OUP user menu

An integrated view of biofilm formation in rhizobia

Luciana V. Rinaudi, Walter Giordano
DOI: http://dx.doi.org/10.1111/j.1574-6968.2009.01840.x 1-11 First published online: 1 March 2010


Biofilms are bacterial communities enclosed within an extracellular matrix of polysaccharides produced by the bacteria, which adhere to a living or an inert macrosurface. In nature, biofilms constitute a protected growth modality allowing bacteria to survive in hostile environments. Studies of environmental isolates have revealed a highly ordered, three-dimensional organization of the extracellular matrix, which has important implications for biofilm physiology. The zone of soil immediately surrounding a plant root where complex biological and ecological processes occur, termed rhizosphere, forms an environment that fulfills the requirements for biofilm formation, including sufficient moisture and supply of nutrients, which are provided by the plant. Biofilm formation on plants appears to be associated with symbiotic and pathogenic responses, but it is unclear how plants regulate the association. Biofilms function as structures resistant against stress factors such as desiccation, UV radiation, predation, and antibiosis, which help create protective niches for rhizobia. However, the role of biofilms in rhizobial–legume symbiosis remains to be clarified. Here, the mechanisms involved in bacterial biofilm formation and attachment on plant roots, and the relation of these mechanisms to rhizobial function and survival are reviewed.

  • rhizobia
  • biofilms
  • legume roots
  • soil
  • rhizosphere
  • exopolysaccharides


The enriched environment around plant roots allows establishment of interactions between soil bacteria and the roots. These relationships can be beneficial, pathogenic, parasitic, or saprophytic, and exert important effects on plant development and productivity. Microorganisms colonize mineral soil particles as well as plant roots. They may cause plant diseases or, in contrast, produce a wide range of beneficial effects, including biocontrol against pathogens, plant growth promotion through nitrogen fixation, phytohormone production, and mobilization of nutrients.

When environmental nitrogen is limited, soil bacteria known as rhizobia interact with roots of leguminous plants to produce symbiotic nodules, inside which atmospheric nitrogen is reduced to ammonium for use by the plant, while the bacteria receive carbohydrates from the plant in a protected environment. Establishment of this symbiosis relies on an exchange of signals between the legume and the rhizobia. Therefore, a particular rhizobia species nodulates a particular group of related legume species. The plant initiates the ‘molecular dialogue’ by producing and secreting flavonoid compounds into the rhizosphere. The bacteria sense these compounds and respond by inducing the expression of nod genes and the production of Nod factors.

During rhizobia–legume symbiosis, bacteria usually invade and colonize roots through structures called ‘infection threads.’ Various types of surface polysaccharides, including exopolysaccharides (EPS), lipopolysaccharides, and capsular polysaccharides, play important roles during the infection and formation of active nodules (Fraysse et al., 2003; Skorupska et al., 2006). Mutants deficient in the production of these polysaccharides fail to induce infection thread formation or to develop effective nodules (Hirsch, 1999). Cyclic glucans, present in bacterial periplasm and secreted into the culture medium, are essential for osmoadaptation of the bacteria, and may play a role in the symbiosis (Zorreguieta et al., 1990).

Bacterial surface components, particularly exopolysaccharides, flagella, and lipopolysaccharides, in combination with the presence of bacterial functional signals, are crucial for the formation of biofilms in all species studied so far. Biofilms are defined as bacterial communities surrounded by a self-produced polymeric matrix, and reversibly attached to an inert or a biotic surface (Costerton et al., 1995). After attachment to the surface, the bacteria multiply, and the communities acquire a three-dimensional structure, in some cases permeated by channels. The channels act as a ‘circulatory system,’ allowing the bacteria to exchange water, nutrients, enzymes, and signals, dispose of potentially toxic metabolites, and enhance metabolic cooperativity (Costerton et al., 1995; Stanley & Lazazzera, 2004). However, it is difficult to draw a clear line between simple aggregates vs. firmly attached biofilms on a surface. It seems that the term ‘biofilm’ is now applied to what were previously described as bacterial aggregation, microcolony, agglutination, and flocculation.

Biofilm composition differs depending on the system. The major components are typically water and bacterial cells. The next most important component is a polysaccharide matrix composed of exopolysaccharides (Sutherland, 2001), which provides a physical barrier against diffusion of compounds such as antibiotics and defense substances from the host, and protection against environmental stress factors such as UV radiation, pH changes, osmotic stress, and desiccation (Flemming, 1993; Gilbert et al., 1997). In Agrobacterium tumefaciens, a plant pathogen that persists as surface-associated populations on plants or soil particles, cellulose overproduction resulted in increased biofilm formation on roots (Matthysse et al., 2005). Minor components include macromolecules such as proteins, DNA, and various products released by lysis (Branda et al., 2005), which also affect the properties of biofilms as a whole.

Bacterial biofilms are widely distributed, and play important roles in many environments. Environments faced by soil rhizobia range from a rhizosphere rich in nutrients and root exudates, to soils deficient in nitrogen, phosphates, water, and nutrients. Numerous microbial species, including rhizobia, form microcolonies or biofilms when they colonize roots. Available data on surface attachment and/or biofilm formation by rhizobia are summarized in Table 1.

View this table:
Table 1

Rhizobia known to attach and colonize different inert surfaces and/or roots of legume and nonlegume plant species

Rhizobial speciesRole described in biofilm formationSource
M. huakiiQuorum sensingWang (2004)
M. tianshanenseExopolysaccharidesWang (2008)
S. melilotiNutritional and environmental conditionsRinaudi (2006)
Succinoglycan and flagellaFujishige (2006)
ExoR and ExoS–ChvI two-component systemWells (2007)
Nod factorsFujishige (2008)
Quorum sensing and EPS IIRinaudi & González (2009)
B. elkaniiBradyrhizobial fungi associationsSeneviratne & Jayasinghearachchi (2003)
Nitrogen fixation in biofilms, improvement of soil fertilityJayasinghearachchi & Sereviratne (2004a, 2004b, 2006); Sereviratne & Jayasinghearachchi (2005)
Bradyrizobium spp.
B. japonicumRhicadhesinDardanelli (2003)
Soybean lectin and exopolysaccharidesPérez-Giménez (2009)
R. leguminosarumAcidic EPS and PrsD–PrsE type I secretion systemRusso (2006)
RapA1 proteinMongiardini (2008)
R. etliGlucomannanWilliams (2008)
Quorum sensingEdwards (2009)
Swarming motilityVerstraeten (2008)
Rhizobium sp.Attachment to nonhost plants independent of EPSSantaella (2008)
  • EPS, exopolysaccharides.

Biofilm formation allows non-spore-forming soil bacteria to colonize surrounding habitat, and to survive common environmental stresses such as desiccation and nutrient limitation. The biofilm mode of life is often crucial for survival of bacteria, as well as for establishment of symbiosis with the legume host. Biofilm formation is believed to occur as a sequential developmental process, culminating in the establishment of these bacterial communities (Fig. 1). Still, an integrated view of biofilm formation in rhizobia has not been presented. In order to organize available information in this review, data are summarized for each of the four major genera: Mesorhizobium, Sinorhizobium, Bradyrhizobium, and Rhizobium.

Figure 1

General model of biofilm formation. Environmental signals, and growth under nutrient limitation, trigger biofilm formation. Flagella are required for approaching and moving across the surface. In some cases, outer membrane proteins such as calcium-binding proteins and adhesins mediate the initial steps of attachment. After formation of microcolonies, the production of AHLs is required for the formation of a mature biofilm. Exopolysaccharides (EPS) (and Nod factors in the case of Sinorhizobium meliloti) provide architectural form and stabilize the three-dimensional structure of biofilms. Dispersion of the biofilm allows bacteria to colonize other surfaces or substrates.


Biofilm formation has been reported in two Mesorhizobium species, Mesorhizobium huakuii and Mesorhizobium tianshanense (Wang et al., 2004, 2008), which, like all members of this genus, show a growth rate intermediate between those described for Rhizobium and Bradyrhizobium.

Quorum sensing is a mechanism allowing bacteria to sense population density and regulate gene expression, leading to activation of specific phenotypes in the population. The process depends on the accumulation in the environment of a signaling molecule termed autoinducer. Many Gram-negative bacteria use N-acylhomoserine lactones (AHLs) as signal molecules, and some have been reported to use other fatty acid derivatives such as 3-hydroxypalmitic acid methyl ester and cis-unsaturated fatty acids. In contrast, many Gram-positive bacteria use amino acids or modified peptides as signal molecules. Both Gram-positive and Gram-negative bacteria use isomers of methyl-2,3,3,4-tetrahydroxytetrahydrofuran (the AI-2 autoinducer) as signals. Signal molecules belonging to other structural classes (indole and its derivatives, quinolones, and (S)-3-hydroxytridecan-4-one) have also been described (Ryan & Dow, 2008). The production of these autoinducers has been described for M. huakuii, which establishes a symbiotic relationship with Chinese milk vetch, Astragalus sinicus (Zhu et al., 2003). Overexpression of the A. tumefaciens quorum regulator TraR in M. huakuii strain Mh93 interfered with the endogenous quorum-sensing system, probably because of competitive binding of TraR proteins to rhizobia AHLs (Wang et al., 2004). A strain overexpressing TraR formed thinner biofilms than the control strain, suggesting that quorum sensing positively regulates biofilm formation in M. huakuii (Wang et al., 2004).

Production of AHLs has also been described for M. tianshanense, a nitrogen-fixing symbiont for at least eight plant species, including species of Glycyrrhiza (licorice), where the MrtR–MrtI quorum-sensing system plays a role in symbiosis (Zheng et al., 2006). Nodulation assays on Glycyrrhiza uralensis with wild-type and quorum-sensing-deficient mutant strains of M. tianshanense showed that mrtI and mrtR mutants were unable to develop nodules on legume roots. This may have been due to poor bacterial attachment by the mutants, because the mrtI strain showed a 60% reduction of root hair attachment efficiency (Zheng et al., 2006).

Exopolysaccharides were recently shown to be involved in biofilm formation in M. tianshanense (Wang et al., 2008). Sequence analysis of nonmucoid strains showed that mutations were located in two gene clusters: the first is similar to pssNOPT of Rhizobium leguminosarum bv. viciae (Young et al., 2006), and the second is similar to the exo5 gene in R. leguminosarum bv. trifolii (Laus et al., 2004). All these genes are conserved among rhizobia and are involved in exopolysaccharide polymerization and translocation (Skorupska et al., 2006). The mtpABCDE genes responsible for exopolysaccharide production in M. tianshanense are regulated by the two-component histidine kinase regulatory system MtpS–MtpR (Wang et al., 2008). The exopolysaccharide-deficient strains mtpC, mtpR, and mtpE failed to nodulate G. uralensis and formed a biofilm with smaller biomass compared with the wild type in the borosilicate attachment assay, suggesting that exopolysaccharides are essential for biofilm formation (Wang et al., 2008).

Quorum-sensing mechanisms control numerous functions in rhizobia, including exopolysaccharide production (Marketon et al., 2003; Hoang et al., 2004; Glenn et al., 2007), motility and nitrogen fixation (Hoang et al., 2004, 2008), and nodulation (Cubo et al., 1992; Rodelas et al., 1999; Daniels et al., 2002; Hoang et al., 2004), all of which are related to symbiosis. The studies cited in this section show clearly that Mesorhizobium is one of the genera of bacteria in which quorum sensing plays an important role in biofilm formation, attachment, colonization, and nodulation of legumes.


Since biofilm formation was first reported in Sinorhizobium meliloti (Fujishige et al., 2005), soil microbiologists have been interested in rhizobial regulatory systems in this species, and conditions for analyzing its ability to produce biofilms. Biofilm formation is clearly an important feature of this species' symbiotic ability, and its resistance to adverse environmental conditions. Biofilm production on abiotic surfaces (glass or plastic) has been used as a model for characterization of bacterial aggregation and attachment (O'Toole & Kolter, 1998b). Use of this approach in S. meliloti has helped clarify the roles of nutritional and environmental conditions (Rinaudi et al., 2006), exopolysaccharides and flagella (Fujishige et al., 2006), ExoR with the ExoS–ChvI two-component system (Wells et al., 2007), nod genes (Fujishige et al., 2008), and regulation of exopolysaccharide biosynthesis (Rinaudi & González, 2009) in biofilm formation.

The sequenced strain of S. meliloti Rm1021 displays reduced biofilm formation on the microplate assay when grown in a rich medium compared with minimal medium (Fujishige et al., 2005). A nutritionally limited environment promotes the transition from a planktonic to a sessile mode of life. Biofilm formation may therefore represent a strategy for survival of bacteria in nutritionally limited environments, because colonization of surfaces provides certain advantages, for example increased capture of nutrients that can be absorbed from the medium (Wimpenny & Colasanti, 1997). In contrast, nutrient abundance in the medium seems to favor biofilm formation in Pseudomonas (O'Toole & Kolter, 1998b; Yousef-Coronado et al., 2008), possibly by increasing bacterial population size and accumulation of autoinducers, which promote biofilm formation. In view of previous findings that the nutrient content of the growth medium regulates the development of biofilms by Pseudomonas species (O'Toole & Kolter, 1998a, b), the effects of various nutrients and environmental conditions on the biofilm formation ability of S. meliloti were tested (Rinaudi et al., 2006). The concentrations of sucrose, phosphate, and calcium were positively correlated with biofilm formation, whereas extreme temperatures and pH values had a negative effect. These findings support the hypothesis that biofilm formation promotes the survival of non-spore-forming rhizobia in soil in the absence of a legume host.

The key regulatory pathways in S. meliloti biofilm formation have been identified. The exoR and exoSchvI two-component system controls many phenotypes, including biofilm formation. Wells (2007) showed that this system affects succinoglycan production, prototrophy, nitrogen fixation, and motility, and also regulates attachment to abiotic surfaces. The exoR95 and exoS96 mutants showed a considerably increased biofilm formation, compared with the wild-type or the other strains tested.

Rhizobium nod genes, and their products, Nod factors, are essential for the development of nitrogen-fixing nodules on legume roots (Lerouge et al., 1990). Microscopic analysis revealed that Nod factors are critical for the establishment of a mature rhizobial biofilm (Fujishige et al., 2008). This is a new function for Nod factors, and is distinct from their established role as a morphogen-inducing legume nodule development. The dual functions of Nod factors, as structural components in biofilms and independently as precursors of host-specific morphogens, imply the existence of two different sets of control mechanisms, one dependent on flavonoids (plant-derived inducers of nod genes in S. meliloti) and the other independent of flavonoids, which regulate Nod factor production (Fujishige et al., 2008).

Bacteria have various mechanisms for movement, including flagellar swimming, swarming, twitching, and gliding motility. In Pseudomonas aeruginosa and many other bacteria, flagella are necessary for the initial surface attachment for biofilm establishment (O'Toole & Kolter, 1998a). Sinorhizobium meliloti in soil has 5–10 flagella per cell (Götz et al., 1982), which are composed of two related (87% identical) flagellins encoded by the closely linked, but separately transcribed genes, flaA and flaB (Pleier & Schmitt, 1989). In rhizobia, flagellar motility allows access to attachment infection sites on the plant. Sinorhizobium meliloti Fla mutants showed a competitive disadvantage compared with the motile, wild-type strains (Ames & Bergman, 1981). Flagellar motility also appears to be involved in biofilm maturation, because flagellar mutants of S. meliloti showed a reduced biofilm formation ability as well as delayed nodule formation (Fujishige et al., 2006).

The biofilm matrix, composed mainly of exopolysaccharides, physically connects cells, and confers many key biofilm features, including resistance to desiccation and other environmental stresses. Depending on the environmental phosphate concentration, S. meliloti produces two different exopolysaccharides, both able to promote symbiosis: succinoglycan (also known as EPS I) and galactoglucan (or EPS II). Low-phosphate (0.1–10 μM) conditions typical of soils (Bieleski, 1973) stimulate the production of EPS II, whereas high-phosphate conditions (up to 100 mM), as found in nodules (Israel, 1987), block EPS II synthesis and induce the production of EPS I.

EPS I, one of the best-understood, symbiotically important exopolysaccharides, is required for invasion of alfalfa roots by S. meliloti Rm1021. EPS I is a polymer of repeating octasaccharide subunits (seven glucose and one galactose), bearing succinyl, acetyl, and pyruvyl substituents (Reuber & Walker, 1993). Mutations affecting EPS I biosynthesis result in a variety of developmental abnormalities during nodule formation, including delayed root hair curling, defective or aborted infection threads, and empty nodules with no bacteria or bacteroids, suggesting a signaling function for EPS I (Finan et al., 1985; Fraysse et al., 2003). Sinorhizobium meliloti EPS I is also required for biofilm formation, because an exoY mutant formed immature biofilms, whereas overproduction of EPS I led to the formation of thicker, but less stable biofilms. Sinorhizobium meliloti exopolysaccharide mutants, in general, display reduced biofilm phenotypes correlated with nodulation ability (Fujishige et al., 2006).

EPS II, another exopolysaccharide produced by S. meliloti, is composed of alternating glucose and galactose residues that are acetylated and pyruvylated, respectively (Spaink, 2000). Under nonstarvation conditions, wild-type laboratory S. meliloti Rm1021 produces detectable quantities of succinoglycan, but not EPS II. The production of EPS II was observed under low-phosphate conditions (Zhan et al., 1991; Mendrygal & González, 2000) and in a mucR mutant (Keller et al., 1995). The presence of a functional expR ORF on a plasmid, or in the genome, is sufficient to promote the production of symbiotically active EPS II, for example in strain Rm8530, which has an intact expR and is therefore termed expR+ (Glazebrook & Walker, 1989). Under laboratory conditions, S. meliloti can form three distinct types of biofilms, termed ‘flat,’‘structured,’ and ‘organized.’ EPS II-producing strain Rm8530, which has a mucoid phenotype, displays a highly structured architectural biofilm, in contrast to the unstructured one formed by non-EPS II-producing strain 1021. In experiments with Medicago sativa (alfalfa), strain Rm8530 expR+ formed biofilms covering the entire surface of the root, including root hairs, whereas strain Rm1021 formed clusters of cells adhering mainly to the main root (Rinaudi & González, 2009).

Exopolysaccharides determine living conditions for microorganisms in biofilms, because they affect the porosity, density, water content, charge, hydrophobicity, and mechanical stability of biofilms (Flemming & Wingender, 2002). In S. meliloti, MucR controls exopolysaccharide production. To clarify the relationship between exopolysaccharide synthesis and biofilm formation, mucR expression was studied using transcriptional fusion to lacZ. The results indicated that mucR does not respond to changes in environmental conditions, and does not play an important role in biofilm formation (Rinaudi et al., 2009). Biofilm formation in the Rm1021 strain is limited, and does not appear to be mediated by the presence of exopolysaccharides. In the Rm8530 expR+ strain, biofilm formation is controlled by the ExpR/Sin quorum-sensing system, through production of EPS II. Levels of biofilm formation and phenotype observed by confocal microscopy in strain Rm1021 mucR are similar to those of wild-type Rm1021, suggesting that the low-molecular-weight fraction of EPS II could control the formation of biofilms both in vivo and in vitro (Rinaudi & González, 2009).

Microscopic examination of S. meliloti cells within curled root hairs revealed small biofilm-type aggregates that could provide inocula for root invasion, and rhizobial cells migrated down infection threads toward the root interior as biofilm-like filaments (Ramey et al., 2004). These authors also showed that Agrobacterium and rhizobia can form dense, structurally complex biofilms on root surfaces.

As explained in the Introduction, it is very difficult to differentiate between structures now known as biofilms to what were previously described as bacterial aggregation, microcolony, agglutination, and flocculation. In this context, the agglutination to glass and the flocculation of R. leguminosarum observed more than two decades ago could be classic biofilms (Smit et al., 1987). Likewise, fluorescence protein-expressing S. meliloti attached to roots and forming infection threads, as documented by Gage (1996), and later by our group (Giordano et al., 2002), are similar to structures referred to as biofilms in later reports by Ramey (2004). Microphotographs of the outer surface of white sweetclover (Melilotus alba) roots show that both DsRed-labeled strain Rm1021 and the green fluorescent protein (GFP)-labeled strain GMI6032 of S. meliloti attached to the root surface, forming aggregates and infection threads that contained only DsRed-labeled cells (Fig. 2a). Infection threads and small aggregates that contained either GFP- or DsRed-expressing rhizobia were observed (Fig. 2b). Where the two strains overlap, fluorescence is yellow. In infection threads containing both, GFP- and DsRed-expressing rhizobia were not randomly intermixed (Fig. 2c).

Figure 2

White sweetclover (Melilotus alba) roots coinoculated with DsRed-labeled Rm1021 and GFP-labeled GMI6032 strains of Sinorhizobium meliloti. (a) Infection threads containing only Rm1021 (red, thin arrow), but GMI6032 cells (green, thick arrow) are found attached on the root surface. (b) Infection threads containing either Rm1021 (red, thin arrow) or GMI6032 (green, thick arrow). (c) Infection threads containing Rm1021 (red, thin arrow) and GMI6032 (green, thick arrow) and both strains overlapping (yellow, asterisks). Scale bars=100 μm. Reproduced from Giordano (2002) with permission.

Surface components are involved in the early stages of nodulation elicited by rhizobia, and are critical for biofilm formation. The change from a planktonic to a biofilm lifestyle in S. meliloti is mediated by numerous environmental signals (Rinaudi et al., 2006). Biofilms are the most common life strategy for bacteria in natural environments, including the rhizosphere, as typified by S. meliloti.


Mycelial colonization and biofilm formation by bradyrhizobia with common soil fungi have been reported (Seneviratne & Jayasinghearachchi, 2003). Such biofilms showed nitrogenase activity (Jayasinghearachchi & Sereviratne, 2004a, b; Sereviratne & Jayasinghearachchi, 2005) and enhanced availability of nitrogen and phosphate when inoculated to soil (Sereviratne & Jayasinghearachchi, 2005).

Heavy mycelial colonization by Bradyrhizobium elkanii SEMIA 5019 was observed in Pleurotus ostreatus-bradyrhizobial biofilms 16 days postincubation (Jayasinghearachchi & Sereviratne, 2004a). Nitrogenase activity was detected in the biofilm, but not in the fungus or Bradyrhizobium alone. This study proved that symbiotic bacteria within biofilms can fix nitrogen, and that the fungi are not responsible for nitrogenase activity, as was claimed previously. Similar findings were reported in a B. elkanii SEMIA 5019/Penicillium spp. system. Shoot, root, and nodule weights of soybean plants treated with a biofilm inoculum were significantly higher than those of control plants under greenhouse conditions (Jayasinghearachchi & Sereviratne, 2004b). Biofilm-inoculated plants also showed significantly higher shoot and root nitrogen accumulation. Therefore, use of nitrogen-fixing biofilms as inoculants may promote soil nitrogen fertility and plant growth. Mycelial growth in the rhizosphere may facilitate the movement of rhizobia, which normally show reduced vertical mobility (McDermott & Graham, 1989), because plants inoculated with a bradyrhizobial-fungal biofilm displayed better nodule distribution than conventionally inoculated plants (Jayasinghearachchi & Sereviratne, 2004b). Application of the B. elkanii SEMIA 5019–Penicillium spp. mix enhanced phosphate mineralization, in addition to increasing nitrogen availability in soil. An application of this microbial association as a biofilm inoculum may be effective for maintaining soil fertility, and survival of rhizobia in soil in the absence of their hosts (Sereviratne & Jayasinghearachchi, 2005).

Biofilm formation by Bradyrhizobium was first described by Sereviratne and Jayasingherachchi (2003). Since then, both bacterial and plant surface molecules have been shown to be involved in the establishment of microbial communities on legume roots.

In the symbiosis between Bradyrhizobium sp. and peanut plant, the attachment level varies depending on the metabolic state of the rhizobia. Optimal attachment was observed when cells were harvested at the late log or the early stationary phase of growth (Dardanelli et al., 2003). A 14-kDa calcium-binding protein is important for bacterial attachment to the plant root, because root incubation with this adhesin before the attachment assay resulted in a significant, dose-dependent decrease of attachment. EDTA treatment of the cells caused the release of the rhicadhesin-like protein from the bacterial surface into the culture medium, and bacterial attachment was restored (Dardanelli et al., 2003).

Plant lectins are proteins that reversibly and nonenzymatically bind specific carbohydrates (De Hoff et al., 2009). They play important roles during the early stages of interaction between the host plant and the symbiotic bacteria, particularly in the initial attachment of rhizobia to root epidermal cells. Soybean lectin causes a dose-dependent increase of attachment and biofilm formation on polystyrene surface by Bradyrhizobium japonicum wild-type USDA 110 cultures (Pérez-Giménez et al., 2009). Preincubation of rhizobia with soybean lectin increases bradyrhizobial adhesion to soybean roots (Lodeiro et al., 2000). Exopolysaccharides seem to be involved in B. japonicum biofilm formation on both inert and biotic surfaces (Pérez-Giménez et al., 2009). A mutant, which lacks UDP-Glc-4′ epimerase activity and produces low levels of a shorter exopolysaccharide lacking galactose, showed biofilm biomass less than that of the wild-type strain. The defective phenotype was not restored by soybean lectin addition to the mutant culture. Adhesion of mutant cells to soybean roots was significantly lower than that of the wild-type strain, indicating that complete exopolysaccharide is required for efficient colonization of B. japonicum on soybean (Pérez-Giménez et al., 2009).


Attachment of R. leguminosarum to plant root hairs has two steps: primary attachment mediated by either bacterial adhesins (Smit et al., 1992) or plant lectins (Dazzo et al., 1984) and then secondary attachment via cellulose fibrils on the bacterial surface (Dazzo et al., 1984).

Rhizobium leguminosarum, like many other bacteria, forms biofilms on sterile inert surfaces (Fujishige et al., 2005, 2006). The biofilm formation ability, assessed by a microtiter plate assay, was much lower in a pSym-deficient mutant than in R. leguminosarum bv. viciae 128C53 wild-type strain, suggesting that factors encoded by the symbiotic pSym plasmid are required for biofilm formation in this species (Fujishige et al., 2005). Biofilm formation in R. leguminosarum was enhanced by nutrient limitation, in this case sucrose-supplemented 1/4-strength Hoagland's medium (which only contains mineral nutrients essential for plant growth) compared with nutrient-rich tryptone–yeast extract medium (Fujishige et al., 2006). Nutrient availability thus appears to play a major role in the transition from a planktonic to a sessile mode of life, similar to the findings for S. meliloti. Rhizobium leguminosarum established a complex, three-dimensional biofilm on an inert surface, and staining of this biofilm allowed the visualization of the exopolysaccharide matrix (Fujishige et al., 2006). However, the pattern observed for the inert surface model cannot be extrapolated to the root surface model. The root surface is a relatively nutrient-rich environment, but still allows the formation of rhizobial biofilms. One possibility is that a yet-unknown signal or factor from the plant promotes biofilm formation and overrides the inhibitory effect of nutrients released from the root.

Rhizobium leguminosarum bv. viciae A34 attaches securely to inert surfaces such as glass and polypropylene, and forms thick biofilm rings at the air–liquid interface of shaken cultures in minimal medium (Russo et al., 2006). Biofilms formed by this strain showed differentiation into three-dimensional structures when evaluated by confocal laser scanning microscopy; later, the microcolonies developed complex, highly organized honeycomb-like biofilms (Russo et al., 2006). Disruption of the PrsD–PrsE type I secretion system led to reduced biofilm formation, and secretion-defective mutants developed an immature biofilm without honeycomb-like structures, suggesting that this system secretes one or more proteins involved in R. leguminosarum biofilm development (Russo et al., 2006). The acidic exopolysaccharide of this rhizobia is depolymerized by two glycanases, PlyA and PlyB, both secreted by the PrsD–PrsE type I secretion system (Finnie et al., 1997, 1998). A plyA mutant showed little difference in the biofilm biomass compared with wild-type strain A34, whereas plyB and plyA/plyB mutants showed a significant reduction. The phenotype of the double mutant was slightly more aberrant than that of the plyB mutant. Both mutant strains displayed an undeveloped biofilm with many small, dense microcolonies, indicating that the PlyA and PlyB glycanases are partially responsible for the phenotypes of the mutants (Russo et al., 2006). Mutation of the pssA gene, which blocks the production of the acidic exopolysaccharide in R. leguminosarum, caused a drastic decrease of biofilm formation in both shaken and static cultures. This mutant strain formed a flat biofilm, and was unable to develop microcolonies or honeycomb-like structures as evaluated by confocal laser scanning microscopy (Russo et al., 2006). Taken together, the above findings suggest that biofilm formation by R. leguminosarum requires a proper size of the acidic exopolysaccharide, and that the PrsD–PrsE type I secretion system exports proteins that are associated with surface attachment and biofilm maturation.

Lipopolysaccharide plays important roles in symbiosis, either as structural components or as signaling molecules (Fraysse et al., 2003). Lipopolysaccharide, a major constituent of the outer membrane of rhizobia, consists of an outer membrane anchor A lipid connected through a core oligosaccharide to a surface-exposed O-chain polysaccharide. Proper O-polysaccharide and core structures appear to be important for symbiosis, and for structural modification during differentiation to bacteroid (Fraysse et al., 2003). In R. leguminosarum bv. viciae, the O-antigen lipopolysaccharide is essential for cell–cell interaction, and the formation of a compact, structured biofilm (D.M. Russo, unpublished data).

Rhizobial adhesion proteins (known as Rap proteins) have been isolated from R. leguminosarum bv. trifolii (Ausmees et al., 2001). RapA1 is an extracellular calcium-binding protein that promotes rhizobial autoaggregation through cell poles, and is involved in attachment and rhizosphere colonization (Mongiardini et al., 2008). A RapA1-overproducing strain, in comparison with the wild-type R200 strain, showed higher adsorption to roots of the legume host red clover, and to nonsymbiotic plants such as common bean, alfalfa, and soybean (Mongiardini et al., 2008). RapA1 protein promoted rhizobial adsorption to root surfaces. However, overproduction of the protein had no effect on attachment to inert surfaces (polystyrene wells, polypropylene beads, sand, and vermiculite), and did not increase nodulation (Mongiardini et al., 2008). These results suggest that RapA1 receptors are located only on the plant surface, and that the function of the protein may be related to early attachment and colonization of roots, but not to nodulation.

Glucomannan, a polysaccharide located on one of the poles of R. leguminosarum cells, is involved in attachment to the root surface through binding to host plant lectin (Laus et al., 2006). Glucomannan-mediated attachment to pea roots is important for competitive nodule infection (Williams et al., 2008). Similar to the findings reported by Russo (2006) for strain A34, the sequenced strain 3841 forms three-dimensional biofilms on glass, with microcolonies surrounded by water channels and clusters of closely packed hexagonal cells (honeycomb-like structures) (Williams et al., 2008). Elimination of the acidic exopolysaccharide by disruption of the pssA gene led to the formation of a flat, unstructured biofilm (Williams et al., 2008), suggesting (like the findings of Russo et al., 2006) that this exopolysaccharide plays an important role in biofilm formation. However, celA, gelA, and gmsA mutants, which are, respectively, defective in the production of cellulose, gel-forming exopolysaccharide, and glucomannan, formed biofilms indistinguishable from those of the wild-type strain (Williams et al., 2008). The glucomannan and cellulose mutants were defective in root colonization when incubated with host plant Vicia hirsuta (vetch), suggesting that interactions between the rhizobia and glass surface are different from those occurring during root cap formation (Williams et al., 2008).

Unlike what has been described in other rhizobial species, disruption of the CinIR quorum-sensing system in R. leguminosarum led to an increase in biofilm formation (Edwards et al., 2009). This effect seemed to be mediated by the transcriptional regulator ExpR as well as the small protein CinS, coexpressed with the autoinducer synthase CinI (Edwards et al., 2009). The introduction of a mutation in the expR or cinS genes caused an enhanced attachment to glass; however, biofilm rings formed by the expR mutant strain were less stable than those of the cinR and cinI quorum-sensing mutants or the cinS-disrupted strain (Edwards et al., 2009). ExpR and CinS regulate expression of the exopolysaccharide glycanase PlyB, responsible for the cleavage of the acidic exopolysaccharide (Zorreguieta et al., 2000). This suggests again that the proper size of the acidic exopolysaccharide is essential for the formation of biofilms in R. leguminosarum.

Although most reports indicate that exopolysaccharides play an important role during biofilm formation, this cannot be considered as a rule. Rhizobium sp. YAS34 was used to study the function of exopolysaccharides in colonization and biofilm formation on roots of two nonlegume plants: Arabidopsis thaliana and Brassica napus (Santaella et al., 2008). In this case, exopolysaccharide production by this strain was not essential for biofilm formation, either on inert surfaces (polypropylene) or on roots of the above normal plants. This bacterial exopolysaccharide did contribute to colonization of specific zones in relation to nutrient availability (Santaella et al., 2008). Thus, in the absence of the legume host, rhizobia are able to attach and colonize roots of other plants, allowing them to take up nutrients and survive in this protected niche until optimal conditions arise for establishment of symbiosis with the host.

As mentioned previously, bacterial motility mechanisms (swimming, swarming, and twitching) are known to play important roles in biofilm formation, including colonization and subsequent expansion into mature structured surface communities. Specifically, swarming motility enables groups of bacteria to move in a coordinated fashion on a solid surface, spreading as a biofilm (Verstraeten et al., 2008). Sequence analysis of various Rhizobium etli mutants defective in swarming showed effects on quorum sensing, polysaccharide composition or export, motility, and metabolism of amino acids and polyamines. Several such mutants showed reduced symbiotic nitrogen-fixing activity (Braeken et al., 2008). Quorum-sensing signal molecules are involved in the control of the swarming behavior of R. etli, and quorum sensing helps regulate dispersion of existing biofilms and interactions between bacteria and higher organisms, for example, in the Rhizobium–bean symbiosis (Daniels et al., 2004).


Experimental models with abiotic surfaces are useful for initial characterization of the structure of rhizobial biofilms, and of the necessary conditions for biofilm formation. Further studies in the natural habitats of rhizobia, i.e. host plant roots and the rhizosphere, are needed to elucidate the complex events affecting passage from a planktonic to a biofilm lifestyle. As with other bacteria, establishment of a biofilm in rhizobia involves several developmental stages. Based on studies of microorganisms that associate with abiotic surfaces, we are ready to extend this biofilm formation model to plant roots (Fig. 1). Environmental signals cue planktonic cells to settle and establish microcolonies on a surface. Upon attachment, the bacteria divide and differentiate to form three-dimensional shapes that characterize a mature biofilm. This process requires the production of AHLs, exopolysaccharides, lipopolysaccharides, and Nod factors. In studies of an abiotic surface model, individual bacteria can leave the biofilm, to function as dispersal units (Russo et al., 2006). This phenomenon may also occur in biofilms formed on host plant roots.

This review has focused on analysis of interactions as related to the rhizobia. If plant root factors had been taken into account, a much more complex analysis would be necessary (Rodríguez-Navarro et al., 2007). For example, surface characteristics vary along the length of the root. Actively growing root tissues typically exhibit higher rates of exudation into the soil, and biofilms are known to be strongly influenced by nutrient release and exudation at different sites. Lectins released from plant roots affect bacterial attachment and biofilm formation. A separate review is needed for analysis of such plant-dependent variables that affect bacterial attachment to the root surface.

A fundamental question is whether the process of biofilm formation significantly affects legume nodulation. Studies to date indicate that the biofilm lifestyle allows rhizobia to survive under unfavorable conditions (temperature and pH extremes, desiccation, UV radiation, predation, and antibiosis). A large number of viable rhizobia may indirectly ensure the success of nodulation, but there is no direct evidence so far that biofilm formation significantly promotes effective symbiosis with the legume host. A challenge for future studies is to determine how rhizobial biofilm formation is integrated with productive symbiosis.


We would like to thank Dr Ann M. Hirsch and Dr Angeles Zorreguieta for stimulating discussions over years of collaborative research. We also thank Dr Simon Silver and the anonymous reviewers for their motivating comments during the preparation of this manuscript. We apologize to colleagues whose work could not be covered and cited because of space limitations. Our research described in this review was supported by the Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET) of the República Argentina and SECyT-UNRC. W.G. is a Career Member of the CONICET. L.V.R. was supported by a fellowship from the CONICET.


  • Editor: Simon Silver


View Abstract