Pseudomonas aeruginosa is a Gram-negative bacterial species that causes several opportunistic human infections. This organism is also found in the environment, where it is renowned (like other Pseudomonads) for its ability to use a wide variety of compounds as carbon and energy sources. It is a model species for studying group-related behaviour in bacteria. Two types of group behaviour it engages in are intercellular signalling, or quorum sensing, and the formation of surface-associated communities called biofilms. Both quorum sensing and biofilm formation are important in the pathogenesis of P. aeruginosa infections. Quorum sensing regulates the expression of several secreted virulence factors and quorum sensing mutant strains are attenuated for virulence in animal models. Biofilms have been implicated in chronic infections. Two examples are the chronic lung infections afflicting people suffering from cystic fibrosis and colonization of indwelling medical devices. This review will discuss quorum sensing and biofilm formation and studies that link these two processes.
The basics of Pseudomonas aeruginosa quorum sensing
In 1993, P. aeruginosa was shown to utilize acyl-homoserine lactone (AHL)-based quorum sensing to regulate the expression of virulence factors (Passador et al., 1993). Currently, it is known to have two primary AHL systems, las and rhl. Each system has its own AHL synthase (LasI and RhlI), an AHL-responsive, DNA-binding regulator (LasR and RhlR) and AHL signal (3-oxo-dodecanoyl homoserine lactone and butyryl homoserine lactone) (Fuqua and Greenberg, 2002; Juhas et al., 2005; Wagner et al., 2006). The las and rhl systems constitute a regulatory cascade, with the rhl system under control of the las system (Fig. 1). In addition to the las and rhl systems, there is another DNA-binding response regulator, QscR (Chugani et al., 2001). Although QscR has no associated AHL synthase, it has been shown to respond to AHL signals and is a key regulator that controls expression of a subset of quorum sensing-controlled genes (Lequette et al., 2006). Finally, there is evidence of another quorum sensing response regulator, VqsR, although it is not known whether it interacts directly with AHL signals (Juhas et al., 2004).
Pseudomonas aeruginosa also possesses a non-AHL extracellular signal, PQS (for Pseudomonas quinolone signal), which is integrated into the AHL signalling circuit. PQS is 2-heptyl-3-hydroxy-4-quinolone and its biosynthesis requires multiple genes (Pesci et al., 1999; Diggle et al., 2003). In addition to PQS, a precursor of PQS biosynthesis, 4-hydroxy-2-heptylquinoline is also secreted from the cell and may act as a signalling molecule (Deziel et al., 2004). Recently, Mashburn et al. demonstrated that the highly hydrophobic PQS signal is exported through membrane vesicles that bleb off of the bacterium (Mashburn and Whiteley, 2005). This is in stark contrast to the AHL signals, which either diffuse directly across the cell membrane (butyryl-HSL) or their transport out of the cell is assisted by an efflux pump (3-oxo-dodecanoyl-HSL) (Pearson et al., 1999).
Several transcriptional profiling studies have demonstrated that quorum sensing is a global regulatory network, as it controls the expression of at least 400 genes either directly or indirectly (Hentzer et al., 2003; Schuster et al., 2003; Wagner et al., 2003). Numerous cellular functions are regulated by quorum sensing, including the production of secreted siderophores, proteases, toxins and a type of surface motility called swarming. Several global regulators, such as RpoS, GacA/S and Vfr, also impact cell-to-cell signalling (Albus et al., 1997; Reimmann et al., 1997; Schuster et al., 2004). Thus, the quorum sensing regulatory network is complex and may be influenced by multiple environmental signals/conditions.
Models for P. aeruginosa biofilm development
In the last 5–10 years, researchers have intensely studied the genes that contribute to P. aeruginosa biofilm formation on abiotic surfaces (e.g., glass and plastics), with the hope that understanding how the bacteria build these communities would aid in the development of antibiofilm therapies. Their work has revealed that biofilm formation proceeds through discrete steps: initial attachment, the formation of mature biofilms and ultimately dispersion, where a subpopulation of the community swims away from the mature biofilm, reinitiating the cycle (Fig. 2). Several factors have been found to influence biofilm formation at different steps. These studies have also demonstrated a couple of important points. The first is that, in many cases, the biofilm culturing conditions can influence the requirement of a particular factor. For example, the requirement of the single polar flagellum for initial adherence to a surface depends upon the carbon source on which P. aeruginosa is grown (Klausen et al., 2003a). Another example is the iron siderophore pyoverdine. Under iron limiting conditions, a pvdA mutant strain is impaired for biofilm formation compared to the wild-type strain, while under iron replete conditions the two strains form biofilms that are indistinguishable (Banin et al., 2005). The second important point is that biofilm formation can be a fairly complex process with subpopulations within the community carrying out different behaviours that contribute to the final biofilm structure. It is not simply just bacteria attaching to a surface and multiplying.
Several groups have demonstrated that P. aeruginosa is capable of forming two general types of biofilms in the laboratory based on their structure. A ‘flat’ biofilm is characterized by a relatively confluent, uniform community of bacteria on the surface (Fig. 2). A ‘structured’ biofilm consists of cell aggregates or ‘mushrooms’ separated by channels or spaces (Fig. 2). This is a bit of an oversimplification as there are a gradient of biofilm structures between these two extremes. Both biofilm-types show enhanced tolerance to antimicrobial treatment relative to planktonic or free-swimming cells – a hallmark of biofilms (Hentzer et al., 2001; Teitzel and Parsek, 2003). However, flat biofilms may be more susceptible to antimicrobials than structured biofilms (Allesen-Holm et al., 2006; Landry et al., 2006), although this remains to be thoroughly tested.
There are features of development specific to either the flat or structured biofilms. The formation of a flat biofilm is characterized by significant surface motility (Singh et al., 2002; Klausen et al., 2003a). Evidence suggests that both twitching and swarming motility may contribute to this process (Shrout et al., 2006). Ultimately, bacteria multiply and continue to move on the surface, forming a confluent mat of cells (Fig. 2).
The development of structured biofilms initially involves the formation of cell aggregates on a surface, after which biofilm formation may proceed in two ways. The first we call ‘structured biofilm I’. Here, a motile subpopulation of the biofilm migrates to the surface of a non-motile subpopulation called a ‘stalk’ (depicted by the orange cells in Fig. 2), forming the ‘cap’ of a mushroom-like structure (Klausen et al., 2003b). Lequette et al. suggested that the stalk population may be producing the surfactant rhamnolipid, which is thought to ‘grease’ the surface of the stalk, facilitating movement of the motile subpopulation onto it (Lequette and Greenberg, 2005). The second we call ‘structured biofilm II’. This involves clonal growth of the small aggregates into large aggregates. This mechanism may be important for the formation of structured biofilms by Pseudomonas putida, as well as biofilm formation by mucoid P. aeruginosa and biofilm formation on a mucin-coated surface (Tolker-Nielsen et al., 2000; Hentzer et al., 2001; Landry et al., 2006).
Several factors have been shown to influence whether P. aeruginosa forms a flat or structured biofilm. The nutritional environment is key in this regard. Klausen et al. demonstrated that P. aeruginosa grown on glucose as the sole carbon source forms structured biofilms, while glutamate- and succinate-grown biofilms were flat (Klausen et al., 2003a). The attachment surface can also influence biofilm development. Landry et al. demonstrated that surfaces coated with the glycoprotein mucin [found in abundance in the airways of cystic fibrosis (CF) patients], caused P. aeruginosa to form structured biofilms (Landry et al., 2006). A mucin adhesin on the tip of the flagellum was shown to bind mucin, immobilizing the cell. Under identical culturing conditions on glass surfaces, the same strain formed flat biofilms characterized by significant surface motility. In addition, exopolysaccharide overproduction has been shown to contribute to the formation of structured biofilms. Alginate overproduction by mucoid P. aeruginosa contributes to the formation of structured biofilms (Hentzer et al., 2001; Nivens et al., 2001). Overexpression of two other polysaccharide loci, pel and psl, also contribute to the formation of structured biofilms (Friedman and Kolter, 2004a,b; Jackson et al., 2004; Matsukawa and Greenberg, 2004). In both these cases, the biofilm developmental pattern follows that of structured biofilm II depicted in Fig. 2 (Kirisits et al., 2005 and data not shown).
Quorum sensing contributes to biofilm formation – sometimes
An initial report linked quorum sensing and biofilms, comparing the biofilm structure of a wild-type P. aeruginosa strain to isogenic rhlI, lasI and lasIrhlI mutant strains (Davies et al., 1998). They found that the wild-type and rhlI mutant strains formed structured biofilms, while the lasI and the lasIrhlI mutant strains formed flat, undifferentiated biofilms. The flat biofilms were susceptible to treatment with the surfactant sodium dodecyl sulphate, while the structured biofilms were resistant. The authors concluded that las-regulated functions were required for biofilm formation.
Subsequent research indicated that quorum sensing's role in P. aeruginosa biofilm formation was not clear-cut. Heydorn et al. reported that wild-type and quorum sensing mutant strains formed flat biofilms that were structurally indistinguishable (Heydorn et al., 2002). On the other hand, Purevdorj showed that under high flow conditions both wild-type and quorum sensing mutant strains formed structured biofilms, although the biofilms differed slightly in microscopic appearance (Purevdorj et al., 2002). M. Givskov's group demonstrated that AHL signal analogues called furanones, known to inhibit P. aeruginosa quorum sensing, impaired biofilm development when added to the growth medium (Hentzer et al., 2002; 2003). In addition, they showed that a lasRrhlR double mutant strain produced biofilms more susceptible to the clinically relevant antibiotic tobramycin than an isogenic wild-type strain (Bjarnsholt et al., 2005).
Several explanations were put forth to explain these discrepancies, including differences in strains, biofilm reactors and culturing conditions. Perhaps the reported differences were not too surprising. Quorum sensing regulates many different functions and its control of these functions can change depending on environmental conditions. Supporting this notion, several different quorum sensing-regulated functions have been shown to impact biofilm formation at different stages.
The quorum sensing-regulated surfactant rhamnolipid is necessary for maintaining the open spaces between cell aggregates in structured biofilms (Davey et al., 2003). In addition, rhamnolipid production may aid in the formation of mature mushroom structures (Lequette and Greenberg, 2005). Another quorum sensing-regulated factor shown to contribute to biofilm formation is the siderophore pyoverdine. As mentioned previously, pyoverdine is a key means for acquiring iron, and mutants unable to make pyoverdine formed flat biofilms, while an isogenic wild-type strain formed structured biofilms (Banin et al., 2005). P. aeruginosa also uses quorum sensing to regulate the production of two sugar-binding lectins, LecA and LecB, which are secreted from the cell. These lectins are expressed in biofilms and both lecA and lecB mutant strains formed aberrant biofilms (Tielker et al., 2005; Diggle et al., 2006).
DNA is known to be an integral part of the extracellular matrix that acts as a scaffolding holding P. aeruginosa biofilm cells together (Whitchurch et al., 2002). Recently, Allesen-Holm demonstrated that DNA is actively secreted in a quorum sensing-dependent fashion (Allesen-Holm et al., 2006). They showed that quorum sensing mutant strains formed aberrant biofilms susceptible to SDS treatment. Finally, recent evidence suggests that swarming, a quorum sensing-regulated surface motility, contributes to biofilm formation. Swarming may play a key role early in biofilm development, with actively swarming cells forming flat biofilms and cells exhibiting reduced swarming forming structured biofilms (Shrout et al., 2006).
Since so many different quorum sensing-regulated functions affect biofilm development, one would expect quorum sensing mutant strains to have profound biofilm defects. Yet several reports indicate that certain culturing conditions produce identical biofilms for wild-type and quorum sensing mutant strains! There are a few potential explanations for this. One is that a quorum sensing response may not be induced or active in biofilms grown under these conditions. There have only been a few studies that have specifically assessed when and where a quorum sensing response occurs in a biofilm community. de Kievit et al. (2001) used lasI-gfp and rhlI-gfp reporter strains to demonstrate that expression is highest near the attachment surface and decreases towards the periphery of the biofilm. However, this analysis was performed for only a single culturing condition. There may be conditions, such as high liquid flow, where signal concentrations may not reach an inducing level in the biofilm. PQS signalling might also explain some of the observed differences in biofilm formation by AHL quorum sensing mutants. PQS signalling has been shown to promote P. aeruginosa biofilm formation and in certain cases can function independently of the AHL signalling systems (Diggle et al., 2003). There might be culturing conditions where the PQS system is active in some AHL regulatory mutant strains, activating expression of quorum sensing controlled genes that contribute to biofilm formation. Another possibility is that the environment may dictate if and which quorum sensing-controlled functions contribute to biofilm formation. The next section will discuss environmental parameters that can influence quorum sensing in biofilm communities.
The impact of environment on quorum sensing's contribution to biofilm formation
Quorum sensing signal production, stability and distribution are influenced by a range of physical, chemical and nutritional factors. For instance, Wagner et al. assessed transcriptional profiles of P. aeruginosa grown under a variety of conditions. A significant portion of the virulence genes thought to require quorum sensing for expression did not for some of the tested conditions (Wagner et al., 2004). Therefore, although quorum sensing may have a clear role in biofilm development under a particular set of conditions, this role may change as environmental conditions change.
Acyl-homoserine lactone signal stability is a key consideration for quorum sensing in biofilms. Quorum sensing induction may be influenced or even precluded by the chemistry of the local environment. For example, the homoserine lactone ring of AHLs is hydrolysed at basic pH values (Schaefer et al., 2000). At a pH of 7, the half-life of an AHL is on the order of hours, while it is minutes at pHs above 8. Incorporating this fact into a mathematical model, Chopp et al. predicted that quorum sensing would not be induced in a P. aeruginosa biofilm above a critical pH threshold (Chopp et al., 2003). In addition to chemical hydrolysis, biological turnover is important. AHLs are rapidly degraded by several microorganisms, including P. aeruginosa and a Ralstonia sp. (Leadbetter and Greenberg, 2000; Zhang et al., 2002; Huang et al., 2003; Lin et al., 2003). This behaviour suggests that in mixed species environments, AHL-degrading organisms could effectively quench quorum sensing. Degradation usually occurs through hydrolysis of the homoserine lactone ring by lactonases, or through acylases that cleave the amide bond linking the homoserine lactone ring and the acyl side chain.
Eukaryotes also appear to have developed mechanisms to disrupt quorum sensing. One example is the macroalga Delisea pulchra, which produces AHL-like halogenated furanones that disrupt quorum sensing (Givskov et al., 1996). This may partially explain why its leaves are not significantly colonized by microorganisms. Primary human respiratory epithelial cells have also been shown to produce an enzyme called paraoxonase, a lactonase which was shown to degrade AHL signals and impair quorum sensing (Chun et al., 2004).
The physical environment can also influence quorum sensing in biofilms. One important aspect of the physical environment is mass transfer. AHLs produced within a biofilm can diffuse out and be removed from the local environment in a system subjected to fluid flow. The hydrodynamic environment could determine the rate at which AHLs are removed from a biofilm. Thus, flow rates could affect the population size of the biofilm needed to produce inducing concentrations of AHL. Certainly hydrodynamic conditions have been shown to affect biofilm structure. For example, Pereira et al. observed a more densely packed biofilm when Pseudomonas fluorescens was grown under turbulent conditions than under laminar conditions (Pereira et al., 2002). Furthermore, it was shown that the structure of a P. aeruginosa biofilm formed under turbulent conditions was more heavily influenced by flow rate than by quorum sensing (Purevdorj et al., 2002), suggesting that certain hydrodynamic conditions might delay or prevent the onset of quorum sensing.
What is the role of P. aeruginosa quorum sensing in environmentally and clinically relevant biofilms?
Many researchers appreciate that biofilm communities are probably an important aspect of P. aeruginosa existence in both natural and clinical settings. An important question is what correlations (if any) are there between the flat and structured laboratory biofilms and ‘real world’ biofilms? This question is impossible to answer at this point in time. However, as we begin to understand how environmental conditions influence biofilm structure, analysis of the specific systems may give clues. We may be able to probe biofilms in situ for specific physiological markers specifically associated with flat or structured biofilms.
The pure culture laboratory work described in the previous sections is crucial for understanding the molecular mechanisms and linkages between quorum sensing and biofilm formation. However, if and how quorum sensing and biofilms are linked in actual clinical and natural environments is unclear. The nature of this linkage might change depending upon environmental context. Quorum sensing may play a crucial role in the formation of biofilm communities (as discussed in the preceding section). On the other hand (or additionally), biofilm formation may represent a means to achieve a quorum. In this section we discuss research that may provide insight into some of these questions.
Quorum sensing, biofilms and chronic infection
Pseudomonas aeruginosa is capable of causing drastically different opportunistic infections. Acute burn wound infections, pneumonia and other primary breaches in host defences are characterized by rapid, explosive growth and systemic spread of the organism, many times leading to sepsis and ultimately death of the host. Chronic infections, such as CF lung infections and colonization of indwelling medical devices (e.g., catheters), are characterized by biofilm formation, localized containment of the bacteria to a specific area of the body and persistence. Recent models have proposed that factors that promote acute and chronic virulence are inversely regulated (Furukawa et al., 2006).
Cystic fibrosis represents a model biofilm infection (Parsek and Singh, 2003). During the course of CF, P. aeruginosa forms biofilms characterized by cellular aggregates embedded within the mucus layer present in the airways (Lam et al., 1980). After chronic colonization occurs, antibiotic therapy can never completely eradicate the bacteria, even though high levels of antibiotics can be achieved in the airway secretions. Colonization of urinary catheters is another example of a P. aeruginosa chronic infection. Stickler et al. demonstrated that AHL signals were produced by P. aeruginosa biofilms growing on the surface of a catheter (Stickler et al., 1998).
Recent research has called into question the importance of quorum sensing in chronic infection. Smith et al. demonstrated that during the course of infection, the CF environment selects for mutations in the quorum sensing regulatory gene, lasR (Smith et al., 2006). Does this suggest that quorum sensing plays little or no role in the maintenance of biofilm communities in CF? As with some laboratory-grown biofilms, perhaps conditions are such that obviate the need for quorum sensing to make and maintain a biofilm. Another possibility is that a lasR mutation serves to minimize production of quorum sensing-regulated acute virulence factors as a means to evade the host immune response. A lasR mutation still leaves the rhl quorum sensing system intact. Singh et al. previously demonstrated that P. aeruginosa growing in laboratory biofilms and in the sputum of CF patients predominantly produce the rhl AHL, butyryl homoserine lactone (Singh et al., 2000). This may indicate that the rhl signalling system plays a more prominent role in CF biofilms and by losing the las system the organism mutes some acute virulence functions, while maintaining expression of necessary rhl-encoded factors for establishing and maintaining chronic biofilms. Although the rhl system requires a functional las system for optimal expression, secondary mutations can restore expression of the rhl system in a lasR mutant background (Van Delden et al., 1998).
How might quorum sensing be useful for P. aeruginosa inhabiting environmental biofilms? The answer may be related to two specific classes of quorum sensing-regulated factors. Neither are related specifically to biofilm growth per se. One class is factors involved in nutrient acquisition, such as the iron siderophores pyoverdine and pyochelin. In addition, secreted proteases would allow for utilization of extracellular proteins. These functions may aid in the competition for common nutritional resources. A second class is quorum sensing-regulated toxins/poisons. These poisons, such as cyanide and the phenazine antimicrobial pyocyanin, might function to impair neighbouring competitors. In line with this observation is that P. aeruginosa will many times produce the antidote for its own poisons. For example, it produces a cyanide-insensitive terminal oxidase.
To initially address the role of quorum sensing in interspecies competition, An et al. studied co-culture biofilms of P. aeruginosa and Agrobacterium tumefaciens, two microorganisms found together in the natural environment (An et al., 2006). Co-culture biofilms inoculated with equal numbers of the two species showed that P. aeruginosa covered A. tumefaciens, causing it to decrease to ∼1% of the biofilm biomass after 1 week. However, A. tumefaciens in a co-culture biofilm inoculated with a P. aeruginosa lasRrhlR mutant strain did not show as significant a decrease. The authors showed that in their binary co-culture format, quorum sensing provided a growth advantage to P. aeruginosa. A similar finding was made in a clinical context. Mashburn et al. demonstrated that P. aeruginosa used quinolones to kill Staphylococcus aureus and utilize iron derived from the lysed cells (Mashburn et al., 2005). This is a relevant partner for P. aeruginosa, as S. aureus is a common cohabitant of the CF airways.
Quorum sensing also appears to protect P. aeruginosa in environmental biofilms from eukaryotic predation. Matz et al. grew parallel biofilms of P. aeruginosa wild type and a P. aeruginosa rhlRasR mutant and exposed them to Rhynchomonas nasuta, a surface-feeding, flagellated protozoan (Matz et al., 2004). Young P. aeruginosa wild-type biofilms demonstrated greater resistance to protozoan grazing than did the quorum sensing mutant. In a subsequent set of experiments, 3-day-old biofilms of P. aeruginosa wild-type and a rhlRlasR mutant strain were grown and then exposed to protozoa. The protozoa did not survive even 24 h after inoculation to the wild-type biofilm, but they readily grew in the quorum sensing mutant biofilm environment (∼107 protozoa ml−1). The authors suggested that quorum sensing-controlled functions are lethal to the protozoa.
Pseudomonas aeruginosa is an effective opportunistic pathogen that can cause a range of chronic and acute infections. One of the reasons for this is its ability to co-ordinate activity as a group. Biofilm formation and quorum sensing are two examples of group behaviour. These two processes can be linked in different ways. For some conditions, quorum sensing may be an integral part in building a biofilm community. Additionally, biofilm formation may allow the high local cell densities necessary to achieve a quorum. Recent research has demonstrated that the link between these two behaviours is very dynamic and dependent upon the environment. This point is not too surprising as environmental conditions influence both quorum sensing and biofilm formation and these processes have been shown to involve differential expression of hundreds of genes. Future research, identifying key environmental parameters that affect the relationship between quorum sensing and biofilm formation will be crucial to understanding these two behaviours in relevant clinical and environmental contexts.
M.J.K. is supported by funding from SERDP and THWRC. M.R.P. is supported by funding from NIH, NSF and CFF. We thank Pradeep Singh for critical reading of the manuscript. We thank E.P. Greenberg and J.H. Lee for supplying Fig. 1.