less robust than the wild type even after 72 hours (Fig. 3, A and B). P. aeruginosa mutants in thelas-quorum sensing pathway are also known to produce thinner, less differentiated biofilms on glass (16), suggesting that the genetic control of biofilm formation on fungal filaments may share some elements with the regulation of biofilm development on inert surfaces. P. aeruginosa PA14 mutants lacking pole-localized type IV pili (pilB andpilC) also do not form mature biofilms on abiotic substrates (17, 18), yet these mutants still formed robust biofilms on filamentous C. albicans (Fig. 3C). Because someP. aeruginosa strains use type IV pili to attach to epithelial cells (17), we assayed their initial attachment to C. albicans filaments and found a decreased rate of adherence (Table 1). These data indicate that pili somehow participate in the initial attachment to C. albicansfilaments, but they are not required for biofilm formation at later time points (compare Fig. 1D with Fig. 3C). The rpoN mutant forms extremely poor biofilms on C. albicans filaments (Fig. 3D), likely owing to multiple factors, including the lack of a flagellum and a decreased growth rate in minimal media (19). With the exception of the rpoN mutant, all P. aeruginosa mutants had the same planktonic growth rate as the wild type (20).

Figure 3

(A to D) Representative phase-contrast images show P. aeruginosa mutant biofilms onC. albicans tup1 after 72 hours of coincubation. (E to G) C. albicans tup1 survival in the presence of isogenic P. aeruginosa mutants. C. albicans tup1 viability (measured as CFU/ml) was followed in the presence of wild-type cells (black) and conditioned medium without cells (black dashed). Same results plotted in (E), (F), and (G). (E)C. albicans viability was monitored in the presence of mutants lacking type IV pili, pilB (blue), pilC(yellow), or lacking the polar flagellum, flgK (red). (F)C. albicans viability was monitored in the presence of mutants defective in the production of virulence factors including phospholipase, plcS (yellow) and plcR (blue); phenazines, phnAB (green); and exotoxin A, toxA(red). (G) C. albicans viability was measured in cultures with P. aeruginosa mutants defective in regulators that control the production of virulence factors including rpoN(green), gacA (blue), rhlR (yellow), andlasR (red). The values plotted represent the averages of four replicate cultures, and the experiment was performed multiple times with similar results. At 42 hours, the differences between the wild type and all mutants except flgK and toxAwere statistically significant at P < 0.05 as determined by a t test analysis.

Pseudomonas aeruginosa mutants were assayed for their virulence toward C. albicans. First, we analyzedP. aeruginosa mutants lacking type IV pili and the polar flagellum. Although the type IV pili mutants made robust biofilms surrounding the filament (Fig. 3C), they did not kill C. albicans tup1 until after 48 hours (Fig. 3E). The mechanisms by which type IV pili influence fungal killing remains unknown, but may involve pilus retraction to bring the bacterium in close contact with the fungal cell or the use of pili as sensors that signal attachment to the fungal surface (17, 21). The flagellar mutant killsC. albicans filaments with kinetics similar to those of the wild type (Fig. 3E), even though it forms slightly smaller biofilms on fungal filaments (Fig. 3A) (22). The second class ofP. aeruginosa mutants was defective in the production of broad-spectrum secreted factors that contribute to virulence toward diverse organisms including mammals, plants, and insects (23, 24). Several of these mutants were also attenuated for their virulence toward C. albicansfilaments. P. aeruginosa mutants unable to produce the hemolytic phospholipase C (because of disruption of eitherplcS, the structural gene, or plcR, which is required for phospholipase C secretion) were significantly attenuated in their ability to kill C. albicans filaments (Fig. 3F) (25, 26). The same was true for thephnAB mutant, which is unable to synthesize phenazine antibiotics (Fig. 3F) (27). A P. aeruginosa mutant lacking exotoxin A, a type II–secreted toxin that targets translation in eukaryotic cells, was not significantly affected in its virulence toward C. albicans (Fig. 3F) (23). Last, inactivation of several virulence-factor regulators including GacA, LasR, RhlR, and RpoN, resulted in either delayed or attenuated virulence toward C. albicans filaments (Fig. 3G). The P. aeruginosa rpoN mutant, which was not capable of biofilm formation on fungal filaments, was also unable to kill C. albicans. P. aeruginosa mutants defective in gacA, lasR, and rpoN are also significantly attenuated in other virulence models (19,23, 28). The less virulent phenotypes of pleiotropic regulatory mutants (gacA, lasR,rhlR, and rpoN) likely resulted from the decreased expression of multiple genes. In addition, some genes, such as those involved in phenazine production, are controlled by multiple regulators (19, 27, 29). However, the attenuated virulence of multiple mutants defective in single traits (plcR, plcS, pilB, pilC, orphnAB) more clearly shows that multiple mechanisms act in concert to kill C. albicans filaments. This explains the eventual decrease in C. albicans CFU/ml in almost all cultures after 60 hours (Fig. 3, E to G).

Our data suggest a link between biofilm formation and the activity of some eukaryotic-specific virulence factors toward fungal cells. Both clinical and environmental isolates of P. aeruginosa produce a similar spectrum of virulence factors including type IV pili, phospholipase C, and phenazines (30,31). Thus, we speculate that antagonism between bacteria and microscopic fungi has contributed to the evolution and maintenance of many pathogenesis-related genes. Furthermore, we propose that the interactions between P. aeruginosa and C. albicans reflect the relationships of bacterial and fungal species that coexist in other environments. A deeper understanding of bacterial-fungal interactions may provide a new perspective on the role of known virulence determinants and may lead to the discovery of new factors involved in pathogenicity in multiple hosts.

  • * To whom correspondence should be addressed. E-mail: rkolter{at}


View Abstract

Navigate This Article