The chemotaxis-like Che1 pathway has an indirect role in adhesive cell properties of Azospirillum brasilense


  • Piro Siuti,

    1. Biosciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA
    2. Graduate School of Genome Science and Technology, Oak Ridge National Laboratory, University of Tennessee, Knoxville, TN, USA
    Search for more papers by this author
  • Calvin Green,

    1. Department of Biochemistry, Cellular and Molecular Biology, University of Tennessee, Knoxville, TN, USA
    Search for more papers by this author
  • Amanda Nicole Edwards,

    1. Biosciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA
    2. Graduate School of Genome Science and Technology, Oak Ridge National Laboratory, University of Tennessee, Knoxville, TN, USA
    Search for more papers by this author
  • Mitchel J. Doktycz,

    1. Biosciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA
    2. Graduate School of Genome Science and Technology, Oak Ridge National Laboratory, University of Tennessee, Knoxville, TN, USA
    3. Center for Nanophase Materials Sciences, Oak Ridge National Laboratory, Oak Ridge, TN, USA
    Search for more papers by this author
  • Gladys Alexandre

    Corresponding author
    1. Department of Biochemistry, Cellular and Molecular Biology, University of Tennessee, Knoxville, TN, USA
    • Graduate School of Genome Science and Technology, Oak Ridge National Laboratory, University of Tennessee, Knoxville, TN, USA
    Search for more papers by this author

Correspondence: Gladys Alexandre, Department of Biochemistry, Cellular and Molecular Biology, F425 Walters Life Sciences Building, 1414 W. Cumberland Avenue, Knoxville, TN 37996, USA. Tel.: +1 865 974 0866; fax: +1 865 974 6306; e-mail:


The Azospirillum brasilense chemotaxis-like Che1 signal transduction pathway was recently shown to modulate changes in adhesive cell surface properties that, in turn, affect cell-to-cell aggregation and flocculation behaviors rather than flagellar-mediated chemotaxis. Attachment to surfaces and root colonization may be functions related to flocculation. Here, the conditions under which A. brasilense wild-type Sp7 and che1 mutant strains attach to abiotic and biotic surfaces were examined using in vitro attachment and biofilm assays combined with atomic force microscopy and confocal microscopy. The nitrogen source available for growth is found to be a major modulator of surface attachment by A. brasilense and could be promoted in vitro by lectins, suggesting that it depends on interaction with surface-exposed residues within the extracellular matrix of cells. However, Che1-dependent signaling is shown to contribute indirectly to surface attachment, indicating that distinct mechanisms are likely underlying flocculation and attachment to surfaces in A. brasilense.


Chemotaxis is a widespread function in motile soil bacteria as it affords cells with the ability to sense and to navigate toward the most favorable niches for growth (Wadhams & Armitage, 2004). At the molecular level, the chemotaxis pathway is the dedicated chemosensory signal transduction system that allows cells to couple detection of physicochemical changes in their surroundings to changes in the swimming pattern (i.e. chemotaxis). Chemotaxis signal transduction has been best studied in Escherichia coli and experimental evidence indicates that this prototypical enteric model is conserved and functions similarly (with some variations on the theme) in phylogenetically diverse motile bacteria. In addition to regulating chemotaxis responses in motile bacteria, chemotaxis-like signal transduction pathways were shown to regulate cellular behaviors other than flagellar rotation in several other bacterial species (Kirby, 2009), including the alphaproteobacterium Azospirillum brasilense, a soil diazotroph (Bible et al., 2008). In only a few cases, however, have the molecular targets of these chemotaxis-like pathways been identified. The A. brasilense Che1 chemotaxis-like pathway has been shown to have a minor, and likely indirect, function in regulating chemotaxis behavior in this species (Hauwaerts et al., 2002; Bible et al., 2008; Edwards et al., 2011). Experimental evidence indicates that Che1 functions to modulate changes in adhesive cell surface properties which impact the propensity for cell-to-cell aggregation and flocculation (Bible et al., 2008). Deletions of cheA1 or cheY1, which each code for central proteins controlling the response output of the signal transduction pathway, yield cells that aggregate and flocculate more than the wild-type strain (Bible et al., 2008). A mutant strain deleted for all of the genes encoded within the che1 gene cluster has a phenotype similar to the strains lacking only CheA1 or CheY1, consistent with a role for Che1 in regulating the ability of cells to flocculate. A strain carrying a mutation that disrupts the function of both CheB1 and CheR1 is severely impaired in flocculation, consistent with CheB1 and CheR1 functioning in a signaling feedback loop that controls chemosensory adaptation (Stephens et al., 2006; Bible et al., 2008). Other possible roles that Che1 may have on functions such as adhesion to surfaces or root colonization, have been previously proposed to be related to flocculation (Burdman et al., 2000ab) but have not yet been investigated. The purpose of the present study was to determine the conditions under which A. brasilense cells attach to surfaces and to examine the effects of Che1-dependent signaling on this ability.

Materials and methods

Bacterial strains and culture conditions

Strains of A. brasilense used in this study are listed in Table 1. Strains AB103 and BS110 were previously shown to have identical phenotypes including growth, motility, chemotaxis as well as flocculation (Stephens et al., 2006; Bible et al., 2008). Except where noted, all Azospirillum strains were routinely maintained on solid tryptone yeast (TY) medium or on minimal medium for A. brasilense (MMAB; Hauwaerts et al., 2002). Flocculation was performed essentially as described in Sadasivan & Neyra (1985) and modified by Bible et al. (2008).

Table 1. Strains used in this study
StrainsRelevant characteristicsReference or source
A. brasilense Sp7Wild-type strainATCC 29145
BS104cheB1cheR1 double mutant of Sp7, Δ(cheB1-cheR1)::Kmr, KmrStephens et al., 2006
BS110che1 mutant of Sp7 by polar insertion in the first gene of the che1 operon, Kmr, TetrStephens et al., 2006
AB101cheA1 nonpolar mutant of Sp7, Δ(cheA1)::gusA-Kmr, KmrBible et al., 2008
AB102cheY1 nonpolar mutant of Sp7, Δ(cheY1)::Kmr, KmrBible et al., 2008
AB103che1 mutant of Sp7 by deletion-insertion of all che1 orfs and insertion of Cmr cassette Δ(cheA1-cheR1)::CmrBible et al., 2008

In vitro attachment assay to glass or polyvinylchloride surfaces

Preliminary experiments identified the following conditions to allow visualization of bacterial attachment. Azospirillum brasilense strains were cultured in TY medium to logarithmic phase and standardized to an OD600 nm of 1.0 using a phosphate buffer (per liter: 1.7 g K2HPO4, 1.36 g KH2PO4, 0.1 mM EDTA). Cells were re-inoculated into Corning 12-well (3.8 cm2) polystyrene containers (Corning, NY Fisher Catalog No. 3512) containing 3 mL liquid TY or MMAB medium, the latter supplemented with combined nitrogen (NH4Cl or NaNO3, as indicated) when applicable and containing 5 mM fructose and 5 mM sodium malate as carbon sources. Attachment to glass (hydrophilic) or polyvinylchloride (hydrophobic) coverslips (2 × 2 cm; Fisher Scientific, Pittsburgh, PA) was tested by placing surface-sterilized coverslips into the wells of a PVLC 96-wells plates prior to adding cells. Attachment of cells to polyvinylchloride or PVLC (hydrophobic surface) was equivalent and further experiments were conducted by measuring attachment to the PVLC wells (Corning). Cells were incubated for 1 and 7 days at 28 °C. To stain the biofilms, the culture was removed from the wells and a 0.01% crystal violet solution (w/v) was added and incubated 20 min. Next, the dye was removed and the excess washed by rinsing three times with sterile water. The remaining dye in the wells (representing attached cells as biofilms) was solubilized with 95% ethanol. Attachment was determined by the absorbance at 600 nm of the crystal violet solubilized (Fujishige et al., 2006).

Atomic force microscopy sample preparation and image acquisition

Samples were prepared on hydrophobic (polystyrene) and hydrophilic (glass) surfaces with polystyrene chips (2 × 2 cm) and glass coverslips (2.2 × 2.2 cm) as described previously (Edwards et al., 2011). Similar preparations were also used with lentil (LcH; Sigma-Aldrich, St. Louis, MO; specificity for α-mannose and/or α-glucose terminal residues) or wheat germ agglutinin (WGA; Sigma-Aldrich; specificity for N-acetylglucosamine terminal residues) lectins. On cleaned and UV-sterilized surfaces, 200 μL of 100 μg mL−1 LcH or WGA were added and allowed to absorb for 2 h at room temperature. After incubation, the excess lectin was removed and 5 mL of normalized cell suspension was added to the treated surfaces, followed by incubation at 28 °C for 24 h without agitation. Next, the surfaces were gently washed and air-dried before being imaged by atomic force microscopy (AFM) using a PicoPlus atomic force microscope (Agilent Technologies, Tempe, AZ). The instrument has a 100-µm multi-purpose large scanner and was operated in contact mode with speeds ranging from 0.5 to 1.0 Hz and 512 pixels per line scan. A Veeco MLCT-E cantilever with a resonant frequency ranging from 26 to 50 kHz and a nominal spring constant of 0.5 N m−1 was used for imaging. Scans were acquired with sizes ranging from 10 to 75 µm for all samples.

Analysis of lectin-dependent adhesion by confocal microscopy

Sterile 55-mm glass bottom petri dishes (MatTek Corp., Columbia, MD) were prepared with lectin prior to inoculation. LcH and WGA lectins, diluted to a final concentration of 100 µg mL−1 in PBS, were added to the glass bottom dishes and incubated for 2 h at room temperature. Next, the liquid was removed and 3 mL of overnight cell cultures in TY, diluted to OD600 nm of 1.0 (approximately 106 CFU mL−1) were immediately placed on the wet glass surface of the petri dish. Dishes were incubated statically at 28 °C for 24 h. SYTO 9 dye (1 µL) (Molecular Probes, Invitrogen Inc., Eugene, OR) was then added for 15 min in the dark to fluorescently label the cells. Images were acquired with laser intensity and gain held constant using a Leica TCS SP2 scanning confocal microscope equipped with a Leica HCX PL APO 63×/1.40–0.60 oil objective lens and Leica LCS software (version 1537, Leica Microsystems Inc., Buffalo Grove, IL). The number of attached cells was assessed using the imagej software to convert the images to a binary format. The pixel area corresponding to the fluorescent cells was identified and calculated as a percentage of the total image area (

Wheat-root attachment and colonization assays

Wheat seeds (Triticum aestivum cv. Jagger) were surface-sterilized and allowed to germinate as described (Greer-Phillips et al., 2004). For the wheat root attachment assay, A. brasilense strains were cultured in TY liquid overnight (28 °C, 200 rpm) and the cultures were normalized to an OD600 nm of 1.0 using sterile phosphate buffer. A volume of 200 μL of each strain prepared as described above was inoculated, in triplicate, into glass tubes containing 9.8 mL sterile phosphate buffer and 0.5 g of sterile roots isolated from 1-week-old plantlets and allowed to incubate for 2 h with shaking. The excised roots were then collected and washed three times with 5 mL of buffer with gentle shaking. Root material was then homogenized in 5 mL of fresh buffer and aliquots of the homogenized slurry were serially diluted and inoculated in triplicate on MMAB plates to determine colony forming units. The fraction of root-attached cells was expressed as percent of total cells inoculated.

Wheat colonization assays were performed as described previously (Greer-Phillips et al., 2004) with cultures inoculated at comparable levels (107 cells mL−1) into 15 mL molten semi-soft (0.4% agar) Fahraeus medium (Zamudio & Bastarrachea, 1994) modified with traces of sodium molybdate. Inoculated plants were incubated in a plant growth chamber with 12 h light cycle, at 28 °C. Five randomly selected plantlets per treatment were collected at 6, 24 or 48 h postinoculation. The root material was weighed before being homogenized in 1 mL of sterile phosphate buffer. Serial dilutions were then inoculated onto sterile MMAB plates placed to incubate at 28 °C. Colonization was estimated as CFU g−1 of fresh root material.

Statistical analysis

A one-tailed t-test assuming equal variances and < 0.05 (Microsoft Excel) was used to assess statistical significance of differences in attachment and colonization between the strains.

Results and discussion

Azospirillum brasilense binds preferentially to hydrophobic abiotic surfaces under diazotrophic conditions

Attachment of A. brasilense to glass or PVLC surfaces was first analyzed under different growth and incubation conditions. Attachment to glass was not significant irrespective of the growth conditions or the incubation time (data not shown). This was confirmed by AFM (Supporting Information, Fig. S1) and topographic analysis of the surfaces (Fig. S2), suggesting that the physical surface properties of glass do not facilitate attachment for A. brasilense. Growth conditions mediating surface attachment (biofilm) in A. brasilense were thus subsequently analyzed only on PVLC surfaces. Attachment was found to increase when the experiments were conducted under low aeration (i.e. nonshaking) conditions with cells transferred from culture in a rich medium (TY) to a minimal media (data not shown). No significant effect of varying the concentrations of either phosphorous or potassium, found to increase attachment in other bacterial species (Danhorn & Fuqua, 2007) could be detected (data not shown). When biofilm formation was monitored in media lacking nitrogen or containing relatively low concentrations (1 mM) of NH4Cl or NaNO3, surface adherence for all strains was greater compared with higher concentrations (10 mM) of NH4Cl or NaNO3, respectively (Table 2). Biofilm formation was the greatest for all strains with low concentrations of sodium nitrate. Differences seen initially between strains remained unchanged over time, although overall biofilm formation was increased at day 7 (Table 2). Nutritional conditions were previously shown to be powerful modulators of the attachment of various bacterial species to surfaces but specific effects of nitrogen availability on attachment have been seldom noted (O'Toole et al., 2000; Rinaudi et al., 2006; Danhorn & Fuqua, 2007). Compared with the parental strain Sp7 and regardless of the incubation conditions, the AB101 and AB102 strains showed a consistent greater attachment to PVLC surfaces. Different attachment abilities detected for these strains was apparent early (day 1) suggesting that the initial surface attachment step was affected (Table 2). Results obtained here strongly indicate that the contribution of Che1 signaling in modulating cell attachment to abiotic surfaces is likely to be indirect because while mutant strains lacking CheA1 and CheY1 attached better to surfaces, a mutant strain deleted for the che1 gene cluster does not. Similar conclusions were made regarding the contribution of Che1-dependent signaling to chemotaxis because mutations in CheA1, CheY1, CheB1 and CheR1 as well as mutations deleting Che1 led to distinct and uncorrelated chemotaxis phenotypes (Stephens et al., 2006; Bible et al., 2008). The results obtained here also indicate that strains lacking CheA1 and CheY1 have a stronger surface attachment response and biofilm forming ability under limiting nitrogen conditions, suggesting that they are more sensitive to the cue(s) that trigger such an attachment response.

Table 2. Effect of the nitrogen source on surface attachment and biofilm formation by Azospirillum brasilenseSp7 and its che1 mutant derivatives, 1 and 7 days postinoculation, using crystal violet staining
 No nitrogenNaNO3 (1 mM)NaNO3 (10 mM)NH4Cl (1 mM)NH4Cl (10 mM)
Strains1 day7 days1 day7 days1 day7 days1 day7 days1 day7 days
  1. The results are average of triplicate experiments.

Sp70.063 ± 0.0130.150 ± 0.0220.065 ± 0.0161.015 ± 0.0340.012 ± 0.0020.193 ± 0.0310.045 ± 0.0060.403 ± 0.0500.012 ± 0.0020.103 ± 0.006
AB1010.074 ± 0.0210.492 ± 0.1710.076 ± 0.0061.497 ± 0.1140.027 ± 0.0010.366 ± 0.0160.049 ± 0.0020.505 ± 0.0090.019 ± 0.0010.183 ± 0.021
BS1040.062 ± 0.0010.158 ± 0.0040.069 ± 0.0150.764 ± 0.0610.012 ± 0.0150.181 ± 0.0210.033 ± 0.0030.329 ± 0.0350.008 ± 0.0010.105 ± 0.018
AB1020.104 ± 0.0100.504 ± 0.0440.082 ± 0.0051.343 ± 0.0460.021 ± 0.0040.361 ± 0.0240.039 ± 0.0040.530 ± 0.0270.017 ± 0.0040.179 ± 0.026
AB1030.042 ± 0.0110.109 ± 0.0080.047 ± 0.0020.503 ± 0.1720.009 ± 0.0010.137 ± 0.0480.022 ± 0.0040.300 ± 0.0540.008 ± 0.0010.090 ± 0.010

Attachment and early colonization of wheat roots have different requirement for Che1 dependent signaling

Similar patterns of attachment between che1 mutant strains were observed on excised sterile wheat roots, with both the AB101 (fraction of root-attached cells, as percent of total cells inoculated were 40.9 ± 1.7%) and AB102 (34.9 ± 4.1%) strains attaching significantly (P < 0.05) more than any other strains tested (Sp7: 15.1 ± 0.8%; AB103: 15.0 ± 1.2%), and strain BS104 (11.0 ± 0.9%) attaching significantly less than the wild-type strain. Attachment to wheat root surfaces may thus not be directly dependent on Che1 signaling activity. The increased ability of strains AB101 and AB102 to attach to excised roots did not correlate with an increased ability to colonize sterile roots (Fig. 1). The mutant strain lacking functional CheB1 and CheR1 (strain BS104) was significantly delayed in root colonization: the earliest population levels detected on the roots (6 h) were at least twofold lower relative to wild-type population levels and remained low after 48 h. A similar significant colonization delay was detected for the mutant strain lacking functional Che1 (Fig. 1). Both mutant strains BS110 and BS104 have comparable colonization phenotypes, suggesting that the colonization defect detected for both strains is related to the lack of functional CheB1 and CheR1. Both strains were previously shown not to have any growth, motility, chemotaxis or aerotaxis defects (Stephens et al., 2006; Bible et al., 2008). Therefore, it is unlikely that any of these functions have contributed to the delayed colonization under these conditions. Attachment to wheat root was performed in a buffer lacking a source of combined nitrogen which could explain the pattern of attachment observed. Nitrogen may not be a limiting nutrient for growth in the wheat rhizosphere under the short-term root colonization conditions used (Fig. 1), thereby eliciting different responses from the A. brasilense cells in the two assays. These results also do not argue against the role for chemotaxis in root colonization, as Che1 does not directly control chemotaxis (Vande Broek et al., 1998; Greer-Phillips et al., 2004; Bible et al., 2008). While Che1 signal transduction functions to modulate the ability of cells to aggregate and flocculate, data obtained here argue against a straightforward correlation between aggregation and flocculation and root colonization abilities that have been previously proposed in A. brasilense (Burdman et al., 2000ab). Instead, it suggests that attachment to wheat root surfaces and Che1-dependent changes in cell surface properties are distinct, although they may partially overlap under nitrogen limiting conditions.

Figure 1.

Colonization of sterile wheat roots by Azospirillum brasilense wild-type strain Sp7 and its Che1 mutant derivatives. The cells were inoculated to sterile germinated wheat seeds at equivalent cell densities. In both panels, the star symbol represents statistically significant differences relative to the wild-type strain (P < 0.05).

Lectins promote A. brasilense attachment to surfaces

The increased attachment of AB101 and AB102 may be partly dependent on changes in cell surface-exposed polysaccharides that are modulated in Che1-dependent manner (Bible et al., 2008; Edwards et al., 2011). To directly evaluate the contribution of specific sugar-binding molecules on promoting attachment and biofilm formation, glass surfaces were treated with LcH or WGA lectins, prior to incubation with A. brasilense cells. AFM imaging indicated that the lectin treatment increased attachment for all strains, with the most significant increase in attachment seen for the AB101, AB102, and AB103 strains on LcH-treated glass surfaces (Fig. 2). The increased attachment was comparable for all strains on WGA-treated glass surfaces (Fig. 2). Although the ability of cells to attach to lectin-treated glass surfaces varied greatly between the strains, no distinctive visible extracellular structure(s), such as flagella, pili or specific patterns in the EPS (exopolysaccharide) matrices, could be attributed to this difference (Fig. S3). This does not account for expression variation in outer membrane proteins (OMPs), polysaccharides, or other adhesions beyond the resolution capabilities of the AFM scans (Fig. S3). Next, confocal microscopy was used to analyze attachment of cells to lectin-treated glass (Fig. 3). Prior to imaging, the lectin-treated surfaces on which cells attached were gently and briefly washed to ensure that only primary attachment to the surface was accounted for and to reduce possible confounding interpretations resulting from secondary attachment events (e.g. to other cells). Under these conditions, the attachment pattern of the Che1 mutant strains on lectin-treated surfaces were similar to that observed by AFM with attachment to LcH-treated glass surfaces, but not WGA treated-glass surface, directly correlating with the flocculation phenotypes of the strains: strains that flocculate more than wild type (AB101, AB102, and AB103) also attached to LcH-treated glass surfaces more (Table 3). Given that cells did not attach to glass in the absence of lectins, the surface attachment detected here is likely via interaction between cell surface exposed sugar residues and the lectins. The two lectins tested mediated different patterns of attachment for the che1 strains tested, suggesting distinct surface-exposed sugar residues between the strains, an observation consistent with similar conclusions reached previously (Edwards et al., 2011). The observation that LcH-dependent attachment correlated with the increased flocculation behavior of some of the che1 mutant strains provides further support to the notion that Che1-dependent changes in cell-to-cell aggregation and flocculation involves remodeling of the extracellular matrix, some of which is shown here to promote surface attachment. Regardless of the exact effects that Che1 signaling has on cell surface changes which are currently investigated in our laboratory, the data obtained here show that attachment of A. brasilense is increased by nitrogen limitation and further suggests that it depends on sugar-exposed residues that have lectin-binding properties, in agreement with the proposition made previously by Mora et al. (2008). Increasing attachment of A. brasilense to root surfaces may thus ultimately depends on fine-tuning metabolic activities, including limiting nitrogen availability that is shown here as a key modulator of attachment to surfaces.

Figure 2.

Attachment of Azospirillum brasilense to LcH and wheat germ agglutinin (WGA) lectins-treated glass surfaces by atomic force microscopy. Scale bars represent 10 µm. LcH-treated glass is shown at the top row and WGA-treated glass at the bottom row.

Figure 3.

Confocal images of Azospirillum brasilense wild type and mutant strains attached to untreated and lectin-treated glass surfaces. (a) Attachment on untreated glass was not apparent for any of the strains. Cells were labeled with Syto9 that stains both live and dead cells and visualized by confocal microscopy (magnification 63×). All cells, regardless of their viability were detected by this method. (b) Attachment to glass treated with LcH lectin after 24 h. Pre-wash refers to surface attachment detected before rinsing the surface with water and postwash were taken immediately after. (c) Attachment to glass treated with wheat germ agglutinin lectin after 24 h. Scale bars represent 20 µm.

Table 3. Surface coverage of Azospirillum brasilenseSp7 and che1 mutant derivatives on lectin-treated and nontreated glass surfaces, analyzed by confocal microscopy
  1. Surface coverage was determined by image analysis (imagej software) and is expressed as percent of total surface area.

Nontreated glass00 00 0
LcH-treated glass0.54 ± 0.053.75 ± 1.9218.86 ± 7.060.56 ± 0.4015.02 ± 2.06
WGA-treated glass2.15 ± 1.012.98 ± 2.035.01 ± 1.971.37 ± 0.161.51 ± 0.06


The authors thank members of the Alexandre's and Doktycz's laboratory for careful comments on the manuscript. This work was supported by a NSF CAREER award (MCB-0622277) and MCB-0919819 to G.A. and by the Genomic Science Program of the Office of Biological and Environmental Research, US DOE. Oak Ridge National Laboratory is managed by UT-Battelle, LLC, for the US Department of Energy under Contract no. DE-AC05-00OR22725.