Many bacteria form biofilms, complex, surface-associated multicellular communities that consist of cells attached to a surface, and cohered to each other. A flurry of research activity over the last two decades has led to a more detailed and nuanced understanding of the process of biofilm formation for diverse microorganisms (Haussler and Parsek, 2010). Bacterial surface features comprised of proteins, polysaccharides and sometimes externalized nucleic acid, can play important roles in the cellular interactions underlying biofilm formation. In a variety of bacteria, large proteins that are exported to the cell surface can impact cell-to-surface and cell-to-cell interactions (Yousef and Espinosa-Urgel, 2007). Mature biofilm structures result from the interplay between functions that mediate attachment to surfaces, those that direct production and integrity of the biofilm matrix, and those that enable cell-to-cell interactions.
In this issue of Molecular Microbiology, Martínez-Gil and colleagues provide intriguing evidence that the second largest protein (6310 aa) encoded within the Pseudomonas putida KT2440 genome, designated LapF (large adhesion protein), is localized to the cell surface, where it is required for the cell-to-cell interactions that mediate cohesion within P. putida biofilms (Martínez-Gil et al., 2010). Their findings extend initial work on the Pseudomonas fluorescens LapA protein, which mediates cell-to-surface interactions; the P. putida LapA homologue is the largest protein (8682 aa) produced by this microorganism (Espinosa-Urgel et al., 2000; Hinsa et al., 2003).
The first indication of the role of lapF in surface colonization came from a genetic screen for transposon mutants with attachment deficiencies on corn seeds, in which a transposon insertion in PP_0806 (mus-20, mutants unattached to seeds) resulted in significantly reduced binding (Espinosa-Urgel et al., 2000). The same screen identified a mutation in PP_0168 (mus-24), which was later also identified by Hinsa et al. as the LapA protein in P. fluorescens, and is conserved among several pseudomonads (Hinsa et al., 2003). LapA was found to be required for stable association with abiotic surfaces and soils and has received significant attention. Several lines of evidence suggest that LapA is required for pseudomonads to transition from reversible surface binding via their poles, to more stable irreversible binding, with cells contacting the surface along their sides (Hinsa et al., 2003; Hinsa and O'Toole, 2006). Studies on LapF were not immediately pursued, perhaps because initial findings suggested that it was dispensable for abiotic surface binding (Espinosa-Urgel et al., 2000). The current study has intensely investigated the function of LapF and elegantly revealed that this protein is not required for single cell attachment to surfaces, but rather for the cell-to-cell interactions that lead to formation of mature, three-dimensionally complex biofilms (Martínez-Gil et al., 2010).
LapA and LapF share a significant set of overall characteristics without having substantial sequence identity. They are both strikingly large, cell surface-localized secreted proteins that contain expansive repeat domains (Fig. 1A). LapF has a single large repeat domain, whereas LapA has two distinct repeat domains. Their C-terminal regions contain predicted von Willebrand factor type A domains (originally identified in eukaryotes, where they function in cell-to-cell interactions), calcium-binding motifs and type I secretion signals. The lapA and lapF genes are well separated from each other in the P. putida genome, but both are closely linked to genes encoding ABC transporter functions (Hinsa et al., 2003; Martínez-Gil et al., 2010) (Fig. 1A). For LapA, these transporter genes are required for secretion (Hinsa et al., 2003). These structural and genetic similarities place the Lap proteins among a functional group of large, repeat-rich, cell surface proteins (Yousef and Espinosa-Urgel, 2007), including Bap (biofilm-associated protein) from Staphylococcus aureus (Cucarella et al., 2001) and FHA (filamentous haemagglutinin) from Bordetella species (Locht et al., 1993), which are often involved in bacteria-to-host cell and bacteria-to-surface interactions.
Martínez-Gil and colleagues provided new insights into Lap protein function and biofilm formation through high-resolution analysis of surface interactions for a lapF mutant (Martínez-Gil et al., 2010). The lapF mutant (mus-20) phenotype is unimpressive by a crude measure of biofilm formation in static cultures (Espinosa-Urgel et al., 2000), manifesting a modest decrease in adherent biomass and a significant, but limited deficiency in competitive colonization of plant roots (Martínez-Gil et al., 2010). Confocal laser scanning microscopy of lapF mutant binding to glass coverslips and fluorescence microscopy of root colonization revealed normal attachment, but significant deficiencies in subsequent biofilm formation. Under flowing conditions, the lapF mutant phenotype is significantly more dramatic, with lapF cells failing to coalesce into microcolonies. Early-stage mixed biofilms contain homogeneous wild-type and heterogeneous microcolonies, but the lapF mutant does not form microcolonies alone, and more mature biofilms are dominated by wild-type cells interspersed with individual lapF cells. These results suggest that lapF mutants cannot participate in the cell-to-cell interactions that lead to dense, multi-layer biofilms but can be passively trapped in the growing biofilm.
LapF associated with the cell surface would be ideally located to mediate interactions between cells. Immunofluorescence with anti-LapF antibodies and non-permeabilized cells revealed LapF within cell clusters and localized as puncta between cells. Exogenous addition of this same antibody diminished biofilm formation by P. putida, reducing multicellular aggregates, providing further evidence of the extracellular location and function of LapF. Taken together, these data provide compelling evidence that it is the externalized and cell-associated form of LapF that promotes multicellular cell-to-cell interactions.
In the same report, the authors examine the regulation of lapF. In biofilms growing on both abiotic and plant surfaces, a lapF::gfp transcriptional fusion is only detectable within cells that are enmeshed in microcolonies and in mature biofilms. Monitoring a lapF–lacZ fusion in liquid cultures reveals expression in late stages of growth. Remarkably, mutants for the stress-responsive sigma factor gene rpoS are found to completely abolish expression of lapF. In Escherichia coli, RpoS can have significant effects on biofilm formation (Ferrieres et al., 2009), although the P. putida rpoS mutants were not tested for biofilm formation in the current study. Therefore, at least one relevant regulator and the overall pattern of lapF gene expression are highly congruent with the observations that LapF protein is only detectable after initial surface colonization and within expanding biofilms. The requirement for LapF is also nutritionally conditional, and the lapF mutant deficiency is only observed in minimal medium or dilute complex medium, i.e. not in full strength complex medium such as LB. Although not further explored in the current study, the RpoS dependence of lapF expression would be a likely mechanism for this nutritional control.
The findings reported by Martínez-Gil et al., combined with the prior work on lapA, suggest an attractive model for how these large surface proteins act in an orchestrated sequence, with LapA driving the transition from polar attachment to irreversible attachment, and then LapF, specifically produced during early surface interactions, assuming the leading role in microcolony growth (Fig. 1B). However, Gjermansen et al. (2009) have reported findings suggesting that LapA might also function at later stages of biofilm maturation. Likewise, numerous images of growing lapF mutant biofilms in the current study (Martínez-Gil et al., 2010) reveal a large number of cells attached by their poles, consistent with a deficiency in the polar to lengthwise surface association. These results hint that the overlap between the LapA and LapF proteins may be more extensive than the strict linear model proposed by Martínez-Gil and colleagues.
Although the current study demonstrates that LapF on the exterior of P. putida cells is required for normal cell-to-cell interactions within the growing biofilm, the mechanisms that underlie the functions of LapF and those of related proteins remain mysterious. Proteins with highly reiterative amino acid sequences are frequently involved in cell-to-cell interactions. Some of the large, repeat-rich surface proteins structurally similar to the Lap proteins, such as FHA from Bordetella pertussis, are known to form filaments (Kajava et al., 2001). It is unclear, however, whether LapF or LapA adopt filamentous structures and, given their weak primary sequence similarity, it is possible that the two proteins form structures distinct from each other. Furthermore, the question of how their repetitive structures and presumptive calcium-binding domains influence the adhesive and cohesive processes for which they are required remains to be addressed. Small proteins on the surface of rhizobial cells, known as Rap adhesins (Rhizobiumadhering proteins) and required for biofilm formation, also appear to bind calcium, suggesting a possible common mechanism (Downie, 2010).
Curiously, Pseudomonas aeruginosa, the most thoroughly studied pseudomonad with respect to biofilm formation does not have lapF or lapA orthologues but it does have large surface proteins that function in biofilm formation. Recently, the P. aeruginosa CdrA (c-di-GMP-regulated protein A), a 2154 aa structural analogue of the FHA-type, cell wall-associated proteins, was reported to stabilize biofilm structures through physical association with the Psl extracellular polysaccharide in the biofilm matrix (Borlee et al., 2010). It is plausible that LapF plays a similar role in the P. putida biofilm, but association with extracellular polysaccharide has not been reported.
Mutations in the ABC transporter homologues, LapB, LapC and LapE, prevent LapA externalization and phenocopy the LapA mutant in P. fluorescens (Hinsa et al., 2003). In P. putida externalization of LapF is abolished by the mus-20 transposon insertion just upstream of several putative type I secretion motifs (GGXGXD), also consistent with secretion via an ABC transporter (Martínez-Gil et al., 2010). Coincidentally, the lapF gene is flanked by ABC transporter homologues, LapH, LapI and LapJ (Fig. 1A), but there is as yet no direct evidence that these proteins secrete LapF.
In P. fluorescens, secretion of LapA is regulated by LapD, a protein with degenerate GGDEF and EAL motifs and that responds to intracellular cyclic diguanosine monophosphate (c-di-GMP) levels to control adhesion (Newell et al., 2009). The lapF gene and the genes for the predicted LapHIJ transporter are located close to a gene encoding a GGDEF-type protein (PP_0798, see Fig. 1A). It seems plausible that c-di-GMP levels also influence lapF function, but a role for PP_0798 has not been reported.
In a broader sense, the genetic variability and stability of large repetitive proteins such as LapF and LapA is of great interest. How constant is the size of each repeat domain among natural populations? It would seem likely that such reiterated sequences would be prone to frequent rearrangements, particularly deletions. As noted above, neither LapA nor LapF has strong homologues in P. aeruginosa. LapF is well conserved among the limited number of different P. putida strains examined, but the length of the protein seems highly variable, consistent with a genetically dynamic repeat structure. LapA is conserved between P. putida and P. fluorescens, species that belong to two major subgroups within the Pseudomonas genus (Yamamoto et al., 2000). However, the P. putida protein has 8682 aa, whereas the P. fluorescens protein has 4920 aa. At least one source of the length variation is the larger size of the P. putida repeat sequence itself. These observations suggest that the Lap proteins and, by extension, similar large repetitive surface proteins, are rapidly evolving via these repeats. More comprehensive population-level analyses will probably reveal interesting evolutionary trends for this remarkable group of proteins.
Many other factors contribute to biofilm formation and attachment in pseudomonads, including exopolysaccharides, nucleic acids, motility functions, dispersal factors and surfactants (Toutain et al., 2004). Martínez-Gil et al. (2010) propose an elegant, stepwise model by which a relay of the LapA and LapF proteins functions to facilitate the early stages of cell-to-surface and then cell-to-cell interactions that can forge the foundation for growing biofilms. The mechanism(s) by which the Lap proteins integrate with the other important biofilm determinants to co-ordinate sessile growth should be a fruitful area for future investigations.