Correspondence: Jennifer L. Morrell-Falvey, Oak Ridge National Laboratory, Biosciences Division, Oak Ridge, TN 37831, USA. Tel.: +1 865 241 2841; fax: +1 865 574 5345; e-mail: firstname.lastname@example.org
To compete in complex microbial communities, bacteria must sense environmental changes and adjust cellular functions for optimal growth. Chemotaxis-like signal transduction pathways are implicated in the regulation of multiple behaviors in response to changes in the environment, including motility patterns, exopolysaccharide production, and cell-to-cell interactions. In Azospirillum brasilense, cell surface properties, including exopolysaccharide production, are thought to play a direct role in promoting flocculation. Recently, the Che1 chemotaxis-like pathway from A. brasilense was shown to modulate flocculation, suggesting an associated modulation of cell surface properties. Using atomic force microscopy, distinct changes in the surface morphology of flocculating A. brasilense Che1 mutant strains were detected. Whereas the wild-type strain produces a smooth mucosal extracellular matrix after 24 h, the flocculating Che1 mutant strains produce distinctive extracellular fibril structures. Further analyses using flocculation inhibition, lectin-binding assays, and comparison of lipopolysaccharides profiles suggest that the extracellular matrix differs between the cheA1 and the cheY1 mutants, despite an apparent similarity in the macroscopic floc structures. Collectively, these data indicate that disruption of the Che1 pathway is correlated with distinctive changes in the extracellular matrix, which likely result from changes in surface polysaccharides structure and/or composition.
Azospirillum brasilense are soil diazotrophic bacteria that colonize the roots of many economically important grass and cereal species (Steenhoudt & Vanderleyden, 2000). Under conditions of high aeration and limiting availability of combined nitrogen, A. brasilense cells differentiate into aggregating cells and form dense flocs that are visible to the naked eye (Sadasivan & Neyra, 1985; Burdman et al., 1998). Flocs are formed by cell-to-cell aggregation between nonmotile cells embedded in a dense extracellular matrix (Burdman et al., 2000b). Flocculation correlates with, and likely requires the production of, arabinose-rich exopolysaccharides (Bahat-Samet et al., 2004). Scanning electron and fluorescence microscopy studies of A. brasilense aggregating cells indicate the presence of fibrillar material connecting cells to each other or to biotic or abiotic substrates (Bashan et al., 1986, 1991). These fibrils seem to be absent in nonaggregating cells or mutant strains that are defective in aggregation, suggesting that they may play a role in promoting this behavior (Burdman et al., 1998; Skvortsov & Ignatov, 1998). The detailed biochemical composition of this fibrillar material remains unknown, although it is possible that it is related to exopolysaccharide production (Bahat-Samet et al., 2004). In support of this idea, the degree of bacterial aggregation appears to correlate with the amount and composition of exopolysaccharide produced by several A. brasilense strains (Burdman et al., 1998).
Chemotaxis is perhaps the most-studied signal transduction pathway in bacteria (reviewed in Sourjik, 2004; Wadhams & Armitage, 2004; Parkinson et al., 2005; Hazelbauer et al., 2008). Despite the identification of homologous chemotaxis systems in phylogenetically distant bacteria and archaeal species, there is a huge diversity in both the number of chemotaxis operons encoded within bacterial genomes and their physiological roles (Wadhams & Armitage, 2004). Recent studies have shown that the functions of chemotaxis-like pathways are not limited to the regulation of motility patterns, but also include the regulation of biofilm formation, exopolysaccharide production, and cell-to-cell interactions (Black & Yang, 2003; Hickman et al., 2005; Yang & Li, 2005; Caiazza et al., 2007). In prototypical chemotaxis, the histidine kinase CheA and the response regulator CheY form a two-component signal transduction system, which ultimately modulate the probability of changes in the direction of rotation of flagellar motors in response to specific environmental cues. Changes in the phosphorylation of CheY regulated by the CheA–CheY phosphorylation cascade modulate the affinity of CheY for the flagellar motor switch complex and thus chemotaxis. Surprisingly, in A. brasilense, strains carrying mutations in components of the Che1 chemotaxis-like pathway were found to be affected in their ability to interact by cell-to-cell aggregation and in flocculation. Mutant strains lacking functional CheA1 or CheY1 aggregate and flocculate significantly more than the wild-type strain, suggesting that Che1 modulates the ability of A. brasilense cells to flocculate. However, the exact mechanism by which the Che1 pathway regulates cellular functions other than chemotaxis is not known (Bible et al., 2008). Initial attempts at identifying extracellular structures produced specifically by the mutant strains lacking CheA1 and CheY1 and thus controlled by the activity of Che1 have failed, but an effect of Che1 on exopolysaccharide production was suggested from differences in Congo Red staining of colonies (Bible et al., 2008). Flocculation in A. brasilense has been correlated previously with changes in the structure and/or the composition of the extracellular matrix (reviewed in Burdman et al., 2000b), and thus the current working hypothesis is that the Che1 pathway affects flocculation by modulating changes in the structure and/or the composition of the extracellular matrix (Bible et al., 2008).
In this study, we tested this hypothesis by applying atomic force microscopy (AFM) techniques to investigate the cell surfaces of wild-type A. brasilense and its Che1 mutant strain derivatives [AB101 (ΔcheA1) and AB102 (ΔcheY1)]. AFM was selected because it allows nanoscale resolution of biological materials without prior sample fixation. Resolution limitations associated with optical imaging methods and the fixation and dehydration procedures typically associated with classical electron microscopy techniques can inhibit visualization of extracellular structures and could have prevented the identification of CheA1- or CheY1-specific extracellular structures produced during flocculation (Dufrene, 2002, 2003; Bible et al., 2008). The data obtained using AFM conclusively identify a distinctive remodeling of the extracellular matrix, likely via changes in exopolysaccharide production, in AB101 (ΔcheA1) and AB102 (ΔcheY1) under flocculation conditions as well as remarkable differences in the structural organization of the aggregates formed by each of these two strains. Further analyses using a lectin-binding assay, flocculation inhibition, and comparison of lipopolysaccharides profiles are consistent with the hypothesis that the Che1 pathway modulates changes in the extracellular matrix that coincide with flocculation, although this effect is likely to be indirect because our data reveal distinct changes in the content or the organization of the extracellular matrix of the ΔcheA1 and ΔcheY1 mutant strains.
Materials and methods
Strains and growth conditions
Azospirillum brasilense wild-type parental strain Sp7 (ATCC29145) and mutant strains defective in CheA1 [AB101 (ΔcheA1)] and CheY1 [AB102 (ΔcheY1)] were used in this study (Stephens et al., 2006; Bible et al., 2008). Strains were grown in nutrient tryptone–yeast extract (TY) and a minimal salt medium (MMAB) (Hauwaerts et al., 2002). To induce flocculation, cells were grown in 20-mL glass culture tubes with 5 mL of flocculation media (MMAB with 20 mM malate and 0.5 mM NaNO3). Cultures were inoculated with 250 μL of an overnight culture normalized to an OD600 nm of 1.0 (approximately 106 CFU mL−1 for all strains), and incubated on a platform shaker (200 r.p.m.) at 28 °C for 24 h or 1 week.
Quantification of flocculation
To quantify flocculation, we modified a protocol described previously (Madi & Henis, 1989; Burdman et al., 1998). Briefly, 1 mL of sample was subjected to mild sonication using a Branson Digital Sonifer Model 102C equipped with a 3.2 mm tapered micro tip. Settings for sonication included sonic pulses of 2 s on and 2 s off, with the amplitude set at 10%. The percentage of flocculation was calculated by (ODa−ODb/ODa) × 100, where ODa is the OD after sonication and ODb the OD before sonication.
AFM samples were prepared as described, with slight modifications (Doktycz et al., 2003). Briefly, 1-mL aliquots of bacteria were harvested by centrifugation (6000 g) after 24 h or 1 week of growth. Cells were resuspended in 100 μL dH2O and then deposited on a freshly cleaved mica surface. Samples were air-dried 8–24 h before imaging with a PicoPlus atomic force microscope (Agilent Technologies, Tempe, AZ) using a 100 μm multipurpose scanner. The instrument was operated in the contact mode at 512 pixels per line scan with speeds ranging from 0.5 to 1.0 Hz. A Veeco MLCT-E cantilever with a nominal spring constant of 0.5 N m−1 was used for imaging. For all samples, first-order flattened topography and deflection scans were acquired with sizes ranging from 1.5 to 75 μm.
Strains were grown in 5 mL cultures as described above. After 24 h, cells were stained with Syto61 (Invitrogen) following the manufacturer's instructions and resuspended in 200 μL phosphate-buffered saline (PBS) (pH 7.4). Fluorescein isothiocyanate (FITC)-conjugated lentil (LcH; Sigma #L9262) or lima bean lectins (LBL; Sigma #L0264) were added at a final concentration of 50 μg mL−1. The cells were incubated at room temperature with shaking for 20 min, harvested at 8000 r.p.m., and washed with PBS. A Leica TCS SP2 scanning confocal microscope was used for image acquisition. imagej was used for image analysis.
Flocculation inhibition assay
An aggregation bioassay described previously (Burdman et al., 1999, 2000a) was used to assess the roles of d-glucose and l-arabinose in flocculation. Briefly, all strains were grown in flocculation medium or in MMAB. After 24 h, flocculating cultures were sonicated for 20 s and then centrifuged (16 000 g, 2min). The supernatant was then added to cells grown in MMAB (nonflocculating) along with 0.05, 0.1, or 0.5 M concentrations of d-glucose or l-arabinose. The cultures were incubated at 28 °C with shaking for 3–4 h. Flocculation was quantified using the protocol described above.
Extraction of lipopolysaccharides
Lipopolysaccharides was extracted from all strains grown in TY and flocculation medium at 24 h and 1 week using an lipopolysaccharides extraction Kit (Intron Biotechnology) following the manufacturer's instructions. Equal aliquots of lipopolysaccharides extract were dissolved in sodium dodecyl sulfate (SDS) sample buffer, boiled for 5 min, and loaded onto a 4–20% Tris-HCl SDS-polyacrylamide gel electrophoresis gel (Bio-Rad). Samples were electrophoresed at 150 V for 1 h. Gels were silver stained as described (Kittelberger & Hilbink, 1993).
Consistent with previous work, we observed differences in the flocculation behavior of A. brasilense strains deficient in CheA1 and CheY1 compared with wild-type cells (Table 1). At 24 h, aggregative structures and flocs were visible for the Che1 pathway mutant strains AB101 (ΔcheA1) and AB102 (ΔcheY1), but were not seen in the wild-type cultures at this time point. The amount of flocculation relative to planktonic cells for AB101 and AB102 was increased after 1 week of incubation (Table 1). Flocculation was significant for the wild-type culture after 1 week of incubation (Table 1). All strains had similar motility before flocculation, and all strains lost motility after significant flocculation occurred, in agreement with previous findings (Burdman et al., 1998; Pereg Gerk et al., 2000; Bible et al., 2008). Taken together, these data are consistent with earlier findings that AB101 and AB102 flocculate earlier than the wild-type strain (Bible et al., 2008).
Table 1. Quantification of flocculation
0.01 ± 0.67
32.3 ± 9.50
52.0 ± 0.07
95.0 ± 0.61
86.3 ± 0.02
93.5 ± 0.38
Comparison of extracellular matrix structures by AFM
Examination of AFM images revealed that the extracellular matrix of AB101 (ΔcheA1) and AB102 (ΔcheY1) contained fibrillar material at 24 h (Fig. 1c and d and Supporting Information, Fig. S1). The extracellular matrix of AB101 (ΔcheA1) and AB102 (ΔcheY1) appeared as a ridged structure on the surface of cells with fibrils protruding from the cells (Fig. 1c and d, Fig. S1). In contrast, the extracellular material surrounding cells of the nonflocculating wild-type strain appeared to be smooth and globular at 24 h (Fig. 1a). Numerous high-resolution scans of wild-type nonflocculating cells failed to reveal fibrillar material (Fig. 1a and data not shown). After 1 week, however, fibrillar material was observed for flocculating wild-type cells (Fig. 1b).
Despite the apparent similarity of the macroscopic flocculation phenotype of the mutant strains, analyses of AFM topography and deflection images revealed a dissimilarity in the organizational pattern of the aggregating cells (Figs 2 and S2). The most striking difference was observed in comparing the extracellular material of AB102 (ΔcheY1) with that of AB101 (ΔcheA1) or wild-type cells. A network of extracellular material is visible between the AB102 (ΔcheY1) cells as early as 24 h (data not shown) and it becomes more distinct after 1 week (Fig. 2c). Line scans across the flocs indicate that AB102 (ΔcheY1) cells are embedded in a matrix that spans approximately 400 nm between cells (Fig. 2f). This tight organization is not observed in flocs formed by AB101 (ΔcheA1) (Fig. 2b). In this strain, as well as in flocculating wild-type cells, individual cells are distinctly defined within the flocs and no obvious features are observed between the cells (Fig. 2a, b, d, and e)
Flocculation inhibition assay
Given the noticeable difference in the floc structures discovered using AFM techniques, we sought to use a flocculation inhibition assay to determine whether glucose and/or arabinose residues, previously shown to contribute to flocculation in A. brasilense (Burdman et al., 2000a), were present and participated in cell-to-cell aggregation and flocculation. The addition of 0.5 M arabinose to flocculating cultures of both mutant strains caused a significant reduction in the total amount of flocculation (Fig. 3a), suggesting that arabinose contributes to the structure and/or the stability of the flocs formed by these strains. In addition, we found that flocs of the AB102 (ΔcheY1) strain were significantly more sensitive to the competitive addition of exogenous arabinose (Fig. 3a) than the flocs of AB101. Similarly, high concentrations of glucose (0.5 M) reduced flocculation in both mutant strains and flocs formed by the AB102 (ΔcheY1) strain appeared to be more sensitive to the addition of glucose, with almost complete inhibition of flocculation after the addition of 0.5 M glucose (Fig. 3b).
To further investigate differences in the extracellular matrix, we used FITC-conjugated lentil lectin (LcH) (affinity for α-mannose and α-glucose) and lima bean lectin (LBL) (affinity for N-acetyl galactosamine) to probe for specific carbohydrates present on or around the cell surface. Wild-type cells did not show any significant binding of either lectin after 24 h of growth as determined by fluorescence imaging and statistical analysis (Fig. 4; Table S1). Both lectins were found to stain AB101 (ΔcheA1) cells and the surrounding material (Fig. 4b and h). In comparison with AB101, AB102 (ΔcheY1) cells displayed reduced staining by both lectins (Fig. 4c and i). When normalized to the fluorescence signal of Syto61 that stains all cells (Fig. 4d–f and j–l), the lectin fluorescence signal detected for AB102 (ΔcheY1) floc structures was significantly (P=0.05) reduced for both lectins with respect to AB101 (Table S1).
Lipopolysaccharides profiles of the Che1 mutants
The lipopolysaccharides profiles of the mutant and wild-type strains grown under flocculating and nonflocculating conditions were compared. Under conditions of growth in rich medium (TY), all strains had similar lipopolysaccharides profiles (Fig. 5). Differences in lipopolysaccharides profiles were detected between the strains as early as 24 h of incubation in flocculation medium, which corresponds to the time at which both mutant strains, but not the wild-type strain, flocculate. Under these conditions and compared with the lipopolysaccharides profile of the wild-type strain, a low-molecular-weight band (arrow 2, Fig. 5) is absent from the profile of both mutant strains while another low-molecular-weight band (arrow 3, Fig. 5) is significantly reduced. A higher molecular weight band (Fig. 5, arrow 1) is also clearly visible for all strains, but more abundant in the lipopolysaccharides profile of both mutant strains at 24 h. After 1 week of incubation, the wild-type strain flocculated and its lipopolysaccharides profile mirrored that of the flocculated mutant strains: the lower molecular weight bands (arrows 2 and 3, Fig. 5) are significantly fainter while a higher molecular weight band (arrow 1, Fig. 5) shows an increase in relative abundance. Collectively, the data suggest that changes in the lipopolysaccharides profiles of flocculated cells are comparable for all strains, and that changes in lipopolysaccharides profiles are correlated and coincident with flocculation.
In this study, we used high-resolution imaging to investigate the cell surface and the surrounding matrix of the A. brasilense AB101 (ΔcheA1) and AB102 (ΔcheY1) mutant cells during flocculation. Several recent investigations support the hypothesis that exopolysaccharides and outer membrane proteins are involved in cell-to-cell aggregation leading to flocculation in Azospirillum spp. These data support a model in which flocculation is accompanied and perhaps triggered by several changes in cell surface properties, because flocculation coincides with remodeling of the cell surface and the extracellular matrix in A. brasilense (Delgallo et al., 1989; Burdman et al., 2000a; Bahat-Samet et al., 2004; Mora et al., 2008; Mulyukin et al., 2009). Comparisons between AFM micrographs, lectin-binding affinities, and lipopolysaccharides profiles of planktonic and flocculating cells performed in the present study are consistent with this model because they collectively show that an increased flocculation phenotype correlates with a set of changes in cell surface characteristics, including the apparent specific production of fibrillar extracellular material at the edge of floc structures. Such fibrillar material was observed in all flocculated strains [AB101 (ΔcheA1) and AB102 (ΔcheY1) at 24 h and in the wild type at 1 week]. Furthermore, more abundant fibrillar material was detected on the surface of the AB102 (ΔcheY1) strain, which correlates with the greater amount of flocculation observed consistently for this strain. Whether this fibrillar material associated with flocculating cells is directly related to or has a role similar to that of the fibrillar structures reportedly formed by A. brasilense Cd cells during aggregative attachment to wheat roots and sand particles remains to be determined (Bashan et al., 1991).
The function of the chemotaxis-like Che1 pathway in modulating flocculation in A. brasilense is unexpected and the underlying molecular mechanism(s) remains to be determined. However, several other chemotaxis-like signal transduction pathways regulate functions other than motility by mechanisms that are yet to be determined (Black & Yang, 2003; Berleman & Bauer, 2005a, b; Hickman et al., 2005). Data obtained here shed light on several important features of Che1-dependent regulation of flocculation behavior in A. brasilense. If the signaling output of the Che1 pathway, in which CheA1 and CheY1 are expected to form a chemotaxis-like two-component system (Hauwaerts et al., 2002; Stephens et al., 2006; Bible et al., 2008), is the direct regulation of molecular target(s) modulating the flocculation behavior, then mutations that impair CheA1 or CheY1 functions should yield similar phenotypes. This study revealed several distinguishing features of the flocs formed by each of the mutant strains that are not consistent with a direct function of Che1 in the regulation of flocculation. First, although cells of both mutant strains were adherent and embedded in a complex matrix apparently comprised of fibrillar material, cell-to-cell contacts within the matrix of the AB102 (ΔcheY1) strain were separated by a thick layer that was visible by AFM after 1 week. This layer formed a tight network around each individual cell within the floc. In contrast, in the flocs of AB101 (ΔcheA1), individual cells were distinctly defined and no obvious connecting features were observed between the cells.
Because it is impossible to determine the composition of this material from imaging alone, we used flocculation inhibition and lectin-binding assays to analyze the different structures observed between the two strains in more detail. The results of the lectin-binding assay suggest that AB101 (ΔcheA1) produces an exopolysaccharide that is more abundant in α-mannose and/or α-glucose, and N-acetyl galactosamine than the exopolysaccharide produced by AB102 (ΔcheY1). Previous studies have shown that the glucose content of exopolysaccharide is significantly lower during flocculation in the wild-type Sp7 strain and in other mutant derivative strains with increased aggregation capacity (Bahat-Samet et al., 2004). Consistent with these data, AB102 (ΔcheY1) strain displays a stronger flocculation phenotype and its extracellular matrix appears to have a reduced mannose and/or a glucose content relative to that of AB101 (ΔcheA1). An alternative explanation for these data is that the structural organization of the AB102 (ΔcheY1) floc reduces the accessibility of the sugar residues to the lectin, thus limiting the amount of lectin that binds to the cells and the surrounding matrix.
Even though the floc structures of the two mutant strains showed different binding affinities for lectins, indicating possible differences in the polysaccharide composition of the exopolysaccharide produced during flocculation, these results do not necessarily demonstrate the contribution of specific polysaccharides to aggregation or flocculation. Previous studies showed that exopolysaccharide composition is modified over time from a glucose-rich exopolysaccharide to an arabinose-rich exopolysaccharide and that this temporal change correlates directly with the timing of flocculation (Bahat-Samet et al., 2004). In agreement with this observation, flocs formed by the ΔcheY1 mutant were more sensitive to the addition of arabinose in the flocculation inhibition assay, suggesting that the sugar residues comprising the matrix of these strains are different in structure and/or composition. This could also suggest that the ΔcheY1 strain, which flocculates quantitatively more, represents a more advanced stage of flocculation compared with the ΔcheA1 mutant strain or the wild-type strain. Comparison of changes in the lipopolysaccharides profiles of the wild-type and mutant strains lends further credence to this possibility because differences in the lipopolysaccharides profiles were seen to occur for all strains, but at different times during the flocculation process. Therefore, the mutant strains lacking cheA1 or cheY1 may be affected in the timing of flocculation, which may result, for example, from an increased sensitivity of the cells to the cues that trigger flocculation or perhaps to other effects. Structural and other differences identified between the flocs formed by ΔcheA1 and ΔcheY1 strains thus collectively suggest that the function of Che1 in modulating flocculation is indirect. Taken together and with data from the literature (Burdman et al., 2000a; Bahat-Samet et al., 2004; Bible et al., 2008), the results obtained here underscore the significant changes of the cell surface and extracellular matrix that occur during flocculation and support a model in which flocculation in A. brasilense is an adaptive behavior that allows the cells to differentiate into resistant forms via extensive remodeling of the cell surface and the extracellular matrix, including lipopolysaccharides and exopolysaccharide.
The authors would like to thank Dave Allison for helpful discussions. This research was funded by the Genomic Science Program of the Office of Biological and Environmental Research, US DOE, and NSF MCB-0919819 to G.A. Oak Ridge National Laboratory is managed by UT-Battelle, LLC, for the US Department of Energy under Contract no. DE-AC05-00OR22725.