Exopolysaccharide Biosynthesis Enables Mature Biofilm Formation on Abiotic Surfaces by Herbaspirillum seropedicae

H. seropedicae associates endophytically and epiphytically with important poaceous crops and is capable of promoting their growth. The molecular mechanisms involved in plant colonization by this microrganism are not fully understood. Exopolysaccharides (EPS) are usually necessary for bacterial attachment to solid surfaces, to other bacteria, and to form biofilms. The role of H. seropedicae SmR1 exopolysaccharide in biofilm formation on both inert and plant substrates was assessed by characterization of a mutant in the espB gene which codes for a glucosyltransferase. The mutant strain was severely affected in EPS production and biofilm formation on glass wool. In contrast, the plant colonization capacity of the mutant strain was not altered when compared to the parental strain. The requirement of EPS for biofilm formation on inert surface was reinforced by the induction of eps genes in biofilms grown on glass and polypropylene. On the other hand, a strong repression of eps genes was observed in H. seropedicae cells adhered to maize roots. Our data suggest that H. seropedicae EPS is a structural component of mature biofilms, but this development stage of biofilm is not achieved during plant colonization.


Introduction
H. seropedicae is a nitrogen-fixing, plant-growth-promoting Betaproteobacterium found attached to and within tissues of important crops such as maize (Zea mays), rice (Oryza sativa), sorghum (Sorghum bicolor) sugar-cane (Saccharum officinarum) and wheat (Triticum aestivum) [1]. The molecular mechanisms of plant recognition, attachment, penetration and endophytic colonization of this microrganism are not well known [1]. EPS are carbohydrate polymers of highly variable composition and structure found outside cells [2]. Bacterial EPS are usually responsible for attachment to solid surfaces and to other bacteria, thus forming microscopic and macroscopic cell aggregates [3]. When the aggregates are neatly organized, they are called biofilms [4]. In these communities the surface-associated microorganisms grow in matrix-enclosed microcolonies separated by a network of open-water channels [5,6]. The presence of a matrix between cells confers a series of selective advantages, such as protection against environmental variations, nutrient and ions retention, resistance to desiccation and mechanical protection [4,7,8].
Most of microorganisms do not occur naturally in planktonic communities, being generally found attached to biological and non-biological surfaces forming biofilms [9]. Initial stages of biofilm formation involves the redistribution of attached cells by surface motility [10][11][12], binary division of attached cells [13] or recruitment of cells from the surrounding fluid to the developing biofilm [14]. The individual adherent cells that initiate biofilm formation on a surface are capable of independent movement [12] before they begin to exude exopolysaccharide and adhere irreversibly [5]. Biofilm maturation results in the generation of a complex architecture with channels, pores, and redistribution of bacteria away from the substrate [15]. As the biofilm matures many cells alter their physiological processes in response to the conditions in their particular niches. The biofilm cells express genes in a pattern that deeply differs from that of their planktonic counterparts [16]. Finally, individual cells or whole microcolonies may detach from the biofilm and colonize other surfaces [17].
EPS and biofilm formation have been associated with the capacity of bacteria to colonize plants in symbiotic, neutral or pathogenic associations. One of the EPS functions in plantbacterial interaction is to permit epiphytic colonization of the plant host [18]. Also, in plant-pathogen interaction EPS helps to create a favorable environment for pathogen survival and growth inside the infected plant, acting as a protective barrier against plant metabolic defenses [19]. The knockout of EPS biosynthesis genes (exo or eps) resulted in loss of virulence by Erwinia stewartii and Xanthomonas axopodis [20]. The mutation of Xanthomonas campestris gumD, which codes for a glucosyltransferase, drastically decreased the pathogenicity of this organism [21]. Also, EPS was the main factor required for bacterial wilt caused by Ralstonia solanacearum [22], where it seems to interfere with plant water transport by clogging the xylem [23]. In the case of diazotrophic symbionts, EPS seems to be indispensable for functional nodule establishment [24,25]. Ensifer meliloti mutant strains deficient in the production of one kind of EPS induce nodule formation, but they do not contains bacteroids [26,27]. The knockout of acidic EPS biosynthesis genes of Ensifer sp. NGR234 also results in pseudonodule formation [28]. The infection and subsequent nodulation of legumes by R. leguminosarum requires bacterial attachment onto root hair, a process that involves EPS production [29]. In the case of associative diazotrophs such as Azospirillum brasilense and Gluconacetobacter diazotrophicus, EPS seems to influence cellular aggregation and biofilm formation on plant root surface [30][31][32]. The knockout of rhamnose biosynthesis in A. brasilense led to a decrease in EPS production, and a decrease in maize colonization [33]. In G. diazotrophicus, exopolysaccharides seem to have a more dramatic effect, where knockout of gumD abolished attachment to rice root surface and endophytic colonization [32].
There is no evidence of the role of H. seropedicae EPS in plant colonization, although scanning electron microscopy revealed production of mucilaginous and fibrillar materials by H. seropedicae during colonization of maize, rice and sorghum root surfaces [34,35]. This material might be EPS. In this work we knocked out the epsB gene which codes for a putative glucosyltransferase of the EPS biosynthesis gene cluster of H. seropedicae. The mutant strain has diminished EPS production and biofilm formation on abiotic surfaces, but showed no alterations on maize colonization profile compared to the wild type.

Growth of bacterial strains, DNA manipulations and mutagenesis
Bacterial strains and their relevant characteristics are listed in Table 1. Herbaspirillum seropedicae strains were grown at 30uC and 120 rpm in NFbHPN medium [41]. Escherichia coli strains were grown at 37uC in LB medium [42]. Antibiotics were added at the following concentrations when required: ampicillin (Ap) 10 mg.mL 21 ; kanamycin (Km) 50 mg.mL 21 ; chloramphenicol (Cm) 30 mg.mL 21 ; tetracycline (Tc) 10 mg.mL 21 ; streptomycin (Sm) 80 mg.mL 21 . The plasmids used in this study are listed in Table 1. Plasmid and total DNA preparations, agarose gel electrophoresis, restriction endonuclease digestion and cloning were performed according to standard protocols [42].
For epsB mutagenesis the primers HSepsB-F (59-gctggaaccgcatatgatcgt-39) and HSepsB-R (59-ccaggtggatccggtcaataa-39) were used to amplify the epsB gene from H. seropedicae genomic DNA, and the amplicon was cloned in pTZ57R/T. The generated plasmid pTZHSwaaL was disrupted in the EcoRV site by the nptI cassette isolated from pKIXX that confers resistance to kanamycin (Km). The disrupted gene was transferred to pSUP202. This construction was electro-transformed in E. coli S17.1, and the transformants were conjugated into H. seropedicae SmR1. The mutant strains were selected and named H. seropedicae EPSEB (epsB 2 ). Insertion of the cassette in the genome of the mutant strain by double crossover event was confirmed by PCR analyses. Wild-type and EPSEB mutant strains were GFP-marked through conjugation with E. coli S17.1 harboring the pHC60 plasmid.

EPS and LPS analyses
For EPS extraction, the H. seropedicae wild type and EPSEB mutant strains were grown in 10 mL of NFbHPN medium [41] at 30C and 120 rpm in the presence of 50 mg of sterile glass fiber. After 12hours, the bacterial cultures together with the glass fiber were transferred to a 50 mL centrifuge tube, and vortexed vigorously for 1 minute to remove glass fiber attached bacteria. The cells and the glass fiber were then removed by centrifugation (15 min, 3000 g) and the supernatant was filtered through a 0.22 mm membrane to remove residual cells. Exopolysaccharides in the filtered supernatant were precipitated with 3 volumes of cold ethanol for 24hours at 20C and centrifuged for 10 minutes at 4uC and 3000 g. The precipitate was vacuum dried, resuspended in MilliQ water and dialyzed against MilliQ water. Ten microliters of dialyzed samples were mixed with sample buffer (120 mM Tris pH 6.8; 3% SDS; 9% b-mercaptoethanol; 30% glycerol; 0.03% bromophenol blue), separated by SDS-PAGE (12% acrylamide) and visualized by silver periodate oxidation staining [43]. Total sugar concentration of the samples was determined with phenol/ sulfuric acid [44], using glucose as standard.
LPS extraction for electrophoretic analysis was performed according to Balsanelli et al. [45] by the proteinase K -SDS method. Four microliters of final mixture were separated by SDS-PAGE (16% acrylamide) and visualized by silver periodate oxidation staining [43].

Biofilm formation on glass fiber
H. seropedicae strains were grown as described for EPS isolation, and biofilm formation was evaluated according to Balsanelli et al. [45]. Briefly, twelve hours after inoculation glass fiber was removed from the medium, stained with 20 mL of crystal violet 1% for 2 minutes, and washed three times with 0.9% saline solution. Then, 1 mL of absolute ethanol was added to remove the dye, and the alcoholic solution was used to determine the OD 550 . The values are expressed as OD 550 of the samples subtracted from the OD 550 of the fiber glass treated culture medium. The results reported represent the average of three independent experiments. Purified wild type EPS (100 mg of glucose equivalents.mL 21 ) was added to the system during incubation with glass fiber to test complementation of the mutant strain phenotype. Samples of stained glass fibers were analyzed by light microscopy for visualization of biofilm structure.

Plant interaction assays
Assays of maize colonization by H. seropedicae strains were performed according to Balsanelli et al. [46]. Briefly, seeds of Zea mays cv. 30F53, Oryza sativa cv Nipponbare or Sorghum bicolor cv A07 were surface-sterilized, germinated and each seedling was inoculated with 10 5 CFU of H. seropedicae strains. The inoculated seedlings were transferred to a hydroponic system containing 30 mL of plant medium [47] and 10 g of sterile culture beads in 100 mL glass tubes. Bacterial counts were made immediately after inoculation to access attached bacteria and 1, 4, 7 or 10 days after inoculation to access endophytic and epiphytic bacteria. The results reported represent the average of at least three independent experiments.
The GFP-marked strains were used as inoculants as described above, and longitudinal root cuts were freshly prepared for visualization. Root attached and 7 d.a.i. epiphytic bacteria were visualized by confocal laser scanning microscopy (CLSM) on a Nikon Ti Microscope. Plant tissues showed DAPI autofluorescence. Snapshots of the tridimensional images were obtained with the NIS-Elements software (Nikon).
Competition assays were performed using as inoculant a mixture of H. seropedicae wild type and epsB strains in 1:1 proportion, with a total of approximately 10, 10 2 , 10 3 , 10 4 , or 10 5 bacteria per seedling. Total bacterial counts were made as described before, and the strains were identified by antibiotic resistance.

EPS biosynthesis gene expression during rhizoplane colonization and biofilm formation
To evaluate eps gene expression during rhizosphere colonization, the H. seropedicae MHS01 [38] (epsG::lacZ) reporter strain was grown in NFbHPN medium for 16 h. After adjusting the culture to OD 600 = 1.0 in saline buffer, 10 8 cells (1 mL) were inoculated onto maize in the hydroponic system described above and incubated at 28uC. After 24 h, bacterial cells were recovered from the liquid medium by centrifugation and attached cells were recovered from root surface and polypropylene spheres by vortexing and centrifugation.
To evaluate eps gene expression during biofilm formation, the H. seropedicae MHS01 reporter strain was grown in the presence of glass fiber as described. After 12 h of growth the free living cells were recovered by centrifugation and attached cells were recovered from glass fiber by vortexing and centrifugation. The b-galactosidase activity of the recovered cells was then measured [48]. Protein determination was carried out according to Bradford [49]. The b-galactosidase activity is reported as nmol of onitrophenol produced per minute and mg of protein. The results reported represent the average of at least three independent experiments. The control containing uninoculated maize seedlings did not show any detectable b-galactosidase activity.

Genomic organization of H. seropedicae EPS biosynthesis genes and mutagenesis
Analyses of H. seropedicae SmR1 genome sequence (CP002039) showed a cluster of 28 genes that code for proteins probably involved in the biosynthesis and secretion of EPS (Fig. S1). The organization of these genes is highly similar to the eps cluster of Herminiimonas arsenicoxydans [50] and Methylobacillus sp. 12S [51], and the encoded proteins share high identity to the homologous proteins of all three microorganisms (Table S1). The EPS produced by Methylobacillus sp. 12S, named metanolan, is a heteropolymer composed of glucose, galactose and mannose in a 3:1:1 molar proportion [52]. The analyses of H. seropedicae eps genes that code for glycosyltransferases and sugar modifying proteins (such as epimerases and phosphatases in Table S1) suggest that the EPS is composed of these same monosaccharides. Indeed, monosaccharide composition analysis of H. seropedicae Z67 T EPS showed galactose, glucose and mannose as constituents at a proportion of 4:3:1, with possible substitutions with tetracarboxylic acids [53].

Knockout of epsB strongly reduces EPS production by H. seropedicae
The production of EPS was initially evaluated by precipitation of H. seropedicae strain culture supernatant with 3 volumes of cold ethanol. When the wild type and EPSEB (epsB) strains were grown in liquid NFbHPN for 24h no EPS was produced in the culture supernatant. Since in many bacteria EPS biosynthesis is induced during biofilm formation [54], the supernatant of H. seropedicae wild type culture grown for 12hours in the presence of glass fiber was processed as above and 0.8 mg.mL 21 of EPS was obtained. In contrast with the wild type strain, no EPS could be detected from the EPSEB strain. The samples were then analyzed by a 12% SDS-PAGE (Fig. 1). Exopolysaccharide from the wild type strain showed three poorly defined bands of different molecular weight/ charge, while supernatant of EPSEB strain had no polysaccharide band. The EPSEB strain LPS electrophoretic profile did not differ from that of the wild type (Fig. S2), suggesting that this glucosyltransferase is specific for EPS biosynthesis.

H. seropedicae EPS is necessary for biofilm formation on glass fiber
To evaluate the role of EPS in biofilm formation, the strains were grown in the presence of glass fiber and biofilm formation was evaluated quantitatively by staining attached bacteria (Table 2), and qualitatively by light microscopy ( Fig. 2A-D). After twelve hours of growth the EPSEB strain showed a 45% reduction in biofilm formation compared to the wild type. Furthermore, microscopic observation showed that the wild type strain formed large tridimensional structures, considered as mature biofilms ( Fig. 2A). On the other hand, the mutant strain did not form mature biofilms, with only few attached cells (Fig. 2B). This phenotype was partially restored by the addition of purified H. seropedicae EPS (Fig. 2D), suggesting that this polysaccharide is required for biofilm development.
The reporter strain MHS01 (epsG::lacZ) was used to determine the regulation of eps genes in glass fiber biofilm formation (Fig. 2E). After 12hours of growth in the above-described system, epsG expression in glass fiber attached cells was about 3 times higher than in planktonic cells. The eps genes up-regulation on bacteria adhered to glass fiber suggests the involvement of EPS in biofilm formation on inert matrix.

Maize colonization by H. seropedicae is not dependent on EPS production
Colonization of H. seropedicae strains on maize roots was followed to evaluate the role of EPS in this interaction. The colonization profile of the EPSEB strain was very similar to that of the wild type (Fig. 3), suggesting that attachment, epiphytic and endophytic colonization are not dependent on epsB gene. Colonization of rice and sorghum by the EPSEB strain was also very similar to that of the wild type strain (Fig. S3), suggesting that EPS production is not required for interaction with poaceous plants. The maize colonization profile of MHS01 was also similar to the wild type one [48], indicating that the eps gene cluster and its product are not involved in plant interaction. The use of smaller numbers of wild type and EPSEB cells in attachment assays on maize roots did not show differences of colonization between the strains (Fig. 4).
CLSM analyses showed that both wild type and epsB mutant strains attach onto the maize root epidermis and root hair as individual cells and in similar numbers (Fig. 5A). Seven days after inoculation (Fig. 5B), the epiphytic population of both strains was still formed of individualized cells, not comprising tridimensional biofilm structures. These results indicate that H. seropedicae do not develop mature biofilms on roots as observed on glass fiber, stressing that EPS production is not required in plant colonization.
H. seropedicae EPS is required for resistance to abiotic stress EPS production has been associated with protection against chemical stress [3,8,33,55]. We tested the resistance of the mutant and parental strain to the flavonoids naringenin and quercetin, to the plant immune metabolites jasmonic and salicylic acids, to phenol and SDS (Fig. 6). The mutant strain's resistance to plant bactericidal compounds was not different from that of the wild   type. On the other hand, the parental strain showed resistance to low concentrations of phenol and SDS, while mutation in epsB gene reduced the survival of the mutant strain by 95%. These results suggest that H. seropedicae EPS is involved in resistance to non-biochemical stress, but not in resistance to plant basal defense.

H. seropedicae eps genes expression is down-regulated during maize colonization
Tadra-Sfeir and coworkers [38] showed by RT-PCR that the expression of epsB and epsG (code for glucosyltransferases) was repressed in the presence of the flavonoid naringenin. The reporter strain MHS01 (epsG::lacZ) was used to determine if the eps genes were regulated during maize colonization (Fig. 7). The results show that epsG is repressed during the first steps of interaction with maize, suggesting that EPS biosynthesis is diminished under this condition. Such repression was observed both in planktonic bacteria free in the hydroponic medium in the presence of the plant roots and in root-attached bacteria, suggesting that H. seropedicae EPS is not required for the attachment on root surface. On the other hand, eps genes were induced (2.5-fold) in the bacteria adhered to the polypropylene spheres of the hydroponic system compared to planktonic bacteria, regardless the plant presence. This result stress the involvement of H. seropedicae EPS in biofilm formation on inert matrices.

Discussion
Exopolysaccharides are important factors that enable cellular aggregation and biofilm formation on solid surfaces. As shown for other plant associative bacteria [30][31][32], mutation of EPS biosynthesis genes in H. seropedicae SmR1 decrease EPS production and consequently biofilm formation, but surprisingly, did not alter maize colonization profile.  The importance of EPS in biofilm formation is supported by the induction of eps genes in the presence of inert substrates such as glass fiber and the polypropylene spheres. On the other hand, no difference was observed between the wild type and mutant strains in maize, rice or sorghum epiphytic colonization capacity. Even when lower numbers of bacteria were used to inoculate maize plants, both strains had similar root attachment patterns. Moreover, the increase and maintenance of the root epiphytic population seemed not to be dependent on EPS production. In agreement with those results, eps gene expression was repressed in H. seropedicae cells colonizing maize root surfaces. A huge impact in attachment and epiphytic colonization was observed by the lack of EPS production in G. diazotrophicus [32], but that seems not to be the case in H. seropedicae SmR1.
EPS can contribute to survival of bacteria within the plant by acting as a barrier against plant defense mechanisms, and creating a favorable microenvironment [55,56]. EPS production seems to be important for H. seropedicae resistance to chemical stress caused by phenol and SDS, but not required for resistance to plant defense metabolites such as flavonoids, jasmonic and salicylic acids. Indeed, the mutant strain was able to cope with the plant chemical defense and endophytically colonize maize roots to the same extend than the wildtype. These results indicate that the product of the eps gene cluster is not necessary for maize root endophytic colonization by H. seropedicae.
The results lead us to propose a model for the early steps of H. seropedicae maize colonization. Upon contact with the rhizosphere environment eps genes are down-regulated, decreasing EPS biosynthesis. On the other hand, LPS biosynthesis is up-regulated, which allows the bacteria to bind to plant lectins on the root surface [46]. In accordance with this suggestion, scanning electron microscopy [34,35] and the CLSM results showed that H. seropedicae cells form a monolayer on maize root surface, not developing to mature biofilm. It seems that H. seropedicae biofilm development is arrested on roots by the reduced biosynthesis of EPS. The loosely attached bacterial cell can then penetrate inner root tissues and colonize them. By avoiding permanent attachment and biofilm maturation H. seropedicae would remain available to seek penetration sites and nutrient sources.
In most plant-interacting bacteria studied so far, including associative, symbiotic or pathogenic, whenever the EPS is involved in biofilm formation it is also required for plant colonization or acts as a virulence factor [32,[55][56][57][58][59][60][61][62][63][64][65][66][67][68][69][70][71][72][73]. In a stark contrast, H. seropedicae SmR1 EPS is necessary for biofilm formation but EPS synthesis is repressed during maize root colonization.  seropedicae MHS-01 (epsG::lacZ) cells were inoculated in the hydroponic system. After 24 hours, the cells from the hydroponic medium were collected by centrifugation. The cells attached to roots or to polypropylene spheres (PP) were removed by vortex and concentrated by centrifugation. For all the samples the b-galactosidase activity was determined, standardized by total protein concentration, and expressed as nmol ONP.(min.mg protein) 21 6 standard deviation. Different letters indicate significant differences (p,0.01, Duncan multiple range test) in epsG expression between the tested conditions. doi:10.1371/journal.pone.0110392.g007 Figure S1 H. seropedicae SmR1 eps gene cluster. The proteins coded by the showed genes were analyzed in Table S1. The indicated probable promoter regions were identified with the BPROM software (SoftBerry). (TIFF) Figure S2 Electrophoretic pattern of LPS isolated from H. seropedicae SmR1 (A) and EPSEB (B). SDS-PAGE was performed with total LPS extracted from 10 7 cells grown in NFbHPN medium by the SDS/proteinase K method, and visualized with silver periodate oxidation staining.