In order to grow as a pure culture under lab conditions, L. pneumophila is remarkably fastidious and requires several different nutrients including iron salts and a number of amino acids such as L-cysteine, which are used as carbon, nitrogen and energy sources (George et al., 1980; Edelstein, 1982; Keen and Hoffman, 1984). The fact that L. pneumophila is commonly detected in oligotrophic aquatic environments (low nutrient content), despite its fastidious nature implies that the bacterium is able to obtain its necessary supply of amino acids and organic carbon from that same environment, more specifically from the microbial consortium located in biofilms.
Biofilms are defined as complex microbial communities characterized by cells that are attached to a substratum or phase boundary and to each other by means of a matrix of self-produced extracellular polymeric substances (EPS) (Donlan and Costerton, 2002). Microcolonies of bacterial cells encased in the EPS matrix are separated from each other by interstitial water channels, allowing transport of nutrients, oxygen, genes and even antimicrobial agents (Prakash et al., 2003). Because of their dynamic character, biofilm communities can continuously change in time and space, providing better survival and growth of the associated microorganisms. For this reason, it is easy to understand that in most natural environments biofilms are the prevailing microbial lifestyle (Watnick and Kolter, 2000). Generally, there are three distinct phases in the ‘biofilm life cycle’ of bacteria: (i) bacterial attachment to a substratum, (ii) biofilm maturation and (iii) detachment from the biofilm and subsequent dispersal in the environment (reviewed by Donlan, 2002). What is currently known about L. pneumophila, in relation to microbial biofilm communities, will be discussed using each of these three phases.
Numerous human pathogens such as Pseudomonas aeruginosa (Whiteley et al., 2001), Staphylococcus epidermidis (Mack et al., 2000), Salmonella enteritidis (Solano et al., 2002), Vibrio cholerae (Nesper et al., 2001), Streptococcus gordonii (Loo et al., 2000) and Burkholderia cepacia (Huber et al., 2001) are able to form a monospecies biofilm. In the case of L. pneumophila, pure cultures have been reported to form biofilms only under well-defined experimental conditions using the nutrient-rich Legionella medium (buffered yeast extract medium, BYE) (Piao et al., 2005; Mampel et al., 2006). Piao and colleagues (2005) demonstrated the ability of L. pneumophila to form biofilms on glass, polystyrene and polypropylene under static conditions in BYE medium at 25°C, 37°C and 42°C. A clear, significant temperature influence on biofilm thickness and structure was detected as seen in Fig. 1. Biofilms grown at 25°C consisted of rod-shaped L. pneumophila cells and possessed structural features typical of biofilms reported to date, i.e. pillar- and mushroom-like structures and what appeared to be water channels within (Fig. 1A). However, biofilms formed at 37°C and 42°C showed an even, thicker and more extensive mat of considerably greater cell density without the commonly observed water channels (Fig. 1B). Also, at higher temperatures the morphology of biofilm-associated L. pneumophila appeared to be filamentous and filaments were multinucleate but non-septate. The study performed by Mampel and colleagues (2006) reported that biofilm formation of L. pneumophila in rich medium at 30°C mostly relies on the adhesion of bacteria grown in the planktonic phase, rather than by clonal replication of sessile cells. Also, biofilms only developed under static conditions and were characterized by rather delicate, fluffy extensions from the surface.
Figure 1. Confocal laser scanning micrographs of GFP-expressing L. pneumophila Philadelphia-1 biofilms formed under static conditions in BYE medium at 25°C and 37°C. A. A 6-day-old L. pneumophila biofilm at 25°C, consisting of rod-shaped cells and structural features typical for biofilms reported to date. B. A 3-day-old L. pneumophila biofilm at 37°C, consisting of filamentous cells embedded in an even, thicker and more extensive structure. Size bar: 10 μm. Reproduced with the permission of Dr Sze Chun Chau.
Download figure to PowerPoint
Taking into account the above-mentioned two studies and the fact that L. pneumophila requires specific and fastidious growth requirements in order to be cultured in the lab, it is logical that under environmental circumstances the bacterium incorporates in pre-established biofilms as a secondary colonizer. This means that instead of docking to a surface and developing a biofilm, the bacterium will form a transient association with other microbes, previously attached to that surface and which have already initiated or formed a biofilm (Watnick and Kolter, 2000). This transient association allows L. pneumophila to search for a place to settle in the community. The concept of one species acting as a primary colonizer and promoting the adhesion and subsequent biofilm formation of another species has been shown for P. aeruginosa promoting the biofilm formation of Escherichia coli O157:H7 (Klayman et al., 2007). Mampel and colleagues (2006) observed adherence of L. pneumophila to monospecies biofilms formed by Empedobacter breve, Microbacterium sp. and Acinetobacter baumanii in BYE medium. In contrast, L. pneumophila did not attach to Pseudomonas spp., Corynebacterium glutamicum or Klebsiella pneumoniae biofilms, which suggests the existence of bacteria that oppose L. pneumophila adherence. Under natural conditions, biofilms occur mostly as complex multispecies communities. Using a more realistic experimental set-up, consisting of multispecies biofilms formed under oligotrophic conditions, it has been demonstrated that L. pneumophila rapidly colonizes present biofilm communities (Murga et al., 2001; Vervaeren et al., 2006; Declerck et al., 2007b). Time periods of less than 2 h were detected (Declerck et al., 2007b).
Whether the interaction between L. pneumophila and a pre-established biofilm is physical or involves cell signalling, better known as quorum sensing, remains unknown at this moment. Previous research has shown that adhesion in biofilms between primary and secondary colonizers may be facilitated by interspecies cell signalling using autoinducer-2 (AI-2) (Rickard et al., 2006). L. pneumophila lacks the widespread AI-2 signalling system, as well as an N-acyl-homoserine lactone (AHL)-based quorum sensing circuit. However, recently Spirig and colleagues (2008) discovered a 3-hydroxypentadecan-4-one signalling system in L. pneumophila designated as Legionella autoinducer-1 (LAI-1). Together with the cholerae autoinducer-1 of Vibrio cholerae (Higgins et al., 2007), LAI-1 forms a family of α-hydroxyketone autoinducer signalling molecules. Intercellular signalling involving α-hydroxyketones might be common among different bacterial species and genera and it warrants further research to investigate if LAI-1 allows interspecies communication between L. pneumophila and other bacteria.
All bacteria produce multiple adhesins permitting organisms to switch from planktonic to sessile forms (or the other way around) triggered by different environmental conditions (An et al., 2000). However, the success of bacterial attachment to an available surface is dictated by a number of variables, including environmental factors (e.g., the flow regime or hydrodynamics) and various properties of the cell surface (Donlan, 2002; Prakash et al., 2003). An increase in flow velocity may equate to increased attachment and it is known from the medical world that bacteria form biofilms preferentially in very high shear environments [Reynolds number (Re) > 10 000], like catheters. Liu and colleagues (2006) used a plumbing model, consisting of three parallel pipes with either a turbulent (Re > 10 000), laminar (Re < 1000) or stagnant (Re = 0) flow regime, to demonstrate a significantly higher biofilm colonization rate of Legionella spp. under turbulent flow conditions compared with laminar conditions as shown in Fig. 2. The lowest biofilm-associated Legionella spp. counts were detected under stagnant flow conditions. As mentioned above, the success rate of surface colonization also depends on various properties of the cell surface. Specific structural components like pili, fimbriae, flagella and surface-associated polysaccharides or proteins have been shown to play a critical role in facilitating bacterial interaction with surfaces (Donlan, 2002). Also, type II secretion, the mechanism by which most Gram-negative organisms export pilus machinery for assembly, translocates virulence factors that may contribute to biofilm maturation (Liles et al., 1999). In contrast to L. pneumophila, many Gram-negative bacterial pathogens have an absolute requirement for type IV pili and the type II secretion system in order to establish a biofilm and allow for maturation (Moorthy and Watnick, 2004; Giltner et al., 2006). Lucas and colleagues (2006) showed that L. pneumophila does not seem to have an absolute requirement for functional pilin expression in order to colonize pre-established biofilms. Using mutants of L. pneumophila, Lucas and colleagues showed that in the absence of suitable protozoan hosts the bacteria were able to utilize pili and or other unidentified pilin-independent adhesins to colonize biofilms. In the presence of protozoans, its unique lifestyle as a facultative intracellular parasite diminishes its need for primary attachment and retention mechanisms.
Figure 2. Comparison of the Legionella spp. concentration in inlet biofilm samples. A significantly higher number of Legionella counts were recovered from the biofilm of the turbulent flow pipe than from the laminar flow pipe, while the lowest counts were recovered from the stagnant flow pipe (P = 0.01). Legends: , turbulent flow; , laminar flow; ●, stagnant flow. Reprinted from Liu and colleagues (2006)J Appl Microbiol101: 437–442, doi:10.1111/j.1365-2672.2006.02970.x, with permission of the publisher (Wiley-Blackwell).
Download figure to PowerPoint
Growth and retention of biofilm-associated Legionella: more than one possible strategy to pick
To become a productive member of society, attached bacteria must differentiate into biofilm-associated cells, which implies an up- and/or downregulation of a number of genes, e.g., repression of flagellum synthesis (Watnick and Kolter, 2000). In the case of P. aeruginosa, 22% of the genes are upregulated in the biofilm state, while 16% are downregulated. Hindré and colleagues (2008) performed the first transcriptome analysis of L. pneumophila biofilm cells. They revealed that a substantial proportion of L. pneumophila genes are differentially expressed, as 2.3% of the 2932 predicted genes exhibited at least a twofold change in gene expression.
In the next phase of the integration process, secondary colonizers start to grow and form microcolonies, which are the basic structural unit of biofilms (Donlan, 2002). Using immunogold and fluorscein immunolabelling, Rogers and Keevil (1992) were the first to detect microcolonies of Legionella in naturally occurring biofilms. Following further growth and an increase in volume the microcolonies become embedded in an EPS matrix, which reinforces the biofilm structure. Different bacterial species produce differing amounts of EPS (Prakash et al., 2003). To date there is only one study, performed by Piao and colleagues (2005), suggesting that L. pneumophila might be able to produce EPS. Mycelial-like biofilms contained a thin layer of ruthenium red-stainable substance coating the cells that might be EPS.
In multispecies biofilms, where bacterial competition for food is high, it is advantageous to have more than one way to obtain the required growth nutrients. In the case of L. pneumophila, there exist two distinct ways. First, nutrients can be delivered by the biofilm environment itself. Numerous publications show that L. pneumophila is capable of obtaining nutrients either directly from other living microorganisms, like algae and some heterotrophic bacteria, producing them in excess, and/or indirectly from decaying organic matter. Second, L. pneumophila is able to infect and replicate inside protozoans. During its stay in the intracellular environment, peptides and proteins of the infected host are degraded and used as a nutrient source. Both ways of obtaining growth nutrients will be considered in detail.
Extracellular replication in the biofilm matrix. Biofilm-associated bacteria distribute themselves geographically based on the neighbours and environment that best suits their needs and requirements (Watnick and Kolter, 2000). The survival and growth of pathogenic microorganisms, generally characterized by a fastidious nature, within biofilms might be enhanced by the association and metabolic interactions with indigenous organisms (Donlan, 2002). Tison and colleagues (1980) were the first to demonstrate that L. pneumophila has the capability to grow on extracellular products provided by other bacteria. L. pneumophila was isolated from an algal-bacterial mat community, growing at 45°C in an anthropogenic thermal effluent. Under lab conditions, the L. pneumophila isolate grew in association with Fisherella sp. (a cyanobacterium) over a pH range of 6.9-7.6 in a mineral salts medium at 45°C. Legionella growth depended upon active photosynthesis of the cyanobacterium, implying the use of released algal products. Other algae that might support growth of L. pneumophila in basal salts medium are Scenedesmus spp., Chlorella spp. and Gleocystis spp. (Hume and Hann, 1984). In addition to algae, some heterotrophic bacteria appear able to support L. pneumophila growth. In complex media lacking L-cysteine or iron salts, and therefore unable to support Legionella growth, L. pneumophila forms satellite colonies around strains of some common aquatic bacteria, including Flavobacterium breve, Pseudomonas spp., Alcaligenes spp. and Acinetobacter spp. (Wadowsky and Yee, 1983; Stout et al., 1985).
In addition to nutritional commensalism, the ability to obtain carbon and energy sources from dead or decaying organic matter would significantly benefit the survival and growth of biofilm-associated L. pneumophila. Dead organic material can result from water disinfection measures (Van der Kooij et al., 2005) or from the biofilm itself, which as a constantly renewing entity can contain nearly 50% dead microbial cells (Tresse et al., 2003). A recent publication by Temmerman and colleagues (2006) demonstrated that L. pneumophila shows necrotrophic growth. Using several experimental set-ups, they clearly proved that L. pneumophila was able to consume and grow on dead Gram-negative bacteria like Pseudomonas putida and E. coli and amoebae like Acanthamoeba castellanii. However, when comparing necrotrophy with intracellular growth in A. castellanii as shown in Fig. 3, it appeared that the major route of L. pneumophila replication under environmental conditions occurs by means of protozoans.
Figure 3. Necrotrophic growth of L. pneumophila at 28°C in filter-sterilized tap water on heat-killed (30 min 60°C) P. putida cells and live A. castellanii cells. Quantification of L. pneumophila cells was performed during 2 weeks using real-time PCR targeting the mip gene. Reprinted from Temmerman and colleagues (2006) Appl Environ Microbiol72: 4323–4328, doi:10.1128/AEM.00070-06, with permission from the American Society for Microbiology.
Download figure to PowerPoint
Competition for substrate is considered to be one of the major evolutionary driving forces in the bacterial world, and numerous experimental data show how different microorganisms may outcompete others, as a result of a better accessibility or utilization of a given energy source. This means that, although a certain microenvironment in the biofilm may harbour sufficient nutrients for L. pneumophila, they might serve for the greater part to amplify faster-growing bacteria like E. coli and P. aeruginosa. By forming multinucleate, non-septate filaments, L. pneumophila may have hit upon another strategy to compete with fast-growing indigenous biofilm-associated bacteria (Piao et al., 2005). Under nutrient-rich conditions, the multinucleate L. pneumophila filaments can form septa. A septum is an inward growth of both the cytoplasmic membrane and cell wall, which separates compartments containing a nucleus. Each compartment represents one future daughter cell. In that way, one multinucleate, filamentous L. pneumophila bacterium can rapidly give rise to numerous daughter cells.
Intracellular replication in protozoans. The high concentration of microorganisms within biofilms provides excellent opportunities for predation by free-living protozoans, more specifically amoebae. This makes them important regulators of biofilm-associated bacterial populations, which contributes to the prokaryotic-eukaryotic co-evolution (Hilbi et al., 2007). It should therefore not be surprising that several biofilm-associated bacterial pathogens such as Mycobacterium spp., Pseudomonas spp., Vibrio spp. and Legionella spp. have evolved and acquired the capability to survive and multiply within free-living amoebae (Greub and Raoult, 2004; Molmeret et al., 2005; Weissenberger et al., 2007). In 1980, Rowbotham described for the first time the ability of L. pneumophila to replicate within amoebae (Rowbotham, 1980). Since then, L. pneumophila has been described to multiply in 14 species of free-living amoebae, two species of ciliated protozoans (Cyclidium spp. and Tetrahymena pyriformis) and one species of slime mold (Dictyostelium discoideum) (Fields et al., 2002).
Studies using flow-through or batch systems, containing a self-assembled or naturally occurring biofilm consortium, demonstrated that amoebae like Hartmanella vermiformis and A. castellanii are used for replication of biofilm-associated L. pneumophila in oligotrophic aquatic environments [Murga et al., 2001; Kuiper et al., 2004; Declerck et al., 2007b (Fig. 4)]. Moreover, in the absence of amoebae, biofilm-associated L. pneumophila numbers did not increase at all. Bacteria were only able to persist in the biofilm community and in some cases entered the viable but non-culturable (VBNC) state in order to promote their survival. Additionally, it has been suggested that biofilm-associated L. pneumophila may spend only a minor portion of its lifespan inside amoebae, using them solely as a host for replication (Declerck et al., 2007b). Forty-eight hours after inoculating A. castellanii, the whole amoeba population was heavily infected. Soon after, amoebic hosts lysed, releasing high numbers of L. pneumophila into the bulk phase of the reactor.
Figure 4. Evolution of biofilm-associated non-Legionella and L. pneumophila. Legionella pneumophila and A. castellanii were inoculated in the rotating annular reactor (RAR) on days 5 and 35 respectively. The RAR was operating in flow-through mode and was fed with sterile dechlorinated tap water. Time-point samplings where L. pneumophila concentrations were below the quantification limit of both the culture and real-time PCR method are indicated by zero values. Reprinted from Declerck and colleagues (2007b).
Download figure to PowerPoint
To date, a lively discussion is ongoing concerning the survival and replication of L. pneumophila in biofilms. Do biofilm-associated L. pneumophila bacteria really require protozoan hosts to grow and increase their cell numbers or are they able to actively replicate in the community and thus obtain their nutrients from the biofilm environment? This discussion continues due to the lack of trustful data from pilot-scale or on-site experiments, using naturally occurring consortia. The problem of working with natural biofilm communities is that biofilms are microbial hotspots and the pathway of trophic interactions is immensely complex. In my opinion, both extracellular and intracellular replication can occur, depending on the environmental circumstances. Apparently, the intracellular pathway remains the most important one as demonstrated by Temmerman and colleagues (2006). Also, numerous benefits are related to intracellular replication in protozoans (reviewed by Molmeret et al., 2005), which have not yet been demonstrated for extracellular replication.
As a member of multispecies biofilms, the survival of L. pneumophila depends not only on its partners but also on its opponents. Toze and colleagues (1990) found that up to 32% of heterotrophic bacteria, e.g., Pseudomonas spp. and Aeromonas spp., isolated from chlorinated drinking water, inhibit growth of Legionella spp. Possible mechanisms of inhibition have not yet been unravelled. However, this effect could be partially attributed to bactericidal proteins, collectively known as bacteriocins that are continuously produced by both Gram-positive and Gram-negative bacteria. Guerrieri and colleagues (2008) performed a study to investigate the effect of bacterial interference on biofilm counts of L. pneumophila. They showed that certain bacteria are able to influence L. pneumophila in its survival and persistence in biofilms through production of bacteriocins. Pseudomonas fluorescens SSD, the best bacteriocin producer, inhibited both the formation and the stability of L. pneumophila biofilms more effectively than the other tested bacterial strains. Pseudomonas aeruginosa, inactive against L. pneumophila, showed a synergistic effect when tested with P. fluorescens. These observations confirm that microbial interference displays a variety of mechanisms, which are only partially known and may combine one with the other. Also, bacteria can influence L. pneumophila replication in the intracellular host environment. In a recent study, we observed that P. aeruginosa and E. coli, when present together with L. pneumophila in the same replication vacuole, significantly decreased Legionella growth (Declerck et al., 2005). This may be because both non-Legionella bacteria can consume important resources provided by the amoeba host, which are normally used by replicating L. pneumophila. Another explanation could also be that, although we did not detect any influence of P. aeruginosa or E. coli on the growth of L. pneumophila when they were incubated together in saline, it is possible that the ‘intracellular amoeba environment’ may induce the production of certain Legionella growth-inhibiting components by the other bacterial species.
Detachment of bacteria from biofilms followed by a subsequent dispersal in the environment is an integral part of the dynamic nature of life in surface-associated microbial communities (Stoodley et al., 2001; Wilson et al., 2004). Biofilm cells may detach and disperse either by shedding of daughter cells from actively growing microbes, as a result of nutrient levels or quorum sensing, or by shearing of biofilm aggregates because of flow effects (Donlan, 2002). Detachment caused by flow effects can be divided into two processes, namely: erosion and sloughing. Erosion is the continual detachment of single cells and small portions of the biofilm, whereas sloughing is induced by unexpected events and is characterized by the rapid, massive loss of biofilm (Characklis, 1990). Storey and colleagues (2004) demonstrated a continual erosion of L. pneumophila from biofilms, present in a pilot-scale water distribution system, under laminar flow conditions. This process turned into sloughing under turbulent conditions, with more than 90% of the L. pneumophila detached from the substratum and mobilized into the adjacent bulk phase. Fluorescent in situ hybridization was used to visualize Legionella bacteria within the detached biofilm and cell numbers ranged from a few in small cell clusters to as many as 100 in larger aggregates. Problems may arise when high numbers of L. pneumophila become detached from the biofilm and mobilize into the bulk water phase. First, such an event can cause a significant decrease in the species richness of the resident planktonic Legionella spp. community (Wéry et al., 2008). Second, detached pathogenic bacteria like L. pneumophila have the potential to reach humans as an infective dose (Storey et al., 2004). Thus, a thorough investigation of bacterial detachment is of fundamental importance to the dissemination of infection and contamination in both clinical and public health settings.
Thanks to its intracellular lifestyle in protozoans, L. pneumophila has some supplementary strategies to become detached from the biofilm community. First, infected amoebae such as Naegleria sp. which possess an additional flagellate stage will swim away from the biofilm when environmental conditions become unfavourable, e.g., nutrient deficiencies. Second, infected amoebae have been shown to release vesicles of respirable size (< 5 μm), each containing 20–200 L. pneumophila (Berk et al., 1998). Such vesicles easily get incorporated into aerosols, which may lead to the transportation of numerous L. pneumophila over considerable distances.